STRUCTURAL ELUCIDATION OF LIGNOCELLULOSE AND WETLAND SOIL BY SOLID-STATE NMR AND DYNAMIC NUCLEAR POLARIZATION By Wancheng Zhao A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Chemistry – Doctor of Philosophy 2023 ABSTRACT Solid-state Nuclear Magnetic Resonance (ssNMR) is a powerful analytical technique that has emerged as a versatile tool for high-resolution detection and characterization of biosolids, such as carbohydrates, proteins, nucleic acids, and organic matter, in their native state. By harnessing the principles of NMR spectroscopy, ssNMR provides detailed insights into the atomic-level structure, dynamics, and interactions of these biomolecules, making it an invaluable resource for advancing our understanding of complex biological systems. This work is dedicated to addressing fundamental yet profoundly significant questions in the fields of sustainable bioenergy development and coastal wetland ecosystem preservation by ssNMR. Firstly, we have unraveled the intricate packing of lignin and carbohydrates within plant cell walls, a critical inquiry that underpins our ability to engineer plants for improved biofuel and materials production. This research holds exceptional significance within the field of bio- renewable energy. Leveraging the power of ssNMR, we have achieved atomic-resolution characterization of the polymorphic structure, dynamical behavior, hydration profiles, and physical arrangement of lignin and polysaccharides. The findings uncover that the extent of glycan- aromatic association increases sequentially across grasses, hardwoods, and softwoods; and lignin principally packs with the xylan in a non-flat conformation via electrostatic interactions and partially binds the junction of flat-ribbon xylan and cellulose surface as a secondary site. Another exciting development in our research is the integration of the sensitivity-enhancing Dynamic Nuclear Polarization (DNP) technique. This innovative approach holds immense promise for the investigation of natural abundance samples. We have successfully demonstrated the feasibility of combining traditional ssNMR with DNP methods to probe the structure and dynamics of carbohydrates in natural abundance rice stems. This streamlined approach enables the efficient screening of a diverse array of biomass materials without the need for labeling. Furthermore, the application of ssNMR and DNP has facilitated the rapid acquisition of 2D 13C/1H-13C correlation spectra to detail the molecular fingerprint and spatial organization of biopolymers in unlabeled wetland soil obtained from a brackish island situated along the Gulf of Mexico coastline. Surprisingly, the lignocellulosic core identified in the plants grown on this island was found to be preserved in the surface layer of soil formed in the brackish marshes, but with tighter physical packing between molecules. Extending the depth from the surface to 2 m covers a geological timeline of eleven centuries, where we found sophisticated changes in the molecular structure and composition of organic matter as well as the bulk properties of the wetland soil. These findings offer a promising avenue toward addressing pressing environmental challenges, including coastal wetland loss and the impact of sea-level rise. Additionally, statistical approaches for the analysis of carbohydrate composition and structure using high-resolution ssNMR data have been introduced. We generate simulated spectra (density maps) using data deposited in our developed and maintained Complex Carbohydrates Magnetic Resonance Database (CCMRD, www.ccmrd.org), to demonstrate the chemical shift dispersion and to aid in the fast identification of important fungal and plant polysaccharides in cell walls. By comparing the projected spectra with experimental data, we highlight the challenges of assigning resonances because distinct carbohydrates from different organisms can produce nearly identical signals. Copyright by WANCHENG ZHAO 2023 To my beloved wife, Thanks for your love, encouragement, and unwavering support. To my dear parents, Thanks for your values, sacrifices, and unshakable faith. v ACKNOWLEDGEMENTS In the tapestry of my academic journey, I must first express my profound gratitude to my advisor, Prof. Tuo Wang, who has not only served as my guiding light but also as the very embodiment of the scholar I aspire to become. If asked to name the single individual I held in the highest regard throughout my entire existence, your name would resound without hesitation. Your vast knowledge has been a wellspring of inspiration, continuously guiding us to the forefront of research. Your unwavering patience and serene encouragement have provided solace during our missteps and diversions from our goals. Your generosity in sharing wisdom, resources, and valuable materials has seamlessly integrated us into the academic community. Your passionate support for our participation in conferences has expanded our intellectual horizons, while your guidance transcends research mechanics, instilling in us the values of a proficient researcher and exemplary research habits. Your exceptional support for our scholarships and heartfelt recommendation letters for our careers have been truly extraordinary. Words alone are inadequate to capture the depth of my gratitude for your contributions. Indeed, it is a rare privilege and a lifetime's honor to have been counted among your students. Only now do I fully appreciate the depth of the saying, “If I have seen a little further, it is by standing on the shoulders of Giants.” To me, you are the very giant whose shoulders I stand upon, affording me a profound view of an expansive world. And it is my fervent hope to be the pillar on which the next generation of scientists can rely, as you have been for me, nurturing their boundless potential. I would like to express my sincere appreciation to the esteemed members of my committees: Prof. Gary Blanchard, Prof. Charles Hoogstraten, Prof. Josh Vermaas, and Prof. David Weliky. Your invaluable contributions, including critical insights, thoughtful feedback, and constructive suggestions, have been instrumental in enhancing the quality of this work. I am vi particularly thankful to Prof. Blanchard for his significant assistance during my transfer process and unwavering support for my Dissertation Completion Fellowship applications. Prof. Hoogstraten, I am grateful for the enlightening discussions we've had at conferences, which have provided me with valuable insights and enriched my academic journey. Prof. Vermaas, your guidance and advice on my research path and future career planning have been invaluable, and I truly appreciate the positive impact you've had on my academic and professional development. Prof. Weliky, your dedication to academic excellence, along with your valuable advice and consistent guidance, has been a cornerstone of my academic journey. I also want to extend my gratitude to my committee members during my time at LSU: Prof. Bin Chen, Prof. Robert Cook, and Prof. Paul South. Your collective wisdom and guidance have broadened my perspective and significantly contributed to the depth and rigor of this research. Prof. Cook, I am especially thankful for your unwavering support in providing recommendation letters, which have played a pivotal role in shaping my research trajectory. Your mentorship and collaborative efforts in the academic realm have greatly enriched my overall experience. My heartfelt thanks go out to my lab colleagues, Dr. Qinghui Cheng, Dr. Jayasubba Yarava, Dr. Krishna Bankala, Isha Gautam, Debkumar Debnath, Anand Jacob, Ankur Ankur, Kalpana Singh, and Priya Sahu. I want to extend my heartfelt appreciation to the former members of the Wang group Dr. Xue Kang, Dr. Alex Kirui, Dr. Malitha Chathuranga, Dr. Liyanage Fernando, Dr. Fabien Deligey, Dr. Nader Ghassemi, Dr. Chandra Shekar, Dr. Alexandre Poulhazan, Dr. Liliana Martínez, Arnab Chakraborty, Irina Gomez, Paninga Muiliya, and Jianyu Guan. Your camaraderie, stimulating discussions, and shared experiences have not only created an intellectually stimulating environment but have also provided a strong support system. Special thanks to Fabien, for your guidance. Most of my NMR knowledge has been acquired from you, especially during the initial vii stages after the pandemic. I've gained valuable insights from you, and my English proficiency has also significantly improved with your help. Special thanks also to Dev and Isha. We came from LSU to MSU together and went through some of the toughest times as a team. Your support during those moments is something I will cherish throughout my life. I extend my gratitude to my scientific collaborators who have been essential to our projects and have engaged in invaluable support and discussions. I deeply appreciate the collaboration with Prof. Daniel Cosgrove from Penn State. Thanks for your wholehearted support during my job search, your patient guidance, and the privilege of learning and exchanging ideas within the CLSF center under your leadership. I am profoundly grateful for the collaboration with Dr. Frederic Mentink-Vigier from the National High Magnetic Field Laboratory (Maglab). Our in-depth discussions, your wholehearted support, and your steadfast guidance in the realm of DNP are sincerely appreciated. Additionally, I would like to recognize Dr. Faith Scott, Dr. Zhenhong Gan, Dr. Ivan Hung, and Dr. Riqiang Fu, all from Maglab, for their valuable guidance and assistance in experimental setup. I would like to express gratitude to Prof. John R. White from LSU for our fruitful collaboration and your valuable insights throughout. I would also like to extend my gratitude to Maglab and the Pacific Northwest National Laboratory for their indispensable instrumental support. I would also like to express my sincere gratitude to the dedicated NMR managers, Dr. Daniel Holmes (MSU), Dr. Li Xie (MSU), Dr. Thomas Weldeghiorghis (LSU), and Dr. Fengli Zhang (LSU). Your unwavering commitment and expertise have been invaluable in ensuring the smooth operation of our NMR facilities. Special thanks to Dr. Holmes and Dr. Xie, your tireless efforts and extensive knowledge in maintaining and optimizing the NMR instruments have greatly contributed to the success of our research endeavors. viii I would like to extend my heartfelt appreciation to my friends, who have been an essential source of support and encouragement during my academic pursuit. To Dr. Xiaobing Chen, as my “buddy” at LSU, your unwavering support facilitated my initial adjustment to the new foreign environment. To Dr. Weiwei Xie, your support and encouragement during my toughest moments were profoundly significant, helping me find strength and hope. I would like to express my deepest appreciation to my family for your endless love, tolerance, and support, which have been instrumental in shaping the person I have become today. To my wife, Tongjia Fu, your understanding and support have been the bedrock of my daily motivation over the past five years of my academic journey in a foreign place. In moments of uncertainty, you've consistently stood by my side, offering me the strength to persevere. Reflecting on our 18 years together, I can't help but feel incredibly fortunate; having you in my life has been an invaluable blessing. To my parents, Xiufen Lang and Xiliang Zhao, I am profoundly thankful for your care, guidance, understanding, and patience. You've consistently been the steadfast foundation of my academic journey. I am always certain that no matter what, your unwavering support and understanding are there for me. To my parents-in-law, Hongmin Fu and Nairu Fu, your consistent encouragement, trust, and belief in me sustained my motivation during this challenging journey. I'm also grateful that you have entrusted me with such an exceptional daughter. Heartfelt appreciation to each and every one of you! ix TABLE OF CONTENTS CHAPTER 1: INTRODUCTION ....................................................................................................1 REFERENCES ....................................................................................................................36 CHAPTER 2: SOLID-STATE NMR OF PLANT AND FUNGAL CELL WALLS: A CRITICAL REVIEW ........................................................................................................................................42 REFERENCES ....................................................................................................................61 CHAPTER 3: CURRENT LIMITATIONS OF SOLID-STATE NMR IN CARBOHYDRATE AND CELL WALL RESEARCH .................................................................................................68 REFERENCES ....................................................................................................................88 CHAPTER 4: POLYSACCHARIDE ASSEMBLIES IN FUNGAL AND PLANT CELL WALLS EXPLORED BY SOLID-STATE NMR .......................................................................................94 REFERENCES ..................................................................................................................115 CHAPTER 5: A THREE-DIMENSIONAL 13C-13C-13C CORRELATION EXPERIMENT FOR HIGH-RESOLUTION ANALYSIS OF COMPLEX CARBOHYDRATES USING SOLID- STATE NMR ...............................................................................................................................125 REFERENCES ..................................................................................................................141 APPENDIX .......................................................................................................................146 CHAPTER 6: EFFECT OF CROSS POLARIZATION RADIOFREQUENCY PHASES ON SIGNAL PHASE .........................................................................................................................153 REFERENCES ..................................................................................................................168 CHAPTER 7: CARBOHYDRATE-AROMATIC INTERFACE AND MOLECULAR ARCHITECTURE OF LIGNOCELLULOSE .............................................................................173 REFERENCES ..................................................................................................................200 APPENDIX .......................................................................................................................210 CHAPTER 8: SOLID-STATE NMR OF UNLABELED PLANT CELL WALLS: HIGH- RESOLUTION STRUCTURAL ANALYSIS WITHOUT ISOTOPIC ENRICHMENT...........275 REFERENCES ..................................................................................................................296 APPENDIX .......................................................................................................................302 CHAPTER 9: CHRONOLOGICAL MOLECULAR FINGERPRINT OF WETLAND SOIL BY SENSITIVITY-ENHANCED SOLID-STATE NMR .................................................................316 REFERENCES ..................................................................................................................340 APPENDIX .......................................................................................................................347 CHAPTER 10: CCMRD: A SOLID-STATE NMR DATABASE FOR COMPLEX CARBOHYDRATES ..................................................................................................................371 REFERENCES ..................................................................................................................382 APPENDIX .......................................................................................................................387 x CHAPTER 11: CHARTING THE SOLID-STATE NMR SIGNALS OF POLYSACCHARIDES: A DATABASE-DRIVEN ROADMAP .......................................................................................390 REFERENCES ..................................................................................................................411 APPENDIX .......................................................................................................................418 xi CHAPTER 1: INTRODUCTION 1.1 Significance of the Research Endeavor Lignocellulosic biomass, derived from sources like forest waste, agricultural crop residues, and algae, represents a vast and sustainable resource with the potential to revolutionize the production of biofuels, biomaterials, and high-value bioproducts1, 2. The structural foundation of this resource lies in plant cell walls, which are essential for maintaining the mechanical strength and flexibility necessary to preserve cellular integrity and form3. The primary cell walls4, 5 of the growing plant consist of three main classes of carbohydrate polymers: cellulose, hemicellulose, and pectin. Cellulose microfibrils, with diameters ranging from 3 to 5 nanometers, are composed of rigid and semi-crystalline β-(1-4)-glucan chains. Hemicellulose, on the other hand, contains shorter and often branched chains. For example, xyloglucan (XyG), a major hemicellulose in the primary walls of dicots like Arabidopsis thaliana, features a β-1,4-glucan backbone adorned with α-xylose (Xyl) sidechains. Pectin, which typically includes homogalacturonan (HG), rhamnogalacturonan-I (RG-I), and RG-II, plays a crucial role in embedding itself within the cellulose-hemicellulose network and performing various functions such as regulating wall porosity, modulating cell expansion, and controlling pH and ionic balance. The secondary cell wall6, 7, formed inside the primary cell wall when the plant ends growing, comprises cellulose, hemicellulose (specifically xylan and glucomannan), and lignin. In the secondary wall, larger cellulose bundles (ranging from 10 to 20 nanometers) are encased by a xylan-lignin crosslink, with lignin being a diverse and hydrophobic biopolymer rich in aromatic subunits. Polysaccharides and lignin, as the fundamental biopolymers in plant cell walls, are held together by numerous covalent and physical interactions8. Nonetheless, these biopolymers are typically insoluble in water, partially indigestible, and often non-crystalline. This presents a 1 significant challenge when attempting to achieve high-resolution structural characterization of intact cell walls. Traditional methods primarily focus on the structural analysis of specific components of the cell wall, involving the hydrolysis of biomolecules into small soluble compounds using organic solvents or ionic liquids. The hydrolysis treatments can significantly disrupt and even modify the biopolymers, causing conventional strategies to fall short in their ability to elucidate the packing and motion within native cell walls. Hence, there is a pressing need for alternative methods that allow for undisturbed detection and analysis of plant cell walls, to benefit more sustainable and efficient biomass utilization. In addition to the more efficient utilization of renewable energy, another critical concern related to global climate issues revolves around mitigating coastal land erosion9-11. The escalating loss of both wetland and coastal land, compounded by the growing threat of rising sea levels, represents an urgent and globally significant environmental crisis. Louisiana holds 40% of the coastal and estuarine wetlands in the contiguous United States but has suffered the highest rate of wetland loss, with around 4,900 square kilometers of coastal land vanishing since 193012. Wetlands, encompassing diverse ecosystems such as salt marshes, mangroves, and seagrass beds, play an indispensable role as natural barriers along coastlines. They provide vital services, including shielding against coastal erosion, flood mitigation, and fostering biodiversity. Remarkably, coastal wetlands, known as blue carbon ecosystems, occupy a mere 0.07 to 0.22% of the Earth's surface, yet they sequester more than 50% of the carbon annually buried in the ocean9, 13. These wetlands serve as major reservoirs of soil organic matter (SOM) and its associated soil organic carbon (SOC), critical in regulating Earth's climate, water systems, and biodiversity. Unfortunately, subsidence and rising sea levels induce the collapse of wetland soils into adjacent aerobic waters, triggering chemical decomposition of SOM components and the release of stored 2 carbon as carbon dioxide and methane into the atmosphere. Beyond their role as carbon sinks, the wetlands of Louisiana hold particular significance, as they offer insights into erosion patterns worldwide, driven by accelerated sea-level rise and subsidence. To protect these invaluable ecosystems and mitigate greenhouse gas emissions, it is imperative to comprehensively grasp, model, and predict SOM carbon dynamics at scale. Detailed molecular insights, including functionality, isomerism, and conformations of this carbon pool, are fundamental prerequisites for this understanding. Presently, there is limited research on this advanced level of carbon sequestration within coastal wetlands due to the complexities and insolubility of soil. Multidimensional 13C–13C/1H correlation ssNMR spectroscopy, with uniformly isotope- labeled (13C) samples, is an emerging technique that unravels the mysteries of solid materials14, 15. It breaks free from the constraints of crystallinity, offering high-resolution structural insights into disordered and insoluble samples, capturing dynamic processes, and preserving sample integrity. Its versatility, ability to decipher a multitude of nuclei, and the capability to conduct in situ studies make it invaluable in various fields, from materials science and chemistry to pharmaceuticals and structural biology. In addition, the introduction of cutting-edge Magic-Angle Spinning Dynamic Nuclear Polarization (MAS-DNP) enables the transfer of electron polarization to NMR-active nuclei, leading to one to two orders of NMR sensitivity enhancements16, 17. This breakthrough allows for the acquisition of high-resolution 2D spectra even from unlabeled samples containing low-abundance isotopes such as carbon-13 (1.1%) and nitrogen-15 (0.4%). My thesis research is dedicated to the integration of cutting-edge multidimensional high- resolution ssNMR techniques, complemented by sensitivity-enhanced DNP, to delve deeply into the intricate structural and dynamic attributes of lignocellulosic biomass and Louisiana wetland samples. The outcomes of this research will have a far-reaching impact, fostering the evolution of 3 next-generation pretreatment techniques and advancing lignin bioengineering for a more sustainable and efficient utilization of biomass. Meanwhile, these advancements will play a crucial role in safeguarding Louisiana's distinctive and fragile coastline, making a substantial and positive contribution towards its preservation. 1.2 SsNMR Methodology 1.2.1 The Basis of NMR Spectroscopy NMR spectroscopy is a sophisticated analytical technique that has had a profound impact on scientific research across a multitude of disciplines18-22. Its origins in the study of nuclear magnetic properties have given rise to its pivotal role as a unifying tool that bridges the gaps between chemistry, physics, biology, and materials science. NMR serves as an indispensable key for unlocking the secrets of the atomic and molecular world, providing profound insights into the structures, dynamics, and interactions at the most fundamental level. This method capitalizes on the intriguing responses of atomic nuclei possessing. When non-zero spins (I≠0, Table 1.1) are subjected to an external magnetic field, they align themselves with this field, much like a compass needle aligns with Earth's magnetic field. The application of a precisely tuned radiofrequency (RF) pulse disrupts this alignment, causing the nuclei to deviate from their equilibrium positions and initiate precession around the magnetic field's direction. This results in an energy difference, denoted as ∆𝐸 and defined by the equation: ∆𝐸 = −ℏ𝛾𝐵0 Eq 1.1 In this equation, 𝛾 represents the gyromagnetic ratio, a fundamental property unique to each isotope. 𝐵0 signifies the strength of the static magnetic field, while ℏ (h-bar) represents the reduced Planck constant. In modern NMR spectroscopy, powerful magnetic fields are employed (5 - 28 T), corresponding to the well-known 1H Larmor frequencies (𝜔0, Figure 1.1)23. 4 Figure 1.1 Basis of NMR Spectroscopy. a, Nuclear spin undergoes precession around a static magnetic field (𝐵0) at a frequency (𝜔0). b, The transition frequency between the energy levels of the spins. c, The resulted single peak spectrum at the Larmor frequency of the spin. Table1.1 Frequently examined atomic nuclei in ssNMR. Nuclei Spin 𝛾 (107∙rad/T∙s) 𝜔0 at 18.8 T Abundance 1H 2H 13C 14N 15N 17O 19F 31P a e- 1/2 1 1/2 1 1/2 5/2 1/2 1/2 1/2 26.75 800 MHz 99.98 % 4.11 6.73 1.93 -2.71 -3.63 25.18 10.84 123 MHz 200 MHz 0.02 % 1.11 % 58 MHz 99.63 % 80 MHz 108 MHz 753 MHz 324 MHz 0.37 % 0.04 % 100 % 100 % -17600 -526 GHz - a The electron constant is provided for the understanding of DNP mechanism. 5 In NMR spectroscopy, several fundamental principles and concepts converge, enabling the profound insights it offers into the world of molecules and materials. One of the key elements is the exquisite sensitivity of NMR frequencies to the electron distribution surrounding the nucleus. This electron shielding affects the nucleus's magnetic environment, resulting in slightly different frequencies due to variations in the shielding constant (σ) among nuclei of the same isotope within a molecule. The frequencies, typically reported as chemical shifts (δ) with a unit part per million (ppm), relative to well-established reference compounds, such as tetramethylsilane (TMS) that is frequently utilized for organic molecules, and sodium trimethylsilyl propane sulfonate (DSS), which is a common reference in protein NMR. The variation in chemical shift, for a specific isotope23, 24, can range from 0 to 10 ppm for 1H, from 0 to 200 ppm for 13C, and up to 900 ppm for 15N. Furthermore, NMR frequencies are subject to the influence of a range of couplings, including spin-spin scalar (J-coupling), spin-spin dipolar, and quadrupolar couplings, which are contingent on factors such as covalent bonding, internuclear distances, and nuclear spin values. These anisotropic interactions22 (Figure 1.2), dependent on the sample's orientation relative to the magnetic field, allow NMR spectra to encode rich three-dimensional structural information, offering valuable insights into the spatial arrangement of atoms within a molecule. Figure 1.2 NMR spin interactions. Dipolar coupling is absent in solution NMR. 6 1.2.2 Solid-State NMR Spectroscopy The stark contrast between ssNMR and solution NMR stems primarily from the absence of rapid molecular tumbling in the former. In solution NMR, the swift tumbling of molecules ensures that the orientation of the shielding/dipolar tensor with respect to the magnetic field (𝐵0) is rapidly averaged over all conceivable values. Conversely, ssNMR, in a static state, is susceptible to a wide array of orientation-dependent interactions, including chemical shift anisotropy (CSA) and heteronuclear dipolar coupling. To address this challenge, the adoption of spinning the sample at the magic angle, set at 54.74° relative to the external magnetic field, is a fundamental strategy. This mechanically induced uniaxial rotation, specifically the magic-angle spinning (MAS)25, 26, is an essential technique extensively utilized in the vast majority of ssNMR experiments (Figure 1.3). Its primary function is to nullify the influence of CSA and assist in mitigating heteronuclear dipolar-coupling effects. Furthermore, it serves to narrow spectral lines originating from quadrupolar nuclei and is increasingly preferred for counteracting homonuclear dipolar coupling in NMR spectra. The chemical shift Hamiltonian can be represented as: 𝐻̂𝑐𝑠 = −𝜎𝑧𝑧. 𝛾ℏ𝐵𝑜. 𝐼̂𝑧 Eq 1.2 𝜎𝑧𝑧, along with other two chemical shielding tensors 𝜎𝑥𝑥 and 𝜎𝑦𝑦, commonly employed to describe the chemical shift interaction within the principal axis system, which is expressed as: 𝜎𝑧𝑧=𝜎𝑖𝑠𝑜 + 1 3 ∑ 𝑗=𝑥,𝑦,𝑧 (3𝑐𝑜𝑠2𝜃 − 1) 𝜎𝑗𝑗 Eq 1.3 Then, to show both isotropic and anisotropic parts, the Hamiltonian can be rephrased as: 𝐻̂𝑐𝑠 = {𝜎𝑖𝑠𝑜. 𝛾ℏ𝐵0 + 1 2 𝛿[3𝑐𝑜𝑠2𝜃 − 1 − 𝜂𝑠𝑖𝑛𝜃2cos (2𝜙)]} 𝐼̂𝑧 Eq 1.4 When a high-speed uniaxial rotation (𝜂 = 0) is employed, it introduces an extra averaging factor, leading to a modification in the frequency of CSA: 7 𝜔𝑎𝑛𝑖𝑠𝑜 = 1 2 𝛿[3𝑐𝑜𝑠2𝛽 − 1 − 𝜂𝑠𝑖𝑛𝛽2cos (2𝛼)] 1 2 (3𝑐𝑜𝑠2𝜃𝑟 − 1) Eq 1.5 Within the rotor frame (Figure 1.3), 𝛽 characterizes the angle between the principal z-axis of the interaction and the rotor's rotation axis, while 𝜃 denotes the angle between 𝐵0 (the external magnetic field) and the primary z-axis of the interaction. When the rotor rotates around an axis tilted at an angle 𝜃𝑟 = 54.74°, then 3𝑐𝑜𝑠2𝜃𝑟 − 1 = 0, resulting in zero CSA. Therefore, the detected frequency matches the isotropic chemical shift in the case of rapid MAS, resulting in well-defined spectral lines instead of the broad powder patterns. In situations where the rotor speed is slow, the chemical shift anisotropy can only be effectively averaged to zero after each rotor period, resulting in spinning sidebands at 𝜔 = 𝜔𝑖𝑠𝑜 ± 𝑛𝜔𝑟 frequencies. Figure 1.3 Magic-angle spinning. a, Illustrated cartoon presenting relationships among magic angle, lab frame, rotor frame, and principal axis system. b, SsNMR spectra (14.4 T) of 13C-glycine at various spinning speeds. The Cross Polarization (CP) technique26 is also a key method used in ssNMR spectroscopy, which serves as both a standalone experiment and a fundamental component to enhance the sensitivity in numerous ssNMR experiments (Figure 1.4a). CP plays a pivotal role in transferring 8 polarization from NMR-sensitive protons to less sensitive X-nuclei like 13C and 15N (Table 1.1), often as the initial step and a common element in an array of 1D, 2D, and 3D ssNMR pulse sequences. CP finds application in both static and MAS conditions and serves diverse purposes. These applications range from reducing experimental time due to the shorter 1H T1 relaxation, to selectively detecting rigid components within the sample. CP is also employed to extract distance information by estimating dipolar interactions based on the build-up of CP intensities. Remarkably, CPMAS itself becomes part of the strategy to transfer polarization from an even richer source (electrons) in indirect DNP experiments. Figure 1.4 Cross polarization. a, SsNMR spectra (14.4 T, 12 kHz) of 13C-glycine with various CP contact time, as well as CP pulse sequence. b, SsNMR spectra (14.4 T, 12 kHz) of 13C-glycine with various heteronuclear dipolar decoupling. Dipolar coupling, also known as dipolar interaction or dipole-dipole coupling, plays a significant role in understanding the structural and dynamic properties of molecules, particularly in ssNMR experiments. In ssNMR, the dipolar coupling can lead to homonuclear (between nuclei of the same type) and heteronuclear (between nuclei of different types) dipolar recoupling, 9 resulting in dipolar-based experiments such as PDSD/DARR, CORD, RFDR, HETCOR, and DIPSHIFT, etc. The dipolar coupling20, 23 can provide crucial information about the spatial arrangement of atoms in a solid-state sample, as it encodes information about the internuclear distances and orientations. Based on Eqs. 1.6 and 1.7, the strength of dipolar coupling is inversely proportional to the cube of the distance (𝑟) between nuclei. 𝐻̂ ℎ𝑒𝑡𝑒𝑟𝑜 = − ( 𝑑𝑑 𝜇0 4𝜋 ) 𝛾𝐼𝛾𝑆ℏ 𝑟3 (3𝑐𝑜𝑠2𝜃 − 1)𝐼̂𝑧𝑆̂𝑧 Eq 1.6 𝐻̂ ℎ𝑜𝑚𝑜 = − ( 𝑑𝑑 𝜇0 4𝜋 ) 𝛾2ℏ 𝑟3 . 1 2 (3𝑐𝑜𝑠2𝜃 − 1)(2𝐼̂1𝑧𝐼̂2𝑧 − 𝐼̂1𝑥𝐼̂2𝑥 − 𝐼̂1𝑦𝐼̂2𝑦) Eq 1.7 Meanwhile, it comes with challenges related to line broadening, orientation dependence, and data interpretation, which may require specialized techniques and expertise to overcome, such as MAS and decoupling. Dipolar interactions can be effectively nullified through MAS when the spinning frequency greatly exceeds the coupling strength. Nevertheless, heteronuclear dipolar coupling remains significant at moderate MAS frequencies, necessitating the use of 1H dipolar decoupling sequences (continuous irradiation on 1H channel) for achieving sharp lines (Figure 1.4b). For 1H-1H homonuclear coupling, besides the application of decoupling techniques, ultra- fast MAS (60-100 kHz) is imperative at present. 1.2.3 Pulse Sequences and Density Operator A pulse sequence is a structured set of RF pulses, time delays, and gradient pulses used in NMR experiments. It serves as a precise set of instructions for manipulating nuclear spins in a sample and acquiring relevant data. In NMR, pulse sequences are tailored to achieve specific research goals, providing insights into relaxation times, spectral information, and molecular interactions. Figure 1.5a shows the pulse sequence of 13C Direct Polarization (DP) experiment with a spin echo (Hahn echo)27, 28. As one of the most renowned pulsed NMR experiments and the most fundamental building blocks, the spin echo plays a crucial role by restoring the coherence of 10 dephased nuclear spins. It can refocus offsets (Ω) and/or chemical shifts and the phase angle by a τ- π- τ sequence, regardless of the initial conditions. As shown in the vector representation of Figure 1.5b, at the end, the magnetization of two spins aligns along the same axis, effectively eliminating the impact of offset. Finally, the time-domain Free Induction Decay (FID) signal collected in the acquisition stage is converted into a frequency-domain spectrum via Fourier Transform (FT). Figure 1.5 Spin echo experiment. a, The pulse sequence of 13C DP with echo and decoupling. b, The vector representation. c, Three distinct spin-echo sequences in heteronuclear spin systems. The utilization of the density operator (product operator) formalism in dissecting NMR experiments signifies a quantum-mechanical strategy, setting it apart from the semiclassical vector model. By harnessing the concept of density operators, we gain the ability to aptly illustrate significant phenomena, such as spin echo. For example, in Figure 1.5b, the original density operator for spin 1 is 𝐼̂𝑧. The 90° pulse at (i) make 𝐼̂𝑧 rotate along x-axis, giving: 𝐼̂𝑧 𝜋 𝐼̂𝑥 2 → 𝑐𝑜𝑠 𝜋 2 𝐼̂𝑧 + 𝑠𝑖𝑛 𝜋 2 𝐼̂𝑦 ≡ 𝐼̂𝑦 Eq 1.8 Note, in this thesis, we follow the left-hand rotation standard. At (ii), a free precession caused by offset result in a rotation about z-axis: 11 𝐼̂𝑦 Ω1𝜏𝐼̂𝑧 → 𝑐𝑜𝑠Ω1𝜏𝐼̂𝑦 + 𝑠𝑖𝑛Ω1𝜏𝐼̂𝑥 Eq 1.9 Then, at (iii), a 180° pulse is applied about x-axis: 𝑐𝑜𝑠Ω1𝜏𝐼̂𝑦 + 𝑠𝑖𝑛Ω1𝜏𝐼̂𝑥 𝜋𝐼̂𝑥 → − 𝑐𝑜𝑠Ω1𝜏𝐼̂𝑦 + 𝑠𝑖𝑛Ω1𝜏𝐼̂𝑥 Eq 1.10 Only the y-component changes sign. Finally, at (iv), second free precession is along z-axis: −𝑐𝑜𝑠Ω1𝜏𝐼̂𝑦 + 𝑠𝑖𝑛Ω1𝜏𝐼̂𝑥 Ω1𝜏𝐼̂𝑧 → −𝑐𝑜𝑠2Ω1𝜏𝐼̂𝑦 − 𝑠𝑖𝑛Ω1𝜏𝑐𝑜𝑠Ω1𝜏𝐼̂𝑥 + 𝑐𝑜𝑠Ω1𝜏𝑠𝑖𝑛Ω1𝜏𝐼̂𝑥 − 𝑠𝑖𝑛2Ω1𝜏𝐼̂𝑦 ≡ −(𝑐𝑜𝑠2Ω1𝜏 + 𝑠𝑖𝑛2Ω1𝜏)𝐼̂𝑦 ≡ −𝐼̂𝑦 Eq 1.11 Hence, we can concisely summarize the overall result of the spin echo sequence as: 𝐼̂𝑧 ° − τ − 180𝑥 90𝑥 → −𝐼̂𝑦 Eq 1.12 ° − τ The final outcome remains constant regardless of the offset (Ω) or the delay (τ), making the spin echo serves as a mechanism that effectively refocusing chemical shifts. Nonetheless, it's crucial to emphasize that the evolution of the scalar coupling may not always experience refocusing. Figure 1.5c illustrates three distinct scenarios. In the first situation, marked as (1), both chemical shifts of spins are refocused, but the J-coupling is not. In situations (2) and (3), the coupling's impact is refocused, while the chemical shifts of spins without the applied 180° pulse remain unrefocused. These assertions find support in the following density operator representations. J-coupling Hamiltonian can be written as 𝐻̂𝐽 = 2𝜋𝐽12𝐼̂1𝑧𝐼̂2𝑧 Eq 1.13 Then, for spin 1 in situation (1): 𝜋 𝐼̂1𝑦 − 2 → 𝐼̂1𝑥 𝐼̂1𝑧 2𝜋𝜏𝐽12𝐼̂1𝑧𝐼̂2𝑧 → 𝑐𝑜𝑠𝜋𝜏𝐽12𝐼̂1𝑥 − 𝑠𝑖𝑛𝜋𝜏𝐽122𝐼̂1𝑦𝐼̂2𝑧 𝜋𝐼̂1𝑥+𝜋𝐼̂2𝑥 → 𝑐𝑜𝑠𝜋𝜏𝐽12𝐼̂1𝑥 − 𝑠𝑖𝑛𝜋𝜏𝐽122𝐼̂1𝑦𝐼̂2𝑧 2𝜋𝜏𝐽12𝐼̂1𝑧𝐼̂2𝑧 → 𝑐𝑜𝑠2𝜋𝜏𝐽12𝐼̂1𝑥 − 𝑠𝑖𝑛2𝜋𝜏𝐽122𝐼̂1𝑦 Eq 1.14 12 Likewise, for spin 2 in scenario (1): 𝐼̂2𝑧 − τ − 180𝑥 ° 90−𝑦 → 𝑐𝑜𝑠2𝜋𝜏𝐽12𝐼̂2𝑥 − 𝑠𝑖𝑛2𝜋𝜏𝐽122𝐼̂2𝑦𝐼̂1𝑧 Eq 1.15 ° − τ The result shows that the coupling simply evolves for 2𝜏 and generates anti-phase magnetization instead of being refocused. However, in situations (2) and (3), the 180° pulse (𝜋𝐼̂1𝑥 ) is only applied to one spin, causing a refocused J-coupling: 𝐼̂1𝑧 𝜋 𝐼̂1𝑦 − 2 → 2𝜋𝜏𝐽12𝐼̂1𝑧𝐼̂2𝑧 → 𝜋𝐼̂1𝑥 → 2𝜋𝜏𝐽12𝐼̂1𝑧𝐼̂2𝑧 → 𝐼̂1𝑥 Eq 1.16 Based on the discussion of spin echo, in heteronuclear systems, we are free to decide whether to permit the evolution of the offset and the coupling. This flexibility grants significant latitude in creating and controlling pivotal anti-phase states in multiple-pulse NMR experiments. 1.2.4 One- and Multi-Dimensional Correlation NMR The progress in ssNMR techniques, including the utilization of high magnetic field spectrometers, high and ultra-high MAS probes, and the implementation of robust pulse sequences, has opened up avenues for acquiring a wide range of atomic-level information. These techniques offer insights into the structure, interactions, dynamics, and polymorphism of cell wall macromolecules. The primary method for distinguishing magnetically inequivalent units and rapidly screening cell wall samples is one-dimensional (1D) 13C ssNMR. By employing the initial magnetization as a dynamic filter, a diverse range of dynamically heterogeneous polysaccharides present in cell walls can be identified. Firstly, the Dipolar-based 1H-13C CP experiment29, 30 primarily detects rigidly coupled molecules, and selective detection can be achieved by adjusting CP contact time (Figure 1.4). Secondly, the DP experiment is also extensively employed for directly polarizing 13C spins, as illustrated in Figure 1.5a. When a long recycle delay, typically set 13 at 5 times of the T1 (spin-lattice relaxation time) value, is employed, it allows for complete relaxation of the nuclear spins. It enables the detection of all molecules within the system, leading to quantitative results with both rigid and mobile molecules information. Conversely, the use of a short recycle delay is designed to selectively detect mobile molecules, such as those with shorter T1 relaxation times. Furthermore, for the analysis of the most dynamic or even solvated molecules within a sample, the scalar-based (J-coupling) 1H-13C refocused-INEPT experiment31 is a powerful and indispensable tool (Figure 1.6a). The 1H magnetization of rigid segments experiences rapid loss, due to fast T2 (spin-spin) relaxation caused by the strong 1H-1H dipolar coupling. In addition, MultiCP (Figure 1.6b) offer a time-efficient alternative to the long recycle delay DP experiment, facilitating quantitative detection, particularly in natural abundance samples32. This method accomplishes extended CP from 1H nuclei, surpassing a duration of 6-10 milliseconds, all while minimizing magnetization losses due to relaxation. It operates with a moderate duty cycle of radio- frequency irradiation, achieved through a sequence of multiple 1-ms CP periods, which are interleaved with 1H spin-lattice relaxation periods designed to repolarize the protons. Multidimensional NMR plays a pivotal role in deciphering complex carbohydrate structures within cell walls, enhancing resolution and simplifying spectral analysis. We can categorize them into two groups: homonuclear and heteronuclear correlation experiments. Heteronuclear resonance assignments are crucial for comprehensive structure determination and intermolecular interactions exploration, such as the two-dimensional (2D) 1H-13C HETCOR33, 34 (Figure 1.6c). During this experiment, 1H homonuclear couplings are removed by Phase- Modulated Lee–Goldburg (PMLG) or frequency-switched Lee-Goldburg (FSLG) decoupling35, and the resulting narrow lines for rigid-solid protons are observed after the t1 evolution. By changing CP contact time that adjusts polarization transfer from 1H to 13C, intra- and 14 intermolecular interactions can be selectively detected. In addition, a mixing period that allows for 1H-1H spin diffusion (SD) can be added before the CP block. The HETCOR method is mainly used in the wetland soil project in Chapter 9. 2D 13C-13C homonuclear correlation experiments fall into two categories: through-space via dipolar coupling, such as CHHC, RFDR, PDSD, DARR, and CORD, and through-bond via J coupling (e.g., CP/DP-J-INADEQUATE). These experiments aid in resonance assignments, covalent linkage pattern determination, and estimating the molar percentage of sugar residues in matrix polysaccharides. Particularly, the CHHC experiment36 is a prime choice for natural-abundance MAS-DNP, due to the 13C–1H–1H–13C pathway for polarization transfer. It effectively resolves the challenge of diagonal peak dominance in the spectrum of unlabeled materials, which dominance arises from the limited probability of observing off-diagonal internuclear correlations at natural 13C abundance levels. The CHHC experiment has been applied to study the unlabeled rice stems in Chapter 8. In order to provide a clearer demonstration of functionality of multidimensional NMR, here I will employ one sugar residue of 3-fold xylan (Xn) and one lignin guaiacyl (G) unit as illustrative examples, which exampled chemical shifts are shown in Figure 1.6e. Typically, at the outset of our projects, following the initial 1D screening, we acquire 2D INADEQUATE spectra to assign chemical shifts for each component within the system (Figure 1.6f). The 2D 13C-J- INADEQUATE method37 is a powerful NMR technique that provides double quantum (DQ)- single quantum (SQ) correlations and offers a comprehensive producing diagonal-free asymmetric connectivity map for all carbon atoms within the molecule, making it a definitive tool for structural assignments of biomolecules. In the generated refocused-INADEQUATE spectrum of xylan, which displays unambiguous through-bond correlations via J-coupling, the signals of two directly bonded carbon atoms are located at coordinates (𝛿Ca, 𝛿Ca + 𝛿Cb) and (𝛿Cb, 𝛿Ca + 𝛿Cb) on the 15 plot (Figure 1.6f). Here, 𝛿Ca and 𝛿Cb represent the SQ chemical shifts of carbon sites a and b, respectively. These values are projected onto the X-axis, corresponding to the direct dimension (𝜔2) in the experimentally measured NMR spectra. The sum of these two SQ chemical shifts (δCa + δCb) is projected onto the Y-axis, representing the DQ chemical shift for this carbon pair. This corresponds to the indirect dimension (𝜔1) in the experimentally measured NMR spectra. Each pyranose unit in xylan is depicted by ten scatter dots based on the chemical shifts of seven carbons, such as the spin pairs of the carbonyl group (AcCO at 174 ppm and AcMe at 21 ppm). Via the evolving of density operator, we can prove how the DQ coherence is created. In the INADEQUATE pulse sequence, the transverse 13C-magnetization is generated by the first 90° pulse (or CP). If we set τ = 1/(4JIS), after the first τ - 180° - τ period we can get: 𝜌(0) τ − 180° − τ → −2𝐼̂𝑥𝑆̂𝑧 − 𝐼̂𝑧𝑆̂𝑥 Eq 1.17 The DQ coherence is created during the following 90° pulse for t1 evolution: −2𝐼̂𝑥𝑆̂𝑧 − 𝐼̂𝑧𝑆̂𝑥 ° 90𝑥 → −2𝐼̂𝑥𝑆̂𝑦 − 𝐼̂𝑦𝑆̂𝑥 Eq 1.18 Finally, the third 90° pulse converts the DQ coherence to SQ one, followed by the second τ - 180° - τ sequence that establishes an in-phase coherence for t2 detection, where τ = 1/(4JIS): −2𝐼̂𝑥𝑆̂𝑦 − 𝐼̂𝑦𝑆̂𝑥 ° 90𝑥 → 2𝐼̂𝑥𝑆̂𝑧 + 𝐼̂𝑧𝑆̂𝑥 τ − 180° − τ → 𝐼̂𝑦 + 𝑆̂𝑦 Eq 1.19 Similar to 1D experiments, the interplay between direct polarization and J-coupling predominantly highlights mobile segments, making the DP J-INADEQUATE experiment inclined toward dynamic signals. In parallel, the analogous version of the CP J-INADEQUATE experiment is designed to exclusively capture rigid signals. 16 Figure 1.6 Several typical ssNMR experiments. a, 1D Insensitive nuclei enhancement by polarization transfer (INEPT) experiment with heteronuclear dipolar decoupling (DD). b, 1D MultiCP. c, 2D 1H–13C heteronuclear chemical shift correlation (HETCOR). d, 2D CHHC. e, Chemical shifts of exampled molecules in (f), (g), and (h). f, 2D CP/DP based J- incredible natural abundance double quantum transfer experiment (INADEQUATE) and a generated spectrum of one xylan residue. g, 2D rotational echo double resonance (REDOR) and a pseudo spectrum of one unit of G lignin. h, 2D 13C–13C correlation, such as proton-driven spin diffusion (PDSD), dipolar assisted rotational resonance (DARR), and combined R2n υ-driven (CORD). In the generated spectrum, intramolecular interactions inside of xylan and G lignin are colored blue and magenta, respectively, while intermolecular peaks are denoted by green dots. 17 Another commonly employed pulse sequence for structural assignments is RFDR38, 39 (Figure 1.6g), which direct recouples 13C-13C dipolar coupling for polarization transfer (commonly referred to as first-order recoupling). Though it is categorized as through-space methods, it is particularly useful for selectively identifying short-range (one or two-bond) correlations by effectively suppressing multi-bond interactions with brief recoupling time (~1-2 ms). This technique utilizes multiple 180° pulses (2n) on the 13C channel to reintroduce 13C-13C homonuclear dipolar couplings through longitudinal polarization transfer. Each pulse should be precisely positioned at the midpoint of a single rotor period. The pseudo-RFDR spectrum illustrates the anticipated one-bond connections within a single G-lignin unit. The graph clearly reveals that every pair of bonded carbon atoms exhibits mutual cross-peaks. However, the OMe group, lacking immediate neighboring carbon atoms, remains devoid of any discernible signals. Upon obtaining structural assignments, we can delve into the exploration of intra- and inter-molecular interactions using 2D 13C-13C correlation experiments that rely on dipolar recoupling during the mixing phase, as illustrated in Figure 1.6h. Short mixing times are valuable for detecting intra-residue correlations, aiding in resonance assignment. In contrast, longer mixing times unveil intermolecular cross peaks, offering insights into sub-nanometer molecular packing. In Figure 1.6h, we generate a pseudo 13C-13C correlation spectrum, where intramolecular interactions within xylan and G lignin marked in blue and magenta, respectively. Possible intermolecular peaks between xylan and G lignin are indicated by green dots, applying a longer mixing time. The foundation of these 2D 13C-13C correlation experiments lies in the reintroduction of 13C-1H dipolar couplings during the mixing phase, often referred to as second-order recoupling. Under moderate MAS conditions and without active recoupling, the residual 13C-13C dipolar 18 couplings are minimal, leading to slow polarization transfer. However, carbon-to-carbon interactions become viable through their interactions with surrounding protons. This is achieved by temporarily deactivating proton decoupling or employing active recoupling with radiofrequency power, thus reintroducing 13C-1H dipolar couplings (known as recoupling). This action broadens the line shape of carbon peaks and enables spectral overlap during the mixing period. Spectral overlap can occur between two carbon peaks with small chemical shift differences or between a carbon peak and the spinning sidebands of a distant carbon peak, preserving energy for polarization transfer. These experiments are less affected by severe dipolar truncation effects, making them particularly suitable for measuring long-range correlations. The thesis explores several common 2D 13C-13C correlation experiments, including PDSD, DARR, and CORD40-42. PDSD, for instance, offers a straightforward experimental setup and the capability to investigate both short-range and long-range correlations. In PDSD, the mixing period does not require radiofrequency irradiation, allowing for extended mixing times of several seconds without concerns about radiofrequency-induced heating. The rate of spin diffusion buildup serves as a qualitative indicator of internuclear distances, providing semi-quantitative insights into spatial proximities. PDSD is most effective at slower spinning frequencies (around 10 kHz) and lower magnetic fields (less than 14 T). However, its efficiency decreases notably at higher MAS frequencies and magnetic fields. To partially address these limitations, the DARR sequence can be employed by applying RF irradiation on the 1H channel (𝜔1 = 𝜔𝑟). Nevertheless, hardware constraints, such as sample heating by the RF pulse, limit the duration of the mixing period, rendering DARR transfer inefficient when the MAS frequency surpasses 30 kHz. The CORD sequence has been developed to facilitate spin diffusion at faster MAS and features broad-band homonuclear dipolar recoupling. CORD spectra exhibit more consistent cross-peak intensities 19 across the spectrum, rendering them advantageous for homonuclear correlation studies in biological samples. The flexibility to combine various sequence blocks makes ssNMR a remarkably versatile technique for structural exploration. Pulse sequences like INADEQUATE, PDSD, and RFDR serve as fundamental building blocks for creating more intricate pulse sequences43, such as 3D DQ-SQ-SQ experiment44 introduced in Chapter 5. It is comprised of a dipolar-base INADEQUATE block followed by a CORD block, which enables unambiguous resonance assignments of complex carbohydrates in cell walls. Moreover, the incorporation of dipolar coupling, J-coupling, and relaxation filters is instrumental in spectral-editing methods, which enable the selective detection of specific sites within a molecule or specific molecules within a sample. For instance, the water-edited experiment4, 45 (Figure 1.7a) has become a valuable tool for understanding the role of water in biological systems. These experiments allow for the selective detection of water-proximal sites or interactions, shedding light on the dynamic behavior of water molecules and their influence on the structure and function of macromolecules. Typically, by using relaxation filters, like the 1H-T2 filter, water signals can be selectively isolated. This is based on the dynamic nature of water’s 1H magnetization, which enables it to persist through a moderately long 1H-T2 filter (in the order of milliseconds). Then, the selected 1H magnetization from water is manipulated and transferred to the neighboring 1H in biomolecules through 1H-1H spin diffusion, further propagated to 13C nuclei through CP. The resulting spectra provide crucial insights into the local environment and interactions of water molecules with biomolecules. Similarly, the Dipolar-Gated PDSD experiment7, 14 is employed to enhance the detection of aromatic signals within lignin, which are typically challenging to discern in plant cell walls due 20 to the substantial CSA in ssNMR. This specialized experiment is designed to selectively capture lignin signals while minimizing interference from proton-rich polysaccharides. The key innovation in this approach involves reintroducing 13C-1H dipolar coupling during an asymmetrical dipolar- dephasing period before t1 evolution. This period comprises two un-decoupled delays with respect to the 180° pulse in the Hahn echo. This strategic design efficiently targets non-protonated carbons within the aromatic lignin, while simultaneously suppressing signals originating from protonated carbons. As illustrated in Figure 1.7b, when compared to conventional 1D CP or 2D DARR experiments, both the 1D and 2D Gated-PDSD experiments exhibit a preference for detecting non- protonated carbons within lignin and alleviate the spectral asymmetry introduced by the heterogeneous proton density within the aromatic motifs. Figure 1.7 Edited-NMR experiments. a, Water-edited experiment with 1H-T2 filter for selective detection of biopolymers in close proximity to water. b, Dipolar-gated 13C-13C correlation experiment with dipolar-dephasing designed to selectively detect non-protonated carbons within the aromatic lignin. 21 1.2.5 Molecular Motions: Rates and Amplitudes Beyond the fundamental principles, NMR encompasses relaxation processes that govern the behavior of nuclear spins and the properties of NMR signals42, 46, 47. These processes are broadly categorized as longitudinal relaxation (T1) and transverse relaxation (T2). T1 relaxation, also called spin-lattice relaxation, dictates the rate at which nuclear spins revert to their equilibrium positions along the magnetic field after being perturbed by RF pulses. The T1 relaxation time offers insights into the dynamics and interactions of molecules, providing information about molecular motion and the correlation times of molecular processes. T2 Relaxation, also named spin-spin relaxation, accounts for the loss of phase coherence among nuclear spins in the transverse plane. This relaxation process directly influences the width of NMR signals and, consequently, the resolution of NMR spectra. T2 plays a vital role in NMR applications where high-resolution spectra and well-defined line shapes are desired. Understanding these relaxation processes is critical for tailoring NMR experiments to the specific properties of the sample under investigation. Various relaxation processes in ssNMR spectroscopy exhibit sensitivity to molecular motions occurring on distinct timescales. 13C-T1 Relaxation primarily reflects motions occurring on the nanosecond timescale, which is commonly measured by the DP-based inversion recovery experiment (Figure 1.8a). In a nutshell, this experiment involves flipping the initial magnetization to the negative z-direction through a 180° pulse. Giving a specific delay period (τ), the magnetization is expected to relax back towards the equilibrium magnetization in the positive z- direction. Then, a following 90° pulse flips the magnetization to the x-y plain for further detection. In the case of a single exponential fitting, the following equation is typically employed to quantitatively find the T1 relaxation constant: 𝐼 = 1 − 2𝑒−𝑡/𝑇1 Eq 1.20 22 13C-T1 Torchia sequence48 is also a powerful tool for the measurement of T1 relaxation times (Figure 1.8b). This method leverages CP to generate transverse magnetization in 13C nuclei, which favors the detection of rigid species within the sample, and subsequently flips magnetization to the positive-z direction, allowing spin-spin relaxation via varying delay τ. After the second 90° pulse, we can monitor how rapidly the magnetization returns to equilibrium, using the equation for a single exponential fitting: 𝐼 = 𝑒−𝑡/𝑇1 Eq 1.21 Figure 1.8 Molecular motion measurements. a, DP-based 13C-T1 inversion recovery. b, CP-based 13C-T1 Torchia. c, 13C-detected 1H-T1ρ pulse sequence with LG-spin lock. d, 2D 1H-13C dipolar chemical-shift (DIPSHIFT) experiment. e, representative 1H-13C dipolar dephasing curves under 7.5 kHz MAS. Corresponding order parameters are labeled based on CH or CH2 group. 1H-T1ρ Relaxation16, also referred to proton spin-lattice relaxation in the rotating frame, provides insights into motions on the microsecond timescale (Figure 1.8c). In the beginning, a 90° pulse followed by a 35° pulse is applied to flip the 1H magnetization to the magic angle. Following, a spin-locking field with a frequency in the range of tens of kHz is applied to the proton channel, which prevents precession in the rotating frame. Then, the LG-CP technique is employed to selectively transfer the 1H polarization to the directly bonded 13C nuclei for the detection. The 23 measurement of 13C-T2 relaxation42 can be accomplished using the CP echo experiment, with the concurrent application of 1H heteronuclear decoupling. The presence of decoupling is essential in this context to prevent the destruction of magnetization due to 13C-1H dipolar coupling. This differs from T1 measurements, where 1H decoupling is not necessary. In ssNMR, the determination of motional amplitudes in polysaccharides is often achieved through the employment of the DIPSHIFT experiment16, 35, as illustrated in Figure 1.8d. The order parameters, denoted as S(CH), derived from this experiment offer valuable information regarding the motional amplitudes of the C-H bonds in the biomolecules under investigation (Figure 1.8e). Notably, a smaller order parameter corresponds to more significant amplitude in molecular motions, signifying increased flexibility or mobility within the C-H bonds. Conversely, a larger order parameter indicates a rigid molecular structure, such as cellulose crystalline regions. In the DIPSHIFT experiment, to realign chemical shifts, a 180° pulse is strategically inserted after a single rotor period. Moreover, the removal of 1H-1H homonuclear couplings is meticulously accomplished using decoupling sequences like FSLG. This precise setup ensures that the 13C nuclei exclusively evolve under the influence of 13C-1H dipolar couplings. The resultant spectrum in this experiment displays 13C isotropic chemical shifts in the direct dimension, while the indirect dimension reveals 13C-1H dipolar dephasing curves. These dephasing curves are subjected to fitting procedures to extract the apparent 13C-1H dipolar couplings. To obtain the true dipolar coupling, accounting for molecular motion, these apparent couplings are divided by scaling factors associated with the homonuclear decoupling sequences, typically set at 0.577 for FSLG. Subsequently, the true dipolar coupling is further divided by the theoretical rigid-limit 13C-1H one- bond dipolar coupling (~ 22.7 kHz) to get the order parameters. It's worth noting that the CH2 group has a deeper dephasing curve when compared to that of the CH group, despite having 24 identical order parameters, as depicted in Figure 1.8e. 1.2.6 Resolution and Sensitivity Two critical aspects of NMR spectroscopy are sensitivity and resolution, which play pivotal roles in the quality and applicability of ssNMR experiments. Resolution in NMR spectroscopy pertains to the ability to distinguish and separate individual resonances in the NMR spectrum. Resolution, often characterized by parameters like the full width at half maximum (FWHM), is inversely related to the apparent T2 relaxation time. Enhancing resolution can be achieved by promoting crystallinity within the sample, thus slowing down intrinsic T2 relaxation. Improved coil design and careful sample positioning contribute to reducing field inhomogeneity, resulting in longer apparent T2 values. Biological samples often face resolution limitations due to peak overlap resulting from the structural similarity of their constituent units. Strategies to address this challenge include selective and sparse labeling of samples. On the spectral side, advanced pulse sequences, including extra dimensions (3D) and spectral editing, offer better resolution and sensitivity, allowing for the analysis of complex samples. Residual anisotropic interactions, such as lingering 1H-1H homonuclear dipolar couplings, can cause peak broadening that standard spinning cannot fully eliminate. Ultra-fast spinning49-51 at rates ranging from 40 to 120 kHz using small rotors (e.g., 0.7 mm) significantly narrows the 1H linewidth, enabling 1H detection of nanomole to femtomole quantities in deuterated proteins under high magnetic fields. In the data analysis stage, there are opportunities to further enhance resolution, although these improvements need to scarify sensitivity. Various window functions can be applied to time domain signals to enhance resolution by focusing on the later part of the FID. The choice of window functions, such as Gaussian, exponential, sine, and qsine, should be made thoughtfully to achieve the desired balance between sensitivity and resolution. Another technique involves zero- 25 filling, which increases digital resolution by appending zeros to the end of the FID. NMR signals originate from a minor surplus of nuclei occupying the lower energy state, with nuclear spins distributed almost evenly at room temperature. This distribution follows the Boltzmann distribution equation: 𝑁+ 𝑁− = 𝑒−∆𝐸/𝑘𝑇 Eq 1.22 Here, k is the Boltzmann constant. At room temperature and within a 10 T magnetic field, the population of the lower energy state (N+) exceeds that of the higher energy state (N-) by a factor of only 1 in 10,000. This small difference in population results in inherently weak NMR signals. Consequently, NMR spectra exhibit low signal-to-noise ratios (S/N) due to these weak signals. 𝑆/𝑁 ∝ 𝑛𝛶𝐸𝑇−1𝑓𝐶𝑃𝜂(𝛶𝑑 3𝐵0 3𝑡) 1 2(𝑉𝑄) 1 2 Eq 1.23 This equation52 offers numerous avenues for enhancing NMR sensitivity. The “n” represents the number of spins. Hence, we can increase the sensitivity by achieving isotope- labeling (12C to 13C), packing more samples into the rotor, or switching to a thin-walled or a larger rotor to contain more spins. Lowering the temperature (T) by employing cryogenic NMR probes and measuring in a higher magnetic field (𝐵0) are another two straightforward approaches. For example, an 800 MHz spectrometer provides a 2.8-time higher sensitivity than a 400 MHz spectrometer. Additionally, “𝑓𝐶𝑃 ”, “ 𝜂 ”, “V”, and “Q” are CP factor, probe-specific factors, signifying the coil filling factor, coil volume, and coil quality factor, respectively. Another strategy for sensitivity improvement is to optimize the effectiveness of the experimental time factor “t”. 4 times of the number of scans (NS) in experiments provide 2-fold sensitivity enhancement. By introducing paramagnetic agents53, 54 into the samples, it is possible to reduce the T1 relaxation times, further, to reduce the recycle delay (typically 5 times T1), through the paramagnetic relaxation enhancement effect. This approach significantly accelerates data 26 acquisition cycles, resulting in time savings of approximately 5-20 folds. Non-uniform sampling55, a technique still under development, holds promise as a method to better allocate the acquisition time during data sampling. This approach enhances sensitivity without sacrificing resolution. In the equation, "𝛶𝐸" and "𝛶𝐷" stand for the gyromagnetic ratios of the excited and detected nucleus, respectively. It is the reason why a ssNMR pulse sequence always starts with a CP block to enhance sensitivity by exciting 1H instead of 13C. Furthermore, direct 1H-detection contributes to increased sensitivity, owing to the high gyromagnetic ratio of the proton nucleus (𝛶1H=4𝛶13𝐶). It is the reason that the 13C-1H HETCOR has an 8 times enhancement when compared to 1H-13C HETCOR. DNP is a cutting-edge innovation that facilitates the transfer of polarization from electrons (excitation) to NMR-active nuclei. As a result, sensitivity enhancements of tens to hundreds of folds can be achieved through DNP56. This also translates to substantial time savings, reducing experimental durations by a factor of 102 to 105. 1.3 Sensitivity-Enhancing DNP In recent developments, the innovative technique of MAS-DNP has demonstrated remarkable potential in addressing a long-standing challenge21, 57-60. It is widely acknowledged that NMR spectroscopy suffers from limited sensitivity due to the minute nuclear spin polarization, accounting for less than 0.007% of 1H, resulting from the small gyromagnetic ratio (γ) of the spins. MAS-DNP has the remarkable capability to facilitate polarization transfer from unpaired electrons, which are approximately 660 times larger in γ than 1H (Table 1.1), to protons when subjected to microwave irradiation (as illustrated in Figure 1.9). This transfer leads to a substantial enhancement in NMR sensitivity, typically ranging from one to two orders of magnitude. The essential microwave resource in this process is the gyrotron, which generates power in the range of 10 to 100 watts. This power is harnessed to induce electron paramagnetic resonance60 (EPR) in 27 mono- or bi-radicals, each possessing one or two unpaired electrons61-63. To facilitate this process, nitroxide-based radicals (for instance, AMUPol64) are commonly dissolved in DNP matrix solutions (such as the mixture of D2O, H2O, and 13C-depleted, d8-glycerol). These solutions are then combined with the solid sample through grinding or vertexing. In order to achieve optimal results, MAS-DNP experiments necessitate cryogenic temperatures within the range of 90–110 K. These low temperatures are essential to circumvent the challenge of the short lifetime of radicals and further improve the S/N ratios by creating a glassy phase within the DNP matrix. This glassy phase efficiently prevents the aggregation of radicals. Figure 1.9 Sensitivity-enhancing DNP. a, DNP instruments. b, DNP mechanism. c, DNP radicals. d, 16-fold sensitivity enhancement of unlabeled soil samples provided by 600 MHz/395 GHz DNP. Numerous features can affect the enhancement factor (𝜀), such as radicals, solvents, and even rotor materials. AsymPolPOK65, an asymmetric biradical that been reported recently, is capable to significantly shorten DNP build-up time, leading to a larger enhancement factor and a shorter experiment time; TEKPol dissolved in ortho-terphenyl enables DNP experiment conducted at near room temperature, which benefits for the detection of matrix polysaccharides that is limited at cryogenic temperature; TinyPol and TEMTriPol-I have a good performance on high-field (18.8T, 800 MHz/527 GHz) DNP instruments with resolution improvement. The DNP matrix protocols also influence the distribution of radicals in cell wall compounds and impact the 28 enhancement factor by formation or separation of the glassy phase of solvents. The broadly used protocols in cell wall studies include 13C-depleted, d8-glycerol/D2O/H2O (60/30/10 vol%), D2O/H2O (90/10), d6-DMSO/D2O/H2O (60/30/10), and even a matrix-free protocol (only D2O), which can effectively avoid radical aggregation66. With the remarkable sensitivity enhancement brought about by the DNP technique, it becomes feasible to perform high-resolution structural analysis of unlabeled biological systems using ssNMR. In this dissertation, with assistance of MAS-DNP, (1) a comprehensive toolbox for systematically illustrating structural models in intact natural abundance plant biomass is established (Chapter 8); (2) molecular composition of organic carbon within wetland soil samples obtained from a brackish island situated along the Gulf of Mexico coastline is analyzed (Chapter 9); (3) unexpected lignin–cellulose interactions in 13C-wood cell walls are confirmed (Chapter 7). 1.4 Sample Preparation for ssNMR and DNP Measurements To conduct ssNMR studies on plant cell wall samples, it is essential to enrich the samples with NMR-active isotopes, such as carbon-13. This can be achieved through two distinct methods, each offering specific advantages based on the research focus. The first method67, 68 typically involves the enrichment of young seedlings (aged up to 2 weeks) for the study of carbohydrates in primary cell walls. These seedlings are grown in darkness in a solution medium containing 13C-glucose or in sealed chambers connected to 13CO2 resources. After harvesting, the seedlings are rapidly frozen and then ground using liquid nitrogen. Subsequently, a 1.5% (w/v) sodium dodecyl sulfate (SDS) solution is used for a 3-hour wash to remove cell membranes, proteins, and inactivate endogenous wall-degrading enzymes. Starch is eliminated by incubating the sample with α-amylase in a sodium metabisulfite (MES) buffer (pH 6.8). To further break down proteins, Pronase is employed in a sodium MES buffer (pH 7.5). 29 Throughout these processes, a solution containing 0.02% (w/v) NaN3 is used to prevent microbial growth. After washing and centrifugation, the resulting never-dried cell walls (approximately 30 mg) can be packed into 3.2-mm rotors for ssNMR experiments. The second isotope-labeling method14, 69 provides greater versatility in terms of plant species and developmental stages, including plants like maize and woody species (several months old) for the study of secondary cell walls. In this approach, plants are cultivated in sealed chambers using a hydroponics system to prevent 12CO2 contamination from soil microbial activity. The composition of the hydroponic solution is adjusted to suit the specific species, supporting minerals, and pH requirements. A gas mixing system combines 99.8% 13CO2 gas with a 20% O2 in N2 gas mix, resulting in a final concentration of 600 ppm of 13CO2. Ideal LED lights with specific wavelengths and intensity are used to minimize heat generation, and humidity and temperature are carefully controlled. After harvesting, fresh plant tissues are cut into small pieces and evenly packed into rotors for ssNMR analysis. Sensitivity-enhanced DNP has opened doors to studying wetland soil or plant samples in their natural and unlabeled state (1.11%-13C). This method57, 70 involves the integration of samples with a suitable DNP solvent, commonly referred to as the "matrix," along with the inclusion of a polarizing agent. These polarizing agents, often based on nitroxides, such as AMUPol, TOTAPOL, and AsymPolPok, are distinguished for their unreactive nature and their affinity for dissolution in various aqueous solvents. In a carefully orchestrated procedure, a concentration typically ranging from 5 to 20 mM of these radicals is expertly dissolved in cryoprotective solvents. These solvents encompass a selection of options, including 13C-depleted d8-glycerol, D2O, and H2O, d6-DMSO in combination with D2O and H2O, d6-DMSO with H2O alone, or a matrix-free approach that relies solely on D2O. The resulting DNP solution, imbued with the distinct properties of these noble 30 radicals, is artfully blended with unlabeled samples, ensuring a deep integration that permits these radicals to permeate the plant cell walls or allows a homogeneous mixing of radicals with soil. These well-blended samples are subsequently packed into 3.2 mm zirconium or sapphire rotors. These rotors embrace a payload of roughly 30 to 50 milligrams, serving as the staging ground for the forthcoming DNP experiment. The optimization of the enhancement factor (εon/off) and the precise calibration of the DNP buildup time are exacting and tailored to the unique characteristics of each sample. In particular, soil samples may require an additional treatment with a 2% hydrofluoric acid (HF) solution30 before mixing with the DNP solution to remove mineral components without altering their chemical structure. 1.5 Thesis Organization My dissertation primarily comprises published papers and preprint manuscripts during my Ph.D. research, focusing on the development and application of ssNMR, with the assistance of DNP, to investigate plant cell walls and wetland soil samples. Besides sections of abstract, acknowledgment, and table of contents, another 11 chapters are also structured into the thesis. An overview of each chapter is specified below. Chapter 1 serves as an introductory chapter, presenting background information from four key aspects: (1) the importance of studying plant cell walls and wetland soil; (2) an introduction to ssNMR theory and methodology; (3) sensitivity-enhanced DNP and natural abundance samples. (4) ssNMR sample preparation and isotope labeling. Additionally, subsections of thesis organization and copyright permissions are included. Chapters 2, 3, and 4 are review papers, as additional introductions, which highlight the recent breakthroughs in using ssNMR to elucidate cell walls and summarize underexplored aspects of cell wall studies. Chapter 2 reviews the research on carbohydrate and cell walls using ssNMR 31 and discusses the potential to address unresolved questions by innovative methods, such as proton detection, quadrupolar nuclei, paramagnetic relaxation enhancement, DNP, ultrahigh field, and database. Chapter 3 focuses on the current limitations of ssNMR in cell wall research, including challenges in resonance assignment due to structural complexity, the need for more effective methods to handle the ever-growing dataset, and the optimization of protocols and sample preparation for the application of high-field DNP. Chapter 4 summarizes how ssNMR has contributed to our understanding of the structure of extracellular matrices, such as the dynamic structure of living fungal cell walls and macro fibrillar structure in intact plant tissues. The chapter also discusses the combination of NMR with other techniques to address unresolved structural questions. Chapters 5 and 6 are published research articles, focusing on pulse programs that serve as the foundation of NMR experiments. In Chapter 5, we introduce a three-dimensional (3D) 13C- 13C-13C correlation experiment designed to enable high-resolution analysis and unambiguous resonance assignments of complex carbohydrates in cell walls. This ssNMR 3D pulse program incorporates a CORD block at the end of an INADEQUATE block, offering an unprecedented resolution for understanding the polymorphic structure of chitin in fungi samples and cellulose in plant cell walls. In Chapter 6, we discuss how the phases of radio frequency pulses affect signal phases during cross polarization. Through a first principles approach, we reveal that the NMR signal is simply dependent on arbitrary CP RF phases under both static and magic angle spinning conditions. Lignin is a polyphenolic biopolymer deposited during the formation of plant secondary cell walls, which provides strength, assists water transporting, and resists microbial attacks. Gaining a comprehensive understanding of the lignin-carbohydrates interactions is pivotal for crop 32 development and enhancing the quality and durability of wood products. In Chapter 7, we achieve the atomic-resolution characterization of the polymorphic structure, dynamical and hydration profile, and physical packing of lignin and polysaccharides using intact plant cells in hardwoods (eucalyptus and poplar) and softwood (spruce). Our findings uncover that the extent of glycan- aromatic association increases sequentially across grasses, hardwoods, and softwoods; and lignin principally packs with the xylan in a non-flat conformation via electrostatic interactions and partially binds the junction of flat-ribbon xylan and cellulose surface as a secondary site. The emerging sensitivity-enhancing technique DNP holds great promise for investigating natural abundance samples, eliminating the need for isotope labeling. Chapters 8 and 9 provide two examples showcasing the application of DNP to unlabeled rice and wetland soil samples, respectively. In Chapter 8, we demonstrate the feasibility of combining traditional ssNMR with DNP methods to examine the structure and dynamics of carbohydrates in natural abundance rice stems. This time-efficient approach allows for the screening of a large collection of biomass materials without labeling. In Chapter 9, DNP enables rapid acquisition of 2D 13C/1H-13C correlation spectra to detail the molecular fingerprint and spatial organization of biopolymers in unlabeled wetland soil. Surprisingly, the lignocellulosic core identified in the plants grown on this island were found to be preserved in the surface layer of soil formed in the brackish marshes, but with tighter physical packing between molecules. Extending the depth from the surface to 2 m covers a geological timeline of eleven centuries, where we found sophisticated changes in the molecular structure and composition of organic matters as well as the bulk properties of the wetland soil. Chapters 10 and 11 focus on statistical approaches for the analysis of carbohydrate composition and structure using high-resolution ssNMR data, which are indexed in a carbohydrate 33 database. Chapter 10 introduces the Complex Carbohydrates Magnetic Resonance Database (CCMRD, www.ccmrd.org), as the first ssNMR database specifically designed for complex carbohydrates. This chapter also presents the assembly of the database, information search functionalities, and the deposition of data into CCMRD. In Chapter 11, We generate simulated spectra (density maps) using data deposited in CCMRD, to demonstrate the chemical shift dispersion and to aid in the fast identification of important fungal and plant polysaccharides in cell walls. By comparing the projected spectra with experimental data, we highlight the challenges of assigning resonances because distinct carbohydrates from different organisms can produce nearly identical signals. 1.6 Copyright Permissions Chapters 2, 3, 4, 5, 6, 7, 8, 9, 10, and 11 consist of reprinted published papers. Elsevier does not necessitate a license for re-use in a dissertation or thesis. Papers published in Biomedcentral (BMC), Springer Nature (Nature Communication), and John Wiley and Sons (Magnetic Resonance in Chemistry) are available in open access, and no permission is necessary for their re-use. Detailed copyright permissions have been secured from the following groups: Chapter 2: Elsevier. Chapter 3: Elsevier. Chapter 4: Elsevier. Chapter 5: Elsevier. Chapter 6: Elsevier. Chapter 7: Springer Nature (Nature Communication), open access. Chapter 8: Biomedcentral (BMC), open access. Chapter 9: Submitted to Nature Geoscience. 34 Preprinted in ChemRxiv (https://doi.org/10.26434/chemrxiv-2023-p4rvg). Chapter 10: Springer Nature. License Number: 5663191296494. Chapter 11: John Wiley and Sons (Magnetic Resonance in Chemistry), open access. 35 REFERENCES 1. Himmel, M. E.; Ding, S.-Y.; Johnson, D. K.; Adney, W. S.; Nimlos, M. R.; Brady, J. W.; Foust, T. D., Biomass Recalcitrance: Engineering Plants and Enzymes for Biofuels Production. Science 2007, 315 (5813), 804-807. 2. Petridis, L.; Smith, J. C., Molecular-level driving forces in lignocellulosic biomass deconstruction for bioenergy. Nat. Rev. Chem. 2018, 2 (11), 382-389. 3. Zhang, Y.; Yu, J.; Wang, X.; Durachko, D. M.; Zhang, S.; Cosgrove, D. J., Molecular insights into the complex mechanics of plant epidermal cell walls. Science 2021, 372 (6543), 706-711. 4. White, P. B.; Wang, T.; Park, Y. B.; Cosgrove, D. J.; Hong, M., Water–Polysaccharide Interactions in the Primary Cell Wall of Arabidopsis thaliana from Polarization Transfer Solid- State NMR. J. Am. Chem. Soc. 2014, 136 (29), 10399-10409. 5. Kirui, A.; Du, J.; Zhao, W.; Barnes, W.; Kang, X.; Anderson, C. T.; Xiao, C.; Wang, T., A pectin Methyltransferase modulates polysaccharide dynamics and interactions in Arabidopsis primary cell walls: Evidence from solid-state NMR. Carbohydr. Polym. 2021, 270, 118370. 6. Gao, Y.; Lipton, A. S.; Wittmer, Y.; Murray, D. T.; Mortimer, J. C., A grass-specific cellulose– xylan interaction dominates in sorghum secondary cell walls. Nat. Commun. 2020, 11 (1), 6081. 7. Kang, X.; Kirui, A.; Dickwella Widanage, M. C.; Mentink-Vigier, F.; Cosgrove, D. J.; Wang, T., Lignin-polysaccharide interactions in plant secondary cell walls revealed by solid-state NMR. Nat. Commun. 2019, 10 (1), 1-9. 8. Zhao, W.; Fernando, L. D.; Kirui, A.; Deligey, F.; Wang, T., Solid-state NMR of plant and fungal cell walls: a critical review. Solid State Nucl. Magn. Reson. 2020, 107, 101660. 9. Sapkota, Y.; White, J. R., Long-term fate of rapidly eroding carbon stock soil profiles in coastal wetlands. Sci. Total Environ. 2021, 753, 141913. 10. Haywood, B. J.; Hayes, M. P.; White, J. R.; Cook, R. L., Potential fate of wetland soil carbon in a deltaic coastal wetland subjected to high relative sea level rise. Sci. Total Environ. 2020, 711, 135185. 11. Roth, V.-N.; Lange, M.; Simon, C.; Hertkorn, N.; Bucher, S.; Goodall, T.; Griffiths, R. I.; Mellado-Vazquez, P. G.; Mommer, L.; Oram, N. J.; Weigelt, A.; Dittmar, T.; Gleixner, G., Persistence of dissolved organic matter explained by molecular changes during its passage through soil. Nat. Geosci. 2019, 2019, 755-761. 12. Osland, M. J.; Gabler, C. A.; Grace, J. B.; Day, R. H.; McCoy, M. L.; McLeod, J. L.; From, A. S.; Enwright, N. M.; Feher, L. C.; Stagg, C. L.; Hartley, S. B., Climate and Plant Controls on Soil Organic Matter in Coastal Wetlands. Glob. Chang. Biol. 2018, 24, 5361-5379. 13. Plaza, C.; Pegoraro, E.; Bracho, R.; Celis, G.; Crummer, K. G.; Hutchings, J. A.; Hicks Pries, C. E.; Mauritz, M.; Natali, S.; Salmon, V. G.; Schadel, C.; Webb, E. E.; Schuur, E. A. 36 G., Direct observation of permafrost degradation and rapid soil carbon loss in tundra. Nat. Geosci. 2019, 12, 627-631. 14. Kirui, A.; Zhao, W.; Deligey, F.; Yang, H.; Kang, X.; Mentink-Vigier, F.; Wang, T., Carbohydrate-aromatic interface and molecular architecture of lignocellulose. Nat. Commun. 2022, 13, 538. 15. Ghassemi, N.; Poulhazan, A.; Deligey, F.; Mentink-Vigier, F.; Marcotte, I.; Wang, T., Solid- State NMR Investigations of Extracellular Matrices and Cell Walls of Algae, Bacteria, Fungi, and Plants. Chem. Rev. 2021, in press. 16. Zhao, W.; Kirui, A.; Deligey, F.; Mentink-Vigier, F.; Zhou, Y.; Zhang, B.; Wang, T., Solid- state NMR of unlabeled plant cell walls: high-resolution structural analysis without isotopic enrichment. Biotechnol. Biofuels 2021, 14 (1), 14. 17. Kirui, A.; Ling, Z.; Kang, X.; Widanage, M. C. D.; Mentink-Vigier, F.; French, A. D.; Wang, T., Atomic resolution of cotton cellulose structure enabled by dynamic nuclear polarization solid-state NMR. Cellulose 2019, 26 (1), 329-339. 18. Barna, J. C. J.; Laue, E. D.; Mayger, M. R.; Skilling, J.; Worrall, S. P., Exponential sampling, an alternative method for sampling in two-dimensional NMR experiments. J. Magn. Reson. 1969, 73 (1), 69-77. 19. Wind, R. A.; Duijvestijn, M. J.; van der Lugt, C.; Manenschijn, A.; Vriend, J., Applications of dynamic nuclear polarization in 13C NMR in solids. Prog. Nucl. Mag. Res. Sp. 1985, 17, 33- 67. 20. Schmidt-Rohr, K.; Spiess, H. W., Multidimensional Solid-State NMR and Polymers. Academic Press: San Diego, 1994; p 13-68. 21. Hall, D. A.; Maus, D. C.; Gerfen, G. J.; Inati, S. J.; Becerra, L. R.; Dahlquist, F. W.; Griffin, R. G., Polarization-Enhanced NMR Spectroscopy of Biomolecules in Frozen Solution. Science 1997, 276 (5314), 930-932. 22. Levitt, M., Spin dynamics: Basics of Nuclear Magnetic Resonance, Second Edition. Wiley- Blackwell: 2008; Vol. 37. 23. Reif, B.; Ashbrook, S. E.; Emsley, L.; Hong, M., Solid-state NMR spectroscopy. Nat. Rev. Methods Primers 2021, 1 (1), 2. 24. Duer, M. J., The contribution of solid-state NMR spectroscopy to understanding biomineralization: atomic and molecular structure of bone J. Magn. Reson. 2015, 253, 98-110. 25. Levitt, M. H.; Grant, D. M.; Harris, R. K., Symmetry-based pulse sequences in magic-angle spinning solid-state NMR. In Solid State NMR Studies of Biopolymers, McDermott, A. P., T., Eds, Ed. 2007; p 229. 26. Haw, J. F.; Maciel, G. E.; Schroeder, H. A., Carbon-13 nuclear magnetic resonance 37 spectrometric study of wood and wood pulping with cross polarization and magic-angle spinning. Anal. Chem. 1984, 56 (8), 1323-1329. 27. Garbow, J. R.; Jacob, G. S.; Stejskal, E. O.; Schaefer, J., Protein dynamics from chemical shift and dipolar rotational spin-echo 15N NMR. Biochemistry 1989, 28 (3), 1362-7. 28. Gullion, T., Introduction to rotational-echo, double-resonance NMR. Concepts Magn. Reson. 1998, 10 (5), 277-289. 29. Yamamoto, H.; Horii, F., CPMAS carbon-13 NMR analysis of the crystal transformation induced for Valonia cellulose by annealing at high temperatures. Macromolecules 1993, 26 (6), 1313-1317. 30. Skjemstad, J.; Clarke, P.; Taylor, J.; Oades, J.; Newman, R., The Removal of Magnetic Materials from Surface Soils - A Solid State 13C CP/MAS NMR Study. Soil Res. 1994, 32, 1215- 1229. 31. Elena, B.; Lesage, A.; Steuernagel, S.; Böckmann, A.; Emsley, L., Proton to Carbon-13 INEPT in Solid-State NMR Spectroscopy. J. Am. Chem. Soc. 2005, 127 (49), 17296-17302. 32. Duan, P.; Schmidt-Rohr, K., Composite-pulse and partially dipolar dephased multiCP for improved quantitative solid-state 13C NMR. J. Magn. Reson. 2017, 285, 68-78. 33. Chen, L.; Kaiser, J. M.; Lai, J.; Polenova, T.; Yang, J.; Rienstra, C. M.; Mueller, L. J., J- based 2D homonuclear and heteronuclear correlation in solid-state proteins. Magn. Reson. Chem. 2007, 45 (S1), S84-S92. 34. Sengupta, I.; Nadaud, P. S.; Helmus, J. J.; Schwieters, C. D.; Jaroniec, C. P., Protein fold determined by paramagnetic magic-angle spinning solid-state NMR spectroscopy. Nat. Chem. 2012, 4 (5), 410-417. 35. Bielecki, A.; Kolbert, A. C.; De Groot, H.; Griffin, R. G.; Levitt, M. H., Frequency-Switched Lee—Goldburg Sequences in Solids. Adv. Magn. Reson. 1990, 14, 111-124. 36. Aluas, M.; Tripon, C.; Griffin, J. M.; Filip, X.; Ladizhansky, V.; Griffin, R. G.; Brown, S. P.; Filip, C., CHHC and 1H–1H magnetization exchange: Analysis by experimental solid-state NMR and 11-spin density-matrix simulations. J. Magn. Reson. 2009, 199 (2), 173-187. 37. Cadars, S.; Sein, J.; Duma, L.; Lesage, A.; Pham, T. N.; Baltisberger, J. H.; Brown, S. P.; Emsley, L., The refocused INADEQUATE MAS NMR experiment in multiple spin-systems: Interpreting observed correlation peaks and optimising lineshapes. J. Magn. Reson. 2007, 188 (1), 24-34. 38. Kono, H.; Numata, Y., Structural investigation of cellulose I α and I β by 2D RFDR NMR spectroscopy: determination of sequence of magnetically inequivalent D-glucose units along cellulose chain. Cellulose 2006, 13 (3), 317-326. 39. Gullion, T.; Schaefer, J., Rotational-echo double-resonance NMR. J. Magn. Reson. 1989, 81 38 (1), 196-200. 40. Hou, G.; Yan, S.; Sun, S.; Han, Y.; Byeon, I.-J. L.; Ahn, J.; Concel, J.; Samoson, A.; Gronenborn, A. M.; Polenova, T., Spin Diffusion Driven by R-Symmetry Sequences: Applications to Homonuclear Correlation Spectroscopy in MAS NMR of Biological and Organic Solids. J. Am. Chem. Soc. 2011, 133 (11), 3943-3953. 41. Hou, G.; Yan, S.; Trébosc, J.; Amoureux, J.-P.; Polenova, T., Broadband homonuclear correlation spectroscopy driven by combined R2nv sequences under fast magic angle spinning for NMR structural analysis of organic and biological solids. J. Magn. Reson. 2013, 232, 18-30. 42. Wang, T.; Williams, J. K.; Schmidt-Rohr, K.; Hong, M., Relaxation-compensated difference spin diffusion NMR for detecting 13C–13C in proteins and polysaccharides. J. Biomol. NMR 2015, 61 (2), 97-107. long-range correlations 43. Li, S.; Zhang, Y.; Hong, M., 3D 13C-13C-13C Correlation NMR for De Novo Distance Determination of Solid Proteins and Application to a Human Alpha Defensin. J. Magn. Reson. 2010, 202, 203-210. 44. Shekar, S. C.; Zhao, W.; Fernando, L. D.; Hung, I.; Wang, T., A 13C three-dimensional DQ- SQ-SQ correlation experiment for high-resolution analysis of complex carbohydrates using solid- state NMR. J. Magn. Reson. 2022, 336, 107148. 45. Gelenter, M. D.; Wang, T.; Liao, S.-Y.; O’Neill, H.; Hong, M., 2H–13C correlation solid-state NMR for investigating dynamics and water accessibilities of proteins and carbohydrates. J. Biomol. NMR 2017, 68 (4), 257-270. 46. Nadaud, P. S.; Helmus, J. J.; Kall, S. L.; Jaroniec, C. P., Paramagnetic ions enable tuning of nuclear relaxation rates and provide long-range structural restraints in solid-state NMR of proteins. J. Am. Chem. Soc. 2009, 131 (23), 8108-8120. 