DEVELOPMENT OF STRUCTURAL LY DEFINED PLATFORMS FOR LONG -RANGE BIOLOGICAL ELECTRON TRANSFER By Jingcheng Huang A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Biochemistry and Molecular Biology - Doctor of Philosophy 2019 ABSTRACT DEVELOPMENT OF STRUCTURAL LY DEFINED PLATFORMS FOR LONG -RANGE BIOLOGICAL ELECTRON TRANSFER By Jingcheng Huang Electron transfer reactions are vital for life: they are the essential s teps in all the major biological energy conservation pathways and the rate of electron transfer sometime s determines the fate of energy flow. While the rates of electron transfer over 1 -2 nanometers in proteins can largely be described by well -known theori es, it is not well understood how these processes scale to microscopic distances, for example, micrometer l ength microbial nanowires. Electron transfer reaction s are known to be highly sensitive to the chemical properties of the electron carriers and distances between carriers, yet , this information is not available for naturally occurring microbial nanowires. On the other hand, microbial nanowires have inspired the d evelopment of novel biological conductive material s and bioelectronics, although these biomimicking materials would significantly benefit from a higher degree of structural definition, which would greatly improve rational redesign . This dissertation work presents two distinct approaches for arranging electron carriers (heme) into structurally defined array s that can facilitate electron transfer: 1) A crystalline lattice of small tetraheme cytochromes that form a well -defined, three -dimens ional network of closely spaced redox centers was used to demonstrate the multi -step electron hopping over a micrometer scale. 2) A heme attachment strategy was developed that allows one to introduce redox active cofactor hemes into non-heme -binding protei ns, while maintain ing the proteins™ original function. Adding hemes to a nanotube -forming self -assembling protein was used to demonstrate the potential of this strategy to form a structurally defined heme array. The first crystal approach provides detailed information about structure and electronic states which can be used as a platform for testing theories, while the second heme -attaching approach is an engineering platform that allows researchers to introduce redox properties into other well -studied prote ins with minimal effort . These two approaches, from two perspectives, lay the foundation of building structurally defined architectures for the understanding of microbial nanowires and the application of biological long -range electron transfer materials. Copyright by JINGCHENG HUANG 2019 v ACKNOWLEDGMENTS During my journey as a graduate student, I had a great fortune to have both Dr. Danny Ducat and Dr. Dave Kramer as my co -advisors. They have allowed me to enjoy the excitements of sci ence as well as trained me to become a scientist. I would like to thank them for all the supports during my Ph.D. study, countless creative ideas for my research, and patient guidance for my career. I would also like to thank members in my guidance commit tee: Dr. Bob Hausinger, Dr. Cheryl Kerfeld, Dr. Eric Hegg and Dr. Michael Feig for their constructive feedbacks as well as criticism. Their advises encouraged me to push myself harder towards a better scientist. I have had a great pleasure to work in Duca t Lab and with all the past and current members. I would like to thank everyone for fostering the friendly and cooperative environment . I also appreciated the opportunity to conduct my dissertation research in communities of Department of Biochemistry and Molecular Biology and Department of Energy Plant Research Laboratory , where I have had chances to interact with many wonderful individuals. I would like to acknowledge the staffs in BMB and PRL for their help during my graduate study. Lastly, my gratitude goes towards my family and friends. I would like to thank my parents and my family for encouraging me to study abroad and supporting me all the time. vi TABLE OF CONTENTS LIST OF TABLES .......................................................................................................... viii LIST OF FIGURES ..........................................................................................................ix LIST OF ALGORITHMS ..................................................................................................xi KEY TO ABBREVIATIONS ............................................................................................ xii CHAPTER 1: O VERVIEW OF SHORT -RANGE ELECTRON TRANSFER MECHANISMS AND EMERGING LONG -RANGE ELECTRON TRANSFER SYSTEMS ........................ 1 Introduction ....................................................................................................... 2 Mechanisms and modeling of short -range electron transfer ............................. 4 Non -adiabatic treatments in biological electron transfer ................................ 4 Electronic coupling factor .............................................................................. 6 Probability of reaching transition state ........................................................... 8 Marcus theory ................................................................................................ 9 Empirical rules for biological ET rates ......................................................... 12 Computer simulation of electron transfer reactions ..................................... 14 Long-range Electron transfer .......................................................................... 17 Long-range ET and ETp found in microbial extracellular nanowires ........... 18 Mechanisms of biological long -range ET ..................................................... 22 Challenges on biological long -range ET studies without defined structures 26 Composition and Stru cture Defined Platforms for long -range ET ................... 29 Cytochrome -based long -range ET platform ................................................. 30 REFERENCES ............................................................................................... 35 CHAPTER 2: MESOSCOPIC ELECTRON TRANSFER BY HOPPING IN A CRYSTAL NETWORK OF CYTOCHROMES ................................................................................. 44 Abstract ........................................................................................................... 45 Introduction ..................................................................................................... 46 Results ............................................................................................................ 50 Heme network and its energy profile in an STC crystal ............................... 50 Photoinduced cytochrome reduction and ET in STC crystals ...................... 53 ET in STC crystals is controlled by a process with high activation enthalpy 66 Discussion ...................................................................................................... 67 Methods .......................................................................................................... 73 Rationale of choosing STC crystal .............................................................. 73 STC expression and purification .................................................................. 74 STC Crystallization ...................................................................................... 74 Microscopy Imaging and Photoreduction setup ........................................... 75 Photoreduction conditions of STC crystals. ................................................. 77 Image analysis. ........................................................................................... 77 Estimation of photoreduction quantum efficiency ........................................ 78 vii Estimation of ket and D in STC crystals using empirical non -adiabatic electron transfer theory ............................................................................................ 78 Curve fitting ................................................................................................. 80 Kinetic Monte Carlo simulation .................................................................... 81 Single crystal spectroscopy ......................................................................... 82 Redox titrations ........................................................................................... 82 REFERENCES ............................................................................................... 85 CHAPTER 3: FUNCTIONALIZATION OF NANOTUBE -FORMING PROTEINS WITH COVALENTLY ATTACHED HEME ............................................................................... 91 Abstract ........................................................................................................... 92 Introduction ..................................................................................................... 92 Results ............................................................................................................ 96 Design of heme attachment sites on shell proteins ..................................... 96 Covalent heme attachment to shell proteins .............................................. 100 Shell -cytochromes form hexamer -like oligomers ....................................... 106 Higher -order structure assembly of shell -cytochromes ............................. 109 Discussion .................................................................................................... 112 Materials and Methods .................................................................................. 118 Identification of possible heme attachment sites for shell proteins based on secondary Structures of the CxxCH motif in natural cytochromes c. ........ 118 Strains and protein expression .................................................................. 118 Molecular mass determination by mass spectrometry ............................... 119 Transmission electron microscopy ............................................................ 120 Molecular dynamic of heme -attached shell proteins .................................. 121 UV-Vis spectroscopy ................................................................................. 121 Redox titrat ion ........................................................................................... 121 Electrophoresis .......................................................................................... 122 Dynamic light scattering ............................................................................ 122 Potassium cyanide titration ........................................................................ 123 REFE RENCES ............................................................................................. 127 CHAPTER 4: CONCLUSION AND PERSPECTIVE .................................................... 133 Overview ....................................................................................................... 134 Comparison of two platfo rms ........................................................................ 135 What makes a highly conductive and robust bionanowire? .......................... 139 Applications of biological conductive polymers ............................................. 144 REFERENCES ............................................................................................. 147 viii LIST OF TABLES Table 2.1 X -ray data collection and refinement statistics .............................................. 83 Table 3.1 List of structures used in Figure 3.1B .......................................................... 123 Table 3.2 Calculated and observed molecular weights of shell cytochromes .............. 126 ix LIST OF FIGURES Figure 1.1 Regimes of Electron Transfer reactions ......................................................... 4 Figure 1.2 Potential energy surface of reactant and product states as a function of the nuclear coordinate ........................................................................................................... 9 Figure 1.3 Potential energy surfaces ............................................................................. 12 Figure 1.4 Scheme of molecular dynamic simulation of an electron transfer reaction ... 17 Figure 1.5 Hypothetical models for the conductive nanowires ...................................... 26 Figure 1.6 Structures of heme b, heme c and heme c with bis -histidine ligands ........... 31 Figure 2.1 Heme network in STC crystals ..................................................................... 49 Figure 2.2 Structure and energy profile of STC crystals ................................................ 51 Figure 2.3 Redox titrations of STC crystals and STC in solution ................................... 53 Figure 2.4 Redistribution of injected electrons within STC crystals ............................... 55 Figure 2.5 Photoreduction of STC crystals .................................................................... 56 Figure 2.6 Microscopy for monochromic imaging and photoreduction .......................... 57 Figure 2.7 Oxidation kinetics of photoreduced STC crysta l ........................................... 59 Figure 2.8 Quantitative analysis of electron diffusion in STC crystals ........................... 60 Figure 2.9 Quantitative analysis of electron diffusion in along a-axis of a STC crystal .. 62 Figure 2.10 Random walk simulations of the impact of crystal defects that inhibit ET on electron diffusion rate .................................................................................................... 64 Figure 2.11 Oxidation of photoreduced STC in solution and STC crystals .................... 65 Figure 2.12 Structure overlap of interprotein heme contact between two adjacent proteins within the crystal as compared to intra -protein contacts within MtrF. ............... 71 Figure 3.1 Design of heme -functionalized BMC -H shell proteins .................................. 95 Figure 3.2 Nanotubes formed by the Y41A variant of MicH during cytosolic overexpression .............................................................................................................. 97 Figure 3.3 Molecular dynamics models of all monoheme shell -cytochromes ................ 99 x Figure 3.4 Modified shell proteins contain redox -active hemes ................................... 102 Figure 3.5 Protein mass spec trometry of purified shell -cytochromes .......................... 103 Figure 3.6 Equilibrium redox titration of representative shell -cytochromes monitored by UV-Vis spectrophotometry .......................................................................................... 104 Figure 3.7 Features of purified shell -cytochrome proteins containing two heme attachment sites .......................................................................................................... 105 Figure 3.8 Shell -cytochromes retain the capacity to oligomerize ................................ 107 Figure 3.9 Shell -cytochromes retain capacity to oligomerize into hexamers and higher -order assemblies ......................................................................................................... 108 Figure 3.10 Cyanide remediates shell -cytochrome aggregation ................................. 111 Figure 3.11 Ho -5815 with Y41A mutation and heme -tag ............................................. 112 Figure 4.1 Graphic overview ....................................................................................... 135 Figure 4.2 Graphic demonstration of a nanowire application ...................................... 146 xi LIST OF ALGORITHMS Equation 1.1 .................................................................................................................. 10 Equation 1.2 .................................................................................................................. 10 Equation 1.3 .................................................................................................................. 10 Equation 1.4 .................................................................................................................. 13 Equation 2.1 .................................................................................................................. 66 Equation 2.2 .................................................................................................................. 66 Equation 2.3 .................................................................................................................. 79 Equation 2.4 .................................................................................................................. 80 Equation 2.5 .................................................................................................................. 80 Equation 2.6 .................................................................................................................. 80 xii KEY TO ABBREVIATIONS BMC Bacterial microcompartment DLS Dynamic light scattering Ea Activation free energy Em' Redox midpoint potential ET Electron transfer ETC Electron transport chain ETp Electron transport G Gibbs free energy ket Electron transfer rate constant MD Molecular dynamics NADPH Reduced nicotinamide adenine dinucleotide phosphate NMR Nuclear magnetic resonance PDB Protein data bank ROS Reactive oxygen species SHE Standard hydrogen electrode STC Small tetraheme cytochrome Reorganization energy 1 CHAPTER 1 : OVERVIEW OF SHORT -RANGE ELECTRON TRANSFER MECHANIS MS AND EMERGING LONG -RANGE ELECTRON TRANSFER SYSTEMS 2 Introduction The importance of electron transfer to biolog ical processes was realized in the mid -20th century. In the 1940s, Szent -Gyorgyi suggested the transfer of electron energy is vital in the core metabolic pathways of respiration and photosynthesis (Szent -Gyorgyi , 1941). Band-gap theory was proposed to be the mechanism since the respiration component s appeared to be solid (found in the insoluble fractions of extracts) and photosynthesis appeared to happen on a complex with thousands of chlorophylls. In the 1960s, Chance and DeVault used pulsed laser s to demonstrate the oxidation of cytochrome in bacteria l photosynthe tic reaction center s is a quantum mechanical tunneling process (DeVault and Chance , 1966). In 1974, Hopfield suggested a thermally -activated model for tunneling and explained the temperature dependency observed in bacteria l photosynthetic reaction centers (Hopfield 1974) . By the 1980s, the involvement of electron transfer in many biochemical processes was becoming clear, and there was a burst of studies about electron transfer on a variety of model proteins, and the mechanism of biological electron transfer was demystified with well -developed theories (reviewed in (Gray and Winkler , 1996; Moser et al ., 1992)). Electron transfer in proteins usually occur s across short distance s (less than 2 nm) and can occur over a wide range of timescale s from picoseconds to milliseconds. Yet many cellular processes require that electrons be carried over much longer distances (micrometers). It is now well -established that soluble electron carriers are often used to transport electrons over longer distances by diffusion of the carrier. However, it has been argued that solid -state electron transfer might have some advantages (Malvankar and Lovley , 2014, 2012) , and new mechanisms for long -range electron transfer have been 3 elucidated within the past two decade s (Dalton et al. , 1990). In 2005, Reguera found that Geobacter can produce extracellular nanowires that can transfer electrons over micrometers in order to use the insoluble metal oxides as terminal electron acceptors (Reguera et al. , 2005). Since then, co nductive extracellular appendages were found in many microb es and these discoveries have brought about a renewed discussion of electron transfer mechanisms that can accommodate nanowires. In this chapter , I will review the well -established theory for shor t distance electron transfer, proposed mechanisms promoting electron transfer on nanowires , and discuss the challenges that the field faces in further characterizing long -distance electron transfer through nanowire -like biological systems . This foundationa l information will be used to contextualize my thesis research, reported in greater detail in the subsequent chapters. For the sake of clarity, I first define a number of specific terminologies for this writing . Historically, short -range electron transfer refer s to single -step electron tunneling between (nearly) contacting donor and acceptors (usually within 1 nm). In comparison, long -range electron transfer is a single -step electron tunneling between donor and acceptors that are separated by ions, water molecule s or protein residues. The molecules between donor and acceptor can support the tunneling of electrons , but they do not form intermediates with the electron localized on the separating molecules. In this dissertation, both historical short -range and long -range electron transfer will be referred as short -range electron transfer (short -range ET). I will use the term filong -range electron transfer fl (long -range ET) to refer to multiple steps of short -range ET or the transfer of electrons through band-gap theory (through a periodical array of identical redox centers). Long-range ET of this type is usually measured from tens of nanometers to millimeter -scale s (Figure 1.1) . 4 When the movement of electrons is driven by an external electric field, usually by electrodes, this motion is referred to electron transport (ETp), which stands in contrast to ET, in which the driving force of electrons is the potential ener gy of each redox cofactor. Figure 1. 1 Regimes of Electron Transfe r reactions Based on the total distances and the number of redox centers involved, ET can be divided into short -range or long -range (purple). Short -range ET can be further divide d into ad iabatic or non -adiabatic treatments depend ing on the strength of electronic coupling (yellow). For non -adiabatic ET, the electronic coupling is supported by through -space tunneling and/or superexchange (orange). The nuclear configuration should meet Franck -Condon principle during ET (pink). In long -range ET, the thermally -activated ET is closely related to the short -range ET , while the delocalized model is unique for long -range ET. Mechanisms and modeling of short -range electron transfer Non -adiabatic treatments in biological electron transfer An exact description of electron transfer (ET) reactions between molecules requires detailed information on the entire system, including the kinetic energies of all nuclei and electrons, the repulsiv e coulomb potential energies between nuclei and 5 between electrons , as well as the attractive interaction between nuclei and electrons. But for practical applications in chemical and biological systems, several approximations are often applied to reduce the complexity and deduce useful estimates (reviewed in (Moser et al. , 1992; Marcus and Sutin , 1985; May and Kühn , 2011)). First, due to the fact of the large mass difference between electrons and nuclei, the configuration of electrons can be considered to respond instantaneously relative to the much slower nuclear movement. Therefore, the potential energy between electrons and nuclei changes adiabatically and stays at a stationary state. It is reasoned that the electronic energy can be calculated as a function of nuclear coordinates separating the entire system into an electronic component and a nuclear component. This assu mption allows the calculation of an electron density distribution of molecules based on fixed (averaged) atomic coordinates, without considering information of nuclear vibrations: this is commonly known as the Born -Oppenheimer approximation . The overlap of the electron density profiles between the electron donor and acceptor molecules provides the electronic coupling for the transition of electronic state from reactant to the product. Similarly, electronic transitions happen on a much faster time scale tha n nuclear motion and thus nuclei are essentially static during ET reactions. According to the Franck ŒCondon principle , the transfer of an electron can only occur when one nuclear configuration transiently allows the reactant and product to have the same en ergy (Marcus , 1964). As the energy level of electron donor and acceptor molecules vibrates, the probability integral of finding configurations with overlapping vibrational energies is called the Franck -Condon facto r. Overall, the rate of ET can be calculat ed by perturbation 6 theory of Fermi™s Golden rule which includes matrix element s of both the perturbation (electronic coupling) and density of states (Franck -Condon factor) (Marcus , 1964). Depending on the strength of electronic coupling, the regimes of ET can be expressed as non -adiabatic or adiabatic representations. When the electronic coupling is strong, the electronic wavefunction moves rapidly back -and-forth across degenerate levels of reactant and products ( i.e. the electron is delocalized across the donor and acceptor) when the complex reaches Franck -Condon crossing -point (Figure 1.2 ). This strong electronic coupling allows the electronic state to follow the reaction coordinate adiabatically, so the electron will localize at the acceptor after the re action coordinate moves towards the product. This reaction is called adiabatic ET. If the electronic coupling is weak, only a fraction of the probability density of the electronic wavefunction moves from donor to acceptor during the short period when the s ystem is at a crossing -point. The remaining fraction will continue to move on the donor side, even if the atomic coordinate is more similar to the product. This type of reaction is called non-adiabatic ET. Both types of ET can occur in biological systems, but fast adiabatic ET only happens rarely when redox centers are extremely close ( e.g. in photosynthetic reaction centers (Huppmann et al ., 2003)). F or long -range ET in proteins, even when some steps in the ET pathway may be within the adiabatic regime, the rate -limiting step is likely to be the slower non -adiabatic ET steps. The following content will focus on ET in the non -adiabatic representation. Electronic coupling factor The electronic coupling factor can be expressed as the o ff-diagonal matrix elements (Hsu , 2009) from the Hamiltonian operation on the two -state diabatic 7 wavefunctions of the donor/acceptor complex. In the simplest picture, where an electron donor and acceptor pair are separated by empty space and the electron f rom the donor tunnels directly to the acceptor, the strength of coupling decays exponentially as separation increases (Hopfield , 1974). But in most cases of biological ET reactions, protein or solvent can provide bridges that assist electron tunneling by s upporting the delocalization of the wavefunction of the electron donor on unoccupied molecular orbitals, which increases the overlap of electron density between donor and acceptor. Two bridge -assisted electron tunneling models have been applied to describe ET in biological systems, namely superexchange and flickering resonance. In the superexchange model (McConnell , 1961) , the electronic coupling between the donor and bridge allows the donor wavefunction to propagate on the bridge, but the energy levels of the bridge molecules are well separated (and much higher) than those of the donor/acceptor states . Since the bridge is off -resonant from donor/acceptor during ET, the electron will not populate the bridge state ( i.e. will not generate a real [bridge + ele ctron] - intermediate), but the presence of bridge states effectively decreases the energy barrier for electron tunneling compared to through -space tunneling (Winkler and Gray , 2014). In the flickering resonance model (Beratan et al ., 2015), the bridging molecules and the donor/acceptor can all simultaneously reach a transient resonance during thermal fluctuations. At this transient moment, when all the electronic transitions between donor and bridges , and between bridges and acceptor , are allowed by the Franck -Condon principle, an electron can move rapidly from donor to acceptor, and the rate of electronic transition between different sites are determined by the electronic couplings between individual sites. Due to this small chance of transient r esonance on 8 bridges, the effective coupling between donor and acceptor is much greater . However, because this mechanism requires states of all the bridge molecules to have the same energy levels ( i.e. can reach resonance) co -incident with resonance of the reactant and product (see below), the probability of resonance decreases as the number of bridges increases. Probability of reaching transition state As described by the Franck -Condon principle, nuclear coordinates do not change during electronic transiti ons, and thus the actual electronic transition s only occur when the react ant and product have essentially the same energy. At physiological temperatures, the energy of the reactant oscillates with thermal fluctuations , as does its corresponding fivirtualfl p roduct (Figure 1.2, reactant and product parabolas), and the probability of ET is related to how often the reactant fivisitsfl the transition state where the reactant and product have the same energy (Figure 1.2, cross -point of two parabolas). Nuclear tunnel ing directly from the reactant vibrational ground state to the product state is also possible in quantum mechani cs theories and has been observed in biological systems (DeVault and Chance , 1966). However, nuclear tunneling is usually a significant pathway only at cryogenic temperatures; around physiological temperatures, thermally -activated ET usually dominate. Since the following content emphasizes ET within living biological systems, I will not consider nuclear tunneling further. 9 Figure 1.2 Potential energy surface of reactant and product states as a function of the nuclear coordinate The donor/acceptor complex ( [Donor - + Acceptor ]) oscillates on the Reactant surface. The probability density of reaching a certain height (potential energy) is illustrated as shading on the curve. Electron transfer can occur at the cross -point between Reactant and Product. The donor/acceptor complex ( [Donor + Acceptor -]) will oscillate on the Product surface afte r electron transfer. The Gibbs free energy (G0), activation free energy ( Ea Marcus theory Neglecting the quantized vibrations, the potential energy surface of reactant and product along the reaction coordi nates can be drawn as a set of two parabolas (Figure 1.2). The first -order rate constant of non -adiabatic ET ( ket) can be given semi -classically as Equation 1.1, also known as Marcus theory . This expression contains the electronic part ( Hab) and the Franck -Condon factor (the exponential component). The exponential 10 component of Equation 1.1 gives the probability of the system to reach the transition state with the classical activation free energy, expressed as the Arrhenius relation in Equati on 1.2 and 1.3. Here, the G0 is the Gibbs free energy, which is the energy difference between the reactant and the product at vibrational ground states. The reorganization energy ( ) corresponds to the energy needed to move the reactant from its equilibri um geometry to the product equilibrium geometry without the actual transfer of an electron. The is a measurement of energies to move all involved atoms during ET, including both the redox cofactor and surrounding dielectric medium. The factor ( )-0.5 is the probability of the transition when the Franck -Condon principle is met, where k is the Boltzmann constant and the is temperature . This factor is related to the time period when the reaction coordinates stay at the crossing -point or the vibronic br oadening (coupling between electronic and vibrational energy, 2 ) and this factor is on the scale of tenths of an electronvolt (eV) for physiological ET. = 214 Equation 1.1 = × Equation 1.2 = (+)4 Equation 1.3 The ET rate approximated by Equation 1.1 provides a practical approach to analyze biological ET reactions. In physiological ranges, the temperature -dependency of ket is mainly from the exponential component and allows the estimation of the activation energy by a classical Arrhenius plot. One of the components in the activation energy, G0 can be experimentally measured, while the originating from the bulk media can be estimated by dielectric continuum theory (Marcus and Sutin , 1985). The effective 11 electronic coupling strength ( ) can be treated as the Arrhenius pre -exponential factor ( A) since the vibronic broadening factor is close to unity. Yet, there are limitations of this calculation derived fro m semi -classical Marcus Theory. For example, this expression assumes ET is coupled to low -frequency vibrational modes (Figure 1.3A), such that the reaction coordinate moves continuously on the potential energy surface. If the transfer of an electron is str ongly coupled to a high -frequency mode (Figure 1.3B), the factor needs to be calculated in the form of separated vibration levels (Sutin , 1982). In other scenarios, where the electronic coupling is very weak, the Born -Oppenheimer approximation may break down (Beratan and Hopfield , 1984), such that electronic coupling changes with vibrational states, thus Hab is no longer separated from nuclear mo tions. It is also important to note that Equation 1.1 is a first -order rate constant that describes ET for molecules at fixed locations. If the ET reaction is bimolecular or require s the transfer of a proton, the observed ket may be limited by other factor s like diffusion, protein conformational gating (Osyczka et al ., 2004), or coupling to proton transfer (Weinberg et al. , 2012). 