CHARACTERIZATION AND DESIGN OF SHEWANELLA ONEIDENSIS AND ESCHERICHIA COLI AS CHASSIS FOR BIOELECTROCHEMICAL SYSTEMS By Megan Caroline Gruenberg Cross A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Biochemistry and Molecular Biology – Doctor of Philosophy 2024 ABSTRACT Microbial electrosynthesis (MES) is an emerging technology which, in the most extreme cases, uses an electroautotrophic biocatalyst to convert CO2 and clean electricity into value-added products. Technologies that combine CO2 remediation efforts with chemical production from fossil-fuel alternatives are promising contributors to the fight against climate change. MES is limited by presently available biocatalysts, which have small product scopes and few genetic engineering tools to improve them. For MES to be an industrially competitive technology, a robust biocatalyst which can be engineered to have a large product scope with high titers must be implemented. This dissertation describes progress toward the development of Shewanella oneidensis and Eschericahia coli as MES hosts. MES uses clean electricity to power the unfavorable reduction reactions that take place in carbon fixation and conversion to fuels and chemicals. Thus, an important aspect of a MES biocatalyst is electroactivity. S. oneidensis is a metal-reducing bacteria that can respire extracellular metals and electrodes. It has a well-characterized extracellular electron transfer pathway, the Mtr pathway, and established genetic engineering tools which make it an attractive candidate for MES. However, S. oneidensis cannot fix CO2. CO2-derived formate assimilation is considered for a CO2-fixing S. oneidensis strain. Formatotrophy would need to be engineered; however, formate is toxic to this organism. As a first step toward engineering formatotrophy in S. oneidensis, several formate tolerant strains were evolved and mutations were characterized (Chapter 2). To drive formate assimilation pathways, low potential electron carriers, such as reduced ferredoxin will be required. Therefore, we characterized the Rnf complex in S. oneidensis, a potential source of low potential electrons, which led to the discovery that Rnf is essential for iron-sulfur cluster biosynthesis and repair in this organism (Chapter 3). This work confirmed that the Rnf complex is a source of reduced ferredoxins in S. oneidensis and could serve as a power source for carbon fixation pathways in the future. In the course of our study of the S. oneidensis Rnf complex, the unusual observation that isopropyl β-D-1-thiogalactopyranoside (IPTG) altered growth of the organism was made. IPTG is often used to induce gene expression in engineered systems and is assumed to have no off-target effects on metabolism or regulation. However, in this work IPTG was found to improve growth of S. oneidensis cultivated with sugar substrate N-acetylglucosamine (Chapter 4). As the use of IPTG is so widespread, researchers in this field should be mindful of our results so false positive correlations can be avoided. Another interesting candidate for a MES biocatalyst is Escherichia coli, one of the most well-studied microorganisms that has been engineered for fermentative biosynthesis of a vast array of compounds. The ability to connect these reactions to CO2 fixation would offer viable means of production of such compounds from a fossil-fuel alternative while helping prevent release of CO2. E. coli requires both the engineering of electroactivity and CO2 fixation to be considered a viable option for MES, as the organism does not natively have either capability. In this work, the biotechnological potential of an electroactive strain of E. coli expressing the Mtr pathway from S. oneidensis is explored (Chapter 5). Overall, this dissertation describes four experimental studies which contribute to overcoming the technical barriers of using S. oneidensis and E. coli as MES biocatalysts. Copyright by MEGAN CAROLINE GRUENBERG CROSS 2024 This work is dedicated to my husband and our growing family. v ACKNOWLEDGEMENTS Graduate school can be a tough time. Ideally, you leave with much more knowledge and practical experience than you started with, and getting there is not an easy task. I am incredibly lucky and grateful that I had support from every front during the completion of my PhD that did away with the stereotype that graduate school can be one of the most difficult times of an academic’s life. My advisor, Michaela, was the foundation of my daily academic support. First, she took a chance, and accepted me into her lab during the first summer of COVID. Due to safety regulations in place at the time, I wasn’t even allowed in the lab for a few months. That time wasn’t wasted. I spent these months reading the literature and creating plans for when I would be allowed to start bench work. Michaela never let me fall between the cracks, even when we only existed to each other on Zoom and Slack. During the course of my PhD, she was always able to give me ideas and papers to read when I was feeling stuck. Michaela was always supportive in the lab, and she was also supportive of my personal goals. For example, when I wanted to take a few weeks off for my wedding and honeymoon during my final year, she was supportive, even though that meant I would be pausing a months-long experiment. The culture of growth and support Michaela cultivated in the lab set the standard for how lab members were expected to treat one another. I worked alongside people who warmly welcomed me and were always willing to help, including Dr. Kody Duhl, Dr. Magdalena Felczak, Dr. Kathryne Ford, Dr. Elhussiny Aboulnaga, Nicholas Tefft, Shaylynn Miller, and Emma Boismier. I will truly miss working with this team of dedicated scientists. vi I also received much appreciated support from my committee members, Dr. Robert Hausinger, Dr. Claire Vieille, Dr. Daniel Ducat, and Dr. Bjoern Hamberger. I always felt as though committee meetings were a productive exchange of ideas that gave me new directions to explore in my research. I am especially grateful to my committee members for advising me more frequently during the summer that Michaela was on maternity leave. While I was pretty confident in my direction at the beginning of the summer, research always has a way of throwing your plans out of the window. Their guidance was instrumental in keeping me productive that summer. Finally, I would like to thank my friends and family for their support from home. They stood behind me when I decided to change directions for my PhD that was going to add a few years onto my time in graduate school. While that was a difficult decision, I am ultimately a more capable scientist with a new outlook on what successful science looks like. My husband, parents, siblings, and close friends all served as much needed personal advisors during that time. I am so grateful that they supported me in taking a leap of faith and that the support I was receiving from my new environment lessened their burden. I am very proud of the work I’ve done over the last four years, especially because of all the wonderful memories that came with it. Thank you to everyone who made my PhD experience so fulfilling. I couldn’t have done it without you. vii TABLE OF CONTENTS LIST OF TABLES ............................................................................................................ x LIST OF FIGURES .......................................................................................................... xi LIST OF ABBREVIATIONS ........................................................................................... xiv CHAPTER 1: INTRODUCTION ....................................................................................... 1 CLIMATE CHANGE ..................................................................................................... 2 RENEWABLE ENERGY .............................................................................................. 4 MICROBIAL ELECTROSYNTHESIS ........................................................................... 5 SHEWANELLA ONEIDENSIS AS A MES HOST ......................................................... 8 OTHER CONSIDERATIONS FOR S. ONEIDENSIS AS A MES HOST ..................... 16 ESCHERICHIA COLI AS A MES HOST .................................................................... 17 REFERENCES .......................................................................................................... 21 CHAPTER 2: A SMALL NUMBER OF POINT MUTATIONS CONFER FORMATE TOLERANCE IN SHEWANELLA ONEIDENSIS ............................................................ 31 PREFACE .................................................................................................................. 32 ABSTRACT ................................................................................................................ 33 INTRODUCTION ....................................................................................................... 34 METHODS ................................................................................................................. 37 RESULTS .................................................................................................................. 43 DISCUSSION ............................................................................................................ 61 FUNDING SOURCES ................................................................................................ 64 ACKNOWLEDGEMENTS .......................................................................................... 65 REFERENCES .......................................................................................................... 66 CHAPTER 3: RNF COMPLEX GENERATES REDUCED FERREDOXIN FOR IRON- SULFUR CLUSTER FORMATION IN SHEWANELLA ONEIDENSIS MR-1 .................. 70 PREFACE .................................................................................................................. 71 ABSTRACT ................................................................................................................ 72 INTRODUCTION ....................................................................................................... 73 METHODS ................................................................................................................. 76 RESULTS .................................................................................................................. 79 DISCUSSION ............................................................................................................ 90 REFERENCES .......................................................................................................... 96 CHAPTER 4: A COMMON INDUCER MOLECULE ENHANCES SUGAR UTILIZATION BY SHEWANELLA ONEIDENSIS MR-1 ...................................................................... 101 PREFACE ................................................................................................................ 102 ABSTRACT .............................................................................................................. 102 INTRODUCTION ..................................................................................................... 103 RESULTS ................................................................................................................ 106 DISCUSSION .......................................................................................................... 115 viii METHODS ............................................................................................................... 118 ACKNOWLEDGEMENTS ........................................................................................ 120 FUNDING SOURCES .............................................................................................. 120 REFERENCES ........................................................................................................ 121 CHAPTER 5: ELECTRO-FERMENTATION INFLUENCES FERMENTATION PRODUCTS IN ENGINEERED ESCHERICHIA COLI ................................................. 124 PREFACE ................................................................................................................ 125 ABSTRACT .............................................................................................................. 126 INTRODUCTION ..................................................................................................... 126 METHODS ............................................................................................................... 131 RESULTS ................................................................................................................ 136 DISCUSSION .......................................................................................................... 145 FUNDING SOURCES .............................................................................................. 149 REFERENCES ........................................................................................................ 150 CHAPTER 6: CONCLUSION ....................................................................................... 155 OBJECTIVES .......................................................................................................... 156 CONTRIBUTION OF PRESENTED WORK TOWARD DEVELOPMENT OF S. ONEIDENSIS AS A MES HOST .............................................................................. 157 FUTURE DIRECTIONS FOR DEVELOPING S. ONEIDENSIS AS A MES HOST ... 160 CONTRIBUTAION OF PRESENTED WORK TOWARD DEVELOPMENT OF E. COLI AS A MES HOST ..................................................................................................... 162 FUTURE DIRECTIONS FOR DEVELOPING E. COLI AS A MES HOST ................. 163 CONCLUDING REMARKS ...................................................................................... 164 REFERENCES ........................................................................................................ 165 ix LIST OF TABLES Table 1. Strains used in experiments described in Chapter 2. ....................................... 38 Table 2. Plasmids used in experiments described in Chapter 2. ................................... 39 Table 3. Primers used in experiments described in Chapter 2....................................... 40 Table 4. Mutations in formate-tolerant strains. ............................................................... 50 Table 5. Predicted binding affinities of bicarbonate and formate to site 1 and site 2 of WT and mutant proteins encoded by SO_3758. ................................................................... 56 Table 6. DeepGO predictions for the protein encoded by SO_1320. ............................. 59 Table 7. Strains described in Chapter 3. ....................................................................... 76 Table 8. Plasmids described in Chapter 3. .................................................................... 77 Table 9. Primers described in Chapter 3. ...................................................................... 77 Table 10. Carrying capacity, growth rates, and their percent differences from the non- targeting strain of knockdowns cultivated in 200 μl of minimal lactate medium. ............. 83 Table 11. Carrying capacity, growth rates, and percent differences from the uninduced conditions of the rnf knockdown strain in 50 ml minimal lactate medium inoculated to different initial OD600 (OD600 i). ....................................................................................... 84 Table 12. Growth rate and carrying capacity of S. oneidensis strains depicted in Figure 31a. ............................................................................................................................. 108 Table 13. Growth rate and carrying capacity of S. oneidensis growth depicted in Figure 34. ............................................................................................................................... 112 Table 14. Strains and plasmids used for experiments described in Chapter 5............. 131 Table 15. Final concentrations and percent differences of metabolites from MFe408 and MFe444 anodic electro-fermentation depicted in Figure 40. ........................................ 139 Table 16. Metabolite concentrations of MFe408 and MFe444 cultivation with glucose, riboflavin, and Fe2O3 depicted in Figure 43. ................................................................. 142 Table 17. Final concentrations and percent differences of metabolites from MFe408 and MFe444 cathodic electro-fermentation depicted in Figure 45. ..................................... 145 x LIST OF FIGURES Figure 1. The greenhouse gas effect. ............................................................................. 3 Figure 2. Microbial electrosynthesis.. .............................................................................. 6 Figure 3. Extracellular electron transfer in S. oneidensis.. .............................................. 9 Figure 4. Proposed pathways for CO2 fixation in S. oneidensis. ................................... 14 Figure 5. Schematic for using cathodic potentials to drive CO2 assimilation in S. oneidensis via the rGly pathway. ................................................................................... 16 Figure 6. Growth curves of MR-1 and JG29597 in 200 μl minimal medium with 20 mM lactate and varying concentrations of formate................................................................ 43 Figure 7. Length of subculturing each strain required for subculturing with increasing concentrations of formate. ............................................................................................. 44 Figure 8. Growth curves of directed evolution subculturing in increasing formate concentrations of parent strains. .................................................................................... 46 Figure 9. Growth curves of single colonies in triplicate isolated from the evolution experiments compared to their associated parent strain. ............................................... 47 Figure 10. Growth curves of formate-tolerant strains cultivated in 50 ml minimal lactate medium with and 100 mM formate. ................................................................................ 47 Figure 11. Metabolite concentrations of parent and formate-tolerant strains. ................ 48 Figure 12. Chemical structures of formate and bicarbonate. ......................................... 49 Figure 13. Growth curves of yadS CRISPRi knockdown (KD) strains and non-targeting control strain. ................................................................................................................. 52 Figure 14. Growth curves of WT S. oneidensis harboring an IPTG-inducible expression vector containing WT or mutant versions of SO_3758 or SO_1320. .............................. 53 Figure 15. Multiple sequence alignment of cyanobacteria SbtA, and the proteins encoded by WT SO_3758 and SO_3758_A269T. ......................................................... 55 Figure 16. Representations of pLDDT scores from the protein sequence encoded by SO_3758 AlphaFold structural predictions. .................................................................... 56 Figure 17. Ligand dock modeling of formate and bicarbonate to the proteins encoded by SO_3758_WT and SO_3758_A269T. ........................................................................... 58 xi Figure 18. Structure prediction of the protein encoded by SO_1320 by Deep TMHMM ........................................................................................................................ 60 Figure 19. AlphaFold modeling of the proteins encoded by WT and mutant versions of SO_1320. ...................................................................................................................... 62 Figure 20. Rnf activities in the ferredoxin-reducing and ferredoxin-oxidizing directions. ...................................................................................................................... 73 Figure 21. Colony forming units (CFU) of knockdown strains plated into LB with increasing concentrations of knockdown inducer IPTG. ................................................ 81 Figure 22. Growth curves of the induced and uninduced non-targeting, rnf, iscS, and nth knockdown strains in 200 μl of minimal lactate medium. ............................................... 82 Figure 23. Growth curves of WT S. oneidensis and Δnth.. ............................................ 82 Figure 24. Growth curves of the induced and uninduced non-targeting, rnf, and iscS knockdown strains in 50 ml minimal lactate medium. .................................................... 84 Figure 25. a) Growth curves and b) OD-corrected ThT fluorescence of uninduced and induced non-targeting, rnf, and iscS knockdown strains. ............................................... 85 Figure 26. Change in metabolite concentrations on a log2 scale from induced rnf knockdown cultures. ...................................................................................................... 87 Figure 27. Change in metabolite concentrations on a log2 scale from induced iscS knockdown cultures. ...................................................................................................... 88 Figure 28. The MEP pathway for biosynthesis of isoprenoid precursors. ...................... 89 Figure 29. Proposed comprehensive model for Rnf function. ........................................ 92 Figure 30. Growth of wild-type S. oneidensis and S. oneidensis CRISPRi non-targeting strain (MR-1 pJMP2846). ............................................................................................. 107 Figure 31. Growth of wild-type S. oneidensis in various conditions. ............................ 108 Figure 32. MR-1 cultures in 50 ml M5 minimal medium with and without 10 mM IPTG in varying conditions. ....................................................................................................... 110 Figure 33. Comparison of S. oneidensis MR-1 cell size during NAG cultivation with or without IPTG. ............................................................................................................... 111 xii Figure 34. 50 ml a) growth curves and b) metabolite analysis of S. oneidensis MR-1 on 40 mM D,L-lactate with or without 10 mM IPTG............................................................ 112 Figure 35. MR-1 growth curves in 200 µl M5 minimal medium with 20 mM NAG and varying IPTG concentrations. ...................................................................................... 113 Figure 36. S. oneidensis growth curves in 50 ml M5 minimal medium with 20 mM NAG and varying IPTG concentrations. ................................................................................ 114 Figure 37. (a) The chemical structures of IPTG and NAG. Their structures differ in the highlighted regions. (b) Growth curves of S. oneidensis growth on 20 mM glucose with varying concentrations of IPTG.................................................................................... 114 Figure 38. Major fermentation pathways in E. coli. ...................................................... 127 Figure 39. Mtr pathway from S. oneidensis expressed in MFe444. ............................. 130 Figure 40. Current (A) and metabolite concentrations (B) of MFe408 and MFe444 during anodic electro-fermentation. ........................................................................................ 138 Figure 41. OD600 measurements of MFe408 and MFe444 during anodic electro- fermentation depicted in Figure 40. ............................................................................. 139 Figure 42. Current of MFe408 and MFe444 anodic electro-fermentation with different riboflavin concentrations. ............................................................................................. 140 Figure 43. Metabolite concentrations during anaerobic cultivation of MFe408 and MFe444 with 20 mM D-glucose as the electron donor and 50 mM Fe2O3 as the electron acceptor....................................................................................................................... 141 Figure 44. Fe2+ and metabolite concentrations during anaerobic cultivation of MFe408 and MFe444 with 20 mM D-glucose as the electron donor and 50 mM Fe(III)-NTA as the electron acceptor. ........................................................................................................ 143 Figure 45. Current (A) and metabolite concentrations (B) of MFe408 and MFe444 during cathodic electro-fermentation. ..................................................................................... 144 Figure 46. OD600 measurements of MFe408 and MFe444 during cathodic electro- fermentation depicted in Figure 45. ............................................................................. 145 xiii LIST OF ABBREVIATIONS 2xYT 2x yeast extract tryptone media 3P-glycerate 3-phosphoglycerate Ace aceA ackA AckA AdhE ADP agxT AMP apo ATP BES C1 CA caa CBB CCCP CctA pyruvate dehydrogenase gene encoding isocitrate lyase gene encoding acetate kinase acetate kinase bifunctional aldehyde-alcohol dehydrogenase adenosine diphosphate gene encoding serine-pyruvate aminotransferase adenosine monophosphate protein without its prosthetic group adenosine triphosphate bioelectrochemical system one carbon carbonic anhydrase casamino acids Calvin Benson Bassham carbonyl cyanide m-chlorophenyl hydrazone periplasmic tetraheme cytochrome c c-di-GMP bis-(3′–5′)-cyclic-dimeric guanosine monophosphate CDP-ME 4-diphosphocytidyl-2-C-methyl-D-erythritol-2-phosphate CDP-MEP methylerythritol cytidyl diphosphate CFU cMEPP CMP CoA colony forming unit 2-C-methyl-D-erythritol-2,4- cyclodiphosphate cytidine monophosphate coenzyme A xiv CRISPRi clustered regularly interspaced short palindromic repeats interference crp CTP CymA DAHP DAP dCas9 dCTP dGMP cAMP-responsive regulator of catabolite repression cytidine triphosphate cytochrome c-type protein 3-Deoxy-D-arabinoheptulosonate 7-phosphate 2,6-diaminopimelic acid inactivated Cas9 deoxycytidine triphosphate deoxyguanosine monophosphate DMAPP dimethylallyl diphosphate DNA dTTP DUF DXP Dxr Dxs e- Eo’ EET Eno FAD deoxyribonucleic acid deoxythymidine triphosphate domain of unknown function 1-deoxy-D-xylylose-5-phosphate 1-deoxy-D-xylulose 5-phosphate reductoisomerase 1-deoxy-D-xylulose-5-phosphate synthase electron standard reduction potential extracellular electron transfer enolase flavin adenine dinucleotide FccA fumarate reductase flavoprotein Fch FDH Fhs methenyltetrahydrofolate cyclohydrolase formate dehydrogenase formate tetrahydrofolate ligase FocA bidirectional formate transporter xv FolD Fpr 10-formyltetrahydrofolate synthase ferredoxin-NADP reductase FrdABCD fumarate reductase FtfL FumBC Fur G3P GapA GCS GlyA gmpA Gpm GTP H4F HEPES HexR HMBPP HPLC IDI IET IMP IPP formate tetrahydrofolate ligase fumarate hydratase ferric uptake regulation protein glyceraldehyde 3-phosphate glyceraldehyde-3-phosphate dehydrogenase glycine cleavage system serine hydroxymethyltransferase gene encoding phosphoglycerate mutase phosphoglycerate mutase guanosine monophosphate tetrahydrofolate 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid 2-keto-3-deoxy-6-phosphogluconate-responsive transcriptional regulator 4-hydroxy-3-methyl-butenyl-1-diphosphate high performance liquid chromatography Isopentenyl-diphosphate delta-isomerase inward electron transfer inosine monophosphate isopentyl diphosphate IPTG Isopropyl β-D-1-thiogalactopyranoside ISC IspD IspE iron sulfur cluster 2-C-methyl-D-erythritol 4-phosphate cytidylyltransferase 4-diphosphocytidyl-2-C-methyl-D-erythritol kinase xvi IspF IspG IspH kbl LacI LB LC-MS LdhA M5 mC Mdh MEP MES ml mM mmol MSA MtdA mV NAD NADP+ NAG nagR NagR NEB Nif 2-C-methyl-D-erythritol 2,4-cyclodiphosphate synthase 4-hydroxy-3-methylbut-2-en-1-yl diphosphate synthase 4-hydroxy-3-methylbut-2-enyl diphosphate reductase gene encoding 2-amino-3-ketobutyrate coenzyme A ligase lac repressor Lysogeny broth liquid chromatography mass spectrometry lactate dehydrogenase minimal medium millicoulombs malate dehydrogenase methylerythritol phosphate microbial electrosynthesis milliliter millimolar millimole multiple sequence alignment methylene tetrahydromethanopterin dehydrogenase millivolt nicotinamide adenine dinucleotide nicotinamide adenine dinucleotide phosphate N-acetylglucosamine gene encoding transcriptional repressor of N-acetylglucosamine utilization transcriptional repressor of N-acetylglucosamine utilization New England Biolabs nitrogenase xvii NTA NTF2 Nth OD600 OET PCR PDH nitrilotriacetic acid nuclear transport factor 2 endonuclease III optical density at 600 nanometers outward electron transfer polymerase chain reaction pyruvate dehydrogenase pdhR gene encoding pyruvate dehydrogenase PEC PEP pfkA pfkB PflB pgi Pgk pLDDT PMF pntAB Ppc ppsR Prk Prs Pta Pyk Q QH2 Rf periplasmic electron carrier phosphoenol pyruvate gene encoding ATP-dependent 6-phosphofructokinase isozyme 1 gene encoding ATP-dependent 6-phosphofructokinase isozyme 2 pyruvate formate lyase gene encoding glucose-6-phosphate isomerase phosphoglycerate kinase predicted local distance difference test proton motive force genes encoding NAD(P) transhydrogenase phosphoenolpyruvate carboxylase gene encoding phosphoenolpyruvate synthase regulatory protein phosphoribulokinase gene encoding ribose-phosphate pyrophosphokinase phosphate acetyltransferase pyruvate kinase quinone quinol riboflavin xviii rGly RID Rnf RnfB RnfC Rpm rpoC reductive glycine refractive index Rhodobacter nitrogen fixation Rhodobacter nitrogen fixation subunit B Rhodobacter nitrogen fixation subunit C rotations per minute gene encoding DNA-directed RNA polymerase subunit beta RuBisCO ribulose-1,5-bisphosphate carboxylase/oxygenase SbtA SdaA serA sgRNA sodium-dependent bicarbonate transporter serine dehydratase gene encoding serine dehydratase single guide ribonucleic acid Shikimate 3P shikimate 3-phosphate SoxR redox-sensitive transcriptional activator SUF TCA TEA ThT sulfur mobilization tricarboxylic acid terminal electron acceptor thiflavin T TRIC trimeric