47. Clore, M. G.; Iwahara, J., Theory, Practice, and Applications of Paramagnetic Relaxation Enhancement for the Characterization of Transient Low-Population States of Biological Macromolecules and Their Complexes. Chem. Rev. 2009, 109 (9), 4108-4139. 48. Torchia, D. A., Measurement of proton-enhanced 13C T1 values by a method which suppresses artifacts. J. Magn. Reson. 1978, 30, 613-616. 49. Zhang, R.; Mroue, K. H.; Ramamoorthy, A., Proton-Based Ultrafast Magic Angle Spinning Solid-State NMR Spectroscopy. Acc. Chem. Res. 2017, 50 (4), 1105-1113. 50. Phyo, P.; Hong, M., Fast MAS 1H–13C correlation NMR for structural investigations of plant cell walls. J. Biomol. NMR 2019, 73 (12), 661-674. 51. Roos, M.; Wang, T.; Shcherbakov, A. A.; Hong, M., Fast Magic-Angle-Spinning 19F Spin Exchange NMR for Determining Nanometer 19F–19F Distances in Proteins and Pharmaceutical Compounds. J. Phys. Chem. B 2018, 122 (11), 2900-2911. 39 52. Ishii, Y.; Tycko, R., Sensitivity Enhancement in Solid State 15N NMR by Indirect Detection with High-Speed Magic Angle Spinning. J. Magn. Reson. 2000, 142 (1), 199-204. 53. Oster, C.; Kosol, S.; Hartlmuller, C.; Lamley, J. M.; Luga, D.; Oss, A.; Org, M.-L.; Vanatalu, K.; Samoson, A.; Madl, T.; Lewandowski, J. R., Characterization of Protein–Protein Interfaces in Large Complexes by Solid-State NMR Solvent Paramagnetic Relaxation Enhancements. J. Am. Chem. Soc. 2017, 139 (35), 12165-12174. 54. Zuckerstätter, G.; Terinte, N.; Sixta, H.; Schuster, K. C., Novel insight into cellulose supramolecular structure through 13C CP-MAS NMR spectroscopy and paramagnetic relaxation enhancement. Carbohydr. Polym. 2013, 93 (1), 122-128. 55. Kaur, M.; Lewis, C. M.; Chronister, A.; Phun, G. S.; Mueller, L. J., Non-Uniform Sampling in NMR Spectroscopy and the Preservation of Spectral Knowledge in the Time and Frequency Domains. J. Phys. Chem. A 2020, 124 (26), 5474-5486. 56. Märker, K.; Pingret, M.; Mouesca, J.-M.; Gasparutto, D.; Hediger, S.; De Paëpe, G., A New Tool for NMR Crystallography: Complete 13C/15N Assignment of Organic Molecules at Natural Isotopic Abundance Using DNP-Enhanced Solid-State NMR. J. Am. Chem. Soc. 2015, 137 (43), 13796-13799. 57. Rankin, A. G. M.; Trébosc, J.; Pourpoint, F.; Amoureux, J.-P.; Lafon, O., Recent developments in MAS DNP-NMR of materials. Solid State Nucl. Magn. Reson. 2019, 101, 116- 143. 58. Mentink-Vigier, F.; Akbey, Ü.; Oschkinat, H.; Vega, S.; Feintuch, A., Theoretical aspects of Magic Angle Spinning - Dynamic Nuclear Polarization. J. Magn. Reson. 2015, 258, 102-120. 59. Mentink-Vigier, F.; Paul, S.; Lee, D.; Feintuch, A.; Hediger, S.; Vega, S.; De Paëpe, G., Nuclear depolarization and absolute sensitivity in magic-angle spinning cross effect dynamic nuclear polarization. Phys. Chem. Chem. Phys. 2015, 17 (34), 21824-21836. 60. Stevanato, G.; Casano, G.; Kubicki, D. J.; Rao, Y.; Esteban Hofer, L.; Menzildjian, G.; Karoui, H.; Siri, D.; Cordova, M.; Yulikov, M.; Jeschke, G.; Lelli, M.; Lesage, A.; Ouari, O.; Emsley, L., Open and Closed Radicals: Local Geometry around Unpaired Electrons Governs Magic-Angle Spinning Dynamic Nuclear Polarization Performance. J. Am. Chem. Soc. 2020, 142 (39), 16587-16599. 61. Mentink-Vigier, F.; Dubroca, T.; Van Tol, J.; Sigurdsson, S. T., The distance between g- tensors of nitroxide biradicals governs MAS-DNP performance: The case of the bTurea family. J. Magn. Reson. 2021, 329, 107026. 62. Mentink-Vigier, F., Optimizing nitroxide biradicals for cross-effect MAS-DNP: the role of g- tensors’ distance. Phys. Chem. Chem. Phys. 2020, 22 (6), 3643-3652. 63. Perras, F. A.; Pruski, M., Large-scale ab initio simulations of MAS DNP enhancements using a Monte Carlo optimization strategy. J. Chem. Phys. 2018, 149 (15), 154202. 40 64. Sauvée, C.; Rosay, M.; Casano, G.; Aussenac, F.; Weber, R. T.; Ouari, O.; Tordo, P., Highly Efficient, Water-Soluble Polarizing Agents for Dynamic Nuclear Polarization at High Frequency. Angew. Chem. Int. Ed. 2013, 52 (41), 10858-10861. 65. Mentink-Vigier, F.; Marin-Montesinos, I.; Jagtap, A. P.; Halbritter, T.; van Tol, J.; Hediger, S.; Lee, D.; Sigurdsson, S. T.; De Paëpe, G., Computationally Assisted Design of Polarizing Agents for Dynamic Nuclear Polarization Enhanced NMR: The AsymPol Family. J. Am. Chem. Soc. 2018, 140 (35), 11013-11019. 66. Chaudhari, S. R.; Wisser, D.; Pinon, A. C.; Berruyer, P.; Gajan, D.; Tordo, P.; Ouari, O.; Reiter, C.; Engelke, F.; Copéret, C.; Lelli, M.; Lesage, A.; Emsley, L., Dynamic Nuclear Polarization Efficiency Increased by Very Fast Magic Angle Spinning. J. Am. Chem. Soc. 2017, 139 (31), 10609-10612. 67. Wang, T.; Park, Y. B.; Cosgrove, D. J.; Hong, M., Cellulose-pectin spatial contacts are inherent to never-dried Arabidopsis primary cell walls: evidence from solid-state nuclear magnetic resonance. Plant Physiol. 2015, 168 (3), 871-884. 68. Wang, T.; Hong, M., Solid-state NMR investigations of cellulose structure and interactions with matrix polysaccharides in plant primary cell walls. J. Exp. Bot. 2016, 67 (2), 503-514. 69. Gao, Y.; Mortimer, J. C., Unlocking the architecture of native plant cell walls via solid-state nuclear magnetic resonance. Methods Cell Biol. 2020, 160, 121-143. 70. Kirui, A.; Dickwella Widanage, M. C.; Mentink-Vigier, F.; Wang, P.; Kang, X.; Wang, T., Preparation of Fungal and Plant Materials for Structural Elucidation Using Dynamic Nuclear Polarization Solid-State NMR. J. Vis. Exp. 2019, (144), e59152. 41 CHAPTER 2: SOLID-STATE NMR OF PLANT AND FUNGAL CELL WALLS: A CRITICAL REVIEW Review paper reprinted with permission: Solid State Nucl. Magn. Reson. 107, 101660 (2020) Authors: Wancheng Zhao, Liyanage D. Fernando, Alex Kirui, Fabien Deligey, and Tuo Wang 2.1 Abstract The cell walls of plants and microbes are a central source for bio-renewable energy and the major targets of antibiotics and antifungal agents. It is highly challenging to determine the molecular structure of complex carbohydrates, protein and lignin, and their supramolecular assembly in intact cell walls. This article selectively highlights the recent breakthroughs that employ 13C/15N solid-state NMR techniques to elucidate the architecture of fungal cell walls in Aspergillus fumigatus and the primary and secondary cell walls in a large variety of plant species such as Arabidopsis, Brachypodium, maize, and spruce. Built upon these pioneering studies, we further summarize the underexplored aspects of fungal and plant cell walls. The new research opportunities introduced by innovative methods, such as the detection of proton and quadrupolar nuclei on ultrahigh-field magnets and under fast magic-angle spinning, paramagnetic probes, natural-abundance DNP, and software development, are also critically discussed. 2.2 Cell Walls: a Medically Important and Energy-Relevant Biomaterial The cell wall is a carbohydrate-rich coating outside the plasma membrane of plants and many microorganisms. The cell walls in photosynthesis systems, such as plants, algae, and green bacteria, are transformed from solar energies and carbon dioxide, with enormous value as a primary source of building materials, textiles, biofuel, nanocomposites, and high-value reagents1. Polysaccharides in the cell walls of fungal pathogens and invading bacteria are absent in human cells; these components trigger immune recognition and serve as the major target of antifungal 42 drugs and antibiotics2, 3. Polysaccharides and other biomolecules (such as protein and lignin) are held together by covalent linkages and physical packing interactions to form a mechanically strong composite, which allows the cell to retain integrity and morphology under external stress. Nevertheless, the numerous interactions between biopolymers also pose a challenge for post- harvest processing and utilization of biomass and make it technically difficult to characterize these biomaterials with high resolution. Recently, magic-angle-spinning (MAS) solid-state NMR spectroscopy has been extensively employed to investigate intact cell walls. Uniformly isotope-labeled samples are produced by feeding the organism of interest with 13CO2 or solid/liquid media containing 13C- glucose and 15N-salts4. Multidimensional 13C-13C/15N correlation spectra collected on whole cells or isolated cell walls provide the atomic resolution needed for determining the polymorphic structure, intermolecular interaction, water contact, and molecular motions of biomolecules in their cellular environment (Figure 2.1). Within the last decade, a large variety of biosystems have been studied: the primary and secondary cell walls of seven plant species, including Arabidopsis thaliana, Brachypodium distachyon and Zea mays (maize), rice, switchgrass, poplar and spruce5- 14; the biofilm or cell walls of bacterial and fungal pathogens such as Aspergillus fumigatus, Cryptococcus neoformans, and Bacillus subtilis15-21; as well as the carbohydrate components in microalgae Chlamydomonas reinhardti22. Here we will review the major findings related to plant and fungal cell walls, emphasize the key questions awaiting investigation, and discuss the future directions enabled by the improved instrumentation and methodology, in the attempt to inspire innovative research in carbohydrate and cell wall NMR. 43 2.3 Recent Advances in Cell Wall Research by Solid-State NMR 2.3.1 Molecular Insight of Plant Primary Cell Walls Since 2010, Hong and colleagues have been pioneering the investigations of primary plant cell walls, a component synthesized in the growing plants (Figure 2.2a-2.2c)23,24. The composition is well known from numerous biochemical studies, and three major types of polysaccharides are present25, 26. Cellulose microfibrils are formed by 18 or more glucan chains (3-4 nm across) and they are highly rigid and partially crystalline. The backbones of pectin, such as rhamnogalacturonan-I (RG-I) and homogalacturonan (HG), are often acidic and responsible for regulating cell wall hydration. Hemicellulose interacts with cellulose and pectin, with a plant- dependent composition: the major hemicellulose is xyloglucan (XyG) in dicots, such as Arabidopsis, but changes to glucuronoarabinoxylan (GAX) and mixed-linkage glucans (MLG) in commelinid monocots (grasses) such as Brachypodium and maize. Using 13C-labeled and isolated cell walls, three ground-breaking discoveries were reported, which have revised and substantiated our limited understanding of primary cell wall architecture. First, cellulose, hemicellulose, and pectin are found to associate noncovalently on the sub- nanometer scale to form an integrated network. In wild-type Arabidopsis (Figure 2.2a, Sample 1), a large number of cross peaks have been identified between pectin and cellulose, which were previously considered to be phase-separated 5. The equilibrium intensity of 13C-13C spin diffusion suggests that 25-50% of cellulose are in close proximity to pectin6, 27. This polymer interaction is independent of the sample’s hydration history28 and can be fully preserved after partial depectination by CDTA and sodium carbonate (Figure 2.2a, Sample 2), which disrupts the calcium crosslinking of HG and consequently removing the interfibrillar HG molecules (40% of all HG) that are not binding cellulose28-30. Due to the loss of immobilized water in the depectinated 44 sample, the rate of 1H-1H polarization transfer from water to polymers have been globally slowed down for all polysaccharides, which can be partially restored by the subsequent digestions of XyG using xyloglucanase and Cel12A enzymes due to the enhanced surface areas of the residual macromolecules (Figure 2.2a, Sample 3)29. In addition, the removal of XyG using an xxt1xxt2xxt5 triple knockout line (Figure 2.2a, Sample 4) markedly enhances the dynamics of the remaining polysaccharides 5, which echoes with the global alternation of 1H spin diffusion observed in the sequentially digested samples, revealing a single network of all polysaccharides. It is noteworthy that a weaker pectin-cellulose interaction is often accompanied by the chemical modification of pectin structure, for example, a higher degree of methyl esterification, an increased occurrence of sidechain branching by arabinan or galactan, a reduced extent of calcium-crosslinking, and promoted HG aggregation. These molecular changes macroscopically correlate with faster growth, for example, in the inflorescence stem of Arabidopsis with a segmentally increasing rate of elongation from the base to the apical region (the tip)31, in the PGX1AT mutant that produces smaller pectin but larger plants (Figure 2.2a, Sample 5)32, and in a low-pH sample that mimics the acid growth condition (Figure 2.2a, Sample 6)33. Figure 2.1 Solid-state NMR strategy for investigating cell wall materials. Second, with the assistance from Dynamic Nuclear Polarization (DNP) and paramagnetic methods, two methods have been developed to reveal how a class of proteins (expansin) unfasten the polysaccharide joints to mediate cell expansion34, 35. Expansins lack the lytic activity expected 45 for wall-loosening enzyme and have been assumed to disrupt the non-covalent contacts between polysaccharides36. Solid-state NMR studies have shown that expansins precisely perturb the cellulose-xyloglucan nexus in Arabidopsis but disrupt the junctions between the highly and lowly substituted GAX in maize (Figure 2.2c); therefore, expansins bind different polysaccharides in the cell walls with distinct composition. Figure 2.2 Cell walls and biomolecules studied by solid-state NMR. a, NMR-derived conceptual 46 Figure 2.2 (cont’d) models of primary cell walls of Arabidopsis (dicot), including the intact and wild-type cell walls at different pH values, as well as multiple mutants that attenuate the structure and content of matrix polysaccharides (pectin or XyG), sequentially digested cell walls that are chemically depectinated followed by enzymatical removal of XyG. The sample numbers are labeled to facilitate discussion. b, Chemical structures of the biomolecular components in primary plant cell walls. c, Structural scheme of primary grass cell walls based on data collected using Brachypodium and maize. d, Plant secondary (2nd) cell wall architecture of Arabidopsis, maize, and softwood spruce. e, Representative structure of polymers that are unique to secondary plant cell walls. f, the structure and polymers of fungal cell walls in A. fumigatus. Third, with the sharp 13C linewidths on high magnetic fields (0.7-1.0 ppm for cellulose on an 800 MHz NMR) and the chemical shift calculations using Density Functional Theory (DFT), we have resolved seven types of glucose units in the cellulose of Arabidopsis and grass primary cell walls, determined their hydroxymethyl conformations via 1H-1H distance measurement, and localized these conformers in the microfibrils37-39. These forms deviate noticeably from the crystallographic structures of Iα and Iβ allomorphs obtained using the highly crystalline cellulose from bacteria and tunicates (a marine animal). In addition, these seven types of glucose residues have been consistently observed in the secondary cell walls of Arabidopsis, maize, switchgrass, and rice8, as well as multiple woody plants such as Eucalyptus, poplar, and spruce (unpublished results). Therefore, the Iα and Iβ model allomorphs are generally absent in most natural resources. So far, the NMR signals of Iα and Iβ structures have only been observed in cotton, thus a large crystallite is a prerequisite for accommodating the model structures40. 47 2.3.2 Lignin-Carbohydrate Packing in Plant Secondary Cell Walls The secondary cell wall is formed once the cell ceases expansion and it comprises the majority of lignocellulosic biomass. In secondary cell walls, cellulose microfibrils aggregate into larger bundles (10–20 nm across), which are further embedded in a matrix containing the aromatic polymer lignin and hemicellulose such as xylan and glucomannan 41. Lignin-carbohydrate interactions confer the biomass with recalcitrance to chemical and enzymatical treatments; therefore, it is of broad interest to understand the chemical principles underlying these polymer interactions. Paul Dupree, Ray Dupree, and colleagues have conducted several studies to recognize the functional relevance of xylan polymorphism in Arabidopsis secondary cell walls. It is found that only the xylan with a 2-fold helical screw symmetry and a regular pattern of acetate or glucuronate substitutions can bind cellulose microfibrils10-12. Stimulated by these discoveries, we have investigated the mature stems of maize, rice, switchgrass, and Arabidopsis, using a series of 2D 13C-13C correlation methods specially designed for enhancing the aromatic signals of lignin and detecting the lignin-carbohydrate interface (Figure 2.3a)8. Hundreds (234) of intermolecular cross-peaks have been identified, which pinpoint six categories of packing interactions between the different functional groups in lignin and carbohydrates as illustrated in Figure 2.3b-2.3g. Strikingly, lignin mainly interacts with xylan rather than cellulose. In addition, the number and intensities of these cross peaks statistically correlate with the number of methyl ether substitutions in lignin residues (Figure 2.3b), which signposts a prevalent role of electrostatic contacts in stabilizing polymer interface. Integrating the information on polymer packing, dynamics, and hydration has resulted in a molecular view of lignocellulosic materials: lignin self-aggregates to form dynamically unique and hydrophobic nanodomains, with surface contact to the non-flat xylan (3-fold) through abundant electrostatic 48 interactions8. This xylan-lignin interface links to the flat-ribbon domain of xylan that is coating the surface of cellulose microfibrils (Figure 2.2d, left). Figure 2.3 Intermolecular interactions in plant secondary cell walls by dipolar-filtered PDSD spectra. a, Overlay of 2D 13C-13C correlation spectra measured with a short (0.1 s, yellow) and long (1.0 s, black) mixing time. Six representative regions of intermolecular cross peaks for the structural illustration in panels b-g are highlighted in dashline circles or rectangle. These interactions happen between. b, xylan acetyl (Ac) group and lignin aromatics (S, G, or H), c, xylan acetyl group and cellulose microfibrils (i: internal glucan chains; s: surface chains), d, lignin methyl group (OMe: methyl ether) and xylan acetyl groups, e, lignin methyl group and cellulose, f, the aromatic carbons of different lignin units, as well as g, lignin aromatics and xylan furanose ring. The blue crosses in panels a and b highlight the missing signal of AcMe-H4 cross peak between xylan and the H-residue of lignin. In the softwood spruce, xylan also binds cellulose through its 2-fold conformer while 49 galactoglucomannan (GGM), a unique hemicellulose in softwoods, binds the surface of cellulose microfibrils in a semi-crystalline manner13. Since both GGM and xylan have shown two domains, one coating cellulose and the other filling interfibrillar space, it is proposed that some GGM and xylan bind to the same microfibril and further associate with lignin (Figure 2.2d, right). Figure 2.4 Polymer packing and hydration in fungal cell walls of A. fumigatus. a, Top: Illustration of the observed cross peaks between α-1,3-glucan (A) and chitin (Ch). Bottom: overlay of 2D 13C- 13C correlation spectra measured with 100 ms DARR (orange, short-range) and 15 ms PAR (black, long-range). b, Overlay of the control (black) and water-edited spectra (orange); the cross section extracted at 69 ppm shows apparent signal dephasing of chitin and α-1,3-glucan in water-edited spectra. c, Site-specific hydration map of biopolymers. 2.3.3 Insights into the Fungal Cell Wall Architecture Recently, we have also initiated a project elucidating the cell wall structure of an airborne fungal pathogen A. fumigatus. The samples are measured alive at room-temperature on an 800 MHz NMR; the 13C linewidths are as narrow as 0.4-0.6 ppm for the rigid polysaccharides and 0.3- 0.5 ppm for the relatively mobile carbohydrates and proteins. With the remarkable resolution, the 13C and 15N signals of 7 types of polysaccharides, including α-1,3-glucan, chitin (a nitrogenated polysaccharide), mannan, and three types of β-glucans, together with their 23 conformers, have 50 been identified (Figure 2.2f)15. Long-range correlation methods, such as 13C-13C and 15N-15N Proton-Assisted Recoupling (PAR)42 as well as NCACX measured with a variable 13C-13C mixing time, have been employed to identify in total 65 intermolecular interactions. Most of these physical interactions occur between chitin and α-1,3-glucans (Figure 2.4a), which also show high hydrophobicity (Figure 2.4b, c) and rigidity. These two molecules are complexed to form a mechanical scaffold that is surrounded by a soft matrix of diversely linked β-glucans and capped by an external shell rich in glycoproteins (Figure 2.2f). This study has established a preliminary structural frame, which requires systematic validation and encourages structural investigations of individual cell wall molecules and their biomedical relevance. 2.4 Biochemical Perspectives: the Unresolved Questions In an earlier Trends article published in 2016, several underexplored areas in plant NMR have been summarized, which mainly include the coalescence of multiple cellulose microfibrils, the functional structure of lignin, and the putative interactions between polysaccharides and structural proteins24. So far, we have already obtained an in-depth understanding of lignin- carbohydrate interactions in secondary cell walls but also discovered more aspects that remain ambiguous. For instance, the sophisticated patterns of covalent linkages between lignin residues and their impact on lignin’s capability of interacting with polysaccharides have not been discussed. Cellulose-lignin interaction is scarce in maize and Arabidopsis, but this concept may not hold for the woody plants with a distinct composition of biomolecules and a more compact packing. The conformational relevance of glucomannan in spruce and other softwood species is not yet understood. We also need to understand the structural origin of the abundant electrostatic interactions between carbohydrates and aromatics and how these physical contacts contribute to the mechanical properties and digestibility of lignocellulosic materials. Inevitably, we also need 51 to figure out a way to integrate solid-state NMR results with the numerous studies using solution- NMR, which are focused on the covalent linkages in extracted residuals. An unexpected finding in the A. fumigatus fungus is the multifaced role of α-1,3-glucans. These molecules are simultaneously in association with chitin for stiffness and existing in the mobile phase15. This observation has countered the biochemical results in which α-1,3-glucans are extractable by strong alkali and thus constantly excluded from the structural core of any prevailing models43. Also, the amount of this polysaccharide is much lower in many other pathogens such as most yeasts (for example, Candida albicans); therefore, it is of great interest to understand the structural and dynamical heterogeneity of polysaccharides across fungal species. Another major polysaccharide, mannan, is found to coexist with proteins in the mobile domain of A. fumigatus cell walls, likely constituting the mannoproteins in the outermost layer as depicted in biochemical studies. Further evidence assessing the structural role of mannan and their covalent linkages with structural proteins are crucial to the understanding of this structurally dynamic shell that regulates cellular recognition and fungal pathogenesis44. It is also important to understand how the microbe re-structures its cell wall in response to antifungal agents, which will explain the origin of drug resistance from a structural perspective. 2.5 Technical Outlook: Opportunities Beyond Conventional Methods The past decade has seen the rapid development of solid-state NMR techniques. Here, we have selectively summarized a few technical advances that could potentially revolutionize the field and establish new research directions. These highlights have extended beyond the conventional 13C/15N-methods by involving other NMR-active nuclei or electrons, with assistance from ultra- fast MAS, ultrahigh magnetic field, DNP, as well as database and software coding. 52 2.5.1 1H and 19F under Fast Spinning: Carbohydrate Structure and Interactions Direct detection of proton resonance provides high-sensitivity due to the high isotope abundance (99.985%) and four-fold higher gyromagnetic ratio over 13C. The strong homonuclear couplings enable distance measurement beyond 1 nm and facilitate structural determination. Although proton detection has already been widely employed to study perdeuterated or protonated proteins45-48, it is rarely applied to the carbohydrates that are rich in protons with complex chemical environments. Recently, Hong and Phyo have conducted a set of proton-detection experiments, such as the 2D CH INEPT, 3D CHH INEPT-TOCSY, and 2D hChH RFDR techniques49-51, to assign the 1H resonances of polysaccharides and to determine their intermolecular packing, for example, through cross peaks between cellulose carbons and matrix polysaccharide protons, in 13C-labeled Arabidopsis primary cell walls52. The protonated material has been back exchanged in D2O, which suppresses the water intensities and reduces the contribution of hydroxyl protons. A moderately fast MAS frequency (30-50 kHz) is chosen to simultaneously enable proton-detection of the mobile matrix and filter out the signals of rigid microfibrils52. The narrow 1H linewidth (0.06 ppm on an 800 MHz spectrometer, Figure 2.5a) and the excellent agreement between solid-state 1H chemical shifts and solution-NMR observables consistently confirmed that the observed matrix polysaccharides are intrinsically mobile in cell walls. In addition, Simorre, Schanda, and coworkers have assigned the resonances of peptidoglycan in intact Bacillus subtilis under 100 kHz MAS, with representative 1H linewidths of 50-120 Hz (0.05-0.13 ppm) on a 950 MHz spectrometer53, 54. Using the 1H-1H RFDR scheme, the authors have identified multiple inter- residue cross peaks, including unambiguous cross peaks between the GlcNAc sugar and the L- alanine residue on the peptide stem, crossing a long distance of at least 5 Å53. These studies have presented a novel strategy for investigating complex biosystems and landed the stage for pursuing 53 1H investigations without labeling. Similarly, 19F has a high gyromagnetic ratio and 100% natural abundance. Adding to these merits is a large range of chemical shifts for resolving various chemical motifs. As demonstrated on pharmaceutical compounds, GB1 protein, and HIV-1 capsid protein, 19F-19F/1H distances can be measured on the nanometer scale (1-2 nm)55-57, which is a major extension from the reach of 13C and 15N methods. Typically, site-specific fluorination causes minimal perturbation to the structures of many proteins and materials56, 58, but may substantially disrupt the hydrogen bonds in carbohydrate polymers. An appropriate labeling scheme is needed to sparsely fluorinate carbohydrates without eliminating their functional structures and assembly59. 2.5.2 17O at Ultrahigh-Field: a New Biochemical Probe Oxygen is another core element that determines the hydrogen bonding and chemical properties of biomolecules. Carbohydrates are particularly rich in oxygen atoms, with at least one oxygen covalently linked to each single carbon site. Recently, the materialization of a world-record 1.5 GHz (35 Tesla) series-connected hybrid (SCH) NMR magnet60 and the commercial ultrahigh field instruments have presented a unique opportunity for high-resolution 17O studies. Griffin and colleagues have revealed the markedly improved resolution of 17O spectra on ultrahigh fields where the line-broadening by second-order quadrupole coupling is attenuated61. They have also collected 2D 13C/15N/1H-17O correlation spectra and determined internuclear distances through recoupling methods such as ZF-TEDOR and REAPDOR61, 62. Back in 2007, Grandinetti and coworkers have already pioneered the measurement of 17O MAS patterns for monosaccharides and disaccharides that are site-specifically labeled at either the hydroxyl or glycosidic oxygen sites (Figure 2.5b) 63. The C–O–H angle and C–O distances, instead of the O–H distances, are found to affect 17O quadrupolar couplings in carbohydrates. In addition, many other quadrupolar nuclei 54 may benefit from the availability of ultrahigh-field magnets. For example, 33S NMR could help characterize many sulfurated carbohydrates (such as the ulvan, carrageenan, and rhamnan sulfate) in marine species64. Another popular molecule is heparin, a sulfated glycosaminoglycan that prevents blood clotting as an anticoagulant agent and induces filament assembly of tau proteins 65. Combining quadrupolar NMR with ultrahigh field magnets provides a novel probe to the biochemically important sites in these carbohydrates. Figure 2.5 Advanced techniques that promote carbohydrate and cell wall research. a, 2D CH INEPT spectrum of 13C labeled Arabidopsis cell walls measured at 50kHz MAS on an 800 MHz NMR. Superscripts denote the different subtypes of monosaccharides and ambiguous assignments are shown in orange. b, Experimental (black) and simulated (blue) 17O MAS spectra at 19.6 T of methyl α-D-galactopyranoside (4-17O), methyl β-D-glucopyranoside (2-17O), and methyl α-D- glucopyranoside (6-17O) at 12 kHz MAS. c, 2D 2H-13C correlation spectrum of H/D exchanged and 13C-labeled Arabidopsis cell wall. The structure of cellulose and 2H-exchanged OD sites are given. d, The 72-ppm 2H cross-sections (black) and best-fit simulation (green) that is a 55 Figure 2.5 (cont’d) supercomposition of components with CQ values of 187 and 50 kHz with weighting factors of 74 and 26%, respectively. e, The structure of β-expansin (EXPB1) with 9 Cys residues shown in magenta. The 3 solvent-accessible cysteines that could be tagged with EDTA-Mn(II) are annotated. f, The PRE effect of Mn(II)-tagged expansin on glucuronoarabinoxylan (GAX). The intensities ratios of 2 spectra measured on the Mn-containing sample and the control sample are shown in red: GAX has significant signal dephasing due to binding to the protein. g, Difference PDSD spectrum of two parent spectra measured on Mn-containing or Mn-free samples: showing no cellulose intensities (iC4) but only the signals of matrix polysaccharides that bind proteins. 2.5.3 2H: Dynamics and Water Accessibility In cell walls, carbohydrate dynamics were primarily evaluated by measuring NMR relaxation and dipolar couplings66, and water-polymer contacts were mainly investigated using 1D/2D 13C-detected, 1H spin diffusion methods and dipolar-filtered heteronuclear 2D correlation techniques like MELODI-HETCOR28, 29, 67 Recently, Hong and coworkers have employed the Rotor Echo Short Pulse IRrAdiaTION mediated cross-polarization (RESPIRATIONCP) technique68, 69 to achieve multi-bond, broadband 2H–13C polarization using an affordable 2H radiofrequency field of ~50 kHz and a short contact time below 1.7 ms70. A rapid trans-gauche isomerization is identified in perdeuterated bacterial cellulose. This hydroxymethyl motion around the C5–C6 bond is absent in the interior glucan chains of cellulose but occurs to the surface chains as revealed by their motionally averaged C6-2H quadrupolar couplings. In H/D exchanged Arabidopsis cell walls, 2D 13C-2H correlation spectra (Figure 2.5c) have shown a mixed quadrupolar pattern that can be best deconvoluted into two components: the quadrupolar coupling constant is 50 kHz for the mobile matrix polysaccharides and 187 kHz for the rigid cellulose, which is a value 56 approaching the hydrogen-bonded rigid deuteroxyl quadrupolar coupling (Figure 2.5d). This robust method can be applied to evaluate the dynamics and water-accessible surface of carbohydrates in various organisms. 2.5.4 PRE: Carbohydrate-Ligand Binding in Cellular Environment In structural biology, carbohydrates are often treated as small ligands attached to large protein complexes, but this concept has been inverted in cell wall studies. In plant cell walls, functional proteins are usually present at low concentrations, but with the capabilities of perturbing polymer nexuses or chemically modifying/digesting certain structural motifs. For the β-expansins that cannot be produced recombinantly, the extracted proteins from grass pollens are tagged with paramagnetic Mn(II) labels via their solvent-accessible Cys residues (Figure 2.5e), and mixed with the 13C-labeled cell walls in maize35. Upon binding to expansin, the hemicellulose glucuronoarabinoxylan (GAX) has shown strong 1H and 13C Paramagnetic Relaxation Enhancement (PRE) effects (Figure 2.5f, g), and its stiff and mobile fractions have become more rigid and dynamic, respectively. Therefore, β-expansins have released the connections between the highly substituted GAX (mobile) that forms the interfibrillar matrix and the rarely branched GAX (rigid) that are packed with cellulose microfibrils. The optimized protocols for incorporating paramagnetic sites and the PRE-enabled distance determination71-74 have made it feasible for revealing the interactions between carbohydrates and many proteins or enzymes that contain carbohydrate-binding modules75. 2.5.5 Natural-Abundance DNP and Database: Accommodate the Growing Field As an emerging technique, natural-abundance MAS-DNP has enabled the measurement of 2D 13C−13C/15N/1H spectra on unlabeled biomolecules (Figure 2.6a). When applied to organic molecules and small peptides, this technique could substantially facilitate NMR crystallography 57 by enabling the determination of 13N-13C distance up to 7 Å and the measurement of 15N-13C correlation at natural isotope abundance76-78. Applications of this method to complicated biosystems allow us to extract long-range distance constraints in polyglutamine (polyQ) amyloid fibrils and nano-assemblies of cyclic peptides77, 79, validate imino acid-aromatic interactions in native collagens80, and identify the compositional and conformational differences of cellulose, hemicellulose, and lignin in various plant species (cotton, rice, and poplar)9, 40, 81. These studies were conducted on medium magnetic fields, the 400 MHz/263 GHz or 600 MHz/395 GHz DNP instruments; due to the limited resolution, only highly ordered systems, such as cellulose microfibrils and amyloid fibrils, or a selected component within whole-cell sample could be studied. Because high-field DNP is still inefficient at this stage but has become a necessity for providing sufficient resolution for studying complex samples, the efforts trying to improve the polarization mechanism and radicals at high fields could substantially strengthen the capability of natural-abundance DNP82-84. A rate-limiting process associated with natural-abundance DNP is to interpret the large number of NMR-observables into structural information. We have recently demonstrated that a heatmap comparing the chemical shifts measured on the cotton cellulose and reported in literature allows us to quickly identify the relevant structures (Figure 2.6b)40. This application benefits from the implementation of Complex Carbohydrates Magnetic Resonance Database (CCMRD) that supports the storage and sharing of information on chemical shifts, dynamics, and structure. CCMRD is freely available to the public at www.ccmrd.org and supports data deposition and data search by NMR chemical shifts, carbohydrate name, and compound class (Figure 2.6c, d)85. By the time of this article, 450 compounds from plants, fungi, bacteria, algae, and engineered biomaterials are indexed by CCMRD, and this platform will accommodate the rapid expansion of 58 the dataset and facilitate the development of statistics-based software40. My vision for carbohydrate ssNMR is to enable high-throughput and semi-automatic analysis of spectra and structure, which requires dedicated efforts in method and software development. Figure 2.6 Natural-abundance DNP of unlabeled material assisted by database development. a, Natural abundance 2D 13C-13C INADEQUATE spectrum of unlabeled cotton. A and A’: glucose units in Iα cellulose allomorph; B and B’: glucose units in Iβ allomorph. b, 13C chemical shift RMSD map for comparisons between cotton and other cellulose sources. The color scale of RMSD (ppm) is shown. c, Search interface of CCMRD database that supports data search by compound name, class, and signal. d, Flowchart of data deposition and the 25 types of entries included for each compound. 2.6 Conclusion Solid-state NMR and DNP have demonstrated their unique capability in understanding the nanoscale assembly of fungal and plant cell walls. The rapid advances in NMR instrumentation and technology have made it possible to address biochemical and structural questions that were previously impossible to answer. The studies of plant and fungal cell walls, combined with the 59 many investigations of other complex biosystems, such as the bacterial cell walls and biofilm, algal polysaccharides, and mammalian carbohydrates, have formed an emerging and unique research direction, which is of high significance to the development of biorenewable energy, biomedical therapies, and high-value products based on carbohydrate polymers. 2.7 Acknowledgments This work was supported by the National Science Foundation (CAREER award MCB- 1942665) and the Center for Lignocellulose Structure and Formation, an Energy Frontier Research Center funded by the US Department of Energy, Office of Science, Basic Energy Sciences under award number DE-SC0001090. 60 REFERENCES 1. Somerville, C.; Youngs, H.; Taylor, C.; Davis, S. C.; Long, S. P., Feedstocks for Lignocellulosic Biofuels. Science 2010, 329 (5993), 790-792. 2. Odds, F. C.; Brown, A. J. P.; Gow, N. A. R., Antifungal agents: mechanisms of action. Trends Microbiol. 2003, 11 (6), 272-279. 3. Romaniuk, J. A.; Cegelski, L., Bacterial cell wall composition and the influence of antibiotics by cell-wall and whole-cell NMR. Phil. Trans. R. Soc. B 2015, 370 (1679). 4. Kirui, A.; Dickwella Widanage, M. C.; Mentink-Vigier, F.; Wang, P.; Kang, X.; Wang, T., Preparation of Fungal and Plant Materials for Structural Elucidation Using Dynamic Nuclear Polarization Solid-State NMR. J. Vis. Exp. 2019, (144), e59152. 5. Dick-Perez, M.; Zhang, Y. A.; Hayes, J.; Salazar, A.; Zabotina, O. A.; Hong, M., Structure and interactions of plant cell wall polysaccharides by two- and three-dimensional magic-angle- spinning solid-state NMR. Biochemistry-Us 2011, 50 (6), 989-1000. 6. Wang, T.; Zabotina, O.; Hong, M., Pectin-cellulose interactions in the Arabidopsis primary cell wall from two-dimensional magic-angle-spinning solid-state nuclear magnetic resonance. Biochemistry-Us 2012, 51 (49), 9846-9856. 7. Wang, T.; Salazar, A.; Zabotina, O. A.; Hong, M., Structure and dynamics of Brachypodium primary cell wall polysaccharides from two-dimensional 13C solid-state nuclear magnetic resonance spectroscopy. Biochemistry-Us 2014, 53 (17), 2840-2854. 8. Kang, X.; Kirui, A.; Dickwella Widanage, M. C.; Mentink-Vigier, F.; Cosgrove, D. J.; Wang, T., Lignin-polysaccharide interactions in plant secondary cell walls revealed by solid-state NMR. Nat. Commun. 2019, 10 (1), 347. 9. Perras, F. A.; Luo, H.; Zhang, X.; Mosier, N. S.; Pruski, M.; Abu-Omar, M. M., Atomic- Level Structure Characterization of Biomass Pre- and Post-Lignin Treatment by Dynamic Nuclear Polarization-Enhanced Solid-State NMR. J. Phys. Chem. A 2017, 121 (3), 623-630. 10. Simmons, T. J.; Mortimer, J. C.; Bernardinelli, O. D.; Poppler, A. C.; Brown, S. P.; deAzevedo, E. R.; Dupree, R.; Dupree, P., Folding of xylan onto cellulose fibrils in plant cell walls revealed by solid-state NMR. Nat. Commun. 2016, 7, 13902. 11. Grantham, N. J.; Wurman-Rodrich, J.; Terrett, O. M.; Lyczakowski, J. J.; Stott, K.; Iuga, D.; Simmons, T. J.; Durand-Tardif, M.; Brown, S. P.; Dupree, R.; Busse-Wicher, M.; Dupree, P., An even pattern of xylan substitution is critical for interaction with cellulose in plant cell walls. Nat. Plants 2017, 3 (11), 859-865. 12. Dupree, R.; Simmons, T. J.; Mortimer, J. C.; Patel, D.; Iuga, D.; Brown, S. P.; Dupree, P., Probing the Molecular Architecture of Arabidopsis thaliana Secondary Cell Walls Using Two- and Three-Dimensional 13C Solid State Nuclear Magnetic Resonance Spectroscopy. Biochemistry-Us 2015, 54 (14), 2335-2345. 61 13. Terrett, O. M.; Lyczakowski, J. J.; Yu, L.; Iuga, D.; Franks, W. T.; Brown, S. P.; Dupree, R.; Dupree, P., Molecular architecture of softwood revealed by solid-state NMR. Nat. Commun. 2019, 10 (1), 4978. 14. Viger-Gravel, J.; Lan, W.; Pinon, A. C.; Berruyer, P.; Emsley, L.; Barder, M.; Luterbacher, J., Topology of Pretreated Wood Fibers Using Dynamic Nuclear Polarization. J. Phys. Chem. C 2019, in press. 15. Kang, X.; Kirui, A.; Muszynski, A.; Widanage, M. C. D.; Chen, A.; Azadi, P.; Wang, P.; Mentink-Vigier, F.; Wang, T., Molecular architecture of fungal cell walls revealed by solid-state NMR. Nat. Commun. 2018, 9, 2747. 16. Chatterjee, S.; Prados-Rosales, R.; Frases, S.; Itin, B.; Casadevall, A.; Stark, R. E., Using Solid-State NMR To Monitor the Molecular Consequences of Cryptococcus neoformans Melanization with Different Catecholamine Precursors. Biochemistry-Us 2012, 51 (31), 6080- 6088. 17. Chatterjee, S.; Prados-Rosales, R.; Itin, B.; Casadevall, A.; Stark, R. E., Solid-state NMR Reveals the Carbon-based Molecular Architecture of Cryptococcus neoformans Fungal Eumelanins in the Cell Wall. J. Biol. Chem. 2015, 290 (22), 13779-13790. 18. Nygaard, R.; Romaniuk, J. A. H.; Rice, D. M.; Cegelski, L., Spectral snapshots of bacterial cell-wall composition and the influence of antibiotics by whole-cell NMR. Biophys. J. 2015, 108 (6), 1380-1389. 19. Takahashi, H.; Ayala, I.; Bardet, M.; De Paëpe, G.; Simorre, J. P.; Hediger, S., Solid-state NMR on bacterial cells: selective cell wall signal enhancement and resolution improvement using dynamic nuclear polarization. J. Am. Chem. Soc. 2013, 135 (13), 5105-10. 20. Kern, T.; Giffard, M.; Hediger, S.; Amoroso, A.; Giustini, C.; Bui, N. K.; Joris, B.; Bougault, C.; Vollmer, W.; Simorre, J. P., Dynamics characterization of fully hydrated bacterial cell walls by solid-state NMR: evidence for cooperative binding of metal ions. J. Am. Chem. Soc. 2010, 132 (31), 10911-9. 21. Thongsomboon, W.; Serra, D. O.; Possling, A.; Hadjineophytou, C.; Hengge, R.; Cegelski, L., Phosphoethanolamine cellulose: A naturally produced chemically modified cellulose. Science 2018, 359 (6373), 334-338. 22. Arnold, A. A.; Bourgouin, J. P.; Genard, B.; Warschawski, D. E.; Tremblay, R.; Marcotte, I., Whole cell solid-state NMR study of Chlamydomonas reinhardtii microalgae. J. Biomol. NMR 2018, 70 (2), 123-131. 23. Wang, T.; Hong, M., Solid-state NMR investigations of cellulose structure and interactions with matrix polysaccharides in plant primary cell walls. J. Exp. Bot. 2016, 67, 503-514. 24. Wang, T.; Phyo, P.; Hong, M., Multidimensional solid-state NMR spectroscopy of plant cell walls. Solid State Nucl. Magn. Reson. 2016, 78, 56-63. 62 25. Carpita, N. C., Structure and biogenesis of the cell walls of grasses. Annu. Rev. Plant Phys. 1996, 47, 445-476. 26. Carpita, N. C.; Gibeaut, D. M., Structural models of primary cell walls in flowering plants: consistency of molecular structure with the physical properties of the walls during growth. The Plant journal : for cell and molecular biology 1993, 3 (1), 1-30. 27. Wang, T.; Williams, J. K.; Schmidt-Rohr, K.; Hong, M., Relaxation-compensated difference spin diffusion NMR for detecting 13C-13C in proteins and polysaccharides. J. Biomol. NMR 2015, 61 (2), 97-107. long-range correlations 28. Wang, T.; Park, Y. B.; Cosgrove, D. J.; Hong, M., Cellulose-Pectin Spatial Contacts Are Inherent to Never-Dried Arabidopsis thaliana Primary Cell Walls: Evidence from Solid-State NMR. Plant Physiol. 2015, 168 (3), 871-884. 29. White, P. B.; Wang, T.; Park, Y. B.; Cosgrove, D. J.; Hong, M., Water-polysaccharide interactions in the primary cell wall of Arabidopsis thaliana from polarization transfer solid-state NMR. J. Am. Chem. Soc. 2014, 136 (29), 10399-10409. 30. Dick-Perez, M.; Wang, T.; Salazar, A.; Zabotina, O. A.; Hong, M., Multidimensional solid- state NMR studies of the structure and dynamics of pectic polysaccharides in uniformly 13C- labeled Arabidopsis primary cell walls. Magn. Reson. Chem. 2012, 50 (8), 539-550. 31. Phyo, P.; Wang, T.; Kiemle, S. N.; O'Neill, H.; Pingali, S. V.; Hong, M.; Cosgrove, D. J., Gradients in Wall Mechanics and Polysaccharides along Growing Inflorescence Stems. Plant Physiol. 2017, 175 (4), 1593-1607. 32. Phyo, P.; Wang, T.; Xiao, C. W.; Anderson, C. T.; Hong, M., Effects of Pectin Molecular Weight Changes on the Structure, Dynamics, and Polysaccharide Interactions of Primary Cell Walls of Arabidopsis thaliana: Insights from Solid-State NMR. Biomacromolecules 2017, 18 (9), 2937-2950. 33. Phyo, P.; Gu, Y.; Hong, M., Impact of acidic pH on plant cell wall polysaccharide structure and dynamics: insights into the mechanism of acid growth in plants from solid-state NMR. Cellulose 2019, 26 (1), 291-304. 34. Wang, T.; Park, Y. B.; Caporini, M. A.; Rosay, M.; Zhong, L. H.; Cosgrove, D. J.; Hong, M., Sensitivity-enhanced solid-state NMR detection of expansin's target in plant cell walls. Proc. Natl. Acad. Sci. U.S.A. 2013, 110 (41), 16444-16449. 35. Wang, T.; Chen, Y. N.; Tabuchi, A.; Cosgrove, D. J.; Hong, M., The Target of beta-Expansin EXPB1 in Maize Cell Walls from Binding and Solid-State NMR Studies. Plant Physiol. 2016, 172 (4), 2107-2119. 36. Cosgrove, D. J., Loosening of plant cell walls by expansins. Nature 2000, 407 (6802), 321- 326. 37. Wang, T.; Yang, H.; Kubicki, J. D.; Hong, M., Cellulose Structural Polymorphism in Plant 63 Primary Cell Walls Investigated by High-Field 2D Solid-State NMR Spectroscopy and Density Functional Theory Calculations. Biomacromolecules 2016, 17 (6), 2210-2222. 38. Phyo, P.; Wang, T.; Yang, Y.; O'Neill, H.; Hong, M., Direct Determination of Hydroxymethyl Conformations of Plant Cell Wall Cellulose Using 1H Polarization Transfer Solid- State NMR. Biomacromolecules 2018, 19 (5), 1485-1497. 39. Yang, H.; Wang, T.; Oehme, D.; Petridis, L.; Hong, M.; Kubicki, J. D., Structural factors affecting 13C NMR chemical shifts of cellulose: a computational study. Cellulose 2018, 25 (1), 23- 36. 40. Kirui, A.; Ling, Z.; Kang, X.; Dickwella Widanage, M. C.; Mentink-Vigier, F.; French, A. D.; Wang, T., Atomic Resolution of Cotton Cellulose Structure Enabled by Dynamic Nuclear Polarization Solid-State NMR. Cellulose 2019, 26, 329-339. 41. Cosgrove, D. J.; Jarvis, M. C., Comparative structure and biomechanics of plant primary and secondary cell walls. Front. Plant. Sci. 2012, 3. 42. De Paepe, G.; Lewandowski, J. R.; Loquet, A.; Bockmann, A.; Griffin, R. G., Proton assisted recoupling and protein structure determination. The Journal of chemical physics 2008, 129 (24). 43. Latge, J. P., The cell wall: a carbohydrate armour for the fungal cell. Mol. Microbiol. 2007, 66 (2), 279-290. 44. Stahl, P. D.; Ezekowitz, R. A. B., The mannose receptor is a pattern recognition receptor involved in host defense. Curr. Opin. Immunol. 1998, 10 (1), 50-55. 45. Andreas, L. B.; Jaudzems, K.; Stanek, J.; Lalli, D.; Bertarello, A.; Le Marchand, T.; Cala- De Paepe, D.; Kotelovica, S.; Akopjana, I.; Knott, B.; Wegner, S.; Engelke, F.; Lesage, A.; Emsley, L.; Tars, K.; Herrmann, T.; Pintacuda, G., Structure of fully protonated proteins by proton-detected magic-angle spinning NMR. Proc. Natl. Acad. Sci. U. S. A. 2016, 113 (33), 9187- 92. 46. Bertini, I.; Emsley, L.; Lelli, M.; Luchinat, C.; Mao, J. F.; Pintacuda, G., Ultrafast MAS Solid-State NMR Permits Extensive 13C and 1H Detection in Paramagnetic Metalloproteins. J. Am. Chem. Soc. 2010, 132 (16), 5558-5559. 47. Zhou, D. H.; Shah, G.; Cormos, M.; Mullen, C.; Sandoz, D.; Rienstra, C. M., Proton-detected solid-state NMR Spectroscopy of fully protonated proteins at 40 kHz magic-angle spinning. J. Am. Chem. Soc. 2007, 129 (38), 11791-11801. 48. Struppe, J.; Quinn, C. M.; Lu, M. M.; Wang, M. Z.; Hou, G. J.; Lu, X. Y.; Kraus, J.; Andreas, L. B.; Stanek, J.; Lalli, D.; Lesage, A.; Pintacuda, G.; Maas, W.; Gronenborn, A. M.; Polenova, T., Expanding the horizons for structural analysis of fully protonated protein assemblies by NMR spectroscopy at MAS frequencies above 100 kHz. Solid State Nucl. Magn. Reson. 2017, 87, 117-125. 49. Elena, B.; Lesage, A.; Steuernagel, S.; Bockmann, A.; Emsley, L., Proton to carbon-13 64 INEPT in solid-state NMR spectroscopy. J. Am. Chem. Soc. 2005, 127 (49), 17296-17302. 50. Agarwal, V.; Reif, B., Residual methyl protonation in perdeuterated proteins for multi- dimensional correlation experiments in MAS solid-state NMR spectroscopy. J Magn Reson 2008, 194 (1), 16-24. 51. Baldus, M.; Meier, B. H., Total correlation spectroscopy in the solid state. The use of scalar couplings to determine the through-bond connectivity. J. Magn. Reson. A 1996, 121 (1), 65-69. 52. Phyo, P.; Hong, M., Fast MAS 1H-13C correlation NMR for structural investigations of plant cell walls. J. Biomol. NMR 2019, 73 (12), 661-674. 53. Bougault, C.; Ayala, I.; Vollmer, W.; Simorre, J. P.; Schanda, P., Studying intact bacterial peptidoglycan by proton-detected NMR spectroscopy at 100 kHz MAS frequency. J. Struct. Biol. 2019, 206 (1), 66-72. 54. Schanda, P.; Triboulet, S.; Laguri, C.; Bougault, C. M.; Ayala, I.; Callon, M.; Arthur, M.; Simorre, J. P., Atomic Model of a Cell-Wall Cross-Linking Enzyme in Complex with an Intact Bacterial Peptidoglycan. J. Am. Chem. Soc. 2014, 136 (51), 17852-17860. 55. Roos, M.; Mandala, V. S.; Hong, M., Determination of Long-Range Distances by Fast Magic- Angle-Spinning Radiofrequency-Driven F-19-F-19 Dipolar Recoupling NMR. The journal of physical chemistry. B 2018, 122 (40), 9302-9313. 56. Shcherbakov, A. A.; Mandala, V. S.; Hong, M., High-Sensitivity Detection of Nanometer 1H- 19F Distances for Protein Structure Determination by H-1-Detected Fast MAS NMR. J. Phys. Chem. B. 2019, 123 (20), 4387-4391. 57. Wang, M.; Lu, M.; Fritz, M. P.; Quinn, C. M.; Byeon, I. L.; Byeon, C. H.; Struppe, J.; Maas, W.; Gronenborn, A. M.; Polenova, T., Fast Magic-Angle Spinning 19F NMR Spectroscopy of HIV-1 Capsid Protein Assemblies. Angew. Chem. Int. Ed. Engl. 2018, 57 (50), 16375-16379. 58. Fry, R. A.; Tsomaia, N.; Pantano, C. G.; Mueller, K. T., 19F MAS NMR quantification of accessible hydroxyl sites on fiberglass surfaces. J. Am. Chem. Soc. 2003, 125 (9), 2378-2379. 59. Meier, B. H., Polarization Transfer and Spin Diffusion in Solid-State NMR. Adv. Magn. Opt. Reson. , 1994. 60. Gan, Z. H.; Hung, I.; Wang, X. L.; Paulino, J.; Wu, G.; Litvak, I. M.; Gor'kov, P. L.; Brey, W. W.; Lendi, P.; Schiano, J. L.; Bird, M. D.; Dixon, L. R.; Toth, J.; Boebinger, G. S.; Cross, T. A., NMR spectroscopy up to 35.2 T using a series-connected hybrid magnet. J Magn Reson 2017, 284, 125-136. 61. Keeler, E. G.; Michaelis, V. K.; Colvin, M. T.; Hung, I.; Gor'kov, P. L.; Cross, T. A.; Gan, Z. H.; Griffin, R. G., 17O MAS NMR Correlation Spectroscopy at High Magnetic Fields. J. Am. Chem. Soc. 2017, 139 (49), 17953-17963. 62. Gullion, T., Measurement of Dipolar Interactions between Spin-1/2 and Quadrupolar Nuclei 65 by Rotational-Echo, Adiabatic-Passage, Double-Resonance Nmr. Chem Phys Lett 1995, 246 (3), 325-330. 63. Sefzik, T. H.; Houseknecht, J. B.; Clark, T. M.; Prasad, S.; Lowary, T. L.; Gan, Z.; Grandinetti, P. J., Solid-state 17O NMR in carbohydrates. Chem. Phys. Lett. 2007, 434 (4), 312- 315. 64. Jiao, G. L.; Yu, G. L.; Zhang, J. Z.; Ewart, H. S., Chemical Structures and Bioactivities of Sulfated Polysaccharides from Marine Algae. Mar. Drugs 2011, 9 (2), 196-223. 65. Zhang, W. J.; Falcon, B.; Murzin, A. G.; Fan, J.; Crowther, R. A.; Goedert, M.; Scheres, S. H. W., Heparin-induced tau filaments are polymorphic and differ from those in Alzheimer's and Pick's diseases. Elife 2019, 8. 66. Wang, T.; Hong, M., Structure and Dynamics of Polysaccharidesin Plant Cell Walls from Solid-State NMR. Royal Society of Chemistry: 2017. 67. Yao, X. L.; Schmidt-Rohr, K.; Hong, M., Medium- and long-distance 1H-13C heteronuclear correlation NMR in solids. J Magn Reson 2001, 149 (1), 139-143. 68. Jain, S.; Bjerring, M.; Nielsen, N. C., Efficient and Robust Heteronuclear Cross-Polarization for High-Speed-Spinning Biological Solid-State NMR Spectroscopy. The journal of physical chemistry letters 2012, 3 (6), 703-8. 69. Wei, D.; Akbey, U.; Paaske, B.; Oschkinat, H.; Reif, B.; Bjerring, M.; Nielsen, N. C., Optimal 2H rf Pulses and 2H-13C Cross-Polarization Methods for Solid-State 2H MAS NMR of Perdeuterated Proteins. J Phys Chem Lett 2011, 2 (11), 1289-94. 70. Gelenter, M. D.; Wang, T.; Liao, S. Y.; O'Neill, H.; Hong, M., 2H-13C correlation solid-state NMR for investigating dynamics and water accessibilities of proteins and carbohydrates. J. Biol. NMR 2017, 68 (4), 257-270. 71. Jaroniec, C. P., Structural studies of proteins by paramagnetic solid-state NMR spectroscopy. J Magn Reson 2015, 253, 50-9. 72. Sengupta, I.; Nadaud, P. S.; Helmus, J. J.; Schwieters, C. D.; Jaroniec, C. P., Protein fold determined by paramagnetic magic-angle spinning solid-state NMR spectroscopy. Nat. Chem. 2012, 4 (5), 410-7. 73. Iwahara, J.; Clore, G. M., Detecting transient intermediates in macromolecular binding by paramagnetic NMR. Nature 2006, 440 (7088), 1227-30. 74. Bertini, I.; Luchinat, C.; Parigi, G.; Pierattelli, R., NMR spectroscopy of paramagnetic metalloproteins. Chembiochem 2005, 6 (9), 1536-1549. 75. Boraston, A. B.; Bolam, D. N.; Gilbert, H. J.; Davies, G. J., Carbohydrate-binding modules: fine-tuning polysaccharide recognition. Biochem. J. 2004, 382, 769-781. 66 76. Marker, K.; Pingret, M.; Mouesca, J. M.; Gasparutto, D.; Hediger, S.; De Paepe, G., A New Tool for NMR Crystallography: Complete 13C/15N Assignment of Organic Molecules at Natural Isotopic Abundance Using DNP-Enhanced Solid-State NMR. .J Am. Chem. Soc. 2015, 137 (43), 13796-13799. 77. Marker, K.; Paul, S.; Fernandez-de-Alba, C.; Lee, D.; Mouesca, J. M.; Hediger, S.; De Paepe, G., Welcoming natural isotopic abundance in solid-state NMR: probing pi-stacking and supramolecular structure of organic nanoassemblies using DNP. Chem. Sci. 2017, 8 (2), 974-987. 78. Olsen, R. A.; Struppe, J.; Elliott, D. W.; Thomas, R. J.; Mueller, L. J., Through-bond 13C- 13C correlation at the natural abundance level: Refining dynamic regions in the crystal structure of vitamin-D-3 with solid-state NMR. J. Am. Chem. Soc. 2003, 125 (39), 11784-11785. 79. Smith, A. N.; Marker, K.; Piretra, T.; Boatz, J. C.; Matlahov, I.; Kodali, R.; Hediger, S.; van der Wel, P. C. A.; De Paepe, G., Structural Fingerprinting of Protein Aggregates by Dynamic Nuclear Polarization-Enhanced Solid-State NMR at Natural Isotopic Abundance. J. Am. Chem. Soc. 2018, 140 (44), 14576-14580. 80. Singh, C.; Rai, R. K.; Aussenac, F.; Sinha, N., Direct Evidence of Imino Acid-Aromatic Interactions in Native Collagen Protein by DNP-Enhanced Solid-State NMR Spectroscopy. The journal of physical chemistry letters 2014, 5 (22), 4044-4048. 81. Zhang, L.; Gao, C.; Mentink-Vigier, F.; Tang, L.; Zhang, D.; Wang, S.; Cao, S.; Xu, Z.; Liu, X.; Wang, T.; Zhou, Y.; Zhang, B., Arabinosyl Deacetylase Modulates the Arabinoxylan Acetylation Profile and Secondary Wall Formation. Plant Cell 2019. 82. Mathies, G.; Caporini, M. A.; Michaelis, V. K.; Liu, Y. P.; Hu, K. N.; Mance, D.; Zweier, J. L.; Rosay, M.; Baldus, M.; Griffin, R. G., Efficient Dynamic Nuclear Polarization at 800 MHz/527 GHz with Trityl-Nitroxide Biradicals. Angew. Chem. Int. Edit. 2015, 54 (40), 11770- 11774. 83. Jaudzems, K.; Bertarello, A.; Chaudhari, S. R.; Pica, A.; Cala-De Paepe, D.; Barbet-Massin, E.; Pell, A. J.; Akopjana, I.; Kotelovica, S.; Gajan, D.; Ouari, O.; Tars, K.; Pintacuda, G.; Lesage, A., Dynamic Nuclear Polarization-Enhanced Biomolecular NMR Spectroscopy at High Magnetic Field with Fast Magic-Angle Spinning. Angew. Chem. Int. Edit. 2018, 57 (25), 7458- 7462. 84. Mentink-Vigier, F.; Marin-Montesinos, I.; Jagtap, A. P.; Halbritter, T.; van Tol, J.; Hediger, S.; Lee, D.; Sigurdsson, S. T.; De Paepe, G., Computationally Assisted Design of Polarizing Agents for Dynamic Nuclear Polarization Enhanced NMR: The AsymPol Family. J. Am. Chem. Soc. 2018, 140 (35), 11013-11019. 85. Kang, X.; Zhao, W.; Dickwella Widanage, M. C.; Kirui, A.; Ozdenvar, U.; Wang, T., CCMRD: A Solid-State NMR Database for Complex Carbohydrates. J. Biomol. NMR 2019, accepted. 67 CHAPTER 3: CURRENT LIMITATIONS OF SOLID-STATE NMR IN CARBOHYDRATE AND CELL WALL RESEARCH Review paper reprinted with permission: J. Magn. Reson. 341, 107263 (2022) Authors: Wancheng Zhao, Fabien Deligey, S. Chandra Shekar, Frederic Mentink-Vigier, and Tuo Wang 3.1 Abstract High-resolution investigation of cell wall materials has emerged as an important application of biomolecular solid-state NMR (ssNMR). Multidimensional correlation experiments have become a standard method for obtaining sufficient spectral resolution to determine the polymorphic structure of carbohydrates and address biochemical questions regarding the supramolecular organization of cell walls. Using plant cellulose and matrix polysaccharides as examples, we will review how the multifaceted complexity of polysaccharide structure is impeding the resonance assignment process and assess the available biochemical and spectroscopic approaches that could circumvent this barrier. We will emphasize the ineffectiveness of the current methods in reconciling the ever-growing dataset and deriving structural information. We will evaluate the protocols for achieving efficient and homogeneous hyperpolarization across the cell wall material using magic-angle spinning dynamic nuclear polarization (MAS-DNP). Critical questions regarding the line-broadening effects of cell wall molecules at cryogenic temperature and by paramagnetic biradicals will be considered. Finally, the MAS-DNP method will be placed into a broader context with other structural characterization techniques, such as cryo-electron microscopy, to advance ssNMR research in carbohydrate and cell wall biomaterials. 68 3.2 Introduction Carbohydrates and glycoconjugates regulate cellular recognition and communication, serve as energy sources, and provide structural support to the cells of many organisms1. These biologically important functions are often driven by the structural characteristics of carbohydrate polymers, such as glycosidic linkages and conformations. However, complex carbohydrates are difficult to characterize in their native physical state and have remained under-investigated, unlike proteins and nucleic acids. Solid-state NMR (ssNMR) has a long history (approximately 50 years) of being employed to characterize cellulose-based materials, such as fibers, derivatives, and lignocellulosic biomass2-4. Early studies typically relied on the extensive use of 1D 13C spectra to resolve different allomorphs, quantify their contents, and determine the crystallinity of cellulosic materials. Another important application, as summarized in a recent review5, is to understand the supramolecular organization of cell walls and biofilms in many organisms across biological kingdoms. In these studies, multidimensional (2D and 3D) correlation approaches are often required to achieve sufficient resolution for investigating cellular samples5-8. Because ssNMR can address questions of polymer structure and packing on the angstrom-to-nanometer length scale, the results complement the information provided by many diffraction and imaging methods that focus on the nanoscale and mesoscale. The high-resolution studies of cell wall materials were built on a toolbox that combines ssNMR methods commonly used in both structural biology and polymer research9. Unlike for purified biomolecules, the complexity and heterogeneity of native biomaterials prevent the construction of atomistic models or of an ensemble of interchanging structures10. As shown by multiple studies of fungal and plant cell walls11-13, only the model-based physical and structural principles that regulate the assembly and function of biomolecules can be obtained, similar to the 69 outcomes expected for polymer ssNMR. However, the ever-growing dataset is evolving striking resemblance to that used in NMR structural biology. For example, a recent study has compared the lignin-carbohydrate packing interfaces in grasses, hardwoods, and softwoods, proposing structural models based on more than a thousand site-specific data of polymer interactions and dynamics11. The goal of this review is to highlight the underexplored directions where method development is urgently needed. We will critically review the intrinsic structural complexity underlying the difficulty in analyzing carbohydrate polymers and recognize the immediate need for calculation methods for converting under-used structural constraints into atomistic models. We will then selectively address questions regarding the applications of magic-angle spinning dynamic nuclear polarization (MAS-DNP) in cell wall research. Sections are also devoted to discussing the potential opportunities opening up for each of the current limitations. 3.3 Overview of Experimental Approach Typically, uniformly 13C/15N-labeled cell wall materials or whole cells are measured using 2D and 3D correlation experiments to examine biopolymer structure and cell wall assembly (Figure 3.1)5. The most useful experiments for resonance assignment are J- or dipolar-based refocused INADEQUATE experiments, which unambiguously display the carbon connectivity in each type of sugar units. Cellular carbohydrates usually come with structural polymorphism to a high extent; a phenomenon that can be closely traced via peak multiplicity in 2D/3D 13C-13C/15N spectra. From the peak intensities, which are represented as the peak volumes or areas, the composition of the cell wall molecules can be quantified. It is always beneficial to cross-compare these values with other analytical methods such as chromatography and mass spectrometry (MS). To examine the sub-nanometer physical packing, ssNMR offers a large collection of long-range 70 correlation experiments, including but not limited to PDSD, CORD, CHHC, and PAR14-16. Molecular dynamics are examined using relaxation, dipolar couplings, and exchange NMR, depending on the timescale of the dynamics of interest. Spectral-editing techniques are effective for selecting molecules with unique structural motifs (for example, aromatics and nitrogen sites), rigidity or mobility, and water association, against the bulk of the cell wall17-19. Depending on the growth condition, plants can be grown in solution media containing 13C- glucose or chambers containing 13CO2 20. The former is usually limited to young seedlings grown in dark to avoid isotope dilution by 12C, while the latter has a broader range of applications regarding the plant species and developmental stages. Well-established protocols are also available for labeling other organisms such as fungi, bacteria, and algae. Sample handling should be carefully designed to minimize structural perturbation, notably by avoiding chemical treatment and harsh physical processing techniques such as ball-milling. Typically, a razor blade is used to cut the tissues into millimetric pieces to allow for an even distribution of weight in the rotor during magic-angle spinning (MAS). The native hydration status is another crucial consideration in biomolecular ssNMR. Because the hydroxyl groups of carbohydrates efficiently associate with water molecules, it is not surprising that highly similar spectral features were found in both never- dried cell walls and in samples submitted a dehydration-rehydration cycle. This has been observed in the primary cell walls of Arabidopsis, the secondary cell walls of spruce and eucalyptus, and microalgae named Parachlorella11, 21, 22. However, moderate irreversible alternations induced by dehydration have been detected in sorghum and pine. The changes mainly happen to the hemicellulose (xylan and mannan) located on the packing interface with other polymers (lignin and cellulose)23, 24. Native and fresh cells or tissues are recommended whenever available. 71 Figure 3.1 Schematic representation of ssNMR studies of cell wall materials. Plants are taken as an example, and only cell wall polysaccharides and lignin are shown here for illustration. Some panels of the figure are adapted from reference25. 3.4 Resonance Assignment Impeded by Structural Complexity Like protein structural determination by ssNMR, analysis of a 13C-labeled cell wall sample begins with resonance assignment, which also turned out to be the most challenging step. The cell wall is typically the most rigid component and can be selected against intracellular molecules using experimental schemes based on cross-polarization (CP). However, the spectrum is still overcrowded due to the coexistence of a large variety of carbohydrates in the cell wall. The structural polymorphism of carbohydrate polymer further leads to peak multiplicity. Over the past decade, we have progressively uncovered the link between the structural and spectroscopic characteristics, which are summarized in Figure 3.2 and detailed below. First, non-cellulosic polysaccharides have highly diverse patterns of covalent linkages. When a linkage occurs to a carbon site, its 13C isotropic chemical shift becomes larger by several ppm (also called downfield shift). In solids, this principle was demonstrated by systematically 72 tracking the signals of matrix polysaccharides in the model grass Brachypodium distachyon26. Figure 3.2a shows an example of arabinose, which is a monosaccharide unit commonly found in hemicellulose and pectin. Arabinose can accommodate linkages at carbons 1, 2, 3, and 5. The 1,5- linked arabinose (also named 5-Ara) is featured with a uniquely large C5 chemical shift of 68-70 ppm, instead of the regular 62-63 ppm value for the terminal arabinose residues (t-Ara). Similarly, an additional linkage at carbon 2 will change the isotropic 13C chemical shifts of C2 from 80-82 ppm to 88-90 ppm. Figure 3.2 Structural complexity and peak multiplicity of plant carbohydrates. a, Linkage diversity demonstrated on the arabinose (Ara) unit. The C5 and C2 regions of Ara are shown using a 2D refocused J-INADEQUATE spectrum measured on Brachypodium. The key linkage sites are labeled in the structure, with the corresponding NMR signals color-coded in the spectra. b, Simplified examples of arabinoxylan and glucuronoxylan containing arabinose and glucuronic acid (GlcA) sidechains. Note that the two-fold and three-fold helical screw symmetries of the xylan backbone have been observed for both arabinoxylan and glucuronoxylan. The C4 region of xyloses 73 Figure 3.2 (cont’d) differentiates signals from two-fold and three-fold xylan in spruce. c, The interior (yellow) and surface (cyan) glucan chains of cellulose have distinct hydroxymethyl conformation and NMR signals (from eucalyptus). The cross-section view of a hypothetical 18-chain model for an elementary cellulose fibril is shown. d, Scheme of a large bundle formed by three elementary cellulose microfibrils. The high-order assembly forms the basis for many different structural environments for the glucose residues. e, Acetylation and methyl-esterification change the 13C chemical shifts of the carbonyl group (measured on Arabidopsis). The structures of three galacturonic acid residues, a component commonly found in pectin, are shown to match the NMR peaks. These panels are reconstituted from three previous publications11, 26, 27. Third, the next factor is the broad distribution of conformations in native polysaccharides. The hemicellulose xylan is typically found in two-fold (flat-ribbon) or three-fold (non-flat) helical screw conformations11, 28-30, which can be resolved using the signals of C4 and C1, the two important carbon sites for the glycosidic bond along the xylan backbone (Figure 3.2b). Moreover, a continuous band of xylose C4 signals has also been observed in dried stems of Arabidopsis and the hydrated stems of hardwoods (eucalyptus and poplar)11, 29, as a result of the coexistence of many intermediate conformations between the two-fold and three-fold conformers. For cellulose, the surface and interior glucan chains have distinct hydroxymethyl conformations and resolvable signals (Figure 3.2c). The dominant conformation is gauche-trans (gt) for surface chains and trans-gauche (tg) for interior chains, as recently determined using 1H-1H distance measurement31. Fourth, the high-order assembly of macromolecules also contributes to peak multiplicity. Since 2016, we have been consistently observing 7 major types of glucose units (namely types a- g) for the cellulose in the native cell walls of Arabidopsis, Brachypodium, maize, switchgrass, rice, 74 spruce, eucalyptus, and poplar (see the representative spectrum in Figure 3.2c). In the cross- section of a cellulose microfibril, the outer layer contains two types of surface chains (types f and g) with differentiable levels of water association, which has been examined using the water-edited 1H polarization transfer method31. As a result, connections have been made between these two types of glucan chains with the concepts of hydrophobic and hydrophilic surfaces widely used in cellulose research, but further investigations are needed to assess this hypothesis. Types a and b are directly underneath the surface layer and exhibit strong cross peaks with surface chains32, 33. The origin of type-c has remained vague for long. Type-c chains failed to show strong cross peaks with surface chains; therefore, it is supposed to be deeply embedded in the core so that it is separated from the surface chains by an intermediate layer (types a and b). However, the prevalent 18-chain model of a cellulose microfibril (as shown in Figure 3.2c) could not accommodate such structural complexity. The presence of type-c chains can be better justified by the existence of larger bundles formed by multiple microfibrils as shown in Figure 3.2d. The other two forms (d and e) have noticeably weaker signals. In addition, there are ongoing efforts using density functional theory calculations to connect structural variables, such as hydrogen bonds and other torsional angles, to the observed 13C chemical shifts of cellulose34, 35. Finally, chemical modifications, primarily acetylation and methyl esterification, are carried out by corresponding enzymes. The well-resolved carbonyl peaks (Figure 3.2e) and methyl signals can be used for distinguishing these functional groups and identifying their occurrence sites in cell wall polysaccharides27, 29, 36, 37. The connection between the NMR fingerprint and the structural complexity is exemplified using plant carbohydrates but the five general principles are appliable to polysaccharides in other organisms. 75 3.5 Biochemical and SsNMR Strategies to Facilitate Resonance Assignment Nowadays, resonance assignment is no longer a barrier for well-studied materials (e.g., bacteria and plants), but additional caution should be taken when dealing with uncharted biosystems. In this scenario, the best strategy is to combine ssNMR with chemical and genomics approaches. Chemical assays, such as compositional and linkage analyses, should be conducted to guide and validate the NMR assignment. When ambiguity arises, the ssNMR assignment can be validated by either generating mutant lines or using chemical extracts that selectively deplete the carbohydrate component of interest, thus rendering the corresponding signals absent in the spectra. An alternative is to compare with other species that are known to lack such molecules. When Arabidopsis thaliana was first studied using 2D and 3D ssNMR, a xyloglucan-deficient mutant and a pectin-depleted cell wall were compared with the wild-type plant38, 39. A similar strategy was applied to the model fungus Aspergillus fumigatus, for which a collection of mutants depleting chitin, α-1,3-glucan, galactomannan, and galactosaminogalactan were generated13. A carbohydrate platform, complex carbohydrate magnetic resonance database (CCMRD), was created in 2019 to index the results from publications containing high-resolution ssNMR spectra and well-validated assignments40. Started from the 450 entries available in 201940, the number of entries has increased to 720 as of Spring 2022. The database was designed to accommodate the prolific growth of datasets in upcoming years, in preparation for implementing automated approaches to encourage more ssNMR studies of carbohydrate systems41, 42. Sequential assignment experiments (e.g., NCACX and NCOCX) have been extensively employed in protein ssNMR, but there are no equivalent methods for carbohydrates. This is partially due to the high degree of polymerization and the lack of nitrogen sites in most carbohydrates. However, novel carbohydrate sequencing strategies, analogous to those used in 76 solution NMR44, might be promising in solids for small polysaccharides with highly variable monosaccharide structures or with nitrogenated sugar units. This possibility needs to be explored. Figure 3.3 3D correlation experiments for resonance assignment and structural determination. a, Representative F1-F3 plane of a 3D refocused-INADEQUATE-CORD spectrum of spruce (with CORD recoupling turned off). b, 2D F2-F3 plane of spruce extracted from a 3D spectrum (with 53 ms CORD) at F1 = 138 ppm. c, Long-range inter-chain interactions in spruce cellulose highlighted using dash line circles. The F2-F3 planes (F1 = 138 ppm) of 3D spectra measured with short (53 ms) and long (300 ms) CORD mixing periods were compared to uncover low-intensity but highly relevant cross peaks reporting interactions. (Adapted from Shekar et al.)43 3D 13C correlation experiments provide additional spectral resolution over the 2D spectra and have been applied to biomass characterization29, 38. Recently, a 3D 13C DQ-SQ-SQ correlation experiment has been successfully employed to resolve the polysaccharide signals in the never- dried stems of spruce43. This experiment is free of the cube’s body-diagonal, and the involvement of the DQ chemical shifts in the F1 dimension allows simultaneous observation of two parallel lines of signals in the F2-F3 planes for each coupled spin pair, e.g., s5-s6 and i5-i6 shown in Figure 3.3a, b. The experiment is also useful for identifying long-range correlation cross peaks when a 77 long-mixing 13C-13C mixing was implemented (Figure 3.3c). Novel 3D experiments, such as those involving underexplored nuclei (e.g. 1H and 17O for carbohydrates)45-47, should be developed to counterbalance the lack of nitrogen sites in regard to proteins. 3.6 Underutilized Structural Data for Determining the Packing Interface The ssNMR technique also provides information on the spatial proximities of biopolymers, which is its most appreciated capability, besides quantification, for investigating cell wall materials. A recent study of the plant secondary cell walls has proposed three comparative models for the secondary cell wall organization and the lignin-carbohydrate interface in the grass, hardwood, and softwood (Figure 3.4)11. The function of hemicellulose xylan is found to be conformation- dependent: coating the even surface of cellulose via its flat-ribbon structure and primarily binding lignin domains via its non-flat conformation using preferential surface contacts stabilized by numerous electrostatic interactions. However, this selectivity is partially compromised in the stems of woody plants, with some of the lignin particles closely packed with the flat-ribbon part of xylan as well as the surface of cellulose fibrils, which is likely forced by molecular crowding in the densely packed stems of trees. Finally, the softwood spruce has the best molecular mixing of biopolymers on the nanoscale, with lignin adapting a dispersive distribution instead of forming nanoparticles. Although over-simplified, these cartoon illustrations were built on the structural concepts derived from more than 1,172 site-specific ssNMR data11, 30. The constantly improving resolution allows us to differentiate more carbon sites, and naturally a larger number of long-range cross peaks between these carbons. The complete dataset includes 508 intermolecular cross peaks (272 cross peaks in woody plants and 234 cross peaks in model plants), 475 13C-T1 and 1H-T1ρ relaxation time constants, and 189 water-edited intensities for various carbon sites. Nowadays, each pair of 78 2D 13C-13C correlation spectra (with long and short mixing times as contrast) allow the identification of 90-100 intermolecular cross peaks in a plant sample (Figure 3.5a). Figure 3.4 Illustrative cartoons of biopolymer packing in different plant species restrained by ssNMR. The scheme summarizes the structural concepts summarized from 1,172 site-specific data regarding lignin (yellow), cellulose (white), xylan in two-fold (red, flat-ribbon) and three-fold (blue, twisted bands) screw conformations, and mannan (GGM, green). The structures of polymers are shown for seven regions. Figure adapted from reference11. As shown in polymer contact maps (Figure 3.5b), intermolecular cross peaks are categorized by their intensities following different chemical or structural motifs, such as the acetyl (Ac) and furanose ring of xylan (Xn) in two-fold or three-fold (2f/3f) screw conformations, the interior (i) and surface (s) chains in cellulose, as well as the ring carbons of guaiacyl (G) and syringyl (S) monolignol units, and the methoxyl (OMe) groups of lignin. In this way it is straightforward to recognize the structure selectivity of molecules in forming physical contacts: lignin primarily interacts with xylan instead of the other carbohydrate component (i.e., cellulose), and the S-residue contributes more to lignin-carbohydrate packing in poplar when compared with the G-residue. In particular, numerous strong cross peaks were identified in spruce, revealing the 79 unique architecture of softwood and the homogeneous mixing of biopolymers happening on the nanoscale11. This systemic treatment of data allowed us to efficiently summarize the principal interactions stabilizing the polymer interface, but we still lack a method of converting the overwhelmingly abundant spectroscopic data into structurally meaningful physical models. Figure 3.5 Structural restraints on polymer packing. a, Overlay of 2D dipolar-gated PDSD spectra with short (0.1 s; orange) and long (1.0 s; cyan) mixing times resolving 98 intermolecular cross peaks in eucalyptus. b, Polymer contact map summarizing 272 intermolecular cross peaks arising among different structural units of xylan (cyan), lignin (lime/yellow), cellulose (red), mannan (light orange), and mixed sugars (purple). All cross peaks are categorized as strong (blue), medium (magenta), and weak (black dash lines) following their intensities. Pictures of the samples used for NMR characterization are shown. Figures adapted from reference11. Unfortunately, the analyses of water-association and motional dynamics are even cruder. We rely on these data to distinguish the rigid and mobile domains of polymers, and to differentiate the hydrophilic matrix from the hydrophobic centers of the cell walls. Two polymers tightly packed to form larger aggregates are expected to show reduced mobility and water permeability, in addition to strong intermolecular cross peaks. These features were spotted for the chitin-α-glucan cores in the fungal cell wall of Aspergillus fumigatus12, 13 and the cellulose-two-fold-xylan core in 80 plant secondary cell walls11, 30. In contrast, molecules well dispersed in the matrix should be better solvated and more dynamic, expecting fewer and weaker cross peaks with other components. This trend has been observed in the β-glucan matrix of fungi and the hemicellulose-pectin matrix of plants. A complication arises as a single polysaccharide may have multiple domains with distinct dynamics. For instance, the pectin in Arabidopsis has a mobile domain forming the matrix that is prone to alkali extraction as well as an alkali-resistant and stiff domain that is packed with cellulose microfibrils21. Data analysis in this polymer/residue-specific manner has made it straightforward to cross- inspect different samples (e.g., from different species or mutant lines), but it has compromised the advantage of high-resolution site-specific data. We are facing the challenge of converting structural restraints into an energy-minimized ensemble of macromolecular structures. Inputs from experts in modeling and structural calculation will be highly valued. 3.7 Critical Factors of MAS-DNP on Cell Wall Samples Nowadays, MAS-DNP is regularly used for characterizing cell walls by providing the needed sensitivity for visualizing the interface between lowly-populated polymers (compared to the bulk of the cell wall) and by enabling high-resolution characterization of unlabeled materials using 2D 13C correlation experiments at natural isotopic abundance (1.1%)48-52. These applications have been reviewed recently53. Here we will mainly address two frequently asked questions regarding 1) how to ensure homogeneous hyperpolarization throughout the biomaterial being studied, and 2) how to mitigate the line-broadening effect due to the low temperature of the sample (e.g., ~100 K for MAS-DNP) and the paramagnetic effect from the radicals. It is expected that samples with homogenous hyperpolarization should show comparable patterns in 1D 13C spectra measured with and without microwave (MW) irradiation (Figure 3.6a). 81 Occasionally, lipid polymers and membranes may be poorly polarized. Lipid polymers mostly form separated domains and their signals are distinguishable from carbohydrate peaks, thus alleviating the interference. Cell walls are porous layers on the surface of the cell that can be freely accessed and easily penetrated by the biradicals dissolved in solution54, which are small molecules only 2-3 nm across55, 56. The polarization of electron spins will be first transferred to the 1H in the biradical57-59, and then to those in the solvent, and finally spread over a very long range via 1H-1H spin diffusion, which ensures homogeneous hyperpolarization, for example, across at least tens of nanometers as estimated using wood fibers60. This is the case for carbohydrates as the nuclear relaxation times are long enough at low temperature, unlike the lipids that still relax quickly. Cellulose can largely retain its linewidth at the cryogenic temperature of DNP due to its high crystallinity (Figure 3.6b, c)50, 61. Indeed, we have observed impressively narrow full width at half maximum (FWHM) linewidths of 0.9-1.0 ppm for cotton cellulose on a 600 MHz/395 GHz MAS-DNP, which allowed us to track the glucose units in the Iα and Iβ model allomorphs frequently observed in model cellulosic materials49. The C-H dipolar order parameters are typically larger than 0.85 for cellulose; therefore, motion is not a major factor to consider for these fibrils. In contrast, hemicellulose and pectin are highly dynamic, and their signals are broadened out at a low temperature as shown in Figure 3.6b. The resulting structure contains conformers with different isotropic chemical shifts, which inhomogeneously broaden the NMR lines. Despite this limitation, some peaks of these mobile components have been recently resolved in 2D 13C-13C correlation MAS-DNP spectra50. Under most circumstances, biradicals mainly stay in the solvent instead of being packed with biomolecules. Therefore, the paramagnetic relaxation enhancement effect from the biradicals is neither pronounced nor an issue for biomaterial DNP (Figure 3.6c), especially for the solutions with a low concentration (5-10 mM) of radicals62. 82 Figure 3.6 Evaluation of sample and biradical conditions for MAS-DNP. a, Uniform hyperpolarization across the cell wall confirmed by the consistent patterns of carbohydrate signals in the microwave-on (MW on) and microwave-off (MW off) spectra when normalized by the interior cellulose carbon 4 peaks (i4). Lipid components are not well polarized. b, Time saving and improved signal-to-noise ratios of DNP over room-temperature NMR (RT-NMR). Blue dash lines indicate the signals of mobile components of rice stems broadened by the cryogenic temperature of DNP. c, 13C spectra of extracted Arabidopsis cell walls at different temperatures and biradical conditions. d, Room‑temperature EPR spectra of AMUPol at 9.6 GHz in four rice stem samples (enhancement factor εon/off of 38-57 on 600 MHz/395 GHz MAS-DNP) prepared using different solvents and procedures. The figures are remade from references50, 61. 