12 Figure 1.3 Potential energy surfaces Potential energy surfaces with (A) low -frequency (classical) or (B) high -frequency vibra tion mode s. (A) If ET is coupled to low -frequency vibrations , the potential energy surfaces are approximately continuous, and ET can occur at the cross point. (B) If ET is strongly coupled to high -frequency vibrations , ET may not occur at the cross point. Empirical rules for biological ET rates Equation 1.1 provides a practical and relative ly accurate theory for ET in biological systems such as proteins and nucleic acid s. However, the electronic coupling and reorganization energy are difficu lt to determine experimentally in many cases. On the other hand, ET reactions in proteins share many common features. For example, the medium for electron tunneling is usually amino acids that are typically arranged into certain organized structures ( e.g. helices), and a small number of redox cofactors are widely reused across many proteins. This section will introduce two widely used empirical rules for ET in proteins that provides a useful first approximation to analyze new systems. 13 Gray , Winkler and cow orkers conducted a series of experiments on photo -activatable ET in ruthenium -labeled redox proteins and analyzed the relationship between electronic coupling and protein structure (Gray and Winkler , 2003). They found that electronic coupling decays nearly perfectly as an exponential function of the separation between redox factors in the copper protein azurin (Gray and Winkler , 2003), and the exponential decay factor was measured at 1.1 Å -1. The decay factors in ET reactions -helic es ( e.g. , cytochromes) showed slightly larger variations (1.0 Œ 1.3 Å -1) (Gray and Winkler , 2003). These studies demonstrated that , in general, the presence of a protein matrix in the ET pathway increases the probability of electron tunneling , compared to the case where the pathway is filled with water . Furthermore, estimates of electronic coupling factor s can be obtained solely based on protein structure information. Moser, Dutton and coworkers analyzed the existing data from a range of measurements and provided an empirical equation of ket, also known as Dutton™s ruler (Moser et al ., 1992; Page et al ., 1999) (Equation 1.4, where R is the edge -to-edge distance between donor a nd acceptor) : log = 15 0.6× 3.1 ×(+)/ Equation 1.4 In their study, a similar exponential decay factor of electronic coupling, which was described as a function of packing density, was given for averaged proteins (1.4 Å -1). They proteins evaluated within a less polar or more polar environment, respectively. Altogether, these empirical rules are very useful for analysis of ET in proteins, including the i nitial examination of novel proteins or a complicate d multistep electron transfer chain. 14 Computer simulation of electron transfer reactions In silico molecular dynamic simulations have been shown to have great merit for understanding and predicti ng biological processes (Karplus and Kuriyan , 2005) . Warshel and coworkers pioneered the application of molecular dynamic s (MD) to Marcus theory, which connected the microscopic molecular movements to the statistical thermodynamic parameters used in Marcus t heory (Warshel , 1982). Over decades, the MD simulation of ET in proteins has been show n to be a powerful tool to explain observed reactions and make predictions (Blumberger , 2015) . In the simple picture of simulating ET (Warshel and Parson , 2001), molecul ar dynamics are performed in both reactant and product states and the trajectory are recorded. During the MD simulation in the reactant state, for example, the extra electron is localized on the donor site and the partial charges of both redox cofactors an d the polarization effects are treated by a proper hybrid Quantum Chemistry/Molecular Mechanics model and the trajectory (usually in electron transfer time -scale) is recorded. At some snapshot on the trajectory, with the atomic coordinate static, the energ y gap between reactant and product electronic states is calculated (Figure 1.4, upper) as well as the electronic coupling. These calculations will also be performed on the product trajectory. With essentially long molecular simulations, the statistical the rmodynamic information can be calculated from the distribution in the trajectory. Electronic coupling Using atomic coordinates and velocity, electronic coupling can be calculated ab initio by a large collection of quantum chemical models which generally e xhibit a trade -off between accuracy and computing expense (Blumberger , 2015). On the other hand, semi -15 empirical approaches, like the pathway model (Beratan, Betts, and Onuchic , 1991), have also been proven to be very useful in MD simulations. The pathway m odel, established by Beratan and coworkers (Beratan, Betts, and Onuchic , 1991) , hypothesized that electron tunneling from donor to acceptor is mediated by a sequence of through -bond superexchange or through -space tunneling. For each step, there is a decay factor that depends on the bond type (covalent or hydrogen bond or no bond) and the step distance. Within all the possible pathways from donor to acceptor, pathways with minimal decay factors are selected to calculate the electronic coupling. Although this classical model cannot capture interference effects accurately, the pathway model can be easily applied to molecular dynamic s trajectory frequently with very little computational cost, which makes it very useful for simulating rapid changes in electronic coupling (Jones, Kurnikov, and Beratan , 2002). Reaction coordinate The potential energy surface in Figure 1.2 uses the nuclei coordinate as the x-axis connecting the reactant and product state. This simple picture represents reactions with one degree of freedom ( e.g. the bond length in a diatomic system) well. For large molecules with multiple degrees of freedom, the potential energy is a multidimensional surface. A reactant moving along some degrees of freedom may lead to it reaching a product energy well, while moving the same reactant along other degrees of freedom will not. Warshel and coworkers chose to use the energy gap between the reactant and product as the reaction coordinate to represent all the motions coupled to electronic transition (Warshel and Parson , 2001). Representing the potential energy surface as a 16 function of the energy gap shows the same shape as Figure 1.2 (Figure 1.4, lower panel) and allows the calculation of energies from microscopic molecular dynamic simulations. The MD simulation of ET not only has been shown to be predictive for single -step electron tunneling (Blumberger , 2015), but also has shown great potential to be scaled to multi -step systems. For example, the decaheme cytochrome MtrF is believed to be an electr on conduit that allows electron flow through it (Clarke et al ., 2011). However, it binds ten chemically identical hemes so that the redox potentials of each heme cannot be determined easily. Using the high -resolution structure, the MD simulation was carrie d out and the redox potentials of each heme were assigned (Byun et al ., 2014; Breuer et al ., 2012). 17 Figure 1.4 Scheme of molecular dynamic s simulation of an electron transfer reaction The electronic transition energy gaps are calculated from MD simulation trajectories, and the distribution of energy gaps should follow a normal distribution (upper panel). The center of the distribution (average energy gap) correspond s to the energy gap a t the ground states (arrows). At the cross -point, the energy gap , by defin ition , is zero. This figure resembles Figure 1.2, but both potential surfaces can be expressed as functions of energy gap. Long -range Electron transfer The highly efficient electron transfer reactions found in the photosynthetic electron transport chain (e.g. Photosystem II) demonstrate the ability and potentials of biolog ical materials to transferring electrons across nanometer scale s (Moser et al. , 2003). For even longer range ET, electrons mostly are carried by diffusive molecules instead of via point -18 to-point electron tunneling between fixed cofactors. Although this diffusion of large molecules limits the possible maximum electron flux, electron carrie rs can protect electrons from unwanted side -reaction s (e.g. generating reactive oxygen species with O 2) and guide electrons to target acceptors through specific interactions. Inside living cells, the potential energy of electrons (redox potential) and the distribution of reduced/oxidized cofactors is part of cellular homeostasis . On the other hand, if a mi crobe needs to transfer electrons to the extracellular environment soluble electron carriers may not be the best option, since these metabolically expensive carriers are hard to recycle once secreted. Long-range ET and ETp found in microbial extracellular nanowires The biological electron transport chain (ETC) is found to extend beyond the boundary of the cell in some microb es that use extracellular material as terminal electron acceptor s of the respirat ory chain (Nealson , 1994). While m itochondria and obligate aerobic microbes use oxygen that diffuse s into the cell as an electron acceptor, bacteria like Shewanella and Geobacter can utilize a wide range of water -soluble and insoluble chemicals as terminal electron acceptor s whe n oxygen is absent (Nealson , 1994). It was found that extracellular pilin proteins are critical for mediating anaerobic respiration of Geobacter sulfurreducens PCA on insoluble metal oxides (Reguera et al. , 2005). It was soon realized that these pili are electronic ally conductive (Regu era et al. , 2005) and transfer electrons from the respirat ory chain to reduce metals far away (up to centimeters (Pfeffer et al. , 2012)) from the cells. Since then, conduct ive efflux of electrons onto extracellular substrates has been detected in a wide range of microb es including Shewanella , Synechocystis and Pelotomaculum (Gorby et al. , 2006) . The extracellular conductive structures observed in these species have been call ed microbial nanowires or 19 bionanowires. Since the discovery of these nanowires in 2005 , extensive studies have been performed mostly in Shewanella oneidensis MR-1 and Geobacter sulfurreducens PCA, however many important features, including the composition, structure and the conductivity mechanisms of the bionanowires are still subject to debate. Conductive fiNanowirefl in Shewanella oneidensis MR-1 Shewanella oneidensis MR-1 (simply Shewanella in this dissertation ) is a Gram -negative bacterium originally found in deposit s from Lake Oneida (Nealson , 1994) . Shewanella showed the ability of using a wide range of electron acceptor s, most noticeably, insoluble metal oxides of Fe(III) and Mn(IV) (Nealson , 1994). These initial observations have led to the detailed characterization of a redox network within Shewanella that connects the respiration chain (e.g. quinol pool) to a variety of periplasmic or extracellular terminal electron acceptors (Myers and Myers , 1993) . A complex cytochrome network was found that shuttles electrons in the periplasm and on the outer membrane. One of the best -studied pathway s in this network is the Mtr (metal reduction) pathway that is believed to transfer electrons to extracellular insoluble Fe(III) -oxide (Pitts et al. , 2003). The Mtr pathway contain s a periplasmic decaheme cytochrome (MtrA), a transmembrane beta -barrel protein (MtrB) , and an outer -membrane decaheme cytochrome MtrC. MtrA receives electrons from the periplasmic pool ( e.g. from other cytochromes) (Pitts et al. , 2003) , docks at MtrB, where MtrA is believed to contact MtrC through the pore of MtrB and shuttles electron s to the extracellular space (Hartshorne et al. 2007) . MtrA and MtrC together combine into a 20-heme short conduit that convey s electrons through the outer membrane (~4 nm): an event which is not likely to happen via single step electron tunneling (Figure 1.5A). This MtrABC pathway can be 20 incorporat ed into E. coli and it allow s the engineered cells to export electrons towards electrodes or insoluble ferric oxide (Jensen et al. , 2010; Teravest and Ajo -Franklin , 2016) , showing the potential of the multiheme cytochromes to be used in non -native systems (Bewley et al ., 2013). The conductivity of Shewanella ™s finanowirefl is also believed to be closely related to outer membrane multi -heme cytochromes (El -Naggar et al. , 2010), though the composition of these finanowiresfl was unknown until Pirbadian et al . found the extracellular extension involves outer membrane vesicles (periplasmic tubules , Figure 1.5B, 1.5 C), which connect to the periplasm and is rich in membrane bound cytochromes (Pirbadian et al. , 2014). This system is distinct to the previously proposed nanowire model, which hypothesized pure protein -based nanowires (see below) . In the current model, electrons diffuse inside/on the outer -membrane vesicles through cytochromes in addition to being transport ed by multi -step electron tunneling along the membrane -bound multiheme cytochrome chain (Subramanian et al. , 2018). Nanowire in Geobacter sulfurreducens PCA The connection between the growth on insoluble Fe(III) -oxide and the extracellular nanowire was first realized in a pilus -deficient mutant of Geobacter sulfurreducens PCA (Reguera et al. , 2005) (Geobacter ). Geobacter is an anaerob ic bacterium (Lin, Coppi, and Lovley , 2004) that possess es an active anaerobic respiration pathway and a broad repertoire of alternative terminal electron acceptors including a variety of insoluble metal oxides (Lovley and Phillips , 1988). A type -IV pilus protein (PilA) was found to be required for the extracellular ET (Reguera et al ., 2005). Type -IV pili are common in micro bes and are known to be involve d in cell mobility and adhesion , and are tightly related to protein 21 secretion pathways (Strom , 1993). Interestingly, the PilA proteins from Geobacters have several conserved aromatic residues near the c -terminus which are absent in other species (Reguera et al. , 2005). Furthermore, t he conductivity of the pili and the rate of Fe(III) -oxide reduction are reduced in mutant s where five aromatic residues were replaced by alanine (Vargas et al. , 2013). It is widely agreed that these aromatic residues are responsible for the conductivity of the nanowire in Geobacter (Lampa -Pastirk et al. , 2016; Malvankar et al. , 2015) (Figure 1.5D, 1.5 E), but the mechanism is under intensive debate. Cytochromes are als o believed to play an important role in the nanowire in Geobacter (Strycharz -Glaven et al. , 2011). The multi -heme outer -surface cytochrome OmcS is found associated with pili (Leang et al ., 2010) , but the spacing between two nearby OmcS are typically beyond the electron tunneling distance. It has been proposed that these extracellular cytochromes are interfaces for the nanowire to interact with Fe(III) -oxide, electrode or other nanowires in a biofilm other than promoting conductivity (Leang et al ., 2010). Most recently, Malvankar and coworkers showed that the Cryo -EM structure of Geobacter™s nanowire can be fit with the homolog model of OmcS and they proposed that one of the highly conductive nanowires is a filament of polymerized multiheme cytochromes (Wang et al ., 2019) . This discovery may bring the composition of Geobacter ™s nanowire back into debate. Nanowires in other microb es Other than the previously mentioned dissimilatory metal -reducing microorganisms, conductive nanowires have been found in ot her genera of bacteria that are not known to have metal respiration. For example, nanowires are found the cyanobacteri um 22 Synechocystis spp. PCC6803 by scanning probe microscopy (Gorby et al ., 2006) and the efflux of the extracellular current is measured from this specie s in a manner dependent on light exposure, suggesting the electrons ultimately originate from the photosynthetic ETC (Cereda et al ., 2014) . Although the physiological importance of exporting electrons and the ET pathway is not clear, finding nanowires in metabolically unrelated microbes may suggest conductive bionanowires could be widespread. Mechanisms of biological long -range ET Electron transport over millimeters on purified (but amorphous) protein sampl es had been repor ted back in the 1970s (Kimura et al. , 1979), but it was the discovery of microbial nanowires that brings the focus back upon understanding the mechanisms behind the long -range ET. Several theories have been proposed to interpret the conduc tivity of bionanowires based on their geometry, chemical composition and measured physical or electrochemical properties . In general, the proposed mechanisms can be categorized as involving one of two regimes: ET through individual redox cofactors (localiz ed), or through band gap (delocalized). The mechanism of long -range ET through localized redox cofactors has been considered as a series of individual short -range ET steps from one redox center to another, often called multi -step hopping (Winkler , 2000; Pirbadian and El -Naggar , 2012). This mechanism has also been proposed for intraprotein ET (Gray and Winkler , 2010). In contrast, if a large number of periodical redox centers are close to each and allow for significant molecular orbital overlap, del ocalized orbitals span across the entire system that can exist as a linear combination of all the participating molecular orbital s. T hese delocalized orbitals , or the fi conductive band fl, allows the probability density of an electron 23 to move very rapidly acr oss these molecules (Grozema and Siebbeles , 2008; Polizzi, Skourtis, and Beratan , 2012). The conductivity of these semiconductor (or metal lic-like) molecules depends on the energy required to put electrons on the conductive band. In proteins, the repeating peptide backbone is considered an insulator since the energy required to excite an electron to the conductive band (band gap) is large (Edwards et al. , 2008). However, aromatic residues tend to have less band gap energy due to their large -orbital s (Amdu rsky , 2013). For example, conductive organic polymers have metallic -like conductivity due to the abundan ce of conjugated orbitals (Pron and Rannou , 2002). Many experiments have been performed in order to test these mechanisms on biological nanowires. The conductivity of these extracellular structures has been studie d at different aspects from living biofilms (Yates et al. , 2015; Strycharz -Glaven et al. , 2011) to isolated individual nanowires (Leung et al. , 2013; Lovley and Malvankar , 2015; E l-Naggar et al. , 2010). Because the living biofilm may contain a network of nanowires, cytochromes and soluble electron carrier s, measurements on individual nanowires can provide more direct insights on the mechanism of the conductivity. A common set of physical and electrochemical methods has been used to extensively scrutinize bionanowires, with the vast majority of experimental evidence reported on the nanowire s produced in Shewanella or Geobacter . The measurements of individual nanowire are performed by microelectrode arrays or scanning probe microscopy (Reguera et al. , 2005; Pulcu et al. , 2012). The conductivity of each nanowire can be tested through the I-V curve , which records the current flow as a function of voltage bias applied across each nanowire. However, at the time of this dissertation , there is no decisive evidence to show these 24 mechanisms are accountable for the long -range ET in bionanowires (Polizzi, Skourtis, and Beratan , 2012; Creasey et al. , 2018). El-Nagger et al . reported that a single Shewanella finanowirefl can have a quite high electrical conductivity ( ket at equivalent 10 9 s-1 per heme chain) (El -Naggar et al ., 2010). The I-V curve measured from experiments was argued to be compatible with the multi -step hopping model with an array of hemes packed closely (heme -to-heme distance is 6.5 Å) (Pirbadian and El -Naggar , 2012). However, the fitting parameters were later argued to be non -physiological and the heme density is too large for membrane -bond cytochromes (Lovley and Malvankar , 2015). Computer simulation of electron flux of a decaheme cytochrome MtrF, which is a homolog to the proposed finanowirefl -bound cytochrome MtrC, was several orders of magnitudes smaller than that reported by the El -Naggar et al . (Breuer, Rosso, and Blumber ger, 2014). These discrepancies have led to the proposal of more unusual mechanisms, such as superexchange between hemes separated by long -distances, or diffusion -assist ed ET (Strycharz -Glaven et al. , 2011). Overall , the mechanism of the high conductivity of the Shewanella ™s finanowirefl is not clear. Adhikari et al . reported t he conductivity of an individual nanowire of Geobacter to be in the same order of magnitude to the Shewanella ™s finanowirefl (Adhikari et al. , 2016) , but the mechanism for ET has led to even more intensive discussion. Aromatic residues are believed to play an important role for the conductivity, but these residues may function as localized redox cofactors, bridg ing molecules , or as delocalized orbitals. There is considerable de bate between these models. Interestingly, t he conductivity of these nanowires was found to increase as temperature was decrease d (Malvankar et al ., 2011), 25 i.e. the opposite o f that expected from a thermally -activated hopping mechanism (Lampa -Pastirk et al. , 2016). On the other hand, temperature dependence measurement results from different groups have not been consistent (e.g. metal lic-like conductivity over the temperature range 2 -65 °C but thermally -activated at lower temperatures) (Creasey et al. , 2018; Lampa -Pastirk et al ., 2016; Malvankar et al ., 2011; Ing, Nusca, and Hochbaum , 2017). This result may be partially due to the variation in sample preparation or the environment during the measurements (Strycharz -Glaven et al ., 2011; Mal vankar, Tuominen, and Lovley , 2012; Strycharz -Glaven and Tender , 2012; Garg et al. , 2018). Indeed, in order to form a conductive band, a large number of aromatic residues need to be very close (~3.6 Å) in a face -to-face orientation (Malvankar et al. , 2011), and even minor structural distortion can break this delicate connection (Yan et al. , 2015). One approach to test these possibilities would be MD simul ation s, a powerful tool to track the fluctuations of the distances and orientation of protein residues . However, the accuracy of MD depends on the reliability of the input structure, and no high -resolution structural information has been reported for assem bled pili. Most recently, an attempt to solv e the cryo -EM structure of the assembled pilus led to the unexpected result that cytochrome OmcS fits to the electron density more readily the than PilA protein proposed to form the core of Geobacter nanowires (Wang et al. , 2019), further increasing the structural uncertainty of the proposed models. In summary, many studies have been published that characterize the properties of ET in bionanowires, however they have not reached a consensus on the mode of ET withi n the nanowires of a specific species. Furthermore, given the differences in the proposed structures of nanowires from different species, it is possible that the mechanism 26 of ET may be distinct between different microbes or within different environments. T he long -range ET found in natur al nanowires may involve multiple redox centers and/or a different type of reaction. Without more knowledge about the composition and structure, the current techniques ( e.g. , temperature -dependence or I-V curves ) cannot be used as decisive evidence to resolve the mechanism of long -range ET. Figure 1.5 Hypothetical models for the conductive nanowires Hypothetical models for the conductive nanowires . (A-C) Shewanella and (D, E) Geobacter . (A) Model of MtrABC pathway on the outer membrane. (B) Proposed outer membrane extension. (C) Multi -step hopping of electrons between hemes. (D) Proposed PilA nanowire find in Geobacter . (E) Closely stacked aromatic residues may convey metallic -like conductivity. Challenges on biological long -range ET studies without defined structures The quantum mechanical theory of electron transfer in organic compounds (cofactors) and in proteins is well -studied and has been show n to account well for such processes as the ultrafast charge se paration in photosynthetic reaction centers (pico -second) and long distance tunneling (microseconds or longer ) (Moser, Anderson, and 27 Dutton , 2010; Gray and Winkler , 2003). However, even with the knowledge of single -step ET, interpreting biological long -range ET is still challenging. This section summarizes the reasons why the lack of structural certainty precludes the understanding of natural nanowires and the engineering of synthetic bionanowires. First, the composition of natural bionanowires is not certain, and this is especially important for the determination of which type(s) of redox cofactor plays the major role in conductivity. There is a large variety of redox cof actors that are commonly found in proteins and they possess different properties in redox chemistry. For example, aromatic residues (e.g. tryptophan ) can be oxidized by a strong oxidant and is often accompanied with deprotonation (Shih et al . 2008) , while hemes can be either ferric or ferrous state in mild environment s. Therefore, aromatic residues are common ly seen as an intermediate residue during electron/hole transfer (Shih et al ., 2008; Warren et al ., 2013) and as an electronic coupling bridge in ETp (Guo et al ., 2016), while hemes are used widely as filong -termfl electron carrier s for electron shuttl ing and storage. However, the smaller size of single residues may allow aromatic residues to stack closely, but the bulky heme moi ety may not stack easily in proteins ( as they are usually found to sit parallel with an offset) (Pereira and Xavier , 2011). The properties of individual redox centers can lead to differences in conductivity mechanisms. Currently, aromatic residues in PilA and hemes in cytochromes ar e believed to be the cofactors responsible for nanowire conductivity (Creasey et al. , 2018). But it cannot be rule d out that future studies may find other cofactors, e.g. flavins (Edwards et al ., 2015) , are involved in bionanowires and /or addition al ET mechanisms, like proton -coupled ET, should be considered. Without the 28 exact knowledge of what compose s the nanowire, fully explain ing the chemistry behind the nanowire will be challenging. Second, structural information, including the distances and relati ve orientations between redox centers is vital for determining the mechanism of ET in nanowires. As mentioned in previous sections, the electronic coupling between redox centers decays exponentially with distance . When the electronic coupling is strong, el ectron density is share d between redox centers. If this delocalized electron orbital span s across the donor and acceptor with a large number of bridges, metallic conductivity is allowed (Malvankar et al. , 2015). If the delocalized orbitals are sparsely loc ated between donor and accepters, the ET is still thermal ly activated, but with higher effective coupling (Ru, Zhang, and Beratan , 2019). If the coupling is even weaker, electrons cannot jump from donor to accepter in one stop, and multiple redox centers a re needed to relay this multi -step hopping. Other than the distances ( e.g. center -to-center), the electronic coupling is also sensitive to relative orientation (Berstis, Beckham, and Crowley , 2015; Smith et al. , 2006) since the distribution of the molecular orbital s is not even in all directions. Therefore, a detailed structure along the possible electron transport pathway will be required to fully explain the mechanism of the long -range ET. With increasing amount of efforts on this topic in recent years, the picture behind the conductivity is getting clear er. Yet, without major breakthroughs on the structure and composition of natural nanowires, the understanding of long -range ET and the applications of nature or engi neered nanowires will be challenging. Generating synthetic analogs that mimic the functional features of natural nanowires , but which have more 29 rigorously defined structural information, has been proposed as a powerful tool to test theories and hypothesis from the fibottom upfl (Creasey et al. , 2018). Composition and Structure Defined Platforms for long -range ET Synthetic platform s have been proven to be greatly useful for the development and verification of electron transfer theories in history. For example , the inverted region of Marcus theory , for which Marcus was awarded the Nobel Prize in 1992, was convinc ingly demonstrated by linking well -defined electron donor and acceptor with a rigid linker (Marcus , 1993). Similarly, the electronic coupling decay thr ough proteins was determined by attaching photosensitizer at defined location s on redox proteins (Meade, Gray, and Winkler , 1989; Durham et al ., 1989). These classical studies illustrate how artificially engineered systems can provide a level of control that is valuable in elucidating fundamental mechanisms of ET in natural systems. However, platforms for studying long -range ET with well -defined structure s are currently very limited. The dimension of single proteins is mostly in the one -to-tens nanometer scale and thus constructing platforms for long -distances ET requires multiple copies of a single protein . One way in which multiple proteins can be physi cally (and electronically) linked is through self -assembl y, as is proposed for the PilA pilus. Self -assembled proteins form a highly organized periodic array (Luo et al. , 2016), therefore the structural information of the entire system can be predicted bas ed on the structure of individual proteins and their periodicity. Short poly peptides and non -proteinogenic amino acids are commonly used for constructing self -assembling nanowires, including the assembling of diphenylalanine (Mason et al. , 2014) and the incorporat ion of the non-natural aromatic amino acid on -amyloid fibril (Xu et al ., 2010) in order to test the -stacking or multi -step hopping 30 mechanisms (Ing, El -Naggar, and Hochbaum , 2018). Although many protein self -assembl y strategies have been devel oped (Bai, Luo, and Liu , 2016), very limited studies have been using natural proteins with proteinogenic residues for constructing structural ly defined nanowires (Ing, El -Naggar, and Hochbaum 2018; Creasey et al ., 2018) . Only recently, have two groups reported adding natural aromatic residues (Kalyoncu et al ., 2017) or rubredoxin (Altamura et al ., 2017) to amyloid - -sheet) in order to produce nanowires. So far, t he diversity of these structural ly defined protein platf orms for studying long -range ET is still limited to the amyloid -like fibers. Also, the lack of a heme -based platform limits the capacity to reproduc e key features of Shewanella ™s outer membrane cytochrome network , or the proposed OmcS nanowire in Geobacter . Other than testing the theories about nature nanowires, structural ly defined platforms will be also useful in the engineering of bioelectronic applications (Creasey et al. , 2018), since the conductivity, charge capacity, and environment -dependency is tig htly bound to the location and orientation of redox centers. Therefore, there is a need for the develop ment of diverse long -range electron transport platforms composed of structur ally defined proteins. Cytochrome -based long -range ET platform Cytochrome s (or hemes) have been proposed to play a major role in Shewanella ™s finanowirefl (Pirbadian et al ., 2014) and arguably they are still a possible main player in Geobacter ™s nanowire or biofilm (Wang et al. , 2019; Strycharz -Glaven et al. , 2011). Cyt ochrome s are proteins that bind cofactor heme, which is a highly versatile cofactor in biology as well as possess a lot of engineering potentials (Reedy and Gibney , 2004; Chapman, Daff, and Munro , 1997) . Heme refer s to a group of iron porphyrin molecules 31 (Figure 1.6) . The word heme is derived from Greek fibloodfl, as depicted by the dark -red color of this group of molecules. Heme possess es a strong absorption at the blue region, which is called the Soret band, and several weaker absorption bands in green -to-red region s (Butt and Keilin , 1962). Based on the spectra, heme s can be catalog ed into different types, including heme a, b, c and o, and these types have different functional group substitution on the porphyrin. Heme b, with two vinyl groups, is found in hemoglobin and myoglobin. In c-type cytochromes, two cysteines react with the vinyl groups and form c-type hemes which ha ve two thioether bonds to the proteins. c-Type cytochromes ( c-type cytochrome is the only type of cytochrome focused on in th is thesis, hereafter it will be referred to as cytochrome) are commonly found as electron shuttles and catalytic enzymes in cells (Chapman, Daff, an d Munro , 1997). Figure 1.6 Structures of heme b, heme c and heme c with bis -histidine ligand s Chemical structure s of heme b (Left) and heme c (Middle) . 3D -structure of heme c covalently bound to protein with two histidine axial ligand s. White: Carbon, Blue: Nitrogen, Red: Oxygen, Yellow: Sulfur and Orange: Iron . (Right) From the protein engineering prospective, compared to other redox cofactors, cytochromes possess several advantages: 32 1) Wide redox midpoint potential ( Em') window. The Em' (or E0') values defines the redox potential energy required for accepting or donating a n electron , under the specific conditions of an experiment (rather than ‚standard™ conditions ( E0) which might be quite different, e.g. pH = 0), and are reported here against the standard hydrogen electrode (SHE). The differences in Em' values between dono r and acceptors can be used to estimate the difference in energies for electrons on starting and ending states, which will partially determine whether a redox protein tend s to be an electron donor or acceptor when interacting with partners. The Em' of heme s found within native cytochromes can range from -400 mV to +300 mV (Liu et al ., 2014). Importantly, many factors that can contribute to poise the Em' at a particular range have already been determined (Hosseinzadeh and Lu , 2016). For example, for heme s, the Em' tend s to be more negative in a polar environment (Reedy and Gibney , 2004) . In the context of an engineered redox carrier, this can enable the design of synthetic cytochromes programmed with a specified electron affinity. 