intracellular cation UDP UTP WT uridine diphosphate uridine triphosphate wild type yadS gene encoding homotrimeric cation channel (TRIC) family protein ybjU zwf μg μl gene encoding threonine aldolase glucose-6-phosphate 1-dehydrogenase microgram microliter xix μM σ micromolar sigma factor xx CHAPTER 1: INTRODUCTION 1 CLIMATE CHANGE Drilling for and burning fossil fuels is one of the greatest contributors to climate change and has led to a 1.1 °C increase in global temperatures since the pre-industrial era.1 Climate change has widespread detrimental effects on the planet, such as changes in global temperatures, ocean acidification, melting glaciers, and extreme weather. Biodiversity loss is a major concern of climate change,2,3 with some arguing that we are heading toward the Earth’s sixth mass extinction.4,5 Plants and animals are not alone – humans are also paying the toll of a changing climate. Wildfires release fine particulate matter which pollutes the air and can cause both long and short-term health issues such as respiratory or cardiovascular complications, eye irritation, and corneal abrasions.6,7 Heat waves can lead to heat stroke, heat exhaustion, cardiovascular complications, and death.8,9 The World Health organization reports that heat-related deaths have increased by 70% in people 65 and older since 2000.10 Disruptions in phenomena such as the timing of flower blooms,11,12 declining bee populations,13,14 and their effect on pollination can lead to difficulties in growing the food required for human and livestock sustenance.15–17 The World Health Organization estimates that 98 million more people experienced food insecurity in 2020 than the average during 1981 – 2010.10 An increase in extreme weather events such as hurricanes, floods, tsunamis, wildfires, and tornadoes have led to humanitarian crises across the globe.18,19 Countries with low emission levels and few means to invest in preventative measures face the worst of this effect, paying the penalties for countries with highest emissions which do not feel the effects as strongly due to greater access to preventative 2 measures and aid.20,21 To prevent the worst effects of climate change, the International Panel on Climate Change recommends limiting warming to below 2 °C.1 The warming of the planet comes from the accumulation of greenhouse gases in the atmosphere. These greenhouse gasses act as a heat trap, keeping energy from the sun in the Earth’s atmosphere.22 Burning of fossil fuels releases greenhouse gases like carbon dioxide (CO2), which is one of the greatest drivers of climate change (Figure 1).1 To limit climate change effects, greenhouse gas emissions must be significantly reduced to prevent their continued accumulation in the atmosphere.1 The challenge with this goal is that CO2 emissions are inevitable. Even a switch to burning renewably derived hydrocarbon fuels would result in the release of CO2, as it is always a product in the combustion of hydrocarbons.23 Therefore, CO2 fixation and recycling are imperative efforts for climate change mitigation.24 Figure 1. The greenhouse gas effect. a) Greenhouse gases help keep the planet warm by trapping heat from the sun. At appropriate levels, the greenhouse gases strike a balance between heat escape into space and heat retention which maintains healthy global temperatures. b) Increased greenhouse gas levels in the atmosphere traps more heat, which raises global temperatures. Created with BioRender.com. 3 If the threat of climate change wasn’t enough to promote development and implementation of greener technology, a 2009 study using a modified Klass model estimated that coal, oil, and natural gas sources will be depleted by 2112, 2040, and 2042, respectively.25 Shifting to new technologies that can produce energy and chemicals from renewable carbon sources must be a priority to keep meeting our energy needs on a planet that’s fit to support life for centuries to come.1,26 RENEWABLE ENERGY Electrical energy is a top contender for replacing fossil fuels. The transition to electrical energy can be seen most prominently with the push towards electric vehicles. For example, the major United States auto manufacturer Chrysler has announced that it plans for 100% of its fleet to be all-electric vehicles by 2028.27 General Motors is expected to do the same by 2035;28 and with the same target date, Ford has pledged for 40 – 50% of its fleet to be electric.29 The Biden Administration also has plans for supporting a multi- million dollar effort to build an electric vehicle charging network across the country.30 Even so, electrical energy cannot fully replace hydrocarbon fuel, especially in the cases of aviation and heat-intensive industrial processes.31,32 Green production of biofuels from renewable carbon sources therefore remains an integral component of the shift away from fossil fuels. Wind, solar, and hydroelectric power have long promised to be renewable, fossil- fuel-free methods of electrical energy generation. While there have been great advancements in these technologies, they are not capable of meeting our electrical energy demands, with over 62% of global electricity still being generated from coal and natural gas.33 Power plants generate electricity as it is needed, with demand driving the real-time 4 production quantities.34 This type of control does not exist with renewable electrical sources because of the intermittent forces that power them.35 If the wind is not blowing or the sun is not shining, there is no electrical output whether there is demand or not. This poses a huge challenge to overcome with the feasibility of replacing fossil-fuel derived electrical energy with renewable sources. Storage of clean electrical energy as chemical bonds in biofuels is a promising method to address two issues. First, the continuing demand for hydrocarbon fuels, and second, the intermittence of electrical energy production from renewable sources. Energy storage is a powerful tool which would allow clean energy to be used even during times when receiving electricity directly from these sources is not possible. MICROBIAL ELECTROSYNTHESIS Microbial electrosynthesis (MES) is a promising technology that has the potential to combine CO2 remediation and clean electricity with the production of biofuels and other useful chemicals (Figure 2).36,37 In this system, carbon from CO2 is fixed by electroactive, autotrophic biocatalysts using clean electricity to power intracellular reduction reactions in a bioelectrochemical system (BES). In this system, electrical input from clean energy sources power the storage of electrons as chemical bonds in the resulting metabolic products. Genetic engineering of these biocatalyst can program biofuels or useful chemicals to be the main metabolic product from CO2 and clean electricity. MES biocatalysts are mainly acetogenic bacteria.38 Acetogens are carbon-fixing microbes, some of which can use electron donors such as hydrogen or a cathode to power the reduction of CO2 via the Wood-Ljungdahl pathway. Acetogens produce acetate as their main metabolic product from acetyl-CoA via substrate level phosphorylation.38,41 As 5 Figure 2. Microbial electrosynthesis. Microbial electrosynthesis converts CO2 and clean electricity into useful chemical products that can be used for replacing hydrocarbons derived from fossil fuels. They could be used in production of biofuels, bioplastics, chemicals, synthetic fibers, and storage of electrical energy. Created with BioRender.com. acetate is a low value and low energy density compound, metabolic engineering would be required to improve product scope and yields for these organisms to be competitive MES hosts.38 Additionally, the mechanism of electron uptake by acetogens is not well understood, and different species may have different mechanisms.41 Due to their limited product scope, little understanding of their electron transfer pathways, and limited genetic engineering tools, acetogens do not provide the power and flexibility required for commercially competitive industrial-scale MES, necessitating improvement to the system.38–41 Some success has been found in supplementing acetogens with mixed bacterial cultures; however, high acetate production remains an issue, and the mixed culture 6 composition is usually unknown, which significantly hampers the possibilities for targeting specific products.42–44 Another issue with mixed cultures is that they are prone to off-target methane production due to increased methanogen populations in the reactors over time.38 More information on current MES biocatalysts can be found on page 38. To overcome challenges with MES, a suitable biocatalyst must be engineered from organisms which have electroactivity, accessible genetic engineering tools, and well understood metabolisms and electron transfer pathways. This dissertation explores two model organisms which have unique strengths for conversion to MES biocatalysts. The first is Shewanella oneidensis, a heterotrophic, facultative anaerobe with the capability of respiring solid, extracellular metals via its well- studied extracellular electron transfer (EET) pathway.45–47 This EET pathway in S. oneidensis makes it capable of interfacing with an electrode.46,47 S. oneidensis’s EET pathway, a good knowledge base of its metabolism, and established genetic engineering tools make this organism an attractive host for use in MES.47–57 The second potential host for improved MES explored in this dissertation is Escherichia coli. As one of the most studied bacterial species, it has a plethora of genetic engineering tools which can be exploited for heterologous expression of CO2 fixation and EET pathways.58–64 E. coli has long been a host for fermentative biosynthesis of biofuels and chemicals.64–70 Transfer of these fermentative biosynthesis systems into an electroautotrophic strain would be a monumental step toward realization of optimizing both the MES host and the possible product scope. 7 SHEWANELLA ONEIDENSIS AS A MES HOST S. oneidensis MR-1 is a dissimilatory metal-reducing bacterium which can respire extracellular heavy metals such as iron, cobalt, and manganese.71–73 This process is conducted by its EET pathways, with the major contributor being the Mtr pathway.51,74,75 Electrons in the inner membrane quinone pool are used for EET. These electrons originate from substrates oxidized by quinone-linked oxidoreductases (e.g. NADH dehydrogenase). The Mtr pathway is initiated by inner membrane protein CymA, which transfers electrons from the quinone pool to the periplasm where they are deposited onto periplasmic electron carriers (PECs).76-79 These PECs shuttle electrons to the other side of the periplasm where the MtrCAB complex transports the electrons across the outer membrane and deposits the electrons on an extracellular terminal electron acceptor (TEA) (Figure 3a).51,78–81 For more information on the function of each component of the Mtr pathway, see pages 39 and 141. Outward electron transfer (OET) is a native capability of S. oneidensis MR-1 which has allowed it to have great respiratory flexibility and gives it the ability to respire solid extracellular TEAs.47,82,83 When cultivated in a BES, an anode can act as the TEA, and the pathway activity generates a current.75,84 This capability has been harnessed for S. oneidensis use as a microbial fuel cell host.46,53,85–87 In these systems, electricity is generated from the cellular oxidation of a substrate.88,89 As a source of clean, renewable electricity, this technology is of great interest, especially when waste can be used as the substrate. For example, S. oneidensis has been used for electricity generation using wastewater in microbial fuel cells.90,91 8 Figure 3. Extracellular electron transfer in S. oneidensis. a) Native outward electron transfer. Electrons are acquired through substrate oxidation by quinone-linked oxidoreductases like NADH dehydrogenase (Nuo) and are passed to the inner membrane quinone pool. CymA transfers electrons from the inner membrane quinone pool to periplasmic electron carriers (PEC). PECs deposit electrons with the outer membrane complex MtrCAB, which exports the electrons to the outer surface of the cell. Electrons are deposited on a terminal electron acceptor (TEA). b) Engineered inward electron transfer. The Mtr pathway is reversed when cultivated under cathodic conditions. The cathode can act as an electron source for the pathway. Electrons are transferred in the reverse order of inward electron transfer, with electrons from NADH being used to catalyze acetoin reduction via a heterologously expressed butanediol dehydrogenase (Bdh). Butanediol levels can act as an indicator of pathway function and efficiency. Reversed activity of proton-pumping Nuo depleted proton motive force (PMF). To replenish PMF, another proton pump must be used. This has been successful with green light-activated proteorhodopsin and native oxidases. 9 Wild-type S. oneidensis is also capable of inward electron transfer (IET). During IET, the electrons follow the same path as the native OET pathway in reverse. This has been demonstrated by applying a negative potential to S. oneidensis biofilm on a graphite electrode and monitoring reduction of exogenously added fumarate.124 The reduction of fumarate was determined to be biotic activity, as shown by the deletion of Mtr pathway components mtrA and cymA decreasing fumarate reduction by 97% and 85%, respectively.124 In this system, electrons derived from EET act as the electron donor for fumarate reduction to succinate by fumarate reductase. Oxygen can also act as a terminal electron acceptor for electrons taken up by IET.48,96 Electrons in the quinone pool derived from EET follow two paths. First, they are used by NADH dehydrogenase to reduce NAD+, a reaction which is coupled with consumption of proton motive force (PMF).96 Second, the electrons are also used by terminal oxidases to generate PMF, complementing the activity of NADH dehydrogenase.48,96 Methods for IET optimization and IET tracking have been engineered in S. oneidensis.48,92–95 Feeding of acetoin and expression of a heterologous butanediol dehydrogenase allows for NADH produced from IET to be monitored as 2,3-butanediol (Figure 3b).93 This serves as proof-of-concept that IET can power intracellular reduction reactions as well as provide a quantifiable readout of the efficiency of the pathway. An important constraint to the IET pathway is the depletion of PMF. In the OET direction, NADH dehydrogenase acts as a proton pump, contributing to PMF. In the IET direction, NADH dehydrogenase uses PMF to power the reduction of NAD+ to NADH (Figure 3).93 If left uncorrected, IET would stop almost immediately because as soon as PMF was depleted, the reaction could no longer proceed. Therefore, another source of PMF is 10 required. This was accomplished in our lab with two methods. The first was the expression of a green light-driven proteorhodopsin.93 This protein localizes to the inner membrane and pumps protons into the periplasm when illuminated with green light. While effective, light penetration decreases with increasing reactor volumes, which would make this technology difficult to scale up. To overcome this limitation, PMF can also be generated with native oxidases.48,96 Since the Mtr pathway is an anaerobic respiration pathway, and because high oxygen levels cause reactive oxygen species formation in BESs, cultivation takes place in anoxic conditions. These anoxic conditions were previously thought to limit the contribution of PMF by oxidases. However, reactor dissolved oxygen concentrations were found to be ~1%, which is enough oxygen for the oxidases to generate PMF.48 Therefore, the use of green light for S. oneidensis IET can be eliminated in microaerobic conditions, a potentially achievable scale-up requirement as complete anaerobic systems are difficult to maintain. The IET pathway in S. oneidensis makes this organism an attractive potential host for MES. The missing component is CO2 fixation. With an accessible genetic toolbox, engineering CO2 fixation could be an achievable feat. There are many natural and synthetic options for engineering CO2 fixation, including carboxylation-first pathways such as the reductive tricarboxylic acid (TCA) cycle and the Calvin Benson Bassham (CBB) cycle, and reduction-first pathways like the reductive glycine (rGly) pathway.97,98 A reduction-first approach has the potential to make CO2 fixation more feasible in S. oneidensis, as reduction-first pathways support greater biomass production and more efficient ATP utilization.98,99 CO2 reduction to formate, where formate is the one-carbon molecule which is assimilated into central metabolism, is an attractive option because 11 engineering CO2 fixation through means such as the reductive TCA cycle or the CBB cycle can be a mountainous task.100 In a reduction-first approach the metabolic engineering of formate assimilation can still be useful if CO2 reduction cannot be achieved in S. oneidensis outright. For example, CO2 could be pre-treated through its electrochemical reduction to formate, and subsequently the formate can act as the substrate for MES. Electrochemical reduction of CO2 is a process which can be conducted at room temperature and pressure, which makes it feasible for inclusion in the MES bioreactor.101 Efforts toward both of these approaches in S. oneidensis are underway. The carboxylation route by RuBisCO and the CBB cycle are being modeled after CO2 efforts in E. coli, which are described in detail on page 20.50,125 Briefly, this approach separates upper and lower glycolysis at the step catalyzed by phosphoglycerate mutase (Gpm) in order to promote CO2 assimilation via heterologously expressed RuBisCO, upper glycolysis, and the pentose phosphate pathway while cellular energy is produced from pyruvate via lower glycolysis and the TCA cycle (Figure 4a).110-112 In S. oneidensis, the first step in separating these components of metabolism was accomplished through the use of CRISPRi and flux balance analysis modeling.125 While Gpm was not essential in E. coli, initial efforts in making gpmA knockouts were unsuccessful. CRISPRi targeting gpmA and flux balance analysis modeling allowed for optimal knockout conditions to be uncovered and used for successful deletion of this conditionally essential gene.125 S. oneidensis ΔgpmA harboring pCBB (a vector which expresses RuBisCO, phosphoribulokinase, and carbonic anhydrase) can grow on a combination of lactate and uridine, which will be utilized for chemostat evolution in a CO2-containing atmosphere.50 Lactate is converted to pyruvate which is used for energy production via the TCA cycle, 12 whereas uridine is used for biomass production in upper glycolysis and the pentose phosphate pathway. By restricting growth through limiting uridine concentrations, pressure will be provided to encourage CO2 assimilation via RuBisCO.50 The reduction-first approach for CO2 assimilation in S. oneidensis is modeled after the rGly pathway engineered in E. coli, which is described in detail on page 21. Briefly, the rGly pathway is a short, linear, energetically favorable pathway which uses CO2-derived formate to generate serine (Figure 4b).100 Conversion of serine to acetyl-CoA by serine dehydratase and pyruvate dehydrogenase can feed the CO2-derived carbon into the TCA cycle. While the rGly pathway generally runs in the oxidative direction for glycine catabolism, all enzymatic steps from formate to serine are reversible.110 The overall ΔrG of this pathway at 1 mM concentrations of formate and CO2 is approximately -6 kJ mol-1, indicating that conversion of formate to serine is dependent upon adequate pathway substrate concentrations.126 In E. coli, cellular growth coupled to reversed action of the rGly pathway was achieved with 30 mM formate and 10% CO2.117-119 This was achieved after streamlining the pathway by deleting all genes whose protein products siphon carbon away from the pathway, expressing missing rGly pathway enzymes, overexpressing limiting native rGly pathway enzymes, and short term adaptive laboratory evolution (see page 21). All the proteins required for a functioning rGly pathway are encoded on the S. oneidensis MR-1 genome and have been detected by transcriptome analysis127,128, indicating this method of CO2 assimilation could be feasible in S. oneidensis. Most of the work toward implementing the rGly pathway in S. oneidensis has been focused on generating the necessary background strain – the glycine auxotroph. Glycine auxotrophy 13 Figure 4. Proposed pathways for CO2 fixation in S. oneidensis. a) Proposed pathway for engineering CO2 fixation via the CBB cycle in S. oneidensis. Pathway substrates are highlighted in black. b) Proposed pathway for engineering formate assimilation via the rGly pathway in S. oneidensis. Arrows highlighted in yellow represent the rGly pathway. Pathway substrate highlighted in black. GCS – glycine cleavage system. 14 is achieved in S. oneidensis through the deletion of genes encodinng threonine aldolase, isocitrate lyase, 2-amino-3-ketobutyrate coenzyme A ligase, serine-pyruvate aminotransferase, and 3-phosphoglycerate dehydrogenase. S. oneidensis becomes more and more resistant to the homologous recombination deletion method after each gene deletion, which makes generating a five-gene knockout strain quite the feat. The glycine auxotroph strain was successfully generated after two years of genetic engineering. This glycine auxotroph cannot use formate to relieve the glycine auxotrophy, indicating that there is not enough reverse activity of the rGly pathway enzymes to support formate assimilation. Therefore, the next necessary step for implementing the rGly pathway in S. oneidensis is to express readily reversible rGly pathway enzymes in a modular format (Figure 4b). The approach taken will be similar to that of the E. coli model, details of which can be found on page 21. S. oneidensis has three quinone-linked formate dehydrogenases and there is some evidence supporting their reversibility under cathodic potentials.102 These promising data indicate that CO2 reduction could occur in vivo and generate the formate required to feed the rGly pathway once it is established in S. oneidensis (Figure 5). As a cathode is already to be used for generating value-added products during MES, it can be used for powering CO2 reduction, NADH generation, and downstream reduction reactions. For this reason, a cathode-driven biotic CO2 reduction method is ideal for pairing with MES technology. Whether CO2 reduction is achieved electrochemically or metabolically, it is clear that formate assimilation would provide S. oneidensis with the ability to utilize CO2 derived compounds as a carbon and electron source for MES. In Chapter 2, progress toward this goal through generation of formate-tolerant S. oneidensis strains is discussed. 15 Figure 5. Schematic for using cathodic potentials to drive CO2 assimilation in S. oneidensis via the rGly pathway. OTHER CONSIDERATIONS FOR S. ONEIDENSIS AS A MES HOST The S. oneidensis respiratory chain is highly branched, with the Mtr pathway being a critical component of anaerobic respiration via EET.74,103 While many aspects of the respiratory chain have been explored, gaps in the literature still remain. For optimal performance of S. oneidensis as an MES host, the electron transfer pathways should be well understood so their influence on EET can be controlled. For example, low potential electrons such as those carried by ferredoxins or flavodoxins could act as yet-unidentified electron donors to the OET pathway. The Rnf complex has not been studied in S. 16 oneidensis. As a respiratory enzyme which connects the ferredoxin and NADH pools104,105, Rnf could be an important source of low-potential electrons that may have influence over EET. Additionally, low potential electrons are needed for processes like carbon and nitrogen fixation. As important reactions to consider for developing an MES-compatible S. oneidensis strain, the generation of low-potential electrons is an important factor to consider. Characterization of the S. oneidensis complex and its broader role across species is discussed in Chapter 3. As discussed above, genetic engineering is required for S. oneidensis to be a suitable MES host. One of the most common methods for testing recombinant proteins for their utility in strain improvement is by overexpressing their genes in the host strain under control of an isopropyl β-D-1-thiogalactopyranoside (IPTG)-inducible promoter.106 An in- depth description of this system can be found on page 114. Since it is such a highly used system, sometimes the influence of IPTG on other cellular components can be overlooked or forgotten.106 For rigorous optimization of MES host strains, it is imperative to understand the effects of not only the heterologous proteins on the system, but also their inducers. In Chapter 4, the influence of IPTG on wild-type S. oneidensis cultivated on N- acetylglucosamine, a common sugar feedstock for this organism, is discussed. ESCHERICHIA COLI AS A MES HOST E. coli is one of the most well-studied microbial production hosts.62,63,107 It is frequently used in fermentative biosynthesis studies with high yields and titers, making it an attractive option for engineering a more robust MES host.108,109 Its success as a production host comes from years of research into its metabolism and gene regulation, giving researchers great insights into its inner workings.63,107,109 Additionally, many robust 17 genetic engineering tools have been developed for this organism, making it easier than ever to control metabolic and regulatory events inside the cell.107 Unlike S. oneidensis, E. coli would need both aspects of MES (CO2 fixation and EET) recombinantly expressed. Efforts into both of these engineering feats have been successfully implemented. CO2 fixation has been difficult to engineer in E. coli. One example of CO2 fixation in recombinant E. coli comes from the expression of RuBisCO110–112, a CO2 fixing enzyme that is responsible for fixing 90% of inorganic carbon globally.113 RuBisCO carboxylates ribulose 1,5-bisphosphate and it is assimilated through the CBB cycle.114 To engineer CO2 fixation via the CBB cycle, upper and lower metabolism in E. coli were separated through the deletion of gpmA and gpmM.110 Expression of RuBisCO, phosphoribulokinase, and carbonic anhydrase allowed for CO2 assimilation via endogenous upper glycolysis and pentose phosphate pathway enzymes in upper metabolism. Pyruvate was used to feed the energy-producing module of lower metabolism comprised of lower glycolysis and the TCA cycle. Additional deletions in pfkA, pfkB, and zwf made the strain dependent on CO2 fixation by RiBisCO for growth on xylose. This strain was continuously evolved on xylose, pyruvate, and CO2. The isolated evolved strain was able to generate all sugars from CO2, a first step in engineering autotrophy.110 Mutations identified to be helpful for CO2 assimilation occurred in genes whose protein products are involved in generating biomass from intermediates of the CBB cycle (prs, serA, and pgi) and in carbon metabolism regulators (crp and ppsR).111 Finally, the energy module fed by pyruvate was replaced by a NAD+-dependent formate dehydrogenase from Pseudomonas sp. 101.112 This allowed all biomass to be generated from CO2 and all energy to be derived from formate, which could eventually be produced from CO2 via a reversible formate dehydrogenase or 18 electrochemical reduction.101,115 The limitation to RuBisCO is its slow rate of activity116, which may not lend itself well to fueling rapid microbial growth. As discussed above, a reduction-first approach to CO2 fixation could be a useful strategy. This method has also been explored in E. coli using the rGly pathway for formate assimilation.117-119 This consisted of expressing reversible rGly pathway enzymes, FtfL, Fch, and MtdA from Methylobacterium extorquens in an E. coli glycine auxotroph.117 This strain was capable of growth using co-substrates formate and glucose. The next iteration of this strain was developed through overexpression of the native glycine cleavage system for greater conversion of formate to glycine, strong expression of glyA and sdaA for pyruvate production from glycine, and a NAD+-dependent formate dehydrogenase from Pseudomonas sp. 101 for energy production from formate.118 After a few rounds of subculturing in high formate concentrations, the strain evolved to grow robustly on formate and CO2, a necessary substrate for the rGly pathway. Two mutations allowed growth on formate alone: one in the promoter of the exogenously expressed formate dehydrogenase which improved energy production and one in the promoter of pntAB which increased availability of the rGly pathway cofactor NADPH.118 Adaptive evolution generated an even more robust strain. The new strain contained mutations in the promoter region of ackA which decreased acetate uptake, pdhR which decreased oxidative flux of the TCA cycle, and rpoC which provided some unknown benefit to the strain.119 These engineering successes show that CO2 fixation by either a carboxylation or a reduction-first method could be a viable means for implementing E. coli as a MES host. EET has also been engineered in E. coli as well, including methods using nanowires or the S. oneidensis Mtr pathway.80,120–122 E. coli MFe444 is a recombinant strain which can 19 participate in EET using the Mtr pathway.122 For an in-depth history of the development of this strain, see page 141. MFe444 participates in both OET and IET using lactate as an electron donor, making the Mtr pathway a viable option for endowing a CO2-fixing strain with electroactivity.58,61,122,123 As E. coli cannot ferment lactate, MFe444 given lactate as an electron donor would be incapable of the fermentative biosynthesis pathways for biofuel or chemical production that make this host so powerful. Therefore, detailed exploration into fermentable carbon and electron sources is necessary to assess its potential utility for MES. In Chapter 5, use of MFe444 as a host for glucose anodic and cathodic electro- fermentation is explored and discussed. 