3.8 Expanding the Application of MAS-DNP in Carbohydrate Research For most cell wall molecules in rice stems, it has been shown that the linewidths in 2D 13C correlation spectra collected on a 600 MHz/395 GHz MAS-DNP were comparable to those collected on a 400 MHz conventional ssNMR50. This is observed not only for cellulose but also 83 for some rigid portions of matrix polymers, such as the backbone of xylan. However, the most dynamic components, such as the arabinose sidechains of xylan, still evades DNP detection. The current resolution allowed for resonance assignment and compositional analysis of unlabeled cell wall materials. In the same study, the MAS-DNP analysis was further combined with room- temperature measurements (by conventional ssNMR) of relaxation and dipolar order parameters, providing a more complete assessment of these unlabeled samples50. However, an unaccomplished task is to develop methods that can efficiently detect intermolecular cross peaks under natural isotopic abundance63. Further questions include whether we can better handle matrix polysaccharides and other mobile components such as intracellular molecules. Finding answers to these questions might introduce new applications of MAS-DNP to biomaterial characterization. It is also unclear though if MAS-DNP at higher fields would help resolve the resolution issues. If the line-broadening is inhomogeneous, higher fields will not help, otherwise higher magnetic fields are desirable. However, in most cases, MAS-DNP at 800 MHz/527 GHz or higher generates lower enhancement in general. Ongoing progress notably in biradical design might improve MAS-DNP at high field56, 64, 65. It should be noted that smaller rotors (e.g., 1.3 mm)66, 67 are needed to improve the enhancement, but the use of a small rotor sacrifices the absolute sensitivity during 13C detection as compared to the conventional 3.2 mm rotors. Overall, high-field DNP bears some promises but still presents major challenges and may not surpass the sensitivity provided by the lower-field DNP setups used today. Unexpectedly, optimizing DNP samples, especially for the mixing protocol of biradical (in the solvent) and the cell wall material, is still fully empirical at this stage. Typically, the most limiting step of a MAS-DNP project is finding a sample preparation maximizing the sensitivity by 84 both a great enhancement factor and a relatively short DNP buildup time. This often requires re- processing the samples with solvents and radicals for multiple cycles, by trial and error. Recently, a benchtop EPR was used to examine a collection of rice samples before subjection to MAS-DNP measurements (Figure 3.6d). The EPR spectra suggests that the biradicals are favorably partitioning into the solvent in the samples with good sensitivity enhancement (εon/off of 38-57). It will be beneficial if a protocol could be developed to rapidly screen the condition of radicals in the samples and pick up the most promising ones. This thematic is presently becoming even more relevant with the introduction of new radicals to the commercially available library. The next question is how to place MAS-DNP into the context of other structural and biophysical techniques. Cryo-EM has established a leading role in protein structure determination and viral research but is not frequently used for carbohydrates or biomaterials. We have developed a strategy for analyzing the structures of unlabeled biomaterials by combining the nanoscale resolution of cryo-electron tomography (CET) with the atomic-level information provided by MAS-DNP (Figure 3.7)68. CET of a cellulose fibril (cross-sectional FWHM width of 5.2-5.7 nm) revealed periodic structure along the fibril axis formed by two wrapped, yet not twisted filaments, each of which generally agrees with an 18-chain model for elementary cellulose microfibril. MAS- DNP resolved the carbon sites of interior and surface glucan chains, and intensity quantification supports the 18-chain arrangement with a relatively low extent of fibrillar bundling. The sample used was cellulose fibrils synthesized in vitro by protein machinery isolated from moss, which provided a simplified target for this successful attempt. It remains to be seen if DNP has synergism with cryo-EM for investigating more sophisticated polysaccharides and biopolymer complexes such as cell walls. 85 Figure 3.7 Synergism of Cryo-EM and DNP for carbohydrate characterization. Subtomogram averages of in vitro fibers and isosurface rendering of the density map (left) reveal a fibril structure formed by two wrapped filaments. Each filament fibril can be fit to an 18-chain cellulose model, where the conformational distribution of glucan chains and the surface-to-interior ratio are analyzed by MAS-DNP (right). A 2D 13C−13C CHHC spectrum is included here to show the spectral resolution. Adapted from Deligey et al.68 3.9 Conclusion Over the last decade, significant advances have been made in the structural characterization of cell walls and carbohydrate-based biomaterials. However, due to the high heterogeneity of these cellular samples and the unique structural complexity of carbohydrates, the tools currently available for polymer research and protein structural biology are not fully applicable here. Development of semi-automatic tools for facilitating resonance assignment, novel pulse sequences and structural calculation methods for carbohydrate structural determination, as well as new radicals and optimized protocols for the application of high-field MAS-DNP may lead to breakthroughs in carbohydrate and biomaterial research. 86 3.10 Acknowledgments This work was supported by the U.S. Department of Energy (grant no. DE-SC0021210). F.D. was supported as part of the Center for Lignocellulose Structure and Formation, an Energy Frontier Research Center funded by the US Department of Energy, Office of Science, Basic Energy Sciences under award number DE-SC0001090. The National High Magnetic Field Laboratory is supported by National Science Foundation through NSF/DMR-1644779 and the State of Florida. 87 REFERENCES 1. Duus, J.; Gotfredsen, C. H.; Bock, K., Carbohydrate Structural Determination by NMR Spectroscopy:  Modern Methods and Limitations. Chem. Rev. 2000, 100, 4589-4614. 2. Atalla, R. H.; Vanderhart, D. L., Native cellulose: a composite of two distinct crystalline forms. Science 1984, 223 (4633), 283-285. 3. El hariri El Nokab, M.; Habibi, M. H.; Alassmy, Y. A.; Abduljawad, M. M.; Alshamrani, K. M.; Sebakhy, K. O., Solid State NMR a Powerful Technique Investigating Sustainable/Renewable Cellulose-Based Materials. Polymers 2022, 14, 1049. for 4. Foston, M., Advances in solid-state NMR of cellulose. Curr. Opin. Biotechnol. 2014, 27, 176- 184. 5. Ghassemi, N.; Poulhazan, A.; Deligey, F.; Mentink-Vigier, F.; Marcotte, I.; Wang, T., Solid- State NMR Investigations of Extracellular Matrices and Cell Walls of Algae, Bacteria, Fungi, and Plants. Chem. Rev. 2021, in press. 6. Reichhardt, C.; Cegelski, L., Solid-state NMR for bacterial biofilms. Mol. Phys. 2014, 112 (7), 887-894. 7. Kelly, J. E.; Chrissian, C.; Stark, R. E., Tailoring NMR experiments for structural characterization of amorphous biological solids: A practical guide. Solid State Nucl. Magn. Reson. 2020, 109, 101686. 8. Arnold, A. A.; Bourgouin, J. P.; Genard, B.; Warschawski, D. E.; Tremblay, R.; Marcotte, I., Whole cell solid-state NMR study of Chlamydomonas reinhardtii microalgae. J. Biomol. NMR 2018, 70 (2), 123-131. 9. Reif, B.; Ashbrook, S. E.; Emsley, L.; Hong, M., Solid-state NMR spectroscopy. Nat. Rev. Methods Primers 2021, 1 (1), 2. 10. Markwick, P. R. L.; Malliavin, T.; Nilges, M., Structural Biology by NMR: Structure, Dynamics, and Interactions. PLoS Comput. Biol. 2008, 4, e1000168. 11. Kirui, A.; Zhao, W.; Deligey, F.; Yang, H.; Kang, X.; Mentink-Vigier, F.; Wang, T., Carbohydrate-aromatic interface and molecular architecture of lignocellulose. Nat. Commun. 2022, 13, 538. 12. Kang, X.; Kirui, A.; Muszynski, A.; Widanage, M. C. D.; Chen, A.; Azadi, P.; Wang, P.; Mentink-Vigier, F.; Wang, T., Molecular architecture of fungal cell walls revealed by solid-state NMR. Nat. Commun. 2018, 9, 2747. 13. Chakraborty, A.; Fernando, L. D.; Fang, W.; Dickwella Widanage, M. C.; Wei, P.; Jin, C.; Fontaine, T.; Latgé, J. P.; Wang, T., A molecular vision of fungal cell wall organization by functional genomics and solid-state NMR. Nat. Commun. 2021, 12, 6346. 88 14. Hou, G.; Yan, S.; Trébosc, J.; Amoureux, J.-P.; Polenova, T., Broadband homonuclear correlation spectroscopy driven by combined R2nv sequences under fast magic angle spinning for NMR structural analysis of organic and biological solids. J. Magn. Reson. 2013, 232, 18-30. 15. Paëpe, G. D.; Lewandowski, J. R.; Loquet, A.; Böckmann, A.; Griffin, R. G., Proton assisted recoupling and protein structure determination. J. Chem. Phys. 2008, 129 (24), 245101. 16. Lange, A.; Luca, S.; Baldus, M., Structural Constraints from Proton-Mediated Rare-Spin Correlation Spectroscopy in Rotating Solids. J. Am. Chem. Soc. 2002, 124 (33), 9704-9705. 17. Ader, C.; Schneider, R.; Seidel, K.; Etzkorn, M.; Becker, S.; Baldus, M., Structural Rearrangements of Membrane Proteins Probed by Water-Edited Solid-State NMR Spectroscopy. J. Am. Chem. Soc. 2009, 131 (1), 170-176. 18. White, P. B.; Wang, T.; Park, Y. B.; Cosgrove, D. J.; Hong, M., Water–Polysaccharide Interactions in the Primary Cell Wall of Arabidopsis thaliana from Polarization Transfer Solid- State NMR. J. Am. Chem. Soc. 2014, 136 (29), 10399-10409. 19. Williams, J. K.; Schmidt-Rohr, K.; Hong, M., Aromatic spectral editing techniques for magic- angle-spinning solid-state NMR spectroscopy of uniformly 13C-labeled proteins. Solid State Nucl. Magn. Reson. 2015, 72, 118-126. 20. Gao, Y.; Mortimer, J. C., Unlocking the architecture of native plant cell walls via solid-state nuclear magnetic resonance. Methods Cell Biol. 2020, 160, 121-143. 21. Wang, T.; Park, Y. B.; Cosgrove, D. J.; Hong, M., Cellulose-pectin spatial contacts are inherent to never-dried Arabidopsis primary cell walls: evidence from solid-state nuclear magnetic resonance. Plant Physiol. 2015, 168 (3), 871-884. 22. Poulhazan, A.; Dickwella Widanage, M. C.; Muszynski, A.; Arnold, A. A.; Warschawski, D. E.; Azadi, P.; Marcotte, I.; Wang, T., Identification and Quantification of Glycans in Whole Cells: Architecture of Microalgal Polysaccharides Described by Solid-State Nuclear Magnetic Resonance. J. Am. Chem. Soc. 2021, 46, 19374-19388. 23. Gao, Y.; Lipton, A. S.; Wittmer, Y.; Murray, D. T.; Mortimer, J. C., A grass-specific cellulose–xylan interaction dominates in sorghum secondary cell walls. Nat. Commun. 2020, 11 (1), 6081. 24. Cresswell, R.; Dupree, R.; Brown, S. P.; Pereira, C. S.; Skaf, M. S.; Sorieul, M.; Dupree, P.; Hill, S., Importance of water in maintaining softwood secondary cell wall nanostructure. Biomacromolecules 2021, 8, 4669-4680. 25. Zhao, W.; Fernando, L. D.; Kirui, A.; Deligey, F.; Wang, T., Solid-state NMR of plant and fungal cell walls: a critical review. Solid State Nucl. Magn. Reson. 2020, 107, 101660. 26. Wang, T.; Salazar, A.; Zabotina, O. A.; Hong, M., Structure and Dynamics of Brachypodium Primary Cell Wall Polysaccharides from Two-Dimensional 13C Solid-State Nuclear Magnetic Resonance Spectroscopy. Biochemistry 2014, 53 (17), 2840-2854. 89 27. Kirui, A.; Du, J.; Zhao, W.; Barnes, W.; Kang, X.; Anderson, C. T.; Xiao, C.; Wang, T., A pectin Methyltransferase modulates polysaccharide dynamics and interactions in Arabidopsis primary cell walls: Evidence from solid-state NMR. Carbohydr. Polym. 2021, 270, 118370. 28. Simmons, T. J.; Mortimer, J. C.; Bernardinelli, O. D.; Pöppler, A.-C.; Brown, S. P.; Deazevedo, E. R.; Dupree, R.; Dupree, P., Folding of xylan onto cellulose fibrils in plant cell walls revealed by solid-state NMR. Nat. Commun. 2016, 7, 13902. 29. Dupree, R.; Simmons, T. J.; Mortimer, J. C.; Patel, D.; Iuga, D.; Brown, S. P.; Dupree, P., Probing the Molecular Architecture of Arabidopsis thaliana Secondary Cell Walls Using Two- and Three-Dimensional 13C Solid State Nuclear Magnetic Resonance Spectroscopy. Biochemistry 2015, 54 (14), 2335-2345. 30. Kang, X.; Kirui, A.; Dickwella Widanage, M. C.; Mentink-Vigier, F.; Cosgrove, D. J.; Wang, T., Lignin-polysaccharide interactions in plant secondary cell walls revealed by solid-state NMR. Nat. Commun. 2019, 10 (1), 1-9. 31. Phyo, P.; Wang, T.; Yang, Y.; O’Neill, H.; Hong, M., Direct determination of hydroxymethyl conformations of plant cell wall cellulose using 1H polarization transfer solid-state NMR. Biomacromolecules 2018, 19 (5), 1485-1497. 32. Wang, T.; Yang, H.; Kubicki, J. D.; Hong, M., Cellulose structural polymorphism in plant primary cell walls investigated by high-field 2D solid-state NMR spectroscopy and density functional theory calculations. Biomacromolecules 2016, 17 (6), 2210-2222. 33. Wang, T.; Hong, M., Solid-state NMR investigations of cellulose structure and interactions with matrix polysaccharides in plant primary cell walls. J. Exp. Bot. 2016, 67 (2), 503-514. 34. Yang, H.; Kubicki, J. D., A density functional theory study on the shape of the primary cellulose microfibril in plants: effects of C6 exocyclic group conformation and H-bonding. Cellulose 2020, 27 (5), 2389-2402. 35. Yang, H.; Wang, T.; Oehme, D.; Petridis, L.; Hong, M.; Kubicki, J. D., Structural factors affecting 13C NMR chemical shifts of cellulose: a computational study. Cellulose 2018, 25 (1), 23-36. 36. Wang, T.; Chen, Y.; Tabuchi, A.; Cosgrove, D. J.; Hong, M., The target of β-expansin EXPB1 in maize cell walls from binding and solid-state NMR studies. Plant Physiol. 2016, 172 (4), 2107- 2119. 37. Terrett, O. M.; Lyczakowski, J. J.; Yu, L.; Iuga, D.; Franks, W. T.; Brown, S. P.; Dupree, R.; Dupree, P., Molecular architecture of softwood revealed by solid-state NMR. Nat. Commun. 2019, 10, 4978. 38. Dick-Pérez, M.; Zhang, Y.; Hayes, J.; Salazar, A.; Zabotina, O. A.; Hong, M., Structure and interactions of plant cell-wall polysaccharides by two-and three-dimensional magic-angle- spinning solid-state NMR. Biochemistry 2011, 50 (6), 989-1000. 90 39. Dick-Perez, M.; Wang, T.; Salazar, A.; Zabotina, O. A.; Hong, M., Multidimensional solid- state NMR studies of the structure and dynamics of pectic polysaccharides in uniformly 13C- labeled Arabidopsis primary cell walls. Magn. Reson. Chem. 2012, 50, 539-550. 40. Kang, X.; Zhao, W.; Widanage, M. C. D.; Kirui, A.; Ozdenvar, U.; Wang, T., CCMRD: a solid-state NMR database for complex carbohydrates. J. Biomol. NMR 2020, 74 (4), 239-245. 41. Li, D.-W.; Hansen, A. L.; Yuan, C.; Bruschweiler-Li, L.; Bruschweiler, R., DEEP picker is a deep neural network for accurate deconvolution of complex two-dimensional NMR spectra. Nat. Commun. 2021, 12, 5229. 42. Cordova, M.; Balodis, M.; Simoes de Almeida, B.; Ceriotti, M.; Emsley, L., Bayesian probabilistic assignment of chemical shifts in organic solids. Sci. Adv. 2021, 7, eabk2341. 43. Shekar, S. C.; Zhao, W.; Fernando, L. D.; Hung, I.; Wang, T., A 13C three-dimensional DQ- SQ-SQ correlation experiment for high-resolution analysis of complex carbohydrates using solid- state NMR. J. Magn. Reson. 2022, 336, 107148. 44. Gray, C. J.; Migas, L. G.; Barran, P. E.; Pagel, K.; Seeberger, P.; Eyers, C. E.; Boons, G.- J.; Pohl, N. L. B.; Compagnon, I.; Widmalm, G.; Flitsch, S. L., Advancing Solutions to the Carbohydrate Sequencing Challenge. J. Am. Chem. Soc. 2019, 141, 14463-14479. 45. Phyo, P.; Hong, M., Fast MAS 1H–13C correlation NMR for structural investigations of plant cell walls. J. Biomol. NMR 2019, 73 (12), 661-674. 46. Bougault, C.; Ayala, I.; Vollmer, W.; Simorre, J. P.; Schanda, P., Studying intact bacterial peptidoglycan by proton-detected NMR spectroscopy at 100 kHz MAS frequency. J. Struct. Biol. 2019, 206 (1), 66-72. 47. Shen, J.; Terskikh, V.; Struppe, J.; Hassan, A.; Monette, M.; Hung, I.; Gan, Z.; Brinkmann, A.; Wu, G., Solid-state 17O NMR study of α-d-glucose: exploring new frontiers in isotopic labeling, sensitivity enhancement, and NMR crystallography. Chem. Sci. 2022, 13, 2591-2603. 48. Takahashi, H.; Lee, D.; Dubois, L.; Bardet, M.; Hediger, S.; De Paëpe, G., Rapid Natural- Abundance 2D 13C–13C Correlation Spectroscopy Using Dynamic Nuclear Polarization Enhanced Solid-State NMR and Matrix-Free Sample Preparation. Angew. Chem. Int. Ed. 2012, 51 (47), 11766-11769. 49. Kirui, A.; Ling, Z.; Kang, X.; Widanage, M. C. D.; Mentink-Vigier, F.; French, A. D.; Wang, T., Atomic resolution of cotton cellulose structure enabled by dynamic nuclear polarization solid-state NMR. Cellulose 2019, 26 (1), 329-339. 50. Zhao, W.; Kirui, A.; Deligey, F.; Mentink-Vigier, F.; Zhou, Y.; Zhang, B.; Wang, T., Solid- state NMR of unlabeled plant cell walls: high-resolution structural analysis without isotopic enrichment. Biotechnol. Biofuels 2021, 14 (1), 14. 51. Berruyer, P.; Gericke, M.; Moutzouri, P.; Jakobi, D.; Bardet, M.; Karlson, L.; Schantz, S.; Heinze, T.; Emsley, L., Advanced characterization of regioselectively substituted methylcellulose 91 model compounds by DNP enhanced solid-state NMR spectroscopy. Carbohydr. Polym. 2021, 262, 117944. 52. Perras, F. A.; Luo, H.; Zhang, X.; Mosier, N. S.; Pruski, M.; Abu-Omar, M. M., Atomic- level structure characterization of biomass pre-and post-lignin treatment by dynamic nuclear polarization-enhanced solid-state NMR. J. Phys. Chem. A 2017, 121 (3), 623-630. 53. Chakraborty, A.; Deligey, F.; Quach, J.; Mentink-Vigier, F.; Wang, P.; Wang, T., Biomolecular complex viewed by dynamic nuclear polarization solid-state NMR spectroscopy. Biochem. Soc. Trans. 2020, 48 (3), 1089-1099. 54. Takahashi, H.; Ayala, I.; Bardet, M.; De Paëpe, G.; Simorre, J. P.; Hediger, S., Solid-state NMR on bacterial cells: selective cell wall signal enhancement and resolution improvement using dynamic nuclear polarization. J. Am. Chem. Soc. 2013, 135 (13), 5105-10. 55. Sauvée, C.; Rosay, M.; Casano, G.; Aussenac, F.; Weber, R. T.; Ouari, O.; Tordo, P., Highly Efficient, Water-Soluble Polarizing Agents for Dynamic Nuclear Polarization at High Frequency. Angew. Chem. Int. Ed. 2013, 52 (41), 10858-10861. 56. Mentink-Vigier, F.; Marin-Montesinos, I.; Jagtap, A. P.; Halbritter, T.; van Tol, J.; Hediger, S.; Lee, D.; Sigurdsson, S. T.; De Paëpe, G., Computationally Assisted Design of Polarizing Agents for Dynamic Nuclear Polarization Enhanced NMR: The AsymPol Family. J. Am. Chem. Soc. 2018, 140 (35), 11013-11019. 57. Mentink-Vigier, F.; Dubroca, T.; Van Tol, J.; Sigurdsson, S. T., The distance between g- tensors of nitroxide biradicals governs MAS-DNP performance: The case of the bTurea family. J. Magn. Reson. 2021, 329, 107026. 58. Mentink-Vigier, F.; Vega, S.; De Paëpe, G., Fast and accurate MAS–DNP simulations of large spin ensembles. Phys. Chem. Chem. Phys. 2017, 19 (5), 3506-3522. 59. Perras, F. A.; Pruski, M., Large-scale ab initio simulations of MAS DNP enhancements using a Monte Carlo optimization strategy. J. Chem. Phys. 2018, 149 (15), 154202. 60. Viger-Gravel, J.; Lan, W.; Pinon, A. C.; Berruyer, P.; Emsley, L.; Bardet, M.; Luterbacher, J., Topology of pretreated wood fibers using dynamic nuclear polarization. J. Phys. Chem. C 2019, 123 (50), 30407-30415. 61. Wang, T.; Park, Y. B.; Caporini, M. A.; Rosay, M.; Zhong, L.; Cosgrove, D. J.; Hong, M., Sensitivity-enhanced solid-state NMR detection of expansin’s target in plant cell walls. Proc. Natl. Acad. Soc. USA 2013, 110 (41), 16444-16449. 62. Takahashi, H.; Hediger, S.; De Paëpe, G., Matrix-free dynamic nuclear polarization enables solid-state NMR 13C–13C correlation spectroscopy of proteins at natural isotopic abundance. Chem. Commun. 2013, 49 (82), 9479-9481. 63. Märker, K.; Paul, S.; Fernández-de-Alba, C.; Lee, D.; Mouesca, J.-M.; Hediger, S.; De Paëpe, G., Welcoming natural isotopic abundance in solid-state NMR: probing π-stacking and 92 supramolecular structure of organic nanoassemblies using DNP. Chem. Sci. 2017, 8 (2), 974-987. 64. Zhai, W.; Lucini Paioni, A.; Cai, X.; Narasimhan, S.; Medeiros-Silva, J.; Zhang, W.; Rockenbauer, A.; Weingarth, M.; Song, Y.; Baldus, M.; Liu, Y., Postmodification via Thiol- Click Chemistry Yields Hydrophilic Trityl-Nitroxide Biradicals for Biomolecular High-Field Dynamic Nuclear Polarization. J. Phys. Chem. B 2020, 124 (41), 9047-9060. 65. Lund, A.; Casano, G.; Menzildjian, G.; Kaushik, M.; Stevanato, G.; Yulikov, M.; Jabbour, R.; Wisser, D.; Renom-Carrasco, M.; Thieuleux, C.; Bernada, F.; Karoui, H.; Siri, D.; Rosay, M.; Sergeyev, I. V.; Gajan, D.; Lelli, M.; Emsley, L.; Ouari, O.; Lesage, A., TinyPols: a family of water-soluble binitroxides tailored for dynamic nuclear polarization enhanced NMR spectroscopy at 18.8 and 21.1 T. Chem. Sci. 2020, 11 (10), 2810-2818. 66. Chaudhari, S. R.; Berruyer, P.; Gajan, D.; Reiter, C.; Engelke, F.; Silverio, D. L.; Copéret, C.; Lelli, M.; Lesage, A.; Emsley, L., Dynamic nuclear polarization at 40 kHz magic angle spinning. Phys. Chem. Chem. Phys. 2016, 18 (15), 10616-10622. 67. Wisser, D.; Karthikeyan, G.; Lund, A.; Casano, G.; Karoui, H.; Yulikov, M.; Menzildjian, G.; Pinon, A. C.; Purea, A.; Engelke, F.; Chaudhari, S. R.; Kubicki, D.; Rossini, A. J.; Moroz, I. B.; Gajan, D.; Copéret, C.; Jeschke, G.; Lelli, M.; Emsley, L.; Lesage, A.; Ouari, O., BDPA- Nitroxide Biradicals Tailored for Efficient Dynamic Nuclear Polarization Enhanced Solid-State NMR at Magnetic Fields up to 21.1 T. J. Am. Chem. Soc. 2018, 140 (41), 13340-13349. 68. Deligey, F.; Frank, M. A.; Cho, S. H.; Kirui, A.; Mentink-Vigier, F.; Swulius, M. T.; Nixon, B. T.; Wang, T., Structure of In Vitro-Synthesized Cellulose Fibrils Viewed by Cryo-Electron Tomography and 13C Natural-Abundance Dynamic Nuclear Polarization Solid-State NMR. Biomacromolecules 2022, in press. 93 CHAPTER 4: POLYSACCHARIDE ASSEMBLIES IN FUNGAL AND PLANT CELL WALLS EXPLORED BY SOLID-STATE NMR Review paper reprinted with permission: Structure 31, 1375-1385 (2023) Authors: Liyanage D. Fernando#, Wancheng Zhao#, Isha Gautam, Ankur Ankur, and Tuo Wang 4.1 Summary #Equally contributions Structural analysis of macromolecular complexes within their natural cellular environment presents a significant challenge. Recent applications of solid-state NMR (ssNMR) techniques on living fungal cells and intact plant tissues have greatly enhanced our understanding of the structure of extracellular matrices. Here we selectively highlight the most recent progress in this field. Specifically, we discuss how ssNMR can provide detailed insights into the chemical composition and conformational structure of pectin, and the consequential impact on polysaccharide interactions and cell wall organization. We elaborate on the use of ssNMR data to uncover the arrangement of the lignin-polysaccharide interface and the macrofibrillar structure in native plant stems or during degradation processes. We also comprehend the dynamic structure of fungal cell walls under various morphotypes and stress conditions. Finally, we assess how the combination of NMR with other techniques can enhance our capacity to address unresolved structural questions concerning these complex macromolecular assemblies. 4.2 Introduction The cell walls of plants and fungi play vital roles in cell shape, mechanics, integrity, adhesion, and extensibility1-4. These macromolecular assemblies are also crucial for energy and carbon storage, as well as antimicrobial resistance, and serve as potential targets for novel antifungal agents5, 6. From a structural perspective, cell walls represent highly sophisticated 94 biomaterials created by nature, inspiring extensive efforts to develop artificial materials that mimic the chemical and physical principles governing macromolecular assembly7, 8. However, the characterization of such nanocomposites in their native state presents significant challenges. Recent advancements in solid-state NMR (ssNMR) techniques have enabled high- resolution investigation of cell wall materials by leveraging methods initially developed for NMR structural biology of proteins, nucleic acids, and polymers9-12. This approach has been adopted due to the sufficient spectroscopic resolution available, which allows for the acquisition of numerous structural constraints to visualize the complex composition of biomolecules within cell walls. Additionally, it addresses the limitations posed by the inherent heterogeneity and complexity of the cell wall by utilizing simplified representations13. Extensive studies have been conducted on numerous plant and fungal species, leading to significant revisions in our understanding of cell wall structure14. In this review, we aim to provide a concise overview of the key principles in ssNMR methodology and the most recent structural findings, with a focus on those published within the past three years. 4.3 Methodology Advances Enabling SsNMR of Cellular Samples One of the major merits of ssNMR lies in its ability to directly characterize living cells or intact tissues15, 16. Since solubilization or extraction is not necessary, the samples fully preserve the inherent physical and chemical characteristics of the biomolecules. The samples analyzed are typically enriched with NMR-active isotopes by incorporating 13C-enriched precursors like 13C- glucose, 13C-maltose, and 13CO2 (for plants), as well as 15N-labeled amino acids and salts17, 18. Occasionally, selective-labeled precursors are used to simplify the spectra and track the biosynthesis of macromolecules19. Deuterated fatty acids can also be incorporated for determining the phospholipid profile in cellular membranes20. The collected material can be directly placed 95 into a magic-angle spinning (MAS) rotor, which can hold approximately between 1 mg (for a 0.7 mm diameter) and 100 mg (for a 4 mm diameter) of material, depending on the requirements of the NMR experiment. A versatile toolbox is available for studying various aspects of biomolecules in cell walls and cellular samples, including molecular composition, structural variations, dynamical distribution, water association, and sub-nanometer packing13, 14, 21. While 1D spectra are useful for quick screening, high-resolution structural characterization often mandates the use of 2D and 3D homonuclear (such as 13C-13C) and heteronuclear (such as 15N/1H-13C) correlation experiments. These techniques rely primarily on 13C detection, which provides excellent spectral dispersion to differentiate many magnetically inequivalent sites22-27. However, recent advancements in 1H detection methods have also yielded valuable insights into the structure and packing of polysaccharides and proteins in plants, fungi, and bacteria28-31. The success of these studies has been facilitated by fast-to-ultrafast magic-angle spinning (MAS) techniques, with MAS rotation frequencies ranging from 60 to 110 kHz, where a higher level of deuteration and the presence of more dynamic molecules allow for reduced reliance on high MAS frequencies. Dynamic Nuclear Polarization (DNP) is another significant advancement that has greatly expanded the capabilities of ssNMR in the structural characterization of macromolecular complexes32-34. In MAS-DNP, polarization of unpaired electrons in stable biradicals are transferred to 1H and subsequently to other nuclei of interest, resulting in enhanced sensitivity35. The prominent biradicals used extensively in structural studies include AMUPol and the recently developed AsympolPOK36, 37. The 600 MHz/395 GHz MAS-DNP instrument strikes a favorable balance between spectral resolution and sensitivity enhancement, and a 30-80-fold boost can be easily achieved for cellular samples. This breakthrough has effectively overcome the sensitivity 96 limitations of NMR, enabling the use of small sample quantities, which is particularly valuable for difficult-to-replicate samples, exploring lowly populated molecules or macromolecular complexes embedded in bulk structures, and even characterizing fully unlabeled biosamples without the need for isotopic enrichment38-43. 4.4 Pectin Methylation Influencing Plant Primary Cell Wall Structure The mechanical roles of cellulose microfibrils, hemicellulose, and pectin in plant primary cell walls have been a topic of ongoing debate for many decades44. One prevailing model suggests the existence of a tethered polymer network, where a single hemicellulose strand (e.g., a xyloglucan) could simultaneously bind to multiple cellulose microfibrils and hold them together to form a load-bearing network45, 46. Pectin is an acidic polymer regulating wall porosity, pH, and ionic balance, and is influencing cell expansion and differentiation47. However, pectin is often considered as a separate component that forms a gel-like matrix, providing reinforcement to the cellulose-hemicellulose network. This perspective was primarily based on biochemical findings, wherein the vast majority of pectin is extractable by strong alkali, and the acidic pectin backbones do not bind to cellulose in vitro48. A significant advancement brought in by ssNMR is the identification of the interaction between pectin and cellulose within intact primary plant cell walls. Pectins are partially interconnected complexes of polysaccharides and proteoglycans47, 49, 50. The major domains include homogalacturonan (HG), rhamnogalacturonan-I (RG-I) with branched arabinan and galactan, as well as RG-II51. By employing ssNMR, even with a simple 1D 13C spectrum, it is possible to discern various specific chemical sites within the galacturonic acid (GalA) residues that comprise HG, the rhamnose (Rha) residues that alternate with GalA along the RG-I backbone, as well as the arabinan sidechains (Figure 4.1a)52. Such spectral resolution forms the basis for tracking the intermolecular cross peaks between major pectin domains and cellulose 97 microfibrils using 2D 13C-13C correlation experiments53, 54. However, RG-II, which is less prevalent and structurally more intricate compared to HG and RG-I, has yet to be investigated using high-resolution ssNMR techniques. Previously, it was believed that pectin backbone and cellulose were physically separated entities, but the numerous pectin-cellulose cross peaks identified by ssNMR have unambiguously revealed their close spatial proximity in the cell wall54, 55. The new insight has contributed to a paradigm shift in our understanding of the architecture of the primary cell wall44, 56, and comprehensive overviews of these ssNMR studies can be found in multiple recent reviews14, 57. Despite these advancements, the precise function of the packing between pectin and cellulose in supporting the structure of the cell wall remains unclear. Recently, mesoscale coarse-grained molecular dynamics (CGMD) simulations conducted on onion epidermal cell walls have shed some light on this matter58. The noncovalent interactions between multiple cellulose fibrils were found to be the primary factor determining the mechanical properties of the cell wall. Notably, the sliding of two aligned cellulose microfibrils against each other was identified as a significant contributor to the plasticity of the cell wall. In contrast, hemicellulose and pectin were suggested to play more indirect roles, potentially by influencing the arrangement of cellulose within the cell wall58. Growing evidence emphasizes the significance of HG methylation in influencing the organization of the cell wall. At first glance, this may appear contradictory, as methylation only involves localized chemical modifications to the carboxyl groups of GalA residues59, 60, while the rearrangement of cell wall polymers happens on the nanoscale. An emerging concept is that methylation alters the charge and physicochemical properties of HG, and may induce repulsion among HG chains, thereby providing the force to expand the interfibrillar space between cellulose 98 microfibrils61. This mechanism has been proposed as a potential regulator of cell expansion and shaping, alongside the more widely accepted turgor-driven mechanism. Dupree, Hong, and coworkers have shown that pectin methylation can change the conformational structure, interactions, and dynamics of polysaccharides in the primary cell wall62. They compared the pectin structure in wild-type Arabidopsis with that of the gosamt1 gosamt2 mutant, which had a reduced degree of methyl esterification in HG. The mutant exhibited an increased abundance of HG in a structure known as the 2-fold screw conformation (21), with its C6 chemical structure as -COO- in this sample (Figure 4.1b). These observations match the characteristics of the egg-box structure that refers to the cross-linking of two anionic GalA units that joined two adjacent HG chains by a Ca2+ ion. The reduced methyl esterification in the mutant promoted the formation of egg-box clusters and enhanced HG aggregation. Conversely, the content of the methylated -COOCH3 motif in the 3-fold screw conformation (31) was reduced in the gosamt1 gosamt2 mutant62. The double mutant also exhibited strong signals indicating a reduction in the chain length of HG, allowing the identification of signals from the reducing ends of the GalA units in both α- and β-configurations. Importantly, this study represents the first successful application of ssNMR to elucidate the detailed structural characteristics of HG in muro, providing direct experimental evidence supporting the widely accepted model of the pectin egg-box structure within the plant cell wall. NMR data showed that HG methylation has an impact on the polysaccharide organization in the cell wall. Pectin backbones showed larger dipolar order parameters and stronger interactions with cellulose in the gosamt1 gosamt2 mutant62. These changes in cell wall structure were found to be associated with impaired polarized growth in the double mutant. Interestingly, these findings appear to contrast with recent observations from other studies using different mutants and 99 experimental conditions. An increase in pectin methylation is often accompanied by stronger pectin-cellulose interactions and impaired plant growth as reported in ssNMR analyses of Arabidopsis primary cell walls and focused on mutants that produced shorter HG chains, cell walls with pH-alteration, or segments along the inflorescences stem63-65. Figure 4.1 Pectin Structure and Its Effect on Primary Plant Cell Wall. a, Representative 1D 13C ssNMR spectra with resolvable signals associated with pectin and cellulose structural motifs. Carbon numbers are shown for key sugar units. b, Selected regions of 2D 13C-13C spectra resolving the conformation and methylation state of GalA units in Arabidopsis pectin. Two-fold (21) and three-fold (31) screw conformations of HG displaying distinct chemical environments at the 100 Figure 4.1 (cont’d) carbonyl site. An egg-box model is shown to depict two cross-linked HG chains. c, Structural rearrangement of the cell wall in Arabidopsis qua2/tsd2 mutant. Figures 1a, c adapted with permission from Kirui et al. Carbohydr. Polym. (2021). Figures 1b adapted with permission from Temple et al. Nat. Plants (2022). The situation becomes more intricate due to another recent study involving two allelic mutants (qua2 and tsd2) of a pectin methyltransferase52. These mutants exhibit similar cell wall composition and pectin methyl esterification levels as the wild-type sample, but, surprisingly, displayed stiffer pectin and stronger interactions with cellulose (Figure 4.1c). Atomic Force Microscopy (AFM) and Field Emission Scanning Electron Microscopy (FESEM) revealed a decrease in cellulose bundling in both mutants66, and the more dispersed microfibril arrangement should have facilitated the extensive interactions as observed by ssNMR analysis. These findings highlight how ssNMR can serve as a valuable tool to bridge the knowledge gaps between the chemical structure of carbohydrates, plant biology, and cell wall organization. However, pectin's role in cell wall structure remains enigmatic. It is now evident that pectin methylation can exert diverse effects on the cell wall structure, and there is no universal principle that applies to all scenarios. It is likely that the structural function of pectin is indirect, influencing the arrangement of cellulose microfibrils and hemicellulose within the cell wall, which requires further investigation. 4.5 Lignin-Carbohydrate Packing in Secondary Plant Cell Walls Within the secondary cell wall, the coalescence of multiple cellulose microfibrils occurs frequently, e.g. an illustration of a 3-microfibril bundle is given in Figure 4.2a, which is accompanied by the association of hemicelluloses (xylan and galactoglucomannan) and lignin to 101 form larger macrofibrils67, 68. Inspired by the seminal work of Dupree and colleagues69, many multidimensional ssNMR studies have been conducted in the past seven years to unravel the molecular architecture of secondary plant cell walls. These studies have led to two significant revisions in our understanding of lignocellulosic biomass. First, the chemical composition and conformational structure of xylan dictate its binding specificity within the cell wall70, 71. Second, the lignin-carbohydrate interface is stabilized by electrostatic interactions, with lignin primarily associated with the non-flat domains of xylan. However, in certain cases, lignin can also interact and co-localize with the xylan-coated cellulose microfibrils as the secondary target72, 73. In dehydrated environments, such as the dried stem of Arabidopsis, xylan exhibits a broad range of helical screw conformations that can be monitored through the continuous band observed in its C4 signals in ssNMR spectra69. In well-hydrated cell walls, including those of Arabidopsis, spruce, maize, switchgrass, rice, and Brachypodium, xylan preferentially adopts either 2-fold or 3- fold helical screw conformations resulting in two distinct C4 signals70, 72-74. This is illustrated by the flat-ribbon and non-flat structures depicted in Figure 4.2a. Notably, a continuous distribution of xylan helical screw conformation was even observed in hydrated hardwood materials, such as eucalyptus and poplar, which is likely induced by factors such as lower hydration levels and/or molecular crowding within hardwood stems73. In the case of sorghum, the secondary cell wall is primarily composed of 3-fold xylan conformations75, while in Brachypodium, both the leaf and root showed two xylan conformers, but the stem shows a continuous range of conformations. The presence of a two-fold flat-ribbon structure is relatively uncommon and was found to be facilitated by xylan’s deposition onto the smooth surface of a cellulose microfibril76. Moreover, it has been discovered that an evenly distributed pattern of substitutions (e.g., by glucuronic acid or arabinose sidechains) along the xylan chain is crucial for maintaining this flat-ribbon structure 102 and promoting its interaction with cellulose71. This also explains why sorghum lacks this conformation as xylan is frequently and irregularly substituted by arabinosyl residues. On the other hand, the three-fold xylan plays a dominant role in carbohydrates’ interactions with lignin, primarily through the polar functional groups, which highlights the significance of electrostatic interactions in stabilizing the interface between lignin nanodomains and carbohydrates (see the grass model in Figure 4.2b)72. The molecular architecture of lignin-carbohydrate interactions varies across different plant species (Figure 4.2b). Comparisons between grasses, hardwoods, and softwoods have revealed a gradual decrease in the domain size and polymer separation between lignin and polysaccharides. Lignin-cellulose packing interactions, which were found to be relatively limited in grass samples, also became abundant in woody plants73. During the analysis of the sparsely populated biopolymer interface, which exhibits structural deviations from the overall equilibrium state of the entire cell wall, ssNMR experiments employing spectral editing (e.g. the selected polarization of lignin aromatics or methoxy group or xylan acetyls) selection and DNP enhancement have proven to be instrumental77, 78. Dupree and coworkers successfully fit the ssNMR data on molecular fraction, polymer conformation, and polymer spatial proximity into a macrofibril observed under Cryo-SEM (Figure 4.2c)74. This large bundle that is tens of nanometers across to accommodate ~50 elementary microfibrils, each of which consists of 18 glucan chains. SsNMR analysis uncovered two distinct conformations for galactoglucomannan (GGM), reminiscent of the functional conformations in xylan. It was found that GGM undergoes conformational changes when bound to cellulose in a semi-crystalline manner. Some of the bound GGMs and two-fold xylan molecules can attach to the surface of the same microfibril, while the other GGMs contribute to the matrix alongside three- 103 fold xylan. The entire carbohydrate core is enveloped by a lignin layer, completing the molecular architecture of the microfibril assembly. Water is another important structural component of the secondary cell wall, and high- temperature (105 °C) oven-drying can induce irreversible alteration to the structure of Monterey pine (Pinus radiata), even after subsequent rehydration79. SsNMR analysis of the rehydrated sample revealed a tighter packing between xylan and cellulose, along with the separation of some mannan into mobile phase, adopting a conformation similar to that observed in solution. These results support the hypothesis that water plays a crucial role in mediating the packing interactions between xylan and cellulose (Figure 4.2d). Importantly, this study provides direct evidence of the irreversible change in biomass ultrastructure caused by harsh drying procedures. It should be noted that another ssNMR study has demonstrated fully reversible changes of spruce and poplar through the process of lyophilization and rehydration73. Li, Kang, Yelle and colleagues have utilized ssNMR to investigate the perturbation of lignin linkages and lignin-polysaccharide packing by termite digestion 80. This was acommplished by feeding 13C-labeled sapwood sections of Canadian poplar (a natural hybrid Populus × canadensis), a hardwood, to a phylogenetically higher termite species (Nasutitermes) and Monterey pine, a softwood, to a lower termite species (Cryptotermes). The abundance of different lignin linkages and lignin-polysaccharide interactions in the termite diet and feces was assessed using 2D 13C-13C radio frequency-driven recoupling (RFDR) and long-range proton-driven spin diffusion (PDSD) spectra. Both termite species were found to effectively dissociate the electrostatic interface between lignin and polysaccharides but only the lower termite species can perturb the structure of the residual lignin. This study highlights the remarkable capability of ssNMR spectroscopy in examining the structural changes occurring in lignocellulosic materials, 104 and the same approach can be applied to investigate biomass degradation by various biological agents such as white/brown-rot fungi, as well as in industrial processes81. Figure 4.2 Macrofibril Structure and Lignin-Carbohydrate Interface. a,Structural features of polysaccharides in plant secondary cell walls resolvable by ssNMR. b, Structural view of the aromatic-carbohydrate packing interface in grass species, hardwood, and softwood. c, Best fit of NMR data collected on spruce secondary cell walls into a macrofibril. A local cellulose microfibril with two domains of glucan chains (light and dark blue) is shown with 2-fold xylan (Xn2f) and galactoglucomannan (GGM) attached. d, Illustration of loosely associated water and tightly bound water in pine secondary cell wall. Figures 2a,b adapted from Kirui et al. Nat. Commun. (2022). Figure 2c modified from Terrett et al. Nat. Commun. (2019). Figure 2d adapted from Cresswell et al. Biomacromolecules (2021). 4.6 Dynamic Structure of Fungal Cell Wall: Challenge and Opportunity Stark and colleagues have been at the forefront of utilizing ssNMR for nearly two decades 105 to study melanin deposition and its interactions with cell wall polysaccharides in Cryptococcus and Saccharomyces species19, 82-85. Recent ssNMR investigations on Aspergillus and Schizophyllum, with a focus on the cell wall structures, were largely inspired by these studies, and were built upon research strategies employed in plant cell wall characterization, albeit with methodological adaptations57, 86. However, an unexpected problem that significantly impeded progress was the dynamic nature of fungal cell walls, characterized by their remarkable ability to adapt to varying growth conditions and environmental contexts2, 87. These microorganisms exhibit an impressive capacity to compensate for the absence of specific polysaccharides through complex biosynthetic reactions, thereby generating a structurally intact cell wall88. Such adaptability, unlike their plant counterparts, challenges the notion that certain cell wall polysaccharides, such as cellulose and pectin in primary cell walls, are essential. The dynamic structural changes in fungal cell walls not only enable the survival and adaptation of these microbes to diverse environments but also present technical barriers to our fundamental understanding of these organelles and the development of antifungal medications to combat invasive infections89. As a result, it is crucial to carefully monitor the status of the samples and ensure the reproducibility of the culture conditions. In many cases, it becomes necessary to prepare fresh batches of samples specifically for lengthy experiments. Complications further arise from the distinct composition and assembly of biomacromolecules in different fungal species and the vast array of mutant strains found in nature. Therefore, it is crucial to first identify the conserved features of fungal cell walls and elucidate the structural functions of their key polysaccharides before investigating the variations across the wide spectrum of fungal species and strains. 4.7 Morphotype-Dependent Structure of Aspergillus Cell Walls During the transition of morphotypes in the fungal life cycle, the cell wall undergoes 106 significant changes in its nanoscale morphology and molecular composition2, 90. These structural transformations play a vital role in many biological processes such as spore germination, hyphal growth, and sporulation. In the case of A. fumigatus, the fungus forms asexual hydrophobic spores called dormant conidia (dorC) on a specialized hyphal structure called conidiophore1. In immunocompromised patients, when the inhaled conidia reach the alveoli in the lungs within 4-6 hours, the germination process takes place. Germination involves the uptake of water by the conidia, resulting in their swelling and transformation into swollen conidia (swoC). Subsequently, the swollen conidia germinate (gerC) into short hyphae known as germ tubes1. Following germination, the growth of hyphae initiates, leading to the formation of a colony. This mycelial colony represents an aggregated form of hyphae that invade the pulmonary tissues, causing a life- threatening disease known as aspergillosis90. SsNMR analysis of the intact Aspergillus mycelium offers valuable insight into the structural organization of the cell wall during the vegetative growth of the fungus in the lungs. These living fungal cells exhibited high spectral resolution, enabling the differentiation of various conformers from six major types of fungal polysaccharides, including chitin, chitosan, α-1,3- glucan, three types of β-glucan, as well as galactomannan (GM), and galactosaminogalactan (GAG) (Figure 4.3a)86, 91. The partially crystalline chitin has a narrow linewidth of 0.5-0.7 ppm, whereas the plant counterpart, cellulose microfibrils, shows a broader linewidth of 0.7-1.0 ppm on high-field magnets92. The mobile molecules present in the fungus and the plant pectin have comparable linewidths of 0.2-0.5 ppm93. Similar to the polymorphic structure observed in plant cellulose 92, chitin also exhibits high polymorphism. The structural arrangement of chitin in fungi is predominantly aligned with the α-allomorph94, characterized by an antiparallel packing of chains95, 96. Through ssNMR analysis, it was discovered that the rigid core of the cell wall 107 comprises α-1,3-glucan, β-glucan, and chitin. Of particular surprise was the identification of α-1,3-glucan's structural function, as it was previously considered insignificant in the cell wall assembly, akin to the underappreciated role of pectin in primary plant cell walls97. Unlike the highly mobile nature of pectin in plants, α-1,3- glucan was found to be the most rigid polysaccharide in Aspergillus fumigatus mycelia. It forms a tightly packed, dehydrated core through extensive physical interactions with chitin. This finding was further strengthened by the observation that α-1,3-glucans are crucial to the stiffness of the cell wall in chitin-deficient mutant (Figure 4.3a). Moreover, α-1,3-glucan has diverse distribution, with its physical presence in both the mobile and rigid domains and chemical existence in both the alkali-soluble and insoluble fractions, as revealed by a combined chemical and NMR approach 91. On the other hand, β-glucans in fungi, similar to xyloglucan in plant primary cell walls, have been proposed to play a crucial role in cross-linking. Chemical analysis has identified the presence of a small fraction of covalently linked mannan-β-1,3-glucan-chitin complex in Aspergillus87. The β-glucans encompass diverse linkages, existing as linear β-1,3-glucan, branched β-1,3/1,6-glucan, and terminal β-1,3/1,4-glucan90, which has not been fully characterized by ssNMR so far. However, structural analysis of β-glucans has shown that these molecules are essential for retaining water in the cell wall and contribute to the formation of a hydrated matrix. Outside of this inner domain, there is a surface shell enriched with highly mobile glycoproteins, GM, and GAG91. Loquet and colleagues reported the evolution of the Aspergillus fumigatus conidial cell wall by examining the structure of dorC, which was the starting material, as well as swoC and gerC, which were prepared by incubating dorC by additional 5 h and 8 h, respectively98. Compared to dorC, the swoC exhibited isotropic growth to double its cellular size, based on which the gerC 108 cells further underwent polarized growth, leading to the formation of germ tubes. The polysaccharides present in the rigid fractions of A. fumigatus conidia cell walls were found to be consistent with those of mycelial cell walls, containing chitin, α-glucan, and β-glucan98. Another study published almost at the same time also reported highly conserved carbohydrate cores in both mycelia and conidia cell walls, but using unlabeled A. fumigatus as enabled by the sensitivity- enhancement of MAS-DNP99. Figure 4.3 NMR-Restrained Structural Models of Fungal Cell Walls. a,Comparative views of mycelial cell walls in the parental strain (left) and chitin-deficient mutant (right) of A. fumigatus. b, Dormant and germinating conidial cell walls of A. fumigatus. c, Mycelial cell walls of S. commune, highlighting the binding polysaccharides (cyan) and proteins that bind Cu(II) ions and antimicrobial peptides. Figure 3a adapted from Chakraborty et al. Nat. Commun. (2021), Figure 3b adapted with permission from Lamon et al. Proc. Natl. Acad. Sci. USA (2023), and Figure 3c adapted with permission from Safeer et al. Eur. J. Chem. (2022) and Ehren et al. Cell Surf. (2020). 109 In the conidial cell wall, the ratio of α-glucan to β-glucan was approximately 1:3 in dorC and gerC, but became almost equal in swoC (Figure 4.3b), suggesting a less stable cell wall composition required for germination98. Both α-glucan and β-glucan exhibited increased hydration levels in swoC and gerC compared to dorC. Interestingly, α-glucan showed high hydration levels in all conidia cell walls98, whereas it was poorly hydrated in mycelia cell walls72, indicating differences in their structural organization. While the chitin content remained the same in the rigid portions of all three cell walls, chitin exhibited a noticeable decrease in both water-contact and structural polymorphism in gerC. This observation can be explained by a model where chitin becomes deeply embedded in the inner core and undergoes phase separation from glucans during the germinating stage (Figure 4.3b)98. In addition, the presence of GAG was detected by ssNMR in both swollen and germinating conidia, but not in dormant conidia, where its biosynthesis has yet to begin. This study demonstrates the capability of cellular ssNMR to effectively uncover the structural reorganization of the fungus during the transition between morphotypes. Triglyceride signals were also identified in the mobile phase of A. fumigatus cells 98. In other fungal species, such as the melanized C. neoformans and S. cerevisiae spore cell wall, triglyceride has been identified by ssNMR as a potential building molecule of stress-resistant cell walls100. The biological relevance of this molecule in A. fumigatus and its location (e.g., as a cell wall component or an intracellular molecule) requires careful evaluation. 4.8 Cell Wall Architecture of Schizophyllum commune Baldus and coworkers conducted two ssNMR investigations of Schizophyllum commune mycelia28,101. The first study involved the analysis of sequentially digested cell walls, progressively removing molecules from the cell wall101. The rigid domain of the cell wall is composed of chitin and β-1,3/1,6-glucan, potentially linked together, as well as α-1,3-glucan and polymeric fucose 110 (Figure 4.3c). The mobile domain contains terminal hexoses included in α- and β-linked glucans. Mannose residues were also identified in the mobile phase, at least partially incorporated in mannosylated proteins. Using proton-detection methods and ultrahigh field magnets (1.2 GHz), the team further examined the binding sites of metals and antimicrobial peptides28. At a low ion concentration of 0.74 mM Cu(II), the paramagnetic relaxation enhancement (PRE) effect revealed signal quenching at the primary targets, which were mainly proteins. However, at a higher ion concentration of 18.5 mM, Cu(II) ions could penetrate into the cell wall region with β-1,3/1,6- glucans. Proton-detection methods were also employed to understand the binding of many other micronutrients, such as Ca2+, Mg2+, and anions, to the S. commune cell wall102,103. Additionally, the antimicrobial peptide Cathelicidin-2, which inhibits S. commune growth, was found to primarily bind to cell wall proteins28. This was demonstrated through chemical shift perturbation (CSP) observed in protein backbone Cα sites and the disappearance of signals from the sidechains of charged amino acids. Interestingly, the signals of GalN/GalNAc were also affected by the peptide28. This observation is intriguing as Cathelicidin-2 is typically positively charged with a theoretical isoelectric point (pI) in the range of 10-12 depending on the origin of the peptide104. At the same time, GalN can exist as cationic residues with an -NH3 + group under the pH of the fungal culture. The unidentified mechanisms underlying the association between the antimicrobial peptide and S. commune cell walls might inspire future studies. 4.9 Integrating NMR Restraints with Other Experimental Approaches Chemical analysis, ssNMR, diffraction methods, imaging techniques can complement each other to complete the understanding of cell wall ultrastructure. Each of these methods rely on distinct physical and chemical principles, offering unique insights into various structural aspects at different length scales. Locally, ssNMR has been combined with chemical assays and sugar 111 analyses to confirm resonance assignment and provide detailed information on the composition and linkage patterns of carbohydrates105. However, the cross-linking between polymers and the extractability as probed by chemical analysis does not directly correlate with the physical packing as investigated by ssNMR97. For example, the majority of pectin and α-1,3-glucan are highly extractable by hot alkali, but they are both tightly associated with the crystalline core of the corresponding cell wall, cellulose in the plant and chitin in the fungus. And it has always been overlooked that a portion of pectin and α-1,3-glucan can never be fully isolated due to their physical entrapment by microfibrillar components55, 91. Similarly, the analysis of lignin also faces challenges in establishing consistency between ssNMR and the more widely used liquid NMR 106, 108. Additionally, when it comes to polymer dynamics, it is not possible to directly combine NMR- reported parameters such as correlation time, relaxation time constants, or dipolar order parameters with values obtained from other techniques like persistence length65. Diffraction methods and imaging techniques are valuable for providing structural information at nanoscale or larger length scales. Consequently, ssNMR data can complement these techniques by offering atomic-to-nanometer structural information109. As we discussed earlier, this strategy has been applied to rationalize the organization of spruce microfibril74. Moreover, a combination of MAS-DNP and Cryo-electron tomography (CryoET) has been employed to examine the structure of in-vitro synthesized cellulose fibrils, revealing that they consist of two wrapped filaments fitting into an 18-chain microfibril110. Cryo-electron microscopy methods have also been employed to examine the fiber arrangement in the cell wall111-113 and the structure of polysaccharide synthases114-116. Another example is the challenge of distinguishing proteins from different sources (e.g., cell walls, membrane proteins, and intracellular sources) using ssNMR alone. In such cases, confocal microscopy and atomic force microscopy (AFM) have been used to 112 investigate the assembly of RodA rodlets in Aspergillus, which were found to interact with melanin on the cell surface of DorC while swoC and gerC did not support rodlet formation98. However, caution is necessary when attempting to bridge the findings obtained from different approaches. 4.10 Conclusions and Perspectives Significant progress has been made in applying ssNMR and other structural methods to unravel the complex structure of polysaccharides and associated biomacromolecules in plant and fungal cell walls. However, there are still numerous unanswered structural questions. For instance, the function of pectin in plant cell wall mechanics remains unknown, the resolution of lignin by ssNMR is insufficient, the role of GGM (galactoglucomannan) is not fully understood, and the identification of lignin-carbohydrate linkages117 in muro remains challenging. Moreover, the fungal cell wall is even less explored, with many structural components, such as exopolysaccharides like glucuronoxylomannan (GXM) and galactoxylomannan (GalXM) in Cryptococcus capsules and β-1,6-glucan and phosphomannan in Candida sp., still requiring investigations. Understanding the mechanisms of cell wall remodeling in response to stress is also crucial. Recent research has unveiled that the halophilic fungus Aspergillus sydowii enhances the hydrophobicity and stiffness of its cell wall to resist osmotic pressure in hypersaline environments118. Similar changes have been observed in A. fumigatus during adaptation to internal stress caused by carbohydrate-deficiency and external stress such as antifungal treatment with caspofungin (unpublished results by Dickwella Widange et al.). It is essential to validate if these cell wall adaptations are universal features across various species during adaptation. Furthermore, high-resolution ssNMR techniques have been employed to address structural challenges in other organisms. An important application is the quantification and compositional analysis of carbohydrate components in algal cells, bacterial cell walls, and biofilms31, 119-122. As ssNMR 113 technology continues to evolve and our understanding of polysaccharide structure and cell wall assembly expands, it is now opportune to tackle important structural problems in the years to come. 4.11 Acknowledgments This work was supported by the National Institutes of Health grant AI173270. The plant cell wall research was supported by the U.S. Department of Energy, Office of Science, Basic Energy Sciences, under award number DE-SC0023702. 114 REFERENCES 1. van de Veerdonk, F. L.; Gresnigt, M. S.; Romani, M. G.; Latge, J. P., Aspergillus fumigatus morphology and dynamic host interactions. Nat. Rev. Microbiol. 2017, 15, 661-674. 2. Gow, N. A. R.; Lenardon, M. D., Architecture of the dynamic fungal cell wall. Nat. Rev. Microbiol. 2022, 21, 248-259. 3. Cosgrove, D. J., Building an extensible cell wall. Plant Physiol. 2022, 189, 1246-1277. 4. Cosgrove, D. J., Growth of the plant cell wall. Nat. Rev. Mol. Cell Biol. 2005, 6, 850-861. 5. Carpita, N. C.; McCann, M. C., Redesigning plant cell walls for the biomass-based bioeconomy. J. Biol. Chem. 2020, 295, 15144-15157. 6. Gow, N. A. R.; Latge, J. P.; Munro, C. A., The Fungal Cell Wall: Structure, Biosynthesis, and Function. Microbiol. Spectr. 2017, 5, FUNK-0035-2016. 7. Sarkar, P.; Bosneaga, E.; Auer, M., Plant cell walls throughout evolution: towards a molecular understanding of their design principle. J. Exp. Bot. 2009, 60, 3615-3635. 8. Paulraj, T.; Wennmalm, S.; Wieland, D. C. F.; Riazanova, A. V.; Dedinaite, A.; Pomorski, T. G.; Cardenas, M.; Svagan, A. J., Primary cell wall inspired micro containers as a step towards a synthetic plant cell. Nat. Commun. 2020, 11, 958. 9. Reif, B.; Ashbrook, S. E.; Emsley, L.; Hong, M., Solid-State NMR Spectroscopy. Nat. Rev. Methos Primers 2021, 1, 2. 10. Kelly, J. E.; Chrissian, C.; Stark, R. E., Tailoring NMR experiments for structural characterization of amorphous biological solids: A practical guide. Solid State Nucl. Magn. Reson. 2020, 109, 101686. 11. Baldus, M., Biological solid-state NMR: Integrative across different scientific disciplines. J. Struct. Biol. X 2022, 6, 100075. 12. van der Wel, P. C. A., New applications of solid-state NMR in structural biology. Emerg. Top. Life Sci. 2018, 2, 57-67. 13. Zhao, W.; Deligey, F.; Shekar, S. C.; Mentink-Vigier, F.; Wang, T., Current limitations of solid-state NMR in carbohydrate and cell wall research. J. Magn. Reson. 2022, 341, 107263. 14. Ghassemi, N.; Poulhazan, A.; Deligey, F.; Mentink-Vigier, F.; Marcotte, I.; Wang, T., Solid- State NMR Investigations of Extracellular Matrixes and Cell Walls of Algae, Bacteria, Fungi, and Plants. Chem. Rev. 2022, 122, 10036-10086. 15. Warnet, X. L.; Arnold, A. A.; Marcotte, I.; Warschawski, D. E., In-Cell Solid-State NMR: An Emerging Technique for the Study of Biological Membranes. Biophys. J. 2015, 109 (12), 2461- 2466. 115 16. Renault, M.; Tommassen-van, B. R.; Bos, M. P.; Post, J. A.; Tommasseen, J.; Baldus, M., Cellular solid-state nuclear magnetic resonance spectroscopy. Proc. Natl. Acad. Sci. USA 2012, 109, 4863-4868. 17. Gao, Y.; Mortimer, J. C., Unlocking the architecture of native plant cell walls via solid-state nuclear magnetic resonance Methods Cell Biol. 2020, 160, 121-143. 18. Kirui, A.; Dickwella Widanage, M. C.; Mentink-Vigier, F.; Wang, P.; Kang, X.; Wang, T., Preparation of Fungal and Plant Materials for Structural Elucidation Using Dynamic Nuclear Polarization Solid-State NMR. J. Vis. Exp. 2019, 144, e59152. 19. Chatterjee, S.; Prados-Rosales, R.; Tan, S.; Phan, V. C.; Chrissian, C.; Itin, B.; Wang, H.; Khajo, A.; Magliozzo, R. S.; Casadevall, A.; Stark, R. E., The melanization road more traveled by: Precursor substrate effects on melanin synthesis in cell-free and fungal cell systems. J. Biol. Chem. 2018, 293 (52), 20157-20168. 20. Laydevant, F.; Mahabadi, M.; Llido, P.; Bourgouin, J. P.; Caron, L.; Arnold, A.; Marcotte, I.; Warschawski, D. E., Growth-phase dependence of bacterial membrane lipid profile and labeling for in-cell solid-state NMR applications. BBA-Biomembranes 2022, 1864, 183819. 21. Shcherbakov, A. A.; Medeiro-Silva, J.; Tran, N.; Gelenter, M. D.; Hong, M., From angstroms to nanometers: measuring interatomic distances by solid-state NMR. Chem. Rev. 2021, 122, 9848- 9879. 22. Heise, H.; Seidel, K.; Etzkorn, M.; Becker, S.; Baldus, M., 3D NMR spectroscopy for resonance assignment and structure elucidation of proteins under MAS: novel pulse schemes and sensitivity considerations. J. Magn. Reson. 2005, 173, 64-74. 23. Li, S.; Zhang, Y.; Hong, M., 3D 13C–13C–13C correlation NMR for de novo distance determination of solid proteins and application to a human α-defensin. J. Magn. Reson. 2010, 202, 203-210. 24. Shekar, S. C.; Zhao, W.; Fernando, L. D.; Hung, I.; Wang, T., A 13C three-dimensional DQ- SQ-SQ correlation experiment for high-resolution analysis of complex carbohydrates using solid- state NMR. J. Magn. Reson. 2022, 336, 107148. 25. Pauli, J.; Baldus, M.; van Rossum, B. J.; de Groot, H.; Oschkinat, H., Backbone and Side- Chain 13C and 15N Signal Assignments of the α-Spectrin SH3 Domain by Magic Angle Spinning Solid-State NMR at 17.6 Tesla. Chem. Biochem. 2001, 2, 272-281. 26. Paëpe, G. D.; Lewandowski, J. R.; Loquet, A.; Böckmann, A.; Griffin, R. G., Proton assisted recoupling and protein structure determination. J. Chem. Phys. 2008, 129 (24), 245101. 27. Cadars, S.; Lesage, A.; Emsley, L., Chemical Shift Correlations in Disordered Solids. J. Am. Chem. Soc. 2005, 127 (12), 4466-4476. 28. Safeer, A.; Kleiburg, F.; Bahri, S.; Beriashvili, D.; Veldhuizen, E. J. A.; van Neer, J.; Tegelaar, M.; de Cock, H.; Wosten, H. A. B.; Baldus, M., Probing Cell-Surface Interactions in 116 Fungal Cell Walls by High-Resolution 1H-Detected Solid-State NMR Spectroscopy. Chem. Euro. J. 2022, 29, e202202616. 29. Phyo, P.; Hong, M., Fast MAS 1H–13C correlation NMR for structural investigations of plant cell walls. J. Biomol. NMR 2019, 73, 661-674. 30. Marchand, T. L.; Schubeis, T.; Bonaccorsi, M.; Paluch, P.; Lalli, D.; Pell, A. J.; Andreas, L. B.; Jaudzems, K.; Stanek, J.; Pintacuda, G., 1H-Detected Biomolecular NMR under Fast Magic-Angle Spinning. Chem. Rev. 2022, 122, 9943-10018. 31. Bougault, C.; Ayala, I.; Vollmer, W.; Simorre, J. P.; Schanda, P., Studying intact bacterial peptidoglycan by proton-detected NMR spectroscopy at 100 kHz MAS frequency. J. Struct. Biol. 2019, 206, 66-72. 32. Ni, Q. Z.; Daviso, E.; Can, T. V.; Markhasin, E.; Jawla, S. K.; Swager, T. M.; Temkin, R. J.; Herzfeld, J.; Griffin, R. G., High Frequency Dynamic Nuclear Polarization. Acc. Chem. Res. 2013, 46 (9), 1933-1941. 33. Rossini, A. J.; Zagdoun, A.; Lelli, M.; Lesage, A.; Copéret, C.; Emsley, L., Dynamic Nuclear Polarization Surface Enhanced NMR Spectroscopy. Acc. Chem. Res. 2013, 46 (9), 1942-1951. 34. Takahashi, H.; Ayala, I.; Bardet, M.; De Paëpe, G.; Simorre, J. P.; Hediger, S., Solid-state NMR on bacterial cells: selective cell wall signal enhancement and resolution improvement using dynamic nuclear polarization. J. Am. Chem. Soc. 2013, 135 (13), 5105-10. 35. Kundu, K.; Mentink-Vigier, F.; Feintuch, A.; Vega, S., DNP Mechanisms. eMagRes 2019, 8 (3), 295-337. 36. Mentink-Vigier, F.; Marin-Montesinos, I.; Jagtap, A. P.; Halbritter, T.; van Tol, J.; Hediger, S.; Lee, D.; Sigurdsson, S. T.; De Paëpe, G., Computationally Assisted Design of Polarizing Agents for Dynamic Nuclear Polarization Enhanced NMR: The AsymPol Family. J. Am. Chem. Soc. 2018, 140 (35), 11013-11019. 37. Sauvée, C.; Rosay, M.; Casano, G.; Aussenac, F.; Weber, R. T.; Ouari, O.; Tordo, P., Highly Efficient, Water-Soluble Polarizing Agents for Dynamic Nuclear Polarization at High Frequency. Angew. Chem. 2013, 52 (41), 10858-10861. 38. Smith, A. N.; Märker, K.; Hediger, S.; De Paëpe, G., Natural Isotopic Abundance 13C and 15N Multidimensional Solid-State NMR Enabled by Dynamic Nuclear Polarization. J. Phys. Chem. Lett. 2019, 10 (16), 4652-4662. 39. Takahashi, H.; Lee, D.; Dubois, L.; Bardet, M.; Hediger, S.; De Paëpe, G., Rapid Natural- Abundance 2D 13C–13C Correlation Spectroscopy Using Dynamic Nuclear Polarization Enhanced Solid-State NMR and Matrix-Free Sample Preparation. Angew. Chem. 2012, 51 (47), 11766-11769. 40. Zhao, W.; Kirui, A.; Deligey, F.; Mentink-Vigier, F.; Zhou, Y.; Zhang, B.; Wang, T., Solid- state NMR of unlabeled plant cell walls: high-resolution structural analysis without isotopic 117 enrichment. Biotechnol. Biofuels 2021, 14 (1), 1-14. 41. Berruyer, P.; Gericke, M.; Moutzouri, P.; Jakobi, D.; Bardet, M.; Karlson, L.; Schantz, S.; Heinze, T.; Emsley, L., Advanced characterization of regioselectively substituted methylcellulose model compounds by DNP enhanced solid-state NMR spectroscopy. Carbohydr. Polym. 2021, 262, 117944. 42. Kirui, A.; Ling, Z.; Kang, X.; Widanage, M. C. D.; Mentink-Vigier, F.; French, A. D.; Wang, T., Atomic resolution of cotton cellulose structure enabled by dynamic nuclear polarization solid-state NMR. Cellulose 2019, 26 (1), 329-339. 43. Viger-Gravel, J.; Lan, W.; Pinon, A. C.; Berruyer, P.; Emsley, L.; Bardet, M.; Luterbacher, J., Topology of pretreated wood fibers using dynamic nuclear polarization. J. Phys. Chem. C 2019, 123 (50), 30407-30415. 44. Cosgrove, D. J., Re-constructing our models of cellulose and primary cell wall assembly. Curr. Opin. Plant Biol. 2014, 22, 122-131. 45. Albersheim, P.; Darvill, A.; Roberts, K.; Sederoff, R.; Staehelin, A., Plant Cell Walls, from Chemistry to Biology. Garland Science: New York, 2011. 46. Pauly, M.; Albersheim, P.; Darvill, A.; York, W. S., Molecular domains of the cellulose/xyloglucan network in the cell walls of higher plants. Plant J. 1999, 20, 629-639. 47. Atmodjo, M. A.; Hao, Z.; Mohnen, D., Evolving views of pectin biosynthesis. Annu. Rev. Plant Biol. 2013, 64, 747-79. 48. Zykwinska, A.; Thibault, J. F.; Ralet, M. C., Organization of pectic arabinan and galactan side chains in association with cellulose microfibrils in primary cell walls and related models envisaged. . J. Exp. Bot. 2007, 58, 1795-1802. 49. Biswal, A. K.; Atmodjo, M. A.; Li, M.; Baxter, D.; Yoo, C. G.; Pu, Y.; Lee, Y. C.; Mazarei, M.; Black, I. M.; Zhang, J. Y.; Ramanna, H.; Bray, A. L.; King, Z. R.; LaFayette, P. R.; Pattathil, S.; Donohoe, B. S.; Mohanty, S. S.; Ryno, D.; Yee, K.; Thompson, O. A.; Rodriguez, M., Jr.; Dumitrache, A.; Natzke, J.; Winkeler, K.; Collins, C.; Yang, X.; Tan, L.; Sykes, R. W.; Gjersing, E.; Ziebell, A.; Turner, G. B.; Decker, S. R.; Hahn, M. G.; Davison, B. H.; Udvardi, M. K.; Mielenz, J. R.; Davis, M. F.; Nelson, R. S.; Parrott, W. A.; Ragauskas, A. J.; Stewart, N. J.; Mohnen, D., Sugar release and growth of biofuel crops are improved by downregulation of pectin biosynthesis. Nat. Biotechnol. 2018, 36, 249-257. 50. White, P. B.; Wang, T.; Park, Y. B.; Cosgrove, D. J.; Hong, M., Water–Polysaccharide Interactions in the Primary Cell Wall of Arabidopsis thaliana from Polarization Transfer Solid- State NMR. J. Am. Chem. Soc. 2014, 136 (29), 10399-10409. 51. Ridley, B. L.; O'Neill, M. A.; Mohnen, D., Pectins: Structure, biosynthesis, and oligogalacturonide-related signaling. Phytochemistry 2001, 57 (6), 929-967. 52. Kirui, A.; Du, J.; Zhao, W.; Barnes, W.; Kang, X.; Anderson, C. T.; Xiao, C.; Wang, T., A 118 pectin Methyltransferase modulates polysaccharide dynamics and interactions in Arabidopsis primary cell walls: Evidence from solid-state NMR. Carbohydr. Polym. 2021, 270, 118370. 53. Wang, T.; Zabotina, O.; Hong, M., Pectin-cellulose interactions in the Arabidopsis primary cell wall from two-dimensional magic-angle-spinning solid-state nuclear magnetic resonance. Biochemistry-Us 2012, 51 (49), 9846-9856. 54. Dick-Perez, M.; Zhang, Y. A.; Hayes, J.; Salazar, A.; Zabotina, O. A.; Hong, M., Structure and interactions of plant cell wall polysaccharides by two- and three-dimensional magic-angle- spinning solid-state NMR. Biochemistry-Us 2011, 50 (6), 989-1000. 55. Wang, T.; Park, Y. B.; Cosgrove, D. J.; Hong, M., Cellulose-pectin spatial contacts are inherent to never-dried Arabidopsis primary cell walls: evidence from solid-state nuclear magnetic resonance. Plant Physiol. 2015, 168 (3), 871-884. 56. Zhang, T.; Tang, H.; Vavylonis, D.; Cosgrove, D. J., Disentangling loosening from softening: insights into primary cell wall structure. Plant J. 2019, 100 (6), 1101-1117. 57. Zhao, W.; Fernando, L. D.; Kirui, A.; Deligey, F.; Wang, T., Solid-state NMR of plant and fungal cell walls: a critical review. Solid State Nucl. Magn. Reson. 2020, 107, 101660. 58. Zhang, Y.; Yu, J.; Wang, X.; Durachko, D. M.; Zhang, S.; Cosgrove, D. J., Molecular insights into the complex mechanics of plant epidermal cell walls. Science 2021, 372 (6543), 706- 711. 59. Pelloux, J.; Rusterucci, C.; Mellerowicz, E. J., New insights into pectin methylesterase structure and function. Trends Plant Sci. 2007, 12 (6), 267-277. 60. Du, J.; Anderson, C. T.; Xiao, C., Dynamics of pectic homogalacturonan in cellular morphogenesis and adhesion, wall integrity sensing and plant development. Nat. Plants 2022, 8, 332-340. 61. Haas, K. T.; Wightman, R.; Meyerowitz, E. M.; Peaucelle, A., Pectin homogalacturonan nanofilament expansion drives morphogenesis in plant epidermal cells. Science 2020, 367, 1003- 1007. 62. Temple, H.; Phyo, P.; Yang, W.; Lyczakowski, J. J.; Echevarria-Poza, A.; Yakunin, I.; Parra-Rojas, J. P.; Terrett, O. M.; Saez-Aguayo, S.; Dupree, R.; Orellana, A.; Hong, M.; Dupree, P., Golgi-localized putative S-adenosyl methionine transporters required for plant cell wall polysaccharide methylation Nat. Plants 2022, 8, 656-669. 63. Phyo, P.; Gu, Y.; Hong, M., Impact of acidic pH on plant cell wall polysaccharide structure and dynamics: insights into the mechanism of acid growth in plants from solid-state NMR. Cellulose 2019, 26 (1), 291-304. 64. Phyo, P.; Wang, T.; Xiao, C.; Anderson, C. T.; Hong, M., Effects of Pectin Molecular Weight Changes on the Structure, Dynamics, and Polysaccharide Interactions of Primary Cell Walls of Arabidopsis thaliana: Insights from Solid-State NMR. Biomacromolecules 2017, 18 (9), 2937- 119 2950. 65. Phyo, P.; Wang, T.; Kiemle, S. N.; O'Neill, H.; Pingali, S. V.; Hong, M.; Cosgrove, D. J., Gradients in Wall Mechanics and Polysaccharides along Growing Inflorescence Stems. Plant Physiol. 2017, 175 (4), 1593-1607. 66. Du, J.; Kirui, A.; Huang, S.; Wang, L.; Barnes, W. J.; Kiemle, S. N.; Zheng, Y.; Rui, Y.; Ruan, M.; Qi, S.; Kim, S. H.; Wang, T.; Cosgrove, D. J.; Anderson, C. T.; Xiao, C., Mutations in the Pectin Methyltransferase QUASIMODO2 Influence Cellulose Biosynthesis and Wall Integrity in Arabidopsis. Plant Cell 2020, 32, 3576-3597. 67. Cosgrove, D. J.; Jarvis, M. C., Comparative structure and biomechanics of plant primary and secondary cell walls. Front Plant Sci 2012, 3, 204. 68. Fernandes, A. N.; Thomas, L. H.; Altaner, C. M.; Callow, P.; Forsyth, V. T.; Apperley, D. C.; Kennedy, C. J.; Jarvis, M. C., Nanostructure of cellulose microfibrils in spruce wood. Proc. Natl. Acad. Sci. USA 2011, 108 (47), E1195-E1203. 69. Dupree, R.; Simmons, T. J.; Mortimer, J. C.; Patel, D.; Iuga, D.; Brown, S. P.; Dupree, P., Probing the Molecular Architecture of Arabidopsis thaliana Secondary Cell Walls Using Two- and Three-Dimensional 13C Solid State Nuclear Magnetic Resonance Spectroscopy. Biochemistry-Us 2015, 54 (14), 2335-2345. 70. Simmons, T. J.; Mortimer, J. C.; Bernardinelli, O. D.; Poppler, A. C.; Brown, S. P.; deAzevedo, E. R.; Dupree, R.; Dupree, P., Folding of xylan onto cellulose fibrils in plant cell walls revealed by solid-state NMR. Nat. Commun. 2016, 7, 13902. 71. Grantham, N. J.; Wurman-Rodrich, J.; Terrett, O. M.; Lyczakowski, J. J.; Stott, K.; Iuga, D.; Simmons, T. J.; Durand-Tardif, M.; Brown, S. P.; Dupree, R.; Busse-Wicher, M.; Dupree, P., An even pattern of xylan substitution is critical for interaction with cellulose in plant cell walls. Nat. Plants 2017, 3 (11), 859-865. 72. Kang, X.; Kirui, A.; Dickwella Widanage, M. C.; Mentink-Vigier, F.; Cosgrove, D. J.; Wang, T., Lignin-polysaccharide interactions in plant secondary cell walls revealed by solid-state NMR. Nat. Commun. 2019, 10 (1), 347. 73. Kirui, A.; Zhao, W.; Deligey, F.; Yang, H.; Kang, X.; Mentink-Vigier, F.; Wang, T., Carbohydrate-aromatic interface and molecular architecture of lignocellulose. Nat. Commun. 2022, 13, 538. 74. Terrett, O. M.; Lyczakowski, J. J.; Yu, L.; Iuga, D.; Franks, W. T.; Brown, S. P.; Dupree, R.; Dupree, P., Molecular architecture of softwood revealed by solid-state NMR. Nat. Commun. 2019, 10 (1), 4978. 75. Gao, Y.; Lipton, A. S.; Wittmer, Y.; Murray, D. T.; Mortimer, J. C., A grass-specific cellulose-xylan interaction dominates in sorghum secondary cell walls. Nat. Commun. 2020, 11 (1), 6081. 120 76. Busse-Wicher, M.; Gomes, T. C. F.; Tryfona, T.; Nikolovski, N.; Stott, K.; Grantham, N. J.; Bolam, D. N.; Skaf, M. S.; Dupree, P., The pattern of xylan acetylation suggests xylan may interact with cellulose microfibrils as a twofold helical screw in the secondary plant cell wall of Arabidopsis thaliana. Plant J. 2014, 79 (3), 492-506. 77. Chakraborty, A.; Deligey, F.; Quach, J.; Mentink-Vigier, F.; Wang, P.; Wang, T., Biomolecular complex viewed by dynamic nuclear polarization solid-state NMR spectroscopy. Biochem. Soc. Trans. 2020, 48, 1089-1099. 78. Addison, B.; Stengel, D.; Bharadwaj, V. S.; Happs, R. M.; Doeppke, C.; Wang, T.; Bomble, Y. J.; Holland, G. P.; Harman-Ware, A. E., Selective One-Dimensional 13C–13C Spin-Diffusion Solid-State Nuclear Magnetic Resonance Methods to Probe Spatial Arrangements in Biopolymers Including Plant Cell Walls, Peptides, and Spider Silk. J. Phys. Chem. B 2020, 124, 9870-9883. 79. Cresswell, R.; Dupree, R.; Brown, S. P.; Pereira, C. S.; Skaf, M. S.; Sorieul, M.; Dupree, P.; Hill, S., Importance of Water in Maintaining Softwood Secondary Cell Wall Nanostructure. Biomacromolecules 2021, 22, 4669-4680. 80. Li, H.; Kang, X.; Yang, M.; Kasseney, B. D.; Zhou, X.; Liang, S.; Zhng, X.; J., W.; Yu, B.; Liu, N.; Zhao, Y.; Mo, J.; Currie, C. R.; Ralph, J.; Yelle, D., Molecular Insight Into the Evolution of Woody Plant Decay in the Gut of Termites. Sci. Ad. 2023, 9, eadg1258. 81. Rytioja, J.; Hilden, K.; Yuzon, J.; Hatakka, A.; de Vries, R. P.; Makela, M. R., Plant- Polysaccharide-Degrading Enzymes from Basidiomycetes. Microbiol. Mol. Biol. Rev. 2014, 78, 614-649. 82. Chatterjee, S.; Prados-Rosales, R.; Itin, B.; Casadevall, A.; Stark, R. E., Solid-state NMR Reveals the Carbon-based Molecular Architecture of Cryptococcus neoformans Fungal Eumelanins in the Cell Wall. J. Biol. Chem. 2015, 290 (22), 13779-13790. 83. Chatterjee, S.; Prados-Rosales, R.; Frases, S.; Itin, B.; Casadevall, A.; Stark, R. E., Using Solid-State NMR To Monitor the Molecular Consequences of Cryptococcus neoformans Melanization with Different Catecholamine Precursors. Biochemistry-Us 2012, 51 (31), 6080- 6088. 84. Zhong, J. Y.; Frases, S.; Wang, H.; Casadevall, A.; Stark, R. E., Following fungal melanin biosynthesis with solid-state NMR: Biopolyrner molecular structures and possible connections to cell-wall polysaccharides. Biochemistry 2008, 47 (16), 4701-4710. 85. Baker, R. P.; Chrissian, C.; Stark, R. E.; Casadevall, A., Cryptococcus neoformans melanization incorporates multiple catecholamines to produce polytypic melanin. J. Biol. Chem. 2022, 298, 101519. 86. Kang, X.; Kirui, A.; Muszynski, A.; Widanage, M. C. D.; Chen, A.; Azadi, P.; Wang, P.; Mentink-Vigier, F.; Wang, T., Molecular architecture of fungal cell walls revealed by solid-state NMR. Nat. Commun. 2018, 9, 2747. 87. Latgé, J.-P., The cell wall: a carbohydrate armour for the fungal cell. Mol. Microbiol. 2007, 66 121 (2), 279-290. 88. Wagener, J.; Striegler, K.; Wagener, N., α- and β-1,3-Glucan Synthesis and Remodeling. In The Fungal Cell Wall: An Armour and a Weapon for Human Fungal Pathogens, Latgé, J. P., Ed. Springer, Cham.: 2020; Vol. 425, pp 53-82. 89. Wagener, J.; Loiko, V., Recent Insights into the Paradoxical Effect of Echinocandins. J. Fungi 2018, 4, 5. 90. Latgé, J. P.; Chamilos, G., Aspergillus fumigatus and Aspergillosis in 2019. Clin. Microbiol. Rev. 2019, 33 (1), e00140-18. 91. Chakraborty, A.; Fernando, L. D.; Fang, W.; Dickwella Widanage, M. C.; Wei, P.; Jin, C.; Fontaine, T.; Latgé, J. P.; Wang, T., A molecular vision of fungal cell wall organization by functional genomics and solid-state NMR. Nat. Commun. 2021, 12, 6346. 92. Wang, T.; Yang, H.; Kubicki, J. D.; Hong, M., Cellulose Structural Polymorphism in Plant Primary Cell Walls Investigated by High-Field 2D Solid-State NMR Spectroscopy and Density Functional Theory Calculations. Biomacromolecules 2016, 17 (6), 2210-2222. 93. Wang, T.; Hong, M., Solid-state NMR investigations of cellulose structure and interactions with matrix polysaccharides in plant primary cell walls. J. Exp. Bot. 2016, 67, 503-514. 94. Fernando, L. D.; Dickwella Widanage, M. C.; Penfield, J.; Lipton, A. S.; Washton, N.; Latge, J.-P.; Wang, P.; Zhang, L.; Wang, T., Structural Polymorphism of Chitin and Chitosan in Fungal Cell Walls from Solid-State NMR and Principal Component Analysis. Front. Mol. Biosci. 2021, 8, 727053. 95. Heux, L.; Brugnerotto, J.; Desbrieres, J.; Versali, M. F.; Rinaudo, M., Solid state NMR for determination of degree of acetylation of chitin and chitosan. Biomacromolecules 2000, 1 (4), 746- 751. 96. Tanner, S. F.; Chanzy, H.; Vincendon, M.; Roux, J. C.; Gaill, F., High-Resolution Solid- State C-13 Nuclear-Magnetic-Resonance Study of Chitin. Macromolecules 1990, 23 (15), 3576- 3583. 97. Latgé, J. P.; Wang, T., Modern Biophysics Redefines Our Understanding of Fungal Cell Wall Structure, Complexity, and Dynamics. mBio 2022, 13, e01145-22. 98. Lamon, G.; Lends, A.; Valsecchi, I.; Wong, S. S. W.; Dupres, V.; Lafont, F.; Tolchard, J.; Schmitt, C.; Mallet, A.; Grelard, A.; Morvan, E.; Dufourc, E.; Habenstein, B.; Guijarro, J. I.; Aimanianda, V.; Loquet, A., Solid-state NMR molecular snapshots of Aspergillus fumigatus cell wall architecture during a conidial morphotype transition. Proc. Natl. Acad. Sci. USA 2023, 120, e2212003120. 99. Fernando, L. D.; Dickwella Widanage, M. C.; Shekar, C. S.; Mentink-Vigier, F.; Wang, P.; Wi, S.; Wang, T., Solid-state NMR analysis of unlabeled fungal cell walls from Aspergillus and Candida species. J. Struct. Biol. X 2022, 6, 100070. 122 100. Chrissian, C.; C.P.-C., L.; Camacho, E.; Casadevall, A.; Neiman, A. M.; Stark, R. E., Unconventional Constituents and Shared Molecular Architecture of the Melanized Cell Wall of C. neoformans and Spore Wall of S. cerevisiae. J. Fungi 2020, 6, 329. 101. Ehren, H. L.; Appels, F. V. W.; Houben, K.; Renault, M. A. M.; Wösten, H. A. B.; Baldus, M., Characterization of the cell wall of a mushroom forming fungus at atomic resolution using solid-state NMR spectroscopy. Cell Surf. 2020, 6, 100046. 102. Bahri, S.; Safeer, A.; Adler, A.; Smedes, H.; van Ingen, H.; Baldus, M., 1H-detected characterization of carbon–carbon networks in highly flexible protonated biomolecules using MAS NMR. J. Biomol. NMR 2023, 77, 111-119. 103. Kleijburg, F. E. L.; Safeer, A.; Baldus, M.; Wosten, H. A. B., Binding of micro-nutrients to the cell wall of the fungus Schizophyllum commune. Cell Surf. 2023, 10, 100108. 104. van Dijk, A.; Molhoek, E. M.; Veldhuizen, E. J. A.; Bokhoven, J. L. M. T.; Wagendrop, E.; Bikker, F.; Haagsman, H. P., Identification of chicken cathelicidin-2 core elements involved in antibacterial and immunomodulatory activities Mol. Immunol. 2009, 46, 2465-2473. 105. Ndukwe, I. E.; Black, I.; Heiss, C.; Azadi, P., Evaluating the Utility of Permethylated Polysaccharide Solution NMR Data for Characterization of Insoluble Plant Cell Wall Polysaccharides. Anal. Chem. 2020, 92, 13221-13228. 106. Mansfield, S. D.; Kim, H.; Lu, F. C.; Ralph, J., Whole plant cell wall characterization using solution-state 2D NMR. Nat Protoc 2012, 7 (9), 1579-1589. 107. Kim, H.; Ralph, J., Solution-state 2D NMR of ball-milled plant cell wall gels in DMSO- d6/pyridine-d5. Org. Biomol. Chem. 2010, 8, 576-591. 108. Wang, W.; Cai, J.; Xu, Z.; Zhang, Y.; Niu, F.; Gao, M.; Jun, Y., Structural characteristics of plant cell wall elucidated by solution-state 2D NMR spectroscopy with an optimized procedure. Green Process. Synth. 2020, 9, 650-653. 109. Reichhardt, C.; Joubert, L. M.; Clemons, K. V.; Stevens, D. A.; Cegelski, L., Integration of electron microscopy and solid-state NMR analysis for new views and compositional parameters of Aspergillus fumigatus biofilms. Med. Mycol. 2019, 57, S239-S244. 110. Deligey, F.; Frank, M. A.; Cho, S. H.; Kirui, A.; Mentink-Vigier, F.; Swulius, M. T.; Nixon, T.; Wang, T., Structure of In Vitro-Synthesized Cellulose Fibrils Viewed by Cryo-Electron Tomography and 13C Natural-Abundance Dynamic Nuclear Polarization Solid-State NMR. Biomacromolecules 2022, 23, 2290-2301. 111. Nicolas, W. J.; Fabler, F.; Dutka, P.; Schur, F. K. M.; Jensen, G.; Meyerowitz, E., Cryo- electron tomography of the onion cell wall shows bimodally oriented cellulose fibers and reticulated homogalacturonan networks. Curr. Biol. 2022, 32, 2375-2389. 112. Metskas, L. A.; Wilfong, R.; Jensen, G., Subtomogram averaging for biophysical analysis and supramolecular context. J. Struct. Biol. X 2022, 6, 100076. 123 113. Harapin, J.; Eibauer, M.; Medalia, O., Structural Analysis of Supramolecular Assemblies by Cryo-Electron Tomography. Structure 2013, 21, 1522-1530. 114. Jimenez-Ortigosa, C.; Jiang, J.; Chen, M.; Kuang, X.; Healey, K. R.; Castellano, P.; Boparai, N.; Ludtke, S. J.; Perlin, D. S.; Dai, W., Preliminary Structural Elucidation of β-(1,3)- glucan Synthase from Candida glabrata Using Cryo-Electron Tomography J. Fungi 2021, 7, 120. 115. Hu, X.; Yang, P.; Chai, C.; Liu, J.; Sun, H.; Wu, Y.; Zhang, M.; Zhang, M.; Liu, X.; Yu, H., Preliminary Structural Elucidation of β-(1,3)-glucan Synthase from Candida glabrata Using Cryo-Electron Tomography Nature 2023, 616, 190-198. 116. Purushotham, P.; Ho, R.; Zimmer, J., Architecture of a catalytically active homotrimeric plant cellulose synthase complex. Science 2020, 369, 1089-1094. 117. Giummarella, N.; Pu, Y.; Ragauskas, A. J.; Lawoko, M., A critical review on the analysis of lignin carbohydrate bonds. Green Chem. 2019, 21, 1573-1595. 118. Fernando, L. D.; Perez-Llano, Y.; Dickwella Widanage, M. C.; Martinez-Avila, L.; Lipton, A. S.; Gunde-Cimerman, N.; Latge, J. P.; Bastista-Garcia, R. A.; Wang, T., Structural Organization of the Cell Wall of Halophilic Fungi. Preprint at bioRxiv 2023, DOI: 10.1101/2023.04.15.537024. 119. Romaniuk, J. A. H.; Cegelski, L., Bacterial cell wall composition and the influence of antibiotics by cell-wall and whole-cell NMR. Philos. Trans. R. Soc. B 2015, 370 (1679), 20150024. 120. Reichhardt, C.; Lim, J. Y.; Rice, D.; Fong, J. N.; Cegelski, L., Structure and Function of Bacterial Biofilms by Solid-State NMR. Biophys. J. 2014, 106 (2), 192a. 121. Poulhazan, A.; Dickwella Widanage, M. C.; Muszynski, A.; Arnold, A.; Warschawski, D. E.; Azadi, P.; Marcotte, I.; Wang, T., Identification and Quantification of Glycans in Whole Cells: Architecture of Microalgal Polysaccharides Described by Solid-State Nuclear Magnetic Resonance. J. Am. Chem. Soc. 2021, 143, 19374-19388. 122. Arnold, A. A.; Bourgouin, J. P.; Genard, B.; Warschawski, D. E.; Tremblay, R.; Marcotte, I., Whole cell solid-state NMR study of Chlamydomonas reinhardtii microalgae. J. Biomol. NMR 2018, 70 (2), 123-131. 124 CHAPTER 5: A THREE-DIMENSIONAL 13C-13C-13C CORRELATION EXPERIMENT FOR HIGH-RESOLUTION ANALYSIS OF COMPLEX CARBOHYDRATES USING SOLID-STATE NMR Research paper reprinted with permission: J. Magn. Reson. 336, 107148 (2022) Authors: S. Chandra Shekar, Wancheng Zhao, Liyanage D. Fernando, Ivan Hung, and Tuo Wang 5.1 Abstract Complex carbohydrates are the key components of the protective cell walls of microbial pathogens and the bioenergy reservoir in plants and algae. Structural characterization of these polymorphic molecules requires assistance from multidimensional 13C correlation approaches. To facilitate the analysis of carbohydrate structure using solid-state NMR, we present a three- dimensional (3D) 13C-13C-13C experiment that includes a double-quantum (DQ) dimension and is thus free of the cube’s body diagonal. The enhanced resolution supports the unambiguous resonance assignment of many polysaccharides in plant and fungal cell walls using uniformly 13C- labeled cells of spruce and Aspergillus fumigatus. Long-range structural restraints were effectively obtained to revisit our understanding of the spatial organization of plant cellulose microfibrils. The method is widely applicable to the investigations of cellular carbohydrates and carbon-based biomaterials. 