STRATEGIES FOR PROFILING AND DISCOVERY OF ACYLSUGAR SPECIALIZED METABOLITES By Steven Michael Hurney A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Chemistry Doctor of Philosophy 2018 A BSTRACT STRATEGIES FOR PROFILING AND DISCOVERY OF ACYLSUGAR SPECIALIZED METABOLITES By Steven Michael Hurney Plant metabolic processes have evolved to produce an enormous array of chemically diverse metabolites that serve important roles in growth, develop ment and resistance, while offering nutritional, medicinal, and economic value for humans. In the post - genome era, comprehensive measurements of plant chemistries, termed plant metabolomics, have experienced rapid growth as a field of research, with a bro ader goal being to provide the foundation for uncovering the relationship between the functions of genes, pro teins and enzymes responsible for the biosynthesis of metabolites. However, researchers in this field face formidable challenges because unlike pro teins and oligonucleotides which are constructed from a limited set of precursors, the building blocks of met abolites are far more varied. As a result, a nnotation and identification of novel metabolites remains the greatest obstacle to understanding the me chanisms responsible for metabolite accumulation and the functional significance of new metabolites. To addre ss these challenges, the goals of this research have been to discover novel specialized metabolites (those that are taxonomically restricted and n ot involved in central metabolism) to facilitate the discovery of enzymatic processes responsible for assembly of plant chemical defenses, and to develop improved methods and technologies for annotation, identification and dereplication of metabolite disco very. Chapters 2 and 3 of this dissertation demonstrate approaches for untargeted metabolite profiling of spec ialized metabolites extracted from leaf surface glandular trichomes (hair - like epidermal cells) from two species of the family Solanaceae, Salpigl ossis sinuata and Solanum quitoense . Liquid chromatography/mass spectrometry (LC/MS) profiling revealed divers e multiply esterified sugar metabolites, known as acylsugars, a family of metabolites known for anti - insect activity. Acylsugar metabolites were p urified and their structures were elucidated using one - and two - dimensional (1D and 2D) nuclear magnetic reson ance spectroscopy (NMR). These efforts established structures of 16 (of more than 400) new acylsucrose metabolites extracted from S. sinuata and e stablished a novel group of acylated myo - inositols and myo - inositol glycosides ( N - acetylglucosaminyl, glucopyranosyl and xylopyranosyl) from S. quitoense (9 structures). These results guided the discovery of previously unidentified acylsugar biosynthetic e nzymes operating in S. sinuata and S. quitoense acylsugar biosynthetic pathways a nd extended our understanding of the evolution of specialized metabolism in the Solanaceae, while providing new analytical approaches for defining acylsugar composition and bi odiversity. Mass spectrometry - based platforms are extremely effec tive tools for metabolite detection and investigations of metabolomes. However, for many metabolites, mass spectrometry alone does not provide unambiguous metabolite identification. While NMR spectroscopy provides more detailed structural information than MS, some researchers avoid its use because it common ly involves time - consuming metabolite purifications. In addition, as more metabolite structures are determined, there is a growing chance t hat researchers will purify and identify compounds that are alrea dy known but thought to be novel isomers. To aid dereplication of metabolite discovery, a homologous set of S - alkyl glutathione (GS - n - alkyl) standards featuring normal saturated chain lengths (1 - 24 carbons) was synthesized and exploited as liquid chromatog raphic (LC) retention index standards (Chapter 4). These standards encompass a wide reversed phase - LC retention range, are easily ionized by electrospray ionization (ESI) in positive - and neg ative - ion modes, and show improved capacity for standardizing chr omatographic retention. A thorough investigation of the dependence of acylsucrose retention index values using GS - n - alkyl standards was performed while altering several important chromatograp hic experiment parameters, including a comparison of columns, sol vent delivery systems, aqueous mobile phase pH, column temperature, LC gradient slope, and organic solvent component. GS - n - alkyl standard LC/MS analysis also shows promise for evaluating RP - H PLC column performance, batch - to - batch column reproducibility, an d degradation of column performance with use. This research has potential to improve interspecies SM metabolite discovery and dereplication. Copyright by STEVEN MICHAEL HURNEY 2018 v ACKNOWLEDG EMENTS This achievement w ould not have been possible without the encouragement, guidance , training and support from several important figures in my life . Foremost, I would like to extend my sincere gratitude to my advisor Professor A. Daniel Jones at Michi gan State University (MSU ). He i s an extraordinary mentor , teacher , scientist and friend . His scientific acumen, passion for teaching , wisdom and warmhearted manner are unmatched. I feel so fortunate to have had the opportunity to learn and grow under his guidance and I consider m yself privileged to have been one of his final students . I will sincerely miss seeing him dail y . Going forward, I am confident th e lessons I have learned from him will continue to permeate all areas of my life . I express my genuin e appreciation to c urrent and former members of the Jones Lab at MSU for their friendship, guidance with research , helpful suggestions projects. I am honored to have been part of such a great group of researcher s. I offer special thanks to group members Dr. Thilani Anthony, Fanny Chu, Dr. Banibrata Ghosh, Dr. Xiaoxiao Liu, Dr. Afrand Kamali, Dr. Sujana Pradhan , Kristen Reese , Dr. Zhenzhen Wang and Dr. Chen Zhang . I extend my sincere appreciation to c urrent and fo rmer members of the Solanaceae Specialized Metabolome Project at MSU , especially P rofessors Robert Last and Cornelius Barry , as well as their students and post - doctoral researchers , whom I ha ve had as collaborators over the years (special thanks go to Matt he w Bedewitz) . I also thank The National Science Foundation (NSF grant IOS - 1546617) for funding our project , making the research presented herei n possible. I thank the staff of the M SU Mass Spectrometry and Metabolomics Core ( par ticular attention to Dr. An th ony Schilmille r ) and the Max T. Rogers NMR facility ( special thanks go to Dr. Daniel Holmes) for technical support , training and helpful suggestions . Finally, I express my gratitude to my committee members, Professors Dana Spence, Kevin Walker and David We liky f or their guidance and helpful suggestions. vi Before I joined MSU, I began my career as an Analytical Chemist as a Pre - Bachelor Analytical College Co - op in Food Packaging Research a t The Dow Chemical Company. I am grateful for the time and effort that A nalytical Chemists Thomas S. Lardie and Vickie Langer devoted to training me, as well as their friendship and encouragement . The skills and experience I acquired during our time together was immensely helpful in graduate school and set the stage for me t o pu rsue my PhD in Chemistry . I also thank Dr. Kenneth Kearns (previously Professor of Analytical Chemistry at Saginaw Valley State University), who suggested during my undergrad uate program that I should pursue an advanced degree in science and gave me co nfid ence to do so. Finally, I am fortunate to have the love and encouragement of wonderful family and friends . During my childhood, my pare nts praised learning and academic achievements . T hey provided me the confidence a nd inspiration needed to seek lofty educ ational goals . This accomplishment would not have been possible without their love and support. T his achievement required considerable time and effort , at the expense of time enjoyed with my parents, siblings, family and friends . I sincerely appreciate the ir patience and repeated encouragement. I am especially grateful for the support and encouragement of my wonderful and loving wife, Dr. Krystyna Kijewska. I could not imagine my life without her. She makes everything that I do worthwhile . me, did all the dishes and took care of me for months while I was writing. She deserves all the credit. Love you. vii TABLE OF CONTENTS LIST OF TABLES ................................ ................................ ................................ ................................ ........ x LIST OF FIGURES ................................ ................................ ................................ ................................ ... xiv Chapter 1: Introduction ................................ ................................ ................................ ................................ . 1 1.1 Ro le of plant metabolomics for functional genomics and systems biology ................................ ............ 1 1.2 The evolution of plant specialized metabolism and natural chemical diversity ................................ ...... 2 1.3 Challenges of identifying plan t speciali zed metabolites ................................ ................................ ......... 3 1.4 Introduction to trichomes and acylsugars of the family Solanaceae ................................ ....................... 4 1.4.1 Trichome structures, f unctions and chemical compositions ................................ ............................ 4 1.4.2 Acylsugars in glandular trichomes of Solanaceae species ................................ .............................. 5 1.4.3 Biosynthesis of acyls ugars ................................ ................................ ................................ ............. 10 1.5 Analytical approaches for profiling and discovery of acylsugar metabolites ................................ ....... 12 1.5.1 Mass spectrometry ................................ ................................ ................................ ......................... 12 1.5.1.1 Gas Chromatography/Mass Spectrometry (GC/MS) ................................ .............................. 12 1.5.1.2 Ultra - High - Performance Liquid Chromatography/Mass Spectrometry (UHPLC/MS) ......... 13 1.5.1.3 Collision In duced Dissociation (CID) ................................ ................................ .................... 15 1.5.1.4 High Resolution Mass Spectrometry ................................ ................................ ...................... 15 1.5.1.5 Tandem Mass Spectrometry (MS/MS) ................................ ................................ ................... 16 1.5.2 Nuclear Magnetic Resonance (NMR) Spectroscopy ................................ ................................ ..... 17 1.5.2.1 Extraction a nd purification of acylsugar metabolites for NMR analysis ............................... 18 1.5.2.2 Identification by 1D and 2D NMR spectroscopy ................................ ................................ ... 18 1.5.3 Chromat ographic Retention Indexing (RI) for annotation of specialized metabolites .................. 19 1.5.3.1 GC/MS Kov á ts retention indexing ................................ ................................ ......................... 20 1.5.3.2 H PLC and LC/MS retention indexing approaches ................................ ................................ . 20 1.6 Summary of research ................................ ................................ ................................ ............................ 21 REFERENCES ................................ ................................ ................................ ................................ ........... 23 Chapter 2: LC/MS profiling and NMR structural elucidation of specialized metabolites from Salpiglossis sinuata reveals extensive acylsucrose diversity including unsaturated and aromatic esters ....................... 30 2.1 Introduction ................................ ................................ ................................ ................................ ........... 30 2.2 Materials and methods ................................ ................................ ................................ .......................... 32 2.2.1 Plant cultivation and metabolite extraction ................................ ................................ ................... 32 2.2.2 Profiling of acylsugar metabolites using UHPLC/MS and MS/MS ................................ .............. 32 2.2.2. 1 Deep profiling of acyl sucrose specialized metabolites by LC/ESI+/MS ............................... 33 2.2.3 Assessment of acyl group diversity by transesterification and GC/MS ................................ ........ 34 2. 2.4 Purification of acylsugar metabol ites by semi - preparative HPLC ................................ ................ 34 2.2.5 Analys is of acylsugars by NMR spectroscopy ................................ ................................ .............. 34 2.3 Re sults and Discussion ................................ ................................ ................................ ......................... 35 2.3.1 UHPLC/ESI/CID/QTof/MS profiling establishes diversity of S. sinuata acylsucroses ................ 35 2.3.2 GC/MS pr ofiling reveals sugar ester composition ................................ ................................ ......... 40 2.3.3 1D and 2D NMR of purified acylsucroses reveals structural diversity ................................ ......... 41 2.3.4 St ructure divers ity of acylsucroses from S. s inuat a ................................ ................................ ....... 46 2.3.4.1 Acyl group diversity in S. sinuata ................................ ................................ .......................... 46 2.3.4.2 Number of acylations in S. sinuata acylsucroses ................................ ................................ ... 47 2.3.4.3 Positions of acyl groups in S. sinuata acylsucroses ................................ ................................ 47 2.4 Conclusions ................................ ................................ ................................ ................................ ........... 48 viii APPENDIX ................................ ................................ ................................ ................................ ................. 49 REFERENCES ................................ ................................ ................................ ................................ ......... 243 Chapter 3: Unexpected diversity in acylsugar metaboli tes: acylinositols from Solanum quitoense ......... 247 3.1 Introduction ................................ ................................ ................................ ................................ ......... 247 3.2 Materials and Methods ................................ ................................ ................................ ........................ 249 3.2.1 Plant Cultivat ion and metabolite extraction ................................ ................................ ................ 249 3.2.2 Profiling of acylsugar metabolites using UHPLC/MS and MS/MS ................................ ............ 249 3.2.3 Purification of acyl sugar metabolites by semi - preparative HPLC ................................ .............. 250 3.2.4 Analysis of acylsugars by NMR spectroscopy ................................ ................................ ............ 251 3.3 Results and Discussion ................................ ................................ ................................ ....................... 251 3.3.1 UHPLC/ESI/CID/QTof/MS profiling and NMR structural elucidation establishes diversity of S. quitoense acylinositols ................................ ................................ ................................ .......................... 251 3.3. 2 LC/MS profiling and NMR structural elucidation establishes acylated myo - inositols ................ 252 3.3.3 Negative - ion mode MS/MS spectra of acyl ated myo - inositols differ from acylglucoses ........... 256 3.3.4 LC/MS profiling and NMR structural elucidation establishes acylated myo - inositol glycosides 258 3.3.4.1 Discovery of 4 - O - N - acetylglucosa minyl (NAG) acylated myo - inositols ............................. 258 3.3.4.2 Discovery of 4 - O - glucopyranosyl (G) acylated myo - inositols ................................ ............. 261 3.3.4.3 Discovery of 4 - O - xyl opyranosyl (X) acylated myo - inositols ................................ ............... 262 3.3.5 Deep profiling of acylinositols by LC/MS ................................ ................................ .................. 26 3 3.3.6 1D and 2D NMR of purifi ed acylinositols reveals acylation positions ................................ ....... 264 3.3.3 Structural diversity of acylinositols from S. quitoense ................................ ................................ 269 3.4 Conclusions ................................ ................................ ................................ ................................ ......... 269 APPENDIX ................................ ................................ ................................ ................................ ............... 271 REFERENCES ................................ ................................ ................................ ................................ ......... 39 7 Chapter 4: S - alkyl glutathione ret ention indexing and column performance evaluation standards for improved annotation, identification and dereplication of metabolite discovery ................................ ....... 400 4.1 Introduction ................................ ................................ ................................ ................................ ......... 400 4.1.1 LC column performance ................................ ................................ ................................ .............. 403 4.2 Materials and Methods ................................ ................................ ................................ ........................ 404 4.2.1 Synthesis of S - alkyl glut athione standards ................................ ................................ .................. 404 4.2.1.1 Synthesis of GS - 2 and GS - 3 ................................ ................................ ................................ . 404 4.2.1.2 Synthesis of GS - 4 to GS - 9 ................................ ................................ ................................ ... 405 4.2.1.3 Synthesis of GS - 10 to GS - 20 and GS - 22 ................................ ................................ ............. 405 4.2.1.4 Synthesis of GS - 21, GS - 2 3 and GS - 24 ................................ ................................ ................ 406 4.2.2 Preparation of retention index standards and their application in LC/MS analyses ................... 407 4.2.2.1 Preparation of GS - n - alkyl mixed stock solution ................................ ................................ .. 407 4.2.2.2 Preparation of S. sinuata retention indexing stock solution ................................ ................. 408 4.2.3 UHPLC/MS me thods and experimental conditions ................................ ................................ .... 409 4.2.3.1 LC/MS instrument configurations ................................ ................................ ........................ 409 4.2.3.2 Retention indexing peak detection para meters ................................ ................................ ..... 409 4.2.3.3 Chr omatographic columns ................................ ................................ ................................ ... 410 4.2.3.4 Column performance and LC systems evaluations ................................ .............................. 411 4.2.3.5 Mobile phase pH dependence of acylsugar metabolite retention index values .................... 411 4.2.3.6 Column temperature dependence of acylsugar retention inde x values ................................ 411 4.2.3.7 Depen dence of acylsugar retention index values on mobile phase gradient slope ............... 411 4.2.3.8 Acylsugar retenti on index values using methanol as organic mobile phase component ...... 412 4.2.3.9 MS/MS spectra of GS - n - alkyl standards ................................ ................................ .............. 412 4.3 Resul ts and discussion ................................ ................................ ................................ ........................ 412 4.3.1 LC /MS of GS - n - alkyl standards ................................ ................................ ................................ .. 413 ix 4.3.2 RI corrects day - to - day chromatographic variation ................................ ................................ ...... 415 4.3.3 Potential for cross - platform RI a pplication ................................ ................................ ................. 417 4.3.4 GS - n - alkyls for use in column performance evaluation ................................ .............................. 419 4.3.5 C18 columns differ in retention selectivity ................................ ................................ ................. 424 4.3.6 RI dependence on mobile phase pH, colu mn temperature and LC gradient ............................... 427 4.3.7 RI dependence with methanol organic compone nt ................................ ................................ ...... 433 4.3.8 GS - n - alkyl standards further applications ................................ ................................ ................... 436 4.4 Conclusions ................................ ................................ ................................ ................................ ......... 439 APPENDIX ................................ ................................ ................................ ................................ ............... 441 REFERENCES ................................ ................................ ................................ ................................ ......... 461 Chapter 5: Closing Thoughts ................................ ................................ ................................ .................... 465 x L IST OF TABLES Table 1.1. Examples of leaf surface acylsugars from Solanaceae species with identified positions of acylations and their residues ................................ ................................ ................................ ......................... 8 Table 2.1. Summary of NMR resolved acylsu croses purified from S. sinuata extracts and percent peak area of [M+NH 4 ] + ion. ................................ ................................ ................................ ................................ ........ 45 Table 2.2. Plant cultivation and metabolite extraction metadata ................................ ................................ 53 Table 2.3. UHPLC/MS metadata ................................ ................................ ................................ ............... 54 Table 2.4. Deep profiling re sults (metabolites highlighted in bold were identified by NMR spectroscopy) ................................ ................................ ................................ ................................ ................................ .... 56 Table 2.5. LC/MS/MS metadata ................................ ................................ ................................ ................. 69 Table 2.6. GC/MS metadata ................................ ................................ ................................ ....................... 87 Table 2.7. GC/MS retention indexing and TIC peak area assessment ................................ ....................... 90 Table 2.8. Purification of ac ylsucroses by semi - preparative HPLC ................................ ........................... 94 Table 2.9. Bruker 900 MHz NMR Instrument Metadata ................................ ................................ ........... 96 Tab le 2.10. Summary of 1 H chemical shifts of sucrose core hydrogen atoms. Chemical shifts labeled in bold indicate acyl substitutions are located at those positions. All spectra were referenced to non - deuterated solvent signal of CDCl 3 H = 7.26 ppm), except for S5:25:4(2,5,5,5,8 P ) which was referenced to non - deuterated solvent signal of acetonitrile - d 3 H = 1.94 ppm). ................................ ................................ ..... 98 Tab le 2.11. Sum mary of 13 C chemical shifts of sucrose core carbon atoms. Chemical sh ifts labeled in bold indicate acyl substitutions are located at those positions. All spectra were referenced to non - deuterated solvent signal of CDCl 3 C = 77.20 ppm), except for S5:25:4 (2,5,5,5,8 P ) which was referenced to non - deuterated solvent signal of acetonitrile - d 3 C = 118.70 ppm). ................................ ................................ . 99 Table 2.12. S4:19:0( 3,5,5,6) Chemical shifts and coupling constants ................................ ..................... 100 Tabl e 2.13. S4:19:0( 2,5,6,6) Chemical shifts and coupling constants ................................ ..................... 109 Table 2.14. S5:20:0( 2,2,5,5,6) Chemical shifts and coupling constants ................................ .................. 118 Table 2.15. S4:20:0( 4,5,5,6) Chemical shifts and coupl ing constants ................................ ..................... 127 Table 2.16. S4:20:0( 3,5,6,6) Chem ical shifts and coupling constants ................................ ..................... 136 Table 2.17. S4:20:0( 2,6,6,6) Chemical shifts and coupling constants ................................ ..................... 145 Tabl e 2.18. S5:22:1( 2,5,5,5,5 T ) Chemical shifts and coupling constants ................................ ................. 154 Table 2.19. S5:21:0( 2,2,5,6,6) Chemical shifts and coupling constants ................................ .................. 162 xi Table 2.20. S4:21:0( 5,5,5,6) Chemical shifts and co upling constants ................................ ..................... 171 Table 2.21. S5:25:4( 2,5,5,5,8 P ) Chemical shifts and coupling constants ................................ ................. 180 Table 2.22. S5:22:0( 2,2,6,6,6) Chemical shifts and coupling constants ................................ .................. 1 89 Table 2.23. S4:22:0( 5,5,6,6) Chemical shifts and coupling constants ................................ ..................... 198 Table 2.24. S5:23:0( 2,5,5 ,5,6) Chemical shifts and coupling constants ................................ .................. 207 Table 2.25. S4:23:0( 5,6,6,6) Chemical shifts a nd coupling constants ................................ ..................... 216 Table 2.26. S5:24:0( 2,5,5 ,6,6) Chemical shifts and coupling constants ................................ .................. 225 Table 2.27. S6:25:0( 2,2,5,5,5,6) Chemical shifts and coupling constants ................................ ............... 234 Table 3.1. Summary of NMR resolved acylinositols purified from S. quito ense extracts and percent peak area of ions by negative ion mode. ................................ ................................ ................................ ........... 268 Table 3.2. Plant cultivation and metabolite extraction metadata ................................ .............................. 273 Table 3.3. UHPLC/MS metadata ................................ ................................ ................................ ............. 274 Table 3.4. Progenesis QI metadata ................................ ................................ ................................ ........... 276 Table 3.5. S. quitoense acylinositol deep profiling results generated by ESI - mode using Waters Progen esis QI software. ................................ ................................ ................................ ................................ ............... 277 Table 3.6. S. quitoense acyli nositol deep profiling results generated by ESI+ mode using Waters Progenesis QI software. ................................ ................................ ................................ ................................ ............... 279 Table 3.7. MS/MS precursor masses and time windows ................................ ................................ .......... 281 Table 3.8. MS/MS precursor masses, time window s and cone voltages ................................ .................. 281 Table 3.9. LC/MS/MS metadata ................................ ................................ ................................ ............... 282 Table 3.10. Purification of acylinositols by semi - preparat ive HPL C ................................ ....................... 302 Table 3.11. Bruker 900 MHz NMR Instrum ent Metadata ................................ ................................ ....... 304 Table 3.1 2. Agilent DDR2 500 MHz NMR Instrument Metadata ................................ ........................... 307 Table 3.13. Varian Inova 600 MHz NMR Instrument Metadata ................................ ............................. 310 Table 3.14. Summary of 1 H c hemical shifts of inositol core hydrogen atoms. Chemical shifts labeled in bold indicate acyl substitutions are located at those positions. All spectra were referenced to non - deuterated solvent signal of acetonitrile - d 3 H = 1.94 ppm). ................................ ................................ ..................... 311 Table 3.15. Summary of 13 C chemical shifts of inositol core carbon atoms. Chemical shifts lab eled in bold indicate acyl substitutions are located at those posit ions. All spectra were referenced to non - deu terated solvent signal of acetonitrile - d 3 C = 118.70 ppm). ................................ ................................ ................. 312 Table 3.16. NAG - I3:22:0( 2,10,10) Chemical shifts and coupling constants ................................ ........... 313 xii Table 3.17. G - I3:22:0( 2,10,10) Chemical shifts and coup ling constants ................................ ................. 32 2 Table 3.18. X - I3:22:0(2,10,10) Chemical shifts and coupling constants ................................ ................. 331 Table 3.19. NAG - I3:24:0(2,10,12) Chemical shifts and coupli ng constants ................................ ........... 340 Table 3.20. G - I3:24:0(2,10,12) Chemical shifts and coupling constants ................................ ................. 350 Table 3.21. I3:22:0(2,10,10) Chemical shifts and coupling constants ................................ ..................... 360 Tab le 3.22. I4:24:0(2,2,10,10) Chemical shifts and coupling constants ................................ .................. 368 Table 3.23. I3:24:0(2,10,12) Chemical shifts and coupling constants ................................ ..................... 377 Table 3.24. I4:26:0(2,2,10,12) Chemical shifts and co upling constants ................................ .................. 385 Table 3.25. I4:24:0(2,2 ,10,10) 1D - TOCSY transfer 2 - - CH 2 (S ) and 2 - CH 3 (S) normalized integrals. 393 Table 3.26. I4:24:0(2,2,10,10) 1D - TOCSY transfer 3 - - CH 2 (S ) and 3 - CH 3 (S) normalized integrals. 394 Table 3.27. I4:26:0(2,2,10,12) 1D - TOCSY tra nsfer 2 - - CH 2 (S ) and 2 - CH 3 (S) normalized integrals. 395 Table 3.28. I4:26:0(2,2,10,12) 1D - TOCSY transfer spectra 3 - - CH 2 excitation at 1.48 ppm (generated using Varian Inova 600 MHz spectrometer). ................................ ................................ ............................ 396 Table 4.1. Preparation of GS - n - alkyl Mixed Stock Solution ................................ ................................ ... 408 Table 4.2. LC/MS RT and RI results for 16 S. sinuata acylsucrose metabolites with structural identifications (column SAE - A, temperature 50°C, aqueous 10 mM ammonium formate pH 2.8 and linear gradient 1 - 100% acetonitrile, slope 1% B min - 1 ) . Analyses were performed in triplicate (two by ESI - and one by ESI+) #2 - 4) described below in Figure 4.3 . Pseudomolecular [M+HCOO] - and [M+NH 4 ] + ions were detected by ESI - and ESI+ ion modes, except for italicized characteristic fragment ions generated at elevated collision potentials (20 V CID by ESI - and 10 V CID by ESI+) . ................................ ................................ ............ 410 Table 4.3. Coefficients for calculating acylsucrose retention index values as a function of aqueous mobile phase pH and column temperature using solvent A = 10 mM aqueous ammonium formate adj usted to pH with formic acid and solvent B = acetonitrile, 1% acetonitrile/min gradient. Column = Ascentis Express C18, 2.1 x 100 mm. Solvent gradient slope at 1% acetonitrile/min. Coefficients were determined by quadratic fit regression of experimental 23 K) as reference. ....................... 432 Table 4.4. LC/MS retention time and retention index values for 16 S. sinuata acylsucrose metabolites with str uctural identifications using methanol organic component (other LC conditions were held const ant as outlined in Section 4.2.3.4). Analyses were performed on an Ascentis Express C18 column in triplicate (two by ESI - and one by ESI+) on four separate dates ( three S4:20:0 isomers (peaks #2 - 4) shown i n Figure 4.3. The same ions listed in Table 4.2 were used for detection. Acylsucrose numbering is in order of elution by acetonitrile using methanol organic modifi er (Table 4.2) ................................ ................................ ................................ ................................ ................. 435 Table 4.5. MS/MS product ion analysis of GS - n - alkyl standards. Ratio of common fragment [C 10 H 12 N 3 O 5 ] - = 254.078 and [M - H] - ion intensities measured using 10 - 40 V CID ramp (0.5 s acquisition time). ........ 458 xiii Table 4.6. MS/MS product ion analysis of GS - n - alkyl standards. Ratio of common frag ment [C 5 H 8 NO 3 S] + = 162.022 and [M+H] + ion intensities measured using 10 - 40 V CID ramp (0.5 s acquisition time). ....... 459 xiv L IST OF FIGURES Figure 1.1. Acylsug a r ester groups and carbohydrates observed within Solanaceae species. C2: acetyl, C3: propionyl, nC4: butyryl, iC4: isobutyryl, nC5: pentanoyl, aiC5: 2 - methylbutanoyl, iC5: 3 - methyl butanoyl, nC6: hexanoyl, aiC6: 3 - methylpentanoyl, iC6: 4 - methylpentanoyl, n C7: heptanoyl, aiC7: 4 - methylhexanoyl, iC7: 5 - methylhexanoyl, nC8: octanoyl, aiC8: 5 - methylheptanoyl, iC8: 6 - methylheptanoyl, iC9: 7 - methyloctanoyl, nC10: decanoyl, iC10: 8 - methyln onanoyl, nC12: dodecanoyl, iC12: 10 - methylundecanoyl, C3 M : malonyl, C5 S : se n ecioyl, C5 T : tiglyl, C8 P : phenylacetyl. ...................... 7 Figure 2.1. UHPLC/ESI/CID/QTof/MS metabolite profiles (left) and mass spectra generated from acylsugar #12 (right): (A) ESI( - ) BPI chromatogram displ ay ing acylsugars (formate adducts) at CID Function 1 = 0 V, purified acylsug ars characterized by NMR analysis are annotated by a number indicative of their order of chromatographic elution; (B) ESI( - ) XIC for C5 and C6 carboxylate fragment anions, m/z 101. 06 and 115.08 combined (Function 4 = 50 V), hexaacylsucrose #16 shows signif icantly weaker signal for C5 and C6 carboxylate anions compared to tetra - and pentaacylsucrose analogues; (C) ESI( - ) XIC for C5 T carboxylate fragment anion, m/z 99.05 (Function 4 = 5 0 V), co - elution of acylsucrose #7 with #8 is evident; (D) ESI( - ) XIC for C8 P carboxylate fragment anion, m/z 135.05 (Function 3 = 25 V), (E) ESI( - ) low energy (Function 1) mass spectrum of acylsucrose #12, S4:22:0(5,5,6,6), displaying [M+HCOO] - adduct, (F ) ESI( - ) at elevated energy (Function 4), displaying neutral losses consis tent with two C6 and two C5 acyl groups and prominent C5 and C6 carboxylate fragments at m/z 101.06 and 115.08; (G) ESI(+) low energy (Function 1 = 0 V) mass spectrum of peak #12 di splaying [M+NH 4 ] + adduct, (H) ESI(+) at elevated energy (Function 2 = 10 V ), displaying cleavage of the glycosidic linkage to reveal a C5 - acylated furanose ring fragment m/z 247.12 and a pyranose ring fragment m/z 443.26 acylated with two C6 and one C5 e st ers. ................................ ................................ ................................ ................................ ..................... 36 Figure 2.2. MS/MS product ion spectra of [M+HCOO] - and [M+NH 4 ] + of acylsugars generated using a linear 5 - 60 V collision energy ramp with 0.5 s scan time; (A) ESI( - ) MS/MS spectrum of pr oduct ions of m/z 763 ([M+HCOO] - ) for acylsugar #7 showing evidence for neutral losses of C2, C5 T , C5 and H 2 O, as well as corresponding carboxylate anions for ester groups C5 T and C5 ( m/z 99 and 101); (B) ESI(+) MS/MS spectrum of product ions of m/z 736 ([ M+NH 4 ] + ) for acylsugar #7 showing cleavage of glycosidic linkage to yield abundant acylated furanose ( m/z 289), and less abundant acylated pyranose ( m/z 413) product ions; (C) ESI( - ) MS/MS spectrum of product ions of m/z 799 ([M+ HCOO] - ) of acylsugar #10 sh owing evidence of neutral losses of C2, C5, C8 P and H 2 O, as well as corresponding carboxylate anions for ester groups C5 ( m/z 101) and C8 P ( m/z 135); (D) ESI(+) MS/MS spectrum of products of m/z 772 ([M+NH 4 ] + ) for acylsugar #10 showing cleavage of glyco sid ic linkage to produce abundant acylated furanose ( m/z 289), and less abundant acylated pyranose ( m/z 449) fragment ions. ................................ ................................ ................ 39 Figure 2.3. Hydroxyl positions on sucrose available for acylation , an d ester groups observed in S. sinuata GT extracts ................................ ................................ ................................ ................................ .................. 41 Figure 2.4. HMBC spectrum (top) and correlations of acylsucrose #10 (bottom); 1 H spectrum is projected on the F2 axis and 13 C spectrum is pr - esters to carbon ................................ 43 Figure 2.5. Micrographs of S. sin u ata Type I/IV glandul ar trichomes located on the surface of a young leaflet. ................................ ................................ ................................ ................................ ......................... 50 Figure 2.6. LC/MS shows evidence for heptaacylsucrose S7:27:0 using positive - ion mode ESI but negligible signal i s detected in negative - ion mode. (A) Positive - ion mode extracted ion chromatogram for xv [M+NH 4 ] + ( m/z 836.427) for S7:27:0 showing one major and three minor chromatographic peaks; (B) Negative - ion mode chromatogram for [M+formate] - ( m/z 863.392) on the sam e absolute y - axis scale (100% = 1.08 x 10 4 ion counts) shows negligible signal for S7:27:0. The structure shown is putative and not solved by NMR, but is pro posed based on our knowledge of acyl group selectivity (established by NMR resolved structures of o th er acylsucroses). The [M+NH 4 ] + consistent with S7:27:0 and fragment ion m/z 373.15 (Function 3 = 25 V) are consistent with an acylsucrose that has a furano se ring substituted by one C5 and three C2 ester groups. ................................ ................................ ................................ ................................ . 51 Figure 2.7 . Sum of extracted ion peak areas of [M+NH 4 ] + ions for S. sinuata acylsucroses, organized by the number of acyl groups on the sucrose core (e.g. S4, S5, and S6 refer to tetra - , penta - , and hexa - acylsucroses). Separate ba rs indicate peak area totals for acylsucroses containing all saturated acyl grou ps (green) and one or two C5 T and C8 P acyl groups (blue), which accounted for ~14% of the total acylsugar peak area. ................................ ................................ ................................ ................................ .................... 52 Figure 2.8 . S4:19:0(3,5,5,6) MS/MS product ion spectra, precursor ion [M+NH 4 ] + (top) & [M+HCOO] - (bottom, magnified 5x over m/z 51 - 664) ................................ ................................ ................................ .... 71 Figure 2.9. S4:19:0(2,5,6,6) MS/MS product io n spe ctra, precursor ion [M+NH 4 ] + (top) & [M+HCOO] - (bottom, magnified 10x over m/z 51 - 664) ................................ ................................ ................................ .. 72 Figure 2.10. S5:20:0(2,2,5,5,6) MS/MS product ion spectra, precursor ion [M+NH 4 ] + (top) & [M+HCOO ] - (bo ttom, magnified 4x over m/z 51 - 692) ................................ ................................ ................................ .... 73 Figure 2.11. S4:20:0(4,5,5,6) MS/MS product ion spectra, precursor ion [M+NH 4 ] + (top) & [M+HCOO] - (bottom, magnified 6x over m/z 51 - 678) ................................ ................................ ................................ .... 74 Figure 2.12. S4:20:0(3,5,6,6) MS/MS product ion spectra, precursor ion [M+NH 4 ] + (top) & [M+HCOO] - (bottom, magnified 3x over m/z 51 - 678) ................................ ................................ ................................ .... 75 Figu re 2.13. S4:20:0(2,6,6,6) MS/MS product ion spectra, precursor ion [M+NH 4 ] + (top) & [M+HCOO] - (bottom, magnified 14x over m/z 51 - 678) ................................ ................................ ................................ .. 76 Figure 2.14. S5:22:1(2,5,5,5,5 T ) MS/MS product ion spect ra, precursor ion [M+NH 4 ] + (to p) & [M+HCOO] - (bottom, magnified 9x over m/z 51 - 718) ................................ ................................ ................................ .... 77 Figure 2.15. S5:21:0(2,2,5,6,6) MS/MS product ion spectra, precursor ion [M+NH 4 ] + (top) & [M+HC OO] - (botto m, magnified 4x over m/z 51 - 706) ................................ ................................ ................................ .... 78 Figure 2.16. S4:21:0(5,5,5,6) MS/MS product ion spectra, precursor ion [M+NH 4 ] + (top) & [M+HCOO] - (bottom, magnified 5x over m/z 51 - 692) ................................ ................................ ................................ .... 79 Figure 2.17. S5:25:4(2,5,5,5,8 P ) MS/MS product ion spectra, precursor ion [M+NH 4 ] + (top) & [M+HCOO] - (bottom, magnified 6x over m/z 51 - 754) ................................ ................................ ................................ .... 80 Figu re 2.18. S5:22:0(2,2,6,6,6) MS/MS product ion spectra, precursor ion [M+NH 4 ] + (top) & [M+HCOO] - (bottom, magnified 7x over m/z 51 - 720) ................................ ................................ ................................ .... 81 Figure 2.19. S4:22:0(5,5,6,6) MS/MS prod uct ion spectra , precursor ion [M+NH 4 ] + (top) & [M+HCOO] - (bottom, magnified 3x over m /z 51 - 706) ................................ ................................ ................................ .... 82 Figure 2.20. S5:23:0(2,5,5,5,6) MS/MS product ion spectra, precursor ion [M+NH 4 ] + (top) & [M +HCOO] - (bottom, magnifi ed 7x over m/z 51 - 734) ................................ ................................ ................................ .... 83 xvi Figure 2.21. S4:23:0(5,6,6,6) MS/MS product ion spectra, precursor ion [M+NH 4 ] + (top) & [M+HCOO] - (bottom, magnified 3x over m/z 51 - 720) ................................ ................................ ................................ .... 84 Figure 2.22. S5:24:0(2,5,5,6,6) MS/MS product ion spectra, precursor ion [M+NH 4 ] + (top) & [M+HCOO] - (bottom, magnified 3x over m/z 51 - 748) ................................ ................................ ................................ .... 85 Figure 2.23. S6:25:0(2,2,5,5,5,6) MS/MS product ion spectra, precursor ion [M+NH 4 ] + (top) & [M+HCOO] - (bottom, magnified 200x over m/z 51 - 776). Note; there is a graph header mistake. This sample ................................ ............ 86 Figure 2.24. GC/MS total ion chromatograms (3 - 23 mins): process blank (bottom), transesterified S. sinuata sample (middle), and retention indexing series C7 - C13 (top). Pea ks are annotated wit h retention time, base peak m/z , and integrated peak area. ................................ ................................ ........................... 89 Figure 2.25. GC/EI/MS of 2 - methylpropanoic acid ethyl ester (iC4) ................................ ........................ 91 Figure 2.26. GC/EI/MS of 2 - methylbutanoic acid ethyl ester (aiC5) ................................ ........................ 91 Figure 2.27. GC/EI/MS of 3 - methylbutanoic acid ethyl ester (iC5) ................................ .......................... 92 Figure 2.28. GC/EI/MS of ethyl tiglate (C5 T ) ................................ ................................ ............................ 92 Figure 2.29. GC/EI/MS of 3 - methylpentanoic acid ethyl ester (aiC6) ................................ ...................... 93 Figu re 2.30. GC/EI/MS of phenylacetic acid ethyl ester (C8 P ) ................................ ................................ .. 93 Figure 2.31. S4:19:0( 3,5,5,6) 1 H NMR ................................ ................................ ................................ .... 102 F igure 2.32. S4:19:0( 3,5,5,6) 13 C NMR ................................ ................................ ................................ ... 103 Figure 2.33. S4:19:0( 3,5,5,6) 1 H - 1 H gCOSY ................................ ................................ ........................... 104 Figure 2.34. S4:19:0( 3,5,5,6) gHSQC ................................ ................................ ................................ ...... 105 Figure 2.35. S4:19:0( 3,5,5,6) gHMBC ................................ ................................ ................................ .... 106 Figure 2.36. S4:19:0( 3,5,5,6) J - resolved ................................ ................................ ................................ .. 107 Figure 2.37. S4:19:0( 3,5,5,6) ROESY ................................ ................................ ................................ ..... 108 Figure 2.38. S4:19:0( 2,5,6,6) 1 H NMR ................................ ................................ ................................ .... 111 Figure 2.39. S4:19:0( 2,5,6, 6) 13 C NMR ................................ ................................ ................................ ... 112 Figure 2.40. S4:19:0( 2,5,6,6) 1 H - 1 H gCOSY ................................ ................................ ........................... 113 Figure 2.41. S4:19:0( 2,5,6,6) gHSQC ................................ ................................ ................................ ...... 11 4 F igure 2.42. S4:19:0( 2,5,6,6) gHMBC ................................ ................................ ................................ .... 115 Figure 2.43. S4:19:0( 2,5,6,6) J - resolved ................................ ................................ ................................ .. 116 Figure 2.44. S4:19:0( 2,5,6,6) ROESY ................................ ................................ ................................ ..... 117 xvii Figure 2.45. S5:20:0( 2,2,5,5,6) 1 H NMR ................................ ................................ ................................ . 120 Figure 2.46. S5:20:0( 2,2,5,5,6) 13 C NMR ................................ ................................ ................................ 121 Figure 2.47. S5:20:0( 2,2,5,5,6) 1 H - 1 H gCOSY ................................ ................................ ........................ 122 Figure 2.48. S5:20:0( 2,2,5,5,6) gHSQC ................................ ................................ ................................ ... 123 Figure 2.49. S5:20:0( 2,2,5 ,5,6) gHMBC ................................ ................................ ................................ . 124 Figure 2.50. S5:20:0( 2,2,5,5,6) J - resolved ................................ ................................ ............................... 125 Figure 2.51. S5:20:0 ( 2,2,5,5,6) ROESY ................................ ................................ ................................ .. 126 Figure 2.52. S4:20:0( 4,5,5,6) 1 H NMR ................................ ................................ ................................ .... 129 Figure 2.53. S4:20:0( 4,5,5,6) 13 C NMR ................................ ................................ ................................ ... 130 Figure 2. 54. S4:20:0( 4,5,5,6) 1 H - 1 H gCOSY ................................ ................................ ........................... 131 Figure 2.55. S4:20:0( 4,5,5,6) gHSQC ................................ ................................ ................................ ...... 132 F igure 2.56. S4:20:0( 4,5,5,6) gHMBC ................................ ................................ ................................ .... 133 Figure 2.57. S4:20:0( 4,5,5,6) J - resolved ................................ ................................ ................................ .. 134 Figure 2.58. S4:20:0( 4,5,5,6) ROESY ................................ ................................ ................................ ..... 135 Figure 2.59. S4:20:0( 3,5,6,6) 1 H NMR ................................ ................................ ................................ .... 138 Figure 2.60. S4:20:0( 3,5,6,6) 13 C NMR ................................ ................................ ................................ ... 139 Figure 2.61. S4:20:0( 3,5,6,6) 1 H - 1 H gCOSY ................................ ................................ ........................... 140 Figure 2.62. S4:20:0( 3,5,6,6) gHSQC ................................ ................................ ................................ ...... 141 Figure 2.63. S4:20:0( 3,5,6, 6) gHMBC ................................ ................................ ................................ .... 142 Figure 2.64. S4:20:0( 3,5,6,6) J - resolved ................................ ................................ ................................ .. 143 Figure 2.65. S4:20:0( 3,5,6,6) ROESY ................................ ................................ ................................ ..... 144 Figure 2.66. S5:20:0( 2,6,6,6 ) 1 H NMR ................................ ................................ ................................ .... 147 Figure 2.67. S5:20:0( 2,6,6,6) 13 C NMR ................................ ................................ ................................ ... 148 Figure 2.68 . S5:20:0( 2,6,6,6) 1 H - 1 H gCOSY ................................ ................................ ........................... 149 Figure 2.69. S5:20:0( 2,6,6,6) gHSQC ................................ ................................ ................................ ...... 150 Figure 2.70. S5:20:0( 2,6,6,6) gHMBC ................................ ................................ ................................ .... 151 Figure 2.71. S5: 20:0( 2,6,6,6) J - resolved ................................ ................................ ................................ .. 152 xviii Figure 2.72. S5:20:0( 2,6,6,6) ROESY ................................ ................................ ................................ ..... 153 Figure 2.73. S5:22:1( 2,5,5,5,5 T ) 1 H NMR ................................ ................................ ............................... 156 Figure 2.74. S5:22:1( 2,5,5,5,5 T ) 13 C NMR ................................ ................................ .............................. 157 Figure 2.75. S5:22:1( 2,5,5,5,5 T ) 1 H - 1 H gCOSY ................................ ................................ ...................... 158 Figure 2.76. S5:22:1( 2,5,5,5,5 T ) gHSQC ................................ ................................ ................................ . 159 Figure 2.77. S5:22:1( 2,5,5,5,5 T ) gHMBC ................................ ................................ ................................ 160 Figure 2.78. S5:22:1( 2,5,5,5,5 T ) J - resolved ................................ ................................ ............................. 161 Figure 2.79. S5:21:0( 2,2,5,6,6) 1 H NMR ................................ ................................ ................................ . 164 Figure 2.80. S5:21:0( 2,2,5,6,6) 13 C NMR ................................ ................................ ................................ 165 Figure 2.81. S5:21:0( 2,2,5,6,6) 1 H - 1 H gCOSY ................................ ................................ ........................ 166 Figure 2.82. S5:21:0( 2,2,5,6,6) g HSQC ................................ ................................ ................................ ... 167 Figure 2.83. S5:21:0( 2,2,5,6,6) gHMBC ................................ ................................ ................................ . 168 Figure 2.84. S5:21:0( 2,2,5,6,6) J - resolved ................................ ................................ ............................... 169 Figure 2.85. S5:21:0( 2,2,5,6,6) ROESY ................................ ................................ ................................ .. 170 Figure 2.86. S4:21:0( 5,5,5,6) 1 H NMR ................................ ................................ ................................ .... 173 Figure 2.87 . S4:21:0( 5,5,5,6) 13 C NMR ................................ ................................ ................................ ... 174 Figure 2.88. S4:21:0( 5,5,5,6) 1 H - 1 H gCOSY ................................ ................................ ........................... 175 Figure 2.89. S4:21:0( 5,5,5,6) gHSQC ................................ ................................ ................................ ...... 176 Figure 2.90. S4:21:0( 5,5,5,6) gHM BC ................................ ................................ ................................ .... 177 Figure 2.91. S4:21:0( 5,5,5,6) J - resolved ................................ ................................ ................................ .. 178 Figure 2.92. S4:21:0( 5,5,5,6) ROESY ................................ ................................ ................................ ..... 179 Figure 2.93. S5:25:4( 2,5,5,5,8 P ) 1 H NMR ................................ ................................ ................................ 182 Figure 2.94. S5:25:4( 2,5,5,5,8 P ) 1 3 C NMR ................................ ................................ .............................. 183 Figure 2.95. S5: 25:4( 2,5,5,5,8 P ) 1 H - 1 H gCOSY ................................ ................................ ....................... 184 Figure 2.96. S5:25:4( 2,5,5,5,8 P ) gHSQC ................................ ................................ ................................ . 185 Figure 2.97. S5:25:4( 2,5,5,5,8 P ) gHMBC ................................ ................................ ................................ 186 Figure 2.98. S5:25:4( 2,5,5,5,8 P ) J - resolved ................................ ................................ ............................. 187 xix Figure 2.99. S5:25:4( 2,5,5,5,8 P ) ROESY ................................ ................................ ................................ . 188 Figure 2.100. S5:22:0( 2,2,6,6,6) 1 H NMR ................................ ................................ ............................... 191 Figure 2.101. S5:22:0( 2,2,6,6,6) 13 C NM R ................................ ................................ .............................. 192 Figure 2.102. S5:22:0( 2,2,6, 6,6) 1 H - 1 H gCOSY ................................ ................................ ...................... 193 Figure 2.103. S5:22:0( 2,2,6,6,6) gHSQC ................................ ................................ ................................ . 194 Figure 2.104. S5:22:0( 2,2,6,6,6) gHMBC ................................ ................................ ............................... 195 Figure 2.105. S5:22:0( 2,2,6,6,6) J - resolved ................................ ................................ ............................. 196 Fig ure 2.106. S5:22:0( 2,2,6,6,6) ROESY ................................ ................................ ................................ 197 Figure 2.107. S4:22:0( 5,5,6,6) 1 H NMR ................................ ................................ ................................ .. 200 Figure 2.108. S4:22:0( 5,5,6,6) 13 C NMR ................................ ................................ ................................ . 201 Figure 2.109. S4:22:0( 5,5,6,6) 1 H - 1 H gCOSY ................................ ................................ ......................... 202 Figure 2.110. S4:22:0( 5,5,6,6) gHSQC ................................ ................................ ................................ .... 203 Figure 2.111. S4:22:0( 5,5,6,6) gHMBC ................................ ................................ ................................ .. 204 Figure 2.112. S4:22:0( 5,5,6,6) J - resolved ................................ ................................ ................................ 205 Figure 2.113. S4:22: 0( 5,5,6,6) ROESY ................................ ................................ ................................ ... 206 Figure 2.114. S5:23:0( 2,5 ,5,5,6) 1 H NMR ................................ ................................ ............................... 209 Figure 2.115. S5:23:0( 2,5,5,5,6) 13 C NMR ................................ ................................ .............................. 210 Figure 2.116. S5:23:0( 2,5,5,5,6) 1 H - 1 H gCOSY ................................ ................................ ...................... 211 Figure 2.117. S5:23:0( 2,5,5,5,6) gHSQC ................................ ................................ ................................ . 212 Figure 2.118. S5:23:0( 2,5,5,5,6) gHMBC ................................ ................................ ............................... 213 Figure 2.119. S5:23:0( 2,5,5,5,6) J - resolved ................................ ................................ ............................. 214 Figure 2.120. S5:23:0( 2,5,5 ,5,6) ROESY ................................ ................................ ................................ 215 Figure 2.121. S4:23:0( 5,6,6,6) 1 H NMR ................................ ................................ ................................ .. 218 Figure 2.122. S4:23:0( 5,6,6,6) 13 C NMR ................................ ................................ ................................ . 21 9 Figure 2.123. S4:23:0( 5,6,6,6) 1 H - 1 H gCOSY ................................ ................................ ......................... 220 Figure 2.124. S4:23:0( 5,6,6,6) gHSQC ................................ ................................ ................................ .... 221 Fig ure 2.125. S4:23:0( 5,6,6,6) gHMBC ................................ ................................ ................................ .. 222 xx Figure 2.126. S4:23: 0( 5,6,6,6) J - resolved ................................ ................................ ................................ 223 Figure 2.127. S4:23:0( 5,6,6,6) ROESY ................................ ................................ ................................ ... 224 Figure 2.128. S5:24:0( 2,5,5,6,6) 1 H NMR ................................ ................................ ............................... 