20 REFERENCES 1. Calvin, K. et al. IPCC, 2023: Climate Change 2023: Synthesis Report. Contribution of Working Groups I, II and III to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change [Core Writing Team, H. Lee and J. Romero (Eds.)]. IPCC, Geneva, Switzerland. (2023) doi:10.59327/IPCC/AR6- 9789291691647. 2. Mantyka‐pringle, C. S., Martin, T. G. & Rhodes, J. R. Interactions between climate and habitat loss effects on biodiversity: a systematic review and meta‐analysis. Glob Chang Biol 18, 1239–1252 (2012). 3. Mooney, H. et al. Biodiversity, climate change, and ecosystem services. Curr Opin Environ Sustain 1, 46–54 (2009). 4. 5. 6. 7. 8. 9. Barnosky, A. D. et al. Has the Earth’s sixth mass extinction already arrived? Nature 471, 51–57 (2011). Cowie, R. H., Bouchet, P. & Fontaine, B. The Sixth Mass Extinction: fact, fiction or speculation? Biol Rev 97, 640–663 (2022). Kinney, P. L. Climate Change, Air Quality, and Human Health. Am J Prev Med 35, 459–467 (2008). Xu, R. et al. Wildfires, Global Climate Change, and Human Health. NEJM 383, 2173–2181 (2020). Luber, G. & McGeehin, M. Climate Change and Extreme Heat Events. Am J Prev Med 35, 429–435 (2008). Kenney, W. L., Craighead, D. H. & Alexander, L. M. Heat Waves, Aging, and Human Cardiovascular Health. Med Sci Sports Exerc 46, 1891–1899 (2014). 10. World Health Organization. Climate change. https://www.who.int/news-room/fact- sheets/detail/climate-change-and-health (2023). 11. Borghi, M., Perez de Souza, L., Yoshida, T. & Fernie, A. R. Flowers and climate change: a metabolic perspective. New Phytologist 224, 1425–1441 (2019). 12. Tun, W., Yoon, J., Jeon, J.-S. & An, G. Influence of Climate Change on Flowering Time. J Plant Biol 64, 193–203 (2021). 13. Goulson, D., Lye, G. C. & Darvill, B. Decline and Conservation of Bumble Bees. Annu Rev Entomol 53, 191–208 (2008). 14. Cameron, S. A. et al. Patterns of widespread decline in North American bumble bees. PNAS 108, 662–667 (2011). 21 15. Walters, J., Zavalnitskaya, J., Isaacs, R. & Szendrei, Z. Heat of the moment: extreme heat poses a risk to bee–plant interactions and crop yields. Curr Opin Insect Sci 52, 100927 (2022). 16. Klein, A.-M. et al. Importance of pollinators in changing landscapes for world crops. Proc R Soc B, 303–313 (2007). 17. Decourtye, A., Alaux, C., Le Conte, Y. & Henry, M. Toward the protection of bees and pollination under global change: present and future perspectives in a challenging applied science. Curr Opin Insect Sci 35, 123–131 (2019). 18. Ebi, K. L. et al. Extreme Weather and Climate Change: Population Health and Health System Implications. Annu Rev Public Health 42, 293–315 (2021). 19. Amirkhani, M., Ghaemimood, S., von Schreeb, J., El-Khatib, Z. & Yaya, S. Extreme weather events and death based on temperature and CO2 emission – A global retrospective study in 77 low-, middle- and high-income countries from 1999 to 2018. Prev Med Rep 28, 101846 (2022). 20. Berlemann, M. & Wenzel, D. Hurricanes, economic growth and transmission channels. World Dev 105, 231–247 (2018). 21. Noy, I. The macroeconomic consequences of disasters. J Dev Econ 88, 221–231 (2009). 22. NRDC. Greenhouse Effect 101. https://www.nrdc.org/stories/greenhouse-effect- 101#whatis (2023). 23. Speight, J. G. Combustion of hydrocarbons. in Handbook of Industrial Hydrocarbon Processes 421–463 (Elsevier, 2020). doi:10.1016/B978-0-12-809923-0.00010-2. 24. Jeffry, L. et al. Greenhouse gases utilization: A review. Fuel 301, 121017 (2021). 25. Shafiee, S., & Topal, E. When will fossil fuel reserves be diminished? Energy Policy, 37(1), 181–189 (2009). 26. Moore, B. et al. Transformations for climate change mitigation: A systematic review of terminology, concepts, and characteristics. WIREs Climate Change 12, (2021). 27. Wayland, M. Chrysler kicks off plans to go all-electric by 2028 with debut of Airflow EV concept car. CNBC (2022). 28. Boudette, N. E. & Davenport, C. G.M. Will Sell Only Zero-Emission Vehicles by 2035. The New York Times (2021). 29. Ford Newsroom. Ford to lead America’s shift to electric vehicles with new mega campus in Tennessee and twin battery plants in Kentucky; $11.4B investment to lineup of advanced EVS. create jobs and power new 11,000 22 https://media.ford.com/content/fordmedia/fna/us/en/news/2021/09/27/ford-to-lead- americas-shift-to-electric-vehicles.html (2021). 30. U.S. Department of Transportation. Biden-Harris Administration Announces $623 Million in Grants to Continue Building Out Electric Vehicle Charging Network. (2024). 31. Sarwer, A. et al. Suitability of Biofuels Production on Commercial Scale from Various Feedstocks: A Critical Review. ChemBioEng Reviews 9, 423–441 (2022). 32. Thiel, G. P. & Stark, A. K. To decarbonize industry, we must decarbonize heat. Joule 5, 531–550 (2021). 33. Leonard, M. D., Michaelides, E. E. & Michaelides, D. N. Energy storage needs for the substitution of fossil fuel power plants with renewables. Renew Energy 145, 951–962 (2020). 34. United States Environmental Protection Agency. U.S. Electricity Grid & Markets. https://www.epa.gov/green-power-markets/us-electricity-grid-markets (2023). 35. Basit, M. A., Dilshad, S., Badar, R. & Sami ur Rehman, S. M. Limitations, challenges, and solution approaches in grid‐connected renewable energy systems. Int J Energy Res 44, 4132–4162 (2020). 36. Dessì, P. et al. Microbial electrosynthesis: Towards sustainable biorefineries for production of green chemicals from CO2 emissions. Biotechnol Adv 46, 107675 (2021). 37. Jourdin, L. & Burdyny, T. Microbial Electrosynthesis: Where Do We Go from Here? Trends Biotechnol 39, 359–369 (2021). 38. Prévoteau, A., Carvajal-Arroyo, J. M., Ganigué, R. & Rabaey, K. Microbial electrosynthesis from CO2: forever a promise? Curr Opin Biotechnol 62, 48–57 (2020). 39. Katsyv, A. & Müller, V. Overcoming Energetic Barriers in Acetogenic C1 Conversion. Front Bioeng Biotechnol 8, (2020). 40. Philips, J. Extracellular Electron Uptake by Acetogenic Bacteria: Does H2 Consumption Favor the H2 Evolution Reaction on a Cathode or Metallic Iron? Front Microbiol 10, (2020). 41. Rosenbaum, M. A. & Henrich, A. W. Engineering microbial electrocatalysis for chemical and fuel production. Curr Opin Biotechnol 29, 93–98 (2014). 42. Ganigué, R., Puig, S., Batlle-Vilanova, P., Balaguer, M. D. & Colprim, J. Microbial electrosynthesis of butyrate from carbon dioxide. Chem Comm 51, 3235–3238 (2015). 23 43. Raes, S. M. T., Jourdin, L., Buisman, C. J. N. & Strik, D. P. B. T. B. Continuous Long‐Term Bioelectrochemical Chain Elongation to Butyrate. ChemElectroChem 4, 386–395 (2017). 44. Arends, J. B. A., Patil, S. A., Roume, H. & Rabaey, K. Continuous long-term electricity-driven bioproduction of carboxylates and isopropanol from CO2 with a mixed microbial community. J CO2 Util 20, 141–149 (2017). 45. Fredrickson, J. K. et al. Towards environmental systems biology of Shewanella. Nat Rev Microbiol 6, 592–603 (2008). 46. Ikeda, S. et al. Shewanella oneidensis MR-1 as a bacterial platform for electro- biotechnology. Essays Biochem 65, 355–364 (2021). 47. Beblawy, S. et al. Extracellular reduction of solid electron acceptors by Shewanella oneidensis. Mol Microbiol 109, 571–583 (2018). 48. Ford, K. C. & TerAvest, M. A. The electron transport chain of Shewanella oneidensis MR-1 can operate bidirectionally to enable microbial electrosynthesis. Appl Environ Microbiol 90, (2024). 49. Brutinel, E. D. & Gralnick, J. A. Shuttling happens: soluble flavin mediators of extracellular electron transfer in Shewanella. Appl Microbiol Biotechnol 93, 41–48 (2012). 50. Ford, K. C. Microbial Electrosynthesis in Shewanella oneidensis MR-1. (Michigan State University, East Lansing, 2023). 51. Breuer, M., Rosso, K. M., Blumberger, J. & Butt, J. N. Multi-haem cytochromes in Shewanella oneidensis MR-1: structures, functions and opportunities. J R Soc Interface 12, 20141117 (2015). 52. Serres, M. H. & Riley, M. Genomic Analysis of Carbon Source Metabolism of Shewanella oneidensis MR-1: Predictions versus Experiments. J Bacteriol 188, 4601 (2006). 53. Kouzuma, A., Kasai, T., Hirose, A. & Watanabe, K. Catabolic and regulatory systems in Shewanella oneidensis MR-1 involved in electricity generation in microbial fuel cells. Front Microbiol 6, (2015). 54. Flynn, C. M., Hunt, K. A., Gralnick, J. A. & Srienc, F. Construction and elementary mode analysis of a metabolic model for Shewanella oneidensis MR-1. Biosystems 107, 120–128 (2012). 55. Pinchuk, G. E. et al. Pyruvate and Lactate Metabolism by Shewanella oneidensis MR-1 under Fermentation, Oxygen Limitation, and Fumarate Respiration Conditions. Appl Environ Microbiol 77, 8234–8240 (2011). 24 56. McLean, J. S. et al. Investigations of structure and metabolism within Shewanella oneidensis MR-1 biofilms. J Microbiol Methods 74, 47–56 (2008). 57. Tang, Y. J., Meadows, A. L., Kirby, J. & Keasling, J. D. Anaerobic Central Metabolic Pathways in Shewanella oneidensis MR-1 Reinterpreted in the Light of Isotopic Metabolite Labeling. J Bacteriol 189, 894–901 (2007). 58. TerAvest, M. A., Zajdel, T. J. & Ajo‐Franklin, C. M. The Mtr Pathway of Shewanella oneidensis MR‐1 Couples Substrate Utilization to Current Production in Escherichia coli. ChemElectroChem 1, 1874–1879 (2014). 59. Goldbeck, C. P. et al. Tuning Promoter Strengths for Improved Synthesis and Function of Electron Conduits in Escherichia coli. ACS Synth Biol 2, 150–159 (2013). 60. Jensen, H. M., TerAvest, M. A., Kokish, M. G. & Ajo-Franklin, C. M. CymA and Exogenous Flavins Improve Extracellular Electron Transfer and Couple It to Cell Growth in Mtr-Expressing Escherichia coli. ACS Synth Biol 5, 679–688 (2016). 61. Lienemann, M. et al. Towards patterned bioelectronics: facilitated immobilization of exoelectrogenic Escherichia coli with heterologous pili. Microb Biotechnol 11, 1184– 1194 (2018). 62. Pontrelli, S. et al. Escherichia coli as a host for metabolic engineering. Metab Eng 50, 16–46 (2018). 63. Hoesl, M. G. & Budisa, N. Recent advances in genetic code engineering in Escherichia coli. Curr Opin Biotechnol 23, 751–757 (2012). 64. Liu, T. & Khosla, C. Genetic Engineering of Escherichia coli for Biofuel Production. Annu Rev Genet 44, 53–69 (2010). 65. Yu, C., Cao, Y., Zou, H. & Xian, M. Metabolic engineering of Escherichia coli for biotechnological production of high-value organic acids and alcohols. Appl Microbiol Biotechnol 89, 573–583 (2011). 66. Chen, X. et al. Metabolic engineering of Escherichia coli: A sustainable industrial platform for bio-based chemical production. Biotechnol Adv 31, 1200–1223 (2013). 67. Wen, M., Bond-Watts, B. B. & Chang, M. C. Production of advanced biofuels in engineered E. coli. Curr Opin Chem Biol 17, 472–479 (2013). 68. Huffer, S., Roche, C. M., Blanch, H. W. & Clark, D. S. Escherichia coli for biofuel production: bridging the gap from promise to practice. Trends Biotechnol 30, 538– 545 (2012). 25 69. Clomburg, J. M. & Gonzalez, R. Biofuel production in Escherichia coli: the role of metabolic engineering and synthetic biology. Appl Microbiol Biotechnol 86, 419–434 (2010). 70. Koppolu, V. & Vasigala, V. K. Role of Escherichia coli in Biofuel Production. Microbiol Insights 9, MBI.S10878 (2016). 71. Bücking, C., Schicklberger, M. & Gescher, J. The Biochemistry of Dissimilatory Ferric Iron and Manganese Reduction in Shewanella oneidensis. in Microbial Metal Respiration 49–82 (Springer Berlin Heidelberg, Berlin, Heidelberg, 2013). doi:10.1007/978-3-642-32867-1_3. 72. Hau, H. H., Gilbert, A., Coursolle, D. & Gralnick, J. A. Mechanism and Consequences of Anaerobic Respiration of Cobalt by Shewanella oneidensis Strain MR-1. Appl Environ Microbiol 74, 6880–6886 (2008). 73. Cooper, R. E., Goff, J. L., Reed, B. C., Sekar, R. & Dichristina, T. J. Breathing Iron: Molecular Mechanism of Microbial Iron Reduction by Shewanella oneidensis. in Manual of Environmental Microbiology 5.2.1-1-5.2.1-13 (ASM Press, Washington, DC, USA, 2015). doi:10.1128/9781555818821.ch5.2.1. 74. Coursolle, D. & Gralnick, J. A. Modularity of the Mtr respiratory pathway of Shewanella oneidensis strain MR‐1. Mol Microbiol 77, 995–1008 (2010). 75. Coursolle, D., Baron, D. B., Bond, D. R. & Gralnick, J. A. The Mtr Respiratory Pathway Is Essential for Reducing Flavins and Electrodes in Shewanella oneidensis. J Bacteriol 192, 467–474 (2010). 76. Marritt, S. J. et al. The roles of CymA in support of the respiratory flexibility of Shewanella oneidensis MR-1. Biochem Soc Trans 40, 1217–1221 (2012). 77. Marritt, S. J. et al. A functional description of CymA, an electron-transfer hub supporting anaerobic respiratory flexibility in Shewanella. Biochemical Journal 444, 465–474 (2012). 78. Sturm, G. et al. A dynamic periplasmic electron transfer network enables respiratory flexibility beyond a thermodynamic regulatory regime. ISME J 9, 1802–1811 (2015). 79. Fonseca, B. M. et al. Mind the gap: cytochrome interactions reveal electron pathways across the periplasm of Shewanella oneidensis MR-1. Biochemical Journal 449, 101–108 (2013). 80. Schuetz, B., Schicklberger, M., Kuermann, J., Spormann, A. M. & Gescher, J. Periplasmic Electron Transfer via the c -Type Cytochromes MtrA and FccA of Shewanella oneidensis MR-1. Appl Environ Microbiol 75, 7789–7796 (2009). 26 81. Firer-Sherwood, M., Pulcu, G. S. & Elliott, S. J. Electrochemical interrogations of the Mtr cytochromes from Shewanella: opening a potential window. JBIC Journal of Biological Inorganic Chemistry 13, 849–854 (2008). 82. Beblawy, S. et al. Extracellular reduction of solid electron acceptors by Shewanella oneidensis. Mol Microbiol 109, 571–583 (2018). 83. Hernandez, M. E. & Newman, D. K. Extracellular electron transfer. Cellular and Molecular Life Sciences 58, 1562–1571 (2001). 84. Roy, J. N. et al. Applied Electrode Potential Leads to Shewanella oneidensis MR-1 Biofilms Engaged in Direct Electron Transfer. J Electrochem Soc 160, H866–H871 (2013). 85. Kumar, R., Singh, L., Wahid, Z. A. & Din, M. F. Md. Exoelectrogens in microbial fuel cells toward bioelectricity generation: a review. Int J Energy Res 39, 1048–1067 (2015). 86. Leung, D. H. L. et al. Engineering S. oneidensis for Performance Improvement of Microbial Fuel Cell—a Mini Review. Appl Biochem Biotechnol 193, 1170–1186 (2021). 87. Zou, L., Huang, Y. hong, Long, Z. er & Qiao, Y. On-going applications of Shewanella species in microbial electrochemical system for bioenergy, bioremediation and biosensing. World J Microbiol Biotechnol 35, 1–9 (2019). 88. Liu, T. et al. Enhanced Shewanella biofilm promotes bioelectricity generation. Biotechnol Bioeng 112, 2051–2059 (2015). 89. Wu, D. et al. Ferric iron enhances electricity generation by Shewanella oneidensis MR-1 in MFCs. Bioresour Technol 135, 630–634 (2013). 90. Dai, H. N. et al. Power generation of Shewanella oneidensis MR-1 microbial fuel cells in bamboo fermentation effluent. Int J Hydrogen Energy 46, 16612–16621 (2021). 91. Nimje, V. R. et al. Comparative bioelectricity production from various wastewaters in microbial fuel cells using mixed cultures and a pure strain of Shewanella oneidensis. Bioresour Technol 104, 315–323 (2012). 92. Yang, Y. et al. Enhancing Bidirectional Electron Transfer of Shewanella oneidensis by a Synthetic Flavin Pathway. ACS Synth Biol 4, 815–823 (2015). 93. Tefft, N. M. & Teravest, M. A. Reversing an Extracellular Electron Transfer Pathway for Electrode-Driven Acetoin Reduction. ACS Synth Biol 8, 1590–1600 (2019). 94. Tefft, N. M., Ford, K. & TerAvest, M. A. NADHdehydrogenases drive inward electron transfer in Shewanella oneidensis MR‐1. Microb Biotechnol 16, 560–568 (2023). 27 95. Ross, D. E., Flynn, J. M., Baron, D. B., Gralnick, J. A. & Bond, D. R. Towards Electrosynthesis in Shewanella: Energetics of Reversing the Mtr Pathway for Reductive Metabolism. PLoS One 6, e16649 (2011). 96. Rowe, A. R. et al. Tracking Electron Uptake from a Cathode into Shewanella Cells: Implications for Energy Acquisition from Solid-Substrate Electron Donors. mBio 9, (2018). 97. François, J. M., Lachaux, C. & Morin, N. Synthetic Biology Applied to Carbon Conservative and Carbon Dioxide Recycling Pathways. Front Bioeng Biotechnol 7, (2020). 98. Cotton, C. A., Edlich-Muth, C. & Bar-Even, A. Reinforcing carbon fixation: CO2 reduction replacing and supporting carboxylation. Curr Opin Biotechnol 49, 49–56 (2018). 99. Bar-Even, A., Noor, E. & Milo, R. A survey of carbon fixation pathways through a quantitative lens. J Exp Bot 63, 2325–2342 (2012). 100. Bar-Even, A., Noor, E., Lewis, N. E. & Milo, R. Design and analysis of synthetic carbon fixation pathways. PNAS 107, 8889–8894 (2010). 101. Philips, M. F., Gruter, G.-J. M., Koper, M. T. M. & Schouten, K. J. P. Optimizing the Electrochemical Reduction of CO2 to Formate: A State-of-the-Art Analysis. ACS Sustain Chem Eng 8, 15430–15444 (2020). 102. Davies, J. Characterisation of the reversible formate dehydrogenases of Shewanella. (University of East Anglia, 2017). 103. Duhl, K. Disentangling the Branched Respiratory Chain of Shewanella oneidensis MR-1. (Michigan State Univeristy, East Lansing, 2022). 104. Biegel, E., Schmidt, S., González, J. M. & Müller, V. Biochemistry, evolution and physiological function of the Rnf complex, a novel ion-motive electron transport complex in prokaryotes. CMLS 68, 613–634 (2011). 105. Biegel, E., Schmidt, S., González, J. M. & Müller, V. Biochemistry, evolution and physiological function of the Rnf complex, a novel ion-motive electron transport complex in prokaryotes. CMLS 68, 613–634 (2011). 106. Gruenberg, M. C. & TerAvest, M. A. A common inducer molecule enhances sugar utilization by Shewanella oneidensis MR-1. J Ind Microbiol Biotechnol 50, (2023). 107. Idalia, V.-M. N. & Bernardo, F. Escherichia coli as a Model Organism and Its Application in Biotechnology. in Escherichia coli - Recent Advances on Physiology, Pathogenesis and Biotechnological Applications (ed. Samie, A.) 253–274 (IntechOpen, Rijeka, 2017). doi:10.5772/67306. 28 108. Huang, T. & Ma, Y. Advances in biosynthesis of higher alcohols in Escherichia coli. World J Microbiol Biotechnol 39, 125 (2023). 109. Yang, D., Park, S. Y., Park, Y. S., Eun, H. & Lee, S. Y. Metabolic Engineering of Escherichia coli for Natural Product Biosynthesis. Trends Biotechnol 38, 745–765 (2020). 110. Antonovsky, N. et al. Sugar Synthesis from CO2 in Escherichia coli. Cell 166, 115– 125 (2016). 111. Herz, E. et al. The genetic basis for the adaptation of E. coli to sugar synthesis from CO2. Nat Commun 8, 1705 (2017). 112. Gleizer, S. et al. Conversion of Escherichia coli to Generate All Biomass Carbon from CO2. Cell 179, 1255-1263.e12 (2019). 113. Berg, I. A. Ecological Aspects of the Distribution of Different Autotrophic CO2 Fixation Pathways. Appl Environ Microbiol 77, 1925–1936 (2011). 114. Federici, F., Orsi, E. & Nikel, P. I. From Rags to Riches: Exploiting the Calvin‐ Benson‐Bassham Cycle for Biomanufacturing. ChemCatChem 15, (2023). 115. Calzadiaz-Ramirez, L. & Meyer, A. S. Formate dehydrogenases for CO2 utilization. Curr Opin Biotechnol 73, 95–100 (2022). 116. Prywes, N., Phillips, N. R., Tuck, O. T., Valentin-Alvarado, L. E. & Savage, D. F. Rubisco Function, Evolution, and Engineering. Annu Rev Biochem 92, 385–410 (2023). 117. Yishai, O., Goldbach, L., Tenenboim, H., Lindner, S. N. & Bar-Even, A. Engineered Assimilation of Exogenous and Endogenous Formate in Escherichia coli. ACS Synth Biol 6, 1722–1731 (2017). 118. Kim, S. et al. Growth of E. coli on formate and methanol via the reductive glycine pathway. Nat Chem Biol 16, 538–545 (2020). 119. Kim, S. et al. Optimizing E. coli as a formatotrophic platform for bioproduction via the reductive glycine pathway. Front Bioeng Biotechnol 11, (2023). 120. Lee, Y., Lee, J. & Kim, S. Direct extracellular electron transfer from Escherichia coli through modified carbon nanoparticles. Electrochim Acta 437, 141497 (2023). 121. Gescher, J. S., Cordova, C. D. & Spormann, A. M. Dissimilatory iron reduction in Escherichia coli: identification of CymA of Shewanella oneidensis and NapC of E. coli as ferric reductases. Mol Microbiol 68, 706–719 (2008). 29 122. Jensen, H. M., TerAvest, M. A., Kokish, M. G. & Ajo-Franklin, C. M. CymA and Exogenous Flavins Improve Extracellular Electron Transfer and Couple It to Cell Growth in Mtr-Expressing Escherichia coli. ACS Synth Biol 5, 679–688 (2016). 123. Baruch, M., Tejedor-Sanz, S., Su, L. & Ajo-Franklin, C. M. Electronic control of redox reactions inside Escherichia coli using a genetic module. PLoS One 16, e0258380 (2021). 124. Ross, D. E. et al. Towards Electrosynthesis in Shewanella: Energetics of Reversing the Mtr Pathway for Reductive Metabolism. PLoS ONE, 6(2), e16649 (2011). 125. Ford, K. C., Kaste, J. A. M., Shachar-Hill, Y., & Teravest, M. A. Flux-Balance Analysis and Mobile CRISPRi-Guided Deletion of a Conditionally Essential Gene in Shewanella oneidensis MR-1. ACS Synthetic Biology, 11(10), 3405–3413 (2022). 126. Yishai O, Bouzon M, Döring V, Bar-Even A In Vivo Assimilation of One-Carbon via a Synthetic Reductive Glycine Pathway in Escherichia coli. ACS synthetic biology, 7(9):2023–2028 (2018). 127. Beliaev, A. S. et al. Global Transcriptome Analysis of Shewanella oneidensis MR-1 Exposed to Different Terminal Electron Acceptors. J Bacteriol 187, 7138–7145 (2005). 128. Gao H, Wang Y, Liu X, Yan T, Wu L, Alm E, Arkin A, Thompson DK, Zhou J Global transcriptome analysis of the heat shock response of Shewanella oneidensis. J bacteriol, 186(22):7796–7803 (2004). 30 CHAPTER 2: A SMALL NUMBER OF POINT MUTATIONS CONFER FORMATE TOLERANCE IN SHEWANELLA ONEIDENSIS 31 PREFACE The overarching goal for the Shewanella components of my doctoral work was to create a formatotrophic strain which would be able to use CO2-derived formate for MES. It is not an easy feat to engineer autotrophy in a heterotroph, evidenced by the lack of successful examples in the literature. Since I had found a source describing the reversible activity of S. oneidensis formate dehydrogenases, I decided to start the engineering at developing formatotrophy, which would allow me to use a soluble source of carbon, making experimental set ups a bit easier. I took two approaches for developing formatotrophy in S. oneidensis. One was an engineering approach in which I would use the E. coli model using rGly pathway for growth on formate for inspiration. I described this approach in the introductory chapter. As many deletions are required to generate the background strain for developing the rGly pathway, almost my entire PhD was complete before I had just the background strain. This is because each subsequent knockout in S. oneidensis using our homologous recombination method becomes more and more difficult and some resistance to the sucrose counterselection is gained while generating each deletion. Needless to say, generating the background strain was a lot of work. As an alternative method, I decided to also attempt to evolve a formatotrophic phenotype. I did this in two different atmospheres. The first was in CO2-enriched air, which was comprised of 10% CO2 and 21% O2 balanced with N2. The strains were continuously subcultured in minimal medium containing lactate and increasing concentrations of formate. The goal was to improve formate tolerance and then start removing lactate. The elevated CO2 levels were to promote reversal of the rGly pathway, as the glycine cleavage 32 system operating in the reductive direction requires CO2 input. This evolution was extremely slow. I saved samples throughout the culturing period of one year (which was actually multiple years because I restarted this evolution twice), and that was imperative because the strains would often die off. This evolution experiment ended toward the end of my PhD with a few strains which were more tolerant to formate. Interestingly, they grew much better in regular air than in CO2-enriched air. Therefore, out of curiosity, this same evolution was performed in regular air. To my great surprise, eight different lines of S. oneidensis were tolerant to 100 mM formate within 30 days. This chapter describes the evolution which took place in a standard air atmosphere and the characterization of the resulting strains. ABSTRACT Microbial electrosynthesis (MES) is a sustainable approach to chemical production from CO2 and clean electricity. However, limitations in the electron transfer efficiency and gaps in understanding of electron transfer pathways in MES systems prevents full realization of this technology. Shewanella oneidensis is a potential MES biocatalyst because it has a well-studied, efficient transmembrane electron transfer pathway. A key first step in MES in this organism could be CO2 reduction to formate. However, wild-type S. oneidensis does not tolerate high levels of formate. In this work, we created and characterized formate-tolerant strains of S. oneidensis for further engineering and future use in MES systems through adaptive laboratory evolution. Separate point mutations in genes encoding a sodium-dependent bicarbonate transporter and a DUF2721-contianing protein separately confer formate tolerance in S. oneidensis. The mutations were further evaluated to understand their role in improving formate tolerance. 33 INTRODUCTION Microbial electrosynthesis (MES) from CO2 is a promising technology that utilizes clean electricity and water for the green production of biofuels, bioplastics, and platform chemicals.1–3 MES is mediated by electrochemically active organisms, which can utilize reducing power from an electrode and carbon from CO2 to generate target molecules through native or engineered metabolic reactions. Three different avenues of MES have been explored to date, each with their own limitations. The first method employs pure cultures of acetogens. At first glance, acetogens are an attractive MES catalyst, because they are autotrophic and some are electrochemically active; however, they suffer from a limited product scope, with acetate being the most abundant product.4–6 Expansion of the product scope is hindered by the small synthetic biology toolbox for acetogens and their slow rate of electron uptake.6,7 Another limitation of acetogens is the narrow understanding of their EET mechanism, which makes it difficult to engineer improvements. At least five mechanisms have been proposed for EET in acetogens, but experimental evidence supporting any of these mechanisms is lacking, and it is not known whether a single EET mechanism is shared by all electrochemically active acetogens.8 Some of these issues are addressed by the second major approach to MES, replacing the acetogen biocatalyst with mixed cultures. Mixed cultures have more success in producing a wider variety of products, although acetate production remains high.9–11 However, utilizing mixed cultures for MES is plagued by off-target methane production due to rising methanogen populations in the reactors over time.5 A third approach uses model organisms in conjunction with exogenous electron carriers for MES. Model organisms like E. coli are more versatile because of its well understood metabolism and abundance of 34 synthetic biology techniques.12 However, scaling up MES reactions which require exogenous electron carriers is likely not feasible due to their cost. To make MES commercially competitive, a better microbial catalyst with well-understood EET mechanisms, metabolism, and synthetic biology tools is needed. One promising candidate is the electroactive bacterium Shewanella oneidensis. The discovery of the inward electron transfer pathway in this organism has shown that, if coupled to a C1 assimilation pathway, S. oneidensis would likely serve as an improved MES biocatalyst.13-15 S. oneidensis is a heterotrophic facultative anaerobe with the ability to utilize a wide variety of terminal electron acceptors (TEAs) including electrodes.16,17 Extracellular electron transfer is made possible by the Mtr pathway which includes the Mtr complex (MtrCAB), CctA, FccA, and CymA.18 Quinone-linked oxidoreductases such as NADH dehydrogenase, formate dehydrogenase, and lactate dehydrogenase pass electrons obtained from their respective oxidation reactions to inner membrane quinones.19–21 CymA, an inner membrane tetraheme cytochrome c, receives electrons from the quinone pool and subsequently passes them to periplasmic c-type cytochromes FccA and CctA to traverse the periplasm.22–24 On the other side of the periplasm, the cytochromes transfer the electrons to the outer membrane Mtr complex (MtrCAB) which transports the electrons to the outer surface of the cell.18 The Mtr complex is composed of a porin (MtrB) which spans the outer membrane and encases two cytochromes c, one that interfaces with the periplasm (MtrA) and one that interfaces with the extracellular space (MtrC).25,26 Once outside the cell, electrons can either directly reduce extracellular TEAs or be shuttled to TEAs via extracellular flavins.16,17 Its well-studied, reversible Mtr pathway makes S. oneidensis a promising host for 35 MES as it can efficiently exchange electrons with an electrode; however, this organism cannot fix CO2. S. oneidensis requires engineered CO2 fixation for its full potential as an effective MES biocatalyst to be realized. There are multiple potential methods for CO2 fixation, with two possibilities for initial conversions: carboxylation, such as in the Calvin cycle and the reductive tricarboxylic acid cycle, and reduction, such as in the Wood- Ljungdahl pathway and the reductive glycine pathway.27–29 Pathways which first reduce CO2 to formate have been found to support higher biomass yields due to utilizing ATP more efficiently.27,29 Because of this, engineering CO2-fixation in S. oneidensis via a reduction-first pathway would be a promising strategy for generating an MES-compatible host. CO2 can be reduced to formate via reversible formate dehydrogenases30,31 and formate can be assimilated through a host of methods such as the reductive glycine pathway, the serine cycle, and reversed activity of pyruvate-formate lyase.32 Additionally, formate-assimilating S. oneidensis could also be utilized for MES when an electrode with adsorbed formate dehydrogenase is supplied, as such electrodes have been shown to freely interconvert CO2 and formate with low overpotentials.31,33 S. oneidensis is not naturally tolerant of very high formate concentrations, necessitating greater tolerance to be acquired if a CO2-reducing electrode is used or a reduction-first CO2 fixation pathway is expressed. In this work, we utilized adaptive laboratory evolution to create formate- tolerant strains of S. oneidensis as a first step toward developing an MES-compatible host. We found two genes that are responsible for increasing formate tolerance in this organism and explore their application toward a future CO2 fixing strain of S. oneidensis. 