2) -system. Iron porphyrin has 26 conjugated -electrons that extend the delocalized molecular orbital almost over the entire the plane (Zerner, Gouterman, and Kobayashi , 1966). In other words, electron tunneling can start from the edge of the heme plane instead of from the iron center, which effectively decrease s the electronic coupling decays (Gilbert and Albinsson , 2015). This delocalized system also distributes the extra cha rge of the ferrous iron to the entire ring, which effectively lower s the reorganization energy compared to iron cations . These properties enable rapid ET reactions to happen on 33 hemes. Indeed, hemes are commonly found in chain s for relaying electrons , like MtrF as an electronic conduit for transport ing electrons (Clarke et al. , 2011). 3) High stability. In c-type cytochromes, hemes are covalently bond to the protein backbone with two thioether bonds (Kranz et al . 2009) . It has been demonstrated that heme w ill stay bound to proteins in many harsh environments (Barker and Ferguson , 1999), including urea and strong acid. This feature not only allow s the engineered cytochromes to be functional under relatively extreme conditions , but also makes chemical modification (e.g. demetallization and metal substitution ) possible on proteins. The iron in heme can be chemically removed and replaced by other metals. Metal -free and zinc porphyrin are highly efficient photosensitizers (Rybicka - et al ., 2016), while cobalt (Firpo et al ., 2018) and manganese (Low et al. , 1998) porphyrins have catalytic properties. Overall, due to the covalent heme binding, cytochromes have the potential to be designed for a wide range of applications. 4) Spectral tractability . Spectroscopic features of cytochromes could be used for rapid diagnosis of a design. Being able to easily inspect and troubleshoot is an important advantage during protein engineering. Cytochromes have strong absorbance in the visib le light range (extinction coefficient is about 100, 000 M -1 cm-1 at a 400 nm band) and have distinctive differen ces in the visible spectra between their reduced and oxidized state s. This visible color feature allows for quick and reliable determination of the presence of hemes and access to the redox status of a design ed protein during operations. For example, the diffusion of electrons is visualized by color changes in Chapter 2. 34 In order to understand the mechanisms behind natural bionan owires and to contribute to the design of conductive biomaterials, this dissertation demonstrated two distinct approaches to construct structurally defined heme networks that can support long - range ET , including a platform of crystallized cytochromes that can support long -range ET over tens to a hundred of micrometer s and a strategy that can introduce hemes to target proteins including self -assemb ling protein scaffolds. 35 REFERENCES 36 REFERENCES Adhikari, R. Y., Malvankar, N. S., Tuominen, M. T., & Lovley, D. R. (2016). Conductivity of individual Geobacter pili. RSC Adv. , 6(10), 8354 Œ8357. https://doi.org/10.1039/C5RA28092C Altamura, L., Horvath, C., Rengaraj, S., Rongier, A., Elouarzaki, K., Gond ran, C., Maçon, A. L. B., Vendrely, C., Bouchiat, V., Fontecave, M., Mariolle, D., Rannou, P., Le Goff, A., Duraffourg, N., Holzinger, M., & Forge, V. (2017). A synthetic redox biofilm made from metalloprotein Œprion domain chimera nanowires. Nat. Chem. , 9(2), 157 Œ163. https://doi.org/10.1038/nchem.2616 Amdursky, N. (2013). Enhanced solid -state electron transport via tryptophan containing peptide networks. Phys. Chem. Chem. Phys. , 15(32), 13479. https://doi.org/10.1039/c3cp51748a Bai, Y., Luo, Q., & Liu, J. (2016). Protein self -assembly via supramolecular strategies. Chem. Soc. Rev. , 45(10), 2756 Œ2767. https://doi.org/10.1039/C6CS00004E Barker, P. D., & Ferguson, S. J. (1999). Still a puzzle: why is haem covalently attached in c-ty pe cytochromes? Structure , 7(12), R281 -90. https://doi.org/10.1016/S0969 -2126(00)88334 -3 Beratan, D. N., Betts, J., & Onuchic, J. (1991). Protein electron transfer rates set by the bridging secondary and tertiary structure. Science , 252(5010), 1285 Œ1288. https://doi.org/10.1126/science.1656523 Beratan, D. N., & Hopfield, J. J. (1984). Failure of the Born ŒOppenheimer and Franck ŒCondon approximations for long distance electron transfer rate calculations. J. Chem. Phys. , 81(12), 5753 Œ5759. https://doi.org/10.1 063/1.447627 Beratan, D. N., Liu, C., Migliore, A., Polizzi, N. F., Skourtis, S. S., Zhang, P., & Zhang, Y. (2015). Charge transfer in dynamical biosystems, or the treachery of (static) images. Acc. Chem. Res. , 48(2), 474 Œ481. https://doi.org/10.1021/ar500 271d Berstis, L., Beckham, G. T., & Crowley, M. F. (2015). Electronic coupling through natural amino acids. J. Chem. Phys. , 143(22). https://doi.org/10.1063/1.4936588 Bewley, K. D., Ellis, K. E., Firer -Sherwood, M. A., & Elliott, S. J. (2013). Multi -heme p roteins: Nature™s electronic multi -purpose tool. Biochim. Biophys. Acta - Bioenerg. , 1827(8Œ9), 938 Œ948. https://doi.org/10.1016/j.bbabio.2013.03.010 Blumberger, J. (2015). Recent Advances in the Theory and Molecular Simulation of Biological Electron Trans fer Reactions. Chem. Rev. , 115(20), 11191 Œ11238. https://doi.org/10.1021/acs.chemrev.5b00298 Breuer, M., Rosso, K. M., & Blumberger, J. (2014). Electron flow in multiheme bacterial cytochromes is a balancing act between heme electronic interaction and redo x potentials. PNAS , 111(2), 611 Œ616. https://doi.org/10.1073/pnas.1316156111 Breuer, M., Zarzycki, P., Shi, L., Clarke, T. A., Edwards, M. J., Butt, J. N., Richardson, D. J., Fredrickson, J. K., Zachara, J. M., Blumberger, J., & Rosso, K. M. (2012). Molecular structure and free energy landscape for electron transport in the decahae m cytochrome MtrF. , 40(6), 1198 Œ1203. https://doi.org/10.1042/BST20120139 Butt, W. D., & Keilin, D. (1962). Absorption spectra and some other properties of cytochrome c and of its compounds with ligands. Proc. R. Soc. London. Ser. B. B iol. Sci. , 156(965), 429 Œ458. https://doi.org/10.1098/rspb.1962.0049 Byun, H. S., Pirbadian, S., Nakano, A., Shi, L., & El -Naggar, M. Y. (2014). Kinetic Monte Carlo 37 simulations and molecular conductance measurements of the bacterial decaheme cytochrome MtrF. ChemElectroChem , 1(11), 1932 Œ1939. https://doi.org/10.1002/celc.201402211 Cereda, A., Hitchcock, A., Symes, M. D., Cronin, L., Bibby, T. S., & Jones, A. K. (2014). A bioelectrochemical approach to characterize extracellular electron transfe r by Synechocystis sp. PCC6803. PLoS One , 9(3), e91484. https://doi.org/10.1371/journal.pone.0091484 Chapman, S. K., Daff, S., & Munro, A. W. (1997). Heme: The most versatile redox centre in biology? Metal Sites in Proteins and Models: Iron Centres (pp. 39 Œ70). https://doi.org/10.1007/3 -540-62870 -3_2 Clarke, T. A., Edwards, M. J., Gates, A. J., Hall, A., White, G. F., Bradley, J., Reardon, C. L., Shi, L., Beliaev, A. S., Marshall, M. J., Wang, Z., Watmough, N. J., Fredrickson, J. K., Zachara, J. M., Butt, J . N., & Richardson, D. J. (2011). Structure of a bacterial cell surface decaheme electron conduit. PNAS , 108(23), 9384 Œ9389. https://doi.org/10.1073/pnas.1017200108 Creasey, R. C. G., Mostert, A. B., Nguyen, T. A. H., Virdis, B., Freguia, S., & Laycock, B. (2018). Microbial nanowires Œ Electron transport and the role of synthetic analogues. Acta Biomater. , 69, 1 Œ30. https://doi.org/10.1016/j.actbio.2018.01.007 Dalton, E. F., Surridge, N. A., Jernigan, J. C., Wilbourn, K. O., Facci, J. S., & Murray, R. W. (1 990). Charge transport in electroactive polymers consisting of fixed molecular redox sites. Chem. Phys. , 141(1), 143Œ157. https://doi.org/10.1016/0301 -0104(90)80026 -T De Vault, D., & Chance, B. (1966). Studies of photosynthesis using a pulsed laser: I. Tem perature dependence of cytochrome oxidation rate in Chromatium . Evidence for tunneling. Biophys. J. , 6(6), 825 Œ847. https://doi.org/10.1016/S0006 -3495(66)86698 -5 Durham, B., Pan, L. P., Long, J. E., & Millett, F. (1989). Photoinduced electron -transfer kinetics of singly labeled ruthenium bis(bipyridine) dicarboxybipyridine cytochrome c derivatives. Biochemistry , 28(21), 8659Œ8665. https://doi.org/10.1021/bi00447a05 7 Edwards, M. J., White, G. F., Norman, M., Tome -Fernandez, A., Ainsworth, E., Shi, L., Fredrickson, J. K., Zachara, J. M., Butt, J. N., Richardson, D. J., & Clarke, T. A. (2015). Redox linked flavin sites in extracellular decaheme proteins involved in mic robe -mineral electron transfer. Sci. Rep. , 5(1), 11677. https://doi.org/10.1038/srep11677 Edwards, P. P., Gray, H. B., Lodge, M. T. J., & Williams, R. J. P. (2008). Electron transfer and electronic conduction through an intervening medium. Angew. Chemie In t. Ed. , 47(36), 6758 Œ6765. https://doi.org/10.1002/anie.200703177 El-Naggar, M. Y., Wanger, G., Leung, K. M., Yuzvinsky, T. D., Southam, G., Yang, J., Lau, W. M., Nealson, K. H., & Gorby, Y. a. (2010). Electrical transport along bacterial nanowires from Shewanella oneidensis MR-1. PNAS , 107 (42), 18127 Œ18131. https://doi.org/10.1073/pnas.1004880107 Firpo, V., Le, J. M., Pavone, V., Lombardi, A., & Bren, K. L. (2018). Hydrogen evolution from water catalyzed by cobalt -mimochrome VI*a, a synthetic mini -protein. Chem. Sci. , 9(45), 8582 Œ8589. https://doi.org/10.1039/C8SC01948G Garg, K., Ghosh, M., Eliash, T., van Wonderen, J. H., Butt, J. N., Shi, L., Jiang, X., Zdenek, F., Blumberger, J., Pecht, I., Sheves, M., & Cahen, D. (2018). Direct evidence for heme -assiste d solid -state electronic conduction in multi -heme c-type cytochromes. Chem. Sci. , 9(37), 7304 Œ7310. https://doi.org/10.1039/C8SC01716F Gilbert, M., & Albinsson, B. (2015). Photoinduced charge and energy transfer in molecular wires. Chem. Soc. Rev. , 44(4), 845 Œ862. https://doi.org/10.1039/C4CS00221K Gorby, Y. A., Yanina, S., McLean, J. S., Rosso, K. M., Moyles, D., Dohnalkova, A., Beveridg e, T. J., Chang, I. S., Kim, B. H., Kim, K. S., Culley, D. E., Reed, S. B., Romine, M. F., Saffarini, D. A., Hill, E. A., 38 Shi, L., Elias, D. A., Kennedy, D. W., Pinchuk, G., Watanabe, K., Ishii, S., Logan, B., Nealson, K. H., & Fredrickson, J. K. (2006). E lectrically conductive bacterial nanowires produced by Shewanella oneidensis strain MR -1 and other microorganisms. PNAS , 103(30), 11358 Œ11363. https://doi.org/10.1073/pnas.0604517103 Gray, H. B., & Winkler, J. R. (1996). Electron Transfer in Proteins. Annu . Rev. Biochem. , 65(1), 537 Œ561. https://doi.org/10.1146/annurev.bi.65.070196.002541 Gray, H. B., & Winkler, J. R. (2003). Electron tunneling through proteins. Q. Rev. Biophys. , 36(3), 341 Œ372. https://doi.org/10.1017/S0033583503003913 Gray, H. B., & Winkl er, J. R. (2010). Electron flow through metalloproteins. Biochim. Biophys. Acta - Bioenerg. , 1797 (9), 1563 Œ1572. https://doi.org/10.1016/j.bbabio.2010.05.001 Grozema, F. C., & Siebbeles, L. D. A. (2008). Mechanism of charge transport in self -organizing org anic materials. Int. Rev. Phys. Chem. , 27(1), 87 Œ138. https://doi.org/10.1080/01442350701782776 Guo, C., Yu, X., Refaely -Abramson, S., Sepunaru, L., Bendikov, T., Pecht, I., Kronik, L., Vilan, A., Sheves, M., & Cahen, D. (2016). Tuning electronic transport via hepta -alanine peptides junction by tryptophan doping. PNAS , 113(39), 10785 Œ10790. https://doi.org/10.1073/pnas.1606779113 Hartshorne, R. S., Jepson, B. N., Clarke, T. a, Field, S. J., Fredrickson, J., Zachara, J., Shi, L., Butt, J. N., & Richardson, D . J. (2007). Characterization of Shewanella oneidensis MtrC: a cell -surface decaheme cytochrome involved in respiratory electron transport to extracellular electron acceptors. JBIC J. Biol. Inorg. Chem. , 12(7), 1083 Œ1094. https://doi.org/10.1007/s00775 -007-0278 -y Hopfield, J. J. (1974). Electron transfer between biological molecules by thermally activated tunneling. PNAS , 71(9), 3640 Œ3644. https://doi.org/10.1073/pnas.71.9.3640 Hosseinzadeh, P., & Lu, Y. (2016). Design and fine -tuning redox potentials of me talloproteins involved in electron transfer in bioenergetics. Biochim. Biophys. Acta - Bioenerg. , 1857(5), 557 Œ581. https://doi.org/10.1016/j.bbabio.2015.08.006 Hsu, C. (2009). The electronic couplings in electron transfer and excitation energy transfer. Acc. Chem. Res. , 42(4), 509 Œ518. https://doi.org/10.1021/ar800153f Huppmann, P., Spörlein, S., Bibikova, M., Oesterhelt, D., Wachtveitl, J., & Zinth, W. (2003). Electron transfer in reaction centers of : Photosynthetic reacti ons approximating the adiabatic regime. J. Phys. Chem. A , 107(40), 8302 Œ8309. https://doi.org/10.1021/jp027845c Ing, N. L., El -Naggar, M. Y., & Hochbaum, A. I. (2018). Going the Distance: Long -range conductivity in protein and peptide bioelectronic materia ls. J. Phys. Chem. B , 122(46), 10403 Œ10423. https://doi.org/10.1021/acs.jpcb.8b07431 Ing, N. L., Nusca, T. D., & Hochbaum, A. I. (2017). Geobacter sulfurreducens pili support ohmic electronic conduction in aqueous solution. Phys. Chem. Chem. Phys. , 19(32), 21791 Œ21799. https://doi.org/10.1039/C7CP03651E Jensen, H. M., Albers, A. E., Malley, K. R., Londer, Y. Y., Cohen, B. E., Helms, B. a, Weigele, P., Groves, J. T., & Ajo -Franklin, C. M. (2010). Engineering of a synthetic electron conduit in living cells. PNAS , 107(45), 19213 Œ19218. https://doi.org/10.107 3/pnas.1009645107 Jones, M. L., Kurnikov, I. V., & Beratan, D. N. (2002). The nature of tunneling pathway and average packing density models for protein -mediated electron transfer. J. Phys. Chem. A , 106 (10), 2002 Œ2006. https://doi.org/10.1021/jp0133743 Kalyoncu, E., Ahan, R. E., Olmez, T. T., & Safak Seker, U. O. (2017). Genetically encoded conductive protein nanofibers secreted by engineered cells. RSC Adv. , 7(52), 32543 Œ32551. 39 https://doi.org/10.1039/C7RA06289C Karplus , M., & Kuriyan, J. (2005). Molecular dynamics and protein function. PNAS , 102(19), 6679 Œ6685. https://doi.org/10.1073/pnas.0408930102 Kimura, K., Nakahara, Y., Yagi, T., & Inokuchi, H. (1979). Electrical conduction of hemoprotein in the solid phase: Anhyd rous cytochrome c3 film. J. Chem. Phys. , 70(7), 3317. https://doi.org/10.1063/1.437916 Kranz, R. G., Richard -Fogal, C., Taylor, J. -S., & Frawley, E. R. (2009). Cytochrome c biogenesis: mechanisms for covalent modifications and trafficking of heme and for h eme -iron redox control. Microbiol. Mol. Biol. Rev. , 73(3), 510 Œ528. https://doi.org/10.1128/MMBR.00001 -09 Lampa -Pastirk, S., Veazey, J. P., Walsh, K. A., Feliciano, G. T., Steidl, R. J., Tessmer, S. H., & Reguera, G. (2016). Thermally activated charge tran sport in microbial protein nanowires. Sci. Rep. , 6(1), 23517. https://doi.org/10.1038/srep23517 Leang, C., Qian, X., Mester, T., & Lovley, D. R. (2010). Alignment of the c-type cytochrome OmcS along pili of Geobacter sulfurreducens . Appl. Environ. Microbiol. , 76(12), 4080 Œ4084. https://doi.org/10.1128/AEM.00023 -10 Leung, K. M., Wanger, G., El -Naggar, M. Y., Gorby, Y., Southam, G., Lau, W. M., & Yang, J. (2013). Shewanella oneidensis MR-1 Bacterial nanowires exhibit p-type, tunable electronic behavio r. Nano Lett. , 13(6), 2407 Œ2411. https://doi.org/10.1021/nl400237p Lin, W. C., Coppi, M. V, & Lovley, D. R. (2004). Geobacter sulfurreducens can grow with oxygen as a terminal electron acceptor. Appl. Environ. Microbiol. , 70(4), 2525 Œ2528. https://doi.org/ 10.1128/AEM.70.4.2525 -2528.2004 Liu, J., Chakraborty, S., Hosseinzadeh, P., Yu, Y., Tian, S., Petrik, I., Bhagi, A., & Lu, Y. (2014). Metalloproteins containing cytochrome, iron Œsulfur, or copper redox centers. Chem. Rev. , 114(8), 4366 Œ4469. https://doi.org/10.1021/cr400479b Lovley, D R, & Phillips, E. J. (1988). Novel mode of microbial energy metabolism: organic carbon oxidation coupled to dissimilatory reduction of iron or manganese. Appl. Environ. Microbiol. , 54(6), 1472 Œ1480. http://www.nc bi.nlm.nih.gov/pubmed/16347658 Lovley, Derek R., & Malvankar, N. S. (2015). Seeing is believing: Novel imaging techniques help clarify microbial nanowire structure and function. Environ. Microbiol. , 17, 2209 Œ2215. https://doi.org/10.1111/1462 -2920.12708 Low, D. W., Abedin, S., Yang, G., Winkler, J. R., & Gray, H. B. (1998). Manganese Microperoxidase -8. Inorg. Chem. , 37(8), 1841 Œ1843. https://doi.org/10.1021/ic971166c Luo, Q., Hou, C., Bai, Y., Wang, R., & Liu, J. (2016). Protein assembly: Versatile approach es to construct highly ordered nanostructures. Chem. Rev. , 116(22), 13571 Œ13632. https://doi.org/10.1021/acs.chemrev.6b00228 Malvankar, N. S., & Lovley, D. R. (2012). Microbial nanowires: A new paradigm for biological electron transfer and bioelectronics. ChemSusChem , 5(6), 1039 Œ1046. https://doi.org/10.1002/cssc.201100733 Malvankar, N. S., & Lovley, D. R. (2014). Microbial nanowires for bioenergy applications. Curr. Opin. Biotechnol. , 27, 88 Œ95. https://doi.org/10.1016/j.copbio.2013.12.003 Malvankar, N. S. , Tuominen, M. T., & Lovley, D. R. (2012). Comment on fiOn electrical conductivity of microbial nanowires and biofilmsfl . Energy Environ. Sci. , 5(3), 6247. https://doi.org/10.1039/c2ee02613a Malvankar, N. S., Vargas, M., Nevin, K. P., Franks, A. E., Leang, C ., Kim, B. -C., Inoue, K., Mester, T., Covalla, S. F., Johnson, J. P., Rotello, V. M., Tuominen, M. T., & Lovley, D. R. (2011). Tunable metallic -40 like conductivity in microbial nanowire networks. Nat. Nanotechnol. , 6(9), 573 Œ579. https://doi.org/10.1038/nnan o.2011.119 Malvankar, N. S., Vargas, M., Nevin, K., Tremblay, P. -L., Evans -Lutterodt, K., Nykypanchuk, D., Martz, E., Tuominen, M. T., & Lovley, D. R. (2015). Structural basis for metallic -like conductivity in microbial nanowires. MBio , 6(2), e00084. https://doi.org/10.1128/mBio.00084 -15 Marcus, R. A. (1964). Chemical and Electrochemical Electron -Transfer Theory. Annu. Rev. Phys. Chem. , 15(1), 155 Œ196. https://doi.org/10.1146/annurev.pc.15.100164.001103 Marcus, R. A. (1993). Electron Transfer R eactions in Chemistry: Theory and Experiment (Nobel Lecture). Angew. Chemie Int. Ed. English , 32(8), 1111 Œ1121. https://doi.org/10.1002/anie.199311113 Marcus, R. A., & Sutin, N. (1985). Electron transfers in chemistry and biology. Biochim. Biophys. Acta - Rev. Bioenerg. , 811(3), 265 Œ322. https://doi.org/10.1016/0304 -4173(85)90014 -X Mason, T. O., Chirgadze, D. Y., Levin, A., Adler -Abramovich, L., Gazit, E., Knowles, T. P. J., & Buell, A. K. (2014). Expanding the solvent chemical space for self -assembly of dipeptide nanostructures. ACS Nano , 8(2), 1243 Œ1253. https://doi.org/10.1021/nn404237f May, V., & Kühn, O. (2011). . https://doi.org/10.1002/9783527633791 McConnell, H. M. (1961). Intramolecular charge transfer in aromatic free radicals. J. Chem. Phys. , 35(2), 508Œ515. https://doi.org/10.1063/1.1731961 Meade, T. J., Gray, H. B., & Winkler, J. R. (1989). Driving -force effects on the rate of long-range electron transfer in ruthenium -modified cytochrome c. J. Am. Chem. Soc. , 111(12), 4353 Œ4356. https://doi.org/10.1021/ja00194a030 Moser, C. C., Anderson, J. L. R., & Dutton, P. L. (2010). Guidelines for tunneling in enzymes. Biochim. Biophys. Acta - Bioenerg. , 1797(9), 1573 Œ1586. https://doi.org/10.1016/j.bb abio.2010.04.441 Moser, C. C., Keske, J. M., Warncke, K., Farid, R. S., & Dutton, P. L. (1992). Nature of biological electron transfer. Nature , 355(6363), 796 Œ802. https://doi.org/10.1038/355796a0 Moser, C. C., Page, C. C., Cogdell, R. J., Barber, J., Wraight, C. A., & Dutton, P. L. (2003). Length, time, and energy scales of photosystems. In Advances in Protein Chemistry (Vol. 63, pp. 71 Œ109). https://doi.org/10.1016/S0065 -3233(03)63004 -4 Myers, C. R., & Myers, J. M. (1993). Ferric reductase is associated with the membranes of anaerobically grown Shewanella putrefaciens MR-1. FEMS Microbiol. Lett. , 108(1), 15 Œ22. https://doi.org/10.1111/j.1574 -6968.1993.tb06066.x Nealson, K. H. (1994). Iron and Manga nese in anaerobic respiration: Environmental significance, physiology, and regulation. Annu. Rev. Microbiol. , 48(1), 311 Œ343. https://doi.org/10.1146/annurev.micro.48.1.311 Osyczka, A., Moser, C. C., Daldal, F., & Dutton, P. L. (2004). Reversible redox ene rgy coupling in electron transfer chains. Nature , 427(6975), 607 Œ612. https://doi.org/10.1038/nature02242 Page, C. C., Moser, C. C., Chen, X., & Dutton, P. L. (1999). Natural engineering principles of electron tunnelling in biological oxidation -reduction. Nature , 402(6757), 47 Œ52. https://doi.org/10.1038/46972 Pereira, I. A. C., & Xavier, A. V. (2011). Multi -heme cytochromes & enzymes. In Encyclopedia of Inorganic and Bioinorganic Chemistry . https://doi.org/10.1002/9781119951438.eibc0260 Pfeffer, C., Larsen , S., Song, J., Dong, M., Besenbacher, F., Meyer, R. L., Kjeldsen, K. U., Schreiber, L., 41 Gorby, Y. A., El -Naggar, M. Y., Leung, K. M., Schramm, A., Risgaard -Petersen, N., & Nielsen, L. P. (2012). Filamentous bacteria transport electrons over centimetre dis tances. Nature , 491(7423), 218 Œ221. https://doi.org/10.1038/nature11586 Pirbadian, S., Barchinger, S. E., Leung, K. M., Byun, H. S., Jangir, Y., Bouhenni, R. A., Reed, S. B., Romine, M. F., Saffarini, D. A., Shi, L., Gorby, Y. A., Golbeck, J. H., & El -Nagg ar, M. Y. (2014). Shewanella oneidensis MR-1 nanowires are outer membrane and periplasmic extensions of the extracellular electron transport components. PNAS , 111(35), 12883 Œ12888. https://doi.org/10.1073/pnas.1410551111 Pirbadian, S., & El -Naggar, M. Y. (2012). Multistep hopping and extracellular charge transfer in microbial redox chains. Phys. Chem. Chem. Phys. , 14(40), 13802. https://doi.org/10.1039/c2cp41185g Pitts, K. E., Dobbin , P. S., Reyes -Ramirez, F., Thomson, A. J., Richardson, D. J., & Seward, H. E. (2003). Characterization of the Shewanella oneidensis MR-1 decaheme cytochrome MtrA. J. Biol. Chem. , 278(30), 27758 Œ27765. https://doi.org/10.1074/jbc.M302582200 Polizzi, N. F., Skourtis, S. S., & Beratan, D. N. (2012). Physical constraints on charge transport through bacterial nanowires. Faraday Discuss. , 155(6), 43 Œ61. https://doi.org/10.1039/C1FD00098E Pron, A., & Rannou, P. (2002). Processible conjugated polymers: from organi c semiconductors to organic metals and superconductors. Prog. Polym. Sci. , 27(1), 135 Œ190. https://doi.org/10.1016/S0079 -6700(01)00043 -0 Pulcu, G. S., Frato, K. E., Gupta, R., Hsu, H. -R., Levine, G. a, Hendrich, M. P., & Elliott, S. J. (2012). The diheme c ytochrome c peroxidase from Shewanella oneidensis requires reductive activation. Biochemistry , 51(5), 974 Œ985. https://doi.org/10.1021/bi201135s Reedy, C. J., & Gibney, B. R. (2004). Heme Protein Assemblies. Chem. Rev. , 104(2), 617 Œ650. https://doi.org/10. 1021/cr0206115 Reguera, G., McCarthy, K. D., Mehta, T., Nicoll, J. S., Tuominen, M. T., & Lovley, D. R. (2005). Extracellular electron transfer via microbial nanowires. Nature , 435 (7045), 1098 Œ1101. https://doi.org/10.1038/nature03661 Ru, X., Zhang, P., & Beratan, D. N. (2019). Assessing possible mechanisms of micrometer -scale electron transfer in heme -free Geobacter sulfurreducens Pili. J. Phys. Chem. B , 123(24), 5035 Œ5047. https://doi.org/10.1021/acs.jpcb.9b01086 Rybicka - photoredox catalysts: Experimental and theoretical studies. J. Am. Chem. Soc. , 138(47), 15451 Œ15458. https://doi.org/10.1021/jacs.6b09036 Shih, C., Museth, A. K., Abrahamsson, M., Blanco -Rodriguez, A. M., Di Bilio, A. J., Sudhamsu, J., Crane, Tryptophan -accelerated electron flow through proteins. Science 320 (5884), 1 760 Œ1762. https://doi.org/10.1126/science.1158241 Smith, D. M. A., Rosso, K. M., Dupuis, M., Valiev, M., & Straatsma, T. P. (2006). Electronic coupling between heme electron -transfer centers and its decay with distance depends strongly on relative orientat ion. J. Phys. Chem. B , 110(31), 15582 Œ15588. https://doi.org/10.1021/jp057068r Strom, M. (1993). Structure -function and biogenesis of the type IV Pili. Annu. Rev. Microbiol. , 47(1), 565 Œ596. https://doi.org/10.1146/annurev.micro.47.1.565 Strycharz -Glaven, S. M., Snider, R. M., Guiseppi -Elie, A., & Tender, L. M. (2011). On the electrical conductivity of microbial nanowires and biofilms. Energy Environ. Sci. , 4(11), 4366. https://doi.org/10.1039/c1ee01753e 42 Strycharz -Glaven, S. M., & Tender, L. M. (2012). Reply to the ‚Comment on fiOn electrical conductivity of microbial nanowires and biofilmsfl™. Energy Environ. Sci. , 5(3), 6250. https://doi.org/10.1039/c2ee03056j Subramanian, P., Pirbadian, S., El -Naggar, M. Y., & Jensen, G. J. (2018). Ultrastructure of Shewanella oneidensis MR-1 nanowires revealed by electron cryotomography. PNAS , 115(14), E3246 ŒE3255. https://doi.org/10.1073/pnas.1718810115 Sutin, N. (1982). Nuclear, electronic, and frequency factors in elec tron -transfer reactions. Acc. Chem. Res. , 15(9), 275 Œ282. https://doi.org/10.1021/ar00081a002 Szent -Gyorgyi, A. (1941). T he study of energy Œlevels in biochemistry . Nature , 148 (3745), 157 Œ159. https://doi.org/10.1038/148157a0 Teravest, M. A., & Ajo -Franklin , C. M. (2016). Transforming exoelectrogens for biotechnology using synthetic biology. Biotechnol. Bioeng. , 113(4), 687 Œ697. https://doi.org/10.1002/bit.25723 Vargas, M., Malvankar, N. S., Tremblay, P., Leang, C., Smith, J. A., Patel, P., Synoeyenbos -West, O., Nevin, K. P., & Lovley, D. R. (2013). Aromatic amino acids required for pili conductivity and long -range extracellular electron transport in Geobacter sulfurreducens . MBio , 4(2), e00105 -13. https://doi.org/10.1128/mBio.00105 -13 Wang, F., Gu, Y., O™Brien, J. P., Yi, S. M., Yalcin, S. E., Srikanth, V., Shen, C., Vu, D., Ing, N. L., Hochbaum, A. I., Egelman, E. H., & Malvankar, N. S. (2019). Structure of microbial nanowires reveals stacked hemes that transport electrons over microme ters. Cell , 177(2), 361 -369.e10. https://doi.org/10.1016/j.cell.2019.03.029 Warren, J. J., Herrera, N., Hill, M. G., Winkler, J. R., & Gray, H. B. (2013). Electron flow through nitrotyrosinate in Pseudomonas aeruginosa azurin. J. Am. Chem. Soc. , 135(30), 1 1151Œ11158. https://doi.org/10.1021/ja403734n Warshel, A. (1982). Dynamics of reactions in polar solvents. Semiclassical trajectory studies of electron -transfer and proton -transfer reactions. J. Phys. Chem. , 86(12), 2218 Œ2224. https://doi.org/10.1021/j100209a016 Warshel, A., & Parson, W. W. (2001). Dyn amics of biochemical and biophysical reactions: insight from computer simulations. Q. Rev. Biophys. , 34(4), 563 Œ679. https://doi.org/10.1017/S0033583501003730 Weinberg, D. R., Gagliardi, C. J., Hull, J. F., Murphy, C. F., Kent, C. A., Westlake, B. C., Paul , A., Ess, D. H., McCafferty, D. G., & Meyer, T. J. (2012). Proton -coupled electron transfer. Chem. Rev. , 112(7), 4016 Œ4093. https://doi.org/10.1021/cr200177j Winkler, J. R. (2000). Electron tunneling pathways in proteins. Curr. Opin. Chem. Biol. , 4(2), 19 2Œ198. https://doi.org/10.1016/S1367 -5931(99)00074 -5 Winkler, J. R., & Gray, H. B. (2014). Long -range electron tunneling. J. Am. Chem. Soc. , 136(8), 2930 Œ2939. https://doi.org/10.1021/ja500215j Xu, H., Das, A. K., Horie, M., Shaik, M. S., Smith, A. M., Luo, Y., Lu, X., Collins, R., Liem, S. Y., Song, A., Popelier, P. L. A., Turner, M. L., Xiao, P., Kinloch, I. A., & Ulijn, R. V. (2010). An investigation of the conductivity of peptide nanotube net works prepared by enzyme -triggered self -assembly. Nanoscale , 2(6), 960. https://doi.org/10.1039/b9nr00233b Yan, H., Chuang, C., Zhugayevych, A., Tretiak, S., Dahlquist, F. W., & Bazan, G. C. (2015). Inter -aromatic distances in Geobacter Sulfurreducens Pili relevant to biofilm charge transport. Adv. Mater. , 27(11), 1908 Œ1911. https://doi.org/10.1002/adma.201404167 Yates, M. D., Golden, J. P., Roy, J., Strycharz -Glaven, S. M., Tsoi, S., Erickson, J. S., El -Naggar, M. Y., Calabrese Barton, S., & Tender, L. M. (2015). Thermally activated long range electron transport in living 43 biofilms. Phys. Chem. Chem. Phys. , 17(48), 32564 Œ32570. https://doi.org/10.1039/C5CP05152E Zerner, M., Gouterman, M., & Kobayashi, H. (1966). Porphyrins. , 6(5), 363 Œ400. https://doi.org/10.1007/BF00528464 44 CHAPTER 2: MESOSCOPIC ELECTRON TRANSFER BY HOPPING IN A CRYSTAL NETWORK OF CYTOCHROMES Co-authors: Jingcheng Huang, Jan Zarzycki, Marilyn R. Gunner, Jan Kern, Junko Yano, Daniel C. Ducat and David M. Kramer I would like to acknowledge Jan Zarzycki, Jan Kern and Junko Yano for the contribution of solving the crystal structures of STC crystals. I would also like to thank Marilyn R. Gunner, Daniel C. Ducat and David M. Kramer for the experimental designs, fruitful discussions and writing of the manuscript. 45 Abstract Rapid and directed electron transfer (ET) between electron carriers is essential for biological processes . While the rates of ET over 1-2 nanometers in proteins can largely be describ ed by simplified non -adiabatic ET theory , it is not known how these processes scale to microscopic distances. Here we present a direct test of whether the assumptions based on short -range ET hold for sequential ET over mesoscopic distances. We utilized a crystalline lattice of Small Tetraheme Cytochromes (STC) that form a well -defined, three -dimensional network of closely -spaced redox centers that appears to be nearly ideal for multistep ET. E lectrons can be injected into STC crystals by dire ct photoreduction and their redistribution was monitored by imaging absorbance changes. The results suggest a hypothetical finanowirefl composed of crystalline STC with a cross -section of about 100 cytochromes could support the anaerobic respiration of a She wanella cell, while insulating the electrons from oxidation by O2. However, the interprotein ET step across 6 Å between hemes in adjacent proteins was about 10 5 s-1, 103-fold slower than expectations based on simplified ET theory . More detailed analyses implied that additional factors, possibly contributed by the crystal lattice itself, may strongly impact mesoscale ET hopping. The results also suggest design strategies for engineering nanowires with rapid ET suitable for future bioelectronic material s. 46 Introduction Life depends on the ability of cells to control the flow of electrons required to harvest energy from redox gradients or produce essential compounds. Electron transfer (ET) reactions must be rapid to maintain high fluxes, and also directed bet ween specific partners in order to avoid energy loss and the formation of harmful side products , especially reactive oxygen species (ROS) (Rutherford, Osyczka and Rappaport, 2012) . Understanding the factors that control ET are critical for diverse fields, including bioenergy, biosynthesis and disease (Page et al. , 1999) . There are currently intensive bioengineering efforts to divert electron flow into high -energy or high -value products. For example, molecular -scale synthetic nanowires have been cons tructed to link the highly efficient photosynthetic reaction center Photosystem I to hydrogenases that can store that energy as a transportable fuel (Lubner et al. , 2010) , providing proof -of-principle for the concept of a biohybrid machine based on control led, targeted ET. In some cases, ET must be conducted over mesoscopic (hundreds of nanometers) to microscopic (micrometers) scales , connecting intra - and extracellular compartments such as membrane systems, eukaryotic organelles or bacterial microcompartme nts (Verburg, 2012; Kerfeld and Erbilgin, 2015) . Recently, solid -state mesoscopic ET systems were discovered in bacteria that produce protein -based nanowires ( bionanowires ) to conduct electronic current between redox couples in the cell to extracellular re dox sinks (Reguera et al. , 2005; Strycharz -Glaven et al. , 2011) . These electrogenic microbes are increasingly under scrutiny for their bionanowires which have potential to overcome diffusion limitations and to direct electrons to specific compartments (Ren garaj et al. , 2017). The ability to form long -range electronic connections opens up possibilities to 47 optimize cellular compartments for diverse (or otherwise incompatible) chemistries (Lassen et al. , 2014; Aussignargues et al. , 2016) , potentially leading to new opportunities for metabolic engineering (Malvankar and Lovley, 2014) , bioenergy storage, synthesis of novel biosynthetic products (Verburg, 2012; Ort et al. , 2015) or biocomputing technologies (Teravest and Ajo -Franklin, 2016) . Conductive bionanowires , produced by a range of proteobacteri a such as Geobacter (Reguera et al. , 2005) and Shewanella (Gorby et al. , 2006) , have been the focus of intense research , yet the mechanism for their long -range ET is unresolved, with diverse prop osed models including the metallic conductivity model (Lovley and Malvankar, 2015; Malvankar et al. , 2015) and the thermal -activated ET models (Strycharz -Glaven et al. , 2011; Pirbadian and El -Naggar, 2012; Yates et al. , 2015; Lampa -Pastirk et al. , 2016) . Recently, Shewanella bionanowires, previously identified as semiconductor -like wires, were shown to be extensions of the outer membranes (Pirbadian et al. , 2014) , suggesting that the availability of the structural and composition information is critical fo r studying ET in proteins. More generally, there is an urgent need for an experimental platform of mesoscopic, organic ET networks that would enable detailed theoretical and systematic studie s of long -range ET (Creasey et al. , 2018) . The rate constants for ET (ket) between nearby biological redox centers (e.g. ~1 nm) can largely be explained using the classical non -adiabatic ET theory with appropriate electron tunneling corrections, or semi -classical theories that include nuclear quantum effect s (Marcus and Sutin, 1985) . Th ese theor ies capture the sensitivity of ET to many factors including donor -acceptor distances (Winkler and Gray, 2014) , orientation (Smith et al. , 2006) , potential -energy barrier s (Moser et al. , 1992) and the solvent environme nt 48 (Bortolotti et al. , 2011) . These factors are encapsulated in an activation energy barrier (for transition state), which Marcus described as a combination of Gibbs free energy ( G0) (Marcus and Sutin, 1985) , and an electronic coupling parameter, which Hopfield described as the overlap of the wave functions of electron donor and acceptor (Hopfield, 1974) . This electronic factor falls off exponentially with distance between ET partners. In one commonly accepted form ulation, this theory has the electronic coupling decay factor appear to be relatively constant within a wide range of proteins (Page et al. , 1999) . It is thus reasonable for long -range ET, e.g. in bionan owires , to be initially analyzed as a series of ET steps, each of which can be approximated by the simplified ET formulation (Pirbadian and El -Naggar, 2012; Polizzi, Skourtis and Beratan, 2012) . The work presented here is the first direct test, using a st ructurally well -defined, solid -state lattice of biological redox centers , to verify whether these assumptions hold over the hundreds to thousands of ET steps needed to deliver electrons over the mesoscopic distances inside crystalline cytochromes with appa rently ideal structural attributes for multistep ET . 