227 Figure 2.129. S5:24:0( 2,5,5,6,6) 13 C NMR ................................ ................................ .............................. 228 Figure 2.130. S5:24:0( 2,5,5,6,6) 1 H - 1 H gCOSY ................................ ................................ ...................... 229 Figure 2.131. S5:24:0( 2,5,5,6,6) gHSQC ................................ ................................ ................................ . 230 Figure 2.132. S5:24:0( 2,5,5,6,6) gHMBC ................................ ................................ ............................... 231 Figure 2.133. S5:24:0( 2,5,5,6,6) J - r esolved ................................ ................................ ............................. 232 Figure 2.134. S5:24:0( 2,5,5,6,6) ROESY ................................ ................................ ................................ 233 Figure 2.135. S6:25:0( 2,2,5,5,5,6) 1 H NMR ................................ ................................ ............................ 236 Figure 2.136. S6:25:0( 2,2,5,5,5,6) 13 C NMR ................................ ................................ ........................... 237 Figure 2.137. S6:25:0( 2,2,5,5,5,6) 1 H - 1 H gCOSY ................................ ................................ ................... 238 Figure 2.138. S6:25:0( 2,2,5,5,5,6) gHSQC ................................ ................................ .............................. 239 Figure 2.13 9. S6:25:0( 2,2,5,5,5,6) gHMBC ................................ ................................ ............................ 240 Figure 2.140. S6:25:0( 2,2,5,5,5,6) J - r esolved ................................ ................................ .......................... 241 Figure 2.141. S6:25:0( 2,2,5,5,5,6) ROESY ................................ ................................ ............................. 242 Figure 3.1. UHPLC/ESI/MS base peak intensity chromatogram of S. quitoense extract generated in negative - ion mode. Annotated acylsugars (formate adducts) detec ted using CID Function 1 = 6 V (1 0x magnification). Abbreviations of sugar groups are as follows: myo - inositol (I), N - acetylglucosaminyl (NAG), glucopyranosyl (G), and xyl opyranosyl (X). One example for abbreviation nomenclature is I:24:0(2,2,10,10), where ore is myo - groups, the numeral the numbers in parentheses descr ibe the number of carbon atoms in each acyl group. ................................ ................................ ............................ 252 Figure 3.2. CID mass spectra of acylsugar at t r = 73.90 mins (later annotated I4:24:0(2,2,10,10) according to NMR results) using 10 - 60 V MS E ramp (0.2 s acquisition times). (A) Negative ion mode. (B) Positive ion mode. ................................ ................................ ................................ ................................ .................. 254 Figure 3.3. Negative ion mode MS/MS product ion spectra of tri - acylated myo - inositol I3:24:0(2,10,12) from S. quitoense and triacylated glucose G3:19:0(4,5,10) from Solanum p ennellii LA0716 (all spectra generated with a linear 5 - 60 V MS E collisio n energy ramp with 0.5 s acquisition time) (A) products of I3:24:0(2,10,12) [M+HCOO] - (spectrum magnified 15x over t he range m/z 51 - 593) (B) products of G3:19:0(4,5,10) [M+HCOO] - . (C) products of I3:24:0(2,10,12) [M - H] - . (D) products of G3:19:0(4,5,10) [M - H] - . ................................ ................................ ................................ ................................ ...................... 257 xxi Figure 3.4. MS/MS product ion spectra of [M +HCOO] - and [M+NH 4 ] + of glycosylated acylinositols in an extract of S. quitoense using a linear 5 - 60 V collision e nergy ramp with 0.5 s acqu isition time. (A) ESI( - ) MS/MS product ion spectrum for m/z 778 ([M+HCOO] - ) for acylsugar NAG - I3:22:0(2,10,10). (B) ESI(+) MS/MS product ion spectrum for m/z 751 ([M+NH 4 ] + ) for acylsugar NAG - I3:22:0(2,10,10). (C) ESI( - ) MS/MS prod uct ion spectrum for m/z 7 37 ([M+HCOO] - ) for acylsugar G - I3:22:0(2,10,10). (D) ESI(+) MS/MS product ion spectrum for m/z 710 ([M+NH 4 ] + ) for a cylsugar G - I3:22:0(2,10,10). (E) ESI( - ) MS/MS product ion spectrum for m/z 707 ([M+HCOO] - ) for acylsugar X - I3:22:0( 2,10,10). (F) ESI(+) MS/MS product ion spectrum for m/z 680 ([M+NH 4 ] + ) for acylsugar X - I3:22:0(2,10,10). ................................ ... 259 Figure 3.5. Chemical structures of core glycosylated myo - inositol metabolites from S. quitoense determined by NMR spectroscopy. R 1 , R 2 , R 3 = acylation at that position. R 4 = H or acylation for myo - inositol monosaccharide. Ac ylations are listed in Table 3.1. ................................ ................................ ................. 260 Figure 3.6. 1 H NMR spectra of (A) I4:24:0 (2,2,10,10) (highlighted regions are - CH 2 positions that were selectively excited for 1D - TOCSY transfer experiments, those spectra are located in Appendix Figures 3.92 - 3.95) and (B) I4:26:0(2,2,10,12) acylinositols. 1D - TOCSY transfer curves for excitation of 2 - - CH 2 and 3 - - CH 2 acyl groups o f (C) I4:24:0(2,2,10,10) and (D) I4:26:0(2,2,10,12) acylinositols. ................ 267 Figure 3.7. S. quitoense images. (A) Picture at ~5 weeks (B) You ng leaflet. (C) Type I - like trichomes [26, 27] on petiol e of a young leaflet (approximately 3 - 6 mm in length). ................................ ...................... 272 Figure 3.8. NA G - I3:22:0(2,10,10) MS/MS product ion spectra, precursor ion [M+NH 4 ] + (top) & [M+H] + (bottom) ................................ ................................ ................................ ................................ .................... 284 Figure 3.9. NAG - I3:22:0(2,10,10) MS/MS product ion spectra, precursor ion [M+HCOO] - (top) & [M - H] - (bottom) ................................ ................................ ................................ ................................ .................... 285 Figure 3.10. G - I3:22:0(2,10,10) MS/MS prod uct ion spectra, precursor ion [M+NH 4 ] + (top) & [M+H] + (bottom) ................................ ................................ ................................ ................................ .................... 286 Figure 3.11. G - I3:22:0(2,10,10) MS/MS product ion spectra, precursor ion [M+HCOO] - (top) & [M - H] - (bottom) ................................ ................................ ................................ ................................ .................... 287 Figure 3.12. X - I3:22:0(2,10,10) MS/MS product ion spectra, precursor ion [M+NH 4 ] + (top) & [M+H] + (bottom) ................................ ................................ ................................ ................................ .................... 288 Figure 3.13. X - I3:22:0(2,10 ,10) MS /MS product ion spectra, precursor ion [M+HCOO] - (top) & [M - H] - (bottom) ................................ ................................ ................................ ................................ .................... 289 Figure 3.14. NAG - I3:24:0(2,10,12) MS/MS product ion spectra, precursor ion [M+NH 4 ] + (top) & [M+H] + (bottom) ................................ ................................ ................................ ................................ .................... 290 Figure 3.15. NAG - I3:24:0(2,10,12) MS/MS product ion spectra, precursor ion [M+HCOO] - (top) & [M - H] - (bottom) ................................ ................................ ................................ ................................ .............. 291 Figure 3.16 . G - I3:24: 0(2,10,12) MS/MS product ion spectra, precursor ion [M+NH 4 ] + (top) & [M+H] + (bottom) ................................ ................................ ................................ ................................ .................... 292 Figure 3.17. G - I3:24:0(2,10,12) MS/MS product ion spectra, precursor ion [M+HCOO] - (top) & [M - H] - (botto m) ................................ ................................ ................................ ................................ .................... 293 Figure 3.18. I3:22:0(2,10,10) MS/MS product ion spectra, precursor ion [M+NH 4 ] + (top) & [M+H] + (bottom) ................................ ................................ ................................ ................................ .................... 294 xxii Fi gure 3.19. I3 :22:0(2,10,10) MS/MS product ion spectra, precursor ion [M+HCOO] - (top, magnified 10x over m/z 51 - 566) & [M - H] - (bottom) ................................ ................................ ................................ ........ 295 Figure 3.20. I4:24:0(2,2,10,10) MS/MS product ion spec tra, precursor ion [M+NH 4 ] + (top) & [M+H] + (bottom) ................................ ................................ ................................ ................................ .................... 296 Figure 3.21. I4:24:0(2,2,10,10) MS/MS product ion spectra, precursor ion [M+HCOO] - (top, magnified 60x over m/z 51 - 608) & [M - H] - (bottom ) ................................ ................................ ................................ . 297 Figure 3.22. I3:24:0(2,10,12) MS/MS product ion spectra, precursor ion [M+NH 4 ] + (top) & [M+H] + (bottom) ................................ ................................ ................................ ................................ .................... 298 Figure 3.23. I 3:24:0(2,10,12) M S/MS product ion spectra, precursor ion [M+HCOO] - (top, magnified 14x over m/z 51 - 594) & [M - H] - (bottom) ................................ ................................ ................................ ........ 299 Figure 3.24. I4:26:0(2,2,10,12) MS/MS product ion spectra, precurs or ion [M+NH 4 ] + (t op) & [M+H] + (bottom) ................................ ................................ ................................ ................................ .................... 300 Figure 3.25. I4:26:0(2,2,10,12) MS/MS product ion spectra, precursor ion [M+HCOO] - (top, magnified 100x over m/z 51 - 590) & [M - H] - (bottom) ................................ ................................ ............................... 301 Figure 3.26. NAG - I 3:22:0( 2,10,10) 1 H NMR ................................ ................................ .......................... 315 Figure 3.27 . NAG - I3:22:0( 2,10,10) 13 C NMR ................................ ................................ ......................... 316 Figure 3.28. NAG - I3:22:0(2,10,10) 1 H - 1 H gCOSY ................................ ................................ ................. 317 Figure 3.29. NAG - I3:22:0( 2,10,10) gHSQC ................................ ................................ ........................... 318 Figure 3.30. NAG - I3:22:0( 2,10,10) gH MBC ................................ ................................ .......................... 319 Figure 3.31. NAG - I3:22:0( 2,10,10) J - resolved ................................ ................................ ........................ 320 Figure 3.32. NAG - I3:22:0( 2,10,10) ROESY ................................ ................................ ........................... 321 Figure 3.33. G - I3:22:0(2,10,10) 1 H NMR ................................ ................................ ................................ 324 Figure 3.34. G - I3:22:0(2,10,10) 13 C NMR ................................ ................................ ............................... 325 Figure 3. 35. G - I3:22:0(2,10,10) 1 H - 1 H gCOSY ................................ ................................ ....................... 326 Figure 3.36 . G - I3:22:0(2,10,10) gHSQC ................................ ................................ ................................ . 327 Figure 3.37 . G - I3: 22:0(2,10,10) gHMBC ................................ ................................ ................................ 328 Figure 3.38 . G - I3:22:0(2,10,10) J - resolved ................................ ................................ ............................. 329 Figure 3.39 . G - I3:22:0(2,10,10) ROESY ................................ ................................ ................................ . 330 Figure 3.40. X - I3:22:0(2,10,10) 1 H NMR ................................ ................................ ................................ 333 Figure 3.41. X - I3:22:0(2,10,10) 13 C NMR ................................ ................................ ............................... 334 Figure 3.42. X - I3:22:0(2,10, 10) 1 H - 1 H gCOSY ................................ ................................ ....................... 335 xxiii Figure 3.43. X - I3:22:0(2,10,10) gHSQC ................................ ................................ ................................ . 336 Figure 3.44. X - I3:22:0(2,10,10 ) gHMBC ................................ ................................ ................................ 337 Figure 3.45. X - I3:22:0(2,10,10) J - resolved ................................ ................................ ............................. 338 Figure 3.46. X - I3:22:0(2,10,10) ROESY ................................ ................................ ................................ . 33 9 Figu re 3.47. NAG - I3:24:0(2,10,12) 1 H NMR ................................ ................................ .......................... 342 Figure 3.48. NAG - I3:24:0(2,10,12) 13 C NMR ................................ ................................ ......................... 343 Figure 3.49. NAG - I3:24:0(2,10,12) 1 H - 1 H gCO SY ................................ ................................ ................. 344 Figure 3.50. NAG - I3:24:0(2,10,12) gHSQC ................................ ................................ ........................... 345 Figure 3.51. NAG - I3:24:0(2,10,12) coupled - gHSQC ................................ ................................ .............. 346 Figure 3.52. NAG - I3:24:0(2,10,12) gHMBC ................................ ................................ .......................... 347 Figure 3.53. NAG - I3:24:0(2,10,12) J - resolved ................................ ................................ ........................ 348 F igure 3.54 . NAG - I3:24:0(2,10,12) ROESY ................................ ................................ ........................... 349 Figure 3.55. G - I3:24:0(2,10,12) 1 H NMR ................................ ................................ ................................ 352 Figure 3.56. G - I3:24:0(2,10,12) 13 C NMR ................................ ................................ ............................... 353 Figure 3.57. G - I3:24:0(2,10,12) 1 H - 1 H gCOSY ................................ ................................ ....................... 354 Figure 3.58. G - I3:24:0(2,10,12 ) gHSQC ................................ ................................ ................................ . 355 Figure 3.59. G - I3:24:0(2,10,12) coupled - gHSQC ................................ ................................ ................... 356 Figure 3.60. G - I3:24:0(2,10,12) gHMBC ................................ ................................ ................................ 357 Figure 3.61. G - I3:24:0(2,10,12) J - resolved ................................ ................................ ............................. 358 Figure 3.62. G - I3:24:0(2,10,12) ROESY ................................ ................................ ................................ . 359 Figure 3.63. I3:22:0(2,10,10) 1 H NMR ................................ ................................ ................................ .... 361 Figure 3.64. I3:22:0(2,10,10) 13 C NMR ................................ ................................ ................................ ... 362 Figure 3.65. I3:22:0(2,10,10) 1 H - 1 H gCOSY ................................ ................................ ........................... 363 Figure 3.66 . I3:22:0(2,10,10) gHSQC ................................ ................................ ................................ ..... 364 Figure 3.67. I3:22:0(2,10,10) gHMBC ................................ ................................ ................................ .... 365 Figure 3.68. I3:22:0(2,10,10) J - resolved ................................ ................................ ................................ . 366 Figure 3.69. I3:22:0(2,10,10) ROESY ................................ ................................ ................................ ..... 367 xxiv Figure 3.70. I4:24:0(2,2, 10,10) 1 H NMR ................................ ................................ ................................ . 369 Figure 3.71. I4:24:0(2,2,10,10) 13 C NMR ................................ ................................ ................................ 370 Figure 3.72. I4:24:0(2,2,10,10) 1 H - 1 H gCOSY ................................ ................................ ........................ 371 Figure 3.73. I4:24:0( 2,2,10,10) gHSQC ................................ ................................ ................................ .. 372 Figure 3.74. I4:24:0(2,2,10,10) gHMBC ................................ ................................ ................................ . 373 Figure 3.75. I4:24:0(2,2,10,10) J - resolved ................................ ................................ .............................. 374 Figure 3.76. I4:24:0(2,2,10,10) ROESY ................................ ................................ ................................ .. 375 Figure 3.77. I4:24:0(2,2,10,10) TOCSY ................................ ................................ ................................ .. 376 Fig ure 3.78. I3:24:0(2,10,12) 1 H NMR ................................ ................................ ................................ .... 378 Figure 3.79. I3:24:0(2,10,12) 13 C NMR ................................ ................................ ................................ ... 379 Figure 3.80. I3:24:0(2,10,12) 1 H - 1 H gCOSY ................................ ................................ ........................... 380 Figure 3.81. I3:24:0(2,10,12) gHSQC ................................ ................................ ................................ ..... 381 Figure 3.82. I3:24 :0(2,10,12) gHMBC ................................ ................................ ................................ .... 382 Figure 3.83. I3:24:0(2,10,12) J - resolved ................................ ................................ ................................ . 383 Figure 3.84. I3:24:0(2,10,12) ROESY ................................ ................................ ................................ ..... 384 Figure 3.85. I4:26:0(2,2,10,12) 1 H N MR ................................ ................................ ................................ . 386 Figure 3.86. I4:26:0(2,2,10,12) 13 C NMR ................................ ................................ ................................ 387 Figure 3.87. I4:26:0(2,2,10,12) 1 H - 1 H gCOSY ................................ ................................ ........................ 388 Figure 3.88. I4:26:0(2,2,10,12) gHSQC ................................ ................................ ................................ .. 389 Figure 3.89. I4:26:0(2,2,10,12) gHMB C ................................ ................................ ................................ . 390 Figure 3.90. I 4:26:0(2,2,10,12) J - resolved ................................ ................................ .............................. 391 Figure 3.91. I4:26:0(2,2,10,12) ROESY ................................ ................................ ................................ .. 392 Figure 3.92. I4:24:0(2,2,10,10) 1D - TOCSY transfer spect ra 2 - - CH 2 excitation at 1.62 ppm (generated using Varian Inova 600 MHz spectrometer). ................................ ................................ ............................ 393 Figure 3.93. I4 :24:0(2,2,10,10) 1D - TOCSY transfe r spectra 3 - - CH 2 excitation at 1.48 ppm (generated us ing Varian Inova 600 MHz spectrometer). ................................ ................................ ............................ 394 Figure 3.94. I4:26:0(2,2,10,12) 1D - TOCSY transfer spectra 2 - - CH 2 excitation at 1.62 ppm (generated using Varian Inova 600 MHz spectrometer). ................................ ................................ ............................ 395 xxv Figure 3.95. I4: 26:0(2,2,10,12) 1D - TOCSY transfer spectra 3 - - CH 2 excitation at 1.48 ppm (generated using V arian Inova 600 MHz spectrometer). ................................ ................................ ............................ 396 Figure 4.1. LC/MS chromatograms of S. sinuata extract containing acylsucroses plus 1.0 µM each GS - n - alkyl RI standards analyzed by ESI+ mode (column SAE - A, te mperature 50°C, aqueous 10 mM ammonium formate pH 2.8 and linear gradient 1 - 100% acet onitrile, slope 1% B min - 1 ). (A) BPI chromatogram of S. sinuata extract containing acylsucroses . (B) Extracted ion chromatogram (XIC) of common fragment ion m/z 162. 022 ( C 5 H 8 NO 3 S + ) generated at 20 V collision potential. Illustration of GS - n - alkyl standard chemical structure and proposed fragmentation positions. ................................ ................................ ....... 414 Figure 4.2. GS - n - alkyl HPLC retention ti mes on a Supelco Ascentis Express C18 column using aqueous ammonium formate/acetonitrile gr adient as a function of alk yl chain length (RI value) (n=3 replicates). ................................ ................................ ................................ ................................ ................................ .. 415 Figure 4.3. RT drift and application of RI correction applied to a group of S4:2 0:0 S. sinuata acylsucrose isomers ([ M+NH 4 ] + , green) with overlaid GS - n - alkyl standards ([M+H] + , blue) using column SAE - A. Peaks #2 - 4 have structural identifications based on NMR spectra. Dates when an alyses were performed are displayed in year, month, day format. (A) 20180203, (B) 20180209, (C) 20180303, (D) 20180322. ..... 416 Figure 4.4. Percent difference comparisons of RT and RI values appli ed to 16 S. sinuata acylsucroses (each marker represents one of the acylsucrose metaboli tes) relative to the first set of measurements (column SAE - A, temperature 50°C, aqueous 10 mM ammonium formate pH 2.8 and linear gradient 1 - 100% acetonitrile, slope 1% B min - 1 ). Results are plotted against RT and RI value for visual purposes. Dates of anal yses are indicated in the legend using year, month, day format. (A) RT percent differences relative to the first analysis. (B) RI percent differences relative to the fi rst analysis. ................................ ............... 417 Figure 4.5. LC system and column dependence assessment using SAE columns. The group of S4:20:0 S. sinuata acylsucrose isomers peaks #1 - 4 is shown ([M+NH 4 ] + , green) with overl aid GS - 15 and GS - 16 ([M+H] + , blue). (A) Column SAE - A, measured using Shimadzu LC system. (B) C olumn SAE - A measured using the Waters LC system. (C) Column SAE - B measured using Shimadzu LC system. ..................... 418 Figure 4.6. LC system and SAE columns RT and RI value percent difference comparison whe n applied to 16 acylsucroses from S. sinuata (percent differences are r elative to column SAE - A measured using Shimadzu LC system). Columns, LC systems and dates of analyses are indicated in the legend. (A) RT percent difference (B) RI percent difference ................................ ................................ ............................ 419 Figure 4.7. Column performance comparison. The group of S. sinuata S4:20:0 acylsucrose isomers is shown ([M+NH 4 ] + , green) with overlaid standards GS - 14, - 15, and - 16 ([M+H] + , blue). (A) Analy sis usi ng column SAE - A. (B) Analysis using heavily used column SAE - C. ................................ .......................... 421 Figure 4.8. Column per formance comparison using XICs for GS - n - alkyl common fragment ion m/z 162.022 (generated at 20 V collision energy). GS - 19 to GS - 24 are b oxed in red. (A) Column SAE - A. (B) Column SAE - C. (C) Plot of RT against RI value using each column. ................................ ..................... 422 Figure 4.9. Analysis of co lumn performance test pa rameters using standard GS - 22 [M+H] + as example. (A) Column SAE - A. (B) Column SAE - B. (C) Column SAE - C. ................................ .............................. 423 Figure 4.10. Predicted underivatized silanol group interactions. Approximate pKa value s were calculated using ChemDraw Professional Software (Version 16.0.1.4). ................................ ................................ ... 424 Figure 4.11. L C/MS chromatograms and retention of GS - n - alkyl standards using C18 columns from three leading manufa cturers (measured using Waters LC system). XICs for common fragment ion m/z 254.078 xxvi (C 10 H 12 N 3 O 5 - ) generated at 20 V collision energy in ESI - mode. (A) Colum n SAE - A. (B) Column AZEP. (C) Column WBEH. (D) Plot of RT against RI value using each column. ................................ .............. 425 Figure 4.12. Column manufacturer selectivity and RI dependence. The group of S4:20:0 S. sinuat a acylsucrose isomers is shown ([M+NH 4 ] + , green) with overlaid GS - 14, - 15, and - 16 ([M+H] + , blue). (A ) Column SAE - A. (B) Column AZEP. (C) Column WBEH. ................................ ................................ ...... 427 Figure 4.13. RI dependence of the gro up of S4:20:0 S. sinuata acylsucrose is omers when chromatographic conditions are altered using column SAE - A. Filled markers are equivalent analyses measured on separate dates. Standard deviations were too small to display error bars (see Appendix Figures 4 .19, 4.21, 4.23) (A) Aqueous mobile pha se pH 2.5 - 4.0, column temp. 50°C, gradient slope 1% acetonitril e · min - 1 (B) Column temperature 30 - 60°C, aqueous mobile phase pH 2.8, gradient slope 1% acetonitrile · min - 1 (C) Linear gradient slope 1, 10/9, 5/4, 10/7, 5/ 3, and 2% acetonitrile · min - 1 , aqueous m obile phase pH 2.8, column temp. 50°C. ................................ ................................ ................................ ................................ ............... 429 Figure 4.14. GS - n - alkyl pH dependence comparisons (XICs of common fragment ion m/z 162.022 by ESI+) using colu mns: (A) SAE - A and (B) SAE - C. ................................ ................................ .................. 430 Fi gure 4.15. LC/MS of S. sinuata acylsucrose sample analyzed by ESI+ mode with methanol organic component. (A) BPI chromatogram of S. sinuata acylsucrose extr act sample. (B) XIC of common fragment ion m/z 162.022 (C 5 H 8 NO 3 S + ) generated at 20 V collision energ y. ( C) RT as a function of RI value. .... 433 Figure 4.16. Retention index dependence of the grou p of S4:20:0 S. sinuata acylsucrose isomers when chromatographic conditions were altered using column SAE - A and methanol was the organic modifier. Filled markers are equivalent analyses measured on separate dates. Standard deviations were too small to disp lay error bars (see Appendix Figures 4.25 - 27) (A) Aqueous mobile phase varied from pH 2.5 - 4.0, column tem p. 50°C, gradient slope 1% methanol · min - 1 (B) Column temperature varied from 30 - 60°C, aqueous mobile phase pH 2.8, gradient slope 1% methanol · min - 1 (C) Linear gradient slope 1, 10/9, 5/4, 10/7, 5/3, and 2% methanol · min - 1 , aqueous mobile phase pH 2.8, colum n temp. 50°C. .......................... 436 Figure 4.17. MS/MS product ion analysis of GS - n - alkyl standards. (A) Selected MS/MS product ion spectra of [M - H] - ions. (B) Ratio of common fragment [C 10 H 12 N 3 O 5 ] - = 254.078 (l abeled by an asterisk in Figure 4.17A) and [M - H] - ion abundances as a function of GS - n - alkyl chain length. ............................. 438 Figure 4.18. LC/MS chromatograms showing GS - n - alkyl pH dependence (acetonitrile or ganic mobile phase component, aqueous component 10 mM ammonium hydroxide adj usted with formic acid, column temperature 50°C, gradient slope 1% aceton trile·min - 1 , column SAE - A) (A) XICs of common fragment ion m/z 162.022 by ESI+. (B) RT as a function of RI value. ................................ ................................ .......... 442 Figure 4.19. RI value pH dependence of 16 S. sinuata acylsucrose metabolit es (acetonitrile organic mobile phase component, aqueous component 10 mM ammonium hydroxide adjusted with fo rmic acid, column temperature 50°C, gradient slope 1% acetontrile · min - 1 , column SAE - A) . ................................ ............... 443 Figure 4.20. LC/MS chromatograms showing GS - n - alkyl temperature dependence (acetonitrile organic mobi le phase component, aqueous component 10 mM ammonium hydroxide adjusted to pH 2.8 with formic acid, gradient slope 1% acetontrile·min - 1 , column SAE - A). (A) XICs of common fragment ion m/z 162.022 by E SI+. (B) RT as a function of RI value. ................................ ................................ ............................... 444 Figure 4.21. RI value column temperature dependence of 16 S. sinuata acylsucrose metab olites (ac etonitrile organic mobile phase component, aqueous component 10 mM ammonium hydroxide adjusted to pH 2.8 with formic acid, gradient slope 1% acetontrile · min - 1 , column SAE - A). ................................ ................. 445 xxvii Figure 4.22. GS - n - alkyl gradient slope dependence (acetonitrile organic mobile phase component, aqueous component 10 mM ammo nium hydroxide adjusted to pH 2.8 with formic acid, column temperature 50°C, column SAE - A). (A) XICs of common fragment ion m/z 254.078 by ESI - . (B) RT as a function of RI value. ................................ ................................ ................................ ................................ ................................ .. 446 Figure 4.2 3. RI value gradient dependence of 16 S. sinuata acylsucrose metabolites (acetonitrile organic mobile phase component, aqueous component 10 mM a mmonium hydroxide adjusted to pH 2.8 with formic acid, column temperature 50°C, gr adient slope 1, 10/9, 5/4, 10/ 7, 5/3, and 2% acetontrile · min - 1 , column SAE - A). ................................ ................................ ................................ ................................ .................... 447 Figure 4.24. GS - n - alkyl dep endence with altered chromatographic conditions using methanol organic component (column SAE - A). (A) RT as a func tion of RI value when pH is altered (aqueous component 10 mM ammonium hydroxide adjusted with formic acid, column temperature 50°C, g radient sl ope 1% acetontrile·min - 1 ). (B) RT as a function of RI value when column temperature is altered (aqueous component 10 mM ammonium hydroxide adjusted to pH 2.8 with formic acid, gradient slope 1% acetontrile·min - 1 ). (C) RT as a function of RI value when the LC gradient is altered (aqueous component 10 mM ammonium hydroxide adjusted to pH 2.8 with formic acid, column te mperature 50°C, gradient slope 1, 10/9, 5/4, 10/7, 5/3, and 2% acetontrile·min - 1 ). ................................ ................................ ........... 448 Figure 4.25. RI value pH dependence of 16 S. sinuata acylsucrose metabolites (methanol organic mobile phas e component, aqueous component 10 mM ammonium hydroxide adjusted with formic acid, column temperature 50°C, gradient slope 1% acetontrile · min - 1 , column SAE - A). ................................ ............... 449 Figure 4.26. RI value column tempera ture dependence of 16 S. sinuata acylsucrose metabolites (methanol organic mobile phase component, aqueous component 10 mM ammonium hy droxide adjusted to pH 2.8 with formic acid, gradient slope 1% acetontrile · min - 1 , column SAE - A). ................................ ................. 450 Figure 4.27. RI value gradient dependence of 16 S. sinuata acylsucrose metabolites (methano l organic mobile phase component, a queous component 10 mM ammonium hydroxide adjusted to pH 2.8 with formic acid, co lumn temperature 50°C, gradient slope 1, 10/9, 5/4, 10/7, 5/3, and 2% acetontrile · min - 1 , column SAE - A). ................................ ................................ ................................ ................................ .................... 451 Figure 4.28. ESI negative mode MS/MS spectra of GS - n - alkyl standards [M - H] - (generate d using Waters SONAR data acquisition platform, 10 - 40 V CID ramp, 0.5 s acquisition time, 10 Da bin widths, spe ctra processed using Waters MS E Data Viewer). ................................ ................................ ............................. 452 Figure 4.29. ESI negative mode M S/MS spectra of GS - n - alkyl standards [M - H] - (generated using Waters SONAR data acquisition platform, 10 - 40 V CID ramp, 0.5 s acquisition t ime, 10 Da bin widths, spectra processed using Waters MS E Data Viewer). ................................ ................................ ............................. 453 Figure 4.30. ESI negative mode MS/MS spectra of GS - n - alkyl standards [M - H] - (generated using Waters SONAR data acqui sition platform, 10 - 40 V CID ramp, 0 .5 s acquisition time, 10 Da bin widths, spectra processed using Waters MS E Data Vi ewer). ................................ ................................ ............................. 454 Figure 4.31. ESI positive mode MS/MS spectra of GS - n - alkyl sta ndards [M+H] + (generated using Waters SONAR data acquisition platform, 10 - 40 V CID ramp, 0.5 s acquisition time, 10 Da b in widths, spectra processed using Waters MS E Data Viewer). ................................ ................................ ............................. 455 Figure 4 .32. ESI positive mode MS/MS spectra of GS - n - alkyl standards [M+H] + (generated using Waters SONAR data acquisition platfo rm, 10 - 40 V CID ramp, 0.5 s acquisition time, 10 Da bin widths, spectra processed using Waters MS E Data Viewer). ................................ ................................ ............................. 456 xxviii Figure 4.33. ESI positive mode MS/MS spectra of GS - n - alkyl standards [M+H] + (generated using Waters SONAR data acquisition platform, 10 - 40 V CID ramp, 0.5 s acquisition time, 10 Da bin widths, spectra processed usi ng Waters MS E Data Viewer). ................................ ................................ ............................. 457 Figure 4.34. MS/MS prod uct ion analysis of GS - n - alkyl standards. Ratio of common fragment [C 5 H 8 NO 3 S] + = 162.022 and [M+H] + ion intensities as a function of GS - n - alkyl chain length (measured using 10 - 40 V CID ramp, 0.5 s acquisition time) ................................ ................................ ..................... 460 1 Ch apter 1: Introduction We depend on plants as sources of food, medicine and renewable energy. Plants produce a boundless array of smal l molecules that have great economic value and play critical roles in plant growth, development and response to stresses [1 ] population requires more productive/higher yielding crops that show greater tolerance to climate changes and r esistance to insects and pathogens [2] . In view of the importance of plants and the chemicals they synthesize, exploration of plant chemistries a nd genetic factors responsible for their biosynthesis could offer opportunities to genetically improve resistan ce and/or modify plants to become biosynthetic factories for production of useful natural products [3] . 1.1 Role of plant metabolomics for funct ional genomics and systems biology Metabolites are the small molecule (<1500 Da) output of biochemical processe s in living things, yet our understanding of the roles that individual metabolic enzymes play in the grand scheme of metabo lism is often unclear. The plant metabolome refers to the complete set of all small molecules produced within an organism and can be defined on all levels of complexity, such as the organism, tissues, cells or cellular compartments [4] . The metabolome is a representation of the chemical phenotype and is the resu lt of expression of genetic code, metabolic networks and molecular activitie s. In the p ost - genome era, plant metabolomics has been a rapidly growing field of research [1, 2, 5] . There is a need for a dvancement of strategies for exploration of plant chemist ries, to identify biological processes responsible for assembly of plant met abolites. Functional genomics and systems biology approaches aim to uncover the functions of genes, proteins, enzymes and t To ac hieve this, it is necessary to integrate genomics, transcriptomics, and othe tabolites remain undefined. As a result, metabolite annot ation and identification continue to present a bottleneck for determining pl ant metabolic gene functions. 2 1.2 The evolution of plant specialized metabolism and natural chemical diversity Plants have evolved to synthesize an array of structurally diverse o rganic compounds including carbohydrates, amino acids, nucleic acids, and li pids to serve essential cellular functions. These central metabolites are produced using fixed carbon derived from CO 2 , nit rogen - , sulfur - , and phosphorus - containing substances tak en up by roots, and other precursors. Selection for increased fitness has re sulted in plant lineages synthesizing distinct sets of metabolites that are not directly involved in primary metabolism. Th ese compounds were typically referred to as secondary met abolites, in large part due to their unknown functions and presence is some species, but not in others. However, their presence or absence in different species make it clear that plants have evolved the ability to synthesize these compounds to satisfy spec ific ecological niches. For example, they serve as attractants for pollinato rs, and defenses against herbivory and microbes, and they provide communication with other plant and non - plant species [6 - 8 ] . In view of their taxonomic restriction and significanc e, more researchers have begun referring to these The diversity of specialized metabolites and the observation of novel compounds in each species drive a true need for improved analytical strategies for defini ng the specialized metabolome in order to discover genes responsible for the ir biosynthesis [5] . Estimates suggest that perhaps 20% of plant genomes are involved in SM production and the number of SM s made by all plant species has been estimated to be in t he hundreds of thousands [9] . However, there are reasons to believe this to be a gross underestimation [1, 8] . It is difficult to estimate the number of SMs in plant species, largely because the acti vities of proteins and enzymes encoded by the majority of genes in any genome are not well understood [8, 10] . Furthermore, existing analytical tools have yet to provide deep information about plant chemistry because most specialized metabolites have yet t o be identified. From metabolite profiling, it is clear that each plant species can only make a small fraction of the total number o f SMs found in the plant kingdom [8] . Most SMs are only found in a limited range of species and tissue types and as a resul t, the genes responsible for their chemistry are not well understood [8] . 3 The evolution of so many different SM is considered the res ult of gene duplication [11] . Plants are organisms that exhibit extensive polyploidy (whole genome duplication, more than t wo sets of chromosomes). In fact, extensive genetic evidence suggests that whole genome duplication in angiosperms (flowering plants) has been common [12] , providi ng potential novel opportunities for evolutionary success. framework for diversification of gene function by altering r edundant copies of important o r essential genes [13] . When gene duplication leads to a situation where one copy of the gene retains the same function, and random mutation of the other copy alters gene product function (known as neofunctionalization), the r esult leads to variability in fitness across a population and serves as the foundation for natural selection. In particular , the presence of a mutant gene that leads to accumulation of SM s in individuals of the same species offers opportunities that reward SM production through advanta geous fitness, particularly with regard to resistance to herbivory of pathogen infection. In practical terms, this may involve a mutant biosynthetic enzyme that catalyzes chemistry similar to the original (central metabolic) enzyme but exhibits altered su bstrate preference and leads to formation of novel SMs. 1.3 Challenges of identifying plant s pecialized metabolites Plants produce far more metabolites than most other organisms and most have yet to be identified [1] . Identifi cation of plant SMs presents m any formidable challenges because unlike proteins and oligonucleotides which are constructed from a limited set of precursors, the building blocks of SMs are far more varied. Solvent extracts of plant tissues often present com plex mixtures with hundreds or thousands of metabolites that exhibit large ranges in relative concentrations (~10 6 ). Many S Ms are present in a limited range of species and tissue types [8, 14] , have numerous isomeric forms and few authentic standards are a vailable for comparison. Analy sis of plant SMs often requires multiple extraction, sample preparation and detection techniq ues because of the wide range of physical properties and chemical stability through extraction protocols. Furthermore, no single tech nology has yet been demonstrat ed that can detect, identify, and quantify all SMs from plants. Rather, a combination of anal ytical approaches is required. Multiple 4 chromatographic and mass spectrometric techniques are typically employed for comprehensive me tabolite profiling. However, t oo often mass spectrometry alone is not sufficient to identify SMs. De novo structural elucid ation of SMs usually requires time - consuming purifications for analysis by nuclear magnetic resonance (NMR) spectroscopy. In addition , analytical structural techni ques require extensive training and experience. As a result, only a small fraction of the tot al number of SMs have been fully identified. 1.4 Introduction to trichomes and acylsugars of the family Solanaceae 1.4.1 Trichome str uctures, functions and chemica l compositions Specialized cell types are prolific at synthesizing SMs. A large fraction of p lant species have glandular epidermal structures called trichomes [15] . These specialized uni - and multicellular epidermal protrusion s are found on the surfaces of leaves, stems and other aerial tissues of plants and may differ in size, shape and morpholog y [15] and chemical barriers against attacking herbivores [16, 17] , as well as guiding pollinators [7] , protecting against UV - B radiation [18] , leaf temperature stress a nd water loss [19] . The widespread abundance of glandular trichomes (GTs), found on ~30% of all vascular plants, suggests that diver se evolutionary forces shape s pecialized metabolism in plants [20] . Plants use GTs to synthesize, store and secrete large a mounts of SMs [15, 20, 21] . The storage compartment of GTs is usually located at the tip of a hair - like cellular structure, permittin g the plant to amass sticky an d/or toxic secretions in a compartment that is remote from remaining plant tissues. GT exudat es are harvested easily by scraping them from plant surfaces, gentle physical contact or dipping leaflets in to an organic solvent [2 2 - 24] , permitting detailed stu dy of their metabolites, as well as genes and proteins responsible for them. In addition, GTs exhibit novel biochemical pathways for specialized metabolism and have been suggested as targets for genetic engineering, with aims including optimization of tric home density, customization of essential oil production, and tuning of biocide activity to en hance crop protection [15, 21] . 5 Humans are unknowingly familiar with GTs and their metabolites, even when we are unaware of their sp ecific chemistries [15] . Trich ome - derived SMs contribute to the aroma and flavor properties of many herbs and spices. For i nstance, members of the plant family Lamiaceae synthesize several SMs in peltate GTs, including terpenoids menthol and limonene in pe ppermint ( Mentha × piperita ) and phenylpropenes eugenol and methylchavicol in basil ( Ocimum basilicum ) [25] . A simple inter net search of the word Cannabis sativa of Cannabaceae, which produce and ac cumulate psychoacti ve cannabinoids and other volatile terpenoid - derived metabolites that contribute to its strong odor [26] . Another Cannabaceae species, hops ( Humulus lupulus ), produces xanthohumol and related prenylf l avonoids that add bitterness and flav or to beer [27] . Tr ichomes are the source of the important medicinal chemical artemisinin of sweet wormwood ( Artemisia annu a ). Artemisinin - based therapies are now standard treatment for malaria and have saved millions of lives [28] . 1.4.2 Acylsugars in gla ndular trichomes of Solanaceae species The family Solanaceae serves as an agronomically important family of flowering plant s that includes many species. The family name is derived from its largest documented genus, Solanum , which includes important cultiva ted crops including tomato ( S. lycopersicum ), potato ( S. tuberosum ) and eggplant ( S. melongena ). Species of the Solanaceae display extensive diversity of trichome types and chemistry. taxonomic survey of tomato and its wild relatives found four major morpholo gical types of GTs: type I consist of a globular and multicellular base with a small and round glandular cell in the trichome tip (6 - 10 cells and 2 - 3 mm long); type IV are similar to type I, but shorter (0.2 - 0.4 mm) and with a flat, uni cellular base; type VI are thick and short, composed of two stalk cells and a head made of 4 secretory cells; type VII are very small (0.05 mm) with a head consisting of 4 - 8 cells [21, 29] . Members of the family Solanaceae synthesize and/or store diverse a rrays of SMs in GTs , including terpenes, phenylpropenes, flavonoids, alkaloids and fatty acid derivatives [15, 21] . Species within the genera Solanum , Petunia , and Nicotiana (among others) produce large quantities of structurally diverse sugar polyesters i n type I & IV trich omes [21] . These SMs are comprised of a carbohydrate core esterified by one 6 or more ester groups (Figure 1.1) at hydroxyl positions around the sugar ring. Various reports have described these SMs as sugar esters, O - acyl sugars or acylsug ars. Documented acy lsugars of the Solanaceae are mostly sucrose or glucose sugar groups, consisting of three to five branch ed and/or straight aliphatic C2 - C12 acyl chains (Table 1.1). These SMs play critical roles for physical and/or chemical defense, incl uding insect resist ance [30 - 32] , oviposition suppression [33] , protection against fungal pathogens [34] , and indirect defen se against herbivory [35] . Most descriptions of acylsugar composition in GTs report only a few acylsugar metabolites per species (se e Table 1 for refer ences). However, it is clear that that this is only the tip of the iceberg, as many species are prolific manufacturers of diverse acylsugars [36] , most of which have not been identified. I n part this is because acylsugars are complex mix tures that display multiple isomeric forms, structural redundancy of similar functional groups and unique acylation pattern s, making structural elucidation formidable [23] . For instance, sucrose - based acylsu gars (acylsucroses) demonstrate up to eight hydro xyl positions available for acylation (Figure 1.1). To the best or our knowledge, evidence of 25 different acylsugar acyl g roups has been reported in published papers and herein (acyl groups and their acrony ms are shown in Figure 1.1). From the known acyl group building blocks alone, the number of possible acylsucrose permutations exceeds 46 billion (Equation 1.1). Only ~100 a cylsugars have been identified from different species within Solanaceae (Table 1.1), but explorations of acylsugar diversity are stil l at an early stage. Since the Solanaceae display capabilities to synthesize precursors, mutations in the enzymes that asse mble acylsugars offer the prospect that an enormous number of acylsugars may exist a cross the Solanaceae, suggesting that diversity o f acylsugars has potential to far exceed the range of terpenoid SMs, documented in 1992 to exceed 22,000 [37] . Equation 1. 1 . Number of possible acylsucrose permutations 7 Figure 1 .1. Acylsugar ester groups and carbohydrates observed within Solanaceae species. C2: acetyl, C3: propionyl, nC4: butyryl, iC4: isobutyryl, nC 5: pentanoyl, aiC5: 2 - methylbutanoyl, iC5: 3 - methylbutanoyl, nC6: hexanoyl, aiC6: 3 - methylpentanoyl, iC6: 4 - methylpen tanoyl, nC7: heptanoyl, aiC7: 4 - methylhexanoyl, iC7: 5 - methylhexanoyl, nC8: octanoyl, aiC8: 5 - methylheptanoyl, iC8: 6 - methylheptanoyl, iC9 : 7 - methyloctanoyl, nC10: decanoyl, iC10: 8 - methylnonanoyl, nC12: dodecanoyl, iC12: 10 - methylundecanoyl, C3 M : malonyl, C5 S : senecioyl, C5 T : tiglyl, C8 P : phenylacetyl. 8 Table 1 .1. Examples of leaf surface acylsugars from Solanaceae sp ecies with identified positions of acylations and their residues Species Carbohydrates No. of acylsugars identified No . of acylations Positions of acylations Acyl residues (certain acyl residues may occur in multiple positions or show position specificity) References Solanum S. lycopersicum M82 sucrose 15 of 24 3,4 2,3,4,3' C2, iC4, iC5, aiC5, iC10, aiC11, nC12, iC12 Ghosh et al. 2013 [38] S. habrochaites LA 1392 and LA1362 sucrose 24 of 24 3,4,5 2,3,4,1',3',6' C2, iC4, iC5, aiC5, iC10, aiC11, nC12 , iC12 Ghosh et al. 2013 [38] S. habrochaites LA 1777 sucrose 22 of 24 3,4,5 2,3,4,1',3',6' C2, iC4, iC5, aiC5, iC10, aiC11, nC12, iC12 Ghosh et al. 2013 [38] S. lycopersicum M82 and IL - 3 sucrose 2 3,4 2,3,4,3' C2, iC5, aiC5 Schilmiller et al. 2010 [24] S. lycopersicum IL11 - 3 sucrose 1 3 3,4,3' iC5, iC5, nC12 Schilmiller et al. 2015 [39] S. pennellii LA0716 glucose, sucrose 2 G3, S3 2,3,4 iC4, iC5, aiC5, iC10 Schilmiller et al. 2016 [40] S. habrochaites (previously Lycopersicon hirsutum ) sucrose 1 4 2, 3,4,1' iC4, aiC5 King et al. 1990 [41] S. habrochaites LA 1777 (previously Lycopersicon typicum ) sucrose 6 3,4 2,3,4,6,3' C2, iC4, aiC5, iC5, iC10 King et al. 1993 [42] S. berthaultii sucrose 1 3 3,4,1' iC4, iC4, nC10 King et al. 1986 [43] S. berthaulti i sucrose 2 4 3,4,6,3' iC4, aiC5, nC10 King et al. 1987 [44] S. neocardenasii sucrose 3 4 2,3,4,3' C2, iC4, nC6, nC10 King and Singh 1988 [45] S. aethiopsicum glucose 3 3 2,3,4 C2, iC4, aiC5, iC9, iC10 King et al. 1988 [46] S . pennellii (previously Lyco persicon pennellii ) glucose N/A 3 2,3,4 iC4, iC5, iC5, aiC5, nC10, iC10 Burke et al. 1987 [47] S . pennellii (previously Lycopersicon pennellii ) glucose 1 3 2,3,4 iC4 Li and Steffens 2000 [48] S . lanceolatum myo - inositol glycosides N/A 1 2 C12 to C20, eve n numbered branched or straight chain (glucopyranosyl or xylopyranosyl at position 1) Herrera - Salgado et al. 2005 [49] S . quitoense myo - inositol glycosides 9 3,4 1,2,3,4 nC2, nC10, nC12 ( N - acetylglucosamine, glucopyranosyl or xylopyranosyl at position 4) Hurney et al. (Chapter 3) ( c on tinue d on following page) 9 Table 1.1. (continued) Species Carbohydrates No. of acylsugars identified No. of acylations Positions of acylations Acyl residues (certain acyl residues may occur in multiple positions or show posit ion specificity) References Petunia P. integrifolia sucrose 21 of 26 4,5 2,3,4,6,1' C3 M , iC4, aiC5, iC5, iC6, iC7, iC8 Liu et al. 2017 [50] P. exserta sucrose 26 of 26 4,5 2,3,4,6,1',6' C3 M , iC4, aiC5, iC5, iC6, iC7, iC8 Liu et al. 2017 [50] P. a xillaris sucrose 21 of 26 4,5 2,3,4,6,1' C3 M , iC4, aiC5, iC5, iC6, iC7, iC8 Liu et al. 2017 [50] P. nyctaginiflora sucrose 2 3 2,3,4 iC6, iC7 Singh et al. 2003 [51] P. nyctaginiflora sucrose 4 3,4 2,3,4 C2, iC6, iC7 Begum et al. 2004 [52] P. nyctaginifl ora sucrose 2 3,4 2,3,4,6' C2, iC7, iC8 Begum et al. 2005 [53] Nicotiana N. tabacum TI - 165 sucrose 1 4 2,3,4,6 C2, iC4, aiC5 Severson et al. 1985 [54] N. glutinosa sucrose N/A 3,4 2,3,4,6,3' C2, C3, iC4, aiC5, iC5, nC5, aiC6, iC6, nC6, aiC 7, iC7, nC7, iC8, nC8 Arrendale et al. 1990 [55] N. acuminata glucose, sucrose N/A G3,G4,G5,S5 1,2,3,4,6,3' C2, C3, iC4, aiC5, iC6, aiC6 and unsaturated aiC5 Matsuzaki et al. 1991 [56] N. pauciflora glucose, sucrose N/A G3,G4,G5,S5 1,2,3,4,6,3' C2, C3, i C4, aiC5, iC6, aiC6, iC7, aiC7 Matsuzaki et al. 1991 [56] Physalis P . philadelphica sucrose 5 2,3,4 2,3,1',3' iC4, aiC5, nC10 Zhang et al. 2016 [57] P . viscosa sucrose 1 2 3,4 C5 S , nC12 Ovenden et al. 2005 [58] Other Solanaceae Datura mete l glucose 2 2,3 1,2,3 nC6 Salpiglossis sinuata sucrose 16 4,5,6 2,3,4,1',3',6' C2, C3, iC4, aiC5, aiC6, C5 T , C8 P Hurney et al. (Chapter 2) Salpiglossis sinuata glucose, sucrose 2 G3,S5 2,3,4,6,3' C2, aiC6 Castillo et al. 1989 [59] *** Updated table [60 ] . N/A indicates no structures were identified, however ester substitution positions were assigned from mixtures using NMR spe ctroscopy. 10 1.4.3 Biosynthesis of acylsugars Acylation of oxygen - containing functional groups serves as a widespread mechanism fo r assembly of plant SMs [61] . Until recently, the functions of individual enzymes required for acylsugar assembly were largely undefined. The combination of metabolite profiling and identification [38, 50] with analysis of gene expression and biochemistry has guided the discovery of trichome - expressed enzymes required for acylsucrose assembly within the Solanaceae [39, 62 - 64] . Acylsugars are synthesized enzymatically by a group of enzymes called BAHD acyltransferases that catalyze the position selective acy lation of sugar cores by acy l - CoAs. This family was named according to the first letter of each of the first four biochemically characterized enzymes of this family (BEAT, AHCT, HCBT, and DAT) [65] . In tomato ( S. lycopersicum ), acylsucrose assembly begins with acylation of the pyran ose ring of sucrose at position 4 by the enzyme AcylSugar AcylTransferase 1 (Sl - ASAT1) using various acyl - CoAs, which serve as acyl group donors, to make monoacylsucroses [63] . ASAT nomenclature has been proposed that numbers th e enzymes in the order they participate in acylsucrose biosynthesis, with ASAT1 being the first. Sl - ASAT2 then adds a second acyl group (iC4, aiC5, iC5, iC6, iC10 or nC12, as defined in Figure 1.1) to position 3 of the Sl - ASAT1 product to make diacylsucros e substrates for Sl - ASAT3 [6 3] . Subsequent catalysis of acylation by Sl - ASAT3 transfers short (four to five carbon) branched acyl chains to the furanose ring at position 3 of diacylsucrose acceptors, producing triacylsucroses [39] , which can be further acylated by Sl - ASAT4 (previous ly Sl - AT2) which catalyzes the transfer of a C2 group to position 2 of tri - acylated sucrose acceptors to form tetra - acylated sucroses [62] . Int erestingly, Sl - ASAT1 showed substrate promiscuity in vitro , acylating sucrose at position 4 using acyl - CoAs with different acyl chain lengths (iC4, iC5, aiC5, nC10 and iC12). However, only iC4 and iC5 acyl groups have been observed at position 4 of acylsuc roses from tomato species [38] . Furthermore, ASAT2 was unable to use iC12 monoacylsucroses to produce diacylsucro ses, suggesting that longer chain monoacylsucroses represent dead - end products that are recycled by an acylsugar acylhydrolase to generate the starting metabolite sucrose [40] . In such manner, 11 acylsucroses may be enzymatically remodeled, leading to accumu lation of a narrower distribution of acylsucroses that do not include less - stable intermediate forms. The orders of action of ASAT enzyme ortho logs, which are those related by descent from a common ancestor, are not conserved between petunia and tomato [64 ] . In contrast to tomato acylsugar biosynthesis which begins by acylation at position 4 by Sl - ASAT1, tetraacylsucrose assembly begins in Petuni a axillaris with acylation of the pyranose ring of sucrose at position 2 by the petunia enzyme Pax - ASAT1 to produ ce a monoacylsucrose. Pax - ASAT2, which is the ortholog of Sl - ASAT1, does not use sucrose as acyl acceptor, but instead acylates the monoacylsuc rose generated by Pax - ASAT1 to add a second acyl group (aiC5) to position 4 to make diacylsucrose Pax - ASAT3 subst rates. Pax - ASAT3 catalyzes the transfer of acyl groups from various acyl - CoAs (aiC5, aiC6, iC6, iC7, iC8) to the pyranose ring at position 3 o f diacylsucrose acceptors, producing triacylsucroses which can be further acylated by Pax - ASAT4 which catalyzes t he transfer of acyl groups from acyl - CoAs (iC4, aiC5 and iC5) to position 6 of the tri - acylated sucrose acceptor to form a tetra - acylated sucro se. Notably, Pax - ASAT1 and Pax - ASAT4 catalyze the acylation of positions 2 and 6 of the pyranose ring of sucrose respectively, and no orthologs to tomato were discovered. Acylsucroses from P. axillaris also showed unusual malonate esters (C3 M ) on the furan ose ring position 1 that had not been observed previously [50] . The identity of the ASAT responsible for adding C3 M group remains unknown. These results demonstrate that while the position - selectivity of ASATs is often conserved across orthologs between t omato and petunia, there is evolutionary divergence in the carbohydrate acyl acceptors of ASATs that alters the o rder of acylsugar assembly in the biosynthetic pathway. Understanding the biochemical processes used by GTs for the production of acylsugar SMs offers the opportunity exploit these genes in industry and agriculture. The full potential of trichomes as proli fic chemical factories has not been fully exploited because our understanding of metabolic pathways and intermediates is still limited. Underst anding the genetic and biochemical factors that lead to substrate specificity and enzymatic promiscuity is centra l to plant improvement efforts [63] . Of particular note is the 12 observation that small changes in amino acid sequences in the acyltransferases c an alter acyl acceptor substrate preferences [63] . As genomic technologies continue to improve, the discovery of acyltransferase functions continues to be limited by our ability to profile and identify acylsugar SMs, particularly with regard to positions o f acylations. Thus, there is a need for advanced analytical strategies that can provide more detailed information about identities of pathway intermediates and products. 1.5 Analytical approaches for profiling and discovery of acylsugar metabolites 1.5.1 M ass spectrometry Metabolite analysis usually begins with mass spectrometry (MS), because MS allows for detection and quantification of a broad range of metabolites with speed, sensitivity and accuracy. Mass spectrometers separate and detect ionized gas - pha se analytes according to their mass - to - charge ratios ( m/z ). Accurate m/z measurements provide information about possible elemental formulas of analytes, which serves as one of the most useful pieces of information for metabolite annotation and discovery. C omprehensive metabolite profiling usually requires multiple MS - based tec hniques. For instance, gas chromatography (GC) coupled with electron ionization (EI) is typically employed for separation and detection of lower molecular weight, volatile, or non - labi le molecules. In contrast, liquid chromatography (LC) techniques coupled with atmospheric pressure chemical ionization (APCI) or electrospray ionization (ESI) are better suited for measuring more polar, labile and non - volatile metabolites without need for derivatization to improve volatility. 