36 METHODS Strains and Plasmids Strains and plasmids used in this study are listed in Table 1 and Table 2, respectively. Plasmids for mutant gene overexpression were prepared by cloning the wild- type (WT) or mutant gene from WT or mutant genomic DNA via PCR with primers (Table 3) that added regions homologous to the backbone vector pRL81434 at the NdeI and HindIII restriction sites. WT copies of SO_3758 and SO_1320 were amplified from WT S. oneidensis, SO_3758_A269T was amplified from MGC002, SO_1320_V106I was amplified from MGC007, SO_1320_V106F was amplified from MGC003, and SO_1320_S52I was amplified from MGC008. Amplified genes were purified via the Wizard PCR clean-up kit (Promega). The pRL814 backbone was digested with NdeI and HindIII- HF (New England Biolabs (NEB)). The digested backbone and amplified genes were ligated using NEBuilder HiFi DNA Assembly Master Mix (NEB) and transformed into chemically competent E. coli WM3064. Plasmids were transferred to WT S. oneidensis via conjugal transfer using E. coli WM3064 as described previously.25 CRISPRi knockdown vectors were prepared from the backbone pJMP2846 as described previously, with sgRNAs targeting yadS.35,36 The CRISPRi vectors are kanamycin resistant and encode dCas9 and the sgRNA. The assembled CRISPRi vectors were transformed into chemically competent E. coli WM3064, a 2,6-diaminopimelic acid (DAP) auxotroph. The CRISPRi vectors were transferred to WT S. oneidensis via Tn7-based triparental conjugation using donor strains E. coli WM3064 with the CRISPRi vector and transposase-containing E. coli WM3064 pJMP2644 as described previously.35 After the mating period, counter-selection was performed on lysogeny broth (LB) agar plates supplemented with 50 μg ml-1 of 37 kanamycin and without DAP. Single S. oneidensis colonies were screened for the correct sgRNAs by PCR using the sgRNA_check and related sgRNA reverse primers (Table 3). Table 1. Strains used in experiments described in Chapter 2. Strain Parent strains Description S. oneidensis MR-1 Wild-type JG2957 JG2955 S. oneidensis with deletion of SO_0101- SO_0103, SO_4509-SO_4511, and SO_4513- SO_4515 S. oneidensis with deletion of SO_4509- SO_4511 and SO_4513-SO_4515 Reference Lab stock 21 21 E. coli WM3064 Conjugal transfer donor, DAP auxotroph Lab stock Formate-tolerant strains MGC001 MGC002 MGC003 MGC004 MGC005 MGC006 MGC007 MGC008 Formate-tolerant strain evolved from S. oneidensis MR-1. (Line A Colony 3) Formate-tolerant strain evolved from S. oneidensis MR-1. (Line B Colony 1) Formate-tolerant strain evolved from S. oneidensis MR-1. (Line C Colony 3) Formate-tolerant strain evolved from S. oneidensis MR-1. (Line D Colony 1) Formate-tolerant strain evolved from S. oneidensis JG2957. (Line A Colony 2) Formate-tolerant strain evolved from S. oneidensis JG2957. (Line B Colony 1) Formate-tolerant strain evolved from S. oneidensis JG2957. (Line C Colony 2) Formate-tolerant strain evolved from S. oneidensis JG2955. (Colony 2) 38 This study This study This study This study This study This study This study This study Table 2. Plasmids used in experiments described in Chapter 2. Strain pRL814 pJMP2846 pJMP2644 Description Broad host plasmid expressing GFP under T7A1 promoter, spectomycin resistance Reference 34 CRISPRi vector expressing dCas9 and sgRNA, kanamycin resistance Transposase-containing vector, ampicillin resistance 35 35 pRL814-SO_3758 WT SO_3758 expression vector, derivative of pRL814, spectomycin resistance This study pRL814- SO_3758_A269T SO_3758_A269T expression vector, derivative of pRL814, spectomycin resistance This study pRL814-SO_1320 WT SO_1320 expression vector, derivative of pRL814, spectomycin resistance This study pRL814- SO_1320_V106I SO_1320_V106I expression vector, derivative of pRL814, spectomycin resistance pRL814- SO_1320_V106F SO_1320_V106F expression vector, derivative of pRL814, spectomycin resistance This study This study pRL814- SO_1320_S52I SO_1320_S52I expression vector, derivative of pRL814, spectomycin resistance This study pJMP2846-yadS- target_1 yadS knockdown vector target 1, derivative of pJMP2846, kanamycin resistance This study pJMP2846-yadS- target_2 yadS knockdown vector target 2, derivative of pJMP2846, kanamycin resistance This study 39 Table 3. Primers used in experiments described in Chapter 2. Primer MG_203 ctttaagaaggagatatacatatgccagatattgtcattgca Sequence (5’->3’) MG_204 aggaattcgatatcattatttatcatcatcatctttgtaatctcccgtaatttgcat tac Description SO_3758 cloning fwd SO_3758 cloning rev MG_205 ctttaagaaggagatatacatatgcatgtttcattaacaac SO_1320 fwd MG_206 aggaattcgatatcattatttatcatcatcatctttgtaatctttactcatctctttg SO_1320 rev at MG_207 TAGTagcggctaaagcgccagaaa MG_208 AAACtttctggcgctttagccgct MG_209 TAGTgtccatacggtggcgaccag MG_210 AAACctggtcgccaccgtatggac sgRNA_ check GGTCCTTGAGCCCCATTATC CRISPRi yadS target 1 forward CRISPRi yadS target 1 reverse CRISPRi yadS target 2 forward CRISPRi yadS target 2 reverse sgRNA foward Culturing Strains were precultured in 2 ml LB overnight at 30 °C with shaking at 250 rpm. The pre-cultures were washed three times and resuspended in basal M5-caa medium (1.29 mM K2HPO4, 1.65 mM KH2PO4, 7.87 mM NaCl, 1.70 mM NH4SO4, 475 μM MgSO4·7H2O, 10 mM HEPES, pH 7.2). Minimal lactate medium was used for all experimental cultures unless otherwise stated and was comprised of basal M5-caa, supplemented with Wolfe’s mineral solution, Wolfe’s vitamin solution without riboflavin, and 20 mM D,L-lactate. 40 Formate and IPTG were added to the appropriate cultures at the specified concentrations. Cultivation took place at 30 °C with shaking at 250-275 rpm. Culture volumes are specified for each experiment. Evolution of formate tolerance and genome sequencing Formate tolerance was evolved by continuously subculturing the parent strains (WT, JG2957 JG2955) in 2 ml minimal lactate medium with increasing concentrations of formate. Pre-cultures of the parent strains were prepared as described above and were used to inoculate three 2 ml cultures of minimal lactate medium with 1 mM formate to an initial OD600 of 0.02. The three or four cultures for each parent strain represented a separate evolution line: line A, line B, line C, and line D. Cultures were incubated at 30 °C with shaking at 250 rpm. 2 μl aliquots were taken daily to monitor OD600 using an Eppendorf G1.0 microcuvette and an Eppendorf BioSpectrometer. When cultures reached an OD600 of 0.2 or higher, they were subcultured into a fresh 2 ml culture with minimal lactate medium and 5 mM formate. Strains were continuously cultured in this way with the formate concentration increasing in each subculture. The formate concentration progression was: 1 mM, 5 mM, 10 mM, 20 mM, 30 mM, 40 mM, 50 mM, 60 mM, 70 mM, 80 mM, 90 mM, 100 mM. Strains spent only one round of culturing in each formate concentration before moving to the next highest concentration. Cryogenic stocks were prepared from each of the populations when the 100 mM cultures reached an OD600 of 0.2 or higher. Cryogenic stocks from each evolution line were streaked onto LB agar plates, and multiple single colonies from each evolution line were saved as pure cryogenic stocks. Genome extraction, library preparation, and Oxford Nanopore sequencing was performed by Plasmidsaurus. 41 Bioinformatics Raw genome sequencing reads were trimmed using Porechop v0.2.4 (https://github.com/rrwick/Porechop.git) to remove Nanopore sequencing adapters. Trimmed sequencing reads were aligned to the S. oneidensis MR-1 reference genome (Accession: NC_004347.2) and reference megaplasmid (Accession: NC_004349.1) and analyzed for mutations using Breseq v0.38.0.37 Deviations of the evolved strains from the reference genome were compared to the deviations of the parent strains from the reference sequences. Deviations which appeared in the evolved strains but not in the parent strains were compiled in Table 4 and classified as mutations. Mutations which appeared in multiple evolved strains were of particular interest. A sodium-dependent bicarbonate transporter (encoded by SO_3758) and a DUF2721-containing protein (encoded by SO_1320) were experimentally investigated for utility in formate tolerance. iPro54-PseKNC and Sigma70Pred were used to search for yadS σ54 and σ70 promoters at the end of the SO_1320. An 81-nucleotide long sequence containing the end of SO_1320 and the first eight nucleotides of yadS were used as the query.38,39 Amino acid sequences of proteins encoded by the WT and mutant versions of SO_3758 were aligned to the sodium-dependent bicarbonate transporter SbtA from cyanobacteria Synechocystis sp. PCC 6803 substr. Kazusa (Accession: P73953) using T- coffee simple multiple sequence alignment.40 Tools DeepGO and DeepTMHMM were used to analyze the amino acid sequence of proteins encoded by SO_1320.41,42 AlphaFold v2.3.2 was used to predict structures of proteins encoded by SO_3758_A269T, SO_1320_V106I, SO_1320_V106F, and SO_1320_S52I.43 Predicted structures for proteins encoded by SO_3758 (Accession: 42 Q8EAY3) and SO_1320 (Accession: Q8EHA9) were downloaded from the AlphaFold database (https://alphafold.com/). Ligand binding prediction was performed with AutoDock Vina v1.1.2.44 PyMOL v2.5.8 was used for visualization of protein and ligand models, identifying polar interactions between the ligand and surrounding amino acids, and calculating distances between interacting atoms.45 RESULTS Two S. oneidensis strains were used in the testing and evolution of formate tolerance: WT and JG2957. S. oneidensis strain JG2957 has three major FDHs deleted (fdhA1B1C1, fdhA2B2C2, and fdnGHI) and was included to prevent loss of formate as CO2 and provide a deeper understanding of formate tolerance. First, to assess the native formate tolerance of S. oneidensis, WT and JG2957 were cultivated in 200 μl minimal lactate medium with varying concentrations of formate (Figure 6). Both strains had the shortest lag phase in medium without formate. With even as little as 1 mM formate, growth was impacted. With increasing formate concentrations, the Figure 6. Growth curves of MR-1 and JG2957 in 200 μl minimal medium with 20 mM lactate and varying concentrations of formate. 43 strains’ lag phase also increased. For WT, formate was tolerated up to a concentration of 10 mM, and for JG2957, formate was tolerated up to a concentration of 20 mM. Formate tolerance in these strains is low, and could present an issue in a S. oneidensis strain which has CO2 reductase activity. Formate toxicity can be a result of cytochrome c oxidase inhibition and/or reduction of proton motive force due to cytoplasm acidification from the protonated acid.46–48 Since these issues are not easily solved with a genetic engineering approach, we opted to generate a formate-tolerant strain of S. oneidensis through adaptive laboratory evolution. This was accomplished through continuous subculturing of the parent strains (WT and JG2957) in minimal lactate medium while increasing the formate concentration in the medium for each subculture, with a goal of improving tolerance to 100 mM formate. Multiple separate evolution studies were conducted for each parent strain. Surprisingly, tolerance to 100 mM formate evolved within 30 days in seven evolution lines (Figure 7), Figure 7. Length of subculturing each strain required for subculturing with increasing concentrations of formate. 44 with the subculturing frequency increasing in the later stages of the evolution, indicating acquired mutation(s) assisting with formate tolerance (Figure 8). Single colonies from the seven evolution lines which could grow on 100 mM formate were isolated. Their formate tolerance was compared to the associated parent strain by culturing in 200 μl minimal lactate medium with 100 mM formate in triplicate (Figure 9). Both parent strains were incapable of growth on lactate when supplemented with 100 mM formate, while almost all the isolated evolved colonies were capable of growth, although not without a long lag phase of 60-90 hours. The evolution study from WT resulted in four lines (A, B, C, and D) which showed different growth phenotypes with lines A, C, and D having a shorter lag phase than line B. Less clear patterns emerge from strains evolved from JG2957. Two line A colonies and one line B colony could grow the most optimally under the conditions tested, while the other colonies from those lines more closely resembled the growth patterns of colonies from line C. One colony from each line was selected for further analysis. These seven new formate-tolerant S. oneidensis strains were renamed MGC001 through MGC007 (Table 1). Each formate-tolerant strain and the parent strains were cultivated in 50 ml minimal lactate medium with 100 mM formate and metabolites were analyzed via HPLC (Figure 10). Cultures were inoculated to a starting OD600 five times greater than previous cultures to allow the parent strains to grow under the high-formate conditions. This allowed comparison between the evolved and parent strains’ metabolite concentrations (Figure 11). Even with the increased inoculum volume, WT and JG2957 had a longer lag phase than the evolved formate-tolerant strains. The evolved formate-tolerant strains did not have as long of a lag phase as when inoculated to a lower OD600. All provided lactate was consumed in all strains. Formate levels remained 45 Figure 8. Growth curves of directed evolution subculturing in increasing formate concentrations of parent strains. a) WT and b) JG2957. Each panel displays growth curves of a replicate evolution line. Evolution experiments were halted after cells were able to tolerate a formate concentration of 100 mM. 46 Figure 9. Growth curves of single colonies in triplicate isolated from the evolution experiments compared to their associated parent strain. Strains were cultured in 200 μl of minimal medium with 20 mM lactate and 100 mM formate. Figure 10. Growth curves of formate-tolerant strains cultivated in 50 ml minimal lactate medium with and 100 mM formate. All lactate is consumed and formate concentrations remain constant throughout the culturing period. 47 Figure 11. Metabolite concentrations of parent and formate-tolerant strains. Strains cultivated in 50 ml minimal medium with 20 mM lactate and 100 mM formate. constant throughout cultivation of both the parent strains and the evolved formate tolerant strains, indicating that the evolved method of formate tolerance does not include formate oxidation or formate assimilation. To determine the mutations that enabled increased formate tolerance, the genomes of the seven formate-tolerant strains and the two parent strains were sequenced. Raw sequencing reads were processed with porechop to remove sequencing adapters and breseq to align the reads to the reference sequences and identify mutations. Mutations that did not exist in the parent strain were flagged (Table 4). A transition mutation of G to A in SO_3758 (SO_RS17510), encoding a sodium-dependent bicarbonate transporter, occurred in three of the seven formate-tolerant strains (MGC002, MGC005, and MGC006), 48 Figure 12. Chemical structures of formate and bicarbonate. resulted in an amino acid sequence replacement A269T, changing a nonpolar, hydrophobic residue to a bulkier polar, hydrophilic residue. Considering that formate and bicarbonate have very similar structures (Figure 12), we hypothesized two potential effects of this mutation. First, the mutation could have provided greater specificity toward formate and provided the protein encoded by SO_3758 with formate efflux activity, preventing toxic intracellular accumulation. Second, the mutation may have decreased formate specificity and prevented the protein encoded by SO_3758 from importing toxic levels of formate. In three of the remaining formate-tolerant strains (MGC003, MGC004, and MGC007), a mutation in SO_1320 (SO_RS06120), encoding a protein of unknown function, was observed. MGC007 had a G to A transition mutation resulting in a V106I amino acid sequence change and MGC003 and MGC004 had a G to T transversion mutation resulting in a V106F amino acid sequence change. We also carried out preliminary adaptive laboratory evolution on using S. oneidensis parent JG2955 which has two of the three formate dehydrogenases deleted (fdhA1B1C1 and fdhA2B2C2). Interestingly, one strain (MGC008) isolated from this evolution also had a mutation in 49 Table 4. Mutations in formate-tolerant strains. Mutations that appeared in multiple strains are bolded and highlighted. strain seq id position mutation annotation gene NC_004347 2,782,334 T→C intergenic (-97/-29) ppsA ← / → SO_RS12190 NC_004347 3,258,800 Δ1 bp intergenic (-86/+14) SO_RS14530 ← / ← SO_RS14535 NC_004347 3,588,820 NC_004347 3,928,289 NC_004349 138,980 NC_004347 3,258,789 NC_004347 3,258,800 NC_004347 3,906,124 NC_004347 1,374,567 N→T G→T Δ3 bp N→T Δ1 bp G→A C→A ?94T (ACN→ACA) SO_RS16095 ← R213L (CGC→CTC) coding (349-351/1041 nt) intergenic (-75/+25) intergenic (-86/+14) SO_RS17595 → SO_RS22825 ← SO_RS14530 ← / ← SO_RS14535 SO_RS14530 ← / ← SO_RS14535 A269T (GCC→ACC) SO_RS17510 → V106F (GTC→TTC) SO_RS06120 ← NC_004347 2,143,924 (CCATGA)10→11 coding (686/1185 nt) SO_RS09420 → NC_004347 3,258,800 NC_004347 3,355,020 NC_004347 4,740,503 NC_004347 1,374,567 Δ1 bp G→A G→A C→A intergenic (-86/+14) P74L (CCA→CTA) W48* (TGG→TGA) V106F (GTC→TTC) SO_RS14530 ← / ← SO_RS14535 flhA ← SO_RS21100 → SO_RS06120 ← NC_004347 2,143,924 (CCATGA)10→11 coding (686/1185 nt) SO_RS09420 → NC_004347 3,355,020 NC_004347 2,353,953 G→A C→T P74L (CCA→CTA) D334N (GAT→AAT) flhA ← ptsP ← NC_004347 3,258,800 Δ1 bp intergenic (-86/+14) SO_RS14530 ← / ← SO_RS14535 NC_004347 3,906,124 NC_004347 2,871,939 NC_004347 3,906,124 NC_004349 138,980 NC_004347 1,374,567 NC_004347 3,258,789 NC_004347 3,258,800 G→A C→T G→A Δ3 bp C→T N→T Δ1 bp A269T (GCC→ACC) SO_RS17510 → D23N (GAC→AAC) tolR ← A269T (GCC→ACC) SO_RS17510 → coding (349-351/1041 nt) V106I (GTC→ATC) intergenic (-75/+25) SO_RS22825 ← SO_RS06120 ← SO_RS14530 ← / ← SO_RS14535 1 0 0 C G M 2 0 0 C G M 3 0 0 C G M 4 0 0 C G M 5 0 0 C G M 6 0 0 C G M 7 0 0 C G M description phosphoenolpyruvate synthase/pyruvate, water dikinase regulatory protein tRNA-Glu/tRNA-Glu IS256-like element ISSod4 family transposase DUF599 domain-containing protein IS481-like element ISSod13 family transposase tRNA-Glu/tRNA-Glu tRNA-Glu/tRNA-Glu sodium-dependent bicarbonate transport family permease DUF2721 domain-containing protein cation diffusion facilitator family transporter tRNA-Glu/tRNA-Glu flagellar biosynthesis protein FlhA alpha/beta hydrolase-fold protein DUF2721 domain-containing protein cation diffusion facilitator family transporter flagellar biosynthesis protein FlhA phosphoenolpyruvate--protein phosphotransferase tRNA-Glu/tRNA-Glu sodium-dependent bicarbonate transport family permease protein TolR sodium-dependent bicarbonate transport family permease IS481-like element ISSod13 family transposase DUF2721 domain-containing protein tRNA-Glu/tRNA-Glu intergenic (-86/+14) SO_RS14530 ← / ← SO_RS14535 tRNA-Glu/tRNA-Glu 50 SO_1320 affecting residue 52 where a serine was converted to an isoleucine. The two residue 106 alterations resulted in a bulkier amino acid in that position, and the S52I alteration replaced a polar, hydrophilic residue with a nonpolar, hydrophobic residue. SO_1320 is upstream of SO_1319 (yadS) and SO_1318. To determine whether the mutations in SO_1320 affected the promoter sequence of downstream yadS, an 81 bp unmutated nucleotide sequence containing the end of SO_1320 and first eight nucleotides of yadS were analyzed by iPro54-PseKNC and Sigma70Pred to search for σ54 and σ70 promoters.38,39 These analyses revealed no σ54 promoter in this location, but Sigma70Pred predicted the region contained a σ70 promoter. The same 81-nucleotide sequence including the G to A or G to T mutations resulting in V106 change to phenylalanine or isoleucine in the formate-tolerant strains were also analyzed by Sigma70Pred. In both cases, the sequence is still predicted to contain a σ70 promoter, suggesting that the mutations do not alter the transcription of downstream yadS. This finding, along with the observation that a mutation elsewhere in SO_1320 also results in formate tolerance, indicates that the mutations in SO_1320 alter this protein’s activity in a way that helps the strains withstand greater formate concentrations. To confirm that the SO_1320 mutations do not impact the transcription of yadS, two S. oneidensis CRISPRi knockdown strains were prepared, each targeting yadS. The yadS knockdown strains and the non-targeting control were cultivated in minimal lactate medium with 20 mM formate and varying levels of induction. The growth curves reveal that knockdown of yadS does not improve S. oneidensis formate tolerance over the non- targeting strain, confirming that the mutations in SO_1320 do not affect yadS expression (Figure 13). 51 Figure 13. Growth curves of yadS CRISPRi knockdown (KD) strains and non- targeting control strain. To confirm that mutations observed in SO_3758 and SO_1320 were responsible for enhanced formate tolerance, we attempted to confer formate tolerance to WT S. oneidensis by overexpressing the mutant sequences. Mutant genes were cloned from the formate-tolerant strains and placed on vector pRL814 under control of an IPTG-inducible promoter. The vectors carrying the mutants were transformed into E. coli and subsequently transferred to WT S. oneidensis via conjugation. The S. oneidensis strains harboring SO_3758, SO_3758_A269T, SO_1320, SO_1320_V106I, SO_1320_V106F, or SO_1320_S52I were cultivated in 200 μl of minimal lactate medium with varying formate concentrations and the WT or mutant gene was overexpressed with 100 μM IPTG (Figure 14). All strains, whether expressing WT or mutant versions of the genes were capable of growth without formate, showing that overexpression of these proteins was not toxic. The strains expressing the WT versions of SO_3758 and SO_1320 were only able to grow robustly in the no formate conditions. The strains expressing SO_3758_A269T, SO_1320_V106I, and SO_1320_S52I grew in elevated formate levels up to 40 mM with a final OD600 similar to that of cells grown in medium without formate. None of the strains tolerated 80 mM formate, indicating that other mutations accumulated during evolution of 52 the formate-tolerant strains could enhance their ability to withstand elevated levels of formate. The strain carrying the SO_1320_V106F mutant did not exhibit as significant growth benefits as the SO_1320_V106I strain. Even during cultivation with no formate it performed worse than the strain expressing SO_1320_WT. These complementation efforts confirmed that the identified mutations in SO_3758 and SO_1320 improved formate tolerance of WT S. oneidensis. To predict the effect of the A269T substitution in the transporter encoded by SO_3758, we modeled the binding of bicarbonate and formate to the WT and mutant versions of the protein. To reveal the binding site for modeling substrate binding, the amino acid sequence of the S. oneidensis protein encoded by SO_3758 was aligned with the only sodium-dependent bicarbonate transporter with a solved structure, SbtA from Figure 14. Growth curves of WT S. oneidensis harboring an IPTG-inducible expression vector containing WT or mutant versions of SO_3758 or SO_1320. Strains cultivated 200 μl minimal lactate medium with 100 μM IPTG and varying formate concentrations. 53 cyanobacteria, using the multiple sequence alignment (MSA) tool T-Coffee 40,49. The alignment revealed that A269 in the protein encoded by SO_3758 corresponds to residue S323 in SbtA (Figure 15). While S323 has not been found to be directly involved in binding sodium or bicarbonate, it is the single residue that separates the two binding pockets, with S322 binding sodium and S324 binding bicarbonate 49,50. These substrate-binding serines are conserved in the protein encoded by SO_3758, with SO_3758 S268 corresponding to SbtA S322 and SO_3758 S270 corresponding to SbtA S324 (Figure 15). The MSA revealed that the protein encoded by SO_3758 residues V99, S100, S270, and I272 are likely part of the bicarbonate binding pocket. Protein models corresponding to the wild-type and mutated versions of SO_3758 were generated using AlphaFold,43 and the binding of bicarbonate and formate were predicted with Autodock Vina.44 The predicted local distance difference test (pLDDT) scores were calculated for each residue in the AlphaFold model, and are depicted in Figure 16. The binding pocket residues were predicted to have pLDDT scores of high (90> pLDDT > 70) to very high (pLDDT > 90) confidence. The docking grid was set to encompass residues predicted by MSA to be part of the binding pocket (V99, S100, S270, I272), A or T269, and nearby residues. The modeling of bicarbonate binding to the protein encoded by SO_3758_WT predicts that it forms polar interactions with V99, A101, Y271, I272, and A273 in the predicted native binding site, referred to here as site 1 (Figure 17a, Figure 16b, Figure 16d). Bicarbonate binding to site 1 has a predicted binding affinity of -3.2 kcal mol-1 (Table 5). Formate is also predicted to bind to site 1 in the protein encoded by SO_3758_WT, forming polar interactions with V99, S270, Y271, and I272 and having a weaker predicted binding affinity of -2.3 kcal mol-1 (Figure 17b). 54 Figure 15. Multiple sequence alignment of cyanobacteria SbtA, and the proteins encoded by WT SO_3758 and SO_3758_A269T. Bracket highlights residues A269 from the protein encoded by WT SO_3758, T269 from the protein encoded by SO_3758_A269T, and corresponding cyanobacteria SbtA residue S323, as well as conserved flanking serine residues involved in the binding of bicarbonate and sodium. 