49 Figure 2.1 Heme network in STC crystals (a) Heme arrangement in the crystal a- and b-axes plane. Each STC contains hour hemes (in the gray outline top left): Heme I, II, III and IV are represented in pink, yellow, blue and green, respectively. One STC backbone is shown in a ribbon model, with the N-terminal and C -terminal colored as blue and yellow, respectively. Interprotein heme contacts along the b-axis ( inset, below left) and along the a-axis ( inset, below right) were measured as the edge-to-edge (conjugated rings, dashed arrows) or as the nearest contact distance (includ ing non-conjugated atoms, solid arrows). (b) Photograph of STC crystals. The corresponding crystal axes of the big crystals are shown. (c) Redox energy profile for ET along the crystal b-axis. Heme I, II, III and IV are shown in red, yellow, blue and green, respectively. Black arrows indicate intraprotein ET, red arrows indicate interprotein ET. The midpoint potentials ( Em') and t he reduction fractions at the first reduction stage of each heme were adapted from (Harada et al. , 2002) , and the Em™ differences between neighboring hemes are labeled and shown as vertical 50 separation between hemes. The Fe -to-Fe distances are given as the horizontal separation between hemes. Results Heme network and its energy profile in an STC crystal Small tetraheme cytochrome (STC ) is a four -heme binding protein found in Shewanella oneidensis . STC was purified , crystal lized , and its structure was solved (Figure 2.1a, 2.2a, Table 2.1). We identified conditions under which the dominant crystal form was a regular, rectangular crystal (Figure 2.1b). The f our c-type hemes in STC were designated as Heme I, II, III and IV in the order of their thioether b onding motifs, located at residues 15 -19, 35 -39, 58 -62 and 75 -79, respectively. All hemes in a single STC unit are partially buried inside the protein backbone and form a chain that runs approximately parallel to the macroscopic b-axis of the crystal (long -edge, Figure 2.1a, 2.1b). These intraprotein hemes are very closely packed, with edge -to-edge distances (all distances between hemes were measured from porphyrin conjugated edges, unless specified otherwise ) smaller than 6 Å, so that intraprotein ET would be expected to proceed rapidly (Harada et al. , 2002; Fonseca et al. , 2009) Heme IV of one STC unit is positioned very near to Heme I of the adjacent STC, with an interprotein distance of 6.4 Å, forming a largescale ET lattice along the b-axis . Similar int erprotein separation was found along the crystal a-axis (short -edge) where Heme s I and II were only 6.9 Å apart from each other. The interprotein heme distance along the crystal c-axis (thickness) was somewhat longer (13. 5 Å, Figure 2. 2b), but still within ET range . Thus, within the crystal, three hemes in one STC protein come nearly in contact with a heme on an adjacent STC, so that the crystal forms a heme lattice in three 51 dimensions that is expected to facilitate multistep ET (Breuer et al. , 2012; Breuer, Rosso and Blumberger, 2014) . Fig ure 2.2 Structure and energy profile of STC crystals (a) Overlapping of the previous ly published STC structure PDB: 1M1R (yellow) and the structure solved in this study (red). (b) Heme arrangement in the crystal b- and c-axis plane. Each STC contains four hemes (in the gray box): Heme I, II, III and IV are represented in pink, yellow, blue and green, respectively. One STC backbone is show n as ribbon model, with N -termin us and C-ter min us colored as blue and yellow, respectively. Interprotein heme contacts along the c-axis were measured in Fe -to-Fe (dashed arrows) or edge -to-edge (solid arrows). (c) Energy profile for ET in three axes. Four hemes in one STC are represented vertically, and Heme I, II, III, IV are shown in red, yellow, blue and green, respectively. The intraprotein heme contacts are shown as strips pairing two hemes with corresponding colors. The midpoint potentials ( Em') of each heme were adapted from (Harada et al. , 2002) , and shown as heights. The shortest interprotein heme contact in each axis are indicated by arrows, and the Fe -to-Fe distances as well as the energy differences are labelled. 52 The microscopic redox midpoint potentials (E m') of all hemes in STC protein solutions have been previously determined (Harada et al. , 2002; Fonseca et al. , 2009) and our redox titration s of STC proteins in solution and as crystals yield ed similar results (Figure 2. 3), indicating that being trapped in a crystal lattice does not significantly affect redox properties. Previous studies also showed that the hemes within a STC interact with each other electrostatically (Harada et al. , 2002; Fons eca et al. , 2009) , such that the E m' should vary as a function of reduction stage. In the following experiments , the photo -injection of electrons into the crystal resulted in the reduction of less than one heme per STC protein on average, and thus we used the microscopic E m' values for the first reduction stage ( -272, -195, -155 and -163 mV for Heme I, II, III and IV respectively, Figure 2.1c). Combining the spatial arrangement of the heme network and the E m' of each heme, we constructed an energy profile f or every ET step in each direction , including both intra - and interprotein ET. For the crystal b-axis , electrons must pass through all four hemes sequentially, resulting in a firoller -coasterfl free energy profile. A similar set of ET steps should occur alon g the a-axis (Figure 2. 2c). The ET in crystals can happen in both direction s (i.e. reversible, Figure 2.1c), and the uphill reactions should slow the ket by a factor of 10 when compared to the reversed downhill reactions. 53 Fig ure 2.3 Redox titrations of STC crystals and STC in solution Redox titration curve s of STC crystals in photoreduction condition (black boxes) and STC proteins in a solution (cyan circles and line). B oth results are normalized to full reduction. Applying these values and assumptions to the simplified ET formulation with Dutton™s parameters (Page et al. , 1999) (see Methods ), we obtained an estimate for the overall ket along the b-axis, presuming the endergonic interprotein ET is the rate -limiting step, to be around 107-108 s-1. Similarly, rapid ket values were estimated using these assumptions for ET along the crystal a-axis ( Figure 2. 2c). While t he predicted ket along the c-axis was only 10 2-103 s-1 (Figure 2. 2b, 2.2 c). Overall , the first -order theoretical estimates suggest that electrons should diffuse quite rapidly along the a- and b-axis, with an expected halftime of about 30 s to move through the long -edge of a typical crystal with 100 µm length. Photoinduced cytochrome reduction and ET in STC crystals The thin rectangular dimensions of STC crystals (Figure 2.1b) were amenable to optical transmission microscopy in the a- and b-axes plane (Figure 2. 4, left column ), but 54 prevented imaging along the c-axis ( i.e. crystal thickness). In air, all hemes were oxidized, as demonstrated by the lack of sharp absorption bands in the 520 -550 nm region (Figure 2.5a). It was previously shown that c-type cytochromes can be photoreduced with blue or green light (Gu et al. , 1993) , and we found that illumination of STC crystals with blue light under micro -aerobic conditions in the presence of a sacrificial electron donor resulted in the appearance of spectral features (Figure 2. 5a) indicative of heme reduction (Eaton and Hochstrasser, 1967) - -bands at 525nm (bandwidth ~20 nm) and 550 nm (bandwidth ~ 10 nm) (Figure 2. 5a, 2. 5b). Imaging with 560 nm (at the isosbestic point for the reduced -oxidized difference spectrum) or 572 nm (at which the reduced heme spectrum shows bleaching) showed the expected absorbance changes (Figure 2.6a, 2.6b), confirming that the signal reflects heme reductio n and not photodamage or other artifacts. In addition, the crystals showed dichroism for both the oxidized and the reduced (either by photoreduction or addition of reductant) states (Figure 2. 5b). The - -bands oriented more strongly along the a-axis (Figure 2. 5b), in agreement with the expected orientation of Hemes II and III in the crystal lattice, which should display transition dipoles more strongly coupled to E// a light (Figure 2.1a ), indicating that the absorbance change at 550 nm 550) signal was specific for hemes in the crystal lattice, and not to the presence of modified or exotic chromophores. We estimate the quantum efficiency for blue light photoreduction of STC hemes in the crystals to be low (~10 -5, Methods ) co nsistent with previous work (Suslick and Watson, 1992) . 55 Figure 2. 4 Redistribution of injected electrons within STC crystals (Left column) STC crystal imaged under 550 nm monochromic light. Scale bar = 10 µm. Crystal axes are shown in blue ( b) and green ( a) arrows. Any visible crystal fractures are indicated (arrows). A cartoon scheme of photoreduced regions (pink) and electron diffusion direction is show n for each example within the yellow outline of a virtual crystal. STC crystals were visualized immediately following photoreduction (middle) and after the indicated length of time (right column) by 550 550 550 = 0, yellow dashed lines: outline of crystal, pink dashed line: fractur e) 56 Fig ure 2.5 Photoreduction of STC crystals (a) UV-Vis spectr a of STC in solution and in crystal form s, normalized at 560 nm. (b) Visible spectra of STC crystals before (0 min) and after (120 min) photoreduction , measured under polarized (electric field vector parallelized) to crystal a- or b-axis. (c) The photoreduction and oxidation cycles of a STC crystal. Top panel shows the crystal and the sampling sections. Scale bar = 10 µm . The lower panel shows the reduced heme density of each section. (d) Comparison of photoreduction rates of ruthenium (tris) bipyridine labeled and unlabeled STC crystals (n = 3, error bar = Std. Dev.). 57 Fig ure 2.6 Microscopy for monochromic imaging and photoreduction (a) Microscope setup and light path scheme for crystal photoreduction. (b) Absorbance shift of STC crystals at 550 nm, 560 nm and 572 nm during photoreduction. Up per-right panel shows the absorption spectr a of oxidized (dotted) and reduced (solid line) STC protein solution with arrows pointing at wavelength s where microscope images were collected . (c) The correlation between reduced heme s with in STC crystal s and the light intensity or exposure time. The inner panel shows the reduction of hemes over time at different light intensity. The outer panel shows the li near response of photoreduction rate to light intensity. (Error bars indicate the 95% confiden ce interval ). By adjusting the size or shape of the microscope field diaphragm for the excitation light (Figure 2. 6a), we could photoreduce hemes at defined locations (with minimal beam ~20 µm) in the crystal (Figure 2. 4). The redistribution of these electrons could then be followed by periodically (2 -10 min -1) imaging the absorbance changes using pulsed 550 58 nm light, polari zed along the crystal a-axis (Figure 2. 4, left column) to coincide with the heme reduction in crystals . The weak measuring pulses by themselves did not cause measurable accumulation of photoreduced STC hemes. Illumination of a small portion of a STC cryst al produced a localized reduction of hemes (Figure 2. 4). The further reduction in this region ceased immediately after switching off illumination , but the region of reduction continued to spread progressively in both the a- and b-axes directions (Figure 2. 4, right column), consistent with diffusion of the electrons throughout the crystal. Extending the photoreduction time led to more complex kinetic profiles of reduced heme redistribution including multiple phase s of heme oxidation (Figure 2. 7) that presuma bly reflected the reduction of multiple hemes with more negative Em', that were more rapidly oxidized by residual O2. To simplify the analyses, we used controlled illumination that on average reduced fewer than one heme per STC , so that STC units should be either in fully oxidized or contain one reduced heme, so that each interprotein ET could be treated independent ly. Because of the high dielectric of the medium, injection of electrons into one part of the crystal should not generate a strong tran s-crystal electric field . Thus, the ET we observed can be viewed as an fielectron diffusionfl process that relaxes from spatial heterogeneity to equilibrium (Figure 2. 8a, 2. 8b, at the gray line ), analogous to the gradient -driven charge transport in redox pol ymers (Dalton et al. , 1990) . 59 Fig ure 2.7 Oxidation kinetic s of photoreduced STC crystal The oxidation (solid line) was fitted with first -order exponential decay (blue dashed line) and second -order exponential decay (red dashed line). 60 Figure 2.8 Quantitative analysis of electron diffusion in STC crystals (a) Each single crystal was subdivided into multiple sections (black boxes) and coded as a function of its position, from blue (top) to red (bottom). The photoreduction illumination was focused on the top part of this crystal. (b) Redox density kinetics are sh own for photoreduction (left of gray line) and in the dark (right of gray line). Reduction densities were estimated for each section by dividing 550 at each time by the initial absorbance (A 550) prior to illumination for thickness correction. The red uction density of each section in (A) are shown in traces with corresponding colors. The black dashed line shown is the averaged reduction density of all sections. (c) Fitting of diffusion model (solid lines) to electron diffusion experimental data (dashed lines). (d) Arrhenius plot of electron transfer rates ( ket) in STC crystals at different temperatures and the linear fitting (dashed line, n = 3 for 36°C and 8°C, n = 4 for 25°C and 0°C , individual measurements: gray circles, error bars = Std. Dev. ). 61 Based on these assumptions, we used a 3 -D electron diffusion model to estimate overall ket along a- or b- crystal axis. Experimentally, we measured electron diffusion in one axis at a time. For example, photoreduction focused on one terminus of a crystal long -edge ( b-axis) resulted in a redox gradient along the b-axis , but an equal reduction across the short -edge. This 1 -D redox gradient allowed us to estimate diffusion coefficient s using a 1-D diffusion equation, but the same results are obtained with a 3 -D model. Figure 2. 8c shows that curve fitting to this model was able to adequately fit our experimental dataset without further assumptions. The calculated diffusion coefficient in the b-axis at room temperature (25°C) was 1.26 (± 0.46) ×10 -9 cm2s-1, which corresponded to an interprotein ket along the b-axis of 7.4 ± 2.7 ×10 4 s-1 (The rate constant is unimolecular at the first reduction stage). Using the same approach, we estimated that the ket along the a-axis was similar to the b-axis (2.4 ± 0.5 × 10 5 s-1, Figure 2. 9). 62 Fig ure 2.9 Quantitative analysis of electron diffusion in along a-axis of a STC crystal (a, c and e) Subdivision of a single crystal. Each section was labeled with a different color. The photoreduction illumination was focused on the left part of this crystal. Scale bar = 10 µm. (b, d and f ) Fitting of the diffusion model (solid lines) to electron diffusion experimental data (dashed lines). The calculated diffusion coefficient for (b)( d)( f) are 2.5, 1.9 and 2.8 ×10 -9 cm2 s-1, respectively, and the corresponding ket are 2.5, 1.9 and 2.8 ×10 5 s-1 (Equation 2.4 in Method s, l = 2.46 nm, n = 3) 63 Both the observed rate of diffusion and the estimated ket values obtained from our diffusion mod el were about 3 orders of magnitude slower than those estimated from Dutton ™s parameterized non -adiabatic ET formulation. W e thus tested for possible artifacts that could impact our estimates . One possible interpretation is that defects within the STC crys tals could restrict ET , thereby decreas ing the overall macroscopic electron diffusion. The redistribution of electrons proceed ed smoothly throughout the entire region s of undamaged crystals ( e.g. Figure 2. 8a) but did not cross gaps or fractures between separate but nearly touching crystals (Figure 2. 4, 3rd and 4th example s). These observations allowed us to rule out the effects of large visible defects. However, it was still possible that uniformly distributed nanoscale imperfections could slow electron diffusion. To predict how such defects would affect diffusion rate, we analyzed a series of kinetic Monte Carlo random walk (Martínez et al. , 2008; Byun et al. , 2014) simulations of electron diffusion in a matrix with varying fractions of randomly -positioned defect, that blocked ET . In 2 -D matrixes, with diffusion only along the a- and b-axes, more than 30% of centers would have to be defective to decrease the ket by a factor of 10 (Figure 2.1 0). Allowing slow ET to occur in the c-axis (with ket 10-fold slower than along the a- and b- axes), largely alleviated the effects of defects (Figure 2.1 0) by opening alternative routes of diffusion, suggesting that the lattice arrangement should be highly robust to defects. Moreover, the hig h degree of disorganization required to substantially slow ET appears to be in contradiction with the highly ordered crystallographic data (Table 2.1 ). We also considered the possibility that the slow rate could be explained by ficongestion effectsfl , i.e. if electrons were occupying adjacent STC proteins, interprotein ET would be 64 prevented. However, our experiments induced only low extents of heme reduction crystals and modulating the extent of photoreduc tion by more than 2 -fold did not greatly alter the ele ctron diffusion rates ( e.g. ket values in Figure 2. 9), implying that congestion effects were not important. Fig ure 2.10 Random walk simulations of the i mpact of crystal defects that inhibit ET on electron diffusion rate All diffusion rates are normalize d to a theoretical ‚perfect™ matrix with no blocking units. Red squares and line represent the simulation with rate in a: b: c axis = 10: 10: 1, cyan circles and line represent the simulation with rate in a: b axes = 1: 1 and c-axis is 0. Another possibility is that the electrons become trapped on high potential electron carriers that are not measured by our 550 signals. In t he absence of oxidation by O 2, 550 across the crystal) of the entire crystal remained constant after the initial photoreduction (Figure 2. 8b), indicating that ET occurred within the crystal l attice without significant electron exchange 550 was linearly dependent on illumination time 65 or in tensity (Figure 2. 6c) and did not display saturation effects. The lack of any notable lag phase in the reduction implie s that there were no electron acceptors with E m' more positive than the STC hemes, the reduction of which would have been favored in the initial stages of illumination. Finally, photoreduction and re -oxidation of STC crystals was reversible and could be repeated for many cycles (Figure 2. 5c), i ndicating that photodamage wa s not a major favor in limiting ET. Interestingly, we found that soluble STC, added to the same environment together with STC crystals, was more than 10-fold more rapidly oxidized by residual O2 than that in the crystals (Figure 2. 11), implying that packing into crystalline networks protected reduced heme from oxidation. This feature may allow a wire constructed from a STC crystal to deliver electrons over long distances while avoiding oxidative damage from the formation of ROS. Figure 2.11 O xidation of photoreduced STC in solution and STC crystals (a) Absorbance measuring sections. Two were in STC solution, and one was on the STC crystal. (b) Absorbance change of each corresponding section in ( a). The absorbance of a STC crystal only (without solution, green trace) was calculated by subtracting absorbance of STC solution (blue and red traces) from observed total absorbance (light -blue trace, contain both STC solution and STC crystal). Note that the lifetime of the oxidative decay of STC soluti on is about 1 min, while the oxidative decay of STC crystal should be much longer. 66 ET in STC crystals is controlled by a process with high activation enthalpy Figure 2. 8d sho ws the temperature dependence, rang ing from 0 to 36 °C, of the ket estimated fro m fitting to our diffusion model. The results were interpreted with the Arrhenius form of the Marcus expression (Bortolotti et al. , 2011) for non -adiabatic ET (Equation 2.1 and 2.2) : = 214 () Equation 2.1 ln()=(+)4 Equation 2.2 where A is the Arrhenius pre -exponential factor related to the electronic coupling matrix , k is the Boltzmann constant, G0 is the Gibbs free energy change and is the reorgan ization energy. The value of A estimated from the plot was in order of 10 12 s-1, in good agreement with typical electronic coupling observed for other biological heme pairs at the range of 6 Å (Moser et al. , 1992; Gray and Winkler, 2009) (~10 11 s-1 based o n activation energy optimized uniform barrier model ). However, the activation energy ( Ea = ) estimated from the plot was quite high, about 0.43 eV (with 90% confidence interval from 0.36 to 0.6 eV). This result support s the argument that the slow electron diffusion is not caused by large number of crystal defects, because these would tend to decrease A (by decreasing the number of centers that were capable of ET) rather than increasing Ea. Because the calculated ket in ST C crystals was slower than the intraprotein ket reported for STC in solution (Harada et al. , 2002; Fonseca et al. , 2009; Jiang et al. , 2017) , the rate -limiting step was most likely the interprotein endergonic ET, from Heme IV to Heme I. Taking into account the Em' differences between Heme I and IV (0.11 V) , we estimated to be ~1.5 eV (with a 90% confidence interval from 1.21 67 eV to 1.77 eV), which is unusually high compared to those reported earlier for similar heme system, which are typically below 1 eV (Muegge et al. , 1997; Alric et al. , 2006; Bortolotti et al. , 2011) . Discussion Electrogenic microbes like Shewanella are reported (El -Naggar et al. , 2010) to eject about 10 5-6 electrons s -1 cell -1 to extracellular electrodes . Assuming that the ET properties of an STC crystal are scalable, achieving the equivalent rate of transport along its b-axis would require a rectangular cross -section of roughly 1 00 proteins (diameter of roughly 30 nm), which is roughly 1/10 th of the diameter of an Escherichia coli cell. It thus seems theoretically plausible that a STC wire could function to move electrons over long distances at physiologically relevant rates. More over, the potential ability to shield diffusing electrons from O 2 (Figure 2. 11), and thus prevent the formation of ROS, could be physiologically beneficial. On the other hand, single natural bionanowires, with a diameter about 10 nm have been reported to transport more than 10 10 electrons s-1 (Polizzi, Skourtis and Beratan, 2012), 10 4 times faster than our hypothetical STC wire. One major difference is that ET in bionanowire s (or in t he form of biofilm s) has usually been studied by following the bias voltage -dependent electrical current with scanning probe microscopy (Reguera et al. , 2005; El -Naggar et al. , 2010; Polizzi, Skourtis and Beratan, 2012) , microelectrodes (El -Naggar et al. , 2010; Pirbadian and El -Naggar, 2012) or cyclic voltammetry (Wang et al. , 2005; Yates et al. , 2015) . In contrast , the movement of electrons over the meso - to microscopic distances in STC crystals was driven predominantly by entropic effects 68 where there is l G0 between the source and the sink. This low driving force limit is likely to be more relevant to the biological environments of bionanowires, where the high dielectric of the medium would tend to dissipate electric field s generated by redox gradients. A strong driving force would imply a large loss of free energy as electrons moved down the wire, so that electrons exiting the distal end of the wire would be less reactive. Thus, the physiological relevance of the fastest me asured bionanowire currents (with large applied voltages) is questionable. Nevertheless, the apparent large differences in electron transport rates suggest that the two systems may behave in fundamentally different ways, and it is instructive to understand the underlying mechanisms so as to design more rapid biological ET systems. The high Ea for ET measured in the crystals (Figure 2. 8d) was consistent with thermal -activation mechanisms, unlike the metallic -like conductor or semi -conductor models proposed f or bionanowires (Leung et al. , 2013; Lovley and Malvankar, 2015) . This conclusion was also consistent with the UV-Vis absorption spectra of the oxidized and reduced crystals which remain very similar to those of the STC in solution (Figure 2. 5a), with no e vidence for strong electronic coupling between hemes. A previous study (Fonseca et al. , 2009) showed only a weak dependence of heme Em' on pH, and nuclear magnetic resonance (NMR ) studies showed that intraprotein ET was quite rapid (Harada et al. , 2002; Fonseca et al. , 2009) between the pairs of hemes likely to be most affected by protonation events. Collectively, these results argue strongly against the participation of a proton -coupled ET mechanism (Fonseca et al. , 2009) . Several mechanisms cou ld contribute to the large Ea for ET in STC crystals and the slow electron diffusion we observed , each with strong implications for constructing 69 more efficient and useable bionanowire s. The E m' of STC hemes vary by about 0.12 V resulting in a redox firoller -coaster fl profile (Figure 2.1c) which create s local energy traps on Hemes III and IV. Such alternating redox potential arrangements have been seen in shorter -chain redox chains found in large bioenergetics complexes, such as the mitochondrial Complex I iro n-sulfur chain (Ohnishi, 1998) and the bacterial reaction center tetraheme donor (Page et al. , 1999; Alric et al. , 2006) . In the STC crystal, ET from high potential hemes can only proceed upon thermal activation, adding incrementally to the Ea for the overall electron diffusion. However, the redox trap in the STC crystals appears to be too small to fully explain the large es timated Ea for electron diffusion in STC crystals . Nevertheless , it may be possible to speed up ET by minimizing the redox traps, though this would require subtle engineering to fine -tune the Em' of redox centers. Completely remov ing the redox traps , but leaving everything else unchanged , should decrease Ea to 0.38 eV, and could theoretically accelerate electron diffusion by about 10 -fold at room temperature. The relative orientations of the electron donor -acceptor might also play a role in determining ket. In the case of hemes in van der Waals contact , the electronic coupling is generally larger when two hemes are stacked on each other (i.e. two heme planes are close and parallel to each other) , but weaker when they are co-plan ar or in a fiT-shape fl as in our crystal (Smith et al. , 2006) . However, we can probably rule out large orientation effects for the following reasons. First, the Arrhenius pre -exponential factor for ET was large, similar to that seen in a uniform barrier model where these effects are not considered. Second, when compared to other heme pairs, t he interprotein heme pair (Heme s I and IV) in STC crystals has a very similar separation and orientation (see 70 overlap in Figure 2.12 ) to some natural heme pairs, e.g. heme 6-7 or heme 1 -2 in decaheme cytochrome c MtrF (Breuer et al. , 2012; Breuer, Rosso and Blumberger, 2014) . The electronic coupling of these two pairs were calculated as Hab = 0.23 × 10-3 eV (heme 6-7) and Hab = 0.22 × 10-3 eV (heme 1-2) (Breuer, Rosso and Bl umberger, 2014) , i.e. weaker than the estimated pre -exponential factor in our crystals ( 8.3 ×10 -3 eV, with a 90% confidence interval from 2.0 ×10 -3 eV to 35.4 ×10 -3 eV). However, further study suggested that these computational results largely underestimated the Hab by about three orders of magnitude when compared to experimental results (Byun et al. , 2014) . Therefore, although there are some uncertainties in our estimates of electronic coupling due to the limited temperature range, we expect the interprotein electronic coupling to be larger than the MtrF counterparts due to the presence of a bridging cysteine (Jiang et al. , 2017) between Heme I and IV (Figure 2.1a). Finall y, our results imply that ket along the a-axis is on the same order of magnitude, perhaps slightly faster (Figure 2. 9), than that along the b-axis. Although the interprotein hemes along the a- and b-axis appear to have similar heme distances (6.4 Å in b-axis vs 6.9 Å in a-axis), solvent exposures and redox traps, they have distinct heme -heme orientations , suggesting that the orientation effects were not the major contributors to the small ket. 71 Fig ure 2.12 Structure overlap of interprotein heme contact between two adjacent proteins within the crystal as compared to intra -protein contacts within MtrF. Structure overlap of interprotein heme contact between Heme s I (pink) and IV (green) in STC crystal s to the intraprotein heme contacts in MtrF (blac k, PDB: 3PMQ). (a) Overlap Heme s I and IV of STC on Heme 1 and 2 of MtrF. (b) Overlap of Heme I and IV of STC on Heme 6 and 7 of MtrF. The temperature dependence measurements for electron diffusion (Figure 2. 8d) imply that ket many biological ET systems (Page et al., repolarization of solvent associated with heme redox changes. We chose STC in part because ex posure of heme edges allows for interprotein contacts needed to form 3 -D heme networks in aggregates or crystals. However, these structural features also expose 72 (Marcus and Sutin, 1985; Tipmanee et al., 2010). But for comparison, the Ru -modified tri -heme cytochrome systems developed by Tiede and coworkers (Kokhan et al., 2015), in which both the cross -linked ruthenium(II) -tris -bipyridine and the three bound hemes were par photoinduced ET systems developed by Gray and coworkers (Meade, Gray and Winkler , 1989) showed a < 1 eV, except for the Ru -(NH 3)-cyt c systems, where = ~1.2 eV (Tipmanee et a l., 2010) , presumably because the ruthenium photosensitizer was nearly fully surrounded by water and its N -ligands are polar (Tipmanee et al. , 2010) . Thus, it also appears unlikely that heme solvent exposure, by itself, can completely explain the large s within individual STC proteins. However, an X-ray structural study (Leys et al. , 2002) as well as NMR structural studies (Harada et al. , 2002; Fonseca et al. , 2009) showed only small structural differences between oxidized and reduced STC, arguing against the involvement of large redox -induced changes in the protein conformation, consistent w ith previous computational work (Jiang et al. , 2017) . We thus propose that at least part of the high Ea may be related to the dynamic of crystal lattice (Meinhold, Merzel and Smith, 2007) , perhaps by introducing new, low -frequency modes of relaxation . Wor k on photosynthetic reaction centers has shown that ET can induce perturbations in protein environment that relax over a wide range of time scales from picoseconds to seconds, and presumably longer, and it has been show n that such vibration relaxations can influence ET (Cherepanov, Krishtalik and Mulkidjanian, 2001). In most short -range biological ET, one can probably ignore vibrational modes with 73 frequencies longer than ET itself (Schenck et al. , 1981; Gunner and Dutton, 1989; Cherepanov, Krishtalik and Mu lkidjanian, 2001) . Packing of proteins in crystals should generate new and damp existing harmonics (Acbas et al. , 2014) that can lead to different ET kinetics compared to our solution -based estimation. Moreover, mesoscopic electron diffusion seen in the STC crystals will occur over a wide range of times scales, from the sub -microsecond intraprotein ET to the diffusion across the crystal on the tens of minutes to hours scale, allowing for interactions between ET steps and these longer relaxation modes, inc luding small changes in the crystal packing and alterations of interaction between protein contents and the solvent both within and surrounding the crystal lattice. Method s Rationale of choosing STC crystal Well studied biological ET chains typically have donor-acceptor cofactor distances (edge -to-edge) between 5 -20 Å (Moser, Anderson and Dutton, 2010) . The ket should decrease approximately exponentially with the distances between redox centers (Winkler and Gray, 2014) . We thus searched the Protein Data Ban k (PDB , www.rcsb.org ) for small cytochromes with multiple heme -binding motifs (Barker and Ferguson, 1999) , that could potentially form crystalline lattice s of hemes within ET distance along multiple crystal axes . We identified 7 non -redundant entries with heme to amino acid quantity ratios greater than 0.04 , including cytochrome c7 from Desulfuromonas acetoxidans (1HH5), small tetraheme cytochrome c (STC) from Shewanella oneidensis MR1 (1M1R) and cytochrome c7 proteins (PpcA -E) from Geobacter sulfurreducens (1OS6, 3BXU, 3H33, 3H4N and 3H34). Of these, STC demonstrated high expression and solubility, and is well characterized in biochemical (Tsapin et al. , 2001; Leys et al. , 2002; Meyer et al. , 2004) , 74 electrochemical (Harada et al. , 20 02; Alves et al. , 2017) and computational (Jiang et al. , 2017) studies. STC expression and purification Recombi nant STC was expressed as a fusion with a Strep -II tag and a factor Xa cleavage site at the N -termin us of the holo cytochrome and encoded in plas mid pTSSX -STC (AddGene ID: 115668). The STC protein was expressed in Escherichia coli BL21 (DE3) ( New England Biolabs , MA, USA ) strain harboring pEC86 (Arslan et al. , 1998) and pTSSX -STC plasmids. Cells were cultured aerobically in 2xYT media (Alpha Biosciences , MD, USA) with 50 µg/m L kanamycin (Sigma -Aldrich , MO, USA) and 40 µg/m L chloramphenicol (Sigma -Aldrich ) at 30 °C and shaken at 140 rpm for 48 h in 4 L flasks. After h arvesting cells, periplasmic components were extracted by osmatic shock, and then captured by anion -exchange chromatography resin (DEAE ŒSepharose, Sigma -Aldrich , MO, USA ) in Tris -Cl (Sigma -Aldrich ) pH 8.0 buffer. This protein extraction was further purifie d with affinity chromatography (Strep -Tactin Superflow Plus, Qiagen , Hilden, Germany ), followed by the removal of affinity tag by protease factor Xa ( New England Biolabs ) treatment and size exclusion chromatography (Sephacryl S -100 HR, GE Healthcare , IL, U SA). STC Crystallization For crystallizing STC, 5 mg/m L of purified cytochrome (in 1 mM HEPES (Sigma -Aldrich ) pH 8.0) was mixed with reservoir buffer (16% polyethylene glycol (PEG) 3350 (Sigma -Aldrich ), 15 mM ZnSO 4 (Sigma -Aldrich )) in 1:1 (v/v) ratio and then placed at 4 °C in sitting drop vapor diffusion dishes. Crystals were harvested after most of the cytochromes in mother liquor were absorbed into crystals (~10 days ). Crystals were 75 transferred to 30% PEG 400 ( Sigma -Aldrich ) before freezing in liquid nitrogen. Diffraction data were collected at the Advanced Light Source of the Lawrence Berkeley National Laboratory (beam line 5.0.1). The data were processed with XDS (Kabsch, 2010) and the CCP4 package (Winn et al. , 2011) . The structure was solve d by molecular replacement using PDB 1M1R (Leys et al. , 2002) as the search model. Molecular replacement was carried out using Phaser of the Phenix software package (Adams et al. , 2010) and refined with Phenix.Refine. Additional modeling, manual refining a nd ligand fitting was done in COOT (Emsley and Cowtan, 2004) . Final positional and B -factor refinements, as well as water -picking for the structures, were performed using Phenix.Refine. The resulting structure was deposited in the Protein Data Bank (PDB) u nder the PDB ID 6EE7. Despite differences in packing, the single -unit protein structure was nearly identical to that previously reported (Leys et al. , 2002) (TM -score= 0.95 (Zhang and Skolnick, 2004) , Figure 2. 