1.5.1.1 Gas Chromatography/Mass Sp ectrometry (GC/MS) GC/MS is the technique of choice for smaller volatile metabolites. The GC separation offers superior chromatographic resolution compared to other chromatographic tec hniques and was coupled with MS long before other inline separations [66 ] . GC/MS instruments often employ electron ionization (EI) at 70 eV electron energy to generate radical M + molecular ions. EI is regarded as a hard ionization technique because it gene rates fragment rich spectra that may lack unambiguous information about molecular mass. The 13 principal advantage of EI derives from reproducibility of mass spectra across instrument platforms which facilitates matching of spectra to spectral databases for i dentification of known metabolites. Analysis of intact acylsugar metab olites by GC/MS has proven to be impractical because acylsugars are present as non - volatile metabolites that usually have multiple free hydroxyl groups that require conversion to GC am enable derivatives (such as trimethylsilyl (TMS) ethers) for GC separati ons [23] . In addition, derivatized disaccharides such as acylsucroses fragment extensively under EI conditions, displaying cleavage at the glycosidic linkage [54] , but usually without molecular mass information in their EI mass spectra. However, GC/MS has proven to be well - suited for the analysis of total acyl composition of acylsugars by first converting acyl groups to alkyl esters such as ethyl [67] or butyl esters [68] by transesteri fication, followed by GC/MS analysis. This method is advantageous becaus e most GC separations resolve their branched alkyl ester isomers (e.g. using a (5% - phenyl) - methylpolysiloxane capillary column), and authentic standards can be purchased or easily made . Comparisons of mass spectra and retention times to authentic standards allow for comprehensive acyl group identification. Several reports have employed similar GC/MS approaches to characterize and quantify the number and types of ester groups that make u p acylsugars in the Solanaceae, including Solanum , Petunia and Nicotiana species [47, 49, 54 - 56, 68] . However, these approaches have a disadvantage in that they do not provide information about the number of acyl groups attached to each acylsugar or the ex act positions of acylation. 1.5.1.2 Ultra - High - Performance Liquid Chroma tography/Mass Spectrometry (UHPLC/MS) The combination of MS with LC separations provides the foundation for chemical analysis of non - volatile metabolites because intact metabolites can UHPLC columns typic ally employ high purity silica - based supports with <2 µm particle sizes and operate at high pressures up to 15,000 psi, providing optimal peak capacity, bandwidth, resolution, sensitiv ity and sample throughput. Some modern columns also utilize superficiall y porous particles that operate at similar pressures to more traditional HPLC systems (up to ~5,000 psi), while achieving chromatographic resolution similar to columns with smaller ful ly - porous particles. Reversed - phase hydrophobic stationary phases (C18 14 a nd C8) are widely used to separate analytes based on their hydrophobicity, and are often the first approach analysis of hydrophobic analytes. Phenyl - containing phases such as pentafluo rophenyl (F5) bonded phases are often used because they interact strongl electron systems) or basic analytes [69, 70] . Hydrophobic interaction liquid chromatography (HILIC) us es amino, diol, amide, a nd silica phases and offer s greater retention of more polar analytes [71 , 72] . The development of atmospheric pressure ionization techniques including atmospheric pressure chemical ionization (APCI) and electrospray ionization (ESI) enabled improved coupli ng of LC to MS and was a major breakthrough for the analysis of intact metaboli tes [73] . ESI is often the method of choice for advanced SM profiling because the ESI process does not require analyte evaporation at high temperatures that may degrade labile i onic compounds [74] . ESI generates ions by applying a high voltage to an LC flo w via a liquid capillary, creating a charged aerosol that is further de - solvated producing gas - phase ions. One or more LC solvents often contain weak acids/bases to facilitate a nalyte ionization, one example is ammonium formate (NH 4 HCOO). Ionization by ESI is dependent on the acidity, basicity, or ion affinities of the analyte, and can be performed in either positive - or negative - ion modes. ESI is considered a soft ionization tec hnique because it generates fewer fragment ions by comparison to ionization tec hniques such as EI. Protonated [M+H] + or deprotonated [M - H] - molecules are observed for analytes with ionizable functional groups. Molecules lacking acidic or basic groups may f orm pseudo - molecular adduct ions including [M+NH 4 ] + and [M+HCOO] - when ammonium formate is added to the LC mobile phase. Selective extraction of acylsugars from trichomes is easily achieved by dipping leaflets into a n extraction solvent such as acetonitril e: isopropanol: water (3:3:2, v/v/v; with 0.1 % formic acid to inhibit rearrangements of acyl group positions) for just 1 - 2 minutes. This solvent extracts a wide range of metabolites from GTs with minimal penetratio n into the surface tissue of leaflets [ 23] . Recent efforts to perform deep profiling of acylsugars using liquid chromatography have employed C18 superficially porous analytical columns with gradient elution using aqueous ammonium formate (10 mM adjusted to pH 2.8 with formic acid) and acetonit rile organic component [38, 50] . These methods often 15 feature long - shallow gradients with run times in excess of 100 mins per sample and are used to probe acylsugar complexity. Results of such analyses demonstrate tha t most acylsugar profiles present multip le isomeric forms for each elemental composition that may not be resolved using more rapid LC gradients. 1.5.1.3 Collision Induced Dissociation (CID) Annotation of acylsugars by LC/MS is facilitated by assigning mole cular masses and formulas, masses of ind ividual acyl groups, the number of acylations, and to the extent possible, the sugar core and positions of ester substituents. Because ESI generates mostly intact pseudomolecular ions from acylsugars under gentle ion ization conditions, their mass spectra l ack information about fragment ion masses, limiting the information available for structure annotation. To circumvent this issue, fragment ion information has been obtained using data - independent nonselective collisi on induced dissociation (CID) that subje cts all ions to collisional activation via collision with a target gas without precursor mass selection [75] . When an ion with high translational energy (up to ~ 100 eV) collides with a target gas such as argon, part translational energy is co nverted to internal vibrational energy which is rapidly (ps - ns) distributed throughout the molecule. Population of excited vibrational states causes a dramatic increase of rates of unimolecular ion dissociation, form ing fragment ions with masses that provi de information about structure. CID may be applied in a multiplexed parallel set of collision conditions, with steps in potential from 0 - 200 V requiring only a few milliseconds per function. At low collision potentia ls, acylsugars produce abundant formate ([M+HCOO] - ) and ammonium ([M+NH 4 ] + ) adduct ions when ammonium formate is used in the mobile phase, as well as [M+Cl] - , [M+NO 3 ] - , [M+Na] + and [M+K] + whose masses facilitate assignments of elemental formulas. CID at el evated collision potentials generates fr agment ions whose masses are used to identify masses of acyl groups, the number of acylations, the mass of the disaccharide sugar core , and, in the case of acylsucroses, assignments of acyl groups to either the fruct ose or pyranose rings. 1.5.1.4 High Res olution Mass Spectrometry High resolution mass analyzers, including time - of - flight (ToF) and orbitrap instruments, provide high mass resolving power and accurate mass measurements ( m/z measurements accurate to ~ 5 parts - per - million or 16 less) that are not ac hieved by nominal resolution mass analyzers such as quadrupoles. As the molecular mass of a metabolite increases, the number of possible elemental compositions for a detected m/z value within a given tolerance window also increases [76] . A major benefit of high resolution mass spectrometry lies in the unique exact - mass of each element and isotope. The combination of exact - mass and natural isotopic abundance measurements usually reduces the number of possible elemental compositions to only a few possibilitie s. The value of high resolution MS is demonstrated in acylsugar profiling. Most acylsugars observed to date in surface extracts of Solanaceae plants consist of homologous series differ ing by the number of acylations and CH 2 units in the acyl groups, observ ed as increases of +14 Da. However, Ghosh et al. demonstrated that isobaric acylsugars (those with different elemental formulas, but with the same nominal (integer) masses) such as a t riacylsucrose of formula C 30 H 52 O 14 and tetraacylsucrose C 29 H 48 O 15 both h ave the same nominal masses of 636 Da [23] . However, their monoisotopic masses differ by 0.036 Da, with the triacylated sucrose having the slightly greater mass. From the molecular mas s measurement alone, modern high resolution MS instruments can distingui sh the number of acyl groups, as well as how many total carbons are in the acyl groups for acylsugars. Accurate mass measurements are also informative for assigning elemental formulas to fragment ions. For instance, positive - ion mode spectra of acylated su croses generated at elevated collision energies shows cleavage of the glycosidic bond. From the exact masses of their pyranose and furanose ring fragment ions, the number of acylations on each ring can be determined [23] , though fragment ion masses have ye t to establish specific positions of acylations because much fragmentation involves neutral losses of acyl groups. 1.5.1.5 Tandem Mass Spectrometry (MS/MS) Even when long and shallow LC gradients are used, LC/MS profiles of most plant extracts present sev eral co - eluting chromatographic peaks, making nonselective CID spectra challenging to interpret because it is not clear which fragment ion is derived from each precursor. Tandem mass s pectrometry (MS/MS) offers multiple steps of mass spectrometry selection with CID fragmentation in between. In the two stage MS/MS 17 product ion scan, the first mass analyzer (often a quadrupole) selects a specific precursor ion for fragmentation in the coll ision cell, filtering away ions of all other m/z values, while the secon d mass analyzer scans through a range of m/z values. For acylsugar analysis, product ion MS/MS spectra may be employed to distinguish coeluting acylsugars that differ in molecular mass by 2 Da or more, such as acylsugars with one or more unsaturated ester groups. Tandem MS/MS spectra may be used to distinguish differences in acyl group carbon lengths and saturations (from those with double bonds or rings), while relative ion abundance i n MS/MS experiments may provide useful clues for distinguishing differen ces among carbohydrate groups that form the foundation of acylsugar metabolites. However, multiple isomeric forms of acylsugars are present in some plant extracts that differ by their acyl group branching or by their sugar substitution positions that are n ot evident from MS/MS spectra [23] . 1.5.2 Nuclear Magnetic Resonance (NMR) Spectroscopy Definitive identification of a metabolite often requires that all of the connections between a toms be catalogued, including their stereochemical configurations. Howev er, mass spectrometry often cannot distinguish structural isomers, leaving many metabolite annotations ambiguous. For example, myo - inositol and glucose (Figure 1.1) are isomers of hexo se (C 6 H 12 O 6 ), sugars that are not readily differentiated by mass spectro metry alone. NMR spectroscopy, especially the combination of one - and two - dimensional (1D and 2D) techniques, provides more detailed structural information of metabolites, and is consi [5] . NMR also has a n added benefit in that it is non - destructive, allowing for further spectroscopic analysis. Unfortunately, one of the major challenges of using NMR is that it is less sensitive compare d to MS techniques, often requiring laborious purifications of milligram quantities of analyte. Some NMR spectrometers are equipped with cryoprobe and capillary probe accessories that make it possible to generate spectra for lower (µg) quantities of purif ied material [77] . Only one of the spectrometers at Michigan State Univ ersity, the Bruker Avance 900, is equipped with cryoprobe capabilities, and none has capillary probe capabilities. 18 1.5.2.1 Extraction and purification of acylsugar metabolites for NM R analysis The amount of plant tissue needed to purify 1 mg of a specifi c metabolite depends on the plant genotype, stage of development, growth conditions, and concentration of the target metabolite in the plant tissue [23] . Several studies demonstrated p urification of acylsugars using a myriad of organic extraction solvents and separation techniques (Table 1.1). Recent reports from our laboratories have documented rapid trichome extractions of tomato and petunia species using organic solvents, followed by purification of concentrated trichome extracts using reversed - phase C18 semi - preparative HPLC with fraction collection [38, 50] . For Petunia species, Liu et al. employed a pre - fractionation step using strong anion exchange (SAX) solid phase extraction (SPE) column for separating a group of anionic malonate ester acylsucroses from neutral acylsucroses that had similar retention behavior using reversed phase HPLC [ 50] . Subsequent separation using reversed phase HPLC yielded fractions of sufficient quantity and purity for structure elucidation using NMR spectra. 1.5.2.2 Identifi cation by 1D and 2D NMR spectroscopy Most plant specialized metabolites are composed only of hydrogen, carbon, nitrogen and oxygen. Of these, 1 H and 13 C isotopes are the most important isotopes for NMR structural elucidation. One - dimensional (1D) NMR provi des useful chemical shifts (indicators of atomic environment), coupling constants (intera ctions between atoms) and relative proton quantification for establishing the connectivity of elements/functional groups and their relative stereochemistry in a compou nd. However, acylsugars are large molecules with many similar functional groups, and thei r 1D spectra exhibit signals with overlapping resonances, or significant second order couplings, making it challenging to interpret the results. In addition, the low n atural abundance of 13 C (1.07% of total carbon) compared to 1 H (99.99% of total hydrogen) makes for weak 13 C signals when amounts of sample are low, which may limit the amount of information available from NMR in such cases. To circumvent these issues, str uctural elucidation of acylsugars is facilitated by a variety 2D NMR spectroscopic techni ques [23, 38, 50] . COSY (COrrelation SpectroscopY) is a homonuclear 1 H - 1 H 19 technique for measuring signals that arise from couplings of neighboring protons, usually up to 4 bonds away. For acylsugar analyses, COSY is particularly useful for assigning ring hydrogen connectivity. The heteronuclear HSQC (Heteronuclear Single Quantum Coherence) technique measures short - range couplings arising from one bond interactions of 1 H - 13 C, revealing which protons are attached to specific carbon atoms. HSQC is also used t o distinguish CH 2 groups from CH/CH 3 , which show negative and positive signal respectively. HMBC (Heteronuclear Multiple - Bond Correlation) is a heteronuclear technique similar to HSQC, however, long - distance couplings of 1 H - 13 C are measured (out to 2 - 4 bon ds). For acylsugars, HMBC is especially useful for confirming ring connectivity, and is a powerful approach for determining the positions of acyl substitutions. Additi onal 2D NMR techniques are also useful for determin ing acylsugar atomic connectivity and relative stereoisomerism. For instance, J - resolved NMR may be used for discerning 1 H - 1 H coupling constants of overlapping 1 H resonances. 2D - TOCSY (TOtal Correlated Sp ectroscopY) spectra are useful for revealing the coupling network of sugar ring hydrogens , as all protons on a given ring will have coupling with all other protons, but not with those that are on other sugar rings. C oupled - HSQC spectra yield 1 H - 13 C couplin g constants that can be used to discern relative stereochemistry of glycosylated positions (i.e. versus glycosylation). 1.5.3 Chromatographic Retention Indexing (RI) for annotation of specialized metabolites Our ability to perform rapid detec tion of compounds in complex extracts using UHPLC/MS far exceeds our ability to generate unambiguous struc ture information about the overwhelming majority of them. As we discover more SMs, the risk of discovering the same compounds rises. Plants with close genetic relationships often produce some of the same or structurally similar metabolites [78] . Annotation of most SMs is initially driven by mass spectra and retention times measured using LC/MS. However, variation in GC and LC retention times across labo ratories, instruments, and experimental protocols compromise retention time comparisons and determination of whether metabolites from two genotypes that exhibit similar masses and chromatographic retention times are the same or different compounds. Greater 20 confidence in retention time comparisons may avoid replication of metabolite purification and structure e lucidation. 1.5.3.1 GC/MS Kov á ts retention indexing In gas chromatographic separations, Kováts retention indices have long been used to convert retent ion time into system independent constants, whereby the retention index of a specific compound is normaliz ed to the retention times of adjacent eluting homologous n - alkanes [79] . RI values are calculated by linear interpolation between retention standards according to Equation 1.2, where, t r is the retention time, and n and N are the number of carbon atoms in the shorter and longer n - alkyl standards bracketing the peak of interest respectively. For metabolite profiling, GC/MS indexing is usually used with electron ionization to form mass spectral libraries that contain retention index information [80] . Tables o f GC retention indices in conjunction with mass spectrometry can help to identify compo unds with known retention indices, often providing additional confidence to metabolite annotations, particularly when mass spectra alone are not particularly discriminat ing. Equation 1. 2 . Calculation of RI values. 1.5.3.2 HPLC and LC/MS retention indexing approaches Even though liquid chromatography is often the preceding step before MS analysis, modified Kováts liquid chromatographic indices are rarely used. In fact, a recent review of o pen - access spectral databases by Johnson et al. recognizes that none of the major open - access spectral databases (e.g. MassBank, METLIN) include retention indexing as an orthogonal measurement parameter for LC/MS/MS data [81] . In part, this is because ther e is no widely - accepted or unified HPLC RI system for reversed - phase (RP) chromatography [73] . A few reports of RI standards have consisted of homologous series of 2 - ketoalkanes, alkyl aryl ketones, and 1 - nitroalkanes, as all have comm on chromophores detec ted using ultraviolet spectroscopy [82 - 84] . However, most lack functional groups that facilitate their detection using mass spectrometry. A 2006 21 dissertation touted a series of linear aliphatic esters of 4 - hydroxybenzoic acid which io nized well in negativ e - ion mode ESI, but a complete set of homologs was not prepared, and the range of retention was limited, as the longest alkyl ester was dodecyl [85] . Because ESI is often the first approach for non - volatile SM discovery, indexing stan dards that are ioniza ble using both positive - and negative - ion modes would be preferable. This way, the same standard series could be used in both ionization modes. Indexing standards should have a wide retention range, have molecular masses that are outsi de of targeted approa ches, be inexpensive to produce, non - toxic and non - reactive [73] . By adding a time - indexed dimension to complex LC/ESI/MS metabolomics datasets, similarities and differences in related plant species can be determined rapidly and effic iently. This information has the potential to accelerate discoveries about how plant s pecialized metabolism has evolved. 1.6 Summary of research In this dissertation, analytical strategies for profiling and discovery of acylsugar specialized metabolites f orm glandular trichomes of plants of the family Solanaceae are outlined. Untargeted reversed - phase UHPLC/ESI/Q - Tof/MS metabolite profiling in posi tive - and negative - ion modes of extracts from leaf surface glandular trichomes of two species of the family So lanaceae, Salpiglossis sinuata and Solanum quitoense , revealed diverse acylsugar metabolites that were different from previously discovered metabo lites of this compound class. Acylsugar metabolites were purified by semi - preparative HPLC and their structure s were elucidated utilizing 1 H, 13 C, COSY, J - resolved, TOCSY, HSQC, HMBC, NOESY and ROESY NMR spectroscopy. These efforts established structures o f new acylsucrose and acylinositol metabolites extracted from S. sinuata and S. quitoense and provide new anal ytical approaches for defining acylsugar composition and biodiversity. Furthermore, these results have guided the discovery of previously unidenti fied acylsugar biosynthetic enzymes operating in acylsugar biosynthetic pathways and have extended our underst anding of the evolution of specialized metabolism in the Solanaceae. As we learn more about the Solanaceae, chances of re - discovering the same com pounds in closely related plant species rises. To address the need for a HPLC RI system, a homologous set of S - alkyl glutathione 22 (GS - n - alkyl) RI standards featuring normal saturated chain lengths (1 - 24 carbons) was synthesized. These RI standards encompass a wide RP - UHPLC retention range, are easily ionized by ESI in positive - and negative - ion modes (yielding yiel ded [M+H] + and [M - H] - ions) and show improved capacity for archiving retention data by gradient elution. To test the performance of this retention indexing approach, GS - n - alkyl standards were added at 1.0 µM each to an extract of S. sinuata containing acyl sugars. A thorough in vestigation of acylsucrose RI dependence by gradient elution using GS - n - alkyl standards was performed while altering several important chromatographic experiment variables, including columns, solvent delivery systems, aqueous mobile ph ase pH, column temper ature, LC gradient and organic solvent component. In addition, GS - n - alkyl standard LC/MS analysis shows promise for evaluating RP - HPLC column performance, batch - to - batch column reproducibility and column lifetime. This research has the potential to improve interspecies SM metabolite discovery and dereplication. With this work, we hope to encourage more metabolomics researchers to adopt retention indexing systems as additional measurement parameters for open - access databases. 23 REFERENCE S 24 REFER ENCES 1. Saito, K. and F. Matsuda, Metabolomics for functional genomics, systems biology, and biotechnology. Annu Rev Plant Biol, 2010. 61 (1): p. 463 - 89. 2. Kumar, R., et al., Metabolomics for Plant Improvement: Status and Prospects. Front Plant Sc i, 2017. 8 (1302): p. 1302. 3. Verpoorte, R. and J. Memelink, Engineering secondary metabolite production in plants. Curr. Opin. Biotechnol., 2002. 13 (2): p. 181 - 187. 4. Fiehn, O., Metabolomics the link between genotypes and phenotypes , in Functional Geno mics , C. Town, Editor. 2002, Springer Netherlands: Dordrecht. p. 155 - 171. 5. Last, R.L., A.D. Jones, and Y. Shachar - Hill, Towards the plant metabolome and beyond. Nat Rev Mol Cell Biol, 2007. 8 (2): p. 167 - 74. 6. Mithofer, A. and W. Boland, P lant defense ag ainst herbivores: chemical aspects , in Annual Review of Plant Biology, Vol 63 , S.S. Merchant, Editor. 2012, Annual Reviews: Palo Alto. p. 431 - 450. 7. Pichersky, E. and J. Gershenzon, The formation and function of plant volatiles: perfumes fo r pollinator at traction and defense. Curr Opin Plant Biol, 2002. 5 (3): p. 237 - 43. 8. Pichersky, E. and E. Lewinsohn, Convergent evolution in plant specialized metabolism , in Annual Review of Plant Biology, Vol 62 , S.S. Merchant, W.R. Briggs, and D. Ort, Ed itors. 2011, An nual Reviews: Palo Alto. p. 549 - 566. 9. Dixon, R.A. and D. Strack, Phytochemistry meets genome analysis, and beyond. Phytochemistry, 2003. 62 (6): p. 815 - 6. 10. Schwab, W., Metabolome diversity: too few genes, too many metabolites? Phytochemi stry, 2003. 62 ( 6): p. 837 - 49. 11. Moghe, G.D. and R.L. Last, Something old, something new: conserved enzymes and the evolution of novelty in plant specialized metabolism. Plant Physiol, 2015. 169 (3): p. 1512 - 23. 12. Van de Peer, Y., S. Maere, and A. Meyer, The evolutiona ry significance of ancient genome duplications. Nat Rev Genet, 2009. 10 (10): p. 725 - 32. 13. Comai, L., The advantages and disadvantages of being polyploid. Nat Rev Genet, 2005. 6 (11): p. 836 - 46. 14. Higashi, Y. and K. Saito, Network analysis for gene disco very in plant - specialized metabolism. Plant Cell Environ, 2013. 36 (9): p. 1597 - 606. 15. Schilmiller, A.L., R.L. Last, and E. Pichersky, Harnessing plant trichome biochemistry for the production of us eful compounds. Plant J, 2008. 54 (4): p. 702 - 11. 16. Levi n, D.A., The Role of Trichomes in Plant Defense. The Quarterly Review of Biology, 1973. 48 (1, Part 1): p. 3 - 15. 25 17. Wagner, G.J., E. Wang, and R.W. Shepherd, New approaches for studying and exploitin g an old protuberance, the plant trichome. Ann Bot, 2004. 93 (1): p. 3 - 11. 18. Landry, L.G., C.C. Chapple, and R.L. Last, Arabidopsis mutants lacking phenolic sunscreens exhibit enhanced ultraviolet - B injury and oxidative damage. Plant Physiol, 1995. 109 (4) : p. 1159 - 66. 19. Ehleringer, J., Ecology and ecophysiolo gy of leaf pubescence in North American desert plants. Biology and chemistry of plant trichomes, 1984. 20. Fahn, A., Secretory tissues in vascular plants. New Phytologist, 1988. 108 (3): p. 229 - 257. 2 1. Glas, J.J., et al., Plant glandular trichomes as targe ts for breeding or engineering of resistance to herbivores. Int J Mol Sci, 2012. 13 (12): p. 17077 - 103. 22. Li, C., Z. Wang, and A.D. Jones, Chemical imaging of trichome specialized metabolites using contact printing and laser desorption/ionization mass spe ctrometry. Anal Bioanal Chem, 2014. 406 (1): p. 171 - 82. 23. Ghosh, B. and A.D. Jones, Profiling, characterization, and analysis of natural and synthetic acylsugars (sugar esters). Analytical Methods, 2017. 9 (6): p. 892 - 905. 24. Schilmiller, A., et al., Mass spectrometry screening reveals widespread diversity in trichome specialized metabolites of tomato chromosomal substitution lines. Plant J, 2010. 62 (3): p. 391 - 403. 25. Gang, D.R., et al., An investi gation of the storage and biosynthesis of phenylpropenes in sweet basil. Plant Physiol, 2001. 125 (2): p. 539 - 55. 26. Turner, C.E., M.A. Elsohly, and E.G. Boeren, Constituents of Cannabis sativa L. XVII. A review of the natural constituents. J Nat Prod, 1980. 43 (2): p. 169 - 234. 27. Stevens, J.F. and J.E. Page, Xa nthohumol and related prenylflavonoids from hops and beer: to your good health! Phytochemistry, 2004. 65 (10): p. 1317 - 30. 28. 2015 Nobel Prize in Physiology or Medicine . 2015, The Nob el Assembly at Karolinska Institutet. 29. Luckwill, L.C., The genus Lycop ersicon. An historical, biological and taxonomic survey of the wild and cultivated tomatoes. PhD thesis, Department of Botany, University of Aberdeen, UK, 1943. 30. van den Oever - van den Elsen, F., et al., Quantitative resistance against Bemisia tabaci in Solanum pennellii: Genetics and metabolomics. J Integr Plant Biol, 2016. 58 (4): p. 397 - 412. 31. Baier, J.E., et al., Indirect selection of industrial tomato genotypes that are resista nt to spider mites (Tetranychus urticae). Genetics and Molecular Research , 2015. 14 (1): p. 244 - 252. 32. Goffreda, J.C., et al., Aphid deterrence by glucose esters in glandular trichome exudate of the wild tomato, Lycopersicon pennellii. J Chem Ecol, 1989. 15 (7): p. 2135 - 47. 33. Leckie, B.M., et al., Differential and synergistic functionality of acylsugars in suppressing oviposition by insect herbivores. PLoS One, 2016. 11 (4): p. e0153345. 34. Luu, V.T., et al., O - acyl sugars protect a wild tobacco from both native fungal pathogens and a specialist herbivore. Plant Physiol, 2017. 174 (1): p. 370 - 386. 26 35. Weinhold, A. and I.T. Baldwin, Trichome - derived O - acyl sugars are a first meal for caterpillars that tags them for predation. Proc Natl Acad Sci U S A, 2011. 108 (19): p. 7855 - 9. 36. Moghe, G.D., et al., Evolutionary routes to bioch emical innovation revealed by integrative analysis of a plant - defense related specialized metabolic pathway. Elife, 2017. 6 : p. e28486. 37. Connolly, J.D. and R.A. Hill, Dictionary of terpenoids . Flavour and Fragrance Journal. 1991, London: Chapman & Hall. 38. Ghosh, B., T.C. Westbrook, and A.D. Jones, Comparative structural profiling of trichome specialized metabolites in tomato (Solanum lycopersicum) and S. habrochaites: acylsugar pr ofiles revealed by UHPLC/MS and NMR. Metabolomics, 2014. 10 (3): p. 496 - 50 7. 39. Schilmiller, A.L., et al., Functionally divergent alleles and duplicated Loci encoding an acyltransferase contribute to acylsugar metabolite diversity in Solanum trichomes. Plant Cell, 2015. 27 (4): p. 1002 - 17. 40. Schilmiller, A.L., et al., Acylsuga r Acylhydrolases: Carboxylesterase - Catalyzed Hydrolysis of Acylsugars in Tomato Trichomes. Plant Physiol, 2016. 170 (3): p. 1331 - 44. 41. King, R.R., et al., Sucrose ester s associated with glandular trichomes of wild Lycopersicon species. Phytochemistry, 199 0. 29 (7): p. 2115 - 2118. 42. King, R.R., et al., Characterization of 2,3,4,3' - tetra - O - acylated sucrose esters associated with the glandular trichomes of Lycopersicon - typi cum. Journal of Agricultural and Food Chemistry, 1993. 41 (3): p. 469 - 473. 43. King, R.R ., et al., 3,4 - Di - O - Isobutyryl - 6 - O - Caprylsucrose - the Major Component of a Novel Sucrose Ester Complex from the Type - B Glandular Trichomes of Solanum - Berthaultii Hawkes (Pi 473340). Journal of the Chemical Society - Chemical Communications, 1986(14): p. 107 8 - 1079. 44. King, R.R., R.P. Singh, and L.A. Calhoun, Isolation and characterization of 3,3',4,6 - tetra - O - acylated sucrose esters from the type - B glandular trichomes of S olanum - berthaultii Hawkes (Pi - 265857). Carbohydrate Research, 1987. 166 (1): p. 113 - 121. 45. King, R.R. and A.K. Singh, - tetra - O - acylated sucrose esters from the type B glandular trichomes of Solunum n eocardenasii Hawkes & HJerting (PI 498129). Carbohydrate Research, 1988. 173 : p. 235 - 24 1. 46. King, R.R., L.A. Calhoun, and R.P. Singh, 3,4 - Di - O - and 2,3,4 - tri - O - acylated glucose esters from the glandular trichomes of nontuberous Solanum species. Phytochem istry, 1988. 27 (12): p. 3765 - 3768. 47. Burke, B.A., G. Goldsby, and J.B. Mudd, Polar ep icuticular lipids of Lycopersicon pennellii. Phytochemistry, 1987. 26 (9): p. 2567 - 2571. 48. Li, A.X. and J.C. Steffens, An acyltransferase catalyzing the formation of di acylglucose is a serine carboxypeptidase - like protein. Proceedings of the National Acad emy of Sciences of the United States of America, 2000. 97 (12): p. 6902 - 6907. 49. Herrera - Salgado, Y., et al., Myo - inositol - derived glycolipids with anti - inflammatory act ivity from Solanum lanceolatum. J Nat Prod, 2005. 68 (7): p. 1031 - 6. 27 50. Liu, X.X., et a l., Profiling, isolation and structure elucidation of specialized acylsucrose metabolites accumulating in trichomes of Petunia species. Metabolomics, 2017. 13 (7): p. 85. 51. Singh, A.P., et al., Two acyl sucroses from Petunia nyctaginiflora. Phytochemistry , 2003. 63 (4): p. 485 - 9. 52. Begum, A.S., et al., New pairs of acyl sucroses from Petunia nyctaginiflora Juss. Journal of the Indian Chemical Society, 2004. 81 (6): p. 49 5 - 500. 53. Begum, A.S., et al., Two novel acyl sucroses from Petunia nyctaginiflora. In dian Journal of Chemistry Section B - Organic Chemistry Including Medicinal Chemistry, 2005. 44 (3): p. 648 - 650. 54. Severson, R.F., et al., Isolation and characterization of the sucrose esters of the cuticular waxes of green tobacco leaf. Journal of Agricult ural and Food Chemistry, 1985. 33 (5): p. 870 - 875. 55. Arrendale, R.F., et al., Characterization of the sucrose ester fraction from Nicotiana - glutinosa. Journal of Agricu ltural and Food Chemistry, 1990. 38 (1): p. 75 - 85. 56. Matsuzaki, T., et al., Leaf surfa ce glycolipids from Nicotiana - acuminata and Nicotiana - pauciflora. Agricultural and Biological Chemistry, 1991. 55 (5): p. 1417 - 1419. 57. Zhang, C.R., et al., New antiinfl ammatory sucrose esters in the natural sticky coating of tomatillo (Physalis philadelph ica), an important culinary fruit. Food Chem, 2016. 196 : p. 726 - 32. 58. Ovenden, S.P., et al., Physaloside A, an acylated sucrose ester from Physalis viscosa. J Nat Prod , 2005. 68 (2): p. 282 - 4. 59. Castillo, M., et al., Partially acylated glucose and sucro se derivatives from Salpiglossis - sinuata (Solanaceae). Journal of Chemical Research - S, 1989(12): p. 398 - 399. 60. Eckart, E., Secondary Metabolites Derived from Fatty Aci ds and Carbohydrates , in Solanaceae and Convolvulaceae: Secondary Metabolites: Biosynth esis, Chemotaxonomy, Biological and Economic Significance (A Handbook) , E. Eich, Editor. 2008, Springer Berlin Heidelberg: Berlin, Heidelberg. p. 525 - 582. 61. D'Auria, J .C., Acyltransferases in plants: a good time to be BAHD. Curr Opin Plant Biol, 2006. 9 ( 3): p. 331 - 40. 62. Schilmiller, A.L., A.L. Charbonneau, and R.L. Last, Identification of a BAHD acetyltransferase that produces protective acyl sugars in tomato trichome s. Proc Natl Acad Sci U S A, 2012. 109 (40): p. 16377 - 82. 63. Fan, P., et al., In vitro reconstruction and analysis of evolutionary variation of the tomato acylsucrose metabolic network. Proc Natl Acad Sci U S A, 2016. 113 (2): p. E239 - 48. 64. Nadakuduti, S.S., et al., Characterization of trichome - expressed BAHD acyltransferases in petunia axi llaris reveals distinct acylsugar assembly mechanisms within the Solanaceae. Plant Physiol, 2017. 175 (1): p. 36 - 50. 65. St - Pierre, B. and V. De Luca, Chapt er Nine Evolution of acyltransferase genes: Origin and diversification of the BAHD superfamily of acy ltransferases involved in secondary metabolism . Vol. 34. 2000. 285 - 315. 28 66. Gohlke, R.S., Time - of - Flight Mass Spectrometry and Gas - Liquid Partition Chromatography. Analytical Chemistry, 1959. 31 (4): p. 535 - 541. 67. Walters, D.S. and J.C. Steffens, Branched chain amino acid metabolism in the biosynthesis of Lycopersicon pennellii glucose esters. Plant Physiol, 1990. 93 (4): p. 1544 - 51. 68. Kro umova, A.B., D. Zaitlin, and G.J. Wagner, Natural variability in acyl moieties of sugar esters produced by certain tob acco and other Solanaceae species. Phytochemistry, 2016. 130 : p. 218 - 27. 69. Bell, D.S., H.M. Cramer, and A.D. Jones, Rational method deve lopment strategies on a fluorinated liquid chromatography stationary phase: mobile phase ion concentration and tempera ture effects on the separation of ephedrine alkaloids. J Chromatogr A, 2005. 1095 (1 - 2): p. 113 - 8. 70. Bell, D.S. and A.D. Jones, Solute at tributes and molecular interactions contributing to "U - shape" retention on a fluorinated high - performance liquid chrom atography stationary phase. J Chromatogr A, 2005. 1073 (1 - 2): p. 99 - 109. 71. LCGC - Editors, HPLC Column Selection , in LCGC Europe . 2013. p. 298. 72. Jandera, P., Stationary and mobile phases in hydrophilic interaction chromatography: a review. Anal Chim Acta , 2011. 692 (1 - 2): p. 1 - 25. 73. Kind, T. and O. Fiehn, Advances in structure elucidation of small molecules using mass spectrometry. Bioana l Rev, 2010. 2 (1 - 4): p. 23 - 60. 74. Bruins, A.P., Mechanistic aspects of electrospray ionization. Journal of Chromatogr aphy A, 1998. 794 (1 - 2): p. 345 - 357. 75. Sleno, L. and D.A. Volmer, Ion activation methods for tandem mass spectrometry. J Mass Spectrom, 2 004. 39 (10): p. 1091 - 112. 76. Xian, F., C.L. Hendrickson, and A.G. Marshall, High resolution mass spectrometry. Anal C hem, 2012. 84 (2): p. 708 - 19. 77. Molinski, T.F., Microscale methodology for structure elucidation of natural products. Curr Opin Biotechnol, 2010. 21 (6): p. 819 - 26. 78. Zhao, N., et al., Studying plant secondary metabolism in the age of genomics. Critical Reviews in Plant Sciences, 2013. 32 (6): p. 369 - 382. 79. Kováts, E., Gas - chromatographische Charakterisierung organisch er Verbindungen. Teil 1: Retentionsindices aliphatischer Halogenide, Alkohole, Aldehyde und Ketone. Helvetica Chimica Acta, 1958. 41 (7): p . 1915 - 1932. 80. Vinaixa, M., et al., Mass spectral databases for LC/MS - and GC/MS - based metabolomics: State of the fi eld and future prospects. Trac - Trends in Analytical Chemistry, 2016. 78 : p. 23 - 35. 81. Johnson, S.R. and B.M. Lange, Open - access metabolom ics databases for natural product research: present capabilities and future potential. Front Bioeng Biotechnol, 2015. 3 : p. 22. 82. Baker, J.K. and C.Y. Ma, Retention index scale for liquid - liquid chromatography. Journal of Chromatography, 1979. 169 (Feb): p. 107 - 115. 29 83. Smith, R.M., Alkylarylketones as a retention index scale in liquid - chromatography. Journal of Chromato graphy, 1982. 236 (2): p. 313 - 320. 84. Bogusz, M. and R. Aderjan, Improved standardization in seversed - phase high - performance liquid - chroma tography using 1 - nitroalkanes as a retention index scale. Journal of Chromatography, 1988. 435 (1): p. 43 - 53. 85. Hanle y Jr, J.C., Chromatography/mass spectrometry methodologies for metabolomics, doctoral dissertation . 2006, The Pennsylvania State Universit y. ProQuest UMI#3229404. 30 Chapter 2: LC/MS profiling and NMR structural elucidation of specialized metabolites from Salpiglossis sinuata reveals extensive acylsucrose diversity including unsaturated and aromatic esters 2.1 Introduction Metabolic processe s that increase fitness in plants evol ved through mutations and led to structurally diverse specialized metabolites (SMs; also referred to as natural products and secondary metabolites) that are not directly involved in central metabolism. Conservative est imates have suggested that the plant k ingdom produces more than 200,000 SMs [1, 2] , with many of these being taxonomically restricted to a limited range of species or genotypes. Accumulation of SMs in specialized cell types or tissues facilitates discovery of candidate biosynthetic genes which are often differentially expressed in these tissues. Glandular trichomes (GTs) serve as convenient model tissues in that they are hair - like epidermal outgrowths (Appendix Figure 2.5), are accessible for sampling and a ccumulate large quantities of SMs. GTs are found on the surfaces of ~30% of vascular plant species [3] and are prolific chemical factories that synthesize, store and secrete large quantities of diverse SMs. GT exudates are easily sampled, facilitating deta iled characterization of their metabol ites as well as transcripts that encode for biosynthetic enzymes [4 - 6] . In addition, GTs exhibit novel biochemical pathways for specialized metabolism and have been suggested as targets for genetic engineering for prod uction of useful natural products and engineering plant resistance against insects and pathogens [6 - 9] . Members of the family Solanaceae, which includes a large number of agronomically important plants including tomato, potato, pepper, and tobacco, synthe size and/or store diverse arrays of SM s in GTs, including terpenes, flavonoids, alkaloids and fatty acid derivatives [6, 7, 10] . Most investigated Solanaceae Type I/IV GTs accumulate structurally diverse sugar esters, known as acylsugars. Common examples i nclude glucose and sucrose with three or four acylations of C2 - C12 branched or straight chain aliphatic esters. These SMs often provide the first line of physical and/or chemical defenses against herbivores [3, 31 11] . GTs of tobacco ( Nicotiana tabacum ), petu nia ( Petunia spp. ), wild potato ( Solan um berthaultii ), cultivated tomato ( Solanum lycopersicum ) and its wild relatives ( S. habrochaites and S. pennellii ) are widely - studied prolific accumulators of acylsugars [12 - 14] . NMR and GC/MS structural characterizat ion have identified > 20 known acylsuga r acyl groups in the Solanaceae [15] . From known acyl group building blocks alone, the number of possible acylsucrose permutations that span mono - to complete - acylation of eight sucrose hydroxyl groups ex tends into the billion s . We estimate that thousands (or more) structurally diverse acylsugars exist across the Solanaceae, with only ~100 acylsugar metabolites fully characterized at present. Acylation serves as a widespread mechanism for assembly of plant SMs [16] . Unt il recent years, functions recent combinations of transcript analysis, LC/ MS metabolite profiling of S. lycopersicum M82 x S. pennellii LA0716 and in vitro biochemistry using recombinant acyltransferases guided the discovery of the entire set of four BAHD acyltransferase (ASAT) enzymes requir ed for assembly of tetraacylsucroses in tomato [17 - 19] . Members of the broader Solanaceae employ multiple ASATs that convert carbohydrates and acyl - CoAs to complex arrays of acylsucroses, but substrate selectivities of ASATs may not be apparent from DNA se quence similarity alone [17, 19 - 21] . Functional assays for ASATs from the gene ra Solanum , Petunia , and more recently, Salpiglossis have shown that small numbers of amino acid substitutions can alter substrate selectivity of ASATs, position - selectivity of a cylation of orthologues, and the order in which individual ASATs are used in a cylsugar biosynthetic pathways [20 - 22] . Several reports have characterized the number, topologies, and lengths of ester groups that make up acylsugars in numerous accessions fro m the Solanaceae, predominantly from the genus Nicotiana [15, 23, 24] . However , outside of the genera Solanum and Petunia [12 - 14] , comprehensive descriptions of acyl substitution positions and acyl chain selectivity at each position of sugar cores are limi ted. Given our recent evidence that the main differences in individual ASAT f unctions derive from varied selectivities of both 32 acyl acceptors and acyl donors [17, 19, 21] , annotation of ASAT functions requires definition of which carbohydrate positions are acylated as well as the diversity of acyl groups at each position. The focu s of the current study is to define acylsugar diversity in Salpiglossis sinuata , a member of an early emerging lineage in the Solanaceae family, estimated to have diverged ~25 mya relative to the carefully studied tomato clade [25] . Trichome extracts were subjected to UHPLC/MS profiling, purification and structural elucidation using NMR spectrosco py for the purpose of uncovering differences in acylsugar composition and to facilitate discovery of genetic factors responsible for their biosynthesis. In doing s o, we have discovered new acylsugars, including those containing unsaturated acyl and aromati c esters; suggesting novel gene functions are at work in acylsugar biosynthesis in S. sinuata . This untargeted approach aims to establish the diversity of acyl gro ups and their position - selective attachment and guided the discovery of biosynthetic enzymes functioning in S. sinuata [21] . 2.2 Materials and methods 2.2.1 Plant cultivation and metabolite extraction S. sinuata seeds were obtained from The New York Botan ical Gardens (NY, USA), and were grown using Jiffy peat pellets in growth chambers at MSU. Fo r metabolite profiling and purification, the complete aerial tissues of 28 ten - week old S. sinuata plants (~0.5 m height; plants were cut at the stems in 10 - 15 cm segments) were dipped into 1.9 L of acetonitrile: isopropanol (AcN:IPA, v/v, 1:1) for 10 mins . A 10 - mL aliquot of this bulk extract solution, with 5.0 µM telmisartan added as internal standard, was used for all metabolite profiling and purification method development. The remaining extract was concentrated for metabolite purification as described below. Additional plant growth and extraction metadata are provided in Appendix Table 2.2. 2.2.2 Profiling of acylsugar metabolites using UHPLC/MS and MS/MS Acylsu gar metabolite profiling was performed using a Waters Xevo G2 - XS quadrupole time - of - flight ma ss spectrometer (QTof/MS) equipped with an Ascentis ® Express C18 Analytical HPLC column (10 cm × 2.1 33 mm, 2.7 µm particle size) and operated using CID in positive - and negative - ion modes. Untargeted CID spectra were acquired over m/z 50 - 1500 in centroid fo rmat using four quasi - simultaneous collision potential functions (0, 10, 25, 50 V, each with 0.1 s acquisition times) to generate fragment ions. MS/MS spectra were generated from purified acylsucrose metabolites using a CID ramp from 5 - 60 V collision poten tial. Additional metadata regarding the UHPLC/MS and MS/MS methods are provided in Appendix Tables 2.3 and 2.5 . MS/MS product ion spectra are provided in Appendix Figures 2.8 - 2.23. 2.2.2.1 Deep profiling of acylsucrose specialized metabolites by LC/ESI+/MS Deep profiling of acylsucrose metabolites was carried out by first calculating exact molecular masses for permutations of acylsucrose pseudomolecular m/z values: [M+NH 4 ] + , [M+Na] + , and [M+K] + . NMR results, as well as LC/MS and GC/MS profiling revealed evidence for eight ester groups: C2, C3, iC4, iC5, aiC5, C5 T , aiC6, and C8 P . Since acylsucroses have up to eight hydroxyl positions available for acylation, all degre es of acylation were considered (mono - to octaacylsucroses) with saturated and un saturated esters, including acylsucroses with unsaturated ester combinations of up to two C5 T and/or C8 P groups (i.e. acylsucroses with 1, 2, 4, 5 and 8 double bond equivalenc ies relative to sucrose saturated ester analogues). For each acylsucrose permutat ion, [M+NH 4 ] + , [M+Na] + , and [M+K] + extracted ion chromatograms were generated (in a stacked format within MassLynx software) using a 40 ppm mass window. Acylsucrose retention times and the presence of all three pseudomolecular masses were used as diagnosti c criteria. For each chromatographic peak, its corresponding mass spectra were then analyzed manually. Since many acylsucrose metabolites are at low levels and co - elute with m ore abundant forms, mass spectra were carefully analyzed for mass accuracy, isoto pe abundancies and centroid satellite peaks to avoid false positives. For instance, an acylsucrose with one unsaturation (containing C5 T ) has M+2 isotope signal that is consis tent with a saturated acylsucrose analogue. Here, examination of only extracted i on chromatograms may lead to an incorrect assignment of two acylsucrose signals instead of one. For example, acylsucroses S5:22 :1( 2,5,5,5,5 T ) with [M+NH 4 +2] + has m/z 738.382 v ersus S5:22 :0( 2,2,6,6,6) with [M+NH 4 ] + has m/z 34 738.391. Also of note, there is a shift towards reduced retention time for acylsucroses containing unsaturations (in the previous example, 52.55 mins versus 56.08 mins respectively). 2.2.3 Assessment of acyl g roup diversity by transesterification and GC/MS To assess the ester composition of S. sinuata acylsugars, a transesterification reaction was performed on a 10 mL aliquot of the bulk extract to convert acyl groups to ethyl esters amenable to GC/MS analysis. Transesterification and GC/MS procedures are d escribed in detail in Appendix Table 2.6. Mass spectra are shown in Appendix Figures 2.25 - 2.30. 2.2.4 Purification of acylsugar metabolites by semi - preparative HPLC For metabolite purification, approximately 1 L of the S. sinuata bulk extract (described i n Section 2.1) was concentrated to dryness under vacuum, dissolved in 5 mL of AcN:IPA, and fractionated by repeated injection of 50 µL samples onto a Thermo Scientific Acclaim TM 120 C18 semi - preparative HPLC c olumn (4.6 x 150 mm, 5 µm particle size) with a utomated fraction collection (Appendix Table 2.8). Acylsugars purified for NMR analysis were isolated with the method described above except the C8 P containing acylsugar, which required secondary purification using a Supelco Ascentis ® Express F5 semi - prepa rative HPLC column (15 cm × 4.6 mm, 2.7 µm particle size). A shortened, but comparable semi - preparative gradient was used for the secondary purification of this acylsugar (Appendix Table 2.8). 2.2.5 Analysis of acylsugars by NMR spectroscopy HPLC fractions of sufficient purity for a single metabolite, as assessed by LC/MS, were combined and concentrated to dryness under N 2 gas. Samples were dissolved in 250 or 300 µL of deuterated NMR solvents CDCl 3 (99.8 atom % D) and acetonitrile - d 3 (99.96 atom % D) and transferred to solvent - matched Shigemi tubes for analysis. 1 H, 13 C, gCOSY, gHSQC, gHMBC, J - resolved 1 H, and ROESY NMR experiments were performed using a Bruker Avance 900 spectrometer equipped with a TCI tripl e resonance probe. All spectra were ref erenced to non - deuterated solvent signals: CDCl 3 H C = 77.20 ppm) and acetonitrile - 35 d 3 H C = 118.70 ppm). More experimental conditions and spectra are located in Appendix Tables 2.9 - 2.27, as well as Appendix Figures 2.31 - 2.141. 2 .3 Results and Discussion 2.3.1 UHPLC/ESI/CID/QTof/MS profiling establishes diversity of S. sinuata acylsucroses Profiling of metabolites in complex mixtures requires sensitive detection, capacity to resolve similar c ompounds including isomers, and accurat e measurements of molecular and fragment masses that serve as the basis for metabolite annotation. UHPLC coupled with QTof/MS detection using ESI in positive - and negative - ion modes aided in resolving acylsugars in co mplex extracts. Base peak intensity chr omatograms of S. sinuata trichome extracts showed hundreds of distinct peaks differing in retention time. A group of compounds eluting between 25 - 75 mins using a 110 - min gradient exhibited retention times, ion masses, and mass defects similar to acylsugars from tomato and its wild relatives [26, 27] . High resolution mass spectra of these peaks were consistent with a complex array of substituted acylsugars. At least 20 of the most abundant acylsugars were observed as pe aks in the base peak intensity (BPI) ch romatogram with retention times in the range of 45 - 65 mins (Figure 2.1A). Further evidence supporting their designation as acylsugars was obtained from examining mass spectra at elevated collision energies. Proposed s tructures of purified metabolites and I nChI keys are presented in Appendix Tables 2.12 - 2.27. As these metabolites were identified without available authentic standards or synthetic confirmation, and their identities are based on NMR and mass spectrometric data alone, their structures should be considered as putative, meeting level 2 criteria of the Metabolomics Standards Initiative guidelines [28] . 36 Figure 2. 1 . UHPLC/ESI/CID/QTof/MS metabolite profiles (left) and mass spectra gener ated from acylsugar #12 (right) : (A) ESI( - ) BPI chromatogram displaying acylsugars (formate adducts) at CID Function 1 = 0 V, purified acylsugars characterized by NMR analysis are annotated by a number indicative of their order of chromatographic elution; (B) ESI( - ) XIC for C5 and C6 carboxy late fragment anions, m/z 101.06 and 115.08 combined (Function 4 = 50 V), hexaacylsucrose #16 shows significantly weaker signal for C5 and C6 carboxylate anions compared to tetra - and pentaacylsucrose analogues; (C) ESI( - ) XIC for C5 T carboxylate fragment anion, m/z 99.05 (Function 4 = 50 V), co - elution of acylsucrose #7 with #8 is evident; (D) ESI( - ) XIC for C8 P carboxylate fragment anion, m/z 135.05 (Function 3 = 25 V), (E) ESI( - ) low energy (Function 1) mass spectrum o f acylsucrose #12, S4:22:0(5,5,6,6), displaying [M+HCOO] - adduct, (F) ESI( - ) at elevated energy (Function 4), displaying neutral losses consistent with two C6 and two C5 acyl groups and prominent C5 and C6 carboxylate fragments at m/z 101.06 and 115.08; ( G) ESI(+) low energy (Function 1 = 0 V) mass spectrum of peak #12 displaying [M+NH 4 ] + adduct, (H) ESI(+) at elevated energy (Function 2 = 10 V), displaying cleavage of the glycosidic linkage to reveal a C5 - acylated furanose ring fragment m/z 247.12 and a p yranose ring f ragment m/z 443.26 acylated with two C6 and one C5 esters. Annotation of acylsugars by LC/MS was facilitated by assigning molecular masses and formulas, the masses of the acyl groups, the number of acylations, and to the extent possible, the sugar core and positions of substituents. This information was obtained using multiplexed CID at various collision potentials. Low collision potentials produced abundant formate and ammonium adduct ions, [M+HCOO] - and [M+NH 4 ] + , as well as [M+Cl] - , [M+NO 3 ] - , [M+Na] + and [M+K] + whose masses facilitate assignments of molecular formulas (MFs). CID at elevated collision potentials generated fragment ions used to identify important differences among acyl groups, the number of acylations, the mass of the disacchar ide sugar core, and assignments of acyl groups to either the fructose or pyranose rings of sucrose. 37 To illustrate acylsugar annotation with an example, a major peak in the BPI chromatogram (Figure 2.1A) at t r = 56.49 min (acylsucrose #12) yielded mass spe ctra showi ng [M+HCOO] - at m/z 751.38 in negative - ion mode (Figure 2.1E) and [M+NH 4 ] + at m/z 724.41 in positive - ion mode (Figure 2.1G). These ion masses are consistent with an acylsucrose of MF C 34 H 58 O 15 (structure in Figure 2.1E). Because sucrose has eleve n oxygen a toms, the four additional oxygens suggest four acylations, as each acylation adds one oxygen atom. Spectra generated at elevated collision energies (Figure 2.1F) show abundant fragment ions at m/z 115.08 and 101.06 indicative of carboxylate (acyl anions) w ith MFs C 6 H 9 O 2 - and C 5 H 7 O 2 - . Further confirmation of C5 and C6 esters comes from observation of two neutral mass losses each from [M - H] - of 84.06 and 98.07 Da (C 5 H 8 O and C 6 H 10 O). Losses of two C6 and two C5 groups leaves m/z 341.11, corresponding to the de protonated sucrose core. In addition, neutral losses of 18.01 Da consistent with H 2 O were regularly observed during fragmentation (one example is m/z 323.10). Positive ion spectra obtained at elevated collision energy (Figure 2.1H) had an abundan t fragment ion m/z 247.12, consistent with one C5 acyl chain on the furanose ring. Cleavage producing a less abundant pyranose fragment ion m/z 443.26 suggests substitution of two C6 and one C5 on this ring. This acylsucrose was annotated as S4:22:0(5,5,6, 6), using groups, the number in parentheses describe the number of carbon atoms in each acyl group, the number 22 reflects the total number of carbon atoms acr or double bonds in the acyl groups. Most acylsugars observed to date in surface extracts of Solanaceae plants consist of homologous series differing by the number of acylations and CH 2 units in the acyl groups, observed as increases of +14 Da [5] . In addition, reverse phase HPLC retention times of acylsugars increase with acyl group chain length. However, the predictability of acylsugar masses and retention times obscures deeper explora tion of low abundance acylsugars. Acyl groups differing by only one degree of unsaturation (or more) frequently co - elute with saturated analogues and may be missed by manual inspection. 38 Metabolite profiles of S. sinuata exhibited extensive complexity, with frequent chromatographic overlap of acylsugar metabolites obscuring less abundant compounds. To improve chromatographic resolution of lower abundance acylsugar metabolites from more abundant forms, a long and shallow g radient of 110 - min duration was emplo yed. The resulting chromatogram resolved peaks with molecular masses that deviate by - 2 Da and - 8 Da from common saturated acyl analogues. This is illustrated by a compound with t r = 52.55 min (#7, Figure 2.1A and 2.1C) , exhibiting [M+HCOO] - at m/z 763.34 and [M+NH 4 ] + at m/z 736.38, consistent with neutral formula C 34 H 54 O 16 . This compound exhibits one more degree of unsaturation than a penta - substituted acylsucrose with saturated esters. Negative mode MS/MS product ion s pectra of m/z 763.34 (Figure 2.2A) sh ow abundant product ions at m/z 101.06 and 99.04 indicative of C5 saturated and monounsaturated acyl anions, the latter consistent with MF C 5 H 5 O 2 - . Given the occurrence of tiglyl esters of tropane alkaloids in the Solan aceae [29] , we initially annotated th is metabolite as a tiglyl (C5 T ) acylsugar (further confirmed by NMR results in Section 2. 3.3). Further support of this conclusion came from discerning combinations of neutral mass losses from [M - H] - of a C5 T group (82.0 4 Da, C 6 H 8 O), and additional product ions provided evidence of three saturated C5 acyl groups and one C2 group (neutral loss 42.01 Da, C 2 H 2 O). Positive mode MS/MS product ion spectra of m/z 736.38 (Figure 2.2B) showed cleavage of the glycosidic bond, producing abundant fragment ion m/z 289.13 , consistent with one saturated C5 and one C2 acyl chain on the furanose ring. The pyranose fragment ion at m/z 413.22 suggests substitutions of two C5 and one C5 T esters on this ring, and support s annotation as S5:22:1(2,5,5,5,5 T ). 39 Figure 2. 