55 site 1 site 2 site 1 site 2 Figure 16. Representations of pLDDT scores from the protein sequence encoded by SO_3758 AlphaFold structural predictions. Residue color depicts pLDDT score. Dark blue represents a very high score (pLDDT > 90). Light blue represents a confident score (90 > pLDDT > 70). Yellow represents a low score (70 > pLDDT > 50). Orange represents a very low score (pLDDT < 50). a) General structure of SO_3758 shown at two angles. View 1 is shown in panels b and c. View 2 is shown in panels d and e. b) View 1 of the protein encoded by WT SO_3758. Formate is colored red and shown binding to site 1. c) View 1 of the protein encoded by SO_3758_A269T. Formate is colored red and shown binding to site 2. d) View 2 of the protein encoded by WT SO_3758. Formate is colored red and shown binding to site 1. c) View 2 of the protein encoded by SO_3758_A269T. Formate is colored red and shown binding to site 2. Table 5. Predicted binding affinities of bicarbonate and formate to site 1 and site 2 of WT and mutant proteins encoded by SO_3758. Binding affinity to the protein encoded by SO_3758 Binding affinity to the protein encoded by SO_3758_A269T Site 1 (kcal mol-1) Site 2 (kcal mol-1) Site 1 (kcal mol-1) Site 2 (kcal mol-1) Bicarbonate Formate -3.2 -2.3 -3.4 -2.9 -3.7 -3.1 -2.9 NA 56 When modeled with the protein encoded by SO_3758_A269T, bicarbonate is still predicted to bind to site 1 and form polar interactions with the same residues with a slightly stronger binding affinity of -3.4 kcal mol-1. However, the greatest predicted binding affinity of -3.7 kcal mol-1 is in a binding pocket with the altered T269 residue, as well as residues T103 and N302, here referred to as site 2 (Figure 17c, Figure 16c, Figure 16e). This indicates that the A269T substitution altered the native bicarbonate binding site 1 slightly to promote tighter binding of the native substrate, even though residue 269 is not itself a part of site 1. More interestingly, it shows that the change in residue 269 alters site 2 to stronger affinity for the substrate. Finally, when formate binding is modeled with the protein encoded by SO_3758_A269T, the greatest predicted binding affinity (-3.1 kcal mol-1) is also in site 2 with T103, N302, and the altered residue T269 (Figure 17d). This brings the predicted binding affinity of formate much closer to that of bicarbonate in site 1 of the native protein encoded by SO_3758 (-3.2 kcal mol-1), suggesting that formate is a suitable substrate for this mutant. The predicted interaction of formate with threonine 269 in the protein encoded by SO_3758_A269T but not with alanine in the WT version also suggests that this alteration was beneficial in promoting formate binding. The two binding sites in the protein encoded by SO_3758_A269T are predicted to bind both substrates, but formate is not predicted to bind to site 2 in the native protein (Table 5). Site 1 binds both bicarbonate and formate better than site 2. The opposite is true for the protein encoded by SO_3758_A269T where site 2 with residue threonine 269 binds both substrates better and increases the predicted binding affinity of formate to a similar level that is predicted for bicarbonate binding in the native protein. Based on this modeling, we anticipated that the A269T substitution in the protein encoded by SO_3758 57 Figure 17. Ligand dock modeling of formate and bicarbonate to the proteins encoded by SO_3758_WT and SO_3758_A269T. Modeled binding of bicarbonate -) to the proteins encoded by SO_3758_WT (a) and SO_3758_A269T (c) and of (HCO3 -) to the proteins encoded by SO_3758_WT (b) and SO_3758_A269T (d). formate (HCO2 Residues in the native bicarbonate binding site (site 1) are shown in pink and residues in site 2 are shown in cyan. Formate and bicarbonate are shown in green. Nitrogen atoms are colored blue, oxygen atoms are colored red, and hydrogen atoms are colored white. Polar interactions between bicarbonate or formate and interacting residues are shown as yellow dotted lines and are labeled with the distance between the interacting atoms. provided MGC002, MGC005, and MGC006 with greater formate efflux capacity to prevent the toxic accumulation of formate inside the cells. The effects of mutations in DUF2721-containing protein encoded by SO_1320 are less clear and more difficult to ascertain. While the function of both DUF2721 and the protein encoded by SO_1320 are unknown, some clues exist to the protein’s potential role. Both YadS and the protein encoded by SO_1328 expressed from the same operon as the protein encoded by SO_1320 are part of membrane protein families. YadS is annotated 58 as a trimeric intracellular cation (TRIC) channel and the protein encoded by SO_1318 is annotated as a nuclear transport factor 2-like (NTF2) protein.51 The proximity of the protein encoded by SO_1320 to these membrane-bound transporter proteins could be an indicator that it shares a similar function. In an attempt to elucidate the potential function of the protein encoded by SO_1320, the amino acid sequence was analyzed using DeepGO, a deep learning tool which predicts localization and function of a protein based on its sequence.52 DeepGO predicted that the protein encoded by SO_1320 is localized to the cell membrane but was unable to predict its function, likely due to a lack of experimentally characterized homologs (Table 6). Further analysis of the amino acid sequence of the protein encoded by SO_1320 with DeepTMHMM,53 a tool for predicting transmembrane proteins using deep neural networks, reiterates the protein’s probable localization to the cell membrane, finding three transmembrane helices (Figure 18). These results suggest that the protein encoded by SO_1320 is associated with the cellular membrane, although it is still unclear what function it serves. Since the formate concentrations are constant throughout culturing of MGC003, MGC004, and MGC007 (formate tolerant strains with mutations in SO_1320), the evolved method of formate tolerance must not involve the oxidation or assimilation of formate. Potential formate tolerance mechanisms of SO_1320 mutants could include enhanced Table 6. DeepGO predictions for the protein encoded by SO_1320. Gene Ontology Cellular Component Description Prediction Score GO:0110165 cellular anatomical entity GO:0016020 membrane 0.392872 0.370649 Molecular Function Biological Process 59 Figure 18. Structure prediction of the protein encoded by SO_1320 by DeepTMHMM. active formate efflux, decreased active formate influx, or membrane stabilization preventing passive formate diffusion. To determine if the protein encoded by SO_1320 could bind formate, the WT and mutated versions of this protein were modeled using AlphaFold43 and formate binding was predicted with AutoDock Vina.44 Modeling results colored with pLDDT scores are shown in Figure 19. The protein encoded by SO_1320 did not appear to have convincing formate binding activity, as the best predicted binding affinity was -2.1 kcal mol-1 at a site on the protein surface. Modeling of formate binding to the proteins encoded by SO_1320_V106I, SO_1320_V106F, and SO_1320_S52I calculated predicted binding affinities to formate of 60 -2.0 kcal mol-1, -2.1 kcal mol-1, and -2.1 kcal mol-1, respectively. The low predicted binding affinity, the location of the predicted binding, and the lack of improvement to the predicted binding affinity in the mutants suggest that neither the WT or mutant version of the protein encoded by SO_1320 are acting as a formate transporter; however, it does not eliminate the possibility that it could be a formate channel or that it could form a multimeric complex with a formate transporter. DISCUSSION This work describes the evolution of seven formate-tolerant S. oneidensis strains for use in systems subjected to high formate concentrations. Two notable mutations arose during evolution of S. oneidensis for greater formate tolerance. The first is a substitution in alanine 269 to a threonine in the sodium-dependent bicarbonate transporter encoded by SO_3758. Every strain exhibiting this mutation in SO_3758 (MGC002, MGC005, and MGC006) shared the same A269T substitution in the encoded protein and was able to withstand greater formate concentrations than the parent strains. The second gene with mutations attributed to greater formate tolerance was SO_1320 encoding a protein of unknown function. Alterations occurred in residue 106, changing the native valine to either isoleucine (MGC007) or phenylalanine (MGC003 and MGC004), and in residue 52 (MGC008), changing the native serine to isoleucine. The evolved strains provided with the capacity to grow robustly on 100 mM formate in minimal lactate medium when the parent strains they are derived from were only capable of growth on 10-20 mM formate. Strains harboring a substitution in V106 of the protein encoded by SO_1320 had a shorter lag phase than strains harboring the A269T substitution in the protein encoded by SO_3758, although all strains grew to a similar final OD600 (Figure 10). Therefore, the 61 Figure 19. AlphaFold modeling of the proteins encoded by WT and mutant versions of SO_1320. Residue color depicts pLDDT score. Dark blue represents a very high score (pLDDT > 90). Light blue represents a confident score (90 > pLDDT > 70). Yellow represents a low score (70 > pLDDT > 50). Orange represents a very low score (pLDDT < 50). Models of proteins encoded by a) SO_1320, b) SO_1320_V106I, c) SO_1320_V106F, and d) SO_1320_S52I. 62 onset of formate tolerance benefits from the SO_1320 mutations are more instantaneous than the SO_3758_A269T mutation. The extra delay in growth of strains with the SO_3758_A269T mutation may be related to the mechanism by which it provides formate tolerance. Modeling of formate binding in the WT and mutant versions of the protein encoded by SO_3758 revealed that the conversion of alanine 269 to threonine increased the binding affinity of formate to a site deeper in the binding pocket. This site is predicted to bind bicarbonate in the native protein with a lower predicted binding affinity, indicating that the mutation did not create a new binding site, but rather strengthened an existing one. The predicted binding affinity of both bicarbonate and formate in the deeper binding site of the protein encoded by SO_3758_A269T is stronger than the binding of both substrates to both the native and deeper binding site in the unmutated protein. The mutation allows formate binding to a similar efficiency of the native bicarbonate binding, suggesting that the mutation has made formate a suitable substrate for the protein encoded by SO_3758_A269T. The ability to better bind formate suggests that the strains harboring this mutation had improved formate efflux capabilities over the WT. Without the benefit of being able to transport formate, the parent strains with WT SO_3758 could not prevent the toxic accumulation of formate in the cells. If the mechanism of formate tolerance in SO_3758 mutants is greater formate efflux rates, the delay in growth may be because formate continues to enter the cell, and the tolerance to the toxic compound only arises when SO_3758_A269T has cleared enough of it from the cytoplasm. The role of changes in DUF2721-containing protein encoded by SO_1320 are less clear as neither this domain nor this protein have any known characterized homologues. SO_1320 is in an operon with genes encoding a membrane protein channel or membrane 63 transport protein (yadS and SO_1318), and sequence analysis tools predict three transmembrane helices. These clues suggest that the protein encoded by SO_1320 is also a membrane protein. Modeling of formate binding to WT and mutant versions of the protein encoded by SO_1320 indicate that the protein does not bind formate as a transporter might. While it is still unclear how SO_1320 mutants are providing cells with greater formate tolerance, our analysis does not rule out the possibilities that the protein of unknown function encoded by SO_1320 is involved in membrane stabilization or acts as a formate channel. Additionally, the protein encoded by SO_1320 could be one subunit in a multimeric formate transporter which binds formate in a different subunit. Expression of these mutant genes in the S. oneidensis WT background on minimal lactate media increased formate tolerance two-fold. The ability to reconstruct the formate tolerant phenotype through expression of these altered proteins further confirms their role in the evolved strains. The simple point mutations which lead to these growth improvements provide two easily accessible methods for transferring formate tolerance to S. oneidensis and potentially other organisms. This could be a valuable tool in engineering biocatalysts that would benefit from reduced intracellular formate or bicarbonate levels, such as in organisms in which these compounds cease metabolic activity prematurely due to toxicity issues. FUNDING SOURCES This work was supported by a 2018 Beckman Young Investigator Award and a National Science Foundation CAREER Award (1750785) to MT. 64 ACKNOWLEDGEMENTS We thank Dr. Jeffrey Gralnick for S. oneidensis strains JG2955 and JG2957. We also thank Drs. Nanye Long and Nicholas Panchy for their assistance with Breseq and AlphaFold. 65 REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. Dessì, P. et al. Microbial electrosynthesis: Towards sustainable biorefineries for production of green chemicals from CO2 emissions. Biotechnol Adv 46, 107675 (2021). Jourdin, L. & Burdyny, T. Microbial Electrosynthesis: Where Do We Go from Here? Trends Biotechnol 39, 359–369 (2021). Jensen, H. M. et al. Engineering of a synthetic electron conduit in living cells. PNAS 107, 19213–19218 (2010). Schiel-Bengelsdorf, B. & Dürre, P. Pathway engineering and synthetic biology using acetogens. FEBS Lett 586, 2191–2198 (2012). Prévoteau, A., Carvajal-Arroyo, J. M., Ganigué, R. & Rabaey, K. Microbial electrosynthesis from CO2: forever a promise? Curr Opin Biotechnol 62, 48–57 (2020). Rosenbaum, M. A. & Henrich, A. W. Engineering microbial electrocatalysis for chemical and fuel production. Curr Opin Biotechnol 29, 93–98 (2014). Katsyv, A. & Müller, V. Overcoming Energetic Barriers in Acetogenic C1 Conversion. Front Bioeng Biotechnol 8, (2020). Philips, J. Extracellular Electron Uptake by Acetogenic Bacteria: Does H2 Consumption Favor the H2 Evolution Reaction on a Cathode or Metallic Iron? Front Microbiol 10, (2020). Ganigué, R., Puig, S., Batlle-Vilanova, P., Balaguer, M. D. & Colprim, J. Microbial electrosynthesis of butyrate from carbon dioxide. Chem Comm 51, 3235–3238 (2015). 10. Raes, S. M. T., Jourdin, L., Buisman, C. J. N. & Strik, D. P. B. T. B. Continuous Long‐Term Bioelectrochemical Chain Elongation to Butyrate. ChemElectroChem 4, 386–395 (2017). 11. Arends, J. B. A., Patil, S. A., Roume, H. & Rabaey, K. Continuous long-term electricity-driven bioproduction of carboxylates and isopropanol from CO2 with a mixed microbial community. J CO2 Util 20, 141–149 (2017). 12. Harrington, T. D. et al. Neutral red-mediated microbial electrosynthesis by Escherichia coli, Klebsiella pneumoniae, and Zymomonas mobilis. Bioresour Technol 195, 57–65 (2015). 13. Tefft, N. M. & Teravest, M. A. Reversing an Extracellular Electron Transfer Pathway for Electrode-Driven Acetoin Reduction. ACS Synth Biol 8, 1590–1600 (2019). 66 14. Ross, D. E., Flynn, J. M., Baron, D. B., Gralnick, J. A. & Bond, D. R. Towards Electrosynthesis in Shewanella: Energetics of Reversing the Mtr Pathway for Reductive Metabolism. PLoS One 6, e16649 (2011). 15. Rowe, A. R. et al. Tracking Electron Uptake from a Cathode into Shewanella Cells: Implications for Energy Acquisition from Solid-Substrate Electron Donors. mBio 9, (2018). 16. Kouzuma, A. Molecular mechanisms regulating the catabolic and electrochemical activities of Shewanella oneidensis MR-1. Biosci Biotechnol Biochem 85, 1572– 1581 (2021). 17. Brutinel, E. D. & Gralnick, J. A. Shuttling happens: soluble flavin mediators of extracellular electron transfer in Shewanella. Appl Microbiol Biotechnol 93, 41–48 (2012). 18. Richardson, D. J. et al. The ‘porin–cytochrome’ model for microbe‐to‐mineral electron transfer. Mol Microbiol 85, 201–212 (2012). 19. Madsen, C. S. & TerAvest, M. A. NADH dehydrogenases Nuo and Nqr1 contribute in transfer by Shewanella oneidensis MR-1 to extracellular electron bioelectrochemical systems. Sci Rep 9, 14959 (2019). 20. Beblawy, S. et al. Extracellular reduction of solid electron acceptors by Shewanella oneidensis. Mol Microbiol 109, 571–583 (2018). 21. Kane, A. L. et al. Formate Metabolism in Shewanella oneidensis Generates Proton Motive Force and Prevents Growth without an Electron Acceptor. J Bacteriol 198, 1337–1346 (2016). 22. Marritt, S. J. et al. A functional description of CymA, an electron-transfer hub supporting anaerobic respiratory flexibility in Shewanella. Biochem J 444, 465–474 (2012). 23. Myers, C. R. & Myers, J. M. Cloning and sequence of cymA, a gene encoding a tetraheme cytochrome c required for reduction of iron(III), fumarate, and nitrate by Shewanella putrefaciens MR-1. J Bacteriol 179, 1143–1152 (1997). 24. Sturm, G. et al. A dynamic periplasmic electron transfer network enables respiratory flexibility beyond a thermodynamic regulatory regime. ISME J 9, 1802–1811 (2015). 25. Coursolle, D. & Gralnick, J. A. Reconstruction of Extracellular Respiratory Pathways for Iron(III) Reduction in Shewanella oneidensis Strain MR-1. Front Microbiol 3, (2012). 26. Coursolle, D., Baron, D. B., Bond, D. R. & Gralnick, J. A. The Mtr Respiratory Pathway Is Essential for Reducing Flavins and Electrodes in Shewanella oneidensis. J Bacteriol 192, 467–474 (2010). 67 27. Cotton, C. A., Edlich-Muth, C. & Bar-Even, A. Reinforcing carbon fixation: CO2 reduction replacing and supporting carboxylation. Curr Opin Biotechnol 49, 49–56 (2018). 28. François, J. M., Lachaux, C. & Morin, N. Synthetic Biology Applied to Carbon Conservative and Carbon Dioxide Recycling Pathways. Front Bioeng Biotechnol 7, (2020). 29. Bar-Even, A., Noor, E. & Milo, R. A survey of carbon fixation pathways through a quantitative lens. J Exp Bot 63, 2325–2342 (2012). 30. Calzadiaz-Ramirez, L. & Meyer, A. S. Formate dehydrogenases for CO2 utilization. Curr Opin Biotechnol 73, 95–100 (2022). 31. Amao, Y. Formate dehydrogenase for CO2 utilization and its application. J CO2 Util 26, 623–641 (2018). 32. Bar-Even, A. Formate Assimilation: The Metabolic Architecture of Natural and Synthetic Pathways. Biochemistry 55, 3851–3863 (2016). 33. Reda, T., Plugge, C. M., Abram, N. J. & Hirst, J. Reversible interconversion of carbon dioxide and formate by an electroactive enzyme. PNAS 105, 10654–10658 (2008). 34. Ghosh, I. N. et al. OptSSeq explores enzyme expression and function landscapes to maximize isobutanol production rate. Metab Eng 52, 324–340 (2019). 35. Peters, J. M. et al. Enabling genetic analysis of diverse bacteria with Mobile- CRISPRi. Nat Microbiol 2019 4:2 4, 244–250 (2019). 36. Ford, K. C., Kaste, J. A. M., Shachar-Hill, Y. & Teravest, M. A. Flux-Balance Analysis and Mobile CRISPRi-Guided Deletion of a Conditionally Essential Gene in Shewanella oneidensis MR-1. ACS Synth Biol 11, 3405–3413 (2022). 37. Deatherage, D. E. & Barrick, J. E. Identification of Mutations in Laboratory-Evolved Microbes from Next-Generation Sequencing Data Using breseq. Methods Mol Biol 1151, 165–188 (2014). 38. Lin, H., Deng, E.-Z., Ding, H., Chen, W. & Chou, K.-C. iPro54-PseKNC: a sequence- based predictor for identifying sigma-54 promoters in prokaryote with pseudo k-tuple nucleotide composition. Nucleic Acids Res 42, 12961–12972 (2014). 39. Patiyal, S., Singh, N., Ali, M. Z., Pundir, D. S. & Raghava, G. P. S. Sigma70Pred: A highly accurate method for predicting sigma70 promoter in Escherichia coli K-12 strains. Front Microbiol 13, (2022). 40. Notredame, C., Higgins, D. G. & Heringa, J. T-coffee: a novel method for fast and accurate multiple sequence alignment. J Mol Biol 302, 205–217 (2000). 68 41. Kulmanov, M., Khan, M. A. & Hoehndorf, R. DeepGO: predicting protein functions interactions using a deep ontology-aware classifier. from sequence and Bioinformatics 34, 660–668 (2018). 42. Hallgren, J. et al. DeepTMHMM predicts alpha and beta transmembrane proteins using deep neural networks. bioRxiv (2022). 43. Jumper, J. et al. Highly accurate protein structure prediction with AlphaFold. Nature 596, 583–589 (2021). 44. Trott, O. & Olson, A. J. AutoDock Vina: Improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading. J Comput Chem 31, 455–461 (2010). 45. Schrödinger LLC. The PyMOL Molecular Graphics System, Version~2.5.8. https://pymol.org/2/ (2015). 46. Nicholls, P. Formate as an inhibitor of cytochrome c oxidase. Biochem Biophys Res Commun 67, 610–616 (1975). 47. Warnecke, T. & Gill, R. T. Organic acid toxicity, tolerance, and production in Escherichia coli biorefining applications. Microb Cell Fact 4, 25 (2005). 48. Cotton, C. A., Claassens, N. J., Benito-Vaquerizo, S. & Bar-Even, A. Renewable methanol and formate as microbial feedstocks. Curr Opin Biotechnol 62, 168–180 (2020). 49. Liu, X.-Y. et al. Structures of cyanobacterial bicarbonate transporter SbtA and its complex with PII-like SbtB. Cell Discov 7, 63 (2021). 50. Fang, S. et al. Molecular mechanism underlying transport and allosteric inhibition of bicarbonate transporter SbtA. PNAS 118, (2021). 51. Heidelberg, J. F. et al. Genome sequence of the dissimilatory metal ion–reducing bacterium Shewanella oneidensis. Nat Biotechnol 20, 1118–1123 (2002). 52. Kulmanov, M., Khan, M. A. & Hoehndorf, R. DeepGO: predicting protein functions interactions using a deep ontology-aware classifier. from sequence and Bioinformatics 34, 660–668 (2018). 53. Hallgren, J. et al. DeepTMHMM predicts alpha and beta transmembrane proteins using deep neural networks. bioRxiv (2022) doi:10.1101/2022.04.08.487609. 69 CHAPTER 3: RNF COMPLEX GENERATES REDUCED FERREDOXIN FOR IRON-SULFUR CLUSTER FORMATION IN SHEWANELLA ONEIDENSIS MR-1 70 PREFACE Early on in my PhD program, my main research focus was engineering-based applied science. I intended to create an electroautotrophic strain of S. oneidensis for use in MES. While that remained the goal throughout my time here, it was a risky endeavor. Although I had two approaches to make this strain, there was no guarantee it would work, and especially no guarantee it would work in the time available to me. So, I added a more exploratory project to my repertoire – elucidating the role of Rnf in S. oneidensis. Originally, Rnf was interesting to us as a prospective source of low potential electrons for carbon fixation. Rnf connects the NADH and ferredoxin pools by ion motive force and has been shown to operate in either the ferredoxin-reducing and ferredoxin- oxidizing direction. Characterization of this enzyme complex in S. oneidensis was expected to give us insight into its catalytic activity and the fate of the ferredoxin. I was able to determine the direction of the activity (it has ferredoxin-reducing activity like we were hoping), but the fate of the reduced ferredoxin surprised us. We were able to confirm our hypothesis through a collaboration with Daniel Amador-Noguez’s lab at University of Wisconsin, Madison. A graduate student of his, Julio Rivera Vazquez, ran LC-MS metabolomics on metabolite extracts that I prepared from our knockdown strains. The data we got from this collaboration was instrumental in confirming the Rnf model that we proposed. This project was a great example of following the data where it leads you, because the model we ultimately proposed for Rnf function in S. oneidensis was not at all on our radar when the project began. It was a helpful exercise for me in combatting confirmation bias and being open-minded about all the potential explanations for an observation. 71 Ultimately, I believe what we discovered has greater impact than our original hypothesis, and I’m excited to see where future research into this topic leads. This chapter describes the experimentation of Rnf in S. oneidensis and presents a new model for Rnf activity across species. ABSTRACT Rnf (Rhodobacter nitrogen fixation) is a membrane protein complex found in diverse bacteria that connects the NADH pool to ferredoxins and couples its redox reaction to ion translocation across the membrane. In the ion motive force-generating direction, Rnf oxidized ferredoxins for NADH production. Alternatively, in the ion motive force-consuming direction, Rnf generates reduced ferredoxins from NADH. This activity has been implicated in a range of pathways, including nitrogen fixation and SoxR reduction. In this work, we characterize the role of the Rnf complex in Shewanella oneidensis, an organism without nitrogen fixation or a Sox system. Insights into the role of Rnf in this host give rise to a comprehensive model for Rnf function across species. We propose the Rnf function is two-pronged. The first is that Rnf generates reduced ferredoxins for enzymes requiring them as cofactors. The second is that one of the most impactful enzymes that uses Rnf- derived ferredoxins is cysteine desulfurase, which transfers elemental sulfur from cysteine to an iron-sulfur cluster during iron-sulfur cluster biosynthesis and repair. The dual functionality of Rnf activity emphasizes its significance in maintaining cellular redox balance and facilitating vital metabolic processes in bacteria. Understanding the mechanisms by which Rnf contributes to iron-sulfur cluster biosynthesis and enzyme activity can provide insights into bacterial physiology. 72 INTRODUCTION The Rnf complex was first described in Rhodobacter capsulatus for its involvement in nitrogen fixation, leading to its moniker Rnf (Rhodobacter nitrogen fixation).1,2 In R. capsulatus, the Rnf complex generates reduced ferredoxin from NADH powered by ion motive force.1,3 The low potential reduced ferredoxin (E0’ = -500 to -420 mV) and ATP are used by nitrogenase (Nif) to reduce nitrogen to ammonia (Figure 20a).1 Since its discovery in R. capsulatus, Rnf has been found in many other bacteria and some archaea, including organisms that do not fix nitrogen.3 For example, Rnf in Escherichia coli does not contribute to nitrogen fixation, but rather returns the oxygen-sensing transcription factor SoxR to its inactivated state by reducing its iron-sulfur cluster (ISC).4 Rnf also operates as a sodium pump driven by ferredoxin oxidation, such as in Acetobacterium woodii, where the first characterization connected Rnf activity to caffeate reduction (Figure 20b).5 Many studies attempt to classify the direction of Rnf activity in a particular organism.1,2,5,6 While useful for identifying certain pathways Rnf participates in, evidence Figure 20. Rnf activities in the ferredoxin-reducing and ferredoxin-oxidizing directions. a) In R. capsulatus, Rnf generates reduced ferredoxin from NADH using sodium or proton motive force. Reduced ferredoxins are used along with ATP to power nitrogen fixation. b) In A. woodii, Rnf oxidizes reduced ferredoxin to pump sodium across the inner membrane. This activity is part of caffeate reduction. 73 suggests Rnf can operate bidirectionally depending on the ferredoxin:NADH ratio. For example, the A. woodii Rnf was originally characterized as operating in the sodium- pumping direction driven by ferredoxin oxidation.5 However, a more recent study found that the complex can also catalyze ferredoxin reduction when NADH levels are greater than that of ferredoxin, such as when the organism is cultivated with low-energy substrates.7 This points to a broader role of Rnf connecting the ferredoxin and NADH pools depending on cellular needs. Additionally, the variety of pathways found to include Rnf activity also suggests an unidentified link connecting Rnf to these diverse processes. Rnf operons have some variation in structure depending on the species.3 R. capsulatus has an operon structure of rnfABCDGEH.2 A common variation on this structure which is shared by organisms such as E. coli and Shewanella oneidensis is the substitution of rnfH with nth, a gene encoding endonuclease III.8,9 While the inclusion of nth in the rnf operon is not uncommon, no link between Rnf and Nth has been established to date. R. capsulatus rnf operon was suggested to be negatively controlled by ferric-uptake regulator (Fur), with less Rnf expression under iron-limiting conditions.2 Decreased Rnf expression under iron-limiting conditions as regulated by Fur was also found in C. difficile.10 Regulation of Rnf by iron availability suggests some connection between Rnf and iron metabolism, although the functional link has yet to be established. Shewanella oneidensis is a heterotrophic, facultative anaerobe, with the capability to respire extracellular electron acceptors using a process called extracellular electron transfer.11–13 This organism can respire a diverse set of terminal electron acceptors such as nitrate, nitrite, dimethyl sulfoxide, vanadate, and iron oxides.14–17 Its respiratory flexibility 74 is the result of a branched electron transfer chain which has been the subject of many studies.18–21 As such, the electron transfer pathways of S. oneidensis are well understood, which makes it a good host to study the wider role of Rnf in electron transfer. Additionally, S. oneidensis does not have a SoxRS system, cannot fix nitrogen, and cannot respire caffeate, so its Rnf likely does not have any of the previously proposed functions. This poises S. oneidensis as a promising host to elucidate the broader role that Rnf plays. The connection of NADH and ferredoxin pools is an important aspect for bioengineering in cases such as carbon dioxide or nitrogen fixation. The reduction potentials of carbon dioxide to carbon monoxide and formate is E0’ = -520 and E0’ = -610 mV, respectively.22 These reduction reactions are especially important in the bioengineering of microbial hosts for conversion of carbon dioxide into value-added products via microbial electrosynthesis. However, they are difficult to achieve because they require electron donors with low potentials. Driving electrons uphill can be a powerful tool in making these reactions more attainable. Previously, we have shown that proton motive force could be used to drive NADH dehydrogenase in reverse in a bioengineered strain of S. oneidensis, moving electrons from the quinone pool to lower potential NADH.23 The ability to drive electrons to an even lower potential (i.e. ferredoxin) would be an important step toward developing more robust hosts for microbial electrosynthesis. In this work, we examine the function of Rnf in S. oneidensis and propose a model which may explain its effect in previously studied systems. 75 METHODS Strains and plasmids Strains and plasmids used in this study are listed in Table 7 and Table 8, respectively. sgRNA in CRISPRi knockdown vector pJMP2846, also carrying dCas9, were prepared to target rnfA, iscS, or nth using previously described methods.24,25 Assembled CRISPRi vectors were inserted into S. oneidensis MR-1 using a Tn7-based triparental conjugation method.25 The donor strains were 2,6-diaminopimelic acid (DAP)-auxotroph E. coli WM3064 carrying the assembled CRISPRi vector and DAP-auxotroph E. coli WM3064 pJMP2644 carrying a transposase. After a mating period of 8 hours on LB agar plates supplemented with 0.3 mM DAP, counter-selection was performed on LB agar plates supplemented with 50 μg ml-1 kanamycin in the absence of DAP. S. oneidensis colonies were screened for correct sgRNA construction with the sgRNA_check and appropriate sgRNA reverse primers (Table 9). Table 7. Strains described in Chapter 3. Strain Description S. oneidensis MR-1 WT S. oneidensis Reference Lab stock Non-targeting WT S. oneidensis harboring pJMP2846 This study rnf knockdown WT S. oneidensis harboring pJMP2846-rnf-88 This study iscS knockdown WT S. oneidensis harboring pJMP2846-iscS-111 This study nth knockdown WT S. oneidensis harboring pJMP2846-nth-141 This study E. coli WM3064 Conjugal transfer donor strain, DAP auxotroph Lab stock 76 Table 8. Plasmids described in Chapter 3. Plasmid pJMP2846 Description Non-targeting CRISPRi vector with dCas9 and sgRNA under control of an IPTG-inducible promoter, kanamycin resistance Reference 25 pJMP2846-rnf-88 pJMP2846 with non-targeting sgRNA replaced with rnfA-targeting sgRNA, kanamycin resistance This study pJMP2846-iscS-111 pJMP2846 with non-targeting sgRNA replaced This study with iscS-targeting sgRNA, kanamycin resistance pJMP2846-nth-141 pJMP2846 with non-targeting sgRNA replaced with nth-targeting sgRNA, kanamycin resistance This study pJMP2644 Contains transposase for triparental conjugation, ampicillin resistance 25 Table 9. Primers described in Chapter 3. Primer MG_087 Sequence (5’->3’) TAGTTTGCTGGAGACCCCCATAAA Description rnfA sgRNA forward MG_088 AAACTTTATGGGGGTCTCCAGCAA rnfA sgRNA reverse MG_110 TAGTCGAGTCGCGGAAAAAATGTT iscS sgRNA forward MG_111 AAACAACATTTTTTCCGCGACTCG iscS sgRNA reverse MG_140 TAGTTTCAGCTCGGTCTGTGGTTT nth sgRNA forward MG_141 AAACAAACCACAGACCGAGCTGAA nth sgRNA reverse sg_RNA check GGTCCTTGAGCCCCATTATC sgRNA forward 77 Culturing Knockdown strains were precultured in 2 ml LB supplemented with 50 μg ml-1 kanamycin overnight at 30 °C with shaking at 250 rpm. The pre-cultures were washed three times and resuspended in basal M5 media (1.29 mM K2HPO4, 1.65 mM KH2PO4, 7.87 mM NaCl, 1.70 mM NH4SO4, 475 μM MgSO4·7H2O, 0.1 g L-1 casamino acids, 10 mM HEPES, pH 7.2). Washed cells were used to inoculate minimal medium cultures to an initial OD600 of 0.02 unless otherwise specified. Minimal lactate medium was comprised of M5 basal medium supplemented with Wolfe’s mineral solution, Wolfe’s vitamin solution without riboflavin, and 20 mM D,L-lactate. Inducer isopropyl β-D-1-thiogalactopyranoside (IPTG) was added to cultures at the specified concentrations. Where specified, thioflavin T (ThT), carbonyl cyanide m-chlorophenyl hydrazone (CCCP), and ETH2120 were added to cultures at working concentrations of 10 μM. Culture volumes are specified for each experiment and were conducted at 30 °C with shaking at 250 - 275 rpm. ThT was excited at 412 nm and fluorescence was monitored at 490 nm. Fluorescence intensity is reported as the ThT fluorescence normalized to the OD600. Colony forming units Cell viability was determined by plating 3 μl spots of washed and serially diluted pre-cultures (prepared as described above) onto LB agar plates supplemented with 50 μg ml-1 kanamycin and the specified amount of IPTG. Plates were incubated at 30 °C for ~16 hours. Colonies were counted from dilutions that produced single colonies. LC-MS Metabolomics The rnf, iscS, and non-targeting knockdown strains were pre-cultured as described above and inoculated into 50 ml minimal lactate medium supplemented with 50 μg ml-1 78 kanamycin to an initial OD600 of 0.035. The cultures were incubated at 30 °C with shaking at 250 rpm for 12 hours. Inoculation of cultures were staggered over the course of an hour, and metabolites were extracted after 12 h of culturing in the same order the cultures were inoculated. This staggering of inoculation and extraction ensured metabolites were extracted as close to the same total cultivation time for each sample as possible. Extraction was performed as previously described.26 Briefly, 5 ml of each culture was harvested via vacuum filtration onto a 0.2 μm nylon membrane filter, and cells were resuspended in ice- cold HPLC-grade extraction solvent (20:20:10 acetonitrile:methanol:H2O). Samples were centrifuged at 16,100 rpm for 10 minutes, and 180 μl of supernatant was dried under a nitrogen stream. LC-MS and metabolomics data analysis was performed as previously described.26 Data analysis Data were analyzed using R version 4.3.1 and packages ggplot2 3.4.2, dplyr 1.1.2, viridis 0.6.4, grid 4.3.1, and gridExtra 2.3.27–32 RESULTS To guide our investigation into the role of Rnf and its link to diverse pathways, we first attempted to establish a link between Rnf and Nth through ferredoxin. Nth does not require ferredoxin for its activity; however, the ISC biosynthesis and repair protein cysteine desulfurase (IscS) does require reduced ferredoxin as an electron donor for ISC assembly.33,34 As an ISC protein, Nth is repaired by IscS in E. coli.35 The potential link between Rnf and IscS was promising not only as a means of connecting Rnf and Nth activity, but also for explaining the regulation of the Rnf operon by Fur in some organisms. 