2a). Microscopy Imaging and Photoreduction set up A scheme of the microscope setup is shown is Figure 2. 6a. An inverted fluorescent microscope (Axio Observer D1, Carl Zeiss , Oberkochen, Germany ) and a metal halide arc light source (X -Cite 120, Excelitas Technologies Corp , MA, USA ) were used in this study. The 550 nm filter (550FSX10 -25, Andover Corporation , NH, USA ) and linear polarizer film (Edmund optics , NJ, USA ) were used for monochromic imaging. Images were captured by a monochromic CCD camera (Axiocam 503 mono, Carl Zeiss). A customized fluores cent filter set, including excitation filter: absorption maximum at 442 nm with bandwidth 150 nm (442HC150 -25, Andover Corporation), dichroic beam splitter: 455 nm (FT 455, Carl Zeiss) and emission filter: absorption maximum at 550 nm with 76 bandwidth 25 nm (BP 550/25 HE, Carl Zeiss), was used for the light -inducing reduction of STC crystals. The size and shape of the photoreduction region was controlled by the confocal aperture (or slit) and the magnification of the objective lens. For all the quantitative a nalysis ( in Figure 2. 8), a circular iris -stop slider (Carl Zeiss) and 20x objective lens (LD Plan -Neofluar 20x/0.4 Korr Ph2, Carl Zeiss) were used, which could provide a minimal photoreduction size of ~100 µm. The 100x objective lens (EC Plan -Neofluar 100x /1.30 Oil, Carl Zeiss) and 50 µm slit (Precision Air Slit, Edmund optics) were used in Figure 2. 4 (2nd, 3rd and 4th examples) . The light shutters and camera for the microscope were controlled by the ‚Macros™ scripting module provided in Zen software (Carl Zeiss). Monochromic images of selected crystals were captured with programed time intervals (2-10 frames per minute, with approximately 100 ms exposure time) throughout the entire experiment. During the initial phase of the experiment, the photoreduction light was on between two image capture 0) was close to the targeted number (~0.15), the photoreduction scripts were manually interrupted. The temperature was measured by a K-type thermocoup le (OMEGA , Surrey, UK) immersed inside the photoreduction buffer within 5 mm from the selected crystals. Samples were cooled (to 0 °C or 8 °C) by streams of cold nitrogen gas evaporated from liquid nitrogen. A lab -build Peltier air -heater was used for heating samples to 36°C temperature. The temperature was controlled manually or by a microcontroller ( Raspberry Pi Foundation , Cambridge, UK ) within ± 0.5°C. 77 Photoreduction conditions of STC crystals. STC crystals were suspended in the degassed photoreductio n buffer (30% (v/v) PEG 400, 200 mM triethanolamine -chloride (TEA) pH 7.9 and 5 mM ZnSO 4, reagents from Sigma -Aldrich) supplied with 10 mM D-glucose (JTBaker , PA, USA) , 50 µg/ml glucose oxidase ( Aspergillus niger , Sigma -Aldrich) and 50 µg/ml catalase (Sigma -Aldrich , MO, USA ) as oxygen scavenger. This crystal suspension was then transferred to a flow -through cuvette (Type 48, 1 mm light path, Fireflysci , NY, USA ) and mounted on the microscope (Figure 2. 6a). Transferring STC crystals from the crystallization condition to the photoreduction condition did not alter the crystal structure (Figure 2. 2a). After crystals were settled, thin (in c-axis) crystals with absorbance (at 550 nm) of ~0.3 were used for photoreduction and imaging. For quantitative experiments, the photoreduction was 550/A550 > 0.15 in order to minimize the accumulation of higher reduction stages. Since the absorbance change for a fully reduced crystal should 550/A550 > 2 (Figure 2. 5 550/A550 = 0.15 was f ar from the beginning of the second reduction stage even in the case of uneven illumination. Image analysis. Crystal images were impor ted into MATLAB (Mathworks , MA, USA ) and processed by a script. The crystal of interest and the subdivisions of that crystal were first manually selected as well as the reference regions for measuring incident light intensity. The script would then align a ll the frames and calculate the absorbance of each section for all the frames. The averaged absorbance of all the pixels in each section was reported and plotted (Figure 2. 8b). 78 Estimation of photoreduction quantum efficiency The excitation light intensity at experimental settings was measured by a photosynthetically active radiation (PAR) probe on LI -250A light meter (LI -COR, NE, USA) as 40 µmole s -1 cm-2. The heme density in STC crystals was estimated based on 8 hemes per unit cell volume (contains 2 STC units) to be ~0.3 mole L -1. The observed photoreduction rate (Figure 2. 6c, 550 = 8 × 10 -4 s-1, initial A 550 = 0.3, reduced oxidized to ~0.1% of total hemes being reduced per second. Assuming all the excitation light (blue) was absorbed by a 100 µm 2 cross -section of a STC crystal with 2 µm thickness, 6 × 10 -17 mole s of hemes would be reduced after the absorption of 4 × 10 -11 mole s of photons every second. This gave us a rough approximation of the photoreduction quantum efficiency on the order of 10 -5 ~ 10 -6. Photoreduction efficiency could be increased by addition o f photosensitizer ( e.g. riboflavin ) into the system, but this approach also introduced diffusive redox carriers that complicate d the analysis of ET in crystals and was therefore abandoned. We also explored the introduction of cross -linked (and thus non -mob ile) ruthenium photosensitizer (Durham et al. , 1989) to STC crystals through a succinimidyl ester functional group, but the modified crystals did not show increased rates of photoreduction (Figure 2. 5d), perhaps because of the inefficient crosslink reactions . Estimation of ket and D in STC crystals using empirical non -adiabatic electron transfer theory Previous work showed that ket between intraprotein hemes is likely to be very rapid, with NMR experiments implying ket >> 104 s-1 and calcu lations based on quantum -79 mechanic s / molecular mechanics suggesting ket > 10 6 s-1 (Harada et al. , 2002; Fonseca et al. , 2009; Jiang et al. , 2017) , we assumed that the rate -limiting step of ET occurs at the interprotein heme -heme step , e.g. between hemes IV and I for the b-axis (as in Figure 2.1c). As a result, in the timescale of interprotein ET, the intraprotein heme redox states will approach equilibrium , i.e. the probability of finding an electron on one of four hemes is determined by th e equilibrium constant (Figure 2.1c). Following this assumption, we concluded that the reversible ( e.g. electron moving left vs right in Figure 2.1c) ET along each axis have identical ket that equals the e ndothermic interprotein step, no matter the directi onality. Because the electrons must transfer to Heme I before the ex othermic interprotein ET to Heme IV, intraprotein ET from Heme IV to Heme I requires the same thermal activation energy as the e ndothermic interprotein ET. In the view of thermodynamic s, at a low reduction stage (fully oxidized with minor one -electron reduced), all STC proteins are identical and there is no 0 or preference for electron s to move along one direction, so the bidirectional ket in each axis should be symmetric. Using the heme -heme distances and redox properties described above, we estimated the rate limiting ket in each crystal axes based on Dutton™s Ruler (Page et al. , 1999): log =150.63.1(+ )/ Equation 2.3 In the case of a- and b-axes (Figure 2.1a), the parameter values used were: heme edge -to-edge distances R = 6 Å (measured from the atomic center -to-center distances between conjugated systems), free energy 0 = +0.12 eV (endothermic ) and reorganization energy eV, resulting a ket = 10 7.5 s-1. For ket in the c-axis (Figure 2. 2c), the parameter 80 values used were : heme edge -to-edge distances R = 14 Å, free energy 0 = +0.08 eV (endothermic ) and reorganization energy eV, resulting a ket = 10 3.0 s-1. The relation between diffusion coefficient D and ket was calculated based on Equation 2.4: =2 Equation 2.4 where the step -length l (3.2 nm for b-axis) corresponding to the repeating distance of STC proteins in STC crystals and dimension n. Figure 2.4 showed the electrons can diffuse in a- and b- axes, while the electron transfer in c-axis was also possible. We thus presumed a three -dimensional diffusion ( n = 3) inside the crystals as the upper limit when we estimat ed the ket from the measured diffusion coefficients D. For estimat ing the one -dimensional (n = 1) mean squared displacement based on theoretical ket (Equation 2.3) , Equation 2.5 was used : =2 Equation 2.5 where is the mean squared displacement and t is the time elapsed. In the scenario of electron diffusion in the b-axis direction across a 100 µm crystal =(100 µ), the ket = 10 7.5 s-1 led to D = 54 µm2 s-1, and the estimated time elapse was approximately 30 s. Curve fitting The absorbance changes after photoreduction were fitted with Equation 2.6: = Equation 2.6 where the change of redox density (u) over time (t) was defined by its one -dimensional spatial ( x) gradient and D. Because molecular oxygen could not be completely excluded 81 in all experiments, a first -order oxidation reaction (rate constant : kox) was added to account for the observed slow rates of electron loss to O 2. The electron density gradient in the first frame of this e xtraction was used as the initial condition. The boundary conditions were =0. Based on these conditions, the time evolution was solved using the MATLAB fipdepefl partial differential equation solver and fitted to the observed kinetics using the MAT LAB filsqcurvefit fl non-linear least squares optimization tool to estimate variables including the diffusion coefficient ( D) and oxidation rate constant ( kox). The ket was then converted from D by Equation 2.4. Kinetic Monte Carlo simulation The effects of random imperfections on diffusion rates were tested by kinetic Monte Carlo (kMC) random -walk simulation in MATLAB . The simulation matrix had a dimension of 50×100×12 units, in a, b and c-axis respectively, which was similar to the aspect ratio of STC crystals. Each unit in the matrix could be either 1 or 0, representing the presence or absence of an electron. A simulated electron in each unit could transfer to an adjacent empty unit or stay with probabilities proportional to th e effective rate constants . In each axis, the probability of moving an ‚electron™ forward or backward were set to be equal, while the probabilities between axes could be customized based on the presumed ket. For mimicking the imperfections, random ly distri buted blocking (or defective) units were incorporated into the simulation matrixes , with fractions of defective sites ranging from 0 to 70% . These blocking units prohibited the transfer of electrons to themselves from any adjacent unit, representing vacanc y of unit cells in an imperfect crystal. Based on the expected ket along each axis of the STC crystal, two scenarios were proposed. If the ET along c-axis was negligible when compared to a- or b-axis (scenario 1), the probabilities 82 were set to be a:b- axes = 1:1 and no ET was allow in c-axis . If ET along the c-axis was comparable to the a- or b-axis (scenario 2), the probabilities were set to be a:b:c- axes = 10:10:1. The initial condition of the simulations was set to have an ‚electron™ density gradient al ong the b-axis, which was extracted from an experimentally photoreduced crystal. Then all ‚electrons™ were randomly moved for 15 ,000 cycles, and the ‚electron™ density redistributions along the b-axis were recorded, subdivided into 10 sections, averaged wi thin each section and fitted with a previous ly described curve fitting script. For scenario s 1 and 2, simulations with defects blocking between 0 and 50% or 70% were tested. Single crystal spectroscopy An u pright fluorescent microscope (LABORLUX 12, Ernst Leitz GmbH , Wetzlar, Germany ) and 10x objective lens (Ph1 10 DL , Nikon, Japan ) were used to find single crystals in a crystal suspension, and the transmitted light through a selected crystal was guided to a spectrometer (LR1, ASEQ -Instrument , Vancouver, Canada ) through fiber optics. A linear polarizer film (Edmund optics) was used for generating polarized incident light. The STC crystal photoreduction on this setup used Filter Cube G (513602, Ernst Leitz Gmb H, Wetzlar, Germany ) and a short arc mercury lamp (USH -102DH, USHIO , Japan ) as light source. Redox titrations The redox titration of an STC solution was performed in 50 mM HEPES pH 8.0 buffer containing a series of redox mediators (Sigma -Aldrich) 10 µM of methyl viologen, anthraquinone -2-sulfonate, anthraquinone -1,5 -disulfonate, 2 -hydroxyl -1,4 -naphthaquinone, 2,5 -dihydroxy -p-benzoquinone, pyocyanin and 1 µM neutral red. 83 Absorbance changes were record ed on a spectrometer ( DU800 , Beckman Coulte r, CA, USA), and a glassy carbon electrode and Ag/AgCl reference electrode were used to record redox potentials. The absorbance change s at 551 nm were plot ted against the ambient redox potential and fitted with a four component s (n = 1) Nernst equation ( Figure 2.3). The redox titration of STC crystals was conducted in a buffer containing 18% PEG 3350, 100 mM TEA pH 7.9 and 5 mM ZnSO 4. High er concentration s (200 µM) of the above redox mediators were used for better stability. The crystal suspension was titrated with so dium dithionite (Sigma -Aldrich) and the redox potential of the system was monitored as above. Once stable redox potentials were achieved, crystal suspensions were rapidly transferred to a flat, N 2-purged, airtight cuvette and the absorbance spectrum of at least two single crystals were obtained using the spectrally -resolved microscope and the setup described above . To avoid sieve effects associated with very large absorbance changes in fully reduced crystals, we used the absorbance changes at 575 nm, rather than the -band peak at 550 nm; over small changes in reduction, these signals were found to be proportional . Table 2.1 X-ray data collection and refinement statistics Beam line ALS 5.0.1 PDB ID 6EE7 Ligands HEM, Zn 2+ Wavelength 1.0 Resolution range (Å) 16.92 Œ 1.39 (1.47 Œ 1.39) Space group P 1 2 1 1 Unit cell dimensions a b c (Å) 24.59 63.86 28.98 (°) 90.00 97.21 90.00 Reflections 83157 (10979) 84 Table 2.1 (cont™d) Multiplicity 4.7 (4.5) Completeness (%) 98.8 (95.2) Mean 10.8 (3.1) Rmerge 0.082 (0.612) Rpim 0.041 (0.310) CC 1/2 0.997 (0.836) Refinement Rwork / Rfree 0.1639 / 0.1832 Number of non -hydrogen atoms 986 macromolecules 671 ligands 175 solvent 140 RMS(bonds) 0.024 RMS(angles) 1.698 Ramachandran favored (%) 98.88 allowed (%) 1.12 outliers (%) 0.00 Average B -factor 20.12 macromolecules 18.42 ligands 14.95 solvent 34.77 Values in parentheses are for the highest -resolution shell. 85 REFERENCES 86 REFERENCES Acbas, G., Niessen, K. A., Snell, E. H., & Markelz, A. G. (2014). Optical measurements of long -range protein vibrations. Nat. Commun. , 5(1), 3076. https://doi.org/10.1038/ncomms4076 Adams, P. D., Afonine, P. V., Bunkóczi, G., Chen, V. B., Davis, I. W., Ech ols, N., Headd, J. J., Hung, L. -W., Kapral, G. J., Grosse -Kunstleve, R. W., McCoy, A. J., Moriarty, N. W., Oeffner, R., Read, R. J., Python -based syst em for macromolecular structure solution. Acta Crystallogr. Sect. D Biol. Crystallogr. , 66(2), 213 Œ221. https://doi.org/10.1107/S0907444909052925 Alric, J., Lavergne, J., Rappaport, F., Verméglio, A., Matsuura, K., Shimada, K., & Nagashima, K. V. P. (2006) . Kinetic performance and energy profile in a roller coaster electron transfer chain: a study of modified tetraheme -reaction center constructs. J. Am. Chem. Soc. , 128(12), 4136 Œ4145. https://doi.org/10.1021/ja058131t Alves, A. S., Costa, N. L., Tien, M., L ouro, R. O., & Paquete, C. M. (2017). Modulation of the reactivity of multiheme cytochromes by site -directed mutagenesis: moving towards the optimization of microbial electrochemical technologies. JBIC J. Biol. Inorg. Chem. , 22(1), 87 Œ97. https://doi.org/1 0.1007/s00775 -016-1409 -0 Arslan, E., Schulz, H., Zufferey, R., Künzler, P., & Thöny -Meyer, L. (1998). Overproduction of the Bradyrhizobium japonicum c-type cytochrome subunits of the cbb 3 oxidase in Escherichia coli . Biochem. Biophys. Res. Commun. , 251 (3), 744 Œ747. https://doi.org/10.1006/bbrc.1998.9549 Aussignargues, C., Pandelia, M. E., Sutter, M., Plegaria, J. S., Zarzycki, J., Turmo, A., Huang, J., Ducat, D. C., Hegg, E. L., Gibney, B. R., & Kerfeld, C. A. (2016). Structure and Function of a Bacteri al Microcompartment Shell Protein Engineered to Bind a [4Fe -4S] Cluster. J. Am. Chem. Soc. , 138(16), 5262Œ5270. https://doi.org/10.1021/jacs.5b11734 Barker, P. D., & Ferguson, S. J. (1999). Still a puzzle: why is haem covalently attached in c-type cytochro mes? Structure , 7(12), R281 -90. https://doi.org/10.1016/S0969 -2126(00)88334 -3 Bortolotti, C. A., Siwko, M. E., Castellini, E., Ranieri, A., Sola, M., & Corni, S. (2011). The reorganization energy in cytochrome c is controlled by the accessibility of the heme to the solvent. J. Phys. Chem. Lett. , 2(14), 1761 Œ1765. https://doi.org/10.1021/jz200734a Breuer, M., Rosso, K. M., & Blumberger, J. (2014). Electron flow in multiheme bacterial cytochromes is a balancing act between heme electronic interaction and redo x potentials. PNAS , 111(2), 611 Œ616. https://doi.org/10.1073/pnas.1316156111 Breuer, M., Zarzycki, P., Shi, L., Clarke, T. A., Edwards, M. J., Butt, J. N., Richardson, D. J., Fredrickson, J. K., Zachara, J. M., Blumberger, J., & Rosso, K. M. (2012). Molecu lar structure and free energy landscape for electron transport in the decahaem cytochrome MtrF. , 40(6), 1198 Œ1203. https://doi.org/10.1042/BST20120139 Byun, H. S., Pirbadian, S., Nakano, A., Shi, L., & El -Naggar, M. Y. (2014). Kinetic Monte Carlo simulations and molecular conductance measurements of the bacterial decaheme cytochrome MtrF. ChemElectroChem , 1(11), 1932 Œ1939. https://doi.org/10.1002/celc.201402211 Cherepanov, D. A., Krishtalik, L. I., & Mulkidjanian, A. Y. (2001). Photosynthetic electron transfer controlled by protein relaxation: Analysis by Langevin stochastic approach. Biophys. J. , 80(3), 1033 Œ1049. https://doi.org/10.1016/S0006 -3495(01)76084 -5 Cr easey, R. C. G., Mostert, A. B., Nguyen, T. A. H., Virdis, B., Freguia, S., & Laycock, B. (2018). 87 Microbial nanowires Œ Electron transport and the role of synthetic analogues. Acta Biomater. , 69, 1 Œ30. https://doi.org/10.1016/j.actbio.2018.01.007 Dalton, E . F., Surridge, N. A., Jernigan, J. C., Wilbourn, K. O., Facci, J. S., & Murray, R. W. (1990). Charge transport in electroactive polymers consisting of fixed molecular redox sites. Chem. Phys. , 141(1), 143Œ157. https://doi.org/10.1016/0301 -0104(90)80026 -T Durham, B., Pan, L. P., Long, J. E., & Millett, F. (1989). Photoinduced electron -transfer kinetics of singly labeled ruthenium bis(bipyridine) dicarboxybipyridine cytochrome c derivatives. Biochemistry , 28(21), 8659Œ8665. https://doi.org/10.1021/bi00447a05 7 Eaton, W. A., & Hochstrasser, R. M. (1967). Electronic spectrum of single crystals of ferricytochrome -c. J. Chem. Phys. , 46(7), 2533 Œ2539. https://doi.org/Doi 10.1063/1.1841081 El-Naggar, M. Y., Wanger, G., Leung, K. M., Yuzvinsky, T. D., Southam, G., Yang, J., Lau, W. M., Nealson, K. H., & Gorby, Y. a. (2010). Electrical transport along bacterial nanowires from Shewanella oneidensis MR-1. PNAS , 107 (42), 18127 Œ18131. https://doi. org/10.1073/pnas.1004880107 Emsley, P., & Cowtan, K. (2004). Coot: model -building tools for molecular graphics. Acta Crystallogr. D. Biol. Crystallogr. , 60, 2126 Œ2132. https://doi.org/10.1107/S0907444904019158 Fonseca, B. M., Saraiva, I. H., Paquete, C. M. , Soares, C. M., Pacheco, I., Salgueiro, C. A., & Louro, R. O. (2009). The tetraheme cytochrome from Shewanella oneidensis MR-1 shows thermodynamic bias for functional specificity of the hemes. J. Biol. Inorg. Chem. , 14(3), 375 Œ385. https://doi.org/10.1007 /s00775 -008-0455 -7 Gorby, Y. A., Yanina, S., McLean, J. S., Rosso, K. M., Moyles, D., Dohnalkova, A., Beveridge, T. J., Chang, I. S., Kim, B. H., Kim, K. S., Culley, D. E., Reed, S. B., Romine, M. F., Saffarini, D. A., Hill, E. A., Shi, L., Elias, D. A., K ennedy, D. W., Pinchuk, G., Watanabe, K., Ishii, S., Logan, B., Nealson, K. H., & Fredrickson, J. K. (2006). Electrically conductive bacterial nanowires produced by Shewanella oneidensis strain MR -1 and other microorganisms. PNAS , 103(30), 11358 Œ11363. htt ps://doi.org/10.1073/pnas.0604517103 Gray, H. B., & Winkler, J. R. (2009). Electron flow through proteins. Chem. Phys. Lett. , 483(1Œ3), 1 Œ9. https://doi.org/10.1016/j.cplett.2009.10.051 Gu, Y., Li, P., Sage, J. T., & Champion, P. M. (1993). Photoreduction of heme proteins: spectroscopic studies and cross -section measurements. J. Am. Chem. Soc. , 115(12), 4993 Œ5004. https://doi.org/10.1021/ja00065a007 Gunner, M. R., & Dutton, P. L. (1989). Temperature and -G° dependence of the electron transfer from BPh·- to QA in reaction center protein from Rhodobacter sphaeroides with different quinones as QA. J. Am. Chem. Soc. , 111(9), 3400 Œ3412. https://doi.org/10.1021/ja00191a043 Harada, E., Kumagai, J., Ozawa, K., Imabayashi, S., Tsapin, A. S., Nealson, K. H., Meyer , T. E., Cusanovich, M. A., & Akutsu, H. (2002). A directional electron transfer regulator based on heme -chain architecture in the small tetraheme cytochrome c from Shewanella oneidensis . FEBS Lett. , 532(3), 333 Œ337. https://doi.org/10.1016/S0014 -5793(02)0 3696-7 Hopfield, J. J. (1974). Electron Transfer Between Biological Molecules by Thermally Activated Tunneling. PNAS , 71(9), 3640 Œ3644. https://doi.org/10.1073/pnas.71.9.3640 Jiang, X., Futera, Z., Ali, M. E., Gajdos, F., von Rudorff, G. F., Carof, A., Bre uer, M., & Blumberger, J. (2017). Cysteine Linkages Accelerate Electron Flow through Tetra -Heme Protein STC. J. Am. Chem. Soc. , jacs.7b08831. https://doi.org/10.1021/jacs.7b08831 Kabsch, W. (2010). Integration, scaling, space -group assignment and post -refi nement. Acta Crystallogr. Sect. D Biol. Crystallogr. , 66(2), 133 Œ144. https://doi.org/10.1107/S0907444909047374 88 Kerfeld, C. A., & Erbilgin, O. (2015). Bacterial microcompartments and the modular construction of microbial metabolism. , 23(1), 22 Œ34. https://doi.org/10.1016/j.tim.2014.10.003 Kokhan, O., Ponomarenko, N. S., Pokkuluri, P. R., Schiffer, M., Mulfort, K. L., & Tiede, D. M. (2015). Bidirectional photoinduced electron transfer in Ruthenium(II) -Tris -bipyridyl -Modified PpcA, a multi -heme c -Type cytochrome from Geobacter sulfurreducens . J. Phys. Chem. B , 119(24), 7612 Œ7624. https://doi.org/10.1021/jp511558f Lampa -Pastirk, S., Veazey, J. P., Walsh, K. A., Feliciano, G. T., Steidl, R. J., Tessmer, S. H., & Reguera, G. (2016). Thermally activated charge transport in microbial protein nanowires. Sci. Rep. , 6(1), 23517. https://doi.org/10.1038/srep23517 Lassen, L. M., Nielsen, A. Z., Ziersen, B., Gnanasekaran, T., Møller, B. L., & Jensen, P. E. (2014). Redirecting Photosynthetic Electron Flow into Light -Driven Synthesis of Alternative Products Including High -Value Bioactive Natural Compounds. ACS Synth. Biol. , 3(1) , 1 Œ12. https://doi.org/10.1021/sb400136f Leung, K. M., Wanger, G., El -Naggar, M. Y., Gorby, Y., Southam, G., Lau, W. M., & Yang, J. (2013). Shewanella oneidensis MR-1 Bacterial Nanowires Exhibit p-Type, Tunable Electronic Behavior. Nano Lett. , 13(6), 2407 Œ2411. https://doi.org/10.1021/nl400237p Leys, D., Meyer, T. E., Tsapin, A. S., Nealson, K. H., Cusanovich, M. a, & Van Beeumen, J. J. (2002). Crystal Structures at Atomic Resolution Reveal the Novel Concept of fiElectron -harvestingfl as a Role for the Small Tetraheme Cytochrome c. J. Biol. Chem. , 277(38), 35703 Œ35711. https://doi.org/10.1074/jbc.M203866200 Lovley, D. R., & Malvankar, N. S. (2015). Seeing is believing: Novel imaging techniques help clarify microbial nanowire structure and function. Environ. M icrobiol. , 17, 2209 Œ2215. https://doi.org/10.1111/1462 -2920.12708 Lubner, C. E., Kn örzer, P., Silva, P. J. N., Vincent, K. A., Happe, T., Bryant, D. A., & Golbeck, J. H. (2010). Wiring an [FeFe] -Hydrogenase with Photosystem I for Light -Induced Hydrogen Pr oduction. Biochemistry , 49(48), 10264 Œ10266. https://doi.org/10.1021/bi1016167 Malvankar, N. S., & Lovley, D. R. (2014). Microbial nanowires for bioenergy applications. Curr. Opin. Biotechnol. , 27, 88 Œ95. https://doi.org/10.1016/j.copbio.2013.12.003 Malvan kar, N. S., Vargas, M., Nevin, K., Tremblay, P. -L., Evans -Lutterodt, K., Nykypanchuk, D., Martz, E., Tuominen, M. T., & Lovley, D. R. (2015). Structural basis for metallic -like conductivity in microbial nanowires. MBio , 6(2), e00084. https://doi.org/10.112 8/mBio.00084 -15 Marcus, R. A., & Sutin, N. (1985). Electron transfers in chemistry and biology. Biochim. Biophys. Acta - Rev. Bioenerg. , 811(3), 265 Œ322. https://doi.org/10.1016/0304 -4173(85)90014 -X Martínez, E., Marian, J., Kalos, M. H., & Perlado, J. M. (2008). Synchronous parallel kinetic Monte Carlo for continuum diffusion -reaction systems. J. Comput. Phys. , 227 (8), 3804 Œ3823. https://doi.org/10.1016/j.jcp.2007.11.045 Meade, T. J., Gray, H. B., & Winkler, J. R. (1989). Driving -force effects on the rate of long-range electron transfer in Ruthenium -modified cytochrome c. J. Am. Chem. Soc. , 111(12), 4353 Œ4356. https://doi.org/10.1021/ja00194a030 Meinhold, L., Merzel, F., & Smith, J. C. (2007). Lattice yynamics of a protein crystal. Phys. Rev. Lett. , 99(13), 138101. https://doi.org/10.1103/PhysRevLett.99.138101 Meyer, T. E., Tsapin, A. I., Vandenberghe, I., De Smet, L., Frishman, D., Nealson, K. H., Cusanovich, M. A., & Van Beeumen, J. J. (2004). Identification of 42 possible cytochrome c genes in the Shewanella oneidensis genome and characterization of six soluble sytochromes. Omi. A J. Integr. Biol. , 8(1), 57 Œ77. 89 https://doi.org/10.1089/153623104773547499 Moser, C. C., Anderson, J. L. R., & Dutton, P. L. (2010). Guidelines for tunneling in enzymes. Biochim. Biophys. Acta - Bioenerg. , 1797(9), 1573 Œ1586. https://doi.org/10.1016/j.bbabio.2010.04.441 Moser, C. C., Keske, J. M., Warncke, K., Farid, R. S., & Dutton, P. L. (1992). Nature of biological electron transfer. Nature , 355(6363), 796Œ802. https://doi.org/10.1038/355796a0 Muegge, I., Qi, P. X., Wand, A. J., Chu, Z. T., & Warshel, A. (1997). The reorganization energy of cytochrome c revisited. J. Phys. Chem. B , 101(5), 825 Œ836. https://doi.org/10.1021/jp962478o Ohnishi, T. (1998). Iron Œsulfur clusters/semiquinones in Complex I. Biochim. Biophys. Acta - Bioenerg. , 1364(2), 186 Œ206. https://doi.org/10.1016/S0005 -2728(98)00027 -9 Ort, D. R., Merchant, S. S., Alric, J., Barkan, A., Blankenship, R. E., Bock, R., Croce, R., Hanson, M. R., Hibberd, J. M., Long, S. P., Moore, T. A., Moroney, J., Niyogi, K. K., Parry, M. A. J., Peralta -Yahya, P. P., Prince, R. C., Redding, K. E., Spalding, M. H., van Wijk, K. J., Vermaas, W. F. J., von Caemmerer, S., Weber, A. P. M., Yeates , T. O., Yuan, J. S., & Zhu, X. G. (2015). Redesigning photosynthesis to sustainably meet global food and bioenergy demand. PNAS , 112(28), 8529 Œ8536. https://doi.org/10.1073/pnas.1424031112 Page, C. C., Moser, C. C., Chen, X., & Dutton, P. L. (1999). Natur al engineering principles of electron tunnelling in biological oxidation -reduction. Nature , 402(6757), 47 Œ52. https://doi.org/10.1038/46972 Pirbadian, S., Barchinger, S. E., Leung, K. M., Byun, H. S., Jangir, Y., Bouhenni, R. A., Reed, S. B., Romine, M. F. , Saffarini, D. A., Shi, L., Gorby, Y. A., Golbeck, J. H., & El -Naggar, M. Y. (2014). Shewanella oneidensis MR-1 nanowires are outer membrane and periplasmic extensions of the extracellular electron transport components. PNAS , 111(35), 12883 Œ12888. https:/ /doi.org/10.1073/pnas.1410551111 Pirbadian, S., & El -Naggar, M. Y. (2012). Multistep hopping and extracellular charge transfer in microbial redox chains. Phys. Chem. Chem. Phys. , 14(40), 13802. https://doi.org/10.1039/c2cp41185g Polizzi, N. F., Skourtis, S . S., & Beratan, D. N. (2012). Physical constraints on charge transport through bacterial nanowires. Faraday Discuss. , 155(6), 43 Œ61. https://doi.org/10.1039/C1FD00098E Reguera, G., McCarthy, K. D., Mehta, T., Nicoll, J. S., Tuominen, M. T., & Lovley, D. R . (2005). Extracellular electron transfer via microbial nanowires. Nature , 435 (7045), 1098 Œ1101. https://doi.org/10.1038/nature03661 Rengaraj, S., Haddad, R., Lojou, E., Duraffourg, N., Holzinger, M., Le Goff, A., & Forge, V. (2017). Interprotein electron transfer between FeS -protein nanowires and oxygen -tolerant NiFe hydrogenase. Angew. Chemie Int. Ed. , 56(27), 7774 Œ7778. https://doi.org/10.1002/anie.201702042 Rutherford, A. W., Osyczka, A., & Rappaport, F. (2012). Back -reactions, short -circuits, leaks and other energy wasteful reactions in biological electron transfer: Redox tuning to survive life in O 2. FEBS Lett. , 586(5), 603 Œ616. https://doi.org/10.1016/j.febslet.2011.12.039 Schenck, C. C., Parson, W. W., Holten, D., Windsor, M. W., & Sarai, A. (1981). Temperature dependence of electron transfer between bacteriopheophytin and ubiquinone in protonated and deuterated reaction centers of Rhodopseudomonas sphaeroides . Biophys. J. , 36(3), 479 Œ489. https://doi.org/10.1016/S0006 -3495(81 )84747 -9 Smith, D. M. A., Rosso, K. M., Dupuis, M., Valiev, M., & Straatsma, T. P. (2006). Electronic coupling between heme electron -transfer centers and its decay with distance depends strongly on relative orientation. J. Phys. Chem. B , 110(31), 15582 Œ15588. https://doi.org/10.1021/jp057068r Strycharz -Glaven, S. M., Snider, R. M., Guiseppi -Elie, A., & Tender, L. M. (2011). On the electrical 90 conductivity of microbial nanowires and biofilms. Energy Environ. Sci. , 4(11), 4366. https://doi.org/10.1039/c1ee0175 3e Suslick, K., & Watson, R. (1992). The photochemistry of chromium, manganese, and iron porphyrin complexes. New. J. Chem , 16, 633 Œ642. https://doi.org/1011/691.9466 Teravest, M. A., & Ajo -Franklin, C. M. (2016). Transforming exoelectrogens for biotechnol ogy using synthetic biology. Biotechnol. Bioeng. , 113(4), 687 Œ697. https://doi.org/10.1002/bit.25723 Tipmanee, V., Oberhofer, H., Park, M., Kim, K. S., & Blumberger, J. (2010). Prediction of reorganization free energies for biological electron transfer: A comparative study of Ru -modified cytochromes and a 4 -helix bundle protein. J. Am. Chem. Soc. , 132(47), 17032 Œ17040. https://doi.org/10.1021/ja107876p Tsapin, A. I., Vandenberghe, I., Nealson, K. H., Scott, J. H., Meyer, T. E., Cusanovich, M. A., Harada, E., Kaizu, T., Akutsu, H., Leys, D., & Van Beeumen, J. J. (2001). Identification of a small tetraheme cytochrome c and a flavocytochrome c as two of the principal soluble cytochromes c in Shewanella oneidensis Strain MR1. Appl. Environ. Microbiol. , 67(7), 3236 Œ3244. https://doi.org/10.1128/AEM.67.7.3236 -3244.2001 Verburg, F. (2012). Plug and play photosynthesis. Chem. Ind. , 76(9), 42 Œ45. https://doi.org/10.1002/cind.7609_12.x Wang, S.F., Chen, T., Zhang, Z.L., Shen, X.C., Lu, Z.X., Pang, D.W., & Wo ng, K.Y. (2005). Direct electrochemistry and electrocatalysis of heme proteins entrapped in agarose hydrogel films in room-temperature ionic liquids. Langmuir , 21(20), 9260 Œ9266. https://doi.org/10.1021/la050947k Winkler, J. R., & Gray, H. B. (2014). Long -range electron tunneling. J. Am. Chem. Soc. , 136(8), 2930 Œ2939. https://doi.org/10.1021/ja500215j Winn, M. D., Ballard, C. C., Cowtan, K. D., Dodson, E. J., Emsley, P., Evans, P. R., Keegan, R. M., Krissinel, E. B., Leslie, A. G. W., McCoy, A. , McNicholas, S. J., Murshudov, G. N., Pannu, N. S., Potterton, E. A., Powell, H. R., Read, R. J., Vagin, A., & Wilson, K. S. (2011). Overview of the CCP 4 suite and current developments. Acta Crystallogr. Sect. D Biol. Crystallogr. , 67(4), 235 Œ242. https: //doi.org/10.1107/S0907444910045749 Yates, M. D., Golden, J. P., Roy, J., Strycharz -Glaven, S. M., Tsoi, S., Erickson, J. S., El -Naggar, M. Y., Calabrese Barton, S., & Tender, L. M. (2015). Thermally activated long range electron transport in living biofil ms. Phys. Chem. Chem. Phys. , 17(48), 32564 Œ32570. https://doi.org/10.1039/C5CP05152E Zhang, Y., & Skolnick, J. (2004). Scoring function for automated assessment of protein structure template quality. Proteins Struct. Funct. Bioinforma. , 57(4), 702 Œ710. htt ps://doi.org/10.1002/prot.20264 91 CHAPTER 3: FUNCTIONALIZATION OF NANOTUBE -FORMING PROTEINS WITH COVALENTLY ATTACHED HEME Co-authors: Jingcheng Huang, Eric J. Young, Bryan H. Ferlez, Clement Aussignargues, Cheryl A. Kerfeld, David M. Kramer, Daniel C. Ducat I would like to thank Clement Aussignargues for generating the Ho-5815 Y41A variant . I would like to acknowledge Eric J. Young, Bryan H. Ferlez and Clement Aussignargues for taking transmission electron microscopy images. I would also like to thank Daniel C. Ducat , David M. Kramer and Cheryl A. Kerfeld for the experimental designs and wr iting of the manuscript. 