2 . MS/MS product ion spectra of [M+HCOO] - and [M+NH 4 ] + of acylsugars generated using a linear 5 - 60 V collision energy ramp with 0.5 s scan time; (A) ESI( - ) MS/MS spectrum of product ions of m/z 763 ([M+HCOO] - ) for acylsugar #7 showing eviden ce for neutral losses of C2, C5 T , C5 and H 2 O, as well as correspondin g carboxylate anions for ester groups C5 T and C5 ( m/z 99 and 101); (B) ESI(+) MS/MS spectrum of product ions of m/z 736 ([M+NH 4 ] + ) for acylsugar #7 showing cleavage of glycosidic linkage to yield abundant acylated furanose ( m/z 289), and less abundant acyl ated pyranose ( m/z 413) product ions; (C) ESI( - ) MS/MS spectrum of product ions of m/z 799 ([M+HCOO] - ) of acylsugar #10 showing evidence of neutral losses of C2, C5, C8 P and H 2 O, as well as corresponding carboxylate anions for ester groups C5 ( m/z 101) and C8 P ( m/z 135); (D) ESI(+) MS/MS spectrum of products of m/z 772 ([M+NH 4 ] + ) for acylsugar #10 showing cleavage of glycosidic linkage to produce abundant acylated furanose ( m/z 289), and l ess abundant acylated pyranose ( m/z 449) fragment ions. Anot her acylsugar, t r = 54.29 min (#10, Figure 2.1A) exhibiting [M+HCOO] - at m/z 799.34 and [M+NH 4 ] + with m/z 772.38 was assigned the formula C 37 H 54 O 16 . In this case, the formula indicates four more d egrees of unsaturation than its corresponding penta - substituted acylsucrose with saturated esters. MS/MS product ion spectra in negative mode (Figure 2.2C) showed abundant fragment ions with masses of C5 sat urated acyl anions, as well as m/z 135.04, consis tent with formula C 8 H 7 O 2 - , suggesting a single acyl group with four unsaturations. Given the precedent of phenylacetyl tropane esters in the Solanaceae [29] , we annotated this metabolite as a phenylacetyl (C 8 P ) acylsugar (further confirmed by NMR in Sectio n 2. 3.3). Support for this assignment derives from neutral losses of 118.04 Da (C 8 H 6 O), three saturated C5 and one C2 from [M - H] - . The positive mode MS/MS product ion spectrum of m/z 772.38 (Figure 2.2D) sho ws cleavage of the glycosidic bond which produced an abundant fragment ion m/z 289.13, consistent with one saturated C5 40 and one C2 acyl chain on the furanose ring, and a pyranose fragment ion, m/z 449.22, with substitutions of two C5 and one C8 P esters on this ring. Thus, this acylsugar was annotated S5: 25:4(2,5,5,5,8 P ). Deep analysis of S. sinuata LC/MS profile data using positive - ion mode revealed 127 different MFs consistent with tri - through hepta - substituted acylsucroses (Appendix Table 2.4). All acyls ugars appear to have at least one C5 or C6 acyl c hain, illustrated in the combined XIC for C5 and C6 acyl anions ( m/z 101 and 115) in Figure 2.1B. We found no evidence of saturated long chain esters exceeding C6 carbon chain lengths. Notably, negative - ion mode CID spectra of hexaacylsucroses (e.g. acylsu crose t r = 62.99 min; #16 in Figure 2.1A), shows weaker fragment ion signals corresponding to C5 and C6 acyl anions (Figure 2.1B) than acylsucroses with fewer acylations under the same CID conditions. Also o f note, heptaacylsucroses were not detected in ne gative ion mode, but were detected at low levels in positive ion mode (Appendix Figure 2.6). The lack of heptaacylsucrose signals in negative - ion mode suggests that at least two non - esterified hydroxyl posit ions facilitate formation of formate adduct ions. In addition, further inspection of XICs for fragment acyl anions of C5 T (Figure 2.1C) and C8 P (Figure 2.1D) revealed a complex series of acylsugars that includes a diverse array of tiglyl and phenylacetyl esters. 2.3.2 GC/MS profiling reveals sugar este r composition As aliphatic ester groups of four or more carbon atoms exhibit isomeric forms th at differ by their branching positions [15] , determining the acyl group composition of acylsugars is aided by gas chromatographic resolution of isomers. This expe rimental approach has been important for distinguishing branching isomers of acyl groups becau se CID spectra often fail to generate fragment ion information that distinguish es branch ing positions. In this study, sugar esters were analyzed via transesterifi cation to generate ethyl esters followed by GC/MS analysis. Acyl group identifications were ba sed on comparisons of their mass spectra to the National Institute of Standards and Technology (NIST11) mass spectrum library, and retention times were referenced to an n - alkane series to generate retention indices. These results demonstrate a high abundan ce of aiC5 and aiC6 acyl groups (Appendix Table 2.7). In comparison, minor signals consistent 41 with iC4, C5 T , iC5 and C8 P were also observed. Taken together, analy sis of LC/MS and GC/MS profiling suggested eight different ester groups: C2, C3, iC4, iC5, aiC 5, C5 T , aiC6, and C8 P (Figure 2.3). Figure 2. 3 . Hydroxyl positions on sucrose available for acylation, and ester groups observed in S. sinuata GT extracts 2.3.3 1D and 2D NMR of purified acylsucroses reveals structural diversity To better understand the diversity of acylsugar metabolites, sixteen abundant acylsugar metabolites (Figure 2.1A and Table 2.1) were purified by HPLC for structur e elucidation by NMR spectroscopy, including eight tetra - substit uted, seven penta - substituted, two of which contain unsaturated esters, and one hexa - substituted acylsugars. 1D ( 1 H and 13 C) and 2D (gCOSY, gHSQC, gHMBC, J - resolved and ROESY) NMR spectroscopi c techniques served as the basis for structure elucidation of pu rified acylsugars. Proton resonances (determined from 1 H, COSY and J - resolved) displayed several informative chemical shift regions (Appendix Table 2.10). Sucrose core hydrogens of acylated po sitions in the range of 4.0 - 5.6 ppm 42 were often apparent due to ~1 ppm downfield shifts compared to hydroxyl containing analogues, wh ich ranged from 3.5 - - 1.9 ppm, while - groups were in the range 2.0 - 2.5 ppm. Similar chemical shift regions and coupling constants were reported for Solanaceae acylsugars with saturated esters [12, 14] . In contrast, characteristic resonances were observed for acylsugars with unsaturated esters . For example, acylsucrose #7 containing a C5 T ester displayed a distinctive multiplet centered at 6.80 ppm for a - H alkene, and tw o overlapping methyl resonances at 1.75 ppm. Acylsucrose #10 containing C8 P ester showed unique multiplets for phenyl - 7.4 ppm and a methylene signal centered at 3.54 ppm (Figure 2.4). COSY and ROESY correlations, alon g with couplings f rom J - on sucrose and atomic connectivity of acyl groups. 43 Figure 2. 4 . HMBC spectrum (top) and correlations of acylsucrose #10 (bottom); 1 H spectru m is projected on the F2 axis and 13 C - represented by colored arrows for clarity. The 13 C, HSQC and HMBC spectra were used to establish atomic connectivity by assignin g carbon chemical shifts of sucrose ring (Appendix Table 2.11) and acyl carbons. HSQC measurements provided carbon attachments to protons and s uggested the number of protons attached to each carbon. HMBC was used to measure atomic connectivity, usually ran ging 2 - 4 bonds, and was vital for the assignment of sucrose and acyl igure 2.4 shows an HMBC spectrum of acylsucrose #10, a pentaacylsucrose with five carbonyl carbon resonances (176.75, 172.11, 176.66, 176.82, 1 71.96 ppm) that show cross peaks for three bond couplings to sucrose r at 4.26 and 4.19 ppm (positions 2 , 3 , 4 , and on sucrose 44 respectively). Similarly, - 30, 2.42 and 2.03 ppm also show cross peaks for two bond correlations to corresponding carbonyl carbons. Additional correlations from 1D and 2D NMR show the se acyl groups are aiC5, C8 P , aiC5, aiC5 and C2 esters at these positions. Characteristic 13 C chemical shifts were also observed for acylsugars with unsaturated esters. For the C8 P ester, 13 C resonances at 135.06, 130.86, 129.82 and 128.47 ppm are consiste nt with phenyl carbons and further confirmed the presence of C8 P ester group. For C5 T , distinc tive 13 C signals at 139.16 and 127.73 ppm for alkene carbons confirmed the presence of a branched monounsaturated ester such as tiglic or angelic acid ( E vs. Z is omer). The - methyl carbon, which showed 13 C chemical shift at 12.08 ppm, agreed with literatu re values for tiglic acid [30] . Table 2. 1 summarizes the structures of NMR resolved acylsucroses purified from S. sinuata extracts (further evidence of these stru ctural assignments is provided in Appendix Tables 2.12 - 2.27 and Appendix Figures 2.31 - 2.141). Of eight hydroxyl positions on sucrose available for acylation (Figure 2.3), the full range of NMR structures identified six of the eight having at least one satu rated or unsaturated esters. The dominant acyl groups in S. sinuata acylsucroses are C2, aiC5 and aiC6. All purified acylsugars contain one or more aiC5 or aiC6 groups, and the pyranose ring was always tri - substituted at 2 , 3 and 4 positions with mostly ai C5 and/or aiC6 acyl chains, though some acylsugars are substituted with C3 and iC4 acylations at position 4 of this ring. In addition, position 3 shows dominant aiC6 acyl substitutions, except for those acylsugars that contain unsaturated acylations of C5 T and C8 P at the 3 position. Finally, the furanose ring is substituted with differing degrees a nd combinations (mono - , di - , tri - ) of C2 and/or aiC5 acylations. The position was acylated with aiC5 or not at all, while the and positions exhibited only C2 acylations or were not acylated. Aside from short chain C2 and C3 acyl groups, all other aliphatic esters were branched iC4, aiC5 and aiC6 acyl chains. 45 Tab le 2. 1 . Summary of NMR resolved acylsucroses purified from S. sinuata extracts and percent peak area of [M+NH 4 ] + ion. Acylsucrose # Acylsugar ID Ret. time (min) Experimental m/z of [M+NH 4 ] + Theoretical m/z of [M+NH 4 ] + Analyte Molecu lar Formula R 2 R 3 R 4 R 1' R 3' R 6' Peak Area (×10 5 ) of [M+NH 4 ] + % of Total Acylsucrose Peak Area Tetraacylsucroses 1 S4:19 :0( 3,5,5,6) 48.21 682.3664 682.3644 C 31 H 52 O 15 aiC5 aiC6 C3 aiC5 H H 3.11 2.8% 2 S4:19 :0( 2,5,6,6) 48.39 682.3675 682.3644 C 31 H 52 O 15 aiC6 aiC6 aiC5 H C2 H 3.52 3.1% 4 S4:20 :0( 4,5,5,6) 50.78 696.3804 696.3801 C 32 H 54 O 15 aiC5 aiC6 iC4 aiC5 H H 0.99 0.9% 5 S4:20 :0( 3,5,6,6) 51.42 696.3815 696.3801 C 32 H 54 O 15 aiC6 aiC6 C3 aiC5 H H 3.02 2.7% 6 S4:20 :0( 2,6,6,6) 51.65 696.3815 696.380 1 C 32 H 54 O 15 aiC6 aiC6 aiC6 H C2 H 2.09 1.9% 9 S4:21 :0( 5,5,5,6) 53.44 710.3993 710.3957 C 33 H 56 O 15 aiC5 aiC6 aiC5 aiC5 H H 8.06 7.2% 12 S4:22 :0( 5,5,6,6) 56.49 724.4154 724.4114 C 34 H 58 O 15 aiC6 aiC6 aiC5 aiC5 H H 9.91 8.8% 14 S4:23 :0( 5,6,6,6) 59.67 738.4266 738.4270 C 35 H 60 O 15 aiC6 aiC6 aiC6 aiC5 H H 2.45 2.2% Pentaacylsucroses 3 S5:20 :0( 2,2,5,5,6) 49.57 710.3621 710.3594 C 32 H 52 O 16 aiC5 aiC6 aiC5 H C2 C2 5.29 4.7% 8 S5:21 :0( 2,2,5,6,6) 52.81 724.3783 724.3750 C 33 H 54 O 16 aiC6 aiC6 ai C5 H C2 C2 10.62 9.5% 11 S5:22 :0( 2,2,6,6,6) 56.08 738.3937 738.3907 C 34 H 56 O 16 aiC6 aiC6 aiC6 H C2 C2 4.93 4.4% 13 S5:23 :0( 2,5,5,5,6) 58.80 752.4092 752.4063 C 35 H 58 O 16 aiC5 aiC6 aiC5 aiC5 H C2 6.98 6.2% 15 S5:24 :0( 2,5,5,6,6) 61.76 766.4236 766.4220 C 36 H 6 0 O 16 aiC6 aiC6 aiC5 aiC5 H C2 4.90 4.4% Pentaacylsucroses w/unsaturated ester 7 S5:22 :1( 2,5,5,5,5 T ) 52.55 736.3743 736.3750 C 34 H 54 O 16 aiC5 C5 T aiC5 aiC5 H C2 2.84 2.5% 10 S5:25 :4( 2,5,5,5,8 P ) 54.29 772.3748 772.3750 C 37 H 54 O 16 ai C5 C8 P aiC5 aiC5 H C2 0.78 0.7% Hexaacylsucrose 16 S6:25 :0( 2,2,5,5,5,6) 62.99 794.4166 794.4169 C 37 H 60 O 17 aiC5 aiC6 aiC5 aiC5 C2 C2 2.34 2.1% % Acylsucrose Peak Area with Structural Identification = 64.1% 46 2.3.4 Structure d iversity of acylsucroses from S. sinuat a Our broader goal is to understand the evolution of genes involved in plant specialized metabolism, focusing on the family Solanaceae. S. sinuata was chosen for this investigation because it is further removed genet ically from tomato and its close relatives for which extensive acylsugar profiling has been performed. To this end, this report describes the isolation and structure elucidation of 16 acylsugars from S. sinuata , relying on NMR spectroscopy to establish cit es of attachment of specific acyl groups. Interpretation of these findings relies on comparison to the published acylsugar structures. 2.3.4.1 Acyl group diversity in S. sinuata Acyl groups aiC5 and aiC6 were the dominant medium chain ester groups, consist ent with isoleucine serving as the major a cyl group precursor as has been observed in other members of the Solanaceae [31] . At least 10 - fold lower levels of iC4 and iC5 were detected, and fragment ion evidence was found for aliphatic esters up to C6. While C2 esters were abundant, lesser amounts o f C3 esters were observed. The observation of acylsugars containing unsaturated C5 T and C8 P in the Solanaceae has not been described before. We postulate that C5 T is derived from isoleucine pathway metabolism and C 8 P arises from phenylalanine by chemistry similar to other amino acids. To our knowledge, the only prior report of an unsaturated C5 acylsugar was in tobacco [24] , and we are not aware of C8 P in acylsuga rs from prior reports. Amazingly, more than 400 acyl sugars were annotated by LC/MS (Appendix Table 2.4) despite only eight of the > 20 acyl groups known in acylsugars of the Solanaceae [15] being detected. This is slightly more than the > 300 peaks detect ed using automated processing of LC/MS data using a steeper gradient [21] , and represents a more diverse group of acylsucroses than has been reported for any plant, with the 73 acylsugars annotated from S. habrochaites being the largest number described to date [32] . In contrast to an earlier report of an a cylglucose identified from S. sinuata [33] , we generated extracted ion chromatograms but found no evidence of acylglucoses, only acylsucroses. 47 2.3.4.2 Number of acylations in S. sinuata acylsucroses The S. sinuata acylsucroses are notably different from t hose produced by cultivated and wild tomato in that the maximum number of acyl groups (7) exceeds the number reported in any accessions in the genera Solanum , Nicotiana , or Petunia . [12, 14] . In contrast, Petunia species show high levels of malonylated pen taacylsucroses [14] . Here, we present NMR structures for as many as six acyl groups on sucrose from S. sinuata , and deep mining of LC/MS data suggests that S. sinuata produces detectable lev els of tri - through heptaacylsucroses (Appendix Table 2.4). Penta - and hexaacylsucroses were of high abundance in S. sinuata (Appendix Figure 2.7) and their structures led to identifications of S. sinuata ASATs in our recent manuscript [21] . 2.3.4.3 Posit ions of acyl groups in S. sinuata acylsucroses The diversity in po sition selectivity of acyl group attachment to the sucrose core is revealed from NMR spectra. S. sinuata acylsucroses are distinguished from those accumulated by tomato and its relatives in that C2 acylation in S. sinuata was found at the and positio ns, but only at the 2 position in tomato [12, 13] . No other acyl groups were observed at the or positions of S. sinuata acylsucroses, in contrast to tomato species, which often have iC 5, iC10, aiC11 or nC12 at the position. Our recent paper found evidence that SsASAT5 catalyzes acetylation at the position as well as positions 6 and [21] . Although less comprehensive, other reports of positions of acyl groups outside of the tomato clade show varied attachment selectivity. Reported wild potato ( Solanum berthaultii ) acylsucroses contrast with S. sinuata in that the former are not known to exhibit acylation at the 2 position, and exhibit acylation by iC4 at the position and C 10 esters at the 6 position [34, 35] . In a previous report, structures of one triacylglucose and one pentaacylsucrose that displayed C2 acylation at the 6 position of sucrose were elucidated from S. sinuata [33] . However, we did not observe any acylsugars substituted at this position from our NMR results. Interestingly, XICs of positive mode spectra at elevated collision potentials did show evidence of low abundance heptaacylsucroses with fully substituted furanose ring at the , , and positions (A ppendix Figure 2.6). As with S. sinuata , reports of petunia acylsugars described at least four 48 tetraacylsucroses substituted with C2 acyl groups at the position [36 - 38] , but a recent manuscript from our laboratory yielded e vidence that these may form by decarboxylation of malonate esters at the position [14] . Petunia species also exhibit iC4, aiC5 and aiC6 esters, in addition to straight chain C6, C7 and iso - branched iC5, iC6, iC7 and iC8 esters at the 2 , 3 , and 4 positio ns. Tobacco species produce ac ylsugars that bear resemblance to those found in S. sinuata [24, 39] . These acylsucroses exhibit extensive aiC6, aiC5 and iC4 at the 2 , 3 or 4 positions (relative positions are uncertain), including C2 groups at the positio n. In addition, acylglucoses f rom these species were found to contain C3 and unsaturated C5 esters, which are presumed to be C5 T [24] . 2.4 Conclusions Identifying novel metabolites from complex profiles continues to present a central challenge in function al genomic investigations of metabolism. In this work, LC/MS profiling and NMR spectroscopy of acylsugar SMs from GTs of S. sinuata revealed diverse acy lsucrose compositions. In addition to documentation of hundreds of acylsucrose metabolites, documentatio n of unsaturated ester groups C5 T and C8 P extends our understanding of acylsugar diversity in the Solanaceae, where only saturated forms were previously reported. Acylsucrose structures provided evidence for enzyme mediated and position - selective attachmen t of ester groups to the sucrose core that differed from reports for other members of the Solanaceae [21] , including evidence for as many as seven acyla tions on a single sucrose core. Considering the roles that acylsugars have in physical and chemical def ense, understanding the mechanisms involved in their formation may provide substantial guidance for engineering acylsugar content of plants. The results herein guided the discovery of four acylsugar biosynthetic enzymes of the S. sinuata pathway and extend ed our understanding of the evolution of specialized metabolism in the Solanaceae. We postulate that promiscuity is an important feature of these SM enz ymes, and thus, additional enzymes that acylate with unsaturated acyl - CoA groups may not be necessary. 49 APPENDIX 50 Figure 2. 5 . Micrographs of S. sinuata Type I/IV glandular trichomes located on the surface of a young leaflet. 51 Figure 2. 6 . LC/MS shows evidence for heptaacylsucrose S7:27:0 using positive - ion mode ESI but negligible signal is detected in negative - ion mode. (A) Positive - ion mode extracted ion chromatogram for [M+NH 4 ] + ( m/z 836.427) for S7:27:0 showing one major and three minor chromatographic peaks; (B) Negative - ion mode chromatogram for [M+ formate] - ( m/z 863.392) on the same absolute y - axis scale (100% = 1.08 x 10 4 ion counts) shows negligible signal for S7:27:0. The structure shown is putative and not solved by NMR, but is proposed based on our knowledge of acyl group selectivity (establish ed by NMR resolved structures of other acylsucroses). The [M+NH 4 ] + consi stent with S7:27:0 and fragment ion m/z 373.15 (Function 3 = 25 V) are consistent with an acylsucrose that has a furanose ring substituted by one C5 and three C2 ester groups. 52 Figu re 2. 7 . Sum of extracted ion peak areas of [M+NH 4 ] + ions for S. sinuata acylsucroses, organized by the number of acyl groups on the sucrose core (e.g. S4, S5, and S6 refer to tetra - , penta - , and hexa - acylsucroses). Separate bars i ndicate peak area totals for acylsucroses containing all saturated a cyl groups (green) and one or two C5 T and C8 P acyl groups (blue), which accounted for ~14% of the total acylsugar peak area. 53 Table 2. 2 . Plant cultivation and meta bolite extraction metadata Species Salpiglossis sinuata Genotype NYBG, S. sinuata seeds were obtained from a single plant growing in The New York Botanical Gardens (NY, USA). The second generation of these seeds were used for the experiments in this study . Organ Aerial tissues Organ specification Aerial tissu es included leaf and stem tissues Cell type Glandular trichomes Growth location Growth chamber at MSU Growth support Seeds were germinated using 500 µM gibberellic acid on Whatman #1 filter paper placed in petri dishes. 3 - 4 day old seedlings were transferred to Jiffy peat pellets. Light 300 µE m - 2 sec - 1 ; 16 h light/8 h dark Humidity 50% relative humidity Temperature 25C day/12C night Watering regime Bottom watering as per requirement Nutritio nal regime solution, once a week Plant growth stage 10 weeks post germination (plants about 0.5 m height) Metabolism quenching method Aerial tissues were extracted in 1.9 L of acetonitrile: isopropanol (AcN:IPA, v/v, 1:1) for 10 mins in a 2 - L beaker with horizontal mixing at 120 rpm. The presence of acylsugars in trichomes were confirmed by dipping leaf and stem tissue into liquid N 2 , and scraping trichomes into a solution of AcN:IPA. Harvest method Plants were cut at the stems in 10 - 15 cm increments Sample storage The extract was decanted into two 1 L glass Wheaton bottles with Teflon lined caps, and stored in a freezer at - 20°C. 54 Table 2. 3 . UHPLC/MS metadata Facility Director Dr. A. Daniel Jones An alyst Steven M. Hurney Instrument Location MSU Mass Spectrometry and Metabolomics Core Facility Instrument Title LCMS: G2 - XS QTof #1 LC System Acquity UPLC I - Class Binary Solvent Manager equipped with Acquity Column Manager Manufacturer Waters Autosam pler 2777C Sample Manager Column Ascentis Express C18 Analytical HPLC, 10 cm x 2.1mm x 2.7µm Column Manufacturer Supelco Catalogue Number 53823 - U Serial Number USRB002977 Packing Lot Number S11020 Injection Volume 10 µL Flow Rate 0.3 mL/min Mobile Phases: A 10 mM ammonium formate in water (pH 2.8, adjusted with formic acid) B Acetonitrile Gradient Profile Hold 1% B at 0 - 1 min, linear 1 - 100% B at 1 - 100 min, hold 100% B at 100 - 107 min, linear 100 - 1% B at 107 - 108 min, and hold 1% B at 108 - 110 min C olumn Oven Temperature 50 ºC Sample Temperature in autosampler 10 °C Inlet Method Name SMH_110min Mass Spectrometer Xevo G2 - XS QTof Manufacturer Waters Software MassLynx v4.1 Ionization Source Electrospray Ionization (ESI) Data Acquisition Sensitivi ty Mode Polarity Positive, Negative Mass Range m/z 50 - 1500 Data Format Centroid Capillary Voltage 2.0 kV Sample Cone 35 V Source Temperature 100 °C Source Offset 80 V Desolvation Temperature 350 °C Cone Gas Flow 50 L/hr (ESI+), 0 L/hr (ESI - ) 55 Tab le 2.3. (continued) Desolvation Gas Flow 600 L/hr Collision Potential Function 1 0 V Function 2 10 V Function 3 25 V Function 4 50 V Scan Duration 0.10 s Inter Scan Delay 0.014 s Collision Cell Pressure 0.06 mbar (ESI+), 0.05 mbar (ESI - ) Lock Spr ay Leu - enkephalin Lock mass ( m/z ) 556.2771 (ESI+), 554.2615 (ESI - ) Lock Spray Scan Time 0.2 s Lock Spray Scan Frequency 10 s MS Method Files SMH_CID_pos_110 min_Lock , SMH_CID_neg_110 min_Lock Sample handling Aerial tissues from 28 plants aged 10 wee ks were harvested (plants were cut at the stems in 10 - 15 cm increments) and extracted in 1.9 L of AcN:IPA for 10 mins in a 2 L beaker with horizontal mixing at 120 rpm. A 10 mL aliquot of this bulk extract solution, with 5.0 µM telmisartan added as interna l standard, was used for all metabolite profiling. Sample Storage - 20 °C in Wheaton glass vessel with PTFE lined cap Protocol when analyzing the samples The instrument was calibrated in ESI+/ - modes using 500 µM sodium formate solution. First, samples we re analyzed in ESI - mode. After column equilibration, a solvent blank sample (AcN:IPA) was analyzed, followed by S. sinuata extract. The two samples were t hen analyzed in ESI+ mode. Deep profiling data analysis parameters: XIC Window 40 ppm Enabled Smo othing Yes Window Size 10 scans Number of Smooths 3 Smoothing Method Savitzky - Golay ApexTrack Peak Integration Yes Peak - to - Peak Baseline Noise Automatic Peak Width at 5% Height (mins) Automatic Baseline Start Threshold 1% Baseline End Threshold 1% Note: some peak integrations were adjusted manually 56 Table 2. 4 . Deep profiling results (metabolites highlighted in bold were identified by NMR spectroscopy) Elution Order Acylsucrose ID # of Isomers [M+NH 4 ] + m/z XIC (±20 p pm) Ret. Time (min) Peak Height Peak Area Triacylsucroses 1 S3:13:0 2 584.291 27.783 5698 895 2 S3:13:0 2 584.291 27.920 5691 815 3 S3:14:0 2 598.307 31.603 15935 3040 4 S3:14:0 2 598.307 32.578 875 139 9 S3:15:0 6 612.323 33.010 7136 1331 10 S3:15: 0 6 612.323 33.494 590 89 11 S3:15:0 6 612.323 33.857 1739 286 12 S3:15:0 6 612.323 35.444 431 77 13 S3:15:0 6 612.323 37.260 747 108 14 S3:15:0 6 612.323 38.682 2180 320 22 S3:16:0 4 626.338 36.398 27512 4412 23 S3:16:0 4 626.338 38.341 6103 956 24 S3:16:0 4 626.338 40.656 8684 1439 25 S3:16:0 4 626.338 41.117 7793 1509 39 S3:17:0 4 640.354 43.697 84074 13052 40 S3:17:0 4 640.354 43.999 6488 1202 41 S3:17:0 4 640.354 44.316 4210 726 42 S3:17:0 4 640.354 44.574 944 153 69 S3:18:0 1 654.370 47.0 01 44022 7234 5 S3:15:1 2 610.307 32.624 375 74 6 S3:15:1 2 610.307 35.505 828 127 28 S3:17:1 1 638.338 47.622 1057 175 51 S3:18:1 1 652.354 51.018 634 107 80 S3:19:1 1 666.370 52.114 1238 208 15 S3:16:4 2 618.276 29.077 1912 289 16 S3:16:4 2 618.27 6 29.220 2004 275 45 S3:18:4 2 646.307 33.849 2039 358 46 S3:18:4 2 646.307 38.024 708 99 71 S3:19:4 4 660.323 37.729 1278 184 72 S3:19:4 4 660.323 40.247 1003 151 73 S3:19:4 4 660.323 41.230 1012 143 74 S3:19:4 4 660.323 41.624 1022 216 99 S3:20:4 2 674.338 43.379 1815 260 100 S3:20:4 2 674.338 44.354 27660 4376 133 S3:22:8 1 694.307 41.193 2423 367 57 Table 2.4. (continued) Elution Order Acylsucrose ID # of Isomers [M+NH 4 ] + m/z XIC (±20 ppm) Ret. Time (min) Peak Height Peak Area Tetraacylsu croses 7 S4:14:0 2 612.286 27.209 699 126 8 S4:14:0 2 612.286 27.481 920 151 19 S4:15:0 3 626.302 31.096 2619 662 20 S4:15:0 3 626.302 33.146 6268 1342 21 S4:15:0 3 626.302 35.959 719 125 30 S4:16:0 9 640.318 34.878 526 86 31 S4:16:0 9 640.318 35.55 9 915 169 32 S4:16:0 9 640.318 36.723 3273 582 33 S4:16:0 9 640.318 37.131 908 155 34 S4:16:0 9 640.318 38.281 511 77 35 S4:16:0 9 640.318 38.674 9308 1529 36 S4:16:0 9 640.318 39.703 11615 1863 37 S4:16:0 9 640.318 39.998 1128 175 38 S4:16:0 9 640. 318 40.142 1247 196 56 S4:17:0 13 654.333 38.205 26244 4426 57 S4:17:0 13 654.333 38.576 710 152 58 S4:17:0 13 654.333 38.977 1479 269 59 S4:17:0 13 654.333 39.461 923 144 60 S4:17:0 13 654.333 40.035 725 114 61 S4:17:0 13 654.333 41.133 550 69 62 S 4:17:0 13 654.333 41.836 7274 1259 63 S4:17:0 13 654.333 42.146 9669 2010 64 S4:17:0 13 654.333 42.441 8626 1791 65 S4:17:0 13 654.333 42.691 3578 402 66 S4:17:0 13 654.333 43.031 48568 9418 67 S4:17:0 13 654.333 44.370 705 133 68 S4:17:0 13 654.333 44.800 391 79 88 S4:18:0 7 668.349 41.450 17944 2885 89 S4:18:0 7 668.349 42.932 8737 1422 90 S4:18:0 7 668.349 43.704 30739 4900 91 S4:18:0 7 668.349 45.163 431891 91237 92 S4:18:0 7 668.349 45.678 125428 39088 93 S4:18:0 7 668.349 46.125 18917 4981 94 S4:18:0 7 668.349 47.523 6581 1122 118 S4:19:0 7 682.364 47.001 19685 3248 119 S4:19:0 7 682.364 47.282 9253 1445 120 S4:19:0 7 682.364 47.599 17753 2456 121 S4:19:0(3,5,5,6) 7 682.364 48.211 1752487 311272 58 Table 2.4. (continued) Elution Order Ac ylsucrose ID # of Isomers [M+NH 4 ] + m/z XIC (±20 ppm) Ret. Time (min) Peak Height Peak Area 122 S4:19:0(2,5,6,6) 7 682.364 48.393 2052739 351859 123 S4:19:0 7 682.364 48.869 222575 47411 124 S4:19:0 7 682.364 50.813 3345 558 153 S4:20:0 5 696.380 5 0.291 681323 136081 154 S4:20:0(4,5,5,6) 5 696.380 50.783 524859 98549 155 S4:20:0(3,5,6,6) 5 696.380 51.418 1552608 301912 156 S4:20:0(2,6,6,6) 5 696.380 51.645 1319806 209027 157 S4:20:0 5 696.380 52.364 1771 298 184 S4:21:0(5,5,5,6) 4 710.396 53.43 8 3509795 806099 185 S4:21:0 4 710.396 53.899 1343764 240461 186 S4:21:0 4 710.396 54.263 10410 1101 187 S4:21:0 4 710.396 54.611 166994 27295 225 S4:22:0(5,5,6,6) 3 724.411 56.494 4000121 991011 226 S4:22:0 3 724.411 57.053 276237 51593 227 S4:22:0 3 724.411 57.401 7417 968 259 S4:23:0(5,6,6,6) 2 738.427 59.670 1282562 245010 260 S4:23:0 2 738.427 60.124 45298 9220 288 S4:24:0 1 752.443 62.756 100656 19118 17 S4:15:1 2 624.286 31.678 1158 156 18 S4:15:1 2 624.286 31.777 1194 156 26 S4:16:1 2 63 8.302 34.500 327 65 27 S4:16:1 2 638.302 37.010 403 68 47 S4:17:1 4 652.318 35.725 526 94 48 S4:17:1 4 652.318 37.238 7592 1267 49 S4:17:1 4 652.318 39.211 1323 237 50 S4:17:1 4 652.318 40.437 3667 548 77 S4:18:1 3 666.333 42.366 80341 13161 78 S4:1 8:1 3 666.333 43.220 1236 241 79 S4:18:1 3 666.333 43.697 530 83 105 S4:19:1 7 680.349 44.475 3920 772 106 S4:19:1 7 680.349 44.733 9162 1507 107 S4:19:1 7 680.349 45.133 3149 467 108 S4:19:1 7 680.349 45.299 10693 1254 109 S4:19:1 7 680.349 45.542 3 3346 8161 110 S4:19:1 7 680.349 45.852 22612 3807 111 S4:19:1 7 680.349 46.419 1297 254 138 S4:20:1 5 694.364 47.138 317379 53412 139 S4:20:1 5 694.364 47.607 5991 1034 140 S4:20:1 5 694.364 47.856 12153 2086 59 Table 2.4. (continued) Elution Order Acy lsucrose ID # of Isomers [M+NH 4 ] + m/z XIC (±20 ppm) Ret. Time (min) Peak Height Peak Area 141 S4:20:1 5 694.364 48.326 22296 4448 142 S4:20:1 5 694.364 48.710 2359 463 172 S4:21:1 4 708.380 50.209 261538 50055 173 S4:21:1 4 708.380 50.852 101670 1 8843 174 S4:21:1 4 708.380 51.282 1263 200 175 S4:21:1 4 708.380 51.570 1888 304 208 S4:22:1 4 722.396 53.037 67327 11935 209 S4:22:1 4 722.396 53.347 15275 2827 210 S4:22:1 4 722.396 53.582 9632 1497 211 S4:22:1 4 722.396 54.051 4834 1374 75 S4:18: 2 1 664.318 40.649 4108 621 132 S4:20:2 1 692.349 45.360 9900 1608 43 S4:17:4 2 646.271 28.622 403 73 44 S4:17:4 2 646.271 28.857 455 82 70 S4:18:4 1 660.286 34.182 1944 474 95 S4:19:4 4 674.302 36.314 162 28 96 S4:19:4 4 674.302 36.912 256 35 97 S4 :19:4 4 674.302 39.052 303 45 98 S4:19:4 4 674.302 39.635 1018 134 125 S4:20:4 6 688.318 38.712 4884 817 126 S4:20:4 6 688.318 39.733 2170 353 127 S4:20:4 6 688.318 41.381 4942 758 128 S4:20:4 6 688.318 41.957 1939 481 129 S4:20:4 6 688.318 42.131 13 95 194 130 S4:20:4 6 688.318 42.743 1865 519 160 S4:21:4 8 702.333 43.129 592 115 161 S4:21:4 8 702.333 43.735 698 109 162 S4:21:4 8 702.333 44.408 30423 5201 163 S4:21:4 8 702.333 44.718 115687 18449 164 S4:21:4 8 702.333 44.937 8303 1032 165 S4:21 :4 8 702.333 45.565 24427 5952 166 S4:21:4 8 702.333 45.905 5795 1020 167 S4:21:4 8 702.333 47.901 725 93 193 S4:22:4 4 716.349 46.858 7578 1697 194 S4:22:4 4 716.349 47.478 18341 3312 195 S4:22:4 4 716.349 47.819 114431 19553 196 S4:22:4 4 716.349 4 8.628 99948 17053 237 S4:23:4 4 730.364 49.209 175130 31121 238 S4:23:4 4 730.364 49.905 35714 7685 60 Table 2.4. (continued) Elution Order Acylsucrose ID # of Isomers [M+NH 4 ] + m/z XIC (±20 ppm) Ret. Time (min) Peak Height Peak Area 239 S4:23:4 4 730 .364 50.246 19092 3845 240 S4:23:4 4 730.364 51.865 10483 1955 267 S4:24:4 4 744.380 52.190 294850 49202 268 S4:24:4 4 744.380 52.742 180719 33086 269 S4:24:4 4 744.380 53.718 48407 11307 270 S4:24:4 4 744.380 54.240 19602 3512 297 S4:25:4 4 758.396 55.177 26389 4796 298 S4:25:4 4 758.396 55.821 7229 1539 299 S4:25:4 4 758.396 56.728 159085 28758 300 S4:25:4 4 758.396 57.265 1592 293 330 S4:26:4 1 772.411 59.776 10590 2229 230 S4:23:5 1 728.349 46.858 9117 1552 262 S4:24:5 1 742.364 50.708 3703 718 243 S4:24:8 2 736.318 44.823 2078 309 244 S4:24:8 2 736.318 45.020 4003 594 61 Table 2.4. (continued) Elution Order Acylsucrose ID # of Isomers [M+NH 4 ] + m/z XIC (±20 ppm) Ret. Time (min) Peak Height Peak Area Pentaacylsucroses 29 S5:15:0 1 640. 281 33.668 251 66 52 S5:16:0 4 654.297 35.876 203 45 53 S5:16:0 4 654.297 36.587 229 36 54 S5:16:0 4 654.297 37.744 372 63 55 S5:16:0 4 654.297 38.364 706 93 81 S5:17:0 7 668.312 36.882 338 65 82 S5:17:0 7 668.312 37.676 329 66 83 S5:17:0 7 668.312 39.551 15768 2761 84 S5:17:0 7 668.312 40.399 7074 1275 85 S5:17:0 7 668.312 40.633 2776 473 86 S5:17:0 7 668.312 40.951 1796 384 87 S5:17:0 7 668.312 41.186 677 104 112 S5:18:0 6 682.328 42.978 5011 731 113 S5:18:0 6 682.328 43.098 5809 998 114 S5: 18:0 6 682.328 43.288 3256 408 115 S5:18:0 6 682.328 43.817 152580 26965 116 S5:18:0 6 682.328 44.060 89101 15755 117 S5:18:0 6 682.328 45.020 5186 790 144 S5:19:0 9 696.344 45.678 4394 601 145 S5:19:0 9 696.344 46.207 153189 26924 146 S5:19:0 9 696. 344 46.435 40705 6064 147 S5:19:0 9 696.344 46.714 161029 30770 148 S5:19:0 9 696.344 47.198 356435 79325 149 S5:19:0 9 696.344 47.773 7117 1350 150 S5:19:0 9 696.344 48.174 4093 758 151 S5:19:0 9 696.344 48.544 3645 597 152 S5:19:0 9 696.344 48.915 3619 709 176 S5:20:0 8 710.359 49.150 265038 47394 177 S5:20:0(2,2,5,5,6) 8 710.359 49.573 2410053 528518 178 S5:20:0 8 710.359 49.989 1325229 279439 179 S5:20:0 8 710.359 50.488 449698 92180 180 S5:20:0 8 710.359 50.791 38080 9636 181 S5:20:0 8 710. 359 51.351 4877 924 182 S5:20:0 8 710.359 51.850 11982 2428 183 S5:20:0 8 710.359 52.152 13242 3344 216 S5:21:0(2,2,5,6,6) 9 724.375 52.810 4393749 1062497 217 S5:21:0 9 724.375 53.226 505090 120334 218 S5:21:0 9 724.375 53.551 956454 157126 62 Table 2. 4. (continued) Elution Order Acylsucrose ID # of Isomers [M+NH 4 ] + m/z XIC (±20 ppm) Ret. Time (min) Peak Height Peak Area 219 S5:21:0 9 724.375 53.756 61114 7127 220 S5:21:0 9 724.375 53.997 7332 1260 221 S5:21:0 9 724.375 54.322 2695 341 222 S5:2 1:0 9 724.375 54.890 2267 466 223 S5:21:0 9 724.375 55.117 2832 489 224 S5:21:0 9 724.375 55.420 5586 1000 252 S5:22:0 7 738.391 54.905 4914 795 253 S5:22:0 7 738.391 55.110 2344 295 254 S5:22:0 7 738.391 55.752 1388972 281505 255 S5:22:0(2,2,6,6,6) 7 738.391 56.077 2131308 492714 256 S5:22:0 7 738.391 56.668 324870 55041 257 S5:22:0 7 738.391 57.559 34545 8165 258 S5:22:0 7 738.391 58.081 4282 932 281 S5:23:0 7 752.406 58.173 270116 47058 282 S5:23:0(2,5,5,5,6) 7 752.406 58.800 2927149 697683 2 83 S5:23:0 7 752.406 59.201 632895 113993 284 S5:23:0 7 752.406 59.504 5069 630 285 S5:23:0 7 752.406 59.746 27467 4375 286 S5:23:0 7 752.406 60.479 9031 1506 287 S5:23:0 7 752.406 60.699 29338 5267 318 S5:24:0 5 766.422 60.494 2909 521 319 S5:24:0 5 766.422 61.515 293690 51296 320 S5:24:0(2,5,5,6,6) 5 766.422 61.758 2166846 489817 321 S5:24:0 5 766.422 62.219 93848 21032 322 S5:24:0 5 766.422 63.762 3372 671 345 S5:25:0 4 780.438 64.117 9318 2095 346 S5:25:0 4 780.438 64.504 43629 6816 347 S5:2 5:0 4 780.438 64.760 197258 44599 348 S5:25:0 4 780.438 65.139 4903 982 374 S5:26:0 3 794.453 66.938 40387 8410 375 S5:26:0 3 794.453 67.347 4532 1040 376 S5:26:0 3 794.453 67.680 16341 3146 396 S5:27:0 1 808.469 69.760 41077 8377 405 S5:28:0 2 822.4 85 72.339 1106 168 406 S5:28:0 2 822.485 72.535 1922 419 76 S5:17:1 1 666.297 37.744 1365 214 101 S5:18:1 4 680.312 40.611 758 143 102 S5:18:1 4 680.312 41.079 1053 211 103 S5:18:1 4 680.312 41.867 3809 810 63 Table 2.4. (continued) Elution Order Acylsu crose ID # of Isomers [M+NH 4 ] + m/z XIC (±20 ppm) Ret. Time (min) Peak Height Peak Area 104 S5:18:1 4 680.312 42.192 6716 1169 134 S5:19:1 4 694.328 43.341 45422 7384 135 S5:19:1 4 694.328 43.764 2531 579 136 S5:19:1 4 694.328 44.150 1122 142 137 S5:19:1 4 694.328 44.959 4218 829 169 S5:20:1 3 708.344 46.578 59103 12792 170 S5:20:1 3 708.344 46.926 5197 715 171 S5:20:1 3 708.344 47.720 337358 63440 202 S5:21:1 6 722.359 49.557 40983 7908 203 S5:21:1 6 722.359 49.792 30656 3756 204 S5:21:1 6 7 22.359 49.928 38048 6191 205 S5:21:1 6 722.359 50.148 83474 16724 206 S5:21:1 6 722.359 50.836 41387 11896 207 S5:21:1 6 722.359 51.660 4663 788 246 S5:22:1(2,5,5,5,5 T ) 4 736.375 52.553 1526524 283707 247 S5:22:1 4 736.375 52.999 24941 5713 248 S5:22 :1 4 736.375 53.196 39069 8149 249 S5:22:1 4 736.375 53.498 13193 4292 273 S5:23:1 4 750.391 55.571 474547 87436 274 S5:23:1 4 750.391 56.016 89748 16109 275 S5:23:1 4 750.391 56.289 1823 207 276 S5:23:1 4 750.391 56.592 3047 395 306 S5:24:1 5 764.40 6 58.339 31285 6439 307 S5:24:1 5 764.406 58.603 20318 3552 308 S5:24:1 5 764.406 58.884 1835 269 309 S5:24:1 5 764.406 59.087 2436 575 310 S5:24:1 5 764.406 59.504 1644 281 338 S5:25:1 2 778.422 60.873 9866 1825 339 S5:25:1 2 778.422 61.259 2892 658 368 S5:26:1 2 792.438 63.785 2461 511 369 S5:26:1 2 792.438 64.163 1062 189 131 S5:19:2 1 692.312 41.911 1032 178 168 S5:20:2 1 706.328 45.769 33592 5785 198 S5:21:2 3 720.344 48.159 5478 1167 199 S5:21:2 3 720.344 48.703 757 129 200 S5:21:2 3 720. 344 49.255 664 92 241 S5:22:2 1 734.359 50.526 97918 17205 158 S5:20:4 2 702.297 40.270 12737 2160 159 S5:20:4 2 702.297 40.679 516 81 64 Table 2.4. (continued) Elution Order Acylsucrose ID # of Isomers [M+NH 4 ] + m/z XIC (±20 ppm) Ret. Time (min) Peak Height Peak Area 188 S5:21:4 5 716.312 42.804 5507 978 189 S5:21:4 5 716.312 43.213 1654 280 190 S5:21:4 5 716.312 43.507 14257 2253 191 S5:21:4 5 716.312 44.090 3708 835 192 S5:21:4 5 716.312 44.611 1273 218 231 S5:22:4 6 730.328 45.414 177030 2872 8 232 S5:22:4 6 730.328 45.928 32721 7624 233 S5:22:4 6 730.328 46.299 1385 226 234 S5:22:4 6 730.328 46.699 2823 395 235 S5:22:4 6 730.328 46.934 12787 2575 236 S5:22:4 6 730.328 47.228 3645 566 263 S5:23:4 4 744.344 48.492 516258 85398 264 S5:23:4 4 744.344 48.953 10415 2311 265 S5:23:4 4 744.344 49.785 379584 82590 266 S5:23:4 4 744.344 50.012 13056 1245 292 S5:24:4 5 758.359 51.592 80150 13944 293 S5:24:4 5 758.359 51.955 41327 8281 294 S5:24:4 5 758.359 52.144 10803 1501 295 S5:24:4 5 758. 359 52.704 58779 11180 296 S5:24:4 5 758.359 52.871 56335 8776 326 S5:25:4 4 772.375 53.551 5425 885 327 S5:25:4(2,5,5,5,8 P ) 4 772.375 54.293 439869 77683 328 S5:25:4 4 772.375 54.974 61766 17401 329 S5:25:4 4 772.375 55.865 2445 438 356 S5:26:4 6 78 6.391 56.637 7389 1084 357 S5:26:4 6 786.391 57.234 233892 43948 358 S5:26:4 6 786.391 57.605 175200 30339 359 S5:26:4 6 786.391 57.930 3136 588 360 S5:26:4 6 786.391 58.725 2224 470 361 S5:26:4 6 786.391 58.997 2821 469 383 S5:27:4 4 800.406 60.116 6104 1177 384 S5:27:4 4 800.406 60.502 1840 551 385 S5:27:4 4 800.406 61.485 5610 1337 386 S5:27:4 4 800.406 61.682 2248 346 409 S5:29:4 2 828.438 65.040 1503 292 410 S5:29:4 2 828.438 66.689 528 113 417 S5:30:4 1 842.453 69.381 434 72 228 S5:22:5 2 728.312 43.402 1111 192 229 S5:22:5 2 728.312 43.704 1131 185 65 Table 2.4. (continued) Elution Order Acylsucrose ID # of Isomers [M+NH 4 ] + m/z XIC (±20 ppm) Ret. Time (min) Peak Height Peak Area 261 S5:23:5 1 742.328 47.312 11112 3176 289 S5:24:5 1 756.344 49.498 5625 1086 323 S5:25:5 1 770.359 51.827 58791 10975 242 S5:23:8 1 736.281 40.920 444 83 301 S5:25:8 1 764.312 45.995 13080 2449 331 S5:26:8 3 778.328 48.884 5016 839 332 S5:26:8 3 778.328 49.490 6146 1002 333 S5:26:8 3 778.328 49.747 94 2 128 362 S5:27:8 1 792.344 51.592 6472 1052 387 S5:28:8 1 806.359 53.846 18840 3149 66 Table 2.4. (continued) Elution Order Acylsucrose ID # of Isomers [M+NH 4 ] + m/z XIC (±20 ppm) Ret. Time (min) Peak Height Peak Area Hexaacylsucroses 143 S6:18:0 1 696.307 42.358 1664 272 212 S6:20:0 4 724.339 47.917 3776 670 213 S6:20:0 4 724.339 48.144 1247 184 214 S6:20:0 4 724.339 48.938 26325 5056 215 S6:20:0 4 724.339 49.309 2004 317 250 S6:21:0 2 738.354 51.758 6158 1234 251 S6:21:0 2 738.354 52.198 107 161 19747 277 S6:22:0 4 752.370 54.519 40410 10452 278 S6:22:0 4 752.370 54.966 45185 10571 279 S6:22:0 4 752.370 55.465 87876 16719 280 S6:22:0 4 752.370 57.038 14385 3932 311 S6:23:0 7 766.386 57.234 16777 3453 312 S6:23:0 7 766.386 57.651 271418 6 9452 313 S6:23:0 7 766.386 57.938 36258 4343 314 S6:23:0 7 766.386 58.203 38668 7533 315 S6:23:0 7 766.386 59.201 6439 1526 316 S6:23:0 7 766.386 59.739 5688 1520 317 S6:23:0 7 766.386 60.207 3750 742 340 S6:24:0 5 780.401 59.859 67633 18369 341 S6: 24:0 5 780.401 60.396 213736 45630 342 S6:24:0 5 780.401 60.880 182116 39975 343 S6:24:0 5 780.401 62.324 61875 12977 344 S6:24:0 5 780.401 62.809 5490 1165 370 S6:25:0(2,2,5,5,5,6) 4 794.417 62.990 1109307 233661 371 S6:25:0 4 794.417 63.527 212097 4 3297 372 S6:25:0 4 794.417 63.875 12480 2476 373 S6:25:0 4 794.417 65.335 38955 7662 393 S6:26:0 3 808.433 66.068 608240 126909 394 S6:26:0 3 808.433 66.416 25085 6380 395 S6:26:0 3 808.433 68.255 2640 488 404 S6:27:0 1 822.448 68.928 54212 11668 41 3 S6:28:0 2 836.464 70.576 520 90 414 S6:28:0 2 836.464 71.688 2224 438 419 S6:29:0 1 850.480 73.450 921 168 201 S6:20:1 1 722.323 46.261 14991 2597 245 S6:21:1 1 736.339 49.724 4238 677 271 S6:22:1 2 750.354 51.547 31806 5434 272 S6:22:1 2 750.354 5 1.880 4027 978 67 Table 2.4. (continued) Elution Order Acylsucrose ID # of Isomers [M+NH 4 ] + m/z XIC (±20 ppm) Ret. Time (min) Peak Height Peak Area 302 S6:23:1 4 764.370 54.270 6205 1081 303 S6:23:1 4 764.370 54.678 6907 1443 304 S6:23:1 4 764.370 55 .185 9032 1813 305 S6:23:1 4 764.370 56.282 1871 333 334 S6:24:1 4 778.386 56.924 23288 4431 335 S6:24:1 4 778.386 57.356 2716 836 336 S6:24:1 4 778.386 57.598 1876 308 337 S6:24:1 4 778.386 57.855 2291 386 363 S6:25:1 5 792.401 59.980 10714 2454 36 4 S6:25:1 5 792.401 60.449 9101 1957 365 S6:25:1 5 792.401 60.691 768 109 366 S6:25:1 5 792.401 60.911 746 122 367 S6:25:1 5 792.401 61.576 1390 232 388 S6:26:1 5 806.417 62.831 3577 709 389 S6:26:1 5 806.417 63.104 951 160 390 S6:26:1 5 806.417 63.4 22 373 75 391 S6:26:1 5 806.417 64.087 485 92 392 S6:26:1 5 806.417 64.586 1819 347 403 S6:27:1 1 820.433 65.607 628 107 197 S6:20:2 1 720.307 44.362 2415 397 290 S6:23:4 2 758.323 48.326 15241 2504 291 S6:23:4 2 758.323 48.635 1308 199 324 S6:24:4 2 772.339 51.010 3317 582 325 S6:24:4 2 772.339 51.471 6050 1083 352 S6:25:4 4 786.354 53.150 8527 1542 353 S6:25:4 4 786.354 53.605 26904 4304 354 S6:25:4 4 786.354 54.141 1616 286 355 S6:25:4 4 786.354 54.481 1585 314 377 S6:26:4 6 800.370 55.699 9 964 1906 378 S6:26:4 6 800.370 56.183 5249 1231 379 S6:26:4 6 800.370 56.372 1347 154 380 S6:26:4 6 800.370 56.675 4658 779 381 S6:26:4 6 800.370 57.129 4649 806 382 S6:26:4 6 800.370 57.590 4217 732 397 S6:27:4 6 814.386 58.165 61542 10899 398 S6:2 7:4 6 814.386 58.483 4300 703 399 S6:27:4 6 814.386 58.710 9016 1654 400 S6:27:4 6 814.386 59.594 3888 719 401 S6:27:4 6 814.386 60.525 3738 655 68 Table 2.4. (continued) Elution Order Acylsucrose ID # of Isomers [M+NH 4 ] + m/z XIC (±20 ppm) Ret. Time (min) Peak Height Peak Area 402 S6:27:4 6 814.386 61.319 851 154 407 S6:28:4 2 828.401 61.175 35416 6602 408 S6:28:4 2 828.401 62.007 8272 1553 415 S6:29:4 2 842.417 63.928 381 82 416 S6:29:4 2 842.417 65.441 1370 330 349 S6:25:5 3 784.339 50.783 105 9 196 350 S6:25:5 3 784.339 51.230 1101 207 351 S6:25:5 3 784.339 52.295 638 110 Heptaacylsucroses 411 S7:27:0 2 836.427 68.806 2230 436 412 S7:27:0 2 836.427 69.276 389 79 418 S7:28:0 1 850.443 71.756 1138 254 69 Table 2. 5 . LC /MS/MS metadata Facility Director Dr. A. Daniel Jones Analyst Steven M. Hurney Instrument Location MSU Mass Spectrometry and Metabolomics Core Facility Instrument Title LCMS: G2 - XS QTof #2 LC System Acquity UPLC I - Class Binary Solvent Manager equipped with Acquity Column Manager Manufacturer Waters Autosampler 2777C Sample Manager Column Ascentis Express C18 Analytical HPLC, 10 cm x 2.1mm x 2.7 µm Column Manufacturer Supelco Catalogue Number 53823 - U Serial Number USRB002977 Packing Lot Number S11 020 Injection Volume 10 µL Flow Rate 0.4 mL/min Mobile Phases: A 10 mM ammonium formate in water (pH 2.8, adjusted with formic acid) B Acetonitrile Gradient Profile Hold 1% B at 0 - 1 min, linear 1 - 40% B at 1 - 2 min, linear 40% - 60% B at 2 - 27 min, linear 60 - 100% B at 27 - 28 min, hold 100% B at 28 - 32 min, linear 100 - 1% B at 32 - 32.01 min, and hold 1% B at 32.01 - 35 min Column Oven Temperature 50 ºC Sample Temperature in autosampler 10 °C Inlet Method Name SMH_C18_Column1_35min_40to60 Mass Spectrometer Xev o G2 - XS QTof Manufacturer Waters Software MassLynx v4.1 Ionization Source Electrospray Ionization (ESI) Data Acquisition Sensitivity Mode (low mass and high mass resolving quadrupole setting at 15) Polarity Positive, Negative Mass Range m/z 50 - 1000 Data Format Centroid Capillary Voltage 3 kV (ESI+), 2 kV (ESI - ) Sample Cone 35 V (ESI+), 40 V (ESI - ) Source Temperature 100 °C Source Offset 80 V 70 Table 2.5. (continued) Desolvation Temperature 350 °C Cone Gas Flow 25 L/hr Desolvation Gas Flow 350 L /hr LM Resolution 15 HM Resolution 15 Collision Potential Ramp Start Potential 5 V End Potential 60 V Scan Duration 0.5 s Inter Scan Delay 0.014 s Collision Cell Pressure 0.2 mbar MS Method Files SMH_MSMS_pos_35 min_Mass A, SMH_MSMS_neg_35 min_ Mass A Sample handling Each acylsucrose NMR sample was concentrated to dryness under N 2 gas in autosampler vials, reconstituted in 0.50 mL AcN:IPA and diluted ~100x by removing a 5.0 µL aliquot and adding it to 0.50 mL of AcN:IPA solution. Sample Storag e - 20 °C in autosampler vials with screw PTFE lined caps Protocol when analyzing the samples The instrument was calibrated in ESI+/ - modes using 500 µM sodium formate solution. First, samples were analyzed in ESI - mode. After column equilibration, a blank sample (AcN:IPA) and a S. sinuata bulk extract sample were analyzed in CID mode, followed by MS/MS analysis of the first eight acylsucrose samples in ascending elution order. Again, the blank and a S. sinuata bulk extract samples were analyzed in CID mode , followed by MS/MS analysis of the remaining eight acylsucrose samples in ascending elution order. Finally, the blank and S. sinuata bulk extract samples were analyzed in CID mode. Subsequently, the sequence was repeated in ESI+ mode. MS/MS product ion sp ectra were generated using a collision potential ramp, targeting the res pective precursor pseudomolecular ion [M+HCOO] - and [M+NH 4 ] + for the entirety of the chromatographic runtime (35 mins). 71 Figure 2. 8 . S4:19:0(3,5,5,6) MS/M S product ion spectra, precursor ion [M+NH 4 ] + (top) & [M+HCOO] - (bottom, magnified 5x over m/z 51 - 664) 72 Figure 2. 9 . S4:19:0(2,5,6,6) MS/MS product ion spectra, precursor ion [M+NH 4 ] + (top) & [M+HCOO] - (bot tom, magnified 10x ove r m/z 51 - 664) 73 Figure 2. 10 . S5:20:0(2,2,5,5,6) MS/MS product ion spectra, precursor ion [M+NH 4 ] + (top) & [M+HCOO] - (bottom, magnified 4x over m/z 51 - 692) 74 Figure 2. 11 . S4:20:0(4,5,5,6) MS/MS product ion spectra, precursor ion [M+NH 4 ] + (top) & [M+HCOO] - (bottom, magnified 6x over m/z 51 - 678) 75 Figure 2. 12 . S4:20:0(3,5,6,6) MS/MS product ion spectra, precursor ion [M+NH 4 ] + (top) & [M+HCOO] - (bottom, magnified 3x over m/z 51 - 6 78) 76 Figure 2. 13 . S4:20:0(2,6,6,6) MS/MS product ion spectra, precursor ion [M+NH 4 ] + (top) & [M+HCOO] - (bottom, magnified 14x over m/z 51 - 678) 77 Figure 2. 14 . S5:22:1(2,5,5,5,5 T ) MS/MS product ion spe ctra, precursor ion [M+NH 4 ] + (top) & [M+HCOO] - (bottom, magnified 9x over m/z 51 - 718) 78 Figure 2. 15 . S5:21:0(2,2,5,6,6) MS/MS product ion spectra, precursor ion [M+NH 4 ] + (top) & [M+HCOO] - (bottom, magnified 4x over m/z 51 - 706) 79 Figure 2. 16 . S4:21:0(5,5,5,6) MS/MS product ion spectra, precursor ion [M+NH 4 ] + (top) & [M+HCOO] - (bottom, magnified 5x over m/z 51 - 692) 80 Figure 2. 17 . S5:25:4(2,5,5,5,8 P ) MS/MS product ion spectra, p recursor ion [M+NH 4 ] + (top) & [M+HCOO] - (bottom, magnified 6x over m/z 51 - 754) 81 Figure 2. 18 . S5:22:0(2,2,6,6,6) MS/MS product ion spectra, precursor ion [M+NH 4 ] + (top) & [M+HCOO] - (bottom, magnified 7x over m/z 51 - 720) 82 Fig ure 2. 19 . S4:22:0(5,5,6,6) MS/MS product ion spectra, precursor ion [M+NH 4 ] + (top) & [M+HCOO] - (bottom, magnified 3x over m/z 51 - 706) 83 Figure 2. 20 . S5:23:0(2,5,5,5,6) MS/MS product ion spectra, precurso r ion [M+NH 4 ] + (top) & [M+HCOO] - (bottom, magnified 7x over m/z 51 - 734) 84 Figure 2. 21 . S4:23:0(5,6,6,6) MS/MS product ion spectra, precursor ion [M+NH 4 ] + (top) & [M+HCOO] - (bottom, magnified 3x over m/z 51 - 720) 85 Figure 2. 22 . S5:24:0(2,5,5,6,6) MS/MS product ion spectra, precursor ion [M+NH 4 ] + (top) & [M+HCOO] - (bottom, magnified 3x over m/z 51 - 748) 86 Figure 2. 23 . S6:25:0(2,2,5,5,5,6) MS/MS product ion spectra, precursor ion [M+NH 4 ] + (top) & [M+HCOO] - (bottom, magnified 200x over m/z 51 - 87 Table 2. 6 . GC/MS metadata Facility Director Dr. A. Daniel Jones Analyst Steven M. Hurney Instrument Location MSU Mass Spectrometry and Metabolomics Core Facility Instrument Title GCMS: Waters GCT GC System Agilent Technologies (model 6890N) Manufacturer Waters Injector Agilent Technologies (mo del V7683B) Column VF - 5ms, 30m x 250µm ID, 0.25µm film thickness (+10m EZ - Guard) Column Manufacturer Agilent Technologies Part Number CP9013 Serial Number 90310154380 Injection Volume 1.0 µL (splitless mode) Flow Rate 1.0 mL/min (constant flow) Carr ier Gas Helium Gradient Profile Initial 36 °C, linear 4 °C/min to 220°C, linear 20 °C/min to 320 °C, and hold at 320 °C for 4 mins Solvent Delay 2.45 - 3.35 mins Inlet Temperature 280 ºC Purge Time 0.75 min Purge Pressure 30.0 kPa Inlet Method Name SMH _splitless_56min_36 - 320C Mass Spectrometer GCT Premier CAB067 Manufacturer Waters Software MassLynx v4.1 Ionization Mode Electron Ionization Mass Range m/z 50 - 600 Data Format Centroid Electron Energy 70.0 eV Source Temperature 200 °C Scan Duration 0.2 s Inter Scan Delay 0.05 s MS Method Files SMH_56min Retention Index Series Supelco C7 - C30 saturated alkane standard series diluted to 40 µg/mL in n - hexane Lot Number XA17133V Sigma Aldrich Product Number 49451 - U 88 Table 2.6. (continued) Sample handling For transesterification, 10 mL of extract solution was concentrated to dryness by purging with N2 gas. 300 µL of 21% (v/v) sodium ethoxide solution was added. The reaction proceeded for 40 mins with horizontal mixing at 140 rpm. 400 µL of n - hexane was added, followed by vortex mixing. The hexane layer was removed for analysis by performing three washes with 500 µL of sodium chloride solution. A 50 µL aliquot of the hexane solution was transferred to an autosampler vial with limited volume insert. Sample Storage - 20 °C in autosampler vials with screw PTFE lined caps Protocol when analyzing the samples The instrument was calibrated in EI+ mode using perfluorotributylamine. First, a hexane blank was analyzed, followed by a process blank (10 mL AcN:IP A concentrated to dryness and taken through the sodium ethoxide reaction and hexane extraction process). Th en, the transesterified S. sinuata sample was analyzed, followed by the 40 ppm C7 - C30 saturated alkanes standard. 89 Figure 2. 24 . GC/MS total ion chromatograms (3 - 23 mins): process blank (bottom), transesterified S. sinuata sample (middle), and retention indexing series C7 - C13 (top). Peaks are annotated with retention time, base peak m/z , and integrated peak area. 90 Table 2. 7 . GC/MS retention indexing and TIC peak area assessment Standard Ret. Time (min) Peak Height Peak Area heptane 3.922 852 78 octane 5.706 1064 84 nonane 8.466 2362 143 decane 11.894 4711 260 undecane 15.531 14196 613 dodecane 19.118 18333 885 tridecane 22.597 46077 2298 Ethyl Ester Retention index % of Ethyl Ester TIC Peak Area C2 ND --- --- --- C3 ND --- --- --- iC4 4.940 BLQ BLQ 757 --- aiC5 7.075 746669 40722 850 38.9% iC5 7.159 22638 802 8 53 0.8% C5 T 9.795 50639 2250 939 2.1% aiC6 10.546 732441 60585 961 57.8% C8 P 20.722 6216 422 1246 0.4% ND = Not detected BLQ = Below Limit of Quantitation in total ion chromatogram 91 Figure 2. 25 . GC/EI/MS of 2 - methy lpropanoic acid ethyl ester (iC4) Figure 2. 26 . GC/EI/MS of 2 - methylbutanoic acid ethyl ester (aiC5) 92 Figure 2. 27 . GC/EI/MS of 3 - methylbutanoic acid ethyl ester (iC5) Figure 2. 28 . GC/EI/MS of ethyl tiglate (C5 T ) 93 Figure 2. 29 . GC/EI/MS of 3 - methylpentanoic acid ethyl ester (aiC6) Figure 2. 30 . GC/EI/MS of phenylacetic acid ethyl ester (C8 P ) 94 Table 2. 8 . Purification of acylsucroses by semi - preparative HPLC Analyst Steven M. Hurney Instrument Location A. Daniel Jones Laboratory Instrument Title Waters semi - prep HPLC (Jones Lab) LC System Waters 2795 Separations Module equipped with LKB Bromma 2211 Superrac Fraction Collector with automated fraction collection Manufacturers Waters Column Thermo Scientific Acclaim 120 C18 semi - preparative HPLC column (4.6 x 150 mm, 5 µm particle size) Column Manufacturer Thermo Scientific Product Number 059148 S erial Number 005992 Packing Lot Number 014 - 25 - 017 Injection Volume 50 µL Flow Rate 1.5 mL/min Mobile Phases: A 0.15% Formic Acid in Water B Acetonitrile C dichloromethane: acetone: methanol (v/v/v, 1;1:1) Gradient Profile (solvents A and B only) Ho ld 5% B at 0 - 1 min, linear 5 - 45% B at 1 - 15 min, linear 45% - 49% B at 15 - 55 min, linear 49 - 52% B at 55 - 70 min, linear 52 - 65% B at 70 - 97 min, linear 65 - 100% B at 97 - 100 min. Column Wash and Re - equilibration Hold 100% B at 100 - 102 min, linear 100% B t o 100% C at 102 - 103 min, hold 100% C at 103 - 111 min, linear 100% C to 100% B at 111 - 112 min, linear 100 - 5% B at 112 - 113 min, and hold 5% B at 113 - 120 min. Column Oven Temperature 50 ºC Sample Temperature in autosampler Room Temperature Method Name SP_56C_70min_ AcN_MeOH Sample handling Approximately 1 L of the S. sinuata bulk extract was added in portions of ~100 mL to a 250 mL round bottom flask, and dried via rotary evaporation under vacuum at ~30°C, leaving a green residue. The residue was reconstituted in 5. 0 mL AcN:IPA with sonication for 5 mins while manually swirli ng. The solution was centrifuged by Eppendorf Centrifuge 5480R at 10000 × g for 10 mins. The supernatant was then transferred to LC autosampler vials for semi - preparative HPLC purification. Samp le Storage - 20 °C in autosampler vials Protocol when analyzi ng the samples One minute fractions were collected in Pyrex glass culture tubes (18 × 150 mm) in eight batches labeled letter A - H. Each batch consisted of six injections each. Fractions were conc entrated to dryness under vacuum using a Thermo Savant SPD 13 1 DDA SpeedVac Concentrator with BOC Edwards XDS Dry Pump. Fractions were labeled by minute, with fractions 33 - 96 containing the most abundant acylsucroses. 95 Table 2.8. (continued) Protocol for testing the fractions In order to test for purity and reproducibility, fractions from each batch were reconstituted in 0.50 mL AcN:IPA, and transferred to LC autosampler vials. 5.0 µL aliquots from each fraction were diluted in autosampler vials containing 0.50 mL of AcN:IPA an d analyzed on an LCT Premier mass spectrometer equipped with Shimadzu LC - 20AD pumps, Shimadzu SIL - 5000 autosampler and a Shimadzu CTO - 20A column oven using a 30 min gradient (0.3 mL min - 1 , Ascentis Express C18 Analytical HPLC, 10cm x 2.1mm x 2.7µm) optimiz ed for separating S. sinuata acylsugars. The elution program is as follows: hold 1% B at 0 - 1 min, linear 1 - 45% B at 1 - 2 min, linear 45% - 65% B at 2 - 27 min, linear 65 - 100% B at 27 - 28 min, hold 100% B at 28 - 32 min, linear 100 - 1% B at 32 - 32.01 min, hold 1% B at 32.01 - 35 min. HPLC fractions were combined according to metabolite purity and by comparison to a S. sinuata bulk extract solution analyses run before and after each group of fractions (i.e. the sample order was blank, S. sinuata bu lk extract, ~20 fracti on samples in ascending elution order, blank, S. sinuata bulk extract, and etc.). Purification of S5:25:4(2,5,5,5,8 P ) Fractions #55 - 56 (A - G) were combined and concentrated to dryness under N 2 gas, and reconstituted in AcN:IPA and tra nsferred to autosampler vials with limited volume insert. Two 200 µL injections were performed for secondary purification using a Supelco Ascentis® Express F5 semi - preparative HPLC column (15 cm × 4.6 mm, 2.7 µm particle size, Cat#53591 - U, BL: S140 29, Col: USBK001496). The elution program is as follows (1.0 mL min - 1 ): hold 5% B at 0 - 1 min, linear 1 - 45% B at 1 - 15 min, linear 45% - 49% B at 15 - 55 min, linear 49 - 100% B at 55 - 56 min, hold 100% B at 56 - 58 min, linear 100% B to 100% C at 58 - 59 min, hold at 100% C a t 59 - 69 mins, linear 100% C to 100% B at 69 - 70 mins, 100 - 5% B at 70 - 71 mins, and hold 5% B at 71 - 80 mins. Fractions 26 - 27 were combined according to purity assessed by LC/MS. 96 Table 2. 9 . Bruker 900 MHz NMR Instrument Meta data Fac ility Supervisor Dr. Daniel Holmes Analyst Steven M. Hurney Instrument Location MSU Max T. Rogers NMR Facility Facility Instrument Title 900 MHz Bruker Time of acquisition November 2015 - March 2016 Manufacturer Bruker Field Frequency Lock Chloroform - d 1 or Acetonitrile - d 3 Additional Solute None Solvent CDCl 3 : 250 - 300 µL and CD 3 CN: 250 µL Chemical Shift Standard CDCl 3 H C = 77.20 ppm) CH 3 CN - d 3 H C = 118.70 ppm) Concentration Standard None Instrument Bruker Avance 900 MHz NMR Geographic location of instrument 42.7164, - 84.4677 Magnet 899.13 - 899.00 MHz Probe Bruker TCI triple - resonance i nverse detection cryoprobe Console Bruker Avance Acquisition Software Topspin 2.1.6 Acquisition Parameters: a) Acquisition parameters file reference 1 H:/opt/topspin216/exp/stan/nmr/lists/pp/zg 13 C:/opt/topspin216/exp/stan/nmr/lists/pp/zgpg30 COS Y:opt/topspin216/exp/stan/nmr/lists/pp/cosygpmfph HSQC:/opt/topspin216/exp/stan/nmr/lists/pp/hsqcedetgpsisp2.2 HMBC: /opt/topspin216/exp/stan/nmr/lists/pp/hmbcgplpndqf J - resolved: /opt/topspin216/exp/stan/nmr/lists/pp/jresqf ROESY: /opt/topspin216 /exp/stan/nmr/lists/pp/roesyetgp b) Sample Details Shigemi (5 mm) NMR tube, Temperature @ 298 K, No Spinning c) Instrument operation details (recorded for each sample independently and are roughly the same for all samples measured; displayed here is the sample containing S4:19:0(3,5,5,6) as an example) Radiation frequency (MHz): 1 H: 899.0000263 13 C: 226.0536927 COSY: 899.0000268 (F2), 899.0000194 (F1) HSQC: 899.0000277 (F2), 226.0536917 (F1) HMBC: 899.0000250 (F2), 226.0536889 (F1) J - resolved: 899.0000266 (F2), 899.0000000 (F1) ROESY: 899.0000271 (F2), 899.0000266 (F1) Acquisition nucleus: 1 H: 90° = 7.59 µs, 13 C: 90° = 25.0 µs 97 Table 2.9. (continued) d) Number of scans (scans, dummy scans) 1 H: 32 - 64, 0 13 C: 1000 - 10000, 8 COSY: 8 - 16, 16 HSQC: 12 - 32, 16 HMBC: 30 - 70, 32 J - resolved: 32 - 64, 4 ROESY: 12 - 16, 8 e) Number of data points acquired (F2, F1) 1 H: 148144 13 C: 65536 COSY: 2048, 200 HSQC: 1024, 400 HMBC: 4096, 480 J - resolved: 2048, 128 ROESY: 1982, 512 e) Dwell t ime (µs) 1 H: 27.0 13 C: 9.225 COSY: 46.4 HSQC: 46.4 HMBC: 46.4 J - resolved: 46.4 ROESY: 46.4 FID and spectral processing parameters: Processing software Topspin 3.5.b.91 pl 7 a) Number of data points in spectrum (F2, F1) 1 H: 262144 13 C: 1310 72 COSY: 4096, 4096 HSQC: 1024, 1024 HMBC: 4096, 2048 J - resolved: 4096, 4096 ROESY: 4096, 2048 b ) Window function details 1 H: exponential multiply, LB = 0 Hz 13 C: exponential multiply, LB = 2 Hz COSY: QSINE, QSINE; LB = - 0.3, - 0.3 Hz; GB = 0 .3, 0; SSB = 2, 2; TM1 = 0, 1; TM2 = 0, 1 HSQC: QSINE, QSINE; SSB = 2, 2; TM1 = 0, 0.1; TM2 = 0, 0.9 HMBC: QSINE, SINE J - resolved: SINE, SINE; LB = - 0.3, - 0.3 Hz; GB = 0.3, 0; TM1 = 0, 1; TM2 = 0, 1 ROESY: QSINE, QSINE; LB = 1.0, 0.3 Hz; GB = 0.35, 0.1; SSB = 2, 2; TM1 = 0, 0.1; TM2 = 0, 0.9 98 Table 2. 10 . Summary of 1 H chemical shifts of sucrose core hydrogen atoms. Chemical shifts labeled in bold indicate acyl substitutions are located at those positions. All spectra were r eferenced to non - deuterated solvent signal of CDCl 3 H = 7.26 ppm), except for S5:25:4(2,5,5,5,8 P ) which was referenced to non - deuterated solvent signal of acetonitrile - d 3 H = 1.94 ppm). Ac ylsugar ID R 1 R 2 R 3 R 4 R 5 R 6 R 1' R 2' R 3' R 4' R 5' R 6' Tetraacyls ucrose S4:19:0(3,5,5,6) 5.84 4.81 5.55 4.93 4.16 3.64, 3.61 4.11, 4.05 --- 4.17 4.31 3.67 3.89, 3.71 S4:19:0(2,5,6,6) 5.64 4.87 5.53 4.92 4.14 3.61 3.66, 3.54 --- 5.17 4.60 3.89 3.89, 3.71 S4:20:0(4,5,5,6) 5.84 4.81 5.57 4.92 4.14 3.62, 3.59 4.12 , 4.06 --- 4.17 4.31 3.71 3.89, 3.70 S4:20:0(3,5,6,6) 5.78 4.87 5.53 4.92 4.16 3.64, 3.60 4.08, 4.06 --- 4.17 4.31 3.71 3.89, 3.71 S4:20:0(2,6,6,6) 5.63 4.87 5.52 4.92 4.14 3.62 3.66, 3.54 --- 5.16 4.61 3.88 3.89, 3.71 S4:21:0(5,5,5,6) 5.87 4.78 5.58 4. 