79 To investigate this model further experimentally, we hoped to characterize the phenotype of an rnf deletion strain. Deletion of rnf from the S. oneidensis genome was attempted multiple times using a homologous recombination method commonly used for this organism but was never successful. The inability to delete this gene suggested that Rnf is essential in S. oneidensis. To study the role of Rnf in S. oneidensis without a knockout and to test essentiality, we created IPTG-inducible CRISPRi knockdown strains using Mobile-CRISPRi.25 We generated a strain to knock down the entire rnf operon including nth, a second to knock down only nth, and a third to knock down iscS. To determine if these genes were essential, we plated the knockdown strains and controls onto LB agar plates supplemented with kanamycin and a range of IPTG concentrations. The controls included a non-targeting negative control, an essential gene control (rpoC), and a non-essential gene control (ccmF). Colony forming units were calculated to determine cell viability for each condition. The rnf and iscS knockdowns displayed similar decreases in cell viability as the essential gene control with increased IPTG concentrations, whereas IPTG had no effect on the nth-only (the last gene in the rnf operon) knockdown and negative controls (Figure 21). This result confirmed that Rnf is essential in S. oneidensis, providing explanation for our inability to generate a knockout. Additionally, IscS also appears to be essential, whereas Nth does not. The knockdown effect is also observed in minimal lactate medium liquid cultures (Figure 22). The knockdown of nth displays only a slight growth deficit compared the non-targeting control, with a 18% decrease in carrying capacity (final OD600); whereas the knockdown of rnf and iscS show substantial growth deficits with a 66% and 79% decrease in carrying capacity, respectively (Table 10). The low impact of the nth knockdown on cell viability suggests 80 Figure 21. Colony forming units (CFU) of knockdown strains plated into LB with increasing concentrations of knockdown inducer IPTG. IPTG concentrations are in mM. that Nth is not essential in S. oneidensis. This was confirmed by the ability to delete nth from the S. oneidensis genome, and the observation that the deletion had little impact on growth (Figure 23). Since the CRISPRi system interferes with transcription, protein levels only begin to drop once the system is induced and the existing protein is diluted out by cell division. This means that the severity of the knockdown effect is dependent on the number of doublings that occur in the culture. Therefore, inoculation to a lower OD600 results in a greater knockdown effect because more doublings occur (Figure 24). In 50 ml minimal lactate medium, the induced rnf knockdown strain displays a 69%, 38%, and 9% decrease in carrying capacity compared to the uninduced controls when inoculated to an initial OD600 of 0.005, 0.02, and 0.05, respectively (Table 11). Due to this observation, the knockdown strains were routinely cultivated by inoculating to an initial OD600 of 0.02 to ensure an adequate balance between phenotype effect and biomass production. 81 Figure 22. Growth curves of the induced and uninduced non-targeting, rnf, iscS, and nth knockdown strains in 200 μl of minimal lactate medium. Figure 23. Growth curves of WT S. oneidensis and Δnth. a) Cultivation of the strains in LB. b) Cultivation of the strains in minimal lactate medium. 82 Table 10. Carrying capacity, growth rates, and their percent differences from the non-targeting strain of knockdowns cultivated in 200 μl of minimal lactate medium. Knockdown strain Non-targeting Carrying Capacity (OD600) 0.409 ± 0.003 Growth Rate (h-1) 0.303 ± 0.019 Carrying Capacity % difference from non-targeting NA Growth Rate % difference from non-targeting NA 0 mM IPTG rnf knockdown 0.421 ± 0.0003 0.274 ± 0.015 iscS knockdown 0.425 ± 0.004 0.327 ± 0.025 nth knockdown 0.434 ± 0.003 0.298 ± 0.018 2.9% 3.9% 6.1% 18.5% 7.9% 1.7% Knockdown strain Non-targeting Carrying Capacity (OD600) 0.351 ± 0.002 Growth Rate (h-1) 0.428 ± 0.022 Carrying Capacity % difference from non-targeting NA Growth Rate % difference from non-targeting NA 10 mM IPTG rnf knockdown 0.118 ± 0.001 0.448 ± 0.002 iscS knockdown 0.073 ± 0.004 0.376 ± 0.020 nth knockdown 0.287 ± 0.001 0.249 ± 0.008 66.4% 79.2% 18.2% 4.7% 12.1% 41.8% 83 Figure 24. Growth curves of the induced and uninduced non-targeting, rnf, and iscS knockdown strains in 50 ml minimal lactate medium. Table 11. Carrying capacity, growth rates, and percent differences from the uninduced conditions of the rnf knockdown strain in 50 ml minimal lactate medium inoculated to different initial OD600 (OD600 i). OD600 i 0.005 0.02 0.05 Carrying Capacity (OD600) Growth Rate (h-1) 0 mM IPTG 0.685 ± 0.059 0.694 ± 0.051 0.620 ± 0.041 0.659 ± 0.415 0.367 ± 0.231 0.912 ± 3.204 10 mM IPTG OD600 i Carrying Capacity (OD600) Growth Rate (h-1) Carrying Capacity % difference from uninduced Growth Rate % difference from uninduced 0.005 0.212 ± 0.011 0.131 ± 0.031 0.02 0.05 0.430 ± 0.016 0.160 ± 0.032 0.566 ± 0.032 0.870 ± 0.831 69.1% 38.0% 8.7% 80.1% 56.4% 4.6% 84 Figure 25. a) Growth curves and b) OD-corrected ThT fluorescence of uninduced and induced non-targeting, rnf, and iscS knockdown strains. Strains cultivated in 200 μl of minimal lactate medium with 10 μM ThT. To determine the direction of Rnf activity under standard culturing conditions, we cultivated the uninduced and induced knockdown strains in minimal lactate medium with ThT, a cationic dye indicator which has been shown to have an inverse correlation to membrane potential in S. oneidensis.36 Since Rnf either contributes to or depletes ion motive force depending on the reaction direction (Figure 20), membrane potential is an indicator of which reaction Rnf was catalyzing under the tested conditions. The induced rnf and iscS knockdown strains displayed greater ThT fluorescence intensity than the rnf and iscS uninduced controls and the induced non-targeting control (Figure 25). These results indicate that the knockdown strain has less depletion of ion motive force, suggesting that Rnf uses ion motive force to power ferredoxin reduction under these conditions. The similar effect observed in the iscS knockdown suggests that Rnf and IscS are associated with the same pathway because IscS on its own does not utilize ion motive force. Because Nth activity also does not interface with ion motive force, the fact similar ThT fluorescence between the induced nth knockdown and the uninduced rnf and iscS knockdowns suggest that Nth is not directly associated with Rnf. 85 Our results suggest that Rnf in S. oneidensis uses ion motive force to generate reduced ferredoxins. To confirm whether this was the case, metabolomic analysis was conducted on the induced and uninduced non-targeting, rnf, and iscS knockdown strains. The metabolomics approach for elucidating the role of Rnf was previously described for Zymomonas mobilis and was the basis for this experiment.26 Pathways involving reduced ferredoxin as a cofactor, such as the methylerythritol phosphate (MEP) pathway, were expected to show accumulated metabolites if the Rnf knockdown was, in fact, producing reduced ferredoxins under the conditions tested. The induced and uninduced knockdown strains were inoculated to an initial OD600 of 0.035 in 50 ml of minimal lactate medium supplemented with kanamycin and grown to mid-exponential phase. A higher initial OD600 in this experiment than others was used to generate enough knockdown strain biomass for metabolites to be detected. The metabolites were extracted after 12 hours of growth and were analyzed by LC-MS (Figure 26 and 27). This analysis revealed that the most accumulated metabolite in both the induced rnf and iscS knockdown strains compared to the uninduced control was 2-C-methyl-D-erythritol-2,4-cyclophosphate (cMEPP). The induced rnf and iscS knockdowns displayed a 91.2-fold and 84.4-fold increase in cMEPP over the uninduced control, respectively. This metabolite is part of the MEP pathway and is the substrate for 4-hydroxy-3-methyl-butenyl-1-diphosphate synthase (IspG). Notably, IspG requires reduced ferredoxin as a cofactor in the reduction of cMEPP, as does IspH which catalyzes the next step in the pathway (Figure 28). In addition, four other MEP pathway metabolites before IspG were also accumulated. CDP-MEP accumulated 2.3-fold in the rnf knockdown and 1.9-fold in the iscS knockdown; CDP-ME accumulated 1.8-fold in the rnf knockdown and 1.4-fold in the iscS knockdown; and both IPP and DMAPP 86 Figure 26. Change in metabolite concentrations on a log2 scale from induced rnf knockdown cultures. Values normalized to metabolite concentrations from the induced non-targeting control. More positive numbers indicate more of that metabolite in the rnf knockdown strain than in the induced non-targeting control. 87 Figure 27. Change in metabolite concentrations on a log2 scale from induced iscS knockdown cultures. Values normalized to metabolite concentrations from the induced non-targeting control. More positive numbers indicate more of that metabolite in the iscS knockdown strain than in the induced non-targeting control. 88 accumulated 1.3-fold in the rnf knockdown and 1.0-fold in the iscS knockdown. Accumulation of MEP pathway metabolites in both the rnf knockdown and the iscS knockdown strains strongly suggest that Rnf is producingreduced ferredoxins during S. oneidensis cultivation on minimal lactate media, and that those ferredoxins are utilized by IscS for ISC biosynthesis and repair. Other notable changes in the metabolomic profiles include decreased levels of TCA cycle metabolites citrate and aconitate. Citrate levels decreased 0.8-fold (i.e. 20% decrease) in the rnf knockdown and 0.7-fold in the iscS knockdown. Aconitate levels decreased 0.9-fold in both the rnf and iscS knockdowns. The oxidized form of antioxidant glutathione (glutathione disulfide) was decreased 0.6-fold in both the rnf and the iscS Figure 28. The MEP pathway for biosynthesis of isoprenoid precursors. Isoprenoid precursors isopentyl diphosphate (IPP) and dimethylallyl diphosphate (DMAPP) are synthesized from pyruvate and glycerol-3-phosphate (G3P). Abbreviations: DXP, 1- deoxy-D-xylylose-5-phosphate; MEP, 2-C-methyl-D-erythritol-4-phosphate; CDP-ME, methylerythritol cytidyl diphosphate; CDP-MEP, 4-diphosphocytidyl-2-C-methyl-D- erythritol-2-phosphate; cMEPP, 2-C-methyl-D-erythritol-2,4-cyclodiphosphate; HMBPP, 4- hydroxy-3-methyl-butenyl-1-diphosphate; CTP, cytidine triphosphate; CDP, cytidine 5’- diphosphate; CMP, cytidine monophosphate; PPi, diphosphate; Fdred, reduced ferredoxin. 89 knockdown strains. Finally, shikimate pathway metabolites shikimate 3-phosphate was increased 27.4-fold and 28.5-fold in the rnf knockdown and the iscS knockdown strain, respectively; and 3-deoxy-D-arabinoheptulosonate 7-phosphate (DAHP) was decreased 0.7-fold in both strains. DISCUSSION In this work, we explored the function of Rnf in S. oneidensis, an organism without pathways Rnf has been proposed to be part of (e.g., nitrogen fixation, SoxR reduction), to elucidate the additional functions of this complex. In S. oneidensis grown on minimal lactate medium, Rnf uses ion motive force to drive ferredoxin reduction from NADH. We found that the rnf knockdown phenotype resembled the iscS knockdown phenotype, and that their metabolomic profiles were similar. Based on these data, we propose a comprehensive model for Rnf that can account for the diverse and seemingly unrelated functions proposed across different species. In our model, Rnf serves as a crucial source of reduced ferredoxins using ion motive force and NADH. The link between Rnf activity and pathways such as nitrogen fixation, SoxR reduction, and isoprenoid precursor biosynthesis is two-pronged. First, as a low potential electron carrier, ferredoxin serves as a necessary cofactor for many cellular reactions, including nitrogen fixation, SoxR reduction, H2 evolution, Fpr activity, ISC formation and repair, and isoprenoid precursor biosynthesis.37–43 This positions Rnf as a vital source of reduced ferredoxins for essential cellular processes (Figure 29). Secondly, since reduced ferredoxins are also an important cofactor in ISC biosynthesis and repair,33,38 the influence of Rnf activity reaches even further than the first tier of directly supplying an essential cofactor. Since many important reactions are catalyzed by ISC proteins which rely on the 90 Isc system for creation of their ISCs, the availability of reduced ferredoxins for cysteine desulfurase activity becomes essential for the activation and maintenance of ISC proteins (Figure 29). This includes many of the enzymes which use reduced ferredoxin as a co- factor in the first tier (e.g. Nif, SoxR, IspG). However, it also connects pathways with ISC enzymes which do not use reduced ferredoxin as a cofactor to Rnf, such as DNA repair, the TCA cycle, the pentose phosphate pathway, nitrate respiration, and NADH dehydrogenase.44–48 In addition to its role in ISC assembly, IscS has also been shown to directly repair oxidized ISCs in ISC proteins, such as Nth, using reduced ferredoxin as an electron donor.33,35,38 Notably, the accumulation of MEP pathway metabolites was also observed via metabolomic analysis of a Z. mobilis rnfE knockdown strain, suggesting that extending this model to include other species is rational.26 Interestingly, Rnf is not an essential protein in E. coli, as determined by the ability to insert transposons into rnf genes and generate rnf knockouts in this organism.4,49,50 This difference in essentiality of Rnf between E. coli and S. oneidensis may be due to the presence of another ISC assembly system in E. coli – the sulfur mobilization (SUF) system. While the Isc proteins are primarily responsible for ISC biosynthesis, the SUF system also participates in this activity, especially under conditions of iron limitation or oxidative stress.51 One important difference between the SUF and Isc systems is the electron donor for iron reduction. The Isc system requires reduced ferredoxin while the SUF system uses FADH2.33,34,52,53 Therefore, the non-essentiality of Rnf in E. coli could be due to a ferredoxin-independent ISC biosynthesis alternative which is not present in S. oneidensis. This is corroborated by the observation that deletion of only the Isc system or only 91 Figure 29. Proposed comprehensive model for Rnf function. Rnf generates reduced ferredoxins from NADH powered by ion motive force. Tier 1: reduced ferredoxins are used as essential cofactors in many cellular processes, including those previously though to include Rnf in their associated pathways (e.g. nitrogen fixation, SoxR reduction). Tier 2: reduced ferredoxins are essential for assembly of iron sulfur cluster (ISC)-containing proteins. Enzymatic reactions catalyzed by ISC proteins require ferredoxin for preparation of their iron-sulfur clusters, even if the proteins themselves do not use reduced ferredoxin as a cofactor. Iron-sulfur clusters are represented by two iron (red sphere) and two sulfur (yellow sphere) atoms. 92 the SUF system in E. coli allows ISC formation and viable cells.54,55 However, deletion of both the Isc and SUF systems results in a strain which is incapable of ISC biosynthesis.56 Interestingly, IspG and IspH of the MEP pathway appear to be the main restrictor of growth without ISCs.57,58 Expression of an alternative ISC-free isoprenoid precursor biosynthesis pathway restores growth of E. coli without the Isc and SUF systems.57 The absence of a SUF system in S. oneidensis further points to the relationship between Rnf and ISC biosynthesis and repair. Application of this model to observations collected in previous Rnf studies can connect seemingly unrelated functions through shared requirements for reduced ferredoxin or ISC proteins. For example, E. coli Rnf was previously shown to be involved in the reduction of the SoxR ISC cluster.4 The link between SoxR and Rnf was most heavily influenced by RnfC, which oxidizes NADH and stabilizes the ferredoxin-reducing domain RnfB. Yet, another study proposed that Fpr, a NADPH: ferredoxin oxidoreductase, was responsible for the reduction of SoxR.42 The overlapping activities of Rnf and Fpr in ferredoxin reduction allows ferredoxin to be connected to both the NADH pool via Rnf and the NADPH pool via Fpr. Our model suggests that the link between Rnf and SoxR is that ferredoxin is required for SoxR reduction. A still-unidentified mediator for SoxR reduction using reduced ferredoxin is likely required, as reduced ferredoxin alone was shown to be insufficient for reducing SoxR.59 While Rnf and Fpr do have some overlapping functions, their activities cannot always complement one another. It was demonstrated that Fpr has no influence over re- activation of oxidized aconitase, 6-phosphogluconate dehydratase, or dihydroxy acid dehydratase.60 This lack of complementation could be observed in the metabolomic 93 analysis of our knockdown strains, where citrate and aconitate were decreased by 0.8-fold and 1.0-fold, respectively in the induced rnf knockdown strain and by 0.7-fold and 0.9-fold, respectively in the iscS knockdown strain (Figure 26 and 27). Citrate and aconitate are substrates for aconitase, the second step in the TCA cycle. Aconitase is an ISC protein, which is likely impacted by decreased Rnf and IscS activity. A decrease in these metabolites rather than an accumulation such as the observed backup in the MEP pathway may be due to the greater complexity in regulation of the TCA cycle. NADPH was not detected in any sample, suggesting that it was oxidized during the extraction process. Therefore, the higher levels of NADP+ may suggest an overall increase in the NADP(H) pool size in the induced knockdown strains. The inhibition of Rnf activity in combination with the downstream inhibition of other ISC proteins or ferredoxin- dependent proteins which use NAD(P)H as a cofactor may lead to an decrease in the NAD(P)H:NAD(P)+ ratio. The increase in NADP(H) pool size could be an attempt to balance the decreasing NAD(P)H:NAD(P)+ ratio resulting from decreased Rnf and IscS activities. Additionally, the NAD+ and NADP+ pools are connected by NAD+ kinase, which may explain the observation of greater NADP(H) pool size even though NAD(H) was not detected. The lower NADPH/NADP+ ratio suggests a more oxidized cellular environment. A shift in the cellular redox balance can alter metabolic pathways and signaling, which can have serious impacts on cellular health.61 Rnf appears to be integral to maintaining redox homeostasis by providing necessary low potential electron carriers. As such, the severe growth impact of the induced rnf knockdown is the culmination of the disruption of redox balance and essential metabolic pathways. Further evidence for a disruption to the cellular redox balance was observed with 94 the decreased levels of glutathione disulfide in both the rnf and iscS knockdown strains. Glutathione disulfide is the oxidized form of antioxidant glutathione, which is an important contributor to maintaining cellular redox balance.62,63 Glutathione reacts with free radicals to form a glutathione radical, which can react with a second glutathione radical to produce glutathione disulfide.63 Altered glutathione disulfide levels indicate a shift in the cellular redox balance, which may have additional impacts on disulfide bonds.64 The second-most accumulated metabolite in both the induced rnf and iscS knockdown strains was shikimate 3-phosphate. This metabolite is part of the shikimate pathway for aromatic amino acid precursor biosynthesis.65 This pathway is not comprised of ISC enzymes nor enzymes using ferredoxin as a cofactor. While the mechanism for a backup in the shikimate pathway as a result of decreased Rnf or IscS activity is not immediately clear, the observation highlights the reach of Rnf activity. Likely, regulatory changes occur due to decreased ferredoxin availability and/or decreased ISC biosynthesis and repair which interfere with essential pathways such as aromatic amino acid biosynthesis. This novel, comprehensive model for Rnf function connects previous explanations of its activity to the broad impact that Rnf has across many cellular functions. Reduced ferredoxins generated by Rnf go on to power essential pathways like isoprenoid precursor biosynthesis and ISC assembly and repair. Disruption of ISC assembly exponentially interferes with other essential processes requiring ISC proteins like those involved in DNA repair, the TCA cycle, and the pentose phosphate pathway. The broad impacts of Rnf activity are central to its classification as an essential protein and explain its effect on reactions such as nitrogen fixation and SoxR reduction. 95 REFERENCES 1. 2. 3. 4. Schmehl, M. et al. Identification of a new class of nitrogen fixation genes in Rhodobacter capsalatus: a putative membrane complex involved in electron transport to nitrogenase. Mol Gen Genet 241–241, 602–615 (1993). Jouanneau, Y., Jeong, H., Hugo, N., Meyer, C. & Willison, J. C. Overexpression in Escherichia coli of the rnf genes from Rhodobacter capsulatus. Eur J Biochem 251, 54–64 (1998). Biegel, E., Schmidt, S., González, J. M. & Müller, V. Biochemistry, evolution and physiological function of the Rnf complex, a novel ion-motive electron transport complex in prokaryotes. CMLS 68, 613–634 (2011). Koo, M. S. et al. A reducing system of the superoxide sensor SoxR in Escherichia coli. EMBO J 22, 2614–2622 (2003). 5. Müller, V., Imkamp, F., Biegel, E., Schmidt, S. & Dilling, S. Discovery of a Ferredoxin:NAD + ‐Oxidoreductase (Rnf) in Acetobacterium woodii. Ann N Y Acad Sci 1125, 137–146 (2008). 6. Kuhns, M., Trifunović, D., Huber, H. & Müller, V. The Rnf complex is a Na+ coupled respiratory enzyme in a fermenting bacterium, Thermotoga maritima. Commun Biol 3, 431 (2020). 7. Westphal, L., Wiechmann, A., Baker, J., Minton, N. P. & Müller, V. The Rnf Complex Is an Energy-Coupled Transhydrogenase Essential To Reversibly Link Cellular NADH and Ferredoxin Pools in the Acetogen Acetobacterium woodii. J Bacteriol 200, (2018). 8. 9. Hayashi, K. et al. Highly accurate genome sequences of Escherichia coli K‐12 strains MG1655 and W3110. Mol Syst Biol 2, (2006). Heidelberg, J. F. et al. Genome sequence of the dissimilatory metal ion–reducing bacterium Shewanella oneidensis. Nat Biotechnol 20, 1118–1123 (2002). 10. Berges, M. et al. Iron Regulation in Clostridioides difficile. Front Microbiol 9, (2018). 11. Hernandez, M. E. & Newman, D. K. Extracellular electron transfer. CMLS 58, 1562– 1571 (2001). 12. Beblawy, S. et al. Extracellular reduction of solid electron acceptors by Shewanella oneidensis. Mol Microbiol 109, 571–583 (2018). 13. Shi, L., Squier, T. C., Zachara, J. M. & Fredrickson, J. K. Respiration of metal (hydr)oxides by Shewanella and Geobacter : a key role for multihaem c ‐type cytochromes. Mol Microbiol 65, 12–20 (2007). 96 14. Gralnick, J. A., Vali, H., Lies, D. P. & Newman, D. K. Extracellular respiration of dimethyl sulfoxide by Shewanella oneidensis strain MR-1. PNAS 103, 4669–4674 (2006). 15. Carpentier, W., De Smet, L., Van Beeumen, J. & Brigé, A. Respiration and Growth of Shewanella oneidensis MR-1 Using Vanadate as the Sole Electron Acceptor. J Bacteriol 187, 3293–3301 (2005). 16. Chen, Y. & Wang, F. Insights on nitrate respiration by Shewanella. Front Mar Sci 1, (2015). 17. Shi, L. et al. Molecular Underpinnings of Fe(III) Oxide Reduction by Shewanella oneidensis MR-1. Front Microbiol 3, (2012). 18. Madsen, C. S. & TerAvest, M. A. NADH dehydrogenases Nuo and Nqr1 contribute in transfer by Shewanella oneidensis MR-1 to extracellular electron bioelectrochemical systems. Sci Rep 9, 14959 (2019). 19. Marritt, S. J. et al. The roles of CymA in support of the respiratory flexibility of Shewanella oneidensis MR-1. Biochem Soc Trans 40, 1217–1221 (2012). 20. Coursolle, D. & Gralnick, J. A. Modularity of the Mtr respiratory pathway of Shewanella oneidensis strain MR‐1. Mol Microbiol 77, 995–1008 (2010). 21. Tiedje, J. M. Shewanella—the environmentally versatile genome. Nat Biotechnol 20, 1093–1094 (2002). 22. Zhu, D. D., Liu, J. L. & Qiao, S. Z. Recent Advances in Inorganic Heterogeneous Electrocatalysts for Reduction of Carbon Dioxide. Adv Mater 28, 3423–3452 (2016). 23. Tefft, N. M. & Teravest, M. A. Reversing an Extracellular Electron Transfer Pathway for Electrode-Driven Acetoin Reduction. ACS Synth Biol 8, 1590–1600 (2019). 24. Ford, K. C., Kaste, J. A. M., Shachar-Hill, Y. & Teravest, M. A. Flux-Balance Analysis and Mobile CRISPRi-Guided Deletion of a Conditionally Essential Gene in Shewanella oneidensis MR-1. ACS Synth Biol 11, 3405–3413 (2022). 25. Peters, J. M. et al. Enabling genetic analysis of diverse bacteria with Mobile- CRISPRi. Nat Microbiol 2019 4:2 4, 244–250 (2019). 26. Enright, A. L. et al. The genetics of aerotolerant growth in an alphaproteobacterium with a naturally reduced genome. mBio 14, (2023). 27. R Core Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing Preprint at https://www.R-project.org/ (2021). 28. H. Wickham. ggplot2: Elegant Graphics for Data Analysis. Preprint at (2016). 97 29. Hadley Wickham, Romain François, Lionel Henry & Kirill Müller. dplyr: A Grammar of Data Manipulation. R package version 1.0.9 Preprint at https://CRAN.R- project.org/package=dplyr (2022). 30. Garnier, S. et al. viridis(Lite) - Colorblind-Friendly Color Maps for R. (2023). 31. Ulrich, J. TTR: Technical Trading Rules. (2021). 32. Auguie, B. gridExtra: Miscellaneous Functions for “Grid” Graphics. Preprint at (2017). 33. Kim, J. H., Frederick, R. O., Reinen, N. M., Troupis, A. T. & Markley, J. L. [2Fe-2S]- Ferredoxin Binds Directly to Cysteine Desulfurase and Supplies an Electron for Iron– Sulfur Cluster Assembly but Is Displaced by the Scaffold Protein or Bacterial Frataxin. J Am Chem Soc 135, 8117–8120 (2013). 34. Takahashi, Y. & Nakamura, M. Functional Assignment of the ORF2-iscS-iscU-iscA- hscB-hscA-fdx-0RF3 Gene Cluster Involved in the Assembly of Fe-S Clusters in Escherichia coli. J Biochem 126, 917–926 (1999). 35. Rogers, P. A., Eide, L., Klungland, A. & Ding, H. Reversible inactivation of E. coli endonuclease III via modification of its [4Fe-4S] cluster by nitric oxide. DNA Repair (Amst) 2, 809–817 (2003). 36. Pirbadian, S., Chavez, M. S. & El-Naggar, M. Y. Spatiotemporal mapping of bacterial membrane potential responses to extracellular electron transfer. PNAS 117, 20171– 20179 (2020). 37. Addison, H., Glatter, T., K. A. Hochberg, G. & Rebelein, J. G. Two distinct ferredoxins are essential for nitrogen fixation by the iron nitrogenase in Rhodobacter capsulatus. mBio (2024) doi:10.1128/mbio.03314-23. 38. Yan, R. et al. Ferredoxin Competes with Bacterial Frataxin in Binding to the Desulfurase IscS*. J Biol Chem 288, 24777–24787 (2013). 39. Srour, B., Gervason, S., Monfort, B. & D’Autréaux, B. Mechanism of Iron–Sulfur Cluster Assembly: In the Intimacy of Iron and Sulfur Encounter. Inorganics (Basel) 8, 55 (2020). 40. Zhou, J., Yang, L., Wang, C., Choi, E.-S. & Kim, S.-W. Enhanced performance of the methylerythritol phosphate pathway by manipulation of redox reactions relevant to IspC, IspG, and IspH. J Biotechnol 248, 1–8 (2017). 41. Buckel, W. & Thauer, R. K. Flavin-Based Electron Bifurcation, Ferredoxin, Flavodoxin, and Anaerobic Respiration With Protons (Ech) or NAD+ (Rnf) as Electron Acceptors: A Historical Review. Front Microbiol 9, (2018). 98 42. Liochev, S. I., Hausladen, A., Beyer, W. F. & Fridovich, I. NADPH: ferredoxin oxidoreductase acts as a paraquatdiaphorase and is a member of the soxRS regulon. PNAS 91, 1328–1331 (1994). 43. Carrillo, N. & Ceccarelli, E. A. Open questions in ferredoxin‐NADP + reductase catalytic mechanism. Eur J Biochem 270, 1900–1915 (2003). 44. Simpson, P. J. L., Richardson, D. J. & Codd, R. The periplasmic nitrate reductase in Shewanella: the resolution, distribution and functional implications of two NAP isoforms, NapEDABC and NapDAGHB. Microbiology (N Y) 156, 302–312 (2010). 45. Sled’, V. D. et al. Bacterial NADH-quinone oxidoreductases: Iron-sulfur clusters and related problems. J Bioenerg Biomembr 25, 347–356 (1993). 46. Gardner, P. R. & Fridovich, I. Superoxide sensitivity of the Escherichia coli 6- phosphogluconate dehydratase. J Biol Chem 266, 1478–1483 (1991). 47. Kennedy, M. C. & Stout, C. D. Aconitase: An Iron—Sulfur Enzyme. Prog Inorg Chem 38, 323–339 (1992). 