5.2 Introduction Three-dimensional (3D) experiments are a standard approach used in biomolecular solid- state NMR studies but have rarely been applied to biomaterials1-3. Baldus, Rienstra, and Hong have established a collection of 3D 13C-13C-13C (CCC) correlation experiments that flexibly combine the detection of single-quantum (SQ) and double-quantum (DQ) 13C chemical shifts with various polarization transfer pathways and frequency filters to fulfill the needs for resonance assignment 125 and obtaining long-range structural restraints4-7. Such 3D correlation experiments also have the potential of assisting the identification of many coexisting polymers and understanding their spatial association in carbon-based biocomposites8-10. Indeed, recent NMR studies of plant cell walls have already benefited from the SQ-SQ-SQ CCC experiment (Figure 5.1a). In 2011, the first 3D CCC spectrum of plant materials was collected on uniformly 13C-labeled primary cell walls of Arabidopsis seedlings, reporting numerous intermolecular cross peaks between cellulose and pectin9. This was the first set of consolidated molecular-level data that support the putative interactions between these two biopolymers. The results also led to a revised concept that the hemicellulose xyloglucan only interacts with cellulose at limited spots9, which was later combined with biochemical evidence to build new conceptual models of primary plant cell walls11-13. In 2015, this experiment was extended to the investigation of secondary plant cell walls10. Demonstrated on the mature stems of the same model plant (Arabidopsis thaliana), the high-resolution data successfully revealed the conformational distribution of the hemicellulose xylan and the interconnectivity of different cellulose domains10. Currently, carbohydrate solid-state NMR has to rely on carbon-13 due to the highly restricted choice of NMR-active nuclei. Nitrogen-15 is widely used in protein studies but is only applicable to a small number of nitrogenated sugars present in nature. Proton detection and oxygen-17 studies have demonstrated their effectiveness in protein and material research14-17, but both are still in the exploratory stage for carbohydrate studies18-20. Therefore, there is an urgent need for more versatile CCC experiments to assist the structural analysis of carbohydrate polymers. 126 Figure 5.1 3D CCC experiments for characterizing carbon-rich biosolids. a, 3D 13C-13C-13C correlation experiments relying on two 13C-13C mixing periods, for example, DARR/PDSD periods. b, 3D 13C-13C-13C (DQ-SQ-SQ) experiment constructed as an INADEQUATE-CORD sequence. The INADEQUATE block is presented in a simplified format as it could be constructed using either J- or dipolar-based pulse sequence. The pulse program used for this study is provided in the Supplementary Data. c, Representative 3D INADEQUATE-CORD solid-state NMR spectrum collected on living cells of A. fumigatus at 800 MHz. The three dimensions (F1, F2, F3) report DQ, SQ, and SQ 13C chemical shifts, respectively. Here we adapt the 3D DQ-SQ-SQ CCC experiment, a scheme first demonstrated by Baldus and coworkers4, to investigate the structure and packing of carbohydrate polymers in cellular environments. After the J-INADEQUATE module21, 22, a COmbined R2nυ-Driven (CORD)23, 24 block was implemented (Figure 5.1b). The consideration is that CORD is a broadband homonuclear correlation sequence that even functions well under fast MAS conditions; therefore, it should remain effective for future applications on high-field magnets under fast spinning. Also, 127 CORD needs very low recoupling power as the radiofrequency (rf) field strength is only needed on the 1H channel and is matching only half of the spinning frequency during the majority (two- thirds) of the mixing period. This can effectively reduce the detrimental effect of rf power on sensitive biological samples such as the living cells of various organisms. The unprecedented resolution provided by this 3D CCC (DQ-SQ-SQ) experiment allows us to systematically document the 13C chemical shift of various glucans and chitin molecules in the cell walls of a pathogenic fungus Aspergillus fumigatus25, 26, using its living cells. Signal multiplicity has been observed for chitin and α-1,3-glucan; the former is attributed to the structural complexity of sugar units when placed in the hydrogen-bonded chitin microfibrils and the latter is correlated with the functional diversity of α-1,3-glucans. The most rewarding improvement is on plant biomass, the research of which has long been hampered by insufficient resolution. We have resolved the signals of previously ambiguous carbon sites and further evaluated how different glucose conformers are packed in the bundled cellulose microfibrils. This experiment has shown its usefulness in the investigations of carbohydrate polymers and should facilitate the ongoing efforts in understanding the carbohydrate structure and assembly in algae, bacteria, fungi, plants, and human cells, as well as biomimetic materials27-32. The pulse program has been provided in the Supplementary Data to expedite future investigations. 5.3 Results and Discussion The 3D INADEQUATE-CORD spectrum of the 3-day-old A. fumigatus mycelium showed remarkable dispersion of the carbon peaks (Figure 5.1c). The sample contains the living cells of this fungus, which should be a highly complex and challenging biosystem for NMR investigation. However, the outstanding spectral resolution (typically 0.3-0.7 ppm for 13C linewidths as reported recently)33, 34 allowed us to use it as a model system for cellular carbohydrates. The conventional 128 INADEQUATE spectrum reports the signals from directly bonded carbons (Figure 5.2a), and all the spin pairs of A. fumigatus carbohydrates were marked using their DQ chemical shifts. As 1H- 13C cross polarization was used for the creation of initial magnetization, the signals detected here mainly originated from four carbohydrate polymers found in the rigid portion of fungal cell walls (Figure 5.2b). These molecules include chitin, α-1,3-glucan, β-1,3-glucan (occasionally with β- 1,6-linkage for branching), and the mixed linked β-1,3/1,4-glucan, which is a terminal segment of the diversely linked β-glucan network35-37, The implementation of a 53-ms CORD mixing period after the INADEQUATE block allowed the 13C magnetization to be transferred to all carbon sites in a molecule that showed up at a given double-quantum chemical shift (Figure 5.3a). For example, the β-1,3-glucan carbon 1 and carbon 2 (B1 and B2) have single-quantum (SQ) chemical shifts of 103.6 ppm and 74.4 ppm, respectively. Therefore, in the INADEQUATE spectrum, their signals showed up at the double- quantum (DQ) chemical shift of 178 ppm, which is the sum of the two SQ chemical shifts. When CORD is implemented, the other carbons of β-1,3-glucan (B3, B4, B5, and B6) also appeared in the same DQ line, with well-resolved signals for resonance assignment. Such spectral feature presents a unique advantage over the conventional 2D correlation spectrum. Both the DQ-SQ correlation (such as INADEQUATE) and SQ-SQ correlation (such as DARR and CORD) spectra often have highly crowded regions, where the carbon connectivity cannot be explicitly tracked. For example, the carbon 1 signals of chitin and β-1,3-glucan are almost identical; therefore, their C1-CX (CX denotes other carbons) cross peaks are severely overlapped signals in the CORD spectrum (Figure 5.3b). However, their C1-C2 spin pairs have distinct DQ chemical shifts: 159 ppm for chitin C1-C2 and 178 ppm for β-1,3-glucan C1-C2 as shown in Figure 5.3a. Therefore, we can unambiguously resolve and assign all their carbon signals 129 in the collapsed INADEQUATE-CORD spectrum, which is free of diagonal and has better dispersion of signals as benefited from the implementation of a DQ dimension. Figure 5.2 2D spectra of fungal cell wall carbohydrates. a, 2D 13C-13C INADEQUATE (DQ-SQ) spectrum of A. fumigatus cells. Key cross sections are marked using the DQ 13C chemical shifts. b, Representative A. fumigatus carbohydrates with the NMR abbreviations labeled. All carbohydrates are color-coded, and the color scheme is used for all fungal figures. When a DQ chemical shift contains only a single pair of carbons, all the carbon sites in the same molecule can be well resolved in the third dimension (Figure 5.4a and Figure S5.1). For example, in the F2-F3 plane of F1 = 178 ppm, both B1 and B2 will have cross peaks with the other carbons in β-1,3-glucan. Similar spectral patterns were also observed in other F2-F3 panels of this molecule, regardless of the F1 DQ chemical shifts, which provided confidence to the resonance assignment. The β-1,4-linked glucose units only account for a minor portion of the β-glucan matrix because the β-1,3/1,4-glucan constitute approximately 10% of all β-glucans38. It is not surprising that β-1,4-linked glucose residues have been particularly difficult to track using NMR. However, its 13C signals can be assigned using the F2-F3 plane extracted at F1 = 138 ppm. 130 Figure 5.3 Collapsed 2D spectra of the 3D INADEQUATE-CORD experiment. The 2D spectra plotted in black represent all the signals of the 3D INADEQUATE-CORD when being projected to a, the F1-F3 dimensions and b, the F2-F3 dimensions. For comparison, a conventional INADEQUATE spectrum (red) is overlaid with the F1-F3 plane, where the extra signals are from multi-bond intramolecular correlations within each molecule. The representative signals of chitin (Ch) and β-1,3-glucan (B) are labeled. For example, Ch1 denotes chitin carbon 1. Both α-1,3-glucan and chitin showed peak multiplicity, with multiple sets of signals for each carbon. This observation further correlates with the functional diversity and structural complexity of these biomolecules. Recent studies have revealed that α-1,3-glucan is widely distributed in all functional domains of fungal cell walls, including the rigid and mobile portions (from the physical perspective) of both the alkali-soluble and alkali-insoluble fractions (from the chemical view)33, 34. Differently, chitin has complex structural environments in the hydrogen- bonded microfibrils, which has been recently revealed by solid-state NMR and principal component analysis39. 131 Figure 5.4 2D planes of the 3D spectra resolving signals of fungal cell wall carbohydrates. The double quantum 13C chemical shifts (in ppm) are marked in panels (a) and (b) to match the signals in Figure 5.2a. a, Well-resolved F2-F3 planes extracted from 3D INADEQUATE-CORD (DQ- SQ-SQ) spectrum of the A. fumigatus sample. The left panels show signals of β-1,3-linked glucans or glucose units (blue) while the right panel shows α-1,3-glucan (green), chitin (orange), and β- 1,4-linked glucose residues. The numbers (1-6) indicate the carbon numbers in each carbohydrate molecule as presented in Figure 5.2b. Two arrows are used to annotate the lines of peaks for β- 1,3-glucan carbon 1 (B1) and carbon 2 (B2), with correlations to each of the six carbons in this molecule. b, F2-F3 planes with multiple molecules observed in each 2D plane. The full views of these panels are presented in Figure S5.1. It is expected that a single DQ chemical shift might represent multiple pairs of carbon. For example, the 160 ppm DQ chemical shift is the sum of 104 ppm and 56 ppm from chitin C1 and C2, as well as 85 ppm and 75 ppm from the C4 and C5 of β-1,4-linked glucose units (Figure 5.4b). All the 13C signals are well resolved even in the mixed planes, for example, those with concurrent contributions from chitin and β-1,3/1,4-glucan (F1 = 160 ppm), β-1,3/1,4-glucan and α-1,3-glucan 132 (F1 = 173 ppm), β-1,3-glucan and α-1,3-glucan (F1 = 154 ppm), as well as chitin and α-1,3-glucan (F1 = 156 ppm). It is obvious that the unprecedented resolution offered by the 3D should facilitate the NMR studies of carbohydrate-rich biomaterials. This 3D CCC experiment is particularly useful for the analysis of highly complex biosystems, such as plant secondary cell walls, whose spectra are much more congested than those of the fungal cells. Even the collapsed 2D INADEQUATE-CORD spectrum measured on a spruce stem sample failed to provide sufficient resolution for resolving a large number of carbohydrate units and conformers present in the cellulose microfibrils and matrix polysaccharides (Figure S5.2). However, the F2-F3 planes extracted at F1 = 138 ppm (cellulose carbon 5 and 6) and F1 = 178 ppm (cellulose carbon 1 and carbon 2) efficiently separated the many carbon peaks in the glucan chains residing on the surface or the internal domains of cellulose microfibrils (Figure 5.5a). Matrix polysaccharides exhibited even better resolution (F1 = 174 ppm) as benefited from their partial mobility. We further spotted the signals from the hydrophobic (sg) and hydrophilic (sf) chains of cellulose surface, the deeply embedded core chains (ic), and an intermediate layer (ia and ib) bridging the core and the surface (Figure 5.5b). These glucose units have been recently identified using high-resolution 2D 13C-13C correlation spectra40, with their hydroxymethyl conformations restrained by 1H-1H distance measurements41. However, the peak congestion of these glucose units caused ambiguity in the previous efforts to complete the resonance assignment40. The resolution presented by the well-isolated signals, such as these C6-C1 cross peaks, has never been achieved before. We can even resolve the carbon 2 signals of internal and surface chains that are apart by less than 0.5 ppm, which is below the 13C FWHM of individual cellulose peaks (0.7-1.0 ppm). 133 Figure 5.5 2D F2-F3 planes of spruce extracted from the 3D spectrum. a, Representative planes of spruce cellulose (F1 = 138 ppm and 178 ppm) and matrix polysaccharides (F1 = 174 ppm). The rows of internal (i) and surface (s) chains are shown using arrows. b, Zoom-in views of the numbered regions boxed by dashlines in panel (a). A bundle formed by three elementary cellulose microfibrils is shown to indicate the different packing environment for internal cellulose (types a, b, c) and surface chains (types f and g). Although the spectrum is diagonal-free considering the 3D cube’s body as shown in Figure 5.1c, the diagonals and anti-diagonals of the 2D F2-F3 planes are clearly visible (Figure 5.6), arising only from the correlated spins in F3-F2 plane at a given F1 (DQ) frequency, as has been noted by Baldus and co-workers4. For example, Figure 5.6 shows the resonance assignment of all signals observed in the 2D plane extracted at F1 = 138 ppm, which is the DQ chemical shifts of interior cellulose C5-C6 and surface cellulose C5-C6 carbon pairs. Therefore, diagonal signals can 134 only be observed for C6-C6 and C5-C5 signals in the type-f and type-g surface chains as well as different types of interior residues. However, these diagonal signals differ from those of the conventional 2D SQ-SQ correlation spectra where all carbon sites present in the biosystem have their corresponding diagonal peaks. The diagonal signals observed in a 2D F2-F3 plane are more discrete as they only originate from the carbons selected by the DQ chemical shift (for example, only cellulose C5 and C6 here). Many other carbon sites in matrix polysaccharides have similar SQ chemical shifts but they are filtered out by their distinct DQ chemical shifts. Also, the diagonal does not span across the whole spectral width (for example, 10-180 ppm for the 13C spectrum of plant cell walls) as typically observed in 2D correlation spectra. The DQ selection has clearly alleviated the spectral crowding issue. Figure 5.6 Diagonal and antidiagonal signals of F2-F3 planes. Resonance assignment is shown for a F2-F3 plane (F1 = 138 ppm; cellulose C5-C6) extracted from the 3D spectrum of the spruce sample. Dashlines show the positions of the diagonal and anti-diagonal cross peaks. The assignments of the diagonal peaks are highlighted in bold. 135 A more promising application of this 3D experiment is to identify long-range correlations for structural determination. This was evidenced by the numerous intermolecular cross peaks reported by the F2-F3 plane (F1 = 138 ppm) of the 3D spectrum measured with a 300-ms long CORD mixing period. Site-specific and conformer-specific information of polymer packing can be obtained. For example, instead of simply reporting a cross peak between surface cellulose carbon 6 and interior cellulose carbon 4 (s6-i4), we identified that the type-a and type-b glucose units (ia,b) were responsible for contacting the surface chains (Figure 5.7). This observation perfectly aligns with the structural scheme presented in Figure 5.5b, as ia,b represents the layer of glucan chains that are directly underneath the surface chains. Always, two types of surface chains, sf and sg , were simultaneously involved in the cross peaks with internal chains (more specifically, the ia,b chains), with comparable intensities. This is understandable as these two types of surface chains, regardless of their capability of retaining water molecules, are both tightly associated with the internal chains. While only a single F2-F3 plane is shown here to demonstrate the technical feasibility, the full long-range correlation 3D spectrum carries a rich pool of data that could possibly revise our understanding of the structure and bundling of cellulose, as well as its interactions with matrix polymers. Both the fungal and plant samples studied here are never-dried materials. The native hydration level is also key to the spectral resolution because the hydration state is particularly important for the mobile and non-crystalline components that rely on motional dynamics to narrow the lines. Fortunately, most carbohydrates could efficiently retain their associations with water molecules. Three recent studies conducted on the primary and secondary plant cell walls as well as the microalgal cells have demonstrated that the 13C resolution of cell wall carbohydrates is largely independent of the hydration history of the sample and can be fully restored upon 136 rehydration27, 42, 43. However, irreversible changes might occur to the nanostructure of water- stabilized biopolymer interface upon dehydration44, 45. Therefore, the sample preparation procedures, in addition to the spectroscopic methods, should be carefully designed and tested to ensure sufficient resolution for investigating carbohydrate-rich biomaterials. Figure 5.7 Long-range intermolecular interactions in spruce cellulose. The comparison of F2-F3 planes (F1 = 138 ppm) showed numerous long-range intermolecular cross peaks. The key interactions between internal and surface chains of cellulose are selectively highlighted using dash line circles. Within each circle, the signals of different types of glucose units can be unambiguously resolved, providing site- and conformer-specific information on the chain packing of cellulose. 5.4 Materials and Methods 5.4.1 Sample Preparation Three uniformly 13C-labeled samples were used in this study, including the intact cells of A. fumigatus (also with uniform 15N-labeling), the never-dried stems of spruce, and the model tri- peptide formyl-Met-Leu-Phe-OH (MLF). For both A. fumigatus and spruce, around 35-40 mg of materials were packed into 3.2 mm pencil rotors. In addition, 7 mg of MLF was center packed into 137 a 4 mm rotor. The fungal sample was freshly prepared following previously established protocols33, 34, 39. Briefly, the mycelium of A. fumigatus was obtained by growing the wild-type strain (RL 578) in a minimum liquid medium containing 13C-glucose (10.0 gram/L) and 15N-sodium nitrate (6.0 gram/L). The culture was incubated at 30 °C for 3 days under 210 rpm shaking condition. The material was then harvested by centrifugation (7,000 × g; 20 mins), and the resulting pellet was washed using 10 mM phosphate buffer (pH 7.4). The 13C-spruce sample was acquired from Isolife (the Netherlands), with the stems debarked manually. The wood material was hand-grinded using a pestle and a mortar to achieve relatively uniform size for the pieces in the range of a few millimeters across. 5.4.2 Experiments The 3D experiment was first verified using MLF on a 400 MHz (9.4 Tesla) using a 4.0 mm probe under 10 kHz MAS. The CORD mixing was 53 ms for the 3D experiment. The recycle delay was 1.6 s, and 8 scans were collected. The acquisition time was 15 ms, 7.0 ms, and 5.2 ms for the F3, F2, and F1, respectively (with respective spectral widths of 497, 99, and 80 ppm). The F2 and F1 dimensions have 140 and 84 points (70 and 42 complex points), respectively. The total experimental time was 43 h. The F2-F3 planes showed good resolution of all carbon sites as demonstrated in Figure S5.3. Both the A. fumigatus and spruce samples were measured on an 800 MHz NMR (18.8 Tesla) Bruker Avance III HD spectrometer at the National High Magnetic Field Laboratory (NHMFL) using a 3.2 mm Low-E HCN probe designed and constructed in-house. The experiments were conducted under 13.5 kHz MAS at 293 K. The 13C chemical shifts were externally referenced to the tetramethylsilane (TMS) scale. The typical radiofrequency field strengths were 83 kHz for 1H decoupling and pulses, and 62.5 kHz for 13C pulses. 138 A 53-ms CORD recoupling time was used for the 3D experiment of A. fumigatus. The spectral widths were set to 497 ppm, 68 ppm, and 66 ppm for the F3, F2, and F1, respectively. The acquisition time was chosen to be 15 ms (F3), 6.0 ms (F2), and 6.0 ms (F1). This corresponds to 164 and 160 slices for the F2 and F1 dimensions. The recycle delay was 1.1 s, and 16 scans were collected. The total experimental time was 134 h. Two INADEQUATE-CORD spectra were collected using the spruce sample, with short (53 ms) and long (300 ms) CORD mixing. Both 3D experiments used spectral widths of 497 ppm, 62 ppm, and 70 ppm for the F3, F2, and F1, respectively. The short-mixing 3D spectrum has acquisition times of 15 ms, 4.1 ms, and 4.1 ms for the F3, F2, and F1, respectively. These values correspond to 102 and 116 slices for the F2 and F1 dimensions. The short-mixing 3D spectrum was collected using 32 scans and 1 s recycle delays, with a total experimental time of 110 h. The number of scans was doubled for the long-mixing 3D spectrum for better signal-to-noise ratios, which is the barrier for observing the weak long-range cross peaks. The recycle delay was further shortened to 0.93 s. The F2 and F1 dimensions have 74 and 84 slices, respectively. The total experimental time for the long-mixing 3D was 136 h. For each sample, two collapsed 2D spectra of the F1-F3 and F2-F3 dimensions were measured separately using identical parameters as used in the 3D experiments. 5.5 Conclusions Because the current solid-state NMR investigations of complex carbohydrates heavily rely on 13C detection, the resolution improvement provided by this 3D DQ-SQ-SQ correlation experiment has significantly expanded our technical capability. This study has demonstrated the unambiguous assignment of all carbon signals in the chitin and glucans of fungal cell walls. The spectra have also shown sufficient resolution for investigating cellulose and matrix 139 polysaccharides in plant biomass. This 3D CCC approach could become a standard experiment for future investigations of the polymorphic structure and spatial association of carbohydrates in cellular systems and functional biomaterials46. Inclusion of novel homonuclear recoupling and third-spin-assisted recoupling methods47-49, as well as the introduction of underexplored nuclei (such as 1H, 2H 15N, 17O, and 19F)18-20, 50-53 might further extend the use of 3D experiments in carbohydrate solid-state NMR research. 5.6 Acknowledgments This work was supported by the National Science Foundation MCB-1942665. The National High Magnetic Field Laboratory is supported by National Science Foundation through NSF/DMR-1644779 and the State of Florida. 140 REFERENCES 1. Pauli, J.; Baldus, M.; van Rossum, B.; de Groot, H.; Oschkinat, H., Backbone and Side-Chain 13C and 15N Signal Assignments of the a-Spectrin SH3 Domain by Magic Angle Spinning Solid- State NMR at 17.6 Tesla. ChemBioChem 2001, 2, 272-281. 2. Reif, B.; Ashbrook, S. E.; Emsley, L.; Hong, M., Solid-state NMR spectroscopy. Nat. Rev. Methods Primers 2021, 1 (1), 2. 3. Hoffmann, J.; Ruta, J.; Shi, C.; Hendriks, K.; Chevelkov, V.; Franks, W. T.; Oschkinat, H.; Giller, K.; Becker, S.; Lange, A., Protein resonance assignment by BSH-CP-based 3D solid-state NMR experiments: A practical guide Magn. Reson. Chem. 2020, 58, 445-465. 4. Heise, H.; Seidel, K.; Etzkorn, M.; Becker, S.; Baldus, M., 3D NMR spectroscopy for resonance assignment and structure elucidation of proteins under MAS: novel pulse schemes and sensitivity considerations. J. Magn. Reson. 2005, 173, 64-74. 5. Zhou, D.; Kloepper, K. D.; Winter, K. A.; Rienstra, C. M., Band-selective 13C homonuclear 3D spectroscopy for solid proteins at high field with rotor-synchronized soft pulses. J. Biomol. NMR 2006, 34, 245-257. 6. Li, S.; Zhang, Y.; Hong, M., 3D 13C-13C-13C Correlation NMR for De Novo Distance Determination of Solid Proteins and Application to a Human Alpha Defensin. J. Magn. Reson. 2010, 202, 203-210. 7. Dregni, A. J.; Mandala, V. S.; Wu, H.; Elkins, M. R.; Wang, H. K.; Hung, I.; DeGrado, W. F.; Hong, M., In vitro 0N4R tau fibrils contain a monomorphic β-sheet core enclosed by dynamically heterogeneous fuzzy coat segments. Proc. Natl. Acad. Sci. USA 2019, 116, 16357- 16366. 8. Duer, M. J., The contribution of solid-state NMR spectroscopy biomineralization: atomic and molecular structure of bone J. Magn. Reson. 2015, 253, 98-110. to understanding 9. Dick-Pérez, M.; Zhang, Y.; Hayes, J.; Salazar, A.; Zabotina, O. A.; Hong, M., Structure and interactions of plant cell-wall polysaccharides by two-and three-dimensional magic-angle- spinning solid-state NMR. Biochemistry 2011, 50 (6), 989-1000. 10. Dupree, R.; Simmons, T. J.; Mortimer, J. C.; Patel, D.; Iuga, D.; Brown, S. P.; Dupree, P., Probing the Molecular Architecture of Arabidopsis thaliana Secondary Cell Walls Using Two- and Three-Dimensional 13C Solid State Nuclear Magnetic Resonance Spectroscopy. Biochemistry 2015, 54 (14), 2335-2345. 11. Cosgrove, D. J., Re-constructing our models of cellulose and primary cell wall assembly. Curr. Opin. Plant Biol. 2014, 22, 122-131. 12. Zhang, Y.; Yu, J.; Wang, X.; Durachko, D. M.; Zhang, S.; Cosgrove, D. J., Molecular insights into the complex mechanics of plant epidermal cell walls. Science 2021, 372 (6543), 706- 711. 141 13. Park, Y. B.; Cosgrove, D. J., A revised architecture of primary cell walls based on biomechanical changes induced by substrate-specific endoglucanases Plant Physiol. 2012, 158, 1933-1943. 14. Struppe, J.; Quinn, C. M.; Lu, M.; Wang, M.; Hou, G.; Lu, X.; Kraus, J.; Andreas, L. B.; Stanek, J.; Lalli, D.; Lesage, A.; Pintacuda, G.; Maas, W.; Gronenborn, A. M.; Polenova, T., Expanding the horizons for structural analysis of fully protonated protein assemblies by NMR spectroscopy at MAS frequencies above 100 kHz. Solid State Nucl. Magn. Reson. 2017, 87, 117- 125. 15. Andreas, L. B.; Jaudzems, K.; Stanek, J.; Lalli, D.; Bertarello, A.; Le Marchand, T.; Cala- De Paepe, D.; Kotelovica, S.; Akopjana, I.; Knott, B.; Wegner, S.; Engelke, F.; Lesage, A.; Emsley, L.; Tars, K.; Herrmann, T.; Pintacuda, G., Structure of fully protonated proteins by proton-detected magic-angle spinning NMR. Proc. Natl. Acad. Sci. USA 2016, 113 (33), 9187- 9192. 16. Su, Y.; Andreas, L.; Griffin, R. G., Magic Angle Spinning NMR of Proteins: High-Frequency Dynamic Nuclear Polarization and 1H Detection. Annu. Rev. Biochem. 2015, 84 (1), 465-497. 17. Keeler, E. G.; Michaelis, V. K.; Wison, C. B.; Hung, I.; Wang, X.; Gan, Z.; Griffin, R. G., High-Resolution 17O NMR Spectroscopy of Structural Water. J. Phys. Chem. B 2019, 123, 3061- 3067. 18. Phyo, P.; Hong, M., Fast MAS 1H–13C correlation NMR for structural investigations of plant cell walls. J. Biomol. NMR 2019, 73 (12), 661-674. 19. Bougault, C.; Ayala, I.; Vollmer, W.; Simorre, J. P.; Schanda, P., Studying intact bacterial peptidoglycan by proton-detected NMR spectroscopy at 100 kHz MAS frequency. J. Struct. Biol. 2019, 206 (1), 66-72. 20. Sefzik, T. H.; Houseknecht, J. B.; Clark, T. M.; Prasad, S.; Lowary, T. L.; Gan, Z.; Grandinetti, P. J., Solid-state 17O NMR in carbohydrates. Chem. Phys. Lett. 2007, 434 (4), 312- 315. 21. Cadars, S.; Sein, J.; Duma, L.; Lesage, A.; Pham, T. N.; Baltisberger, J. H.; Brown, S. P.; Emsley, L., The refocused INADEQUATE MAS NMR experiment in multiple spin-systems: Interpreting observed correlation peaks and optimising lineshapes. J. Magn. Reson. 2007, 188 (1), 24-34. 22. Cadars, S.; Lesage, A.; Emsley, L., Chemical Shift Correlations in Disordered Solids. J. Am. Chem. Soc. 2005, 127 (12), 4466-4476. 23. Hou, G.; Yan, S.; Sun, S.; Han, Y.; Byeon, I.-J. L.; Ahn, J.; Concel, J.; Samoson, A.; Gronenborn, A. M.; Polenova, T., Spin Diffusion Driven by R-Symmetry Sequences: Applications to Homonuclear Correlation Spectroscopy in MAS NMR of Biological and Organic Solids. J. Am. Chem. Soc. 2011, 133 (11), 3943-3953. 24. Hou, G.; Yan, S.; Trébosc, J.; Amoureux, J.-P.; Polenova, T., Broadband homonuclear 142 correlation spectroscopy driven by combined R2nv sequences under fast magic angle spinning for NMR structural analysis of organic and biological solids. J. Magn. Reson. 2013, 232, 18-30. 25. Brown, G. D.; Denning, D. W.; Gow, N. A. R.; Levitz, S. M.; Netea, M. G.; White, T. C., Hidden Killers: Human Fungal Infections. Sci. Transl. Med. 2012, 4 (165), 165rv13. 26. Latgé, J.-P.; Chamilos, G., Aspergillus fumigatus and Aspergillosis in 2019. Clin. Microbiol. Rev. 2019, 33 (1), e00140-18. 27. Poulhazan, A.; Dickwella Widanage, M. C.; Muszynski, A.; Arnold, A. A.; Warschawski, D. E.; Azadi, P.; Marcotte, I.; Wang, T., Identification and Quantification of Glycans in Whole Cells: Architecture of Microalgal Polysaccharides Described by Solid-State Nuclear Magnetic Resonance. J. Am. Chem. Soc. 2021, 46, 19374-19388. 28. Kelly, J. E.; Chrissian, C.; Stark, R. E., Tailoring NMR experiments for structural characterization of amorphous biological solids: A practical guide. Solid State Nucl. Magn. Reson. 2020, 109, 101686. 29. Ghassemi, N.; Poulhazan, A.; Deligey, F.; Mentink-Vigier, F.; Marcotte, I.; Wang, T., Solid- State NMR Investigations of Extracellular Matrices and Cell Walls of Algae, Bacteria, Fungi, and Plants. Chem. Rev. 2021, in press. 30. Thongsomboon, W.; Serra, D. O.; Possling, A.; Hadjineophytou, C.; Hengge, R.; Cegelski, L., Phosphoethanolamine cellulose: A naturally produced chemically modified cellulose. Science 2018, 359 (6373), 334-338. 31. Overall, S. A.; Barnes, A. B., Biomolecular Perturbations in In-Cell Dynamic Nuclear Polarization Experiments. Front. Mol. Biosci. 2021, 8, 743829. 32. Ghosh, R.; Xiao, Y.; Kragelj, J.; Frederick, K. K., In-Cell Sensitivity-Enhanced NMR of Intact Living Mammalian Cells. J. Am. Chem. Soc. 2021, 143 (44), 18454-18466. 33. Kang, X.; Kirui, A.; Muszynski, A.; Widanage, M. C. D.; Chen, A.; Azadi, P.; Wang, P.; Mentink-Vigier, F.; Wang, T., Molecular architecture of fungal cell walls revealed by solid-state NMR. Nat. Commun. 2018, 9, 2747. 34. Chakraborty, A.; Fernando, L. D.; Fang, W.; Dickwella Widanage, M. C.; Wei, P.; Jin, C.; Fontaine, T.; Latgé, J. P.; Wang, T., A molecular vision of fungal cell wall organization by functional genomics and solid-state NMR. Nat. Commun. 2021, 12, 6346. 35. Latgé, J.-P.; Beauvais, A.; Chamilos, G., The Cell Wall of the Human Fungal Pathogen Aspergillus fumigatus: Biosynthesis, Organization, Immune Response, and Virulence. Annu. Rev. Microbiol. 2017, 71 (1), 99-116. 36. Latgé, J.-P., The cell wall: a carbohydrate armour for the fungal cell. Mol. Microbiol. 2007, 66 (2), 279-290. 37. Lee, M. J.; Sheppard, D. C., Recent advances in the understanding of the Aspergillus fumigatus 143 cell wall. . J. Microbiol. 2016, 54, 232-242. 38. Gastebois, A.; Clavaud, C.; Aimanianda, V.; Latgé, J.-P., Aspergillus fumigatus: cell wall polysaccharides, their biosynthesis and organization. Future Microbiol. 2009, 4 (5), 583-595. 39. Fernando, L. D.; Dickwella Widanage, M. C.; Penfield, J.; Lipton, A. S.; Washton, N.; Latge, J.-P.; Wang, P.; Zhang, L.; Wang, T., Structural Polymorphism of Chitin and Chitosan in Fungal Cell Walls from Solid-State NMR and Principal Component Analysis. Front. Mol. Biosci. 2021, 8, 727053. 40. Wang, T.; Yang, H.; Kubicki, J. D.; Hong, M., Cellulose structural polymorphism in plant primary cell walls investigated by high-field 2D solid-state NMR spectroscopy and density functional theory calculations. Biomacromolecules 2016, 17 (6), 2210-2222. 41. Phyo, P.; Wang, T.; Yang, Y.; O’Neill, H.; Hong, M., Direct determination of hydroxymethyl conformations of plant cell wall cellulose using 1H polarization transfer solid-state NMR. Biomacromolecules 2018, 19 (5), 1485-1497. 42. Wang, T.; Park, Y. B.; Cosgrove, D. J.; Hong, M., Cellulose-pectin spatial contacts are inherent to never-dried Arabidopsis primary cell walls: evidence from solid-state nuclear magnetic resonance. Plant Physiol. 2015, 168 (3), 871-884. 43. Kirui, A.; Zhao, W.; Deligey, F.; Yang, H.; Kang, X.; Mentink-Vigier, F.; Wang, T., Carbohydrate-aromatic interface and molecular architecture of lignocellulose. Nat. Commun. 2022, 13, 538. 44. Cresswell, R.; Dupree, R.; Brown, S. P.; Pereira, C. S.; Skaf, M. S.; Mathias, S.; Dupree, P.; Hill, S., Importance of Water in Maintaining Softwood Secondary Cell Wall Nanostructure. Biomacromolecules 2021, 22, 4669-4680. 45. Gao, Y.; Lipton, A. S.; Wittmer, Y.; Murray, D. T.; Mortimer, J. C., A grass-specific cellulose–xylan interaction dominates in sorghum secondary cell walls. Nat. Commun. 2020, 11 (1), 6081. 46. Hariri El Nokab, M. E.; van der Wel, P. C. A., Use of Solid-state NMR spectroscopy for Investigating Polysaccharide-based Hydrogels: A Review. Carbohydr. Polym. 2020, 240, 116276. 47. Gelenter, M. D.; Dregni, A. J.; Hong, M., Pulsed Third-Spin-Assisted Recoupling NMR for Obtaining Long-Range 13C–13C and 15N–13C Distance Restraints. J. Phys. Chem. B 2020, 124 (33), 7138-7151. 48. Paul, S.; Takahashi, H.; Hediger, S.; De Paepe, G., Third Spin-Assisted Recoupling in SSNMR: Theoretical Insights and Practicable Application to Biomolecular Structure Determination. Annual Reports on NMR Spectroscopy 2015, 85, 93-142. 49. Paëpe, G. D.; Lewandowski, J. R.; Loquet, A.; Böckmann, A.; Griffin, R. G., Proton assisted recoupling and protein structure determination. J. Chem. Phys. 2008, 129 (24), 245101. 144 50. Wu, G., 17O NMR studies of organic and biological molecules in aqueous solution and in the solid state. Progress in Nuclear Magnetic Resonance Spectroscopy 2019, 114, 1335-191. 51. Fritz, J. H.; Kraus, J.; Quinn, C. M.; Yap, G. P. A.; Struppe, J.; Sergeyev, I. V.; Gronenborn, A. M.; Polenova, T., Measurement of accurate interfluorine distances in crystalline organic solids: a high-frequency magic angle spinning NMR approach. J. Phys. Chem. B 2019, 123, 10680-10690. 52. Shcherbakov, A. A.; Medeiros-Silva, J.; Tran, N.; Gelenter, M. D.; Hong, M., From Angstroms to Nanometers: Measuring Interatomic Distances by Solid-State NMR. Chem. Rev. 2021, in press. . 53. Gelenter, M. D.; Wang, T.; Liao, S.-Y.; O’Neill, H.; Hong, M., 2H–13C correlation solid-state NMR for investigating dynamics and water accessibilities of proteins and carbohydrates. J. Biomol. NMR 2017, 68 (4), 257-270. 145 APPENDIX Figure S5.1 F2-F3 planes of A. fumigatus 3D spectra. The ω1 double-quantum (DQ) 13C chemical shifts were labeled for each 2D plane. Bottom panel contains folded peaks for chitin CO and CH3. 146 Figure S5.2 3D spectra of spruce stems with highly crowded spectra. a, 3D INADEQUATE- CORD of never-dried spruce measured on an 800 MHz NMR. b, The 2D F1-F3 plane of 3D INADEQUATE-CORD spectrum (black). A conventional INADEQUATE spectrum (red) is compared. Blue dashlines indicate the F1 positions at which 2D F2-F3 planes are extracted in Figure 5.4. 147 Figure S5.3 F2-F3 planes extracted from the 3D spectrum of MLF. Representative 2D planes were shown for Met and Leu. The signals of these two residues are well resolved from each other. The rows of Met and Leu carbons are indicated using blue. The data were collected on a 400 MHz NMR. 148 Supplementary Data: Pulse Program of DQ-SQ-SQ 3D CCC Experiment --------------------------------------------------------------------------------------------------------------------- ;p1 : 13C 90 ;p2 : 13C 180 ;p3 : 1H 90 ;p15: CP Contact Time ;p4 :CORD basic block ;p31:P3*170/90=TPPM dec.pulse ;pl1 : 13C CP Contact ;pl2 : 1H excitation ;pl3 : 1H CP Contact ;pl11: C-C J Mixing ;pl12: 1H Dec for acq ;pl22: 1H dec for evoln and J-mixing ;cpdprg2: 1H Dec4Inadq+t1: tppm15_p26_pl22 ;cpdprg3: 1H Dec Acq, use tppm15_p31_pl12 ;cnst31 : spinning speed (Hz) ;l6=cnst31*(d2+p1/2+p2/2+1u) ;l1=td1/2, for States cosine/sine ;l2=td2/2, for States cosine/sine ;l4:CORDloop# mixing=l4*24*tauR<500m ;d0 : t1=ntr ;d2=l6*1s/cnst31-p1/2-p2/2-1u ;1/2 of J evolution, up to 5 ms ;d10 : t2=ntr ;d11=30m ;#include ;#include "l1=td1/2" ; for States (t1) "in0=inf1" "l2=td2/2" ; for States (t1 & t2) "in10=inf2" aqseq 321 "p4=(1s/cnst31)/2" "p30=p31-0.4u" "d2=l6*1s/cnst31-p1/2-p2/2-1u" ; 1/2 of J evolution, up to 5 ms "d11=30m" dccorr ze 149 1 d11 ; t1-loop ; d11 ; t1 States loop ; d11 ; t2-loop ; d11 ; t2 States-loop ; d11 ; go-loop d1 do:f2 1m rpp17 10u reset:f1 reset:f2 6u setnmr3|28 \n 4u setnmr3^28 10u pl1:f1 10u pl2:f2 p3:f2 ph19 (p15:spf0 ph1):f1 (p15 pl3 ph20):f2 1u cpds2:f2 pl22:f2 d2 pl11:f1 ; C-C J mixing p2:f1 ph2 ; X 180 pulse d2 1u p1:f1 ph3 ; X 90 mixing d0 ; t1=ntr p1:f1 ph4 ; X 90 mixing d2 ; refocus p2:f1 ph5 d2 d10 p1:f1 ph6 ; X 90 mixing 1u do:f2 ; cord 2 (p4 pl13 ph17^):f2 (p4 pl13 ph17^):f2 (p4 pl13 ph17^):f2 (p4 pl13 ph17^):f2 (p4*2 pl14 ph17^):f2 (p4*2 pl14 ph17^):f2 (p4*2 pl14 ph17^):f2 (p4*2 pl14 ph17^):f2 (p4 pl13 ph17^):f2 150 (p4 pl13 ph17^):f2 (p4 pl13 ph17^):f2 (p4 pl13 ph17^):f2 (p4*2 pl14 ph17^):f2 (p4*2 pl14 ph17^):f2 (p4*2 pl14 ph17^):f2 (p4*2 pl14 ph17^):f2 (p4 pl13 ph17^):f2 (p4 pl13 ph17^):f2 (p4 pl13 ph17^):f2 (p4 pl13 ph17^):f2 (p4*2 pl14 ph17^):f2 (p4*2 pl14 ph17^):f2 (p4*2 pl14 ph17^):f2 (p4*2 pl14 ph17^):f2 (p4 pl13 ph17^):f2 (p4 pl13 ph17^):f2 (p4 pl13 ph17^):f2 (p4 pl13 ph17^):f2 (p4*2 pl14 ph17^):f2 (p4*2 pl14 ph17^):f2 (p4*2 pl14 ph17^):f2 (p4*2 pl14 ph17^):f2 lo to 2 times l4 1u do:f2 p1:f1 ph7 ; X 90 mixing 1u cpds3:f2 pl12:f2 go=1 ph31 30u do:f2 ; decoupler off d11 wr #0 if #0 zd ; write2disk,incrmntFilePntr,0data 1m ip4 1m ip5 lo to 1 times 2 ; States detection 1m rp4 1m rp5 9m id10 ; increment d0 by in0 = tr lo to 1 times l2 10m rd10 1m ip1 1m ip2 1m ip3 151 lo to 1 times 2 ; States detection 1m rp1 1m rp2 1m rp3 9m id0 ; increment d0 by in0 = tr lo to 1 times l1 10m rd0 HaltAcqu, 1m exit ph1 = (8) 2 4 6 0 ;CP pulse on X ph2 = (8) 0 2 4 6 ;first 180 ph3 = (8) 0 2 4 6 ;DQ generation ;DQ reconversion ph4 = 0 ph5 = {0}*8 {2}*8 ;second 180 ph6 = 2 ph7 = 0 ph17= 1 3 2 2 1 3 2 2 2 0 3 3 2 0 3 3 3 1 0 0 3 1 0 0 0 2 1 1 0 2 1 1 ph19= {1}*4 {3}*4 ;1H excitation ph20= 0 ;1H CP ph31= 0 2 0 2 2 0 2 0 ; NS=8*n 152 CHAPTER 6: EFFECT OF CROSS POLARIZATION RADIOFREQUENCY PHASES ON SIGNAL PHASE Research paper reprinted with permission: Solid State Nucl. Magn. Reson. 117, 101771 (2022) Authors: S. Chandra Shekar#, Wancheng Zhao#, Thomas K. Weldeghiorghis, and Tuo Wang 6.1 Abstract #Equal contribution Utilizing phases of radio frequency (RF) pulses to manipulate spin dynamics is routine in NMR and MRI, leading to spectacular techniques like phase cycling. In a very different area, cross polarization (CP) also has a long history as part of a vast number of solid-state NMR pulse sequences. However, a detailed study devoted to the effect of CP RF phases on NMR signal, seems not to be readily available. From the first principles, we arrive at a simple dependence of NMR signal on arbitrary CP RF phases, for static and MAS conditions, accompanied by experimental verification. In the process, the CP propagator emerges as a product of RF “pulses” and a period of “free precession”, conforming to coherence transfer pathway theory. The theoretical expressions may lend confidence for dealing with CP blocks with tunable phases in pulse sequences. 6.2 Introduction Effect of phases of radio frequency (RF) pulses on spin dynamics is a deeply studied topic in NMR and MRI1. It has led to spectacularly successful techniques such as phase cycling2, 3 to select certain coherence transfer pathways (CTP) while blocking the undesirable ones, for spectral editing and suppression of image artifacts in countless number of NMR and MRI pulse sequences. This in turn has revolutionized NMR via pulse sequences such as COSY, NOESY, TOCSY, TROSY etc., as well as innumerable 3D pulse sequences built upon them. Cross polarization (CP)4-7 technique is employed, both as a standalone experiment, as well 153 as a building block in an innumerable number of solid-state NMR experiments8. Often as a first step, CP is an integral part of solid-state NMR pulse sequences9 to transfer polarization from NMR sensitive protons (1H) to insensitive X-nuclei such as 13C and 15N10. It is a regular feature of numerous 1D, 2D, and 3D solid-state NMR pulse sequences. CP is used in static as well as magic- angle spinning (MAS) conditions, including (by no means limited to) widely differing applications such as reduction of experimental time driven by the much shorter 1H T1, selective detection of rigid components in the sample11, extraction of distance information by estimating the dipolar interaction based on the CP intensity build up curves12, etc. Also, there has been a steady effort to extend applicability by expanding its capability: combining CP with MAS (resulting in CPMAS)13, CP at fast MAS14, 15 and ultra-fast MAS16, 17, tilted angle CP (such as DOTA-CP)18, Lee-Goldburg CP19, broad band CP20 in an attempt to broaden the Hartmann-Hahn (HH) match, CP for ’wide- line’ NMR (BRAIN-CP)21, exploiting chemical shift offset to reduce the power requirements for CP10, 22, selective CP (SPECIFIC-CP)23 for X-nuclei at specific chemical shifts as an assignment strategy, and repetitive CP blocks for quantitative detection (Multi-CP)24. Interestingly CPMAS itself becomes part of the strategy to transfer from even richer source of polarization (electrons) in indirect DNP experiments25, 26. Yet, there does not seem to be an in-depth study examining the effect of CP RF phases on the NMR signal (notwithstanding simple and standard phases ±𝑥 and ± 𝑦27-29), to the best of our knowledge. The lack of related studies is further evidenced by the results of literature search. For example, what would be the effect of CP S-channel RF phase of 30o, under static and MAS conditions? We explore this elementary question below in a general manner. Both thermodynamic4-6, 27, 30 and quantum mechanical descriptions1, 14, 15, 20-23, 27, 30-40 are available to understand the CP phenomenon, which has been essential because the technique has undergone 154 many major developments to broaden its applicability since its inception. Drawing upon this established literature, we arrive at a surprisingly simple relationship between the phases of CP RF 1 fields and the NMR signal for a heteronuclear spin- 2 pair, for the pulse sequence shown in Figure 6.1, and followed by extensive experimental verification, carried out on several compounds by employing 1H-X (X= 13C, 15N, 31P) CP. Such expressions may come in handy to design pulse schemes combining transients resulting from different sets of RF phases during CP. Figure 6.1 Effect of RF phases during cross polarization on NMR signal. In the pulse sequence, 𝜙0 and 𝜙𝐼 are respectively RF phases during 𝜋 2 - pulse and CP on I-spins. Under the constraint 𝜙0 = 𝜙𝐼 + 𝜋 2 , S-spin signal acuires same phase as that of RF phase 𝜙𝑆 during CP, experimentally confirming Eq. (20). 𝜙𝐼 𝑎𝑛𝑑 𝜙𝑆 are independently cycled according to: 𝜙𝐼𝑘 = 𝑘2𝜋/𝑁𝐼 ; 𝑘 = 0,1, … . , 𝑁𝐼 − 1. Each 𝜙𝐼𝑘 was accompanied by 𝜙𝑆(𝑙) = 𝜙𝑆 0 + 𝜙𝑆 = 𝜙𝑆 0 + 𝑙2𝜋 𝑁𝑆 ; 𝑙 = 0,1, … . , 𝑁𝑆 − 1. 𝜙𝑆 0 is the constant overall phase shift. Receiver phase 𝜙𝑅 → 𝜙𝑅𝑙 = 𝜙𝑆𝑙., thus, selecting the observable -1QC for the S-spins (see text). 155 6.3 Methods To test the theoretical predictions ensuing from the derivations in this work, CP experiments were performed on four samples, packed into 4 Bruker 4-mm zirconium rotors: Rotors (1) and (2): 10 mg and 75 mg of U(13C,15N) glycine (from Cambridge Isotope Laboratories), respectively; Rotor (3): DMPC multilamellar vesicles, prepared by dispersing 25mg of the lipid (from Avanti) in 40 μL of water; Rotor (4): around 5 mg of U(13C,15N) labeled tripeptide formyl- Methionine-Leucine-Phenylalanine-OH (MLF). All data were acquired on a Bruker Avance 400 MHz (9.4 Tesla) spectrometer, equipped with a multinuclear 4-mm MAS probe as detailed previously, along with experimental details, unless noted otherwise in this work41-43. Typical RF field strengths were, 83 kHz for 1H 𝜋 2 -pulse, 71 kHz for 1H TPPM decoupling (with associated pulse widths of 6.7 and 6.3 s), 50-63 kHz for 1H CP, typically either nominally matched to the CP RF field of the X (13C, 15N, 31P) nucleus ± spinning speed (corresponding respectively to DQ or ZQ HH match), or experimentally optimized. No ramp was applied on either RF channel during CP. A recycle delay of 3s was employed between free induction decay transients, and 4 to 256 transients were accumulated for signal averaging. 6.4 Theory In Figure 6.1, the phase 𝜙0 of the excitation 𝜋 2 -pulse and 𝜙𝐼 of the CP ‘pulse’ (on I-spins) are constrained by44 𝜙0 = 𝜙𝐼 + 𝜋 2 Eq 6.1 Starting with the density matrix corresponding to I-spin equilibrium magnetization at time t = 0, 𝝈(0) = 𝐈𝐳 Eq 6.2 P, a 𝜋 2 -pulse with phase 𝜙𝐼 + 𝜋 2 on I-spins (see Figure 6.1), is given by1, 45-47 156 𝐏 ( 𝜋 2 , 𝜙𝐼 + 𝜋 2 ) = 𝐑𝑧(𝜙𝐼) 𝐑𝑦( 𝜋 2 ) 𝐑𝑧 † Eq 6.3 𝐑𝑘(𝜃) = 𝑒−𝑖𝜃𝐈𝐤 Eq 6.4 is applied. († symbolizes adjoint operation defined by the relationship between matrix elements as follows: 𝐴𝑘𝑙 † = 𝐴𝑙𝑘 ∗ , with * signaling complex conjugation). Equation (Eq.) (4) represents a rotation in spin space about a direction specified by the unit vector 𝐤̂ . The pulse rotates Eq. (2) into transverse magnetization at an angle 𝜙𝐼 relative to the x-axis in the rotating frame, given by the density matrix, 𝝈(𝟏)1, 45-47, 𝝈(𝟏) = 𝐑𝑧(𝜙𝐼) 𝐈𝐱 𝐑𝑧 † Eq 6.5 Evolution of the density matrix during CP, ignoring relaxation effects for simplicity, is governed by1, 22, 24, 38, 45-48 𝝈(𝑡) = 𝐔(𝜙𝐼𝜙𝑆; 𝑡) 𝝈(𝟏) 𝐔† Eq 6.6 𝐔(𝜙𝐼𝜙𝑆; 𝑡) = 𝐑 𝐮(𝑡) 𝐑† Eq 6.7 Its elements are determined by the RF field strengths on both I and S channels and the heteronuclear orientation dependent dipolar coupling38. Above, 𝐑 = 𝐑𝑧(𝜙𝐼𝜙𝑆) 𝐑𝑦 ( 𝑆 𝜋 2 𝐼 𝜋 2 ) Eq 6.8 𝐑𝑧(𝜙𝐼𝜙𝑆) = 𝐑𝑧(𝜙𝐼) 𝐑𝑧(𝜙𝑆) = 𝑒−𝑖(𝜙𝐼𝐈𝐳+𝜙𝑆𝐒𝐳) Eq 6.9 𝐑𝑦 ( 𝑆 𝜋 2 𝐼 𝜋 2 ) = 𝐑𝑦 ( 𝐼 𝜋 2 ) 𝐑𝑦 ( ) = 𝑒−𝑖 𝜋 2 𝑆 𝜋 2 (𝐈𝐲+𝐒𝐲) Eq 6.1 signaling1, 14, 38, 45-47, (i) a transformation from the laboratory frame (with B0 || z) to a doubly rotating frame (at respective Larmor frequencies for the I and S spins), (ii) followed by a transformation to a doubly rotated frame along z, (iii) and a subsequent transformation to a doubly tilted frame along y; 157 Using Eqs. (5) and (7) in (6), 𝝈(𝑡) = 𝐑 𝐮 𝐑† 𝐑𝐳(𝜙𝐼) 𝐈𝐱 𝐑𝐳 †(𝜙𝐼) 𝐑 𝐮† 𝐑† since (AB)† = B † A †. But from Eqs. (8), (9), and (10), 𝐑† 𝐑𝐳(𝜙𝐼)𝐈𝐱 𝐑𝐳 † ( †(𝜙𝐼)𝐑 = 𝐑𝐲 𝑆 π 2 𝐼 π 2 )𝐑𝐳 †(𝜙𝐼𝜙𝑆)𝐑𝐳(𝜙𝐼)𝐈𝐱 𝐑𝐳 †(𝜙𝐼)𝐑𝐳(𝜙𝐼𝜙𝑆)𝐑𝐲 = 𝐑𝐲 † ( 𝑆 π 2 𝐼 π 2 ) 𝐑𝐳 † ( †(𝜙𝑆)𝐈𝐱𝐑𝐳(𝜙𝑆)𝐑𝐲 = 𝐑𝐲 𝐼 π 2 )𝐈𝐱𝐑𝐲 = 𝐈𝐳, yielding, 𝝈(𝑡) = 𝐑 𝐃(𝑡) 𝐑† Eq 6.11 𝐃(𝑡) = 𝐮(𝑡) 𝐈𝐳 𝐮† Eq 6.12 The amount of an operator 𝐀 present in the density matrix at a time 𝑡 can be extracted by the following trace operation1, 45-48 〈𝐀(𝑡)〉 = 𝐓𝐫{𝐀 𝛔(𝑡)} = 𝐓𝐫{𝐑† 𝐀 𝐑 𝐃(𝑡)} Eq 6.13 Using46 𝐃𝐑†𝐒𝐱𝐑 = 𝐒𝐳 cos 𝜙𝑆 − 𝐒𝐲 sin 𝜙𝑆 𝐑†𝐒𝐲𝐑 = 𝐒𝐳 sin 𝜙𝑆 + 𝐒𝐲 cos 𝜙𝑆 𝐑†𝐒𝐳𝐑 = −𝐒𝐱 Eq 6.14 in Eq. (13) yields 〈𝐒𝐱(𝑡)〉 = cos 𝜙𝑆 𝑃(𝑡) − sin 𝜙𝑆 𝑄(𝑡) 〈𝐒𝐲(𝑡)〉 = sin 𝜙𝑆 𝑃(𝑡) + cos 𝜙𝑆 𝑄(𝑡) 〈𝐒𝐳(𝑡)〉 = −𝐓𝐫{𝐒𝐱𝐃(𝑡)} Eq 6.15 with 𝑃(𝑡) ≡ 𝐓𝐫{𝐒𝐳𝐃(𝑡)} 𝑄(𝑡) ≡ 𝐓𝐫{𝐒𝐲𝐃(𝑡)}(𝑡)} Eq 6.16 158 Distinction should be drawn between the scalars 〈𝐒𝐱(𝑡)〉, 〈𝐒𝐲(𝑡)〉 and the 4×4 matrices 1 or ±𝑖 𝐒𝐱, 𝐒𝐲, each with just 4 non-zero elements = 2 1 . Moreover in Eq. (13), when A is one of the 2 16 possible product operators (PO’s)1, 47, 49-51, 𝐑†𝐀𝐑 is mapped back into one of the 16 PO’s. Tabulates non-vanishing traces of the PO’s with D. This enables evaluation of expectation values for all 16 PO’s by Eq. (13). Remarkably, 𝑄(𝑡) ≡ 𝐓𝐫{𝐒𝐲𝐃(𝑡)} = 0 Eq 6.17 as shown by Eq. (S31), leaving us with 〈𝐒𝐱(𝑡)〉 = cos 𝜙𝑆 𝑃(𝑡) 〈𝐒𝐲(𝑡)〉 = sin 𝜙𝑆 𝑃(𝑡) Eq 6.18 Using the ladder operators1, 45-48 𝐒± = 𝐒𝐱 ± 𝑖𝐒𝐲 Eq 6.19 on Eq. (18), signal observed in quadrature is given by1-3, 45-47, 52, 53 𝑓(𝑡) ≡ 〈𝐒+(𝑡)〉 = 𝑒𝑖𝜙𝑆𝑃(𝑡) Eq 6.20 It can be seen that 𝑃(𝑡) is real. It depends only on the CP RF amplitudes and the orientation dependent 𝐼 − 𝑆 dipolar interaction, contained in the matrix elements of u(t). But the phase 𝜙𝑆, imparted to the signal is impervious to such details, and Eq. (20) is valid for any regime (Hartmann-Hahn (HH) match, HH mismatch, HH anti match) of CP RF fields, and any distribution of orientations (crystal, partial order, powder). Some of these features have emerged in theoretical simulations previously38, extending additional support for Eq. (20). Though the effect of canonical phases of RF fields during CP have been treated before in excellent treatises27-29, those pioneering works did not deal with arbitrary RF phases. For example, the outcome of phase cycling, employing RF and receiver phases ±𝑥, ± 𝑦 only, can be found in a 159 very helpful table, but without actual derivation27-29; the underlying mechanism is not elucidated. It may be possible to arrive at such a dependence purely by trial-and-error experiments, and/or the intuition based on the effect of resonant RF pulses on NMR signal phase (see below). Phase acquired1, 47 by a CTP of coherence jump Δm = −1, under an RF “pulse” of phase fS , e-i Dm fS = ei fS, is the same as in Eq. (20). At first glance, this seems surprising, as CTP theory requires1 that the propagator (the ‘block’) comprise of a sequence of RF pulses, including free precession intervals. Since the CP Hamiltonian is a sum of the non-commuting RF and dipolar Hamiltonians, the ensuing evolution cannot be factored into a simple product of exponential operators, apparently failing to meet CTP theory requirements. But, upon closer inspection, Eqs. (7) & (8) reveal that the CP evolution, is indeed a product of actual and equivalent RF ’pulses’ sandwiching a free precession interval. The propagator’s phases 𝜙𝐼, and 𝜙𝑆 stem from the phases of the actual RF irradiation on I and S-spins during CP. The diagonalizing y-rotations in Eq. (S20) play the role of RF ’pulses’ along y-axis, and the z-rotation that of a free precession interval. Thus, the CTP theory indeed applies to CP evolution. Notice that, under MAS there is an additional effective free precession interval associated with the sample rotation, as well as additional orientation (ϒ) dependent phase factors, in addition to those imparted due to CP RF phases. 6.5 Results and Discussion First experimental verification of Eq. (20) is furnished in Figure 6.2, which shows static 13C NMR CP spectrum of U (13C,15N) glycine. (Please see the section on methods for additional experimental details.) The spectra were acquired at 13C carrier frequency of 100.673 MHz with a spectral width of 50 kHz and 294 complex sampling points in quadrature for an acquisition time of 29.4 ms. 𝜙𝐼 and 𝜙𝑆 were independently phase cycled, with 𝜙𝑆 0 as the constant overall phase 𝜋 shift, as described in Figure 6.1. Explicitly in units of 2 , 𝜙𝑆 − 𝜙𝑆 0 = 𝜙𝑅 = 0,1,2,3; 𝜙0 = 𝜙𝐼 + 160 1; 𝜙𝐼 = 0,0,0,0, 1,1,1,1, 2,2,2,2, 3,3,3,3. The time domain data was subject to Gaussian multiplication. The time domain data was subsequently zero filled to 16384 complex points prior to Fourier Transformation (FT). Phase correction1, 45-47 on spectra was carried out with the same 1st order phase correction 𝜙1 𝑐, regardless of the value of 𝜙𝑆 0. On the other hand, the 0th order phase correction 𝜙0 𝑐 , depended on 𝜙𝑆 0 . In fact, for spectra acquired with 𝜙𝑆 0 ≠ 0 , corrections of 𝜙0 𝑐(𝜙𝑆 0) = 𝜙0 𝑐(0) + 𝜙𝑆 0 did the trick, completing the experimental confirmation of Eq. (20). What is more, 𝜙𝑆 0 dependence survives phase cycling of 𝜙𝐼 and 𝜙𝑆, further consolidating the validity of Eq. (20). Even though Eq. (20) was derived for the static regime, we found that it held for MAS at 10, 5, and 2.5 kHz, for the same sample as above. We also found that, for a powder sample of the tripeptide U(13C,15N) formyl-Methionine- Leucine-Phenylalanine-OH (MLF), the signal phase obeyed Eq. (20) for 13C CP MAS at 10 kHz. In addition, we found that Eq. (20) held not only for the phase cycling employed previously, but also under various other schemes (maintaining 𝜙0 = 𝜙𝐼 + 𝜋 2 ), such as no phase cycling of 𝜙𝐼 and/or 𝜙𝑆, 𝑁𝐼 and/or 𝑁𝑆 =2 or 4 steps. For the same sample, the 𝜙𝑆 dependence was also tested for 15N CP MAS at 2.5 kHz (Figure 6.3) and 10 kHz, adding to the growing list of experiments validating Eq. (20). Further affirmation of Eq. (20) followed, by 15N CP MAS at 5 kHz, 1 kHz, and 0 kHz, once again on U(13C,15N) glycine. We extended the investigation of 𝜙𝑆 dependence to 31P CP MAS of DMPC multilamellar vesicles that contain 25 mg of lipids and 40 mg of water. The MAS at 1 kHz and 5 kHz demonstrate the phase dependence described by Eq. (20), yet again. 161 Figure 6.2 Effect of 𝜙𝑆 (Figure 6.1) on 13C CP static NMR of U(13C,15N) glycine. B0=9.4 T and Larmor Frequency ν0(13C)=100 MHz. 𝜙𝐼 and 𝜙𝑆 were independently cycled with NI = NS = 4. Phase correction1, 45-47 on spectra was carried out with 1st order phase correction 𝜙1 c = 107°, regardless of the value of 𝜙𝑆 0! On the other hand, the 0th order phase correction 𝜙0 c, is a function of 𝜙𝑆 0 (Figure 6.1). All spectra were acquired by accumulating N=128 shots. Top row: All with same 𝜙0 c = 172°. Left to Right: 𝜙𝑆 0=0°; 45° and 90° respectively. Bottom row: Same as top row, but with respective corrections 𝜙0 c = 172° + 𝜙𝑆 0 = 172°; 217° and 262°. 𝜙𝑆 0 =0 spectrum was superposed in the middle and right panels. In each case, as the displayed difference spectrum shows, the overlays are virtually identical, validating Eq. (20). The 𝜙𝑆 0 dependence persists despite cycling 𝜙𝐼 and 𝜙𝑆, further supporting Eq. (20). See text for further details. 6.5.1 Validation on Phospholipid Membranes Unlike the MAS spectra, the static CP of the same sample in Figure 6.4, displays a broad dip, that is deep in the vicinity of isotropic CS, which has been observed previously.54-58 Figure 6.4 displays a superposition of static 31P NMR obtained without employing CP — dubbed as direct polarization (DP), under CP, and under CP MAS at 1 kHz. 162 Figure 6.3 Effect of 𝜙𝑆 (Figure 6.1) on 15N CP MAS (at 2.5 kHz) spectrum of U(13C,15N) tripeptide MLF. Unless otherwise noted, parameters are the same as in Figure 6.2 Larmor frequency ν0 (15N)=40 MHz. N=256 and 𝜙1 c = 95°. Top row: 𝜙0 c = −114°. Left to Right: 𝜙𝑆 0=0°; 45° and 90° respectively. Bottom row: same as top row but with respective 𝜙0 c = −114° + 𝜙𝑆 0 = −114°, −69°, and − 24°. 𝜙𝑆 0 = 0 spectrum was superposed in the middle and right panels, and the difference displayed. In each case, the overlays are virtually identical, Like in Figure 6.2, 𝜙𝑆 dependence is indifferent to phase cycling. See text for further details. When molecular rotation about an axis is rapid on the NMR time scale, the 2nd-rank CS and dipolar (D) tensors are averaged down in magnitude and projected onto the rotation axis as axially symmetric tensors27, 49-51, 59. This is borne out by the DP spectrum, which is characteristic of an effectively axially symmetric PAS for CSA, even if the tensor were asymmetric prior to averaging. The position of the underlying resonances that make up the powder pattern, accompanied by a nominal relative amplitude (1/𝑐𝑜𝑠𝛽𝑀𝐿). 163 For the static CP spectrum, these still determine the resonance position, but the amplitude depends on the residual dipolar interaction available for CP, by 𝜔𝐷 ≡ 𝑑 = 𝑑00 2 (𝛽𝑃𝑀)𝑑00 2 (𝛽𝑀𝐿)𝑑𝑃 Eq 6.21 Figure 6.4 Static 31P NMR (black), under CP (red), and under CP-MAS at 1 kHz (blue) of the same sample. The static powder pattern obtained without CP (dubbed as DP spectrum), is characteristic of an isotropic distribution of an effectively axially symmetric CSA tensors around B0, brought about by fast (on the NMR time scale) rotation about the lipid axis, as shown by Eqs. (S62) and (S76). A nominal relative intensity within the spectrum = 1/𝑐𝑜𝑠𝛽𝑀𝐿. On the other hand, for the static CP spectrum, for the same resonance position, the actual amplitude is determined by Eq. (21). At magic angle, the resonance is at isotropic shift. For DP, the nominal intensity at this shift is √3x that at 𝛽𝑀𝐿 = 0 . However, at magic angle, from Eq. (21), the residual dipolar interaction =0, resulting in a null signal from Table S2 and Table S3. The overall effect is that of a broad pit that is deep at this shift, in the CP spectrum. On the other hand, in the CP-MAS (at 1 kHz) spectrum, the averaged CSA tensor of size of ~ 8 kHz, breaks up the powder pattern into sidebands. See text for further details. 164 Which determines the CP amplitude. Above, 𝛽𝑃𝑀 (see Eq. (S68)) specifies the fixed orientation of the internuclear vector in M-frame, while 𝛽𝑀𝐿 specifies the orientation of the rotation axis relative to B0. At the magic angle, 𝛽𝑀𝐿 = 𝜃𝑀, the resonance is at the isotropic shift, and 𝑑 = 0. This results in 𝐓𝐫(𝐒𝐳 𝐃) = 0, and in turn, leads to a null signal, corroborated by current experiments (Figure 6.4) and consistent with earlier observations54-58. For one last time, we interrogated the 𝜙𝑆 dependence, on the static 31P CP spectrum of this sample, and confirmed its good agreement with Eq. (20). The totality of experimental findings, thus far, can be summarized as follows: First and foremost the phase of the S-spin NMR signal obeys Eq. (20). Always (not just for 0o & 180 o and 90 o & 270 o, corresponding to +/-x and +/- y). No dependence on 𝜙𝐼 in the pulse sequence of Figure 6.1, as predicted by Eq. (20). Hence, independent of 𝜙𝐼 cycling details. When 𝜙𝑆 is cycled, overall shift 𝜙𝑆 0 replaces 𝜙𝑆 in Eq. (20). Hence, independent of 𝜙𝑆 cycling details. It can be shown rather easily theoretically that cycling f0 between fI +/- p/2 (instead of imposing Eq. (1), for the purposes of eliminating thermal equilibrium magnetization of S-spins), still preserves Eq. (20), and was confirmed experimentally as well. 6.5.2 Consideration of CP-MAS Attention was drawn earlier to the fact that Eq. (20) was derived under the assumption of static condition. Yet, it was found to be valid under MAS, from 1 to 10 kHz, across different samples and for different observed nuclei (13C, 15N, and 31P). Initially, we entertained the idea that this is due to slow spinning relative to the CP RF amplitudes (~50kHz), and that Eq. (20) may break down as the spinning speed catches up to the CP RF amplitudes. However, we reexamined 165 the problem under the light of pioneering analytical treatments of CP MAS14, 30, 39. Closely following Zilm and coworkers14, we rederived the MAS equivalents of the evolution operator u, and the operator D, occurring crucially in Eqs. (11), (12), and (13). It can be seen that uMAS and DMAS have identical structures as their static counterparts. Thus, remain valid under MAS, with elements of D replaced by the elements of DMAS. In particular, all remain valid under MAS, along with Eqs. (17) and (20), experimentally validated time and again throughout this manuscript. Also, since seem to account for changes in spinning speed, at least the phase dependence given by Eq.(20), may survive even ultra-fast MAS rates that are available currently (> 100 kHz), but remains to be experimentally confirmed. 6.6 Conclusions We have examined the effect of RF phases during CP on the evolution operator for a heteronuclear spin- 1 pair. The process uncovers that the emergent propagator is factored into a 2 product of rotations due to equivalent RF pulses and free precession intervals (under both static and MAS conditions), in compliance with the CTP theory requirements.1 We have provided expressions for the effect of these phases on all 16 PO’s and found both theoretically and experimentally that their dependencies remain intact under both static and MAS conditions. We can also use this treatment to evaluate the effect of the CP RF phases on any of the 16 single transition operators |𝑝⟩⟨𝑞| via Eq. (11), if needed. This work may also help to incorporate the CP block with arbitrary and nonstandard 𝜙𝐼 and 𝜙𝑆 in pulse sequences and combine the resulting signals to design experiments for desired outcomes. For example, suppose the pulse sequence actually starts with CP, followed by a block of pulses, RF irradiations, delays etc. Again, Eq. (11) can be used to compute the entire density matrix (not just Sx and Sy) at the end of the 1st CP stage for further calculations. E.g., if the next stage is a 2nd CP involving a 3rd spin-1/2 nucleus, we 166 may ’feed’ the result of the 1st stage as the σ(1) parameter to Eq. (6), albeit treating the two active nuclei as the spin-1/2 ’CP pair’, to evaluate the entire density matrix after the 2nd stage. Additionally, as mentioned earlier, since the phase dependence due to CP Hamiltonian, given by Eq. (20), coincides with that due to an “ordinary” RF pulse, similar applications may become available utilizing the present work. Knowledge on the effect of arbitrary phases will guide the development of novel pulse sequences to expand the capability of NMR technique. For long we had difficulty locating a systematic description of the CP phase dependence, which stimulated us to undertake this endeavor. We think a complete presentation can be of some interest and might benefit many ssNMR researchers. A possible future direction is to find the equivalent of Eq. (20), which was derived for a spin-1/2 pair, for the case of CP from 1H to quadrupolar nuclei (static as well as MAS and ultra-fast MAS), which may or may not have an analytical solution as it is an open-ended physics question. However, a systematic numerical analysis approach may be mounted, if analytical solution is out of reach. 6.7 Acknowledgements This work was supported by the National Science Foundation MCB-1942665. 167 REFERENCES 1. Ernst, R. R.; Bodenhausen, G.; Wokaun, A., Principles of nuclear magnetic resonance in one and two dimensions. Oxford: Clarendon press: 1987; Vol. 14. 2. Wokaun, A.; Ernst, R. R., Selective detection of multiple quantum transitions in NMR by two- dimensional spectroscopy. Chem. Phys. Lett. 1977, 52 (3), 407-412. 3. Bodenhausen, G.; Kogler, H.; Ernst, R. R., Selection of coherence-transfer pathways in NMR pulse experiments. J. Magn. Reson. (1969) 1984, 58 (3), 370-388. 4. Pines, A.; Gibby, M. G.; Waugh, J. S., Proton‐enhanced nuclear induction spectroscopy. A method for high resolution NMR of dilute spins in solids. J. Chem. Phys. 1972, 56 (4), 1776-1777. 5. Pines, A.; Gibby, M. G.; Waugh, J. S., Proton‐enhanced NMR of dilute spins in solids. J. Chem. Phys. 1973, 59 (2), 569-590. 6. Hartmann, S. R.; Hahn, E. L., Nuclear double resonance in the rotating frame. Phys. Rev. 1962, 128 (5), 2042. 7. Engelke, F.; Steuernagel, S., Cross Polarization in Rotating Solids: Spin-1/2 Nuclei * Update based on original article by Frank Engelke, Encyclopedia of Magnetic Resonance, ©1996, John Wiley & Sons Ltd. In eMagRes, 2011. 8. Bryce, D. L.; Bernard, G. M.; Gee, M.; Lumsden, M. D.; Eichele, K.; Wasylishen, R. E., Practical aspects of modern routine solid-state multinuclear magnetic resonance spectroscopy: one-dimensional experiments. Can. J. Anal. Sci. Spectrosc. 2001, 46 (2), 46-81. 9. Lee, D.; Duong, N. T.; Lafon, O.; De Paëpe, G., Primostrato Solid-State NMR Enhanced by Dynamic Nuclear Polarization: Pentacoordinated Al3+ Ions Are Only Located at the Surface of Hydrated γ-Alumina. J. Phys. Chem. C 2014, 118 (43), 25065-25076. 10. Shekar, S. C.; Lee, D.-K.; Ramamoorthy, A., An experimental strategy to dramatically reduce the RF power used in cross polarization solid-state NMR spectroscopy. J. Am. Chem. Soc. 2001, 123 (30), 7467-7468. 11. Nowacka, A.; Bongartz, N. A.; Ollila, O. H. S.; Nylander, T.; Topgaard, D., Signal intensities in 1H–13C CP and INEPT MAS NMR of liquid crystals. J. Magn. Reson. 2013, 230, 165-175 12. Kolodziejski, W.; Klinowski, J., Kinetics of cross-polarization in solid-state NMR: A guide for chemists. Chem. Rev. 2002, 102 (3), 613-628. 13. Stejskal, E. O.; Schaefer, J.; Waugh, J. S., Magic-angle spinning and polarization transfer in proton-enhanced NMR. J. Magn. Reson. (1969) 1977, 28 (1), 105-112. 14. Wu, X. L.; Zilm, K. W., Cross polarization with high-speed magic-angle spinning. J. Magn. Reson., Ser. A 1993, 104 (2), 154-165. 168 15. Wi, S.; Gan, Z.; Schurko, R.; Frydman, L., Cross-polarization phenomena in the NMR of fast spinning solids subject to adiabatic sweeps. J. Chem. Phys. 2015, 142 (6), 064201. 16. Paluch, P.; Pawlak, T.; Amoureux, J.-P.; Potrzebowski, M. J., Simple and accurate determination of X–H distances under ultra-fast MAS NMR. J. Magn. Reson. 2013, 233, 56-63. 17. Paluch, P.; Trébosc, J.; Nishiyama, Y.; Potrzebowski, M. J.; Malon, M.; Amoureux, J. P., Theoretical study of CP-VC: A simple, robust and accurate MAS NMR method for analysis of dipolar C–H interactions under rotation speeds faster than ca. 60kHz. J. Magn. Reson. 2015, 252, 67-77. 18. Matsunaga, T.; Matsuda, I.; Yamazaki, T.; Ishii, Y., Decoherence optimized tilted-angle cross polarization: A novel concept for sensitivity-enhanced solid-state NMR using ultra-fast magic angle spinning. J. Magn. Reson. 2021, 322, 106857. 19. Venkatesh, A.; Hung, I.; Boteju, K. C.; Sadow, A. D.; Gor’kov, P. L.; Gan, Z.; Rossini, A. J., Suppressing 1H Spin Diffusion in Fast MAS Proton Detected Heteronuclear Correlation Solid- State NMR Experiments. Solid State Nucl. Magn. Reson. 2020, 105, 101636. 20. Baldus, M.; Meier, B. H., Broadband polarization transfer under magic-angle spinning: application to total through-space-correlation NMR spectroscopy. J. Magn. Reson. 1997, 128 (2), 172-193. 21. Harris, K. J.; Lupulescu, A.; Lucier, B. E. G.; Frydman, L.; Schurko, R. W., Broadband adiabatic inversion pulses for cross polarization in wideline solid-state NMR spectroscopy. J. Magn. Reson. 2012, 224, 38-47. 22. Shekar, S. C.; Lee, D.-K.; Ramamoorthy, A., Chemical shift anisotropy and offset effects in cross polarization solid-state NMR spectroscopy. J. Magn. Reson. 2002, 157 (2), 223-234. 23. Baldus, M.; Petkova, A. T.; Herzfeld, J.; Griffin, R. G., Cross polarization in the tilted frame: assignment and spectral simplification in heteronuclear spin systems. Mol. Phys. 1998, 95 (6), 1197-1207 24. Johnson, R. L.; Schmidt-Rohr, K., Quantitative solid-state 13C NMR with signal enhancement by multiple cross polarization. J. Magn. Reson. 2014, 239, 44-49. 25. Wang, Z.; Hanrahan, M. P.; Kobayashi, T.; Perras, F. A.; Chen, Y.; Engelke, F.; Reiter, C.; Purea, A.; Rossini, A. J.; Pruski, M., Combining fast magic angle spinning dynamic nuclear polarization with indirect detection to further enhance the sensitivity of solid-state NMR spectroscopy. Solid State Nucl. Magn. Reson. 2020, 109, 101685. 26. Lilly Thankamony, A. S.; Wittmann, J. J.; Kaushik, M.; Corzilius, B., Dynamic nuclear polarization for sensitivity enhancement in modern solid-state NMR. Prog. Nucl. Magn. Reson. Spectrosc. 2017, 102-103, 120-195. 27. Schmidt-Rohr, K.; Spiess, H. W., Multidimensional solid-state NMR and polymers. Elsevier: 2012. 169 28. Duer, M. J., Introduction to solid-state NMR spectroscopy. Oxford, UK: 2004. 29. Apperley, D. C.; Harris, R. K.; Hodgkinson, P., Solid-state NMR: Basic principles and practice. Momentum Press: 2012. 30. Meier, B. H., Cross polarization under fast magic angle spinning: thermodynamical considerations. Chem. Phys. Lett. 1992, 188 (3-4), 201-207. 31. Müller, L.; Kumar, A.; Baumann, T.; Ernst, R. R., Transient oscillations in NMR cross- polarization experiments in solids. Phys. Rev. Lett. 1974, 32 (25), 1402. 32. Müller, L.; Ernst, R. R., Coherence transfer in the rotating frame: Application to heteronuclear cross-correlation spectroscopy. Mol. Phys. 1979, 38 (3), 963-992. 33. Levitt, M. H.; Suter, D.; Ernst, R. R., Spin dynamics and thermodynamics in solid‐state NMR cross polarization. J. Chem. Phys. 1986, 84 (8), 4243-4255 34. Zhang, S.; Wu, X., A new model for cross polarization in the rotating frame in solids. Chem. Phys. Lett. 1989, 156 (4), 333-336 35. Levitt, M. H., Heteronuclear cross polarization in liquid‐state nuclear magnetic resonance: Mismatch compensation and relaxation behavior. J. Chem. Phys. 1991, 94 (1), 30-38 36. Hediger, S.; Meier, B. H.; Kurur, N. D.; Bodenhausen, G.; Ernst, R. R., NMR cross polarization by adiabatic passage through the Hartmann—Hahn condition (APHH). Chem. Phys. Lett. 1994, 223 (4), 283-288 37. Ernst, M.; Meier, B. H.; Tomaselli, M.; Pines, A., Time-reversal of cross-polarization in nuclear magnetic resonance. J. Chem. Phys. 1998, 108 (23), 9611-9613. 38. Shekar, S. C.; Ramamoorthy, A., The unitary evolution operator for cross-polarization schemes in NMR. Chem. Phys. Lett. 2001, 342 (1-2), 127-134. 39. Marica, F.; Snider, R. F., An analytical formulation of CPMAS. Solid State Nucl. Magn. Reson. 2003, 23 (1-2), 28-49 40. Marks, D.; Vega, S., A Theory for Cross-Polarization NMR of Nonspinning and Spinning Samples. J. Magn. Reson., Ser. A 1996, 118 (2), 157-172. 41. Chakraborty, A; Fernando, L.D.; Fang, W.; Dickwella Widanage, M. C.; Wei, P.; Jin, C.; Fontaine, T.; Latge, J.P.; Wang, T. Nat. Commun. 2021 12, 6346. 42. Kang, X.; Kirui, A.; Muszyński, A.; Dickwella Widanage, M. C.; Chen, A.; Azadi, P.; Wang, P.; Mentink-Vigier, F.; Wang, T., Molecular architecture of fungal cell walls revealed by solid-state NMR. Nat. Commun. 2018, 9 (1), 1-12 43. Kang, X.; Kirui, A.; Dickwella Widanage, M. C.; Mentink-Vigier, F.; Cosgrove, D. J.; Wang, T., Lignin-polysaccharide interactions in plant secondary cell walls revealed by solid-state NMR. 170 Nat. Commun. 2019, 10 (1), 1-9 44. Heise, H.; Seidel, K.; Etzkorn, M.; Becker, S.; Baldus, M., 3D NMR spectroscopy for resonance assignment and structure elucidation of proteins under MAS: novel pulse schemes and sensitivity considerations. J. Magn. Reson. 2005, 173 (1), 64-74. 45. Abragam, A., The principles of nuclear magnetism. Oxford university press: 1961. 46. Slichter, C. P., Principles of magnetic resonance. Springer Science & Business Media: 2013; Vol. 1 47. Levitt, M. H., Spin dynamics: basics of nuclear magnetic resonance. John Wiley & Sons: 2013. 48. Dirac, P. A. M., The principles of quantum mechanics. Oxford university press: 1981. 49. Shekar, S. C.; Rong, P.; Jerschow, A., Irreducible spherical tensor analysis of quadrupolar nuclei. Chem. Phys. Lett. 2008, 464 (4-6), 235-239. 50. Jaccard, G.; Wimperis, S.; Bodenhausen, G., Multiple‐quantum NMR spectroscopy of S= 3/2 spins in isotropic phase: a new probe for multiexponential relaxation. J. Chem. Phys. 1986, 85 (11), 6282-6293 51. Shekar, S. C.; Tang, J. A.; Jerschow, A., Dynamics of I=3\over2 nuclei in isotropic slow motion, anisotropic and partially ordered phases. Concepts Magn. Reson., Part A 2010, 36 (6), 362-387 52. States, D.; Haberkorn, R.; Ruben, D., A two-dimensional nuclear Overhauser experiment with pure absorption phase in four quadrants. J. Magn. Reson. (1969) 1982, 48 (2), 286-292. 53. Marion, D.; Ikura, M.; Tschudin, R.; Bax, A. D., Rapid recording of 2D NMR spectra without phase cycling. Application to the study of hydrogen exchange in proteins. J. Magn. Reson. 1989, 85 (2), 393-399. 54. Akutsu, H.; Kyogoku, Y., Phase transitions of phosphatidylcholine bilayers as revealed by cross-polarization efficiency of 31P-NMR. Biochimica et Biophysica Acta 1984, 774 (2), 293-295. 55. Akutsu, H., Detection of slow-motional changes in the phosphatidylcholine bilayers by 1H- 31P cross-polarization dynamics. J. Magn. Reson. (1969) 1986, 66 (2), 250-263. 56. Salmon, A.; Wiedemann, T., The effect of lung surfactant apolipoproteins B/C on the chemical shift anisotropy of sn-2-[1-13C]dipalmitoylphosphatidylcholine. Chem. Phys. Lipids 1991, 58 (1), 55-62. 57. Mavromoustakos, T.; Theodoropoulou, E., A combined use of 13C-cross polarization/magic angle spinning, 13C-magic angle spinning and 31P-nuclear magnetic resonance spectroscopy with differential scanning calorimetry to study cannabinoid-membrane interactions. Chem. Phys. Lipids 1998, 92 (1), 37-52. 171 58. Tsvetkova, N. M.; Phillips, B. L.; Crowe, L. M.; Crowe, J. H.; Risbud, S. H., Effect of Sugars on Headgroup Mobility in Freeze-Dried Dipalmitoylphosphatidylcholine Bilayers: Solid-State 31P NMR and FTIR Studies. Biophys. J. 1998, 75 (6), 2947-2955. 59. Harris, R. K.; Wasylishen, R. E.; Duer, M. J.; Shekar, S. C.; Jerschow, A., Tensors in NMR in 'NMR crystallography'. John Wiley & Sons: 2009; Vol. 4 172 CHAPTER 7: CARBOHYDRATE-AROMATIC INTERFACE AND MOLECULAR ARCHITECTURE OF LIGNOCELLULOSE Research paper reprinted with permission: Nat. Commun. 13, 538 (2022) Authors: Alex Kirui#, Wancheng Zhao#, Fabien Deligey#, Hui Yang, Xue Kang, Frederic Mentink-Vigier, and Tuo Wang #Equal contribution 7.1 Abstract Plant cell walls constitute the majority of lignocellulosic biomass and serve as a renewable resource of biomaterials and biofuel. Extensive interactions between polysaccharides and the aromatic polymer lignin make lignocellulose recalcitrant to enzymatic hydrolysis, but this polymer network remains poorly understood. Here we interrogate the nanoscale assembly of lignocellulosic components in plant stems using solid-state nuclear magnetic resonance (NMR) and dynamic nuclear polarization (DNP) approaches. We show that the extent of glycan-aromatic association increases sequentially across grasses, hardwoods, and softwoods. Lignin principally packs with the xylan in a non-flat conformation via non-covalent interactions and partially binds the junction of flat-ribbon xylan and cellulose surface as a secondary site. All molecules are homogeneously mixed in softwoods; this unique feature enables water retention even around the hydrophobic aromatics. These findings unveiled the principles of polymer interactions underlying the heterogeneous architecture of lignocellulose, which will guide the rational design of more digestible plants and more efficient biomass-conversion pathways. 7.2 Introduction With solar energy and carbon dioxide transformed into carbohydrate-rich cell walls, terrestrial plants constitute eighty percent of the biomass distributed in the biosphere1. The 173 secondary cell wall is a lignocellulosic composite deposited once the cellular expansion has ceased, which has evolved into a major source of biopolymers and biofuels2, 3. Lignification mechanically strengthens secondary walls, however, the presence of these intractable polyphenols and their association with carbohydrate components contributes to the biomass recalcitrance that renders the feedstock resistant to enzymatic hydrolysis during its conversion to liquid transportation fuel4, 5. To cost-effectively access structural polysaccharides for ethanol fermentation, vast efforts have been dedicated to tailoring plants to produce more digestible walls and optimizing deconstruction procedures in biorefineries6-8. These efforts have not yet reached the full potential due to our limited understanding of cell wall architecture. The secondary cell wall is assembled by carbohydrate and aromatic constituents, with remarkable complexity and variability. Each elementary cellulose microfibril contains eighteen 1,4-β-glucan chains, which are held together by a hydrogen-bonding network9, 10. The exact organization of these glucan chains is unresolved, but recent density functional theory (DFT) calculations suggest a six-layered organization, likely with 2, 3, 4, 4, 3, and 2 chains in each layer (Figure 7.1a)11. Elementary microfibrils frequently coalesce, forming large fibrils that often span across tens of nanometers12, 13. Hemicelluloses, such as xylan, glucuronoxylan, arabinoxylan, and glucomannan, are highly variable in their monosaccharide composition and linkage pattern. Xylan is among the most found hemicelluloses, and its backbone comprises β-1,4-xylose units in a wide range of conformations, with substitutions by arabinose (Ara) or glucuronic acid (GlcA), and modifications by acetyl (Ac) groups. Lignin contains guaiacyl (G), syringyl (S), and p- hydroxyphenyl (H) phenolic residues, which are interconnected by different types of covalent linkers such as β-O-4 ether-O-aryl, β-β’ resinol, and β-5’ phenylcoumaran14, 15. 174 Conceptually, the mechanical scaffold of crystalline cellulose is dispersed in a matrix formed by hemicellulose and lignin16. Our understanding of cell wall organization is supported by many studies that employed diffraction methods to reveal the spatial arrangement of cellulose microfibrils, imaging techniques to map out cell wall meshes and the microscopic distribution of lignin, and solution NMR spectroscopy to identify lignin-carbohydrate linkages17-19. However, the interface between lignin and polysaccharides, the focus of this study, is not yet well understood. This is partially due to the hardly accessible length scale (angstrom to nanometer) and the requirement of both chemical and atomic resolutions. In addition, only a small number of molecules reside on this lignin-carbohydrate interface, which needs to be deconvoluted from the bulk of the cell wall. As both lignin and polysaccharide exist in the solid state, conventional separation methods often perturb their structures and interactions, making it difficult to investigate this polymer interface. Recently, multidimensional solid-state NMR spectroscopy of Arabidopsis and Zea mays (maize) has spotlighted a structure-function relationship of the molecules involved in the lignocellulosic interface20, 21. Lignin tends to form hydrophobic and disordered nanodomains, the surface of which binds the xylan in a three-fold helical screw conformation (three sugar residues per helical turn: a non-flat structure; Figure 7.1a) through non-covalent interactions. The three- fold domain is connected to its two-fold flat-ribbon region, which is coating the smooth surface of cellulose microfibrils22-25. To generalize these structural principles, we need to examine other plant systems to evaluate three critical aspects: (i) the conformational bias of hemicellulose’s function, (ii) the absence of cellulose-lignin contact, and (iii) the self-aggregating nature of aromatic polymers. 175 The combination of solid-state NMR and DNP methods has allowed us to unveil the structural and chemical principles underlying the formation of lignocellulosic materials. We investigated the 13C-labeled stems of two hardwoods, eucalyptus (Eucalyptus grandis) and poplar (Populus x canadensis), and the softwood spruce (Picea abies). These plants are non-food energy candidates for the development of second-generation biofuels to reduce our dependence on grain crops26. Despite lignin’s preference for binding non-flat xylan, direct contacts are also observed between the aromatics and the junctions of cellulose surface and flat-ribbon xylan, which is probably enforced by molecular crowding. In these woody plants, molecules experience a highly homogeneous mixing on the nanoscale, which inevitably involves polymer interpenetration and entanglement. Consequently, the counter-intuitive hydration of aromatics is observed, where water is retained and stabilized by the polysaccharides closely packed to lignin. These molecular insights have brought our understanding of lignocellulose architecture to an unprecedented level of detail, allowing us to envision better biomass-conversion schemes for sustainable energy production. 7.3 Results 7.3.1 Polymorphic Structure of Carbohydrates in Woody Stems We obtained 13C-labeled plant stems by providing poplar, eucalyptus (gum tree), and spruce with 13CO2. Free from chemical treatments, the hydrated lignocellulosic materials were directly analyzed using solid-state NMR. Therefore, all biomolecules have fully retained their chemical structure and physical packing. Atomic-level information of polysaccharide structure was obtained using two-dimensional (2D) 13C-13C correlation experiments (Figure S7.1), which spotted 69 allomorphs of monosaccharide residues in three samples (Tables S7.1 and S7.2). These sugar units, with a wide range of linkages and conformations, were mainly found in cellulose and 176 four hemicelluloses, including glucuronoxylan, arabinoxylan, galactoglucomannan, and a very low amount of xyloglucan (Figure 7.1a). Figure 7.1 Solid-state NMR resolves the polymorphic structure of polysaccharides using intact wood cells. a, Representative structures of cellulose and major types of hemicellulose (xylan and galactoglucomannan). The cross-section of an elementary cellulose microfibril is shown, with a close view of an individual glucan chain and the hydroxymethyl conformation of surface (gauche- trans, gt) and internal (trans-gauche, tg) glucan chains. Xylan has two-fold or three-fold helical screw conformation in the backbone, with glucuronic acid (GlcA) or arabinose (Ara or A) sidechains. Acetyl motifs (Ac) are present in hemicelluloses. b, Representative 2D 13C J- INADEQUATE spectrum of eucalyptus, which is based on 13C cross polarization selecting rigid 177 Figure 7.1 (cont’d) molecules. Dash lines track the carbon connectivity of xylan conformers. c, A model of three elementary cellulose microfibrils fitting the NMR observables. The hydrophilic (sf) and hydrophobic (sg) surfaces, the embedded chains (ic), and the middle layer (ia,b) are color-coded. d, Carbon-4 regions measured using 13C direct polarization (orange spectra selecting mobile molecules) and cross polarization (grey spectra selecting rigid molecules). Both two- and three- fold xylan backbones are rigid while arabinose is mobile. e, Arabinose carbon-1 and carbon-2 regions. The only rigid arabinose is the terminal one in xylan sidechain. f, Eucalyptus has a high content of GlcA and spruce is rich in mannose (M). g, Composition of the rigid molecules in cell walls determined by peak volumes from 2D spectra. Source Data files are provided. The remarkable resolution is evidenced by the narrow 13C linewidths of 0.5-0.9 ppm, which allowed us to inspect the multifaceted polymorphism of polysaccharide structures. Locally, the structural variation of cellulose happens to the glucose hydroxymethyl conformations defined by the O5−C5−C6−O6 (χ) torsional angle (Figure 7.1a)27. The surface (s) and internal (i) chains primarily and respectively adopt gauche-trans (gt, χ=+60°) and trans-gauche (tg, χ=180°) conformations27, which resulted in well-resolved signals (Figure 7.1b). Upon bundling, the averaged structure of a fibril was restrained using the number of glucan chains residing in different environments, including the hydrophilic (sf) or hydrophobic (sg) surface, the inaccessible core (ic), and the middle layer (ia,b) sandwiched in between (Figure 7.1c)28. A satisfactory fit includes three elementary microfibrils, with 15, 12, 6, and 21 chains for the sf , sg , ic , and ia,b forms, respectively. Assuming uniform interfibrillar association, this averaged structure only designates the minimal bundle size, without considering loose packing. Other arrangements disagree with spectral observables (Figure S7.2). 178 Hemicellulose structure is highly complex as evidenced by the peak multiplicity. The backbones of hardwood xylan encompass two-fold (Xn2f), three-fold (Xn3f), and mixed (Xn) conformations (Figure 7.1d). The mixed form is absent in spruce; therefore, softwood xylan has higher homogeneity. Xylan was found to be rigid, suggestive of a close association with the stiff cellulose microfibrils. This observation differs from previous findings in Arabidopsis, where three- fold xylan remained mobile due to its spatial separation from cellulose23. Four types of arabinose (A) signals were identified, likely induced by the varied linkage sites at carbon 1, 2, 3, and 5 (Figure 7.1e). Only a single type is partially rigid, which is attributed to the terminal arabinose of xylan sidechains in secondary cell walls, whereas the mobile ones are from pectic polymers in primary walls. Hardwoods showed a high content of glucuronic (GlcA), while the softwood spruce exhibited unique signals of mannoses (M), one of the major constituents of galactoglucomannan (Figure 7.1f)29. Analysis of the spectral intensities led to the molar composition of rigid polysaccharides (Figure 7.1g). Around three quarters (74-81%) are cellulose, making it the most abundant polymer in woody stems (Table S7.3). Xylan makes up 18-26% of hardwood polysaccharides. Spruce has equal shares of xylan and mannan, each accounting for 10-12%. Consistent across these samples, the amount of two-fold xylan has doubled that of the three-fold counterpart, likely due to their promoted interactions with cellulose. Xylan sidechains are predominantly GlcA (therefore, glucuronoxylan) in hardwoods but mainly Ara (that is, arabinoxylan) in spruce. 