93 4.15 3.63, 3.60 4.13, 4.06 --- 4.17 4.31 3.71 3.88, 3.71 S4:22:0(5,5,6,6) 5.81 4.85 5.54 4.93 4.14 3.63, 3.58 4.09, 4.07 --- 4.17 4.31 3.69 3.88, 3.71 S4:23:0(5,6,6,6) 5.78 4.86 5.53 4.92 4.12 3.64 4.09, 4.06 --- 4.17 4.31 3.71 3.88, 3.70 Pentaacylsucrose S5:20:0(2,2,5,5,6) 5.65 4.88 5.58 4.98 4.06 3.66, 3.58 3.63, 3.55 --- 5.22 4.36 4.07 4.41, 4.28 S5:21:0(2,2,5,6,6) 5.62 4.90 5.55 4.97 4.05 3.66, 3.58 3.63, 3.55 --- 5.20 4.35 4.08 4.41, 4.28 S5:22:0(2,2,6,6,6) 5.62 4.90 5.54 4.98 4.04 3.67, 3.60 3.63, 3.55 --- 5.18 4.35 4.08 4.43, 4.28 S5:23:0(2,5,5,5,6) 5.67 4.90 5.56 4.94 4.16 3.65, 3.59 4.18, 4.15 --- 4.14 4.11 3.87 4.39, 4.24 S5:24:0(2,5,5,6,6) 5.66 4.93 5.54 4.94 4.14 3.66, 3.59 4.16 --- 4.14 4.11 3.90 4.42, 4.22 Pentaacylsucrose w/unsaturated ester S5:22:1(2,5,5,5,5 T ) 5.65 5.02 5.63 5.03 4.15 3.68, 3.61 4.21, 4.17 --- 4.15 4.11 3.90 4.42, 4.23 S5:25:4(2,5,5,5,8 P ) 5.58 4.90 5.49 5.03 4.12 3.57, 3.45 4.01 --- 3.98 3.97 3.80 4.26, 4.19 Hexaacylsucrose S6:25:0(2,2,5,5,5,6) 5.70 4.87 5.57 4.97 4.07 3.66, 3.59 4.14, 4.04 --- 5.23 4.35 4.04 4.40, 4.30 99 Table 2. 11 . Summary of 13 C chemical shifts of sucrose core carbon atoms. Chemical shifts labeled in bold indicate acyl substitutions are located at those positions. All spectra were referenced to non - deuterated solvent signal of CDCl 3 C = 77.20 ppm), except for S5:25:4(2,5,5,5,8 P ) which was referenced to non - deuterated solvent signal of acetonitrile - d 3 C = 118.70 ppm). Acylsugar ID R 1 R 2 R 3 R 4 R 5 R 6 R 1' R 2' R 3' R 4' R 5' R 6' Tetraacylsucrose S4:19:0(3,5,5,6) 88.96 70.87 68.8 8 68.82 72.05 61.55 63.88 103.61 78.40 72.66 81.10 59.71 S4:19:0(2,5,6,6) 89.75 70.53 68.86 68.62 71.94 61.86 64.63 104.15 80.13 71.54 82.38 59.86 S4:20:0(4,5,5,6) 89.03 70.92 68.72 68.61 72.09 61.49 63.93 103.66 78.49 72.89 81.16 59.74 S4:20:0(3,5,6,6) 89.14 70.56 68.98 68.82 72.19 61.57 63.63 103.45 78.45 72.79 81.33 59.77 S4:20:0(2,6,6,6) 89.78 70.52 68.87 68.71 71.99 61.91 64.59 104.16 80.31 71.52 82.32 59.76 S4:21:0(5,5,5,6) 88.79 71.11 68.76 68.58 71.99 61.51 64.07 103.74 78.43 72.56 80.98 59.71 S4:22:0(5,5,6,6) 89.00 70.81 68.82 68.55 72.15 61.52 63.80 103.58 78.39 72.74 81.22 59.78 S4:23:0(5,6,6,6) 89.15 70.69 68.83 68.66 72.21 61.55 63.69 103.52 78.43 72.93 81.35 59.81 Pentaacylsucrose S5:20:0(2,2,5,5,6) 89.52 70.55 68. 9 a 68.9 a 71.11 61.64 64.34 104.58 79.23 73.72 80.38 64.08 S5:21:0(2,2,5,6,6) 89.52 70.48 68.98 68.85 71.13 61.64 64.27 104.65 79.30 73.90 80.43 64.27 S5:22:0(2,2,6,6,6) 89.58 70.49 69.0 b 69.0 b 71.17 61.67 64.23 104.66 79.56 74.04 80.39 64.27 S5:23:0(2,5 ,5,5,6) 89.36 70.24 68.88 68.94 71.28 61.65 63.70 103.63 78.11 74.69 78.63 63.49 S5:24:0(2,5,5,6,6) 89.51 70.22 68.93 68.90 71.31 61.64 63.49 103.57 78.16 75.03 78.80 63.69 Pentaacylsucrose w/unsaturated ester S5:22:1(2,5,5,5,5 T ) 89 .99 70.06 69.12 68.73 71.49 61.58 63.39 103.58 78.31 75.33 78.79 63.62 S5:25:4(2,5,5,5,8 P ) 90.22 71.37 71.49 69.70 72.08 61.95 64.02 104.55 78.29 75.16 80.59 65.60 Hexaacylsucrose S6:25:0(2,2,5,5,5,6) 89.41 70.37 68.87 68.97 71.15 6 1.67 64.10 102.98 78.95 73.45 80.22 63.76 a - Two 13 C signals not resolved in 2D spectra (68.91, 68.93 ppm) b - Two 13 C signals not resolved in 2D spectra (68.96, 68.97 ppm) 100 Table 2. 12 . S4:19:0( 3,5,5,6) Chemical shifts and coupl ing constants Molecular Formula: C 31 H 52 O 15 110 min Retention Time (ESI+): 48.21 mins HRMS: (ESI+) m/z calculated for C 31 H 56 NO 15 + ([M+NH 4 + ]): 682.3644, found: 682.3664 Fra ction, Batch: #34, A - D Sample mass for NMR analysis: 0.8 mg NMR Solvent: CDCl 3 InChi Key: HGYUMQHIZULVOD - AOJWIRPASA - N Carbon # (group) 1 H (ppm) 13 C (ppm) 1 (CH) 5.84 (d, J = 4.0 Hz) 88.96 2 (CH) 4.81 ( dd, J = 10.4, 4.0 Hz) 70.87 - 1(CO) 177.40 - 2(CH) 2.42 (sextet, J = 7.0 Hz) 40.76 - 3(CH 3 ) 1.14 (d, J = 7.0 Hz) 16.16 - 4(CH 2 ) 1.63 (m), 1.45 (m) 26.80 - 5(CH 3 ) 0.87 (t, J = 7.4 Hz) 11.60 3 (CH) 5.55 (dd, J = 10.6, 9.2 Hz) 68.88 - 1(CO) 172.24 - 2(CH 2 ) 2.25 (dd, J = 15.2, 5.8 Hz), 2.03 (dd, J = 15.2, 8.3 Hz) 41.26 - 3(CH) 1.80 (m) 31.72 - 4(CH 3 ) 0.88 (d, J = 6.7 Hz) 19.30 - 5(CH 2 ) 1.31 (m), 1.19 (m) 29.35 - 6(CH 3 ) 0.86 (t, J = 7.4 Hz) 11.35 4 (CH) 4.93 (dd, J = 10.7, 9.2 Hz) 68.82 - 1(CO) 173.84 - 2(CH 2 ) 2.35 (m), 2.29 (m) 27.56 - 3(CH 3 ) 1.12 (t, J = 7.6 Hz) 9.08 5 (CH) 4.16 ( ddd, J = 10.4, 6.5, 2.5 Hz) 72.05 6 (CH 2 ) 3.64 (dd, J = 12.5, 2.4 Hz), 3.61 (dd, J = 12.5, 6.5 Hz) 61.55 1' (CH 2 ) 4.11 (d, J = 11.6 Hz), 4.05 (d, J = 11.6 Hz) 63.88 - 1(CO) 175.92 - 2(CH) 2.41 (sextet, J = 7.0 Hz) 41.13 - 3(CH 3 ) 1.15 (d, J = 7.0 Hz) 16.73 - 4(CH 2 ) 1.68 (m), 1.49 (m) 26.88 - 5(CH 3 ) 0.91 (t, J = 7.4 Hz) 11.74 101 Table 2.12. (continued) 2' (C) 103.61 3' (CH) 4.17 (d, J = 8.8 Hz) 78.40 4' (CH) 4.31 (t, J = 8.9 Hz) 72.66 5' (CH) 3.67 (dt, J = 9.0, 2.3 Hz) 81.10 6' (CH 2 ) 3.89 (dd, J = 13.4, 2.4 Hz), 3.71 (dd, J = 13.4, 2.3 Hz) 59.71 102 Figure 2. 31 . S4:19:0( 3,5,5,6) 1 H NMR 103 Figure 2. 32 . S4:19:0( 3,5,5,6) 13 C NMR 104 Fig ure 2. 33 . S4:19:0( 3,5,5,6) 1 H - 1 H gCOSY 105 Figure 2. 34 . S4:19:0( 3,5,5,6) gHSQC 106 Figure 2. 35 . S4:19:0( 3,5,5,6) gHMBC 107 Figure 2. 36 . S4:19:0( 3,5,5,6) J - r esolved 108 Figure 2. 37 . S4:19:0( 3,5,5,6) ROESY 109 Table 2. 13 . S4:19:0( 2,5,6,6) Chemical shifts and coupling constants Molecular Formula: C 31 H 52 O 15 110 min Ret ention Time (ESI+): 48.39 mins HRMS: (ESI+) m/z calculated for C 31 H 56 NO 15 + ([M+NH 4 + ]): 682.3644, found: 682.3675 Fraction, Batch: #35, A - D Sample mass for NMR analysis: 1.3 mg NMR Solvent: CDCl 3 InChi Key: IWACBCFUOAGBIU - QRRKJAOQSA - N Carbon # (group) 1 H (ppm) 13 C (ppm) 1 (CH) 5.64 (d, J = 3.7 Hz) 89.75 2 (CH) 4.87 (dd, J = 10.4, 3.7 Hz) 70.53 - 1(CO) 172.78 - 2(CH 2 ) 2.32 (dd, J = 15.7, 5.7 Hz), 2.06 (dd , J = 15.7, 8.3 Hz) 40.99 - 3(CH) 1.82 (m) 31.58 - 4(CH 3 ) 0.88 (d, J = 6.7 Hz) 19.37 - 5(CH 2 ) 1.34 (m), 1.20 (m) 29.34 - 6(CH 3 ) 0.88 (t, J = 7.4 Hz) 11.4 a 3 (CH) 5.53 (dd, J = 10.7, 9.1 Hz) 68.86 - 1(CO) 172.3 b - 2(CH 2 ) 2.23 (dd, J = 15.5, 5.6 Hz), 2.02 ( dd, J = 15.5, 8.4 Hz) 41.22 - 3(CH) 1.79 (m) 31.64 - 4(CH 3 ) 0.88 (d, J = 6.7 Hz) 19.43 - 5(CH 2 ) 1.31 (m), 1.18 (m) 29.38 - 6(CH 3 ) 0.85 (t, J = 7.4 Hz) 11.4 a 4 (CH) 4.92 (dd, J = 10.7, 9.2 Hz) 68.62 - 1(CO) 176.12 - 2(CH) 2.36 (sextet, J = 7.0 Hz) 41.08 - 3(CH 3 ) 1.10 (d, J = 7.0 Hz) 16.53 - 4(CH 2 ) 1.67 (m), 1.44 (m) 26.63 - 5(CH 3 ) 0.90 (t, J = 7.4 Hz) 11.82 5 (CH) 4.14 (ddd, J = 10.2, 5.5, 2.8 Hz) 71.94 6 (CH 2 ) 3.61 (m) 61.86 1' (CH 2 ) 3.66 (d, J = 12.3 Hz), 3.54 (d, J = 12.3 Hz) 64.63 2' (C) 104.15 110 Table 2.13. (continued) 3' (CH) 5.17 (d, J = 7.8 Hz) 80.13 - 1(CO) 172.3 b - 2(CH 3 ) 2.26 (s) 21.00 4' (CH) 4.60 (t, J = 7.8 Hz) 71.54 5' (CH) 3.89 (m) 82.38 6' (CH 2 ) 3.89 (m), 3.71 (m) 59.86 a - Two 13 C signals not resolved in 2D spectra (11.37, 11.39 ppm) b - Two 13 C signals not resolved in 2D spectra (172.27, 172.29 ppm) 111 Figure 2. 38 . S4:19:0( 2,5,6,6) 1 H NMR 112 Figure 2. 39 . S4:19:0( 2,5,6,6) 13 C NMR 113 Figure 2. 40 . S4:19:0( 2,5,6, 6) 1 H - 1 H gCOSY 114 Figure 2. 41 . S4:19:0( 2,5,6,6) gHSQC 115 Figure 2. 42 . S4:19:0( 2,5,6,6) gHMBC 116 Figure 2. 43 . S4:19:0( 2,5,6,6) J - resolved 117 Figure 2. 44 . S4:19:0( 2,5,6,6) ROESY 118 Table 2. 14 . S5:20:0( 2,2,5,5,6) Chemical shifts and coupling constants Molecular Formula: C 32 H 52 O 16 110 min Retention Time (ESI+): 49.57 mins HRMS : (ESI+) m/z calculated for C 32 H 56 NO 16 + ([M+NH 4 + ]): 710.3594, found: 710.3621 Fraction, Batch: #38, A - D Sample mass for NMR analysis: 1.5 mg NMR Solvent: CDCl 3 InChi Key: DVSXLPKXNQNXQO - CJWUYPCVS A - N Carbon # (group) 1 H (ppm) 13 C (ppm) 1 (CH) 5.65 (d, J = 3.6 Hz) 89.52 2 (CH) 4.88 (dd, J = 10.5, 3.7 Hz) 70.55 - 1(CO) 175.86 - 2(CH) 2.36 (sextet, J = 7.0 Hz) 41.10 - 3(CH 3 ) 1.10 (d, J = 7.0 Hz) 16.47 - 4(CH 2 ) 1.67 (m), 1.45 (m) 26 .69 - 5(CH 3 ) 0.90 (t, J = 7.4 Hz) 11.82 3 (CH) 5.58 (dd, J = 10.7, 9.1 Hz) 68.9 a - 1(CO) 172.29 - 2(CH 2 ) 2.24 (dd, J = 15.8, 5.4 Hz), 2.01 (dd, J = 15.8, 8.6 Hz) 41.15 - 3(CH) 1.78 (m) 31.45 - 4(CH 3 ) 0.88 (d, J = 6.7 Hz) 19.47 - 5(CH 2 ) 1.31 (m), 1.18 (m) 29.43 - 6(CH 3 ) 0.85 (t, J = 7.4 Hz) 11.39 4 (CH) 4.98 (dd, J = 10.6, 9.2 Hz) 68.9 a - 1(CO) 176.40 - 2(CH) 2.35 (sextet, J = 7.0 Hz) 40.73 - 3(CH 3 ) 1.11 (d, J = 7.0 Hz) 16.08 - 4(CH 2 ) 1.61 (m), 1.43 (m) 26.83 - 5(CH 3 ) 0.86 (t, J = 7.4 Hz) 11.67 5 (CH) 4.06 (ddd, J = 10.4, 5.1, 2.1 Hz) 71.11 6 (CH 2 ) 3.66 (dd, J = 12.8, 2.5 Hz), 3.58 (dd, J = 12.8, 5.1 Hz) 61.64 1' (CH 2 ) 3.63 (d, J = 12.2 Hz), 3.55 (d, J = 12.2 Hz) 64.34 2' (C) 104.58 119 Table 2.14. (continued) 3' (CH) 5.22 (d, J = 7.7 Hz) 79.23 - 1(CO) 172. 05 - 2(CH 3 ) 2.22 (s) 20.99 4' (CH) 4.36 (t, J = 7.8 Hz) 73.72 5' (CH) 4.07 (ddd, J = 7.8, 6.1, 3.7 Hz) 80.38 6' (CH 2 ) 4.41 (dd, J = 12.1, 6.0 Hz), 4.28 (dd, J = 12.2, 3.5 Hz) 64.08 - 1(CO) 171.69 - 2(CH 3 ) 2.13 (s) 21.01 a - Two 13 C signals not resolved i n 2D spectra (68.91, 68.93 ppm) 120 Figure 2. 45 . S5:20:0( 2,2,5,5,6) 1 H NMR 121 Figure 2. 46 . S5:20:0( 2,2,5,5,6) 13 C NMR 122 Figure 2. 47 . S5:20:0( 2,2,5,5,6) 1 H - 1 H gCOSY 123 Figure 2. 48 . S5:20:0( 2,2,5,5,6) gHSQC 124 Figure 2. 49 . S5:20:0( 2,2,5,5,6) gHMBC 125 Figure 2. 50 . S5:20:0( 2,2,5,5,6) J - resolved 126 Figure 2. 51 . S5:20:0( 2,2,5,5,6) ROESY 127 Table 2. 15 . S4:20:0( 4,5,5,6) Chemical shifts and coupling constants Molecular Formula: C 32 H 54 O 15 110 min Retention Time (ESI+): 50.78 mins HRMS: (ESI+) m/z calcula ted for C 32 H 58 NO 15 + ([M+NH 4 + ]): 696.3801, found: 696.3804 Fraction, Batch: #42, A - D Sample mass for NMR analysis: 0.2 mg NMR Solvent: CDCl 3 InChi Key: JSZQPBVIKRWRRQ - IRCRNTFMSA - N Carbon # (group) 1 H (ppm) 13 C (ppm) 1 (CH) 5.84 (d, J = 4.0 Hz) 89.03 2 (CH) 4.81 (dd, J = 10.3, 4.1 Hz) 70.92 - 1(CO) 177.36 - 2(CH) 2.41 (sextet, J = 7.0 Hz) 40.76 - 3(CH 3 ) 1.15 (d, J = 7.0 Hz) 16.15 - 4(CH 2 ) 1.64 (m), 1.45 (m) 26.76 - 5(CH 3 ) 0.87 (t , J = 7.4 Hz) 11.62 3 (CH) 5.57 (dd, J = 10.6, 9.3 Hz) 68.72 - 1(CO) 172.11 - 2(CH 2 ) 2.24 (dd, J = 15.6, 5.6 Hz), 2.02 (dd, J = 15.6, 8.4 Hz) 41.1 a - 3(CH) 1.80 (m) 31.59 - 4(CH 3 ) 0.89 (d, J = 6.7 Hz) 19.41 - 5(CH 2 ) 1.31 (m), 1.19 (m) 29.38 - 6(CH 3 ) 0.86 (t, J = 7.4 Hz) 11.38 4 (CH) 4.92 (dd, J = 10.7, 9.3 Hz) 68.61 - 1(CO) 176.56 - 2(CH) 2.54 (septet, J = 7.0 Hz) 34.10 - 3(CH 3 ) 1.16 (d, J = 7.0 Hz) 19.08 - 4(CH 3 ) 1.13 (d, J = 7.0 Hz) 18.83 5 (CH) 4.14 (ddd, J = 10.5, 6.1, 1.9 Hz) 72.09 6 (CH 2 ) 3.62 (dd, J = 12.7, 2.1 Hz), 3.59 (dd, J = 12.7, 6.1 Hz) 61.49 128 Table 2.15. (continued) 1' (CH 2 ) 4.12 (d, J = 11.7 Hz), 4.06 (d, J = 11.7 Hz) 63.93 - 1(CO) 175.91 - 2(CH) 2.41 (sextet, J = 7.0 Hz) 41.1 a - 3(CH 3 ) 1.16 (d, J = 7.0 Hz) 16.74 - 4(CH 2 ) 1.68 (m), 1.49 (m) 26.89 - 5(CH 3 ) 0.91 (t, J = 7.4 Hz) 11.75 2' (C) 103.66 3' (CH) 4.17 (d, J = 8.7 Hz) 78.49 4' (CH) 4.31 (t, J = 8.8 Hz) 72.89 5' (CH) 3.71 (dt, J = 9.0, 2.3 Hz) 81.16 6' (CH 2 ) 3.89 (dd, J = 13.3, 2.4 Hz), 3.70 (dd, J = 13.3, 2.3 Hz) 59.74 a Two 1 3 C signals not resolved in 2D spectra (41.13, 41.15 ppm) 129 Figure 2. 52 . S4:20:0( 4,5,5,6) 1 H NMR 130 Figure 2. 53 . S4:20:0( 4,5,5,6) 13 C NMR 131 Figure 2. 54 . S4:20:0( 4,5,5,6) 1 H - 1 H gCOSY 132 Figure 2. 55 . S4:20:0( 4,5,5,6) gHSQC 133 Figure 2. 56 . S4:20:0( 4,5,5,6) gHMBC 134 Figure 2. 57 . S4:20:0( 4,5,5,6) J - resolved 135 Figure 2. 58 . S4:20: 0( 4,5,5,6) ROESY 136 Table 2. 16 . S4:20:0( 3,5,6,6) Chemical shifts and coupling constants Molecular Formula: C 32 H 54 O 15 110 min Retention Time (ESI+): 51.42 mins HRMS: (ESI+) m/z calculated for C 32 H 58 NO 15 + ([M+NH 4 + ]): 696.3801, found: 696.3815 Fraction, Batch: #44, A - D Sample mass for NMR analysis: 0.7 mg NMR Solvent: CDCl 3 InChi Key: BYNYVKQGYIXALZ - ZWCYDAJKSA - N Carbon # (group) 1 H (ppm) 13 C (ppm) 1 (CH) 5.78 (d, J = 4.0 Hz) 89.14 2 (CH) 4.87 (dd, J = 10.4, 4.0 Hz) 70.56 - 1(CO) 173.62 - 2(CH 2 ) 2.38 (dd, J = 15.5, 5.7 Hz), 2.13 (dd, J = 15.5, 8.3 Hz) 41.17 - 3(CH) 1.84 (m) 31.73 - 4(CH 3 ) 0.90 (d, J = 6.7 Hz) 19.37 - 5(CH 2 ) 1.34 (m), 1.20 (m) 29.3 a - 6(CH 3 ) 0.88 (t, J = 7.4 Hz) 11.4 b 3 (CH) 5.53 (dd, J = 10.6, 9.2 Hz) 68.98 - 1(CO) 172.33 - 2(CH 2 ) 2.25 (dd, J = 15.0, 5.9 Hz), 2.04 (dd, J = 15.0, 8.3 Hz) 41.31 - 3(CH) 1.80 (m) 31.83 - 4(CH 3 ) 0.88 (d, J = 6.7 Hz) 19.28 - 5(CH 2 ) 1.31 (m), 1.19 (m) 29.3 a - 6(CH 3 ) 0.86 (t, J = 7.4 Hz) 11.4 b 4 (CH) 4.92 (dd, J = 10.7, 9.3 Hz) 68.82 - 1(CO) 173.85 - 2(CH 2 ) 2.35 (m), 2.31 (m) 27.57 - 3(CH 3 ) 1.12 (t, J = 7.6 Hz) 9.09 5 (CH) 4.16 (ddd, J = 10.3, 5.9, 2.6 Hz) 7 2.19 6 (CH 2 ) 3.64 (dd, J = 12.7, 2.4 Hz), 3.60 (dd, J = 12.7, 6.3 Hz) 61.57 137 Table 2.16. (continued) 1' (CH 2 ) 4.08 (d, J = 11.6 Hz), 4.06 (d, J = 11.6 Hz) 63.63 - 1(CO) 175.90 - 2(CH) 2.41 (sextet, J = 7.0 Hz) 41.10 - 3(CH 3 ) 1.15 (d, J = 7.0 Hz) 16.73 - 4(CH 2 ) 1.68 (m), 1.49 (m) 26.87 - 5(CH 3 ) 0.91 (t, J = 7.4 Hz) 11.75 2' (C) 103.45 3' (CH) 4.17 (d, J = 8.7 Hz) 78.35 4' (CH) 4.31 (t, J = 8.7 Hz) 72.79 5' (CH) 3.71 (m) 81.33 6' (CH 2 ) 3.89 (dd, J = 13.7, 2.8 Hz), 3.71 (m) 59.77 a - Two 13 C signals not resolved in 2D spectra (29.32, 29.33 ppm) b - Two 13 C signals not resolved in 2D spectra (11.35, 11.36 ppm) 138 Figure 2. 59 . S4:20:0( 3,5,6,6) 1 H NMR 139 Figure 2. 60 . S4:20:0( 3,5,6,6) 13 C NMR 140 Figure 2. 61 . S4:20:0( 3,5,6,6) 1 H - 1 H gCOSY 141 Figure 2. 62 . S4:20:0( 3,5,6,6) gHSQC 142 Figure 2. 63 . S4:20:0( 3,5,6,6) gHMBC 143 Figure 2. 64 . S4:20:0( 3,5,6,6) J - resol ved 144 Figure 2. 65 . S4:20:0( 3,5,6,6) ROESY 145 Table 2. 17 . S4:20:0( 2,6,6,6) Chemical shifts and coupling constants Molecular Formula: C 32 H 54 O 15 110 min Retention Time (ESI+): 51.65 mins HRMS: (ESI+) m/z calculated for C 32 H 58 NO 15 + ([M+NH 4 + ]): 696.3801, found: 696.3815 Fraction, Batch: #45, A - D Sample mass for NMR analysis: 0.5 mg NMR Solvent: CDCl 3 I nChi Key: WLTSSQZLQMTYDB - MPGSVAJHSA - N Carbon # (group) 1 H (ppm) 13 C (ppm) 1 (CH) 5.63 (d, J = 3.6 Hz) 89.78 2 (CH) 4.87 (dd, J = 10.4, 3.7 Hz) 70.52 - 1(CO) 172.71 a - 2(CH 2 ) 2.31 (dd, J = 15.7, 5.8 Hz), 2.06 (dd, J = 15.7, 8.3 Hz) 41.01 - 3(CH) 1.82 (m) 31.60 - 4(CH 3 ) 0.89 (d, J = 6.7 Hz) 19.4 b - 5(CH 2 ) 1.35 (m), 1.20 (m) 29.34 - 6(CH 3 ) 0.88 (t, J = 7.4 Hz) 11.4 c 3 (CH) 5.52 (dd, J = 10.7, 9.2 Hz) 68.87 - 1(CO) 172.28 - 2(CH 2 ) 2.24 (dd, J = 15.3, 5.7 Hz), 2.02 (dd, J = 15.3, 8.3 Hz) 41.32 - 3(CH) 1.79 (m) 31.73 - 4(CH 3 ) 0.88 (d, J = 6.7 Hz) 19.4 b - 5(CH 2 ) 1.31 (m), 1.19 (m) 29.4 d - 6(CH 3 ) 0.86 (t, J = 7.4 Hz) 11.4 c 4 (CH) 4.92 (dd, J = 10.6, 9.2 Hz) 68.71 - 1(CO) 172.71 a - 2(CH 2 ) 2.29 (dd, J = 15.3, 5.8 Hz), 2.11 (dd, J = 15.3, 8.4 Hz) 41 .21 - 3(CH) 1.83 (m) 31.81 - 4(CH 3 ) 0.91 (d, J = 6.7 Hz) 19.4 b - 5(CH 2 ) 1.32 (m), 1.20 (m) 29.4 d - 6(CH 3 ) 0.88 (, J = 7.4 Hz) 11.4 c 5 (CH) 4.14 (ddd, J = 10.3, 5.4, 2.8 Hz) 71.99 6 (CH 2 ) 3.62 (m) 61.91 146 Table 2.17. (continued) 1' (CH 2 ) 3.66 (d, J = 12.4 Hz ), 3.54 (d, J = 12.4 Hz) 64.59 2' (C) 104.16 3' (CH) 5.16 (d, J = 7.9 Hz) 80.31 - 1(CO) 172.38 - 2(CH 3 ) 2.27 (s) 21.02 4' (CH) 4.61 (t, J = 8.2 Hz) 71.52 5' (CH) 3.88 (dt, J = 8.6, 2.4 Hz) 82.32 6' (CH 2 ) 3.89 (d, J = 13.4, 2.5 Hz), 3.71 (d, J = 13.4, 2 .8 Hz) 59.76 a - Two 13 C signals overlapping b - Three 13 C signals not resolved in 2D spectra (19.36, 19.39, 19.43 ppm) c - Three 13 C signals not resolved in 2D spectra (11.38, 11.39, 11.42 ppm) d - Two 13 C signals not resolved in 2D spectra ( 29.40, 29.42 ppm) 147 Figure 2. 66 . S5:20:0( 2,6,6,6) 1 H NMR 148 Figure 2. 67 . S5:20:0( 2,6,6,6) 13 C NMR 149 Figure 2. 68 . S5:20:0( 2,6,6,6) 1 H - 1 H gCOSY 150 Figure 2. 69 . S5:20:0( 2,6,6,6) gHSQC 151 Figure 2. 70 . S5:20:0( 2,6,6,6) gHMBC 152 Figure 2. 71 . S5:20:0( 2,6,6,6) J - resolved 153 Figure 2. 72 . S5:20:0( 2,6,6,6) ROESY 154 Table 2. 18 . S5:22:1( 2,5,5,5,5 T ) Chemical shifts and coupling constants Molecular Formula: C 34 H 54 O 16 110 min Retention Time (ESI+): 52.55 mins HRMS: (ESI+) m/z calculated for C 34 H 58 NO 16 + ([M+NH 4 + ]): 7 36.3750, found: 736.3743 Fraction, Batch: #47, D Sample mass for NMR analysis: 0.2 mg NMR Solvent: CDCl 3 InChi Key: AHDZAIMCFBHKJA - QZUIWNAASA - N Carbon # (group) 1 H (ppm) 13 C (p pm) 1 (CH) 5.65 (d, J = 3.9 Hz) 89.99 2 (CH) 5.02 (dd, J = 10.4, 3.9 Hz) 70.06 - 1(CO) 175.85 - 2(CH) 2.37 (sextet, J = 7.0 Hz) 40.81 - 3(CH 3 ) 1.08 (d, J = 7.0 Hz) 16.25 - 4(CH 2 ) 1.55 (m), 1.39 (m) 26.77 - 5(CH 3 ) 0.77 (t, J = 7.4 Hz) 11.38 3 (CH) 5.63 (dd , J = 10.6, 9.3 Hz) 69.12 - 1(CO) 166.85 - 2 (C) 127.73 - 3(CH 3 ) 1.75 (m) 12.08 - 4(CH) 6.80 (m) 139.16 - 5(CH 3 ) 1.75 (m) 14.63 4 (CH) 5.03 (dd, J = 10.6, 9.3 Hz) 68.73 - 1(CO) 176.43 - 2(CH) 2.32 (sextet, J = 7.0 Hz) 41.43 - 3(CH 3 ) 1.04 (d, J = 7.0 Hz) 16.87 - 4(CH 2 ) 1.61 (m), 1.41(m) 26.62 - 5(CH 3 ) 0.84 (t, J = 7.4 Hz) 11.88 5 (CH) 4.15 (m) 71.49 6 (CH 2 ) 3.68 (dd, J = 12.7, 3.1 Hz), 3.61 (dd, J = 12.7, 5.6 Hz) 61.58 155 Table 2.18. (continued) 1' (CH 2 ) 4.21 (d, J = 11.8 Hz), 4.17 (d, J = 11.8 Hz) 63.39 - 1(CO) 175.96 - 2(CH) 2.41 (sextet, J = 7.0 Hz) 41.11 - 3(CH 3 ) 1.16 (d, J = 7.0 Hz) 16.69 - 4(CH 2 ) 1.68 (m), 1.49 (m) 26.88 - 5(CH 3 ) 0.92 (t, J = 7.4 Hz) 11.75 2' (C) 103.58 3' (CH) 4.15 (m) 78.31 4' (CH) 4.11 (td, J = 8.4, 3.3 Hz) 75.33 5' (CH) 3.90 ( ddd, J = 8.4, 5.3, 3.1 Hz) 78.79 6' (CH 2 ) 4.42 (dd, J = 12.3, 5.2 Hz), 4.23 (dd, J = 12.3, 3.0 Hz) 63.62 - 1(CO) 171.69 - 2(CH 3 ) 2.15 (s) 21.01 156 Figure 2. 73 . S5:22:1( 2,5,5,5,5 T ) 1 H NMR 157 Figure 2. 74 . S5:22:1( 2,5,5,5,5 T ) 13 C NMR 158 Figure 2. 75 . S5:22:1( 2,5,5,5,5 T ) 1 H - 1 H gCOSY 159 Figure 2. 76 . S5:22:1( 2,5,5,5,5 T ) gHSQC 160 Figure 2. 77 . S5:22:1( 2,5,5,5,5 T ) gHMBC 161 Figure 2. 78 . S5:22:1( 2,5,5,5,5 T ) J - resolved 162 Table 2. 19 . S5:21:0( 2,2,5,6,6) Chemical shifts and coupling constants Molecular Formula: C 33 H 54 O 16 110 min Retention Time (ESI+ ): 52.81 mins HRMS: (ESI+) m/z calculated for C 33 H 58 NO 16 + ([M+NH 4 + ]): 724.3750, found: 724.3783 Fraction, Batch: #48, A - D Sample mass for NMR analysis: 5.0 mg NMR Solvent: CDCl 3 InChi Key: OBSFKETUAQFCKT - KUQBJEAISA - N Carbon # (group) 1 H (ppm) 13 C (ppm) 1 (CH) 5.62 (d, J = 3.6 Hz) 89.52 2 (CH) 4.90 (dd, J = 10.4, 3.7 Hz) 70.48 - 1(CO) 172.3 a - 2(CH 2 ) 2.30 (dd, J = 15.5, 5.8 Hz), 2.05 (dd, J = 15.5, 8.3 Hz) 41.02 - 3(CH) 1.82 (m) 31.62 - 4(CH 3 ) 0.88 (d, J = 6.7 Hz) 19.34 - 5(CH 2 ) 1.33 (m), 1.20 (m) 29.35 - 6(CH 3 ) 0.87 (t, J = 7.4 Hz) 11.38 b 3 (CH) 5.55 (dd, J = 10.7, 9.2 Hz) 68.98 - 1(CO) 172.3 a - 2(CH 2 ) 2.23 (dd, J = 15.6, 5.5 Hz), 2.02 (dd, J = 15.5, 8.4 Hz) 41.26 - 3(CH) 1 .79 (m) 31.65 - 4(CH 3 ) 0.87 (d, J = 6.7 Hz) 19.41 - 5(CH 2 ) 1.30 (m), 1.17 (m) 29.37 - 6(CH 3 ) 0.85 (t, J = 7.4 Hz) 11.38 b 4 (CH) 4.97 (dd, J = 10.5, 9.2 Hz) 68.85 - 1(CO) 176.36 - 2(CH) 2.35 (sextet, J = 7.0 Hz) 41.13 - 3(CH 3 ) 1.10 (d, J = 7.0 Hz) 16.53 - 4(CH 2 ) 1.66 (m), 1.44 (m) 26.67 - 5(CH 3 ) 0.90 (t, J = 7.4 Hz) 11.83 5 (CH) 4.05 (ddd, J = 10.3, 5.2, 2.1 Hz) 71.13 6 (CH 2 ) 3.66 (dd, J = 12.8, 2.3 Hz), 3.58 (dd, J = 12.8, 5.3 Hz) 61.64 1' (CH 2 ) 3.63 (d, J = 12.3 Hz), 3.55 (d, J = 12.3 Hz) 64.27 c 2' (C) 104.65 163 Table 2.19. (continued) 3' (CH) 5.20 (d, J = 7.6 Hz) 79.30 - 1(CO) 172.01 - 2(CH 3 ) 2.22 (s) 20.99 4' (CH) 4.35 (t, J = 7.6 Hz) 73.90 5' (CH) 4.08 (ddd, J = 7.6, 6.1, 3.6 Hz) 80.43 6' (CH 2 ) 4.41 (dd, J = 12.1, 6.2 Hz), 4.28 (dd, J = 12.1, 3.6 Hz) 6 4.27 c - 1(CO) 171.72 - 2(CH 3 ) 2.12 (s) 21.01 a - Two 13 C signals not resolved in 2D spectra (172.34, 172.35 ppm) b - Two 13 C signals overlapping c - Two 13 C signals overlapping 164 Figure 2. 79 . S5:21:0( 2,2,5,6,6) 1 H NMR 165 Figure 2. 80 . S5:21:0( 2,2,5,6,6) 13 C NMR 166 Figure 2. 81 . S5:21:0( 2,2,5,6,6) 1 H - 1 H gCOSY 167 Figure 2. 82 . S5:21:0( 2,2,5,6,6) gHSQC 168 Figure 2. 83 . S5:21:0( 2,2,5,6,6) gHMBC 169 Figure 2. 84 . S5:21:0( 2,2,5,6,6) J - resolved 170 Figure 2. 85 . S5:21:0( 2,2,5,6,6) ROESY 171 Table 2. 20 . S4:21:0( 5,5,5,6) Chemical shifts and coupling constants Molecular Formula: C 33 H 56 O 15 110 min Retention Time (ESI+): 53.44 mins HRMS: (ESI+) m/z calculated for C 33 H 60 NO 15 + ([M+NH 4 + ]): 710.3957, found: 710.3993 Fraction, Batch: #52, A - D Sample mass for NMR analysis: 1.1 mg NMR Solvent: CDCl 3 InChi Key: XBHZETAMEBHMPK - LLOMWPEHSA - N Carbon # (group) 1 H (ppm) 13 C (ppm) 1 (CH) 5.87 (d, J = 4.0 Hz) 88.79 2 (CH) 4.78 (dd, J = 10.3, 4.1 Hz) 71.11 - 1 (CO) 177.57 - 2(CH) 2.41 (sextet, J = 7.0 Hz) 40.78 - 3(CH 3 ) 1.14 (d, J = 7.0 Hz) 16.21 - 4(CH 2 ) 1.64 (m), 1.45 (m) 26.73 - 5(CH 3 ) 0.87 (t, J = 7.4 Hz) 11.64 3 (CH) 5.58 (dd, J = 10.6, 9.3 Hz) 68.76 - 1(CO) 172.15 - 2(CH 2 ) 2.25 (dd, J = 15.8, 5.4 Hz), 2. 02 (dd, J = 15.8, 8.5 Hz) 41.1 a - 3(CH) 1.79 (m) 31.53 - 4(CH 3 ) 0.89 (d, J = 6.7 Hz) 19.45 - 5(CH 2 ) 1.31 (m), 1.19 (m) 29.40 - 6(CH 3 ) 0.86 (t, J = 7.4 Hz) 11.39 4 (CH) 4.93 (dd, J = 10.7, 9.3 Hz) 68.58 - 1(CO) 176.15 - 2(CH) 2.36 (sextet, J = 7.0 Hz) 41.1 a - 3(CH 3 ) 1.10 (d, J = 7.0 Hz) 16.42 - 4(CH 2 ) 1.66 (m), 1.44 (m) 26.64 - 5(CH 3 ) 0.90 (t, J = 7.4 Hz) 11.82 5 (CH) 4.15 (ddd, J = 10.6, 6.2, 2.5 Hz) 71.99 6 (CH 2 ) 3.63 (dd, J = 12.4, 2.4 Hz), 3.60 (dd, J = 12.4, 6.5 Hz) 61.51 172 Table 2.20. (continued) 1' (CH 2 ) 4.13 (d, J = 11.6 Hz), 4.06 (d, J = 11.6 Hz) 64.07 - 1(CO) 175.92 - 2(CH) 2.41 (sextet, J = 7.0 Hz) 41.1 a - 3(CH 3 ) 1.15 (d, J = 7.0 Hz) 16.73 - 4(CH 2 ) 1.68 (m), 1.49 (m) 26.89 - 5(CH 3 ) 0.91 (t, J = 7.4 Hz) 11.74 2' (C) 103.74 3' (CH) 4.17 (d, J = 8.9 Hz) 78.43 4' (CH) 4.31 (t, J = 8.9 Hz) 72.56 5' (CH) 3.71 (dt, J = 9.2, 2.3 Hz) 80.98 6' (CH 2 ) 3.88 (dd, J = 13.4, 2.4 Hz), 3.71 (dd, J = 13.4, 2.3 Hz) 59.71 a - Three 13 C signals not resolved in 2D spectra (41.08, 41.10, 41.13 ppm) 173 Figure 2. 86 . S4:21:0( 5,5,5,6) 1 H NMR 174 Figure 2. 87 . S4:21:0( 5,5,5,6) 13 C NMR 175 Figure 2. 88 . S4:21:0( 5,5,5,6) 1 H - 1 H gCOSY 176 Figure 2. 89 . S4:21:0( 5,5,5,6) gHSQC 177 Figure 2. 90 . S4:21:0( 5,5,5,6) gHMBC 178 Figure 2. 91 . S4:21:0( 5,5,5,6) J - resolved 179 Figure 2. 92 . S4:21:0( 5,5,5,6) ROESY 180 Table 2. 21 . S5:25:4( 2,5,5,5,8 P ) Chemical shifts and coupling constants Molecular Formula: C 37 H 54 O 16 110 min Retention Time (ESI+): 54.29 mins HRMS: (ESI+) m/z calculated for C 37 H 58 NO 16 + ([M+NH 4 + ]): 772.3750, found: 772.3748 Fraction, Batch: #55 - 56, A - G & 2nd purification w/F5 column #26 - 27 combined Sample mass for NMR analysis: < 0.1 mg NMR Solvent: d 3 - acetonitrile InChi Key: ZOZCRBKJAIYBPH - IJNFVYORSA - N Carbon # (group) 1 H (ppm) 13 C (ppm) 1 (CH) 5.58 (d, J = 3.9 Hz) 90.22 2 (CH) 4.90 (dd, J = 10.4, 3.9 Hz) 71.37 - 1(CO) 176.75 - 2(CH) 2.30 (sextet, J = 7.0 Hz) 41.74 a - 3(CH 3 ) 1.03 (d, J = 7.0 Hz) 16.54 - 4(CH 2 ) 1.53 (m), 1.35 (m) 27.71 - 5(CH 3 ) 0.83 (t, J = 7. 4 Hz) 12.23 3 (CH) 5.49 (dd, J = 10.7, 9.2 Hz) 71.49 - 1(CO) 172.11 - 2(CH 2 ) 3.55 (d, J = 16.0 Hz), 3.53 (d, J = 16.0 Hz) 41.73 a - 3 (C) 135.06 - 4(2 x CH) 7.19 (m) 130.86 - 5(2 x CH) 7.30 (m) 129.82 - 6(CH) 7.26 (m) 128.47 4 (CH) 5.03 (dd, J = 10.6, 9.2 Hz) 69.70 - 1(CO) 176.66 - 2(CH) 2.25 (sextet, J = 7.0 Hz) 42.06 b - 3(CH 3 ) 0.98 (d, J = 7.0 Hz) 17.19 - 4(CH 2 ) 1.57 (m), 1.37 (m) 27.40 - 5(CH 3 ) 0.85 (t, J = 7.4 Hz) 12.26 5 (CH) 4.12 ( J = 10.3, 5.7, 2.6 Hz) 72.08 6 (CH 2 ) 3.57 (dd, J = 12.2, 2.9 Hz), 3.45 (dd, J = 12.2, 5.7 Hz) 61.95 181 Table 2.21. (continued) 1' (CH 2 ) 4.01 (s) 64.02 - 1(CO) 176.82 - 2(CH) 2.42 (sextet, J = 7.0 Hz) 42.06 b - 3(CH 3 ) 1.11 (d, J = 7.0 Hz) 17.25 - 4(CH 2 ) 1.63 (m), 1.48 (m) 27.88 - 5(CH 3 ) 0.89 (t, J = 7.4 Hz) 12.21 2' (C) 104 .55 3' (CH) 3.98 (m) 78.29 4' (CH) 3.97 (t, J = 8.9 Hz) 75.16 5' (CH) 3.80 (ddd, J = 8.9, 6.7, 2.6 Hz) 80.59 6' (CH 2 ) 4.26 (dd, J = 12.2, 2.7 Hz), 4.19 (dd, J = 12.2, 7.0 Hz) 65.60 - 1(CO) 171.96 - 2(CH 3 ) 2.03 (s) 21.39 a - Two 13 C signals not resolved i n 2D spectra (41.73, 41.74 ppm) b - Two 13 C signals overlapping 182 Figure 2. 93 . S5:25:4( 2,5,5,5,8 P ) 1 H NMR 183 Figure 2. 94 . S5:25:4( 2,5,5,5,8 P ) 13 C NMR 184 Figure 2. 95 . S5:25:4 ( 2,5,5,5,8 P ) 1 H - 1 H gCOSY 185 Figure 2. 96 . S5:25:4( 2,5,5,5,8 P ) gHSQC 186 Figure 2. 97 . S5:25:4( 2,5,5,5,8 P ) gHMBC 187 Figure 2. 98 . S5:25:4( 2,5,5,5,8 P ) J - resolved 188 Figure 2. 99 . S5:25:4( 2,5,5,5,8 P ) ROESY 189 Table 2. 22 . S5:22:0( 2,2,6,6,6) Chemical shifts and coupling constants Molecular Formula: C 34 H 56 O 16 110 min Retention Time (ESI+): 56.08 mi ns HRMS: (ESI+) m/z calculated for C 34 H 60 NO 16 + ([M+NH 4 + ]): 738.3907, found: 738.3937 Fraction, Batch: #63, A - D Sample mass for NMR analysis: 1.0 mg NMR Solvent: CDCl 3 InChi Key: VVXKFHPDDJU GBH - NHDNHSBXSA - N Carbon # (group) 1 H (ppm) 13 C (ppm) 1 (CH) 5.62 (d, J = 3.6 Hz) 89.58 2 (CH) 4.90 (dd, J = 10.5, 3.6 Hz) 70.49 - 1(CO) 172.31 - 2(CH 2 ) 2.30 (dd, J = 15.5, 5.8 Hz), 2.05 (dd, J = 15.4, 8.3 Hz) 41.04 - 3(CH) 1.80 (m) 31.7 a - 4(CH 3 ) 0.88 (d, J = 6.7 Hz) 19.4 b - 5(CH 2 ) 1.33 (m), 1.20 (m) 29.4 c - 6(CH 3 ) 0.87 (t, J = 7.4 Hz) 11.4 d 3 (CH) 5.54 (dd, J = 10.7, 9.3 Hz) 69.0 e - 1(CO) 172.35 - 2(CH 2 ) 2.24 (dd, J = 15.3, 5.7 Hz), 2.03 (dd, J = 15.3, 8.4 Hz) 41.35 - 3(CH) 1.82 (m) 31.7 a - 4(CH 3 ) 0.88 (d, J = 6.7 Hz) 19.4 b - 5(CH 2 ) 1.33 (m), 1.20 (m) 29.4 c - 6(CH 3 ) 0.87 (, J = 7.4 Hz) 11.4 d 4 (CH) 4.98 (dd, J = 10.5, 9.3 Hz) 69.0 e - 1(CO) 172.96 - 2(CH 2 ) 2.29 (dd, J = 15.3, 5.8 Hz), 2.10 (dd, J = 15.3, 8.4 Hz) 41.27 - 3(CH) 1.83 (m) 31. 7 a - 4(CH 3 ) 0.91 (d, J = 6.7 Hz) 19.4 b - 5(CH 2 ) 1.33 (m), 1.20 (m) 29.4 c - 6(CH 3 ) 0.87 (, J = 7.4 Hz) 11.4 d 5 (CH) 4.04 (ddd, J = 10.2, 5.1, 2.2 Hz) 71.17 6 (CH 2 ) 3.67 (dd, J = 12.8, 2.6 Hz), 3.60 (dd, J = 12.8, 5.1 Hz) 61.67 190 Table 2.22. (continued) 1' (CH 2 ) 3.63 (d, J = 12.4 Hz), 3.55 (d, J = 12.4 Hz) 64.23 2' (C) 104.66 3' (CH) 5.18 (d, J = 7.6 Hz) 79.56 - 1(CO) 172.15 - 2(CH 3 ) 2.22 (s) 20.0 f 4' (CH) 4.35 (t, J = 7.6 Hz) 74.04 5' (CH) 4.08 (ddd, J = 7.7, 6.2, 3.6 Hz) 80.39 6' (CH 2 ) 4.43 (dd, J = 12.1, 6.2 Hz), 4.28 (dd, J = 12.1, 3.6 Hz) 64.27 - 1(CO) 171.68 - 2(CH 3 ) 2.13 (s) 20.0 f a - Three 13 C signals not resolved in 2D spectra (31.64, 31.74, 31.85 ppm) b - Three 13 C signals not resolved in 2D spectra (19.34, 19.38, 19.42 ppm) c - Three 13 C signal s not resolved in 2D spectra (29.35, 29.40, 29.42 ppm) d - Three 13 C signals not resolved in 2D spectra (11.39, 11.39, 11.42 ppm) e - Two 13 C signals not resolved in 2D spectra (68.96, 68.97 ppm) f - Two 13 C signals not resolved in 2D spectra (21.01, 21 .02 ppm) 191 Figure 2. 100 . S5:22:0( 2,2,6,6,6) 1 H NMR 192 Figure 2. 101 . S5:22:0( 2,2,6,6,6) 13 C NMR 193 Figure 2. 102 . S5:22:0( 2,2,6,6,6) 1 H - 1 H gCOSY 194 Figure 2. 103 . S5:22:0( 2,2,6,6,6) gHSQC 195 Figure 2. 104 . S5:22:0( 2,2,6,6,6) gHMBC 196 Figure 2. 105 . S5:22:0( 2,2,6,6,6) J - resolved 197 Figure 2. 106 . S5:22:0( 2,2,6,6,6) ROESY 198 Table 2. 23 . S4:22:0( 5,5,6,6) Chemical shifts and coupling constants Molecular Formula: C 34 H 58 O 15 110 min Retention Time (ESI+): 56.49 mins HRMS: (ESI+) m/z calculated for C 34 H 62 NO 15 + ( [M+NH 4 + ]): 724.4114, found: 724.4154 Fraction, Batch: #66, A - B Sample mass for NMR analysis: 1.3 mg NMR Solvent: CDCl 3 InChi Key: CACGEXBARUEDGB - FJNLHFDOSA - N Carbon # (group) 1 H (ppm) 13 C (ppm) 1 (CH) 5.81 (d, J = 4.0 Hz) 89.00 2 (CH) 4.85 (dd, J = 10.4, 4.0 Hz) 70.81 - 1(CO) 173.79 - 2(CH 2 ) 2.39 (dd, J = 15.7, 5.6 Hz), 2.12 (dd, J = 15.7, 8.4 Hz) 41.1 a - 3(CH) 1.83 (m) 31.65 - 4(CH 3 ) 0.88 (d, J = 6.7 Hz) 19.4 b - 5(CH 2 ) 1.35 (m), 1.21 (m) 29.32 - 6(CH 3 ) 0.87 (t, J = 7.4 Hz) 11.36 3 (CH) 5.54 (dd, J = 10.6, 9.2 Hz) 68.82 - 1(CO) 172.23 - 2(CH 2 ) 2.24 (dd, J = 15.5, 5.6 Hz), 2.03 (dd, J = 15.5, 8.4 Hz) 41.1 a - 3(CH) 1.80 (m) 31.72 - 4(CH 3 ) 0.88 (d, J = 6.7 Hz) 19.4 b - 5(CH 2 ) 1.31 (m), 1. 19 (m) 29.37 - 6(CH 3 ) 0.86 (t, J = 7.4 Hz) 11.39 4 (CH) 4.93 (dd, J = 10.7, 9.3 Hz) 68.55 - 1(CO) 176.18 - 2(CH) 2.37 (sextet, J = 7.0 Hz) 41.1 a - 3(CH 3 ) 1.10 (d, J = 7.0 Hz) 16.48 - 4(CH 2 ) 1.67 (m), 1.44 (m) 26.63 - 5(CH 3 ) 0.89 (t, J = 7.4 Hz) 11.83 5 (C H) 4.14 (ddd, J = 10.3, 6.3, 1.9 Hz) 72.15 6 (CH 2 ) 3.63 (d, J = 12.7, 2.5 Hz), 3.58 (d, J = 12.7, 6.3 Hz) 61.52 199 Table 2.23. (continued) 1' (CH 2 ) 4.09 (d, J = 11.7 Hz), 4.07 (d, J = 11.7 Hz) 63.80 - 1(CO) 175.90 - 2(CH) 2.41 (sextet, J = 7.0 Hz) 41.1 a - 3(CH 3 ) 1.15 (d, J = 7.0 Hz) 16.74 - 4(CH 2 ) 1.68 (m), 1.49 (m) 26.88 - 5(CH 3 ) 0.91 (t, J = 7.4 Hz) 11.75 2' (C) 103.58 3' (CH) 4.17 (d, J = 8.8 Hz) 78.39 4' (CH) 4.31 (t, J = 8.8 Hz) 72.74 5' (CH) 3.69 (dt, J = 8.9, 2.3 Hz) 81.22 6' (CH 2 ) 3.88 (dd, J = 13 .3, 2.4 Hz), 3.71 (dd, J = 13.3, 2.3 Hz) 59.78 a - Four 13 C signals not resolved in 2D spectra (41.11, 41.11, 41.14, 41.19 ppm) b - Two 13 C signals not resolved in 2D spectra (19.40, 19.41 ppm) 200 Figure 2. 107 . S4:22:0( 5,5,6,6 ) 1 H NMR 201 Figure 2. 108 . S4:22:0( 5,5,6,6) 13 C NMR 202 Figure 2. 109 . S4:22:0( 5,5,6,6) 1 H - 1 H gCOSY 203 Figure 2. 110 . S4:22:0( 5,5,6,6) gHSQC 204 Figure 2. 111 . S4:22:0( 5,5,6,6) gHMBC 205 Figure 2. 112 . S4:22:0( 5,5,6,6) J - resolved 206 Figure 2. 113 . S4:22:0( 5,5,6,6) ROESY 207 Table 2. 24 . S5:23:0( 2,5,5,5,6) Chemical shifts and coupling constant s Molecular Formula: C 35 H 58 O 16 110 min Retention Time (ESI+): 58.80 mins HRMS: (ESI+) m/z calculated for C 35 H 62 NO 16 + ([M+NH 4 + ]): 752.4063, found: 752.4092 Fraction, Batch: #75, A - D Sample mass for NMR analysis: 1.5 mg NMR Solvent: CDCl 3 InChi Key: RTBHWDIHLBPIRS - YGWKQMRCSA - N Carbon # (group) 1 H (ppm) 13 C (ppm) 1 (CH) 5.67 (d, J = 3.9 Hz) 89.36 2 (CH) 4.90 (dd, J = 10.4, 3.9 Hz) 7 0.24 - 1(CO) 175.83 - 2(CH) 2.37 (sextet, J = 7.0 Hz) 40.80 - 3(CH 3 ) 1.12 (d, J = 7.0 Hz) 16.31 - 4(CH 2 ) 1.62 (m), 1.42 (m) 26.7 b - 5(CH 3 ) 0.86 (t, J = 7.4 Hz) 11.70 3 (CH) 5.56 (dd, J = 10.7, 9.2 Hz) 68.88 - 1(CO) 172.29 - 2(CH 2 ) 2.24 (dd, J = 15.8, 5.4 Hz), 2.01 (dd, J = 15.8, 8.6 Hz) 41.1 a - 3(CH) 1.78 (m) 31.51 - 4(CH 3 ) 0.88 (d, J = 6.7 Hz) 19.44 - 5(CH 2 ) 1.30 (m), 1.18 (m) 29.40 - 6(CH 3 ) 0.85 (t, J = 7.4 Hz) 11.38 4 (CH) 4.94 (dd, J = 10.7, 9.2 Hz) 68.94 - 1(CO) 176.33 - 2(CH) 2.34 (sextet, J = 7.0 Hz) 41.1 a - 3(CH 3 ) 1.10 (d, J = 7.0 Hz) 16.47 - 4(CH 2 ) 1.67 (m), 1.44 (m) 26.7 b - 5(CH 3 ) 0.90 (t, J = 7.4 Hz) 11.83 5 (CH) 4.16 (m) 71.28 6 (CH 2 ) 3.65 (dd, J = 12.5, 2.4 Hz), 3.59 (dd, J = 12.5, 6.2 Hz) 61.65 208 Table 2.24. (continued) 1' (CH 2 ) 4.18 (d, J = 11.7 Hz), 4.15 (d, J = 11.7 Hz) 63.70 - 1(CO) 176.01 - 2(CH) 2.41 (sextet, J = 7.0 Hz) 41.1 a - 3(CH 3 ) 1.16 (d, J = 7.0 Hz) 16.69 - 4(CH 2 ) 1.68 (m), 1.49 (m) 26.88 - 5(CH 3 ) 0.92 (t, J = 7.4 Hz) 11.74 2' (C) 103.63 3' (CH) 4.14 (d, J = 8.8 Hz) 78.11 4 ' (CH) 4.11 (t, J = 8.6 Hz) 74.69 5' (CH) 3.87 (ddd, J = 8.3, 5.2, 2.8 Hz) 78.63 6' (CH 2 ) 4.39 (dd, J = 12.4, 5.2 Hz), 4.24 (dd, J = 12.4, 2.9 Hz) 63.49 - 1(CO) 171.74 - 2(CH 3 ) 2.15 (s) 20.99 a - Three 13 C signals not resolved in 2D spectra (41.12, 41.12, 41.13 ppm) a - Two 13 C signals not resolved in 2D spectra (26.65, 26.66 ppm) 209 Figure 2. 114 . S5:23:0( 2,5,5,5,6) 1 H NMR 210 Figure 2. 115 . S5:23:0( 2,5,5,5,6) 13 C NMR 211 Figure 2. 116 . S5:23:0( 2,5,5,5,6) 1 H - 1 H gCOSY 212 Figure 2. 117 . S5:23:0( 2,5,5,5,6) gHSQC 213 Figure 2. 118 . S5:23:0( 2,5,5,5,6) gHMBC 214 Figure 2. 119 . S5:23:0( 2,5,5,5,6) J - resolved 215 Figur e 2. 120 . S5:23:0( 2,5,5,5,6) ROESY 216 Table 2. 25 . S4:23:0( 5,6,6,6) Chemical shifts and coupling constants Molecular Formula: C 35 H 60 O 15 110 min Retention Time (ESI+): 59. 67 mins HRMS: (ESI+) m/z calculated for C 35 H 64 NO 15 + ([M+NH 4 + ]): 738.4270, found: 738.4266 Fraction, Batch: #80, A - D Sample mass for NMR analysis: 0.4 mg NMR Solvent: CDCl 3 InChi Key: HTYLQRCONXFYCT - ZVU FDNGLSA - N Carbon # (group) 1 H (ppm) 13 C (ppm) 1 (CH) 5.78 (d, J = 3.9 Hz) 89.15 2 (CH) 4.86 (dd, J = 10.3, 4.0 Hz) 70.69 - 1(CO) 173.65 - 2(CH 2 ) 2.39 (dd, J = 15.7, 5.6 Hz), 2.12 (dd, J = 15.7, 8.4 Hz) 41.2 a - 3(CH) 1.83 (m) 31.7 b - 4(CH 3 ) 0. 90 (d, J = 6.7 Hz) 19.4 c - 5(CH 2 ) 1.33 (m), 1.21 (m) 29.4 d - 6(CH 3 ) 0.87 (t, J = 7.4 Hz) 11.4 e 3 (CH) 5.53 (dd, J = 10.6, 9.2 Hz) 68.83 - 1(CO) 172.22 - 2(CH 2 ) 2.25 (dd, J = 15.3, 5.7 Hz), 2.04 (dd, J = 15.3, 8.4 Hz) 41.2 a - 3(CH) 1.83 (m) 31.7 b - 4(CH 3 ) 0.90 (d, J = 6.7 Hz) 19.4 c - 5(CH 2 ) 1.33 (m), 1.21 (m) 29.4 d - 6(CH 3 ) 0.87 (, J = 7.4 Hz) 11.4 e 4 (CH) 4.92 (dd, J = 10.7, 9.3 Hz) 68.66 - 1(CO) 172.82 - 2(CH 2 ) 2.29 (dd, J = 15.3, 5.8 Hz), 2.11 (dd, J = 15.3, 8.4 Hz) 41.2 a - 3(CH) 1.83 (m) 31.7 b - 4(CH 3 ) 0.90 (d, J = 6.7 Hz) 19.4 c - 5(CH 2 ) 1.33 (m), 1.21 (m) 29.4 d - 6(CH 3 ) 0.87 (, J = 7.4 Hz) 11.4 e 5 (CH) 4.12 (ddd, J = 10.4, 5.8, 2.0 Hz) 72.21 6 (CH 2 ) 3.64 (dd, J = 12.8, 2.2 Hz), 3.61 (dd, J = 12.6, 6.1 Hz) 61.55 217 Table 2.25. (continued) 1' (CH 2 ) 4.09 (d, J = 11.7 Hz), 4.06 (d, J = 11.7 Hz) 63.69 - 1(CO) 175.89 - 2(CH) 2.41 (sextet, J = 7.0 Hz) 41.2 a - 3(CH 3 ) 1.16 (d, J = 7.0 Hz) 16.74 - 4(CH 2 ) 1.68 (m), 1.49 (m) 26.88 - 5(CH 3 ) 0.92 (t, J = 7.4 Hz) 11.75 2' (C) 103.52 3' (CH) 4.17 (d, J = 8.7 Hz) 78.43 4' (CH) 4.31 (t, J = 8.7 Hz) 72.93 5' (CH) 3.71 (m) 81.35 6' (CH 2 ) 3.88 (dd, J = 14.2, 3.3 Hz), 3.70 (m) 59.81 a - Four 13 C signals not resolved in 2D spectra (41.10, 41.14, 41.23, 41.26 ppm) b - Three 13 C signals not resolved in 2D spectra (31.66, 31.7 8, 31.82 ppm) c - Three 13 C signals not resolved in 2D spectra (19.37, 19.39, 19.43 ppm) d - Three 13 C signals not resolved in 2D spectra (29.33, 29.39, 29.42 ppm) e - Three 13 C signals not resolved in 2D spectra (11.36, 11.39, 11.41 ppm) 218 Figure 2 . 121 . S4:23:0( 5,6,6,6) 1 H NMR 219 Figure 2. 122 . S4:23:0( 5,6,6,6) 13 C NMR 220 Figure 2. 123 . S4:23:0( 5,6,6,6) 1 H - 1 H gCOSY 221 Figure 2. 124 . S4:23:0( 5,6,6,6) g HSQC 222 Figure 2. 125 . S4:23:0( 5,6,6,6) gHMBC 223 Figure 2. 126 . S4:23:0( 5,6,6,6) J - resolved 224 Figure 2. 127 . S4:23:0( 5,6,6,6) ROESY 225 Table 2. 26 . S5:24:0( 2, 5,5,6,6) Chemical shifts and coupling constants Molecular Formula: C 36 H 60 O 16 110 min Retention Time (ESI+): 61.76 mins HRMS: (ESI+) m/z calculated for C 36 H 64 NO 16 + ([M+NH 4 + ]): 766.4220, found: 766 .4236 Fraction, Batch: #86, A - D Sample mass for NMR analysis: 1.1 mg NMR Solvent: CDCl 3 InChi Key: XPWKILWQFLWJSO - QBISVFFXSA - N Carbon # (group) 1 H (ppm) 13 C (ppm) 1 (CH) 5.66 ( d, J = 3.8 Hz) 89.51 2 (CH) 4.93 (dd, J = 10.5, 3.9 Hz) 70.22 - 1(CO) 172.38 - 2(CH 2 ) 2.34 (dd, J = 15.4, 5.8 Hz), 2.08 (dd, J = 15.4, 8.4 Hz) 41.07 - 3(CH) 1.83 (m) 31.68 - 4(CH 3 ) 0.89 (d, J = 6.7 Hz) 19.33 - 5(CH 2 ) 1.34 (m), 1.20 (m) 29.36 a - 6(CH 3 ) 0.8 6 (t, J = 7.4 Hz) 11.36 3 (CH) 5.54 (dd, J = 10.6, 9.2 Hz) 68.93 - 1(CO) 172.33 - 2(CH 2 ) 2.23 (dd, J = 15.4, 5.6 Hz), 2.03 (dd, J = 15.4, 8.4 Hz) 41.23 - 3(CH) 1.80 (m) 31.72 - 4(CH 3 ) 0.88 (d, J = 6.7 Hz) 19.39 - 5(CH 2 ) 1.31 (m), 1.18 (m) 29.36 a - 6(CH 3 ) 0.86 (t, J = 7.4 Hz) 11.38 4 (CH) 4.94 (dd, J = 10.6, 9.3 Hz) 68.90 - 1(CO) 176.35 - 2(CH) 2.35 (sextet, J = 7.0 Hz) 41.18 - 3(CH 3 ) 1.10 (d, J = 7.0 Hz) 16.53 - 4(CH 2 ) 1.67 (m), 1.44 (m) 26.65 - 5(CH 3 ) 0.90 (t, J = 7.4 Hz) 11.85 5 (CH) 4.14 (m) 71.31 6 (C H 2 ) 3.66 (dd, J = 12.6, 2.8 Hz), 3.59 (dd, J = 12.6, 6.0 Hz) 61.64 226 Table 2.26. (continued) 1' (CH 2 ) 4.16 (d, J = 11.6 Hz), 4.16 (d, J = 11.6 Hz) 63.49 - 1(CO) 175.95 - 2(CH) 2.41 (sextet, J = 7.0 Hz) 41.10 - 3(CH 3 ) 1.16 (d, J = 7.0 Hz) 16.71 - 4(CH 2 ) 1 .68 (m), 1.49 (m) 26.88 - 5(CH 3 ) 0.92 (t, J = 7.4 Hz) 11.75 2' (C) 103.57 3' (CH) 4.14 (m) 78.16 4' (CH) 4.11 (t, J = 8.4 Hz) 75.03 5' (CH) 3.90 (ddd, J = 8.4, 5.5, 3.0 Hz) 78.80 6' (CH 2 ) 4.42 (dd, J = 12.3, 5.3 Hz), 4.22 (dd, J = 12.3, 3.0 Hz) 63.69 - 1 (CO) 171.75 - 2(CH 3 ) 2.15 (s) 21.00 a - Two 13 C signals overlapping 227 Figure 2. 128 . S5:24:0( 2,5,5,6,6) 1 H NMR 228 Figure 2. 129 . S5:24:0( 2,5,5,6,6) 13 C NMR 229 Figure 2. 130 . S 5:24:0( 2,5,5,6,6) 1 H - 1 H gCOSY 230 Figure 2. 131 . S5:24:0( 2,5,5,6,6) gHSQC 231 Figure 2. 132 . S5:24:0( 2,5,5,6,6) gHMBC 232 Figure 2. 133 . S5:24:0( 2,5,5,6,6) J - resolved 233 Figure 2. 134 . S5:24:0( 2,5,5,6,6) ROESY 234 Table 2. 27 . S6:25:0( 2,2,5,5,5,6) Chemical shifts and coupling constants Molecular Formula: C 37 H 60 O 17 110 min Retention Time (ESI+) : 62.99 mins HRMS: (ESI+) m/z calculated for C 37 H 64 NO 17 + ([M+NH 4 + ]): 794.4169, found: 794.4166 Fraction, Batch: #89 - 90, A - G Sample mass for NMR analysis: 0.7 mg NMR Solvent: CDCl 3 InC hi Key: RRWBGWSVUMNBBH - NACAQSIGSA - N Carbon # (group) 1 H (ppm) 13 C (ppm) 1 (CH) 5.70 (d, J = 3.6 Hz) 89.41 2 (CH) 4.87 (dd, J = 10.4, 3.7 Hz) 70.37 - 1(CO) 175.94 - 2(CH) 2.40 (sextet, J = 7.0 Hz) 40.71 - 3(CH 3 ) 1.12 (d, J = 7.0 Hz) 16.1 1 - 4(CH 2 ) 1.63 (m), 1.44 (m) 26.77 - 5(CH 3 ) 0.86 (t, J = 7.4 Hz) 11.70 3 (CH) 5.57 (dd, J = 10.7, 9.2 Hz) 68.87 - 1(CO) 172.18 - 2(CH 2 ) 2.24 (dd, J = 15.8, 5.4 Hz), 2.01 (dd, J = 15.8, 8.6 Hz) 41.1 a - 3(CH) 1.78 (m) 31.47 - 4(CH 3 ) 0.88 (d, J = 6.7 Hz) 19 .46 - 5(CH 2 ) 1.30 (m), 1.18 (m) 29.42 - 6(CH 3 ) 0.85 (t, J = 7.4 Hz) 11.39 4 (CH) 4.97 (dd, J = 10.3, 9.2 Hz) 68.97 - 1(CO) 176.48 - 2(CH) 2.36 (sextet, J = 7.0 Hz) 41.1 a - 3(CH 3 ) 1.10 (d, J = 7.0 Hz) 16.48 - 4(CH 2 ) 1.66 (m), 1.45 (m) 26.70 - 5(CH 3 ) 0.90 ( t, J = 7.4 Hz) 11.83 5 (CH) 4.07 (ddd, J = 10.2, 5.1, 2.2 Hz) 71.15 6 (CH 2 ) 3.66 (dd, J = 12.8, 2.3 Hz), 3.59 (dd, J = 12.8, 5.1 Hz) 61.67 235 Table 2.27. (continued) 1' (CH 2 ) 4.14 (d, J = 11.8 Hz), 4.04 (d, J = 11.8 Hz) 64.10 - 1(CO) 175.81 - 2(CH) 2.44 ( sextet, J = 7.0 Hz) 41.1 a - 3(CH 3 ) 1.17 (d, J = 7.0 Hz) 16.66 - 4(CH 2 ) 1.70 (m), 1.51 (m) 26.90 - 5(CH 3 ) 0.93 (t, J = 7.4 Hz) 11.65 2' (C) 102.98 3' (CH) 5.23 (d, J = 8.1 Hz) 78.95 - 1(CO) 171.46 - 2(CH 3 ) 2.20 (s) 20.91 4' (CH) 4.35 (t, J = 8.1 Hz) 73.4 5 5' (CH) 4.04 (ddd, J = 8.2, 5.7, 3.6 Hz) 80.22 6' (CH 2 ) 4.40 (dd, J = 12.2, 5.6 Hz), 4.30 (dd, J = 12.2, 3.5 Hz) 63.76 - 1(CO) 171.65 - 2(CH 3 ) 2.13 (s) 20.99 a - Three 13 C signals not resolved in 2D spectra (41.10, 41.12, 41.14 ppm) 236 Figure 2. 135 . S6:25:0( 2,2,5,5,5,6) 1 H NMR 237 Figure 2. 136 . S6:25:0( 2,2,5,5,5,6) 13 C NMR 238 Figure 2. 137 . S6:25:0( 2,2,5,5,5,6) 1 H - 1 H gCOSY 239 Figure 2. 138 . S6:25:0( 2,2 ,5,5,5,6) gHSQC 240 Figure 2. 139 . S6:25:0( 2,2,5,5,5,6) gHMBC 241 Figure 2. 140 . S6:25:0( 2,2,5,5,5,6) J - resolved 242 Figure 2. 141 . S6:25:0( 2,2,5,5,5,6) ROESY 243 REFERENCES 244 REFERENCES 1 . Dixon, R.A. and D. Strack, Phytochemistry meets genome analysis, and beyond. Phytochemistry, 2003. 62 (6): p. 815 - 6. 2. Pichersky, E. and E. Lewinsohn, Convergent evolution in plant specialized metabolism , in Annual Review of Plant Biology, Vol 62 , S.S. M erchant, W.R. Briggs, and D. Ort, Editors. 2011, Annual Reviews: Palo Alto. p. 549 - 566. 3. Weinhold, A. and I.T. Baldwin, Trichome - derived O - acyl sugars are a first meal for caterpillars that tags them for predation. Proc Natl Acad Sci U S A, 2011. 108 (19) : p. 7855 - 9. 4. Balcke, G.U., et al., Multi - omics of tomato glandular trichomes reveals distinct features of central carb on metabolism supporting high productivity of specialized metabolites. Plant Cell, 2017. 29 (5): p. 960 - 983. 5. Ghosh, B. and A.D. Jones , Profiling, characterization, and analysis of natural and synthetic acylsugars (sugar esters). Anal. Methods, 2017. 6. S chilmiller, A.L., R.L. Last, and E. Pichersky, Harnessing plant trichome biochemistry for the production of useful compounds. Plant J, 2008. 54 (4): p. 702 - 11. 7. Glas, J.J., et al., Plant glandular trichomes as targets for breeding or engineering of resist ance to herbivores. Int J Mol Sci, 2012. 13 (12): p. 17077 - 103. 8. Kortbeek, R.W.J., et al., Engineering of tomato glandular trichomes f or the production of specialized metabolites , in Synthetic Biology and Metabolic Engineering in Plants and Microbes, Pt B : Metabolism in Plants , S.E. O'Connor, Editor. 2016, Elsevier Academic Press Inc: San Diego. p. 305 - 331. 9. Tissier, A., Glandular tric homes: what comes after expressed sequence tags? Plant J, 2012. 70 (1): p. 51 - 68. 10. Simmons, A.T. and G.M. Gurr, Trichomes of Lycopersicon species and their hybrids: effects on pests and natural enemies. Agricultural and Forest Entomology, 2005. 7 (4): p. 265 - 276. 11. Alba, J.M., M. Montserrat, and R. Fernandez - Munoz, Resistance to the two - spotted spider mite (Tetrany chus urticae) by acylsucroses of wild tomato (Solanum pimpinellifolium) trichomes studied in a recombinant inbred line population. Exp Appl Ac arol, 2009. 47 (1): p. 35 - 47. 12. Ghosh, B., T.C. Westbrook, and A.D. Jones, Comparative structural profiling of tr ichome specialized metabolites in tomato (Solanum lycopersicum) and S. habrochaites: acylsugar profiles revealed by UHPLC/MS and NMR. Metabolo mics, 2014. 10 (3): p. 496 - 507. 13. King, R.R., et al., Characterization of 2,3,4,3' - tetra - O - acylated sucrose ester s associated with the glandular trichomes of Lycopersicon - typicum. Journal of Agricultural and Food Chemistry, 1993. 41 (3): p. 469 - 473. 14. Li u, X.X., et al., Profiling, isolation and structure elucidation of specialized acylsucrose metabolites accumulatin g in trichomes of Petunia species. Metabolomics, 2017. 13 (7): p. 85. 245 15. Kroumova, A.B., D. Zaitlin, and G.J. Wagner, Natural variability in a cyl moieties of sugar esters produced by certain tobacco and other Solanaceae species. Phytochemistry, 2016. 130 : p. 218 - 27. 16. D'Auria, J.C., Acyltransferases in plants: a good time to be BAHD. Curr Opin Plant Biol, 2006. 9 (3): p. 331 - 40. 17. Fan, P., et al., In vitro reconstruction and analysis of evolutionary variation of the tomato acylsucrose metabolic network. Proc Natl Acad Sci U S A, 2016. 113 (2): p. E239 - 48. 18. Schilmiller, A.L., A.L. Charbonneau, and R.L. Last, Identification of a BAHD acetyltra nsferase that produces protective acyl sugars in tomato trichomes. Proc Natl Acad Sci U S A, 2012. 109 (40): p. 163 77 - 82. 19. Schilmiller, A.L., et al., Functionally divergent alleles and duplicated Loci encoding an acyltransferase contribute to acylsugar m etabolite diversity in Solanum trichomes. Plant Cell, 2015. 27 (4): p. 1002 - 17. 20. Nadakuduti, S.S., et al., Chara cterization of trichome - expressed BAHD acyltransferases in petunia axillaris reveals distinct acylsugar assembly mechanisms within the Solanac eae. Plant Physiol, 2017. 175 (1): p. 36 - 50. 21. Moghe, G.D., et al., Evolutionary routes to biochemical innovation revealed by integrative analysis of a plant - defense related specialized metabolic pathway. Elife, 2017. 6 . 22. Fan, P., et al., Evolution of a flipped pathway creates metabolic innovation in tomato trichomes through BAHD enzyme promiscuity. Nat Commun, 2017. 8 (1): p. 2080. 23. Arrendale, R.F., et al., Characterization of the sucrose ester fraction from Nicotiana - glutinosa. Journal of Agricultur al and Food Chemistry, 1990. 38 (1): p. 75 - 85. 24. Matsuzaki, T., et al., Leaf surface glycolipids from Nicotiana - acuminata and Nicotiana - pauciflora. Agricultural and Biological Chemistry, 1991. 55 (5): p. 1417 - 1419. 25. Sarkinen, T., et al., A phylogenetic framework for evolutionary study of the nightshades (Solanaceae): a dated 1000 - tip tre e. BMC Evol Biol, 2013. 13 : p. 214. 26. Ghosh, B. and A.D. Jones, Dependence of negative - mode electrospray ionization response factors on mobile phase composition and mol ecular structure for newly - authenticated neutral acylsucrose metabolites. Analyst, 201 5. 140 (19): p. 6522 - 31. 27. Ekanayaka, E.A., M.D. Celiz, and A.D. Jones, Relative mass defect filtering of mass spectra: a path to discovery of plant specialized metaboli tes. Plant Physiol, 2015. 167 (4): p. 1221 - 32. 28. Sumner, L.W., et al., Proposed minim um reporting standards for chemical analysis Chemical Analysis Working Group (CAWG) Metabolomics Standards Initiative (MSI). Metabolomics, 2007. 3 (3): p. 211 - 221. 29. Rob ins, R.J., et al., Esterification reactions in the biosynthesis of tropane alkaloids i n transformed root cultures. Plant Cell Tissue and Organ Culture, 1994. 38 (2 - 3): p. 241 - 247. 30. Josephnathan, P., J.R. Wesener, and H. Gunther, A two - dimensional NMR - stu dy of angelic and tiglic acid. Organic Magnetic Resonance, 1984. 22 (3): p. 190 - 191. 246 31 . Kroumova, A.B. and G.J. Wagner, Different elongation pathways in the biosynthesis of acyl groups of trichome exudate sugar esters from various solanaceous plants. Plant a, 2003. 216 (6): p. 1013 - 21. 32. Treutler, H., et al., Discovering regulated metabolit e families in untargeted metabolomics studies. Anal Chem, 2016. 88 (16): p. 8082 - 90. 33. Castillo, M., et al., Partially acylated glucose and sucrose derivatives from Salp iglossis - sinuata (Solanaceae). Journal of Chemical Research - S, 1989(12): p. 398 - 399. 34. King, R.R., et al., 3,4 - di - O - isobutyryl - 6 - O - caprylsucrose - the major component of a novel sucrose ester complex from the type - B glandular trichomes of Solanum - berthau ltii Hawkes (Pi 473340). Journ al of the Chemical Society - Chemical Communications, 1986(14): p. 1078 - 1079. 35. King, R.R., R.P. Singh, and L.A. Calhoun, Isolation and characterization of 3,3',4,6 - tetra - O - acylated sucrose esters from the type - B glandular tri chomes of Solanum - berthaultii Hawkes (Pi - 265857). Carbohydrate Research, 1987. 166 (1): p. 113 - 121. 36. Begum, A.S., et al., New pairs of acyl sucroses from Petunia nyctaginiflora Juss. Journal of the Indian Chemical Society, 2004. 81 (6): p. 495 - 500. 37. Be gum, A.S., et al., Two novel a cyl sucroses from Petunia nyctaginiflora. Indian Journal of Chemistry Section B - Organic Chemistry Including Medicinal Chemistry, 2005. 44 (3): p. 648 - 650. 38. Singh, A.P., et al., Two acyl sucroses from Petunia nyctaginiflora. Phytochemistry, 2003. 63 (4): p . 485 - 9. 39. Severson, R.F., et al., Isolation and characterization of the sucrose esters of the cuticular waxes of green tobacco leaf. Journal of Agricultural and Food Chemistry, 1985. 33 (5): p. 870 - 875. 247 Chapter 3: Unexpect ed diversity in acylsugar m etabolites: acylinositols from Solanum quitoense 3. 1 Introduction Plants across the family Solanaceae have proven to be prolific at synthesizing diverse arrays of acylsugar specialized metabolites (SMs) in epidermal cells known a s glandular trichomes (GTs) [1] . Documented acylsugars in the genus Solanum include glucose a nd sucrose decorated with branched and straight chain aliphatic ester groups [2, 3] , malonyl esters in Petunia [4] , and acylsucroses esterified by saturated and un saturated ester groups including the new report of phenylacetyl and tiglyl esters in Salpiglo ssis sinuata (Chapter 2 of this dissertation). Recent efforts to elucidate the biochemical pathways responsible for the production of acylsugars have identified a class of BAHD acyltransferase enzymes expressed in GTs of Solanaceae species that use CoA thi oesters and sucrose substrates to catalyze multi - step position - selective acylation of sucrose substrates [5 - 8] . Of particular note is the observation that small c hanges in amino acid sequences in the acyltransferases can alter acyl acceptor substrate pref erences. Variations in acyltransferase substrate preference raises the potential that mutations in BAHD acyltransferase sequences might allow these enzymes to cata lyze acylations of alternative carbohydrate cores. To explore acylsugar chemistry in the Sola naceae more broadly, the focus of this study has been to define acylsugar diversity from GTs of Solanum quitoense (also known as naranjilla or lulo) (images in Ap pendix Figure 3.7), a species that is genetically removed from oft - studied acylsugar producer s of the family Solanaceae, such as wild tomato ( Solanum pennellii and Solanum habrochaites ), tobacco ( Nicotiana tabacum ), and Petunia species [2 - 4, 9, 10] . S. q uitoense is a little - known species native to South America that is widely appreciated for the pleasant aroma and intense flavor of its fruits [11, 12] , and has been suggested as a promising source for new food flavors. Investigations of S. quitoense chemis tries have targeted flavor and aroma producing specialized metabolites from fruits. Forero et al. found that odor - active volatile compounds (Z) - 3 - hexenal, ethyl butanoate, 3 - sulphanylhexyl acetate, and ethyl hexanoate were key aroma compounds [11] . Process ing fruit 248 (juicing or drying) has been used to extend post - harvest shelf life, but spermidine derivatives formed during processing contributed to undesirable bitterness [13] . A series of C 13 - norisoprenoid glucoconjugates were isolated from homogenized leav es of S. quitoense , and are believed to contribute to flavor and aroma after they are transpo rted to the fruit and deglycosylated [14] . Further studies of volatile compounds from glycosidic precursors isolated from fruit peelings act as precursors for mono terpenes (linalool, hotrienol, nerol oxide and linalool oxides), which contribute to the frui t flavor [14] . At this stage, little is known about other specialized metabolites that are produced outside the fruit, particularly what plant chemical defenses ma y be employed in leaf surfaces of S. quitoense and how they may differ from previous examinat ions of species of the Solanaceae. In this investigation, the specialized metabolites of S. quitoense aerial tissues were explored by subjecting extracts of S. qu itoense leaves and stems to untargeted metabolite profiling using UHPLC/MS. The findings present evidence of acylsugars (believed to be GT derived) with fewer elemental formulas and fewer isomers than prolific acylsugar accumulators in the genus Solanum [2 , 15] , and these displayed unusual molecular masses and fragment ion spectra. To uncover the novelty in acylsugar composition from S. quitoense and facilitate discovery of genetic factors responsible for their biosynthesis, several acylsugar metabolites we re puri fied and their structures identified using NMR spectroscopy. In doing so, acylsugars with unexpected carbohydrate chemistries were discovered , which include a group of acylated myo - inositols and myo - inositol glycosides ( N - acetylglucosaminyl, glucopy ranosyl and xylopyranosyl), suggesting novel enzyme substrate selectivities are at work in acylsugar biosynthesis in S. quitoense . Where previous reports about acylsugars from the family Solanaceae have identified novel ester gr oups and differences among a cyl group substitution positions on glucose - and sucrose - based acylsugars, the current findings indicate that structures of core carbohydrates in acylsugars are more variable than is widely appreciated. 249 3. 2 Materials and Methods 3. 2.1 Plant Cultivation an d metabolite extraction S. quitoense seeds were obtained from The New York Botanical Gardens (NY, USA), and were grown using Jiffy peat pellets in growth chambers at MSU. Additional plant growth and extraction metadata are provi ded in Appendix Table 3.2. F or metabolite profiling and purification, aerial tissues of 20 ten - week - old S. quitoense plants (~0.25 m height) were extracted in 1.9 L of acetonitrile: isopropanol (AcN:IPA, v/v, 1:1) for 10 mins (plants were cut at the stems and stem junctions). For met abolite profiling, a 100 µL aliquot of this bulk extract solution was diluted in 800 µL AcN:IPA containing 0.1 µM telmisartan added as internal standard and 100 µL of aqueous 0.046 M formic acid (to inhibit rearrangements of acy l group positions). A previo us study from our laboratories showed that gentle extraction of young leaflets of S. quitoense revealed acylsugar metabolites attributed to trichomes [1] . For comparison of chromatographic retention times and mass spectra, a Sol anum pennellii LA0716 leaf d ip extract was prepared by removing two leaflets mins. 3. 2.2 Profiling of acylsugar metabolites using UHPLC/MS and MS/MS Metabo lite profiling was performed using a Waters Xevo G2 - XS quadrupole time - of - flight mass spectrometer (QTof/MS) equipped with an Ascentis Express C18 Analytical HPLC column (10 cm × 2.1 mm, 2.7 µm particle size) and operated using CID in positive - and negativ e - ion modes. Untargeted CID spectra were acquired over m/z 50 - 1500 using MS E [16] in continuum mode (6 V and 10 - 60 V collision potential ramp to generate fragment ions, each with 0.2 s acquisition times). Additional metadata regarding the UHPLC/MS methods are provided in Appendix Tab le 3.3. Deep profiling of metabolites was performed using Waters Progenesis QI software platform. Additional metadata regarding Progenesis QI settings and strategy are provided in Appendix Table 3.4. Acylsugar annotations from n egative - and positive - ion mo de LC/MS data are provided in Appendix Tables 3.5 and 3.6. 250 MS/MS product ion spectra were generated using a CID ramp from 5 - 60 V collision potential (0.5 s per spectrum). Acylsugar precursor adducts [M+HCOO] - and [M+NH 4 ] + were selected for MS/MS analysis in positive and negative - ion modes respectively (MS/MS precursor mass and time windows are shown in Appendix Tables 3.7 and 3.8), with low mass and high mass resolving quadrupole resolution settings at 20 (Waters system specific parameters) and sensitivity mode data acquisition. MS/MS product ion spectra of [M - H] - and [M+H] + were generated by altering cone voltages to produce those precursor ions. Additional source parameters and LC/MS metadata are located in Appendix Table 3.9. MS/MS product ion spectra of acylinositol metabolites identified by NMR analysis are located in Appendix Figures 3.8 - 3.25. Annotation of S. quitoense acylsugars by LC/MS was carried out from ESI mass spectra generated in positive - and negative - ion modes by assigning molecular mass an d formula, the masses of acyl groups and the number of acylations, and by assigning the elemental formulas (EF) of the carbohydrate cores and relative ring substitutions when pertinent. This information was assembled using untar geted MS E acquisition with t wo quasi - simultaneous collision potential functions (6 V and 10 - 60 V collision potential ramp, 0.2 s acquisition times). Abundant low energy ammonium [M+NH 4 ] + and formate [M+HCOO] - adduct ions, as well as sodium [M+Na] + and chlo rine [M+Cl] - ions assisted a ssignments of EFs. Collision induced dissociation (CID) at elevated potentials was used to generate useful fragment ion data for assigning acyl group carbon lengths, the number of acylations, the mass of the carbohydrate core an d the relative ring substitu tion of acyl groups. Further evidence supporting structural assignment was performed by examination of high energy CID spectra followed by metabolite purification and NMR analysis. 3. 2.3 Purification of acylsugar metabolites by semi - preparative HPLC For m etabolite purification, approximately 1 L of the S. quitoense bulk extract (described in Section 3. 2.1) was concentrated to dryness under vacuum, dissolved in 5 mL of AcN:IPA, and fractionated by repeated injection of 200 - µL samples onto a Thermo Scientifi c Acclaim 120 C18 HPLC column (4.6 x 150 mm, 5 µm particles) with automated fraction collection (Supplemental Materials). Additional methods and metadata are provided in Appendix Table 3.10. 251 3. 2.4 Analysis of acylsugar s by NMR spectroscopy HPLC fractions o f sufficient purity of a single metabolite, as assessed by LC/MS, were combined and concentrated to dryness under a flow of N 2 gas. Samples were dissolved in acetonitrile - d 3 (99.96 atom % D) and transferred to solvent - matched Shigemi tubes or Kontes tubes for NMR analysis. 1 H, 13 C, gCOSY, gHSQC, coupled - gHSQC, gHMBC, J - resolved 1 H, TOCSY and ROESY NMR experiments were performed using a Bruker Avance 900 MHz spectrometer equipped with a TCI triple resonance probe or an A gilent DDR2 500 MHz spectrometer equip ped with OneNMR probe (with Protune accessory for hands - off tuning). 1D - TOCSY transfer experiments were performed using a Varian Inova 600 MHz spectrometer equipped with a Nalorac 5 mm PFG switchable probe (pretuned f or 1 H and 13 C). All spectra were refer enced to non - deuterated solvent signals: acetonitrile - d 3 H C = 118.70 ppm). NMR spectra were processed using TopSpin 3.5pl7 or MestReNova 12.0.0 software. Additional NMR metadata are located in Appendix T ables 3.11 - 3.13. NMR spectra, chemical shifts and coupling constants are located in Appendix Figures 3.26 - 3.95 and Tables 3.14 - 3.28. 1D ( 1 H, 13 C and TOCSY) and 2D (gCOSY, gHSQC, gHMBC, J - resolved, TOCSY and ROESY) NMR spectroscopic techniques served as th e basis for structure elucidation of n ine purified S. quitoense acylsugars. Because these metabolites were identified without authentic standards or synthetic confirmation, their structures should be considered putative, meeting level 2 criteria of the Met abolomics Standards Initiative guideli nes [17] . 3. 3 Results and Discussion 3. 3.1 UHPLC/ESI/CID/QTof/MS profiling and NMR structural elucidation establishes diversity of S. quitoense acylinositols Profiling of S. quitoense metabolites using UHPLC/QTof MS re vealed a group of S. quitoense metabol ites eluting between 35 - 86 mins using a 110 - min gradient that exhibited molecular masses, relative mass defects and fragment ion spectra consistent with acylsugar SMs [4, 18, 19] . Compared to previous reports of acylsu gar profiles in Solanum habrochaites a ccessions [2] and three Petunia species [4] , LC/MS 252 chromatograms of S. quitoense exhibited fewer peaks annotated as acylsugars. As shown in Figure 3.1, at least 9 abundant acylhexoses and acyl disaccharides exhibited retention times in the range of 57 - 81 m ins and are evident in the base peak intensity (BPI) chromatogram. Figure 3. 1 . UHPLC/ESI/MS base peak intensity chromatogram of S. quitoense extract generated in negative - ion mode. Annotated acylsugars (formate adducts) detected using CID Function 1 = 6 V (10x magnification). Abbreviations of sugar groups a re as follows: myo - inositol (I), N - acetylglucosaminyl (NAG), glucopyranosyl (G) , and xylopyranosyl (X). One example for abbreviation - ester s or double bonds in the acyl groups, and th e numbers in parentheses describe the number of carbon atoms in each acyl group. 3. 3.2 LC/MS profiling and NMR structural elucidation establishes acylated myo - inositols Profiling and annotation of S. quitoense acylsugars by LC/MS was carried out by generat ing ESI mass spectra in positive - and negative - ion modes. To illustrate this, the most abundant peak in the base peak intensity (BPI) chromatogram (Figure 3.1) at t r = 73.90 mins yielded mass spectra with ions annotated as [M+HCOO] - at m/z 617.35 in negati ve - ion mode and [M+NH 4 ] + at m/z 590.39 in positive - ion mode (Figure 253 3.2A - B). These ion masses are consistent with an acylated hexose of EF C 30 H 52 O 10 . Because hexoses have six oxygen atoms, the four additional oxygens suggested four acylations, as each acyl ation adds one oxygen atom. Negative - ion spectra generated at elevated collision potentials showed a fragment ion at m/z 171.14, consistent with C10 carboxylate (acyl anion) of EF C 10 H 19 O 2 - (Figure 3.2A). Acyl ated hexoses from S. quitoense shared molecular masses with calculated masses of acylated glucoses. Information obtained from negative - ion mode CID spectra alone usually facilitates assignments of the number of carbon atoms of each ester group [18] . Howeve r, their negative - ion mode CID spectra generate d using a collision potential ramp from 10 - 60 V exhibited lower abundances of fragment ions than observed for acylglucoses or acylsucroses (Figure 3.2A). Positive - ion mode spectra obtained at elevated collisio n energy showed abundant fragment ions that wer e informative for annotating the number acylations and the number of carbons in each ester group, whereas this information was not readily obtained from negative - ion mode spectra. Further confirmation of C10 c arboxylic acid esters comes from observation of the m/z 155.14 acylium fragment ion with EF C 10 H 19 O + and neutral mass losses from [M+H] + of 172.15 Da (decanoic acid: C 10 H 20 O 2 ) yielding m/z 401.22 (Figure 3.2B). Neutral loss of 60.02 Da (acetic acid: C 2 H 4 O 2 ) from [M+H] + generated in - source yielded m/z 5 13.34, consistent with an acetate (C2) ester. The neutral losses of C2 or C10 groups from [M - C10 - C2] + at m/z 341.20 revealed mono - esterified fragment masses consistent with [M - C10 - 2C2] + at m/z 281.17 and [M - 2C10 - C2] + at m/z 169.05 respectively. Thus, the acylated hexose was assigned two C2 and two C10 groups. Using only LC/MS - based analyses, mass spectra of S. quitoense acylated hexoses shared resemblance to previously discovered acylglucose derivatives [20, 21] . This is because the hexose core (C 6 H 12 O 6 ) s hares the molecular formula of glucose and all other carbons, hydrogen and oxygen atoms were accounted for b y acyl group assignments. 