48. Fu, W., O’Handley, S., Cunningham, R. P. & Johnson, M. K. The role of the iron- sulfur cluster in Escherichia coli endonuclease III. A resonance Raman study. J Biol Chem 267, 16135–16137 (1992). 49. Baba, T. et al. Construction of Escherichia coli K‐12 in‐frame, single‐gene knockout mutants: the Keio collection. Mol Syst Biol 2, (2006). 50. Goodall, E. C. A. et al. The Essential Genome of Escherichia coli K-12. mBio 9, (2018). 51. Pérard, J. & Ollagnier de Choudens, S. Iron–sulfur clusters biogenesis by the SUF machinery: close to the molecular mechanism understanding. J Biol Inorg Chem 23, 581–596 (2018). 52. Wollers, S. et al. Iron-Sulfur (Fe-S) Cluster Assembly. J Biol Chem 285, 23331– 23341 (2010). 53. Saini, A., Mapolelo, D. T., Chahal, H. K., Johnson, M. K. & Outten, F. W. SufD and SufC ATPase Activity Are Required for Iron Acquisition during in Vivo Fe-S Cluster Formation on SufB. Biochemistry 49, 9402–9412 (2010). 54. Tokumoto, U. & Takahashi, Y. Genetic Analysis of the isc Operon in Escherichia coli Involved in the Biogenesis of Cellular Iron-Sulfur Proteins. J Biochem 130, 63–71 (2001). 99 55. Outten, F. W., Djaman, O. & Storz, G. A suf operon requirement for Fe–S cluster assembly during iron starvation in Escherichia coli. Mol Microbiol 52, 861–872 (2004). 56. Tokumoto, U. Interchangeability and Distinct Properties of Bacterial Fe-S Cluster Assembly Systems: Functional Replacement of the isc and suf Operons in Escherichia coli with the nifSU-Like Operon from Helicobacter pylori. J Biochem 136, 199–209 (2004). 57. Tanaka, N. et al. Novel features of the ISC machinery revealed by characterization of Escherichia coli mutants that survive without iron–sulfur clusters. Mol Microbiol 99, 835–848 (2016). 58. Esquilin-Lebron, K., Dubrac, S., Barras, F. & Boyd, J. M. Bacterial Approaches for Assembling Iron-Sulfur Proteins. mBio 12, (2021). 59. Gaudu, P. & Weiss, B. SoxR, a [2Fe-2S] transcription factor, is active only in its oxidized form. PNAS 93, 10094–10098 (1996). 60. Krapp, A. R. et al. The Flavoenzyme Ferredoxin (Flavodoxin)-NADP(H) Reductase Modulates NADP(H) Homeostasis during the soxRS Response of Escherichia coli. J Bacteriol 184, 1474–1480 (2002). 61. Sporer, A. J., Kahl, L. J., Price-Whelan, A. & Dietrich, L. E. P. Redox-Based Regulation of Bacterial Development and Behavior. Annu Rev Biochem 86, 777– 797 (2017). 62. Knoke, L. R. et al. The role of glutathione in periplasmic redox homeostasis and oxidative protein folding in Escherichia coli. Redox Biol 64, 102800 (2023). 63. Smirnova, G. V. & Oktyabrsky, O. N. Glutathione in Bacteria. Biochemistry (Moscow) 70, 1199–1211 (2005). 64. Vašková, J., Kočan, L., Vaško, L. & Perjési, P. Glutathione-Related Enzymes and Proteins: A Review. Molecules 28, 1447 (2023). 65. Herrmann, K. M. & Weaver, L. M. The Shikimate Pathway. Annu Rev Plant Physiol Plant Mol Biol 50, 473–503 (1999). 100 CHAPTER 4: A COMMON INDUCER MOLECULE ENHANCES SUGAR UTILIZATION BY SHEWANELLA ONEIDENSIS MR-1 This chapter is adapted from a publication in The Journal of Industrial Microbiology & Biotechnology.1 101 PREFACE As part of my investigation into the role of Rnf in S. oneidensis, I had attempted to find culturing conditions in which I could generate knockouts of the rnf operon and iscS. During this process I discovered that cultivating the rnf and iscS knockdown strains on sugar substrate N-acetylglucosamine improved their carrying capacity over cultivation on lactate. While at first this result was exciting (maybe I could use this condition for generating the difficult knockouts), it quickly became concerning when the non-targeting control also showed improved growth on the substrate. Was the dCas9 or the sgRNA somehow improving the fitness of cells growing on N-acetylglucosamine? The CRISPRi components turned out to play no part in the phenotype, because we soon discovered that wild-type S. oneidensis also displayed enhanced growth on N- acetylglucosamine when cultivated with the same concentration of IPTG used for induction of the CRIPSRi system. This was an unexpected observation that we were interested in exploring further – leading to this study on the effect of IPTG on S. oneidensis growth on N-acetylglucosamine. This work was published in August 2023 in The Journal of Industrial Microbiology and Biotechnology in the special issue on “Frontiers of Industrial Microbiology and Biotechnology”.1 ABSTRACT Shewanella oneidensis MR-1 is an electroactive bacterium that is a promising host for bioelectrochemical technologies, which makes it a common target for genetic engineering, including gene deletions and expression of heterologous pathways. Expression of heterologous genes and gene knockdown via CRISPRi in S. oneidensis are 102 both frequently induced by isopropyl β-D-1-thiogalactopyranoside (IPTG), a commonly used inducer molecule across many model organisms. Here, we report and characterize an unexpected phenotype; IPTG enhances the growth of wild-type S. oneidensis MR-1 on the sugar substrate N-acetylglucosamine. IPTG improves the carrying capacity of S. oneidensis growing on N-acetylglucosamine while the growth rate remains similar to cultures without the inducer. Extracellular acetate accumulates faster and to a higher concentration in cultures without IPTG than those with it. IPTG appears to improve acetate metabolism which combats the negative effect that acetate accumulation has on the growth of S. oneidensis with N-acetylglucosamine. We recommend using extensive experimental controls and careful data interpretation when using both N- acetylglucosamine and IPTG in S. oneidensis cultures. INTRODUCTION Shewanella oneidensis MR-1 is a facultatively anaerobic bacterium with great potential as a host for bioelectrochemical technologies due to its well characterized extracellular electron transfer pathway and diverse genetic toolbox.2,3 Some of these technologies include using S. oneidensis as a biosensor for arsenic or toxic water4,5 and for electricity generation from waste.6,7 However, S. oneidensis natively has limited options for carbon substrates and does not naturally accumulate high-value compounds.8 These limitations make pathway engineering a crucial component of this organism’s expanded utility in bioelectrochemical systems. Sugar metabolism is particularly limited in S. oneidensis, with several essential genes for sugar utilization missing or nonfunctional. For sugar transport, the S. oneidensis MR-1 genome encodes one set of phosphotransferase system genes for an unknown 103 substrate, but phosphoenolpyruvate is not sufficiently produced to sustain the energy requirements for the system to function.9,10 Additionally, the genome does not encode an annotated glucose permease, glucose kinase, or the important glycolysis enzyme 6- phosphofructokinase, which prevents S. oneidensis from utilizing most six carbon sugars.10,11 N-acetylglucosamine (NAG) is a rare exception to this rule, as S. oneidensis MR-1 converts NAG to fructose 6-phosphate with a NAG permease, NAG kinase, NAG 6- phosphate deacetylase, and glucosamine 6-phosphate deaminase.12 With a limited pool of compounds which can be utilized for carbon and energy, S. oneidensis and its engineered strains are typically grown using lactate, pyruvate, acetate, or NAG as substrates. Another prominent feature of S. oneidensis metabolism is acetate accumulation. Under anoxic conditions, S. oneidensis MR-1 excretes high levels of acetate because of limited TCA cycle activity.13 Under oxic conditions, acetate accumulation and subsequent uptake are also observed in S. oneidensis cultures growing on NAG or lactate.14,15 Acetate is a common product of overflow metabolism, and S. oneidensis begins to use the excreted acetate as a secondary substrate via the glyoxylate shunt when the main substrate is depleted.15 The accumulation of acetate is usually accompanied with a drop in medium pH which can affect the health of the cells.16 In S. oneidensis, acetate is produced from acetyl-CoA and vice versa by the combined, reversible activities of acetate kinase and phosphate acetyltransferase; additionally, acetyl-CoA can be produced from acetate via acetyl-coenzyme A synthetase.17 To improve S. oneidensis as a host for use in bioelectrochemical systems, expansion of substrate utilization and product scope are critical. Gene expression under 104 the control of an inducible promoter is one of the most common methods for experimenting with both native and non-native pathways in vivo. One tool that uses IPTG-inducible expression for studying gene function and for pathway engineering in S. oneidensis is the Mobile CRISPRi knockdown system.18 Promoters inducible by IPTG are utilized for both protein synthesis via common S. oneidensis-compatible vectors and CRISPRi knockdown systems. For example, Ford et al. deleted a gene previously thought to be essential in S. oneidensis (gpmA) by uncovering optimal knockout conditions using an IPTG-inducible CRISPRi system.19 S. oneidensis has been engineered to utilize substrates such as xylose and glucose via IPTG-inducible systems.20,21 Improved extracellular electron transfer in S. oneidensis has been achieved through various IPTG-inducible methods including using two types of transcriptional logic gates to fine-tune expression of extracellular electron transfer genes,22 expression of NAD+ biosynthesis genes for increased intracellular NAD(H) levels,23 expression of a synthetic flavin pathway from Bacillus subtilis to increase secreted flavins,24 and expression of diguanylate cyclase from Escherichia coli for enhanced S. oneidensis biofilm formation.25 IPTG-inducible systems have also been used for production of 5-aminolevulinic acid26 and n-butanol27 in S. oneidensis, increasing the organism’s potential product scope. The use of IPTG for induction of heterologous genes is widespread due to the efficiency and simplicity of the system. The synthetic inducer is used to regulate the lac repressor (LacI) which regulates the lac operon in E. coli. Natively, the lac operon is expressed when lactose is present to produce lactose metabolic genes β-galactosidase, lactose permease, and a transacetylase.28,29 Lactose is detected indirectly by its isomer allolactose which can bind to LacI.28 LacI binds to the lac operator when allolactose 105 concentrations are low, which physically blocks transcription of the lac operon. Allolactose induces transcription of the lac operon by binding to LacI and releasing it from the DNA.30 IPTG can be used as a replacement for allolactose since it isn’t hydrolyzed by cells, making it a gratuitous inducer whose concentration does not change during the course of the experiment.29 When used in the context of heterologous gene expression, IPTG induces expression of a synthetic operon which is regulated by LacI. IPTG is an indispensable tool for genetic engineering of many organisms, S. oneidensis included. In this work, we characterize the unexpected phenotype of wild-type S. oneidensis MR-1 in which growth on the sugar substrate NAG is enhanced by the presence of IPTG and advise circumspection when using NAG and IPTG together in evaluation of engineered S. oneidensis strains. RESULTS IPTG-enhanced growth of wild-type S. oneidensis was discovered during experimentation with CRISPRi knockdown strains. With this system, a small guide RNA is engineered to target a gene of interest. The small guide RNA forms a complex with a catalytically inactive Cas9 which binds to the target gene and blocks its transcription.18 We observed that when grown on NAG, inducing the non-targeting control plasmid (pJMP2846) of the CRISPRi system with IPTG improved growth. To determine if this was due to the induction of the non-targeting control, we cultured wild-type S. oneidensis harboring pJMP2846 on 20 mM NAG with and without 10 mM IPTG and compared it to wild-type S. oneidensis without the plasmid under the same conditions (Figure 30). We found that the improvement to growth was not exclusive to the CRISPRi non-targeting strain and explored the phenotype further in wild-type S. oneidensis carrying no heterologous plasmid. 106 Figure 30. Growth of wild-type S. oneidensis and S. oneidensis CRISPRi non- targeting strain (MR-1 pJMP2846). Strains cultivated in 200 µl minimal medium with 20 mM NAG with or without 10 mM IPTG. Y-axis is on a logarithmic scale. Lines represent the average of three biological replicates and transparent ribbons represent standard error. To determine if IPTG influenced S. oneidensis growth on various substrates, the wild-type strain was cultivated with 20 mM of NAG, lactate, or acetate with and without 10 mM IPTG in 96-well plates (Figure 31a). When S. oneidensis was grown on 20 mM NAG, the carrying capacity (final OD600) nearly doubled from 0.372 ± 0.002 without IPTG to 0.648 ± 0.004 with IPTG while the growth rate remained similar at 0.181 ± 0.005 h-1 without IPTG and 0.168 ± 0.005 h-1 with IPTG (Figure 31a, Table 12). IPTG slightly decreased the carrying capacity during growth on lactate but did not significantly affect the growth rate of MR-1 on either lactate or acetate (Table 12). To determine whether the carrying capacity increased because S. oneidensis can consume IPTG as a carbon substrate, the wild-type strain was cultivated with 10 mM IPTG as the only available carbon source. S. oneidensis was incapable of growth on IPTG alone. Additionally, S. oneidensis cultivated on 30 mM NAG did not share the same growth phenotype as when cultivated with 20 mM NAG and 10 mM IPTG (Figure 31b), indicating 107 Figure 31. Growth of wild-type S. oneidensis in various conditions. (a) Growth of S. oneidensis MR-1 cultivated on 20 mM of different carbon substrates with or without 10 mM IPTG in 200 µl M5 minimal medium in 96-well plates. (b) S. oneidensis MR-1 cultivated on NAG and/or IPTG in 200 µl M5 minimal medium. Y-axis is on a logarithmic scale. Lines represent the average of three biological replicates and transparent ribbons represent standard error. Table 12. Growth rate and carrying capacity of S. oneidensis strains depicted in Figure 31a. Substrate Growth Rate (h-1) Carrying capacity (OD600) - IPTG + IPTG - IPTG + IPTG NAG 0.181 ± 0.005 0.168 ± 0.005 0.372 ± 0.002 0.648 ± 0.004 Lactate 0.250 ± 0.01 0.266 ± 0.028 0.447 ± 0.003 0.322 ± 0.005 Acetate 0.120 ± 0.008 0.193 ± 0.017 0.175 ± 0.005 0.165 ± 0.004 108 that changes in osmotic pressure were not responsible for the increased carrying capacity. This observation also indicates that IPTG does not act as a secondary substrate, demonstrated by the lack of carrying capacity increase by 30 mM NAG as compared to the increase in carrying capacity of 20 mM NAG and 10 mM IPTG. We next investigated whether IPTG influenced metabolic product formation in S. oneidensis grown on NAG. S. oneidensis was cultivated in 50 ml of minimal medium in 250-ml flasks with 20 mM NAG either with or without 10 mM IPTG (Figure 32a). We observed a similar increase in carrying capacity in the presence of IPTG, although the effect was diminished in flasks versus in 96-well plates. We confirmed by microscopy that cell size was the same in both conditions, indicating that the difference in OD600 reflects a difference in cell density (Figure 33). Samples from these cultures were analyzed via HPLC to detect changes in metabolic products. In cultures with IPTG, acetate accumulation was minimal and was quickly consumed (Figure 32a). NAG and IPTG co- eluted in our HPLC method and their peaks could not be resolved. However, no NAG/IPTG peak was detected after 18 hours in the no IPTG condition, and the peak area was equal to that of 10 mM IPTG in the IPTG-containing cultures after 16 hours. Further, we did not detect IPTG degradation as measured by galactose levels in the media. This indicates complete utilization of NAG both with and without IPTG. The pH trends of these cultures were consistent with the observed acetate concentrations, with the pH dropping when acetate concentrations were high (Figure 32a). We also investigated the effect of IPTG on cultures grown with 40 mM D,L-lactate and found that IPTG had no effect (Figure 34). 109 Figure 32. MR-1 cultures in 50 ml M5 minimal medium with and without 10 mM IPTG in varying conditions. a) 20 mM NAG growth curves, acetate concentrations, and pH. b) 20 mM NAG + 10X buffer growth curves, acetate concentrations, and pH. c) 10 mM NAG growth curves, acetate concentrations, and pH. Lines represent the average of three biological replicates and error bars represent standard error. Y-axis of growth curves are on a logarithmic scale. 110 Figure 33. Comparison of S. oneidensis MR-1 cell size during NAG cultivation with or without IPTG. a) Growth curve of S. oneidensis MR-1 in 20 mM NAG with or without 10 mM IPTG in 50 ml cultures. b) Average cell area of S. oneidensis MR-1 at various times during the growth curve. +/- IPTG values at each timepoint are not significantly different (t-test, p > 0.05). The acidification of the cultures and accumulation of acetate could be prevented when S. oneidensis was cultivated in the same conditions except with a 10X buffering capacity (Figure 32b) or with half the original concentration of NAG (Figure 32c). The difference in carrying capacity between +/- IPTG conditions was reduced with higher buffering capacity or lower starting NAG concentrations (Table 13), although acetate accumulation remained lower in the cultures with IPTG. This observation indicates that IPTG still significantly alters metabolism of S. oneidensis even when the growth appears unaffected. NAG was also completely utilized in cultures with either 10 mM NAG or with 10X buffering capacity, and no IPTG was detected. Taken together, the trends of growth, acetate accumulation, and culture pH suggest that acetate accumulation stalls growth of S. oneidensis and that IPTG counteracts the negative effects of acetate accumulation in the cultures. 111 Figure 34. 50 ml a) growth curves and b) metabolite analysis of S. oneidensis MR-1 on 40 mM D,L-lactate with or without 10 mM IPTG. Table 13. Growth rate and carrying capacity of S. oneidensis growth depicted in Figure 34. Strains cultivated in 50 mL M5 minimal medium. Values were calculated from the growth curves using the R package “growthCurver” using default values. Growth Rate (h-1) Carrying Capacity (OD600) -IPTG +IPTG -IPTG +IPTG 0.393 ± 0.062 0.560 ± 0.045 2.644 ± 0.162 3.345 ± 0.082 0.956 ± 0.663 1.034 ± 0.553 2.257 ± 0.255 2.072 ± 0.183 1.658 ± 0.085 1.759 ± 0.11 0.633 ± 0.163 1.326 ± 0.978 Conditions 20 mM NAG 1X buffering capacity (Figure 34a) 20 mM NAG 10X buffering capacity (Figure 34b) 20 mM NAG 1X buffering capacity (Figure 34c) 112 Figure 35. MR-1 growth curves in 200 µl M5 minimal medium with 20 mM NAG and varying IPTG concentrations. Y-axis is on a logarithmic scale. Lines represent the average of three biological replicates and transparent ribbons represent standard error. IPTG at concentrations of 1.0 and 10 mM significantly increased the carrying capacity (p ≤ 0.05) for cultures with 20 mM NAG in 200 µl cultures in a 96-well plate (Figure 35). However, IPTG concentrations of 1 mM or less did not enhance MR-1 growth on 20 mM NAG in 50-ml cultures (Figure 36). Since IPTG is similar in structure to NAG (Figure 37a), we hypothesized that IPTG induced expression of NAG metabolic genes. The key genes involved in metabolism of NAG are controlled by the NagR repressor, which is released from the operator site in the presence of NAG.31 Deletion of nagR in S. oneidensis results in a strain that is capable of growth on glucose due to the unregulated, constitutive expression of NAG permease and NAG kinase which can perform the functions of the missing glucose permease and kinase to some degree.11 To determine whether IPTG induced the NagR regulon, we cultivated S. oneidensis in minimal medium with 20 mM D-glucose and varying concentrations of IPTG. Our results show that higher IPTG concentrations diminish the ability for S. oneidensis to grow on glucose, suggesting that IPTG did not induce the NagR regulon 113 Figure 36. S. oneidensis growth curves in 50 ml M5 minimal medium with 20 mM NAG and varying IPTG concentrations. Y-axis is on a logarithmic scale. Lines represent the average of three biological replicates and error bars represent standard error. Figure 37. (a) The chemical structures of IPTG and NAG. Their structures differ in the highlighted regions. (b) Growth curves of S. oneidensis growth on 20 mM glucose with varying concentrations of IPTG. Y-axis is on a logarithmic scale. Lines represent the average of three biological replicates and error bars represent standard error. 114 (Figure 37b). S. oneidensis cannot natively grow on glucose; however, the ability to utilize glucose evolves quickly because only a single mutation is required for the phenotype to arise.11 This accounts for the growth observed on glucose after prolonged culturing. DISCUSSION Our results indicate that IPTG reduced acetate accumulation by MR-1 during growth on NAG. Rapid accumulation of acetate in the absence of IPTG was accompanied by a drop in culture pH. Acetate accumulation in E. coli cultures inhibits growth by causing decreased expression of TCA cycle and glyoxylate shunt genes.16 Acetate accumulation also appears to impact growth of S. oneidensis in our experiments, although it is not known whether acetate decreases expression of TCA cycle genes in this organism. Higher buffering capacity and lower NAG concentrations both reduced acetate accumulation to a rate that did not inhibit growth. With more buffering capacity, the pH presumably did not inhibit the expression of TCA cycle or glyoxylate shunt genes to the same extent as in cultures with the standard buffering capacity, thus allowing acetate to be more quickly metabolized. Lower NAG concentrations similarly prevent such a notable drop in pH; however, in this case it is due to less accumulation of acetate. The presence of IPTG in both cases displayed a compounding effect in which the acetate accumulation was even lower. These data point to a role of IPTG in improving acetate metabolism in S. oneidensis MR-1, perhaps through the induction of TCA cycle and/or glyoxylate shunt genes. Although we initially hypothesized that IPTG may interact with the native NagR regulator, another possibility is that it induces the HexR regulon, which includes genes of the glyoxylate shunt. Since the glyoxylate shunt is responsible for assimilating acetate as a carbon and energy source, its upregulation could increase the rate of acetate 115 assimilation.15 Although outside the scope of this study, future work could explore the influence of IPTG on the HexR regulon and interaction with the HexR protein. High-throughput experimentation with induction of heterologous genes in a 96-well plate is a common practice. Using this culturing method, we found that as little as 1 mM IPTG can cause a significant effect on growth. Higher IPTG concentrations of 10 mM were required for the effect to be seen in 250-ml shake flasks. While this is a high concentration of the inducer, it remains within the standard working range. The difference in required IPTG to observe the enhanced growth between these two culturing methods may be caused by oxygen availability. Larger cultures with more shaking are more efficiently aerated,32 and in S. oneidensis a higher availability of oxygen leads to a decrease in acetate production and higher carbon flux toward the TCA cycle.33 The differences in S. oneidensis metabolism caused by oxygen levels suggest that in the smaller cultures, acetate accumulates to an even greater level than we report from the shake flasks, potentially explaining the greater impact that IPTG has on growth in these conditions. Unexpectedly, we found that at the 200 µl scale, S. oneidensis grown on 20 mM NAG and or on 20 mM lactate without IPTG have similar final cell density, despite the greater size of the NAG molecule. We speculate that this was due to incomplete NAG utilization under this condition, although we could not measure NAG because of the small culture size. As culture size decreases, aeration and oxygen availability also decrease.32 Oxygen limitation reduces the substrate utilization of S. oneidensis.38 At the 50 ml scale, we observe the expected growth yield difference between S. oneidensis growth on NAG vs D,L-lactate (Figure 34). 116 To our knowledge, the phenotype of wild-type S. oneidensis displaying enhanced growth on NAG with IPTG has not been reported in the scientific literature. The enhanced growth of wild-type S. oneidensis on NAG when cultivated with IPTG is an important observation to be mindful of when testing inducible systems in this organism because of this inducer’s capability of altering S. oneidensis metabolism. The impact of this was showcased by our discovery of the phenotype during experimentation with S. oneidensis CRISPRi control strains. This observation may also be buried in the literature and could raise questions regarding results without thorough controls. For example, Jeon et al. report optimal n-butanol production by an engineered strain of S. oneidensis on 2% NAG, 0.3% butyrate and 0.1 mM IPTG under microaerobic conditions.27 This strain overexpresses recombinant proteins CoA transferase, acetyl-CoA synthase, and alcohol dehydrogenase from an IPTG-inducible promoter. Since the carbon source was NAG, it is unclear whether the IPTG aided in n-butanol production because of expression of these proteins, or if it was decreasing acetate accumulation in the media. The n-butanol-producing strain generated the highest titers under microaerobic conditions. Our data suggest less aeration (therefore lower oxygen concentrations) exacerbates the effect of IPTG on NAG-cultivated S. oneidensis. So, one could speculate that this effect is responsible for the higher production of n-butanol under microaerobic conditions. We encourage researchers working with S. oneidensis to be mindful of these observations during their efforts to engineer this organism so that false positive correlations can be avoided. 117 METHODS Bacterial strains and culturing conditions Wild-type Shewanella oneidensis MR-1 was used for all experiments in this study. Plasmid pJMP2846 was used as the CRISPRi non-targeting control.18 S. oneidensis was grown at 30 °C with shaking in M5 minimal medium (1.29 mM K2HPO4, 1.65 mM KH2PO4, 7.87 mM NaCl, 1.70 mM NH4SO4, 475 μM MgSO4·7H2O, 10 mM HEPES, 0.01% (w/v) casamino acids, 1× Wolfe’s mineral solution, and 1× Wolfe’s vitamin solution without riboflavin, pH 7.2). For cultures with 10X buffering capacity, HEPES concentration was increased to 100 mM. M5 was supplemented with either sodium D,L-lactate, NAG, or sodium acetate to final concentrations of 20 mM unless otherwise stated. IPTG was added to cultures at the specified concentrations. Growth experiments were conducted in either 96-well plates containing 200 µl of media or 250-ml shake flasks containing 50 ml of media. Culture volumes are specified for each experiment. OD600 and pH Measurement OD600 was measured every 15-30 minutes using an H1M BioTek Plate Reader for 200 µl cultures. OD600 of 50-ml cultures were measured on an Eppendorf BioSpectrometer using a 1 ml culture sample in a disposable cuvette with a 1-cm pathlength (diluted 1:10 when OD600 was above 1.0). A 2-ml culture sample was taken from 50-ml cultures for pH measurement. The sample was centrifuged at 8,000 rpm for 5 minutes to remove cells from the media. The pH of the supernatant was measured using a Mettler Toledo FiveEasy Plus FP20 pH meter. 118 HPLC Analysis HPLC samples were prepared by centrifuging 1 ml aliquots from 50 ml cultures at 16,100 rpm for 10 minutes in a Minispin Plus Eppendorf microcentrifuge. 700 µl of the supernatant was transferred to a 2 ml glass HPLC vial. Sodium acetate standards were prepared at concentrations of 30, 20, 15, 7, and 3 mM in Milli-Q water. Samples were refrigerated at 10 °C in the HPLC autosampler during the run. HPLC analysis was performed on a Shimadzu 20A HPLC using an Aminex HPX- 87H column (BioRad, Hercules, CA) with a Microguard Cation H+ guard column (BioRad, Hercules, CA) at 50 °C. Analysis was conducted at a 0.6 ml min-1 flow rate in 5 mM sulfuric acid and a 30-minute run time. HPLC eluent was prepared by diluting 98% HPLC-grade sulfuric acid solution in Milli-Q water. The eluent was degassed at 37 °C for 3 to 5 days prior to use. Acetate was detected by refractive index (RID-20A). NAG concentration could not be measured due to co-elution with IPTG. Data analysis Growth and HPLC data were analyzed using RStudio using packages ggplot2, dplyr, grid, and gridExtra.34–36 Carrying capacity and growth rates of S. oneidensis cultures were determined using the R package “growthCurver” using default values.37 Microscopy Samples (2 µl) of 50 ml growth curves were periodically imaged on a Leica DM 1000 LED microscope and images were taken from a Leica ICC50 W camera and Leica app. Average cell areas were calculated with Image J.39 119 ACKNOWLEDGEMENTS We thank Drs. Jason Peters and Amy Banta for Mobile-CRISPRi materials and consultation, Nicholas Tefft for assistance with HPLC measurements, and Shaylynn Miller for assistance with Image J. FUNDING SOURCES This material is based upon work supported by a 2018 Beckman Young Investigator Award to MT. 120 REFERENCES 1. Gruenberg, M. C. & TerAvest, M. A. A common inducer molecule enhances sugar utilization by Shewanella oneidensis MR-1. J Ind Microbiol Biotechnol 50, (2023). 2. 3. Kouzuma, A. Molecular mechanisms regulating the catabolic and electrochemical activities of Shewanella oneidensis MR-1. Biosci Biotechnol Biochem 85, 1572– 1581 (2021). Zou, L., Huang, Y. Hong, Long, Z. er & Qiao, Y. On-going applications of Shewanella species in microbial electrochemical system for bioenergy, bioremediation and biosensing. World J Microbiol Biotechnol 35, 1–9 (2019). 4. Webster, D. P. et al. An arsenic-specific biosensor with genetically engineered Shewanella oneidensis in a bioelectrochemical system. Biosens Bioelectron 62, 320–324 (2014). 5. 6. Zang, Y. et al. A bio-electrochemical sensor based on suspended Shewanella oneidensis MR-1 for the sensitive assessment of water biotoxicity. Sens Actuators B Chem 341, 130004 (2021). Lin, T. et al. Synthetic Saccharomyces cerevisiae-Shewanella oneidensis consortium enables glucose-fed high-performance microbial fuel cell. AIChE Journal 63, 1830–1838 (2017). 