7.3.2 Complex Structure and Linkage of Lignin Wood lignin mainly contains guaiacyl (G) and syringyl (S) residues, with a single and two methoxyl groups, respectively (Figure 7.2a, b). The unsubstituted ring, p-hydroxyphenyl (H), was not observed in solid-state NMR spectra due to its low abundance in these plants. The aromatic 179 signals are dispersed over an extensive range of chemical shifts. For instance, four types of S/S’ residues and four forms of G units were identified in eucalyptus and spruce, respectively. The multiplicity should be triggered by the varied oxidation states (for example, S’ is a Cα-oxidized form of S unit), the assorted linkages, and presumably, the wide-ranging conformation and packing in native solids. Aromatic rings are interconnected by a diverse array of covalent linkers. For example, β- ethers (A), with a characteristic one-bond cross peak between carbons β and γ at (85, 59) ppm (Figure 7.2a), have the β-O-4 linkages that are readily cleavable during degradation. Other signals emerged from resinol (B) that has β-β connections, phenylcoumaran (C) that encompasses β-5 and α-O-4 linkages and spirodienone (E) that contains β-1 and α-O-α links (Figure 7.2c). These signals are detectable by two experimental schemes based on 1H-13C cross polarization (CP) and 13C direct polarization (DP) methods, indicating the distribution of these linkers in the rigid and mobile phases of lignin (Figure S7.3). Despite the low intensity, we managed to detect several cross peaks between these linkers and aromatic carbons (Figure 7.2a). For example, the connection between the carbon α of β-ethers and the carbon 1 of S residues yielded a cross peak at (72, 135) ppm in eucalyptus and poplar. Similar correlations were also found between the carbon 1 of S residues and the carbon α of B and E linkers. In spruce, analogous connections were observed for G units. These junctions typically evade solution-NMR characterization because the aromatic carbon 1 of lignin is non-protonated. These results demonstrated the feasibility of using solid-state NMR to characterize lignin structure and linkage, but the technical capability still requires further development. Hardwood samples are rich in S residues (63-86 mol%) while the softwood only contains G units (Figure 7.2d). This observation agrees with our freshly collected solution NMR data of 180 ball-milled and dissolved lignocellulose (Figure S7.4 and Table S7.4), average S/G ratios found in the literature (Table S7.5) as well as the quantification using deconvoluted 1D solid-state NMR spectra (Figure S7.5 and Table S7.6). Both solid-state and solution NMR results (Figure 7.2e) suggest that these woody plants contain a large amount of β-ethers30. Solution NMR spectra also show considerable signals of both resinol and phenylcoumaran in poplar and spruce (Figure 7.2e), while the phenylcoumaran peaks become very weak in eucalyptus15, 30-36. Since these linkers respond differently to degradation, such analysis helps identify the plant candidates for saccharification. Figure 7.2 Chemical structure of lignin determined using solid-state and liquid NMR. Lignin signals are resolved using the aromatic regions of a, 2D 13C-13C RFDR spectra and b, 2D 13C J- INADEQUATE spectra of intact plant cells of eucalyptus, poplar, and spruce. c, The structure of 181 Figure 7.2 (cont’d) lignin units and linkages resolved in wood stems. The observed 13C chemical shifts are labeled for each carbon site. Key inter-residual linkages are highlighted. d, Composition of lignin units quantified using the integration of one-bond cross peaks in 2D 13C RFDR spectra (left panel) and using solution NMR as reported in the literature (right panel)30, 32, 33, 35, 37-39. Eucalyptus and poplar are S-rich while spruce is G-rich. Source Data files are provided. e, Molar fractions of the major linkage types quantified using the integration of one-bond cross peaks in 2D 13C RFDR spectra. e, Representative lignin linkages detected by 2D 1H-13C HSQC spectra of ball-milled biomass in d6- DMSO. 7.3.3 Distinct Patterns of Lignin-Carbohydrate Packing Across Plant Species The supramolecular architecture of lignocellulose differs dramatically in hardwood and softwood. In both eucalyptus and poplar, the use of a long-mixing period (1.0 s) in 2D 13C-13C correlation spectra has generated many long-range intermolecular cross peaks that are absent in the short-mixing (0.1 s) spectrum (Figure 7.3a and Figure S7.6). Puzzlingly, these two spectra showed a comparable pattern in spruce, which signified that 13C magnetization was already equilibrated among polymers within 0.1 s. This is not caused by variations of spin diffusion coefficients or polymer dynamics in hardwood and softwood as validated experimentally (Figure S7.7 and S7.8). Therefore, the rapid equilibrium observed in spruce indicates that lignin and polysaccharides are well-mixed on the nanoscale in spruce but stay apart in hardwoods. The spatial association of polymers was pinpointed by 272 intermolecular cross peaks, each of which represents a sub-nanometer, physical contact between two carbons from different molecules (Figure 7.3b and Figure S7.9). For example, the methyl carbon in the acetyl group of xylan (AcMe) exhibited cross peaks with the carbon 3 or 5 of S-lignin (S3/5) at (21, 153) ppm and 182 with the carbon 3 of G-lignin (G3) at (21, 148) ppm (Figure 7.3b). These interactions happen between i) acetyl groups and lignin, ii) xylan and lignin, iii) acetyl and cellulose, iv) acetyl and lignin methyl, v) lignin methoxyl group and cellulose, and vi) different lignin residues (Figure S7.10)21. Figure 7.3 DNP-assisted detection of long-range interactions defines lignin-carbohydrate packing. a, Overlay of long-range (1.0 s mixing; grey) and short-range (0.1 s mixing; orange) 2D 13C-13C correlation spectra. The long- and short-range spectra showed similar spectral patterns in spruce: softwood polymers are homogeneously mixed on the molecular level. b, Dipolar-gated 2D 13C- 13C correlation spectra showing 98 intermolecular cross peaks in eucalyptus. c, Summary of 272 intermolecular interactions identified in eucalyptus, poplar, and spruce. d, The count of 183 Figure 7.3 (cont’d) lignin-lignin, cellulose-lignin, and xylan-lignin interactions in eucalyptus, poplar, spruce, Arabidopsis, and grasses (maize, rice, and switchgrass). Lignin-cellulose interactions are scarce in Arabidopsis and grasses but become abundant in woods. e, DNP enhancing NMR sensitivity of eucalyptus by 24-fold. Inset shows a picture of the plant material in a sapphire DNP-NMR rotor. f, Comparison of the equilibrium spectrum (bottom) of eucalyptus that detects all components with the aromatic-edited spectrum (top) that only shows lignin-bound molecules. The lignin-bound polysaccharides include the two-fold, three-fold, and mixed conformers of xylan as well as the surface and interior cellulose. The deeply embedded glucan chains in cellulose (ic) are absent as highlighted using the dash line circle. g, DFT energy-minimized structures showing the possible packing between lignin units and polysaccharides. A statistical view of the number and intensities of intermolecular contacts (Tables S7.7- S7.10) designated xylan as the major interactor with lignin, which is manifested by the extensive correlations between G/S and xylan carbons, including the carbonyl and methyl carbons of acetyl (AcCO and AcMe) and xylose ring carbons (Figure 7.3c). This result echoes with the recent findings in Arabidopsis and commelinid monocots (grasses, for example, maize and switchgrass)24, validating the principal role of xylan-lignin interaction in stabilizing lignocellulose. Mannan has a small number of resolvable sites; therefore, it only showed a few cross peaks with lignin. However, the equilibrated pattern in Figure 7.3a supports a widespread association of mannan and lignin in spruce. Unexpectedly, wood lignin also exhibited extensive correlations with cellulose (Figure 7.3c), which accounts for 20-30% of all intermolecular cross peaks (Figure 7.3d). Such interactions are particularly abundant in spruce, consistent with the homogeneous mixing of 184 molecules in this sample. In contrast, cellulose-lignin contacts only account for 9-14% of polymer interactions in Arabidopsis and grasses24. Thus, cellulose-lignin interaction is a unique feature of woody plants, which should contribute, at least in part, to the mechanical strengths of their stiff stems. 7.3.4 Visualization of the Polysaccharide-Lignin Interface The unanticipated lignin-cellulose interactions were verified using the sensitivity- enhancing DNP technique40-42. The eucalyptus stem showed a 24-fold enhancement of NMR intensity (Figure 7.3e), which was achieved by transferring the polarization of electrons in the stable biradical AMUPol to the protons of biopolymers. This sensitivity enhancement shortens the measurements by 576 times (a 1.5-year experiment in a day), allowing us to take snapshots of the lowly populated boundaries between polysaccharides and lignin. Eucalyptus cellulose was found in the lignin-bound part of polysaccharides (Figure 7.3f), but such signals were missing in maize24. Only the surface glucan chains of cellulose (s) or those chains right underneath the surface layer (ia,b) correlated with lignin, while the embedded core (ic) lacked such interaction. The three-fold xylan (Xn3f) is still favorable for binding lignin, showing stronger signals than the two-fold (Xn2f) and the mixed (Xn) conformers. Favorable lignin-polysaccharide interactions were revealed using quantum mechanical geometry optimizations conducted using the DFT method in continuum solvation models (Figure 7.3g). The three-fold xylan formed a packet to enclose aromatics. Conversely, the flat chains from cellulose surface or two-fold xylan relied on their pyranose/furanose rings to align with an aromatic unit. Examination of recently reported DFT structures43 showed that S-units preferentially aligned to carbohydrate rings as stabilized by the two methoxyl groups on both sides of the aromatic ring (Figure S7.11). The G unit, on the other hand, typically has its single methoxyl 185 group closer to the carbohydrates. Such orientational preference supports the strong interactions experimentally observed between lignin methoxyl groups and xylan acetyls (Figure S7.10). Consequently, the methoxyl-rich S-residues correlate better with xylan in space. Therefore, non- covalent interactions between these polar groups are essential to the existence of lignin-xylan complex. 7.3.5 Landscape of Biomolecular Hydration and Dynamics Biopolymers have sophisticated dynamics and variable water-association in their native environments. Among the three plants, poplar turned out to be the worst hydrated sample as shown by its slowest water-to-polymer 1H polarization transfer buildup curves, which were consistent for both cellulose and xylan (i4 and Xn3f1, Figure 7.4a), as well as lignin (Figure S7.12). The relative intensities (S/S0) of a hydration heatmap reflect the degree of water association (Figure 7.4b and Figure S7.13). Plots of the S/S0 ratios against 127 carbon sites show that the hydration level increases sequentially from poplar to eucalyptus and spruce (Figure 7.4c). Within each sample, polymer hydration generally increases from cellulose to xylan, and then to mannan, if present. Surprisingly, wood lignin retained high S/S0 ratios, typically on the same range as those of xylan, indicating a high level of hydration (Figure 7.4c). This is intriguing because these aromatic polymers are perceived as relatively hydrophobic, which by expectation, should hinder water retention. For example, maize lignin was previously found to self-aggregate, forming nanodomains that are largely separated from water24. The water contact observed here should originate from lignin’s tight association with those carbohydrates that kept water localized. Likewise, wood biopolymers should experience frequent entanglement and interpenetration instead of domain separation. 186 Figure 7.4 Site-specific hydration and dynamic landscape of biomolecules. a, Representative water-to-biomolecule 1H polarization transfer buildup curves. The data of interior cellulose carbon 4 (i4) and three-fold xylan carbon 1 (Xn3f1) are compared across plants. Poplar has slow buildup due to the limited water contact. b, Hydration map on top of a 2D spectrum plotting the ratio (S/S0) of water-edited intensity (S) to the equilibrium intensity (S0). A larger S/S0 ratio indicates better polymer hydration. c, Distribution of the relative water-edited intensities (S/S0) of polysaccharides and lignin in eucalyptus, poplar, and spruce. Molecules with better water retention show higher water-edited intensities. Poplar is poorly hydrated. d, 13C-T1 relaxation times of rigid (blue rectangles) and all molecules (colored as yellow, red, purple, and magenta for different molecules) in three woody plants, which represent nanosecond timescale motions. e, 1H-T1ρ relaxation times reflecting microsecond timescale dynamics. The average value and standard deviation (error bars) are presented for each violin plot in panels c-e, the dataset of which are tabulated in Tables S7.11- 7.15. Source Data files are provided. 187 Polymer dynamics were examined using 13C-T1 and 1H-T1ρ relaxation measurements, which generated 150 relaxation curves (Figure S7.14 and S7.15). When all molecules were considered, the 13C-T1 relaxation time decreased in the order of cellulose, lignin, xylan, and mannan, if any (Figure 7.4d). The short 13C-T1 time constants of lignin and hemicellulose revealed the efficient 13C-T1 relaxation in these non-cellulosic polymers, and furthermore, their enhanced motion on the nanosecond timescale. Such differences became indistinguishable if only rigid molecules were focused on, for example, all rigid molecules of eucalyptus showed 13C-T1 time constants of 4-5 s. This observation contradicts previous results in which maize lignin showed substantially longer 13C-T1 than any polysaccharide24. This can be comprehended using the efficient spin-exchange between lignin and carbohydrates (mediated by the better molecular mixing), which has averaged the 13C-T1 relaxation times in wood. On the microsecond timescale, cellulose had the longest 1H-T1ρ relaxation times (30-40 ms) due to the restricted motion in the massively hydrogen-bonded microfibrils (Figure 7.4e). Both lignin and hemicellulose had short 1H-T1ρ times of 10-15 ms, like the counterparts in maize24. Compared to G residues, S units exhibited slower 1H-T1ρ relaxation, indicating attenuated dynamics due to interactions with polysaccharides (Figure S7.15). Among the three samples, spruce displayed the shortest 13C-T1 and 1H-T1ρ relaxation times, revealing a unique profile of molecular dynamics in softwood. 7.3.6 Effects of Rehydration on Water-Retention and Polymer Dynamics Water molecules are important for stabilizing the nanostructure of cell wall biomaterials, and the complete removal of water could potentially cause irreversible changes. In a recent NMR study of the softwood Pinus radiata, oven-drying (for complete dehydration) and rehydration were found to promote polymer association (e.g. xylan-cellulose and lignin-cellulose packing) and irreversibly alter the dynamics and conformation of mannan in the secondary cell walls44. Similarly, 188 the lyophilized-rehydrated grass sorghum (Sorghum bicolor L. Moench) has shown permanent changes of hemicellulose, with enhanced mobility for the three-fold xylan backbone and some arabinose residues45. Differently, dehydration only caused reversible structural changes to the primary cell walls of Arabidopsis, likely due to the high content of pectic polymers that can efficiently associate with water molecules46. We measured the never-dried samples of spruce and eucalyptus, followed by re- examination of the same materials after lyophilization and rehydration. In spruce, the lignocellulosic complex was not dramatically affected by the rehydration process, which differs from lipids and proteins (Figure 7.5a). The structure of the carbohydrates and lignin, as well as their distribution in dynamically distinct domains, can be efficiently retained in the rehydrated biomass, occasionally with minor intensity variations that are well below 10% for each carbohydrate or lignin carbon site (Figure 7.5a). The retention of water molecules can also be fully restored, as shown by the identical spectral patterns in 1D water-edited spectra (Figure 7.5b) and the comparable 2D heatmap representation for both carbohydrates and aromatics in the never- dried and rehydrated samples (Figure 7.5c). The nanosecond-timescale motion and the subnanometer polymer packing (reflected by spin-exchange) are unaffected since there is no deviation of 13C-T1 relaxation within the well-controlled error margin (Figure 7.5d). Both lignin and carbohydrates showed a 10-20% increase in the 1H-T1ρ relaxation time constants; therefore, the only notable change happened to the microsecond scale motions of biopolymers (Figure 7.5e). Eucalyptus behaves differently: this plant fully restored the structural and dynamical features but moderately improved polymer-water association after rehydration (Figure S7.16). 189 Figure 7.5 Molecular-level dynamics and water association after dehydration and rehydration. A never-dried spruce sample was first measured, and then freeze-dried and rehydrated for direct comparison. a, Overlay of spectra collected using never-dried (black) and lyophilized and then rehydrated (orange) spruce stem samples showing high similarity for polysaccharides and lignin. From top to bottom are the CP, quantitative DP, and short recycle-delay DP spectra. The key regions of lignin, carbohydrate, and lipid are marked. b, 1D water-to-carbohydrate polarization transfer spectra revealing similar water association for never-dried and rehydrated samples. The well-hydrated molecules are selected using a short 1H mixing time of 4 ms while the equilibrated state is measured using 36 ms 1H mixing. c, Water-edited 2D hydration map showing comparable water-edited intensities for both lignin and carbohydrate regions in the never-dried and the rehydrated samples. d, Average 13C-T1 relaxation plot, using the integrals of spectral regions for carbohydrates (left) and lignin (right). e, Average T1ρ relaxation of carbohydrates and lignin in 190 Figure 7.5 (cont’d) never-dried (black) and rehydrated (orange) samples. Error bars are standard deviations of the fitting derived parameters. 7.4 Discussion The abundant molecular-level evidence presented four novel features of lignocellulosic materials, each exploring a structural or chemical foundation of the supramolecular architecture (Figure 7.6). First, even though three-fold xylan is favored for binding lignin, in wood, other xylan conformers can also coexist with lignin in part. Accompanying with this functional resemblance is the hydration equivalence of these xylan forms (Figure 7.4c). These results have extended the conceptual model of lignocellulose derived from maize, in which lignin mainly interacts with the three-fold xylan, and vice versa. Actually, interactions between three-fold xylan and cellulose have been reported in a grass Sorghum45, which also implies the interchangeable roles of xylan conformers. Uniquely, wood xylan is mainly in two-fold helical screw, which is energetically unfavorable unless forced by the binding to cellulose surface. Therefore, the unanticipated proximity of lignin to the two-fold xylan, as well as its associated cellulose surface, might be a destined consequence of spatial crowding in densely packed lignocellulosic materials. Second, the packing between cellulose and lignin is plant-specific and only serves as a secondary interaction. In Arabidopsis and grasses, lignin and cellulose are spatially separated24. The situation has changed for woody biomass, where lignin and cellulose become colocalized. Although Arabidopsis is widely used as a model system for investigating the biophysical traits of hardwoods, it becomes apparent that these plants differ in their cell wall organization. Our DFT results and a recent molecular simulation study47 consistently suggest that lignin is mainly docked on the hydrophobic surface of cellulose, with aligned phenyl and pyranose rings. The increased 191 coverage of cellulose surface by aromatic polymers might reduce enzyme accessibility and contribute to biomass recalcitrance. Third, the structural feature of aromatic polymer needs reconsideration. Lignin nanodomains observed in maize are absent in woody plants due to the promoted polymer mixing on the nanoscale (Figure 7.3). Accordingly, polymer entanglements and inter-penetration, rather than superficial contact between domain surfaces, should govern lignin-carbohydrate interactions in wood48. Once well-mixed with polysaccharides, even the aromatics could effectively retain and immobilize water molecules (Figure 7.4c). The thermodynamical driving force of this peculiar phenomenon awaits inputs from modeling methods. Figure 7.6 Comparative schemes of secondary cell walls in grass, hardwood, and softwood. The figure shows the spatial arrangement of lignin (yellow), cellulose (white fibrils), two-fold xylan (red flat ribbons), three-fold xylan (blue twisted ribbons), mixed forms of xylan (magenta), and mannan (GGM; green) in secondary plant cell walls. Numbered spheres highlight the structural features of (1) lignin-xylan interaction in all plants, (2) cellulose bundles in woody plants, 192 Figure 7.6 (cont’d) (3-6) twofold and threefold xylan with different sidechains, and (7) galactoglucomannan. The molecular fraction of polysaccharides is considered in the depiction, but the illustration may not be strictly to scale. The model of grass cell wall is generated using the data recently reported24, for comparison with the models of woody plants. Lignin and carbohydrates are much better mixed in woody plants than in grass, resulting in the binding of lignin to both three-fold and two-fold xylan as well as to cellulose microfibrils. The structural assembly of woody cell walls is thus different from the domain-separation scheme of grass cell walls. Finally, the ultrastructure of softwood is featured by homogeneous molecular mixing (Figure 7.3a). This structural hallmark has revised the prevailing model in which well-aligned and partially crystalline glucomannans were perceived as being sandwiched between an internal cylinder of cellulose microfibrils and an outside tubular domain of lignin, with xylan forming an outermost layer49. Actually, our results better align with a recent study reporting that glucomannan and xylan could potentially associate with both cellulose and lignin29. Although the focus of this exploratory study is to understand the chemical and physical principles underlying polymer packing and lignocellulose architecture, the structural foundation and methodology established here would inspire more in-depth investigations for understanding the structural diversity and mechanical properties presented by numerous plant species and mutants, various cell types, and different growth stages50. These structural insights will guide the utilization of forestry resources for the production of biomaterials and biofuels51, and the spectroscopic toolbox will stimulate structural investigations of polymer assemblies in other organisms, such as bacteria, fungi, and algae, as well as bio-inspired materials52-60. 193 7.5 Methods 7.5.1 Plant Material. Uniformly 13C-enriched stems (97% 13C) of eucalyptus (Eucalyptus grandis; age 16 weeks), poplar wood (Populus x canadensis; 27 weeks), and spruce (Picea abies; 16 weeks) were obtained for structural characterization from IsoLife (The Netherlands). Eucalyptus and spruce were obtained from seeds while poplar was obtained from stem cutting. Poplar plants were saplings of close to 1 m height at harvest. Immediately after removing the plants from the growth chamber, plant shoots were dissected into leaves and stems. The stem-sections were split, cut, and debarked after freeze-drying. Debarking was strived to be complete, which was conducted through the longitudinal cutting of the bark followed by separation from the xylem at the cambium interface. Debarked stems were hydrated for NMR analysis. The material was briefly hand-grinded using a pestle and a mortar, resulting in small pieces on the range of a few mm across. The homogeneity helps to avoid potential issues during magic-angle spinning. The technical details of isotope labeling are presented in the Supplementary Methods. To test the effect of freeze-drying, we obtained never-dried 13C-labeled spruce and eucalyptus wood from IsoLife, without lyophilization. The debarking process is finished manually by carefully removing the outside soft bark layer with tweezers. Around 104 mg of debarked spruce was cut into mm scale pieces, packed into 4-mm rotor, and measured under 10 kHz MAS on a 400 MHz Bruker NMR spectrometer. The never-dried samples were first measured freshly. After ssNMR measurements, overnight lyophilization was conducted, giving 43.5 mg of freeze- dried spruce stem. The sample is then sufficiently rehydrated (with a final weight of 107 mg), and all ssNMR experiments were conducted again on the rehydrated material for direct comparison with never-dried samples (Figure 7.5). Besides, never-dried (96 mg) eucalyptus stem, and the 194 dehydrated and then rehydrated material (94mg including 68 wt% H2O) were also examined (Figure S7.16). 7.5.2 Solid-State NMR Experiments for Assignment. The plant stems were packed into MAS rotors for measurements on 600 MHz and 400 MHz Bruker spectrometers using 3.2-mm and 4-mm MAS probes, respectively. Most experiments were collected under 10-14 kHz MAS at 293 K unless otherwise stated. 13C chemical shifts were referenced to the tetramethylsilane (TMS) scale. Radiofrequency field strengths were 80-100 kHz for 1H decoupling, 62.5 kHz for 1H CP contact pulse, and 50–62.5 kHz for 13C. The technical parameters of all solid-state NMR experiments were summarized in Tables S7.16 and S7.17 for the initial batch of rehydrated samples, and in Table S7.18 for the more recently obtained never- dried and the subsequent freeze-dried and rehydrated samples. To assign NMR signals, 2D correlation spectra were recorded using the refocused CP J- INADEQUATE sequence61, which was coupled with 13C DP for detecting mobile molecules or 1H-13C CP for detecting rigid polymers. A 2D 13C radio frequency-driven recoupling (RFDR) correlation experiment was measured under 10 kHz MAS and 83 kHz 1H decoupling to assign intra-residue cross peaks62. A recoupling time of 1.6 ms was used to detect one-bond 13C−13C cross peaks. An additional set of 13C-DP PDSD experiments were conducted to examine the structure of lignin in the mobile phase, which were selected through the short recycle delay of 2 s. The standard flow chart of resonance assignment is presented in Figure S7.17. All chemical shifts are validated by comparison with literature-reported values (Table S7.19). To enhance aromatic signals, we measured a dipolar-coupling-gated version of the proton- driven spin diffusion (PDSD) experiment24, 63, 64. A dipolar-dephasing period of 48 µs was employed to efficiently suppress the signals of protonated carbons, allowing the preferential 195 detection of non-protonated carbons in the aromatic lignin. This period was asymmetric with respect to the π pulse in the Hahn echo, containing two undecoupled delays of 32 µs and 16 µs. A 0.1-s PDSD mixing was used to report intramolecular cross peaks in all samples. A 20-ms mixing was also used for spruce. The observed chemical shifts were compared with values deposited in the Complex Carbohydrate Magnetic Resonance Database to facilitate resonance assignment65. 7.5.3 Solid-State NMR Experiments for Structural Analysis. The 2D gated PDSD experiment was also conducted to determine lignin-carbohydrate packing, with a long mixing time (1 s) for intermolecular polarization transfer. We have identified 98, 80, and 94 cross peaks for eucalyptus, poplar, and spruce, respectively. These 272 correlations were categorized as 112 strong, 75 medium, and 85 weak restraints according to their relative intensities (area of a peak relative to that of the 1D cross-section), with cutoffs set to >4.0% (strong), 2.0–4.0% (medium), and <2.0 (weak) as tabulated in Tables S7.7-S7.10. Polymer hydration was determined using water-edited 2D 13C−13C correlation experiment at 277 K (Figure 7.4 a, b)66, 67, which generated 44, 48, and 37 datasets for eucalyptus, poplar, and spruce, respectively. This experiment used a 1H-T2 relaxation filter of 120 µs × 2 to suppress the polysaccharide signals to <5% and retain >85% of water magnetization. Water-polarization was transferred to spatially proximal polymers using a 4-ms 1H mixing period, followed by a 1-ms CP for 13C detection. A 100-ms DARR mixing was used for both water-edited and control spectra. 1D buildup curves were measured at 277 K using a 1H-T2 filter of 120 µs × 2 and a 1H mixing of 0-81 ms for spruce, 0-121 ms for eucalyptus, and 0- 169 ms for poplar. To probe dynamics, we measured 13C spin-lattice relaxation (T1) and 1H rotating frame spin-lattice relaxation (T1ρ). 13C-T1 was measured using Torchia-CP 68 and standard 13C DP inversion recovery schemes. Torchia-CP preferentially detected rigid molecules, with a z-filter of 196 0-12 s. A 30-s recycle delay was used in the DP inversion recovery experiment for quantitatively detecting all molecules, with a z-period of 0-35 s. 1H–T1ρ was measured using an effective 1H spin- lock field of 62.5 kHz with a duration of 0-19 ms. Relaxation data were fit using single exponential functions (Tables S7.14 and S7.15). Also, the dipolar-chemical shift (DIPSHIFT) experiment69 was conducted under 7.5 kHz MAS, to report dipolar order parameters of biopolymers in three wood samples. Frequency-Switched Lee Goldburg (FSLG) 1H homonuclear decoupling sequence was utilized, the scaling factor was verified to be 0.577. 7.5.4 Solution NMR Sample Preparation and Experiments. To validate the lignin assignment obtained using the solid-state method, we conducted solution NMR analysis, which has well-documented chemical shifts of lignin polymers. A summary of the lignin chemical shifts is included in Table S7.4. Wood samples were ball-milled for 1 h with a motor running at 1750 rpm. About 30 mg of each powdered wood sample was dissolved in 10 mL DMSO-d6 with constant stirring using a magnetic stirrer for 2 h at 60°C for solution NMR experiments. To identify linkages in lignin, we conducted 2D 1H-13C HSQC spectra conducted on a Bruker Avance III 500 MHz 1H Larmor frequency and equipped with a 5 mm z- gradient Prodigy TCI probe. Both the 1H and the 13C chemical shifts gave indications of the linkage types using the chemical shifts reported previously39, 70-74. 7.5.5 MAS-DNP Experiment. About 30-35 mg of eucalyptus stems were impregnated into 50 µL DNP matrix solution containing 10 mM AMUPol75 radical in d8-glycerol/D2O/H2O (60/30/10 Vol%). The material was softly ground for 10-15 min to allow radical distribution and packed into a 3.2-mm sapphire rotor. The NMR sensitivity was enhanced by 24-fold with microwave irradiation on a 600 MHz/395 GHz MAS-DNP spectrometer under 10 kHz MAS. DNP analysis has been applied to various 197 carbohydrate and plant systems using different protocols76-78. Recently, we have demonstrated that the radical can effectively and homogeneously polarize all molecules in cell wall materials79. Lignin had a slightly shorter DNP buildup time (1.8 s) compared with polysaccharides (2.2 s). The recycle delay was set to 2.3 s to facilitate the selection of lignin against carbohydrates in the 2D experiments. A lignin-edited 2D 13C-13C experiment was measured to detect the signals of lignin- bound carbohydrates, with a 0.5-s PDSD mixing to transfer lignin polarization to carbohydrates, followed by a 20-ms PDSD mixing for 2D 13C-13C correlation24. For comparison, a 20-ms PDSD spectrum was measured as a control. 7.5.6 DFT Calculation. DFT calculations were used to investigate the interactions between the S or G units in lignin and cellulose, or xylan with two- and three-fold conformations. Tetramers of β-1,4-linked glucose and xylose were built to represent cellulose and two-fold xylan, respectively. Hexamers of β-1,4-linked xylose were built to represent three-fold xylan, based on a crystal structure of three- fold xylopentaose (PDB ID: 1GNY).80 Six models were constructed: G unit-cellulose, G unit- xylan (two-fold), G unit-xylan (three-fold), S unit-cellulose, S unit-xylan (two-fold), and S unit- xylan (three-fold). Previously, it has been shown that for lignin-cellulose and lignin-xylan interactions, the stacked configurations tend to have the largest interaction energies43. Therefore, the initial starting configurations of the six models were constructed by positioning the G/S units parallel to the surface of the glucose/xylose residues, ~4 Å away. The geometry optimizations were conducted using the DFT method, M05-2X/6-31+G*81 with the integral equation formalism for the polarizable continuum model (IEFPCM)82 solvation model in Gaussian 0983. Although other methods may be available, this exchange-correlation functional with dispersion corrections performs reasonably well for carbohydrate and carbohydrate-aromatic interactions11, 84, 85. 198 7.6 Acknowledgments The solid-state and solution NMR analyses of wood cell walls were conducted by AK, XK, and WZ, who were supported by the U.S. Department of Energy (grant no. DE-SC0021210). DFT calculation and compositional analysis conducted by HY and FD were supported as part of the Center for Lignocellulose Structure and Formation, an Energy Frontier Research Center funded by the US Department of Energy, Office of Science, Basic Energy Sciences under award no. DE- SC0001090. The National High Magnetic Field Laboratory is supported by National Science Foundation through NSF/DMR-1644779 and the State of Florida. The MAS-DNP system at NHMFL is funded in part by NIH S10 OD018519 and NSF CHE-1229170. The authors thank Dr. Daniel Cosgrove for the helpful discussion. 199 REFERENCES 1. Bar-On, Y. M.; Phillips, R.; Milo, R., The biomass distribution on Earth. P Natl Acad Sci USA 2018, 115 (25), 6506-6511. 2. Somerville, C.; Youngs, H.; Taylor, C.; Davis, S. C.; Long, S. P., Feedstocks for Lignocellulosic Biofuels. Science 2010, 329 (5993), 790-792. 3. Moon, R. J.; Martini, A.; Nairn, J.; Simonsen, J.; Youngblood, J., Cellulose nanomaterials review: structure, properties and nanocomposites. Chem Soc Rev 2011, 40 (7), 3941-3994. 4. Himmel, M. E.; Ding, S. Y.; Johnson, D. K.; Adney, W. S.; Nimlos, M. R.; Brady, J. W.; Foust, T. D., Biomass Recalcitrance: Engineering Plants and Enzymes for Biofuels Production. Science 2007, 315, 804-805. 5. McCann, M. C.; Carpita, N. C., Biomass recalcitrance: a multi-scale, multi-factor, and conversion-specific property. J. Exp. Bot. 2015, 66 (14), 4109-18. 6. Loque, D.; Scheller, H. V.; Pauly, M., Engineering of plant cell walls for enhanced biofuel production. Curr Opin Plant Biol 2015, 25, 151-161. 7. Ragauskas, A. J.; Beckham, G. T.; Biddy, M. J.; Chandra, R.; Chen, F.; Davis, M. F.; Davison, B. H.; Dixon, R. A.; Gilna, P.; Keller, M.; Langan, P.; Naskar, A. K.; Saddler, J. N.; Tschaplinski, T. J.; Tuskan, G. A.; Wyman, C. E., Lignin Valorization: Improving Lignin Processing in the Biorefinery. Science 2014, 344 (6185), 709-+. 8. Cai, Y.; Zhang, K.; Kim, H.; Hou, G.; Zhang, X.; Yang, H.; Feng, H.; Miller, L.; Ralph, J.; Liu, C. J., Enhancing digestibility and ethanol yield of Populus wood via expression of an engineered monolignol 4-O-methyltransferase. Nature communications 2016, 7, 11989. 9. Purushotham, P.; Ho, R.; Zimmer, J., Architecture of a catalytically active homotrimeric plant cellulose synthase complex. Science 2020, 369 (6507), 1089-1094. 10. Jarvis, M., Chemistry: cellulose stacks up. Nature 2003, 426 (6967), 611-612. 11. Yang, H.; Kubicki, J., A density functional theory study on the shape of the primary cellulose microfibril in plants: effects of C6 exocyclic group conformation and H-bonding. Cellulose 2020, 27, 2389-2402. 12. Fernandes, A. N.; Thomas, L. H.; Altaner, C. M.; Callow, P.; Forsyth, V. T.; Apperley, D. C.; Kennedy, C. J.; Jarvis, M. C., Nanostructure of cellulose microfibrils in spruce wood. P Natl Acad Sci USA 2011, 108 (47), E1195-E1203. 13. Thomas, L. H.; Altaner, C. M.; Forsyth, V. T.; Mossou, E.; Kennedy, C. J.; Martel, A.; Jarvis, M. C., Nanostructural deformation of high-stiffness spruce wood under tension. Scientific reports 2021, 11 (1), 453. 14. Vanholme, R.; Demedts, B.; Morreel, K.; Ralph, J.; Boerjan, W., Lignin biosynthesis and 200 structure. Plant Physiol. 2010, 153 (3), 895-905. 15. Wang, H. M.; Ma, C. Y.; Li, H. Y.; Chen, T. Y.; Wen, J. L.; Cao, X. F.; Wang, X. L.; Yuan, T.-Q.; Sun, R.-C., Structural Variations of Lignin Macromolecules from Early Growth Stages of Poplar Cell Walls. ACS Sustainable Chem. Eng. 2020, 8 (4), 1813-1822. 16. Cosgrove, D. J.; Jarvis, M. C., Comparative structure and biomechanics of plant primary and secondary cell walls. Front Plant Sci 2012, 3. 17. Du, X.; Perez-Boada, M.; Fernandez, C.; Rencoret, J.; del Rio, J. C.; Jimenez-Barbero, J.; Li, J.; Gutierrez, A.; Martinez, A. T., Analysis of lignin-carbohydrate and lignin-lignin linkages after hydrolase treatment of xylan-lignin, glucomannan-lignin and glucan-lignin complexes from spruce wood. Planta 2014, 239, 1079-1090. 18. Giummarella, N.; Pu, Y.; Ragauskas, A. J.; Lawoko, M., A critical review on the analysis of lignin carbohydrate bonds. Green Chem. 2019, 21, 1573-1595. 19. Zhang, T.; Vavylonis, D.; Durachko, D. M.; Cosgrove, D. J., Nanoscale movements of cellulose microfibrils in primary cell walls. Nat Plants 2017, 3 (5). 20. Reif, B.; Ashbrook, S. E.; Emsley, L.; Hong, M., Solid-State NMR Spectroscopy. Nat. Rev. Methods Primers 2021, 1, 2. 21. Zhao, W.; Fernando, L. D.; Kirui, A.; Deligey, F.; Wang, T., Solid-state NMR of plant and fungal cell walls: A critical review. Solid state nuclear magnetic resonance 2020, 107, 101660. 22. Grantham, N. J.; Wurman-Rodrich, J.; Terrett, O. M.; Lyczakowski, J. J.; Stott, K.; Iuga, D.; Simmons, T. J.; Durand-Tardif, M.; Brown, S. P.; Dupree, R.; Busse-Wicher, M.; Dupree, P., An even pattern of xylan substitution is critical for interaction with cellulose in plant cell walls. Nat Plants 2017, 3 (11), 859-865. 23. Simmons, T. J.; Mortimer, J. C.; Bernardinelli, O. D.; Poppler, A. C.; Brown, S. P.; deAzevedo, E. R.; Dupree, R.; Dupree, P., Folding of xylan onto cellulose fibrils in plant cell walls revealed by solid-state NMR. Nature communications 2016, 7, 13902. 24. Kang, X.; Kirui, A.; Dickwella Widanage, M. C.; Mentink-Vigier, F.; Cosgrove, D. J.; Wang, T., Lignin-polysaccharide interactions in plant secondary cell walls revealed by solid-state NMR. Nature communications 2019, 10 (1), 347. 25. Dupree, R.; Simmons, T. J.; Mortimer, J. C.; Patel, D.; Iuga, D.; Brown, S. P.; Dupree, P., Probing the Molecular Architecture of Arabidopsis thaliana Secondary Cell Walls Using Two- and Three-Dimensional 13C Solid State Nuclear Magnetic Resonance Spectroscopy. Biochemistry-Us 2015, 54 (14), 2335-2345. 26. Simmons, B. A.; Loque, D.; Blanch, H. W., Next-generation biomass feedstocks for biofuel production. Genome biology 2008, 9 (12), 242. 27. Phyo, P.; Wang, T.; Yang, Y.; O'Neill, H.; Hong, M., Direct Determination of 201 Hydroxymethyl Conformations of Plant Cell Wall Cellulose Using 1H Polarization Transfer Solid- State NMR. Biomacromolecules 2018, 19 (5), 1485-1497. 28. Wang, T.; Yang, H.; Kubicki, J. D.; Hong, M., Cellulose Structural Polymorphism in Plant Primary Cell Walls Investigated by High-Field 2D Solid-State NMR Spectroscopy and Density Functional Theory Calculations. Biomacromolecules 2016, 17 (6), 2210-2222. 29. Terrett, O. M.; Lyczakowski, J. J.; Yu, L.; Iuga, D.; Franks, W. T.; Brown, S. P.; Dupree, R.; Dupree, P., Molecular architecture of softwood revealed by solid-state NMR. Nature communications 2019, 10 (1), 4978. 30. Rencoret, J.; Gutierrez, A.; Nieto, L.; Jimenez-Barbero, J.; Faulds, C. B.; Kim, H.; Ralph, J.; Martinez, A. T.; Del Rio, J. C., Lignin Composition and Structure in Young versus Adult Eucalyptus globulus Plants. Plant Physiol. 2011, 155, 667-682. 31. Rinaldi, R.; Jastrzebski, R.; Clough, M. T.; Ralph, J.; Kennema, M.; Bruijnincx, P. C. A.; Weckhuysen, B. M., Paving the Way for Lignin Valorisation: Recent Advances in Bioengineering, Biorefining and Catalysis. Angew Chem Int Edit 2016, 55 (29), 8164-8215. 32. Rico, A.; Rencoret, J.; del Rio, J. C.; Martinez, A. T.; Gutierrez, A., In-Depth 2D NMR Study of Lignin Modification During Pretreatment of Eucalyptus Wood with Laccase and Mediators. Bioenergy Res. 2015, 8, 211-230. 33. Ibarra, D.; Chavez, M. I.; Rencoret, J.; Del Rio, J. C.; Gutierrez, A.; Romero, J.; Camarero, S.; Martinez, M. J.; Jimenez-Barbero, J.; Martinez, A. T., Structural modification of eucalypt pulp lignin in a totally chlorine-free bleaching sequence including a laccase-mediator stage. Holzforschung 2007, 61, 634-646. 34. Li, M.; Cao, S.; Meng, X.; Studer, M.; Wyman, C. E.; Ragauskas, A. J.; Pu, Y., The effect of liquid hot water pretreatment on the chemical–structural alteration and the reduced recalcitrance in poplar. . Biotechnology for biofuels 2017, 10, 237. 35. Samuel, R.; Foston, M.; Jaing, N.; Cao, S.; Allison, L.; Studer, M.; Wyman, C.; Ragauskas, A. J., HSQC (heteronuclear single quantum coherence) 13C–1H correlation spectra of whole biomass in perdeuterated pyridinium chloride–DMSO system: An effective tool for evaluating Fuel 2011, 90 (9), 2836-2842. 36. Samuel, R.; Cao, S.; Das, B. K.; Hu, F.; Pu, Y.; Ragauskas, A. J., Investigation of the fate of poplar lignin during autohydrolysis pretreatment to understand the biomass recalcitrance. RSC Adv. 2013, 3 (16), 5305-5309. 37. Rencoret, J.; Marques, G.; Gutierrez, A.; Nieto, L.; Santos, J. I.; Jimenez-Barbero, J.; Martinez, A. T.; del Rio, J. C., HSQC-NMR analysis of lignin in woody (Eucalyptus globulus and Picea abies) and non-woody (Agave sisalana) ball-milled plant materials at the gel state 10th EWLP, Stockholm, Sweden, August 25–28, 2008. Holzforschung 2009, 63, 691-698. 38. Wen, J. L.; Sun, S. L.; Yuan, T. Q.; Xu, F.; Sun, R. C., Structural elucidation of lignin polymers of Eucalyptus chips during organosolv pretreatment and extended delignification. 202 Journal of agricultural and food chemistry 2013, 61 (46), 11067-11075. 39. Yuan, T.-Q.; Sun, S.-N.; Xu, F.; Sun, R.-C., Characterization of Lignin Structures and Lignin– Carbohydrate Complex (LCC) Linkages by Quantitative 13C and 2D HSQC NMR Spectroscopy. Journal of agricultural and food chemistry 2011, 59, 10604-10614. 40. Su, Y.; Andreas, L.; Griffin, R. G., Magic angle spinning NMR of proteins: high-frequency dynamic nuclear polarization and 1H detection. Annu. Rev. Biochem. 2015, 84, 465-97. 41. Mentink-Vigier, F.; Akbey, U.; Oschkinat, H.; Vega, S.; Feintuch, A., Theoretical aspects of Magic Angle Spinning - Dynamic Nuclear Polarization. J Magn Reson 2015, 258, 102-120. 42. Rossini, A. J.; Zagdoun, A.; Lelli, M.; Lesage, A.; Coperet, C.; Emsley, L., Dynamic nuclear polarization surface enhanced NMR spectroscopy. Accounts of chemical research 2013, 46 (9), 1942-51. 43. Yang, H.; Watts, H. D.; Gibilterra, V.; Weiss, T. B.; Petridis, L.; Cosgrove, D. J.; Kubicki, J. D., Quantum Calculations on Plant Cell Wall Component Interactions. Interdisciplinary sciences, computational life sciences 2019, 11 (3), 485-495. 44. Cresswell, R.; Dupree, R.; Brown, S. P.; Pereira, C. S.; Skaf, M. S.; Sorieul, M.; Dupree, P.; Hill, S., Importance of Water in Maintaining Softwood Secondary Cell Wall Nanostructure. Biomacromolecules 2021, in press. 45. Gao, Y.; Lipton, A. S.; Wittmer, Y.; Murray, D. T.; Mortimer, J. C., A grass-specific cellulose-xylan interaction dominates in sorghum secondary cell walls. Nature communications 2020, 11 (1), 6081. 46. Wang, T.; Park, Y. B.; Cosgrove, D. J.; Hong, M., Cellulose-Pectin Spatial Contacts Are Inherent to Never-Dried Arabidopsis thaliana Primary Cell Walls: Evidence from Solid-State NMR. Plant Physiol. 2015, 168, 871-884. 47. Vermaas, J. V.; Crowley, M. F.; Beckham, G. T., A Quantitative Molecular Atlas for Interactions Between Lignin and Cellulose. ACS Sustainable Chem. Eng. 2019, 7 (24), 19570- 19583. 48. Langan, P.; Petridis, L.; O'Neill, H. M.; Pingali, S. V.; Foston, M.; Nishiyama, Y.; Schulz, R.; Lindner, B.; Hanson, B. L.; Harton, S.; Heller, W. T.; Urban, V.; Evans, B. R.; Gnanakaran, S.; Ragauskas, A. J.; Smith, J. C.; Davison, B. H., Common processes drive the thermochemical pretreatment of lignocellulosic biomass. Green Chem. 2014, 16 (1), 63-68. 49. Terashima, N.; Yoshida, M.; Hafren, J.; Fukushima, K.; Westermark, U., Proposed supramolecular structure of lignin in softwood tracheid compound middle lamella regions. Holzforschung 2012, 66 (8), 907-915. 50. Berglund, J.; Mikkelsen, D.; Flanagan, B. M.; Dhital, S.; Gaunitz, S.; Henriksson, G.; Lindstrom, M. E.; Yakubov, G. E.; Gidley, M. J.; Vilaplana, F., Wood hemicelluloses exert distinct biomechanical contributions to cellulose fibrillar networks. Nature communications 2020, 203 11 (1), 4692. 51. Ramage, M. H.; Burridge, H.; Busse-Wicher, M.; Fereday, G.; Reynolds, T.; Shah, D. U.; Wu, G. L.; Yu, L.; Fleming, P.; Densley-Tingley, D.; Allwood, J.; Dupree, P.; Linden, P. F.; Scherman, O., The wood from the trees: The use of timber in construction. Renew Sust Energ Rev 2017, 68, 333-359. 52. Kang, X.; Kirui, A.; Muszynski, A.; Widanage, M. C. D.; Chen, A.; Azadi, P.; Wang, P.; Mentink-Vigier, F.; Wang, T., Molecular architecture of fungal cell walls revealed by solid-state NMR. Nat. Commun. 2018, 9, 2747. 53. Ehren, H. L.; Appels, F. V. W.; Houben, K.; Renault, M. A. M.; Wosten, H. A. B.; Baldus, M., Characterization of the cell wall of a mushroom forming fungus at atomic resolution using solid-state NMR spectroscopy. Cell surface 2020, 6, 100046. 54. Chrissian, C.; Camacho, E.; Kelly, J. E.; Wang, H.; Casadevall, A.; Stark, R. E., Solid-state NMR spectroscopy identifies three classes of lipids in Cryptococcus neoformans melanized cell walls and whole fungal cells. The Journal of biological chemistry 2020, 295 (44), 15083-15096. 55. Martin, R. W.; Kelly, J. E.; Kelz, J. I., Advances in instrumentation and methodology for solid-state NMR of biological assemblies. J. Struct. Biol. 2019, 206, 73-89. 56. Bougault, C.; Ayala, I.; Vollmer, W.; Simorre, J. P.; Schanda, P., Studying intact bacterial peptidoglycan by proton-detected NMR spectroscopy at 100 kHz MAS frequency. J. Struct. Biol. 2019, 206 (1), 66-72. 57. Mao, H.; Tang, J.; Chen, J.; Wang, J.; Hou, K.; Peng, Y.; Halat, D. M.; Xiao, L.; Zhang, R.; Lv, X.; Yang, A.; Cui, Y.; Reimer, J. A., Designing hierarchical nanoporous membranes for highly efficient gas adsorption and storage. Sci. Adv. 2020, 6, eabb0694. 58. Thongsomboon, W.; Serra, D. O.; Possling, A.; Hadjineophytou, C.; Hengge, R.; Cegelski, L., Phosphoethanolamine cellulose: A naturally produced chemically modified cellulose. Science 2018, 359 (6373), 334-338. 59. Chakraborty, A.; Fernando, L. D.; Fang, W.; Dickwella Widanage, M. C.; Wei, P.; Jin, C.; Fontaine, T.; Latge, J. P.; Wang, T., A molecular vision of fungal cell wall organization by functional genomics and solid-state NMR. Nature communications 2021, 12, 6346. 60. Poulhazan, A.; Dickwella Widanage, M. C.; Muszynski, A.; Arnold, A.; Warschawsi, D.; Parastoo, A.; Isabelle, M.; Wang, T., Identification and Quantification of Glycans in Whole Cells: Architecture of Microalgal Polysaccharides Described by Solid-State Nuclear Magnetic Resonance. J Am Chem Soc 2021, in press. 61. Lesage, A.; Auger, C.; Caldarelli, S.; Emsley, L., Determination of through-bond carbon- carbon connectivities in solid-state NMR using the INADEQUATE experiment. J Am Chem Soc 1997, 119 (33), 7867-7868. 62. Bennett, A. E.; Rienstra, C. M.; Griffiths, J. M.; Zhen, W. G.; Lansbury, P. T.; Griffin, R. 204 G., Homonuclear radio frequency-driven recoupling in rotating solids. J Chem Phys 1998, 108 (22), 9463-9479. 63. Williams, J. K.; Schmidt-Rohr, K.; Hong, M., Aromatic spectral editing techniques for magic- angle-spinning solid-state NMR spectroscopy of uniformly (13)C-labeled proteins. Solid state nuclear magnetic resonance 2015, 72, 118-26. 64. Schmidt-Rohr, K.; Fritzsching, K. J.; Liao, S. Y.; Hong, M., Spectral editing of two- dimensional magic-angle-spinning solid-state NMR spectra for protein resonance assignment and structure determination. Journal of biomolecular NMR 2012, 54 (4), 343-353. 65. Kang, X.; Zhao, W.; Dickwella Widanage, M. C.; Kirui, A.; Ozdenvar, U.; Wang, T., CCMRD: A Solid-State NMR Database for Complex Carbohydrates. Journal of biomolecular NMR 2019, in press. 66. White, P. B.; Wang, T.; Park, Y. B.; Cosgrove, D. J.; Hong, M., Water-polysaccharide interactions in the primary cell wall of Arabidopsis thaliana from polarization transfer solid-state NMR. J Am Chem Soc 2014, 136 (29), 10399-10409. 67. Ader, C.; Schneider, R.; Seidel, K.; Etzkorn, M.; Becker, S.; Baldus, M., Structural Rearrangements of Membrane Proteins Probed by Water-Edited Solid-State NMR Spectroscopy. J Am Chem Soc 2009, 131 (1), 170-176. 68. Torchia, D. A., Measurement of Proton-Enhanced 13C-T1 Values by a Method Which Suppresses Artifacts. J Magn Reson 1978, 30 (3), 613-616. 69. Munowitz, M. G.; Griffin, R. G.; Bodenhausen, G.; Huang, T. H., Two-dimensional rotational spin-echo nuclear magnetic-resonance in solids: correlation of chemical-shift and dipolar interactions. J Am Chem Soc 1981, 103, 2529-2533. 70. Terashima, N.; Akiyama, T.; Ralph, S.; Evtuguin, D.; Neto, C. P.; Parkas, J.; Paulsson, M.; Westermark, U.; Ralph, J., 2D-NMR (HSQC) difference spectra between specifically C-13- enriched and unenriched protolignin of Ginkgo biloba obtained in the solution state of whole cell wall material. Holzforschung 2009, 63 (4), 379-384. 71. Yelle, D. J.; Ralph, J.; Frihart, C., Characterization of nonderivatized plant cell walls using high-resolution solution-state NMR spectroscopy. Magn. Reson. Chem. 2008, 46, 508-517. 72. Mansfield, S. D.; Kim, H.; Lu, F. C.; Ralph, J., Whole plant cell wall characterization using solution-state 2D NMR. Nat Protoc 2012, 7 (9), 1579-1589. 73. Grabber, J. H.; Ralph, J.; Hatfield, R. D., Model studies of ferulate-coniferyl alcohol cross- product formation in primary maize walls: implications for lignification in grasses. Journal of agricultural and food chemistry 2002, 50 (21), 6008-16. 74. Grabber, J. H.; Ralph, J.; Hatfield, R. D., Cross-linking of maize walls by ferulate dimerization and incorporation into lignin. Journal of agricultural and food chemistry 2000, 48 (12), 6106-13. 205 75. Sauvee, C.; Rosay, M.; Casano, G.; Aussenac, F.; Weber, R. T.; Ouari, O.; Tordo, P., Highly Efficient, Water-Soluble Polarizing Agents for Dynamic Nuclear Polarization at High Frequency. Angew Chem Int Edit 2013, 52 (41), 10858-10861. 76. Takahashi, H.; Lee, D.; Dubois, L.; Bardet, M.; Hediger, S.; De Paepe, G., Rapid Natural- Abundance 2D 13C-13C Correlation Spectroscopy Using Dynamic Nuclear Polarization Enhanced Solid-State NMR and Matrix-Free Sample Preparation. Angew Chem Int Edit 2012, 51 (47), 11766-11769. 77. Perras, F. A.; Luo, H.; Zhang, X.; Mosier, N. S.; Pruski, M.; Abu-Omar, M. M., Atomic- Level Structure Characterization of Biomass Pre- and Post-Lignin Treatment by Dynamic Nuclear Polarization-Enhanced Solid-State NMR. The journal of physical chemistry. A 2017, 121 (3), 623- 630. 78. Viger-Gravel, J.; Lan, W.; Pinon, A. C.; Berruyer, P.; Emsley, L.; Bardet, M.; Luterbacher, J., Topology of Pretreated Wood Fibers Using Dynamic Nuclear Polarization. J Phys Chem C 2019, 123 (50), 30407-30415. 79. Zhao, W.; Kirui, A.; Deligey, F.; Mentink-Vigier, F.; Zhou, Y.; Zhang, B.; Wang, T., Solid- state NMR of unlabeled plant cell walls: high-resolution structural analysis without isotopic enrichment. Biotechnology for biofuels 2021, 14 (1), 14. 80. Szabo, L.; Jamal, S.; Xie, H.; Charnock, S. J.; Bolam, D. N.; Gilbert, H. J.; Davies, G. J., Structure of a family 15 carbohydrate-binding module in complex with xylopentaose. Evidence that xylan binds in an approximate 3-fold helical conformation. The Journal of biological chemistry 2001, 276 (52), 49061-5. 81. Rassolov, V. A.; Ratner, M. A.; Pople, J. A.; Redfern, P. C.; Curtiss, L. A., 6‐31G* basis set for third‐row atoms. J. Comput. Chem. 20001, 22, 976-984. 82. Cances, E.; Mennucci, B.; Tomasi, J., A new integral equation formalism for the polarizable continuum model: Theoretical background and applications to isotropic and anisotropic dielectrics. . J Chem Phys 1997, 107, 3032-3041. 83. Frisch, M. e. a., (Gaussian, Inc., Wallingford CT, 2009). 84. Raju, R. K.; Ramraj, A.; Hillier, I. H.; Vicent, M. A.; Burton, N. A., Carbohydrate–aromatic π interactions: a test of density functionals and the DFT-D method. Phys. Chem. Chem. Phys. 2009, 11, 3411-3416. 85. Yang, H.; Zimmer, J.; Yingling, Y. G.; Kubicki, J. D., How Cellulose Elongates-A QM/MM Study of the Molecular Mechanism of Cellulose Polymerization in Bacterial CESA. J. Phys. Chem. B 2015, 119, 6525-6535. 86. Smakman, G.; Rinie, H., Energy metabolism of Plantago lanceolata, as affected by change in root temperature Physiol. Plant. 1982, 56, 33-37. 87. De Visser, R.; Vianden, H.; Schnyder, H., Kinetics and relative significance of remobilized 206 and current C and N incorporation in leaf and root growth zones of Lolium perenne after defoliation: assessment by 13C and 15N steady-state labelling. . Plant Cell Environ. 1997, 20, 37- 46. 88. Wang, H. M.; Ma, C. Y.; Li, H. Y.; Chen, T. Y.; Wen, J. L.; Cao, X. F.; Wang, X. L.; Yuan, T.-Q.; Sun, R.-C., Structural Variations of Lignin Macromolecules from Early Growth Stages of Poplar Cell Walls. ACS Sustainable Chem. Eng. 2020, 8 (4), 1813-1822. 89. Rencoret, J.; Gutierrez, A.; Nieto, L.; Jimenez-Barbero, J.; Faulds, C. B.; Kim, H.; Ralph, J.; Martinez, A. T.; Del Rio, J. C., Lignin Composition and Structure in Young versus Adult Eucalyptus globulus Plants. Plant Physiol. 2011, 155, 667-682. 90. Purushotham, P.; Ho, R.; Zimmer, J., Architecture of a catalytically active homotrimeric plant cellulose synthase complex. Science 2020, 369 (6507), 1089-1094. 91. Yang, H.; Kubicki, J., A density functional theory study on the shape of the primary cellulose microfibril in plants: effects of C6 exocyclic group conformation and H-bonding. Cellulose 2020, 27, 2389-2402. 92. Phyo, P.; Wang, T.; Yang, Y.; O'Neill, H.; Hong, M., Direct Determination of Hydroxymethyl Conformations of Plant Cell Wall Cellulose Using 1H Polarization Transfer Solid- State NMR. Biomacromolecules 2018, 19 (5), 1485-1497. 93. Wang, T.; Yang, H.; Kubicki, J. D.; Hong, M., Cellulose Structural Polymorphism in Plant Primary Cell Walls Investigated by High-Field 2D Solid-State NMR Spectroscopy and Density Functional Theory Calculations. Biomacromolecules 2016, 17 (6), 2210-2222. 94. Massiot, D.; Fayon, F.; Capron, M.; King, I.; Le Calve, S.; Alonso, B.; Durand, J. O.; Bujoli, B.; Gan, Z. H.; Hoatson, G., Modelling one- and two-dimensional solid-state NMR spectra. Magn Reson Chem 2002, 40 (1), 70-76. 95. Yang, H.; Watts, H. D.; Gibilterra, V.; Weiss, T. B.; Petridis, L.; Cosgrove, D. J.; Kubicki, J. D., Quantum Calculations on Plant Cell Wall Component Interactions. Interdisciplinary sciences, computational life sciences 2019, 11 (3), 485-495. 96. Rencoret, J.; Marques, G.; Gutierrez, A.; Nieto, L.; Santos, J. I.; Jimenez-Barbero, J.; Martinez, A. T.; del Rio, J. C., HSQC-NMR analysis of lignin in woody (Eucalyptus globulus and Picea abies) and non-woody (Agave sisalana) ball-milled plant materials at the gel state 10th EWLP, Stockholm, Sweden, August 25–28, 2008. Holzforschung 2009, 63, 691-698. 97. Rico, A.; Rencoret, J.; del Rio, J. C.; Martinez, A. T.; Gutierrez, A., In-Depth 2D NMR Study of Lignin Modification During Pretreatment of Eucalyptus Wood with Laccase and Mediators. Bioenergy Res. 2015, 8, 211-230. 98. Ibarra, D.; Chavez, M. I.; Rencoret, J.; Del Rio, J. C.; Gutierrez, A.; Romero, J.; Camarero, S.; Martinez, M. J.; Jimenez-Barbero, J.; Martinez, A. T., Structural modification of eucalypt pulp lignin in a totally chlorine-free bleaching sequence including a laccase-mediator stage. Holzforschung 2007, 61, 634-646. 207 99. Wen, J. L.; Sun, S. L.; Yuan, T. Q.; Xu, F.; Sun, R. C., Structural elucidation of lignin polymers of Eucalyptus chips during organosolv pretreatment and extended delignification. Journal of agricultural and food chemistry 2013, 61 (46), 11067-11075. 100. Yuan, T.-Q.; Sun, S.-N.; Xu, F.; Sun, R.-C., Characterization of Lignin Structures and Lignin–Carbohydrate Complex (LCC) Linkages by Quantitative 13C and 2D HSQC NMR Spectroscopy. Journal of agricultural and food chemistry 2011, 59, 10604-10614. 101. Li, M.; Cao, S.; Meng, X.; Studer, M.; Wyman, C. E.; Ragauskas, A. J.; Pu, Y., The effect of liquid hot water pretreatment on the chemical–structural alteration and the reduced recalcitrance in poplar. . Biotechnology for biofuels 2017, 10, 237. 102. Samuel, R.; Foston, M.; Jaing, N.; Cao, S.; Allison, L.; Studer, M.; Wyman, C.; Ragauskas, A. J., HSQC (heteronuclear single quantum coherence) 13C–1H correlation spectra of whole biomass in perdeuterated pyridinium chloride–DMSO system: An effective tool for evaluating Fuel 2011, 90 (9), 2836-2842. 103. Samuel, R.; Cao, S.; Das, B. K.; Hu, F.; Pu, Y.; Ragauskas, A. J., Investigation of the fate of poplar lignin during autohydrolysis pretreatment to understand the biomass recalcitrance. RSC Adv. 2013, 3 (16), 5305-5309. 104. Lagerquist, L.; Pranovich, A.; Sumerskii, I.; von Schoultz, S.; Vahasalo, L.; Willfor, S.; Eklund, P., Structural and thermal analysis of softwood lignins from a pressurized hot water extraction biorefinery process and modified derivatives. . Molecules 2019, 24 (2), 335. 105. Michalak, L.; Knutsen, S. H.; Aarum, I.; Westereng, B., Effects of pH on steam explosion extraction of acetylated galactoglucomannan from Norway spruce. Biotechnology for biofuels 2018, 11 (1), 1-12. 106. Du, X.; Perez-Boada, M.; Fernandez, C.; Rencoret, J.; del Rio, J. C.; Jimenez-Barbero, J.; Li, J.; Gutierrez, A.; Martinez, A. T., Analysis of lignin-carbohydrate and lignin-lignin linkages after hydrolase treatment of xylan-lignin, glucomannan-lignin and glucan-lignin complexes from spruce wood. Planta 2014, 239, 1079-1090. 107. Wang, T.; Salazar, A.; Zabotina, O.; Hong, M., Structure and Dynamics of Brachypodium Primary Cell Walls Polysaccharides from 2D 13C Solid-State NMR Spectroscopy. Biochemistry- Us 2014, (53), 2840-2854. 108. Kang, X.; Kirui, A.; Dickwella Widanage, M. C.; Mentink-Vigier, F.; Cosgrove, D. J.; Wang, T., Lignin-polysaccharide interactions in plant secondary cell walls revealed by solid-state NMR. Nature communications 2019, 10 (1), 347. 109. Gao, Y.; Lipton, A. S.; Wittmer, Y.; Murray, D. T.; Mortimer, J. C., A grass-specific cellulose-xylan interaction dominates in sorghum secondary cell walls. Nature communications 2020, 11 (1), 6081. 110. Simmons, T. J.; Mortimer, J. C.; Bernardinelli, O. D.; Poppler, A. C.; Brown, S. P.; deAzevedo, E. R.; Dupree, R.; Dupree, P., Folding of xylan onto cellulose fibrils in plant cell 208 walls revealed by solid-state NMR. Nature communications 2016, 7, 13902. 111. Duan, P.; Kaser, S. J.; Lyczakowski, J. J.; Phyo, P.; Tryfona, T.; Dupree, P.; Hong, M., Xylan Structure and Dynamics in Native Brachypodium Grass Cell Walls Investigated by Solid- State NMR Spectroscopy. ACS Omega 2021, 6 (23), 15460-15471. 112. Terrett, O. M.; Lyczakowski, J. J.; Yu, L.; Iuga, D.; Franks, W. T.; Brown, S. P.; Dupree, R.; Dupree, P., Molecular architecture of softwood revealed by solid-state NMR. Nature communications 2019, 10 (1), 4978. 113. Phyo, P.; Wang, T.; Xiao, C.; Anderson, C.; Hong, M., Effects of Pectin Molecular-Weight Changes on the Structure, Dynamics and Polysaccharide Interactions of Primary Cell Walls of Arabidopsis thaliana: Insights from Solid-State NMR. Biomacromolecules 2017, 18, 2937-2950. 114. Ibarra, D.; Chávez, M. I.; Rencoret, J.; Del Río, J. C.; Gutiérrez, A.; Romero, J.; Camarero, S.; Martínez, M. J.; Jiménez-Barbero, J.; Martínez, A. T., Lignin Modification during Eucalyptus globulus Kraft pulping followed by totally chlorine-free bleaching: A two-dimensional nuclear magnetic resonance, fourier transform infrared, and pyrolysis− gas chromatography/mass spectrometry study. Journal of agricultural and food chemistry 2007, 55 (9), 3477-3490. 115. Rencoret, J.; Neiva, D.; Marques, G.; Gutiérrez, A.; Kim, H.; Gominho, J.; Pereira, H.; Ralph, J.; Del Río, J. C., Hydroxystilbene glucosides are incorporated into Norway spruce bark lignin. . Plant Physiol. 2019, 180, 1310-1321. 209 APPENDIX S7.1 Supplementary Methods S7.1.1 Isotope Labeling. Plant samples from eucalyptus, poplar, and spruce were produced under identical growth conditions in custom-designed, air-tight, high-irradiance labeling chambers of the Experimental Soil Plant Atmosphere System at IsoLife (Wageningen, The Netherlands). Under regulated environmental conditions, these plants were grown hydroponically in a closed atmosphere containing 97 atom% 13CO2 from germinated seed or rooted stem cutting till harvest: photosynthetic photon flux density (PPFD) 800 µmol m–2 s–1 (top of plants); CO2 concentration (Day); 400 ppm (v/v); 15 h (eucalyptus), 21 h (poplar), and 16 h (spruce) day lengths; day/night temperature of 22/16 °C (eucalyptus), 22/15 °C (poplar), and 24/20 °C (spruce). The day/night relative humidity is 75/75% for eucalyptus, 70/80% for poplar, and 70/75% for spruce. Minerals and water were supplied as aerated modified Hoagland nutrient solutions with micronutrients and iron86,87, maintaining nitrogen concentration between 25 and 200 mg/L; pH close to 5; EC between 0.4 – 0.7 mS/cm; 25% of total nitrogen was supplied as ammonium. Plant shoots were dissected into leaves and stems after removing the plants from the growth chamber. After freeze-drying, these plant stems were kept stored at -20°C, in the dark with silica-gel drying bags. All plants were rehydrated for solid-state NMR analysis. S7.1.2 Effects of Plant Age on Cell Walls. Given our ever-increasing knowledge on plant secondary cell wall, impacts of the plant age at which the material is harvested relative to its natural growth cycle can be considered. Our 6.5-months old poplar can be put in perspective of 3-, 6- and 18-month-old poplar samples reported in literature88. Overall, there is no significant change in cellulose to hemicellulose proportions, and 210 cellulose crystallinity has become a stable parameter after 6 months. However, a slight increase in lignin content with age has been consistently reported for both poplar and eucalyptus89. Solution NMR data have demonstrated that lignin structure undergoes substantial evolution over a longer period of time, in both unit nature. For example, S lignin is deposited at later stages and the proportions of different linkage types. In-depth analyses of polysaccharide structure, lignin composition and linkage types, the consequential effect on lignocellulose architecture is thus of significant interest for future solid-state NMR research. 211 Figure S7.1 2D 13C-13C spectra of polysaccharides in softwood and hardwood. a, 2D 13C-13C RFDR spectra showing the rigid components of eucalyptus, poplar, and spruce. Uniquely, spruce has mannan and arabinose signals in the CP-based RFDR spectrum. b, 2D 13C DP J- INADEQUATE spectra of eucalyptus, poplar, and spruce. The combination of 13C DP with a short recycle delay of 2 s selectively detects the mobile molecules. Arabinose (A) signals dominate the spectra, and the four major types of arabinose are distinguished by the superscript, from Aa to Ad. The carbon connectivity of arabinose is highlighted by yellow lines. Poplar has relatively weak signals of arabinose. The 13C chemical shifts are summarized in Tables S7.1 and S7.2. 212 Figure S7.2 Cellulose bundling models that violate the NMR constraints. The cartoon illustrations include a and b, two elementary microfibrils, c, four elementary microfibrils, and d, five elementary microfibrils. Each elementary microfibril is depicted to contain 18 glucan chains following the current biochemical evidence90, 91. The number of hydrophobic surface chains (sg), hydrophilic surface chains (sf), middle layer internal chains (ia,b), and the deeply embedded core chains (ic) are labeled92, 93. The matrix polymers are added to better fit the NMR constraints, but major violations still exist for all these models. 213 Figure S7.3 Lignin distributed in the mobile and rigid phases. The top row is the quantitative- detected DP (orange), mobile molecule detected DP, rigid component detected CP spectra of a, eucalyptus, b, poplar, and c, spruce. All the three difference spectra of 13C DP spectra with long and short recycle delays showing peaks of the rigid components, are comparable to the corresponding CP spectra. Both 2-s DP and CP spectra give well-resolved aromatic peaks, indicating dynamically heterogeneous lignin components. The bottom row is DP-PDSD spectra (1.7s recycle delay) of d, eucalyptus, e, poplar, and f, spruce. Abundant signals of aromatics and linkers are identified, suggesting that these linkers are also revolved at mobile phases of lignin. 214 Figure S7.4 Solution NMR of lignin and polysaccharides in woody stems. 2D 1H-13C HSQC spectra showing selected regions of a, mainly lignin signals and b, mainly polysaccharide peaks. Arabinose (Ara) is more predominant in spruce than in eucalyptus and poplar. c, Full HSQC spectra of the three samples. The 13C and 1H chemical shifts are documented in Table S7.4 and supported by a literature surveil compiled in Table S7.5. 215 Figure S7.5 Spectral deconvolution reveals the lignin composition. Deconvolution is conducted on 1D quantitative 13C DP spectra collected with long recycle delays of 35-40 s. The experimentally measured spectra (black) is overlaid with the simulated spectra (red) for eucalyptus, poplar, and spruce. The bottom panels show the deconvoluted peaks. Attributed resonances are plotted in blue and uncertain resonances are in purple. Deconvolution is conducted using the DMfit software94, using a Lorentzian model on a 92-195 ppm chemical shift window. The fitting parameters are summarized in Table S7.6. 216 Figure S7.6 1D 13C cross sections of lignin region reveal the polymer mixing patterns. Representative cross sections are obtained from dipolar gated 2D 13C-13C correlation spectra (Figure 7.3a) of a, eucalyptus, b, poplar, and c, spruce. For eucalyptus and poplar, the first column shows the cross sections from S-lignin units and the second column shows those from G units. All the cross sections are normalized by the diagonal peaks (blue asterisks). The more similar the spectral pattern between 0.1 s and 1.0 s, the more homogeneous the polymers are mixed on the sub-nanometer scale. The spruce cross sections are more equilibrated. 217 Figure S7.7 Comparable spin diffusion rates of biopolymers in three wood samples. a, Aromatics regions (top row) and carbohydrates regions (second row) of 0.1s PDSD spectra of three woody plants. b, Representative cross peaks for cellulose, xylan, and lignin are obtained from 0.1s PDSD spectra, which are normalized by the diagonal peaks (red asterisks). The similar spectral patterns among three wood samples indicates the spin diffusion rates of lignin and carbohydrates are comparable in all three woody stems. 218 Figure S7.8 13C–1H dipolar order parameters of biopolymers in wood cell walls. a, Dipolar dephasing curves of carbohydrates and lignin processed from DIPSHIFT spectra. The spectra were collected under 7.5 kHz MAS with a 0.577 scaling factor for C-H bond. b, Summary of C–H dipolar order parameters with error bars. Spruce shows a relatively smaller order parameter for xylan. c, The first (0 μs dipolar dephasing) and middle (67 μs dipolar dephasing) slices of the 2D DIPSHIFT experiment imply how much the spectra are suppressed at a half rotor period. Source data are provided as a Source Data file. 219 Figure S7.9 Intermolecular interactions of polymers in woody plants. Dipolar-gated 2D 13C-13C correlation spectra measured with 1.0 s and 100 ms (poplar), 1.0 s and 20 ms (spruce) are overlaid for a, poplar and b, spruce. 100 ms (poplar) and 20 ms (spruce) spectra mainly detect the intramolecular correlations, and the 1 s spectra show many intermolecular cross peaks. Only the 220 Figure S7.9 (cont’d) intermolecular cross peaks are labeled. c, The bar diagrams show the number of interactions between different molecules in the woods. The interactions are categorized according to the peak intensity. The details of the short-range and long-range cross peaks are summarized in Tables S7.7- 7.10. Source data are provided as a Source Data file. 221 Figure S7.10 NMR restraints of polymer packing in secondary cell walls. Thick lines, thin lines, and dash lines are used to represent strong, medium, and weak cross peaks observed in solid-state NMR spectra, respectively. Three major types of interactions happen between xylan acetyl and lignin (brown), between lignin methyl ether and carbohydrates (orange), and between lignin aromatics and carbohydrates (yellow). The details of these cross peaks are tabulated in Tables S7.7-7.10. 222 Figure S7.11 DFT structures of lignin-carbohydrate packing. The structure includes the complex formed between G (yellow) or S (cyan) unit of lignin and three types of polysaccharides. The structures were adapted from an earlier study. 95 223 Figure S7.12 Buildup curves for water-to-polysaccharides/lignin. The water 1H spin diffusion curves for cellulose, xylan, and lignin for a, Eucalyptus, b, poplar and c, spruce. Dash lines indicate the intensities of 4-ms 1H mixing. Poplar has the slowest spin diffusion from water among the three woods, with ~15-20% of the equilibrium intensity detected at 4-ms 1H mixing. Eucalyptus and spruce have the fastest spin diffusion with ~45-60% of the equilibrium intensity detected at 4- ms 1H mixing. d, Water-edited spectra measured with different mixing times. The 4-ms spectrum 224 Figure S7.12 (cont’d) detects the hydrated molecules and the 36-ms (eucalyptus and spruce) and 169-ms (poplar) spectra report equilibrium intensities. The difference spectrum only shows well-hydrated molecules. Source data are provided as a Source Data file. 225 Figure S7.13 Water-edited 2D 13C-13C correlation spectra of polysaccharides. Overlay of water- edited and equilibrium spectra showing hydration maps with S/S0 values for a, Eucalyptus, b, poplar, and c, spruce. The blue dash lines indicate the positions at which the 13C cross sections are extracted. The representative 1D 13C cross sections of cellulose (interior: i4; surface: s4) and xylan (Xn4 and Xn5) are shown for a, Eucalyptus, b, poplar, and c, spruce. The water-edited spectra and the equilibrium spectra are plotted in orange and black, respectively. The 3-fold and 2-fold xylans have enhanced intensity in the water-edited spectra, indicating better interactions with water molecules. e, Hydration maps of lignin regions with S/S0 values for poplar (top) and spruce (bottom). Spruce is well hydrated compared to poplar as it shows higher S/S0 values. The water- edited intensities are summarized in Tables S7.11-7.13. 226 Figure S7.14 The 13C-T1 relaxation curves of polysaccharides and lignin. The 13C-T1 relaxation curves of a, cellulose (interior and surface glucan chains), b, hemicellulose (2-fold and 3-fold xylan; mannan), and c, lignin (S and G units) are shown. The left, middle, and right columns are for eucalyptus, poplar, and spruce, respectively. The data are fitted using a single exponential equation. The fit parameters are summarized in Table S7.14. Source data are provided as a Source Data file. 227 Figure S7.15 1H-T1ρ relaxation curves of polysaccharides and lignin in woody plants. The 1H-T1ρ relaxation curves of a, polysaccharides and b, lignin data are shown for eucalyptus, poplar, and spruce. The data are fitted using a single exponential equation. Xylan shows faster relaxation times compared to cellulose. Within lignin, the G residue has faster 1H-T1ρ relaxation than the S unit. The fit parameters are summarized in Table S7.15. Source data are provided as a Source Data file. 228 Figure S7.16 Effect of freeze-drying and rehydration on hardwood eucalyptus. a, CP (rigid components selection), quantitative DP and DP (mobile components selection) normalized to i4 peak at 89 ppm. Distinct chemical shift intervals for lignin, carbohydrate, and lipid are highlighted. b, T2-filtered (water-edited) 1D experiments. Signal maxima observed after 64 ms of spin diffusion, in comparison to the previous 36 ms and 26% signal difference in the absence of diffusion upon rehydration are consistent with both a sample less prone to deep hydration and a slight betterment of its surface water access after freeze-drying. c, Average 13C T1 relaxation curves, using an evolution of integrals of spectral regions established in panel (a). d, Average 1H- T1ρ relaxation using the same two integrated intervals. Error bars are standard deviations of the fitting derived parameters. Source data are provided as a Source Data file. All these spectra and fitted relaxation parameters are very close for the fresh and rehydrated samples, demonstrating that structure and dynamics are not significantly altered by this sample preparation method. Their overall consistency with the main text eucalyptus sample also validates experimental reproducibility. 229 Figure S7.17 Representative flow chart for resonance assignment. Shading and numbering on the left highlight the four main steps, while red coloring denotes essential sub-steps of the protocol followed for assigning chemical shifts. 1) Collection of reference chemical shifts for expected resolved peaks from solid-state NMR literature and databases. For each carbon site, we report associated references in Table S7.19. 2) Acquisition of 2D solid-state NMR spectra. Subsequent assignments rely primarily on the resolution provided by the DQ dimension of INADEQUATE spectra. 3) Sequential assignment (boxes from right to left), from well-established cellulose and xylan conformers to mobile primary cell wall polysaccharides, ending with lignin and its linkages. For each polymer, we indicate a well-established site from which assignment can be easily initiated, while less resolved sites are either identified via cross-peaks connections or from literature reported values. 4) Safeguarding steps taken to provide accountability in the final assignments. Particularly for the complex lignin assignment, a good match is demanded with solution NMR spectra, as summarized in Tables S7.5 and S7.19. 230 Table S7.1. 13C chemical shifts of rigid polysaccharides and lignin in wood secondary cell walls. The rigid components are identified from 13C CP-based INADEQUATE and RFDR spectra. Superscripts are used to denote different allomorphs. Unidentified signals are indicated as (-). Not applicable (/) G Ga Gb Gc Gd Ga Gb Sa Sb Sa Sb S’a S’b A(S) A (G) A(S) B B’ B B’ B C’ C’ C E Lignin guaiacyl syringyl Cα-oxidized syringyl β-O-4’ γ-acylated β-β’ β-β’ β-5’ γ-acylated β-5’ β’-1’ C1 - C2 C3 C4 C5 - - - - - - - - 112.1 148.2 144.9 116.1 113.8 148.2 144.9 116.1 111.1 148.9 144.9 115.9 130.7 113.4 152.2 145.3 116.3 133.5 110.5 149.7 146.1 134.2 113.5 147.6 144.1 136.0 109.5 151.6 142.2 57.0 56.7 120.2 57.0 119.2 57.0 121.1 57.0 56.6 56.6 134.9 102.8 153.3 133.5 153.3 102.8 57.0 109.7 153.2 137.2 153.2 109.7 57.0 134.2 104.0 153.1 134.6 153.1 104.0 56.6 137.1 102.5 152.7 135.8 152.7 102.5 56.6 134.1 109.2 144.8 136.3 144.8 109.2 56.5 136.5 112.5 145.5 138.5 145.5 112.5 56.5 / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / 231 C6 OMe C7 (α) C8(β) C9(γ) / - / - / / / / / / / / / / / / / Plant Eucalyptus Spruce Spruce Spruce Spruce Poplar Poplar Eucalyptus Eucalyptus Poplar Poplar Eucalyptus Eucalyptus Eucalyptus Spruce Poplar Eucalyptus Eucalyptus Spruce Spruce Poplar Spruce Spruce Spruce Eucalyptus Spruce / / / / / / / 59.4 59.9 60.4 71.2 74.4 70.3 74.1 70.5 / / / / / / / 72.3 72.4 72.0 84.4 83.9 84.8 90.2 86.7 88.1 81.1 81.4 / / / / / / / 84.5 84.3 86.3 53.5 53.5 53.9 52.9 52.6 49.9 49.6 54.2 59.6 60.0 Table S7.1 (cont’d) Xn2f Xn3f Xn Xn2f Xn3f Xn2f Xn3f x GlcAa GlcAb GlcAc GalA M MAc A (t-Ara) ia ib ic sf sg ia ib ic sf sg ia ib sf sg Carbohydrate 2-fold xylan 3-fold xylan xylan mix 2-fold xylan 3-fold xylan 2-fold xylan 3-fold xylan xylose in XyG GlcA in GAX GlcA in GAX GlcA in GAX galacturonic acid mannan acetylated mannan arabinose interior cellulose surface cellulose interior cellulose surface cellulose interior cellulose surface cellulose / C1 104.8 102.5 - 104.8 102.2 104.8 103.5 - 101.5 99.2 98.5 100.2 101.2 101.6 108.1 104.9 104.9 104.9 104.9 104.9 104.8 104.8 104.8 104.8 104.8 104.8 104.8 104.8 104.8 / C2 72.2 73.4 - 72.1 73.6 72.2 73.9 - 71.3 69.9 71.6 68.1 71.3 72.1 82.1 72.1 72.1 72.1 72.1 72.1 72.1 72.1 72.1 72.1 72.1 72.2 72.2 72.2 72.2 / C3 74.0 74.1 74.2 73.2 74.2 73.3 74.2 74.1 79.8 - 80.8 - 75.1 75.1 - 75.2 75.2 74.1 74.6 73.6 75.0 75.0 74.3 75.7 75.2 75.3 75.3 75.1 74.3 / C4 82.2 77.4 78.5 82.0 77.9 82.2 77.8 70.4 69.9 - - 79.6 - 80.7 85.6 88.7 88.7 87.9 84.4 83.5 88.9 88.9 87.9 84.3 83.4 89.0 89.0 84.5 84.3 232 / C5 64.0 63.3 64.1 63.9 64.1 63.9 65.4 62.5 76.6 - - - 75.1 75.1 62.1 72.3 72.3 71.1 74.6 73.6 72.4 72.4 70.1 75.0 73.7 72.4 72.4 75.0 74.3 80.3 / / C6 AcCO AcMe 21.0 174.0 / 21.0 174.0 / 21.0 174.0 / 21.0 174.0 / 21.0 174.0 / 21.0 174.0 / 21.0 174.0 / / / / - - - - 62.1 62.1 / 65.0 65.0 65.7 62.3 61.4 64.9 64.9 65.5 62.3 61.3 65.0 65.0 62.1 61.4 59.7 Poplar Eucalyptus Eucalyptus Poplar Spruce All samples All samples All samples All samples All samples Spruce Spruce Spruce Eucalyptus Eucalyptus Poplar Poplar Spruce Spruce Table S7.2. 13C chemical shifts of mobile molecules in woods. The mobile components are identified from 13C DP-based INADEQUATE spectra measured with short recycle delays of 2 s. Superscripts are used to denote different allomorphs. Not applicable (/). Unidentified (-). carbohydrate C1 arabinose galactose α-glucose β-glucose xylose in XyG cellulose lignin guaiacyl syringyl guaiacyl syringyl guaiacyl 107.8 107.8 109.7 106.9 104.9 92.7 96.6 100.3 - - - - - - - C1 136.3 136.3 136.7 135.5 130.1 - - - - C2 81.8 80.1 81.8 85.3 72.3 - 72.3 - - - - - - - C2 - - - - - - - - - C3 77.2 - 77.4 - - - - - 75.5 75.5 75.5 75.5 - - - C3 C4 84.6 - 82.9 - - - - 71.9 88.9 88.9 88.9 88.9 - - - C4 C5 61.8 - 67.6 - - - - 61.3 72.4 70.9 72.4 70.9 75.0 73.7 73.7 C5 144.1 145.2 152.9 151.7 147.8 152.6 151.9 148.9 152.3 144.1 145.2 - - 146.4 - - 144.7 - - - 152.9 151.7 - 152.6 151.9 - 152.3 C6 / / / / - - - / 64.9 65.8 64.9 65.8 61.4 61.2 61.2 C6 - - - - - - - - - Aa Ab Ac Ad Gal α-Glc β-Glc x ia,b ic ia,b ic sf sg sg Ga Gb Sa Sb G Sa Sb G G Plants All samples All samples All samples Eucalyptus All samples Spruce Spruce All samples Eucalyptus Eucalyptus Poplar Poplar All samples Eucalyptus Poplar Plants Eucalyptus Poplar Spruce 233 Table S7.3. Molar composition of polysaccharides in wood secondary cell walls. For each plant, the composition of cellulose, xylan, mannan, and the primary cell wall component xyloglucan (XyG) is given. Cellulose contains interior glucan chains (ia,b: middle layer; ic: embedded core chains) and surface glucan chains (sf: hydrophilic surface; sg: hydrophobic surface). Xylan contains the backbone (Xn2f: two-fold xylan; Xn3f: three-fold xylan; Xn: mixed conformation) and sidechains (GlcA and/or Ara). For Mannan, only the mannose residues (M: unacetylated; MAc: acetylated) are resolved, while the actual fraction of mannan should be higher than the value reported here due to the presence of Glc residues in the backbone. The sidechain α- Mannan - - - Mannan - - XyG 0.5% x 0.5% XyG 0.6% x 0.6% xylose (x) is used to denote xyloglucan. Cellulose 74.0% interior chains 38.2% surface chains 35.8% ic sf sg Xn2f ia,b 28.3% Cellulose 81.4% interior chains 45.0% surface chains 36.4% ia,b ic sf sg Xn2f Eucalyptus Xylan 25.5% Poplar Xylan 18.0% backbone 16.8% sidechains 8.7% Xn Xn3f GlcA Ara M MAc 9.8% 21.2% 14.6% 10.0% 2.0% 4.9% 8.7% - - - backbone 13.3% sidechains 4.8% Xn Xn3f GlcA Ara M MAc 32.7% 12.3% 21.9% 14.6% 7.9% 1.6% 3.8% 4.8% - - - Cellulose 75.8% interior chains 37.1% surface chains 38.7% ia,b 28.1% ic Xn2f 8.9% 23.6% 15.2% 6.9% sg sf Spruce Xylan 11.8% Mannan 10.5% XyG 1.9% backbone 10.4% sidechains 1.5% Xn Xn3f GlcA Ara M MAc - 3.5% 0.5% 1.0% 6.2% 4.3% 1.9% x The area of the following well-resolved peak pairs in CP J-INADEQUATE spectra are used: ia,b : the average of i5a,b - i6a,b, i4a,b - i5a,b , and i3a,b - i4a,b. 234 Table S7.3 (cont’d) ic : the average of i5c - i6c, i4c - i5c , and i3c - i4c. sf : the average of s5f - s6f, s4f - i5/3f. The integration of s4f - i5/3f has been doubled because 2 peak pairs are included (carbon 4-carbon 5 and carbon 4-carbon 3). sg : the average of s5g - s6g, s4g - i5/3g. The integration of s4g - i5/3g has been doubled because 2 peak pairs are included (carbon 4-carbon 5 and carbon 4-carbon 3). Xn2f: the average of Xn42f and Xn52f. Xn: the average of Xn4 and Xn5. Xn3f: the average of Xn43f and Xn53f. GlcA: the average of GlcA1 and GlcA2. The value reported are the sum of GlcAa, GlcAb and GlcAc. Ara: the average of A1 and A2. M: the average of M1 and M2. MAc, the average of M1Ac and M2Ac. x: the average of x4 and x5. 235 Table S7.4. Solution NMR HSQC 1H and 13C chemical shifts of lignin aromatics and linkages. Unidentified or unresolved (-). Eucalyptus Poplar Spruce 13C (ppm) 1H (ppm) 13C (ppm) 1H (ppm) G2 G5 guaiacyl guaiacyl G6 G’2 G’6 S2/6 S’2/6 guaiacyl oxidized guaiacyl oxidized guaiacyl syringyl oxidized syringyl H2/6 Aα (G) Aα (S) Aβ(G) Aβ(S) Aγ Bα Bβ Bγ Cα Cβ Cγ β-O-4’ β-O-4’ β-O-4’ β-O-4’ β-O-4’ β-β’ (resinol) β-β’ (resinol) β-β’ (resinol) β-5’ (phenylcoumaran) β-5’ (phenylcoumaran) β-5’ (phenylcoumaran) 111.6 114.9 & 115.9 119.7 112.2 123.6 - - 7.0 6.77 & 6.93 6.86 7.51 7.57 - - - - 71.9 4.77 84.6 4.31 - - 85.6 53.9 70.3 4.65 3.14 3.77 13C (ppm) 111.9 115.5 & 115.8 119.9 112.2 124.2 104.3 107.3 1H (ppm) 6.99 6.67 111.6 115.7 & 6.97 & 115.8 6.99 7.53 7.16 6.71 7.32 & 7.12 119.8 112.4 123.8 104.6 107.2 126.9 71.8 7.20 4.65 128.6 71.8 7.00 6.78 & 6.97 6.83 7.53 7.60 6.73 7.24 & 7.09 7.22 4.65 4.88 72.6 4.88 72.6 84.5 86.7 60.5 85.8 53.8 71.7 87.7 53.8 62.7 84.6 86.7 - 84.6 54.3 70.5 4.31 4.14 - 4.76 3.10 3.73 4.31 4.14 3.42 & 3.67 4.68 3.14 3.88 & 4.20 5.49 87.7 5.49 87.6 5.49 3.14 54.1 3.49 54.2 3.42 3.44 - - - - & 3.59 236 Table S7.4 (cont’d) Dα Dβ Eα Eα’ Eβ Eβ’ E2/2’ E6/6’ Iα Iβ Iγ Jα Jβ J6 5-5’ (dibenzodioxocin) 5-5’ (dibenzodioxocin) β-1’ (spirodienone) β-1’ (spirodienone) β-1’ (spirodienone) β-1’ (spirodienone) β-1’ (spirodienone) β-1’ (spirodienone) cinnamyl alcohol cinnamyl alcohol cinnamyl alcohol cinnamaldehyde cinnamaldehyde cinnamaldehyde PB2/6 p-hydroxybenzoate Stα/β stilbene 83.8 85.6 81.8 86.9 60.4 - 114.9 119.5 128.3 129.7 62.4 154.5 126.7 122.4 - - 4.90 - - - - 3.91 85.6 3.92 84.9 3.88 5.07 4.41 2.79 - 6.24 6.12 6.46 6.25 4.11 7.65 6.81 7.21 - - 81.9 87.0 60.7 78.7 114.2 119.2 129.0 129.5 62.3 154.2 127.1 123.9 132.1 - 5.08 4.42 2.80 4.14 6.23 6.15 6.45 6.27 4.12 7.64 6.80 7.22 7.70 - 83.7 - 61.8 - 114.3 119.5 129.5 129.1 62.5 154.4 127.4 124.3 - 127.1 5.02 - 2.8 - 6.21 6.11 6.48 6.32 4.13 7.66 6.88 7.20 - 6.88 237 Table S7.5. Average solution NMR chemical shifts of lignin from literature. The reference (Ref.) for each chemical shift is provided. Not detected (-). 