254 Figure 3. 2 . CID mass spectra of acylsugar at t r = 73.90 mins (later annotated I4:24:0(2,2,10, 10) according to NMR results) using 10 - 60 V MS E ramp (0.2 s acquisition times). (A) Negative ion mode. (B) Positive ion mode. While LC/MS - based techniques provide annotations of acylsugar metabolites in terms of the number of acyl groups and the acyl group chain lengths, these analyses yield limited in formation about the atomic connectivity and stereochemistry of atoms that define each carbohydrate group. A surprising finding from structural characterization of purified S. quitoense acylsugars by NMR spectr oscopy revealed chemical shifts that differed f rom acylglucoses [20, 21] . For acylsugar at t r = 73.90 min, the absence of a characteristic downfield doublet for the anomeric ( or ) ring hydrogen was inconsistent with glucoses [20] . 1 H spectra of hexose core hydrogens were consistent with the conclusi ons that acylated hexose positions exhibited chemical shifts in the range of H 4.7 - 5.5 ppm that were attributed to ~1 ppm downf ield shifts compared to non - esterified analogues, which ranged from H 3.3 - 3.8 ppm (ring hydrogen chemical shifts are summarized in Appendix Table 3.14). In addition, HSQC spectra (used to determine the number of protons attached to 255 each carbon) were absen t of methylene (CH 2 ) ring hydrogens and suggested a cyclitol ring syst em, further confirmed by 2D - TOCSY, which showed a complete set of 1 H - 1 H spin - spin coupled ring hydrogens. Ring 1 H - 1 H couplings determined from 1 H and J - resolved spectra (and confirmed by COSY spectra) showed decisive coupling constants for discerning the r elative stereochemistry of hexose ring hydrogens. A hydrogen labeled H - 2 was shifted furthest downfield and served as the starting point for tracing the connectivity and relative stereoc hemistry of the cyclitol ring system (Figure 3.2). An equatorial H - 2 s ignal at 5.49 ppm showed couplings to axial proton H - 1 at 4.81 ppm and H - 3 at 4.97 ppm ( J 1,2 = J 2,3 = 3.0 Hz), both of which were shifted downfield to suggest esters at these positions. The H - 1 signal showed coupling to axial H - 6 at 3.77 ppm ( J 1,6 = 10.2 H z) which lacked an ester group, while H - 3 signal showed coupling to axial H - 4 at 5.21 ppm ( J 3,4 = 10.0 Hz) at the site of another ester group. Finally, H - 5 signal at 3.50 ppm (position n ot esterified) was consistent with axial couplings to H - 4 and H - 6 ( J 4, 5 = J 5,6 = 9.5 Hz), confirming a six - membered cyclitol ring. Thus, from NMR results the hexose core was assigned as a myo - inositol tetraester. Assignments of acyl group branching and att achment positions of acylinositols are discussed in Section 3. 3.6. Based on the MS and NMR results, the acylinositol whose mass spectrum is shown in Figure 3.2A was arbohydrate core is myo - and the numbers in pare ntheses describe the number of carbon atoms in each acyl group. Furt her evidence of this structural assignment is provided in Appendix Table 3.24 and Appendix Figures 3.70 - 3.77. The basis for the clockwise carbon numbering system used for assignment of myo - inositols in this study traces from the biosynthetic conversion of D - glucose 6 - phosphate to 1L - myo - inositol - 1 - phosphate [22] . Due to the plane of symmetry of the myo - inositol ring system about the C2 and C5 positions, an enantiomeric assignment is possibl e by substituting the C1 and C3 positions (1D - myo - inositol), followe d 256 by counterclockwise numbering. Due to the scarcity of purified material (most < 3 mg), absolute stereochemical configurations were not determined. 3. 3.3 Negative - ion mode MS/MS spectra of acylated myo - inositols differ from acylglucoses Mass spectra of m yo - inositol - based acylsugars generated at elevated collision potentials in negative - ion mode showed few fragment ions in the middle of the spectrum corresponding to neutral losses of acyl groups, unlike previously reported acylglucose analogues [20, 21] . T o further explore their differences, Figure 3.3A - B compares MS/MS product ion spectra of [M+HCOO] - for tri - acylated inositol I3:24:0(2,10,12) from S. quitoense and triacylglucose G3:19:0(4 ,5,10) from Solanum pennellii LA0716 (structure in Figure 3.3B and 3 .3D provided by Dr. Banibrata Ghosh and Dr. Xiaoxiao Liu). Both spectra display carboxylate anions for C10 ester groups at m/z 171, however acylglucoses differ from acylinositols in that t he former yield prominent product ions derived from acyl group neutr al losses, including displacement of the C10 acid (neutral loss of C 10 H 20 O 2 , 172.15 Da) by formate (CHOO, 45.00 Da) to the glucose ring ( m/z 361.15, [M - C10+HCOO] - ). This suggests that prox imity of charge to ester groups promotes generation of either carbox ylate anions from ester groups or neutral losses of ketenes, and is more significant in acylglucoses than acylinositols. Another acylglucose ion showed the neutral loss of H 2 O (18.01 Da) a nd C10 (154.14 Da) as a neutral ketene ( m/z 315.14, [M - C10 - H 2 O] - ). A neutral ketene loss is expected to result from removal of an alpha proton ( - H) of an ester group via an alkoxide anion. The intermediate negative ion can displace the oxygen where the acyl group is attached, liberating a neutral ketene group such as C10. The observation of a fragment ion at m/z 315.14 suggests the ion is further stabilized by neutral loss of H 2 O, though concerted elimination of the neutral carboxylic acid cannot be ruled out. In the case of acylated myo - inositol, it is likely that the energy barrier for an alkoxide ion to reach across the molecule to remove an es ter group alpha proton is much gr eater than in the case of glucose (where the charge can migrate to one of the primary alcohol oxygens that can reach across the rings with a lower activation energy barrier). Interestingly, the MS/MS product ion spectrum of [M - H] - of acylinositol (Figure 3 .3C) shows similar relative ion abundances for C10, C12 ( m/z 199.17) and [M - H] - by comparison 257 to MS/MS product ions of [M+HCOO] - (Figure 3.3A), suggesting the pathway for generating carboxylate ions goes through neutral los s of formic acid, formation of al koxide ion, followed by liberation of carboxylate anion (perhaps via displacement by an alkoxide ion on an adjacent carbon). In contrast, the MS/MS product ion spectrum of [M - H] - of acylglucose (Figure 3.3D) exhibits differ ent abundances of neutral losses relative to carboxylate formation when compared to MS/MS product ions from [M+HCOO] - in that lesser abundant neutral loss fragment ions are observed in the center of the spectrum (Figure 3.3B). These results suggest there are multiple fragmentation pathwa ys for ion activation of acylglucoses, the formate adduct plays a role in this process by driving charge localization, and the acylsugar core structure influences CID fragmentation in ways that may discriminate different ca rbohydrate cores in acylsugars. Figure 3. 3 . Negative ion mode MS/MS product ion spectra of tri - acylated myo - inositol I3:24:0(2,10,12) from S. quitoense and triacylated glucose G3:19:0(4,5,10) from Solanum pennellii LA0716 (all sp ectra generated with a linear 5 - 6 0 V MS E collision energy ramp with 0.5 s acquisition time) (A) products of I3:24:0(2,10,12) [M+HCOO] - (spectrum magnified 15x over the range m/z 51 - 593) (B) products of G3:19:0(4,5,10) [M+HCOO] - . (C) products of I3:24:0(2,1 0,12) [M - H] - . (D) products of G3:19:0(4,5,10) [M - H] - . 258 3. 3.4 LC/MS profiling and NMR structural elucidation establishes acylated myo - inositol glycosides 3. 3.4.1 Discovery of 4 - O - N - acetylglucosaminyl (NAG) acylated myo - inositols Al though acylated myo - inosito ls were of high abundance in the S. quitoense extracts, a group of less abundant myo - inositol glycosides were also observed. Documented neutral acylsugar metabolites within the Solanaceae are present as derivatives of carbon, hyd rogen and oxygen [2 - 4] . The refore, a peak in the S. quitoense extract chromatogram at t r = 58.00 min (NAG - I3:22:0(2,10,10) in Figure 3.1) that yielded mass spectra showing [M+HCOO] - at m/z 778.42 (Figure 3.4A) and [M+H] + at m/z 734.43 (Figure 3.4B) stood out because its even ion mas ses are consistent with a nitrogen - containing acylated myo - inositol glycoside of EF C 36 H 63 NO 14 . Unlike acylinositol monosaccharide mass spectra, glycosylated inositol forms showed abundant fragment io ns in negative - ion CID mass spectra. Negative mode MS/MS product ion spectra of m/z 778.42 (Figure 3.4A) show abundant neutral mass losses from [M - H] - ( m/z 732.42) such as loss of C10 ketene (loss of 154 Da, m/z 578.27) and a product ion resulting from an additional loss of C2 ketene ( m/z 536.27). In addition, losses of 18.01 Da consistent with H 2 O were regularly observed, and are attributed to charge - remote losses of carboxylic acids rather than ketenes. Losses of two C10 and one C2 ketene groups leaves th e product ion at m/z 382.13, corresponding to the deprot onated disaccharide core with EF C 14 H 24 NO 11 - . Positive - ion mode MS/MS product ion spectra of [M+NH 4 ] + at m/z 751.46 (Figure 3.4B) showed cleavage of the glyosidic linkage, producing abundant product i on m/z 204.09, consistent with an N - acetylglucosamine (N AG) group (C 8 H 14 NO 5 + ). A less abundant fragment ion at m/z 513.34 (C 28 H 49 O 8 + ) suggests substitutions of two C10 and one C2 ester groups on the myo - inositol ring and no esters on the NAG ring. However, the presence of a NAG glycoside required further confir mation by NMR analysis. 259 Figure 3. 4 . MS/MS product ion spectra of [M+HCOO] - and [M+NH 4 ] + of glycosylated acylinositols in an extract of S. quitoense using a linear 5 - 60 V collision energy ramp with 0.5 s acquisition time. (A) ES I( - ) MS/MS product ion spectrum for m/z 778 ([M+HCOO] - ) for acylsugar NAG - I3:22:0(2,10,10). (B) ESI(+) MS/MS product ion spectrum for m/z 751 ([M+NH 4 ] + ) for acylsugar NAG - I3:2 2:0(2,10,10). (C) ESI( - ) MS/MS product ion spectrum for m/z 737 ([M+HCOO] - ) for a cylsugar G - I3:22:0(2,10,10). (D) ESI(+) MS/MS product ion spectrum for m/z 710 ([M+NH 4 ] + ) for acylsugar G - I3:22:0(2,10,10). (E) ESI( - ) MS/MS product ion spectrum for m/z 707 ( [M+HCOO] - ) for acylsugar X - I3:22:0(2,10,10). (F) ESI(+) MS/MS product ion spectru m for m/z 680 ([M+NH 4 ] + ) for acylsugar X - I3:22:0(2,10,10). 260 1D and 2D NMR spectroscopic techniques served as the basis for structure elucidation of glycosides of acylinositols (disaccharide chemical structures and numbering are located in Figure 3.5). An an omeric doublet at H - observed at 4.3 - 4.5 ppm served as th e starting point. The orientation of glycosides was confirmed by 1 H - 1 H couplings (d, J 1 ,2 = 7.5 - 8.3 Hz) and 1 H - 13 C couplings by coupled - HSQC or from HMBC breakthrough signal ( 1 J CH = 161 - 163 Hz ) [23] . In all instances, HMBC spectra showed three bond coupling from the anomeric H - to carbon resonances of myo - inositol at position 4 , which ranged from 81.5 - 83.2 ppm (ring carbon chemical shifts are summarized in Appendix Table 3.15). Figure 3. 5 . Chemical structures of core glycosylated myo - inositol metabolites from S. quitoense determined by NMR spectroscopy. R 1 , R 2 , R 3 = acylation at that position. R 4 = H or acylation for myo - inositol monosaccharide. Acylations are liste d in Table 3.1. For acylsugar t r = 58.00 min (NAG - I3:22:0(2,10,10) in Figure 3.1), the axial H - signal observed at 4.50 ppm (d, J = 8.4 Hz) exhibited coupling to axial H - at 3.49 (m) by COSY. HMBC at H - showed three bond correlation to an a cetamido carbonyl carbon signal at 171.55 ppm. Similarly, a - CH 3 singlet of acetamide centered at 1.82 ppm showed two bond correlations to the carbonyl carbon. Characteristic proton resonances observed for an acetamido - N - H at 6.39 ppm and carbon resonance C - 2 at 57.5 ppm were also telling of an acetylglucosaminyl group. HMBC at H - showed two bond coupli ng to C - at 102.40 ppm and C - at 75.63 ppm. The remainder of the ring system could be traced by 1 H, COSY and J - resolved from chemical shifts and coupling constants. Axial H - signal observed at 3.44 ppm (t, J = 9.1 Hz) had coupling to H - and axial H - at 3.30 ppm (dd, J = 10.0, 8.1 Hz), H - had coupling to axial H - at 3.33 ppm (ddd, J = 10.2, 5.8, 2.8 Hz), and H - had coupling to diastereotopic methylene hydrogens at H 2 - at 3.78 ppm (dd, J = 11.8, 2.9 Hz) and 3.61 ppm (dd, J = 11.8, 5.8 Hz). T hus, the 4 - O - - N - acetylglucosaminyl acylated myo - inositols core structure was confirmed, and supports the annotati on as NAG - I3:22:0(2,10,10). 261 Assignment of acyl group branching and attachment positions of acylinositols are discussed in Section 3. 3.6. Furth er evidence of this structural assignment is provided in Appendix Table 3.16 and Appendix Figures 3.26 - 3.32. 3. 3.4 .2 Discovery of 4 - O - glucopyranosyl (G) acylated myo - inositols Another glycosylated acylinositol at t r = 58.83 min (G - I3:22:0(2,10,10) in Figur e 3.1), exhibiting [M+HCOO] - at m/z 737.40 and [M+NH 4 ] + at m/z 710.43, was assigned the formula C 34 H 60 O 14 . Negative mode MS/MS product ion spectra showed neutral mass losses from [M - H] - of two C10 and one C2 ketene ester groups leaving m/z 341.11 (Figure 3 .4C), corresponding to EF C 12 H 21 O 11 , consistent with previousl y reported sucrose - based acylsugars. However, initial results would prove misleading. Positive - ion mode MS/MS product ion spectra of [M+NH 4 ] + at m/z 710.43 (Figure 3.4D) showed product ions attr ibuted to cleavage of the glyosidic linkage, producing abundan t product ions m/z 531.35 and 513.34, consistent with EFs C 28 H 51 O 9 + and C 28 H 49 O 8 + . Those ions are consistent with neutral mass losses of an anhydrohexose (162.05 Da, C 6 H 10 O 5 ) and hexose (180.06 Da, C 6 H 12 O 6 ) from [M+H] + , and indicate that all the acyl group s (two C10 and one C2) are on the myo - inositol ring. From the molecular mass and fragment ion spectra, this acylsugar was consistent with an acylsucrose. However, structural elucidation by NMR s pectroscopy showed that this sugar core consisted of a glycosi de of myo - inositol. For t r = 58.83 min (G - I3:22:0(2,10,10) in Figure 3.1), the axial H - signal observed at 4.33 ppm (d, J = 7.9 Hz) had coupling to axial H - at 3.11 (m) by COSY. HMBC at H - showed two bond coupling to C - 1 at 104.69 ppm and C - at 74.36 ppm. The remainder of the ring system could be traced by 1 H, COSY and J - resolved from chemical shifts and coupling constants. Axial H - signal observed at 3.27 ppm (m) had coupling to H - and axial H - at 3.27 ppm (m), H - had coupling to axial H - at 3.33 ppm (ddd, J = 9.8, 5.8, 2.7 Hz), and H - had coupling to diastereotopic me thylene hydrogens at H 2 - at 3.77 ppm (dd, J = 11.9, 2.8 Hz) and 3.61 ppm (dd, J = 11.9, 5.9 Hz). Thus, the 4 - O - - glucopyranosyl (G) acylated myo - inositols core structure was confirmed, and supports the annotation as G - I3:22:0(2,10,10). Assignment of 262 acyl group branching and attachment positions of acylinositols are discussed in Section 3. 3.6. Further evidence of this structural assignment is pr ovided in Appendix Table 3.17 and Appendix Figures 3.33 - 3.39. 3. 3.4.3 Discovery of 4 - O - xylopyranosyl (X) acylated myo - inositols A third type of glycosylated acylinositol, t r = 61.28 min (X - I3:22:0(2,10,10) in Figure 3.1) exhibiting [M+HCOO] - at m/z 707.39 and [M+NH 4 ] + at m/z 680.42 was assigned the formula C 33 H 58 O 13 . Product ion MS/MS spectra were unlike most documen ted sucrose - based acylsugars from the Solanaceae. Negative mode MS/MS product ion spectra showed neutral mass losses of two C10 and one C2 ester groups as ketenes from [M - H] - , this time leaving m/z 311.10 (Figure 3.4E), corresponding to EF C 11 H 19 O 10 - , cons istent with a pentosyl myo - inositol. Positive - ion mode MS/MS product ion spectra of [M+NH 4 ] + at m/z 680.42 (Figure 3.4F) mirrored G - I3: 22:0(2,10,10) (Figure 3.4D), showing cleavage of the glyosidic linkage, again producing abundant fragment ions m/z 531.35 and 513.34, consistent with EFs C 28 H 51 O 9 + and C 28 H 49 O 8 + . Neutral mass losses of ammonia plus either 132.04 and 150.05 Da from [M+H] + w ere consistent with a pentose group (C 5 H 8 O 4 and C 5 H 10 O 5 ) observed during fragmentation. These results again suggest subst itutions of two C10 and one C2 ester groups on the myo - inositol ring. For t r = 61.28 min (X - I3:22:0(2,10,10) in Figure 3.1), the ring s ystem could be traced by 1 H, COSY and J - resolved from chemical shifts and coupling constants. The axial H - signal observed at 4.28 ppm (d, J = 7.5 Hz) had couplings to axial H - at 3.11 (dd, J = 9.1, 7.4 Hz). Axial H - signal observed at 3.27 ppm (t, J = 8.8 Hz) had coupling to H - and axial H - at 3.47 ppm (ddd, J = 10.3, 8.4, 5.4 Hz), H - had coupli ng to diastereotopic methylene ring hydrogens at H 2 - at 3.90 ppm (dd, J = 11.4, 5.3 Hz) and at 3.21 ppm (dd, J = 11. 6, 10.0 Hz). Thus, the 4 - O - - xylopyranosyl (X) acylated myo - inositols core structure was confirmed, and supports the annotation as X - I3:22 :0(2,10,10). Assignment of acyl group branching and attachment positions of acylinositols are discussed in Section 3. 3 .6. Further evidence of this structural assignment is provided in Appendix Table 3.18 and Appendix Figures 3.40 - 3.46. 263 3. 3.5 Deep profiling of acylinositols by LC/MS Deep profiling of S. quitoense acylsugars using UHPLC/MS in negative - and positive - ion mode s revealed 29 different EFs (Appendix Tables 3.5 and 3.6) consistent with acylsugars, and all are annotated as acylated myo - inositols and myo - inositol glycosides ( N - acetylglucosaminyl, glucopyranosyl and xylopyranosyl). Notably, acylinositols were not obse rved in tissues of mature leaflets. The isomeric diversity of S. quitoense acylinositols was comparatively less than other Solanaceae spec ies, with only 36 chromatographic peaks annotated using both ionization modes. Most acylinositols and their glycosides were detected in both ionization modes and their molecular masses were consistent with combinations of two medium - length (C10 or C12) est er groups, and varying numbers of C2 ester groups. To our knowledge, di - substituted medium - chain acyl groups ( > 8 carbo ns) were not previously detected during profiling of Solanaceae species [2, 4] . The acylinositol peak area with structural identification (nine structures in Table 3.1) was 98.2% by negative - ion mode and 87.6% by positive - ion mode (Appendix Tables 3.5 and 3.6). Positive - ion mode mass spectra suggest that glycosylated myo - inositols were only esterifie d on the inositol moiety and were compris ed of tri - and tetra - esters. Acylated myo - inositol monosaccharide EFs were consistent with di - esters through fully esterified hexa - esters. For example, an annotation at t r = 83.68 with m/z 674.41 had [M+NH 4 ] + consistent with EF C 34 H 56 O 12 a hexa - ester of my o - inositol consistent with I6:28:0(2,2,2,2,10,10) (Appendix Table 3.6). However, evidence for penta - and hexa - acylated myo - inositol monosaccharides was provided only in positive - ion mode ( likely because positive adducts such as ammonium have affinity for e ster groups). We interpret the lack of detection of penta - and hexa - acylated myo - inositol monosaccharides in negative - ion mode to suggest that at least two non - esterified hydroxyl position s are necessary for formation of [M+HCOO] - adduct anions, presumably because anions have low affinity for ester groups relative to hydroxyl groups [18] . 264 3. 3.6 1D and 2D NMR of purified acylinositols reveals acylation positions Using only LC/MS, S. quitoen se myo - inositol and glucopyranosyl myo - inositol acylsugars resembled previously discovered glucose and sucrose derivatives, in that they had molecular masses consistent with those acylsugars, while N - acetylglucosaminyl and xylopyranosyl myo - inositol acylsu gars had ambiguous core structures that were inconsistent with molec ular masses of glucose or sucrose. Because the long - term goal of this work is to understand the diversity of acylsugars from the Solanaceae and their biosynthesis, assignment of the carboh ydrates as well as the acyl group branching and attachment positions assists the discovery of acylsugar acyltransferase enzymes responsible for their biosynthesis. In view of this, a total of nine abundant acylinositols were purified by semi - preparative HP LC for structural elucidation by NMR spectroscopy, including four ac ylated myo - inositols, and two 4 - O - - N - acetylglucosaminyl, two 4 - O - - glucopyranosyl and one 4 - O - - xylopyranosyl acylated myo - inositol glycosides (Figure 3.5). HMBC was used to measure atom ic connectivity, usually ranging 2 - 4 bonds, and was vital for the assignment of acyl positions on the myo - myo - inositol ylations. For instance, the H - 2 position of I:24:0(2,2,10,10) showed three bond correlation to a carbo nyl carbon resonance at 174.11 ppm. Similarly, 2 - - CH 2 and 2 - - CH 2 of this acyl group centered at 2.38 and 1.62 ppm showed two and three bond correlations to the carbonyl carbon (resolved acyl hydrogens were annotated with a number prefix to indicate their positions in Figure 3.6A). From the combination of mass spectra a nd absence of - CH acyl aliphatic units by HSQC we can deduce that the aliphatic ester group at position 2 is a normal 3 by HMBC correlations displayed - CH 2 and - CH 2 signals cent ered at 2.18 and 1.48 ppm, overlapping - CH 2 (polymethylene) signals from both nC10 acyl groups in the range 1.2 - 1.4 ppm, and - CH 3 C2 acyl groups displayed - CH 3 singlet signals at ~2.0 ppm and were as signed to positio ns 1 and 4 by HMBC correlations consistent with those ring hydrogen positions. The structure of I:24:0(2,2,10,10) is summarized in Table 3.1. A similar approach was used for identifying ester branching and acylation position 265 assignments fo r each acylinosit ol identified in this study. However, four acylinositols identified nC10 and nC12 ester groups at positions 2 and 3 of myo - inositol that were not differentiated with this approach. Therefore, an alternative 1D - TOCSY NMR method was used to characterize thos e structures. Positional assignment of acylinositols esterified with nC10 and nC12 ester groups by NMR presents a daunting analytical challenge, as the spectra of these groups are very similar and 1 H resonances overlap. Figure 3.6A - B compa res the 1 H NMR sp ectra of I4:24:0(2,2,10,10) and I4:26:0(2,2,10,12) (Figure 3.1, t r = 80.27 min) acylinositols, with the only appreciable difference being the shape of a non - resolved polymethylene peak in the range of 1.2 - 1.4 ppm (due to a mixture of unres olved resonances from both the nC10 and nC12 ester groups). To circumvent this issue, a series of 1D - TOCSY transfer experiments were employed. Sachleben et. al. (2014) showed that a 1D - TOCSY technique could be used to identify aliphatic chain length by exc itation and trans fer magnetization from a resolved methylene resonance, and that the time - dependence of this transfer to another resolved site, such as a terminal methyl group, depends on the number of aliphatic carbons separating the two sites [24] . For t his study, the resolved methylene hydrogen at positions 2 - - CH 2 (1.62 ppm) and 3 - - CH 2 (1.48 ppm) of acyl groups of I4:24:0(2,2,10,10) and I4:26:0(2,2,10,12) acylinositols (Figure 3.6A - B) were selectively excited at their respective methylene resona nces (excitation bandwidth of ~53 Hz) and 1 H NMR collected over a series of mixing times (0 ms, then every 20 ms for 100 - 300 ms). As the mixing time was increased, the transfer of magnetization to the terminal - CH 3 unit with signal at ~0.9 ppm showed incre asing intensity (Appendix Figures 3.92 - 3.95). The - CH 3 signal (S) at each mixing time was integrated and normalized to the integral of the excitation peak at 2 - - CH 2 or 3 - - CH 2 s of TOCSY transfer experiments (Appendix Tables 3.25 - 3.28). Figure 3.6C - D illustrates the 1D - TOCSY transfer curves for each of the four experiments. Because I4 :24:0(2,2,10,10) must have two nC10 ester groups (nC10 and nC12 assigned from HSQC results), it was used for comparison to I4:26:0(2,2,10,12). Figure 3.6C shows that the nC10 esters at positions 2 and 3 of have similar transfer curves. In contrast, acylino sitol I4:26:0(2,2,10,12) in Figure 3.6D shows that the amount of magnetization transferred is le ss 266 for the acylation at position 2 as the mixing time increases. Thus, the nC12 ester groups was assigned to position 2 . Notably, we interpret the differences i - to - noise when measuring different sample quantities (48.6 µM ver sus 2.5 µM). The remaining nC10 and nC12 containing acylinositols and acylinositol glycosides were assigned by inference from this group of measurements. Table 3.1 summarizes the structures of NMR resolved acylinositols purified from S. quitoense extracts (further evidence of these structural assignments is provided in Appendix Tables 3.14 - 3.28 and Appendix Figures 3.26 - 3.95). 267 Figure 3. 6 . 1 H NMR spectra of (A) I4:24:0(2,2,10,10) (highlighted regions are - CH 2 positions that were selectively excited for 1D - TOCSY transfer experiments, those spectra are located in Appendix Figures 3.92 - 3.95) and (B) I4:26:0(2,2,10,12) acylinositols. 1D - TOCSY transfer curves for ex citation of 2 - - CH 2 and 3 - - CH 2 acyl groups o f (C) I4:24:0(2,2,10,10) and (D) I4:26:0(2,2,10,12) acylinositols. 268 Table 3. 1 . Summary of NMR resolved acylinositols purified from S. quitoense extracts and percent peak area of ions by negative ion mode. Acylsugar ID Retention tim e (min) Experimental m/z a Adducts Detected b Analyte Molecular Formula Mass Error (ppm) R 1 R 2 R 3 R 4 Sample Peak Area (×10 4 ) % of Total Acylsugar Peak Area c Acyl myo - inositols I3:22:0(2,10,10) 68.18 575.3435 M+Cl, M+HCOO , M+NO 3 C 28 H 50 O 9 - 0.82 C2 nC10 nC10 H 4.42 6.3% I4:24:0(2,2,10,10) 73.90 617.3539 M+Cl, M+HCOO , M+NO 3 C 30 H 52 O 10 - 0.62 C2 nC10 nC10 C2 52.7 75.1% I3:24:0(2,10,12) 75.14 603.3745 M+Cl, M+HCOO C 30 H 54 O 9 0.71 C2 nC12 nC10 H 1.07 1.5% I4:26:0(2,2,10,12) 80.27 645.3857 M+Cl, M+HCOO , M+NO 3 C 32 H 56 O 10 0.29 C2 nC12 nC10 C2 2.95 4.2% Acyl ß - N - acetylglucosaminyl myo - inositols NAG - I3:22:0(2,10,10) 58.00 778.4241 M - H, M+Cl, M+HCOO , M+NO 3 C 36 H 63 NO 14 0.20 C2 nC10 nC10 NAG 2.84 4.1% NAG - I3:24:0(2,10,12) 64.28 806.4550 M - H, M+Cl, M+HCOO C 38 H 67 NO 14 0.56 C2 nC12 nC10 NAG 0.69 1.0% Acyl ß - glucopyranosyl myo - inositol G - I3:22:0(2,10,10) 58.83 737.3965 M - H, M+Cl, M+HCOO , M+NO 3 C 34 H 60 O 14 - 1.07 C2 nC10 nC10 G 2.77 3.9% G - I3:24:0(2,10,12) 65.37 765.4278 M - H, M+HCOO C 36 H 64 O 14 - 0.17 C2 nC12 nC10 G 0.78 1.1% ß - xyl opyranosyl myo - inositol X - I3:22:0(2,10,10) 61.28 707.3864 M - H, M+Cl, M+HCOO , M+NO 3 C 33 H 58 O 13 0.69 C2 nC10 nC10 X 0.68 1.0% % Acylinositol Peak Area with Structural Identification = 98.2% a. Monoisotopic m/z value with greatest ion abundance. b. Bolded a dduct ion with greatest ion abundance. c. Sample peak area values were calculated by combining integrations of ions detected 269 3. 3.3 Structural diversity of acylinositols from S. quitoense Where S. quitoense was comparatively lacking in total number of acyl sugars, it showed unexpected structural diversity regarding sugar core structures. The presence of acylated myo - inositols and their glycosides widens our understanding of acylsugar chemical diversity. To our knowledge, the only prior r eport of acylated myo - inositols within the family Solanaceae was in Solanum lanceolatum [25] , which showed acylated glucopyranosyl and xylopyranosyl myo - inositols, named lanceolitols. In contrast to S. quitoense, reported lanceolitols were glycosylated at position 1 (or 3 depe nding on the nomenclature system that is used) of the myo - inositol ring and had only one acylation at position 2 (even numbered C12 to C20 branched or straight chain aliphatic ester groups). S. quitoense acylinositols showed limited st ructural variability in acylation patterns compared to other acylsugar analogues [2 - 4] . Purified acylinositols were esterified at position 4 by C2 groups or glycosylated by N - acetylglucosaminyl, glucopyranosyl or xylopyranosyl groups. Only C2 esters were a ttached at position 1 , whereas nC10 or nC12 esters were observed at position 2 and nC10 at position 3. 3. 4 Conclusions S. quitoense had fewer abundant acylated SMs, but showed unexpected structural diversity with regard to carbohydrate chemistries. Previo us reports have mostl y identified glucose and acylsucrose, whereas a group of acylated myo - inositols and myo - inositol glycosides ( N - acetylglucosaminyl, glucopyranosyl and xylopyranosyl) were discovered in S. quitoense (9 structures). Acylsugars containing two or more long - chai n acyl groups (>8 carbons) were not previously detected during profiling of Solanaceae species. Likewise, nitrogen - containing acylsugars are telling of the surprising chemical diversity observed within this species. Using only LC/MS - ba sed techniques, many S. quitoense acylinositols resemble previously discovered sucrose and glucose derivatives, while other acylsugars had ambiguous core structures that were inconsistent with molecular masses of glucose and sucrose. This result demonstrat es the important marr iage between mass spectrometry and NMR spectroscopy for structural characterization of unknown SMs, which is particularly important for distinguishing isomeric acylsugars. 270 Investigations of BAHD acylsugar acyltransferases have shown t hat glandular trichom e - specific enzymes catalyze acylation across a host of available acyl - CoAs and sugar substrates. It is unlikely that S. quitoense does not produce sucrose. Rather, one hypothesis suggests that sucrose is not readily available in the sa me cellular or subcel lular compartments as Acyl - CoAs, inositols or glycoinositols, and acyltransferase enzymes. Therefore, acyltransferase enzymes have evolved to synthesize acylsugars via other methods. For acyltransferases, small changes in amino acid se quences can alter acy l acceptor substrate preferences [1, 6, 8] . A more likely scenario is that acyltransferase enzymes in S. quitoense have different substrate specificities and cannot acylate sucrose. The limited number of acylinositols from S. quitoense and the apparent lim ited promiscuity of acylation positions suggest that S. quitoense acyltransferase enzymes are more selective for acyl group acceptor substrates. Exploration of candidate biosynthetic genes operating within S. quitoense will be needed and are under way to a scertain this hypothesis. 271 APPENDIX 272 Figure 3. 7 . S. quitoense images. (A) Picture at ~5 weeks (B) Young leaflet. (C) Type I - like trichomes [26, 27] on petiole of a young leaflet (approximately 3 - 6 mm in length). 273 Table 3. 2 . Plant cultivation and metabolite extraction metadata Species Solanum quitoense Genotype NYBG Organ Aerial tissues Organ specification Aerial tissues included leaf and stem tissues Cell type Glandular trichomes Growth location Growt h chamber at MSU Growth support Seeds were germinated using MilliQ deionized water on Whatman #1 filter paper placed in petri dishes. One - week old seedlings were t ransferred to Jiffy peat pellets. Light 300 µE m - 2 sec - 1 ; 17 h light/7 h dark Humidity 50 % relative humidity Temperature Watering regime Bottom watering as per requirement Nutritional regime solution, once a week Date(s) of plant establishment 5/28/2015 Plant growth stage 10 weeks post germination (plants about 0.25 m height), plants extracted on 8/6/2015 Metabolism quenching method Aerial tissues were extracted in 1.9 L of acetonitrile: isopropanol (AcN:IPA, v/v, 1:1) for 10 mins in two 1 - L Erlenmeyer flasks with horizontal mixing at 120 rpm Harvest method Plants were cut at the stems; larger leaflets were cut at nodes to reduce bulk volume. Sample storage The extract was decanted into two 1 L glass Wheaton bottles with Teflon lined caps and stored in a freezer at - 20°C. ~1 L of solution was concentrated to dryness via ro tary evaporation under vacuum in a 1 L round bottom flask (4 × 250 mL), leaving a dark green residue. The residue was dissolved in 5 mL of AcN:IPA with sonication, transferred to 15 mL polypropylene centrifuge tubes, and centrifuged by Eppendorf Centrifuge 5480R at 10000 × g for 10 mins, the supernatant was transferred to 1 mL autosampler vials and stored at - 20°C prior to purification by semi - preparative HPLC. 274 Table 3. 3 . UHPLC/MS metadata Facility Director Dr. A. Daniel Jones A nalyst Steven M. Hurney Instrument Location A. Daniel Jones Laboratory Facility Instrument Title LCMS: Jones Lab G2 - XS QTof LC System Shimadzu LC - 20AD pumps and Shimadzu CTO - 20A column oven. Manufacturers Shimadzu / Waters Autosampler Shimadzu SIL - 500 0 autosampler Column Ascentis Express C18 Analytical HPLC, 10 cm x 2.1mm x 2.7 µm Column Manufacturer Supelco Catalogue Number 53823 - U Serial Number USRB004698 Packing Lot Number S17138 Injection Volume 10 µL (5 µL sample loop) Flow Rate 0.3 mL/min Mobile Phases: A 10 mM ammonium formate in water (pH 2.8, adjusted with formic acid) B Acetonitrile Gradient Profile Hold 1% B at 0 - 1 min, linear 1 - 100% B at 1 - 100 min, hold 100% B at 100 - 105 min, linear 100 - 1% B at 105 - 106 min, and hold 1% B at 106 - 11 0 min Column Oven Temperature 50 ºC Sample Temperature in autosampler Room Temperature Inlet Method Name SMH_110min_C18_1to100_linear_1to100mins_AcN_Pump C_50C Mass Spectrometer Xevo G2 - XS QTof Manufacturer Waters Software MassLynx v4.1 Ionization S ource Electrospray Ionization (ESI) Data Acquisition Sensitivity Mode Polarity Positive, Negative Mass Range m/z 50 - 1500 Data Format Continuum Capillary Voltage 3.06 kV (ESI+), 2.5 kV (ESI - ) Sample Cone 36 V (ESI+), 40 V (ESI - ) Source Temperature 10 0 °C Source Offset 80 V Desolvation Temperature 300 °C Cone Gas Flow 50 L/hr 275 Table 3.3 (continued) Desolvation Gas Flow 600 L/hr Collision Potential Function 1 6.0 V Function 2 10 to 60 V Scan Duration 0.20 s Inter Scan Delay 0.014 s Collision Ce ll Pressure 0.085 mbar (ESI+), 0.083 mbar (ESI - ) Lock Spray Leu - enkephalin Lock mass ( m/z ) 556.2771 (ESI+), 554.2615 (ESI - ) Lock Spray Scan Time 0.2 s Lock Spray Scan Frequency 10 s MS Method Files SMH_MSe_110 min_10to60V_ST02s_ESI+, SMH_MSe_110 m in_10to60V_ST02s_ESI - Sample handling Aerial tissues from 20 plants aged 10 weeks were harvested (plants were cut at the stems and stem junctions) and extracted in 1.9 L of AcN:IPA for 10 mins in two 1 L Erlenmeyer flasks with horizontal mixing at 120 rpm . For metabolite profiling, the bulk extract solution was diluted to 1/10th concentration, by adding 100 µL to 800 µL of AcN:IPA (containing 0.1 µM telmisartan internal standard) an d adding 100 µL of 0.046 M formic acid (pH ~4.0, inhibits of acylinositol r earrangement) Sample Storage - 20 °C in Wheaton glass vessel with PTFE lined cap Protocol when analyzing the samples The instrument was calibrated in ESI+/ - modes using 500 µM sodi um formate solution. First, the column was equilibrated by analyzing two bl anks using a 22 min gradient. A blank and S. quitoense samples were analyzed in ESI - mode, followed by blank and sample in ESI+ mode. 276 Table 3. 4 . Progenesis QI metadata Peak Picking Parameters: Sensitivity Automatic, Value = 5 Ignore Ions after 90 mins Adducts ESI(+): M+H - H 2 O, M+H, M+NH 4 , M+Na ESI ( - ): M - H, M+Cl, M+HCOO, M+NO 3 Elemental Composition Calculation Parameters: Composition Hydrogen (10 - 100), Carbon (6 - 100), Nitrogen (0 - 1), Oxygen (7 - 30) Precursor tolerance 20 pp m Isotope similarity 80% Tag filters (used to exclude markers): Chromatographic Peak Width < 0.05 min Sample > Blank Integrals of sample peak must be > 3x blank m/z range ESI(+): 240 < m/z < 850 ESI( - ): 267 < m/z < 877 No Elemental Composition Peak s that could not be assigned a molecular formula using the designated elemental composition calculation parameters were excluded Assessment of acylinositol peak annotations To further assess peak annotations, the m/z values of acylinositol homologues were calculated. For example, myo - inositol (C 6 H 12 O 6 , 180.0634 Da) with C2 acyl group (C 8 H 14 O 7 , 222.0740 Da) would have [M+NH 4 ] + = 240.1078 and [M+HCOO] - = 267.0722. The homologues of the following acylinositols and their adducts were considered: 1) myo - inosito l with C2 (as indicated); 2) N - acetylglucosaminyl myo - inositol with C2 (C 16 H 27 NO 12 , [M+H] + = 426.1606 and [M+HCOO] - = 425.1533); 3) glucopyranosyl myo - inositol with C2 (C 14 H 24 O 12 , [M+NH 4 ] + = 402.1606 and [M+HCOO] - = 429.1250); and 4) xylopyranosyl myo - inos itol with C2 (C 13 H 22 O 11 , [M+NH 4 ] + = 372.1500 and [M+HCOO] - = 399.1144). Annotations that could not be assigned as homologues with m/z values of +14.0456 Da (consistent with - CH 2 - saturated acyl groups) were excluded. Annotations not consistent with homolog ues of multiply esterified acylinositols were also exclude d. For example, diacylated myo - inositol with two C2 groups would have formula C 10 H 16 O 8 and [M+NH 4 ] + = 282.1183 and [M+HCOO] - = 309.0827). Permutations of m/z values for all acylinositol and acylinos itol glycosides with one to fully esterified hydroxyl grou ps were considered. Those that did not have m/z values consistent with these permutations were excluded. The results are shown in Appendix Tables 3.5 and 3.6. 277 Table 3. 5 . S. quitoense acylinositol deep profiling results generated by ESI - mode using Waters Progenesis QI software. Peak # a Metabolite Annotation b Retention time (min) m/z c Adducts Detected d Formula Mass Error (ppm) Isotope Similarity (%) Sample Peak Area e Relativ e Abundance 1 I2:12:0 35.06 421.2070 M+HCOO C 18 H 32 O 8 - 2.47 93.74 353 0.05% 2 G - I3:16:0 39.37 653.3015 M+HCOO C 28 H 48 O 14 - 1.85 90.48 134 0.02% 3 G - I4:18:0 41.72 695.3129 M+HCOO C 30 H 50 O 15 - 0.39 90.11 231 0.03% 5* NAG - I3:22:0(2,10,10) 58.00 778.4241 M - H, M +Cl, M+HCOO , M+NO 3 C 36 H 63 NO 14 0.20 99.16 28418 4.05% 6* G - I3:22:0(2,10,10) 58.83 737.3965 M - H, M+Cl, M+HCOO , M+NO 3 C 34 H 60 O 14 - 1.07 99.13 27655 3.94% 8* X - I3:22:0(2,10,10) 61.28 707.3864 M - H, M+Cl, M+HCOO , M+NO 3 C 33 H 58 O 13 0.69 97.99 6823 0.97% 11* NAG - I3 :24:0(2,10,12) 64.28 806.4550 M - H, M+Cl, M+HCOO C 38 H 67 NO 14 0.56 98.01 6909 0.99% 12* G - I3:24:0(2,10,12) 65.37 765.4278 M - H, M+HCOO C 36 H 64 O 14 - 0.17 98.86 7774 1.11% 13* X - I4:24:0 66.78 749.3986 M+HCOO C 35 H 60 O 14 3.01 89.59 171 0.02% 14* I4:22:0 67.13 589. 3220 M+HCOO C 28 H 48 O 10 - 1.69 94.79 1510 0.22% 15 I3:22:0 67.25 575.3429 M+HCOO C 28 H 50 O 9 - 1.41 98.21 2141 0.31% 16* X - I3:24:0 68.07 735.4174 M+HCOO C 35 H 62 O 13 0.27 96.13 1801 0.26% 17* I3:22:0(2,10,10) 68.18 575.3435 M+Cl, M+HCOO , M+NO 3 C 28 H 50 O 9 - 0.82 98.6 2 44175 6.30% 18* I4:23:0 70.58 603.3375 M+HCOO C 29 H 50 O 10 - 1.93 92.22 426 0.06% 19* NAG - I3:26:0 70.89 834.4857 M+HCOO C 40 H 71 NO 14 0.03 89.12 491 0.07% 20* I4:24:0 72.00 617.3534 M+HCOO C 30 H 52 O 10 - 1.42 92.50 636 0.09% 21* G - I3:26:0 72.18 793.4592 M+HCOO C 38 H 68 O 14 0.14 89.87 632 0.09% 22 I4:24:0 73.23 617.3538 M+HCOO C 30 H 52 O 10 - 0.78 88.29 192 0.03% 24* I4:24:0(2,2,10,10) 73.90 617.3539 M+Cl, M+HCOO , M+NO 3 C 30 H 52 O 10 - 0.62 97.25 526844 75.1% 26 X - I3:26:0 74.94 763.4491 M+HCOO C 37 H 66 O 13 0.74 82.12 56 0.01% 27* I3:24:0(2,10,12) 75.14 603.3745 M+Cl, M+HCOO C 30 H 54 O 9 0.71 98.69 10658 1.52% 28* I4:25:0 77.16 631.3700 M+HCOO C 31 H 54 O 10 0.25 90.95 402 0.06% 31* I5:26:0 79.56 659.3603 M+HCOO C 32 H 54 O 11 - 7.35 92.00 160 0.02% 32* I4:26:0(2,2,10,12) 80.27 645.3857 M +Cl, M+HCOO , M+NO 3 C 32 H 56 O 10 0.29 98.52 29461 4.20% 278 Table 3.5. (continued) Peak # a Metabolite Annotation b Retention time (min) m/z c Adducts Detected d Formula Mass Error (ppm) Isotope Similarity (%) Sample Peak Area e Relative Abundance 34 I3:26:0 81.73 631.4050 M+HCOO C 32 H 58 O 9 - 2.26 93.38 1031 0.15% 36* I4:28:0 86.13 673.4126 M+HCOO C 34 H 60 O 10 - 6.72 92.28 2105 0.30% a. Peak number indicates elution order of S. quitoense acylinositol annotation. Peak numbers with asterisks (*) were annotated in both ESI - and ESI+ ion modes. Acronyms: myo - inositol ( I ), glucopyranosyl ( G ) , N - acetylglucosaminyl ( NAG ) and xylopyranosyl ( X ). b. Metabolite annotations in bold were identified by NMR analysis. c. Monoisotopic m/z value with greatest ion abundance. d. Bolded adduc t ion had greatest ion abundance. e. Sample peak area values were calculated by combining ions detected. 279 Table 3. 6 . S. quitoense acylinositol deep profiling results generated by ESI+ mode using Waters Progenesis QI software. Peak # a Metabolite Annotation b Retention time (min) m/z c Adducts Detected d Formula Mass Error (ppm) Isotope Similarity (%) Sample Peak Area e Relative Abundance 4 I5:22:0 54.92 576.3355 M+NH 4 , M+Na C 28 H 46 O 11 - 4.10 94.05 966 0.09% 5* NAG - I3:22:0(2,10,10) 57.98 734.4296 M+H , M+Na C 36 H 63 NO 14 - 3.48 98.44 17785 1.73% 6* G - I3:22:0(2,10,10) 58.82 710.4291 M+NH 4 , M+Na C 34 H 60 O 14 - 4.33 99.47 21817 2.13% 7 I5:20:0 59.49 548.3062 M+H, M+NH 4 , M+Na C 26 H 42 O 11 2.65 93.16 2976 0.29% 8* X - I3:22:0(2,10,10) 61.25 680.4176 M+NH 4 , M+Na C 33 H 58 O 13 - 5.95 97.78 5908 0.58% 9 X - I3:22:0 61.98 680.4191 M+NH 4 C 33 H 58 O 13 - 3.65 91.59 228 0.02% 10 I5:20:0 62.23 562.3216 M+NH 4 C 27 H 44 O 11 - 1.17 92.10 215 0.02% 11* NAG - I3:24:0(2,10,12) 64.27 762.4595 M+H C 38 H 67 NO 14 - 5.22 97.05 4279 0.42% 12* G - I3:24:0(2,10,12) 65.37 738.4597 M+NH 4 , M+Na C 36 H 64 O 14 - 5.20 97.53 6441 0.63% 13* X - I4:24:0 66.76 722.4296 M+NH 4 C 35 H 60 O 14 - 3.62 88.63 223 0.02% 14* I4:22:0 67.14 562.3550 M+NH 4 , M+Na C 28 H 48 O 10 - 6.54 91.99 3471 0.34% 16* X - I3:24:0 68.08 708.4477 M+NH 4 C 35 H 62 O 13 - 7.52 95.98 1823 0.18% 17* I3:22:0(2,10,10) 68.16 548.3777 M+H - H 2 O, M+H, M+NH 4 , M+Na C 28 H 50 O 9 - 7.45 97.64 43496 4.24% 18* I4:23:0 70.58 576.3701 M+NH 4 , M+Na C 29 H 50 O 10 - 7.46 88.66 1272 0.12% 19* NAG - I3:26:0 70.85 790.4894 M+H C 40 H 71 NO 14 - 6.69 94 .31 425 0.04% 20* I4:24:0 71.96 595.3428 M+H - H 2 O, M+NH 4 , M+Na C 30 H 52 O 10 - 12.63 88.37 6854 0.67% 21* G - I3:26:0 72.11 766.4872 M+NH 4 C 38 H 68 O 14 - 10.06 92.91 484 0.05% 23 I4:25:0 73.85 604.3987 M+NH 4 C 31 H 54 O 10 - 11.70 82.08 2692 0.26% 24* I4:24:0(2,2,10,10) 73.89 590.3892 M+H - H 2 O, M+H, M+NH 4 , M+Na C 30 H 52 O 10 - 2.92 95.16 799389 77.9% 25 I4:26:0 73.89 618.4181 M+NH 4 C 32 H 56 O 10 - 5.09 95.75 12437 1.21% 27* I3:24:0(2,10,12) 75.13 576.4077 M+H - H2O, M+NH 4 , M+Na C 30 H 54 O 9 - 12.80 91.61 11086 1.08% 28* I4:25:0 77.08 6 04.3985 M+NH 4 C 31 H 54 O 10 - 11.94 93.16 818 0.08% 29 I5:26:0 78.24 632.3978 M+H - H 2 O, M+H, M+NH 4 , M+Na C 32 H 54 O 11 - 4.32 94.93 18652 1.82% 30 I5:28:0 78.26 660.4273 M+NH 4 C 34 H 58 O 11 - 6.91 80.54 930 0.09% 31* I5:26:0 79.51 632.3971 M+H - H 2 O, M+NH 4 , M+Na C 32 H 54 O 1 1 - 5.49 84.09 6393 0.62% 280 Table 3.6. (continued) Peak # a Metabolite Annotation b Retention time (min) m/z c Adducts Detected d Formula Mass Error (ppm) Isotope Similarity (%) Sample Peak Area e Relative Abundance 32* I4:26:0(2,2,10,12) 80.26 618.4181 M+H - H 2 O , M+H, M+NH 4 , M+Na C 32 H 56 O 10 - 5.10 96.20 46869 4.57% 33 I4:28:0 80.29 646.4453 M+NH 4 C 34 H 60 O 10 - 11.44 82.95 968 0.09% 35 I6:28:0 83.68 674.4079 M+NH 4 , M+Na C 34 H 56 O 12 - 4.80 90.70 4490 0.44% 36* I4:28:0 86.07 646.4481 M+NH 4 C 34 H 60 O 10 - 7.02 86.52 3192 0.31 % a. Peak number indicates elution order of S. quitoense acylinositol annotation. Peak numbers with asterisks (*) were annotated in both ESI - and ESI+ ion modes. Acronyms: myo - inositol ( I ), glucopyranosyl ( G ) , N - acetylglucosaminyl ( NAG ) and xylopyranosyl ( X ). b. Metabolite annotations in bold were identified by NMR analysis . c. Monoisotopic m/z value with greatest ion abundance. d. Bolded adduct ion had greatest ion abundance . e. Sample peak area values were calculated by combining ions detected. 281 Table 3. 7 . MS/MS precursor masses and time windows Acylinositol ID [M+HCOO] - [M+NH 4 ] + Start (min) Stop (min) Cone (V) NAG - I3:22:0(2,10,10) 778.423 751.459 55.00 57.00 40 G - I3:22:0(2,10,10) 737.397 710.432 57.01 58.50 40 X - I3:22:0(2,10,10 ) 707.386 680.422 58.51 61.00 40 NAG - I3:24:0(2,10,12) 806.454 779.490 61.01 63.25 40 G - I3:24:0(2,10,12) 765.428 738.463 63.26 65.00 40 I3:22:0(2,10,10) 575.344 548.379 65.01 68.00 40 I4:24:0(2,2,10,10) 617.354 590.390 70.00 73.00 40 I3:24:0(2,10,12) 6 03.375 576.411 73.01 75.00 40 I4:26:0(2,2,10,12) 645.386 618.421 78.00 80.00 40 Table 3. 8 . MS/MS precursor masses, time windows and cone voltages Acylinositol ID [M - H] - [M+H] + Start (min) Stop (min) [M - H] - Cone (V) [M+H] + Cone (V ) NAG - I3:22:0(2,10,10) 732.418 734.432 55.00 57.00 120 40 G - I3:22:0(2,10,10) 691.391 693.406 57.01 58.50 120 80 X - I3:22:0(2,10,10) 661.380 663.395 58.51 61.00 120 80 NAG - I3:24:0(2,10,12) 760.449 762.464 61.01 63.25 140 60 G - I3:24:0(2,10,12) 719.422 72 1.437 63.26 65.00 120 80 I3:22:0(2,10,10) 529.338 531.353 65.01 68.00 120 120 I4:24:0(2,2,10,10) 571.349 573.363 70.00 73.00 100 120 I3:24:0(2,10,12) 557.370 559.384 73.01 75.00 100 120 I4:26:0(2,2,10,12) 599.380 601.395 78.00 80.00 100 120 282 Table 3. 9 . LC/MS/MS metadata Facility Director Dr. A. Daniel Jones Analyst Steven M. Hurney Instrument Location A. Daniel Jones Laboratory Facility Instrument Title LCMS: Jones Lab G2 - XS QTof LC System Shimadzu LC - 20AD pumps and Shimadz u CTO - 20A column oven. Manufacturers Shimadzu / Waters Autosampler Shimadzu SIL - 5000 autosampler Column Ascentis Express C18 Analytical HPLC, 10 cm x 2.1mm x 2.7 µm Column Manufacturer Supelco Catalogue Number 53823 - U Serial Number USRB003996 Packin g Lot Number S15013 Injection Volume 10 µL Flow Rate 0.3 mL/min Mobile Phases: A 10 mM ammonium formate in water (pH 2.8, adjusted with formic acid) B Acetonitrile Gradient Profile Hold 1% B at 0 - 1 min, 1 - 100% B at 1 - 100 min, 100% B at 100 - 105 min, 1 00 - 1% B at 105 - 106 min, and 1% B at 106 - 110 min. Column Oven Temperature 50 ºC Sample Temperature in autosampler Room Temperature Inlet Method Name SMH_110min_C18_1to100_linear_1to100mins_Pump C Mass Spectrometer Xevo G2 - XS QTof Manufacturer Waters S oftware MassLynx v4.1 Ionization Source Electrospray Ionization (ESI) Data Acquisition Sensitivity Mode Polarity Positive, Negative Mass Range m/z 50 - 1000 Data Format Centroid Capillary Voltage 3 kV (ESI+), 3 kV (ESI - ) Sample Cone See Appendix Table s 7 and 8 Source Temperature 100 °C Source Offset 80 V Desolvation Temperature 350 °C Cone Gas Flow 50 L/hr (ESI - ), 20 L/hr (ESI+) Desolvation Gas Flow 600 L/hr (ESI - ), 500 L/hr (ESI+) LM Resolution 20 HM Resolution 20 283 Table 3.9. (continued) Col lision Potential Ramp Start Potential 5 V End Potential 60 V Scan Duration 0.5 s Inter Scan Delay 0.014 s Collision Cell Pressure 0.1 mbar MS Method Files SMH_MSMS_5 - 60V_S quitoense_ESI - , SMH_MSMS_5 - 60V_S quitoense_ESI+, SMH_MSMS_5 - 60V_S quitoense _ESI - _M - H, SMH_MSMS_5 - 60V_S quitoense_ESI+_M+H, Sample handling S. quitoense bulk extract solution as described in Appendix Table 2 Sample Storage - 20 °C in Wheaton vessel with PTFE lined cap Protocol when analyzing the samples The instrument was calib rated in ESI+/ - modes using 500 µM sodium formate solution for all analyses. After column equilibration, an S. quitoense bulk extract sample was analyzed by MS/MS using precursor adducts [M+HCOO] - and [M+NH 4 ] + according to Appendix Table 7. An S. quitoense bulk extract sample was analyzed in CID mode with increasing cone potentials (40, 60, 80, 100, 120, and 140 V). Cone v oltages shown in Appendix Table 8 were selected according to maximum [M - H] - and [M+H] + ion formation. MS/MS product ion spectra were gen erated using a collision potential ramp, targeting the respective precursor pseudomolecular ion. 284 Figure 3. 8 . NAG - I3:22:0(2,10,10) MS/MS product ion spectra, precursor ion [M+NH 4 ] + (top) & [M+H] + (bottom) 285 Figure 3. 9 . NAG - I3:22:0(2,10,10) MS/MS product ion spectra, precursor ion [M+HCOO] - (top) & [M - H] - (bottom) 286 Figure 3. 10 . G - I3:22:0(2,10,10) MS/MS product ion spectra, precursor ion [M+NH 4 ] + (top) & [M+H] + (bottom) 287 Figu re 3. 11 . G - I3:22:0(2,10,10) MS/MS product ion spectra, precursor ion [M+HCOO] - (top) & [M - H] - (bottom) 288 Figure 3. 12 . X - I3:22:0(2,10,10) MS/MS product ion spectra, precursor ion [M+NH 4 ] + (top) & [M+H] + (bo ttom) 289 Figure 3. 13 . X - I3:22:0(2,10,10) MS/MS product ion spectra, precursor ion [M+HCOO] - (top) & [M - H] - (bottom) 290 Figure 3. 14 . NAG - I3:24:0(2,10,12) MS/MS product ion spectra, precursor ion [M+NH 4 ] + (to p) & [M+H] + (bottom) 291 Figure 3. 15 . NAG - I3:24:0(2,10,12) MS/MS product ion spectra, precursor ion [M+HCOO] - (top) & [M - H] - (bottom) 292 Figure 3. 16 . G - I3:24:0(2,10,12) MS/MS product ion spectra, precursor i on [M+NH 4 ] + (top) & [M+H] + (bottom) 293 Figure 3. 17 . G - I3:24:0(2,10,12) MS/MS product ion spectra, precursor ion [M+HCOO] - (top) & [M - H] - (bottom) 294 Figure 3. 18 . I3:22:0(2,10,10) MS/MS product ion spectra, precursor ion [M+NH 4 ] + (top) & [M+H] + (bottom) 295 Figure 3. 19 . I3:22:0(2,10,10) MS/MS product ion spectra, precursor ion [M+HCOO] - (top, magnified 10x over m/z 51 - 566) & [M - H] - (bottom) 296 Figure 3. 20 . I4:2 4:0(2,2,10,10) MS/MS product ion spectra, precursor ion [M+NH 4 ] + (top) & [M+H] + (bottom) 297 Figure 3. 21 . I4:24:0(2,2,10,10) MS/MS product ion spectra, precursor ion [M+HCOO] - (top, magnified 60x over m/z 51 - 608) & [M - H] - (bottom) 298 Figure 3. 22 . I3:24:0(2,10,12) MS/MS product ion spectra, precursor ion [M+ NH 4 ] + (top) & [M+H] + (bottom) 299 Figure 3. 23 . I3:24:0(2,10,12) MS/MS product ion spectra, precursor ion [M+HCOO] - (top, magnified 14x over m/z 51 - 594) & [M - H] - (bottom) 300 Figure 3. 24 . I4:26:0(2,2,10,12) MS/MS product ion spectra, precursor ion [M+NH 4 ] + (top) & [M+H] + (bottom) 301 Figure 3. 25 . I4:26:0(2,2,10,12) MS/MS product ion spect ra, precursor ion [M+HCOO] - (top, magnified 100x over m/z 51 - 590) & [M - H] - (bottom) 302 Table 3. 10 . Purification of acylinositols by semi - preparative HPLC Analyst Steven M. Hurney Instrument Location A. Daniel Jones Laboratory Instru ment Title Waters semi - prep HPLC (Jones Lab) LC System Waters 2795 Separations Module equipped with LKB Bromma 2211 Superrac Fraction Collector with automated fraction collection Manufacturers Waters Column Thermo Scientific Acclaim 120 C18 semi - prepara tive HPLC column (4.6 x 150 mm, 5 µm particle size) Column Manufacturer Thermo Scientific Product Number 059148 Serial Number 005992 Packing Lot Number 014 - 25 - 017 Injection Volume 200 µL Flow Rate 1.5 mL/min Mobile Phases: A 0.15% Formic Acid in Wa ter B Acetonitrile C Methanol Gradient Profile (solvents A and B only) Hold 1% B at 0 - 1 min, linear 1 - 56% B at 1 - 2 min, linear 56% - 68% B at 2 - 40 min, linear 68 - 90% B at 40 - 53 min, linear 90 - 100% B at 53 - 54 min (followed by a Column Wash). Column Wash a nd Re - equilibration Linear 100% B to 100% C at 54 to 55 min, hold 100% C at 55 - 65 min, linear 100% C to 100% B at 65 - 66 min, Linear 100% B to 1% B (99% A) at 66 - 67, hold 1% B at 67 - 70 min. Column Oven Temperature 50 ºC Sample Temperature in autosampler R oom Temperature Method Name SP_56C_70min_AcN_MeOH Sample handling Approximately 1 L of the S. quitoense bulk extract was added in portions of ~250 mL to a 1000 mL round bottom flask, and dried via rotary evaporation under vacuum at ~30°C, leaving a green residue. The residue was reconstituted in 5.0 mL AcN:IPA with sonic ation for 5 mins while manually swirling. The solution was centrifuged by Eppendorf Centrifuge 5480R at 10000 × g for 10 mins. The supernatant was then transferred to LC autosampler vials for semi - preparative HPLC purification. Sample Storage - 20 °C in au tosampler vials Protocol when analyzing the samples One minute fractions were collected in Pyrex glass culture tubes (18 × 150 mm) in five batches labeled letter A - E. Each batch consisted of six injections each. Fractions were concentrated to dryness unde r vacuum using a Thermo Savant SPD 131 DDA SpeedVac Concentrator with BOC Edwards XDS Dry Pump. Fractions were labeled by minute, with fractions 15 - 55 containing the most abundant acylinos itols. 303 Table 3.10. (continued) Protocol for testing the fractions In order to test for purity and reproducibility, fractions from each batch were reconstituted in 0.50 mL AcN:IPA, and transferred to LC autosampler vials. 5.0 µL aliquots from each fractio n were diluted in autosampler vials containing 0.50 mL of AcN:IPA and analyzed on an LCT Premier mass spectrometer equipped with Shimadzu LC - 20AD pumps, Shimadzu SIL - 5000 autosampler and a Shimadzu CTO - 20A column oven using a 30 min gradient (0.3 mL min - 1 , Ascentis Express C18 Analytical HPLC, 10cm x 2.1mm x 2.7µ m) optimized for separating S. quitoense acylsugars. The elution program is as follows: hold 1% B at 0 - 1 min, linear 1 - 55% B at 1 - 2 min, linear 55% - 85% B at 2 - 22 min, linear 85 - 100% B at 22 - 22.01 m in, hold 100% B at 22.01 - 27 min, linear 100 - 1% B at 27 - 27. 01 min, hold 1% B at 27.01 - 30 min. HPLC fractions were combined according to metabolite purity and by comparison to a S. quitoense bulk extract solution analyses run before and after each group of f ractions (i.e. the sample order was blank, S. quitoense bu lk extract, ~20 fraction samples in ascending elution order, blank, S. quitoense bulk extract, and etc.). 304 Table 3. 11 . Bruker 900 MHz NMR Instrument Metadata Facility Super visor Dr. Daniel Holmes Analyst Steven M. Hurney Instrument Location MSU Max T. Rogers NMR Facility Facility Instrument Title 900 MHz Bruker Time of acquisition October 2015 - February 2016 Manufacturer Bruker Field Frequency Lock Acetonitrile - d 3 Ad ditional Solute None Solvent 300 µL CD 3 CN Chemical Shift Standard CH 3 CN - d 3 H C = 118.70 ppm) Concentration Standard None Instrument Bruker Avance 900 MHz NMR Geographic location of instrument 42.7164, - 84.4677 Magnet 899.13 - 899.00 MHz Probe Bruker TCI triple - resonance inverse detection cryoprobe Console Bruke r Avance Acquisition Software Topspin 2.1.6 Acquisition Parameters: a) Acquisition parameters file reference 1 H:/opt/topspin216/exp/stan/nmr/lists/pp/zg 13 C:/opt/topspin216/exp/stan/nmr/lists/pp/zgpg30 COSY:opt/topspin216/exp/stan/nmr/lists/pp/co sygpmfph HSQC:/opt/topspin216/exp/stan/nmr/lists/pp/hsqcedetgpsisp2.2 coupledHSQC:/opt/topspin216/exp/stan/nmr/lists/pp/hsqcedetgpsisp2.2nd HMBC: /opt/topspin216/exp/stan/nmr/lists/pp/hmbcgplpndqf J - resolved: /opt/topspin216/exp/stan/nmr/lists/pp/j resqf ROESY: /opt/topspin216/exp/stan/nmr/lists/pp/roesyetgp b) Sample Details Shigemi (5 mm) NMR tube, Temperature @ 298 K, No Spinning c) Instrument operation details (recorded for each sample independently and are roughly the same for all samples m easured; displayed here is the sample containing NAG - I3:22:0(2,10,10) as an example) Radiation frequency (MHz): 1 H: 899.1300264 13 C: 226.0861132 COSY: 899.1300240 (F2), 899.1300049 (F1) HSQC: 899.1300246 (F2), 226.0861186 (F1) HMBC: 899.1300260 ( F2), 226.0860984 (F1) J - resolved: 899.1300262 (F2), 899.1340467 (F1) ROESY: 899.1300266 (F2), 899.1300267 (F1) Acquisition nucleus: 1 H: 90° = 7.25 µs, 13 C: 90° = 20.0 µs 305 Table 3.11. (continued) d) Number of scans (scans, dummy scans) 1 H: 32, 0 13 C: 2000 - 4000, 8 COSY: 4 - 8, 16 HSQC: 8 - 16, 16 Coupled - HSQC: 12 - 16, 16 HMBC: 20 - 40, 32 J - resolved: 16 - 32, 4 ROESY: 8 - 16, 8 e) Number of data points acquired (F2, F1) 1 H: 148144 13 C: 65536 COSY: 2048, 200 HSQC: 1024, 400 Coupled - HSQC: 4 308, 400 HMBC: 4096, 480 J - resolved: 2048, 128 ROESY: 1982, 512 e) Dwell time (µs) 1 H: 27.0 13 C: 9.225 COSY: 46.4 HSQC: 46.4 Coupled - HSQC: 46.4 HMBC: 46.4 J - resolved: 46.4 ROESY: 46.4 FID and spectral processing parameters: Processin g software Topspin 3.5 pl 7 a) Number of data points in spectrum (F2, F1) 1 H: 262144 13 C: 131072 COSY: 4096, 4096 HSQC: 1024, 1024 Coupled - HSQC: 8192, 1024 HMBC: 4096, 2048 J - resolved: 4096, 4096 ROESY: 4096, 2048 306 Table 3.11. (continued) b ) Window function details 1 H: exponential multiply, LB = 0 Hz 13 C: exponential multiply, LB = 2 Hz COSY: QSINE, QSINE; LB = - 0.3, - 0.3 Hz; GB = 0.3, 0; SSB = 2, 2; TM1 = 0, 1; TM2 = 0, 1 HSQC: QSINE, QSINE; SSB = 2, 2; TM1 = 0, 0.1; TM2 = 0, 0.9 C oupled - HSQC: QSINE, QSINE; SSB = 2, 2; TM1 = 0, 0.1; TM2 = 0, 0.9 HMBC: QSINE, SINE J - resolved: SINE, SINE; LB = - 0.3, - 0.3 Hz; GB = 0.3, 0; TM1 = 0, 1; TM2 = 0, 1 ROESY: QSINE, QSINE; LB = 1.0, 0.3 Hz; GB = 0.35, 0.1; SSB = 2, 2; TM1 = 0, 0.1; TM2 = 0, 0.9 307 Table 3. 12 . Agilent DDR2 500 MHz NMR Instrument Metadata Facility Supervisor Dr. Daniel Holmes Analyst Steven M. Hurney Instrument Location MSU Max T. Rogers NMR Facility Facility Instrument Titles Ahriman and Ormuzd Time of acquisition October 2015 Manufacturer Agilent Field Frequency Lock Acetonitrile - d 3 Additional Solute None Solvent CD 3 CN: 500 µL Chemical Shift Standard CH 3 CN - d 3 H C = 118.70 ppm) Concentration Standard None Instrument Agilent DDR2 500 MHz with 7600AS 96 autosamplers Geographic location of instrument 42.7288, - 84.4745 Magnet 499.91 MHz Probe OneNMR Probe with Protune accessory for hands - off tuni ng Acquisition Software VnmrJ 3.2A Acquisition Parameters: a) Acquisition parameters file reference 1 H: VnmrJ/Experiment Selector/Common/PROTON 13 C: VnmrJ /Experiment Selector/Common/CARBON COSY: VnmrJ/Experiment Selector/Common/(HH)gCOSY HSQC: VnmrJ/Experiment Selector/Common/(HC)gHSQCAD HMBC: VnmrJ/Experiment Selector/Common/(HC)gHMBCAD J - resolved: VnmrJ/Experiment Selector/Liquid/JSpectra/ HOMO2DJ TOCSY: VnmrJ/Experiment Selector/Liquid/TOCSY ROESY: VnmrJ/ Experiment Selector/Common/RO ESY b) Sample Details Kontes NMR tube, 8 in, Temperature @ 298 K, No Spinning c) Instrument operation details Radiation frequency (MHz): 1 H: 499.907 13 C: 125.713 COSY: 499.700 (F2), 499.700 (F1) HSQC: 499.701 (F2), 125.661 (F1) HMBC: 499.701 ( F2), 125.662 (F1) J - resolved: 499.906 (F2), 499.906 (F1) TOCSY: 499.906 (F2), 499.906 (F1) ROESY: 499.906 (F2), 499.906 (F1) 1 H: 90° = 7.80 µs, 13 C: 90° = 9.50 µs 308 Table 3.12. (continued) d) Number of scans 1 H: 512 13 C: 4800 COSY: 4 HSQC: 4 HMBC: 8 J - resolved: 8 TOCSY: 8 ROESY: 16 e) Number of data points acquired (F2, F1) 1 H: 16384 13 C: 65536 COSY: 864, 800 HSQC: 1500, 512 HMBC: 1500, 512 J - resolved: 3584, 400 TOCSY: 860, 512 ROESY: 860, 400 e) Acquisition time (s) 1 H: 1.6384 13 C: 1.2321 COSY: 0.15 HSQC: 0.15 HMBC: 0.15 J - resolved: 0.6250 TOCSY: 0.15 ROESY: 0.15 FID and spectral processing parameters: Processing software MestReNova v12.0.0 - 20080 a) Number of data points in spectrum (F2, F1) 1 H: 131072 13 C: 131072 COSY: 1024, 1024 HSQC: 2048, 2048 HMBC: 2048, 2048 J - resolved: 4096, 4096 TOCSY: 1024, 1024 ROESY: 1024, 1024 309 Table 3.12. (continued) b ) Window function details 1 H: none 13 C: exponential, LB = 0.5 Hz COSY: F1 Sine Square II , 0%, 50%; F2 Sine Square II, 0%, 50% HSQC: F1 Gaussian 56.43; F2 Gaussian 7.65 HMBC: F1 Gaussian 67.70; F2 Sine Square II, 0%, 50% J - resolved: F1 Sine Square II, 0%, 50%; F2 Gaussian 1.84, Sine Bell II - 0.3%, 50%; First Point 0.51 TOCSY: F1 Gaussi an 17.99; F2 Gaussian 7.65; First Point 0.50 ROESY: F1 Gaussian 8.24; F2 Gaussian 7.65; First Point 0.50 310 Table 3. 