92 Abstract Heme is a versatile redox cofactor that has considerable potential for synthetic biology and bioelectronic applications. Being able to functionalize non-heme -binding proteins with covalently bound heme moieties in vivo could expand the variety of bioelectronic materials, particularly if hemes could be attached at defined locations so as to facilitate position -sensitive processes like electron transfer. In this study, we utilized cytochrome maturation system I to incorporat e hemes on self -assembling shell proteins. We found substitution of three amino acids within the backbone of the protein promoted heme attachment with high occupancy. Spectroscopy measurements suggested these modified proteins bind c-type low -spin hemes wi th low redox midpoint potentials ( -210 mV vs SHE). Heme -modified shell proteins partially retained their self -assembly properties, including hexamerization. However, some higher -order assembly features were altered in heme -bound shell proteins, potentially as a result of changes in the most favorable forms of hexamer -hexamer contacts. Introduction Efficie nt electron transport through protein redox carriers is required for many energy -converting biological processes. In order to achieve rapid electron transfer from electron donor towards acceptors, the redox cofactors are precisely placed at defined locations and fine -tuned with specific redox energy profiles . The electron transport chain in photosystems is an example (Bl ankenship , 2014). It is well -known that t he positioning of redox cofactors must be controlled with Angstrom precision, otherwise the rate of electron transfer can decrease by orders of magnitude , dramatically influencing the 93 number and proportion of electr ons that reach the terminal end -product (Rutherford, Osyczka, and Rappaport , 2012). Accurate positioning of redox cofactors is also critical when designing redox proteins (Page et al. , 1999) for the growing field of bioelectronics, where biological molecules are used as components (e.g. bionanowires, biocapacitors, light capture or biosensors) in novel electronics devices. One design principle for producing a long -range electron transport dev ice, inspired by the natural microbial nanowire, is to use a self -assembl ing protein scaffold to spatially organize redox domains int o close proximity and with suitable orientations. For example, the redox biofilm reported by Forge and coworkers utilized a fusion protein containing elements of a rubredoxin (redox domain) and prion (self -assembling domain ) (Altamura et al. , 2017) . This approach produced a redox active biofilm but required a flexible linker to connect two protein domains, and such unstructured domains can cause unwanted mobility between the tethered redox centers (Altamura et al. , 2017). This dynamic positioning creates additional layers of complexity in the analysis and optimization of long -range electron transport in bioelectronic ma terials. Therefore, strategies that allow positioning of redox co -factors in structurally defined locations of a biomaterial are desired . In this study, we attempted to attach heme as a redox cofactor onto targeted positions within the existing backbone o f self -assembling proteins in vivo . Heme is a universal biological redox cofactor that is utilized in a diverse array of endogenous electron transfer reactions (Chapman, Daff, and Munro , 1997). c-Type heme, found in cytochrome c, has two peripheral substitution s (position 2 and 4) that form thioether bonds with two cysteine residues within the heme binding motif (mostly CxxCH, where C 94 is cysteine, H is histidine and x can be any residue, Figure 3.1A ). The histidine in the motif acts as one of two axial ligands to the iron center of the heme . This covalent linkage is believed to influence protein stability, tune redox potentials, and support conformations that allow hemes to be surface -exposed (Bowman and Bren , 2008). The covalent attachment of heme to the binding motif is catalyzed by a dedicated system called the cytochrome c maturation (Ccm) system (Kranz et al. , 2009). Ccm system I (Ccm -I) is found in Gram -negative bacteria (Thony -Meyer et al. , 1995), has the broadest protein -substrate diversity, and can recognize the heme binding motif within a large variety of apo-cytochromes (Allen and Ferguson , 2006), making it a promising posttranslational modification tool for protein engineering. Ccm -I requi res the Sec protein secretion pathway to translocate nascent polypeptide chains through the plasma membrane into the periplasm before a heme will be attached (Kranz et al. , 2009) . It has been hypothesized that Ccm -I will attempt to attach a heme to any Cxx CH motif of any protein that is delivered by the Sec pathway (Allen and Ferguson , 2006). Thus, it would be feasible to attach a heme to any target protein that is secreted by mutating the codons for three amino acids, so they are encoding CxxCH, assuming t hat the residue substitutions and attachment of the heme permit the protein to fold properly. This approach would potentially allow redox features to be appended onto functional proteins at defined locations and would avoid the sophisticated de novo designing of heme binding pockets (Figure 3.1A). In this work, I utilized the Ccm -I to attach hemes onto the self -assembling shell proteins of the bacterial microcompartment (BMC). BMCs are proteinaceous bacterial organelles that encapsulate enzymes that catalyze specialized reactions by the shell 95 proteins (Kerfeld et al. , 2018; Cheng et al. , 2008; Yeates et al. , 2008). Aside from spherical compartments, these shell proteins can also self -assemble into other super -mol ecular structures naturally ( e.g. sheets, strips , or tubes; Figure 3.1A), many of which could be ideal scaffolds for other biological and nanotechnology applications (Young et al. , 2017; Lassila et al. , 2014; Noël, Cai, and Kerfeld , 2016; M. J. Lee et al. , 2018). Towards the construction of synthetic bionanowires, I target ed two shell proteins known to form nanotubes in vivo and in vitro . This Chapter demonstrate s the attachment of one or two hemes directly onto the shell protein subunits, and present evide nce that appending these redox cofactors does not eliminate their self -assembly capabilities. Figure 3.1 Design of heme -functionalized BMC -H shell proteins 96 (A) Cartoon schem atic showing a hypothetical pathway of a heme from attachment on a shell protein to the formation of large structures . (B) Ramachandran plot of torsion angles within the heme binding motif s of natural ly-occurring cytochrome c structures (n = 249 heme -binding motifs) . Each color represents one amino acid in the motif . (C) Parti al p rotein sequences of Ho-5815 and RmmH shell proteins . Secondary structures are highlighted (green: helix, yellow: loop, cyan: sheet) . Three amino acid residues at the indicated location s (black ticks) at one of the mutation sites 1,2,3 and 5 were change d to a heme binding motif (CxxCH) . Six amino acids were inserted at mutation site 4 to introduce the heme binding motif at the loop region. (D) Molecul ar dynamic s model BMC -H variant R1 (heme at mutation site 1 in RmmH). Hemes are shown in red. Results Design of h eme attach ment sites on shell proteins Among the BMC shell proteins, one protein family is known to form homo -hexamers (BMC -H) that natively self -assemble into large structures (e.g. , the facets of endogenous BMCs) (Kerfeld et al. , 2018; Sutter et al ., 2017; Kerfeld et al. , 2005; Axen, Erbilgin, and Kerfeld , 2014). Many reports have shown that when BMC -H proteins are expressed individually outside of their native context, they c an form other useful higher -order protein assemblies, such as tubes or sheets (Young et al. , 2017; Noël, Cai, and Kerfeld , 2016; Sutter et al. , 2016; Lassila et al. , 2014). However, it is also known that these higher -order assemblies are sensitive to minor changes within individual shell proteins, especially to the amino acids that mediate the lateral hexamer -hexamer interactions (Young et al. , 2017). For example, the wild -type (WT) BMC -H protein from Haliangium ochraceum (Ho-5815) forms sheets in vitro (Sutter et al. , 2016), but the Y41A variant forms nanotubes (Figure 3. 2). We selected two nanotube -forming BMC -H proteins as candidates for appending heme cofactors, the Y41A variant of Ho-5815 and MSM_0272 (also known as RmmH, Rhodococcus and Mycoba cterium microcompartment BMC -H) (Mallette and Kimber , 2017; Noël, Cai, and Kerfeld , 2016). 97 Figure 3.2 Nanotubes formed by the Y41A variant of MicH during cytosolic overexpression (A-C) Three representative transmission electron micrographs of thin secti on images of E. coli are shown. ( D) The Y41 residue of WT Ho -5815 is highlighted with red stick -representation in a hexameric structure model (pdb: 5djb). To determine potential heme attachment sites, we first searched the preferred secondary structure o f heme binding motifs within natural cytochromes deposited in the Protein Databank (PDB) to gain insight into structural features that naturally support a ligated heme. The attachment of a rigid heme to the CxxCH motif on an unfolded protein likely constrains the possible secondary structures (Negron, Fufezan, and Koder , 2009). 98 The torsion angles of the residues in the heme binding motif (x 1C1x2x3C2H, where x 1 is the residue before the heme binding motif, and the x 2, x3 are the first and the second residue s between the two C) of 180 c-type cytochromes within the PDB were plotted in a Ramachandran plot (Figure 3.1B, Table 3.1). As shown, the torsion angles of residue s C1x2x3C2 in the motif are likely to be a helical structure ( -50°< ° -50°), although the torsion angles in x 1 or H have more variability (Figure 3.1B; red or dark blue , respectively ). Based on this information about torsion angles in natural cytochromes, we searched for similar secondary structures within the conserved pfam0 0936 domain of BMC -H proteins. After checking for potential intra -/inter -protein clashes after heme attachment, we identified four substitution sites that are compatible with the introduction of heme binding motifs (Figure 3.1C). At each mutation site, three (two for RmmH site 5 ) point -substitution were introduced to create the CxxCH m otif, while the residues between the cysteines remained unmodified. In addition, a loop region (site 4) (Jorda et al. , 2016) was chosen for inserting the heme binding motif. Molecular dynamic s simulations (Figure 3. 3A) of the predicted motion of hemes atta ched at these 5 sites ( i.e. , the RMSD of the central iron atom) were comparable to the flexibility of atoms within the protein backbone itself (~1 -2 Å in 800 ps of simulated time; Figure 3. 3B), suggesting these sites offer a relatively structurally -defined location for heme attachment. 99 Figure 3.3 Molecular dynamics models of all monoheme shell -cytochromes (A) Top-view, side -view, cross -section, and zoom -in view (top to bottom) of the predicted structures for each hexameric shell -cytochrome. In the top -view, Mutation sites 1 -4 have the convex side facing out, Mutation site 5 have the concave side facing out. In the side -view, the convex site is oriented upwards. (B) Root -mean -square deviation (RMSD) of atomic positions over 0.8 ns of simulation for hexameric shell -cytochromes . Cqf, Cqg, Cqh, Cqi and Cqj are the averaged RMSD 100 of qf-carbon to qj-carbon (if it exists) of each residue . Fe is the averaged RMSD of iron in six porphyrins for each hexameric shell -cytochrome . We performed mutagenesis on DNA encoding two BMC -H candidates: the Y41A variant of Ho-5815 (H-variants) and the WT RmmH ( R-variants) (Figure 3.1C). Based on crystal structures of the wild -type shell proteins , sites 1,2,3 and 4 are located at the convex side of the hexamer (Figure 3.1D), while a heme coordinated at site 5 is predicted to face the concave side (Figure 3. 3A). An N-terminal periplasmic targeti ng signal peptide from the small tetraheme cytochrome (STC) in Shewanella oneidensis MR-1 (SO_2727) (Leys et al. , 2002) and a C-termin us 6xHis -tag were also appended to complete each expression construct (Table 3.2). Each construct was expressed in Escherichia coli BL21 (E. coli ) with the Ccm -I gene cassette encoded by the pEC86 plasmid (Arslan et al. , 1998). Covalent heme attachment to s hell proteins E. coli expressing four of the five modified shell proteins of both the H- and R- variants (sites 1,2,4 and 5) exhibited vivid red pigmentation of the cell pellets. Site 3 for both H- and R- variants did not yield red pigment and thus were excluded from further investigation. A ffinity chromatography was used to isolate each of the eight recombinant proteins, and the purified fractions exhibited characteristic spectral features of a covalently -bound, low -spin, c-type heme in a ferric state (Butt and Keilin , 1962) , including an intense Soret band at 40 7 nm as well as a broad Q -band peaking at 530 nm (Figure 3.4A). As expected, upon reducing the purified protein with sodium dithionite, th e Soret peak shifted to 41 6 nm and two Q -bands emerged at 52 1 nm and 550 nm. In the red region, H1, H2 and R1 (Each sub script indicates one heme binding motif at the corresponding site in Figure 3.1C) each possesses a small absorbance band around 650 101 nm, suggesting the presence of a high -spin heme with weak -field ligands ( e.g. , water), whereas this band is absent in other samples, indicating H4, H5, R2, R4 and R5 contain mainly low -spin hemes with strong -field ligand s. Although the functional group that could serve as the sixth heme ligand was not certain, protein residues ( e.g. , His, Lys) or imidazole in the purification buffer could potentially serve as the strong -field axial ligand. Altogether, these spectral features indicate the presence of c-type hemes (Butt and Keilin , 1962) in these modified shell proteins. (Hereafter, these recombinant proteins are referred to as shell -cytochromes). The identity of shell -cytochromes and the occupancy of covalently attached heme s were further confirmed by intact protein mass spectrometry (MS) under denaturing conditions (Figure 3. 4B). In all eight purif ied shell -cytochromes, the expected molecular weights of monomeric holo -proteins were observed (Figure 3. 4B). Additionally, MS revealed that some purified shell -cytochrome samples contained a relatively minor fraction of apo-proteins without attached heme (Figure 3. 5). In H1, H2, H5, R1 and R5, the relative amount of apo-protein was measurable and the percentages of the holo -protein in the total full -length shell proteins were 66%, 69%, 86%, 88% and 76% respectively. In H4, R2 and R4 shell -cytochrome sample s, a minor amount of apo-protein was detectable, but it could not be quantified accurately by spectrometry . 102 Figure 3.4 Modified shell proteins contain redox -active hemes (A) UV-Vis spectra of purified shell -cytochromes under oxidiz ing (upper) or reduc ing (lower) condition . The 480 nm Œ 575 nm region and the 580 nm Œ 680 nm region are also plo tted with 5x or 50x amplification. There are vertical offsets between traces. (B) Pro tein mass spectrometry of purified shell -cytochromes. Expected holo -protein ions are marked with charge states in red triangles. The percentage of holo -protein in all full -length proteins of detectable samples are labeled under the sample name. See Figure 3.5 for unlabeled peaks. 103 Figure 3.5 Protein mass spectrometry of purified shell -cytochromes The mass spectra s how the same results as Figure 3.4 B but marked for the apo -protein (blue triangles) . Additionally, a few shell -cytochromes designed using RmmH as the base BMC -H exhibited degradation products lacking the first 27 amino acids on the N -terminus of the protein (magenta triangles). The redox midpoint potentials of two representative shell -cytochromes R1 and R5 were determined by s pectrophotometric redox titrations (Leslie Dutton , 1978). Fitting the titration curve data with a one -component single -electron (n = 1) Nernst equation yielded a Em' (vs SHE) of -215 ± 2 mV and -219 ± 3 mV, for R1 and R5 respectively ( Figure 3. 6). These E m' values are close to the model heme peptide microperoxidase with bis -imidazole ligand s (-206 mV (Zamponi et al ., 1990)), suggesting a fully solvent exposed c-type heme with two imidazole (or similar) ligands. However, the titration cu rves appeared slightly broader than the one -component fitting curve, potentially indicating the hemes 104 were in an inhomogeneous microenvironment (Battistuzzi et al. , 2002) (e.g. , some hemes may have contained a different axial ligand). Figure 3.6 Equilibrium redox titration of representative shell -cytochromes monitored by UV -Vis spectrophotomet ry (A/C) UV -Vis absorption spectra of R1 and R5 at different redox potentials ( vs SHE). Shell -cytochromes were first t itrated with sodium dithionite to complete ly reduc e the attached heme (dashed traces) , then potassium ferricyanide was used to oxidize the sample (solid traces). The absorbance values at wavelengths between 500 Œ 650 nm are magnified to highlight these spectral features ( 5x; inset) . (B/D) Titration curves of R1 (top) and R5 (bottom) . The signal of reduced 105 heme was calculated by the intensity of A551 Œ A561, with the minimal absorption normalized to zero). Both potassium ferricyanide (O) and sodium single component (black dashed line ) or two components ( black solid line ). Each component of the two -component fitting was plotted in blue (low potential) or green (high potential) with the estimated ratio of each single component displayed . The two -component Em' for R1 are -144 ± 12 mV (23%) and -231 ± 4 mV (77%); for R5, -138 ± 7 mV (32%) and -246 ± 4 mV (68%). To increase heme den sity, shell protein variants with two heme binding motifs in each monomer were constructed ( H2+5 , H2+4 , H4+5 , R2+5 , R2+4 and R4+5 ). Upon expression, three ( H2+5 , R2+5 and R4+5 ) variants yield vivid red cell pellets and the shell -cytochromes were purified from these samples for further analysis. UV -Vis spectra suggested these diheme shell -cytochromes also possess c-type low -spin hemes, with comparable features to the shell -cytoc hromes ligated with a single heme (Figure 3. 7). Figure 3.7 Features of purified s hell -cytochrome proteins contain ing two heme attachment sites A) UV-Vis spectra of the indicated shell -cytochromes under oxidizing (upper) or reducing (lower) condition s. Insets correspond to magnified views of the wavelengths between 480 -520 nm (5X magnification) and 580 -700 nm (50X). (B) Protein mass spectrometry of purified shell -cytochromes H2+5 (top), R2+5 (middle), and R4+5 (bottom) . Expected holo -protein ions are marked with charge states in red stars . The calculated percentage s of diheme protein (red star), monoheme 106 protein (red triangle) and apo-protein (blue triangle) detected in all full -length protein s of each sample are displayed . Shell -cytochromes form hexamer -like oligomers BMC -H proteins typically form stable, 6 -fold rotational symmetric hexamers (Figure 3.1A), and the formation of these hexagonal tiles is the first step before self -assembly into higher -order structures (Young et al. , 2017; Kerfeld et al. , 2005; Sutter et al ., 2017; Greber, Sutter, and Kerfeld , 2019; Kerfeld et al. , 2018). We next examined whether shell -cytochromes retain the capacity to oligomerize like WT shell proteins. On native polyacrylamide gel electrophoresis (PAGE), the heme -attached variants have similar or slightly greater mobilities in comparison to their WT ( i.e. , unmodified) variant s (Figure 3.8A, Figure 3.9A). After calculating the theoretical net charges for all samples, we found the mobilities of the major bands of shell -cytochrome samples followed the trend of net charges. This result suggested the shell -cytochromes formed structures with similar size s to the hexa meric WT proteins . 107 Figure 3.8 Shell -cytochromes retain the capacity to oligomerize (A) Native PAGE and (B) native PAGE with 4 M urea show the mobility of H-/R-variants and their corresponding WT proteins. Lanes of H-/R-variants are show n as unstained where the red is the natural color of hemes. WT Ho -5815 and RmmH as well as the molecule weight marker BSA ( Bovine Serum Albumin , monomer: 66 kDa, dimer: 132 kDa) were overlapped in the same gel after staining. The calculated net charge for shell -cytochrome monomers are plotted as black bars . The ‚0™ charge is aligned to the beginning of resolving gel and the vertical length is stretched to allow RmmH to align to the gel band. The pink arrows point to protein degradation products. (C) DLS of H-/R-variants compare d with corresponding WT proteins and STC. The heights of bars represent the relative mass distribution. 108 Figure 3.9 Shell -cytochromes retain capacity to oligomerize into hexamers and higher -order assemblies (A) Native PAGE and (B) native PAGE with 4 M urea show the mobility of H-/R-variants and their corresponding wild -type proteins . Red arrows point to expected full -length holo -shell -cytochromes. Pink arrows point to degraded holo -shell -cytochromes detected with in the RmmH -derived samples ( see Figure 3.8). Cyan arrows point to putative full -length apo-shell -cytochromes. (C) DLS of H-/R-variants in the presence of 4 M urea were compare d against the corresponding WT BMC -H proteins . The heights of bars represent the relative mass distribution s. The expected 3.5 nm radii for a single hexamer is shown as a red dash -line. We noticed that most of the shell -cytochromes (except H5) did not yield sharp bands on native PAGE (Figure 3. 8A). To disrupt any potential lateral, inter -hexamer associations, we supplemented 4 M urea to the gels, and found this modification yielded sharp migration bands with a similar net charge -dependent pattern (Figure 3. 8B, Figure 3.9B). Other than the expected red bands, we observed several secon dary bands on the native PAGE, including a fast -moving heme -containing product (Figure 3. 8A, 3.3B, Figure 3.9A, 3.9B pink arrows ), which MS analysis confirmed to be a degradation product lacking the first 27 amino acids from N -terminal (Figure 3. 5, Table 3.2). Finally, a band that did 109 not contain heme was observed, consistent with the fraction of apo-proteins also identified by MS ( Figure 3.9 A, 3.9 B cyan arrows, Table 3.2). We utilized dynamic light scattering (DLS) for further evaluation of the par ticle size distribution of shell -cytochrome samples. All the shell -cytochrome s have similar ( e.g. H5) or slightly larger average particle size compared to the WT hexamer, which has a radius of 3.5 nm based on X -ray crystallography structures (Mallette and Kimber , 2017) (Figure 3.8C). By contrast, the radius of STC , which has a similar molecular weight (12,122 Da) to a BMC -H monomer, but does not oligomerize, was measured to be < 2 nm by DLS (Figure 3. 8C). Nearly all of the shell -cytochrome samples (expect H5 and R5) exhibited fractions of proteins with properties consistent with much larger radii (>10 nm) and attributed to higher -order shell protein assembly. By contrast, DLS analysis of the same samples in a buffer that disrupts hexamer -hexamer interactions ( i.e. , with 4 M urea) favored a single sharp peak of particle radii at 3 -4 nm across for all shell -cytochromes (Figure 3.9C). Taken together, DLS measurem ents agreed with the PAGE results that shell -cytochromes retained the ability to form hexamers. The presence of large particles in DLS and slow -migrating bands in native PAGE suggested there were higher -order assemblies that were preferentially disrupted b y the addition of 4 M urea. However, these analyses could not indicate if the same types of higher -order assemblies ( i.e. , nanotubes) were formed by isolated shell -cytochromes. Higher -order structure assembly of shell -cytochromes The unmodified RmmH and Y4 1A Ho -5815 shell proteins readily form nanotubes that can be observed by electron microscopy (Noël, Cai, and Kerfeld , 2016) (Figure 3. 2), yet shell -cytochrome preparations do not exhibit such clear higher -order structures. On 110 the other hand, when analyzed by DLS and PAGE , shell -cytochromes samples exhibit features consistent with the formation of assemblies larger than single hexamers (Figure 3.8, Figure 3.9). Studies of microperoxidase have show n that both intermolecular coordination of the central iron at om and non-covalent -interaction s can occur between exposed heme s (Lombardi, Nastri, and Pavone , 2001) . To block potential interprotein coordination of hemes, we titrated potassium cyanide (KCN) into purified shell -cytochromes and analyze d them by DLS (Figure 3.10). We observed the presence of R4 particle sizes larger than a single hexamer was inversely proportional with the concentration of KCN (Figure 3.10A, 3.10B). Indeed, R4 shell -cytochrome s evaluated under conditions with an approxima tely equimolar concentration of CN - anions to heme exhibited average particle radii consistent with single hexamers (3.5 Œ 4 nm, Figure 3.10 B). Particle sizes of other shell -cytochromes also decreased when 3 mM KCN was added (Figure 3.10C). These results w ere consistent with the hypothesis that inter -protein interactions mediated by hemes may compete with the intrinsic hexamer -hexamer interactions formed by BMC -H proteins, such as a binding between the 6xHis affinity tag and 5-coordinated heme Fe ions. 111 Figure 3.10 Cyanide remediates shell -cytochrome aggregation (A) Titration of potassium cyanide to 200 M R4 shell -cytochrome (based on heme concentration) at pH 7.8 was monitored by DLS. (B) Titration curve of (A), showing the mass -weighted average particle sizes in different concentration of KCN. (C) Particle sizes of 200 M shell -cytochromes (based on heme concentration) for R1, R2 and R5 in the presence of 3 mM KCl or KCN. It was also po ssible that proteins may not fold properly after being processed by Sec and heme maturation pathway s. We constructed a n H-variant with a c -terminus flexible heme -tag (Braun, Rubio, and Thony -Meyer , 2005) (Hc). In the heme -tag, a 6xHis -tag is adjacent to th e heme binding motif ( Figure 3.11 A), so intra -protein coordination is 112 preferred rather than inter -protein coordination compared to shell -cytochromes. Hc forms nanotubes after purification ( Figure 3.11B, 3.11C), suggesting the periplasmic expression will no t alter the self -assembly capability of shell proteins. Overall, the incapability for shell -cytochromes for forming highly ordered structures may due to the strong heme -mediated protein -protein interactions. Figure 3.11 Ho -5815 with Y41A mutation and heme -tag (A) Sequence of matured HC. Residue number 1 is the methionine in WT. In HC, a signal peptide was added to the N -terminus, but this signal peptide was removed by a peptidase, leaving alanine as the first residue in the matured protein . The heme -tag is highlighted. (B) & (C) transmission electron micrographs of purified HC showing the formation of nanotubes. Discussion Naturally -occurring h eme proteins exhibit a large range of functionalities within living cells, including oxygen tran sport, energy conversion, detoxication (Chapman, Daff, and Munro , 1997) and signaling (Liu et al ., 1996). In recent years, proteins that bind heme or heme derivatives have become bioengineering targets for conferring novel functions in 113 biological and biohybrid systems , including biosensing (Beissenhirtz, Scheller, and Lisdat , 2004), biomemory (Lee et al ., 2010; Güzel et al ., 2018), biofuel cell s (Katz, Willner, and Kotlyar , 1999; Ramanavicius and Ramanaviciene , 2009), catalyst s (Kleingar dner, Kandemir, and Bren 2014; Firpo et al ., 2018) and artificial photosynthesis (Anderson et al. , 2014; Kubie et al. , 2017; Koshiyama et al ., 2011). A versatile platform that enables rapid design and synthesis of heme -binding proteins could be of consider able benefit to both fundamental and applied research . In this work, we demonstrated a simple in vivo approach t hat allows c-type hemes to be incorporated into non -heme protein s with defined locations (Figure 3. 3B). While our approach would be compatible with inserting a heme attachment site at any location in the primary sequence of a target protein, our results suggest important considerations that must be accounted for in order to preserve the structure and higher -order ass embly capabilities of a target protein . The approach we describe d above exploits the substrate fipromiscuityfl of CCM -I that may incorporate heme through thioether bonds into virtually any C xxCH motifs to which it is exposed (Li, Bonkovsky, and Guo , 2011). Four of the five unique sites we chose on the two non -heme shell proteins achieved high -occupancy of heme incorporation, suggesting CCM-I is capable of recogniz ing heme binding motifs mostly independently of other protein contextual features. While the majo rity of heterologous shell proteins were determined to have a heme ligated at the appropriate location, a proportion of each preparation (<10% to 3 4%) was the apo-protein without a heme attached. However, our purification methodology cannot rule out the po ssibility that these apo-proteins are partially a result of proteins that failed to translocate to the periplasm, and therefore remain inaccessible to CCM -I activity. In support of this possibility , in preparations with 114 lower heme incorporation ( e.g . H1, H2, R5; Figure 3. 4B), we observed a protein band that did not contain hemes on the native gel s that migrated consistently slower relative to the homo -hexamer of holo -proteins (Figure 3.9A, cyan arrows). This result would be consistent with a cytosolic pool of homo -hexamer of apo-proteins, as hetero -hexamers with a random incorporation of apo -proteins and holo -proteins would be expected in a mixed population (Sommer et al ., 2019). Our evidence suggested that shell -cytochromes retained many structural features of the unmodified parent, although we also find evidence that the higher -order assembly of shell hexamers may be influenced by this modification. Their capacity to hexameriz e appeared to be preserved , as indicated by having similar mobilities to hexameric wild -type shells on native PAGE and having 3 -4 nm particle sizes as determined in DLS (Figure 3 .8). However, we did not observe the formation of nanotubes of the shell -cytochromes . The covalent attachment of hemes might have changed the geometry or conformation of key residues of hexamer tiles and thus the higher -order structure was altered as described in many mutagenesis studies in which a single amino acid substitution can lead to dramatic change s in higher -order structure (Sinha et al. , 2014; Young et al. , 2017; Pang et al. , 2014; Sutter et al. , 2016). Another probable cause, demonstrated by KCN titra tion , is that the exposed distal face of heme may mediate unwanted protein -protein interactions that overpowers the driving force of self -assembl y (e.g. histidine -ligands to metal bonds have a comparable dissociation energy to shell -to-shell interaction (Goff and Morgan , 1976; Greber, Sutter, and Kerfeld , 2019)). Minimizing surface histidine s or covering the heme distal face by designed protein structures could reduce the unwanted interactions. 