13C (ppm) 111.3 115.2 119.3 112.4 124.0 104.2 106.5 - 71.6 72.2 83.9 86.4 59.8 85.2 53.7 71.4 87.1 53.5 62.7 83.6 85.9 81.4 86.6 60.0 79.5 112.4 119.1 - - G2 G5 G6 G’2 G’6 S2/6 S’2/6 H2/6 Aα(G) Aα (S) Aβ(G) Aβ(S) Aγ Bα Bβ Bγ Cα Cβ Cγ Dα Dβ Eα Eα’ Eβ Eβ’ E2/2’ E6/6’ Iα Iβ Eucalyptus 1H (ppm) Ref. Poplar 1H (ppm) 13C (ppm) 6.98 6.72 6.94 6.80 7.50 7.00 6.99 7.31 - 4.74 4.87 4.29 4.09 3.39 3.69 4.66 3.08 3.82 5.46 3.45 3.67 4.83 3.88 5.09 4.39 2.74 4.11 6.26 6.07 - - 89, 96-99 89, 96-99 89, 96-99 89, 96-99 98 98 89, 96-99 89, 96-99 - 89, 96-99 89, 96-99 89, 96-99 89, 96-99 96-99 96-99 89, 96-99 96-99 96-99 89, 96-99 96-99 96-99 96 96 96, 97 96 96, 97 96, 97 96, 97 96, 97 - - 111.1 115.0 - 119.1 111.4 123.3 103.8 106.1 127.9 71.5 71.9 83.7 a 85.8 59.9 85.0 53.6 70.6 87.0 53.2 62.5 81.4 - 81.2 - 59.7 79.5 113.3 118.9 128.4 128.2 6.98 6.80 - 6.8 7.51 7.60 6.67 7.22 7.19 4.75 4.83 4.29 4.13 3.45 4.64 3.07 3.81 5.47 3.48 3.74 4.7 - 5.07 - 2.77 4.12 6.22 6.07 6.44 6.25 238 Spruce 1H (ppm) 6.94 6.85 - 6.83 - - - - - 4.75 - 4.29 - 3.24 3.61 4.65 3.07 3.74 5.47 3.46 3.72 4.84 3.88 5.03 - 2.77 - - - - - ref 104, 105 104, 105 - 104, 105 98, 100 100 - - - 104, 106 104, 106 104, 106 104, 106 104, 106 104, 106 104, 106 104, 106 104, 106 104, 106 104, 106 104, 106 104, 106 104, 106 106 - 106 - 96, 97, 100 96, 97, 100 - - 13C (ppm) 110.3 115.2 - 118.9 - - - - - 71.1 - 84.0 - 59.9 85.0 53.6 70.9 87.0 53.1 62.8 83.3 85.4 81.7 - 59.3 - - - - - ref 88, 100-103 88, 100-103 - 88, 100-103 100 100 88, 100-103 88, 100-103 100 101 88, 101-103 88, 100, 101 88, 100-103 100-103 100, 101 88, 100-103 100-103 100, 101 88, 100-103 100-102 100, 101 101 - 100 - 100 100 100 100 100 100 Table S7.5 (cont’d) Iγ Jα Jβ J6 PB2/6 Stα/β 61.7 - - - - 126 4.09 - - - - 7.0 96-98 - - - - 99 61.4 153.5 126.1 - 131.1 - 4.10 7.60 6.76 - 7.66 - 100 100 100 - 88, 100-103 - 61.5 153.7 126.3 123.2 - 128.3 4.08 7.61 6.77 7.22 - 7.12 104, 106 104 104 104 104 104 239 Table S7.6. Spectral deconvolution of quantitative 13C spectra for lignin compositional analysis. The attributed chemical shifts are labeled. Unidentified (-). Chemical Shift [ppm] 180.6 177.9 173.8 171.8 169.4 168.4 167.5 157.3 152.2 147.7 144.8 138.4 136.4 132.2 129.8 128.2 125.0 119.9 114.7 109.4 107.6 104.6 102.2 100.8 98.5 96.4 92.6 Chemical Shift [ppm] 181.0 173.6 171.7 152.6 148.2 145.2 136.9 134.2 Eucalyptus Amplitude Width [ppm] 117 170 956 211 306 314 385 36 401 171 550 63 341 149 79 128 133 70 195 185 229 2261 75 444 281 209 135 4.2 2.6 3.9 3.2 1.5 0.4 0.4 0.1 3.5 3.4 2.6 2.6 4.4 5.5 0.6 1.9 2.0 3.8 6.0 4.9 0.9 2.6 0.3 2.4 1.9 0.8 0.8 Poplar Amplitude Width [ppm] 109 1037 173 744 217 151 165 315 6.7 3.2 1.1 3.8 3.3 2.3 2.9 3.8 240 Integral [%] 2.2 1.9 16.2 3.0 2.0 0.5 0.7 0.02 6.2 2.6 6.3 0.7 6.6 3.6 0.2 1.1 1.2 1.2 5.2 4.0 0.9 25.5 0.1 4.6 2.4 0.7 0.5 Integral [%] 3.1 14.1 0.8 12.2 3.0 1.5 2.0 5.1 Attribution - - AcCO - - - - - S3/5 G3 G4/S’3 S’b4 S1/S’a4 - - - - G5 S’2/6 A1 i/s/Xn12f Xn13f Man1 - - - Attribution AcCO S3/5 G3 G4 Sb4 Sa4 3.2 0.4 6.8 0.1 40.1 0.1 7.2 0.1 Integral [%] 1.2 1.3 1.9 9.5 1.7 0.1 0.6 6.3 9.2 3.8 0.9 1.6 4.5 1.7 1.7 1.4 0.6 20.6 2.1 18.4 1.6 4.6 2.8 1.2 0.7 G5 i/s/Xn12f Xn13f Attribution AcCO Ga3/4 Gb3/4 Gb1 Gc/d1 Ga1 G6 G2/5 A1 M1 β-Glc1 α-Glc1 Table S7.6 (cont’d) 131.2 120.0 115.5 108.9 104.7 102.4 100.6 96.5 Chemical Shift [ppm] 183.0 181.7 178.0 174.1 171.9 169.5 157.4 153.4 148.2 145.5 137.1 134.6 132.1 129.9 128.6 125.0 120.1 114.7 107.7 104.8 101.8 100.4 98.8 96.5 92.6 193 42 188 16 3650 60 632 62 3.9 2.4 8.4 1.9 2.6 0.5 2.7 0.3 Spruce Amplitude Width [ppm] 96 248 206 836 224 36 375 323 466 289 89 163 322 389 266 52 57 389 623 2103 275 907 492 477 308 3.3 1.4 2.5 3.1 2.1 1.1 0.4 5.3 5.3 3.6 2.8 2.6 3.8 1.2 1.7 7.6 2.9 14.2 0.9 2.6 1.6 1.4 1.5 0.7 0.7 241 Table S7.7. The Intensities of intermolecular cross peaks of eucalyptus. In sum, 98 intermolecular interactions are identified, including 45 strong, 29 medium, and 24 weak restraints. In gated 1-s PDSD spectra, a peak higher than 4% is categorized as a strong restraint (s, in bold), between 2% - 4% for medium (m) restraint, and below 2% for weak (w) restraint. All the peaks shown in the gated 100-ms PDSD spectrum are categorized as strong restraints. The intensity is a relative ratio of the peak, which is normalized by the integral of the 13C cross-section. Atom1 (ω1) Atom2 (ω2) Gated 100 ms PDSD (%) cellulose-xylan cellulose- lignin lignin-lignin AcCO AcMe AcCO AcMe G3 G4,S’3/5 G5 S1/4a,S’1a S/S’4b S3/5 OMe G3 G3 G4,S’3/5 G4,S’3/5 G5 S1/4a,S’1a S1/4a,S’1a S3/5 S3/5 S/S’4b S/S’4b OMe OMe G3 G3 G3 G3 G4,S’3/5 G4,S’3/5 i4 i4 s4 s4 i4 i4 i4 i4 i4 i4 i4 s4 s6 s4 s6 s4 s4 s6 s4 s6 s4 s6 s4 s6 S1/4a,S’1a S2/6b S3/5 S/S’4b S2/6a S2/6b 242 2.11 0.67 0.94 1.23 0.98 1.34 Gated 1s PDSD (%) 0.13 1.40 0.16 0.85 1.84 0.49 3.29 7.81 1.28 1.27 0.99 2.39 4.05 1.24 3.49 0.83 2.35 5.49 1.34 4.21 1.24 5.25 1.77 3.39 3.42 2.41 7.00 3.37 1.34 3.45 Type w w w w w w m s w w s m s w m w m s w s w s s s m m s m w m Table S7.7 (cont’d) xylan-lignin G4,S’3/5 G5 G5 G5 G5 S1/4a,S’1a S3/5 S3/5 S3/5 S3/5 S3/5 S/S’4b S/S’4b AcCO AcCO AcCO AcCO AcCO G3 G4,S’3/5 G5 S1/4a,S’1a S3/5 S/S’4b OMe AcMe AcMe AcMe AcMe G3 G4,S’3/5 G5 S1/4a,S’1a S3/5 S/S’4b OMe G3 G3 G4,S’3/5 G4,S’3/5 G4,S’3/5 G5 G5 G5 S3/5 S1/4a,S’1a S2/6a S3/5 S/S’4b G3 G2 G3 G4,S’3/5 G5 S’1b/4a G3 G5 G4,S’3/5 G5 S1/4a,S’1a S3/5 OMe AcCO AcCO AcCO AcCO AcCO AcCO AcCO G4,S’3/5 S3/5 S/S’4b OMe AcMe AcMe AcMe AcMe AcMe AcMe AcMe Xn42f Xn52f, 3f Xn42f Xn43f Xn52f, 3f Xn42f Xn43f Xn52f, 3f 243 1.36 3.12 2.36 2.65 0.99 1.19 1.68 4.75 7.34 3.26 5.18 3.91 3.87 2.40 1.88 1.88 0.48 3.39 2.58 2.20 2.26 0.47 1.83 2.01 4.61 3.85 8.98 6.87 9.32 3.58 8.08 6.01 1.56 2.33 1.95 2.85 5.12 6.31 6.94 4.26 3.60 4.27 4.34 3.34 3.34 1.59 1.19 5.03 2.11 2.81 1.33 s s m s m m m w w w m m m m w w m s m s s s m s s w m w m s s s s m s s m m w w s m m w Table S7.7 (cont’d) mixed sugar- lignin S1/4a,S’1a S1/4a,S’1a S1/4a,S’1a S3/5 S3/5 S/S’4b S/S’4b S/S’4b OMe OMe OMe G3 G3 G4,S’3/5 G4,S’3/5 G5 G5 S1/4a,S’1a S3/5 S3/5 S/S’4b S/S’4b OMe OMe Xn42f Xn43f Xn52f, 3f Xn42f Xn52f, 3f Xn42f Xn43f Xn52f, 3f Xn42f Xn43f Xn52f, 3f i/s/Xn12f Xn2/33f,Xn32f,i2/5,s2/3 i/s/Xn12f Xn2/33f,Xn32f,i2/5,s2/3 i/s/Xn12f Xn2/33f,Xn32f,i2/5,s2/3 i/s/Xn12f i/s/Xn12f Xn2/33f,Xn32f,i2/5,s2/3 i/s/Xn12f Xn2/33f,Xn32f,i2/5,s2/3 i/s/Xn12f Xn2/33f,Xn32f,i2/5,s2/3 0.64 0.96 1.34 1.67 2.61 4.36 3.96 7.36 2.71 2.62 5.29 2.71 5.64 3.42 4.92 2.86 4.07 4.15 3.66 1.55 3.85 6.08 23.70 1.23 18.03 5.84 22.50 8.47 7.68 26.12 2.22 17.08 8.71 14.56 s s s m s m s s s s s s s w s s s s s s m s s s 244 Table S7.8. The Intensities of intermolecular cross peaks of poplar wood. In sum, 80 intermolecular interactions are identified, including 22 strong, 33 medium, and 25 weak restraints. In gated 1-s PDSD spectra, a peak higher than 4% is categorized as a strong restraint (s, in bold), between 2% - 4% for medium (m) restraint, and below 2% for weak (w) restraint. All the peaks shown in the gated 100-ms PDSD spectrum are categorized as strong restraints. The intensity is a relative ratio of the peak, which is normalized by the integral of the 13C cross-section. Atom1 (ꞷ1) Atom2 (ꞷ2) Gated 100 ms PDSD (%) cellulose-xylan cellulose-lignin lignin-lignin AcMe AcCO AcMe AcMe G4 S1a S1b S3/5 OMe S1a S1a S1b S1b S3/5 S3/5 G4 G4 OMe OMe S1a S1a S1a S1a S1b S1b S1b S1b S1b G4 i4 s4 s4 s6 i4 i4 i4 i4 i4 s4 s6 s4 s6 s4 s6 s4 s6 s4 s6 G2a G3 G4 G5 G2a G2b G3 G4 G5 S1a 245 Gated 1s PDSD (%) 2.16 3.37 1.76 2.13 1.80 2.36 2.51 1.11 0.90 1.78 4.26 1.55 5.32 3.21 6.04 2.52 4.01 1.89 3.07 2.15 2.88 1.98 2.62 1.86 1.27 1.85 1.81 1.31 4.36 Type m m w m w m m w w w s w s m s m s w m m m w m w w w w w s Table S7.8 (cont’d) xylan-lignin G4 G4 G4 S3/5 S3/5 S3/5 AcCO AcCO AcCO AcCO AcCO AcCO AcCO AcCO AcCO G4 S1a S1b S3/5 OMe AcMe AcMe AcMe AcMe AcMe AcMe AcMe G4 S1a S1b S3/5 OMe G4 G4 G4 S1a S1a S1a S1b S1b S1b S3/5 S3/5 S3/5 S1b S2/6a S3/5 G3 G4 G5 G2a G3 G4 G5 S1a S1b S2/6b S3/5 OMe AcCO AcCO AcCO AcCO AcCO G2b G4 S1b S2/6b S3/5 S4 OMe AcMe AcMe AcMe AcMe AcMe Xn42f Xn43f Xn52f, 3f Xn42f Xn43f Xn52f, 3f Xn42f Xn43f Xn52f, 3f Xn42f Xn43f Xn52f, 3f 246 2.97 1.44 1.62 1.17 1.67 3.67 9.76 2.48 1.13 2.20 2.02 1.28 1.12 1.38 1.37 0.57 2.44 2.71 5.09 3.91 2.48 6.49 4.58 6.99 0.80 0.65 0.51 4.91 3.08 1.39 3.48 3.48 3.81 5.00 2.87 3.91 2.37 3.93 2.55 2.43 3.97 3.58 1.72 2.28 5.16 2.15 3.36 4.66 w m s m w m m w w w w w m m s m m s s s w w w s m w m m m s s s m m m m m s w m s m m s Table S7.8 (cont’d) mixed sugar- lignin OMe OMe OMe G4 S1a S3/5 OMe Xn42f Xn43f Xn52f, 3f i/s/Xn12f i/s/Xn12f i/s/Xn12f i/s/Xn12f 1.88 1.56 2.09 2.47 1.95 3.29 2.53 4.13 2.33 2.92 5.04 11.46 s s s m m s s 247 Table S7.9. The Intensities of intermolecular cross peaks of poplar spruce. In sum, 94 intermolecular interactions are identified, including 51 strong, 9 medium, and 34 weak restraints. In gated 1-s PDSD spectra, a peak higher than 4% is categorized as a strong restraint (s, in bold), between 2% - 4% for medium (m) restraint, and below 2% for weak (w) restraint. All the peaks shown in the gated 100-ms PDSD spectrum are categorized as strong restraints. The intensity is a relative ratio of the peak, which is normalized by the integral of the 13C cross-section. Atom1 (ꞷ1) Atom2 (ꞷ2) Gated 100 ms PDSD (%) cellulose-xylan cellulose- lignin AcCO AcCO AcMe AcMe AcCO AcCO AcMe AcMe G1a G1a G1b,c G1d G3a G3a G3c G4a,b,c OMe G1a G1a G1b,c G1b,c G1d G3a G3a G3c G3c G4a,b,c G4a,b,c OMe OMe i4 i6 i4 i6 s4 s6 s4 s6 i4 i6 i4 i4 i4 i6 i4 i4 i4 s4 s6 s4 s6 s4 s4 s6 s4 s6 s4 s6 s4 s6 248 0.96 2.06 1.21 2.23 1.12 0.69 0.36 0.68 1.03 0.78 2.66 3.98 2.36 1.22 2.94 2.36 1.07 1.98 1.17 2.32 3.63 4.01 Gated 1s PDSD (%) 0.66 3.06 1.38 2.08 3.08 13.45 3.92 9.84 2.50 2.26 1.10 1.52 2.93 1.29 1.47 4.48 9.61 2.20 5.98 2.39 8.96 1.82 1.44 5.05 2.21 8.07 3.57 6.63 9.99 23.53 Type w m w m s s s s s s s s s s w s s m s s s w w s s s s s s s Table S7.9 (cont’d) xylan-lignin AcCO AcCO AcCO AcCO AcCO AcCO AcCO AcCO AcCO AcCO AcCO AcCO G1a G1b,c G1d G3a G3c G4a,b,c OMe AcMe AcMe AcMe AcMe AcMe AcMe AcMe AcMe AcMe AcMe G1a G1b,c G1d G3a G3c G4a,b,c OMe G1a G1a G1b,c G1b,c G1d G1d G1d G3a G1a G1b,c G1d G2a,c G2d G3a G3b G3c G3d G4a,b,c G6c OMe AcCO AcCO AcCO AcCO AcCO AcCO AcCO G1a G1b,c G2a,c G3a G3c G3d G4a,b,c G4d G6c OMe AcMe AcMe AcMe AcMe AcMe AcMe AcMe Xn42f Xn43f Xn42f Xn43f Xn42f Xn43f Xn52f, 3f Xn42f 249 0.73 0.67 0.93 1.98 1.95 4.90 0.11 2.61 0.24 3.22 1.57 7.82 6.31 3.06 1.92 5.09 2.09 1.55 3.49 1.33 0.30 1.65 2.26 3.00 0.67 1.66 1.67 1.64 5.37 3.40 1.36 1.20 0.41 0.72 1.59 2.29 3.88 1.13 2.75 2.27 2.09 3.06 6.47 2.82 w w w w w s w m w m w s s s w s s s s w w w m m w w w w s m w w w w w s s w s s s s s s 1.06 2.59 2.36 1.29 2.22 1.03 0.69 1.19 1.49 1.20 0.86 0.98 1.12 1.21 1.51 2.67 1.36 Table S7.9 (cont’d) mannan-lignin mannan-xylan mixed sugar- lignin G3a G3a G3c G3c G4a,b,c G4a,b,c G4a,b,c OMe OMe G3c OMe AcCO G1a G1b,c G1d G3a G3c G4a,b,c OMe OMe Xn43f Xn52f, 3f Xn42f Xn43f Xn42f Xn43f Xn52f, 3f Xn42f Xn43f Man1 Man1 Man1 i/s/Xn12f i/s/Xn12f i/s/Xn12f i/s/Xn12f i/s/Xn12f i/s/Xn12f i/s/Xn12f Xn2/33f,Xn32f,i2/5,s2/3 1.54 1.11 2.01 1.86 0.87 0.99 4.69 3.96 1.32 0.66 2.61 3.39 4.14 2.46 4.43 1.94 4.48 2.48 2.79 13.7 9.92 1.19 2.59 2.04 0.94 0.60 0.58 1.53 1.43 1.53 5.99 15.46 s s s w s s s s s w m s w w w w w s s s 250 Table S7.10. Intermolecular interactions of polymers in intact plant stems. In the summary of intermolecular cross-peaks between different polymers of the secondary cell wall of woods, a total of 272 cross-peaks have been identified and categorized as 112 strong, 75 medium, and 85 weak interactions. Eucalyptus Poplar Spruce Interactions cellulose-xylan cellulose-lignin lignin-lignin xylan-lignin mixed sugar- lignin cellulose-xylan cellulose-lignin lignin-lignin xylan-lignin mixed sugar- lignin cellulose-xylan cellulose-lignin lignin-lignin xylan-lignin Mannan-lignin Mannan-xylan mixed sugar- lignin s 0 8 4 22 11 0 4 2 14 2 4 18 0 25 0 1 3 m 0 4 11 13 1 3 5 6 17 2 2 1 0 5 1 0 0 Total 112 75 w 4 8 4 7 1 1 6 8 10 0 2 3 0 23 1 0 5 85 Total 4 20 19 42 13 4 15 16 41 4 8 22 0 53 2 1 8 272 251 Table S7.11. Water-edited intensities of polysaccharide and lignin of Eucalyptus. The intensity ratio is obtained by comparing the water-edited and control 2D 13C-13C correlation spectra. Standard deviations of NMR signal-to-noise ratios are used as error bars. The numbers (bold) in parentheses indicate the average intensities under each type of the polysaccharide or lignin. Type interior cellulose (0.24) surface cellulose (0.36) xylan2f (0.31) xylan3f (0.44) lignin (0.51) cross peaks i4-1 i4-3a i4-3b,c i4-2/5 i4-6 i6-1 i6-4 i6-3a,b i6-3c i6-2/5,2 s4-1 s4-3/5 s4-2 s4-6 Xn2f4-1 Xn2f4-3 Xn2f4-5 Xn2f4-2 Xn3f2/3-1 Xn3f2/3-4 Xn3f2/3-5 Xn3f5-1 Xn3f5-4 S3/5-S1b S3/5-S1a S3/5-S2/6a S3/5-S2/6b S3/5-OMe S’3/5-4 S’3/5-2, G4-2 S’3/5-OMe, G4-OMe S4-S3/5 S4-S2/6 S4-S1 S4-OMe 252 Intensity 0.44±0.03 042±0.05 0.42±0.03 0.42±0.09 0.40±0.07 0.48±0.19 0.4±0.2 0.5±0.1 0.5±0.1 0.5±0.1 0.55±0.09 0.6±0.1 0.6±0.1 0.5±0.1 0.52±0.09 0.55±0.08 0.47±0.09 0.7±0.2 0.7±0.2 0.6±0.2 0.6±0.2 0.5±0.1 0.5±0.1 0.50±0.08 0.48±0.03 0.61±0.04 0.49±0.09 0.53±0.08 0.61±0.09 0.5±0.1 0.5±0.1 0.46±0.09 0.44±0.08 0.47±0.08 0.51±0.09 Table S7.11 (cont’d) lignin OMe (0.42) S2/6-S3/5 S2/6-S1 S2/6-S4 S2/6-OMe OMe-S3/5 OMe-G3 OMe-S1 OMe-S4 OMe-S2/6 0.6±0.2 0.5±0.2 0.5±0.1 0.51±0.09 0.42±0.06 0.43±0.07 0.40±0.09 0.5±0.1 0.4±0.1 253 Table S7.12. Water-edited intensities of polymers in poplar. The intensity ratio is obtained by comparing the water-edited and control 2D 13C-13C correlation spectra. Standard deviations of NMR signal-to-noise ratios are used as error bars. The numbers (bold) in parentheses indicate the average intensities under each type of the polysaccharide or lignin. Type interior cellulose (0.11) surface cellulose (0.11) xylan2f (0.30) xylan3f (0.25) guaiacyl (0.19) syringyl (0.16) cross peaks i4-1 i4-3a i4-3b,c i4-2/5 i6-1 i6-4 i6-3a,b i6-3c i6-2/5 s4-1 s4-3/5 s4-2 s4-6 Xn42f-1 Xn42f-3 Xn42f-2 Xn42f-5 Xn3f2/3-1 Xn3f2/3-4 Xn3f2/3-5 Xn3f5-1 Xn3f5-4 G5-3 G5-1 G5-OMe G2-3 G2-1 G2-6 G2-OMe S3/5-1b S3/5-1a S3/5-2/6 S3/5-OMe S1b-S3/5 S1b-S2/6 S1b-OMe 254 Absolute Intensity 0.11±0.04 0.09±0.07 0.09±0.07 0.10±0.08 0.12±0.05 0.12±0.03 0.11±0.03 0.11±0.04 0.11±0.02 0.12±0.06 0.10±0.04 0.11±0.04 0.11±0.03 0.36±0.06 0.42±0.06 0.20±0.05 0.21±0.06 0.23±0.05 0.39±0.09 0.15±0.03 0.31±0.07 0.18±0.09 0.14±0.04 0.14±0.03 0.18±0.02 0.28±0.06 0.16±0.05 0.21±0.04 0.21±0.03 0.16±0.09 0.21±0.06 0.15±0.04 0.14±0.05 0.14±0.07 0.20±0.06 0.15±0.06 Table S7.12 (cont’d) syringyl/guaicyl OMe (0.14) S1a-S3/5 S1a-OMe S2/6-S3/5 S2/6-S1b S2/6-OMe OMe-S3/5 OMe-G3 OMe-S1 OMe-G5 OMe-S2/6 0.13±0.03 0.11±0.03 0.17±0.04 0.20±0.06 0.11±0.03 0.13±0.04 0.14±0.03 0.14±0.04 0.15±0.05 0.15±0.03 255 Table S7.13. Water-edited intensities of polysaccharides and lignin of spruce. The intensity ratio is obtained by comparing the water-edited and control 2D 13C-13C correlation spectra. Standard deviations of NMR signal-to-noise ratios are used as error bars. The numbers (bold) in parentheses indicate the average intensities under each type of the polysaccharide or lignin. Type interior cellulose (0.38) surface cellulose (0.43) xylan2f (0.57) xylan 3f (0.40) mannan (0.76) guaiacyl (0.47) guaiacyl OMe (0.42) cross peaks i4-1 i4-3a i4-3b,c i4-2/5 i6-1 i6-4 i6-3 i6-2/5,2 s4-1 s4-2/5 s4-3 s4-6 Xn2f4-1 Xn2f4-3 Xn2f4-5 Xn2f4-1 Xn2f4-2 Xn3f2/3-1 Xn3f2/3-4 Xn3f2/3-5 Xn3f5-4 Xn3f5-1 Man1-4 Man1-5 Man1-6 G3-1 G3-2 G3-OMe G1-3 G6-3 G6-OMe G2-3b,c G2-3a G2-OMe OMe-G1 OMe-G2 256 Absolute Intensity 0.44±0.02 0.37±0.04 0.37±0.03 0.39±0.03 0.42±0.07 0.4±0.1 0.32±0.05 0.4±0.1 0.5±0.1 0.44±0.03 0.4±0.2 0.46±0.03 0.7±0.1 0.43±0.04 0.57±0.04 0.67±0.05 0.56±0.06 0.4±0.1 0.32±0.07 0.4±0.1 0.44±0.05 0.48±0.04 0.86±0.06 0.63±0.06 0.79±0.05 0.4±0.1 0.47±0.04 0.52±0.09 0.5±0.1 0.40±0.06 0.35±0.03 0.51±0.06 0.65±0.05 0.54±0.15 0.40±0.06 0.42±0.07 Table S7.14. 13C-T1 relaxation times of lignin and polysaccharides in the three woods. The data are fit using single exponential equations 𝐼 (𝑡) = 1 − 2𝑒−𝑡/𝑇for T1, DP (inversion recovery) and 𝐼 (𝑡) = 𝑒−𝑡/𝑇 for T1, CP (Torchia CP). Standard deviations of the fitting parameters are used as error bars. Unidentified (-). Eucalyptus Poplar Spruce T1, Atom (ppm) CP (s) 174 2.1±0.1 - - 148 4.6±0.2 145 4.7±0.3 136 4.5±0.2 134.5 4.7±0.3 - - - - 57 4.6±0.3 105 4.1±0.1 89 4.8±0.2 84 3.9±0.1 65 3.7±0.3 62 2.7±0.3 102.5 3.4±0.2 AcCO S3/5 G3 G4/ S’3/5 S1 S4 G1 G6 OMe i/s/Xn2f1 i4 s4 i6 s6 Xn3f1 Man Xn2f4 Xn3f4 - - 82 3.7±0.2 78 3.6±0.2 Xn2f/3fAcMe 21.5 3.9±0.1 Atom (ppm) 174 153 - 145 136 134.5 - - 57 105 89 84 65 61 102.5 - 82 78 21.5 T1, DP (s) 1.6±0.2 3.1±0.5 - 1.8±0.2 2.2±0.3 1.9±0.3 - - 2.5±0.2 2.3±0.7 4.4±0.2 2.3±0.3 2.9±0.3 - 2.1±0.3 - 1.4±0.2 1.2±0.2 2.2±0.2 Atom (ppm) 174 - 148 145 - 134.5 - - 57 105 89 84 65 62 102.5 - 82 78 21.5 T1, CP (s) 2.83±0.05 - 4.2±0.2 4.1±0.1 - 4.2±0.1 - - 4.4±0.3 4.1±0.1 4.8±0.2 3.7±0.2 3.6±0.3 2.7±0.3 3.4±0.2 - 3.7±0.2 3.6±0.2 4.7±0.2 Atom (ppm) 174 153 - 145 136 134.5 - - 56.5 105 89 84 65 - 102.5 - 82 78 21.5 T1, DP (s) 2.4±0.2 1.3±0.1 - 2.4±0.2 2.7±0.2 2.7±0.2 - - 2.8±0.2 3.7±0.2 4.3±0.2 3.1±0.2 3.4±0.3 - 2.2±0.2 - 2.1±0.4 1.4±0.3 1.5±0.1 Atom (ppm) 174 - 148 145 - - 130.8 120.5 57 105 89 84 65 62 102.5 100.2 81.8 78 21.5 T1, CP (s) 1.4±0.2 - 3.7±0.1 3.6±0.1 - 2.9±0.1 3.3±0.1 3.4±0.3 3.45±0.04 3.97±0.06 3.12±0.09 3.0±0.2 2.0±0.2 2.7±0.1 2.3±0.1 2.8±0.1 2.5±0.1 2.1±0.2 Atom (ppm) 174 - 148 145 - - 130.8 120.5 57 105 89 84 65 - 102.5 100 82 78 21.5 T1, DP (s) 1.3±0.1 - - 2.4±0.2 - 0.9±0.1 2.6±0.3 2.1±0.2 3.2±0.2 3.9±0.2 1.5±0.2 2.4±0.3 - 1.7±0.2 0.8±0.1 1.1±0.2 0.8±0.1 1.1±0.1 257 Table S7.15. 1H-T1ρ relaxation times of lignin and polysaccharides in three samples. The data are fit using a single exponential equation: 𝐼 (𝑡) = 𝑒−𝑡/𝑇 . Standard deviations of the fitting parameters are used as error bars. Unidentified (-). Eucalyptus Spruce Poplar Atom (ppm) 153.9 - - - 148.1 - 145.6 135.9 - 130.1 - 115.3 - 110.9 57 105.0 89.0 84.2 65.0 61.5 102.5 - 82.0 78.1 21.0 T1 (ms) 16.8±0.8 - - - 10±1 - 11.4±0.8 12±1 - 11.1±0.7 - 8.0±0.7 - 10.1±0.9 14.2±0.7 28.1±0.9 35±1 22 ±1 26±2 16±1 12.8±0.9 - 14.6±0.9 11.5±0.9 17.6±0.6 Atom (ppm) - 131 134 136 152.2 147.6 - - 134.2 - 119.0 116.3 113.5 - 57 105.0 89.0 84.1 65.0 61.5 102.3 101 82.0 78.0 21.0 S3/5 G1a G1b/c G1d G3a G3c S’3/5/G4 S4 G1 S’b4 G6 G5 G2 S’1 OMe i/s/Xn12f i4 s4 i6 s6 Xn13f Man1 Xn42f Xn43f Xn2f,3fAcMe T1 (ms) - 8.3±0.6 11.6±0.5 10.1±0.7 13±1 12.8±0.9 - 8.0±0.5 8.9±0.7 - 8.8±0.6 8.2±0.7 8.3±0.6 - 12.5±0.7 29±2 35±1 20±1 26±2 15±1 12±1 10.4±0.9 13.7±0.9 10.2±0.9 13±1 Atom (ppm) 153.2 - - - 148.1 - 145.0 - - - - 116.2 - - 57 105.0 89.0 84.5 64.8 61.5 102.4 - 81.8 78.0 21.5 T1 (ms) 19.5±0.6 - - - 14±1 - 13.8±0.9 - - - - 9.1±0.8 - - 15.1±0.8 36±2 43±2 28±2 32±2 19±1 13.0±0.8 - 17±1 12±1 19.5±0.5 258 Table S7.16. Parameters of 1D solid-state NMR experiments measured on the three wood samples. Recycle delay (d1); number of scans (NS); number of points of time domain for the direct (td2) and indirect (td1) dimensions; the acquisition time of the direct dimension (aq2); the evolution time of indirect dimension (aq1); excitation frequency for proton (ν1H) and carbon (ν13C) channels. Quantitative DP 13C-T1 Torchia Field (T) Temp. (K) MAS (kHz) d1(s) NS td2 aq2 (ms) ν1H (kHz) ν13C (kHz) CP (1H/13C) (kHz) CP contact time (ms) 1H decoupling (kHz) CP 14.1 294 14 1.7 128 2048 14 83.3 62.5 62.5 1 DP 14.1 294 14 2 128 2800 19.6 - 62.5 - - 14.1 294 14 35/40a 64 2800 19.6 - 62.5 - - 83.3 83.3 83.3 13C-T1 Inversion- recovery 9.4 300 10 30 128 1600 16 - 62.5 - - 83.3 1H-T1ρ 1D Water- edited 9.4 300 10 2 256/128/64b 1400 14 62.5 - 51c 1.0 62.5 9.4 278 10 1.8 512 1400 14 83.3 62.5 62.5 1.0 83.3 9.4 300 10 2 256 1600 16 62.5 62.5 62.5 1 62.5 Processing GM (-7, 0.07) GM (-7, 0.07) GM (-7, 0.07) GM (-7, 0.07) GM (-7, 0.07) GM (-5, 0.1) GM (-5, 0.1) aRecycle delays are 40s (for eucalyptus) and 35s (for the other two). bTo protect the probe during longer LG-SL time (14 ms and 19 ms), smaller NSs are used (128/64). cFor 1H-T1ρ, 51 kHz for CP matching means w1eff is 62.5 kHz during LG-SL and LG-CP. 259 Table S7.17. Parameters of 2D NMR experiments measured on the three wood samples. Recycle delay (d1); number of scans (NS); number of points of time domain for the direct (td2) and indirect (td1) dimensions; the acquisition time of the direct dimension (aq2); the evolution time of indirect dimension (aq1); excitation frequency for proton (ν1H) and carbon (ν13C) channels. Field (T) Temp. (K) MAS (kHz) d1(s) NS td2 td1 aq2 (ms) aq1 (ms) ν1H (kHz) ν13C (kHz) CP (1H/13C) (kHz) CP contact time (ms) 1H decoupling (kHz) Mixing time (ms) CP-INADQ DP- INADQ RFDR DARR DIPSHIFT DP- PDSD HSQC 14.1 294 14 1.6 32 1400 340 14 4.5 83.3 62.5 62.5 1 14.1 294 14 1.6 16 1800 112 18 4.5 - 62.5 - - Dipolar- gated PDSD 14.1 294 14 1.48 160 1400 180 14 3.42 83.3 62.5 62.5 2D Water- edited 9.4 278 10 1.65 128/256 a 1400 152 14 4.94 83.3 62.5 62.5 9.4 298 10 2.0 128 1600 280 16 7.00 71.4 62.5 62.5 9.4 278 10 1.65 128 1400 152 14 4.94 83.3 62.5 62.5 9.4 298 7.5 2.0 128 1600 13 16 0.14 83.3 62.5 62.5 1 1 1 1 1 9.4 298 12 1.7 192 1600 200 16 4.20 - 62.5 - - 11.8 - - 1 224 2048 256 12.78 3.91 31.3 20.8 - - 83.3 83.3 71.4 83.3 83.3 83.3 83.3 83.3 31.3 - - 1.6 20/100/1 000 b 50 50 - 100 0.86c 260 Table S7.17 (cont’d) Processing QSINE (ssb:2.8) QSINE (ssb:3) GM (-10, 0.05) QSINE (ssb:4)e GM (-15, 0.04) QSINE (ssb:3)e GM (-15, 0.04) QSINE (ssb:4.5) e GM (-30, 0.03) QSINE (ssb:4.5) e GM (-5, 0.1) GM (-30, 0.03) QSINE (ssb:4)e QSINC (ssb:3.5) QSINE (ssb:3.5)d a 128 NS for control one and 256 NS for 4ms-SD water edited one. b 1000ms/100ms for long and short mixing time and spruce also has 20-ms mixing time spectrum. c 1/(8JCH). d Along direct and indirect dimensions respectively. For the later forward linear prediction is applied with 32 coefficients. e GM for lignin region and QSINE for carbohydrate region. 261 Table S7.18. Parameters of ssNMR experiments measured for never-dried spruce and eucalyptus samples. The same experiments with same parameters were measured again for both two wood samples after lyophilization and rehydration. Never-dried spruce was involved in all 1D and 2D experiments, and never-dried eucalyptus was only involved in 1D experiments. Recycle delay (d1); number of scans (NS); number of points of time domain for the direct (td2) and indirect (td1) dimensions; the acquisition time of the direct dimension (aq2); the evolution time of indirect dimension (aq1); excitation frequency for proton (ν1H) and carbon (ν13C) channels. CP DP Quantitative DP 13C-T1 Torchia 1H-T1ρ Field (T) Temp. (K) MAS (kHz) d1(s) NS td2 td1 aq2 (ms) aq1 (ms) ν1H (kHz) ν13C (kHz) CP match (1H/13C) (kHz) CP contact time (ms) 9.4 298 10 2 1k 1600 1 16 - 83.3 62.5 62.5 1.0 9.4 298 10 2 1k 1600 1 16 - - 62.5 - - 9.4 298 10 35 512 1600 1 16 - - 62.5 - - 9.4 298 10 2 512 1600 1 16 - 83.3 62.5 62.5 1.0 9.4 298 10 2 512/256/128a 1600 1 16 - 62.5 - 51b 1.0 262 2D Water- edited CP- INADQ DP- INADQ 53-ms CORD 1D Water- edited 9.4 280 10 2 1k/512 1400 1 14 - 83.3 62.5 9.4 280 10 1.7 96 1600 166 16 4.15 83.3 62.5 9.4 298 10 1.8 192 1400 90 14 5.31 83.3 62.5 62.5 62.5 62.5 1.0 1.0 1.0 9.4 298 10 2 192 1600 90 16 5.31 - 62.5 - - 9.4 298 10 2 128 1600 166 16 4.15 83.3 62.5 62.5 1.0 Table S7.18 (cont’d) 1H decoupling (kHz) Processing 83.3 83.3 83.3 83.3 62.5 83.3 83.3 83.3 83.3 83.3 GM (-5, 0.1) GM (-10, 0.05) GM (-10, 0.05) GM (-5, 0.1) GM (-5, 0.1) GM (-5, 0.1) GM (-20, 0.04) QSINE (ssb:4)c QSINE (ssb:4) QSINE (ssb:4) GM (-30, 0.03) a To protect the probe during longer LG-SL time (14 ms and 19 ms), smaller NSs are used (256/128). b For 1H-T1ρ, 51 kHz for CP matching means w1eff is 62.5 kHz during LG-SL and LG-CP. c GM for lignin region and QSINE for carbohydrate region. 263 Table S7.19. Comparison of 13C chemical shifts with literature-reported values. The comparison results are shown for carbohydrates, lignin aromatic carbons as well as covalent linkers. The references are indicated for each comparison. Carbohydrate Eucalyptus Poplar Spruce reference 1 reference 2 Wang et al. 2014107 Kang et al. 2019108 ia ib ic sf C1 C2 C3 C4 C5 C6 C1 C2 C3 C4 C5 C6 C1 C2 C3 C4 C5 C6 C1 C2 C3 C4 104.9 72.1 75.2 88.7 72.3 65.0 104.9 72.1 75.2 88.7 72.3 65.0 104.9 72.1 74.1 87.9 71.1 65.7 104.9 72.1 74.6 84.4 104.8 72.1 75.0 88.9 72.4 64.9 104.8 72.1 75.0 88.9 72.4 64.9 104.8 72.1 74.3 87.9 70.1 65.5 104.8 72.1 75.7 84.3 104.8 72.2 75.3 89.0 72.4 65.0 104.8 72.2 75.3 89.0 72.4 65.0 104.8 72.2 74.3 88.1 71.0 65.6 104.8 72.2 75.1 84.5 105.8 71.5 75.8 89.1 72.5 64.9 Kang et al. 2019 105.1 72.9 74.1 89.1 72.5 65.0 Kang et al. 2019 104.2 71.1 - 87.9 - 65.8 Kang et al. 2019 105.2 72.7 75.8 84.3 105.0 72.9 75.8 89.2 72.9 65.2 Wang et al. 2014 105.0 72.5 75.6 88.8 72.5 65.1 Wang et al. 2014 105.0 72.5 75.4 88.8 72.5 65.1 Wang et al. 2014 105.0 72.5 75.6 84.7 264 reference 3 Gao et al. 2020109 105.2 72.8 75.1 88.1 72.8 64.8 Gao et al. 2020 105.2 72.8 75.8 89.0 72.6 65.4 Gao et al. 2020 105.2 72.8 75.8 89.5 73.0 65.3 Gao et al. 2020 105.2 72.8 75.1 83.8 Table S7.19 (cont’d) C5 C6 C1 C2 C3 C4 C5 C6 C1 C2 C3 C4 C5 ACCO ACMe C1 C2 C3 C4 C5 ACCO ACMe 74.6 62.3 104.9 72.1 73.6 83.5 73.6 61.4 104.8 7.2 74.0 82.2 64.0 174.0 21.0 102.5 73.4 74.1 77.4 63.3 174.0 21.0 75.0 62.3 75.0 62.1 104.8 72.1 75.2 83.4 73.7 61.3 104.8 72.1 73.2 82.0 63.9 174.0 21.0 102.2 73.6 74.2 77.9 64.1 174.0 21.0 104.8 72.2 74.3 84.3 74.3 61.4 104.8 72.2 73.3 82.2 63.9 174.0 21.0 103.5 73.9 74.2 77.8 65.4 174.0 21.0 75.7 62.9 Wang et al. 2014 105.0 72.5 75.6 84.7 75.9 62.0 75.4 62.5 Kang et al. 2019 106.1 73.0 73.1 83.5 74.6 61.5 Simmons et al. 201625 Kang et al. 2019 105.2 72.3 75.2 82.2 64.3 173.6 21.6 Simmons et al. 2016 102.6 73.7 74.7 77.4 63.9 173.6 21.6 Kang et al. 2019 105.1 72.5 73.5 82.3 64.6 174.0 21.0 Kang et al. 2019 102.5 73.5 74.5 78.0 62.5 174.0 21.0 sg Xn2f Xn3f Xn 75.1 61.8 Gao et al. 2020 105.2 72.8 75.4 84.1 75.4 62.2 Gao et al. 2020 105.1 72.3 75.3 82.1 64.1 174.0 21.4 Gao et al. 2020 102.5 73.7 74.7 77.3 63.8 174.0 21.4 265 Table S7.19 (cont’d) C1 C2 C3 C4 C5 ACCO ACMe C1 C2 C3 C4 C5 C1 C2 C3 C4 C5 C1 C2 C3 C4 C5 C6 - - 74.2 78.5 64.1 101.5 71.3 79.8 69.9 76.6 99.2 69.9 - - - 98.5 71.6 80.8 - - - / / / / / 102.5 73.5 74.5 78.0 62.5 174.0 101.5 71.3 79.8 69.9 76.6 101.5 71.3 79.8 69.9 76.6 99.2 69.9 - - - 98.5 71.6 80.8 - - - 99.2 69.9 - - - 98.5 71.6 80.8 - - - Duan et al. 2021 111 99 72 Simmons et al. 2016 98.6 72.4 Wang et al. 2014 98.6 72.4 78.3 - 72.6 177.5 Wang et al. 2014 GlcAa GlcAb GlcAc GalA 266 Simmons et al. 2016 98.6 72.4 Terrett et al. 2019112 Phyo et al. 2017113 Table S7.19 (cont’d) C1 C2 C3 C4 C5 C6 C1 C2 C3 C4 C5 C6 C1 C2 C3 C4 C5 C6 C1 C2 C3 C4 C5 C1 C2 100.2 68.1 - 79.6 - - 100.2 68.1 - 79.6 - - / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / 100.2 68.1 - 79.6 - - 101.2 71.3 75.1 - 75.1 62.1 101.6 72.1 75.1 80.7 75.1 62.1 108.1 82.1 - 85.6 62.1 107.8 81.8 107.8 81.8 107.8 81.8 M MAc A(t-Ara) Aa 100.2 69.3 - - - 101.1 69.8 - - - 99.9 69.0 69.8 78.8 71.3 171.6 Terrett et al. 2019 101.9 72.0 - 80.4 75.8 61.6 Terrett et al. 2019 108.6 82.1 78.3 86.1 62.8 Phyo et al. 2017 107.2 87.8 Wang et al. 2014 100.2 69.3 - - - - Terrett et al. 2019 100.9 71.9 75.9 80.4 75.8 61.6 Wang et al. 2014 110.0 82.2 77.7 84.9 62.3 Wang et al. 2014 107.1 89.8 267 Duan et al. 2021 109 82 Table S7.19 (cont’d) C3 C4 C5 C1 C2 C3 C4 C5 C1 C2 C3 C4 C5 C1 C2 C3 C4 C5 77.2 84.6 61.8 107.8 80.1 - - - 109.7 81.8 77.4 82.9 67.6 106.9 85.3 - - - 77.2 84.6 61.8 107.8 80.1 - - - 109.7 81.8 77.4 82.9 67.6 / / / / / 77.2 84.6 61.8 107.8 80.1 - - - 109.7 81.8 77.4 82.9 67.6 / / / / / Ab Ac Ad Lignin G Eucalyptus Poplar Spruce C1 - / / 76.8 78.6 70.7 Wang et al. 2014 107.3 87.8 72.6 - - Wang et al. 2014 110.0 82.2 77.7 84.9 62.3 Wang et al. 2014 109.2 82.1 78.2 85.7 62.4 reference 1 Kang et al. 2019 131.6 76.2 82.4 67.3 Phyo et al. 2017 108.4 85.8 81.1 81.9 67.0 Phyo et al. 2017 108.4 81.8 77.8 83.1 67.8 Phyo et al. 2017 107.7 82.2 77.5 84.9 62.1 reference 2 Ralph et al. 2009 130.5 Other solution NMR studies 268 Table S7.19 (cont’d) C2 111.1 / / - 112.2 C3 C4 148.9 144.9 C5 115.9 / / / / / / 148.2 148.2 150.3 150.1 115.5 112.2 111.5 96 110.8 97 111.7 114 111.6 89 110.9 99 110.9 100 110.9 101 111.4 102 111.4 103 110.9 88 111.0 100 110.3 104 115.2 96 115.4 114 115.4 89 114.7 99 114.9 100 115.1 101 115.0 102 115.4 103 114.6 88 114.9 105 115.2 104 269 119.5 96 118.8 97 119.6 114 119.5 89 119.0 99 119.0 100 118.8 101 119.5 102 119.3103 119.188 118.6 105 118.9 104 Table S7.19 (cont’d) C6 - / / 119.6 120.8 56.1 OMe 57.0 / / C1 C2 C3 C4 C5 C6 OMe C1 C2 C3 C4 C5 C6 OMe C1 C2 / / / / / / / / / / / / / / / / - 112.1 48.2 144.9 116.1 - 56.6 - 113.8 148.2 144.9 116.1 - 56.6 / / 130.7 113.4 152.2 145.3 116.3 - 56.7 133.5 110.5 149.7 146.1 - 120.2 57.0 134.2 113.5 56.9 Kang et al. 2019 132.1 114.5 146.2 146.2 114.5 - 56.7 Kang et al. 2019 134.8 111.5 147.4 147.4 115.8 120.4 57.0 Kang et al. 2019 134.8 111.5 270 Ga Gb Gc Table S7.19 (cont’d) C3 C4 C5 C6 OMe C1 C2 C3 C4 C5 C6 OMe C1 C2 C3 C4 C5 C6 OMe C1 C2 C3 C4 C5 C6 OMe / / / / / / / / / / / / 134.9 102.8 153.3 133.5 153.3 102.8 57.0 - 109.7 153.2 137.2 153.2 109.7 57.0 / / / / / / / / / / / / 134.2 104.0 153.1 134.6 153.1 104.0 56.6 137.1 102.5 152.7 135.8 152.7 102.5 56.6 C1 134.1 / 147.6 144.1 - 119.2 57.0 136.0 109.5 151.6 142.2 - 121.1 57.0 / / / / / / / / / / / / / / / Gd Sa Sb S’a 147.4 147.4 115.8 120.4 57.0 Kang et al. 2019 134.8 111.5 147.4 147.4 115.8 120.4 57.0 Kang et al. 2019 136.2 105.5 152.7 134.5 152.7 105.5 56.7 Kang et al. 2019 136.2 105.5 152.7 134.5 152.7 105.5 56.7 Rencoret et al. 2011 Yuan et al. 2011 271 Table S7.19 (cont’d) C2 C3 C4 C5 C6 OMe C1 C2 C3 C4 C5 C6 OMe 109.2 144.8 136.3 144.8 109.2 56.5 136.5 112.5 145.5 138.5 145.5 112.5 56.5 / / / / / / / / / / / / / / / / / / / / / / / / / / 106.6 - - - 106.6 - 106.2 - - - 106.2 - Rencoret et al. 2011 Yuan et al. 2011 106.6 - - - 106.6 - 106.2 - - - 106.2 - S’b Lignin Linkages Eucalyptus Poplar Spruce C7(α) 72.3 72.0 72.4 A(G) / A(S) C8(β) 84.5 86.3 84.3 C9(γ) 59.4 60.4 59.9 Chemical shifts from solution NMR spectra Eucalyptus 71.6 / 72.5 96 - / 71.8 97 71.4 / 72.6 114 71.7 / 72.489 - / 71.8 99 84.1 / 86.5 96 83.6 / 85.7 97 84.1 / 86.4; 87.3 114 84.1 / 86.4 89 83.6 / 85.8 99 60.0 96 59.4 97 60.1 114 59.6 99 Poplar 71.5 / 71.9 101 - / 72.0 102 - / 71.8 103 - / 71.8 88 83.9 / 85.9 100 83.4 / 85.8 101 - / 85.8 103 83.8 / 85.8 88 59.6 100 59.8 101 60.2 103 272 Spruce 71.1 104 71.0 106 84.0 104 83.9 106 59.9 104 59.8 106 Table S7.19 (cont’d) C7(α) 84.4 84.8 - C8(β) 53.5 52.6 53.9 C9(γ) 71.2 70.5 70.3 C7(α) C8(β) C9(γ) C7(α) C8(β) C9(γ) C7(α) C8(β) C9(γ) - 53.5 74.4 81.1 59.6 / 75.8 78.4 59.8 83.9 52.9 74.1 80.3 59.7 / 77.2 79.0 60.7 / / / 81.4 60.0 / 76.2 78.9 60.1 C7(α) / / 88.1 B B’ E V C 84.8 100 85.2 101 85.7 102 84.7 103 84.8 88 53.5 100 53.6 101 53.6 103 71.0 100 70.1 101 81.2 100 59.7 100 86.8 100 87.0 101 87.4 102 87.1 103 86.8 88 85.4 96 84.7 97 85.4 114 85.5 89 84.8 14 53.7 96 53.3 97 54.1 114 53.5 99 71.7 96 71.0 97 71.7 114 71.3 99 81.7 96 81.0 97 60.3 96 59.7 97 87.7 96 86.4 97 87.5 114 87.4 89 86.7 99 273 84.9 104 85.0 106 53.5 104 53.6 106 70.8 104 70.9 106 81.7 106 59.3 106 75.8 115 78.0 115 87.0 104 87.0 106 Table S7.19 (cont’d) C8(β) C9(γ) C7(α) C8(β) C9(γ) / / / / / / / / / / 54.2 / 90.2 49.9 / C’ 53.7 96 53.1 97 53.7 114 53.5 90 63.3 96 62.3 97 62.5 99 53.3100 53.2 101 53.2 102 62.5100 62.5 101 53.1 104 53.0 106 62.8 104 62.7 106 274 CHAPTER 8: SOLID-STATE NMR OF UNLABELED PLANT CELL WALLS: HIGH- RESOLUTION STRUCTURAL ANALYSIS WITHOUT ISOTOPIC ENRICHMENT Research paper reprinted with permission: Biotechnol. Biofuels 14:14, (2021) Authors: Wancheng Zhao, Alex Kirui, Fabien Deligey, Frederic Mentink-Vigier, Yihua Zhou, Baocai Zhang, and Tuo Wang 8.1 Abstract Multidimensional solid-state nuclear magnetic resonance (ssNMR) spectroscopy has emerged as an indispensable technique for resolving polymer structure and intermolecular packing in primary and secondary plant cell walls. Isotope (13C)-enrichment provides feasible sensitivity for measuring 2D/3D correlation spectra, but this time-consuming procedure and its associated expenses have restricted the application of ssNMR in lignocellulose analysis. Here, we present a method that relies on the sensitivity-enhancing technique Dynamic Nuclear Polarization (DNP) to eliminate the need for 13C-labeling. With a 26-fold sensitivity enhancement, a series of 2D 13C-13C correlation spectra were successfully collected using the unlabeled stems of wild-type Oryza sativa (rice). The atomic resolution allows us to observe a large number of intramolecular cross peaks for fully revealing the polymorphic structure of cellulose and xylan. NMR relaxation and dipolar order parameters further suggest a sophisticated change of molecular motions in a ctl1 ctl2 double mutant: both cellulose and xylan have become more dynamic on the nanosecond and microsecond timescale, but the motional amplitudes are uniformly small for both polysaccharides. By skipping isotopic labeling, the DNP strategy demonstrated here is universally extendable to all lignocellulose materials. This time-efficient method has landed the technical foundation for understanding polysaccharide structure and cell wall assembly in a large variety of plant tissues and species. 275 8.2 Introduction The past decade has witnessed the rapid advances in multidimensional solid-state NMR (ssNMR) capabilities that have enabled high-resolution characterization of intact plant cell walls. This spectroscopic method provides a wealth of atomic-level information on the conformational structure of polysaccharides, covalent linkage patterns of matrix polysaccharides, dynamical profile and water contact, as well as cellulose-matrix packing on the subnanometer scale1. With a rapidly expanding territory, from eudicotyledons (Arabidopsis thaliana) to commelinid monocotyledons (Zea mays, Brachypodium distachyon, etc.)2-5, from primary to secondary cell walls6-10, and from plants to algal and fungal species11-13, ssNMR has progressively evolved into a vital tool for characterizing carbohydrate-rich biomaterials. The molecular information of cell wall architecture can serve as the structural basis for improving the current technologies of biofuel production using lignocellulosic biomass14. Enriching the cell walls with NMR-active isotopes (such as 13C and 15N) is a prerequisite for measuring two- and three-dimensional (2D/3D) correlation experiments, which provides the spectral resolution required for resolving numerous carbon and nitrogen sites in cell wall polymers. Two strategies can be employed: plants can be grown in the dark, using a medium containing 13C- labeled glucose2; otherwise, 13CO2 can be continuously supplied to the plants grown in a day-night cycle4, 15. Depending on the developmental stage and the tissue of interest, labeling can be time- consuming and costly. In-vitro replication procedures also weaken the merit of ssNMR as an analytical technique targeting natural tissues. The recent development of Magic-Angle Spinning Dynamic Nuclear Polarization (MAS- DNP) methods has presented a unique opportunity for circumventing these drawbacks16-19. MAS- DNP enhances NMR sensitivity by tens to hundreds of folds, which allows us to skip isotope- 276 enrichment and use unlabeled samples to measure 2D 13C-13C/15N correlation spectra for high- resolution structural characterization20-22. Regarding the plant biomass, three exploratory studies have been conducted to reveal the restructuring effect of ball-milling on cotton cellulose23, the alternation of xylan conformations induced by genetic mutations of rice24, and the compositional changes of lignin in high-S and low-S poplar25. With the rapid development of DNP instrumentation26-29 and radical design30-33, this is certainly a direction of great potential but not yet explicitly explored for plant materials. This methodology article aims at establishing a universally applicable toolbox for characterizing polymer structure and assembly in unlabeled plant biomass. This is achieved by combining a series of DNP-enabled experiments that probe the composition and conformational structure of polymers with conventional ssNMR measurements that examine the rate and amplitudes of molecular motions. Implementation of this method will expand the ssNMR capabilities and enable high-resolution investigations of unlabeled cell walls, which, at least in part, provides a replacement and upgrade to the conventional methods that rely on 13C-enriched materials. Most of the structural aspects previously investigated using 13C-enriched samples, such as the composition, conformation, packing, and motion of cell wall polymers, can be studied using unlabeled materials via a blend of ssNMR and MAS-DNP methods (Table 8.1). This technical advance will eliminate the threshold that has long been impeding lignocellulose characterization, which will immediately benefit the research communities of plant biology, biomaterials, and bioenergy. 277 Table 8.1 Capabilities of solid-state NMR and DNP. Technical aspects are categorized as being fully capable (Yes), partially limited by insufficient resolution or sensitivity (In part), and infeasible or unsuitable (No). Samples and Polymer Structural Intermolecular Molecular Water Techniques Composition Polymorph Packing Motion Contact 13C-labeled; ssNMR Unlabeled; ssNMR Unlabeled; MAS-DNP Yes Yes Yes Yes Yes In part No No In part In part Yes Yes In part Noa Noa aMAS-DNP is conducted at cryogenic temperature (~100 K); therefore, it is unsuitable for investigating molecular motions. It is also better to investigate polymer-water contacts at room temperature. 8.3 Results 8.3.1 Sensitivity Enhancement by MAS-DNP For decades, one-dimensional (1D) 13C spectra have been conducted using unlabeled materials to analyze the structure and composition of polysaccharides at natural isotope abundance (1.1% for 13C)34-37. Figure 8.1 shows the 1D quantitative spectra of four rice samples including the wild-type (WT) stems, two single mutants, which are ctl1 (harboring an identical mutation of brittle culm 15) and ctl2, as well as a double mutant ctl1 ctl2. These mutations happen to rice CHITINASE-LIKE1 (CTL1) and its homolog CTL2, which have been suggested a role in controlling cellulose biosynthesis thereby affecting wall integrity38-41, as ctl1 and ctl1 ctl2 plants 278 exhibited obvious brittleness phenotypes. Quantitative detection of all carbons is achieved using a recently developed pulse sequence, MultiCP, which counts on 1H T1 relaxation to repeatedly restore 1H magnetization between the many CP blocks included in each individual scan of the experiment42-44. The rice stems are collected from the field, without any dissolution procedures or chemical digestions; therefore, the cell walls being analyzed are still native and intact. Various peaks of cellulose (interior chains: i; surface chains: s) and xylan (Xn) have been observed (Figure 8.1a), indicating the predominance of secondary cell wall components in these mature stems. Different from the WT stems, the double mutant has a higher amount of methyl ether (- OMe) groups, which is a chemical substitution frequently occurring in lignin. The ctl1 ctl2 sample shows a higher peak at 56 ppm for methyl ether groups (Figure 8.1a) but comparable intensities for aromatic carbons (Figure S8.1). This observation indicates a higher fraction of guaiacyl (G) and syringyl (S) units that are heavily substituted by methyl ether groups. The ctl1 ctl2 sample has shown stronger xylan peaks, for example, the carbon 1 of 3-fold xylan at 102 ppm (Xn13f), the carbon 4 of two-fold xylan at 82 ppm (Xn42f), and the methyl carbon of acetyl group at 21 ppm (AcMe). Subtraction of the WT and ctl1 ctl2 spectra has generated a difference spectrum, which contains well-resolved signals from 2-fold and 3-fold xylan, revealing a higher amount of xylan in the double mutant. The spectral patterns of the two single mutants are intriguing: the spectrum of ctl1 resembles that of the double mutant while ctl2 is similar to the WT sample (Figure 8.1b). Combined with the cell wall defects detected in bc1541, the results have suggested a stronger role of the ctl1 mutation in modulating cell wall composition and structure. While useful information can be obtained by following this conventional 1D analysis, only a few carbon sites can be resolved. This resolution issue can be partially alleviated by spectral subtraction partially, but significant improvement is still needed for characterizing these polysaccharides and cell walls. 279 Figure 8.1 1D 13C spectra of unlabeled cell walls of wild-type rice and mutants. a, 1D 13C MultiCP spectra of wild-type sample (WT; black) and ctl1 ctl2 double mutant (orange) yielding quantitative detection of all molecules. Abbreviations of carbohydrate assignments are shown in Figure 8.3, except for xylan acetyl group (Ac). The difference spectrum is obtained by subtraction of the two parent spectra after normalization by i4 peaks. The difference spectrum mainly contains xylan signals, revealing the higher hemicellulose content in ctl1 ctl2 mutant. The insert picture on the right top side is the representative photo of the two kinds of rice stems. b, 1D 13C MultiCP spectra of ctl1 and ctl2 single mutants. The spectral pattern of ctl1 mutant is similar to that of the ctl1 ctl2 double mutant while ctl2 is closer to the WT sample. The rice material is then impregnated in a solvent mixture of 13C-depleted, d8- glycerol/D2O/H2O, which contains 10 mM of biradical AMUPol30. 13C-depletion of glycerol is a necessity for investigating unlabeled samples where both biomolecules and glycerol are at natural 13C abundance: it can effectively eliminate the glycerol signals that overlap with the carbohydrate signals. This protocol, when compared with a matrix-free method that was recently applied to cotton23, 45, 46, better retains polymer hydration in cell walls. The payoff is a moderate decline in the signal-to-noise ratios because the solvent occupies some volume in the MAS rotor, which will 280 otherwise be filled with more biomolecules. Figure 8.2 Sensitivity enhancement from DNP. a, DNP enhances the NMR sensitivity by 26-fold on the wild-type sample. This value is obtained by quantifying the intensity ratio between the spectra collected with microwave on (MW on) and microwave off (MW off). The DNP-enhanced spectrum of ctl1 ctl2 (orange dashline) is compared with the wild-type (WT) sample. b, Lignin region of the room-temperature spectrum (RT; top), DNP-enhanced spectrum (MW on; middle), and non-DNP spectrum (MW off; bottom). The experimental time is labeled for each panel. The “×2” symbol means the intensities of the room-temperature spectrum is manually increased by two folds. DNP significantly saves experimental time and enhances NMR sensitivity. c, Spectral deconvolution of the interior cellulose carbon 4 (i4) and surface cellulose carbon 4 (s4) region (dashline square in panel a) reveals the crystallinity change in rice cellulose. The experimentally measured spectra (black) is overlaid with the simulated spectra (blue), with the deconvoluted peaks underlying the spectral envelope. The interior-to-surface (i/s) ratio of cellulose is quantified for both samples, which has shown an increase in cellulose crystallinity in the ctl1 ctl2 mutant. The fit parameters are listed in Tables S8.1 and S8.2. Under microwave irradiation, polarization of the electrons in the biradical will thereafter 281 be transferred to 1H spins in unlabeled rice stem, and then to the natural abundance 13C for detection. The sensitivity enhancement (εon/off) is 26-fold for the wild-type sample and 22-fold for the ctl1 ctl2 double mutant (Figure 8.2a, Figure S8.2a), which respectively shortens the experimental time by factors 676 and 484 times. Unlike the optimal performance of 13C-depleted, d8-glycerol/D2O/H2O mixture, a sample prepared using a different solvent of d6-DMSO/D2O/H2O only shows an enhancement of 5-fold, thus being excluded from the protocol (Figure S8.2b). The same gain of sensitivity also occurs to lignin. We have observed strong signals of aromatic carbons (120-160 ppm) within a short time of 0.5 h (Figure 8.2b). These signals, however, are invisible in conventional ssNMR even after 42.5 h of measurements at room temperature. The DNP experiment preferentially detects the molecules with a highly ordered structure, for example, the cellulose microfibrils in plants (Figure S8.3) and the highly microcrystalline chitin in fungi10, 12. At the cryogenic temperature (~100 K) of MAS-DNP, the conformational distribution of mobile molecules, such as xylan and small molecules, will be fully trapped, resulting in broader lines and lower intensity. Recently, concerns have also been raised regarding the radical distribution in cell walls, which were supposed to induce a biased detection of molecules in spatial proximity to the radicals. This argument is invalid because all molecules within tens of nanometer to a radical can be efficiently polarized via 1H-1H spin diffusion, which is due to the inherently long relaxation times of the 1Hs in cell walls47. This mechanism ensures a homogenous polarization of all molecules throughout the cell wall, which has been confirmed by the consistent spectral envelopes with and without microwave irradiation (Figure S8.4). In addition, the crystallinity of cellulose becomes lower in the double mutant. Cellulose crystallinity is evaluated using the intensity ratio between the carbon 4 of interior glucan chains (i4; 89 ppm) and the carbon 4 of surface glucan chains (s4; 84 ppm) in cellulose. When normalized 282 by the i4 signal, the s4 peak is lower in the double mutant than in the WT sample as consistently observed at room temperature (Figure 8.1a) and under the DNP condition (Figure 8.2a). Spectral deconvolution allows us to resolve the underlying resonances and quantify the interior-to-surface (i/s) ratios, which are 0.68 for WT and 0.81 for ctl1 ctl2 (Figure 8.2c, Figure S8.5, Tables S8.1 and S8.2). The higher cellulose crystallinity in the double mutant may originate from an increased degree of cellulose bundling in the cell wall. The observation here actually counters previous X- ray analysis of Arabidopsis, which has shown reduced content of crystalline cellulose in a ctl1 ctl2 mutant40. The discrepancy can be attributed to the different plant species or the distinct structural aspects being examined in diffraction and spectroscopic techniques for determining the ambiguously defined “cellulose crystallinity”48. Our unpublished X-ray diffraction data on the same rice stems have observed a higher relative crystallinity index (RCI) for cellulose in the double mutant., which has confirmed the NMR observation. 8.3.2 2D 13C-13C Correlation Spectra of Unlabeled Cell Walls Benefiting from the substantial gain of NMR sensitivity, we have successfully collected a 2D 13C-13C J-INADEQUATE spectrum49, 50 using unlabeled stems of WT rice (Figure 8.3a). Because of the low abundance (1.1%) of 13C in nature and the even lower probability (0.01%) for observing a cross peak between two carbons, such a 2D experiment has long been deemed impossible, but can now be finished now within 35 h using the rice stems. We have observed eight types of glucose units in cellulose: types a-e for the glucan chains embedded in the core of microfibrils and types f-h for those exposed on the surface. Their 13C chemical shifts are consistent with those observed in Arabidopsis, Zea mays, and Brachypodium distachyon, revealing the polymorphic nature of cellulose structure51. Hemicellulose shows weak signals, for example, the carbon 4 and carbon 5 in three-fold xylan (Xn43f and Xn53f) as well as 283 the carbon 3 and carbon 4 in two-fold conformers (Xn32f and Xn42f). In addition, the α-xylose (x) of xyloglucan also exhibits some weak signals, indicative of a small portion of primary cell walls. Besides polysaccharides, we have also observed some self-correlation cross peak of the CH2 acyl chains in lipid polymers. Therefore, the remarkable resolution of the DNP-enabled 2D spectrum allows us to unambiguously resolve many carbon sites in cellulose, matrix polysaccharides, and lipids, which is otherwise impossible. The validity of natural-abundance MAS-DNP is confirmed by the consistent spectral pattern between unlabeled (Figure 8.3a) and 13C-labeled (Figure 8.3b) rice stems. In addition, a few peaks are omitted in the natural-abundance DNP spectrum. These signals are primarily from the arabinose residues of xylan sidechains, some carbon sites in 2-fold, 3-fold and mixed forms of xylan backbones, as well as the acetyl groups (Figure 8.3c). Peak counting has confirmed MAS- DNP’s preferential detection of ordered molecules: we can effectively detect all cellulose carbons and part of xylan backbones, but not for arabinose sidechains (Figure 8.3d). These mobile molecules, when trapped at 100 K, bear a wide distribution of conformations, which has broadened out their signals. While the J-INADEQUATE experiment only probes through-bond correlations, we have also performed a CHHC experiment to detect long-range and through-space correlations (Figure 8.4a). The CHHC sequence relies on 1H-1H spin diffusion to transfer polarization, which is followed by cross-polarization (CP) from 1H to 13C for site-specific detection52. The Full Width at Half Maximum (FWHM) linewidths are around 3 ppm and the representative signal-to-noise ratios are ranging from 4 to 41 (Figure S8.6). The off-diagonal cross peaks are mostly intramolecular correlations within cellulose, for example, the i6-3 cross peak at (65, 75 ppm) and the s3-6 cross peak at (75, 62 ppm) in the 1-ms CHHC spectrum. Elongating the mixing time to 2 ms allows us 284 to extend the reach; therefore, many additional cross peaks now become observable (Figure 8.2b). The new signals are primarily from cellulose, such as the i6-1 cross peak at (65, 105 ppm) and i2,5-4 cross peak at (72, 89 ppm). We have also observed some signature signals of xylan, for example, Xn2f2-4 at (73, 82 ppm) and Xn2f3-4 at (78, 82 ppm). Figure 8.3 DNP enables 2D correlation spectroscopy using unlabeled and native cell walls. a, 2D 13C–13C INADEQUATE spectrum of unlabeled WT rice collected on a 600 MHz/395 GHz DNP system. DNP effectively detects and resolves the signals from cellulose and hemicellulose at natural isotope abundance. Superscripts are used to denote the eight types of glucose units in interior and surface cellulose and the 2-fold and 3-fold conformations in xylan. Inset shows the lipid signals that are folded in the indirect dimension. b, 2D 13C–13C INADEQUATE spectrum of 13C labeled WT rice stems measured on a 400 MHz NMR spectrometer. The spectral pattern is comparable to the DNP-assisted natural-abundance spectrum in panel a. c, Arabinose and acetyl signals are detected using 13C-labeled samples but become invisible in the DNP spectrum of unlabeled materials. d, Number of peaks detected using labeled and unlabeled samples, which are tabulated in Table S8.3. Asterisks indicate the components that are poorly detected using DNP. 285 Figure 8.4 DNP-enabled 2D CHHC spectra of unlabeled rice stems. The CHHC spectra of unlabeled WT sample are measured with a, 1 ms and b, 2 ms 1H-1H mixing times. c, Summary of through-space intramolecular cross peaks. Solid lines denote the cross peaks observable in both 1 and 2 ms spectra while dashlines indicate the cross peaks that are observed only in the 2-ms spectrum. Arrows are used to indicate whether the cross peaks are bidirectional (e.g. C1-to-C2 and C2-to-C1 cross peaks) or unidirectional (e.g. only from C1 to C2). d, Comparison of polysaccharide signals in the 2-ms CHHC spectrum of the unlabeled sample and a 5-ms PDSD spectrum of 13C-labeled rice. All the inter-carbon correlations are summarized in Figure 8.4c. For cellulose, the large 286 number of cross peaks allows us to assign the 13C chemical shifts of all carbons in interior and surface glucan chains. The flat-ribbon 2-fold xylan has shown 4 distinguishable signals while the non-flat 3-fold xylan only exhibits a single cross peak from carbon 1 to carbon 2 (Xn3f1-2). The spectral pattern of CHHC resembles those PDSD spectra that measured using 13C-labeled plant biomass (Figure 8.4d, Figure S8.7). Therefore, natural-abundance MAS-DNP has the full capability of investigating cellulose structure; it is also partially prepared for investigating matrix polysaccharides. It should be noted that the CHHC experiment is specifically chosen for natural-abundance MAS-DNP. For most other 2D correlation methods, such as PDSD and DARR, the diagonal peaks will dominate the spectrum of unlabeled materials, because the probability of observing off- diagonal internuclear correlation is very low at natural 13C abundance. The CHHC experiment can effectively avoid this issue due to the 13C-1H-1H-13C pathway used for polarization transfer53. Alternatively, dipolar homonuclear recoupling schemes could also be used for characterizing unlabeled biomaterials22, 54. 8.3.3 Motional Rates and Amplitudes of Cell Wall Polysaccharides To understand the effect of ctl1 ctl2 mutation on polymer dynamics, we have measured 13C-T1 and 1H-T1ρ relaxation times using the unlabeled stems of WT and ctl1 ctl2 samples. As the most rigid component of cell walls, cellulose exhibits very slow 13C-T1 relaxation (typically on the order of 10-40 s), revealing its lack of motion on the nanosecond timescale (Figure 8.5a, b, Table S8.4 and Figure S8.8). In contrast, the 13C-T1 of 3-fold xylan, the conformer that fills the interfibrillar space, is shorter by 4-5 times: 7.7 s for carbon 1 and 10 s for carbon 4 in the WT sample. The 2-fold xylan, a unique form induced by its packing with cellulose surface, has shown an intermediate 13C-T1 of 10.5 s for its resolved peak of carbon 4. These 13C-T1 time constants are 287 substantially longer than the values reported for uniformly 13C-labeled materials (4-6 s for cellulose and 1-2 s for matrix polysaccharides)55. This can be attributed to the lack of spin exchange between rigid and mobile motifs in unlabeled materials. The 1H-T1ρ data is highly consistent with the 13C-T1 results. Cellulose remains as the most rigid component on the slower, microsecond timescale, with a representative 1H-T1ρ time of 20-50 ms in WT stems (Figure 8.5c, d). The 2-fold and 3-fold conformers of xylan have more pronounced dynamics, showing 1H-T1ρ values of 18 ms and 8-16 ms, respectively. Interestingly, most polysaccharides have considerably faster 13C-T1 and 1H-T1ρ relaxations in the ctl1 ctl2 mutant, indicative of a flawed cell wall formed by polysaccharides that are highly mobile on both nanosecond and millisecond timescales. To examine the motional amplitudes of these polysaccharides, we have conducted the 13C- 1H dipolar-chemical shift (DIPSHIFT) correlation experiment (Figure 8.6)56. The 13C-1H dipolar couplings (in kHz) are measured at representative carbon sites of cellulose and xylan, which are further converted to dipolar order parameters. A near-unity value indicates restricted motional amplitudes for the C-H bonds while a small order parameter suggests large-amplitude motions. Strikingly, all polysaccharides are highly rigid and exhibit large order parameters above 0.90 (Figure 8.6b). This is very different from the results collected using Arabidopsis and Brachypodium primary cell walls, in which the order parameters ranges from 0.3 to 0.6 for matrix polymers5, 55. Cellulose has identical order parameters in WT and ctl1 ctl2 samples. The hemicellulose, however, has shown subtle changes: the order parameter has slightly decreased for the carbon 1 of 3-fold xylan (Xn13f) but increased for the carbon 4 of 2-fold xylan (Xn42f). These two peaks are the best-resolved signals of xylan, and the discrepancy observed here suggests a small extent of phase separation in the double mutant: the cellulose-packed 2-fold xylan becomes 288 even more rigid in the mutant while the interfibrillar 3-fold conformer undergoes even larger amplitude motions. The relaxation and DIPSHIFT experiments can systematically assess the motional characteristics of polysaccharides, which complement the 2D MAS-DNP data to provide a complete view of polymer composition, structure, and dynamics in unlabeled biomass. Figure 8.5 Carbohydrate dynamics probed by NMR relaxation using unlabeled rice stems. a, Representative 13C-T1 relaxation curves of cellulose (top row) and xylan (second row). The ctl1 ctl2 double mutant has shown substantially faster relaxation than the wild-type sample, indicating more pronounced motions on the nanosecond timescale. b, Bar diagram summarizing the 13C-T1 relaxation times of cellulose (pink) and xylan (blue). The x-axis represents different carbon sites, which can be found in Table S8.4. c, Representative 1H-T1ρ relaxation curves. For most peaks, the double mutant shows faster 1H-T1ρ relaxation than the wild-type sample, indicating enhanced 289 Figure 8.5 (cont’d) dynamics on the microsecond timescale. All the 1H-T1ρ relaxation time constants are summarized in panel d. Figure 8.6 Dipolar order parameters of polysaccharides in unlabeled rice cell walls. a, The first slice of the 2D DIPSHIFT experiment measured under 7.5 kHz MAS, which shows the signals of cellulose and xylan. b, 13C−1H dipolar evolution curves extracted from 2D DIPSHIFT experiment. The fit dipolar coupling values (in kHz) and the calculated dipolar order parameters are labeled for each panel. Most carbon sites, except for some xylan peaks, have shown comparable order parameters in the wild-type and double mutant samples. 8.4 Discussion Using a common agriculture lignocellulosic biomass, rice stems, we have demonstrated the feasibility of employing ssNMR and MAS-DNP methods to investigate the structure and dynamics of cell wall polysaccharides without 13C enrichment. The strategy and the experiments described above can be directly applied to many other biomass samples. The method is time-efficient: the total experimental time is 254 h for the wild-type stems, which includes 132 h of measurements on a 600 MHz/395 GHz DNP instrument and 122 h on a conventional 400 MHz ssNMR spectrometer. 290 Multiple factors can impact the experimental time, for example, the mixing protocol that determines biradical distribution in cell wall materials, the sample pH that affects the lifetime and stability of radicals, and the choice of DNP solvent. As the beginning step, it is crucial to collect a set of 1D spectra to screen several samples prepared using different protocols and conditions, each of which only takes a few minutes to measure. By monitoring the absolute sensitivity and the DNP gain, the optimal condition will be identified for measuring 2D experiments. At present, natural-abundance DNP still has major limitations. First, it is impossible to derive a detailed pattern of intermolecular contacts using only unlabeled materials (Table S8.1). This is due to the demanding requirement of sensitivity for probing longer-range correlations, which can even be challenging for some 13C-labeled samples. Progress is being made: long-range correlations, typically on the range of 3-5Å, have been observed recently using model cyclic peptides22. Once successfully extended to cellular samples, these techniques will fully enable structural determination. Partial information could be obtained if the polymer contact depends on the conformational structure. For example, the flat-ribbon structure of 2-fold xylan is responsible for coating cellulose surface; therefore, we have tracked its amount to estimate the extent of xylan- cellulose packing in unlabeled rice stems24. Second, molecules with a high disorder suffer from intensity loss under MAS-DNP conditions. Fortunately, we still have distinguishable carbon sites for different xylan conformers, which can be used for structural and compositional analysis. Tremendous efforts have been devoted to further improving MAS-DNP efficiency. Multiple biradicals have been developed recently, such as the AsymPolPOK family that shortens the DNP buildup time33 and the Tinypol and TEMTRIPol-I radicals that have shown improved performance over AMUPol on high-field (18.8 and 21.1 Tesla) DNP instruments31, 57, 58. The resolution improvement at high-field MAS-DNP (e.g. 800 MHz/527 GHz) will allow us to better 291 analyze biomolecular structures, at least for the molecules bearing structural order59, 60. The rapidly evolving MAS-DNP technology has a great potential to bring biomass analysis to a new level. 8.5 Materials and Methods 8.5.1 Rice Stem Preparation The mutant ctl1 was generated by backcrossing a previously reported mutant brittle culm 15 (A213L mutation in OsCTL1) into Oryza sativa cv. Nipponbare background41. The ctl2 is an insertional mutant (NC2596 from rice Tos17 insertion mutant database) with a Tos17 insertion at the 1633 bp downstream of the OsCTL2 (LOC_Os08g41100) coding sequence. The double mutant ctl1 ctl2 was created by crossing ctl1 and ctl2 and then screening out from the F2 progenies via molecular identification. All the WT and mutant rice plants (Oryza sativa) were grown in the experimental fields at the Institute of Genetics and Developmental Biology in Beijing (China). At least five mature stems from different plants of each genotype were harvested for the measurement. 8.5.2 Dynamic Nuclear Polarization Sample Preparation The unlabeled rice stems were sliced using a razor into pieces that are on the dimension of a few millimeters. The materials were mixed with a stock solution, which contains 10 mM AMUPol radical30 and a solvent mixture of 13C-depleted, d8-glycerol/D2O/H2O (60/30/10 vol%). Around 50 mg of the rice material was impregnated in 100 μL of the AMUPol solution and then ground for 20-30 min, which allows the radicals to penetrate the porous cell walls. Around 33 mg of the plant material was transferred to a 3.2-mm sapphire rotor for DNP experiments. In addition, a different sample has been prepared using DMSO/D2O/H2O (60/30/10 vol%) solvent to evaluate the effect of different solvents on DNP efficiency. A video protocol can be found in reference61. 8.5.3 Dynamic Nuclear Polarization Experiment All DNP experiments were conducted on a 600 MHz (14.1 Tesla)/395 GHz MAS-DNP 292 instrument at National High Magnetic Field Laboratory, Tallahassee, Florida27. The DNP spectra were collected using a 3.2-mm probe under 8 kHz MAS. All 13C chemical shifts were reported on the TMS scale. The microwave irradiation is around 12 W. The temperature was 104 K with microwave irradiation and decreased to 94 K when the microwave was turned off. The DNP build up time was measured to be 2.80 s for the wild-type sample and 2.63 s for the ctl1 ctl2 sample. Therefore, the recycle delay was typically 3.6 and 3.2 s for the wild-type and ctl1 ctl2 samples, respectively. 1D 13C CP spectra were collected using a 1 ms contact time. A total number of 256 and 512 scans were collected on the wild-type, and ctl1 ctl2 samples, respectively. The total experimental time ranges from 16 to 29 mins for each sample. Spectral deconvolution of the cellulose peaks was performed using DmFit software62. Two types of 2D 13C-13C correlation experiments were conducted on the rice stems: a 2D CP J-INADEQUATE experiment that probes through-bond correlations49 and a 2D CHHC experiment that probes spatial correlations52. For the INADEQUATE spectra, a total number of 320 scans were recorded within 35 h for the wild-type sample. The indirect dimension includes 110 points. CHHC spectrum was much more time-consuming and was only collected on the wild- type sample. For 1 ms CHHC, a total of 320 scans were collected over 36 h, with 64 points in the indirect dimension, and for 2 ms CHHC, a total of 608 scans were collected over 60 h, with the indirect dimension varying from 50- 60 points. The CHHC experiment relied on 3 CP blocks to transfer polarization from 1H to 13C, back to 1H to enable 1H spin diffusion, and then back to 13C for site-specific detection. The corresponding contact times of the three CP blocks were 1 ms, 0.5 ms, and 0.5 ms, respectively. 8.5.4 Room-Temperature Solid-State NMR Multiple experiments were conducted at room-temperature to compare with the DNP data. 293 Around 65 mg of unlabeled rice stems were packed into a 4-mm ZrO2 rotor for measurements. All non-DNP experiments were conducted on a 400 MHz (9.4 Tesla) Bruker spectrometer using a 4- mm HCN probe. 1D 13C quantitative spectra were collected using the recently developed MultiCP pulse sequence at room temperature under 10 kHz 43. This experiment allows for the quantitative detection of all 13C signals with enhanced sensitivity. Within each experiment, 8 CP blocks were used for WT and ctl1 ctl2 double mutant sample, and 9 blocks were used for ctl1 and ctl2 single mutant rice, with a z-filter time of 0.9 s between two CP blocks for repolarization. In total, 20k, 23k, 23k, and 20k scans were collected for the WT, ctl1, ctl2, and ctl1 ctl2 samples, respectively. To investigate the molecular motion of polysaccharides in unlabeled rice stems, we have measured 13C-T1 and 1H-T1ρ relaxation. 13C-T1 was measured using a CP-based experiment63. 1H-T1ρ was measured using a spinlock field of 62.5 kHz. For 13C T1 relaxation, 4k and 3k scans are collected for each time point of WT and ctl1 ctl2 samples, respectively. Meanwhile, 6k (WT) and 2k (ctl1 ctl2) scans are collected for each time point of the 1H-T1ρ relaxation experiment. In addition, we have conducted the dipolar-chemical shift (DIPSHIFT) correlation experiment under 7.5 kHz MAS56. Frequency-switched Lee Goldburg (FSLG) sequence was used for 1H homonuclear decoupling, with a transverse field of 83 kHz and an effective field of 102 kHz. The scaling factor is 0.577 as verified using a model tripeptide MLF. In addition, we have measured a series of 2D PDSD experiments using 13C-labeled rice stems to compare with the DNP-assisted 2D CHHC experiment collected on unlabeled samples. Around 85 mg 13C-labeled rice material was packed into a 4-mm ZrO2 rotor for the solid-state NMR experiment. The mixing times were chosen to be 1, 2, 3, and 5 ms for four different spectra. 8.6 Conclusion We have presented a novel strategy that integrates natural-abundance DNP with dynamics 294 measurement to analyze unlabeled plant cell walls. Because this protocol no longer requires isotopic enrichment, it now becomes possible to rapidly screen a large collection of lignocellulose materials found in nature or engineered in vitro. Consequently, the large-scale and high-resolution biomass characterization will provide the structural foundation for improving the biotechnologies of biofuel production. 8.7 Acknowledgments The DNP analysis of plant cell walls is supported by the U.S. Department of Energy (grant no. DE-SC0021210). Room-temperature experiments and spectral deconvolution were conducted by FD, who is supported as part of the Center for Lignocellulose Structure and Formation, an Energy Frontier Research Center funded by the U.S. Department of Energy, Office of Science, Basic Energy Sciences, under Award number DE-SC0001090. YZ and BZ were supported by the National Nature Science Foundation of China (31571247). The National High Magnetic Field Laboratory is supported by National Science Foundation through NSF/DMR-1644779 and the State of Florida. The MAS-DNP instrument at NHMFL is funded in part by NIH S10 OD018519, NIH P41-GM122698-01, and NSF CHE-1229170. 295 REFERENCES 1. Zhao, W.; Fernando, L. D.; Kirui, A.; Deligey, F.; Wang, T., Solid-state NMR of plant and fungal cell walls: A critical review. Solid State Nucl. Magn. Reson. 2020, 107, 101660. 2. Dick-Perez, M.; Zhang, Y. A.; Hayes, J.; Salazar, A.; Zabotina, O. A.; Hong, M., Structure and interactions of plant cell wall polysaccharides by two- and three-dimensional magic-angle- spinning solid-state NMR. Biochemistry-Us 2011, 50 (6), 989-1000. 3. Wang, T.; Zabotina, O.; Hong, M., Pectin-cellulose interactions in the Arabidopsis primary cell wall from two-dimensional magic-angle-spinning solid-state nuclear magnetic resonance. Biochemistry-Us 2012, 51 (49), 9846-9856. 4. Wang, T.; Chen, Y. N.; Tabuchi, A.; Cosgrove, D. J.; Hong, M., The Target of beta-Expansin EXPB1 in Maize Cell Walls from Binding and Solid-State NMR Studies. Plant Physiol. 2016, 172 (4), 2107-2119. 5. Wang, T.; Salazar, A.; Zabotina, O. A.; Hong, M., Structure and dynamics of Brachypodium primary cell wall polysaccharides from two-dimensional 13C solid-state nuclear magnetic resonance spectroscopy. Biochemistry-Us 2014, 53 (17), 2840-2854. 6. Terrett, O. M.; Lyczakowski, J. J.; Yu, L.; Iuga, D.; Franks, W. T.; Brown, S. P.; Dupree, R.; Dupree, P., Molecular architecture of softwood revealed by solid-state NMR. Nat. Commun. 2019, 10 (1), 4978. 7. Grantham, N. J.; Wurman-Rodrich, J.; Terrett, O. M.; Lyczakowski, J. J.; Stott, K.; Iuga, D.; Simmons, T. J.; Durand-Tardif, M.; Brown, S. P.; Dupree, R.; Busse-Wicher, M.; Dupree, P., An even pattern of xylan substitution is critical for interaction with cellulose in plant cell walls. Nat. Plants 2017, 3 (11), 859-865. 8. Simmons, T. J.; Mortimer, J. C.; Bernardinelli, O. D.; Poppler, A. C.; Brown, S. P.; deAzevedo, E. R.; Dupree, R.; Dupree, P., Folding of xylan onto cellulose fibrils in plant cell walls revealed by solid-state NMR. Nat. Commun. 2016, 7, 13902. 9. Dupree, R.; Simmons, T. J.; Mortimer, J. C.; Patel, D.; Iuga, D.; Brown, S. P.; Dupree, P., Probing the Molecular Architecture of Arabidopsis thaliana Secondary Cell Walls Using Two- and Three-Dimensional 13C Solid State Nuclear Magnetic Resonance Spectroscopy. Biochemistry-Us 2015, 54 (14), 2335-2345. 10. Kang, X.; Kirui, A.; Dickwella Widanage, M. C.; Mentink-Vigier, F.; Cosgrove, D. J.; Wang, T., Lignin-polysaccharide interactions in plant secondary cell walls revealed by solid-state NMR. Nat. Commun. 2019, 10 (1), 347. 11. Arnold, A. A.; Bourgouin, J. P.; Genard, B.; Warschawski, D. E.; Tremblay, R.; Marcotte, I., Whole cell solid-state NMR study of Chlamydomonas reinhardtii microalgae. J. Biomol. NMR 2018, 70 (2), 123-131. 12. Kang, X.; Kirui, A.; Muszynski, A.; Widanage, M. C. D.; Chen, A.; Azadi, P.; Wang, P.; 296 Mentink-Vigier, F.; Wang, T., Molecular architecture of fungal cell walls revealed by solid-state NMR. Nat. Commun. 2018, 9, 2747. 13. Chrissian, C.; Camacho, E.; Kelly, J. E.; Wang, H.; Casadevall, A.; Stark, R. E., Solid-state NMR spectroscopy identifies three classes of lipids in C. neoformans melanized cell walls and whole fungal cells. J Biol Chem 2020. 14. Smith, P. J.; Wang, H. T.; York, W. S.; Pena, M. J.; Urbanowicz, B. R., Designer biomass for next-generation biorefineries: leveraging recent insights into xylan structure and biosynthesis. Biotechnol. Biofuels 2017, 10. 15. Gao, Y.; Mortimer, J. C., Unlocking the architecture of native plant cell walls via solid-state nuclear magnetic resonance. Methods Cell Biol. 2020, 160, 121-143. 16. Rossini, A. J.; Zagdoun, A.; Lelli, M.; Lesage, A.; Coperet, C.; Emsley, L., Dynamic Nuclear Polarization Surface Enhanced NMR Spectroscopy. Acc. Chem. Res. 2013, 46 (9), 1942-1951. 17. Ni, Q. Z.; Daviso, E.; Can, T. V.; Markhasin, E.; Jawla, S. K.; Swager, T. M.; Temkin, R. J.; Herzfeld, J.; Griffin, R. G., High Frequency Dynamic Nuclear Polarization. Acc. Chem. Res. 2013, 46, 1933-1941. 18. Lilly Thankamony, A. S.; Wittmann, J. J.; Kaushik, M.; Corzilius, B., Dynamic nuclear polarization for sensitivity enhancement in modern solid-state NMR. Prog Nucl. Magn. Reson. Sp. 2017, 102-103, 120-195. 19. Lee, D.; Hediger, S.; De Paepe, G., Is solid-state NMR enhanced by dynamic nuclear polarization? Solid State Nucl. Magn. Reson. 2015, 66-67, 6-20. 20. Smith, A. N.; Marker, K.; Hediger, S.; De Paepe, G., Natural Isotopic Abundance 13C and 15N Multidimensional Solid-State NMR Enabled by Dynamic Nuclear Polarization. The journal of physical chemistry letters 2019, 10 (16), 4652-4662. 21. Takahashi, H.; Lee, D.; Dubois, L.; Bardet, M.; Hediger, S.; De Paepe, G., Rapid Natural- Abundance 2D 13C-13C Correlation Spectroscopy Using Dynamic Nuclear Polarization Enhanced Solid-State NMR and Matrix-Free Sample Preparation. Angew. Chem. Int. Edit. 2012, 51 (47), 11766-11769. 22. Marker, K.; Paul, S.; Fernandez-de-Alba, C.; Lee, D.; Mouesca, J. M.; Hediger, S.; De Paepe, G., Welcoming natural isotopic abundance in solid-state NMR: probing pi-stacking and supramolecular structure of organic nanoassemblies using DNP. Chem. Sci. 2017, 8 (2), 974-987. 23. Kirui, A.; Ling, Z.; Kang, X.; Dickwella Widanage, M. C.; Mentink-Vigier, F.; French, A. D.; Wang, T., Atomic Resolution of Cotton Cellulose Structure Enabled by Dynamic Nuclear Polarization Solid-State NMR. Cellulose 2019, 26, 329-339. 24. Zhang, L.; Gao, C.; Mentink-Vigier, F.; Tang, L.; Zhang, D.; Wang, S.; Cao, S.; Xu, Z.; Liu, X.; Wang, T.; Zhou, Y.; Zhang, B., Arabinosyl Deacetylase Modulates the Arabinoxylan Acetylation Profile and Secondary Wall Formation. Plant Cell 2019. 297 25. Perras, F. A.; Luo, H.; Zhang, X.; Mosier, N. S.; Pruski, M.; Abu-Omar, M. M., Atomic- Level Structure Characterization of Biomass Pre- and Post-Lignin Treatment by Dynamic Nuclear Polarization-Enhanced Solid-State NMR. J. Phys. Chem. A 2017, 121 (3), 623-630. 26. Sergeyev, I. V.; Aussenac, F.; Purea, A.; Reiter, C.; Bryerton, E.; Retzloff, S.; Hesler, J.; Tometich, L.; Rosay, M., Efficient 263 GHz magic angle spinning DNP at 100 K using solid-state diode sources. Solid State Nucl. Magn. Reson. 2019, 100, 63-69. 27. Dubroca, T.; Smith, A. N.; Pike, K. J.; Froud, S.; Wylde, R.; Trociewitz, B.; Mckay, J.; Mentink-Vigier, F.; van Tol, J.; Wi, S.; Brey, W.; Long, J. R.; Frydman, L.; Hill, S., A quasi- optical and corrugated waveguide microwave transmission system for simultaneous dynamic nuclear polarization NMR on two separate 14.1 T spectrometers. J Magn Reson 2018, 289, 35-44. 28. Matsuki, Y.; Idehara, T.; Fukazawa, J.; Fujiwara, T., Advanced instrumentation for DNP- enhanced MAS NMR for higher magnetic fields and lower temperatures. J Magn Reson 2016, 264, 107-115. 29. Chaudhari, S. R.; Berruyer, P.; Gajan, D.; Reiter, C.; Engelke, F.; Silverio, D. L.; Coperet, C.; Lelli, M.; Lesage, A.; Emsley, L., Dynamic nuclear polarization at 40 kHz magic angle spinning. Phys Chem Chem Phys 2016, 18 (15), 10616-22. 30. Sauvee, C.; Rosay, M.; Casano, G.; Aussenac, F.; Weber, R. T.; Ouari, O.; Tordo, P., Highly Efficient, Water-Soluble Polarizing Agents for Dynamic Nuclear Polarization at High Frequency. Angew. Chem. Int. Edit. 2013, 52 (41), 10858-10861. 31. Lund, A.; Casano, G.; Menzildjian, G.; Kaushik, M.; Stevanato, G.; Yulikov, M.; Jabbour, R.; Wisser, D.; Renom-Carrasco, M.; Thieuleux, C.; Bernada, F.; Karoui, H.; Siri, D.; Rosay, M.; Sergeyev, I. V.; Gajan, D.; Lelli, M.; Emsley, L.; Ouari, O.; Lesage, A., TinyPols: a family of water-soluble binitroxidestailored for dynamic nuclear polarization enhancedNMR spectroscopy at 18.8 and 21.1 T. Chem. Sci. 2020, 11 2810. 32. Stevanato, G.; Casano, G.; Kubicki, D. J.; Rao, Y.; Esteban Hofer, L.; Menzildjian, G.; Karoui, H.; Siri, D.; Cordova, M.; Yulikov, M.; Jeschke, G.; Lelli, M.; Lesage, A.; Ouari, O.; Emsley, L., Open and Closed Radicals: Local Geometry around Unpaired Electrons Governs Magic-Angle Spinning Dynamic Nuclear Polarization Performance. J. Am. Chem. Soc. 2020. 33. Mentink-Vigier, F.; Marin-Montesinos, I.; Jagtap, A. P.; Halbritter, T.; van Tol, J.; Hediger, S.; Lee, D.; Sigurdsson, S. T.; De Paepe, G., Computationally Assisted Design of Polarizing Agents for Dynamic Nuclear Polarization Enhanced NMR: The AsymPol Family. J. Am. Chem. Soc. 2018, 140 (35), 11013-11019. 34. Atalla, R. H.; Vanderhart, D. L., Native cellulose: a composite of two distinct crystalline forms. Science 1984, 223 (4633), 283-285. 35. Ha, M. A.; Apperley, D. C.; Evans, B. W.; Huxham, M.; Jardine, W. G.; Vietor, R. J.; Reis, D.; Vian, B.; Jarvis, M. C., Fine structure in cellulose microfibrils: NMR evidence from onion and quince. The Plant journal : for cell and molecular biology 1998, 16 (2), 183-190. 298 36. Du, X.; Gellerstedt, G.; Li, J., Universal fractionation of lignin–carbohydrate complexes (LCCs) from lignocellulosic biomass: an example using spruce wood. The Plant journal : for cell and molecular biology 2013, 74 (2), 328-338. 37. Bardet, M.; Gerbaud, G.; Giffard, M.; Doan, C.; Sabine, H.; Le Page, L., 13C high-resolution solid-sate NMR for structural elucidation of archaeological woods. Prog. Nucl. Magn. Reson. Sp. 2009, 55, 199-224. 38. Hermans, C.; Porco, S.; Vandenbussche, F.; Gille, S.; De Pessemier, J.; Van Der Straeten, D.; Verbruggen, N.; Bush, D. R., Dissecting the role of CHITINASE-LIKE1 in nitrate-dependent changes in root architecture. Plant Physiol. 2011, 157 (3), 1313-26. 39. Jiao, S.; Hazebroek, J. P.; Chamberlin, M. A.; Perkins, M.; Sandhu, A. S.; Gupta, R.; Simcox, K. D.; Yinghong, L.; Prall, A.; Heetland, L.; Meeley, R. B.; Multani, D. S., Chitinase- like1 Plays a Role in Stalk Tensile Strength in Maize. Plant Physiol. 2019, 181 (3), 1127-1147. 40. Sanchez-Rodriguez, C.; Bauer, S.; Hematy, K.; Saxe, F.; Ibanez, A. B.; Vodermaier, V.; Konlechner, C.; Sampathkumar, A.; Ruggeberg, M.; Aichinger, E.; Neumetzler, L.; Burgert, I.; Somerville, C.; Hauser, M. T.; Persson, S., Chitinase-like1/pom-pom1 and its homolog CTL2 are glucan-interacting proteins important for cellulose biosynthesis in Arabidopsis. Plant Cell 2012, 24 (2), 589-607. 41. Wu, B.; Zhang, B.; Dai, Y.; Zhang, L.; Shang-Guan, K.; Peng, Y.; Zhou, Y.; Zhu, Z., Brittle culm15 encodes a membrane-associated chitinase-like protein required for cellulose biosynthesis in rice. Plant Physiol. 2012, 159 (4), 1440-52. 42. Duan, P.; Schmidt-Rohr, K., Composite-pulse and partially dipolar dephased multiCP for improved quantitative solid-state (13)C NMR. J Magn Reson 2017, 285, 68-78. 43. Johnson, R. L.; Schmidt-Rohr, K., Quantitative solid-state 13C NMR with signal enhancement by multiple cross polarization. J. Magn. Reson. 2014, 239, 44-49. 44. Bernardinelli, O. D.; Lima, M. A.; Rezende, C. A.; Polikarpov, I.; deAzevedo, E. R., Quantitative C-13 MultiCP solid-state NMR as a tool for evaluation of cellulose crystallinity index measured directly inside sugarcane biomass. Biotechnol. Biofuels 2015, 8. 45. Takahashi, H.; Hediger, S.; De Paepe, G., Matrix-free dynamic nuclear polarization enables solid-state NMR C-13-C-13 correlation spectroscopy of proteins at natural isotopic abundance. Chem. Commun. 2013, 49 (82), 9479-9481. 46. Fernandez-de-Alba, C.; Takahashi, H.; Richard, A.; Chenavier, Y.; Dubois, L.; Maurel, V.; Lee, D.; Hediger, S.; De Paepe, G., Matrix-Free DNP-Enhanced NMR Spectroscopy of Liposomes Using a Lipid-Anchored Biradical. Chem. Eur. J. 2015, 21 (12), 4512-4517. 47. Viger-Gravel, J.; Lan, W.; Pinon, A. C.; Berruyer, P.; Emsley, L.; Bardet, M.; Luterbacher, J., Topology of Pretreated Wood Fibers Using Dynamic Nuclear Polarization. J. Phys. Chem. C. 2019, 123 (50), 30407-30415. 299 48. Ling, Z.; Wang, T.; Makarem, M.; Cintron, M. S.; Cheng, H. N.; Kang, X.; Bacher, M.; Potthast, A.; Rosenau, T.; King, H.; Delhom, C. D.; Nam, S.; Edwards, J. V.; Kim, S. H.; Xu, F.; French, A. D., Effects of ball milling on the structure of cotton cellulose. Cellulose 2019, 26 (1), 305-328. 49. Lesage, A.; Auger, C.; Caldarelli, S.; Emsley, L., Determination of through-bond carbon- carbon connectivities in solid-state NMR using the INADEQUATE experiment. J. Am. Chem. Soc. 1997, 119 (33), 7867-7868. 50. Lesage, A.; Bardet, M.; Emsley, L., Through-Bond Carbon−Carbon Connectivities in Disordered Solids by NMR. J. Am. Chem. Soc. 1999, 121 (47), 10987-10993. 51. Wang, T.; Yang, H.; Kubicki, J. D.; Hong, M., Cellulose Structural Polymorphism in Plant Primary Cell Walls Investigated by High-Field 2D Solid-State NMR Spectroscopy and Density Functional Theory Calculations. Biomacromolecules 2016, 17 (6), 2210-2222. 52. Aluas, M.; Tripon, C.; Griffin, J. M.; Filip, X.; Ladizhansky, V.; Griffin, R. G.; Brown, S. P.; Filip, C., CHHC and 1H-1H magnetization exchange: Analysis by experimental solid-state NMR and 11-spin density-matrix simulations. J Magn Reson 2009, 199 (2), 173-187. 53. Kobayashi, T.; Slowing, I. I.; Pruski, M., Measuring Long-Range 13C–13C Correlations on a Surface under Natural Abundance Using Dynamic Nuclear Polarization-Enhanced Solid-State Nuclear Magnetic Resonance. J. Phys. Chem. C 2017, 121, 24687-24691. 54. Teymoori, G.; Pahari, B.; Stevensson, B.; Eden, M., Low-power broadband homonuclear dipolar recoupling without decoupling: Double-quantum C-13 NMR correlations at very fast magic-angle spinning. Chem Phys Lett 2012, 547, 103-109. 55. Wang, T.; Park, Y. B.; Cosgrove, D. J.; Hong, M., Cellulose-Pectin Spatial Contacts Are Inherent to Never-Dried Arabidopsis thaliana Primary Cell Walls: Evidence from Solid-State NMR. Plant Physiol. 2015, 168 (3), 871-884. 56. Hong, M.; Gross, J. D.; Rienstra, C. M.; Griffin, R. G.; Kumashiro, K. K.; Schmidt-Rohr, K., Coupling amplification in 2D MAS NMR and its application to torsion angle determination in peptides. J Magn Reson 1997, 129 (1), 85-92. 57. Mentink-Vigier, F.; Mathies, G.; Liu, Y.; Barra, A. L.; Caporini, M.; Lee, D.; Hediger, S.; Griffin, R. G.; De Paepe, G., Efficient cross-effect dynamic nuclear polarization without depolarization in high-resolution MAS NMR. Chem. Sci. 2017, 8, 8150. 58. Zhai, W.; Lucini Paioni, A.; Cai, X.; Narasimhan, S.; Medeiros-Silva, J.; Zhang, W.; Rockenbauer, A.; Weingarth, M.; Song, Y.; Baldus, M.; Liu, Y., Postmodification via Thiol- click Chemistry Yields Hydrophilic Trityl-nitroxide Biradicals for Biomolecular High-Field Dynamic Nuclear Polarization. The journal of physical chemistry. B 2020. 59. Jaudzems, K.; Bertarello, A.; Chaudhari, S. R.; Pica, A.; Cala-De Paepe, D.; Barbet-Massin, E.; Pell, A. J.; Akopjana, I.; Kotelovica, S.; Gajan, D.; Ouari, O.; Tars, K.; Pintacuda, G.; Lesage, A., Dynamic Nuclear Polarization-Enhanced Biomolecular NMR Spectroscopy at High 300 Magnetic Field with Fast Magic-Angle Spinning. Angew. Chem. Int. Edit. 2018, 57 (25), 7458- 7462. 60. Berruyer, P.; Bjorgvinsdottir, S.; Bertarello, A.; Stevanato, G.; Rao, Y.; Karthikeyan, G.; Casano, G.; Quari, O.; Lelli, M.; Reiter, C.; Engelke, F.; Emsley, L., Dynamic Nuclear Polarization Enhancement of 200 at 21.15 T Enabled by 65 kHz Magic Angle Spinning. The journal of physical chemistry letters 2020, 11, 8386-8391. 61. Kirui, A.; Dickwella Widanage, M. C.; Mentink-Vigier, F.; Wang, P.; Kang, X.; Wang, T., Preparation of Fungal and Plant Materials for Structural Elucidation Using Dynamic Nuclear Polarization Solid-State NMR. J. Vis. Exp. 2019, 144, e59152. 62. Massiot, D.; Fayon, F.; Capron, M.; King, I.; Le Calve, S.; Alonso, B.; Durand, J. O.; Bujoli, B.; Gan, Z. H.; Hoatson, G., Modelling one- and two-dimensional solid-state NMR spectra. Magn. Reson. Chem. 2002, 40 (1), 70-76. 63. Torchia, D. A., Measurement of Proton-Enhanced 13C-T1 Values by a Method Which Suppresses Artifacts. J Magn Reson 1978, 30 (3), 613-616. 301 APPENDIX Table S8.1 Fit parameters of 13C CP spectrum of wild-type (WT) sample. The peaks are classified in three groups according to their influence on the C4 region, as by integrating, interior and surface cellulose contribution is shown to respectively be 40.5% and 59.5% of the cellulose content. δ [ppm] Assignment Amplitude Width [ppm] Integral [%] 90.0 89.0 88.0 87.3 84.3 83.1 81.9 80.5 76.1 105.8 104.9 103.9 74.9 74.0 i4 s4 Cellulose C4 0.4 6.2 1.7 1.0 6.8 4.1 0.8 1.2 1.2 1.0 1.6 1.2 Other peaks close to cellulose C4 2.7 1.3 13.4 1.8 1.8 2.6 Other peaks far from cellulose C4 1.1 1.1 1.2 1.2 1.0 13.4 11.9 10.3 20.8 17.0 302 0.1 3.3 0.9 0.4 4.7 2.2 2.1 1.0 10.3 6.4 5.7 5.4 7.5 4.8 Table S8.1 (cont’d) 72.1 71.3 64.9 64.4 63.1 61.9 59.6 40.4 3.9 11.0 7.0 6.5 9.5 3.1 2.3 0.4 1.6 0.7 1.4 1.7 2.4 27.4 0.4 5.2 1.5 2.8 4.9 3.1 303 Table S8.2 Fit parameters of 13C CP spectrum of ctl1 ctl2 double mutant. The peaks are classified in three groups according to their influence on C4 region, as by integrating, interior and surface cellulose contribution is shown to respectively be 44.6% and 55.4% of the cellulose content. δ [ppm] Assignment Amplitude Width [ppm] Integral [%] 90.0 89.0 88.0 87.4 84.3 83.5 83.0 81.9 81.0 80.0 76.1 105.8 104.9 103.8 100.8 74.9 i4 s4 Cellulose C4 1.2 6.2 1.8 0.9 4.9 0.7 3.4 0.8 1.0 1.2 1.0 1.5 0.8 1.5 Other peaks close to cellulose C4 3.2 1.3 1.5 13.4 1.8 1.0 1.4 2.6 Other peaks far from cellulose C4 0.8 1.5 1.3 4.4 1.2 9.5 13.0 10.3 1.5 23.8 304 0.4 2.5 0.9 0.3 2.9 0.2 1.9 2.3 0.5 0.8 11.3 3.2 7.3 5.3 2.4 7.7 Table S8.2 (cont’d) 74.0 72.1 71.3 70.5 64.9 64.4 63.2 61.9 59.6 19.0 39.1 7.1 6.0 13.1 3.3 8.3 6.9 4.3 1.0 2.3 0.5 0.7 1.7 1.0 1.6 1.7 3.5 4.9 23.9 1.0 1.1 5.9 0.9 3.5 3.2 5.6 305 Table S8.3 Peak numbers of INADEQUATE spectra shown in Figure 8.3. interior surface xylose in xylose in arabinose cellulose (i) cellulose (s) xylan (Xn) xyloglucan (x) WT ctl1 ctl2 18 16 16 15 6 18 6 2 (A) 0 10 306 Table S8.4 13C-T1 and 1H-T1r relaxation times of cellulose and xylan in WT and ctl1 ctl2 samples. The data is fit using single exponential equation 𝐼 (𝑡) = 𝑒−𝑡/𝑇, where T could be T1 or T1r. Error bars are standard deviations of the fitting parameters. CS: 13C chemical shift. Unidentified (-). Rice WT Rice ctl1 ctl2 double mutant Assignments CS T1(CP) CS T1ρ CS T1(CP) CS T1ρ (ppm) (s) (ppm) (ms) (ppm) (s) (ppm) (ms) i/s/Xn2f1 105.6 25±1 105.5 43±2 105.5 11.3±0.8 105.3 31±2 i4 s4 i6 s6 89.3 35±5 89.3 53±2 89.3 21±2 89.1 42±2 84.2 20±1 84.1 33±2 84.1 9.3±0.6 84.0 29±3 65 13±1 64.7 28±3 65.2 7.4±0.8 65 24±2 62.9 4.2±0.5 62.8 24±2 63.1 2.3±0.4 62.8 18±2 Xn-AcCO 173.7 - 173.7 11±1 174.3 8.3±0.7 174.3 - Xn13f Xn42f Xn43f 102.0 7.7±0.4 102.5 7.9±0.8 102.7 4.8±0.4 101.8 8±1 82.2 10.5±0.9 82.2 18±2 82.2 7.5±0.6 82.2 18±2 77.7 10±2 78.0 16±2 77.5 6.9±0.5 78.0 10±1 Xn-AcMe 21.6 7.8±0.7 21.7 10±2 21.7 4.6±0.3 21.7 9±1 307 Figure S8.1 Lignin has increased methyl ether substitution in the double mutant. The spectra of wild-type (black) and ctl1 ctl2 double mutant (yellow) are normalized by the interior cellulose carbon 4 (i4) peak. The lignin methyl ethers (lignin -OMe) has a doubled intensity in ctl1 ctl2 but the lignin aromatics have a comparable intensity in both samples. 308 Figure S8.2 Effect of solvents on DNP enhancement. a, ctl1 ctl2 sample in the solvent of 13C- depleted, d8-glycerol/D2O/H2O (60/30/10 vol%) has shown a 22-fold enhancement of sensitivity. b, ctl1 mutant in d6-DMSO/D2O/H2O (60/30/10 vol%) matrix only exhibits a 5-fold of sensitivity gain. 309 Figure S8.3 Timesaving by DNP on carbohydrate signals in unlabeled rice stems. Top row: The room-temperature spectrum collected on a 400 MHz NMR gives a signal-to-noise (S/N) ratio of 37 for the highest peak after 43 h of measurement. Bottom row: 600 MHz/395 GHz DNP provides a S/N ratio of 618 after only 0.5 h of measurement. Cellulose peaks are well reserved, but intensity suppression has been observed for xylan signals, lignin methyl ether (-OMe) and small molecules (Glc; glucose). 310 Figure S8.4 DNP polarization is uniform across the cell wall. The microwave-on (MW on) and microwave-off (MW off) spectra are normalized by the interior cellulose carbon 4 peaks (i4) to compare the spectral pattern. The consistent spectral envelope clearly demonstrate that the polarization is uniform across the whole cell wall. 311 Figure S8.5 The experimental and simulated spectra have a good match. The 120 to 50 ppm regions of a, wild-type sample (left) and b, ctl1 ctl2 double mutant (right) are shown. All numerical parameters used to obtain the fits are summarized in Tables S8.1 and S8.2. Color code follows peak classification in these tables: i4 cellulose in red, s4 in magenta, close peaks in dark yellow, and others in grey. 312 Figure S8.6 1D cross sections of DNP-enabled 2D CHHC spectrum. Representative slices were extracted from the 2 ms CHHC spectrum of unlabeled wild-type rice stem. The 13C FWHM linewidths and signal-to-noise (S/N) ratios are shown for the major peaks. 313 Figure S8.7 2D PDSD spectra of 13C-labeled rice stems. The spectra are collected using a, 3 ms and b, 5 ms mixing times. Assignments of cellulose and xylan peaks are annotated on the spectra. 314 Figure S8.8 NMR relaxation curves of polysaccharides in unlabeled rice stems. The a, 13C-T1 and b, 1H-T1ρ data are plotted separately for wild-type and ctl1 ctl2 samples. Cellulose signals (red) generally exhibit faster relaxation than xylan peaks (blue). The exceptions in panel a only occur to the 62 ppm s6/x5 peak, which has mixed contribution from both cellulose and matrix polymers. 315 CHAPTER 9: CHRONOLOGICAL MOLECULAR FINGERPRINT OF WETLAND SOIL BY SENSITIVITY-ENHANCED SOLID-STATE NMR Research paper submitted to Nature Geoscience. Preprinted in ChemRxiv (https://doi.org/10.26434/chemrxiv-2023-p4rvg). Authors: Wancheng Zhao, Elizabeth C Thomas, Debkumar Debnath, Faith Scott, Frederic Mentink-Vigier, John R. White, Robert L. Cook, and Tuo Wang 9.1 Abstract Soil organic matter (SOM) plays a major role in mitigating greenhouse gas emission and thereby regulating Earth’s climate, carbon and water cycles, and biodiversity. Wetland soils contain the highest stores of soil carbon on the planet on an areal basis, accounting for one-third of all the SOM, yet our understanding of carbon sequestration within wetlands lags behind that of upland soils. Here we show the molecular-level fingerprints of wetland soils spanning eleven centuries using advanced solid-state nuclear magnetic resonance (ssNMR) spectroscopy. Remarkably, combining dynamic nuclear polarization (DNP) with SOM enrichment allowed up to an 8,000-fold time-saving over conventional NMR approaches. This innovative approach for SOM characterization revealed that the parent herbaceous plant core molecular structures are preserved, with the aromatic and carbohydrate motifs becoming tightly packed, even after a millennium. Such preserved cores occur alongside molecules from the decomposition of loosely packed parent biopolymers and biogeochemical processing driven by geological global and anthropogenic changes, adding to the chemical diversity of SOM. These findings reveal that particulate organic matter (POM) should be a major focus for wetlands, and other soils with high organic matter content, especially when considering the fate of coastal wetland SOM when exposed to oxygenated water due to erosion. 