13 . Varian Inova 600 MHz NMR Instrument Metadata Facility Supervisor Dr. Daniel Holmes Analyst Steven M. Hurney Instrument Location MSU Max T. Rogers NMR Facility Facility Instrument Titles Sobek Time of acquisition March 2016 Manufacturer Varian Field Frequency Lock Acetonitrile - d 3 Additional Solute None Solvent CD 3 CN: 300 µL Chemical Shift Standard CH 3 CN - d 3 H C = 118.70 ppm) Concentration Standard None Instrument Varian Inova 600 MHz Geographic location of instrument 42.7288, - 84.4745 Magnet 599.77 MHz Probe Nalorac 5mm PFG switchable probe pretuned for 1H, 13C Acquisition Software CentO S 5.6 with VnmrJ 3.2A Acquisition Parameters: a) Acquisition parameters file reference 1D - TOCSY: VnmrJ/Experiment Selector/Liquid/TOCSY1D b) Sample Details Shigemi (5 mm) NMR tube, Temperature @ 298 K, No Spinning c) Instrument operation details Rad iation frequency (MHz): 1 H: 599.77 1 H: 90° = 16.15 µs Selective band center: 3.89 ppm; width: 61.6 Hz Selective band center: 1.49 ppm; width: 53.6 Hz Selective band center: 1.63 ppm; width: 53.2 Hz d) Number of scans 1 H: 552 e) Number of data points acquired 1 H: 32764 e) Acquisition time (s) 1 H: 1.7074 FID and spectral processing parameters: Processing software MestReNova v12.0.0 - 20080 a) Number of data points in spectrum 1 H: 131072 b ) Window function details 1 H: none 311 Table 3. 14 . Summary of 1 H chemical shifts of inositol core hydrogen atoms. Chemical shifts labeled in bold indicate acyl substitutions are located at t hose positions. All spectra were referenced to non - deuterated solvent signal of acetonitrile - d 3 H = 1.94 ppm). Acylinositol ID R 1 R 2 R 3 R 4 R 5 R 6 R 1' R 2' R 3' R 4' R 5' R 6' NAG - I3:22:0(2,10,10) 4.78 5.41 4.93 3.80 3.44 3.75 4.50 3.49 3.44 3.30 3.33 3.78, 3.61 G - I3:22:0(2,10,10) 4.82 5.47 4.96 3.78 3.44 3.77 4.33 3.11 3.27 3.27 3.33 3.77, 3 .61 X - I3:22:0(2,10,10) 4.80 5.47 4.94 3.79 3.40 3.75 4.28 3.11 3.27 3.47 3.90, 3.21 --- NAG - I3:24:0(2,10,12) 4.78 5.41 4.93 3.80 3.44 3.75 4.50 3.49 3.43 3.29 3.33 3.78, 3.61 G - I3:24:0(2,10,12) 4.82 5.47 4.96 3.78 3.44 3.76 4.33 3.10 3.27 3.27 3.32 3.77 , 3.59 I3:22:0(2,10,10) 4.76 5.44 4.78 3.68 or 3.69 3.30 3.68 or 3.69 --- --- --- --- --- --- I4:24:0(2,2,10,10) 4.81 5.49 4.97 5.21 3.50 3.77 --- --- --- --- --- --- I3:24:0(2,10,12) 4.75 5.44 4.77 3.68 3.30 3.68 --- --- --- --- --- --- I4:26:0(2,2,10 ,12) 4.81 5.49 4.97 5.21 3.50 3.77 --- --- --- --- --- --- 312 Table 3. 15 . Summary of 13 C chemical shifts of inositol core carbon atoms. Chemical shifts labeled in bold indicate acyl substitutions are located at those positions. A ll spectra were referenced to non - deuterated solvent signal of acetonitrile - d 3 C = 118.70 ppm). Acylinositol ID R 1 R 2 R 3 R 4 R 5 R 6 R 1' R 2' R 3' R 4' R 5' R 6' NAG - I3:22:0(2,10,10) 72.25 69.43 71.33 81.50 73.91 71.81 102.40 57.46 75.63 72.06 77.72 62.78 G - I3:22:0(2,10,10) 72.07 69.19 70.79 82.86 73.70 71.51 104.69 74.36 77.4 0 71.05 77.58 62.42 X - I3:22:0(2,10,10) 72.43 69.53 71.34 81.82 73.79 71.90 105.53 74.37 77.61 70.75 66.91 --- NAG - I3:24:0(2,10,12) 72.26 69.44 71.32 81.57 73.91 71.83 102.38 57.47 75.65 72.08 77.73 62.82 G - I3:24:0(2,10,12) 72.38 69.50 71.09 83.19 74.00 71.83 104.98 74.69 77.72 71.37 77.91 62.74 I3:22:0(2,10,10) 72.84 69.61 72.53 71.80 or 71.75 75.69 71.80 or 71.75 --- --- --- --- --- --- I4:24:0(2,2,10,10) 72.47 69.42 70.36 72.66 73.50 71.86 --- --- --- --- --- --- I3:24:0(2,10,12) 72.84 69.61 72.53 7 1.80 or 71.75 75.70 71.80 or 71.75 --- --- --- --- --- --- I4:26:0(2,2,10,12) 72.48 69.42 70.36 72.67 73.52 71.87 --- --- --- --- --- --- 313 Table 3. 16 . NAG - I3:22:0( 2,10,10) Chemical shifts and coupling constants Molecular Formula: C 36 H 63 NO 14 110 min Retention Time (ESI - ): 58.00 mins HRMS: (ESI - ) m/z calculated for C 37 H 64 NO 16 - ([M+HCOO] - ): 778.4231, found: 778.4241 110 min Retention Time (ESI+): 57.98 mins HRMS: (ESI+) m/z calculated for C 36 H 64 NO 14 + ([M+H] + ): 734.4321, found: 734.4296 Instrument: Bruker Avance 900 MHz NMR Fraction: #16 Sample mass for NMR analysis: 2.3 mg NMR Solvent: D 3 CN InChi Key: CSNOFVDMFKBZIV - SHTLPGHTSA - N Carbon # (group) 1 H (ppm) 13 C (ppm) 1 (CH) 4.78 (dd, J = 10.3, 2.9 Hz) 72.25 - 1(CO) 171.45 - 2(CH 3 ) 1.96 21.36 2 (CH) 5.41 (t, J = 2.9 Hz) 69.43 - 1(CO) 174.1 a - 2(CH 2 ) 2.37 (m) 35 .0 b - 3(CH 2 ) 1.61 (p, J = 7.3 Hz) 26.26 - 4 to 9(CH 2 ) 1.36 - 1.25 (m) 30.7 - 30.1 c , 33.00 d , 23.78 e - 10(CH 3 ) 0.89 (t, J = 7.2 Hz) 14.78 f 3 (CH) 4.93 (dd, J = 10.2, 3.0 Hz) 71.33 - 1(CO) 174.1 a - 2(CH 2 ) 2.33 (m), 2.27 (m) 35.0 b - 3(CH 2 ) 1.54 (m), 1.50 (m) 25.9 0 - 4 to 9(CH 2 ) 1.36 - 1.25 (m) 30.6 - 30.1 c , 33.00 d , 23.78 e - 10(CH 3 ) 0.88 (t, J = 7.2 Hz) 14.78 f 4 (CH) 3.80 (dd, J = 10.4, 8.8 Hz) 81.50 5 (CH) 3.44 (t, J = 9.1 Hz) 73.91 6 (CH) 3.75 (dd, J = 10.5, 8.8 Hz) 71.81 1' (CH) 4.50 (d, J = 8.3 Hz) 102.40 (H - C, J = 163 Hz) g 2' (CH) 3.49 (m) 57.46 NH 6.39 (d, J = 8.4 Hz) --- - 1(CO) 171.55 - 2(CH 3 ) 1.82 23.82 314 Table 3.16. (continued) 3' (CH) 3.44 (t, J = 9.1 Hz) 75.63 4' (CH) 3.30 (dd, J = 10.0, 8.1 Hz) 72.06 5' (CH) 3.33 (ddd, J = 10.2, 5.8, 2.8 Hz) 77.72 6' (CH 2 ) 3.78 (dd, J = 11.8, 2.9 Hz), 3.61 (dd, J = 11.8, 5.8 Hz) 62.78 a - Two 13 C signals not resolved in 2D spectra (174.14, 174.12 ppm) b - Two 13 C signals not resolved in 2D spectra (35.05, 34.99 ppm) c - 13 C signals for CH 2 carbon positions 4 to 7 (30.63, 30.56, 30.47, 30.45, 30.44, 30.44, 30.12, 30.08 ppm) d - Overlapping 13 C signals for CH 2 carbon position 8 e - Overlapping 13 C signals for CH 2 carbon position 9 f - Overlapping 13 C signals for CH 3 carbon position 10 g - 1 J CH determined from HMBC break through signal 315 Figure 3. 26 . NAG - I3:22:0( 2,10,10) 1 H NMR 316 Figure 3. 27 . NAG - I3:22:0( 2,10,10) 13 C NMR 317 Figure 3. 28 . NAG - I3:22:0(2,10,10) 1 H - 1 H gCOSY 318 Figure 3. 29 . NAG - I3:22:0( 2,10,10) gHSQC 319 Figure 3. 30 . NAG - I3:22:0( 2,10,10) gHMBC 320 Figure 3. 31 . NAG - I3:22:0( 2,10,10) J - resolved 321 Figure 3. 32 . NAG - I3:22:0( 2,10,10) ROESY 322 Table 3. 17 . G - I3:22:0( 2,10,10) Chemical shifts and coupling constants Molecular Formula: C 34 H 60 O 14 110 min Retention Time (ESI - ): 58.83 mins HRMS: (ESI - ) m/z calculated for C 35 H 61 O 16 - ([M+HCOO] - ): 737.3965, found: 737.3965 110 min Retention Time (ESI+): 58.82 mins HRMS: (ESI+) m/z calculated for C 34 H 64 NO 14 + ([M+NH 4 ] + ): 710.4321, found: 710.4291 Instrument: Bruker Avance 900 MHz NMR Fraction: #18 Sample mass for NMR analysis: 3.3 mg NMR Solvent: D 3 CN InChi Key: LCZLGZXAZPZJHW - QWUQPGGFSA - N Carbon # (group) 1 H (ppm) 13 C (ppm) 1 (CH) 4.82 (dd, J = 10.3, 2.9 Hz) 72.07 - 1 (CO) 171.18 - 2(CH 3 ) 1.97 21.06 2 (CH) 5.47 (t, J = 2.9 Hz) 69.19 - 1(CO) 173.79 - 2(CH 2 ) 2.37 (t, J = 7.2 Hz) 34.71 a - 3(CH 2 ) 1.61 (p, J = 7.3 Hz) 25.99 - 4 to 9(CH 2 ) 1.37 - 1.24 (m) 30.3 - 29.8 b , 32.71 c , 23.47 d - 10(CH 3 ) 0.89 (t, J = 7.2 Hz) 14.48 e 3 (CH) 4.96 (dd, J = 10.2, 3.0 Hz) 70.79 - 1(CO) 173.85 - 2(CH 2 ) 2.24 (m) 34.71 a - 3(CH 2 ) 1.52 (m) 25.38 - 4 to 9(CH 2 ) 1.37 - 1.24 (m) 30.3 - 29.8 b , 32.71 c , 23.47 d - 10(CH 3 ) 0.88 (t, J = 7.2 Hz) 14.48 e 4 (CH) 3.78 (dd, J = 10.5, 8.8 Hz) 82.86 5 (CH) 3.44 (t, J = 9.1 Hz) 73.70 6 (CH) 3.77 (dd, J = 10.6, 8.8 Hz) 71.51 1' (CH) 4.33 (d, J = 7.9 Hz) 104.69 (H - C, J = 162 Hz) f 2' (CH) 3.11 (m) 74.36 3' (CH) 3.27 (m) 77.40 4' (CH) 3.27 (m) 71.05 5' (CH) 3.33 (ddd, J = 9.8, 5.8, 2.7 Hz) 77.58 323 Table 3.17. (continued) 6' (CH 2 ) 3.77 (dd, J = 11.9, 2.8 Hz), 3.61 (dd, J = 11.9, 5.9 Hz) 62.42 a - Overlapping 13 C signals b - 13 C signals for CH 2 carbon positions 4 to 7 (30.34, 30.27, 30.16, 30.16, 30.14, 30.13, 29.82, 29.76 ppm) c - Overlapping 13 C signals for CH 2 carbon position 8 d - Overlapping 13 C signals for CH 2 carbon position 9 e - Overlapping 13 C signals for CH 3 carbon position 10 f - 1 J CH determined from HMBC breakthrough signal 324 Figure 3. 33 . G - I3:22:0(2,10,10) 1 H NMR 325 Figure 3. 34 . G - I3:22:0(2,10,10) 13 C NMR 326 Figure 3. 35 . G - I3:22:0(2,10,10) 1 H - 1 H gCOSY 327 Figure 3. 36 . G - I3:22:0(2,10,10) gHSQC 328 Figure 3. 37 . G - I3:22:0(2,10,10) gHMBC 329 Figure 3. 38 . G - I3:22:0(2,10,10) J - resolved 330 Figure 3. 39 . G - I3:22:0(2,10,10) ROESY 331 Table 3. 18 . X - I3:22:0(2,10,10) Chemical shifts and coupling constants Molecular Formula: C 33 H 58 O 13 110 min Retention Time (ESI - ): 61.28 mins HRMS: (ESI - ) m/z calculated for C 34 H 59 O 15 - ([M+HCOO] - ): 707.3859, found: 707.3864 110 min Retention Time (ESI+): 61.25 mins HRMS: (E SI+) m/z calculated for C 33 H 62 NO 13 + ([M+NH 4 ] + ): 680.4216, found: 680.4176 Instrument: Bruker Avance 900 MHz NMR Fraction: #21 Sample mass for NMR analysis: 1.3 mg NMR Solvent: D 3 CN I nChi Key: LSJLPMMRXZQJNM - OTPYBMEMSA - N Carbon # (group) 1 H (ppm) 13 C (ppm) 1 (CH) 4.80 (dd, J = 10.2, 2.9 Hz) 72.43 - 1(CO) 171.48 - 2(CH 3 ) 1.96 21.37 2 (CH) 5.47 (t, J = 2.9 Hz) 69.53 - 1(CO) 174.08 - 2(CH 2 ) 2.37 (t, J = 7.2 Hz) 35.0 a - 3(CH 2 ) 1.61 (p, J = 7.3 Hz) 26.32 - 4 to 9(CH 2 ) 1.37 - 1.23 (m) 30.7 - 30.1 b , 33.02 c , 23.79 d - 10(CH 3 ) 0.89 (t, J = 7.2 Hz) 14.79 e 3 (CH) 4.94 (dd, J = 10.3, 3.0 Hz) 71.34 - 1(CO) 174.02 - 2(CH 2 ) 2.24 (m) 35.0 a - 3(CH 2 ) 1.52 (m) 25.69 - 4 to 9(CH 2 ) 1.37 - 1.2 3 (m) 30.7 - 30.1 b , 33.02 c , 23.79 d - 10(CH 3 ) 0.88 (t, J = 7.2 Hz) 14.79 e 4 (CH) 3.79 (dd, J = 10.5, 8.9 Hz) 81.82 5 (CH) 3.40 (t, J = 9.2 Hz) 73.79 6 (CH) 3.75 (dd, J = 10.5, 8.9 Hz) 71.90 1' (CH) 4.28 (d, J = 7.5 Hz) 105.53 (H - C, J = 163 Hz) f 2' (CH) 3.11 ( dd, J = 9.1, 7.4 Hz) 74.37 3' (CH) 3.27 (t, J = 8.8 Hz) 77.61 4' (CH) 3.47 (ddd, J = 10.3, 8.4, 5.4 Hz) 70.75 332 Table 3.1 8 . (continued) 5' (CH) 3.90 (dd, J = 11.4, 5.3 Hz), 3.21 (dd, J = 11.6, 10.0 Hz) 66.91 a - Two 13 C signals not resolved in 2D spectra (35.04, 35.01 ppm) b - 13 C signals for CH 2 carbon positions 4 to 7 (30.66, 30.58, 30.48, 30.47, 30.45, 30.45, 30.14, 30.06 ppm) c - Overlapping 13 C signals for CH 2 carbon position 8 d - Overlapping 13 C signals for CH 2 carbon position 9 e - Overlapping 13 C signals for CH 3 carbon position 10 f - 1 J CH determined from HMBC breakthrough signal 333 Figure 3. 40 . X - I3:22:0(2,10,10) 1 H NMR 334 Figure 3. 41 . X - I3:22:0(2,10,10) 13 C NMR 335 Figure 3. 42 . X - I3:22:0(2,10,10) 1 H - 1 H gCOSY 336 Figure 3. 43 . X - I3:22:0(2,10,10) gHSQC 337 Figure 3. 44 . X - I3:22:0(2,10,10) gHMBC 338 Figure 3. 45 . X - I3:22:0(2,10,10) J - resolved 339 Figure 3. 46 . X - I3:22:0(2,10,10) ROESY 340 Table 3. 19 . NAG - I3:24 :0( 2,10,12) Chemical shifts and coupling constants Molecular Formula: C 38 H 67 NO 14 110 min Retentio n Time (ESI - ): 64.28 mins HRMS: (ESI - ) m/z calculated for C 39 H 68 NO 16 - ([M+HCOO] - ): 806.4544, found: 806.4550 110 min Retention Time (ESI+): 64.27 mins HRMS: (ESI+) m/z calculated for C 38 H 68 NO 14 + ([M+H] + ): 762.4634, found : 762.4595 Instrument: Bruker Avance 900 MHz NMR Fraction: #27 Sample mass for NMR analysis: 1.0 mg NMR Solvent: D 3 CN InChi Key: BOJNLBZVYDCKSX - VYHRCMGRSA - N Carbon # (group) 1 H (ppm) 13 C (ppm) 1 (CH) 4.78 (dd, J = 10.3, 2.9 Hz) 72.26 - 1(CO) 171.45 - 2(CH 3 ) 1.96 21.36 2 (CH) 5.41 (t, J = 2.9 Hz) 69.44 - 1(CO) 174.1 a - 2(CH 2 ) 2.37 (m) 35.0 b - 3(CH 2 ) 1.61 (p, J = 7.3 Hz) 26.25 - 4 to 11(CH 2 ) 1.36 - 1.23 (m) 30.8 - 30.1 c , 33.0 d , 23.79 e - 12(CH 3 ) 0.88 (t, J = 7.2 Hz) 14.78 f 3 (CH) 4.93 (dd, J = 10.2, 3.0 Hz) 71.32 - 1(CO) 174.1 a - 2(CH 2 ) 2.33 (m), 2.27 (m) 35.0 b - 3(CH 2 ) 1.54 (m), 1.50 (m) 25.91 - 4 to 9(CH 2 ) 1.36 - 1.23 (m) 30.8 - 30.1 c , 33.0 d , 23.79 e - 10(CH 3 ) 0.88 (t, J = 7.2 Hz) 14.78 f 4 (CH) 3.80 (dd, J = 10.4, 8.8 Hz) 81.57 5 (CH) 3.44 (t, J = 9.1 Hz) 73.91 6 (CH) 3.75 (dd, J = 10.5, 8.9 Hz) 71.83 1' (CH) 4.50 (d, J = 8.3 Hz) 102.38 (H - C, J = 162.2 Hz) g 2' (CH) 3.49 (m) 57.47 NH 6.38 (d, J = 8.3 Hz) --- - 1(CO) 171.51 - 2(CH 3 ) 1.82 23.83 3' (CH) 3.43 (dd, J = 10.2, 8.5 Hz) 75.65 4' (CH) 3.29 (dd, J = 10.0, 8.2 Hz) 72.08 341 Table 3.1 9 . (continued) 5' (CH) 3.33 (ddd, J = 10.0, 5.9, 2.7 Hz) 77.73 6' (CH 2 ) 3.78 (dd, J = 11.9, 2.8 Hz), 3.61 (dd, J = 11.9, 5.8 Hz) 62.82 a - Two 13 C signals not resolved in 2D spectra (174.14, 174.13 ppm) b - Two 13 C signals not resolved in 2D spectra (35.07, 34.99 ppm) c - 13 C signals for CH 2 carbon positions 4 to 9, 4 to 7 (30.81, 30.75, 30.68, 30.57, 30.48, 30.45, 30.45, 30.45, 30.13, 30.09 p pm) d - Two 13 C signals for CH 2 carbon position 10 or 8 (33.03, 33.02 ppm) e - Overlapping 13 C signals for CH 2 carbon position 11 or 9 f - Overlapping 13 C signals for CH 3 carbon position 12 or 10 g - 1 J CH determined from coupled - HSQC measurement 342 Figure 3. 47 . NAG - I3:24:0(2,10,12) 1 H NMR 343 Figure 3. 48 . NAG - I3:24:0(2,10,12) 13 C NMR 344 Figure 3. 49 . NAG - I3:24:0(2,10,12) 1 H - 1 H gCOSY 345 Figure 3. 50 . NAG - I3:2 4:0(2,10,12) gHSQC 346 Figure 3. 51 . NAG - I3:24:0(2,10,12) coupled - gHSQC 347 Figure 3. 52 . NAG - I3:24:0(2,10,12) gHMBC 348 Figure 3. 53 . NAG - I3:24:0(2,10,12) J - resolved 349 Figure 3. 54 . NAG - I3:24:0(2,10,12) ROESY 350 Table 3. 20 . G - I3:24:0(2,10,12) Chemical shifts and coupling constants Molecular Formula: C 36 H 64 O 14 110 min Retention Time (ESI - ): 65.37 m ins HRMS: (ESI - ) m/z calculated for C 37 H 65 O 16 - ([M+HCOO] - ): 765.4278, found: 765.4278 110 min Retention Time (ESI+): 65.37 mins HRMS: (ESI+) m/z calculated for C 36 H 68 NO 14 + ([M+H] + ): 738.4634, found: 738.4597 Instrument: Bruker Avance 900 MHz NMR Fraction: #29 Sample mass for NMR analysis: 1.0 mg NMR Solvent: D 3 CN InChi Key: CSZWVWXZOGWQNT - FZFLAQSVSA - N Carbon # (group) 1 H (ppm) 13 C (ppm) 1 (CH) 4.82 (dd, J = 10.3, 2.9 Hz) 72.38 - 1(CO) 171.48 - 2(CH 3 ) 1.97 21.37 2 (CH) 5.47 (t, J = 2.9 Hz) 69.50 - 1(CO) 174.10 - 2(CH 2 ) 2.37 (t, J = 7.3 Hz) 35.03 a - 3(CH 2 ) 1.61 (p, J = 7.3 Hz) 26.30 - 4 to 11(CH 2 ) 1.36 - 1.23 (m) 30.8 - 30.1 b , 33.03 c , 23.79 d - 12(CH 3 ) 0.88 (t, J = 7.2 Hz) 14.78 e 3 (CH) 4.96 (dd, J = 10.3, 3.0 Hz) 71.09 - 1(CO) 174.16 - 2(CH 2 ) 2.24 (m) 35.03 a - 3(CH 2 ) 1.52 (m) 25.69 - 4 to 9(CH 2 ) 1.36 - 1.23 (m) 30.8 - 30.1 b , 33.03 c , 23.79 d - 10(CH 3 ) 0.88 (t, J = 7.2 Hz) 14.78 e 4 (CH) 3.78 (dd, J = 10.5, 8.8 Hz) 83.19 5 (CH) 3.44 (t, J = 9.1 Hz) 74.00 6 (CH) 3.76 (dd, J = 10.6, 8.8 Hz) 71.83 1' (CH) 4.33 (d, J = 7.8 Hz) 104.98 (H - C, J = 161.4 Hz) f 2' (CH) 3.10 (m) 74.69 3' (CH) 3.27 (m) 77.72 4' (CH) 3.27 (m) 71.37 5' (CH) 3.32 (ddd, J = 9 .8, 5.9, 2.5 Hz) 77.91 351 Table 3. 20 . (continued) 6' (CH 2 ) 3.77 (dd, J = 11.9, 2.6 Hz), 3.59 (dd, J = 11.9, 5.9 Hz) 62.74 a - Overlapping 13 C signals b - 13 C signals for CH 2 carbon positions 4 to 9, 4 to 7 (30.81, 30.76, 30.70, 30.59, 30.48, 30.46, 30.46, 30.45, 30.14, 30.07 ppm) c - Overlapping 13 C signals for CH 2 carbon position 10 or 8 d - Overlapping 13 C signals for CH 2 carbon position 11 or 9 e - Overlapping 13 C signals for CH 3 carbon position 12 or 10 f - 1 J CH determined from coupled - HSQC measure ment 352 Figure 3. 55 . G - I3:24:0(2,10,12) 1 H NMR 353 Figure 3. 56 . G - I3:24:0(2,10,12) 13 C NMR 354 Figure 3. 57 . G - I3:24:0(2,10,12) 1 H - 1 H gCOSY 355 Figure 3. 58 . G - I 3:24:0(2,10,12) gHSQC 356 Figure 3. 59 . G - I3:24:0(2,10,12) coupled - gHSQC 357 Figure 3. 60 . G - I3:24:0(2,10,12) gHMBC 358 Figure 3. 61 . G - I3:24:0(2,10,12) J - resolved 359 Figure 3. 62 . G - I3:24:0(2,10,12) ROESY 360 Table 3. 21 . I3:22:0(2,10,10) Chemical shifts and coupling constants Molecular Formula: C 28 H 50 O 9 110 min Retention Time (ESI - ): 68.18 mins HRMS: (ESI - ) m/z calculated for C 29 H 51 O 11 - ([M+HCOO] - ): 575.3437, found: 575.3435 110 min Retention Time (ESI+): 68.16 mins HRMS: (ESI+) m/z calculated for C 28 H 54 NO 9 + ([M+NH 4 ] + ): 548.3793, found: 548.3777 Instrument: Bruker Avance 900 MHz NMR Fraction: #34 Sample mass for NMR analysis: 3.6 mg NMR Solvent: D 3 CN InChi Key: ROFZSOBDHCZVAN - WBKISLEQSA - N Carbon # (group) 1 H (ppm) 13 C (ppm ) 1 (CH) 4.76 (dd, J = 10.2, 3.0 Hz) 72.84 - 1(CO) 171.49 - 2(CH 3 ) 1.96 21.39 2 (CH) 5.44 (t, J = 2.9 Hz) 69.61 - 1(CO) 174.1 a - 2(CH 2 ) 2.37 (t, J = 7.3 Hz) 35.0 b - 3(CH 2 ) 1.60 (p, J = 7.3 Hz) 26.30 - 4 to 9(CH 2 ) 1.36 - 1.23 (m) 30.6 - 30.1 c , 33.01 d , 23.78 e - 10(CH 3 ) 0.88 (t, J = 7.2 Hz) 14.78 f 3 (CH) 4.78 (dd, J = 10.2, 3.0 Hz) 72.53 - 1(CO) 174.1 a - 2(CH 2 ) 2.24 (m) 35.0 b - 3(CH 2 ) 1.53 (m) 25.95 - 4 to 9(CH 2 ) 1.36 - 1.23 (m) 30.6 - 30.1 c , 33.01 d , 23.78 e - 10(CH 3 ) 0.88 (t, J = 7.2 Hz) 14.78 f 4 (CH) 3.68 or 3.69 (dd, J = 10.5, 9.0 Hz) 71.8 g 5 (CH) 3.30 (t, J = 9.2 Hz) 75.69 6 (CH) 3.68 or 3.69 (dd, J = 10.5, 9.0 Hz) 71.8 g a - Two 13 C signals not resolved in 2D spectra (174.15, 174.13 ppm) b - Two 13 C signals not resolved in 2D spectra (35.08, 35.04 ppm) c - 13 C signals for CH 2 carbon positions 4 to 7 (30.61, 30.56, 30.46, 30.43, 30.42, 30.41, 30.09, 30.05 ppm) d - Overlapping 13 C signals for CH 2 carbon position 8 e - Overlapping 13 C signals for CH 2 carbon position 9 f - Overlapping 13 C signals for CH 3 carbon position 10 g - Two 13 C signals not resolved in 2D spectra (71.80, 71.75 ppm) 361 Figure 3. 63 . I3:22:0(2,10,10) 1 H NMR 362 Figure 3. 64 . I3:22:0(2,10,10) 13 C NMR 363 Figure 3. 65 . I3:2 2:0(2,10,10) 1 H - 1 H gCOSY 364 Figure 3. 66 . I3:22:0(2,10,10) gHSQC 365 Figure 3. 67 . I3:22:0(2,10,10) gHMBC 366 Figure 3. 68 . I3:22:0(2,10,10) J - resolved 367 Figure 3. 69 . I3:22:0(2,10,10) ROESY 368 Table 3. 22 . I4:24:0(2,2,10,10) Chemical shifts and coupling constants Molecular Formula: C 30 H 52 O 10 110 min Retention Time (ESI - ): 73.90 mins HRMS: (ESI - ) m/z calculated for C 31 H 53 O 12 - ([M+HCOO] - ): 617.3543, found: 617.3539 110 min Retention Time (ESI+): 73.89 mins HRMS: (ESI+) m/z calculated for C 30 H 56 NO 10 + ([M+NH 4 ] + ): 590.3899, found: 590.3892 Instrument: Agilent DDR2 500 MHz NMR Fraction: #44 Sample mass for NMR analysis: 27.8 mg NMR Solvent: D 3 CN InChi Key: QSRBSPJUNLXRLY - XBSCKGQLSA - N Carbon # (group) 1 H (ppm) 13 C (ppm) 1 ( CH) 4.81 (dd, J = 10.2, 2.9 Hz) 72.47 - 1(CO) 171.47 - 2(CH 3 ) 1.97 21.37 2 (CH) 5.49 (t, J = 3.0 Hz) 69.42 - 1(CO) 174.11 - 2(CH 2 ) 2.38 (t, J = 7.2 Hz) 35.07 - 3(CH 2 ) 1.62 (p, J = 7.2 Hz) 26.36 - 4 to 9(CH 2 ) 1.37 - 1.21 (m) 30.6 - 30.1 a , 33.0 b , 23.8 c - 10(CH 3 ) 0.88 (t, J = 7.2 Hz) 14.8 f 3 (CH) 4.97 (dd, J = 10.6, 3.0 Hz) 70.36 - 1(CO) 173.80 - 2(CH 2 ) 2.18 (m) 35.00 - 3(CH 2 ) 1.48 (p, J = 7.3 Hz) 25.89 - 4 to 9(CH 2 ) 1.37 - 1.21 (m) 30.6 - 30.1 a , 33.0 b , 23.8 c - 10(CH 3 ) 0.88 (t, J = 7.2 Hz) 14.8 f 4 (CH) 5.21 (t, J = 10.0 Hz) 72.66 - 1(CO) 171.14 - 2(CH 3 ) 2.01 21.54 5 (CH) 3.50 (t, J = 9.5 Hz) 73.50 6 (CH) 3.77 (t, J = 9.7 Hz) 71.86 a - 13 C signals for CH 2 carbon positions 4 to 9, 4 to 7 (30.60, 30.52, 30.46, 30.30.46, 30.41, 30.36, 30.06, 30.05 ppm) b - Two 13 C si gnals for CH 2 carbon position 10 or 8 (33.03, 33.00 ppm) c - Two 13 C signals for CH 2 carbon position 11 or 9 (23.78, 23.77 ppm) d - Two 13 C signals for CH 3 carbon position 12 or 10 (14.79, 14.78 ppm) 369 Figure 3. 70 . I4:24:0(2,2 ,10,10) 1 H NMR 370 Figure 3. 71 . I4:24:0(2,2,10,10) 13 C NMR 371 Figure 3. 72 . I4:24:0(2,2,10,10) 1 H - 1 H gCOSY 372 Figure 3. 73 . I4:24:0(2,2,10,10) gHSQC 373 Figure 3. 74 . I4:24:0(2,2,10,10) gHMBC 374 Figure 3. 75 . I4:24:0(2,2,10,10) J - resolved 375 Figure 3. 76 . I4:24:0(2,2,10,10) ROESY 376 Figure 3. 77 . I4:24:0(2,2,10,10) TOCSY 377 Table 3. 23 . I3:24:0(2,10,12) Chemical shifts and coupling constants Molecular Formula: C 30 H 54 O 9 110 min Retention Time (ESI - ): 75.14 mins HRMS: (ESI - ) m/z calculated for C 31 H 55 O 11 - ([M+HCOO] - ): 60 3.3750, found: 603.3745 110 min Retention Time (ESI+): 75.13 mins HRMS: (ESI+) m/z calculated for C 30 H 58 NO 9 + ([M+NH 4 ] + ): 576.4106, found: 576.4077 Instrument: Bruker Avance 900 MHz NMR Fraction: #48 Sample mass for NMR analysis: 1.1 mg NMR Solvent: D 3 CN InChi Key: NCEAPMSZCADWMM - XBSCKGQLSA - N Carbon # (group) 1 H (ppm) 13 C (ppm) 1 (CH) 4.75 (dd, J = 10.2, 3.0 Hz) 72.84 - 1(CO) 171.49 - 2(CH 3 ) 1.96 21.38 2 (CH) 5.44 (t, J = 2.9 Hz) 69.61 - 1(CO) 174.1 a - 2(CH 2 ) 2.35 (t, J = 7.3 Hz) 35.1 b - 3(CH 2 ) 1.60 (p, J = 7.3 Hz) 26.30 - 4 to 11(CH 2 ) 1.36 - 1.23 (m) 30.8 - 30.1 c , 33.0 d , 23.78 e - 12(CH 3 ) 0.88 (t, J = 7.2 Hz) 14.77 f 3 (CH) 4.77 (dd, J = 10.2, 3.0 Hz) 72.53 - 1(CO) 174.1 a - 2(CH 2 ) 2.24 (m) 35.1 b - 3(CH 2 ) 1.53 (m) 25.95 - 4 to 9(CH 2 ) 1.36 - 1.23 (m) 30.8 - 30.1 c , 33.0 d , 23.78 e - 10(CH 3 ) 0.88 (t, J = 7.2 Hz) 14.77 f 4 (CH) 3.68 (dd, J = 10.5, 9.0 Hz) 71.8 g 5 (CH) 3.30 (t, J = 9.2 Hz) 75.70 6 (CH) 3.68 (d d, J = 10.5, 9.0 Hz) 71.8 g a - Two 13 C signals not resolved in 2D spectra (174.15, 174.13 ppm) b - Two 13 C signals not resolved in 2D spectra (35.09, 35.04 ppm) c - 13 C signals for CH 2 carbon positions 4 to 9, 4 to 7 (30.78, 30.74, 30.65, 30.56, 30.47, 30.42, 30.42, 30.42, 30.10, 30.05 ppm) d - Two 13 C signals for CH 2 carbon position 10 or 8 (33.03, 30.01 ppm) e - Overlapping 13 C signals for CH 2 carbon position 11 or 9 f - Overlapping 13 C signals for CH 3 carbon position 12 or 10 g - Two 13 C signals n ot resolved in 2D spectra (71.80, 71.75 ppm) 378 Figure 3. 78 . I3:24:0(2,10,12) 1 H NMR 379 Figure 3. 79 . I3:24:0(2,10,12) 13 C NMR 380 Figure 3. 80 . I3:24:0(2,10,12) 1 H - 1 H gCOSY 381 Figu re 3. 81 . I3:24:0(2,10,12) gHSQC 382 Figure 3. 82 . I3:24:0(2,10,12) gHMBC 383 Figure 3. 83 . I3:24:0(2,10,12) J - resolved 384 Figure 3. 84 . I3:24:0(2,10,12) ROESY 385 Table 3. 24 . I4:26:0(2,2,10,12) Chemical shifts and coupling constants Molecular Formula: C 32 H 56 O 10 110 min Retention Time (ESI - ): 80.27 mins HRMS: (ESI - ) m/z calculated for C 33 H 57 O 12 - ([M+HCOO] - ): 645.3856, found: 645.3857 110 min Retention Time (ESI+): 80.26 mins HRMS: (ESI+) m/z calculated for C 32 H 60 NO 10 + ([M+NH 4 ] + ): 618.4212, found: 618.4181 Instrument: Bruker Avance 90 0 MHz NMR Fraction: #53 Sample mass for NMR analysis: 1.5 mg NMR Solvent: D 3 CN InChi Key: RAWFQSLFZCSMTH - YBYHVSLFSA - N Carbon # (group) 1 H (ppm) 13 C (ppm) 1 (CH) 4.81 (dd, J = 10.2, 3.0 Hz) 72.48 - 1(CO) 171.46 - 2(CH 3 ) 1.97 21.36 2 (CH) 5.49 (t, J = 2.9 Hz) 69.42 - 1(CO) 174.11 - 2(CH 2 ) 2.38 (m) 35.07 - 3(CH 2 ) 1.62 (p, J = 7.3 Hz) 26.36 - 4 to 11(CH 2 ) 1.37 - 1.21 (m) 30.8 - 30.1 a , 33.0 b , 23.8 c - 12(CH 3 ) 0.88 (t, J = 7.2 Hz) 14.77 d 3 (CH) 4.97 (dd, J = 10.6, 3.0 Hz) 70.36 - 1(CO) 173.80 - 2(CH 2 ) 2.18 (m) 35.00 - 3(CH 2 ) 1.48 (p, J = 7.5 Hz) 25.89 - 4 to 9(CH 2 ) 1.37 - 1.21 (m) 30.8 - 30.1 a , 33.0 b , 23.8 c - 10(CH 3 ) 0.88 (t, J = 7.2 Hz) 14.77 f 4 (CH) 5.21 (dd, J = 10.8, 9.4 Hz) 72.67 - 1(CO) 171.14 - 2(CH 3 ) 2.01 21.54 5 (CH) 3.50 (t, J = 9.5 Hz) 73.52 6 (CH) 3.77 (dd, J = 10.5, 9.0 Hz) 71.87 a - 13 C signals for CH 2 carbon positions 4 to 9, 4 to 7 (30.78, 30.75, 30.64, 30.51, 30.47, 30.44, 30.41, 30.05, 30.05, 30.05 ppm) b - Two 13 C signals for CH 2 ca rbon position 10 or 8 (33.04, 32.99 ppm) c - Two 13 C signals for CH 2 carbon position 11 or 9 (23.78, 23.77 ppm) d - Overlapping 13 C signals for CH 3 carbon position 12 or 10 386 Figure 3. 85 . I4:26:0(2,2,10,12) 1 H NMR 387 Figure 3 . 86 . I4:26:0(2,2,10,12) 13 C NMR 388 Figure 3. 87 . I4:26:0(2,2,10,12) 1 H - 1 H gCOSY 389 Figure 3. 88 . I4:26:0(2,2,10,12) gHSQC 390 Figure 3. 89 . I4:26:0(2,2,10,12 ) gHMBC 391 Figure 3. 90 . I4:26:0(2,2,10,12) J - resolved 392 Figure 3. 91 . I4:26:0(2,2,10,12) ROESY 393 Figure 3. 92 . I4:24:0(2,2,10,10) 1D - TOCSY transfer spectra 2 - - CH 2 excitation at 1 .62 ppm (generated using Varian Inova 600 MHz spectrometer). Table 3. 25 . I4:24:0(2,2,10,10) 1D - TOCSY transfer 2 - - CH 2 (S ) and 2 - CH 3 (S) normalized integrals. mix ing time (ms) 2 - ß - CH 2 excitation at 1.62 ppm 2 - CH 3 S/S o at 0.88 ppm 0 1.0000 - 0.0023 100 0.1992 0.0065 120 0.2397 0.0092 140 0.1706 0.0134 160 0.1441 0.0179 180 0.1607 0.0213 200 0.1445 0.0243 220 0.1380 0.0265 240 0.1427 0.0278 260 0.1498 0.0332 280 0.1330 0.0337 300 0.1249 0.0326 394 Figure 3. 93 . I4:24:0(2,2,10,10) 1D - TOCSY transfer spectra 3 - - CH 2 excitation at 1.48 ppm (generated using Varian Inova 600 MHz spectrometer). Table 3. 26 . I4:24:0(2,2,10,10) 1D - TOCSY transfer 3 - - CH 2 (S ) and 3 - CH 3 (S) normalized integrals. mixing time (ms) 3 - ß - CH 2 excitation at 1.48 ppm 2 - C H 3 S/S at 0.88 ppm 0 1.0000 0.0001 100 0.2283 0.0056 120 0.2651 0.0093 140 0.1761 0.0129 160 0.1518 0.0155 180 0.1816 0.0199 200 0.1482 0.0227 220 0.1433 0.0253 240 0.1640 0.0251 260 0.1498 0.0309 280 0.1376 0.0321 300 0.1239 0.0319 395 Figure 3. 94 . I4:26:0(2,2,10,12) 1D - TOCSY transfer spectra 2 - - CH 2 excitation at 1.62 ppm (generated using Varian Inova 600 MHz spectrometer). Table 3. 27 . I4:26:0(2,2,10,12) 1D - TOCSY transfer 2 - - CH 2 (S ) and 2 - CH 3 (S) normalized integrals. mixing time (ms) 2 - ß - CH 2 excitation at 1.62 ppm 2 - CH 3 S/S o at 0.88 ppm 0 1.0000 0.0008 100 0.1673 0.0025 120 0.2161 0.0035 140 0.2020 0.0071 160 0.1435 0.0081 180 0.1366 0.0092 200 0.1532 0.0132 220 0.1395 0.0144 240 0. 1391 0.0155 260 0.1275 0.0167 280 0.1147 0.0162 300 0.1081 0.0194 396 Figure 3. 95 . I4:26:0(2,2,10,12) 1D - TOCSY transfer spectra 3 - - CH 2 excitation at 1.48 ppm (generated using Varian Inova 600 MHz spectrometer). Table 3. 28 . I4:26:0(2,2,10,12) 1D - TOCSY transfer spectra 3 - - CH 2 excitation at 1.48 ppm (generated us ing Varian Inova 600 MHz spectrometer). mixing t ime (ms) 3 - ß - CH 2 excitation at 1.48 ppm 2 - CH 3 S/S at 0.88 ppm 0 1.0000 - 0.0009 100 0.1753 0.0043 120 0.2411 0.0072 140 0.2113 0.0109 160 0.1482 0.0127 180 0.1505 0.0151 200 0.1592 0.0199 220 0.1437 0.0206 240 0.1580 0.0224 260 0.1425 0.0222 280 0.1227 0.0236 300 0.1148 0.0248 397 REFERENCES 398 REFERENCES 1. Moghe, G.D., et al., Evolutionary routes to biochemical innovation revealed by integrative analysis of a plant - defense related specialized metabolic pathway. Elife, 2017. 6 : p. e28486. 2. Ghosh, B., T.C. Westbrook, and A.D. Jones, Comparative structural profiling of trichome specialized metabolites in tomato (Solanum lycopersicum) and S. habrochaites: acylsug ar profiles revealed by UHPLC/MS and NMR. Metabolomics, 2014. 10 (3): p. 496 - 507. 3. King, R.R., et al., Characterization of 2,3,4,3' - tetra - O - acylated sucrose esters associated with the glandular trichomes of Lycopersicon - typicum. Journal of Agricultural an d Food Chemistry, 1993. 41 (3): p. 469 - 473. 4. Liu, X.X., et al., Profiling, isolation and structure elucidation of specialized acylsucrose metabolites accumulating in trichomes of Petunia species. Metabolomics, 2017. 13 (7): p. 85. 5. Schilmiller, A.L., A.L. Charbonneau, and R.L. Last, Identification of a BAHD acetyltransferase that produces pr otective acyl sugars in tomato trichomes. Proc Natl Acad Sci U S A, 2012. 109 (40): p. 16377 - 82. 6. Fan, P., et al., In vitro reconstruction and analysis of evolutionary variation of the tomato acylsucrose metabolic network. Proc Natl Acad Sci U S A, 2016. 113 (2): p. E239 - 48. 7. Schilmiller, A.L., et al., Functionally divergent alleles and duplicated Loci encoding an acyltransfer ase contribute to acylsugar metabolite diversity in Solanum trichomes. Plant Cell, 2015. 27 (4): p. 1002 - 17. 8. Nadakuduti, S.S., et al., Characterization of trichome - expressed BAHD acyltransferases in petunia axillaris reveals distinct acylsugar assembly m echanisms within the Solanaceae. Plant Physiol, 2017. 175 (1): p. 36 - 50. 9. Schilmiller, A., et al., Mass spectrometry screening rev eals widespread diversity in trichome specialized metabolites of tomato chromosomal substitution lines. Plant J, 2010. 62 (3): p. 391 - 403. 10. Kroumova, A.B., D. Zaitlin, and G.J. Wagner, Natural variability in acyl moieties of sugar esters produced by cert ain tobacco and other Solanaceae species. Phytochemistry, 2016. 130 : p. 218 - 27. 11. Forero, D.P., et al., Chemical and sensor y comparison of fresh and dried lulo (Solanum quitoense Lam.) fruit aroma. Food Chem, 2015. 169 : p. 85 - 91. 12. Gancel, A.L., et al. , Identifying carotenoids and phenolic compounds in naranjilla (Solanum quitoense Lam. var. Puyo hybrid), an Andean fruit. J Agric Food Chem, 2008. 56 (24): p. 11890 - 9. 13. Forero, D.P., et al., Spermidine derivatives in lulo (Solanum quitoense Lam.) fruit: sensory (taste) versus biofunctional (ACE - Inhibition) properties. J Agric Food Chem, 2016. 64 (26): p. 5375 - 83. 14. Osorio, C ., C. Duque, and F. Batista - Viera, Studies on aroma generation in lulo (Solanum quitoense): enzymatic hydrolysis of glycosides from leaves. Food Chemistry, 2003. 81 (3): p. 333 - 340. 399 15. Treutler, H., et al., Discovering regulated metabolite families in unta rgeted metabolomics studies. Anal Chem, 2016. 88 (16): p. 8082 - 90. 16. Plumb, R.S., et al., UPLC/MS (E) ; a new approach for generatin g molecular fragment information for biomarker structure elucidation. Rapid Commun Mass Spectrom, 2006. 20 (13): p. 1989 - 94. 1 7. Sumner, L.W., et al., Proposed minimum reporting standards for chemical analysis Chemical Analysis Working Group (CAWG) Metabolo mics Standards Initiative (MSI). Metabolomics, 2007. 3 (3): p. 211 - 221. 18. Ghosh, B. and A.D. Jones, Profiling, characterizat ion, and analysis of natural and synthetic acylsugars (sugar esters). Analytical Methods, 2017. 9 (6): p. 892 - 905. 19. Ekanayaka, E. A., M.D. Celiz, and A.D. Jones, Relative mass defect filtering of mass spectra: a path to discovery of plant specialized meta bolites. Plant Physiol, 2015. 167 (4): p. 1221 - 32. 20. Burke, B.A., G. Goldsby, and J.B. Mudd, Polar epicuticular lipids of Lycopers icon pennellii. Phytochemistry, 1987. 26 (9): p. 2567 - 2571. 21. Matsuzaki, T., et al., Leaf surface glycolipids from Nicotiana - acuminata and Nicotiana - pauciflora. Agricultural and Biological Chemistry, 1991. 55 (5): p. 1417 - 1419. 22. Loewus, F.A. and P.P.N. Murthy, myo - inositol metabolism in plants. Plant Science, 2000. 150 (1): p. 1 - 19. 23. Duus, J., C.H. Gotfredsen, and K. Bock, Carbohydrate structural determination by NMR spectroscopy: modern methods and limitations. Chem Rev, 2000. 100 (12): p. 4589 - 614. 24 . Sachleben, J.R., et al., Aliphatic chain length by isotropic mixing (ALCHIM): determining composition of complex lipid samples by (1)H NMR spectroscopy. J Biomol NMR, 2014. 59 (3): p. 161 - 73. 25. Herrera - Salgado, Y., et al., Myo - inositol - derived glycolipi ds with anti - inflammatory activity from Solanum lanceolatum. J Nat Prod, 2005. 68 (7): p. 1031 - 6. 26. Luckwill, L.C., The genus Lycopersicon. An historical, biological and taxonomic survey of the wild and cultivated tomatoes. PhD thesis, Department of Botan y, University of Aberdeen, UK, 1943. 27. Glas, J.J., et al., Plant glandular trichomes as targets for breeding or enginee ring of resistance to herbivores. Int J Mol Sci, 2012. 13 (12): p. 17077 - 103. 400 Chapter 4 : S - alkyl glutathione retention indexing and co lumn performance evaluation standards for improved annotation, identification and dereplication of metabolite disco very 4.1 Introduction Annotation of novel metabolites remains as the greatest obstacle to understanding the mechanisms responsible for metabo lite accumulation and the functional significance of new metabolites. The combination of mass spectrometry (MS) with high - performance liquid chromatography (HPLC) separations provides the foundation for chemical analysis of non - volatile small molecules in cluding drugs, pesticides, endogenous metabolites and other complex mixtures, yet only a small fraction of substances detected by LC/MS have been identified. LC/MS approaches exploit relationships between chromatographic retention times (RT), mass - to - charg e ( m/z ) measurements and fragment ion spectra for identification and quantification, or for det ermination of unknown elemental compositions and partial structural information [1] . - high perform ance LC (UHPLC) separations with phenomenal chromatographic resolution and peak capacities [2] , the number of metabolites detected in some untargeted metabolite analyses far exceeds chromatographic peak capacities. Coupled with high resolution mass analyze rs, modern UHPLC/MS instruments provide high acquisition speeds, mass accuracy and sensitive de tection [3] . Despite great advancements in LC/MS technologies, annotation and identification of unknowns continues to present a bottleneck. Perhaps in no other f ields is this more evident than in natural product discovery and metabolomics research, where t he immense array of chemical diversity in living organisms presents formidable challenges [4] . Our ability to measure metabolites within complex extracts far exc eeds our ability to generate unambiguous structure information about most of them. For mass spe ctral interpretation, GC/MS electron ionization (EI) and LC/MS tandem (MS/MS) collision induced dissociation (CID) mass spectral database searches and molecular networking approaches aim to match mass spectra to a library or to provide structural insights 401 from compounds that nearly match [1, 5] . A preferred method for identifying metabolites involves matching mass spectra and retention time data using authentic re ference materials. However, it not practical or possible to have reference standards available for all metabolites, and of course, these do not exist for novel metabolites not described before. All too often, mass spectra alone are not enough to fully char acterize and/or identify metabolites, and final structure confirmation is performed using indep endent approaches including one - and two - dimensional (1D and 2D) NMR spectroscopy or x - ray crystallography, and these techniques work best after compounds have b een purified from mixtures. While NMR spectroscopy provides more detailed information about con nections between individual atoms than MS, it is frequently avoided because it requires time - consuming purifications, adequately pure samples and far greater sam ple quantities than MS for analysis (~80% purity, ~1 mg). In addition, as more metabolite struc tures are determined, there is a growing chance that researchers will purify and identify compounds that are already known but not in mass spectrum libraries. I n view of this, dereplication of metabolite discovery by LC/MS approaches is essential for impr oving the efficiency of metabolite identification. Mass spectral databases are widely used for LC/MS [6, 7] , yet the addition of RT data to MS/MS libraries has found limited success because RTs often vary across laboratories, instruments, and experimental protocols. In GC/MS analyses, this problem has been addressed through Kováts retention indexing (RI), which converts gas chromatographic RTs into system - indepen dent constants by normalizing RTs to those of a homologous series of n - alkanes, generating RIs that are comparable across platforms [8] . RI values are calculated by linear interpolation between retention standards according to Equation 4.1, where, t r is th e retention time, and n and N are the number of carbon atoms in the shorter and longer n - alkyl standards bracketing the peak of interest respectively. Equation 4. 1 . Calculatio n of RI values. 402 By comparison to GC/MS databases, LC/MS approaches have yet to find widespread use. In part, this is because MS/MS spectral databases are less complete, and LC protocols vary, leading to retention being less reproducible across labor atories, and there are no widely - accepted RP - HPLC RI standards [1] . A few reports of RI standards have consisted of homologous series of 2 - ketoalkanes, alkyl aryl ketones, and 1 - nitroalkanes with ultraviolet spectroscopic detection [9 - 11] , but most lack fu nctional groups that fa cilitate their detection using mass spectrometry. Since electrospray ionization (ESI) is often the first approach for metabolite discovery by LC/MS, indexing standards that ionize well using both positive - and negative - ion modes woul d be preferable. This w ay, the same standard series could be used in both ionization modes. Two recent reports use combinations of N - alkyl acetamides and N - alkylbenzamides homologous standards to characterize HPLC gradients and project retention times acro ss laboratories and ins truments. However, it is not clear whether those standards are suitable for detection by both ionization modes [12, 13] . A recent patent by Quilliam (U.S. Patent No. US9594063B2) documents 1 - alkylpyridinesulfonic acids that are ioniz able using both modes, however, those retention standards have not been described in peer - reviewed publication [14, 15] . To address these issues, a homologous set of S - alkyl derivatives of the tripeptide glutathione (GS - n - alkyl RI standards ) featuring norm al saturated chain leng ths (1 - 24 carbons) was synthesized. Experimental factors that influence RI values for 16 previously identified acylsucrose metabolites from Salpiglossis sinuata (outlined in Chapter 2) were examined by varying chromatographic condit ions including brands o f RP C18 columns, solvent delivery systems, which may differ in delays in delivering mixed solvents to columns, mobile phase pH, column temperature, steepness of LC gradients, and variation in organic solvent modifiers. Acylsugars ar e diverse sugar polyest er metabolites that accumulate in many plants, especially within the family Solanaceae [16] . Hundreds of unique acylsugars have been detected and resolved by UHPLC/MS [17 - 19] . Moreover, they exist in numerous isomeric and isobaric fo rms in a single plant e xtract, many of which are not distinguished using MS/MS, making them suitable to test the performance of LC chromatographic retention indexing to distinguish isomers. 403 4.1.1 LC column performance Silica and silica hybrid supports wit h bonded groups based o n octadecyl (C18) groups are the most widely used materials for RP - HPLC applications [20] . Compounds that exhibit ideal chromatographic behavior are expected to elute in a Gaussian profile that results from statistical behavior of in dividual molecules trav eling through the column. However, it is common that chromatographic elution appears in the form of asymmetric, or tailing, peaks. Columns based on silica supports often exhibit severe peak tailing of basic compounds, limited pH rang e compatible with colum n stability, and irreproducibility of retention for the same column chemistries [21] . There are several factors that influence column reproducibility, including particle shape and size, pore diameter, surface area, stationary phase l oading, bonded phase de nsity, metal impurities, endcapping, and more [20] . Approaches to column performance troubleshooting have identified four main causes of poor chromatographic performance: 1) blockage of the frit of the column, 2) void formation, 3) a dsorbed sample impuriti es that alter analyte partitioning between mobile and stationary phases, or 4) chemical attack on the silica or bonded phase [22] . Symptoms of column - related problems manifest as increased column backpressure, peak fronting/tailing, reduced column plate nu mber, and changes in retention including selectivity. For reversed - phase alkyl silica columns, hydrophobic factors dominate analyte partitioning between mobile and stationary phases and retention, particularly for neutral analytes. H owever, contributions f rom other solute - column interactions described by the Hydrophobic Subtraction Model are widely considered responsible for column selectivity and peak tailing [23] . Principally, underivatized silanol groups, which are always present o n silica column surface s, play a key role, and can interact with retained solutes [21] . For example, protonated bases can interact with ionized silanol groups ( - Si - O - ), and neutral silanol groups ( - Si - OH) can hydrogen bond with proton acceptor solutes. End - capping of silanol groups, often with trimethylsilyl (TMS) groups, reduces the number of underivatized silanol groups and decreases the silanol accessibility because of steric hindrance between underivatized silanol and TMS and/or C18 groups, but endcappi ng fails to achieve complete inactivation of silanols [21] . Hydrolysis of the bonded phase may 404 occur as the column ages, or more aggressively due to operating at very low pH (usually pH < 2). Under moderately basic conditions (usually pH > 8), the silica s upport reacts with hydroxide ion at more substantial rates and may dissolve, reducing the structural integrity of the silica support. Dissolution of the silica suppor t may release the bonded stationary phase and create voids in the column due to particle c ollapse. In any case, these processes give rise to more underivatized silanol groups, changes in peak shape, retention and selectivity [24] . 4.2 Materials and Methods 4.2.1 Synthesis of S - alkyl glutathione standards Preparation of GS - n - alkyl standards was performed by modifications to methods previously reported [25] . An abbreviated nomenclature of S - alkylglutathiones is suggested as GS - n, where n equals the number of carbon atoms in a linear alkyl group. Authentic standards of GS - 1 (used for preparation o f mixed stock solution), GS - 6 and GS - 8 were purchased from Sigma - Aldrich and had matching HPLC retention times and fragment ion spectra generated by nonselective coll ision - induced dissociation compared to synthesis products. The purity of reaction products was qualitatively assessed using LC/MS. 4.2.1.1 Synthesis of GS - 2 and GS - 3 To a 250 - mL round bottom flask, ~10 mmoles reduced glutathione (GSH, Sigma - Aldrich) was ad ded. A 10 - mL volume of MilliQ water was added, followed by addition of 10 mL of 2.0 M aque ous NaOH solution. Dissolution of the glutathione was promoted by stirring using a magnetic stir bar at room temperature. Once the glutathione had dissolved, 75 mL of 95% ethanol was added, and the solution turned cloudy. An equimolar quantity of ethyl iod ide or propyl iodide (Sigma - Aldrich) was added. The vessel was sealed with a rubber septum, then purged by bubbling N 2 gas through a syringe and venting while mixing for ~10 mins. The reaction was left to proceed overnight (12 - 16 h) at room temperature (~ to ~4.5 by dropwise addition of 12 M HCl and the solution concentrated under rotary evaporation (~40°C) to a volume of ~10 mL where precipi tates were observed. The mixture was further recrystallized by cooling in an ice bath, fol lowed by vacuum filtration to recover precipitates. The precipitates were washed with 405 methanol and dried. The standards were further purified by repeating the recryst allization in water, recovered by vacuum filtration and washed with methanol and dried ove rnight in a 40°C drying cabinet. Purity was checked by LC/MS. The yield was approximately 10% and 25% for GS - 2 and GS - 3 respectively (most material was lost during r ecrystallization and washing steps). 4.2.1.2 Synthesis of GS - 4 to GS - 9 Syntheses of homolo gs GS - 4 to GS - 9 were performed according to Vince et. al. (1971) with minor deviations. To a 40 - mL glass vial, ~2.0 mmoles of GSH was added. A 2.0 - mL volume of MilliQ water was added, followed by addition of 2.0 mL of 2.0 M NaOH solution. Dissolution of th e glutathione was promoted by stirring using a magnetic stir bar at room temperature. After dissolution, ~15 mL of 95% ethanol was added until the solution turned clo udy. An equimolar quantity of alkyl halide RX (X = iodine for butyl through octyl; X = bro mine for nonyl; Sigma - Aldrich) was added, vials were sealed with Teflon lined caps and the reaction was left to proceed 12 - e often visible within a few hours. The pH was adjusted to ~3.5 by dropwise addition of 1 2 M HCl and the vials were placed in a - 20°C freezer for 12 - 16 h. Insoluble products were recovered by vacuum filtration, with alternating washes using water and hexa nes, and then dried overnight in a 40°C drying cabinet. The yields ranged from 55 - 80%. 4.2 .1.3 Synthesis of GS - 10 to GS - 20 and GS - 22 To a 100 - mL round bottom flask, ~0.4 mmoles of alkyl bromide (from Sigma - Aldrich, except docosyl bromide from TCI America) was added. A 45 - mL volume of 95% ethanol and stir bar was added. The mixture was sealed wi th a rubber septum, then purged by bubbling N 2 gas through a syringe with venting (~10 min) while heating in a sand bath at ~55°C. In a separate vial, a 1.1 molar eq uivalent of GSH was added to a centrifuge tube, along with 2.0 mL of MilliQ water. A volum e of 2.0 M NaOH corresponding to 2:1 molar ratio of NaOH to GSH was added to the vial and vortexed to dissolve GSH. Its contents were added dropwise to the round bott om flask via a syringe while mixing and purging with N 2 gas. The N 2 gas line was removed f rom the septum, and the sealed flask was placed in an orbital mixer at 55°C and 140 406 rpm. The reaction was left to proceed for 12 - 16 h. To stop the reaction, the flask was cooled to room temperature and pH was adjusted to ~3.5 with dropwise addition of 12 M HCl. The flask was placed in a - 20°C freezer overnight. The insoluble products were recovered by vacuum filtration, with alternating washes using water and hexanes, then dried overnight in a 40°C drying cabinet. The yields were ~65%. 4.2.1.4 Synthesis of GS - 21, GS - 23 and GS - 24 Synthesis of GS - 21, GS - 23 and GS - 24 was performed using the same method described in Section 4.2.1.3. However, longer chain 1 - bromoalkanes were not readily available for purchase and were synthesized from alcohol analogues: 1 - heneico sanol (Ultra Scientific), 1 - tricosanol (Ultra Scientific) and 1 - tetracosanol (TCI America). Their synthesis is described below. All glassware was dried in an oven >10 0 °C overnight before use. To a 100 - mL three neck round bottom flask, ~200 mg of long chai n alcohol was added. The reactor was sealed with rubber septa, placed in a sand bath at ~60 °C and purged with N 2 gas. Twenty - five mL of toluene: tetrahydrofuran (1:1 , v:v) dried with 3 Å molecular sieves was added to the reactor. A 10% molar excess of pho sphorous tribromide (PBr 3 ) was added and the reaction left overnight. The reaction was quenched with 25 mL of 0.01 M aqueous KBr solution. The mixture was added to a 125 - mL separatory funnel and the water layer removed. The organic layer was washed three t imes with ~20 mL MilliQ water. The organic layer was added to a 100 - mL three neck round bottom reactor and dried under N 2 gas. Colorless solids were visible. The crud e product was placed in an oven (> 100 °C) and dried overnight. The crude product underwen t a second treatment with PBr 3 described above to raise the yield. After washing, the organic layer was dried under N 2 gas. The products were dissolved in 2.0 mL hexa nes and purified using 500 mg - 3 mL Avantor Bakerbond silica gel solid phase extraction c olumns. The columns were conditioned with 12 mL of hexanes, followed by the addition of 2.0 mL of crude reaction products, then eluted with 6.0 mL of hexanes. The elu ate was added to a 100 - mL round bottom flask and dried under N 2 gas leaving white solids. Identities of 1 - bromoalkane synthesis products were confirmed by analysis using GC - MS. Typical yields were ~50%. 407 4.2.2 Preparation of retention index standards and their application in LC/MS analyses 4.2.2.1 Preparation of GS - n - alkyl mixed stock solution Individual GS - n - alkyl stocks were prepared by weighing ~60 µmoles of each standard into 40 - mL glass vials (see Table 4.1 for stock preparation details). For vials containing GSH to GS - 9, a 22.5 - mL volume of 3:3:2 acetonitrile: isopropanol: water (AcN:IPA:H 2 O, v:v:v) was added to each. To vials containing GS - 10 to GS - 24, an 18.0 - mL volume of 3:3:2 AcN:IPA:H 2 O was ad ded to each. To all vials, 75 µL of ammonium hydroxide (NH 4 OH, ACS, EMD Millipore, assay percentage range 28 - 30%) was added. Finally, 4.5 mL of c hloroform was added to vials containing GS - 10 to GS - 24. In most cases the standards had not dissolved completel y (especially for longer chain lengths). The vials were placed on an orbital shaker at 55 °C and 150 rpm. The stock solutions were occasionally r emoved, vigorously shaken by hand and visually examined for undissolved material. When all standard material ha d dissolved completely, an aliquot of each stock was combined to a mixed stock solution (according to the values listed in Table 4.1) such that e ach standard was present at 107.5 µM. The final volume of the mixed stock was adjusted to 100 µM each by additi on of 1.834 mL of 3:3:2 AcN:IPA:H 2 O solution. The stock solutions and mixed stock were stable when stored in a - 20 °C freezer for more than 6 mon ths and showed no apparent degradation of the GS - n - alkyl standards based on LC/MS analyses. However, reduced GS H in the mixed stock was mostly converted to the oxidized form (GSSG) when tested after a few months. 408 Table 4. 1 . Preparat ion of GS - n - alkyl Mixed Stock Solution GS - # Mass Weighed ( g ) Vol. 3:3:2 (mL) Vol. CHCl 3 (mL) Vol NH 4 OH (mL) Total Vol. Stock 1 (mL) Stock 1 Conc. (µM) Vol. from Stock 1 (mL) Mix Stock Conc. (µM) H 0.0186 22.5 --- 0.075 22.575 2681 0.981 107.5 1 0.0219 2 2.5 --- 0.075 22.575 3019 0.871 107.5 2 0.0201 22.5 --- 0.075 22.575 2655 0.990 107.5 3 0.0213 22.5 --- 0.075 22.575 2700 0.974 107.5 4 0.0223 22.5 --- 0.075 22.575 2718 0.967 107.5 5 0.0228 22.5 --- 0.075 22.575 2676 0.983 107.5 6 0.0232 22.5 --- 0.0 75 22.575 2625 1.002 107.5 7 0.0237 22.5 --- 0.075 22.575 2589 1.016 107.5 8 0.0260 22.5 --- 0.075 22.575 2745 0.958 107.5 9 0.0272 22.5 --- 0.075 22.575 2779 0.946 107.5 10 0.0261 18.0 4.5 0.075 22.575 2583 1.018 107.5 11 0.0280 18.0 4.5 0.075 22.575 2687 0.979 107.5 12 0.0288 18.0 4.5 0.075 22.575 2682 0.980 107.5 13 0.0291 18.0 4.5 0.075 22.575 2632 0.999 107.5 14 0.0299 18.0 4.5 0.075 22.575 2629 1.000 107.5 15 0.0309 18.0 4.5 0.075 22.575 2644 0.995 107.5 16 0.0321 18.0 4.5 0.075 22.575 2674 0.983 107.5 17 0.0327 18.0 4.5 0.075 22.575 2654 0.991 107.5 18 0.0340 18.0 4.5 0.075 22.575 2690 0.977 107.5 19 0.0377 18.0 4.5 0.075 22.575 2910 0.904 107.5 20 0.0345 18.0 4.5 0.075 22.575 2600 1.011 107.5 21 0.0356 18.0 4.5 0.075 22.575 2620 1.004 107.5 22 0.0376 18.0 4.5 0.075 22.575 2704 0.972 107.5 23 0.0375 18.0 4.5 0.075 22.575 2637 0.997 107.5 24 0.0398 18.0 4.5 0.075 22.575 2738 0.960 107.5 Stock 2 Vol (mL) = 24.458 Volume 3:3:2 solution added to adjust final Conc. to 100 µM (mL)= 1.8 34 4.2.2.2 Preparation of S. sinuata retention indexing stock solution To evaluate the performance of GS - n - alkyls for retention indexing, the mixed stock solution was added to achieve a concentration of 1.0 µM to a diluted extract from Salpiglossis sinua ta containing acylsucrose metabolites. The sample was prepared by diluting 1.00 mL of S. sinuata bulk extract (described in Chapter 2, Section 2.2.1) into 8.80 mL of 3:3:2 AcN:IPA:H 2 O. Followed by a 100 µL addition of GS - n - alkyl mixed stock solution (warme d to 50 °C and shaken vigorously prior to addition). Finally, an equimolar concentration of formic acid relative to NH 4 OH in the solution was added (100 µL of 0.046 M). 409 4.2.3 UHPLC/MS methods and experimental conditions 4.2.3.1 LC/MS instrument configurati ons Retention indexing was performed u sing two Waters Xevo G2 - XS quadrupole time - of - flight mass spectrometers equipped with either Shimadzu or Waters LC systems. The Shimadzu LC system was equipped with LC - 20AD pumps, CTO - 20A column oven and SIL - 5000 autos ampler. The Waters LC system consisted of Acquity UPLC I - Class Binary Solvent Manager and 2777C Sample Manager. The instruments were operated using electrospray ionization in positive - and negative - ion modes (ESI+ and ESI - yzer parameters. All retention indexin g experiments were performed in triplicate, with two analyses by ESI - and one by ESI+. Nonselective CID spectra were acquired over m/z 50 - 1500 in centroid format using four quasi - simultaneous collision potential functi ons (0, 10, 20, 40 V, each with 0.1 s acquisition times) to generate fragment ions. Lockspray mass correction was applied using Leu - enkephalin as a lock mass reference standard (0.1 s acquisition time, 10 s scan frequency). 4.2.3.2 Retention indexing peak detection parameters Chromatographic retention times were determined using TargetLynx software (Waters). For GS - n - alkyls, [M - H] - and [M+H] + monoisotopic ions were detected. For acylsucrose metabolites, formate ([M+HCOO] - ) and ammonium ([M+NH 4 ] + ) monoisot opic adduct ions were detected, except when acylsucrose isomers were unresolved. For those instances, distinguishing fragment ions from high energy CID spectra were used (see Table 4.2). Peak integrations were adjusted manually as needed, the integration p arameters are as follo ws: 1) smoothing method mean; 2) smoothing iterations 5; 3) smoothing width 3; 4) chromatographic mass window 0.0 2 Da. Retention time data were exported and processed using Microsoft Excel. 410 Table 4. 2 . L C/MS RT and RI results for 16 S. sinuata acylsucrose metabolites with structural identifications (column SAE - A, temperature 50°C, aqueous 10 mM ammonium formate pH 2.8 and linear gradient 1 - 100% acetonitrile, slope 1% B min - 1 ) . Analyses were performed in t riplicate (two by ESI - are three S4:20:0 isomers (peaks #2 - 4) described below in Figure 4.3. Pseudomolecular [M+HCOO] - and [M+NH 4 ] + ions w ere detected by ESI - and ESI+ ion modes, except for italicized characteristic fragment ions generated at elevated collision potentials (20 V CID by ESI - and 10 V CID by ESI+). Acylsucrose # Acylsucrose ID Ion detected by ESI - ( m/z ) Ion detected by ESI+ ( m/z ) Avg Ret. Time (min) Ret. Ti me SD (min) Ret. Time RSD (%) RI Value RI Value SD RI Value RSD (%) 1 S4:19 :0( 3,5,5,6) 579.266 401.217 50.26 0.384 0.76 1446 0.6 0.04 2 S4:19 :0( 2,5,6,6) 621.313 443.264 50.41 0.386 0.77 1451 0.6 0.04 3 S5:20 :0( 2,2,5,5,6) 737.324 710.359 51.54 0.392 0.76 1487 0.4 0.03 4 S4:20 :0( 4,5,5,6) 723.345 696.380 52.85 0.398 0.75 1529 0.4 0.03 5 S4:20 :0( 3,5,6,6) 579.266 415.233 53.44 0.401 0.75 1548 0.4 0.03 6 S4:20 :0( 2,6,6,6) 635.328 457.280 53.64 0.402 0.75 1555 0.3 0.02 7 S5:22 : 1 ( 2,5,5,5,5 T ) 763.339 736.375 5 4.54 0.404 0.74 1583 0.3 0.02 8 S5:21 :0( 2,2,5,6,6) 751.343 724.375 54.75 0.406 0.74 1590 0.4 0.03 9 S4:21 :0( 5,5,5,6) 737.360 710.396 55.47 0.411 0.74 1613 0.4 0.02 10 S5:25 : 4 ( 2,5,5,5,8 P ) 799.339 772.375 56.18 0.416 0.74 1635 0.3 0.02 11 S5:22 :0( 2,2,6,6 ,6) 765.355 738.391 57.94 0.424 0.73 1691 0.2 0.01 12 S4:22 :0( 5,5,6,6) 751.376 724.411 58.46 0.424 0.73 1708 0.3 0.02 13 S5:23 :0( 2,5,5,5,6) 779.372 752.406 60.65 0.412 0.68 1776 0.4 0.02 14 S4:23 :0( 5,6,6,6) 765.391 738.427 61.54 0.412 0.67 1804 0.4 0.02 15 S5:24 :0( 2,5,5,6,6) 793.389 766.422 63.53 0.408 0.64 1865 0.5 0.03 16 S6:25 :0( 2,2,5,5,5,6) 821.381 794.417 64.70 0.411 0.64 1901 0.3 0.02 4.2.3.3 Chromatographic columns All separations were performed using C18 columns with dimensions 10 cm × 2.1 mm . Three column types from leading manufacturers were chosen for comparison: 1) three Supelco Ascentis Express C18 (SAE; 2.7 µm particle size, superficially porous particles with monofunctional C18 bonding chemistry and endcapping) from separate lots and di ffering stages of use were measured (column S AE - A had Bin Lot: S17138, column SA E - B had BL: S17156 and column SAE - C had BL: S15013). Columns SAE - A and - B were new columns, while column SAE - C had been heavily used before evaluation; 2) Waters Acquity BEH C1 8 (WBEH; Lot No. 0308372931, 1.7 µm particle size, fully porous particle, 18% carbon loading, 411 trifunctional bonding chemistry with endcapping); and 3) Agilent Zorbax Eclipse Plu s (AZEP; L.N. B17300, 1.8 µm particle size, fully porous particle, 9% carbon lo ading, monofunctional bonding chemistry with endcapping). 4.2.3.4 Column performance and LC systems evaluations Comparisons of columns and LC systems were performed using linear gradient elution, aqueous 10 mM ammonium hydroxide adjusted to pH 2.8 with for mic acid (Solvent A) and acetonitrile (Solvent B). Solvents were not degassed. The elution profile used for GS - n - alkyl retention indexing of acylsugars consisted of a 110 - min me thod: hold at 1% B from 0 - 1 min, linear gradient from 1 - 100% B over 1 - 100 min ( slope 1% Solvent B min - 1 ), hold at 100% B at 100 - 105 min, linear 100 - 1% B over 105 - 106 min, and hold at 1% B from 106 - 110 min. Injections of 5 µL (using a 5 µL sample loop) were performed with mobile phase flow rate of 0.3 mL min - 1 and column oven temperat ure at 50 °C. The method described herein was used for all measurements unless otherwise noted. 4.2.3.5 Mobile phase pH dependence of acylsugar metabolite retention index values To assess pH dependence of GS - n - alkyl standards and acylsucrose RI values, the pH of aqueous mobile phase Solvent A (starting with 10 mM aqueous ammonium hydroxide) was adjusted from pH 2.5 - 4.0, in steps of 0.3 pH units with formic acid (Solvent A). Colum n SAE - A was used for all analyses of pH dependence. All other LC conditions are as outlined in Section 4.2.3.4. 4.2.3.6 Column temperature dependence of acylsugar retention index values To assess column temperature dependence of GS - n - alkyl standards and a cylsucrose RI values, the column temperature was adjusted from 30 to 60 °C in s teps of 5 °C. Column SAE - A was used for all experiments. All other LC conditions are as outlined in Section 4.2.3.4. 4.2.3.7 Dependence of acylsugar retention index values on mobile phase gradient slope To assess gradient slope dependence of GS - n - alkyl st andards and acylsucrose RI values, the slope of the linear gradient was adjusted to values of 1, 10/9, 5/4, 10/7, 5/3 and 2% Solvent B min - 1 (wash and 412 equilibration times are described in Section 4.2.3.4; total run times: 110, 100, 90, 80, 70 and 60 mins r espectively). Column SAE - A was used for all experiments. All other LC conditions are as outlined in Section 4.2.3.4. 4.2.3.8 Acylsugar retention index values using methanol as organic mobile phase component To examine the effect of organic component depe ndence of GS - n - alkyl standards and acylsucrose RI values, methanol (Solvent B) was substituted for acetonitrile. Column SAE - A was used for all experime nts. All other LC conditions are as outlined in Section 4.2.3.4. 