115 The BMC proteins are believed to be a powerful tool for synthe tic biology, including protein scaffolding (Lee et al ., 2018), nanoreactors (Plegaria and Kerfeld , 2018) and biomaterials (Noël, Cai, and Kerfeld , 2016). A previous study has demonstrated the introduction a redox active iron -sulfur cluster at the pore (cen ter) of a trimeric BMC shell protein (BMC -T) through the mutagenesis of the residues at the pore (Aussignargues et al., 2016). The pores of BMC shell proteins are believed to be selective channels for the influx or efflux of metabolites to the BMCs (Klein et al ., 2009; Thompson et al. , 2015). Being able to transport electrons across the BMC shell proteins potentially expand s the range of reactions that can occur in the microcompartment (Plegaria and Kerfeld , 2018). The incorporation of metal centers at a pore of the hexameric shell has some clear advantages , especially for building an in vivo nanoreactor, for example the metal ce nter can be inserted in the cytosol and the se modification s at pore have less negative impact on the higher -order assembly (Aussignargues et al. , 2016). However, the pore -to-pore distances in the hexametric BMC shell proteins are too far (>5 nm (Sutter et al. , 2017)) for hexamer -to-hexamer electron transfer (Page et al. , 1999). This work demonstrated a proof -of-concept strategy towards constructing a self -assembling conductive architecture. Although compared to naturally occurring multiheme cytochromes, the separations between hemes in the shell -cytochromes are still too far ( e.g. 2.1 nm iron -to-iron in H5) for supporting satisfying electronic current. For example, the STC has four hemes per 91 amino acid residues, resulting in a closely positioned heme chain through the protein (Leys et al. , 2002). In comparison, the monoheme shell -cytochromes have about 2 -4 nm separation between hemes within a hexamer. To reach a similar heme density as the STC, a BMC -H monomer need to bind four hemes. This study identified four possible 116 heme binding sites on the shell proteins and showed the CCM -I could add two hemes per shell -cytochrome monomer, while attaching more hemes with high occupancy still need s further investigat ion . On the other hand, though we examined with the BMC -H shell proteins, this heme -attaching strategy could be applied to other targets, for example smaller self -assembling scaffolds ( e.g. tube -forming coiled -coil peptides (Sharp et al. , 2012)) for achieving higher redox cofactor density. Using heme -attachment to a non -heme binding protein, our study demonstrated a new platform to design novel heme proteins. Over the decades, numerous novel heme -binding proteins have been crafted by severa l unique design principles (Reedy and Gibney , 2004). First, one can perform modifications on natur ally occurring heme proteins, like myoglobin, peroxidase and cytochrome c. This approach focuse d on generat ing specific variations in heme environment and dem onstrated the influences of key factors, such as axial ligands (Shimizu et al. , 1988), solvent exposure (Bortolotti et al. , 2011), hydrogen bonding network (Goodin and McRee , 1993) , on the functions of the heme proteins. This design strategy creates new proteins from existing heme -binding scaffolds, and depends on the goals, where the scope of redesigning could vary from the substitution of a single residue to rebuild ing most of the proteins (Isogai et al. , 1999). Another design principle, on the other en d of the spectrum, attempt s to isolate the heme -binding site from the inherent complexity of natur al protein folding by using short polypeptides or de novo designed protein scaffolds. Miniature metalloporphyrinyl -peptides such as microperoxidases (Verbaro et al. , 2009), peptide -sandwiched mesoheme (Benson et al. , 1995) and mimochromes (Lombardi et al. , 1998) consist of a short polypeptide (usually from 8 -30 residues) that does not form defined tertiary 117 structures and a covalent ly bound heme or heme analog. These de novo heme -binding proteins commonly adopt a four -helix bundle as the scaffold for the heme -binding pocket and provide metal ligands at the core of the bundles (Reedy and Gibney , 2004). One successful family of de novo designed heme proteins is the maquette (Robertson et al. , 1994). An underutilized design principle that redesign s a non -heme natur al protein to bind a heme, has been rarely explored in the past. We only find two examples, including the designing of the he me binding ROP protein (Wilson, Caruana, and Gilardi , 2003), whose overall folding is a four -helix bundle and the heme -tag (Braun, Rubio, and Thony -Meyer , 2005) developed for heme -affinity chromatography that attaches the sequence of microperoxidase to a p rotein of interest for a heme -specific protein purification (Asher and Bren , 2010) . As the demand for redox active components in the synthetic biology toolbox increase s, being able to add heme to functional protein modules creates potentials to build sophi sticate d architectures with redox and catalytic capabilities (Bostick et al ., 2018). In this study, we demonstrated that introducing heme binding motifs to the allowed region of shell proteins would likely produce a heme protein with native folding. Since the geometric constrictions of the heme binding motif have been systematically studied (Kozak, Gray, and Wittung -Stafshede , 2018; Fufezan, Zhang, and Gunner , 2008), this approach can be applied to other protein s of interest. More importantly, the capacity to add heme to a rigid location with well -defined structure is a critical consideration when the applications are related to electron transfer (Matyushov , 2013). Conversely, this system is limited in practice to the introduction of hemes upon the surface of a protein, although protein surface -exposed hemes can be very useful in designs requiring a high 118 accessibility of heme moiety , such as electron transfer (Ruzgas, Gaigalas, and Gorton , 1999) and catalys is (Klei ngardner, Kandemir, and Bren , 2014). Materials and Methods Identification of possible heme attachment sites for shell proteins based on secondary Structures of the CxxCH motif in natural cytochrome s c. Table 3.1 list 180 cytochrome Protein Data Bank (PD B) entries used for calculating the torsion angles of 249 CxxCH motifs. Based on the Ramchandra plot , the following gating parameters were used for searching for similar secondary structures in the Ho-5815 shell protein: - X1 < -50°, -180° < X1 < +180°; - C1 < -50°, -100° < C1 < +0°; - X2 < -25°, -75° < X2 < +50°; - X2 < -50°, -75° < X2 < +0°; - C2 < -50°, -75° < C2 < +25°. Strains and protein expression The CxxCH motifs were introduced into the wild -type shell prot eins by designing primers with mutated codons. Heme -attached shell protein s were expressed in Escherichia coli BL21 (DE3) (New England Biolabs, Massachusetts , USA) strain harboring pEC86 and plasmid encoding modified shell protein s. Fresh transformants were first inoculated in 1 mL LB medium, which was then used to inoculate 1 L 2xYT medium (Alpha Biosciences, M aryland , USA) . 50 µg/mL kanamycin (Sigma -Aldrich, Missouri , USA) and 40 µg/mL chloramphenicol (Sigma -Aldrich) were added to a ll the growth media in this study. Cells were first grown in 2xYT media at 3 7 °C and shaken at 140 rpm for ~6 hours or OD 600 = 0.8 and then amended with a final concentration of 10 µg / mL of 5-119 aminolevulinic acid (Frontier Scientific, Inc. , Utah , USA) an d 1 mM i -D-1-thiogalactopyranoside (GoldBio, Missouri , USA) for induction. Induced cultures were grown at 30 °C and 80 rpm for 12 h. Cells were allowed to settle overnight before discarding the supernatant and harvesting the cells . Cells were lysed by sonication (80% Energy 15 min, Model 120 Sonic Dismembrator , Fisher Scientific , Massachusetts, USA) in 50 mM sodium phosphate buffer pH 7 and the soluble fraction was removed ( except H5, which was soluble and loaded onto resin directly). Red pellets were resuspended in buffer containing 8 M urea 50 mM ,Tris and 10 mM imidazole at pH 8 and loaded on to 10 mL Ni -NTA super -flow resin (Qiagen, Hilden, Germany) equilibrated with the same buffer. Samples were washed with buffer c ontaining 4 M urea, 50 mM Tris, 500 mM NaCl and 50 mM imidazole at pH 8, and eluted with buffer containing 4 M urea , 250 mM imidazole and 25 mM Tris pH 8. Purified proteins were dialy zed against buffer containing 50 mM Tris , 100 mM imidazole and 150 mM NaC l pH 8, then concentrated with Gel -Absorbent ( Spectrum Laboratories, Inc. , California, USA) and finally dialy zed against buffer contains 50 mM Tris , 100 mM imidazole , 150 mM NaCl and 10% glycerol pH 8. Concentrated samples were flash frozen in liquid nitro gen and stored at -80 °C. Molecular mass determination by mass spectrometry Purified shell -cytochromes were diluted to approximately 5 µ M (base d on their Soret absorbance) in NH 4Ac. 5 µL of sample was injected onto a Hypersil Gold cyanopropyl guard column (1x10 mm) (Thermo Scientific , Massachusetts , USA ) using a Waters Acquity UHPLC system (Waters , Massachusetts , USA ). A binary gradient flowing at 0.1 ml min -1 to elute the protein was run as follows: initial conditions 98% A (water + 120 0.15% formic acid) / 2% B (acetonitrile), hold at initial conditions for 5 min to wash away salts and buffer, linear ramp to 75% B from 5 to 10 min, hold at 75% B until 12 min, return to starting conditions of 98% A / 2% B at 12.01 min and hold until 15 min to re -equil ibrate the column for the next injection. Proteins were analyzed on a Waters G2 -XS quadrupole -time -of-flight mass spectrometer using electrospray ionization operating in positive ion mode and scanning a mass range of m/z 200 -2000 with 1 scan per second. Ca pillary voltage was 3 kV, sample cone voltage was 35 V, source temperature was 100°C, desolvation temperature was 350 °C and desolvation gas flow was 600 L hr-1. Elution peaks at 7 -9 min were integrated and the molecular mass spectra were deconvoluted usin g the MaxEnt 1 algorithm in the Waters Masslynx software package. Transmi ssion electron microscopy Thin sections of E. coli overexpressing H o-5815 Y41A were prepared as described Young et al ., 2017. A negative stain of the Hc higher -order protein assembly was prepared following protein purification. Briefly, a ~5 µL droplet was floated on 400 -mesh copper grids with supporting carbon film (Ted Pella, California, USA ) for 5 min, then 1% uranyl acetate was applied to stain the sample for another 5 min. Finall y, Reynolds lead citrate was applied for 5 min, the sample was then washed with ultrapure water and dried by wicking with a filter paper. Thin section images were imaged on a JEM 100CX II (JEOL , Tokyo, Japan ) equipped with a Prius SC200 -830 CCD camera ( Gatan, California, USA ), and purified protein samples were imaged on JEM -1400Flash (JEOL ) equipped with a fiMatataki Flash fl sCMOS camera (JEOL) . 121 Molecular dynamic of heme -attached shell proteins Mutations on Site 1,2,3 and 5 were built by the Mutage nesis Plugin in PyMol (github.com/schrodinger/pymol -open-source ). Loop extensions at site 4 were produced by SWISS -MODEL Homology Modelling webserver (Waterhouse et al. , 2018) . Hemes were loaded to the close proximate to the heme binding motif in PyMol, an d then bound to the cysteines and histidine in VMD (Humphrey, Dalke, and Schulten , 1996) using a modified topology / parameter file based on CHARMM36 (Huang and MacKerell , 2013) toppar_all36_prot_heme.str . The stereochemistry of the carbon -sulfur bonds was manually added. All solvated models were minimized for 5000 steps and the molecular dynamic trajectory for 1 ns were calculated by NAMD (Phillips et al ., 2005) (version 2.13 with CUDA) on a lab PC. The snapshot at 1 ns is show n in Figure 3.1D, an d the last 0.8 nanoseconds were used to calculat e the RMSD in Figure 3. 3B. UV-Vis spectroscopy Frozen samples were d ilute d to have an a bsorbance at 408 nm of 0.7 -0.8 (1 cm) with buffer contain ing 50 mM Tris , 150 mM NaCl and 100 mM imidazole pH 8. Oxidized samples were measure d as prepared. To reduce hemes, several crystals of sodium dithionite were added to the sample cuvettes. All spectra were recorded by using a DU800 spectrophotometer (Beckman Coulter, C alifornia , USA ). Spectra in Figure 3. 4 were normalized to A Soret = 1. Redox titration A cocktail of redox mediators (2 µM methyl viologen, 2 µM benzyl viologen, 2 µM neutral red, 10 µM anthraquinone -2-sulfonate, 10 µM anthraquinone -2,5 -disulfonate, 10 µM 2 -hydroxy -1,4 -naphthaquinone, 10 µM 2,5 -dih ydroxy -p-benzoquinone, 15 µM 122 pyocyanin, 10 µM 5 -hydroxy -1,4 -naphthaquinone and 10 µM potassium ferrocyanide) was added to degassed buffer containing 50 mM Tris and 100 mM imidazole pH 8.0. A glassy carbon working electrode and Ag/ AgCl reference electrode w ere used. The reference electrode was calibrated with saturated quinhydrone solution at pH 7 (E m' (Ag/ AgCl) = +190 mV). Reductant (sodium dithionite, 5 mM or 50 mM) or oxidant (potassium ferricyanide, 5 mM or 50 mM) were titrated into 5 mL of sample that was stirred and protected from oxygen by purging with N2. UV-Vis spectra were recorded approximately 2 min after each titration (or until the redox potential reading stabilized within 2 mV) by spectrophotometer ( LAMBDA 650 , PerkinElmer , Massachusetts, USA ). Electrophoresis Native polyacrylamide gels were casted with 4% acrylamide/ bis-acrylamide (29:1) pH 8.3 as stacking gel and 12% acrylamide/ bis-acrylamide (29:1) pH 8. 8 as the resolving gel. pH 8.3 was used in the stacking gel because the isoelectric point of samples is close to pH 6.8, which is commonly used in PAGE protocols. Frozen samples were concentrated with 30 kDa Amicon 0.5 mL spin columns and then diluted to an absorbance at 408 nm of 10 (for a 1 cm pathl ength ). 7 µL samples were loaded on to the gel. After e lectrophoresis , gels were first imaged before staining to show the heme bands. Dynamic light scattering The same diluted samples for electrophoresis were used for DLS. For the native condition, samples were used as purified ; for the 4 M urea condition, 1:1 mixture of sample with 8 M urea and 50 mM Tris -Cl pH 8.0 was measured. All samples were 123 measured in DynaPro NanoStar (Wyatt Technology , California, USA) with 5 s exposure and 25 sample averaging. Potassium cyanide titration Shell cytochromes were concentrated by 50 kDa cutoff Amicon ultrafiltration spin columns, and diluted in to 50 mM Tris, 100 mM NaCl containing 0 to 10 mM KCN pH 7.8 to an absor bance at 408 nm of approximate ly to 20 ( for a 1 cm pathlength ). DLS were measured with 2 s exposure and 50 sample averaging. The mass -weighted averaged radii of particles were calculated using the distribution of sizes (radius) between 1 to 100 nm. Potassium chloride was used as control. The approximate h eme concentrations were 200 qq 408 nm = 100 x 10 3 M-1 cm-1). The approximate concentration ionized cyanide anion (CN -) was 120 qqM for 3 mM KCN at pH 7.8 (pK aHCN = 9.21). Table 3.1 List of structures used in Figure 3.1B The first and forth column s are the Protein Data Bank entries , the second and fifth column s are the chain identifier . Description Description 19hc A Protein (Nine -Haem Cytochrome c) 2oz1 A Diheme Cytochrome c 1aqe A Cytochrome c3 2p0b A Cytochrome c-Type Protein Nrfb 1bvb A Cytochrome c-554 2rdz A Cytochrome c-552 1czj A Cytochrome c3 2rf7 A Cytochrome c-552 1d4c A Flavo Cytochrome c Fumarate Reductase 2vhd A Cytochrome c551 Peroxidase 1d4e A Flavo Cytochrome c Fumarate Reductase 2vr0 A Cytochrome c Nitrite Reductase, Catalytic Subunit Nfra 1duw A Nonaheme Cytochrome c 2wjn C Photosynthetic Reaction Center Cytochrome c Subunit 1e39 A Fumarate Reductase Flavoprotein Subunit 2x5v C Photosynthetic Reaction Center Cytochrome c Subunit 1eb7 A Cytochrome c551 Peroxidase 2yxc A Cytochrome c3 1etp A Cytochrome c4 2yyw A Cytochrome c3 1fcd C Flavo Cytochrome c Sulfide Dehydrogenase (Cytochrome Subunit) 2yyx A Cytochrome c3 1fgj A Hydroxylamine Oxidoreductase 2z47 A Cytochrome c3 1fs8 A Cytochrome c Nitrite Reductase 2zo5 A Eight -Heme Nitrite Reductase 1fs9 A Cytochrome c Nitrite Reductase 3bng A Cytochrome c-552 124 Table 3.1 (Cont™d) 1ft6 A Cytochrome c554 3bnj A Cytochrome c-552 1gmb A Cytochrome c3 3bxu A Cytochrome c3 1gu6 A Cytochrome c552 3cao A Cytochrome c3 1gyo A Cytochrome c3, A Dimeric Class Iii C -Type Cytochrome 3cyr A Cytochrome c3 1h1o A Cytochrome c-552 3d1i A Eight -Heme Nitrite Reductase 1h21 A Split -Soret Cytochrome c 3f29 A Eight -Heme Nitrite Reductase 1h29 A High -Molecular -Weight Cytochrome c 3fo3 A Eight -Heme Nitrite Reductase 1h31 A Diheme Cytochrome c 3gm6 A Eight -Heme Nitrite Reductase 1h32 A Diheme Cytochrome c 3h33 A Cytochrome c7 1h33 A Diheme Cytochrome c 3h34 A Cytochrome c7 1hh5 A Cytochrome c7 3h4n A Cytochrome c7 1i77 A Cytochrome c3 3hq6 A Cytochrome c551 Peroxidase 1iqc A Di -Heme Peroxidase 3hq7 A Cytochrome c551 Peroxidase 1j0p A Cytochrome c3 3hq8 A Cytochrome c551 Peroxidase 1jmx A Amine Dehydrogenase 3hq9 A Cytochrome c551 Peroxidase 1jni A Diheme Cytochrome c Nap B 3ir6 B Respiratory Nitrate Reductase 1 Beta Chain 1jrx A Flavo Cytochrome c 3l1t A Cytochrome c-552 1jry A Flavo Cytochrome c 3lg1 A Eight -Heme Nitrite Reductase 1jrz A Flavo Cytochrome c 3lgq A Eight -Heme Nitrite Reductase 1kss A Flavo Cytochrome c 3mk7 B Cytochrome c Oxidase, Cbb 3-Type, Subunit O 1ksu A Flavo Cytochrome c 3mm o A Eight -Heme Nitrite Reductase 1lj1 A Flavo Cytochrome c3 3o5a B Diheme Cytochrome c Nap B 1m1p A Small Tetraheme Cytochrome c 3o5c A Cytochrome c551 Peroxidase 1m1r A Small Tetraheme Cytochrome c 3oue A Cytochrome c Family Protein 1m64 A Flavo Cytochrome c3 3ouq A Cytochrome c Family Protein 1m6z A Cytochrome c4 3ov0 A Cytochrome c Family Protein 1m70 A Cytochrome c4 3owm A Eight -Heme Nitrite Reductase 1mdv A Cytochrome c3 3pmq A Decaheme Cytochrome c Mtrf 1nml A Di -Haem Cytochrome c Peroxidase 3rkh A Eight -Heme Nitrite Reductase 1oah A Cytochrome c Nitrite Reductase 3s7w A Eight -Heme Nitrite Reductase 1ofw A Nine -Heme Cytochrome c 3sce A Eight -Heme Nitrite Reductase 1ofy A Nine Heme Cytochrome c 3sel X Cytochrome c7 1ogy B Diheme Cytochrome c Nap B Molecule: Nitrate Reductase 3sj0 X Cytochrome c7 1p2e A Flavo Cytochrome c3 3sj1 X Cytochrome c7 Table 3.1 (Cont™d) 1p2h A Flavo Cytochrome c3 3sj4 X Cytochrome c7 1pby A Quinohemoprotein Amine Dehydrogenase 60 Kda Subunit 3sxq A Eight -Heme Nitrite Reductase 1q16 B Respiratory Nitrate Reductase 1 Beta Chain 3t6d C Photosynthetic Reaction Center Cytochrome c Subunit 125 Table 3.1 (Cont™d) 1q9i A Flavo Cytochrome c3 3tor A Cytochrome c Nitrite Reductase 1qdb A Cytochrome c Nitrite Reductase 3ttb A Eight -Heme Nitrite Reductase 1qo8 A Flavo Cytochrome c3 Fumarate Reductase 3u99 A Diheme Cytochrome c 1r27 B Respiratory Nitrate Reductase 1 Beta Chain 3ubr A Cytochrome c-552 1rwj A Cytochrome c Family Protein 3uu9 A Eight -Heme Nitrite Reductase 1rz6 A Cytochrome c Peroxidase 3vrd A Flavo Cytochrome c Heme Subunit 1sp3 A Cytochrome c, Putative 4aal A Cytochrome c551 Peroxidase 1up9 A Cytochrome c3 4aan A Cytochrome c551 Peroxidase 1vrn C Photosynthetic Reaction Center Cytochrome c Subunit 4aao A Cytochrome c551 Peroxidase 1w7o A Cytochrome c3 4fas A Hydroxylamine Oxidoreductase 1wad A Cytochrome c3 4l38 A Eight -Heme Nitrite Reductase 1wr5 A Cytochrome c3 4l3x A Eight -Heme Nitrite Reductase 1y0p A Fumarate Reductase Flavoprotein Subunit 4l3y A Eight -Heme Nitrite Reductase 1y5l B Respiratory Nitrate Reductase 1 Beta Chain 4l3z A Eight -Heme Nitrite Reductase 1yqd A Sinapyl Alcohol Dehydrogenase 4lm8 A Extracellular Iron Oxide Respiratory System Surface Decaheme Cytochrome c Component Mtr C 1yqx A Sinapyl Alcohol Dehydrogenase 4lmh A Extracellular Iron Oxide Respiratory System Surface Decaheme Cytochrome c Component Omc C 1z1n X Sixteen Heme Cytochromes 4n4k A Hydroxylamine Oxidoreductase 1zzh A Cytochrome c Peroxidase 4n4m A Hydroxylamine Oxidoreductase 2a3p A Cog3005: Nitrate/Tmao Reductases, Membrane -Bound Tetraheme Cytochrome c Subunit 4n4n A Hydroxylamine Oxidoreductase 2b7r A Fumarate Reductase Flavoprotein Subunit 4n4o A Hydroxylamine Oxidoreductase 2b7s A Fumarate Reductase Flavoprotein Subunit 4q0t A Eight -Heme Nitrite Reductase 2bq4 A Basic Cytochrome c3 4q17 A Eight -Heme Nitrite Reductase 2c1u A Di -Haem Cytochrome c Peroxidase 4q1o A Eight -Heme Nitrite Reductase 2c1v A Di -Haem Cytochrome c Peroxidase 4q4u A Eight -Heme Nitrite Reductase 2cdv A Cytochrome c3 4q5b A Eight -Heme Nitrite Reductase 2cth A Cytochrome c3 4q5c A Eight -Heme Nitrite Reductase 2cvc A High -Molecular -Weight Cytochrome c Precursor 4qo5 A Hypothetical Multiheme Protein 2cy3 A Cytochrome c3 4rwm A Similar To Hydroxylamine Oxidoreductase HAO 2czs A Cytochrome c, Putative 4v2k A Thiosulfate Dehydrogenase 2e81 A Cytochrome c-552 4wjy A Cytochrome c-552 2e84 A High -Molecular -Weight Cytochrome c 4wq8 A Thiosulfate Dehydrogenase 2ewi A Cytochrome c3 4wqa A Thiosulfate Dehydrogenase 2ewk A Cytochrome c3 4wqc A Thiosulfate Dehydrogenase 2ewu A Cytochrome c3 4wqe A Thiosulfate Dehydrogenase 2ffn A Cytochrome c3 5c2w A Hydrazine Synthase Alpha Subunit 2fw5 A Dhc, Diheme Cytochrome c 5djq B Cbb 3-Type Cytochrome c Oxidase Subunit II 126 Table 3.1 (Cont™d) 2fwt A Dhc, Diheme Cytochrome c 5lo9 A Cytochrome c 2j7a A Cytochrome c Nitrite Reductase Nrf A 5m7j A Photosynthetic Reaction Center Cytochrome c Subunit 2ot4 A Eight -Heme Nitrite Reductase 5m7l A Photosynthetic Reaction Center Cytochrome c Subunit Table 3.2 Calculated and observed molecular masses of shell cytochromes Calculated mass (Da) Observed mass (Da) Sample apoprotein holo protein Full -length apoprotein Full -length holo protein Truncated holo protein H1 10873 11489 10871.5 11490 n.d. H2 10872 11488 10871 11489 n.d. H4 11345 11961 11343 11960.5 n.d. H5 10887 11503 10886 11504 n.d. H2+5 10960 11576/12193 10955 11572.5/12191.5 n.d. R1 10549 11165 10547.5 11165.5 8575 R2 10534 11150 10533 11150 8560.5 R4 11006 11622 11004.5 11622.5 9033.5 R5 10468 11084 10466 11083.5 n.d. R2+5 10540 11156/11773 10535.5 11153.5/11771.5 9182 R4+5 11012 11629/12245 11008.5 11626/12243.5 9654.5 n.d. Not detected. 127 REFERENCE S 128 REFERENCES Allen, J. W. A., & Ferguson, S. J. (2006). What is the substrate specificity of the System I cytochrome c biogenesis apparatus? , 34(1), 150 Œ151. https://doi.org/10.1042/BST0340150 Altamura, L., Horvath, C., Rengaraj, S., Rongier, A., E louarzaki, K., Gondran, C., Maçon, A. L. B., Vendrely, C., Bouchiat, V., Fontecave, M., Mariolle, D., Rannou, P., Le Goff, A., Duraffourg, N., Holzinger, M., & Forge, V. (2017). A synthetic redox biofilm made from metalloprotein Œprion domain chimera nanowi res. Nat. Chem. , 9(2), 157 Œ163. https://doi.org/10.1038/nchem.2616 Anderson, J. L. R., Armstrong, C. T., Kodali, G., Lichtenstein, B. R., Watkins, D. W., Mancini, J. a, Boyle, A. L., Farid, T. a, Crump, M. P., Moser, C. C., & Dutton, P. L. (2014). Construc ting a man -made c-type cytochrome maquette in vivo : electron transfer, oxygen transport and conversion to a photoactive light harvesting maquette. Chem. Sci. , 5(2), 507 Œ514. https://doi.org/10.1039/C3SC52019F Arslan, E., Schulz, H., Zufferey, R., Künzler, P., & Thony -Meyer, L. (1998). Overproduction of the Bradyrhizobium japonicum c-type cytochrome subunits of the cbb 3 oxidase in Escherichia coli . Biochem. Biophys. Res. Commun. , 251 (3), 744 Œ747. https://doi.org/10.1006/bbrc.1998.9549 Asher, W. B., & Bren, K . L. (2010). A heme fusion tag for protein affinity purification and quantification. Protein Sci. , 19(10), 1830 Œ1839. https://doi.org/10.1002/pro.460 Aussignargues, C., Pandelia, M. E., Sutter, M., Plegaria, J. S., Zarzycki, J., Turmo, A., Huang, J., Ducat , D. C., Hegg, E. L., Gibney, B. R., & Kerfeld, C. A. (2016). Structure and function of a bacterial microcompartment shell protein engineered to bind a [4Fe -4S] cluster. J. Am. Chem. Soc. , 138(16), 5262Œ5270. https://doi.org/10.1021/jacs.5b11734 Axen, S. D., Erbilgin, O., & Kerfeld, C. A. (2014). A taxonomy of bacterial microcompartment loci constructed by a novel scoring method. PLoS Comput. Biol. , 10(10), e1003898. https://doi.org/10.1371/journal.pcbi.1003898 Battistuzzi, G., Borsari, M., Cowan, J. A., Ranieri, A., & Sola, M. (2002). Control of cytochrome c redox potential: Axial ligation and protein environment effects. J. Am. Chem. Soc. , 124(19), 5315 Œ5324. https://doi.org/10.1021/ja017479v Beissenhirtz, M. K., Scheller, F. W., & Lisdat, F. (2004). A superoxide sensor based on a multilayer cytochrome c electrode. Anal. Chem. , 76(16), 4665 Œ4671. https://doi.org/10.1021/ac049738f Benson, D. R., Hart, B. R., Zhu, X., & Doughty, M. B. (1995). Des ign, synthesis, and circular dichroism investigation of a peptide -sandwiched Mesoheme. J. Am. Chem. Soc. , 117(33), 8502 Œ8510. https://doi.org/10.1021/ja00138a002 Blankenship, R. E. (2014). Molecular mechanisms of photosynthesis . John Wiley & Sons. Bortolotti, C. A., Siwko, M. E., Castellini, E., Ranieri, A., Sola, M., & Corni, S. (2011). The reorganization energy in cytochrome c is controlled by the accessibility of the heme to the solvent. J. Phys. Chem. Lett. , 2(14), 1761 Œ1765. https://doi.org/10. 1021/jz200734a Bostick, C. D., Mukhopadhyay, S., Pecht, I., Sheves, M., Cahen, D., & Lederman, D. (2018). Protein bioelectronics: a review of what we do and do not know. Reports Prog. Phys. , 81(2), 026601. https://doi.org/10.1088/1361 -6633/aa85f2 Bowman, S . E. J., & Bren, K. L. (2008). The chemistry and biochemistry of heme c: functional bases for covalent attachment. Nat. Prod. Rep. , 25(6), 1118. https://doi.org/10.1039/b717196j 129 Braun, M., Rubio, I. G., & Thony -Meyer, L. (2005). A heme tag for in vivo synt hesis of artificial cytochromes. Appl. Microbiol. Biotechnol. , 67(2), 234 Œ239. https://doi.org/10.1007/s00253 -004 -1804 -2 Butt, W. D., & Keilin, D. (1962). Absorption spectra and some other properties of cytochrome c and of its compounds with ligands. Proc. R. Soc. London. Ser. B. Biol. Sci. , 156(965), 429 Œ458. https://doi.org/10.1098/rspb.1962.0049 Chapman, S. K., Daff, S., & Munro, A. W. (1997). Heme: The most versatile redox centre in biology? . Metal Sites in Proteins and Models: Iron Centres (pp. 39 Œ70). https://doi.org/10.1007/3 -540-62870 -3_2 Cheng, S., Liu, Y., Crowley, C. S., Yeates, T. O., & Bobik, T. A. (2008). Bacterial microcompartments: their properties and paradoxes. BioEssays , 30(11 Œ12), 1084 Œ1095. https://doi.org/10.1002/bies.20830 Dutton, P. L. (1978). Redox potentiometry: Determination of midpoint potentials of oxidation -reduction components of biological electron -transfer systems. In Methods in enzymology (Vol. 54, pp. 411 Œ435). https://doi.org/10.1016/S0076 -6879(78)54026 -3 Firpo, V., Le, J. M., Pavone, V., Lombardi, A., & Bren, K. L. (2018). Hydrogen evolution from water catalyzed by cobalt -mimochrome VI*a, a synthetic mini -protein. Chem. Sci. , 9(45), 8582 Œ8589. https://doi.org/10.1039/C8SC01948G Fufezan, C., Zhang, J., & Gunner, M. R. (2008). Ligand preference and orientation in b- and c-type heme -binding proteins. Proteins Struct. Funct. Bioinforma. , 73(3), 690 Œ704. https://doi.org/10.1002/prot.22097 Goff, H., & Morgan, L. O. (1976). Amino acid -substituted iron porphyrins. 2. The rmodynamic studies of ligand binding. Inorg. Chem. , 15(9), 2069 Œ2076. https://doi.org/10.1021/ic50163a011 Goodin, D. B., & McRee, D. E. (1993). The Asp -His -iron triad of cytochrome c peroxidase controls the reduction potential electronic structure, and cou pling of the tryptophan free radical to the heme. Biochemistry , 32(13), 3313 Œ3324. https://doi.org/10.1021/bi00064a014 Greber, B. J., Sutter, M., & Kerfeld, C. A. (2019). The plasticity of molecular interactions governs bacterial microcompartment shell assembly. Structure , 27(5), 749 -763.e4. https://doi.org/10.1016/j.str.2019.01.017 -hemoglobin film based sextet state biomemory device by cross -linked photosensitive hapten monomer. , 176(April 2017), 85 Œ91. https://doi.org/10.1016/j.talanta.2017.08.014 Huang, J., & MacKerell, A. D. (2013). CHARMM36 all -atom additive protein force field: Validation based on comparison to NMR data. J. Comput. Chem. , 34(25), 2135 Œ2145. https://doi.org/10.1 002/jcc.23354 Humphrey, W., Dalke, A., & Schulten, K. (1996). VMD: Visual molecular dynamics. J. Mol. Graph. , 14(1), 33Œ38. https://doi.org/10.1016/0263 -7855(96)00018 -5 Isogai, Y., Ota, M., Fujisawa, T., Izuno, H., Mukai, M., Nakamura, H., Iizuka, T., & Ni shikawa, K. (1999). Design and synthesis of a globin fold. Biochemistry , 38(23), 7431 Œ7443. https://doi.org/10.1021/bi983006y Jorda, J., Leibly, D. J., Thompson, M. C., & Yeates, T. O. (2016). Structure of a novel 13 nm dodecahedral nanocage assembled from a redesigned bacterial microcompartment shell protein. Chem. Commun. , 52(28), 5041 Œ5044. https://doi.org/10.1039/C6CC00851H Katz, E., Willner, I., & Kotlyar, A. B. (1999). A non -compartmentalized glucose O2 biofuel cell by bioengineered electrode surfaces . J. Electroanal. Chem. , 479(1), 64 Œ68. https://doi.org/10.1016/S0022 -0728(99)00425 -8 Kerfeld, C. A., Aussignargues, C., Zarzycki, J., Cai, F., & Sutter, M. (2018). Bacterial microcompartments. Nat. Rev. Microbiol. https://doi.org/10.1038/nrmicro.2018.10 130 Kerfeld, C. A., Sawaya, M. R., Tanaka, S., Nguyen, C. V, Phillips, M., Beeby, M., & Yeates, T. O. (2005). Protein structures forming the shell of primitive bacterial organelles. Science , 309(5736), 936 Œ938. https://doi.org/10.1126/science.1113397 Klein, M. G., Zwart, P., Bagby, S. C., Cai, F., Chisholm, S. W., Heinhorst, S., Cannon, G. C., & Kerfeld, C. A. (2009). Identification and structural analysis of a novel carboxysome shell protein with implications for metabolite transport. J. Mol. Biol. , 392(2), 319Œ333. https://doi.org/10.1016/j.jmb.2009.03.056 Kleingardner, J. G., Kandemir, B., & Bren, K. L. (2014). Hydrogen evolution from neutral water under aerobic conditions catalyzed by cobalt microperoxidase -11. J. Am. Chem. Soc. , 136(1), 4 Œ7. https://doi.o rg/10.1021/ja406818h Koshiyama, T., Shirai, M., Hikage, T., Tabe, H., Tanaka, K., Kitagawa, S., & Ueno, T. (2011). Post -crystal engineering of zinc -substituted Myoglobin to construct a long-lived photoinduced charge -separation system. Angew. Chemie Int. Ed ., 50(21), 4849 Œ4852. https://doi.org/10.1002/anie.201008004 Kozak, J. J., Gray, H. B., & Wittung -Stafshede, P. (2018). Geometrical description of protein structural motifs. J. Phys. Chem. B , 122 (49), 11289 Œ11294. https://doi.org/10.1021/acs.jpcb.8b07130 Kranz, R. G., Richard -Fogal, C., Taylor, J. -S., & Frawley, E. R. (2009). Cytochrome c biogenesis: mechanisms for covalent modifications and trafficking of heme and for heme -iron redox control. Microbiol. Mol. Biol. Rev. , 73(3), 510 Œ528, Table of Contents. https://doi.org/10.1128/MMBR.00001 -09 Kubie, L., Amori, A. R., Chakraborty, S., Bren, K. L., & Krauss, T. D. (2017). Photoinduced charge separation in single -walled carbon nanotube/protein integrated systems. Nanoscale Horizons , 2(3), 163 Œ166. https://doi. org/10.1039/C6NH00172F Lassila, J. K., Bernstein, S. L., Kinney, J. N., Axen, S. D., & Kerfeld, C. A. (2014). Assembly of robust bacterial microcompartment shells using building blocks from an organelle of unknown function. J. Mol. Biol. , 426(11), 2217Œ2228. https://doi.org/10.1016/j.jmb.2014.02.025 Lee, M. J., Mantell, J., Hodgson, L., Alibhai, D., Fletcher, J. M., Brown, I. R., Frank, S., Xue, W.F., Verkade, P., Woolfson, D. N., & Warren, M. J. (2018). Engineered synthetic scaffolds for organizing proteins within the bacterial cytoplasm. Nat. Chem. Biol. , 14(2), 142 Œ147. https://doi.org/10.1038/nchembio.2535 Lee, T., Kim, S. U., Min, J., & Choi, J. W. (2010). Multilevel biomemory device consisting of recombinant azurin/cytochrome c. Adv. Mater. , 22(4), 510 Œ514. https://doi.org/10.1002/adma.200902288 Leys, D., Meyer, T. E., Tsapin, A. S., Nealson, K. H., Cusanovich, M. a, & Van Beeumen, J. J. (2002). Crystal structures at atomic resolution reveal the novel concept of fi electron -harvestingfl as a role f or the Small Tetraheme Cytochrome c. J. Biol. Chem. , 277(38), 35703 Œ35711. https://doi.org/10.1074/jbc.M203866200 Li, T., Bonkovsky, H. L., & Guo, J. (2011). Structural analysis of heme proteins: implications for design and prediction. BMC Struct. Biol. , 11(1), 13. https://doi.org/10.1186/1472 -6807 -11-13 Liu, X., Kim, C. N., Yang, J., Jemmerson, R., & Wang, X. (1996). Induction of apoptotic program in cell -free extracts: Requirement for dATP and cytochrome c. Cell , 86(1), 147 Œ157. https://doi.org/10.1016/S0092 -8674(00)80085 -9 Lombardi, A., Nastri, F., & Pavone, V. (2001). Peptide -based heme -protein models. Chem. Rev. , 101(10), 3165Œ3189. https://doi.org/10.1021/cr000055j Lombardi, A., Nastri, F., Sanseverino, M., Maglio, O., Pedone, C., & Pavone, V. (1998). Miniaturized hemoproteins: design, synthesis and characterization of mimochrome II. Inorganica Chim. Acta , 275Œ276(i), 301 Œ313. https://doi.org/10.1016/S0020 -1693(97)06180 -X 131 Mallette, E., & Kimber, M. S. (2017). A complete structu ral inventory of the Mycobacterial microcompartment shell proteins constrains models of global architecture and transport. J. Biol. Chem. , 292(4), 1197 Œ1210. https://doi.org/10.1074/jbc.M116.754093 Matyushov, D. V. (2013). Protein electron transfer: Dynami cs and statistics. J. Chem. Phys. , 139(2). https://doi.org/10.1063/1.4812788 Negron, C., Fufezan, C., & Koder, R. L. (2009). Geometric constraints for porphyrin binding in helical protein binding sites. Proteins Struct. Funct. Bioinforma. , 74(2), 400 Œ416. https://doi.org/10.1002/prot.22143 Noël, C. R., Cai, F., & Kerfeld, C. A. (2016). Purification and characterization of protein nanotubes assembled from a single bacterial microcompartment shell subunit. Adv. Mater. Interfaces , 3(1), 1 Œ6. https://doi.org/10.1002/admi.201500295 Page, C. C., Moser, C. C., Chen, X., & Dutton, P. L. (1999). Natural engineering principles of electron tunnelling in biological oxidation -reduction. Nature , 402(6757), 47 Œ52. https://doi.org/10.1038/46972 Pang, A., F rank, S., Brown, I., Warren, M. J., & Pickersgill, R. W. (2014). Structural insights into higher order assembly and function of the bacterial microcompartment protein PduA. J. Biol. Chem. , 289(32), 22377 Œ22384. https://doi.org/10.1074/jbc.M114.569285 Phill ips, J. C., Braun, R., Wang, W., Gumbart, J., Tajkhorshid, E., Villa, E., Chipot, C., Skeel, R. D., Kalé, L., & Schulten, K. (2005). Scalable molecular dynamics with NAMD. J. Comput. Chem. , 26(16), 1781 Œ1802. https://doi.org/10.1002/jcc.20289 Plegaria, J. S., & Kerfeld, C. A. (2018). Engineering nanoreactors using bacterial microcompartment architectures. Curr. Opin. Biotechnol. , 51, 1 Œ7. https://doi.org/10.1016/j.copbio.2017.09.005 Ramanavicius, A., & Ramanaviciene, A. (2009). Hemoproteins in design of bio fuel cells. Fuel Cells , 9(1), 25Œ36. https://doi.org/10.1002/fuce.200800052 Reedy, C. J., & Gibney, B. R. (2004). Heme Protein Assemblies. Chem. Rev. , 104(2), 617 Œ650. https://doi.org/10.1021/cr0206115 Robertson, D. E., Farid, R. S., Moser, C. C., Urbauer, J. L., Mulholland, S. E., Pidikiti, R., Lear, J. D., Wand, A. J., DeGrado, W. F., & Dutton, P. L. (1994). Design and synthesis of multi -haem proteins. Nature , 368(6470), 425 Œ432. https://doi.org/10.1038/368425a0 Rutherford, A. W., Osyczka, A., & Rappaport , F. (2012). Back -reactions, short -circuits, leaks and other energy wasteful reactions in biological electron transfer: Redox tuning to survive life in O 2. FEBS Lett. , 586(5), 603 Œ616. https://doi.org/10.1016/j.febslet.2011.12.039 Ruzgas, T., Gaigalas, A., & Gorton, L. (1999). Diffusionless electron transfer of microperoxidase -11 on gold electrodes. J. Electroanal. Chem. , 469(2), 123 Œ131. https://doi.org/10.1016/S0022 -0728(99)00194 -1 Sharp, T. H., Bruning, M., Mantell, J., Sessions, R. B., Thomson, A. R., Zaccai, N. R., Brady, R. L., Verkade, P., & Woolfson, D. N. (2012). Cryo -transmission electron microscopy structure of a gigadalton peptide fiber of de novo design. PNAS , 109(33), 13266 Œ13271. https://doi.org/10.1073/pnas.1118622109 Shimizu, T., Hirano, K., Takahashi, M., Hatano, M., & Fujii -Kuriyama, Y. (1988). Site -directed mutageneses of rat liver cytochrome P -450d: axial ligand and heme incorporation. Biochemistry , 27(11), 4138Œ4141. https://doi.org/10.1021/bi00411a035 Sinha, S., Cheng, S., Sung, Y. W., McNamara, D. E., Sawaya, M. R., Yeates, T. O., & Bobik, T. A. (2014). Alanine scanning mutagenesis identifies an Asparagine ŒArginine ŒLysine triad essential to assembly of the shell of the Pdu microcompartment. J. Mol. Biol. , 426(12 ), 2328 Œ2345. https://doi.org/10.1016/j.jmb.2014.04.012 132 Sommer, M., Sutter, M., Gupta, S., Kirst, H., Turmo, A., Lechno -Yossef, S., Burton, R. L., Saechao, C., Sloan, N. B., Cheng, X., Chan, L.J. G., Petzold, C. J., Fuentes -Cabrera, M., Ralston, C. Y., & K erfeld, C. A. (2019). Heterohexamers formed by CcmK3 and CcmK4 increase the complexity of Beta carboxysome shells. Plant Physiol. , 179(1), 156 Œ167. https://doi.org/10.1104/pp.18.01190 Sutter, M., Faulkner, M., Aussignargues, C., Paasch, B. C., Barrett, S., Kerfeld, C. A., & Liu, L.N. (2016). Visualization of bacterial microcompartment facet assembly using high -speed atomic force microscopy. Nano Lett. , 16(3), 1590 Œ1595. https://doi.org/10.1021/acs.nanolett.5b04259 Sutter, M., Greber, B., Aussignargues, C., & Kerfeld, C. A. (2017). Assembly principles and structure of a 6.5 -MDa bacterial microcompartment shell. Science , 356(6344), 1293 Œ1297. https://doi.org/10.1126/science.aan3289 Thompson, M. C., Cascio, D., Leibly, D. J., & Yeates, T. O. (2015). An allosteric model for control of pore opening by substrate binding in the EutL microcompartment shell protein. Protein Sci. , 24(6), 956 Œ975. https://doi.org/10.1002/pro.2672 Thony -Meyer, L., Fischer, F., Künzler, P., Ritz, D., & Hennecke, H. (1995). Escherichia coli genes required for cytochrome c maturation. J. Bacteriol. , 177 (15), 4321 Œ4326. http://www.ncbi.nlm.nih.gov/pubmed/7635817 Verbaro, D., Hagarman, A., Kohli, A., & Schweitzer -Stenner, R. (2009). Microperoxidase 11: a model system for porphyrin networks and heme Œprotein interactions. J. Biol. Inorg. Chem. , 14(8), 1289 Œ1300. https://doi.org/10.1007/s00775 -009-0574 -9 Waterhouse, A., Bertoni, M., Bienert, S., Studer, G., Tauriello, G ., Gumienny, R., Heer, F. T., de Beer, T. A. P., Rempfer, C., Bordoli, L., Lepore, R., & Schwede, T. (2018). SWISS -MODEL: homology modelling of protein structures and complexes. Nucleic Acids Res. , 46(W1), W296 ŒW303. https://doi.org/10.1093/nar/gky427 Wils on, J. R., Caruana, D. J., & Gilardi, G. (2003). Engineering redox functions in a nucleic acid binding protein. Chem. Commun. , (3), 356 Œ357. https://doi.org/10.1039/b209443f Yeates, T. O., Kerfeld, C. a, Heinhorst, S., Cannon, G. C., & Shively, J. M. (2008 ). Protein -based organelles in bacteria: carboxysomes and related microcompartments. Nat. Rev. Microbiol. , 6(9), 681 Œ691. https://doi.org/10.1038/nrmicro1913 Young, E. J., Burton, R., Mahalik, J. P., Sumpter, B. G., Fuentes -Cabrera, M., Kerfeld, C. A., & D ucat, D. C. (2017). Engineering the bacterial microcompartment domain for molecular scaffolding applications. Front. Microbiol. , 8(JUL), 1 Œ9. https://doi.org/10.3389/fmicb.2017.01441 Zamponi, S., Santucci, R., Brunori, M., & Marassi, R. (1990). A spectroel ectrochemical study of microperoxidase at bare and gold -plated RVC thin -layer electrodes. Biochim. Biophys. Acta - Gen. Subj. , 1034(3), 294 Œ297. https://doi.org/10.1016/0304 -4165(90)90054 -Z 133 CHAPTER 4: CONCLUSION AND PERSPECTIVE 134 Overview This dissertation work presents two distinct approaches that potentially arrange redox cofactor hemes into a highly organized array that could support electron transfer (ET) through multi -step hopping mechanism (Figure 4.1) . In Chapter 1, the theoretical importance of having structural precision for biological ET and the demand for structural ly defined platforms for protein -based ET are elaborated upon . In order to provide a platform for ET with defined structural information, Chapter 2 demonstrates the use of atomic -level structures to describe long -range ET in a cytochrome crystal. Chapter 3, conceived from an engineering prospective, proposes an approach for designing protein platforms capable of long -range ET propert ies by functionalizing a self -assembl ing scaffold protein with hemes. With these two distinct approaches in mind , this Chapter will compare the values and limitations of each platform and discuss potential applications of long -range ET devices. 