316 9.2 Introduction Soil organic matter (SOM) and associated soil organic carbon (SOC), critically represents one of the major reservoirs of carbon on the planet and is critical to ecosystem services, from the molecular to the global scale1, 2. Just the top 1 m of the world’s soil contains 1500 petagrams (Pg, billion tons) of carbon, approximately twice the carbon pool contained in the entire atmosphere3 and more than the atmosphere and vegetation combined4. Soil has been a sink to about 210 ± 45 Pg of carbon between the years 1850 and 2021, mitigating around 100 ppm of atmospheric CO2 levels5. Wetland soils contain approximately one-third of this SOC despite occupying only 5-6% of the Earth’s land surface area6. Coastal wetlands, known as the blue carbon ecosystems, occupy only 0.07-0.22% of the Earth’s surface but sequester 0.08-0.22 Pg carbon each year, accounting for more than half of the carbon buried in the oceans annually7-9, and store the vast majority of this carbon for hundreds to thousands of years, with minimal methane emissions due to a poised redox potential10. The anaerobic conditions of wetland soil cause a reduced metabolic efficiency of microbes in acquiring energy during decomposition of complex organic carbon compounds, which leads to increased preservation of carbon in soil profile unless hydrologic conditions change11. Subsidence, sea level rise and wave energy promote wetland soil collapse into the surrounding shallow, estuarine aerobic water column, resulting in the oxidation of SOM and releasing the stored carbon back to the atmosphere as primarily carbon dioxide12-14. To preserve and manage wetlands, minimize greenhouse gas emissions, and maximize carbon sequestration potential, it is crucial to understand, model, and accurately predict the dynamics of SOM, and hence SOC, in a scalable and applicable manner15. Thus, detailed molecular information, including functionality, isomerism, and conformations of the sequestered carbon pool is a fundamental requirement. Currently, there 317 is limited work on this advanced level of carbon sequestration within coastal wetlands, as most work has been focused on upland soils. The persistence of SOC was considered to occur through the formation of recalcitrant macromolecules by polymerization during a process called humification16. Presently, the dynamic stability concept prevails, whereby organic carbon is preserved as an organic entity via molecular preservation for periods of time, in concert with dynamically changing organic carbon speciation17. This new view also breaks SOM down into two major pools: particulate organic matter (POM), which includes occluded SOM and plant materials, and mineral associated organic matter (MAOM), which contains mostly small organic molecules and biological metabolites18, 19. Over time, these two more recalcitrant pools may become accessible, and therefore can be viewed as shorter-term stable carbon deposits in the context of geologic time20 (Swift and Hayes, 2020). Most research on this new paradigm was focused on upland or mineral soils, with the MAOM fraction receiving the bulk of the attention, although the focus is beginning to shift19, 21, 22. Most coastal wetlands are POM-dominated systems with a high organic matter to mineral matter ratio. Globally, per unit area, coastal wetlands are considered the most carbon-rich dynamic reservoirs of organic carbon, warranting a closer examination of this critical and most vulnerable to loss carbon pool, as sea level continues to rise. Here, we examine coastal wetland soil samples from the Mississippi River Delta, USA that have been deposited and preserved over the past 11 centuries, but whose soil C stores are being lost at an accelerated rate due to the high relative sea level rise rate. Sensitivity-enhanced ssNMR is applied, for the first time, to whole soil samples to enable rapid acquisition of 2D 13C/1H-13C correlation spectra, providing atomic-level structural information of the carbon fraction of SOC. This method allows for a detailed characterization of soil on the fine granular and molecular level, 318 beyond the moiety characterization level (Supplementary Discussion), which reveals that: 1) in the initial stages of SOC sequestration, core structures of the parent herbaceous plant materials are preserved but the aromatic and carbohydrates motifs become tightly packed, with non- carbohydrate components being concentrated in the soil, 2) some structural cores of herbaceous plant biopolymers survive anaerobic microbial degradation, with their original structure and physical packing preserved up to 1000 years since deposition, 3) molecules from the decomposition of loosely packed parent biopolymers and biogeochemical processing are present and add to the diversity of SOC chemical nature, and 4) changes in the carbon speciation during the sequestration processes are driven by both natural geological changes (e.g., delta lobe switching) as well as anthropogenic changes (e.g., river levees). Besides these multifaceted conceptual advances, this high-resolution and rapid technological platform also opens a new research avenue for SOC analysis in undisturbed soil materials. 9.3 Results 9.3.1 Molecular Fingerprinting of Carbohydrate and Aromatics in Wetland Soil We examined seven soil samples (down to almost 2 m) collected from a salt marsh Spartina alterniflora-vegetated island in the Barataria Bay, Louisiana, located along the Gulf of Mexico coastline (Figure S9.1a-d). Each year, Louisiana experiences a loss of over 65 km2 of coastal wetlands, leading to an annual release of over 1 million metric tons of stored carbon from Barataria Basin alone23, 24. This island underwent shoreline erosion rates of over 1.5 m y-1 and became entirely eroded in 2021 (Figure S9.1e, f). A two-step protocol was developed to improve the NMR sensitivity by a factor of 90, reducing experimental duration 8,100 times, essentially reducing a 22-year-long experiment down to one day. This was achieved by employing hydrofluoric acid (HF) treatment, which depleted the mineral component, concentrated SOM, and enhanced their 319 signals by a factor of 5 (Figure 9.1a) without chemically perturbing its structure as described in Supplementary Methods25-27, and DNP, which relies on microwave irradiation to transfer electron polarization to NMR-active nuclei in the soil, resulting in an additional 18 enhancement (Figure 9.1b). Together, this allowed for carbon connectivity to be directly mapped by a 2D 13C- 13C correlation spectrum on unlabeled soil (Figure 9.1c), a task previously impossible by conventional techniques, but now achievable within 16 hours of analytical time. The carbohydrate signals of the top 10-cm layer of soil (sample 1) were predominantly from cellulose (Figure 9.1c), including the glucan chains residing on both the surface and internal domains of the microfibrils (Figure 9.1d). Multiple forms of glucan chains are identifiable within cellulose. First, we observed two distinct sets of signals from surface chains (sf and sg in Figure 9.1c), which have been proposed to contribute to distinctly hydrated exteriors of the microfibrils, namely the concept of hydrophobic and hydrophilic surfaces28. Second, in addition to the dominant interior conformers (ia and ib), a third form of ic was also identified in a 2D 1H-13C correlation spectrum, collected within an hour (Figure 9.1e), with a unique C4 chemical shift of 87.5 ppm. Type-c glucan has been found in the native cellulose across many different grass and woody plant species and have been attributed to the deeply embedded and inaccessible core of the larger bundle formed by multiple microfibrils (Figure 9.1d)29-31. The identification of these five glucan types demonstrates that the native cellulosic material has been preserved in surface soil. Further structural preservation is indicated by the unexpected identification of 2-fold and 3-fold xylan (Xn2f and Xn3f) in the soil (inset of Figure 9.1c). The 2-fold and 3-fold refers to the helical screw conformation and indicate the number of sugar residues needed for finishing a 360° helical rotation along the chain (Figure 9.1f)32. Recent studies of the lignocellulosic plant biomass have revealed the distinct functions of these two xylan conformers, with the flat-ribbon 2-fold 320 xylan coating the smooth surface of cellulose microfibrils33, 34 and the zigzag 3-fold xylan preferentially packing with disordered aromatics, namely the lignin domains in plants30. Evidently, at least a portion of the structural core of plant lignocellulosic biomass is preserved, intact, in the wetland soil. Figure 9.1 Molecular fingerprint and spatial organization of biopolymers in wetland soil. a, Comparison of quantitative 13C spectra of the HF and non-HF treated soil sample 1 (surface layer) at room temperature. HF treatment concentrates the organic phase and gives a 5-fold sensitivity boost. b, Comparison of 1D 13C spectra with and without microwave (MW) irradiation. The sensitivity enhancement (εon/off) provided by DNP is 18-fold. Inset shows pictures of the soil sample and a sapphire rotor containing the soil sample. c, DNP-enabled 2D 13C-13C correlation spectrum (refocused J-INADEQUATE) of soil sample 1 measured in 16 h, resolving the carbon connectivity of carbohydrate components. Overlay of the measured spectrum (black) with the 321 Figure 9.1 (cont’d) probability map constructed using 412 carbohydrate units from Complex Carbohydrate Magnetic Resonance Database36 indicates the best match with cellulose (magenta dots). Assignments of the interior (i) and surface (s) glucan chains of cellulose microfibrils and the xylose units of 2- and 3- fold xylans (Xn2f and Xn3f) are labeled. Inset shows the xylose carbon 4 region processed with large line-broadening to show xylan signals. d, Representative cellulose structure with each microfibril containing eighteen β-1,4-glucan chains on the surface and interior domains. Multiple microfibrils aggregate to form larger bundles that accommodate different forms of glucan chains (ia,b, ic, sf, and sg), which are further wrapped by matrix non-cellulosic polymers. e, Single-hour DNP 2D 1H-13C correlation spectrum resolves the signals from aliphatic carbons, carbohydrates, and aromatics. Key signals of polymethylene (-CH2-)n, methoxyl (-OCH3), cellulose, and xylan are labeled. f, Structure of 2- and 3-fold xylan conformers. g, Overlay of 2D 1H-13C correlation spectrum measured with short (0.1 ms; blue) and long (1.0 ms; yellow) CP contact times. The symbolic representations correspond to the carbons in lignin monolignol units. Dashed lines show the key 1H positions. Many non-carbohydrate molecules were also identified, including the aromatics and methoxy substitutions in lignin, the acyl chains (or polymethylene35) in lipid polymers, and other aliphatic motifs (Figure 9.1e). This characterization was achieved using a single-hour 2D experiment that relies on a short CP contact time (0.1 ms) to emphasize one-bond 1H-13C correlations. Protonated carbons in lignin exhibited signals in the 13C chemical shift range of 110- 125 ppm (blue spectrum in Figure 9.1g). The 105 ppm 13C signals have dual contributions from both carbohydrates and aromatics, including the carbon 1 of cellulose and 2-fold xylan, as well as the carbons 2 and 6 of the S-unit, thus showing one-bond correlations with both carbohydrate and 322 aromatic protons. A range of nonprotonated carbons were also observed with a longer CP contact (1.0 ms) that extended the reach of 1H-13C correlation (yellow spectrum of Figure 9.1g; Figure S9.2). The spotted signals included carbonyl groups (CO) and monolignols, such as the carbon 3 and 5 of the syringyl (S) unit at 154 ppm and the carbon 3 and 4 of the guaiacyl (G) unit at 145-149 ppm. Their chemical nature was confirmed by their strong peaks in dipolar-dephasing spectra that removed all protonated carbon signals (Figure S9.3). The NMR observations of SOC unveiled a complex composition in terms of plant polysaccharides, lignin, and lipid polymers. 9.3.2 Domain Distribution of Polymers in Soil It is noticeable that the aromatic carbons (13C chemical shifts of 100-140 ppm) not only show cross peaks with the aromatic protons at 6-7 ppm, but also cross-talked with carbohydrate and aliphatic protons that resonate at 3-5 ppm, revealing the co-localization of aromatics, aliphatics, and carbohydrates on the nanoscale, consistent with previous work37. This concept of molecular mixing is also supported by the cross peaks between carbohydrate carbon sites (13C chemical shifts of 70-110 ppm) with aromatic protons (1H chemical shifts of 6-7 ppm). The only exception was observed in polymethylene (-CH2-), which failed to show correlations with other carbons or protons, providing a clear indication of domain separation for lipid polymers. This finding corroborates earlier ssNMR results35, 37 where poly-methylene were found to form large aggregates to resist further degradation, which is a characteristic commonly shared by diverse soil materials in nature. 9.3.3 Preserved Structural Core in Plant Material and Surface Soil The composition of plant detritus inputs to the soil as well as the redox status of the soil are among the key external factors that affect the rate of carbon sequestration38. The plant tissues 323 gathered at the soil collection site retain highly similar carbohydrate signals when compared to the surface soil (Figure 9.2a). Cellulose crystallinity is unchanged, as evidenced by both soil and plant samples showing comparable intensity ratios between the interior cellulose C4 at 89 ppm and the surface cellulose C4 at 84 ppm. With a 24-fold DNP enhancement (Figure S9.4), we unambiguously detected the varied signals from multiple cellulose forms and xylan conformers in 2D 13C-13C correlation spectrum of these plants (Figure 9.2b), which demonstrated a similar pattern to the soil spectrum. The soil exhibits elevated levels of carbonyls, methoxyls, aromatics, and aliphatics, as revealed by the difference of two parental spectra (Figure 9.2a). These components might have accumulated due to their slower decomposition rate when compared to carbohydrates. Figure 9.2 Conserved structural core of plant material and wetland soil. a, DNP-enhanced 13C spectra of surface soil (yellow; sample 1) and the plant on top of the soil (blue). The bottom panel shows the difference of the two spectra, revealing a signature pattern of lignin, aliphatic and polymethylene, as well as matrix polysaccharides. b, Carbohydrate region of DNP enhanced 324 Figure 9.2 (cont’d) 13C-13C refocused J-INADEQUATE spectrum showing signals from cellulose and xylan. Inset picture shows the original plant material characterized here. c, Aromatic amino acids resolved from the plant sample, with chemical shifts labeled on the structure. d, 2D 1H-13C correlation spectrum measured using 0.1 ms CP resolving key signals of cellulose, xylan, lipid polymers. Lignin methoxyls (-OCH3 or -OMe) and xylan acetyls (Ac) are also observed, with structures presented. e, 2D 1H-13C correlation spectrum measured using 1 ms CP resolving signals of aromatic and carbonyl carbons from proteins and lignin. Dash lines in blue and red annotate the key positions of methoxyl and aromatic protons, respectively. Black dashed line represents the anticipated correlations between lignin aromatic carbons and carbohydrate anomeric protons, which are notably less numerous than the signals observed in the soil. Strong signals of aromatic amino acid residues have been spotted, which align with the chemical shifts of histidine or tryptophan (Figure 9.2c). These molecules are uniquely abundant in the plant material and are not present in the soil underneath it, likely due to being vulnerable to rapid microbial degradation. Though the carbohydrates signals are highly consistent with the soil, the aliphatic region shows a dramatically simplified pattern (Figure 9.2d), with only 3 peaks from the methoxyl group of lignin, the CH2 groups likely from the acyl chain of lipids in the membrane or from the cutan or suberin, and the acetyl group that serves as an important modification of matrix polysaccharides, such as xylan. The aromatic region is simpler (Figure 9.2e). The key carbon sites of both S and G units are still detectable, with these carbons mainly correlating with the methoxyl and aromatic protons and lacking the cross peak with the anomeric protons of carbohydrates at 4.3-4.5 ppm, suggesting substantially reduced interactions between carbohydrates and aromatic polymers. 325 Despite the conserved structures of individual carbohydrate and lignin components, soil has two unique structural features that are absent in plant materials. First, non-carbohydrate components are highly concentrated, which is likely caused by the faster degradation rate of polysaccharides compared with lignin and polymethylene polymers. Phenols serve as antioxidant during degradation reactions, while polymethylene polymers contain crystalline domains of aliphatic chains, conferring these polymers with high stability39, 40. This trend was also confirmed by the analysis of another plant sample collected 30 m inland on the same island (Figure S9.5). Secondly, the aromatics and carbohydrates are more tightly packed in the surface soil than in plants. This finding might originate from the faster decay of primary cell walls that only contain cellulose and soft matrix polysaccharides but does not contain lignin, and/or the removal of intra- and extracellular components leading to tighter packing of residual lignocellulosic components. 9.3.4 Mapping Carbon Composition and Packing Along Depth The molecular composition of the soil matter changes with the depth. In the quantitative 1D 13C spectra41, the content of polymethylene carbons, marked by the intensities of two adjacent peaks at 33 and 31 ppm, increased substantially from sample 1 (0-10 cm interval) to sample 2 (40- 50 cm interval), as shown in Figure 9.3a. These two peaks in polymethylene, also observed in many other soil organic matter samples, correspond to crystalline (CH2)n chains in all-trans conformation (type-a; 33 ppm) and amorphous regions accommodating both trans and gauche conformations without long-range order (type-b, 31 ppm)35. The self-aggregated nature and limited accessibility of polymethylene might have prevented it from being degraded after deposition in the soil. However, when normalized to the aromatic signal, the polymethylene, together with carbohydrates and carbonyls decrease sequentially as one moves deeper in the soil profile (samples 2-7; age range from 1963AD to 945AD). These spectral observations were quantified through a 326 deconvolution protocol applied to 1D quantitative spectra (Figure S9.6), using the carbon sites resolved from high-resolution 2D dataset. Carbohydrates exhibit a substantial reduction in their proportion, decreasing from 29% in sample 1 to 20% in sample 7, while the content of aromatics are being enriched from 29% to 36% when moving deeper from the surface (Figure 9.3b). Figure 9.3 Structural changes of SOM in relation to depth. a, Overlay of 1D quantitative 13C spectra of seven soil samples with normalization by the major aromatic peak (asterisk). The top panel compares samples 1 and 2, and the bottom panel includes samples 2-7. The two key peaks (types a and b) of polymethylenes are marked, with an illustration of the crystalline all-trans domain and the amorphous domain that include both trans and gauche conformations. Three conformers, gauche(+), gauche(-), and trans, are illustrated, with methylene groups represented as yellow circles and hydrogen atoms as open circles. The wetland condition of each soil sample 327 Figure 9.3 (cont’d) (saltwater, brackish, or freshwater) are also labeled. b, Quantification of the four major carbon types. The details of spectral deconvolution are documented in Figure S9.6, 1H-13C correlation spectra of two untreated soil samples measured with 0.1 ms (yellow), 0.5 ms (light blue), and 1 ms (grey) CP contact times. d, molecular and physical evolution of wetland soil over time. The figure depicts a geological timeline through soil depth based on 14C dating results and historic events in the area. Molecular profiling of 7 soil samples reveals the content of major carbon types. Detailed information differentiates different types of polymethylene carbons, cellulose chains, and monolignol units. Bulk density and loss-on-ignition measurements were also taken on 18 soil samples immediately following collection. A higher bulk density indicates a higher mineral content, while a lower loss-on-ignition % suggests a lower organic matter content. Regarding the ratio between carbohydrate and aromatic moieties, there is a vertical distribution based on salinity as the wetland transitioned from freshwater, to brackish, to a current day salt marsh environment (Figure 9.3a)42. A similar trend can also be seen regarding the polymethylene-to-aromatics ratio moving from the deeper soil profile to the present-day surface. This finding suggests that, as sea level rises, salinity and resulting shifts in plant species is likely affecting how the sequestered carbon is stored and ultimately degraded in the soil profile. Native, untreated samples were also investigated using DNP-enhanced 2D 1H-13C correlation experiments (Figure 9.3c). It is intriguing that the spectral pattern was consistently maintained in the untreated samples 4 and 6 (Figure 9.3c), and in both HF-treated and untreated materials of sample 1 (Figure S9.7 and Figure 9.1g); therefore, HF did not perturb the native structure of the SOC core. This similarity is also an indication that at least a significant portion of the lignocellulosic cores were preserved, supported by the observation of internal cellulose, whose 328 carbon 4 (i4) shows a major decline starting from sample 2 (Figure 9.3a) but still exhibits some weak signals in samples 4 and 6 (Figure 9.3c). While most cellulose was decomposed rapidly in the surface layers, a fraction of these crystalline cores were preserved for centuries in this soil. Moreover, the close spatial proximities between aromatics and carbohydrates observed in samples 4 and 6 (Figure 9.3c) resemble those identified in the surface soil (Figure 9.1g). The polymethylene also shows the same self-aggregation features in these deeper samples. Therefore, the decomposition of SOM did not happen homogeneously. Some biopolymeric structural cores, like self-aggregated polymethylene and densely packed lignin-polysaccharide domains, have efficiently withstood microbial degradation and maintained their original structure and physical packing after approximately 500 years (at 100 cm depth) and even up to 1000 years (at 180 cm depth). This preservation has been maintained despite the presence of microbial extracellular enzymes present throughout the profile43. 9.3.5 Natural and Anthropogenic Influences on Carbon Speciation Over a Millennium The observed nondirectional changes of molecular composition and structure (Figure 9.3d) are not expected based on conventional soil aging and humification or the activity of microbial communities at depth. As tracked by the 14C dating, the soil material collected across the ~2-m depth covers a geological timeline of 11 centuries43, 44; therefore, the interplays of the delta lobes and water salinity in the Barataria Bay should also play a key role. The Mississippi River watershed is the dominant surface hydrologic feature in North America, which collects runoff from 40% of the continental US between the Rocky Mountain region to the Appalachian Mountains42. River deltas are dynamic systems with continually shifting lobe formation and abandonment over time42. Therefore, these environmental shifts, over time, can influence the physical and chemical characteristics of the accreted carbon pool based on hydrodynamics and 329 salinity. One anthropogenic driver has been the construction of river levees over the past century, essentially separating the coastal basins from the river, preventing the historical freshwater and sediment-subsidies from occurring45. The continuing decrease of loss-on-ignition (LOI; see Online Methods) and total carbon (TC) and the gradual increase of bulk density (BD) from 1 m depth to the surface of the soil is due to marsh fragmentation (Figure 9.3d; Supplementary Methods). As the continuous marsh platform begins to erode from all edges, the interior of the marsh becomes closer to the shoreline. Consequently, fine-grained sediments present in the bay are transported into the marsh during storm events.46, 47. The soil at 40 cm depth formed under hypersaline condition also shows the most unique chemical characteristics that violate the trends on the molecular level. It shows the highest content of aliphatics and polymethylene, and the polymethylene has a unique structure that is rich in the carbon site resonating at 34 ppm (named type-a polymethylene; the crystalline domain in all-trans conformation). The lignin amount is low but contains a high level of S-monolignols with a high degree of methoxy substitutions. Cellulose crystallinity also becomes low: only 10% of glucan chains are now in a crystalline interior environment, while the remaining majority are disordered. 9.3.6 Paradigm of Wetland POM Preservation and Carbon Sequestration It is imperative to understand the connection between the chemical stability of wetland SOC and its carbon structure, given wetlands' crucial role in global carbon stocks and their ability to sequester more carbon per unit area compared to other soil types1, 6. It is also crucial to differentiate between molecular and carbon speciation, microbial transformation, and preservation17, 18 when considering SOC persistence as sea level continues to rise globally8, 12. In this pursuit, 13C ssNMR spectroscopy48-50, along with the sensitivity enhancement yielded by 330 DNP51, 52, has been introduced to minimize the biases introduced by the extraction, solubilization, and relaxation present in liquid state NMR (Supplementary Discussion), while leveraging the detailed molecular view NMR spectroscopy allows, especially multidimensional techniques. This technique has allowed new insights into SOC sequestration and the importance of preservation for organic rich and POM-dominated blue carbon systems. This study reveals robust preservation of the polymeric assembly in the top 10 cm of soil echoing the core structure of plant parent materials, evident in the comparable interior and surface cellulose signals, the preservation of multiple forms of cellulose, xylan conformers, and both the S and G monolignols in lignin (Figure 9.1c, e, f). This result can be attributed to the tight packing of some lignin and carbohydrate components, which could be induced by the decay of bulky cell wall cellulose and soft matrix polysaccharides, as evidenced by lower content of these moieties within the whole soil (Figure 9.2a). Further evidence for decay taking place in parallel with preservation is the absence of aromatic amino acids within the SOM (Figure 9.2c). This new insight allows for a refocusing on the concept of molecular preservation in the form of the conservation of the structural core of parent biopolymers, i.e., recalcitrance, as an important component of carbon storage, especially for high SOM systems, such as we find in wetlands, and counters the concept that free POM, as a whole, is a less than stable form of carbon53, 54. The preservation of recalcitrant lignocellulosic domains in soil POM involves maintaining both the molecular structure and supramolecular assembly of participating biopolymers (Figure 9.3c). This concurs with the biomolecular transformation of more-accessible molecules on the millennium timescale. The rapid decay of carbohydrates can be explained by their natural preferential utilization by microbes over other molecules, such as aromatic compounds, as both an energy and a nutrient source under the anaerobic soil conditions53. The better preservation of 331 aromatics over polymethylene is likely related to the reduced soil conditions. While soil microbes can produce extracellular compounds such as phenol oxidase, these metalloenzyme require oxygen to oxidize phenol compounds. Hence, in the anaerobic wetland soil profile, these compounds are stable11. Additionally, it has been found that high phenolic compounds strongly inhibit hydrolases further muting microbial decomposition of SOM55, thus we posit that to a limited extent additional aromatic moieties are synthesized by biotic/abiotic processing of the loosely associated lignocellulose56-58. The dynamics of deltaic systems on C sequestration were revealed through quantitative 13C NMR data, with major transformations found to occur within the SOC pool over a millennium (Figure 9.3a, d). The first important takeaway is that the conditions of the wetland under which the SOC is initially preserved play a major role in the decomposition of the SOC pool. This is evidenced by an increased preservation of the carbohydrate when the wetland transitioned from freshwater, to brackish, to a saltwater-dominated wetland (Figure 9.3a); the same general trend can also be seen regarding polymethylene. The relative proportion of preservation only changed with time with the change in the depositional environment, highlighting the environmental controls on plant species as the dominant factor. The initial transition from a freshwater to brackish water wetland was due to geological influences in the form of lobe transition (Lafourche to Plaquemine lobe and Plaquemine to Balize lobe, respectively), while the transition from a brackish water to a salt marsh was influenced by both a geological lobe relocation and levee construction, starving the wetlands of freshwater and sediment inputs42. 9.4 Methods and Materials 9.4.1 Collection of Soil Material Soil cores (2 m in length) were extracted with a polycarbonate core tube from a brackish 332 Spartina alterniflora-dominated island in Barataria Bay, Louisiana, USA (GPS coordinates: 29.44358, -89.899722). Two cores were extracted at different distances (1 m and 2 m, respectively) from the shoreline of the island (Figure S9.1d). The extracted materials were divided into 10 cm sections based on depth. The samples were stored on ice during transportation and then kept at 4 °C for storage until analyzed. 9.4.2 Hydrofluoric Acid Treatment Visible plant matter was removed from the dried soil samples. Each sample was ground with a mortar and pestle set until the material can pass through a 125 μm sieve. Around 600 mg of ground material was transferred into a 15 mL centrifuge tube, and 10 mL of 2% HF solution was added. The tube was capped and turned end-over-end in a rotary mixer throughout for 9 different time intervals in the following sequence: five 1 h intervals then a 16 h interval followed by two 24 intervals and a finally 72 h interval. In between these intervals, the tubes were placed into a benchtop centrifuge and spun at 2000 rpm for 20 min at room temperature. After centrifugation, the 2% HF solution was decanted and replaced with freshly prepared 2% HF solution. The soil samples were then vacuum filtered with 18 M water for three times to remove the excess HF and freeze-dried for 24 h. This protocol was modified from a previously reported method26. Comparison of soil materials before and after HF treatment is shown in Figures 9.1a and S9.8a. 9.4.3 Solid-State NMR Spectroscopy For each soil sample, 95-105 mg of HF-treated material were packed into a 4-mm zirconium rotor and measured on a Bruker Avance 400 MHz (9.4 Tesla) NMR spectrometer. All experiments were conducted using a 4 mm probe under 14 kHz MAS at 298 K. 1D quantitative spectra were measured using the MultiCP pulse sequence41, with 11 CP blocks applied. Each CP block used 1.1 ms contact time, with a delay of 0.6 s between blocks. The acquisition time was set 333 to 25 ms, and the recycle delay was 1 s. For each sample, 16,384 scans were recorded within 35 h. The field strengths of the radiofrequency pulses were 71.4 kHz for both 13C and 1H hard pulses, and 62.5 kHz for 1H decoupling. The 13C chemical shifts were externally referenced to the tetramethylsilane (TMS) scale by calibrating the adamantane CH2 peak to 38.48 ppm. In this work, all ssNMR and DNP spectra were collected using the software Topspin 4.0 and analyzed in Topspin 4.1 version. Graphs were plotted using Origin Pro 2019b software and Adobe Illustrator CC Cs6 V16.0.0. To analyze the content of different carbon pools, deconvolution was performed on the 1D quantitative MultiCP 13C spectra using DMfit59 (20200306 version) following the positions of the peaks resolved from 2D DNP spectra, as detailed in Figure S9.6 and Table S9.1. This allowed us to convert peak intensities into carbon percentages for different structural motifs including carbohydrates, aromatic, carbonyl, and aliphatic components, as well as the ratios of different carbon sites within each category (Table S9.2). 1D rotor-synchronized non-quaternary suppression (NQS) spectra were collected under 14 kHz to identify quaternary carbons60. Signals from the protonated carbons were dephased using two delays (30 µs × 2) without heteronuclear decoupling. The CP contact time was 2 ms. The acquisition time and the recycle delay was set to 41 ms and 2 s, respectively. In addition, 1D conventional 13C CP spectra were collected to compare with NQS spectra, with identical experimental parameters. The NQS and CP spectra are shown in Figure S9.3. 9.4.4 Preparation of Soil and Plant Samples for MAS-DNP A stock solution, which is often referred as the DNP juice was prepared using a mixture of D2O and H2O (90:10 Vol%) and 10 mM AsymPolPOK biradical (Catalogue# C015P01, CortecNet)61. Another two stock solutions were also prepared with the same radical concentration 334 but using different solvents of d6-DMSO/D2O/H2O (10/80/10 Vol%) and d6-DMSO/H2O (90/10 Vol%). The D2O (Catalogue# DLM-4DR-PK) and d6-DMSO (Catalogue# DLM-10TC-PK) were from Cambridge Isotope Laboratories. The details parameters of DNP juice composition used for each sample and the setup parameters of all experiments were listed in Table S9.3. The stock solutions were mixed with three types of materials, including HF-treated and non-treated soil as well as plant materials. Around 50 mg of HF-treated soil material was impregnated in 150 μL of the stock solution and vortexed briefly. The mixture was ground mildly for 20 min using a mortar and pestle to allow the radicals to penetrate the porous components of the soil. 30 mg of the final material were then packed into a 3.2 mm sapphire rotor for measurement. For comparison, the two plant samples (Spartina alterniflora) collected from the edge of the island (on top of the soil extraction site) and 30 m inland were also processed for MAS- DNP measurement. Around 30 mg of each plant sample was subjected to the same protocol described above to mix with 10 mM AsymPolPOK. For nontreated soil samples, the protocol was modified regarding the concentration of the bi-radical, which has increased to 30 mM to gain more enhancement. The DNP enhancements and Electron paramagnetic resonance (EPR) spectra (EMX Nano benchtop EPR) measured on the plant and soil samples were shown in Figures S9.4 and S9.8. The EPR spectra were plotted by MATLAB R2020a with a toolbox Easyspin (6.0.0). The evaluation of the inhomogeneity was explained in Supplementary Methods. 9.4.5 2D 13C/1H-13C Correlation Experiments Enabled by MAS-DNP In unlabeled samples, the natural-abundance of 13C isotope is very low (1.1%), and the probability of observing connectivity between two carbon-13 nuclei in a 2D 13C-13C correlation spectrum is inhibitory (0.01%). To obtain sufficient sensitivity for measuring 2D correlation experiments51, the soil and plant samples were measured on a Bruker 600 MHz/395 GHz MAS- 335 DNP system at National High Magnetic Field Laboratory, with the microwave irradiation power set to 12 W. The sample temperature was 104 K and 100 K when the microwave was on and off, respectively. The DNP buildup time was 1.3-4.5 s for all the MAS-DNP samples, including the HF-treated and untreated soil samples as well as the plant materials collected 30 m inland and at the edge of the island. Recycle delays were typically set to be 1.3-fold of the DNP buildup time constant for each sample. 1D 13C CP experiments were measured with and without microwave irradiation under 8 kHz for soil and 10.5 kHz for plant samples, with the CP contact time set to 1 ms. The experimental parameters for all 1D and 2D NMR and MAS-DNP experiments are documented in Table S9.3. 2D 1H-13C HETCOR experiments were carried out under 8 kHz or 10.5 kHz MAS frequencies. 1H-1H homonuclear decoupling was achieved using either Phase-Modulated Lee– Goldburg (PMLG)62 or Frequency-Switched Lee-Goldberg (FSLG) sequence63 with a 1H transverse field strength of 100 kHz, corresponding to an effective field strength of 122 kHz. To vary the range of detection between the proton and carbon sites, 1H magnetization was transferred to 13C using a Hartmann-Hahn (HH) CP block with a variable length, with 0.1 ms for primarily one-bond correlations, 0.5 ms for intermediate range of correlations, and 1.0 ms for long range correlations. 2D 13C-13C correlation experiments were carried out using the refocused INADEQUATE scheme64. The experiment was dipolar-based, using the broadband dipolar recoupling SPC5 sequence65. The MAS frequencies were 10.5 kHz for the HF-treated soil sample 1 and inland plants and changed to 8 kHz for the plant samples collected at the island edge. For the direct dimension (ω2), the acquisition time was 17 ms for all soil and plant samples. The acquisition time of the indirect dimension (ω1) was 2.7 ms and 1.7 ms for soil and plants, respectively. The indirect 336 dimensions of the spectra were set to 200 ppm (50 - 250 ppm) to effectively cover the double- quantum chemical shifts of carbohydrate and aromatic polymers. For each sample, 100 increments were collected for the indirect dimension. 320 scans were collected for the soil sample in 16 h, and 160 scans were collected for each of the two plant samples, with experimental time of 13 h and 23 h for the plants on the edge and inland, respectively. To rapidly identify the key carbohydrate components in soil, a probability map was built by extracting 412 datasets of plant carbohydrates from the Complex Carbohydrate Magnetic Resonance Database36 following a recently reported protocol66. All 13C and 1H chemical shifts of identified polymers are documented in Table S9.4. 9.4.6 14C Dating Prior to 14C dating, the soil sample was pretreated with an acid/alkali/acid solution to avoid potential effect of the secondary carbon components (roots, bacteria) on the determined age of the sample43, 44. The decayed plants in the soil were used for 14C dating, which was calibrated to radiocarbon age (years Before Present, yBP) and calendar years (cal AD). The analysis was performed using BetaCal 3.21, INTCAL13 database, and high probability density range method (HPD). The dataset is summarized in Table S9.5. 9.4.7 Bulk Density (BD) and Loss-On-Ignition (LOI) The BD was determined by drying the soil at 60 °C for 24 h in a muffle furnace and then calculated as oven-dry wight per unit volume at field moisture capacity.9 To determine the LOI, the dried material was ground with a mortar and pestle, and placed into a muffle furnace at 550 °C for 4 h. The mass difference before and after the combustion was divided by the original dry mass to get the percentage value of LOI ratio13, which represents the relative fraction of organic matter in the sample. The results of these bulk property measurements were detailed in Supplementary Methods. 337 9.4.8 Total Carbon Percentage The dried sample was ground using a mortar and pestle and sieved with 125 μm sieve to ensure equal particle size. 10 mg of soil were weighed into ceramic crucibles, which were placed into a total organic carbon analyzer (Shimadzu TOC SSM-5000A) to analyze the content of total carbon. Information on the physiochemical property is documented in Table S9.5. 9.5 Conclusions and Perspectives On aggregate, SOC sequestration in the studied coastal wetland is a combination of molecular—including biomolecular—preservation, recalcitrant carbon, and carbon stabilization through dynamic carbon speciation (biological carbon turnover). The hydrogeomorphic setting changed over the 1000 years during which this sequestration has taken place, transitioning from an active freshwater delta to an abandoned freshwater delta lobe, then to a brackish and, eventually, a salt marsh system, as sea level has continued to rise. Despite these drastic surface changes, preservation, via tighter packing, of parent biopolymers has been consistent over time. A new framework of terminology can be derived, in which preservation can be viewed as molecular preservation and sequestration can be viewed as carbon storage regardless of speciation, with preservation being a subcategory of sequestration. This study provides strong evidence for giving equal weight to POM, just as what has been done for MAOM in regard to global SOC management19, with POM being the major focus for organic soils, such as wetlands which contain ~1/3 of the planet’s SOC. Recalcitrance should also be a major part of the focus within the preservation of high organic soil, especially as this preserved SOC becomes quickly processed and converted to greenhouse gases when exposed to highly oxygenated water due to erosion12, 13, 67. Therefore, POM and molecular recalcitrance, including biopolymeric structures, are important and may become the main drivers in SOC sequestration for about 1/3 or more of the planet’s soil 338 organic carbon pool. 9.6 Acknowledgments This research was supported by the U.S. Department of Energy under the grant no. DE- SC0021210 to T.W. and the National Science Foundation Chemical Oceanography grants OCE- 1636052 and OCE-2054935 to J.R.W. and R.L.C. The National High Magnetic Field Laboratory is supported by the National Science Foundation through NSF/DMR-1644779 and DMR-2128556 and the State of Florida. The MAS-DNP system at NHMFL is funded in part by NIH P41 GM122698 and NIH RM1-GM148766. The authors thank Dr. Alex Kirui for initial data analysis, and M.P. Hayes and Benjamin Haywood for assistance with sample collection and bulk property measurements. 339 REFERENCES 1. Nahlik, A. M.; Fennessy, M. S., Carbon Storage in US Wetlands. Nat. Commun. 2016, 7, 13835. 2. Beillouin, D.; Corbeels, M.; Demenois, J.; Berre, D.; Boyer, A.; Fallot, A.; Feder, F.; Cardinael, R., A global meta-analysis of soil organic carbon in the Anthropocene. Nat. Commun. 2023, 14, 3700. 3. Lal, R., Soil Carbon Sequestration Impacts on Global Climate Change and Food Security. Science 2004, 304, 1623-1627. 4. Georgiou, K.; Jackson, R. B.; VVinduskova, O.; Abramoff, R. Z.; Ahlstrom, A.; Feng, W.; Harden, J. W.; Pellegrini, A. F.; Polley, H. W.; Soong, J. L.; Riley, W. J.; Tom, M. S., Global stocks and capacity of mineral-associated soil organic carbon. Nat. Commun. 2022, 13, 3797. 5. Ruehr, S.; Keenan, T. F.; Williams, C.; Zhou, Y.; Lu, X.; Bastos, A.; Canadell, J. G.; Prentice, I. C.; Sitch, S.; Terrer, C., Evidence and attribution of the enhanced land carbon sink. Nat. Rev. Earth Environ. 2023, 4, 518-534. 6. Murray, B.; Pendleton, L.; Jenkins, W.; Sifleet, S., Green Payments for Blue Carbon: Economic Incentives for Protecting Threatened Coastal Habitats Nicholas Institute for Environmental Policy Solutions, Duke University, Durham, 2011. 7. Spivak, A. C.; Sanderman, J.; Bowen, J. L.; Canuel, E. A.; Hopkinson, C. S., Global-change controls on soil-carbon accumulation and loss in coastal vegetated ecosystems. Nat. Geosci. 2019, 12, 685-692. 8. Wang, F.; Lu, X.; Sanders, C.; Tang, J., Tidal wetland resilience to sea level rise increases their carbon sequestration capacity in United States. Nat. Commun. 2019, 10, 5434. 9. Duarte, C. M.; Losada, I. J.; Hendriks, I. E.; Mazarrasa, I.; Marbà, N., The role of coastal plant communities for climate change mitigation and adaptation. Nat. Clim. Chang. 2013, 3, 961– 968. 10. Poffenbarger, H. J.; Needelman, B. A.; Megonigal, J. P., Salinity Influence on Methane Emissions from Tidal Marshes. Wetlands 2011, 31, 831–842. 11. Reddy, K. R.; DeLaune, R. D.; Inglett, P. W., Biogeochemistry of Wetlands: Science and Application. CRC Press: Boca Raton, 2008. 12. Haywood, B. J.; White, J. R.; Cook, R. L., Investigation of an early season river flood pulse: Carbon cycling in a subtropical estuary. Sci. Total Environ. 2018, 635, 867-877. 13. Haywood, B. J.; Hayes, M. P.; White, J. R.; Cook, R. L., Potential fate of wetland soil carbon in a deltaic coastal wetland subjected to high relative sea level rise. Sci. Total Environ. 2020, 711, 135185. 14. Pendleton, L.; Donato, D. C.; Murray, B. C.; Crooks, S.; Jenkins, W. A.; Sifleet, S.; Craft, 340 C.; Fourqurean, J. W.; Kauffman, J. B.; Marba, N.; Megonigal, P.; Pidgeon, E.; Herr, D.; Gordon, D.; Baldera, A., Estimating Global “Blue Carbon” Emissions from Conversion and Degradation of Vegetated Coastal Ecosystems. PLOS One 2012, 7, e43542. 15. Moomaw, W. R.; Chmura, G. L.; Davies, G. T.; Finlayson, C. M.; Middleton, B. A.; Natali, S. M.; Perry, J. E.; Roulet, N.; Sutton-Grier, A. E., Wetlands In a Changing Climate: Science, Policy and Management. Wetlands 2018, 38, 183-205. 16. Stevenson, F. J., Humus Chemistry: Genesis, Composition, Reactions. John Wiley & Sons: New York, NY, 1994. 17. Schmidt, M. W. I.; Torn, M. S.; Abiven, S.; Dittmar, T.; Guggenberger, G.; Janssens, I. A.; Kleber, M.; Kogel-Knabner, I. K.; Lehmann, J.; Manning, D. A. C.; Nannipieri, P.; Rasse, D. P.; Weiner, S.; Trumbore, S. E., Persistence of soil organic matter as an ecosystem property. Nature 2011, 478, 49-56. 18. Lehmann, J.; Kleber, M., The Contentious Nature of Soil Organic Matter. Nature 2015, 528, 60-68. 19. Angst, G.; Mueller, K. E.; Castellano, M. J.; Vogel, C.; Wiesmeier, M.; Mueller, C. W., Unlocking complex soil systems as carbon sinks: multi-pool management as the key. Nat. Commun. 2023, 14, 2967. 20. Lehmann, J.; Hansel, C. M.; Kaiser, C.; Kleber, M.; Maher, K.; Manzoni, S.; Nunan, N.; Reichstein, M.; Schimel, J. P.; Torn, M. S.; Wieder, W. R.; Kogel-Knabner, I. K., Persistence of soil organic carbon caused by functional complexity. Nat. Geosci. 2020, 13, 529-534. 21. Kleber, M.; Bourg, I. C.; Coward, E. K.; Hansel, C. M.; Myneni, S. C. B.; Nunan, N., Dynamic interactions at the mineral–organic matter interface. Nat. Rev. Earth Environ. 2021, 2, 402-421. 22. Hall, S. J.; Ye, C.; Weintraub, S. R.; Hockaday, W. C., Molecular trade-offs in soil organic carbon composition at continental scale. Nat. Geosci. 2020, 13, 687-692. 23. Boesch, D. F.; Jocelyn, M. N.; Mehta, A. J.; Morris, J. T.; Nuttle, W. K.; Simenstad, C. A.; Swift, D. J. P., Scientific Assessment of Coastal Wetland Loss, Restoration and Management in Louisiana. J. Coast. Res. 1994, 20, 1–103. 24. Jankowski, K. L.; Törnqvist, T. E.; Fernandes, A. M., Vulnerability of Louisiana’s Coastal Wetlands to Present-Day Rates of Relative Sea-Level Rise. Nat. Commun. 2017, 8, 14792. 25. Schmidt, M. W. I.; Knicker, H.; Hatcher, P. G.; Kogel-Knabner, I. K., Improvement of 13C and 15N CPMAS NMR Spectroscopy of Bulk Soils, Particle Size Fractions and Organic Material by Treatment with 10% Hydrofluoric Acid. Eur. J. Soil Sci. 1997, 48, 319-328. 26. Skjemstad, J.; Clarke, P.; Taylor, J.; Oades, J.; Newman, R., The Removal of Magnetic Materials from Surface Soils - A Solid State 13C CP/MAS NMR Study. Soil Res. 1994, 32, 1215- 1229. 341 27. Salati, S.; Adani, F.; Cosentino, C.; Torri, G., Studying Soil Organic Matter Using 13C CP- MAS NMR: The Effect of Soil Chemical Pre-Treatments on Spectra Quality and Representativity. Chemosphere 2008, 70, 2092-2098. 28. Phyo, P.; Wang, T.; Yang, Y.; O’Neill, H.; Hong, M., Direct determination of hydroxymethyl conformations of plant cell wall cellulose using 1H polarization transfer solid-state NMR. Biomacromolecules 2018, 19 (5), 1485-1497. 29. Wang, T.; Yang, H.; Kubicki, J. D.; Hong, M., Cellulose structural polymorphism in plant primary cell walls investigated by high-field 2D solid-state NMR spectroscopy and density functional theory calculations. Biomacromolecules 2016, 17 (6), 2210-2222. 30. Kang, X.; Kirui, A.; Dickwella Widanage, M. C.; Mentink-Vigier, F.; Cosgrove, D. J.; Wang, T., Lignin-polysaccharide interactions in plant secondary cell walls revealed by solid-state NMR. Nat. Commun. 2019, 10 (1), 347. 31. Kirui, A.; Zhao, W.; Deligey, F.; Yang, H.; Kang, X.; Mentink-Vigier, F.; Wang, T., Carbohydrate-aromatic interface and molecular architecture of lignocellulose. Nat. Commun. 2022, 13, 538. 32. Busse-Wicher, M.; Gomes, T. C. F.; Tryfona, T.; Nikolovski, N.; Stott, K.; Grantham, N. J.; Bolam, D. N.; Skaf, M. S.; Dupree, P., The pattern of xylan acetylation suggests xylan may interact with cellulose microfibrils as a twofold helical screw in the secondary plant cell wall of Arabidopsis thaliana Plant. J. 2014, 79, 492-506. 33. Simmons, T. J.; Mortimer, J. C.; Bernardinelli, O. D.; Poppler, A. C.; Brown, S. P.; deAzevedo, E. R.; Dupree, R.; Dupree, P., Folding of xylan onto cellulose fibrils in plant cell walls revealed by solid-state NMR. Nat. Commun. 2016, 7, 13902. 34. Grantham, N. J.; Wurman-Rodrich, J.; Terret, O. M.; Lyczakowski, J. J.; Stott, K.; Iuga, D.; Simmons, T. J.; Durand-Tardif, M.; Brown, S. P.; Dupree, R.; Busse-Wicher, M.; Dupree, P., An even pattern of xylan substitution is critical for interaction with cellulose in plant cell walls. Nat. Plants 2017, 859-865. 35. Hu, W.-G.; Mao, J.; Xing, B.; Schmidt-Rohr, K., Poly(Methylene) Crystallites in Humic Substances Detected by Nuclear Magnetic Resonance. Environ. Sci. Technol. 1999, 34, 530-534. 36. Kang, X.; Zhao, W.; Widanage, M. C. D.; Kirui, A.; Ozdenvar, U.; Wang, T., CCMRD: a solid-state NMR database for complex carbohydrates. J. Biomol. NMR 2020, 74 (4), 239-245. 37. Lattao, C.; Birdwell, J.; Wang, J. J.; Cook, R. L., Studying Organic Matter Molecular Assemblage within a Whole Organic Soil by Nuclear Magnetic Resonance. J. Environ. Qual. 2008, 37, 1501-1509. 38. Yan, J.; Wang, L.; Hu, Y.; Tsang, Y. F.; Zhang, Y.; Wu, J.; Fu, X.; Sun, Y., Composition Selects Different Soil Microbial Structures and in Turn Drives Different Litter Decomposition Pattern and Soil Carbon Sequestration Capability. Geoderma 2018, 319, 194-203. 342 39. Lovley, D. R.; Coates, J. D.; Blunt-Harris, E. L.; Phillips, E. J.; Woodward, J. C., Humic Substances as Electron Acceptors for Microbial Respiration. Nature 1996, 382, 445-448. 40. Kögel-Knabner, I.; de Leeuw, J. W.; Hatcher, P. G., Nature and Distribution of Alkyl Carbon the Origin and Humification of Aliphatic in Forest Soil Profiles: Implications for Biomacromolecules. Sci. Total Environ. 1992, 117-118, 175-185. 41. Johnson, R. L.; Schmidt-Rohr, K., Quantitative solid-state 13C NMR with signal enhancement by multiple cross polarization. J. Magn. Reson. 2014, 239, 44-49. 42. Coleman, J. M.; Roberts, H. H.; Stone, G. W., Mississippi River Delta: an Overview. J. Coast. Res. 1998, 14, 698-716. 43. Sapkota, Y.; White, J. R., Long-term fate of rapidly eroding carbon stock soil profiles in coastal wetlands. Sci. Total Environ. 2021, 753, 141913. 44. Sapkota, Y.; White, J. R., Marsh edge erosion and associated carbon dynamics in coastal Louisiana: A proxy for future wetland-dominated coastlines world-wide. Estuar. Coast. Shelf Sci. 2019, 226, 106289. 45. Alexander, J. S.; Wilson, R. C.; Green, W. R. A brief history and summary of the effects of river engineering and dams on the Mississippi River system and delta: U.S. Geological Survey Circular 1375; 2012; p 43 p. 46. DeLaune, R. D.; Nyman, J. A.; Patrick, W. H., Peat Collapse, Ponding and Wetland Loss in a Rapidly Submerging Coastal Marsh. J. Coast. Res. 1994, 10, 1021-1030. 47. Odum, W. E., Comparative Ecology of Tidal Freshwater and Salt Marshes. Annu. Rev. Ecol. Syst. 1988, 19, 147-176. 48. Simpson, A. J.; McNally, D. J.; Simpson, M. J., NMR spectroscopy in environmental research: From molecular interactions to global processes. Prog. Nucl. Magn. Reson. Spectr. 2011, 58, 97- 175. 49. Mao, J.; Cao, X.; Olk, D. C.; Chu, W.; Schmidt-Rohr, K., Advanced solid-state NMR organic matter. Prog. Nucl. Magn. Reson. Spectr. 2017, 100, 17-51. 50. Cook, R. L., Coupling NMR to NOM. Anal. Bioanal. Chem. 2004, 378, 1484-1503. 51. Marker, K.; Paul, S.; Fernandez-de-Alba, C.; Lee, D.; Mouesca, J. M.; Hediger, S.; De Paëpe, G., Welcoming natural isotopic abundance in solid-state NMR: probing π-stacking and supramolecular structure of organic nanoassemblies using DNP. Chem. Sci. 2017, 8, 974-987. 52. Ni, Q. Z.; Daviso, E.; Can, T. V.; Markhasin, E.; Jawla, S.; Swager, T. M.; Temkin, R. J.; Herzfeld, J.; Griffin, R. G., High frequency dynamic nuclear polarization. Acc. Chem. Res. 2013, 46, 1933-1941. 53. Angst, G.; Mueller, K. E.; Nierop, K. G. J.; Simpson, M. J., Plant- or microbial-derived? A 343 review on the molecular composition of stabilized soil organic matter. Soil Biol. Biochem. 2021, 156, 108189. 54. Lavallee, J. M.; Soong, J. L.; Cotrufo, M. F., Conceptualizing soil organic matter into particulate and mineral‐associated forms to address global change in the 21st century. Glob. Change Biol. 2020, 26, 261-273. 55. Freeman, C.; Ostle, N.; Kang, H., An enzymic 'latch' on a global carbon store. Nature 2001, 409, 149. 56. Gerke, J., Concepts and misconceptions of humic substances as the stable part of soil organic matter: A review. Agronomy 2018, 2018, 18258. 57. Cao, X.; Schmidt-Rohr, K., Abundant Nonprotonated Aromatic and Oxygen-Bonded Carbons Make Humic Substances Distinct from Biopolymers. Environ. Sci. Technol. Lett. 2018, 5, 476- 480. 58. De Nobili, M.; Bravo, C.; Chen, Y., The spontaneous secondary synthesis of soil organic matter components: A critical examination of the soil continuum model theory. App. Soil Ecol. 2020, 154, 103655. 59. Massiot, D.; Fayon, F.; Capron, M.; King, I.; Le Calvé, S.; Alonso, B.; Durand, J. O.; Bujoli, B.; Gan, Z.; Hoatson, G., Modelling one and two-dimensional solid-state NMR spectra. Magn. Reson. Chem. 2002, 40, 70-76. 60. Opella, S. J.; Frey, M. H., Selection of nonprotonated carbon resonances in solid-state nuclear magnetic resonance. J. Am. Chem. Soc. 1979, 101, 5854-5856. 61. Mentink-Vigier, F.; Marin-Montesinos, I.; Jagtap, A. P.; Halbritter, T.; van Tol, J.; Hediger, S.; Lee, D.; Sigurdsson, S. T.; De Paëpe, G., Computationally Assisted Design of Polarizing Agents for Dynamic Nuclear Polarization Enhanced NMR: The AsymPol Family. J. Am. Chem. Soc. 2018, 140 (35), 11013-11019. 62. Vinogradov, E.; Madhu, P. K.; Vega, S., High-resolution proton solid-state NMR spectroscopy by phase-modulated Lee–Goldburg experiment. Chem. Phys. Lett. 1999, 314, 443- 450. 63. Bielecki, A.; Kolbert, A. C.; De Groot, H.; Griffin, R. G.; Levitt, M. H., Frequency-Switched Lee—Goldburg Sequences in Solids. Adv. Magn. Reson. 1990, 14, 111-124. 64. Lesage, A.; Bardet, M.; Emsley, L., Through-bond carbon-carbon connectivities in disordered solids by NMR. J. Am. Chem. Soc. 1999, 188, 10987-10993. 65. Hohwy, M.; Jakobsen, H. J.; Eden, M.; Levitt, M. H.; Nielsen, N. C., Broadband dipolar recoupling in the nuclear magnetic resonance of rotating solids: A compensated C7 pulse sequence. J. Chem. Phys. 1998, 108, 2686-2694. 66. Zhao, W.; Debnath, D.; Gautam, I.; Fernando, L. D.; Wang, T., Charting the Solid-State 344 NMR Signals of Polysaccharides: A Database-Driven Roadmap. Magn. Reson. Chem. 2023, in press. 67. Steinmuller, H. E.; Hayes, M. P.; Hurst, N. R.; Sapkota, Y.; Cook, R. L.; White, J. R.; Xue, Z.; Chambers, L. G., Does edge erosion alter coastal wetland soil properties? A multi-method biogeochemical study. Catena 202, 187, 104373. 68. Mentink-Vigier, F.; Marin-Montesinos, I.; Jagtap, A. P.; Halbritter, T.; van Tol, J.; Hediger, S.; Lee, D.; Sigurdsson, S. T.; De Paëpe, G., Computationally Assisted Design of Polarizing Agents for Dynamic Nuclear Polarization Enhanced NMR: The AsymPol Family. J. Am. Chem. Soc. 2018, 140 (35), 11013-11019. 69. Viger-Gravel, J.; Lan, W.; Pinon, A.; Berruyer, P.; Emsley, L.; Bardet, M.; Luterbacher, J., Topology of Pretreated Wood Fibers Using Dynamic Nuclear Polarization. J. Phys. Chem. C. 2019, 123, 30407-30415. 70. Jepsen, R.; Roberts, J.; Lick, W., Effects of Bulk Density on Sediment Erosion Rates. Wat. Air Soil Pollut. 1997, 99, 21-31. 71. Craft, C., Freshwater Input Structures Soil Properties, Vertical Accretion, and Nutrient Accumulation of Georgia and U.S Tidal Marshes. Limnol. Oceanogr. 2007, 52, 1220-1230. 72. DeLaune, R. D.; Nyman, J. A.; Patrick, W. H., Peat Collapse, Ponding and Wetland Loss in a Rapidly Submerging Coastal Marsh. J. Coast. Res. 1994, 10, 1021-1030. 73. Qu, W.; Li, J.; Han, G.; Wu, H.; Song, W.; Zhang, X., Effect of Salinity on the Decomposition of Soil Organic Carbon in a Tidal Wetland. J. Soils Sediments 2018, 19, 487-496. 74. Remusat, L.; Hatton, P. J.; Nico, P. S.; Zeller, B.; Kleber, M.; Derrien, D., NanoSIMS Study of Organic Matter Associated with Soil Aggregates: Advantages, Limitations, and Combination with STXM. Environ. Sci. Technol. 2012, 46, 3943-3949. 75. Lehmann, J.; Solomon, D.; Kinyangi, J.; Dathe, L.; Wirick, S.; Jacobsen, C., Spatial complexity of soil organic matter forms at nanometre scales. Nat. Geosci. 2008, 1, 238-242. 76. Lattao, C.; Birdwell, J.; Wang, J. J.; Cook, R. L., Studying Organic Matter Molecular Assemblage within a Whole Organic Soil by Nuclear Magnetic Resonance. J. Environ. Qual. 2008, 37, 1501-1509. 77. Masoom, H.; Courtier-Murias, D.; Farooq, H.; Soong, R.; Kelleher, B. P.; Zhang, C.; Maas, W. E.; Fey, M.; Kumar, R.; Monette, M.; Stronks, H. J.; Simpson, M. J.; Simpson, A. J., Soil Organic Matter in Its Native State: Unravelling the Most Complex Biomaterial on Earth. Environ. Sci. Technol. 2016, 50, 1670-1680. 78. Kelleher, B. P.; Simpson, A. J., Humic Substances in Soils:  Are They Really Chemically Distinct? Environ. Sci. Technol. 2006, 40, 4605-4611. 79. Schmidt, M. W. I.; Torn, M. S.; Abiven, S.; Dittmar, T.; Guggenberger, G.; Janssens, I. A.; 345 Kleber, M.; Kogel-Knabner, I. K.; Lehmann, J.; Manning, D. A. C.; Nannipieri, P.; Rasse, D. P.; Weiner, S.; Trumbore, S. E., Persistence of soil organic matter as an ecosystem property. Nature 2011, 478, 49-56. 80. Simon, C.; Duhrkop, K.; Petras, D.; Roth, V. N.; Bocker, S.; Dorrestein, P. C.; Gleixner, G., Mass Difference Matching Unfolds Hidden Molecular Structures of Dissolved Organic Matter. Environ. Sci. Technol. 2022, 56, 11027-11040. 81. Hertkorn, N.; Frommberger, M.; Witt, M.; Koch, B. P.; Schmitt-Kopplin, P.; Perdue, E. M., Natural Organic Matter and the Event Horizon of Mass Spectrometry. Anal. Chem. 2008, 80, 8908- 8919. 346 S9.1 Supplementary Methods APPENDIX Effect of HF treatment on soil. The application of 2% HF solution efficiently depleted iron and mineral component in soil, allowing for a substantially higher concentration of SOM within a MAS rotor. Comparative analysis of ssNMR spectra showed no noticeable difference between HF- treated and untreated samples, whether at room temperature (Figure 9.1a) or at the cryogenic temperature used for DNP measurements (Figure S9.8a). Discrepancies were observed between MultiCP and DNP-enhanced CP spectra (Figure 9.1a, b), which were not caused by HF treatment but resulted from the application of higher magnetic fields and MAS frequencies, reducing the signals of aromatic motifs with significant chemical shift anisotropy (CSA). Inhomogeneity of DNP in soil samples. For DNP measurements, soil materials were doped with 10 mM of a stable biradical named AsymPolPok68. This biradical is water-soluble and has a small dimension of 1-2 nm, enabling penetration in porous materials. When the electron polarization is transferred to 1H in the DNP matrix, it relies on 1H-1H relayed transfer, ensuring efficient and homogenous polarization across a range of at least a hundred nanometers, as recently reported69. Typically, it is straightforward to achieve homogeneous polarization throughout the sample. However, within each soil sample, polarization levels varied among molecules and structural motifs. In the case of HF-treated sample 1, all carbohydrate signals exhibited a uniform 18-fold enhancement, whereas aromatic and aliphatic carbons experienced a comparatively lower enhancement of 7-9-fold (Figure S9.8b). Hence, the predominant portions of carbohydrates appear to be more accessible, which may partially explain their accelerated degradation rate through microbial processes in soil. Alternatively, carbohydrates may exhibit enhanced retention of solvent and biradicals, probably due to the rich polar functionalities present in carbohydrates. 347 Similar patterns were observed in non-treated soil sample 1 (Figure S9.8c) and non-treated samples #4 and #6 (Figure S9.8e, f), where carbohydrates consistently exhibited the highest enhancement, while other components displayed slightly lower enhancements. Remarkably, the most effective DNP performance was noted in non-treated soil sample 4, demonstrating a 33-fold enhancement for carbohydrates and a 23-fold enhancement for all other molecules (Figure S9.8e). Additionally, it is noteworthy that even in plant samples, polymethylene peaks exhibited minimal DNP enhancement (Figure S9.3b, d), indicating the presence of lipid polymers in self-aggregated domains that are challenging to penetrate, not only in soil but also in plants. This could also be due to faster 1H relaxation, leading to inadequate preservation of hyperpolarization. Bulk properties of wetland soil samples. During the Lafourche delta lobe formation, the Mississippi River ran through Barataria Bay and deposited silt and mineral matter into the surrounding wetland soils. The soil at 1.5-1.8 m depth formed during that period, exhibiting very low loss-on-ignition (LOI) ratios of 1.16-1.40 and total carbon (TC) percentage of 8.1-11.2%, but a high bulk density (BD) of 0.14-0.31 g/cm3 (Table S9.5). High BD indicates a high concentrations of large mineral particles, with the soil expected to erode more slowly70, while a low LOI ratio represent a low content of organic matter. When it came to the Plaquemines delta lobe, the Barataria Bay became a fresh marsh and expected to contain more organic matter71-73. Consistently, the soil at the depth of 1-1.5 m showed a high LOI ratio of 1.51-1.65, a high TC percentage of 10.7-14.0%, and a low BD of 0.11-0.18 g/cm3. The geological and historical timestamps explain the observed changes in the chemical composition of soil organic matter. Across all geological stages listed here, the soil formed during the Lafourche delta lobe is featured with a high content of aromatic carbons and a low amount of carbohydrates. This trend has been flipped in the soil associated with the Balize delta lobe. 348 S9.2 Supplementary Discussion S9.2.1 An Incomplete Molecular Characterization of SOM Advances in analytical instrumentation, in terms of resolution, sensitivity, and less harsh sample pre-treatment, in concert with some highly insightful experimental design, has allowed for new data and insights to emerge and a rethinking of SOM preservation. While there is no argument that these new techniques have allowed a major leap forward in our understanding of SOM and the associated SOC fraction, care must be taken when extrapolating the data interpretation to SOM as a whole. In addition to the bias on upland/mineral soil, as described in the Introduction, two guiding principles are also useful: 1) granularity (speciation versus molecular) and 2) bias on extraction and small molecule and the need for complementary characterization techniques. S9.2.2 Granularity Methods that interrogate whole soils, to date, provide data at an elemental to moiety level, e.g., aliphatic (alkyl), carbohydrate (O-alkyl), aromatic (aryl), aromatic with oxygen functional groups (O-aryl), phenolic, carbonyl group, etc., but not at the molecular level, which identifies exact molecular structures. This challenge is a result of the complexity of SOM as well as the limited resolution of each individual technique applied to whole soils to date. Some of the key techniques include: Nano Secondary Ion Mass Spectrometry (NanoSIMS), near-edge X-ray spectromicroscopy, and NMR, including 2D methods74-78. Consequently, unlike in other research fields, while we are able to interrogate a system and gain a general moiety (structural categories) level understanding, we are unable to ask direct questions regarding the more informative and scalable molecular level. 349 S9.3.3 Need for Complementary Technique for Soil Characterization Two techniques that have been able to overcome the granularity limit mentioned above are FT-ICR-MS and multidimensional liquid state NMR, each with their own limitations brought about by three biases. The extraction bias arises from the SOM being extracted and, since all too often, such extraction is incomplete79, thus these techniques are blind to the non-extracted fraction of SOM. The solubility bias stems from the fact that not all SOM, even extracted SOM, is equally soluble, with the more soluble fraction being more highly weighted. The third bias is the small molecule bias. For liquid state NMR, this bias is a result of molecular rotation, which leads to differences in relaxation behavior, especially T2 relaxation, with smaller molecules yielding sharper peaks and hence more intense signals compared to larger molecules that yield broader spectral peaks and tend to be lost due to the broad peak convolution for samples as complex as SOM extracts. On the FT-ICR-MS front80, 81, this small molecule bias, all other things being equal, results in a small molecule ionization preference. This means that the high-resolution FT-ICR and liquid state NMR data are biased towards smaller molecules, with an understanding that the absence of evidence of larger molecules does not mean they are not there. In this context, DNP- enabled high-resolution ssNMR can serve as a complementary technique to provide information on the large macromolecules in soil materials. 350 Table S9.1 Spectral deconvolution parameters of quantitative 13C spectra for HF-treated soil samples. The attributed chemical shifts (CSs) are labeled. The integral value (I), linewidth (L), and relative amplitude (A) are listed. Uncertain (U). Soil 1 Soil 2 Soil 3 Soil 4 Soil 5 Soil 6 Soil 7 Assign. CSs(ppm); I(%) CSs(ppm); I(%) CSs(ppm); I(%) CSs(ppm); I(%) CSs(ppm); I(%) CSs(ppm); I(%) CSs(ppm); I(%) [L(ppm), A(%)] [L(ppm), A(%)] [L(ppm), A(%)] [L(ppm), A(%)] [L(ppm), A(%)] [L(ppm), A(%)] [L(ppm), A(%)] CO Carbonyl CO 182.3; 0.28 183.8; 0.46 178.7; 1.89 181.0; 1.17 180.2; 0.92 179.4; 1.63 183.2; 0.66 [6.5, 0.2] [4.1, 0.5] [8.7, 1.2] [9.6, 0.7] [9.0, 0.6] [9.2, 1.0] [8.0, 0.5] 176.1; 3.06 176.2; 3.75 175.8; 2.78 175.7; 3.42 176.1; 2.70 175.7; 2.57 176.2; 3.00 [5.7, 2.8] [6.2, 3.0] [4.9, 3.0] [5.6, 3.6] [5.3, 3.0] [4.9, 3.0] [6.3, 3.0] CO H4 S3/5b 173.4; 3.29 173.6; 3.43 173.5; 3.40 173.1; 3.78 173.3; 4.92 173.1; 3.60 173.3; 4.21 [4.4, 3.9] [4.3, 3.9] [4.2, 4.3] [4.6, 4.5] [5.1, 5.7] [4.8, 4.4] [5.3, 5.1] 161.5; 0.16 162.8; 0.4 161.6; 0.07 161.5; 0.08 161.1; 0.07 162.7; 0.06 162.3; 0.12 [2.4, 0.4] [2.9, 0.8] [1.5, 0.2] [2.3, 0.2] [1.5, 0.3] [1.1, 0.3] [1.6, 0.5] 158.4; 0.23 158.1; 0.69 158.4; 0.36 157.8; 0.61 157.8; 0.66 159.0; 0.43 158.5; 0.43 [3.1, 0.4] [4.5, 0.7] [0.7, 2.9] [3.9, 0.9] [3.5, 1.1] [4.7, 0.5] [4.7, 0.6] Aromatic S3/5a G3 G4 154.1; 3.07 154.2; 2.24 154.2; 2.58 153.6; 2.35 153.9; 2.56 154.4; 3.11 154.4; 2.95 [5.2, 3.1] [4.4, 2.5] [4.8, 2.9] [5.0, 2.6] [5.2, 2.9] [5.5, 3.3] [5.5, 3.4] 148.7; 3.02 148.9; 1.84 149.4; 2.05 148.4; 2.51 148.5; 2.42 148.8; 3.08 148.3; 3.34 [6.5, 2.4] [4.6, 1.9] [4.7, 0.3] [5.8, 2.4] [6.0, 2.4] [6.4, 2.8] [7.2, 2.9] 144.4; 0.17 146.0; 0.50 146.6; 1.12 144.5; 0.64 144.6; 0.82 145.0; 0.40 141.8; 0.87 [2.0, 0.4] [4.0, 0.6] [5.3, 1.1] [6.8, 0.5] [5.6, 0.9] [4.2, 0.5] [9.1, 0.6] 351 Table S9.1 (cont’d) S1 G1/H1 H2/6 G6 G5, H3/5 G2 i/s1 Xn3f1 i4 s4 Xn2f4 136.6; 4.28 137.4; 2.99 136.2; 5.06 136.1; 4.87 137.5; 3.04 137.1; 4.24 137.2; 3.48 [7.9, 2.8] [7.5, 2.0] [9.2, 2.9] [9.0,3.0] [7.6, 2.4] [7.6, 3.2] [6.4, 3.4] 131.7; 4.63 132.6; 3.93 130.7; 5.86 130.9; 5.69 133.4; 3.56 133.0; 3.71 132.6; 4.70 [6.8, 3.6] [7.5, 2.6] [7.9, 3.9] [7.6, 4.2] [7.0, 3.0] [6.3, 3.4] [6.6, 4.5] 127.8; 3.52 128.9; 3.21 125.8; 2.28 126.5; 3.34 129.6; 6.25 129.4; 5.88 128.7; 6.38 [6.7, 2.8] [6.5, 2.4] [6.7, 1.8] [7.4, 2.5] [7.4, 5.0] [7.1, 4.7] [7.4, 5.4] 123.4; 2.51 123.9; 3.47 121.4; 2.27 122.3; 2.39 123.8; 4.36 123.8; 4.48 123.1; 4.98 [6.4, 2.0] [7.9, 2.6] [6.8, 1.8] [6.7, 2.0] [8.4, 3.1] [7.8, 3.3] [8.2, 3.9] 117.0; 4.52 117.1; 4.01 116.6; 3.52 116.7; 3.98 117.0; 4.42 116.9; 5.91 116.6; 4.16 [7.4, 3.4] [7.9, 2.6] [6.5, 3.0] [7.3, 3.2] [7.6, 3.6] [8.1, 4.4] [7.3, 3.8] 111.4; 1.95 110.8; 1.90 111.4; 1.93 111.2; 1.84 111.3; 1.58 111.0; 1.6 111.0; 2.17 [6.1, 1.7] [6.3, 1.5] [5.6, 1.8] [6.2, 1.6] [6.6, 1.4] [5.9, 1.6] [6.5, 2.1] 105.9; 3.59 105.7; 3.47 106.0; 2.36 105.3; 2.56 105.5; 1.95 106.1; 2.20 105.2; 2.52 [3.9, 4.8] [4.3, 4.0] [3.7, 3.4] [4.4, 3.2] [4.4, 2.6] [4.0, 3.2] [4.7, 3.4] 102.7; 0.86 101.9; 1.17 103.1; 0.90 101.9; 0.71 102.3; 0.83 103.1; 0.71 101.0; 0.08 [3.0, 1.5] [3.5, 1.6] [3.1, 1.6] 88.8; 0.46 88.2; 0.07 88.5; 0.12 [2.3, 1.3] [1.4, 0.3] [1.8, 0.4] [3.2, 1.2] 88.4; 0.1 [1.6, 0.4] [3.2, 1.5] [2.7, 1.5] [1.2, 0.4] 88.6; 0.01 88.6; 0.03 88.5; 0.03 [0.5, 0.1] [0.9, 0.2] [0.8, 0.3] 84.9; 1.14 84.9; 0.62 85.1; 0.67 84.7; 0.61 84.0; 0.82 85.0; 0.81 85.0; 0.73 [3.0, 2.0] [2.6, 1.2] [2.7, 1.3] [2.8, 1.2] [3.4, 1.4] [3.0, 1.6] [3.1, 1.5] 82.4; 1.22 82.6; 0.99 82.4; 0.98 82.2; 0.92 80.2; 0.97 81.7; 1.03 81.8; 0.82 [4.0, 1.6] [3.7, 1.3] [3.7, 1.4] [3.6, 1.4] [4.2, 1.3] [4.2, 1.4] [4.3, 1.2] Carbohydr ate 352 Table S9.1 (cont’d) C2,3,5 C2,3,5 C6 75.9; 7.10 76.0; 5.97 76.0; 5.42 75.9; 5.87 75.8; 4.12 76.2; 4.72 76.2; 3.29 [5.0, 7.4] [5.4, 5.4] [5.2, 5.5] [5.7, 5.7] [5.3, 4.5] [5.4, 5.0] [5.3, 3.9] 73.0; 7.90 72.9; 9.62 72.9; 8.10 72.4; 9.84 72.1; 7.13 72.7; 7.48 72.8; 7.97 [5.3, 7.8] [6.7, 7.0] [6.2, 6.9] [6.9, 7.9] [6.8, 7.1] [6.6, 6.5] [7.4, 6.8] 63.8; 6.32 63.5; 4.62 63.2; 4.90 62.6; 4.84 62.5; 7.13 63.1; 4.49 62.7; 3.48 [7.4, 4.5] [6.7, 3.3] [6.7, 3.8] [7.2, 3.7] [7.1, 3.2] [7.3, 3.6] [6.9, 3.2] -OCH3 b or 56.8; 6.43 56.8; 5.34 56.8; 6.55 56.4; 5.97 56.4; 5.61 56.6; 7.08 56.6; 5.95 protein Cα [4.7, 7.1] [4.5, 5.7] [4.8, 7.2] -OCH3 a or 51.1; 3.15 51.4; 3.29 50.8; 4.27 protein Cα [8.2, 2.0] [8.3, 2.0] [8.5, 2.7] [5.0, 6.6] 50.7; 3.7 [7.6, 2.7] [5.0, 6.6] [5.1, 8.0] [5.2, 7.2] 51.0; 4.80 50.7; 3.90 51.0; 3.08 [8.4, 3.4] [8.4, 2.7] [7.3, 2.7] R3CH 43.1; 3.0 43.1; 2.34 43.1; 2.30 43.2; 2.43 42.8; 3.30 43.3; 2.78 43.9; 2.27 & R4C [7.6, 2.1] [6.4, 1.8] [5.6, 2.2] [6.2, 2.2] [7.2, 2.7] [6.6, 2.4] [7.8, 1.9] Aliphatic -CCH -(CH2)n-a -(CH2)n-b 38.6; 1.46 39.3; 0.61 39.2; 1.35 38.6; 2.25 38.2; 2.76 38.9; 2.84 39.1; 4.28 [4.8, 1.6] [3.2, 1.0] [4.2, 1.7] [5.9, 2.1] [6.5, 2.5] [6.8, 2.4] [8.9, 3.0] 33.9; 3.52 33.9; 6.45 33.8; 4.45 33.5; 3.57 33.5; 2.97 33.9; 2.57 33.6; 2.61 [2.2, 8.5] [1.9, 16.8] [2.0, 12.0] [2.1, 9.3] [2.2, 7.7] [2.3, 6.5] [2.6, 6.4] 30.9; 5.45 31.0; 8.96 31.0; 6.81 30.6; 7.25 30.6; 6.70 31.0; 5.49 30.8; 4.04 [3.4, 8.4] [3.7, 11.6] [3.6, 9.8] [3.7, 10.9] [4.0, 9.8] [3.7, 8.5] [3.4, 7.5] -CH2- 27.0; 4.66 26.6; 4.31 26.6; 3.89 26.1; 4.44 26.1; 3.73 27.3; 2.98 27.6; 3.52 CH=CH- [6.3, 3.9] [5.2, 4.0] [5.4, 3.8] [5.9, 4.2] [5.7, 3.9] [5.6, 3.1] [6.2, 3.6] AcMe 23.9; 1.27 24.0; 1.93 23.9; 1.23 23.6; 0.83 23.6; 1.55 24.1; 2.66 23.6; 4.00 [3.2, 2.1] [3.6, 2.6] [3.3, 2.0] [2.5, 1.8] [3.6, 2.5] [4.5, 3.4] [6.2, 4.1] 353 Table S9.1 (cont’d) -CH3 -CH3 U 20.6; 1.98 20.4; 1.11 20.5; 2.21 20.3; 1.75 20.0; 1.98 20.4; 1.36 19.3; 1.13 [5.3, 2.0] [3.5, 1.5] [5.5, 2.1] [4.7, 2.0] [5.2, 2.2] [4.4, 1.8] [5.1, 1.4] 15.3; 2.54 15.5; 3.65 15.5; 2.21 15.6; 2.15 15.3; 2.56 15.8; 3.01 15.1; 2.88 [7.8, 1.7] [8.4, 2.1] [7.4, 1.6] [7.2, 1.7] [7.8, 1.9] [8.4, 2.1] [9.5, 1.9] 354 Table S9.2 Molecular composition of HF-treated soils based on spectral deconvolution results. Soil Depth Carbohyd Aromatic Aliphatic CO i4/s4 a S/G b c-(CH2)n- da/b -OCH3 Era # 1 2 3 4 5 6 7 (cm) 0-10 29.6% 29.0% 34.6% 6.8% 0.40 40-50 27.3% 25.9% 39.0% 7.9% 0.11 80-90 25.0% 28.9% 37.6% 8.6% 0.17 100-110 26.4% 29.3% 35.6% 8.7% 0.17 130-140 21.9% 31.3% 37.8% 9.0% 0.01 160-170 22.2% 34.0% 35.8% 8.1% 0.03 170-180 20.1% 35.7% 35.9% 8.4% 0.04 1.03 1.25 0.93 0.94 0.99 1.02 0.80 9.3 0.65 9.9% 2018 AD 15.8 0.72 8.9% 1963 AD 12 0.65 11.5% 1562 AD 11.2 0.49 10.0% 10.2 0.44 10.9% 1169 AD 8.3 7.1 0.47 11.3% 0.65 9.6% 945 AD a The ratio of interior cellulose to surface cellulose. Using integral of i4 peak over integrals of s4 peak. b Lignin aromatic ring S unit to G unit ratio. Using integrals of S3/5 peaks over integrals of G3/4 peaks. c Summation of integrals of aliphatic polymethylene -(CH2)n-a at 33 ppm and -(CH2)n-b at 31 ppm. d Ratio of integrals of aliphatic polymethylene -(CH2)n-a at 33 ppm and -(CH2)n-b at 31 ppm. 355 Table S9.3 Parameters of ssNMR experiments measured for soil and plant samples. HF-treated soil samples (Soil #HF); Non-treated soil samples (Soil #Non); NS: number of scans; number of points of time domain for the direct (td2) and indirect (td1) dimensions; the acquisition time of the direct dimension (aq2); the evolution time of indirect dimension (aq1); d1: recycle delay. N/A: not applicable. Sample DNP juice Experiment Time Figure # MAS (kHz) NS td2 td1 aq2 aq1 (ms) (ms) ssNMR #1-7HF 9.4 T 298 K #1Non #1HF #1-7HF N/A N/A N/A N/A 13C-MultiCP 35 h 1a, 3a, S6 16k 2494 24.9 13C-MultiCP 13C-CP 13C-NQS 35 h 9.5 h 9.5 h 1a S3a S3 14 16k 2494 16k 4096 N/A #1HF D2O/H2O, 9/1 13C-CP 2 min 1b, 2a, S8a, S8b d1 (s) 1 2 2 2 1.8 2.1 N/A 41 41 41 22.9 10.2 10.2 22.9 22.9 1.7 3 5.2 128 2048 N/A 10.2 N/A 2.0 16k/8k 4096 64 4096 1k 2048 512 2048 0.6 h S8a, S8c 5 min 0.25 h S8e S8f 0.5 h 2a, S4a, S4b 10.5 512 4096 0.75 h S4c, S4d 10.5 512 4096 8 8 8 8 8 #1HF D2O/H2O, 9/1 13C dipolar-INADEQUATE 16 h 1c 10.5 320 2048 98 17.2 2.7 1.8 plant (edge) D2O/H2O, 9/1 13C dipolar-INADEQUATE 13 h 2b, 2c 8 160 2048 100 17.2 1.7 2.9 plant (inland) D2O/H2O, 9/1 13C dipolar-INADEQUATE 23 h 10.5 160 2048 100 17.2 1.7 5.2 #1HF #1HF D2O/H2O, 9/1 1H-13C HETCOR (PMLG) 1 h d6-DMSO/D2O/H2O 1H-13C HETCOR (FSLG) 4.5 h 1g, S2 10.5 356 S5a 1e 40 96 4096 4096 38 64 22.9 3.8 1.8 22.9 2.4 2.6 #1Non #4Non #6Non d6-DMSO/H2O, 9/1 d6-DMSO/H2O, 9/1 d6-DMSO/H2O, 9/1 plant (edge) D2O/H2O, 9/1 plant (inland) D2O/H2O, 9/1 MAS DNP 14.1 T 100 K 13C-CP 13C-CP 13C-CP 13C-CP 13C-CP Table S9.3 (cont’d) #1Non #1Non #2HF #4Non #6Non #6Non d6-DMSO/H2O, 9/1 1H-13C HETCOR (FSLG) 12 h S7a, S7b d6-DMSO/H2O, 9/1 1H-13C HETCOR (FSLG) 24 h D2O/H2O, 9/1 1H-13C HETCOR (FSLG) 3 h S7a S7b d6-DMSO/H2O, 9/1 1H-13C HETCOR (FSLG) 1.4 h 3c, S7b d6-DMSO/H2O, 9/1 1H-13C HETCOR (FSLG) 10 h 3c, S7b d6-DMSO/H2O, 9/1 1H-13C HETCOR (FSLG) 7.3 h 3c 8 8 10.5 8 8 8 512 2048 1k 32 64 2048 2048 512 2048 512 2048 40 40 10.2 1.5 2.1 10.2 1.5 2.1 4096 128 22.9 4.8 2.6 40 40 30 10.2 1.5 2.0 10.2 1.5 1.7 10.2 1.1 1.7 plant (edge) D2O/H2O, 9/1 1H-13C HETCOR (FSLG) 1.7 h 2d, 2e 10.5 plant (inland) D2O/H2O, 9/1 1H-13C HETCOR (FSLG) 1.7 h S5b, S5c 10.5 8 8 4096 128 22.9 4.8 5.8 4096 128 22.9 4.8 5.8 357 Table S9.4 13C and 1H chemical shifts of soil and plant samples. The assignments are identified from 13C INADEQUATE spectra and 1H-13C HETCOR spectra. The ppm values are shown as 13C (1H), e.g., 88.9 (3.3) represents 13C (1H) chemical shifts (ppm) of cellulose i4. Aromatic rings include guaiacyl (G), syringyl (S), p-hydroxyphenyl (H) units. Not applicable (/). Unidentified (-). Uncertain (U). Carbohydrate Cellulose: Interior (i); Surface (s) Xylan: 2-fold (Xn2f); 3-fold (Xn3f) CO Sample 173.4 175.0 #1HF 179.6 173.3 C1 C2, 3, 5 iC4 sC4 iC6 sC6 C13f C42f C43f C52f C53f 104.9 74.7 (3.4) (4.3) 72.8 (3.5) 90.2 (3.6)e 85.5 (3.5)h 89.5 (3.9)a,b 84.4 (3.5)f 87.8 (3.8)c 83.3 (3.5)g 65.9 (3.5)c 62.4 (3.6)f 103.1 81.2 79.3 63.6 63.2 63.7 (3.5)a,b 60.6 (3.6)g (4.4) (3.6) (3.5) (-) (-) 105.1 (4.3) 75.3 (3.9) 90.2 (4.0)e 74.0 (3.6) 88.5 (4.8)a,b 72.5 (3.4) 87.6 (5.1)c 86.0 (4.9)h 65.4 (3.4)c 62.3 (3.4)f 102.7 82.0 79.0 83.7 (3.9)g 63.8 (3.7)a,b 61.1 (3.2)g (4.2) (4.2) (4.3) - - 175.3 #1Non 180.7 172.3 104.8 74.6 (3.3) 89.5 (3.4)a,b 86.7 (3.9)h 66.0 (3.8)c 63.0 (3.6)f 103.1 82.7 79.4 (4.3) 72.8 (3.4) 88.2 (3.7)c 83.9 (3.5)f,g 64.7 (3.5)a,b 61.6 (3.2)g (4.4) (3.7) (3.4) - - 175.6 #2HF 105.2 (4.4) 75.7 (3.7) 74.6 (3.6) 73.3 (3.7) 89.3 (4.2)a,b 88.0 (3.9)c 85.7 (3.9)h 84.3 (4.0)f 83.2 (3.9)g 65.7 (3.7)c 64.3 (3.7)a,b 62.7 (3.4)f 103.3 81.5 77.6 (4.7) (4.0) (3.7) - - 175.6 #4Non 180.1 180.4 174.0 358 Table S9.4 (cont’d) 75.8 (3.5) 105.0 (4.9) 74.6 (3.7) 72.5 (3.5) 89.4 (4.0)a,b 86.0 (3.7)h 66.2 (3.1)c 103.4 81.1 79.0 87.4 (3.9)c 83.9 (3.8)f,g 64.1 (3.1)a,b (4.9) (4.4) (4.3) 62.3 (3.3)f - - 176.3 #6Non 173.8 180.9 174.5 173.1 plant edge plant inland 105.0 74.5 (3.3) 88.8 (3.3)a,b 85.8 (3.7)h 101.9 81.4 78.8 65.2 63.2 64.5 (3.6)a,b 62.4 (3.6)f (4.2) 72.3 (3.3) 87.7 (3.7)c 83.7 (3.3)f,g (4.3) (3.5) (3.5) (-) (-) 105.0 74.7 (3.3) 89.0 (3.3)a,b 84.3 (3.3)f (4.3) 72.2 (3.4) 87.4 (3.7)c 83.0 (3.4)g 64.6 (3.6)a,b 62.7 (3.4)f 101.9 81.4 78.9 (4.3) (3.5) (3.4) - 63.6 (-) Aromatic S2/6 G2 G5, H3/5 G6 H2/6 G1/H1 S4 S1 G4 G3 S3/5 H4 Sample 105.1 111.5 114.9 117.9 (6.6) 124.7 128.5 132.3 136.5 145.6 148.1 153.6 (6.0) (6.1) (6.3) 121.0 (6.6) (6.7) (/) (/) (/) (/) (/) (/) - (/) 104.6 110.1 114.3 118.9 (6.5) 126.8 129.4 133.4 136.4 145.4 147.7 152.6 162.1 (6.0) (6.0) (6.2) 122.2 (6.5) (5.8) (/) (/) (/) (/) (/) (/) (/) 105.1 111.3 115.8 117.7 (6.7) 125.9 129.7 132.4 136.1 143.4 147.1 153.4 163.3 (6.1) (6.3) (6.4) 121.7 (6.7) (6.9) (/) (/) (/) (/) (/) (/) (/) 106.9 111.1 115.5 118.9 (6.1) 126.7 130.0 133.1 136.8 143.9 148.4 152.5 159.7 (6.2) (6.1) (6.0) 122.2 (6.4) (6.0) (/) (/) (/) (/) (/) (/) (/) #1HF #1Non #2HF #4Non 359 Table S9.4 (cont’d) 105.7 113.0 115.6 118.4 (6.2) 127.3 130.8 133.7 136.8 144.5 147.7 152.9 161.6 (6.1) (6.0) (6.7) 122.2 (6.1) (6.1) (/) (/) (/) (/) (/) (/) (/) - - 111.9 116.0 121.6 127.4 130.6 132.8 137.0 144.2 147.6 153.0 (6.7) (6.6) (6.8) (6.7) (/) (/) (/) (/) (/) (/) 111.8 116.0 121.7 126.6 131.8 133.8 145.1 147.9 152.8 (6.3) (6.5) (7.3) (6.8) (/) (/) (/) (/) (/) - - - #6Non plant edge plant inland -CH3 -OCOCH3 -CH2- CH=CH- -CH2-b -CH2-a -CCH R3CH & R4C -OCH3 a -OCH3 b Sample Aliphatic 17.9 (0.8) 18.7 (0.6) 18.0 (0.7) 18.5 (0.5) 19.5 (0.7) 23.2 (1.6) 23.8 (1.3) 23.1 (1.1) 22.4 (1.6) 23.2 (1.6) 25.7 (1.4) 25.9 (1.6) 26.4 (1.1) 26.7 (1.3) 27.9 (1.1) 30.6 (1.4) 30.1 (1.4) 30.1 (1.4) 30.2 (1.1) 31.0 (1.2) 33.4 (1.3) 33.6 (1.3) 33.4 (1.3) 34.0 (1.3) 34.0 (1.3) 40.8 (1.8) 38.2 (2.8) 39.6 (1.8) 39.4 (2.8) 40.1 (2.4) 360 43.6 (2.9) 43.3 (3.3) 44.3 (3.1) 43.5 (2.8) 44.6 (2.2) 50.8 (4.1) 50.0 (2.5) 50.0 (4.2) 50.6 (3.2) 51.8 (2.5) 56.6 (3.5) 56.5 (3.3) 56.3 (3.6) 56.3 (3.2) 56.4 (2.9) #1HF #1Non #2HF #4Non #6Non Table S9.4 (cont’d) 22.0 - 18.2 (1.0) (1.9) 22.1 (1.0) 25.8 (1.4) 26.0 (1.4) 30.2 (1.8) 30.5 (1.3) 32.6 (1.3) 32.8 (1.3) 37.9 (5.2) 38.0 (5.2) - - - - 56.2 (3.5) 58.0 (3.4) plant edge plant inland 361 Table S9.5 Physiochemical properties of soil samples based on the depth. Depth is an average value, for example, 5 cm depth means the sample is collected from 0-10 cm section. Samples #1 to #7 are the samples measured by ssNMR and DNP. Average Bulk Density Loss on Ignition Total Carbon Soil Sample # Depth (cm) (g/cm3) (Wt %) 1 2 3 4 5 6 7 5 15 25 35 45 55 65 75 85 95 105 115 125 135 145 155 165 175 20.5 27.1 17.9 28.5 19.9 13.7 40.0 43.2 40.9 31.1 51.6 59.4 35.4 64.8 60.7 51.6 39.7 16.1 0.265 0.240 0.318 0.247 0.294 0.368 0.141 0.166 0.166 0.207 0.152 0.117 0.182 0.114 0.119 0.146 0.135 0.306 362 (%) 8.356 10.826 12.326 14.039 10.717 11.165 8.114 Figure S9.1 Wetland soil from a recently vanished brackish island. a, The island is 55 km southeast of New Orleans, and 160 km away from Baton Rouge, the capital city of the state of Louisiana, USA. b, Soil materials were collected in February 2018 from a brackish island in Barataria Bay (GPS coordinates: 29°26'36.9"N, 89°53'59.0"W). c, Picture of the Spartina alterniflora-dominated island with the two sample sites marked. d, Location and depth of seven soil samples used for structural characterization. Two poles were used to extract the soil materials, which were divided into 10-cm sections, based on the depth. e, Timetable summarizing the landscape change of the island and adjacent lands over two decades. The catastrophic hurricanes that affected this island and the category (Cat.) numbers of these hurricanes are labeled. Blue dashline arrows indicate the ten time points where pictures of the landscape are provided. f, Zoom- in view of the dashline boxes in panel (e). Positions of the two sample sites are marked using magenta circles to guide the comparison. The island has been finally disappeared in 2021. 363 Figure S9.2 DNP 2D 1H-13C correlation spectra of unlabeled HF treated soil. a, 2D 1H-13C correlation spectra measured with short (0.1 ms), medium (0.5 ms), and long (1.0 ms) of CP contact times of unlabeled soil sample 1. The blue dash lines show the key proton positions of aliphatics, carbohydrates, and aromatics. The spectrum with 1 ms CP shows intermolecular cross peaks between aromatics and carbohydrates. b, Representative 1H cross sections extracted at different carbon sites from the 0.1 ms (black) and 1 ms (red) CP contact times. c, Representative 13C cross sections from the 0.1 ms (black) and 1 ms (red) CP contact times. The spectra were measured at 10.5 kHz. 364 Figure S9.3 1D non-quaternary suppression selectively detecting non-protonated carbons. a, Comparison of 13C CP and NQS spectra of the HF treated soil sample #1 at room temperature. The difference spectrum shows only protonated carbons. b, Overlay of 1D NQS 13C spectra of seven soil samples. All spectra are normalized by the CO peak (asterisk). QNS spectra mainly shows non-protonated carbons, with methyl carbons as an exception due to their rapid molecular motions. 365 Figure S9.4 DNP and EPR spectra of plant samples. a, DNP enhances the sensitivity by 24-fold for the plant on top of soil (on the edge of the island). b, Magnification of the microwave (MW) off spectra showed overall consistent pattern with the MW-on spectra, revealing homogeneous polarization by DNP, except for the polymethylene peaks. c and d, The 30 m inland plant sample also showed 18-fold of DNP enhancement, with homogeneous DNP of carbohydrates and aromatics as shown in panel e. f, Room temperature EPR spectra of AsymPolPOK (D2O/H2O, 90/10 Vol%) at 9.6 GHz for these inland plants. 366 Figure S9.5 2D 13C/1H-13C spectra of unlabeled plants 30-m inland. a, Carbohydrate region of DNP enhanced 13C-13C refocused INADEQUATE spectrum of unlabeled plant samples collected 30-m inland. Signals are resolved for cellulose and xylan. b, 2D Carbohydrate and aliphatic region of 2D 1H-13C HETCOR spectrum of the 30-m inland plant. A short 0.1 ms CP was used to emphasize the one-bond correlations. c, The aromatic region collected with long (1 ms) CP contact to show aromatic-aliphatic correlations. No cross peaks were observed with carbohydrates. The spectroscopic features are largely consistent in the plants collected at different locations of the island. 367 Figure S9.6 Spectral deconvolution of quantitative 13C spectra for molecular composition. For each sample, the simulated spectra (dark red) fit the experimentally measured 1D 13C MultiCP spectra (black). Underneath are the individual peaks that contribute to carbohydrate (orange), aliphatic (brown), aromatic (green) and carbonyl sites (blue). The peak list is guided by the resolvable sites obtained from high-resolution 2D data. Information on the deconvolution was documented in Table S9.1. 368 Figure S9.7 2D 1H-13C correlation DNP spectra of untreated soil. a, 2D 1H-13C correlation spectra of untreated soil sample 1 measured with 0.1 ms (yellow), 0.5 ms (blue), and 1 ms (grey) CP contact times. b, Zoom-in regions of carbohydrate and aliphatic signals in three untreated soil samples (1, 4, and 6). c, Additional 2D 1H-13C spectra of HF-treated soil sample 2. The key signals of carbohydrates and aromatics are observable. Top and bottom panels show the aliphatic/carbohydrate and aromatic signals, respectively. 369 Figure S9.8 1D DNP 13C spectra and EPR of soil samples. a, Comparison of 13C spectra of the HF and non-HF treated materials of soil sample #1 under DNP enhancement. b, Comparison of 13C spectra with and without microwave (MW) irradiation collected on HF-treated soil sample 1. The enhancement is 18-fold for carbohydrates, 9-fold for aromatics, 7-fold for CO, 9-fold for most aliphatic carbons, and 7-fold for the polymethylene CH2 peak. c, MW-on and MW-off spectra of non-treated soil sample 1. The enhancement is 13-fold for carbohydrates, 9-fold for aromatics, 8- fold for CO, and 8-fold for most aliphatic carbons, and 9-fold for the polymethylene CH2 peak. d, EPR spectra of soil sample #1 with (top) and without (bottom) HF treatment, hydrated using the d6-DMSO/D2O/H2O matrix. e, 13C spectra with and without microwave (MW) irradiation collected on native soil sample #4 without HF treatment, showing enhancement factor of 33-fold for carbohydrate and 23-fold for all other carbon sites. f, MW-on and MW-off spectra of untreated soil sample #6, showing enhancement of 15 for carbohydrate and 9-11 for other carbon sites. g, EPR spectra of AsymPolPOK at 9.6 GHz for untreated sample #4 (green) and #6 (blue), with a solvent of d6-DMSO/D2O (90:10 Vol%). 370 CHAPTER 10: CCMRD: A SOLID-STATE NMR DATABASE FOR COMPLEX CARBOHYDRATES Research paper reprinted with permission: J. Biomol. NMR 74, 239-250 (2020) Authors: Xue Kan, Wancheng Zhao, Malitha C. Dickwella Widanage, Alex Kirui, Uluc Ozdenvar, and Tuo Wang 10.1 Abstract Carbohydrates are essential to various life activities in living organisms and serve as the central component in many biomaterials. As an emerging technique with steadily improving resolution, solid-state Nuclear Magnetic Resonance (NMR) spectroscopy has the unique capability in revealing the polymorphic structure and heterogeneous dynamics of insoluble complex carbohydrates. Here, we report the first solid-state NMR database for complex carbohydrates, Complex Carbohydrates Magnetic Resonance Database (CCMRD). This database currently holds the chemical shift information of more than four hundred solid-state NMR compounds and expects rapid expansion. CCMRD provides open portals for data deposition and supports search options based on NMR chemical shifts, carbohydrate names, and compound classes. With the timely implementation, this platform will facilitate spectral analysis and structure determination of carbohydrates and promote software development to benefit the research community. The database is freely accessible at www.ccmrd.org. 10.2 Introduction Complex carbohydrates play central roles in many biological processes such as energy storage, structural building, and cell recognition1-3. These biomacromolecules also form the basis for novel biomaterials such as the scaffolds developed for tissue engineering and the carriers for drug delivery4, 5. Polysaccharides are the polymers of the monosaccharide building blocks linked 371 by glycosidic bonds. Their structural complexity is multifaceted, including substantial variations in the monosaccharide composition, the glycosidic linkages and the anomeric configuration of subunits, the branching pattern of the backbone by sidechains, chemical modifications such as acetylation and methyl esterification, hydrogen bonding patterns, and more subtly, torsional conformations (Figure 10.1)6-11. The structural complexity determines the physical and chemical properties and leads to the highly heterogeneous dynamics, hydration and intermolecular interactions of carbohydrates in biomaterials. Figure 10.1 Representative structure of complex carbohydrates from different origins. Most of these polysaccharides have been characterized by solid-state NMR. The structures of polysaccharides in their cellular environments are often more complicated and irregular. 