4.2.3.9 MS/MS spectra of GS - n - alkyl sta ndards MS/MS spectra of GS - n - alkyl products were generated using Waters SONAR data acquisition platform using a 10 - 40 V collision potential ramp, 10 Da quadrupole transmission bin widths and 0.5 s acquisition times. The LC method was as described in Sectio n 4.2.3.4. However, methanol was used as organic component instead of acetonitrile. Measurements were performed in ESI - and ESI+ modes using S. sinuata bulk extract containing 10 µM GS - n - alkyl standards. Ion abundance measurements were determined using Wat ers MS E Data Viewer Software (Version 1.4). Peak detection parameters were as follows: 1) chromatographic peak width auto; 2) MS resolution 0.05 Da ; 3) Lock mass enabled Leu - Enkephalin ( m/z 554.262); 3) low energy filtering threshold 100 counts; 4) h igh energy filtering threshold 10 counts; 5) chromatographic peak width factor 1.0; 6) quadrupole time peak width factor 1.0; and XIC window for high energy data 0.05 Da. 4.3 Results and discussion The performance of using S - alkylglutathione homolo gs for retention indexing was examined by adding equimolar GS - n - alkyl standards from a mixed stock solution to an extract of S. sinuata containing neut ral acylsucrose metabolites and performing LC/MS analyses using linear gradient elution with acetonitrile as organic mobile phase component (Figure 4.1). The experimental conditions used for data presented herein are given in Section 4.2.3.4, unless noted otherwise (Ascentis Express C18 Column A (SAE - A), aqueous 10 mM ammonium formate pH 2.8 and linear gradie nt 1 - 100% acetonitrile, slope 1% B min - 1 ). RI values 413 were evaluated under different chromatographic conditions using linear gradient elution and by alt ering columns, solvent delivery systems, mobile phase pH, column temperature, linear gradient slope and o rganic mobile phase component. 4.3.1 LC/MS of GS - n - alkyl standards GS - n - alkyl RI standards encompass a wide RP - UHPLC retention range and are ionized by ESI in positive - and negative - ion modes (yielding [M+H] + and [M - H] - ions). Extracted ion chromatograms ( XICs) for characteristic fragment ions such as m/z 162.022 in positive - ion mode (C 5 H 8 NO 3 S + , Figure 4.1B) and m/ z 254.078 (C 10 H 12 N 3 O 5 - ) in negative - ion mode allow for their selective detection using a single narrow m/z window for each polarity, simplifying reporting of standard RTs when they have been added to a sample. GSH and GS - 1 standards were not retained and eluted as unretained so lutes. GS - 2 was poorly retained and displayed peaks that eluted at the solvent front and a smaller peak that eluted at ~3 m ins (when detected, the later peak was reported). This behavior is attributed to having the standard dissolved in a strong solvent t hat contributed to double chromatographic peaks. Standards GS - 3 to GS - 5 showed peak fronting/splitting that was attributed - n - alkyl functional groups at injection (the later/larger peak was r eported). Standards < GS - 6 carbon chain length showed lower signal intensity than the longer chain homologs, a portion of w hich was attributed to sampling of broadened/split peaks and lower ionization efficiency when the aqueous mobile phase was the domina nt mobile phase component at those elution times [26] . GS - 6 to GS - 24 showed single peaks that broadened with more tailing a s the series chain length increased. This may be attributed to reduced solvation of charged functional groups at high acetonitrile so lvent compositions. Figure 4.2 displays a polynomial fit for standards GSH (GS - 0) to GS - 7, relating retention times to corr esponding RI values, while standards GS - 7 to GS - 24 exhibit a linear relationship (~3.3 mins per GS - n - alkyl standard) between retentio n time and RI value. Thus, the GS - n - alkyl elution order is predictable. 414 Figure 4. 1 . LC/MS chrom atograms of S. sinuata extract containing acylsucroses plus 1.0 µM each GS - n - alkyl RI standards analyzed by ESI+ mode (column SAE - A, temperature 50°C, aqueous 10 mM ammonium formate pH 2.8 and linear gradient 1 - 100% acetonitrile, slope 1% B min - 1 ). (A) BPI chromatogram of S. sinuata extract containing acylsucroses . (B) Extracted ion chromatogram (XIC) of common fragment ion m/z 162.022 (C 5 H 8 NO 3 S + ) generated at 20 V collision potential. Illustration of GS - n - alkyl standard chemical structure and proposed fra gmentation positions. 415 Figure 4. 2 . GS - n - alkyl HPLC retention times on a Supelco Ascentis Express C18 column using aqueous ammonium formate/acetonitrile gradient as a function of alkyl chain length (RI value) (n=3 replicates). 4.3 .2 RI corrects day - to - day chromato graphic variation Even when chromatographic protocols are followed (aiming to standardize LC methods, instrument, column and solvents), day - to - day RT variations compromise RT comparisons across multiple samples, particular ly when multiple isomeric forms ar e detected. Figure 4.3A - D illustrates chromatographic retention drift over the course of a 7 - week period. Extracted ion chromatograms (XICs) for a group of S4:20:0 S. sinuata acylsucrose isomers labeled peaks #1 - 4 were mea sured on four separate dates (C 32 H 54 O 15 , detected as [M+NH 4 ] + at m/z 696.380, peaks #2 - 4 have structural identifications outlined in Chapter 2). Inspection of overlaid XICs for adjacently eluting GS - n - alkyl standards, GS - 15 and GS - 16 (C 25 H 47 N 3 O 6 S and C 26 H 49 N 3 O 6 S , [M+H] + at m/z 518.326 and 532.341), shows parallel RT drift of standards and analytes. For instance, acylsucrose peak #4 (S4:20 :0( 2,6,6,6)) exhibits significant RT variability. Relative to the first measurement (Figure 4.3A), RTs varied relative t o the first measurement by - 0.95, - 0.83 and - 0.38 mins (Figure 4.3B - D), or - 1.7, - 1.5 and - 0.7% different respectively. However, the absolute percent deviations of RI values for peak #4 from the first RI measurement were < 0.03%. 416 Figure 4. 3 . RT drift and application of RI correction applied to a group of S4:20:0 S. sinuata acylsucrose isomers ([M+NH 4 ] + , green) with overlaid GS - n - alkyl standards ([M+H] + , blue) using column S AE - A. Peaks #2 - 4 have structural identifications based on NMR spectra. Dates when analyses were performed are displayed in year, month, day format. (A) 20180203, (B) 20180209, (C) 20180303, (D) 20180322. RTs of the GS - n - alkyl standards and 16 S. sinuata acylsucrose metabolites were determined and the RI values c alculated for each of the acylsucrose metabolites (Table 4.2). Values of standard deviation (SD) in RT of twelve measurements were approximately ±0.4 mins, or ~0.7% relative standard deviation (RSD), while the SD of RI values was approximately ±0.4, or ~0. 02% RSD. To il lustrate the day - to - day differences among the 16 acylsucroses, Figure 4.4A - B displays RT and RI percent differences for each of four triplicate analyses, with results displayed relative to the first date of analysis. This example shows that t he differences in acylsucrose RI values are much smaller compared to variability of RT and demonstrates how GS - n - alkyl retention indexing has capacity for reducing day - to - day and/or sample batch - to - batch chromatographic variability. 417 Figure 4. 4 . Percent difference comparisons of RT and RI values applied to 16 S. sinuata acylsucroses (each marker represents one of the acylsucrose metabolites) relative to the first set of measurements (column SAE - A, temperature 50°C, aque ous 10 mM am monium formate pH 2.8 and linear gradient 1 - 100% acetonitrile, slope 1% B min - 1 ). Results are plotted against RT and RI value for visual purposes. Dates of analyses are indicated in the legend using year, month, day format. (A) RT percent diffe rences relat ive to the first analysis. (B) RI percent differences relative to the first analysis. 4.3.3 Potential for cross - platform RI application GS - n - alkyl standards show improved capacity for comparing retention data across laboratories. To demonstrate RI applica tion across laboratory conditions, the S. sinuata extract was analyzed using two different LC solvent delivery systems (Shimadzu LC - 20AD and Waters Acquity I - class pumps; same column SAE - A), and by performing measurements on the same make and model columns (SA E ) produced from separate lots and at different stages of use (same Shimadzu LC system; columns SAE - A and SAE - B were new columns, while c olumn SAE - C had been heavily used). For comparison, all results are referenced to column SAE - A measured on the Shim adzu LC system. Figure 4.5A - B demonstrates a significant shift to earlier RTs when the same column (SAE - A) is transferred to the Waters LC sy stem, consistent with lower extracolumn dead volumes or differences in mixing of solvents in gradient formation. Fo r instance, Peak #4 had RTs that differed by - 1.62 mins, or - 3.1% different. However, when RI was applied (values of 1555 and 1552 respective ly for a single acylsucrose peak #4 (S4:20 :0( 2,6,6,6)), the difference is reduced to only - 0.1%. Figures 4.5A and 4 .5C compare two SAE columns prepared from separate lots (SAE - A and - 418 B) measured using the Shimadzu LC system. Peak #4 had RTs that differed b y - 0.72 min, or - 1.3%, while the RI value (1559) only differed by 0.3%. Figure 4. 5 . LC s ystem and column dependence assessment using SAE columns. The group of S4:20:0 S. sinuata acylsucrose isomers peaks #1 - 4 is shown ([M+NH 4 ] + , green) with overlaid GS - 15 and GS - 16 ([M+H] + , blue). (A) Column SAE - A, measured using Shimadzu LC system. (B) Colum n SAE - A measured using the Waters LC system. (C) Column SAE - B measured using Shimadzu LC system. RI values on different LC systems and colu mns were calculated for the 16 acylsucrose metabolites. Figure 4.6A - B illustrates percent differences relative to co lumn SAE - A, measured using the Shimadzu LC system. Acylsucrose RTs showed substantial differences across LC systems, with as much as - 3.5% d ifference for earlier eluting acylsucroses (Figure 4.6A). In contrast, their RI values differed over a range of - 0.2 to 0.05% (Figure 4.6B). When the column was changed from SAE - A to SAE - was applied, their values ranged f rom 0.3 to 0.7% different. The greater differences observed for columns SAE - A to SAE - B are attributed to batch - to - ba tch column variability. The RI results from a third column SAE - C, showed reduced capacity for RI correction. These findings are discussed fu rther in Section 4.3.4. 419 Figure 4. 6 . LC system and SA E columns RT and RI value percent dif ference comparison when applied to 16 acylsucroses from S. sinuata (percent differences are relative to column SAE - A measured using Shimadzu LC system). Columns, LC systems and dates of analyses are indicated in the legend. (A) RT percent difference (B) RI percent difference 4.3.4 GS - n - alkyls for use in column performance evaluation Silica - based LC columns eventually show reduced performance with time and use owing to dissolution of the silica support [27] , which dissolves slightly in the range of pH 2 - 7 a nd more substantially at higher pH [28] . Symptoms of column - related problems manifest as inc reased column backpressure, peak asymmetry, reduced column plate number, retention selectivity changes and retention time changes. To illustrate this, Figures 4.7A - B compare the group of S4:20:0 S. sinuata acylsucrose isomers analyzed with SAE columns at v arious stages of use. Column SAE - A was new (<15 hours use at the time of analysis), while column SAE - C had been heavily used (estimated ~900 hours), involving analy sis of hundreds of acidic and basic sample injections (pH < 2 and pH > 8) and had been brief ly operated at low pH while at temperatures above manufacturer - recommended conditions (aqueous phase pH 2.5, up to 70°C for a period > 20 hours). Inspection of pea ks #1 - 4 in Figure 4.7 reveals lower chromatographic resolution, peak fronting, as well as ch anges in selectivity for the heavily - used column SAE - C (notably, column backpressures of SAE - A, - B and - C were roughly the same). Without this direct comparison to column SAE - A, the reduced column 420 performance of column SAE - C could be overlooked. A surprisi ng finding was that acylsucroses had RTs that were more consistent than RI values (peak #4 RTs showed - 0.8% difference, while RI values were - 1.5% different). These results are further illustrated in Figures 4.6A - B, where the 16 acylsucrose metabolites had RTs that differed by - 0.7 to - 0.9%, while RI values differed by - 1.4 to - 1.8%. In this instance, GS - n - alkyls RI values revealed evidence of column degradation in t he form of changed RI values. In this context, GS - n - alkyl standards demonstrate value as a column performance test mixture. Comparison of SAE - A and SAE - C high energy XICs for GS - n - alkyl common fragment ion m/z 162.022 in Figure 4.8A - B reveals distinguishin g peak shape differences for each column. Column SAE - C shows substantial peak broadening/ta iling and later RTs with increasing GS - n - alkyl chain length (Figure 4.8B). Interestingly, this - 19. It may be notable that these standards had alkyl groups that exceeded the length of the octadecyl - C18 bonded phase. The peak width of standard GS - 24 demonstrates this most clearly. It had w 50% of ~0.25 mins for column SAE - A, while column SAE - C was ~2.0 mins (Figure 4.8A - B). In addition, its RTs differed by ~5.5 mins. By comparison, RT as a f unction of RI value for column SAE - C reveals an upward inflection starting near GS - 19 (Figure 4.8C). This feature was not observed for column SAE - A (or SAE - B, not sh own). These results suggest that GS - n - alkyl standard peak shapes and RTs could be used as t ool for evaluating column performance and column equivalency and identifying when a column should be replaced. 421 Figure 4. 7 . Column performa nce comparison. The group of S. sinuata S4:20:0 acylsucrose isomers is shown ([M+NH 4 ] + , green) with overlaid standards GS - 14, - 15, and - 16 ([M+H] + , blue). (A) Analysis using column SAE - A. (B) Analysis using heavily used column SAE - C. 422 Figure 4. 8 . Column performance comparison using XICs for GS - n - alkyl common fragment ion m/z 162.022 (generated at 20 V collision energy). GS - 19 to GS - 24 are boxed in red. (A) Column SAE - A. (B) Column SAE - C. (C) Plot of RT against RI value using each c olumn. Considering these results, it is recommended that researchers establish a set of column equivalency acceptance parameters before making comparisons between GS - n - alkyl RI values, even when the same LC method conditions and column make/model is used. Exploration of predefined limits for column performance parameters such as RT, theoretical plate number (N) or w 50% , peak asymmetry factor (A s ) and tailing factor (TF) may be used to establish column lifetime and/or batch - to - batch column equivalency. To il lustrate this, Figure 4.9A - C displays XICs for GS - 22 (C 32 H 61 N 3 O 6 S , [M+H] + at m/ z 616.435) measured using columns SAE - A, - B and - C respectively. Columns SAE - A and - B had ~4.5 times more theoretical plates than the heavily - used column SAE - C. Perhaps the most perilous evidence for reduced column performance using column SAE - C are values of A s and TF. Column SAE - A had values of A s = 1.56 and TF = 1.58, while heavily - used column SAE - C had A s = 3.10 and TF = 2.49 respectively. 423 Figure 4. 9 . Analysis of column performance test parameters using standard GS - 22 [M+H] + as example. (A) Column SAE - A. (B) Column SAE - B. (C) Column SAE - C. The increased peak asymmetry and tailing of GS - n - alkyl standards using column SAE - C is believed to be the produ ct of interactions of the standard with underivatized silanol groups. These include cation - exchange interactions between the positively - charged amino group with negatively - charged ionized silanol groups (Figure 4.10). Furthermore, vicinal silanol groups, c omp rised of adjacently bonded silanols may contribute to bidentate hydrogen bonding with the carboxylic acids of the GS - n - alkyl standards (Figure 4.10). These and other interactions (such as adsorbed impurities) produce changes in retention and selectivity wi th column age and use [23] . Accordingly, the examination of RT and peak shape parameters of GS - n - alkyl standards offers a new approach for evaluating column performance, column lifetime and manufacturer batch - to - batch column equivalency. 424 Figure 4. 10 . Predicted underivatized silanol group interactions. Approximate pKa values were calculated using ChemDraw Professional Software (Version 16.0.1.4). 4.3.5 C18 columns differ in retention selectivity While hydrophobic solute - colu mn and solute - solvent interactions dominate the C18 RP - HPLC retention process, other solute - column interactions contribute to differences in retention and selectivity across columns [23] . Differences among column chemistries, such as bonded phase chemistry (m onomeric versus polymeric), silica supports (silica versus hybrid silica), ligand density (µmoles·m - 2 ), particle pore diameter, extent of end - capping (or residual silanols) and particle packing all play important roles in column retention and selectivit y. Consequently, it is improbable that a single RI standard series could be developed that would yield identical RI values on all types of reversed - phase columns, even when the column bonded phase is of common alkyl length (e.g. C18). To illustrate this, t hre e C18 column types from leading manufacturers were chosen for side - by - side evaluation of RI reproducibility. Columns were chosen for their differences in particle technology, carbon loading levels (percent by weight of carbon on the stationary phase) an d C 18 bonding chemistry. SA E superficially porous columns (pore size 90 Å, surface area 120 m 2 ·g - 1 ) were chosen because they can function at similar pressures to more traditional HPLC columns, while achieving comparable separation 425 efficiencies for a 2.7 µm pa rticle size as UHPLC separations performed on sub - 2 µm particles. An Agilent Zorbax Eclipse Plus (AZEP) UHPLC column was selected because it is manufactured using fully porous silica particles with monomeric C18 bonded phase but with lower carbon loadin g l evels of 9% (pore size 95 Å, surface area 160 m 2 ·g - 1 ). Finally, a Waters Acquity BEH (WBEH) UPLC column was selected because it is manufactured using fully porous bridged ethylene hybrid (BEH) silica particles with polymeric (trifunctional) C18 bonded p has e and has higher carbon loading levels of 17.7% (pore size 130 Å, surface area 185 m 2 ·g - 1 ). Figure 4. 11 . LC/MS chromatograms and retention of GS - n - alkyl standards using C18 columns from three leading manufacturers (measured usi ng Waters LC system). XICs for common fragment ion m/z 254.078 (C 10 H 12 N 3 O 5 - ) generated at 20 V collision energy in ESI - mode. (A) Column SAE - A. (B) Column AZEP. (C) Column WBEH. (D) Plot of RT against RI value using each column. Measurements using each col umn were performed in sequence on the same day (to reduce day - to - day chromatographic variance) using the Waters Acquity LC system with identical LC conditions. Figu re 4.11A - C demonstrates universal detection of the GS - n - alkyl series with ESI - mode common f ragment ion m/z 254.078 (C 10 H 12 N 3 O 5 - ) measured using columns SA E - A, AZEP and WBEH respectively. The RTs for each column are plotted against RI value in Figure 4.11D . Notably, no other columns had an upward 426 inflection like the heavily used column SAE - C (Fig ure 4.9C). While RTs differed, the GS - n - alkyl standards had predictable elution order and similar peak shape for all three column manufacturers. Figure 4.12A - C comp ares the retention indexing of the group of S4:20:0 S. sinuata acylsucrose isomers analyzed on each column. Like the GS - n - alkyl standards, the peak shape and width of resolved acylsucroses were comparable. However, the selectivity is noticeably different f or each column. By comparison, isomeric peaks #3 and #4 had reduced retention selectivity wh en column AZEP was used (Figure 4.12B), but were baseline resolved using column WBEH (Figure 4.12C). By comparison, it appears there are upwards of seven isomers de tected using column WBEH, where columns SAE - A and AZEP resolved only four isomers. According ly, RI values were noticeably different across the columns. For instance, peak #4 had RI values of 1552 (SAE - A), 1583 (AZEP) and 1498 (WBEH) respectively. Thus, is not appropriate to assume that all C18 columns will have the same selectivity and RI values generated from different column manufacturers may not be equivalent. 427 Figure 4. 12 . Column manufacturer selectivity and RI dependence. The group of S4:20:0 S. sinuata acylsucrose isomers is shown ([M+NH 4 ] + , green) with overlaid GS - 14, - 15, and - 16 ([M+H] + , blue). (A) Column SAE - A. (B) Column AZEP. (C) Column WBEH. 4.3.6 RI dependence on mobile phase pH, column temperature and LC gradient The pH of the mobile phase, column temperature and LC gradient play important roles for retenti on and selectivity of analytes by RP - HPLC. Therefore, RI reproducibility is likely to be dependent on alterat ion of these chromatographic variables. As the mobile phase pH increases, silanol ionization (pKa ~3.5 - 4.5) and the negative charge on the column i ncreases [23] . Furthermore, the mobile phase pH may influence the ionization state of solutes. As the column temperature is increased, the analyte specific rate of exchange between the stationary and mobile phase is increased and RTs may be shortened. Furt hermore, RP - HPLC analyses are used to analyze a wide range of compounds by changing aqueous and organic mobil e phase components during gradient elution. As the mobile phase solvent strength increases with higher organic content, analyte specific interactio ns with the stationary phase may be reduced, solubility in the mobile phase increases, and the analyte advanc es through the column. 428 In this study, RI values were calculated for neutral acylsucrose metabolites, while GS - n - alkyl standards are comprised of i onizable amine and carboxylic acid functional groups that may change ionization state depending on mobile phase pH (pKa values of GS - n - alkyl standards were predicted from ChemDraw Software; 2.1, 3.6 and 9.4 respectively, see Figure 4.10). To evaluate the p H dependence of acylsucrose RI values when using ammonium formate aqueous mobile phase component (b uffer range ~2.8 - 4.8), the pH was altered from 2.5 - 4.0 by 0.3 units (Appendix Figure 4.18, other LC conditions were held constant as outlined in Section 4.2. 3.4). Figure 4.13A demonstrates the RI dependence on pH when applied to the group of S4:20:0 S. sin uata acylsucrose isomers (peaks #2 - 4). RI values exhibit a polynomial fit that rises sharply f all 16 acylsucrose metabolites followed a similar trend (Appendix Figure 4.19). The increase in R I va lues at pH>3 is due to earlier elution of the standard series (Figure 4.14A) relative to the acylsucrose metabolites. These observations are attributed t o increased ionization of the residues of the GS - n - alkyl standard series as the mobile phase pH is incr eased, producing more negatively charged and water - soluble forms. These results show that mobile phase pH is an important consideration when attempting t o reproduce RI values. Consequently, RI values were generated at pH 2.8 where the retention of the homo logous standard series exhibited less pH dependence. 429 Figure 4. 13 . RI dependence of the group of S4:20:0 S. sinuata acylsucros e isomers when chromatographic conditions are altered using column SAE - A. Filled markers are equiva lent analyses measured on separate dates. Standard deviations were too small to display error bars (see Appendix Figures 4.19, 4.21, 4.23) (A) Aqueous mobile phase pH 2.5 - 4.0, column temp. 50°C, gradient slope 1% acetonitrile · min - 1 (B) Column temperature 30 - 60 °C, aqueous mobile phase pH 2.8, gradient slope 1% acetonitrile · min - 1 (C) Linear gradient slope 1, 10/9, 5/4, 10/7, 5/3, and 2% acetonitrile · min - 1 , aqueo us mobile phase pH 2.8, column temp. 50°C. Examination of the GS - n - alkyl standards mobile phase pH depe ndence also provides useful information for evaluating symptoms of reduced column performance. Figure 4.14A - B compares the elution profile of the standar d series using columns SAE - A and SAE - C. For column SAE - A, earlier RTs wer e observed for GS - n - alkyl stan dards as the mobile phase pH was increased (Figure 4.14A). However, column SAE - C showed excessive peak broadening/tailing and later RTs for GS - n - alkyl st andards that was exacerbated by increased pH (Figure 4.14B). Surprisingly , as the Solvent A pH was incr eased to 4.0, the last standard to elute as a recognizable peak was GS - 22, while standards GS - 23 and GS - 24 no longer eluted as resolved peaks. These resu lts further illustrate reduced performance for column SAE - C, attributed t o greater roles of underivatiz ed silanol groups. 430 Figure 4. 14 . GS - n - alkyl pH dependence comparisons (XICs of common fragment ion m/z 162.022 by ESI+) using columns: (A) SAE - A and (B) SAE - C. To evaluate the dependence of acylsucrose RI values on column tempera ture, the column temperature was varied from 30 - 60°C, in steps of 5°C (Appendix Figure 4.20, other LC conditions were held constant as outlined in Sectio n 4.2.3.4). Figure 4.13B shows tha t RI values display unique U - shaped (concave up) behavior for the gro up of S4:20:0 S. sinuata acylsucrose isomers (peaks #2 - 4), reaching minima between ~45 - - specific depe ndence on column temperature. Inte restingly, each of the 16 acylsucrose metabolites in this study displ ayed unique polynomial t emperature dependence (Appendix Figure 4.21), suggesting two or more temperature - dependent retention processes (mixed retention modes) influence acylsucrose select ivity (and their RI values), such as hydrophobic and silanophilic int eractions and changes in conformation of acylsucroses interacting at the surface of the stationary phase [29] . We expect the pKa of ionizable groups on GS - n - alkyl standards to be lowered by increased temperature producing more negatively charged and water - soluble forms [30] . Howe ver, this 431 does not account for the reversal of RI values with temperature for some acylsucroses, but not others. For instance, all tetra - esters exhibited U - shape b ehavior reaching minima between 40 - - and hexa - esters displayed monotonic dec reases as temperature increased (Appendix Figure 4.21) . The temperature dependence behavior of RI values offers a novel approach for metabolite annotation that ou ght to be further explored. Table 4.3 summarizes the pH and temperatu re - dependence of RI valu es for the 16 acylsucrose metabolites in this study by applying a quadratic fit regression of experimental retention The pH and temperature - dependence illustrate how much (or how little) RI ch ange occurs with varied conditions and different compounds. For instance, a 10°C change in column temperature (column temperature at 40°C) drives relatively small differences in RI for th e tetra - esters, but were somewhat larger changes for penta - and hexa - esters. For most RP - HPLC analyses of complex mixtures, gradients are preferred over isocratic elution because they offer superior peak shapes, peak capacity and ability to resolve anal ytes that have varied partition behavior within a reasonable time frame. To explore the gradient dependence of acylsucrose RI values, the gradient slope was varied using values of 1, 10/9, 5/4, 10/7, 5/3, and 2% acetonitrile·min - 1 (Appendix Figure 4.22, ot her LC conditions were held constant as outlined in Section 4.2.3.4). Fi gure 4.13C demonstrates RI gradient dependence for the group of S4:20:0 S. sinuata acylsucrose isomers (peaks #2 - 4). RI values for the set of isomers rise with increasing gradient slop e. However, this feature is gradually transformed from a monotonic incre ase in RI, to an upside - down U - shape (concave down), to a monotonic decrease in the order of later eluting acylsucroses (Appendix Figure 4.23). These results demonstrate the gradient d ependence of RI values, and are in accordance with investigations of gra dient elution that have demonstrated the rate of change of analyte retention with gradient elution is solute - dependent [31] . 432 Table 4. 3 . Coefficients for calculating acylsucrose retention index values as a function of aqueous mobile phase pH and column temp erature using solvent A = 10 mM aqueous ammonium formate adjusted to pH with formic acid and solvent B = acetonitrile, 1% acetonitrile/ min gradient. Column = Ascentis Express C18, 2.1 x 100 mm. Solvent gradient slope at 1% acetonit rile/min. Coefficients pH dependence RI (323K) = A(pH) 2 +B(pH)+C Temperature dependence RI (T, K) /RI (323K) = D(T) 2 +E(T)+F Acylsugar # Number of acyl groups Acylsugar annotation A B C Calculated RI (323K) at pH 2.8, D E F Calculated RI at pH (313K) RI a 1 4 Ss_S4:19:0(3,5,5,6) 53.327 - 290.82 1841.4 1445.2 1.303E - 05 - 8.139E - 03 2.269 1442.3 - 2.8 2 4 Ss_S4:19:0(2,5,6,6) 53.947 - 295.09 1853.3 1450.0 7.900E - 06 - 5.018E - 03 1.797 1450.4 0.5 3 5 Ss_S5:20:0(2,2,5,5,6) 53.654 - 291.17 1880.6 1486.0 1.426E - 05 - 9.492E - 03 2.578 1491.9 5.9 4 4 Ss_S4:20:0(4,5,5,6) 55.568 - 303.39 1942.0 1528.2 9.412E - 06 - 5.971E - 03 1.947 1528.4 0.2 5 4 Ss_S4:20:0(3,5,6,6) 56.279 - 307.92 1968.2 1547.3 1.022E - 05 - 6.544E - 03 2.047 1547.2 - 0.1 6 4 Ss_S4:20:0(2,6,6,6) 56.872 - 311.83 1981.0 1553.8 8.296E - 06 - 5.399E - 03 1.878 1555.0 1.3 7 5 Ss_S5:22:1(2,5,5,5,5 T ) 55.879 - 303.14 1992.9 1582.2 1.640E - 05 - 1.075E - 02 2.762 1588.3 6.1 8 5 Ss_ S5:21:0(2,2,5,6,6) 56.429 - 307.03 2006.3 1589.0 1.372E - 05 - 9.215E - 03 2.545 1596.6 7.5 9 4 Ss_S4:21:0(5,5,5,6) 58.067 - 317.77 2046.5 1612.0 1.073E - 05 - 6.889E - 03 2.105 1611.9 - 0.1 10 5 Ss_S5:25:4(2,5,5,5,8 P ) 57.311 - 311.00 2055.8 1634.3 1.561E - 05 - 1.035E - 0 2 2.715 1641.9 7.6 11 5 Ss_S5:22:0(2,2,6,6,6) 59.123 - 322.21 2128.9 1690.2 1.262E - 05 - 8.535E - 03 2.441 1700.1 9.8 12 4 Ss_S4:22:0(5,5,6,6) 60.358 - 330.68 2159.6 1706.9 1.071E - 05 - 6.904E - 03 2.113 1709.1 2.2 13 5 Ss_S5:23:0(2,5,5,5,6) 59.992 - 325.95 2217.1 1774.8 1.336E - 05 - 8.882E - 03 2.476 1783.2 8.4 14 4 Ss_S4:23:0(5,6,6,6) 62.015 - 339.55 2267.5 1803.0 1.026E - 05 - 6.657E - 03 2.080 1805.6 2.7 15 5 Ss_S5:24:0(2,5,5,6,6) 62.097 - 338.11 2323.7 1863.8 1.172E - 05 - 7.774E - 03 2.288 1869.2 5.3 16 6 Ss_S6:25:0(2,2,5 ,5,5,6) 62.301 - 338.22 2358.1 1899.5 1.168E - 05 - 7.910E - 03 2.338 1911.6 12.1 a 433 4.3.7 RI dependence with methanol organic component While acetonitrile is a widely used RP - HPLC organic solvent component, other solvents such as methanol may provide alternatives in retention and selectivity. Figure 4.15A - B demonstrates LC/MS analysis of the S. sinuata extract using methanol organic compon ent instead of acetonitrile (Appendix Figure 4.24, other LC conditions were held constant as outlined in Section 4.2.3.4). Once mor e, the GS - n - alkyl standard series encompasses a wide RP - UHPLC retention range (Figure 4.15B). In contrast to the mostly linea r elution profile observed for acetonitrile (Figure 4.2), GS - n - alkyl standards showed a sigmoidal relationship between chain length and retention time, with large gaps in RT for earlier eluting members of the standard series (Figure 4.15C). The gaps in RT increased to an inflection point near GS - 7, after which RT density increased as the standard series alkyl chain length increased. R I values calculated for the set of 16 acylsucrose metabolites had SD in RT of approximately ±0.2 mins, or ~0.25% RSD, while t he SD of RI values was approximately ±1.4, or ~0.1% RSD (Table 4.4). Figure 4. 15 . LC/MS of S. sinuata ac ylsucrose sample analyzed by ESI+ mode with methanol organic component. (A) BPI chromatogram of S. sinuata acylsucrose extract sample. (B) XIC of common fragment ion m/z 162.022 (C 5 H 8 NO 3 S + ) generated at 20 V collision energy. (C) RT as a function of RI val ue. 434 RTs and selectivity of the 16 S. sinuata acylsucroses showed significant differences compared to aceto nitrile. Acylsucrose RTs were shifted longer by 13.7 - 20.5 mins using methanol (Table 4.3). Additionally, the relative elution order of several acylsuc roses had changed. Although the acylsucrose isomers in this study did not change elution order, several ac ylsucroses that had different numbers of ester groups changed elution order. Interestingly, the differences in retention and selectivity between the t wo organic components offers yet another approach for RI annotation. Relative to acetonitrile measurements , RI values were shifted to lower values that ranged from 228 - 445 difference (Table 4.4). 435 Table 4. 4 . LC/MS retention time a nd retention index values for 16 S. sinuata acylsucrose metabolites with structural identifications using methanol organic component ( other LC conditions were held constant as outlined in Section 4.2.3.4 ) . Analyses were performed on an Ascentis Express C18 column in triplicate (two by ESI - and one by ESI+) on four separate dates (n=12 total - 4) shown in Figure 4.3. The same ions listed in Table 4.2 were used for de tection. Acylsucrose numbering is in order of elution by acetonitr ile using methanol organic modifier (Table 4.2) Acylsucrose # Acylsucrose ID Avg Ret. Time ± SD (min) Ret. Time RSD (%) Methanol RI Value ± SD RI Value RSD (%) 1 S4:19 :0( 3,5,5,6) 70.13 ± 0. 17 0.25 1202 ± 1.2 0.10 2 S4:19 :0( 2,5,6,6) 70.87 ± 0.18 0.25 1223 ± 1.2 0.10 3 S5:20 :0( 2,2,5,5,6) 69.50 ± 0.17 0.25 1186 ± 1.0 0.08 4 S4:20 :0( 4,5,5,6) 72.31 ± 0.17 0.24 1263 ± 1.2 0.10 5 S4:20 :0( 3,5,6,6) 73.17 ± 0.18 0.24 1287 ± 1.2 0.09 6 S4:20 :0( 2,6 ,6,6) 73.82 ± 0.18 0.25 1306 ± 1.7 0.13 7 S5:22 : 1 ( 2,5,5,5,5 T ) 71.36 ± 0.18 0.25 1236 ± 1.2 0.10 8 S5:21 :0( 2,2,5,6,6) 72.50 ± 0.18 0.24 1268 ± 1.2 0.09 9 S4:21 :0( 5,5,5,6) 74.42 ± 0.19 0.26 1325 ± 1.4 0.11 10 S5:25 : 4 ( 2,5,5,5,8 P ) 73.77 ± 0.19 0.25 1305 ± 1.5 0.11 11 S5:22 :0( 2,2,6,6,6) 75.22 ± 0.19 0.25 1350 ± 1.4 0.10 12 S4:22 :0( 5,5,6,6) 76.96 ± 0.20 0.26 1405 ± 1.6 0.11 13 S5:23 :0( 2,5,5,5,6) 76.92 ± 0.20 0.26 1404 ± 1.6 0.11 14 S4:23 :0( 5,6,6,6) 79.18 ± 0.20 0.25 1483 ± 1.5 0.10 15 S5:24 :0( 2,5,5,6,6) 79.12 ± 0.20 0.26 1481 ± 1.6 0.11 16 S6:25 :0( 2,2,5,5,5,6) 78.40 ± 0.20 0.26 1456 ± 1.6 0.11 Acylsucrose RI dependence on chromatographic parameters was also examined by altering mobile phase pH, column temperature and LC gradient. Figure 4.16A - C demonst rates their RI dependence when applied to the group of S4:20:0 S. sinuata isomers. RI values di splayed polynomial U - shaped pH dependence with minima between pH 3.1 - 3.4 (Figure 4.16A). All 16 S. sinuata acylsucroses followed a similar trend (Appendix Figure 4.25). Unlike the RI determinations for varied column temperatures using acetonitrile, which d isplayed polynomial dependence, methanol RI values for all 16 S. sinuata acylsucroses exhibited linear relationship between column temperature and RI (Figure 4.1 6C), varying in slopes and y - intercept values (Appendix Figure 4.26). Finally, RI values declin ed with increasing gradient slope for the set of 16 acylsucroses (Figure 4.16C, Appendix Figure 4.27). 436 Figure 4. 16 . Retention index de pendence of the group of S4:20:0 S. sinuata acylsucrose isomers when chromatographic conditions were altered using column SAE - A and methanol was the organic modifier. Filled markers are equivalent analyses measured on separate dates. Standard deviations we re too small to display error bars (see Appendix Figures 4.25 - 27 ) (A) Aqueous mobile phase varied from pH 2.5 - 4.0, column temp. 50°C, gradient slope 1% methanol · min - 1 (B) Column temperature varied from 30 - 60°C, aqueous mobile phase pH 2.8, gradient slope 1 % methanol · min - 1 (C) Linear gradient slope 1, 10/9, 5/4, 10/7, 5 /3, and 2% methanol · min - 1 , aqueous mobile phase pH 2.8, column temp. 50°C. 4.3.8 GS - n - alkyl standards further applications In addition to retention indexing and column performance evaluations, GS - n - alkyl standards offer additional LC/MS applications. The s tandards may be used as mass axis calibration standards, or to examine LC/MS system performance and instrument sensitivity. Furthermore, RI values may be combined with MS/MS spectral databases to improve confidence in metabolite annotations, particularly w hen MS/MS spectra fail to distinguish isomeric compounds. Interestingly, MS/MS spectra of the standards exhibit common fragment ions that differ in relative parent and product ion abundances a ccording to their alkyl chain length (Appendix Figures 4.28 - 33). For example, Figure 4.17A displays ESI - mode MS/MS spectra of standards GS - 5, - 10, - 15 and - 20 using a 10 - 40 V CID ramp (generated using Waters SONAR data acquisition platform, 10 Da bins wid ths, 0.5 s acquisition time). As the GS - n - alkyl chain length is increased, the product ion abundances are increased relative to the [M - H] - ion. Such relationship runs in contrast to the center - of - mass collision energies, which dictate that the maximum amou nt of energy deposited by a single ion - molecule collision should decrease as the precursor ion mass increases [32] . An illustration of 437 the relationship between ion abundances and chain length is presented in Figure 4.17B, which shows an increasing relation ship (linear regression R 2 = 0.9626) when the ratio of common product ion [C 10 H 12 N 3 O 5 ] - at m/z 254.078 and [M - H] - ion abundances are plotted as a function of GS - n - alkyl chain length (Appendix Table 4.5). These results suggest that the standard set could op erate as a fragmentation e nergy index for standardizing collision energy conditions or normalizing MS/MS fragment ion abundances generated across instruments (comparable to the robust 70 eV electron ionization spectral libraries used by GC/MS) [33, 34] . Co mparison of the ratio of c ommon product ion [C 5 H 8 NO 3 S] + at m/z 162.022 and [M+H] + ion abundances as a function of GS - n - alkyl chain length also showed an increasing relationship with chain length (Appendix Table 4.6 and Figure 4.34). The observation of incr eased relative product ion abundance is attributed to greater collision cross - section for longer chain lengths and higher probability for ion - molecule collision in the collision cell. As the GS - n - alkyl alkyl chain length increases so does the collisional c ross section. The potentia l predictability of GS - n - alkyl collisional cross sections also suggests that this group of standards may be suitable as ion mobility reference standards [35] . 438 Figure 4. 17 . MS/MS product ion analysis of G S - n - alkyl standards. (A) Selected MS/MS product ion spectra of [M - H] - ions. (B) Ratio of common fragment [C 10 H 12 N 3 O 5 ] - = 254.078 (labeled by an asterisk in Figure 4.17A) and [M - H] - ion abundances as a function of GS - n - alkyl chain length. 439 4.4 Conclusions GS - n - alkyl retention indexing standards offer an increased level of confidence for metabolite identification and discovery. The standards encompass a wide RP - UHPLC retention range, are ionized by ESI in positive - and negative - ion modes and have characteris tic fragment (product) ions that allow for their recognition using narrow window XICs for a single m/z value of a given polarity, simplifying reporting of standard RTs when added directly to a sample with minimal interference by sample constituents. This R I ap proach shows capacity for reducing day - to - day and/or sample batch - to - batch chromatographic variability and displays improved capacity for archiving retention data across laboratories. In this study, RI values were calculated for acylsucrose metabolite s. H owever, this RI approach is applicable for other analytes as well. The application of retention indexing across laboratories is highly - to - batch column variance. It is not appropriate to assume that all C18 columns will have the same selectivity. Thus, RI values generated from different column manufacturers and models may not be equivalent. Using the same manufacturer and model column is recommended for RI comparison s. H owever, even when the same manufacturer/model of LC column is used, it may be beneficial to evaluate GS - n - alkyl retention and peak shape parameters before attempting to draw accurate comparisons between RI values. In this regard, GS - n - alkyl standard LC /MS analysis shows promise for evaluating RP - HPLC column performance, batch - to - batch column reproducibility and column lifetime. This report presents a thorough investigation of RI dependence by gradient elution using GS - n - alkyl standards while altering se vera l important chromatographic experiment variables. Mobile phase pH, column temperature, LC gradients and organic solvent component all influenced RI values. We acknowledge that there are other factors that will influence RI values, including mobile phas e io nic strength, buffer composition, mobile phase components, column length, flow rates and more. The ability of the analyst to reproduce and document LC experimental variables as accurately as possible will improve RI reproducibility. While RI values are not accurate enough to provide unambiguous determination of 440 structures on their own, they may act as refinement filter for comprehensive structural elucidation workflows, dereplicating metabolite discovery. Lastly, we hypothesize that GS - n - alkyl standards hav e applications that are not limited to retention indexing, such as fragmentation energy index standards and as ion mobility reference standards. 441 APPEN DIX 442 Figure 4. 18 . LC/MS chromatograms showing GS - n - alkyl pH dependence (ace tonitrile organic mobile phase component, aqueous component 10 mM ammonium hydroxide adjusted with formic acid, column temperature 50°C, gradient slope 1% acetontrile· min - 1 , column SAE - A) (A) XICs of common fragment ion m/z 162.022 by ESI+. (B) RT as a function of RI value. 443 Figure 4. 19 . RI value pH dependence of 16 S. sinuata acylsucrose metabolites (acetonitrile organic mobile phase component, aqueous component 10 mM ammonium hydroxide adjusted with formic acid, column temper ature 50°C, gradient slope 1% acetontrile · min - 1 , column SAE - A). 444 Figure 4. 20 . LC/MS chromatograms showing GS - n - alkyl temperature dependence (acetonitrile organic mobile phase component, aqueous component 10 mM ammonium hydroxide adjusted to pH 2.8 with formic acid, gradient slope 1% acetontrile·min - 1 , column SAE - A). (A) XICs of common fragment ion m/z 162.022 by ESI+. (B) RT as a function of RI value. 445 Figure 4. 21 . RI value column temperature depende nce of 16 S. sinuata acylsucrose metabolites (acetonitrile organic mobile phase component, aq ueous component 10 mM ammonium hydroxide adjusted to pH 2.8 with formic acid, gradient slope 1% acetontrile · min - 1 , column SAE - A). 446 Figure 4. 22 . GS - n - alkyl gradient slope dependence (acetonitrile organic mobile phase component, aqueous component 10 mM ammonium hydroxide adjusted to pH 2.8 with formic acid, column temperature 50°C, column SAE - A). (A) XICs of common fragment ion m/z 254.078 b y ESI - . (B) RT as a function of RI value. 447 Figure 4. 23 . RI value gradient dependence of 16 S. sinuata acylsucrose metabolites (acetonitrile organic mobile phase component, aqueous component 10 mM ammonium hydroxide adjusted to pH 2.8 with formic acid, column temperature 50°C, gradient slope 1, 10/9, 5/4, 10/7, 5/3, and 2% acetontrile · min - 1 , column SAE - A). 448 Figure 4. 24 . GS - n - alkyl dependence with altered chromatographic conditions using methanol organi c component (column SAE - A). (A) RT as a function of RI value when pH is altered (aqueous component 10 mM ammonium hydroxide adjusted with formic acid, column temperature 50°C, gradient slope 1% acetontrile·min - 1 ). (B) RT as a function of RI value when colu mn temperature is altered (aque ous component 10 mM ammonium hydroxide adjusted to pH 2.8 with formic acid, gradient slope 1% acetontrile·min - 1 ). (C) RT as a function of RI value when the LC gradient is altered (aqueous component 10 mM ammonium hydroxide ad justed to pH 2.8 with formic ac id, column temperature 50°C, gradient slope 1, 10/9, 5/4, 10/7, 5/3, and 2% acetontrile·min - 1 ). 449 Figure 4. 25 . RI value pH dependence of 16 S. sinuata acylsucrose metabolites (methanol organic mobi le phase component, aqueous component 10 mM ammonium hydroxide adjusted with formic acid, column temperature 50°C, gradient slope 1% acetontrile · min - 1 , column SAE - A). 450 Figure 4. 26 . RI value column temperature dependence of 16 S . sinuata acylsucrose metabolites (methanol organic mobile phase component, aqueous component 10 mM ammonium hydroxide adjusted to pH 2.8 with formic acid, gradient slope 1% acetontrile · min - 1 , column SAE - A). 451 Figure 4. 27 . RI va lue gradient dependence of 16 S. sinuata acylsucrose metabolites (methanol organic mobile phase component, aqueous component 10 mM ammonium hydroxide adjusted to pH 2.8 with formic acid, column temperature 50° C, gradient slope 1, 10/9, 5/4, 10/7, 5/3, and 2% acetontrile · min - 1 , column SAE - A). 452 Figure 4. 28 . ESI negative mode MS/MS spectra of GS - n - alkyl standards [M - H] - (generated using Waters SONAR data acquisition platform, 10 - 40 V CID ramp, 0.5 s acquisition time, 10 Da bin width s, spectra processed using Waters MS E Data Viewer). 453 Figure 4. 29 . ESI negative mode MS/MS spectra of GS - n - alkyl sta ndards [M - H] - (generated using Waters SONAR data acquisition platform, 10 - 40 V CID ramp, 0.5 s acquisition time, 10 Da bin widths, spectra processed using Waters MS E Data Viewer). 454 Figure 4. 30 . ESI negative mode MS/MS spectra of GS - n - alkyl standards [M - H] - (generated using Waters SONAR data acquisition platform, 10 - 40 V CID ramp, 0.5 s acquisi tion time, 10 Da bin widths, spectra processed using Waters MS E Data Viewer). 455 Figure 4. 31 . ESI positive mode MS/MS spectra of GS - n - alkyl standards [M+H] + (generated using Waters SONAR data acquisition platform, 10 - 40 V CID ramp, 0.5 s acquisition time, 10 Da bin widths, spectra processed using Waters MS E Data Viewer) . 456 Figure 4. 32 . ESI positive mode MS/MS spectra of GS - n - alkyl standards [M+H] + (generated using Waters SONAR data acquisition platform, 10 - 4 0 V CID ramp, 0.5 s acquisition time, 10 Da bin widths, spectra processed using Waters MS E Data Viewer). 457 Figure 4. 33 . ESI positive mode MS/MS spectra of GS - n - alkyl st andards [M+H] + (generated using Waters SONAR data acquisition platform, 10 - 40 V CID ramp, 0.5 s acquisition time, 10 Da bin widths, spectra processed using Waters MS E Data Viewer). 458 Table 4. 5 . MS/MS product ion analysis of GS - n - alkyl standards. Ratio of common fragment [C 10 H 12 N 3 O 5 ] - = 254.078 a nd [M - H] - ion intensities measured using 10 - 40 V CID ramp (0.5 s acquisition time). GS - n - alkyl Carbon # [M - H] Peak Abundance [C H N O ] = 254.078 Peak Abundance A [C H N O ] / A [M - H] 1 9310 1599 0.172 2 7534 2315 0.307 3 1809 650 0.359 4 6439 2660 0.413 5 24612 8149 0.331 6 54754 22600 0.413 7 79870 33296 0.417 8 66882 31786 0.475 9 60941 32356 0.531 10 82873 45326 0.547 11 116219 55879 0.481 12 125116 76766 0.614 13 103254 78101 0.756 14 97177 79983 0.823 15 107991 98996 0.917 16 101188 89185 0.881 17 95678 89548 0.936 18 89609 90648 1.012 19 84461 79556 0.942 20 66093 64989 0.983 21 52703 57815 1.097 22 52975 58557 1.105 23 41919 46109 1.100 24 43747 53183 1.216 459 Table 4. 6 . MS/MS product ion anal ysis of GS - n - alkyl standards. Ratio of common fragment [C 5 H 8 NO 3 S] + = 162.022 and [M+H] + ion intensities measured using 10 - 40 V CID ramp (0.5 s acquisition time). GS - n - alkyl Carbon # [M+H] Peak Abundance [C H NO S] = 162.022 Peak Abundance A [C H NO S] / A [M+H] 1 6021 ND --- 2 5788 ND --- 3 1114 ND --- 4 2812 1464 0.521 5 7400 5434 0.734 6 9971 16624 1.667 7 9401 40365 4.294 8 8550 47700 5.579 9 5558 70572 12.70 10 5374 90319 16.81 11 5979 127259 21.28 12 4195 143271 34.15 13 3697 188873 51 .09 14 4420 220424 49.87 15 3053 259937 85.14 16 3214 233103 72.53 17 2964 231604 78.14 18 1964 206838 105.3 19 1807 200967 111.2 20 1786 188049 105.3 21 1375 161634 117.6 22 1299 176077 135.5 23 1143 188927 165.3 24 922 185657 201.4 460 Figur e 4. 34 . MS/MS product ion analysis of GS - n - alkyl standards. Ratio of common fragment [C 5 H 8 NO 3 S] + = 162.022 and [M+H] + ion intensities as a function of GS - n - alkyl chain length (measured using 10 - 40 V CID ramp, 0.5 s acquisition time ) 461 REFERENCES 462 REFERENCES 1. Kind, T. and O. Fiehn, Advances in structure elucidation of small molecules using mass spectrometry. Bioanal Rev, 2010. 2 (1 - 4): p. 23 - 60. 2. Forcisi, S., et al., Liquid chromatography - mass spectrometry in metabolomics research: mass analyzers in ultra high pressure liquid chromatography coupling. J Chromatogr A, 2013. 1292 : p. 51 - 65. 3. Xian, F., C.L. He ndrickson, and A.G. Marshall, High resolution mass spectrometry. Anal Chem, 2012. 84 (2): p. 708 - 19. 4. Last, R.L., A.D. Jones, a nd Y. Shachar - Hill, Towards the plant metabolome and beyond. Nat Rev Mol Cell Biol, 2007. 8 (2): p. 167 - 74. 5. Nguyen, D.D., et a l., MS/MS networking guided analysis of molecule and gene cluster families. Proc Natl Acad Sci U S A, 2013. 110 (28): p. E2611 - 20 . 6. Johnson, S.R. and B.M. Lange, Open - access metabolomics databases for natural product research: present capabilities and fut ure potential. Front Bioeng Biotechnol, 2015. 3 : p. 22. 7. Vinaixa, M., et al., Mass spectral databases for LC/MS - and GC/MS - bas ed metabolomics: State of the field and future prospects. Trac - Trends in Analytical Chemistry, 2016. 78 : p. 23 - 35. 8. Kováts, E. , Gas - chromatographische Charakterisierung organischer Verbindungen. Teil 1: Retentionsindices aliphatischer Halogenide, Alkohol e, Aldehyde und Ketone. Helvetica Chimica Acta, 1958. 41 (7): p. 1915 - 1932. 9. Baker, J.K. and C.Y. Ma, Retention index scale for liquid - liquid chromatography. Journal of Chromatography, 1979. 169 (Feb): p. 107 - 115. 10. Smith, R.M., Alkylarylketones as a ret ention index scale in liquid - chromatography. Journal of Chromatography, 1982. 236 (2): p. 313 - 320. 11. Bogusz, M. and R. Aderjan, Improved standardization in seversed - phase high - performance liquid - chromatography using 1 - nitroalkanes as a retention index sca le. Journal of Chromatography, 1988. 435 (1): p. 43 - 53. 12. Magee, M.H., et al., "Measure Your Gradient": a new way to measure gr adients in high performance liquid chromatography by mass spectrometric or absorbance detection. J Chromatogr A, 2014. 1369 : p. 73 - 82. 13. Abate - Pella, D., et al., Retention projection enables accurate calculation of liquid chromatographic retention times across labs and methods. J Chromatogr A, 2015. 1412 (Supplement C): p. 43 - 51. 14. Quilliam, M., et al., A new retention index sys tem for liquid chromatography - mass spectrometry , in 42nd International Symposium on High Performance Liquid Pha se Separations and Related Techniques (HPLC2015) . 2015: Geneva, Switzerland. 463 15. Quilliam, M., Retention index standards for liquid chromatograp hy , U.S. Patent, Editor. 2017, National Research Council of Canada . 16. Ghosh, B. and A.D. Jones, Profiling, ch aracterization, and analysis of natural and synthetic acylsugars (sugar esters). Analytical Methods, 2017. 9 (6): p. 892 - 905. 17. Treutler, H., e t al., Discovering regulated metabolite families in untargeted metabolomics studies. Anal Chem, 2016. 88 (16): p. 8082 - 90. 18. Ghosh, B., T.C. Westbrook, and A.D. Jones, Comparative structural profiling of trichome specialized metabolites in tomato (Solanum lycopersicum) and S. habrochaites: acylsugar profiles revealed by UHPLC/MS and NMR. Metabolomics, 2014. 10 (3): p. 496 - 507. 19. Liu, X.X., et al., Profiling, isolation and structure elucidation of specialized acylsucrose metabolites accumulating in trichom es of Petunia species. Metabolomics, 2017. 13 (7): p. 85. 20. Poole, C.F., Chromatographic test methods for chara cterizing alkylsiloxane - bonded silica columns for reversed - phase liquid chromatography. J Chromatogr B Analyt Technol Biomed Life Sci, 2018. 109 2 : p. 207 - 219. 21. Nawrocki, J., The silanol group and its role in liquid chromatography. Journal of Chromatogra phy A, 1997. 779 (1 - 2): p. 29 - 71. 22. Dolan, J.W., Detective work part 1: simplify the choices. LCGC Europe, 2015. 28 (11): p. 612 - 615. 23. Snyder , L.R., J.W. Dolan, and P.W. Carr, The hydrophobic - subtraction model of reversed - phase column selectivity. J Chr omatogr A, 2004. 1060 (1 - 2): p. 77 - 116. 24. Dolan, J.W., Detective work, part 4: chemical problems with the column: chemical attack. LCGC Europe, 2016. 29 (2): p. 82 - 84. 25. Vince, R., S. Daluge, and W.B. Wadd, Studies on the inhibition of glyoxalase I by S - substituted glutathiones. J Med Chem, 1971. 14 (5): p. 402 - 4. 26. Kostiainen, R. and T.J. Kauppila, Effect of eluent on the ionization process in liquid chromatography - mass spectrometry. J Chromatogr A, 2009. 1216 (4): p. 685 - 99. 27. Kirkland, J.J., M.A. van Straten, and H.A. Claessens, High pH mobile - phase eEffects on silica - based reversed - phase high - performance liquid - chromatographic columns. Jour nal of Chromatography A, 1995. 691 (1 - 2): p. 3 - 19. 28. Claessens, H.A. and M.A. van Straten, Review on the chemic al and thermal stability of stationary phases for reversed - phase liquid chromatography. J Chromatogr A, 2004. 1060 (1 - 2): p. 23 - 41. 29. Li, J., E ffect of temperature on selectivity in reversed - phase liquid chromatography, a thermodynamic analysis. Analytica Chimica Acta, 1998. 369 (1): p. 21 - 37. 30. Dolan, J.W., Temperature selectivity in reversed - phase high performance liquid chromatography. J Chro matogr A, 2002. 965 (1 - 2): p. 195 - 205. 31. Shoenmakers, P.J., H.A.H. Billiet, and L. De Galan, Influenc e of organic modifiers on the rentention behaviour in reversed - phase liquid chromatography and its consequences for gradient elution. Journal of Chromatog raphy A, 1979. 185 : p. 179 - 195. 464 32. Sleno, L. and D.A. Volmer, Ion activation methods for tandem mass spectrometry. J Mass Spectrom, 2004. 39 (10): p. 1091 - 112. 33. Bristow, A.W., et al., Reproducible product - ion tandem mass spectra on various liquid chroma tography/mass spectrometry instruments for the development of spectral libraries. Rapid Commun Mass Spectrom, 2004. 18 (13): p. 1447 - 54. 34. Hopley, C., et al., Towards a universal product ion mass spectral library - reproducibility of product ion spectra a cross eleven different mass spectrometers. Rapid Commun Mass Spectrom, 2008. 22 (1 2): p. 1779 - 86. 35. Paglia, G., et al., Ion mobility derived collision cross sections to support metabolomics applications. Anal Chem, 2014. 86 (8): p. 3985 - 93. 465 Chapter 5: Clos ing Thoughts Plants produce a truly enormous array of structurally comple x and varied metabolites, and annotation and discovery of their chemical structures will continue to present formidable challenges. C omprehensive metabolite profiling requires multipl e chromatographic and mass spectrometric platforms to separate and detect their chemical complexity, while the important relationship between mass spectrometry and nuclear magnetic resonance (NMR) spectroscopy will continue to be central to de novo structu ral elucidation of metabolites. The analytical challenges of metabolomics are well - demonstrated in the case of acylsugars. In Chapter 2, more than 400 acylsucroses were manually annotated from LC/MS of Salpiglossis sinuata , while slightly more than ~ 300 p eaks were detected using automated processing by collaborator Dr. Gaurav Moghe. A point of significance is that these numbers are from a single plant genotype, at a specific stage of development, in a single set of environmental conditions, and from a sing le tissue type (trichomes). These results suggest that the acylsugar meta bolome is far more complex than imagined. Even by manual annotation, no matter what we do it seems we are still underestimating the range of acylsugars that can be produced by a singl e plant. For instance, acylsucrose mono - and di - esters must be made to fo rm tri - , tetra - , penta - esters, and so on. However, the early pathway precursors were not detected and are likely at very low concentrations, but must be intermediates in the formation of more heavily acylated forms. Furthermore, column performance comparis ons in Chapter 4 showed that reversed - phase C18 columns from different manufacturers differ in acylsucrose retention selectivity, revealing coeluting isomers that we would never know existed (four isomers resolved on one column, and seven isomers for the s ame extract analyzed on another column). Plant metabolic complexity is even more multifaceted when we consider that individual acyltransferase enzymes can use several different acyl - CoAs as substrates, and that mutations can alter their sugar acyl accepto r substrates . To our knowledge, there are 25 documented acylsugar ester groups (Chapter 1), including two new ester groups (tiglyl and phenylacetyl) outlined in Chapter 2. With eight positions 466 available on sucrose for acylation, there are billions of possi ble combinations for the compound class of acylsucroses alone. Many of these metabolites are expected to exist as isomeric forms and/or may be of minor abundance. Furthermore, plants have capacity to make far more acyl - CoA groups, so it would not be surpr ising to find new acylsugar chemistr ies as new plants are investigated. For instance, where previous reports have mostly identified acylglucose s and acylsucrose s from the Solanaceae, Chapter 3 showed a group of acylated myo - inositols and myo - inositol glyco sides discovered in S olanum quitoense , revealing the carbohydrate core can also vary. There are certainly many other carbohydrates that exist in nature that could be acylated too. Mor eover, those sugar groups also have potential to exist as isomers, differ ing by stereochemical configuration and linkage positions. Therefore, the challenges to characterize acylsugars remain to be enormous. In this regard, it is not feasible to purify and perform NMR on all metabolites, and no combination of current analytical technologies could be used to resolve them all. Therefore, the chemical complexity of nature drives a need for continued improvements in analytical technologies, particularly for tho se that can be used to resolve and identify isomeric metabolites. The fur ther advancement of pre - fractionation, two - dimensional LC and ion mobility mass spectrometry separations offers opportunities to improve analytical resolving capacity to advance futur e discovery of isomeric metabolites. Though application of GS - n - alkyl retention indexing outlined in Chapter 4 does not improve instrumental resolution of metabolites, it improves confidence in annotation, identification and dereplication of discovery of s pecialized metabolites and offers a second level of information about metabolites, particularly for those isomeric metabolites that are not readily differentiated by their mass spectra but exhibit differential changes in retention behavior (e.g. retention index temperature - dependence). Liquid chromatography retention indexing has yet to find widespread use, and for retention indexing to become more commonly adopted, researchers must incorporate approaches that embrace the idea that more experimental variabl es influence liquid chromatographic retention than is the case for gas chromatography. The development of a computational framework that models the effects of experimental parameters on retention index values of individual compounds and/or classes has pote ntial to improve recognition of these effects and enhance 467 confi dent annotations of unknowns. As a follow - up to this work, we hope to make the GS - n - alkyl standards available to the community to encourage more metabolomics researchers to adopt retention indexing systems as additional parameters in open - access metabolite databases.