135 Figure 4.1 Graphic overview Graphic scheme of this thesis. Top panel shows small structural differences may lead to large conductivity variations. Lower panels show the strategies of building structurally defined platforms described in Chapter 2 & 3. Comparison of two platforms In order to form larg e and organized structures that can support long -range ET, it is necessary to bring individual redox proteins into close proximity with periodical protein -136 protein interactions. Both crystallization and self -assembl y are possible methods to achieve such a goal, yet there are advantages and limit ations of each method. To form a crystal, proteins must be organized in a lattice with one of the sixty -five possible crystal space group s (Wukovitz and Yeates , 1995), and the crystal lattice defines the overall shape of the crystals as well as interprotein contacts. Due to the highly repeated arrangement, the atomic coordinates can be determined precisely by X -ray scattering, which provides Angstrom -level structures f or the entire crystal which extends over hundreds of micrometers (Giegé , 2013). In addition to the precision of structur al information , the highly organized and periodical nature of the crystal possesses many optical features, for example, optical transpar ency, linear and circular dichroism and birefringence (Ronda et al. , 2015). Since many of the biological redox cofactors have absorption within the UV-Vis range and many of these optical transitions are polarized (Hofrichter and Eaton , 1976), much ET-relat ed information can be measured by far -field optical approaches. For instance, the STC crystal in Chapter 2 shows a linear dichroism since the absorption coefficient of hemes are polarized ( e.g. the Soret band is polarized to the heme plane), and it is poss ible to distinguish the redox states of each heme by the polarization ratio between crystal a-/b- axes , which is impossible to measure for STC in solution. In the meantime, due to the large dimension of protein crystals, near -field optical measurement is a lso possible. For example, the vibrational anisotropy of a protein crystal can be measured by Terahertz spectroscopy (Acbas et al. , 2014). In summary , since many ET related parameters can be determined unambiguously, this crystal platform is thus potentially a powerful testbed for analyzing long -range ET theories. 137 The major limitation for this crystal platform is, quite obviously, the lack of control on the crystallization process. Although the screening and optimization of crystal growth conditions have been streamlined, it is still difficult to predict which protein can be crystallized or what interprotein contacts will be generate d from the crystallization process (Gavezzotti , 1994). STC was chosen based on its high heme density, which increase s the possibility of possessing multiple interprotein heme to heme interaction s. In contrast the hypothetical natural finano wire fl-forming decaheme cytochrome MtrF does not form a crystal -wide interprotein heme network supporting long -range ET (Clarke et al ., 2011). Fortunately, many novel approaches , for example using computational design (Lanci et al., 2012) or nanobodies (Bukowska and Grütter , 2013) have been develope d to assist the formation of protein crystals. These approaches seek to rationally design protein - protein interactions and guide the protein monomers to self -assemble towards the formation of crystal lattices. The designs of self -assembling structures will be discussed below. The development of the heme -functionalized self -assembling protein in Chapter 3 was an attempt to find a compromise between increasing designing plasticity and minimizing structural flexibility. Protein self -assembly has been a widely used strategy to generate large biological architecture s with highly organized structures (Luo et al ., 2016). Although crystallization is one form of self -assembly, in this chapter we will focus on the concept of specific and physiological intera ctions that bring individual proteins to associate with each other . Protein -protein association can be mediated by ligand -receptor interaction, electrostatic attraction, shape compl ementary, etc . Self -assembly brings proteins into a periodic array through repeating p rotein -protein associations. Although t he 138 atomic coordinates of protein assembly cannot be determined directly by X -ray scattering, it can be estimated if the structure of the individual protein and the assembl y pattern are known (King et al ., 2014). Since t he st ructural basis of hypothetical self -assembl ing nanowires ( e.g . PilA (Reardon and Mueller , 2013) or OmcS (Wang et al ., 2019) ) is not clear at the time of this writing, utilizing current knowledge of protein self -assemb ly is a plausible solution to build bio nanowires for applications or as an analog for a microbial nanowire . With the goal of defined structure in mind, Chapter 3 suggests an approach to introduce a redox cofactor to a family of self -assembl ing protein s. Compared to commonly used protein fusions , e.g. (Altamura et al ., 2017), the heme -attaching approach in Chapter 3 not only is able to introduce one or more hemes to the middle of a polypeptide, but also can diminish the structural uncertainty associate d with the flexibility of the N -/C-ter minus or linker motifs. Although the results in Chapter 3 fail to reproduce the formation of nanotubes, the developed heme attachment strategy can be widely applied to a huge variety of self -assembly proteins (see below) . The major limitation of this appro ach is that there is a very limited type of redox cofactor that can be covalently connected to a protein in vivo . c-Type heme was chosen due to its simple binding motif and the presence of efficient and fipromiscuousfl maturation pathway. Redox cofactor s tha t do not bind protein tightly (e.g. non-heme iron or copper centers), or are not stable ( e.g. iron -sulphur cluster s), require protein d esign with multiple spheres of a binding network in order to stabilize the cofactor (Peacock , 2013), which needs much mor e engineering efforts than the strategy presented here. Cofactors like flavin (Decker , 1993) or phycocyanin (Mancini et al ., 2018) 139 (for exciton transfer) can form stable covalent bond s and might be potential engineering target s to attach to self -assembling proteins. The two platforms presented in this dissertation work demonstrated the merits of using structural ly defined platforms for studying long -range ET and the potential of using self -assembly strategies to construct such platforms. Yet, extensive improvements are still required in future work in order to enhance the conductivity and plasticity of such systems . What makes a highly conductive and robust bionanowire? Organic materi als , including biological matters , are usually found as insulators, but conducting organic polymers are widely used in daily life as well. Organic (semi -)conductors are commonly made with heavily conjugated p-orbitals that allow a metallic conductivity (Pr on and Rannou , 2002; MacDiarmid , 2001). This mechanism would provide high conductivity, but it is also hard to achieve for a biologist since biological building blocks like nucleic acid, lipid or polypeptide chain are not fully conjugated. Other than conne cting directly by conjugated covalent bonds, p-orbital electrons can interact and delocalize by inter - -stacking between aromatic compounds (Xu et al ., 2010) -Stacking is found between bases in nucleic acid (Kelley , 1999) and has been proposed as the ET mechanism in the PilA protein -based nanowire (Malvankar et al ., 2011) -orbital overlap, the distances between each aromatic molecule need to be short, e.g. less than 3.6 Å (Xu et al. , 2010) for two parallel molecules. This small distance may pose challenges for engineering since this distance is about th e same size as amino acids and is even smaller than the repeating 140 unit in some common secondary structures ( e.g. the alpha -helix pitch is 5.4 Å ). In the model of PilA nanowire s, aromatic residues from multiple polypeptide chains stack alternatively (Reardo n and Mueller , 2013; Malvankar et al ., 2015). Because the conductivity mechanism of Geobacter ™s nanowire is not fully elucidated and the precision required to build metallic -like bionanowires by proteins is beyond the scope of this dissertation, constructi ng metallic -like wires will not be discussed in the following design principles. Although c onstructing metallic -like bionanowires can be a formidable task , synthesizing bionanowires for multistep hopping using in vivo biosynthesized proteins has been demo nstrated (Altamura et al ., 2017). In the regime of multistep nonadiabatic ET, the rate of ET generally depends on two factors, the distance between redox centers and the activation energy (see Chapter 1) . In a protein polymer, assuming redox centers are evenly distributed (Dalton et al ., 1990), the distance between redox centers can be approximated by redox factor density. Multiheme cytochromes with the high heme to amino acid ratios are found in STC or MtrA of Shewanella (Meyer et al ., 2004), with about 4 hemes per 100 amino acids. In STC, this density gives an approximate 1 heme per nanometer ( or 0.5 nm edge -to-edge distance). This heme density in theory can support 109 s-1 electron flow if the activation energy is low (see Chapter 2 ). Therefore, from an engineering prospective, identifying the small redox cofactor binding domains and small self -assembling domains could be an initial step to create nanowires with high redox factor density. In the work of Altamura et al ., they used the smallest rubredoxin (44 residues with one Fe center ) for making electron tunneling possible (Altamura et al . 2017) . For hemes, the minimal heme binding motif that can be matured by CCM -I in vivo was 141 identified by Th ony-Meyer et al ., to be about 15 amino acids (Braun and Th ony-Meyer , 2004). Based on this finding, this dissertation proposed a similar approach of using CCM -I for adding hemes without introducing extra amino acids. With these small redox domains and a small and robust self -assembly domain, there are many potentia l combinations for producing conductive biomaterials with different properties. Besides electronic coupling, lowering the activation energy ( Ea) will also improve the rate of ET. Based on Marcus theory, the activation - G0 = - 0 must be negative or exothermic (Page et al . 1999) . Such a reaction accelerates ET with the expense of free energy at each step, which is not applicable for multistep hopping without an external electric field ( i.e. driven by electrodes). Chapter 2 demonstr ates that if an ET pathway is composed of redox cofactors with different midpoint potentials ( Em'), the energetically uphill steps can be rate limiting, assuming the electronic couplings are similar. It is thus optimal to have all the redox centers with the same Em', so the activation G0 = 0). This can be simply accomplished by using one kind of redox protein repeatedly. Another potential advantage of usi ng redox centers with the same Em' is that there will be a higher chance for flickering resonance to occur. As described in Chapter 1, the flickering resonance requires bridging redox centers to reach the same energy level transiently. As a parallel mechan ism to multistep hopping, flickering resonance supports additional electron tunneling through multiple redox centers with one thermal activation step and accelerates the overall ET rate. In the meantime, rease the electron flow. One possible design principle for decreas ing the surface -exposed area of redox centers (Bortolotti et al ., 2011), for example, to shield a redox center through the use of the protein 142 medium . The preliminary design of a nanowire by a shell -cytochrome in Chapter 3 does not consider this factor, but future improvement could complete th is design by adding an extra domain that can act as both sixth axial ligand as well as a hydrophobic pocket. In addition, the shielding of redox cofactors from solvent will potentially reduce side reactions. For instance, molecular oxygen is a widespread electron acceptor and can react with low -potential electrons inside nanowires. In Chapter 2, the STC crystals show the capacity to decrea se the oxidation rate compare d to fully solvent -exposed STC molecules. With these ET parameters in mind, I propose several forward -thinking ideas for constructing protein nanowires. Although both electronic coupling and activation energy influence the fin al rate of ET, due to the exponential decay of electronic coupling over distance, bringing redox cofactors close and increasing the redox cofactor density should be the dominant priority when designing the fifirst -generationfl of protein nanowires. Protein s elf -assembly is a potential approach for rational design of a large -scale closely positioned protein scaffold, and the strategies of protein self -assembly can be found in several reviews (Luo et al ., 2016; Bai, Luo, and Liu , 2016) . Here, I will highlight several strategies that could be handy for designing protein nanowires. To organize existing redox proteins (or redox domains) into a periodic array, the protein surface can be reengineered to mediate high -affinity protein -protein interaction. Computer assisted design has been shown to be a powerful tool for introducing self -assembling properties to naïve proteins (Stranges et al ., 2011; Parmeggiani et al ., 2015). This approach allows the rational design of the interfaces for redox centers as well as for the ET pathway in the system. With proper design of protein -protein interaction geometry, 143 self -assembling proteins can create 2 -D and even 3 -D protein crystals (Gonen et al ., 2015; Lanci et al ., 2012). Alternatively, attaching redox domains to miniature self -assembling proteins is another potentially feasible approach to make highly conductive protein nanowires. The research in Chapter 3 chose one self -assembling shell protein as the scaffold to demonstrate the heme -attachment method s. However, because shell proteins are relatively large (about 100 residues), obtaining heme density similar to STC will be challenging. In future studies, attaching hemes to smaller self -assembling proteins can be a more practical approach to increase the heme density. Well -studied small self -assembling domains, like coiled -coil s and amyloid - -sheets (Luo et al ., 2016), are among the best candidates for heme attachment. Figure 3.1B shows that the heme binding motif tends to adapt to a helical secondar y structure. Thus, attaching hemes to coiled -coil structures should be achievable. The diheme cytochrome maquette developed by Anderson and Dutton (Anderson et al ., 2014) can be used as the template for introducing heme binding motifs and their axial histidine ligands, while modify ing its four helical bundles to promote interprotein coiled -coil assembly (Sharp et al ., 2012). The covalent attachment of heme to -sheet s may not be possible since the two cysteines and the histidine in the heme binding motif (CxxCH) must be close ( i.e. at the same side of a -sheet ). However, inserting a heme binding motif in a loop region between two -strands should be possible, as suggested by the design of heme attachment site 4 in Chapter 3. 144 Applications of biological conductive polymers Electronics is an essential component in the Information Era. Biological conductive polymers are believed to have values in many applications including m edical (Cohen -Karni, Langer, and Kohane , 2012) and bioenergy (Lee et al ., 2013). Although the conductivity of biological material is hardly comparable with metals, key advantages including biocompatibi lity and biodegradab ility makes them ideal material for implants (Tan et al ., 2016). The potential of generating renewable energy using bioelectronic material or living cells has also been investigated including microbial fuel cells (Li, Lesnik, and Liu , 2017), microbial photovoltaic (Schuergers et al ., 2017) and bio -inspired photosynthesis (Lee et al ., 2016; Mancini et al ., 2018; Nam et al ., 2010) . In vivo applications of bionanowires are seldomly mentioned. One potential application for biological nanowires is to construct living photovoltaics by phot osynthetic microorganisms . Photosynthetic reaction centers generate high energy (low -redox potential) electrons with high efficiency (Blankenship et al ., 2011). Being able to redirect these electrons to other high -value products ( e.g. biofuel, plastic or m edicine) or to electrodes for electricity can reduce the need for fossil fuel and lower the carbon footprint. The presence of biological nanowires can be beneficial for this synthetic biology goal. In the simple unicellular oxygenic photosynthetic bacteria like cyanobacteria, the photosynthetic reaction centers are embed ded in the thylakoid membrane, which is further surrounded by two membranes, plasma membrane and outer membrane. To draw electrons from reaction centers towards electrodes, these electrons n eed to first be solvat ed and transport ed on soluble electron carriers ( e.g. ferredoxin, NADPH) within the cytosol, then transfer red to lipid -soluble carriers like quinol in the plasma membrane, and 145 once again transfer red to water -soluble carriers ( e.g. cytochrome, flavin) to pass through the porous outer -membrane to the extracellular environment. Since the redox potentials of cytosol, periplasm and plasma membrane are tightly regulated, the energy of a single electron wil l be much lower than the high energy electron produced by reaction centers. For example, electron s redirected from Photosystem I can drive hydrogen production (Lubner et al. , 2010), while only low energy (more positive potential) electron s can be obtain ed from extracellular electrode contact (Cereda et al. , 2014). Therefore, building an orthogonal ET pathway that is isolate d from the cellular redox pool would be essential for this application and bionanowires could be an ideal method. For example, f uture st udies could use photosystem subunits like PsaC or PsaE protein to anchor alternative electron acceptors to redirect electrons into nanowires (Figure 4.2) . In summary, natural microbial nanowires demonstrate that biological material can be highly electroni cally conductive. Inspired by nature and previous studies on naturally occurring microbial nanowires, my dissertation research attempted to reproduce the conductivity of natural nanowires from two aspects. Although the design of current platforms cannot su pport similar conductivity to natural bionanowires, these strategies are general design principles and can be applied to a great many of other potential target proteins. I hope the results in this dissertation work could inspire future studies on creating biological conductive materials. 146 Figure 4.2 Graphic demonstration of a nanowire application An artistic image of a possible application of bionanowire that can transfer electrons from Photosystem I to a cytosolic nanoreactor. 147 REFERENCES 148 REFERENCES Acbas, G., Niessen, K. A., Snell, E. H., & Markelz, A. G. (2014). Optical measurements of long -range protein vibrations. Nat. Commun. , 5(1), 3076. https://doi.org/10.1038/ncomms4076 Altamura, L., Horvath, C., Rengaraj, S., Rongier, A., Elouarzaki, K., Gondran, C., Maçon, A. L. B., Vendrely, C., Bouchiat, V., Fontecave, M., Mariolle, D., Rannou, P., Le Goff, A., Duraffourg, N., Holzinger, M., & Forge, V. (2017). A synthetic redox biofi lm made from metalloprotein Œprion domain chimera nanowires. Nat. Chem. , 9(2), 157 Œ163. https://doi.org/10.1038/nchem.2616 Anderson, J. L. R., Armstrong, C. T., Kodali, G., Lichtenstein, B. R., Watkins, D. W., Mancini, J. a, Boyle, A. L., Farid, T. a, Crump , M. P., Moser, C. C., & Dutton, P. L. (2014). Constructing a man -made c-type cytochrome maquette in vivo : electron transfer, oxygen transport and conversion to a photoactive light harvesting maquette. Chem. Sci. , 5(2), 507 Œ514. https://doi.org/10.1039/C3S C52019F Bai, Y., Luo, Q., & Liu, J. (2016). Protein self -assembly via supramolecular strategies. Chem. Soc. Rev. , 45(10), 2756 Œ2767. https://doi.org/10.1039/C6CS00004E Blankenship, R. E., Tiede, D. M., Barber, J., Brudvig, G. W., Fleming, G. R., Ghirardi, M., Gunner, M. R., Junge, W., Kramer, D. M., Melis, A., Moore, T. A., Moser, C. C., Nocera, D. G., Nozik, A. J., Ort, D. R., Parson, W. W., Prince, R. C., & Sayre, R. T. (2011). Comparing photosynthetic and photovoltaic efficiencies and recognizing the potential for improvement. Science , 332(6031), 805 Œ809. https://doi.org/10.1126/science.1200165 Bortolotti, C. A., Siwko, M. E., Castellini, E., Ranieri, A., Sola, M., & Corni, S. (2011). The reorganization energy in cytochrome c is controlled by the accessib ility of the heme to the solvent. J. Phys. Chem. Lett. , 2(14), 1761 Œ1765. https://doi.org/10.1021/jz200734a Braun, M., & Thony -Meyer, L. (2004). Biosynthesis of artificial microperoxidases by exploiting the secretion and cytochrome c maturation apparatuses of Escherichia coli . PNAS , 101(35), 12830 Œ12835. https://doi.org/10.1073/pnas.0402435101 Bukowska, M. A., & Grütter, M. G. (2013). New concepts and aids to facilitate crystallization. Curr. Opin. Struct. Biol. , 23(3), 409 Œ416. https://doi.org/10.1016/j.sbi.2013.03.003 Cereda, A., Hitchcock, A., Symes, M. D., Cronin, L., Bibby, T. S., & Jones, A. K. (2014). A bioelectrochemical approach to characterize extracellular electron transfer by Synechocystis sp. PCC6803. PLoS One , 9(3), e91484. https://doi.org/10.1371/journal.pone.0091484 Clarke, T. A., Edwards, M. J., Gates, A. J., Hall, A., White, G. F., Bradley, J., Reardon, C. L., Shi, L., Beliaev, A. S., Marshall, M. J., Wang, Z., Watmough, N. J., Fredrickson, J. K., Zachara, J. M., Butt, J. N., & Richardson, D. J. (2011). Structure of a bacterial cell surface decaheme electron conduit. PNAS , 108(23), 9384 Œ9389. https://doi.org/10.1073/pnas.1017200108 Cohen -Karni, T., Langer, R., & Kohane, D. S. (2012). The Smartest materials: The future of nanoelectronics in medicine. ACS Nano , 6(8), 6 541 Œ6545. https://doi.org/10.1021/nn302915s Dalton, E. F., Surridge, N. A., Jernigan, J. C., Wilbourn, K. O., Facci, J. S., & Murray, R. W. (1990). Charge transport in electroactive polymers consisting of fixed molecular redox sites. Chem. Phys. , 141(1), 143Œ157. https://doi.org/10.1016/0301 -0104(90)80026 -T Decker, K. F. (1993). Biosynthesis and function of enzymes with covalently bound flavin. Annu. Rev. Nutr. , 13(1), 17 Œ41. https://doi.org/10.1146/annurev.nu.13.070193.000313 Gavezzotti, A. (1994). Are crystal structures predictable? Acc. Chem. Res. , 27(10), 309 Œ314. 149 https://doi.org/10.1021/ar00046a004 Giegé, R. (2013). A historical perspective on protein crystallization from 1840 to the present day. FEBS J., 280(24), 6456 Œ6497. https://doi.org/10.1111/febs.12580 Gonen, S., DiMaio, F., Gonen, T., & Baker, D. (2015). Design of ordered two -dimensional arra ys mediated by noncovalent protein -protein interfaces. Science , 348(6241), 1365 Œ1368. https://doi.org/10.1126/science.aaa9897 Hofrichter, J., & Eaton, W. A. (1976). Linear dichroism of biological chromophores. Annu. Rev. Biophys. Bioeng. , 5(1), 511 Œ560. ht tps://doi.org/10.1146/annurev.bb.05.060176.002455 Kelley, S. O. (1999). Electron transfer between bases in double helical DNA. Science , 283 (5400), 375 Œ381. https://doi.org/10.1126/science.283.5400.375 King, N. P., Bale, J. B., Sheffler, W., McNamara, D. E. , Gonen, S., Gonen, T., Yeates, T. O., & Baker, D. (2014). Accurate design of co -assembling multi -component protein nanomaterials. Nature , 510 (7503), 103Œ108. https://doi.org/10.1038/nature13404 Lanci, C. J., MacDermaid, C. M., Kang, S. -g., Acharya, R., N orth, B., Yang, X., Qiu, X. J., DeGrado, W. F., & Saven, J. G. (2012). Computational design of a protein crystal. PNAS , 109(19), 7304 Œ7309. https://doi.org/10.1073/pnas.1112595109 Lee, J. -H., Lee, J. H., Lee, Y. J., & Nam, K. T. (2013). Protein/peptide bas ed nanomaterials for energy application. Curr. Opin. Biotechnol. , 24(4), 599 Œ605. https://doi.org/10.1016/j.copbio.2013.02.004 Lee, S. -H., Blake, I. M., Larsen, A. G., McDonald, J. A., Ohkubo, K., Fukuzumi, S., Reimers, J. R., & Crossley, M. J. (2016). Syn thetically tuneable biomimetic artificial photosynthetic reaction centres that closely resemble the natural system in purple bacteria. Chem. Sci. , 7(10), 6534 Œ6550. https://doi.org/10.1039/C6SC01076H Li, C., Lesnik, K. L., & Liu, H. (2017). Stay connected: Electrical conductivity of microbial aggregates. Biotechnol. Adv. , 35(6), 669 Œ680. https://doi.org/10.1016/j.biotechadv.2017.07.010 Lubner, C. E., Kn örzer, P., Silva, P. J. N., Vincent, K. A., Happe, T., Bryant, D. A., & Golbeck, J. H. (2010). Wiring an [FeFe] -Hydrogenase with Photosystem I for light -induced hydrogen production. Biochemistry , 49(48), 10264 Œ10266. https://doi.org/10.1021/bi1016167 Luo, Q., Hou, C., Bai, Y., Wang, R., & Liu, J. (2016). Protein assembly: Versatile approaches to construct highly ordered nanostructures. Chem. Rev. , 116(22), 13571 Œ13632. https://doi.org/10.1021/acs.chemrev.6b00228 MacDiarmid, A. G. (2001). fiSynthetic met alsfl: a novel role for organic polymers. Curr. Appl. Phys. , 1(4Œ5), 269Œ279. https://doi.org/10.1016/S1567 -1739(01)00051 -7 Malvankar, N. S., Vargas, M., Nevin, K. P., Franks, A. E., Leang, C., Kim, B. -C., Inoue, K., Mester, T., Covalla, S. F., Johnson, J. P., Rotello, V. M., Tuominen, M. T., & Lovley, D. R. (2011). Tunable metallic -like conductivity in microbial nanowire networks. Nat. Nanotechnol. , 6(9), 573 Œ579. https://doi.org/10.1038/nnano.2011.119 Malvankar, N. S., Vargas, M., Nevin, K., Tremblay, P. -L., Evans -Lutterodt, K., Nykypanchuk, D., Martz, E., Tuominen, M. T., & Lovley, D. R. (2015). Structural basis for metallic -like conductivity in microbial nanowires. MBio , 6(2), e00084. https://doi.org/10.1128/mBio.00084 -15 Mancini, J. A., Sheehan, M., Koda li, G., Chow, B. Y., Bryant, D. A., Dutton, P. L., & Moser, C. C. (2018). De novo synthetic biliprotein design, assembly and excitation energy transfer. J. R. Soc. Interface , 15(141), 20180021. https://doi.org/10.1098/rsif.2018.0021 150 Meyer, T. E., Tsapin, A . I., Vandenberghe, I., De Smet, L., Frishman, D., Nealson, K. H., Cusanovich, M. A., & Van Beeumen, J. J. (2004). Identification of 42 possible cytochrome c genes in the Shewanella oneidensis genome and characterization of six soluble cytochromes. Omi. A J. Integr. Biol. , 8(1), 57 Œ77. https://doi.org/10.1089/153623104773547499 Nam, Y. S., Shin, T., Park, H., Magyar, A. P., Choi, K., Fantner, G., Nelson, K. A., & Belcher, A. M. (2010). Virus -te mplated assembly of porphyrins into Light -harvesting nanoantennae. J. Am. Chem. Soc. , 132(5), 1462 Œ1463. https://doi.org/10.1021/ja908812b Page, C. C., Moser, C. C., Chen, X., & Dutton, P. L. (1999). Natural engineering principles of electron tunnelling in biological oxidation -reduction. Nature , 402(6757), 47 Œ52. https://doi.org/10.1038/46972 Parmeggiani, F., Huang, P. -S., Vorobiev, S., Xiao, R., Park, K., Caprari, S., Su, M., Seetharaman, J., Mao, L., Janjua, H., Montelione, G. T., Hunt, J., & Baker, D. (2015). A general computational approach for repeat protein design. J. Mol. Biol. , 427(2), 563 Œ575. https://doi.org/10.1016/j.jmb.2014.11.005 Peacock, A. F. A. (2013). Incorporating metals into de novo proteins. Curr. Opin. Chem. Biol. , 17(6), 934 Œ939. https://doi.org/10.1016/j.cbpa.2013.10.015 Pirbadi an, S., & El -Naggar, M. Y. (2012). Multistep hopping and extracellular charge transfer in microbial redox chains. Phys. Chem. Chem. Phys. , 14(40), 13802. https://doi.org/10.1039/c2cp41185g Pron, A., & Rannou, P. (2002). Processible conjugated polymers: fro m organic semiconductors to organic metals and superconductors. Prog. Polym. Sci. , 27(1), 135 Œ190. https://doi.org/10.1016/S0079 -6700(01)00043 -0 Reardon, P. N., & Mueller, K. T. (2013). Structure of the Type IVa major Pilin from the electrically conductive bacterial nanowires of Geobacter sulfurreducens . J. Biol. Chem. , 288(41), 29260 Œ29266. https://doi.org/10.1074/jbc.M113.498527 Ronda, L., Bruno, S., Bettati, S., Storici, P., & Mozzarelli, A. (2015). From protein structure to function via single crystal optical spectroscopy. Front. Mol. Biosci. , 2(April), 12. https://doi.org/10.3389/fmolb.2015.00012 Schuergers, N., Werlang, C., Ajo -Franklin, C. M., & Boghossian, A. A. (2017). A synthetic biology approach to engineering living photovoltaics. Energy Environ. Sci. , 10(5), 1102 Œ1115. https://doi.org/10.1039/C7EE00282C Sharp, T. H., Bruning, M., Mantell, J ., Sessions, R. B., Thomson, A. R., Zaccai, N. R., Brady, R. L., Verkade, P., & Woolfson, D. N. (2012). Cryo -transmission electron microscopy structure of a gigadalton peptide fiber of de novo design. PNAS , 109(33), 13266 Œ13271. https://doi.org/10.1073/pna s.1118622109 Stranges, P. B., Machius, M., Miley, M. J., Tripathy, A., & Kuhlman, B. (2011). Computational design of a symmetric homodimer using -strand assembly. PNAS , 108(51), 20562 Œ20567. https://doi.org/10.1073/pnas.1115124108 Tan, M. J., Owh, C., Che e, P. L., Kyaw, A. K. K., Kai, D., & Loh, X. J. (2016). Biodegradable electronics: cornerstone for sustainable electronics and transient applications. J. Mater. Chem. C , 4(24), 5531 Œ5558. https://doi.org/10.1039/C6TC00678G Wang, F., Gu, Y., O™Brien, J. P., Yi, S. M., Yalcin, S. E., Srikanth, V., Shen, C., Vu, D., Ing, N. L., Hochbaum, A. I., Egelman, E. H., & Malvankar, N. S. (2019). Structure of microbial nanowires reveals stacked hemes that transport electrons over micrometers. Cell , 177(2), 361 -369.e10. https://doi.org/10.1016/j.cell.2019.03.029 Wukovitz, S. W., & Yeates, T. O. (1995). Why protein crystals favour some space -groups over others. Nat. Struct. Biol. , 2(12), 1062 Œ1067. http://www.ncbi.nlm.nih.gov/pubmed/8846217 151 Xu, H., Das, A. K., Horie, M., S haik, M. S., Smith, A. M., Luo, Y., Lu, X., Collins, R., Liem, S. Y., Song, A., Popelier, P. L. A., Turner, M. L., Xiao, P., Kinloch, I. A., & Ulijn, R. V. (2010). An investigation of the conductivity of peptide nanotube networks prepared by enzyme -trigger ed self -assembly. Nanoscale , 2(6), 960. https://doi.org/10.1039/b9nr00233b