372 Complex carbohydrates are significantly under-investigated compared with other biomacromolecules such as nucleic acids and proteins. Polysaccharides are mostly insoluble in water, partially indigestible, and often non-crystalline, making it practically difficult to analyze their high-resolution structure12. The conventional methods typically need to hydrolyze the digestible polysaccharides to smaller, soluble segments13, 14 and grind or ball-mill the indigestible portion into fine particles to enable structural analysis by solution-NMR and mass spectrometry 15-18. The biochemical treatments could considerably perturb and even restructure the biomolecules. Due to these technical hurdles, our knowledge of the structure and dynamics of carbohydrates remains inadequate, which has impeded the development of carbohydrate-based biorenewable energy biomaterials15. Solid-state NMR (ssNMR) spectroscopy is an emerging technique that is capable of elucidating the molecular structure and dynamics of insoluble polysaccharides in their native cells or environments, without the need for pretreatments. This technique has long been employed to reveal the polymorphic structure of cellulose and matrix polysaccharides, which heavily relies on one-dimensional 13C spectra 19, 20. With better resolution, multidimensional ssNMR has further revealed biopolymer interactions in plant primary and secondary cell walls, the supramolecular architecture and pigment deposition of pathogenic fungal cell walls and biofilms, the composition of mammalian cells, the structure of bacterial peptidoglycans and lipopolysaccharides and their interactions with antibiotics, the difference of galactolipids, storage polysaccharides and cell wall components in algae, as well as naturally modified or artificially functionalized polysaccharides21- 42. With the readily improving resolution from the improvement and development of ultrahigh- field magnets43, the sensitivity enhancement from dynamic nuclear polarization (DNP) and solid- state CryoProbe44-48, and the assistance from various NMR methods such as ultrafast spinning and 373 proton detection49-51 as well as paramagnetic and spectral editing techniques52-54, many long- standing questions regarding the structure and dynamics of complex carbohydrates now become feasible and await systematic investigations. At the same time, the rapid expansion of ssNMR applications in carbohydrate research inevitably necessitates the development of a databank to facilitate information storage and sharing as well as statistical analysis and software development. Here we developed the Complex Carbohydrates Magnetic Resonance Database (CCMRD) together with its web interface for public access, information search, and data deposition. The database is freely available at www.ccmrd.org. 10.3 Methods 10.3.1 Database Assembly The CCMRD system is constructed with a three-tier architecture: server, client and database. The data records are stored in MySql database55. The initial data were collected from studies that were published over the past four decades. The database is regularly maintained and updated on a weekly basis. The structure of the database is designed to accommodate the complexity of polysaccharide structure. The data are organized in multiple levels to include as many details as possible. The base unit/monosaccharide serves as a central entity, which is directly related to the NMR chemical shifts and the branching pattern. Such information belongs to a compound, which is an upper-level central entity. The compound is connected to the information of organism, references and experimental details (Figure 10.2a). The organizational schema is shown in Figure S10.1. 10.3.2 Server/Web Interface The server/web interface was programmed using PHP/Laravel framework56. The website is currently hosted on Amazon Web Services. The web server uses Amazon cloud EC2 service, 374 and the mysql database uses Amazon cloud RDS service. A physical server is being implemented for long-term maintenance and development. The web interface provides three options for data search based on NMR chemical shifts, polysaccharide names, and compound classes. A data submission portal is also available, which contains a multi-page process for collecting all the required information. All the submitted data are reviewed before the final deposition. Figure 10.2 Overview of the carbohydrate database. a, Structure of the database. b, Search interface (left) and a representative entry (right). More detailed instructions for the search and data deposition are in Figure S10.2, S10.3. c, For possible applications, a cellulose 13C chemical shift heatmap (left) and a synthesized spectrum of 40 conformers from 9 plant carbohydrates (right) are shown as examples. 10.4 Results and Discussion To date, CCMRD stores the ssNMR chemical shifts of 435 complex carbohydrates from various organisms (Figure 10.3a). Plants account for half of the deposited compounds while fungi, 375 algae and bacteria share one-third of the publications. The NMR chemical shifts are obtained predominantly from high-resolution 2D/3D correlation solid-state NMR spectra that allows the unambiguous identification of carbon-carbon connectivity in each carbohydrate26, 57. Glycosyl linkage and carbohydrate compositional analysis, genetic mutants or chemical extraction of certain components, as well as Density Functional Theory (DFT) calculations are often coupled to implement and effectively verify the NMR assignments. Critical breakthroughs of ssNMR technology have substantially promoted high-resolution studies over the last decade as evidenced by the upsurge of compound numbers (Figure 10.3b). 10.4.1 Data Search The web interface provides three options for data search (Figure S10.2): (1) Search by chemical shifts: This feature is specially designed to help users with unknown signals that are identified in NMR spectra. Users need to specify three parameters include the nucleus of interest, the chemical shift and the threshold. The nucleus is chosen from 13C, 1H, 15N, and 31P. The threshold determines how accurate the search results should match the targeted value. The default threshold is set to 1 ppm for 13C, 3 ppm for 15N/31P and 0.5 ppm for 1H based on the typical ssNMR resolution and these values can be changed by the users. The returned result is a list of compounds with matched signals, with the detailed information of carbohydrate structure and experimental conditions underneath each CCMRD compound ID (Figure 10.2b). In addition, CCMRD also support double-signal search using two chemical shifts from the same compound. By inputting two chemical shifts separated by a comma, the server will return the entries that match both the inputs within the specified range of tolerance. For each returned entry, its ID number, compound name, residue name and chemical shifts are listed, with the 376 matched chemical shifts highlighted. The double-signal search can efficiently eliminate irrelevant compounds. Figure 10.3 Statistics of compounds in CCMRD. The data are categorized by a, taxonomical context and b, years. (2) Search by polysaccharide names: The users can conduct a search using a trivial name (Figure S10.2). There is no restriction on the input, and both full and partial names are accepted. For example, the users can use either “cel” or “cellulose”. CCMRD will perform a search for the name pattern and return all relevant entries with their ID numbers, with the full details accessible by clicking each ID. (3) Search by compound classes: Based on the currently available entries, the data deposited in CCMRD are categorized as six classes to facilitate the data search: polysaccharide, oligosaccharide, monosaccharide, lipopolysaccharide, N-acetylglucosamine, and peptidoglycan. This option is useful when the compound type is already known. To eliminate the potential ambiguity in the atom position labeling of carbohydrates and to provide a more intuitive way to present the data, we have included the molecular structure of each residue with carbon atom numbers so that users can access the data with ease and accuracy. The Symbol Nomenclature for Glycans (SNFG) is also included for full structure presentation but only for those entries with unambiguous information of the linkage patterns58, 59. 377 10.4.2 Data Deposition The web server of CCMRD also provides a portal for data submission (Figure 10.4 and Figure S10.3). The users will need to fill a multi-step form for multiple instances of all the complex carbohydrates in a single publication. The submitted data will be saved in a temporary table and an email will be generated automatically informing the users of the successful submission. An administrator will review, verify, and deposit the submitted record, and a second email containing the CCMRD ID number(s) will be sent automatically to the user to confirm the success of data deposition. The users will need to prepare the following information for data submission: (1) Provide an email. The email address is only used for communications before the final deposition process. (2) NMR experimental conditions. The detailed NMR experimental conditions are required to be specified, including magnetic field strengths, sample pH, temperature, NMR reference scale, and sample processing protocols. As some data from the earlier stages of ssNMR has relatively low resolution as measured on low-field magnets, the experimental conditions will assist the research community in evaluating data reliability. Taxonomic domain of the organism and the biological source with full name of genus and species are also required. NCBI TaxID from the NCBI taxonomy database is used for cross-reference. Fields with asterisks are required while the others are optional (Figure 10.4). (3) Compound information. Users are required to provide both the trivial name and linear code of the compound. For the linear code format, we followed the IUPAC recommendations for future compatibility60. Compound class also need to be specified, for example, as polysaccharide, monosaccharide, or oligosaccharide. 378 (4) Residue/Unit. For each residue, the users need to provide the residue/unit name, anomeric configuration, absolute configuration and ring size. Select “unknown” if any field is undetermined. After inputting all the chemical shifts for this residue, additional residues can be added by clicking the “add new residue” button for heteropolysaccharide. (5) Connection table. The patterns of covalent linkage between different residues need to be clarified by specifying the connecting residues (chosen from previously entered residues in the residues/unit section) and the corresponding connecting position/site. Multiple connection entries can be added for heteropolysaccharide. Until here, user will finish a compound. To add more compounds from the same publication, use “add new compound” and repeat steps (3) – (5). (6) Reference. The reference type needs to be specified as book chapters or journal articles. Information (title, authors, year etc.) can be entered using the corresponding forms. Both PubMed ID and DOI number are required for cross-referencing to other bibliographic databases. (7) Review submission. A summary of the information will be provided to the users for review before completing the submission. 10.4.3 Current Status and Limitation With the recent advances in instrumentation and methodology, ssNMR has become a powerful tool for characterizing the structure and dynamics of complex carbohydrates in-vivo and in-situ. Multidimensional ssNMR can efficiently identify the constituent monosaccharide units and partially reveal the pattern of covalent linkages of the polysaccharides 6. The current studies mainly focus on the polysaccharides with highly repetitive structures, including the linear polymers of a single monosaccharide unit or multiple residue types but with an alternating pattern, as well as those polysaccharides with identical patterns of branching. Solving the full structure of more irregular carbohydrates remains challenging. At this stage, only high-resolution data (mostly with 379 2D/3D correlation spectra) are indexed in CCMRD, and we only accept data from peer-reviewed publications, with every record manually examined for reliability. With the rapid progress in carbohydrate ssNMR, we will also be able to develop statistics-based tools that automatically validate the new entries. Figure 10.4 Flowchart of data deposition. The entries listed here will be needed for depositing data into CCMRD. The step-by-step user guide is shown in Figure S10.3. 10.5 Conclusion and Outlook CCMRD presents a timely implementation to the existing databases of carbohydrates, such as Bacterial Carbohydrate Structure DataBase (BCSDB)61 and the GLYCOSCIENCES.de portal62, which collected data from traditional methods including x-ray, solution NMR and mass spectrometry. Together, these databases integrate and promote carbohydrate research and its relevant applications in the field of biomedical sciences, biomaterials and bioenergy. We will 380 continue updating CCMRD with trustworthy data sources and introducing auxiliary functions and software. The ongoing efforts include the implementation of new portals for including additional types of structural data (such as dynamics), the addition of web services in both json and xml formats, and the coding of novel functions for generating chemical shift heatmap63 and artificial spectra (Figure 10.2c). These efforts will facilitate the statistical analysis of carbohydrate composition and structure, which will ultimately lead to high-throughput spectral analysis using various algorithms including deep-learning neural networks. 10.6 Acknowledgement This work is supported by the National Science Foundation NSF OIA-1833040. 381 REFERENCES 1. Lowe, J. B.; Marth, J. D., A genetic approach to mammalian glycan function. Annu. Rev. Biochem. 2003, 72 (1), 643-691. 2. Raman, R.; Raguram, S.; Venkataraman, G.; Paulson, J. C.; Sasisekharan, R., Glycomics: an integrated systems approach to structure-function relationships of glycans. Nat. Methods 2005, 2 (11), 817. 3. Himmel, M. E.; Ding, S.-Y.; Johnson, D. K.; Adney, W. S.; Nimlos, M. R.; Brady, J. W.; Foust, T. D., Biomass recalcitrance: engineering plants and enzymes for biofuels production. Science 2007, 315 (5813), 804-807. 4. Collins, M. N.; Birkinshaw, C., Hyaluronic acid based scaffolds for tissue engineering-A review. Carbohyd Polym 2013, 92 (2), 1262-1279. 5. Kang, B.; Opatz, T.; Landfester, K.; Wurm, F. R., Carbohydrate nanocarriers in biomedical applications: functionalization and construction. Chem Soc Rev 2015, 44 (22), 8301-8325. 6. Wang, T.; Phyo, P.; Hong, M., Multidimensional solid-state NMR spectroscopy of plant cell walls. Solid State Nucl Mag 2016, 78, 56-63. 7. Buchanan, B. B.; Gruissem, W.; Jones, R. L., Biochemistry and Molecular Biology of Plants. American Society of Plant Physiologists, Rockville, MD: 2000. 8. Latge, J. P., The cell wall: a carbohydrate armour for the fungal cell. Mol. Microbiol. 2007, 66 (2), 279-90. 9. Aspinall, G. O., The Polysaccharides. 1st ed.; Academic Press, New York, NY: 1983. 10. Lindberg, B., Components of bacterial polysaccharides. Adv. Carbohydr. Chem. Biochem. 1990, 48, 279-318. 11. Albersheim, P.; Darvill, A.; Roberts, K.; Sederoff, R.; A., S., Plant Cell Walls. Garland Science, New York 2011. 12. Rytioja, J.; Hildén, K.; Yuzon, J.; Hatakka, A.; de Vries, R. P.; Mäkelä, M. R., Plant- polysaccharide-degrading enzymes from basidiomycetes. Microbiol. Mol. Biol. Rev. 2014, 78 (4), 614-649. 13. Wormald, M. R.; Petrescu, A. J.; Pao, Y.-L.; Glithero, A.; Elliott, T.; Dwek, R. A., Conformational studies of oligosaccharides and glycopeptides: complementarity of NMR, X-ray crystallography, and molecular modelling. Chem. Rev. 2002, 102 (2), 371-386. 14. Sluiter, J. B.; Ruiz, R. O.; Scarlata, C. J.; Sluiter, A. D.; Templeton, D. W., Compositional analysis of lignocellulosic feedstocks. 1. Review and description of methods. J. Agric. Food Chem. 2010, 58 (16), 9043-9053. 382 15. Duus, J. Ø.; Gotfredsen, C. H.; Bock, K., Carbohydrate structural determination by NMR spectroscopy: modern methods and limitations. Chem. Rev. 2000, 100 (12), 4589-4614. 16. Bubb, W. A., NMR spectroscopy in the study of carbohydrates: Characterizing the structural complexity. Concept Magn. Reson. A 2003, 19 (1), 1-19. 17. Cheng, K.; Sorek, H.; Zimmermann, H.; Wemmer, D. E.; Pauly, M., Solution-state 2D NMR spectroscopy of plant cell walls enabled by a dimethylsulfoxide-d6/1-ethyl-3-methylimidazolium acetate solvent. Anal. Chem. 2013, 85 (6), 3213-21. 18. Mansfield, S. D.; Kim, H.; Lu, F. C.; Ralph, J., Whole plant cell wall characterization using solution-state 2D NMR. Nat. Protoc. 2012, 7 (9), 1579-1589. 19. Atalla, R. H.; Vanderhart, D. L., Native cellulose: a composite of two distinct crystalline forms. Science 1984, 223 (4633), 283-285. 20. Newman, R. H.; Davies, L. M.; Harris, P. J., Solid-State 13C Nuclear Magnetic Resonance Characterization of Cellulose in the Cell Walls of Arabidopsis thaliana Leaves. Plant Physiol. 1996, 111 (2), 475-485. 21. Bardet, M.; Emsley, L.; Vincendon, M., Two-dimensional spin-exchange solid-state NMR studies of 13C-enriched wood. Solid State Nucl Mag 1997, 8 (1), 25-32. 22. Renault, M.; Tommassen-van Boxtel, R.; Bos, M. P.; Post, J. A.; Tommassen, J.; Baldus, M., Cellular solid-state nuclear magnetic resonance spectroscopy. Proceedings of the National Academy of Sciences of the United States of America 2012, 109 (13), 4863-8. 23. Phyo, P.; Wang, T.; Xiao, C.; Anderson, C. T.; Hong, M., Effects of Pectin Molecular Weight Changes on the Structure, Dynamics, and Polysaccharide Interactions of Primary Cell Walls of Arabidopsis thaliana: Insights from Solid-State NMR. Biomacromolecules 2017, 18 (9), 2937- 2950. 24. Wang, T.; Yang, H.; Kubicki, J. D.; Hong, M., Cellulose Structural Polymorphism in Plant Primary Cell Walls Investigated by High-Field 2D Solid-State NMR Spectroscopy and Density Functional Theory Calculations. Biomacromolecules 2016, 17 (6), 2210-22. 25. Simmons, T. J.; Mortimer, J. C.; Bernardinelli, O. D.; Poppler, A. C.; Brown, S. P.; deAzevedo, E. R.; Dupree, R.; Dupree, P., Folding of xylan onto cellulose fibrils in plant cell walls revealed by solid-state NMR. Nat. Commun. 2016, 7, 13902. 26. Kang, X.; Kirui, A.; Muszyński, A.; Widanage, M. C. D.; Chen, A.; Azadi, P.; Wang, P.; Mentink-Vigier, F.; Wang, T., Molecular architecture of fungal cell walls revealed by solid-state NMR. Nat. Commun. 2018, 9 (1), 2747. 27. Kang, X.; Kirui, A.; Widanage, M. C. D.; Mentink-Vigier, F.; Cosgrove, D. J.; Wang, T., Lignin-polysaccharide interactions in plant secondary cell walls revealed by solid-state NMR. Nat. Commun. 2019, 10 (1), 347. 383 28. Takahashi, H.; Ayala, I.; Bardet, M.; De Paëpe, G.; Simorre, J. P.; Hediger, S., Solid-state NMR on bacterial cells: selective cell wall signal enhancement and resolution improvement using dynamic nuclear polarization. J. Am. Chem. Soc. 2013, 135 (13), 5105-10. 29. Kim, S. J.; Chang, J.; Singh, M., Peptidoglycan architecture of Gram-positive bacteria by solid-state NMR. Bba-Biomembranes 2015, 1848 (1), 350-362. 30. Thongsomboon, W.; Serra, D. O.; Possling, A.; Hadjineophytou, C.; Hengge, R.; Cegelski, L., Phosphoethanolamine cellulose: A naturally produced chemically modified cellulose. Science 2018, 359 (6373), 334-338. 31. Werby, S. H.; Cegelski, L., Spectral comparisons of mammalian cells and intact organelles by solid-state NMR. Journal of structural biology 2018. 32. Chatterjee, S.; Prados-Rosales, R.; Itin, B.; Casadevall, A.; Stark, R. E., Solid-state NMR Reveals the Carbon-based Molecular Architecture of Cryptococcus neoformans Fungal Eumelanins in the Cell Wall. J. Biol. Chem. 2015, 290 (22), 13779-13790. 33. White, P. B.; Wang, T.; Park, Y. B.; Cosgrove, D. J.; Hong, M., Water-Polysaccharide Interactions in the Primary Cell Wall of Arabidopsis thaliana from Polarization Transfer Solid- State NMR. J Am Chem Soc 2014, 136 (29), 10399-10409. 34. Grantham, N. J.; Wurman-Rodrich, J.; Terrett, O. M.; Lyczakowski, J. J.; Stott, K.; Iuga, D.; Simmons, T. J.; Durand-Tardif, M.; Brown, S. P.; Dupree, R.; Busse-Wicher, M.; Dupree, P., An even pattern of xylan substitution is critical for interaction with cellulose in plant cell walls. Nat. Plants 2017, 3 (11), 859-865. 35. Arnold, A. A.; Bourgouin, J. P.; Genard, B.; Warschawski, D. E.; Tremblay, R.; Marcotte, I., Whole cell solid-state NMR study of Chlamydomonas reinhardtii microalgae. J. Biol. NMR 2018, 70, 123-131. 36. Cadars, S.; Lesage, A.; Emsley, L., Chemical shift correlations in disordered solids. J Am Chem Soc 2005, 127 (12), 4466-76. 37. Arnold, A. A.; Genard, B.; Zito, F.; Tremblay, R.; Warschawski, D. E.; Marcotte, I., Identification of lipid and saccharide constituents of whole microalgal cells by 13C solid-state NMR. Biochim. Biophys. Acta 2015, 1848 (1 Pt B), 369-77. 38. Laguri, C.; Silipo, A.; Martorana, A. M.; Schanda, P.; Marchetti, R.; Polissi, A.; Molinaro, A.; Simorre, J. P., Solid State NMR Studies of Intact Lipopolysaccharide Endotoxin. ACS chemical biology 2018, 13 (8), 2106-2113. 39. Terrett, O. M.; Lyczakowski, J. J.; Yu, L.; Iuga, D.; Franks, W. T.; Brown, S. P.; Dupree, R.; Dupree, P., Molecular architecture of softwood revealed by solid-state NMR. Nat. Commun. 2019, 10 (1), 4978. 40. Kern, T.; Giffard, M.; Hediger, S.; Amoroso, A.; Giustini, C.; Bui, N. K.; Joris, B.; Bougault, C.; Vollmer, W.; Simorre, J. P., Dynamics characterization of fully hydrated bacterial 384 cell walls by solid-state NMR: evidence for cooperative binding of metal ions. J Am Chem Soc 2010, 132 (31), 10911-9. 41. Kern, T.; Hediger, S.; Müller, P.; Giustini, C.; Joris, B.; Bougault, C.; Vollmer, W.; Simorre, J. P., Toward the characterization of peptidoglycan structure and protein-peptidoglycan interactions by solid-state NMR spectroscopy. J. Am. Chem. Soc. 2008, 130 (17), 5618-9. 42. Wang, T.; Park, Y. B.; Cosgrove, D. J.; Hong, M., Cellulose-Pectin Spatial Contacts Are Inherent to Never-Dried Arabidopsis Primary Cell Walls: Evidence from Solid-State Nuclear Magnetic Resonance. Plant Physiol. 2015, 168 (3), 871-84. 43. Gan, Z. H.; Hung, I.; Wang, X. L.; Paulino, J.; Wu, G.; Litvak, I. M.; Gor'kov, P. L.; Brey, W. W.; Lendi, P.; Schiano, J. L.; Bird, M. D.; Dixon, L. R.; Toth, J.; Boebinger, G. S.; Cross, T. A., NMR spectroscopy up to 35.2 T using a series-connected hybrid magnet. J Magn Reson 2017, 284, 125-136. 44. Rossini, A. J.; Zagdoun, A.; Lelli, M.; Lesage, A.; Coperet, C.; Emsley, L., Dynamic Nuclear Polarization Surface Enhanced NMR Spectroscopy. Acc. Chem. Res. 2013, 46 (9), 1942-1951. 45. Takahashi, H.; Lee, D.; Dubois, L.; Bardet, M.; Hediger, S.; De Paepe, G., Rapid Natural- Abundance 2D 13C-13C Correlation Spectroscopy Using Dynamic Nuclear Polarization Enhanced Solid-State NMR and Matrix-Free Sample Preparation. Angew Chem Int Edit 2012, 51 (47), 11766-11769. 46. Smith, A. N.; Long, J. R., Dynamic Nuclear Polarization as an Enabling Technology for Solid State Nuclear Magnetic Resonance Spectroscopy. Anal Chem 2016, 88 (1), 122-132. 47. Sergeyev, I. V.; Itin, B.; Rogawski, R.; Day, L. A.; McDermott, A. E., Efficient assignment and NMR analysis of an intact virus using sequential side-chain correlations and DNP sensitization. P Natl Acad Sci USA 2017, 114 (20), 5171-5176. 48. Saliba, E. P.; Sesti, E. L.; Scott, F. J.; Albert, B. J.; Choi, E. J.; Alaniva, N.; Gao, C. K.; Barnes, A. B., Electron Decoupling with Dynamic Nuclear Polarization in Rotating Solids. J Am Chem Soc 2017, 139 (18), 6310-6313. 49. Andreas, L. B.; Le Marchand, T.; Jaudzems, K.; Pintacuda, G., High-resolution proton- detected NMR of proteins at very fast MAS. J Magn Reson 2015, 253, 36-49. 50. Struppe, J.; Quinn, C. M.; Lu, M. M.; Wang, M. Z.; Hou, G. J.; Lu, X. Y.; Kraus, J.; Andreas, L. B.; Stanek, J.; Lalli, D.; Lesage, A.; Pintacuda, G.; Maas, W.; Gronenborn, A. M.; Polenova, T., Expanding the horizons for structural analysis of fully protonated protein assemblies by NMR spectroscopy at MAS frequencies above 100 kHz. Solid State Nucl Mag 2017, 87, 117- 125. 51. Phyo, P.; Hong, M., Fast MAS 1H-13C correlation NMR for structural investigations of plant cell walls. J Biomol Nmr 2019. 52. Jaroniec, C. P., Structural studies of proteins by paramagnetic solid-state NMR spectroscopy. 385 J Magn Reson 2015, 253, 50-59. 53. Schmidt-Rohr, K.; Fritzsching, K. J.; Liao, S. Y.; Hong, M., Spectral editing of two- dimensional magic-angle-spinning solid-state NMR spectra for protein resonance assignment and structure determination. J Biomol Nmr 2012, 54 (4), 343-353. 54. Wang, T.; Chen, Y.; Tabuchi, A.; Cosgrove, D. J.; Hong, M., The Target of b-Expansin EXPB1 in Maize Cell Walls from Binding and Solid-State NMR Studies. Plant physiology 2016, 172 (4), 2107-2119. 55. Greenspan, J.; Bulger, B., MySQL/PHP database applications. John Wiley & Sons, Inc.: 2001. 56. Stauffer, M., Laravel: up and running: a framework for building modern PHP apps. " O'Reilly Media, Inc.": 2016. 57. Wang, T.; Salazar, A.; Zabotina, O. A.; Hong, M., Structure and Dynamics of Brachypodium Primary Cell Wall Polysaccharides from Two-Dimensional C-13 Solid-State Nuclear Magnetic Resonance Spectroscopy. Biochemistry-Us 2014, 53 (17), 2840-2854. 58. Thieker, D. F.; Hadden, J. A.; Schulten, K.; Woods, R. J., 3D implementation of the symbol nomenclature for graphical representation of glycans. Glycobiology 2016, 26 (8), 786-7. 59. Neelamegham, S.; Aoki-Kinoshita, K.; Bolton, E.; Frank, M.; Lisacek, F.; Lutteke, T.; O'Boyle, N.; Packer, N. H.; Stanley, P.; Toukach, P.; Varki, A.; Woods, R. J.; Group, S. D., Updates to the Symbol Nomenclature for Glycans guidelines. Glycobiology 2019, 29 (9), 620-624. 60. McNaught, A. D., Nomenclature of carbohydrates (IUPAC Recommendations 1996). Pure Appl. Chem. 1996, 68 (10), 1919-2008. 61. Toukach, P. V.; Egorova, K. S., Carbohydrate structure database merged from bacterial, archaeal, plant and fungal parts. Nucleic Acids Res. 2015, 44 (D1), D1229-D1236. 62. Böhm, M.; Bohne-Lang, A.; Frank, M.; Loss, A.; Rojas-Macias, M. A.; Lütteke, T., Glycosciences. DB: an annotated data collection linking glycomics and proteomics data (2018 update). Nucleic Acids Res. 2018, 47 (D1), D1195-D1201. 63. Kirui, A.; Ling, Z.; Kang, X.; Dickwella Widanage, M. C.; Mentink-Vigier, F.; French, A. D.; Wang, T., Atomic resolution of cotton cellulose structure enabled by dynamic nuclear polarization solid-state NMR. Cellulose 2018. 386 APPENDIX Figure S10.1 Organizational schema of CCMRD. This figure shows the primary organization and how the different entries (NMR signals, compounds, publications, species, etc.) are connected. 387 Figure S10.2 Data Search by compound name and NMR signals. Part or the full name of the polysaccharides is needed for searching by the compound name. In the example, 112 entries have been listed for searching cellulose. Two opinions are available for data searching by the NMR signals: a single NMR chemical shift or two chemical shifts in the same compound. The threshold represents the tolerance allowed for the search (e.g. 105 ± 1 ppm for the example shown in the figure). Both the search functions will finally lead to the details of each entry, with information on the compound trivial name, linear code, origin, species, NMR chemical shifts and experimental conditions. 388 Figure S10.3 User deposition for CCMRD. The flow chart, interface, and an example of step-by- step deposition are given. The user deposition requires approval by the administrator before the record is indexed by the database. 389 CHAPTER 11: CHARTING THE SOLID-STATE NMR SIGNALS OF POLYSACCHARIDES: A DATABASE-DRIVEN ROADMAP Research paper reprinted with permission: Magn. Reson. Chem., in press (2023) Authors: Wancheng Zhao, Debkumar Debnath, Isha Gautam, Liyanage D. Fernando, Tuo Wang 11.1 Abstract Solid-state NMR (ssNMR) measurements of intact cell walls and cellular samples often generate spectra that are difficult to interpret due to the presence of many coexisting glycans and the structural polymorphism observed in native conditions. To overcome this analytical challenge, we present a statistical approach for analyzing carbohydrate signals using high-resolution ssNMR data indexed in a carbohydrate database. We generate simulated spectra to demonstrate the chemical shift dispersion and compare this with experimental data to facilitate the identification of important fungal and plant polysaccharides, such as chitin and glucans in fungi and cellulose, hemicellulose, and pectic polymers in plants. We also demonstrate that chemically distinct carbohydrates from different organisms may produce almost identical signals, highlighting the need for high-resolution spectra and validation of resonance assignments. Our study provides a means to differentiate the characteristic signals of major carbohydrates and allows us to summarize currently undetected polysaccharides in plants and fungi, which may inspire future investigations. 11.2 Introduction Polysaccharides are high molecular weight biopolymers that typically consist of hundreds to thousands of sugar units. Elucidating the structure of complex carbohydrates with solid-state NMR (ssNMR) spectroscopy is a challenging task. Early pioneers in this field explored the use of different ssNMR methods, such as one-dimensional (1D) experiments for identifying the structure and quantifying the content of carbohydrate polymers and allomorphs, and Rotational Echo 390 Double Resonance (REDOR) and multidimensional correlation techniques for revealing polymer cross-linking and spatial proximities1-6. There has been a substantial increase in high-resolution ssNMR data that rely on 13C and 15N 2D/3D correlation experiments7-9, as well as 1H detection10- 12, to provide detailed insights into the structure and function of carbohydrates in a variety of living organisms. Recent high-resolution ssNMR studies have led to numerous conceptual breakthroughs, including but not limited to the identification of cellulose-pectin spatial contacts in the primary plant cell walls (mostly the seedlings)13-16, the conformation-dependent function of xylan in associating with lignin and cellulose in the secondary cell walls of grasses and woody stems17-22, the multi-layered assembly of chitin and glucans in fungal cell walls that provide remarkable structural dynamics to support microbial survival23-28, as well as the quantification of dynamically heterogeneous carbohydrate components in algal cells and bacterial cell wall and biofilms29-33. There is also a growing interest in utilizing ssNMR techniques for the characterization of carbohydrate-based materials, including hydrogels34, 35. By taking advantage of the repetitive nature of monosaccharides in a polysaccharide, the structural view of these polymers can be simplified, and their spectral features can be explored. Currently, there are five distinct structural variations of carbohydrates that can be closely tracked in the solid-state, which include the variation in covalent linkages, the compositional differences in cell walls and monosaccharide units that form polysaccharides, the conformational distribution of native polysaccharides, the spatial assembly, and the chemical modifications such as methylation and methyl esterification, as documented in a recent review aricle36. The majority of existing ssNMR studies depict polysaccharides in a simplified manner, reducing their complex structures to symbolic representations and only highlighting the most significant constituents 391 showing strong signals7. While this approach enables researchers to concentrate on the primary polysaccharides found in a sample, it overlooks the many sparsely distributed molecules, minor side chains, and irregular substitutions, which may also contribute significantly to the structural stability and functional properties of both the polysaccharide and the cell wall. In 2020, a complex carbohydrate magnetic resonance database (CCMRD) was established to accommodate the prolific growth of datasets in the upcoming years37. Upon its inception, the small-scale database contained 420 entries of high-resolution ssNMR data on carbohydrates. As of early 2023, the number of entries has increased to approximately 850. The database facilitates data search using single or paired chemical shifts, polysaccharide names, and compound classes, and it also supports user deposition. This platform also serves as a vital preparatory step for implementing semi-automated tools to streamline spectral analysis and resonance assignment of carbohydrate systems in the future as the size of the dataset increases. The present study aims to accomplish three objectives by utilizing the available data. Firstly, we aim to construct artificial NMR spectra of commonly employed experiments to summarize the polysaccharide signals present in plants and fungi, which is akin to the amino acid maps available for protein NMR38. Secondly, we will utilize these spectroscopic maps to identify the unique signals that can unequivocally recognize specific polysaccharides. Finally, we will provide a synopsis of the key factors currently constraining our ability and give an overview of the undetected carbohydrates that necessitate further exploration. The pursuit of these objectives intends to encourage the use of high-resolution ssNMR in carbohydrate characterization. 392 11.3 Materials and Methods 11.3.1 Generation of Simulated 2D Spectra from Chemical Shift Data We used a dataset containing 569 entries of plant and fungal carbohydrates sourced from CCMRD37 to construct artificial 2D NMR spectra. Each entry within the dataset comprised 13C chemical shift data for a carbohydrate unit sourced from a peer-reviewed publication. These publications featured high-resolution solid-state NMR data, typically presented in the form of 2D or 3D correlation spectra. The raw 13C chemical shift data were transformed into different scatter plots depending on the targeted type of 2D spectra following the protocol described below. For the refocused INADEQUATE spectrum that exhibits through-bond correlations4, the signals of two directly bonded carbons were situated at coordinates (δCa, δCa+δCb) and (δCb, δCa+δCb) on the plot. Here, δCa and δCb denote the single-quantum (SQ) chemical shifts of carbon sites a and b, respectively, and these values were projected to the X-axis of the simulated plot, corresponding to the direct dimension (ω2) of experimentally measured spectra. The sum of two SQ chemical shifts (δCa+δCb) were projected to the Y-axis, which represents the value of the double-quantum (DQ) chemical shift of this pair of carbons, corresponding to the indirect dimension (ω1) of experimentally measured spectra. Each pyranose unit is represented by 10 scatter dots based on the chemical shifts of six carbons, and an illustrative example is presented in Figure 11.1a, showcasing the distribution of these scatter dots for a glucose unit in cellulose. In comparison, each furanose unit will result in 8 scatter dots, due to the lack of C5-C6 spin pair. More scatter dots are needed for plotting 2D 13C SQ-SQ correlation spectra, such as PDSD/DARR, CORD39, and PAR40. This is because each carbon atom can potentially correlate with all the other carbons within the same monosaccharide unit. Each scatter dot now represents a distinct intramolecular cross peak, with the two coordinates representing the SQ chemical shifts of 393 two carbons. When examining a pyranose, the anticipated SQ-SQ correlation spectrum consists of 36 derived data points, as illustrated in Figure 11.1b for a glucose unit of cellulose. Among these points, 6 are positioned along the spectral diagonal. In contrast, a furanose's spectrum contains 25 data points, with 5 residing on the diagonal. The chemical shift coordinates utilized in generating the scatter plots shown in Figure 11.1 are provided in Table S11.1. Our analysis extended to encompass 412 entries from plant cell walls and 157 datasets from fungal cell walls obtained from CCMRD. Scatter plots were generated to depict both DQ-SQ and SQ-SQ correlation spectra for various sugar units. Subsequently, density maps were generated from these scatter plots using Kernel Density Estimate in Origin 2019b. These density maps provide an informative overview of the average trends represented by the scatter plots, allowing us to identify the specific regions and patterns related to different sugar units within 2D NMR spectra. Figures S11.1-S11.26 in the Supplementary Information present individual plots for various plant and fungal carbohydrates. These figures can be used for direct comparison with experimental data, streamlining the analysis process. 11.3.2 Sample Preparation and Packing Three uniformly 13C-enriched samples were prepared for ssNMR measurement. These samples include the fungal culture of Aspergillus fumigatus, and the stems of two plant species: Arabidopsis thaliana and Picea abies (Norway spruce)21. The fungal mycelium was cultivated for two weeks on a liquid minimal medium containing 13C-glucose as the only carbon source26. The spruce plant was grown within a closed atmosphere featuring 97 atom% 13CO2. This controlled environment was maintained from germination until harvest, spanning approximately four months. Similarly, 13C-enriched Arabidopsis was grown in a custom-built transparent chamber for 4 weeks. The harvested fresh plant stems were cut into millimeter-scale pieces by razor blade for better 394 distribution of mass during MAS. Figure 11.1 Scatter plots of chemical shift values for a glucose unit of cellulose. The plots correspond to a, DQ-SQ and b, SQ-SQ 13C correlation experiments, with the former showing through-bond correlations and the latter showing all intramolecular through-space correlations. See Table S11.1 for comprehensive details on the dataset used to create these spectral projections. Isotope-labeling is essential for acquiring high-resolution 2D/3D correlation spectra, particularly in dealing with the significant signal overlap encountered in cellular samples or complex biomaterials. Natural-abundance 2D 13C-13C correlation experiments of unlabeled samples becomes attainable with the utilization of the sensitivity-enhancing technique MAS-DNP, as recently demonstrated on lignocellulosic plant biomass41-43, fungal mycelia and conidia44, and functionalized cellulose materials45-48. However, it's important to note that at the low temperatures required for MAS-DNP, there is a significant sacrifice in spectral resolution. Therefore, this technique is primarily employed when isotope-labeling presents challenges. 395 Around 30 mg of A. fumigatus was packed into a 3.2 mm MAS rotor and measured on an 800 MHz Bruker Avance Neo Spectrometer at National High Magnetic Field Laboratory using a 3.2 mm HCN MAS probe. Similarly, around 35 mg of Arabidopsis sample was packed into a 3.2 mm MAS rotor and measured on a 700 MHz Bruker Avance Neo Spectrometer at Louisiana State University using a 3.2 mm HCN MAS probe. For spruce, approximately 100 mg of material was packed into a 4 mm MAS rotor for measurement on a 400 MHz Bruker Avance spectrometer at Louisiana State University using a 4.0 mm HXY MAS probe. 11.3.3 Solid-State NMR Experiments All experiments were performed with spinning speeds ranging from 10-13.5 kHz. The radiofrequency field strengths were set to 83.3 kHz for both 1H excitation and decoupling, and 62.5 kHz for 13C hard pulses. The recycle delays were 1.6-2.0s. 2D experiments were recorded over durations of 4 to 30 h, while 1D data were collected in timeframes varying from 5 to 140 min. The 2D refocused J-INADEQUATE scheme was integrated with either direct polarization (DP) and short recycle delays of 1.8 s for the selective detection of mobile components, or with cross- polarization (CP) for the detection of rigid components. A 2D 13C-13C correlation experiment was also measured using the CORD scheme with a 53-ms recoupling period. Within each spectrum, both the sidechains and primary constituents of a polysaccharide, if they exhibit similar dynamics, were detected simultaneously. The acquisition time was 12-29 ms, and the evolution time for the indirect dimension of 2D experiments was 4-10 ms. The key parameters used in these ssNMR experiments are listed in Table S11.2. Accurate experimental chemical shift data are essential for structural analysis49. All 13C NMR spectra are consistently referenced to the tetramethylsilane (TMS) scale by externally calibrating the CH2 signal of adamantane to 38.48 ppm at room temperature, which is a standard 396 calibration procedure widely employed in solid-state NMR studies50, including carbohydrate research. In the chemical shift data deposited in CCMRD, NMR chemical shifts are only reported on the TMS scale, the information of which is included in each entry37. DSS scale is another reference scale that yields a 13C chemical shift difference of 2.0 ppm compared to the TMS scale. Chemical shifts reported on the DSS scale can be adjusted by subtracting 2.0 ppm for the purpose of comparison with data on the TMS scale, which has been reported in recent studies of amyloid fibrils51. However, it finds primary usage in protein solid-state NMR and solution NMR studies. It should be noted that variations in chemical shifts may arise due to different calibration procedures employed by various research groups, as reported in recent cellulose studies49. 11.4 Results and Discussion 11.4.1 Unique Signals of Different Carbohydrates Initial assignments of carbohydrate signals heavily rely on 2D 13C refocused INADEQUATE correlation experiments4 that correlate DQ and SQ chemical shifts to obtain diagonal-free spectra. We extracted 412 plant carbohydrate entries and constructed artificial INADEQUATE spectra (Figure 11.2a). The scatter plots were derived from data deposited in CCMRD, with all through-bond correlations simulated for each carbohydrate unit. For example, for each hexose, the DQ and SQ chemical shifts were calculated for six spin pairs of C1-C2, C2- C3, C3-C4, C4-C5, C5-C6, while the last spin pair is absent in each pentose. Furthermore, we created corresponding color-coded contour lines based on computed Kernel Density Estimate to represent the probability of different sugar residues in certain spectral regions. These maps included sugar residues found in context of a plant polysaccharide, such as the glucose units (Glc) found in cellulose and seven sugar residues commonly found in the matrix polysaccharides (hemicellulose and pectin), namely xylose (Xyl), galacturonic acid (GalA), glucuronic acid 397 (GlcA), arabinose (Ara), galactose (Gal), rhamnose (Rha), and mannose (Man) (Figure 11.2b). Figure 11.2 Density map of plant polysaccharides projected into INADEQUATE spectra. a, 13C chemical shift data of 412 datasets plotted as either scatter plot (left) or density map (middle), or both (right). The data of eight types of monosaccharide units were color coded. b, Representative monosaccharide units with carbon numbers labeled. Two glucose units were shown for cellulose. c, Unique spectral regions of the C5-C6 signals of Rha (top) and acidic residues (bottom), where the methyl and carbonyl groups can help to track the signals of these carbohydrates. In particular, the Ara units are typically resolved by their C1-C2 spin pair, with unique DQ chemical shifts ranging from 188 ppm to 198 ppm. Although the cellulose signals are tightly clustered with peaks from other carbohydrates, their C3-C4 and C4-C5 spin pairs can still be used as the starting point to track their signals. Xyl signals have a broad distribution but can be tracked 398 using the C4-C5 spin pair: α-linked Xyl has a unique DQ chemical shift at 130-135 ppm while β- linked Xyl typically shows up at 140-150 ppm, although the latter depends on the helical screw conformation. Rha can be unambiguously tracked starting from its C6, which is a methyl group, and the corresponding C5-C6 spin pair (Figure 11.2c). Acidic sugars, such as GalA and GlcA, rely on the spin pair between C5 and C6, which is the carbonyl group, for initiating the resonance assignment and tracking the other carbons within the same sugar unit. It should be noted that the chemical shift of the carbonyl groups in acidic sugars can vary significantly depending on the state of protonation, the deprotonation and ion coordination (e.g., with Calcium in HG and with Boron in RG-II)15, 52, as well as the degree of methyl esterification. Peak multiplicity in this spectral region has been consistently observed in many plant species, such as Arabidopsis, Zea mays, and Brachypodium16, 53-55, with 176-178 ppm for carboxylate (-COO- or -COOH), 173-175 ppm for acetyl (-OCOCH3), and 170-172 ppm for methyl ester (-COOCH3). Due to the low bulk pKa of GalA (∼3.5)56, the carboxylate is in the deprotonated form under the near-neutral pH of most plant cell wall samples. However, the carboxylate signal vanishes GalA residues are neutralized during cell wall acidification, which has been observed at pH 4.0 in an Arabidopsis sample57. This further impedes GalA’s capability to coordinate Ca2+ for crosslinking adjacent HG chains. Therefore, the pKa of these sugars and the pH of the samples being investigated should be considered and the carbonyl signals should be carefully tracked. Signals of Gal residues are difficult to separate from the peaks of other sugar residues when using 2D 13C-13C correlation spectra; therefore, it requires special attention in analysis. The SQ chemical shift of carbon 4 site (~78 ppm) and the DQ chemical shifts of C3-C4 and C4-C5 (152- 155 ppm) offer partial resolution in the refocused INADEQUATE spectrum, serving as an initial reference for Gal resonance assignment, especially for those in the mobile phase55. Proton- 399 detection techniques such as 2D TOCSY and 3D HCC INEPT-TOCSY have been employed to distinguish Gal signals from other sugar residues in mobile matrix polysaccharides10. We also presented these datasets as a 2D 13C SQ-SQ correlation spectrum, as shown in Figure 11.3. This spectrum encompasses all intramolecular cross-peaks, resembling the spectral patterns commonly observed with ~100 ms mixing periods under experimental schemes like CORD39, DARR/PDSD, or other analogous approaches conducted at moderately slow magic- angle spinning (MAS) frequencies, e.g., in the 10-20 kHz range. The simulated spectrum is more congested than the INADEQUATE spectra shown earlier due to the presence of numerous multibond correlations. Nevertheless, discernable spectral regions are still available for almost every type of carbohydrate, except for Gal, where complementary techniques like the fast MAS INEPT spectra with 1H detection, as demonstrated in a recent study, may become necessary10. Figure 11.3 Projected NMR signals of plant polysaccharides as the 2D SQ-SQ correlation spectra using the same datasets for plotting Figure 11.2. 400 11.4.2 Reconstituted Spectra for Polysaccharide Identification Monosaccharide units can associate with others to form various polysaccharides. We synthesized artificial spectra to depict the potential range of chemical shift dispersion for polysaccharides (Figure 11.4a). The first spectrum cataloged all the signals identified in plant cellulose, which possesses a highly polymorphic structure characterized by up to seven types of Glc subforms within each plant sample58. The second spectrum encompassing Xyl and Ara signals, contained the spectra from various matrix polysaccharides, including hemicellulose arabinoxylan, pectic sidechain arabinan, as well as the sidechains of xyloglucan. Notably, the α-Xyl units of xyloglucan and β-Xyl residues of xylan were not differentiated in this plot, but the distinction can be easily achieved, if necessary. The third spectrum displayed the signals of acidic residues and Rha, which covered two primary pectic polymers: homogalacturonan (HG) and rhamnogalacturonan-I (RG-I), as well as the GlcA sidechains of the hemicellulose glucuronoxylan. The last spectrum displayed Gal and Man residues, which could be part of mannan polymers like galactoglucomannan (GGM), or the galactan sidechains of RG-I. The corresponding carbohydrate structures are summarized in Figure 11.4b. The comparison of simulated carbohydrate maps with experimentally measured 2D spectra can aid in the initial evaluation of the sample and facilitate the identification of the most probable carbohydrates. As an example, we analyzed the refocused J-INADEQUATE spectrum of 13C- labeled Arabidopsis primary cell walls, which were measured using DP and short recycle delays of 1.8 s to select the mobile molecules with rapid 13C-T1 relaxation (Figure 11.5a). Consequently, only dynamic matrix polysaccharides were detected, including the well-resolved signals of Rha, acidic sugars, and Ara from pectin, as well as the signature C4-C5 signals of α-linked Xyl units from xyloglucan. 401 Figure 11.4 Simulated NMR spectra of individual carbohydrate units. a, Chemical shift dispersion and density maps of cellulose glucose units, Xyl, Ara, GalA and GlcA, Rha, Gal and Man. b, Simplified structure of polysaccharides containing these residues widely found in plant primary and secondary cell walls. Ara units are distinguishable by their distinct signals falling within the DQ 13C chemical shift range of 188-200 ppm. Ara peaks are also evident in all other discernible spectral regions indicated by the yellow regions in the simulated density maps, which include the DQ chemical shift range of 160-170 ppm for C2-C3 and C3-C4, as well as the DQ chemical shift range of 145- 155 ppm for C4-C5. These observations confirmed the presence of Ara residues in Arabidopsis. Similarly, the presence of Rha is verified by its characteristic C5-C6 signals, where C6 represents a methyl carbon, situated within the DQ chemical shift range of 80-90 ppm. The existence of acidic 402 residues (GalA/GlcA) is substantiated by their unique signals with DQ chemical shifts extending beyond 240 ppm, corresponding to the covalently linked C5 and carbonyl group. Cellulose is not detected in this spectrum, as evidenced by the absence of peaks in the magenta regions, where DQ chemical shifts typically fall within the range of 160-170 ppm, a characteristic range for C3-C4 and C4-C5 carbon pairs in cellulose. This outcome aligns with expectations since cellulose is a rigid component and is not found in this spectrum, which is designed to detect mobile components. Intensities are observed in other magenta regions, but these sharp signals originate from mobile polysaccharides exhibiting similar chemical shifts at these indistinguishable carbon positions. The rigid portion of the spruce secondary cell wall was examined using 13C CP-based INADEQUATE and CORD experiments, where the spectra were dominated by cellulose signals (Figure 11.5a, b). Weak peaks were also spotted for the β-linked Xyl from xylan backbones and the Ara residues from xylan sidechains, as well as some Man signals from GGM. The α-Xyl residues, which are indicative of xyloglucans, are not observed in this spectrum as shown by the absence of C4-C5 signals within the DQ chemical shift range of 130-136 ppm. This outcome is anticipated, considering that the spruce sample under examination is primarily composed of secondary cell wall component and the current spectrum selectively targets rigid polysaccharides, whereas xyloglucan, being partially mobile, exists in primary cell walls. The use of carbohydrate fingerprints and the combination of selected datasets from certain glycans or organisms turned out to be an efficient approach for identifying polysaccharides, which should be applicable to both purified biopolymers and cellular samples. We have included a collection of individual scatter and density plots as Figures S11.1-S11.26 for the purpose of enabling direct comparisons with experimentally acquired spectra. This qualitative analysis will 403 aid in identifying potential carbohydrates in the plant sample of interest as the initial step, which will require further validation or complete resonance assignment as subsequent procedures. As the Arabidopsis and spruce spectra were collected on 400-700 MHz NMR instruments, it is evident that a moderate-to-high magnetic field is adequate for the implementation of this method. A current limitation is the lack of quantitative methods to assess the probability of specific resonance assignments, a capability akin to what is available for proteins and small molecules59, 60. Further development of new software and algorithms61, 62 will be essential. Figure 11.5 Overlay of carbohydrate density map and experimental spectra for a, refocused J- INADEQUATE and b, 53 ms CORD spectra. The DP-based J-INADEQUATE spectrum measured on Arabidopsis primary cell walls, and the CP-based J-INADEQUATE and CORD experiments were conducted on spruce stems that are rich in secondary cell walls. Arabidopsis and spruce samples were measured on 700 MHz and 400 MHz NMR spectrometers, respectively. 11.4.3 Plant Carbohydrates Evading Detection Many polysaccharides remain elusive to solid-state NMR detection, even for the polysaccharides in plants, which have been relatively well studied. Although xyloglucan (XyG) is the primary hemicellulose in the primary cell wall, most studies have to rely on its α-Xyl sidechains 404 to track XyG signals and examine interactions with other cell wall components13, 63. Its β-1,4- glucan backbone, as well as the occasional substitutions of β-galactose (Gal) and α-fucose (Fuc)64, have not been unambiguously identified. The signals corresponding to the β-1,4-glucan backbones are predominantly shielded by the surface glucan chains of cellulose, which are chemically identical, while the Gal and Fuc residues are relatively scarce. In addition, the third abundant pectic polymer, RG-II, has not been detected in intact cell wall samples either. RG-II contains at least 12 sugar types with complex linkages and has a relatively lower content compared with HG and RG- I, making it very difficult to characterize65, 66. Moreover, chemical data suggest that pectic polymers exist in integrated networks containing polysaccharides and proteoglycans and might be covalently interconnected and further crosslinked to other cell wall macromolecules such as hemicellulose and protein components66-68. At this moment, such structural features have not been recognized using intact cell walls. The integration of solid-state NMR analysis with functional genomics, incorporating engineered mutants that selectively remove individual carbohydrate components, holds promise for the unambiguous identification of previously unresolved polysaccharides. This approach has recently proven successful in fungal species, particularly in resolving the ambiguity associated with the identification of galactosaminogalactan (GAG) and galactomannan (GM) within the cell walls of Aspergillus species25. Augmenting this approach with MAS-DNP can further mitigate detection issues associated with the low levels of certain carbohydrate components, such as RG-II and the Gal and Fuc sidechains of xyloglucan. Improvement in 2D/3D correlation experiments to achieve better signal dispersion69, 70, coupled with the assessment of 1H chemical shifts using ultrafast MAS techniques10, 71, are poised to surmount the existing resolution limitations. 405 11.4.4 Limited Dataset of Fungal Carbohydrates Fungal glycans have received inadequate attention, and most of the available data has been collected in the past five years, as summarized in Figure 11.6. We can unambiguously identify the signals from several amino sugars, such as the N-acetylglucosamine (GlcNAc) from chitin and the glucosamine (GlcN) from chitosan. Both chitin and chitosan have exhibited broad distribution owing to the high level of structural polymorphism in their crystallites72, 73. However, the relationship between the polymorphism of chitin and chitosan and the diversity of chitin synthase gene families remains unclear74. β-glucans are prominent molecules widely distributed in different fungi, and have a high level of linkage complexity and can exist as different structures, such as linear β-1,3-glucan, branched β-1,3/1,6-glucan, linear β-1,3/1,4-glucan, and linear β-1,6-glucan, depending on the location within the cell wall and the fungal species being studied75, 76. Different linkages can exist in the same fungi with variable percentages. For example, the β-glucans of yeast consist (~85%) of branched β-1,3-glucan, which contains ~3% β-1,3,6 interchain branching sites, with a smaller amount of linear β-1,6-glucans (~15%)76. Meanwhile, the filamentous mold Aspergillus fumigatus also have predominantly branched β-1,3-glucan accounts with ~4% β-1,6 branching sites, but the minor component has changed to β-1,3/1,4-glucan, which constitutes approximately 10% of the cell wall β-glucans in this species75. Solid-state NMR studies mainly focus on the function of β-1,3-glucan and ongoing research aims to identify the signals of each linkage type and relate them to their structural functions in the cell wall. Recent solid-state NMR studies have mainly focused on the function of β-1,3-glucan and ongoing research aims to identify the signals of each linkage type and relate them to their structural functions in the cell wall. Recent studies have revealed the roles of α-1,3-glucan in stabilizing the cell wall assembly, but information regarding the α-1,4-linked glucose residue is lacking77. Other spotted 406 monosaccharide units, such as galactofuranose (Galf), galactosamine (GalN), N- acetylgalactosamine (GalNAc), Man, and Fuc, belong to galactosaminogalactan (GAG), galactomannan (GM), phosphomanan, mannoproteins, and other minor cell wall polysaccharides, which can further complicated spectral analysis. The contribution of glycoproteins and lipid components to the cell wall architecture requires further investigation78. Figure 11.6. Summary of 13C chemical shifts from fungal carbohydrates. The simulated spectra were presented as a, DQ-DQ and b, SQ-SQ correlation spectra. The current plot contains data from 157 entries, which is relatively limited and cannot efficiently represent the complex nature of fungal polysaccharides. The name of each polysaccharide is shown in bold followed by the monosaccharide units including GlcNAc, GlcN, Glc, Man, Galf, GalN, Galp, and GalNAc. Compared to their plant counterparts, fungal polysaccharides exhibit much higher structural diversity, and extensive solid-state NMR studies are necessary to document the carbohydrate signals across a wide range of fungal species to facilitate antifungal development and the biotechnology applications of these microorganisms. These endeavors may encompass the 407 utilization of multidimensional correlation experiments involving 13C, 15N, and 1H nuclear spins to elucidate the carbohydrate signals in significant human pathogenic species, such as Aspergillus, Candida, and Cryptococcus species, as well as non-pathogenic species that could potentially be used for nutritional resources and biomaterials28. 11.4.5 Distinct Carbohydrates Can Have Identical Signals Caution should be taken when utilizing NMR maps to investigate novel species with unknown composition. In certain instances, two distinct carbohydrates may demonstrate strictly overlapping signals, for example, between arabinofuranose (Araf) and Galf25. While Araf (typically called Ara in our studies) is a common sugar forming the side chains of pectin and hemicellulose in plants, Galf is a rare form of galactose and is sporadically identified in some fungal species, for example, as a constituent of the GM present in the cell wall79-81. Although Galf is a hexose and Ara is a pentose, both sugar units possess five-membered rings, resulting in identical 13C chemical shifts in their first four carbons, with representative chemical shifts of approximately 108 ppm, 82 ppm, 77 ppm, and 84 ppm, respectively. The only unique signals of Galf in a 2D INADEQUATE spectrum come from its C5-C6 carbon pair, however, it showed up in a heavily crowded spectral region and cannot be utilized as a reliable tracker of molecules (Figure 11.7a). In a standard 1D spectrum, the sole distinctive signal for Galf, when compared to Araf, is the C5 signal at 69 ppm (Figure 11.7b). However, this peak falls within a densely populated region, making it impractical for distinguishing signals from other carbohydrates. The occurrence of such indistinguishable spectral patterns between two chemically and structurally distinct monosaccharide units was unanticipated as we extend our reach of carbohydrate ssNMR and expand our knowledge and understanding of these complex biopolymers. Hence, it is essential to carefully consider the source of the samples being analyzed, be it from plants, fungi, or other 408 organisms. In the case of a novel species, it becomes imperative to complement NMR analysis with the chemical assays that provide information on the linkage patterns and chemical composition of carbohydrate constituents. Moreover, the reliance on 1D spectra for analyzing previously uncharted samples should be minimized. Figure 11.7 Comparison of the NMR signals of Galf and Ara. a, The C1-C2-C3-C4-C5 signals of Galf in the 2D DP refocused J-INADEQUATE spectrum of A. fumigatus falls within the expected signals for arabinose residues from plant cell walls. Only the C5-C6 carbon pair of Galf is unique. b, Comparison of 1D 13C quantitative DP spectrum measured on Arabidopsis (top) and A. fumigatus (bottom). Orange and green lines indicate the signals from Ara (only in Arabidopsis) and Galf residues (only in A. fumigatus), respectively. 11.5 Conclusions and Perspectives Significant advances have been made in solid-state NMR analysis of carbohydrates over the past decade, allowing us to identify key spectroscopic features of major plant and fungal glycans. It is critical to ensure that there is sufficient spectral resolution to track the minor 409 differences in molecules, carefully track the full carbon connectivity in 2D/3D correlation spectra without missing any carbon in the linkage pattern, seek additional validation from other techniques, such as mass spectrometry, and cross-compare with mutants or other strains lacking the carbohydrate of interest. Further datasets are required before statistical analysis and automated tools can be developed to facilitate the analysis of these complex carbohydrate polymers. It is anticipated that tools similar to those already available for other molecules, such as the DEEP picker for protein 2D spectral deconvolution and peak picking, and the probabilistic assignment tool for organic crystals59, 82, will lead to significant progress in solid-state glycoNMR. 11.6 Acknowledgments This work was supported by the National Science Foundation (NSF) MCB-2308660. Part of the NMR data was collected on a 700 MHz NMR spectrometer funded by NSF Major Research Instrumentation Program under the award number DBI-2019046. A portion of this work was performed at the National High Magnetic Field Laboratory, which is supported by National Science Foundation Cooperative Agreement No. DMR-2128556 and the State of Florida. 410 REFERENCES 1. Fukamizo, T.; Kramer, K. J.; Mueller, D. D.; Schaefer, J.; Carbow, J.; Jacob, G. S., Analysis of Chitin Structure by Nuclear Magnetic Resonance Spectroscopy and Chitinolytic Enzyme Digestion. Arch. Biochem. Biophys. 1986, 249, 15-26. 2. Schaefer, J.; Stejskal, E. O.; Beard, C. F., Carbon-13 Nuclear Magnetic Resonance Analysis of Metabolism in Soybeans Labeled by 13CO2. Plant Physiol. 1975, 55, 1048-1053. 3. Bardet, M.; Emsley, L.; Vincendon, M., Two-dimensional spin-exchange solid-state NMR studies of 13 C-enriched wood Solid State Nucl. Magn. Reson. 1997, 8, 25-32. 4. Lesage, A.; Bardet, M.; Emsley, L., Through-bond carbon-carbon connectivities in disordered solids by NMR. J. Am. Chem. Soc. 1999, 188, 10987-10993. 5. Tong, G.; Pan, Y.; Dong, H.; Pryor, R.; Wilson, G. E.; Schaefer, J., Structure and Dynamics of Pentaglycyl Bridges in the Cell Walls of Staphylococcus aureus by 13C-15N REDOR NMR. Biochemistry 1997, 36, 9859-9866. 6. Atalla, R. H.; Vanderhart, D. L., Native cellulose: a composite of two distinct crystalline forms. Science 1984, 223, 283-285. 7. Ghassemi, N.; Poulhazan, A.; Deligey, F.; Mentink-Vigier, F.; Marcotte, I.; Wang, T., Solid- State NMR Investigations of Extracellular Matrices and Cell Walls of Algae, Bacteria, Fungi, and Plants. Chem. Rev. 2022, 122, 10036–10086. 8. Reif, B.; Ashbrook, S. E.; Emsley, L.; Hong, M., Solid-state NMR spectroscopy. Nat. Rev. Methods Primers 2021, 1 , 2. 9. Zhao, W.; Fernando, L. D.; Kirui, A.; Deligey, F.; Wang, T., Solid-state NMR of plant and fungal cell walls: a critical review. Solid State Nucl. Magn. Reson. 2020, 107, 101660. 10. Phyo, P.; Hong, M., Fast MAS 1H-13C correlation NMR for structural investigations of plant cell walls. J. Biomol. NMR 2019, 73, 661-674. 11. Safeer, A.; Kleiburg, F.; Bahri, S.; Beriashvili, D.; Veldhuizen, E. J. A.; van Neer, J.; Tegelaar, M.; de Cock, H.; Wosten, H. A. B.; Baldus, M., Probing Cell-Surface Interactions in Fungal Cell Walls by High-Resolution 1H-Detected Solid-State NMR Spectroscopy. Chem. Euro. J. 2022, 29, e202202616. 12. Bougault, C.; Ayala, I.; Vollmer, W.; Simorre, J. P.; Schanda, P., Studying intact bacterial peptidoglycan by proton-detected NMR spectroscopy at 100 kHz MAS frequency. J. Struct. Biol. 2019, 206, 66-72. 13. Dick-Perez, M.; Zhang, Y.; Hayes, J.; Salazar, A.; Zabotina, O. A.; Hong, M., Structure and Interactions of Plant Cell-Wall Polysaccharides by Two- and Three-Dimensional Magic- Angle-Spinning Solid-State NMR. Biochemistry 2011, 50, 989-1000. 411 14. White, P. B.; Wang, T.; Park, Y. B.; Cosgrove, D. J.; Hong, M., Water–Polysaccharide Interactions in the Primary Cell Wall of Arabidopsis thaliana from Polarization Transfer Solid- State NMR. J. Am. Chem. Soc. 2014, 136, 10399-10409. 15. Temple, H.; Phyo, P.; Yang, W.; Lyczakowski, J. J.; Echevarria-Poza, A.; Yakunin, I.; Parra-Rojas, J. P.; Terrett, O. M.; Saez-Aguayo, S.; Dupree, R.; Orellana, A.; Hong, M.; Dupree, P., Golgi-localized putative S-adenosyl methionine transporters required for plant cell wall polysaccharide methylation Nat. Plants 2022, 8, 656-669. 16. Wang, T.; Park, Y. B.; Cosgrove, D. J.; Hong, M., Cellulose-pectin spatial contacts are inherent to never-dried Arabidopsis primary cell walls: evidence from solid-state nuclear magnetic resonance. Plant Physiol. 2015, 168, 871-884. 17. Terrett, O. M.; Lyczakowski, J. J.; Yu, L.; Iuga, D.; Franks, W. T.; Brown, S. P.; Dupree, R.; Dupree, P., Molecular architecture of softwood revealed by solid-state NMR. Nat. Commun. 2019, 10, 4978. 18. Simmons, T. J.; Mortimer, J. C.; Bernardinelli, O. D.; Pöppler, A.-C.; Brown, S. P.; Deazevedo, E. R.; Dupree, R.; Dupree, P., Folding of xylan onto cellulose fibrils in plant cell walls revealed by solid-state NMR. Nat. Commun. 2016, 7, 13902. 19. Dupree, R.; Simmons, T. J.; Mortimer, J. C.; Patel, D.; Iuga, D.; Brown, S. P.; Dupree, P., Probing the Molecular Architecture of Arabidopsis thaliana Secondary Cell Walls Using Two- and Three-Dimensional 13C Solid State Nuclear Magnetic Resonance Spectroscopy. Biochemistry 2015, 54, 2335-2345. 20. Kang, X.; Kirui, A.; Widanage, M. C. D.; Mentink-Vigier, F.; Cosgrove, D. J.; Wang, T., Lignin-polysaccharide interactions in plant secondary cell walls revealed by solid-state NMR. Nat. Commun. 2019, 10, 347. 21. Kirui, A.; Zhao, W.; Deligey, F.; Yang, H.; Kang, X.; Mentink-Vigier, F.; Wang, T., Carbohydrate-aromatic interface and molecular architecture of lignocellulose. Nat. Commun. 2022, 13, 538. 22. Gao, Y.; Lipton, A. S.; Wittmer, Y.; Murray, D. T.; Mortimer, J. C., A grass-specific cellulose–xylan interaction dominates in sorghum secondary cell walls. Nat. Commun. 2020, 11, 6081. 23. Kelly, J. E.; Chrissian, C.; Stark, R. E., Tailoring NMR experiments for structural characterization of amorphous biological solids: A practical guide. Solid State Nucl. Magn. Reson. 2020, 109, 101686. 24. Ehren, H. L.; Appels, F. V. W.; Houben, K.; Renault, M. A. M.; Wösten, H. A. B.; Baldus, M., Characterization of the cell wall of a mushroom forming fungus at atomic resolution using solid-state NMR spectroscopy. Cell Surf. 2020, 6, 100046. 25. Chakraborty, A.; Fernando, L. D.; Fang, W.; Dickwella Widanage, M. C.; Wei, P.; Jin, C.; Fontaine, T.; Latgé, J. P.; Wang, T., A molecular vision of fungal cell wall organization by 412 functional genomics and solid-state NMR. Nat. Commun. 2021, 12, 6346. 26. Kang, X.; Kirui, A.; Muszynski, A.; Widanage, M. C. D.; Chen, A.; Azadi, P.; Wang, P.; Mentink-Vigier, F.; Wang, T., Molecular architecture of fungal cell walls revealed by solid- state NMR. Nat. Commun. 2018, 9, 2747. 27. Chrissian, C.; Lin, C. P.-C.; Camacho, E.; Casadevall, A.; Neiman, A. M.; Stark, R. E., Unconventional Constituents and Shared Molecular Architecture of the Melanized Cell Wall of C. neoformans and Spore Wall of S. cerevisiae. J. Fungi 2020, 6, 329. 28. Pylkkanen, R.; Werner, D.; Bishoyi, A.; Weil, D.; Scoppola, E.; Wagermaier, W.; Safeer, A.; Bahri, S.; Baldus, M.; Paananen, A.; Penttila, M.; Szilvay, G.; Mohammadi, P., The complex structure of Fomes fomentarius represents an architectural design for high- performance ultralightweight materials. Sci. Adv. 2023, 8, ade5417. 29. Nygaard, R.; Romaniuk, J. A. H.; Rice, D. M.; Cegelski, L., Spectral snapshots of bacterial cell-wall composition and the influence of antibiotics by whole-cell NMR. Biophys. J. 2015, 108, 1380-1389. 30. Arnold, A. A.; Bourgouin, J. P.; Genard, B.; Warschawski, D. E.; Tremblay, R.; Marcotte, I., Whole cell solid-state NMR study of Chlamydomonas reinhardtii microalgae. J. Biomol. NMR 2018, 70, 123-131. 31. Romaniuk, J. A. H.; Cegelski, L., Bacterial cell wall composition and the influence of antibiotics by cell-wall and whole-cell NMR. Philos. Trans. R. Soc. B 2015, 370, 20150024. 32. Poulhazan, A.; Dickwella Widanage, M. C.; Muszynski, A.; Arnold, A. A.; Warschawski, D. E.; Azadi, P.; Marcotte, I.; Wang, T., Identification and Quantification of Glycans in Whole Cells: Architecture of Microalgal Polysaccharides Described by Solid-State Nuclear Magnetic Resonance. J. Am. Chem. Soc. 2021, 143, 19374-19388. 33. Reichhardt, C.; Cegelski, L., Solid-state NMR for bacterial biofilms. Mol. Phys. 2013, 7, 887- 894. 34. El Hariri El Nokab, M.; Lasorsa, A.; Sebakhy, K. O.; Pincchioni, F.; van der Wel, P. C. A., Solid-state NMR spectroscopy insights for resolving different water pools in alginate hydrogels. Food Hydrocol. 2022, 127, 107500. 35. El Hariri El Nokab, M.; van der Wel, P. C. A., Use of solid-state NMR spectroscopy for investigating polysaccharide-based hydrogels: A review. Carbohydr. Polym. 2020, 240, 116276. 36. Zhao, W.; Deligey, F.; Shekar, S. C.; Mentink-Vigier, F.; Wang, T., Current limitations of solid-state NMR in carbohydrate and cell wall research. J. Magn. Reson. 2022, 341, 107263. 37. Kang, X.; Zhao, W.; Widanage, M. C. D.; Kirui, A.; Ozdenvar, U.; Wang, T., CCMRD: a solid-state NMR database for complex carbohydrates. J. Biomol. NMR 2020, 74, 239-245. 413 38. Fritzsching, K. J.; Hong, M.; Schmidt-Rohr, K., Conformationally selective multidimensional chemical shift ranges in proteins from a PACSY database purged using intrinsic quality criteria. J. Biomol. NMR 2016, 64, 115-130. 39. Hou, G.; Yan, S.; Trebosc, J.; Amoureux, J. P.; Polenova, T., Broadband Homonuclear Correlation Spectroscopy Driven by Combined R2nv Sequences under Fast Magic Angle Spinning for NMR Structural Analysis of Organic and Biological Solids. J. Magn. Reson. 2013, 232, 18-30. 40. De Paepe, G.; Lewandowski, J. R.; Loquet, A.; Bockmann, A.; Griffin, R. G., Proton assisted recoupling and protein structure determination. J. Chem. Phys. 2008, 129, 245101. 41. Zhao, W.; Kirui, A.; Deligey, F.; Mentink-Vigier, F.; Zhou, Y.; Zhang, B.; Wang, T., Solid- state NMR of unlabeled plant cell walls: high-resolution structural analysis without isotopic enrichment. Biotechnol. Biofuels 2021, 14, 14. 42. Viger-Gravel, J.; Lan, W.; Pinon, A. C.; Berruyer, P.; Emsley, L.; Bardet, M.; Luterbacher, J., Topology of Pretreated Wood Fibers Using Dynamic Nuclear Polarization. J. Phys. Chem. C 2019, 123, 30407-30415. 43. Perras, F. A.; Luo, H.; Zhang, X.; Mosier, N. S.; Pruski, M.; Abu-Omar, M. M., Atomic- Level Structure Characterization of Biomass Pre- and Post-Lignin Treatment by Dynamic Nuclear Polarization-Enhanced Solid-State NMR. J. Phys. Chem. A 2017, 121, 623-630. 44. Fernando, L. D.; Dickwella Widanage, M. C.; Shekar, S. C.; Mentink-Vigier, F.; Wang, P.; Wi, S.; Wang, T., Solid-state NMR analysis of unlabeled fungal cell walls from Aspergillus and Candida species. J. Struct. Biol. X 2022, 6, 100070. 45. Deligey, F.; Frank, M. A.; Cho, S. H.; Kirui, A.; Mentink-Vigier, F.; Swulius, M. T.; Nixon, B. T.; Wang, T., Structure of In Vitro-Synthesized Cellulose Fibrils Viewed by Cryo-Electron Tomography and 13C Natural-Abundance Dynamic Nuclear Polarization Solid-State NMR. Biomacromolecules 2022, 23, 2290-2301. 46. Takahashi, H.; Lee, D.; Dubois, L.; Bardet, M.; Hediger, S.; De Paepe, G., Rapid Natural- Abundance 2D 13C-13C Correlation Spectroscopy Using Dynamic Nuclear Polarization Enhanced Solid-State NMR and Matrix-Free Sample Preparation. Angew. Chem. Int. Ed. 2012, 51, 11766-11769. 47. Kumar, A.; Watbled, B.; Baussanne, I.; Hediger, S.; Demeunynck, M.; De Paepe, G., Optimizing chemistry at the surface of prodrug-loaded cellulose nanofibrils with MAS-DNP. Commun. Chem. 2023, 6, 58. 48. Berruyer, P.; Gericke, M.; Moutzouri, P.; Jakobi, D.; Bardet, M.; Karlson, L.; Schantz, S.; heinze, T.; Emsley, L., Advanced characterization of regioselectively substituted methylcellulose model compounds by DNP enhanced solid-state NMR spectroscopy. Carbohydr. Polym. 2021, 262, 117944. 49. Brouwer, D. H.; Mikolajewski, J. G., Resolving the discrepancies in reported 13C solid state 414 NMR chemical shifts for native celluloses. Cellulose 2023, 30, 4827-4839. 50. Hoffman, R., Solid-state chemical-shift referencing with adamantane. J. Magn. Reson. 2022, 340, 107231. 51. Elkins, M. R.; Wang, T.; Nick, M.; Jo, H.; Lemmin, T.; Prusiner, S. B.; DeGrado, W. F.; Stohr, J.; Hong, M., Structural polymorphism of Alzheimer’s β-amyloid fibrils as controlled by an E22 switch: A solid-state NMR study. J. Am. Chem. Soc. 2016, 138, 9840-9852. 52. Kirui, A.; Du, J.; Zhao, W.; Barnes, W.; Kang, X.; Anderson, C. T.; Xiao, C.; Wang, T., A pectin methyltransferase modulates polysaccharide dynamics and interactions in Arabidopsis primary cell walls: Evidence from solid-state NMR. Carbohydr. Polym. 2021, 270, 118370. 53. Wang, T.; Chen, Y.; Tabuchi, A.; Cosgrove, D. J.; Hong, M., The target of β-expansin EXPB1 in maize cell walls from binding and solid-state NMR studies. Plant Physiol. 2016, 172, 2107- 2119. 54. Wang, T.; Salazar, A.; Zabotina, O. A.; Hong, M., Structure and Dynamics of Brachypodium Primary Cell Wall Polysaccharides from Two-Dimensional 13C Solid-State Nuclear Magnetic Resonance Spectroscopy. Biochemistry 2014, 53, 2840-2854. 55. Phyo, P.; Wang, T.; Kiemle, S. N.; O'Neill, H.; Pingali, S. V.; Hong, M.; Cosgrove, D. J., Gradients in wall mechanics and polysaccharides along growing inflorescence stems. Plant Physiol. 2017, 175, 1593-1607. 56. Kohn, R.; Kovac, P., Dissociation-constants of D-galacturonic and D-glucuronic acid and their O-methyl derivatives. Chem. Pap. 1978, 32, 478-485. 57. Phyo, P.; Gu, Y.; Hong, M., Impact of acidic pH on plant cell wall polysaccharide structure and dynamics: insights into the mechanism of acid growth in plants from solid-state NMR. Cellulose 2018, 26, 291-304. 58. Wang, T.; Yang, H.; Kubicki, J. D.; Hong, M., Cellulose structural polymorphism in plant primary cell walls investigated by high-field 2D solid-state NMR spectroscopy and density functional theory calculations. Biomacromolecules 2016, 17, 2210-2222. 59. Cordova, M.; Balodis, M.; de Almeida, B. S.; Ceriotti, M.; Emsley, L., Bayesian probabilistic assignment of chemical shifts in organic solids. Sci. Adv. 2021, 7, abk2341. 60. Fritzsching, K. J.; Yang, Y.; Schmidt-Rohr, K.; Hong, M., Practical use of chemical shift databases for protein solid-state NMR: 2D chemical shift maps and amino-acid assignment with secondary-structure information. J. Biomol. NMR 2013, 56, 155-167. 61. Yang, Y.; Fritzsching, K. J.; Hong, M., Resonance assignment of the NMR spectra of disordered proteins using a multi-objective non-dominated sorting genetic algorithm. J. Biomol. NMR 2013, 57, 281-296. 62. Hu, K. N.; Qiang, W.; Tycko, R., A general Monte Carlo/simulated annealing algorithm for 415 resonance assignment in NMR of uniformly labeled biopolymers. J. Biomol. NMR 2011, 50, 267-276. 63. Dick-Perez, M.; Wang, T.; Salazar, A.; Zabotina, O. A.; Hong, M., Multidimensional solid- state NMR studies of the structure and dynamics of pectic polysaccharides in uniformly 13C- labeled Arabidopsis primary cell walls. Magn. Reson. Chem. 2012, 50, 539-550. 64. Scheller, H. V.; Ulvskov, P., Hemicelluloses. Annu. Rev. Plant Biol. 2010, 61, 263-289. 65. Ridley, B. L.; O'Neill, M. A.; Mohnen, D., Pectins: Structure, biosynthesis, and oligogalacturonide-related signaling. Phytochemistry 2001, 57, 929-967. 66. Atmodjo, M. A.; Hao, Z.; Mohnen, D., Evolving views of pectin biosynthesis. Annu. Rev. Plant Biol. 2013, 64, 747-779. 67. Biswal, A. K.; Atmodjo, M. A.; Li, M.; Baxter, H. L.; Geun Yoo, C.; Pu, Y.; Lee, Y. C.; Mazarei, M.; Black, I. M.; Zhang, J. Y.; Ramanna, H.; Bray, A. L.; King, Z. R.; LaFayette, P. R.; Pattathil, S.; Donohoe, B. S.; Mohanty, S. S.; Ryno, D.; Yee, K.; Thompson, O. A.; Rodriguez, M.; Dumitrache, A.; Natzke, J.; Winkeler, K.; ......; Mohnen, D., Sugar release and growth of biofuel crops are improved by downregulation of pectin biosynthesis. Nat. Biotechnol. 2018, 36, 249-257. 68. Tan, L.; Zhang, L.; Black, I. M.; Glushka, J.; Urbanowicz, B.; Heiss, C.; Azadi, P., Most of the rhamnogalacturonan-I from cultured Arabidopsis cell walls is covalently linked to arabinogalactan-protein. Carbohydr. Polym. 2023, 301, 120340. 69. Shekar, S. C.; Zhao, W.; Fernando, L. D.; Hung, I.; Wang, T., A 13C three-dimensional DQ- SQ-SQ correlation experiment for high-resolution analysis of complex carbohydrates using solid-state NMR. J. Magn. Reson. 2022, 336, 107148. 70. Li, S.; Zhang, Y.; Hong, M., 3D 13C–13C–13C correlation NMR for de novo distance determination of solid proteins and application to a human α-defensin. J. Magn. Reson. 2010, 202, 203-210. 71. Bahri, S.; Safeer, A.; Adler, A.; Smedes, H.; Ingen, H. v.; Baldus, M., 1H-detected characterization of carbon-carbon networks in highly flexible protonated biomolecules using MAS NMR. J. Biomol. NMR 2023, 77, 111-119. 72. Fernando, L. D.; Dickwella Widanage, M. C.; Penfield, J.; Lipton, A. S.; Washton, N.; Latge, J. P.; Wang, P.; Zhang, L.; Wang, T., Structural Polymorphism of Chitin and Chitosan in Fungal Cell Walls From Solid-State NMR and Principal Component Analysis. Front. Mol. Biosci. 2021, 8, 727053. 73. Pillai, C. K. S.; Paul, W.; Sharma, C. P., Chitin and Chitosan Polymers: Chemistry, Solubility and Fiber Formation. Prog. Polym. Sci. 2009, 34, 641-678. 74. Lenardon, M. D.; Whitton, R. K.; Munro, C. A.; Marshall, D.; Gow, N. A. R., Individual Chitin Synthase Enzymes Synthesize Microfibrils of Differing Structure at Specific Locations 416 in the Candida Albicans Cell Wall. Mol. Microbiol. 2007, 66, 1164-1173. 75. Latgé, J.-P., The cell wall: a carbohydrate armour for the fungal cell. Mol. Microbiol. 2007, 66, 279-290. 76. Rui-Herrera, J.; Ortiz-Castellanos, L., Cell Wall Glucans of Fungi. A Review. Cell Surf. 2019, 5, 100022. 77. Grun, C. H.; Hochstenbach, F.; Humbel, B.; Verkleij, A. J.; Sietsma, J. H.; Klis, F. M.; Kamerling, J. P.; Vliegenthart, J. F. G., The structrue of cell wall a-glucan from fission yeast. Glycobiology 2005, 15, 245-257. 78. Gow, N. A. R.; Lenardon, M. D., Architecture of the dynamic fungal cell wall. Nat. Rev. Microbiol. 2022, 21, 248-259. 79. Fontaine, T.; Latgé, J. P., Galactomannan Produced by Aspergillus fumigatus: An Update on the Structure, Biosynthesis and Biological Functions of an Emblematic Fungal Biomarker. J. Fungi 2020, 6, 283. 80. Latge, J. P., Galactofuranose containing molecules in Aspergillus fumigatus. Med. Mycol. 2009, 47, S104-S109. 81. Heiss, C.; Skowyra, M. L.; Liu, H.; Klutts, J. S.; Wang, Z.; Williams, M.; Srikanta, D.; Beverley, S. M.; Azadi, P.; Doering, T. L., Unusual galactofuranose modification of a capsule polysaccharide in the pathogenic yeast Cryptococcus neoformans. J. Biol. Chem. 2013, 288, 10994-11003. 82. Li, D. W.; Hansen, A. L.; Yuan, C.; Bruschweiler-Li, L.; Bruschweiler, R., DEEP picker is a deep neural network for accurate deconvolution of complex two-dimensional NMR spectra. Nat. Commun. 2021, 12, 5229. 417 APPENDIX Table S11.1 An example of simulated data for a glucose unit in cellulose. The representative dataset of actual chemical shift values of cellulose includes the six 13C chemical shifts sequentially from C1 to C6: 105.8 ppm, 71.5 ppm, 75.8 ppm, 89.1 ppm, 72.5 ppm, 64.9 ppm. The resultant INADEQUATE and SQ-SQ correlation data could be employed to construct graphical depictions portraying the projected 2D NMR spectra. The X and Y coordinates are indicative of the chemical shifts (ppm) of NMR signals/cross-peaks within a 2D spectrum. Not applicable (/). a Cellulose DQ-SQ Correlation (ppm) X Y X Y C1 C1+C2 C2 C1+C2 SQ-SQ Correlation X C1 Y C1 X C1 Y C2 X C1 Y C3 X C1 Y C4 X C1 Y C5 X C1 Y C6 C1 105.8 C2 71.5 C3 75.8 C4 89.1 C5 72.5 C6 64.9 105.8 177.3 71.5 177.3 105.8 105.8 105.8 71.5 105.8 75.8 105.8 89.1 105.8 72.5 105.8 64.9 C2 C2+C3 C3 C2+C3 C2 C1 C2 C2 C2 C3 C2 C4 C2 C5 C2 C6 71.5 147.3 75.8 147.3 71.5 105.8 71.5 71.5 71.5 75.8 71.5 89.1 71.5 72.5 71.5 64.9 C3 C3+C4 C4 C3+C4 C3 C1 C3 C2 C3 C3 C3 C4 C3 C5 C3 C6 75.8 164.9 89.1 164.9 75.8 105.8 75.8 71.5 75.8 75.8 75.8 89.1 75.8 72.5 75.8 64.9 C4 C4+C5 C5 C4+C5 C4 C1 C4 C2 C4 C3 C4 C4 C4 C5 C4 C6 89.1 161.6 72.5 161.6 89.1 105.8 89.1 71.5 89.1 75.8 89.1 89.1 89.1 72.5 89.1 64.9 C5 C5+C6 72.5 177.3 C6 C5+C6 64.9 177.3 / C5 C1 C5 C2 C5 C3 C5 C4 C5 C5 C5 C6 72.5 105.8 72.5 71.5 72.5 75.8 72.5 89.1 72.5 72.5 72.5 64.9 C6 C1 C6 C2 C6 C3 C6 C4 C6 C5 C6 C6 64.9 105.8 64.9 71.5 64.9 75.8 64.9 89.1 64.9 72.5 64.9 64.9 a The chemical shifts data for the six carbons of glucose in cellulose is from CCMRD. 418 Table S11.2 Parameters of ssNMR experiments applied for plant and fungi samples. Corresponding figure numbers are indicated. Not applicable (/). Sample Experiment Experiment Time Number Scans Field (T) Spinning Speed (kHz) Temperature (K) Recycle Dalay (s) 1H Excitation (kHz) 13C Excitation (kHz) CP Contact Time (ms) 1H Decoupling (kHz) Acquisition Time (aq2; ms) Evolution Time (aq1; ms) Number of Points For Direct Dimensions (td2) Number of Points For Indirect Dimensions (td1) Figure 11.5a Arabidopsis thaliana DP J- INADEQUATE Figure 11.5a Spruce CP J-INADEQUATE Figure 11.5b Spruce CORD Figure 11.7a Aspergillus fumigatus DP J-INADEQUATE 30 h 128 16.5 20 293 1.8 / 62.5 / 83.3 12.8 5.0 1600 464 12 h 128 9.4 10 298 2.0 83.3 62.5 1.0 83.3 16.0 4.2 1600 166 9 h 192 9.4 10 298 1.8 83.3 62.5 1.0 83.3 14.0 5.3 1400 90 419 4 h 8 18.8 13.5 298 1.55 83.3 62.5 / 83.3 19.4 10.2 2600 1024 Figure 11.7b Aspergillus fumigatus Figure 11.7b Arabidopsis thaliana 1D 2s-DP 1D 2s-DP 5 min 140 min 128 18.8 13.5 298 2.0 / 62.5 / 83.3 28.7 / 2048 16.5 20 293 2.0 / 62.5 / 83.3 12.8 / 4096 1600 / / Figure S11.1 Projected 2D INADEQUATE correlation spectra of plant glucose residue in cellulose. 420 Figure S11.2 Projected 2D SQ-SQ 13C correlation spectra of glucose residues in cellulose. 421 Figure S11.3 Projected 2D INADEQUATE DQ-SQ 13C correlation spectra of plant xylose residue. 422 Figure S11.4 Projected 2D SQ-SQ 13C correlation spectra of plant xylose residue. 423 Figure S11.5 Projected 2D INADEQUATE DQ-SQ 13C correlation spectra of plant arabinose residue. 424 Figure S11.6 Projected 2D SQ-SQ 13C correlation spectra of plant arabinose residue. 425 Figure S11.7 Projected 2D 13C INADEQUATE correlation spectra of plant galactose residue. 426 Figure S11.8 Projected 2D SQ-SQ correlation spectra of plant galactose residue. 427 Figure S11.9 Projected 2D 13C INADEQUATE correlation spectra of plant galacturonic acid and glucuronic acid residues. 428 Figure S11.10 Projected 2D 13C SQ-SQ correlation spectra of plant galacturonic acid and glucuronic acid residues. 429 Figure S11.11 Projected 2D 13C INADEQUATE correlation spectra of plant rhamnose residue. 430 Figure S11.12 Projected 2D 13C SQ-SQ correlation spectra of plant rhamnose residue. 431 Figure S11.13 Projected 2D 13C INADEQUATE correlation spectra of plant mannose residue. 432 Figure S11.14 Projected 2D 13C SQ-SQ correlation spectra of plant mannose residue. 433 Figure S11.15 Projected 2D 13C INADEQUATE correlation spectra of fungi chitin. 434 Figure S11.16 Projected 2D 13C SQ-SQ correlation spectra of fungi chitin. 435 Figure S11.17 Projected 2D 13C INADEQUATE correlation spectra of fungi chitosan. 436 Figure S11.18 Projected 2D 13C SQ-SQ correlation spectra of fungi chitosan. 437 Figure S11.19 Projected 2D 13C INADEQUATE correlation spectra of fungi α-glucans. 438 Figure S11.20 Projected 2D 13C SQ-SQ correlation spectra of fungi α-glucans. 439 Figure S11.21 Projected 2D 13C INADEQUATE correlation spectra of fungi β-glucans. 440 Figure S11.22 Projected 2D 13C SQ-SQ correlation spectra of fungi β-glucans. 441 Figure S11.23 Projected 2D 13C INADEQUATE correlation spectra of fungi galactomannan (GM). 442 Figure S11.24 Projected 2D 13C SQ-SQ correlation spectra of fungi galactomannan (GM). 443 Figure S11.25 Projected 2D 13C INADEQUATE correlation spectra of fungi galactosaminogalactan (GAG). 444 Figure S11.26 Projected 2D 13C SQ-SQ correlation spectra of fungi galactosaminogalactan (GAG). 445