7. Min, D. et al. Enhancing Extracellular Electron Transfer of Shewanella oneidensis MR-1 through Coupling Improved Flavin Synthesis and Metal-Reducing Conduit for Pollutant Degradation. Environ Sci Technol 51, 5082–5089 (2017). 8. 9. Venkateswaran, K. et al. Polyphasic taxonomy of the genus Shewanella and description of Shewanella oneidensis sp. nov. Int J Syst Bacteriol 49, 705–724 (1999). Serres, M. H. & Riley, M. Genomic Analysis of Carbon Source Metabolism of Shewanella oneidensis MR-1: Predictions versus Experiments. J Bacteriol 188, 4601 (2006). 10. Rodionov, D. A. et al. Genomic encyclopedia of sugar utilization pathways in the Shewanella genus. BMC Genomics 11, 494 (2010). 11. Chubiz, L. M. & Marx, C. J. Growth Trade-Offs Accompany the Emergence of Glycolytic Metabolism in Shewanella oneidensis MR-1. J Bacteriol 199, (2017). 12. Yang, C. et al. Comparative Genomics and Experimental Characterization of N- Acetylglucosamine Utilization Pathway of Shewanella oneidensis. J Biol Chem 281, 29872–29885 (2006). 121 13. Brutinel, E. D. & Gralnick, J. A. Anomalies of the anaerobic tricarboxylic acid cycle in Shewanella oneidensis revealed by Tn-seq. Mol Microbiol 86, 273–283 (2012). 14. Duhl, K. L., Tefft, N. M. & TerAvest, M. A. Shewanella oneidensis MR-1 Utilizes both Sodium- and Proton-Pumping NADH Dehydrogenases during Aerobic Growth. Appl Environ Microbiol 84, (2018). 15. Feng, X., Xu, Y., Chen, Y. & Tang, Y. J. Integrating Flux Balance Analysis into Kinetic Models to Decipher the Dynamic Metabolism of Shewanella oneidensis MR- 1. PLoS Comput Biol 8, e1002376 (2012). 16. Orr, J. S., Christensen, D. G., Wolfe, A. J. & Rao, C. V. Extracellular acidic pH inhibits acetate consumption by decreasing gene transcription of the tricarboxylic acid cycle and the glyoxylate shunt. J Bacteriol 201, 410–428 (2019). 17. Flynn, C. M., Hunt, K. A., Gralnick, J. A. & Srienc, F. Construction and elementary mode analysis of a metabolic model for Shewanella oneidensis MR-1. Biosystems 107, 120–128 (2012). 18. Peters, J. M. et al. Enabling genetic analysis of diverse bacteria with Mobile- CRISPRi. Nat Microbiol 2019 4:2 4, 244–250 (2019). 19. Ford, K. C., Kaste, J. A. M., Shachar-Hill, Y. & TerAvest, M. A. Flux-Balance Analysis and Mobile CRISPRi-Guided Deletion of a Conditionally Essential Gene in Shewanella oneidensis MR-1. ACS Synth Biol 11, 3405–3413 (2022). 20. Li, F. et al. Engineering Shewanella oneidensis enables xylose-fed microbial fuel cell. Biotechnol Biofuels 10, 1–10 (2017). 21. Cheng, L. et al. Developing a base-editing system to expand the carbon source utilization spectra of Shewanella oneidensis MR-1 for enhanced pollutant degradation. Biotechnol Bioeng 117, 2389–2400 (2020). 22. Dundas, C. M., Walker, D. J. F. & Keitz, B. K. Tuning Extracellular Electron Transfer by Shewanella oneidensis Using Transcriptional Logic Gates. ACS Synth Biol 9, 2301–2315 (2020). 23. Li, F. et al. Modular engineering to increase intracellular NAD(H/+) promotes rate of extracellular electron transfer of Shewanella oneidensis. Nat Comm 2018 9:1 9, 1– 13 (2018). 24. Yang, Y. et al. Enhancing Bidirectional Electron Transfer of Shewanella oneidensis by a Synthetic Flavin Pathway. ACS Synth Biol 4, 815–823 (2015). 25. Liu, T. et al. Enhanced Shewanella biofilm promotes bioelectricity generation. Biotechnol Bioeng 112, 2051–2059 (2015). 122 26. Yi, Y.-C. & Ng, I.-S. Redirection of metabolic flux in Shewanella oneidensis MR-1 by CRISPRi and modular design for 5-aminolevulinic acid production. Bioresour Bioprocess 2021 8:1 8, 1–11 (2021). 27. Jeon, J. M. et al. Butyrate-based n-butanol production from an engineered Shewanella oneidensis MR-1. Bioprocess Biosyst Eng 41, 1195–1204 (2018). 28. Lewis, M. A tale of two repressors – a historical perspective. J Mol Biol 409, 14 (2011). 29. Wilson, C. J., Zhan, H., Swint-Kruse, L. & Matthews, K. S. The lactose repressor system: Paradigms for regulation, allosteric behavior and protein folding. CMLS 64, 3–16 (2007). 30. Jobe, A. & Bourgeois, S. lac repressor-operator interaction: VI. The natural inducer of the lac operon. J Mol Biol 69, 397–408 (1972). 31. Novichkov, P. S. et al. RegPrecise 3.0 – A resource for genome-scale exploration of transcriptional regulation in bacteria. BMC Genomics 14, 745 (2013). 32. Somerville, G. A. & Proctor, R. A. Cultivation conditions and the diffusion of oxygen into culture media: The rationale for the flask-to-medium ratio in microbiology. BMC Microbiol 13, 9 (2013). 33. Tang, Y. J., Hwang, J. S., Wemmer, D. E. & Keasling, J. D. Shewanella oneidensis MR-1 Fluxome under Various Oxygen Conditions. Appl Environ Microbiol 73, 718– 729 (2007). 34. R Core Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing Preprint at https://www.R-project.org/ (2021). 35. H. Wickham. ggplot2: Elegant Graphics for Data Analysis. Preprint at (2016). 36. Hadley Wickham, Romain François, Lionel Henry & Kirill Müller. dplyr: A Grammar of Data Manipulation. R package version 1.0.9 Preprint at https://CRAN.R- project.org/package=dplyr (2022). 37. Sprouffske, K. & Wagner, A. Growthcurver: An R package for obtaining interpretable metrics from microbial growth curves. BMC Bioinformatics 17, 1–4 (2016). 38. Pinchuk, G. E., Hill, E. A., Geydebrekht, O. v., de Ingeniis, J., Zhang, X., Osterman, A., Scott, J. H., Reed, S. B., Romine, M. F., Konopka, A. E., Beliaev, A. S., Fredrickson, J. K., & Reed, J. L. Constraint-Based Model of Shewanella oneidensis MR-1 Metabolism: A Tool for Data Analysis and Hypothesis Generation. PLoS Comput Biol, 6(6), e1000822 (2010). 39. Schneider, C. A., Rasband, W. S., & Eliceiri, K. W. NIH Image to ImageJ: 25 years of image analysis. Nat Methods, 9(7), 671–675 (2012). 123 CHAPTER 5: ELECTRO-FERMENTATION INFLUENCES FERMENTATION PRODUCTS IN ENGINEERED ESCHERICHIA COLI 124 PREFACE This investigation began as summer projects for two undergraduate students: Carrie Gregg and Andrew Scheil. The goal of the project was to determine whether an electroactive strain of E. coli developed by my advisor’s postdoc lab could participate in EET with a fermentable substrate. The previous studies with the organism used the non- fermentable substrate lactate in order to study the EET pathway. Now, with confirmation that EET does occur, we were interested in exploring the biotechnological potential of the strain by using the fermentable substrate glucose. The project began with Carrie and I experimenting with solid and soluble iron as the terminal electron acceptor while Andrew and his graduate student mentor, Shay Miller, were experimenting with an anode as the terminal electron acceptor. Over the course of the summer, it became clear that we were seeing the same results in both cases – the fermentation product ratios were shifting toward more oxidized products when the electroactive E. coli was participating in EET. We had some promising preliminary data showing there was an interesting story here, but the summer was ending. Carrie returned to her home institution and Andrew started on a different project. I took over primary ownership of this project and repeated experiments both with iron and an anode as the terminal electron acceptor. I was also able to take this project a step further by reversing the EET pathway using a cathode as an electron source and demonstrated the opposite shift in fermentation product ratios. I performed some additional analysis of the electron and carbon flow through the system which allowed us to make some further conclusions about the efficiency of the system and areas which could be improved upon. This is by far one of the most collaborative projects I had the 125 pleasure of working on during my PhD, and I really enjoyed having multiple hands and minds contributing toward the same goal. While I do enjoy independent work, this project was a nice change of pace from some of the others which were solely my responsibility from their inception. This chapter describes the work that was done on investigating anodic and cathodic electro-fermentation with an electroactive strain of E. coli. ABSTRACT Escherichia coli is a popular host for fermentative biosynthesis of small molecule products due to its well understood metabolism and large genetic toolbox. A major limitation to product scope and yield using fermentation is the requirement that oxidation and reduction reactions be balanced. In this work, we demonstrate unbalanced electro- fermentation using an electroactive strain of E. coli expressing the Mtr pathway from Shewanella oneidensis. The Mtr pathway could act as an electron sink by transporting intracellular electrons derived from substrate oxidation to an extracellular electron acceptor, thus favoring production of oxidized fermentation products. We also show that the Mtr pathway is reversible in E. coli and could alternatively act as an electron source by importing electrons from a cathode, therefore favoring reduced fermentation products. This proof-of-concept suggests that electro-fermentation could be a useful tool for improving fermentations that are currently restrained by redox balance. INTRODUCTION Fermentation is the metabolic process of balancing substrate oxidation and reduction reactions in the absence of oxygen or other electron acceptors, which many organisms can use to conserve energy.1 During this process, a substrate is oxidized to 126 generate an intermediate and reducing equivalents (e.g., NADH, FADH2).1,2 For a balanced reaction, the intermediate must be subsequently reduced for the regeneration of the oxidized cofactor (i.e. NAD+, FAD).1,2 These balanced redox reactions enable ATP generation via substrate level phosphorylation.1 For example, Escherichia coli produces ethanol, lactate, succinate, formate, and acetate as major fermentation products from glucose (Figure 38), and cellular redox balance is maintained by generating different ratios of these products.3 Metabolic engineering techniques have been harnessed for production of value- added chemicals, nutritional supplements, pharmaceuticals, and proteins via fermentation.4–9 Major constraints of fermentation processes include limited product scopes and low product yields, which arise from the requirement for redox equivalents to be balanced between substrates and products during fermentation.10,11 Electro- Figure 38. Major fermentation pathways in E. coli. Boxed metabolites indicate common fermentation products. Abbreviations: PEP, phosphoenolpyruvate; ppc, PEP carboxylase; pykAF, pyruvate kinase; ldhA, lactate dehydrogenase; pflB, pyruvate formate lyase; adhE, aldehyde dehydrogenase, pta, phosphate acetyl transferase; ackA, acetate kinase; mdh, malate dehydrogenase; fumBC, fumarate hydratase; frdABCD, fumarate reductase; e-, electron. Created with Biorender.com. 127 fermentation is an emerging technology in which an electrode can act as an electron source or sink, allowing unbalanced fermentation by alleviating the constraint of redox balance.12 Electro-fermentation is possible with microorganisms capable of extracellular electron transfer (EET), which allows electron transport across the cellular membrane to or from an electrode. Electroactive microorganisms can participate in EET through direct or mediated electron transfer.13–18 Direct EET requires direct contact between the cell and the extracellular electron acceptor.16–18 Indirect EET involves electron shuttling from the cell surface to the extracellular electron acceptor by an electron carrier.13–15 Mixed microbial cultures are used as the biocatalyst in most electro-fermentation systems; however, they are limited in their potential for improvement of substrate scope and product yield, purity, and scope due to their intrinsic unknown composition.12 As an alternative, electro-fermentation could be engineered using well-studied strains with large synthetic biology toolboxes. For example, E. coli is a well-understood model organism with many synthetic biology tools available for strain improvement, making it an appealing host for engineering electroactivity.19 To take advantage of the benefits that come with using E. coli as a host, electroactive E. coli strains have been engineered to reduce extracellular electron acceptors. Some strains utilize synthetic electron conduits for extracellular electron transfer such as modified carbon nanoparticles which are added to wild-type E. coli and traverse the cell membrane.20 Other strains express membrane cytochromes from the Mtr pathway in Shewanella oneidensis, which are products of work pioneered by research groups led by Seward, Spormann, Gescher, and Ajo-Franklin.21–24 The S. oneidensis Mtr pathway is composed of CymA and MtrCAB. CymA is an inner membrane tetraheme cytochrome c which transfers electrons from the inner 128 membrane quinone pool to MtrCAB.25,26 The MtrCAB complex is composed of MtrC, an outer membrane cytochrome c; MtrA, a periplasmic cytochrome c; and MtrB, a β-barrel protein which spans the outer membrane and houses MtrA and MtrC.27,28 The native function of the Mtr pathway in S. oneidensis is to transfer intracellular electrons to extracellular electron acceptors; however, this pathway can also operate in reverse and power intracellular reduction reactions.29–32 Seward and Gescher conducted initial work exploring the heterologous expression of S. oneidensis cytochromes in E. coli.21–23 The first of these studies showed that MtrA could be matured by the native E. coli cytochrome c maturation pathway (Ccm) and could reduce soluble iron.21 E. coli expressing MtrA alone could not reduce solid iron, which was an indicator that there was no native E. coli mechanism which could transport electrons to the cell surface.21,33 Gescher et al. showed that expressing CymA alone or CymA and MtrA together both allowed E. coli to reduce chelated iron; however, these strains also were incapable of reducing solid electron acceptors.22,23,33 The electroactive E. coli MFe444 developed by Ajo-Franklin and colleagues expresses CymA and MtrCAB from S. oneidensis and the native E. coli cytochrome maturation system.24 In this strain, the Mtr pathway allows electrons derived from substrate oxidation by quinone-linked oxidoreductases (e.g., lactate dehydrogenase) to be passed to the outer surface of the cell via the Mtr pathway. Electron transfer from the cell surface to a solid electron acceptor in this system is largely dependent on flavins as electron shuttles, as observed in S. oneidensis (Figure 39). MFe444 reduces Fe3+ with or without riboflavin when provided lactate as the electron donor; however, riboflavin improves electron transfer to solid iron minerals by 2.5-fold.24 MFe444 also generates a higher current in bioelectrochemical 129 Figure 39. Mtr pathway from S. oneidensis expressed in MFe444. CymA and MtrCAB from S. oneidensis are expressed from an IPTG-inducible plasmid. Electrons from substrate oxidation by a quinone-linked (Q-linked) oxidoreductase are passed through the inner membrane quinone pool to CymA, then transported across the outer membrane via MtrCAB. During flavin-independent EET, electrons are deposited directly into the terminal electron acceptor (TEA). In flavin-dependent EET, riboflavin (Rf) shuttles electrons to the TEA. Created with Biorender.com. systems than a control strain expressing only the cytochrome maturation genes.24,34,35 When connected to the electrode during lactate metabolism, the spectrum of metabolites generated by MFe444 is shifted toward more oxidized products, such as acetate, and away from reduced products, such as ethanol. However, less than 20% of available lactate was consumed under this condition because the current electron transfer rates were not sufficient to enable growth via respiration. These findings indicate the functional expression of the Mtr pathway; however, E. coli does not ferment lactate. MFe444 can also use the Mtr pathway for inward electron transfer, as shown by its ability to reduce fumarate or nitrate with electrons supplied from a cathode.36 In this work, we explore the biotechnological potential for E. coli MFe444 by utilizing it as a host for glucose electro-fermentation and demonstrate control in both 130 directions over the ratio of oxidized and reduced E. coli fermentation products via an electrode. METHODS Strains and plasmids Strains and plasmids used in this study are listed in Table 14. Table 14. Strains and plasmids used for experiments described in Chapter 5. Strains Parent Plasmids Source MFe408 C43(DE3) pEC86, pSB1ET2 MFe444 C43(DE3) pEC86, I5049 54 24 Plasmids Genes Antibiotic Resistance Source pEC86 ccmA-H Chloramphenicol pSB1ET2 none Kanamycin I5049 cymA, mtrCAB Kanamycin 55 56 24 Culturing MFe408 and MFe444 were precultured aerobically in 2 ml LB (Neogen) supplemented with 50 μg ml-1 kanamycin and 30 μg ml-1 chloramphenicol overnight at 30 °C with shaking at 250 rpm. To induce expression of the Mtr pathway and cytochrome c maturation proteins, MFe408 and MFe444 were subcultured in 50 ml of 2xYT (Difco) supplemented with 50 μg ml-1 kanamycin and 30 μg ml-1 chloramphenicol. 2xYT cultures were inoculated to an initial OD600 of 0.02 and were incubated aerobically at 30 °C with 131 shaking at 250 rpm. OD600 was monitored and IPTG was added to a final concentration of 10 μM when cultures reached an OD600 of 0.2-0.3. After approximately 18 hours, cells were harvested via centrifugation (9,000 rpm, 5 min, 22 °C), the supernatant was discarded, and cell pellets were washed with 25 ml of basal M5 high HEPES medium (1.29 mM K2HPO4, 1.65 mM KH2PO4, 7.87 mM NaCl, 1.70 mM NH4SO4, 475 μM MgSO4·7H2O, 100 mM HEPES, 0.01% (w/v) casamino acids, pH 7.2). Cell pellets were resuspended in 25 ml basal M5 high HEPES medium and were inoculated into experimental cultures to an initial OD600 of 0.2. Cultures for iron-reduction experiments were conducted in an anaerobic chamber with an atmosphere of 5% H2 and 95% N2. All media and culture tubes were placed into the chamber one day prior to inoculation. Iron-reduction medium contained basal M5 high HEPES medium supplemented with Wolfe’s mineral solution, Wolfe’s vitamin solution without riboflavin, 20 mM D-glucose, either 50 mM Fe(III)-NTA or 50 mM Fe2O3, and 50 μg ml-1 kanamycin. Where specified, 5 μM riboflavin was added. Fe2+ quantification Fe2+ quantification was performed using the ferrozine method.37 Acidified samples and standards were prepared by adding 25 μl of standard or sample to 75 μl of anaerobic 2 M HCl in an anaerobic chamber. Samples and standards were removed from the chamber after acidification. A 10 μl aliquot of acidified sample or Fe2+ standard prepared from iron (II) chloride tetrahydrate was combined with 190 μl of 1 g L-1 FerroZine reagent in a 96-well plate. Absorbance was read at 562 nm on a H1M BioTek Plate Reader. 132 HPLC analysis HPLC samples were prepared by centrifuging 700 μl aliquots at 16,100 rpm for 10 minutes in a Minispin Plus Eppendorf microcentrifuge. 600 µl of the supernatant was transferred to a 2 ml glass HPLC vial. Samples were refrigerated at 10 °C in the HPLC autosampler during the run. HPLC analysis was performed on a Shimadzu 20A HPLC using an Aminex HPX- 87H column (BioRad, Hercules, CA) with a Microguard Cation H+ guard column (BioRad, Hercules, CA) at 50 °C. Analysis was conducted at a 0.6 ml min-1 flow rate in 5 mM sulfuric acid and a 30-minute run time. HPLC eluent was prepared by diluting 98% HPLC-grade sulfuric acid solution in Milli-Q water. The eluent was degassed at 37 °C for 3 to 5 days prior to use. Bioelectrochemical system set-up Bioelectrochemical measurements were performed using two-chamber bioelectrochemical reactors (Adams and Chittenden Scientific Glass Coop, custom design), separated by a cation exchange membrane (Membranes International, Catalog: CMI-7000S). The three-electrode system consisted of a carbon felt working electrode adhered to a titanium wire via CCC carbon adhesive (Electron Microscopy Sciences, Catalog: 12664), a graphite rod as the counter electrode (Electron Microscopy Science, Catalog: 07200), and an Ag/AgCl reference electrode in saturated KCl, with a magnesia stick frit (Sigma Aldrich, Catalog: 31408-1EA). The working chamber was filled with a final volume of 100 ml of M5 high HEPES medium prior to autoclaving. After autoclaving, the medium was supplemented with Wolfe’s mineral solution, Wolfe’s vitamin solution without riboflavin, 50 mg ml-1 kanamycin, 50 mM D-glucose, 10 μM IPTG, and 5 μM riboflavin. The 133 electrodes were connected to a potentiostat (VMP, BioLogic USA), and the working electrode was poised at +200 mVAg/AgCl for anodic experiments or -500 mVAg/AgCl for cathodic experiments. Current was measured every 1 second. Precultures were prepared as described above, and bioelectrochemical system working chambers were inoculated with washed cells to an initial OD600 of 0.2. Data analysis Data were analyzed using R version 4.3.1 and packages ggplot2 3.4.2, dplyr 1.1.2, TTR 0.24.3, viridis 0.6.4, grid 4.3.1, and gridExtra 2.3.38–43 Moles of electrons to or from the electrode during electro-fermentations were calculated by integrating the area under the curve and converting to moles using Faraday’s constant using Equation 1. Percent of electrons from glucose routed to the anode were calculated by dividing the moles of electrons by the total moles of electrons input using Equation 2. Percent of electrons from glucose routed to fermentation products during anodic electro-fermentation were calculated by dividing the total number of electrons in the fermentation products by the total number of electrons in glucose consumed following Equation 3. Percent of electrons from the cathode routed to fermentation products were calculated by dividing the total number of electrons in the fermentation products by the total amount of electrons input (from glucose and electrode) using Equation 4. Percent of carbon routed to fermentation products were calculated by dividing the moles of carbon in the fermentation products by the moles of carbon from glucose consumed using Equation 5. electrode e− moles = ∫ current mC s × 1 s × C 1000 mC × mole 96,485 C Equation 1. Conversion of mC s to moles of electrons. 134 % e− to anode = ( electrode e− moles ∆glucose moles L (0.08 L × ) × 24 e− ) × 100 Equation 2. Percent of electrons routed to electrode in anodic electro- fermentation. 𝑓(𝑒) = 𝑒− 𝑚𝑜𝑙𝑒𝑠 𝑜𝑓 𝑝𝑟𝑜𝑑𝑢𝑐𝑡 = (0.08 L × product moles L ) × e− Product Lactate Formate Acetate Succinate Ethanol e- 12 2 8 13 12 % e− to fermentation products = ( (0.08 L × n 𝑓(𝑒, 𝑖) Σi=1 ∆glucose moles L ) × 100 ) × 24 e− Equation 3. Percent of electrons routed to fermentation products in anodic electro- fermentation. 𝑒− 𝑚𝑜𝑙𝑒𝑠 𝑜𝑓 𝑝𝑟𝑜𝑑𝑢𝑐𝑡 = (0.08 L × product moles L ) × e− Product Lactate Formate Acetate Succinate Ethanol e- 12 2 8 13 12 135 % e− to fermentation products n 𝐞𝐪𝐮𝐚𝐭𝐢𝐨𝐧 𝟓, i Σi=1 ((0.08 L × ∆glucose moles L ) × 24 e−) + electrode e− moles = ( × 100 ) Equation 4. Percent of electrons routed to fermentation products in cathodic electro-fermentation. 𝑓(𝑐) = 𝑒− 𝑚𝑜𝑙𝑒𝑠 𝑜𝑓 𝑝𝑟𝑜𝑑𝑢𝑐𝑡 = (0.08 L × product moles L ) × C Product Lactate Formate Acetate Succinate Ethanol C 3 1 2 4 2 % C to fermentation products = ( n 𝑓(𝑐, 𝑖) Σi=1 ∆glucose moles L ) × 100 ) × 6 C (0.08 L × Equation 5. Percent of carbon routed to fermentation products. RESULTS The engineered electroactive E. coli strain MFe444 can use the non-fermentable electron donor lactate to reduce extracellular iron and electrodes.24 Here, we tested whether MFe444 can also use a fermentable substrate to power electrode reduction and whether the presence of the electrode would alter the ratio of fermentation products. We 136 cultivated MFe444 (E. coli expressing S. oneidensis CymA, MtrCAB, and E. coli cytochrome c maturation proteins) and MFe408 (E. coli expressing only the cytochrome c maturation proteins) in a bioelectrochemical system with 50 mM D-glucose, 5 μM riboflavin, and an anode poised at +200 mVAg/AgCl. MFe444 produced a higher current than MFe408, indicating more electron transfer occurring in the Mtr-expressing strain (Figure 40a). OD600 in the working chamber of the bioelectrochemical system was monitored throughout the cultivation period. Both strains reached a similar maximum OD600 after one day of cultivation. MFe408 maintained a relatively consistent OD600 from days 1-6 whereas MFe444 displays a drop in OD600 after day 1 (Figure 41). This indicates that the difference in current between the strains is due to Mtr pathway activity and not a difference in cell density, as MFe444 displayed a higher current than MFe408 on day one when both strains had a similar OD600. We measured the concentrations of fermentation products throughout the anodic electro-fermentation and observed that MFe444 produced 19.0% and 28.5% less of the more reduced fermentation products lactate and succinate, respectively. MFe444 produced 10.4% and 16.4% more of the oxidized fermentation products formate and acetate, respectively (Figure 40b, Table 15). Fumarate concentrations were not assessed because it is not a major fermentation product and due to low succinate production.3,44,45Previous studies have shown that riboflavin improves EET to solid electron acceptors in MFe444 oxidizing lactate.24 To determine whether riboflavin was required for EET during glucose electro-fermentation, MFe408 and MFe444 were cultivated in bioelectrochemical systems with 50 mM D-glucose as the carbon source and electron donor and an anode poised at +200 mVAg/AgCl as the extracellular electron acceptor. When the bioelectrochemical systems were inoculated, the reactors contained 137 Figure 40. Current (A) and metabolite concentrations (B) of MFe408 and MFe444 during anodic electro-fermentation. Strains cultivated in a bioelectrochemical system with 50 mM D-glucose, 5 μM riboflavin, and an anode poised at +200 mVAg/AgCl. Inoculation occurred at time 0. P-values were calculated at each timepoint between the MFe408 and MFe444 conditions. * 0.05 < p < 0.1, ** 0.01 < p < 0.05, *** 0.001 < p < 0.01, **** p < 0.001. P values not calculated for glucose. 138 Table 15. Final concentrations and percent differences of metabolites from MFe408 and MFe444 anodic electro-fermentation depicted in Figure 40. Lactate (mM) Succinate (mM) Formate (mM) Acetate (mM) Ethanol (mM) MFe408 14.7 ± 0.3 0.7 ± 0.1 15.4 ± 0.1 6.7 ± 0.1 6.2 ± 0.2 MFe444 11.9 ± 0.2 0.5 ± 0.0 17.0 ± 0.4 7.8 ± 0.1 5.5 ± 0.6 Percent difference 19.0% 28.5% 10.4% 16.4% 11.3% Figure 41. OD600 measurements of MFe408 and MFe444 during anodic electro- fermentation depicted in Figure 40. Strains cultivated in a bioelectrochemical system with 50 mM D-glucose, 5 μM riboflavin, and an anode poised at +200 mVAg/AgCl. 0.2 μM riboflavin. After ~24 hours, riboflavin was increased to a final concentration of 5 μM. With 0.2 μM riboflavin, current production by the two strains was similar. A divergence in current values was observed when the riboflavin concentration was increased, with MFe444 displaying a greater current than MFe408 (Figure 42). 139 Figure 42. Current of MFe408 and MFe444 anodic electro-fermentation with different riboflavin concentrations. Strains were cultivated with 50 mM D-glucose and an anode poised at +200 mVAg/Agcl with 0.2 or 5 μM riboflavin (Rf). Sufficient riboflavin concentration is required to support EET between the electroactive E. coli and the electrode. Inoculation occurred at time 0. Since electro-fermentation requires specialized equipment, we tested whether we could achieve the same effect using a cheaper and more accessible method. To do this, we cultured MFe408 and MFe444 anaerobically in minimal medium containing 20 mM D- glucose as a carbon and electron source, either soluble iron (Fe(III)-NTA) or insoluble iron (Fe2O3) as an electron acceptor, with 5 μM riboflavin. Iron reduction was monitored as an indicator of extracellular electron transfer (Figure 43a). Under these conditions, MFe444 reduced 225% more Fe2O3 than the control strain (Table 16). Glucose was completely consumed, and fermentation product generation was complete within the first 24 hours of cultivation (Figure 43b). All fermentation product concentrations were 0 mM at the time of inoculation, and this measurement was omitted from the plots to ease visualization of differences. Consistent with the anodic electro-fermentation results, MFe444 produced 27.6% less lactate and 2.6% more acetate than the control strain (Table 16). Formate 140 Figure 43. Metabolite concentrations during anaerobic cultivation of MFe408 and MFe444 with 20 mM D-glucose as the electron donor and 50 mM Fe2O3 as the electron acceptor. Note that y-axis scales differ between panels. Metabolite concentrations are reported starting at day 1 to promote ease of trend visualization. All fermentation product concentrations were measured and found to be 0 mM at the time of inoculation. P-values were calculated at each timepoint between the MFe408 and MFe444 conditions. * 0.05 < p < 0.1, ** 0.01 < p < 0.05, *** 0.001 < p < 0.01, **** p < 0.001. P values for glucose were not determined. 141 Table 16. Metabolite concentrations of MFe408 and MFe444 cultivation with glucose, riboflavin, and Fe2O3 depicted in Figure 43. Fe(II) (mM) Lactate (mM) Succinate (mM) Formate (mM) Acetate (mM) MFe408 0.04 ± 0.01 1.45 ± 0.03 0.71 ± 0.02 8.96 ± 0.22 6.05 ± 0.05 MFe444 0.13 ± 0.01 1.05 ± 0.09 0.76 ± 0.02 7.40 ± 0.32 6.21 ±0.06 Percent difference 225% 27.6% 7.0% 17.4% 2.6% concentrations began to drop after the first day, while the other fermentation products (lactate, succinate, and acetate) remained at consistent levels. Because iron reduction does not occur until later timepoints where formate concentrations are decreasing, it is possible that electrons derived from formate oxidation participate in the iron reduction. With soluble iron provided as the electron acceptor, there was no difference between the two strains’ iron reduction capabilities (Figure 44). To determine if the fermentation product ratio could be influenced in the opposite direction, the strains were cultivated in bioelectrochemical systems with 50 mM D-glucose, 5 μM riboflavin, and a cathode poised at -500 mVAg/AgCl as an extracellular electron donor. Under these conditions, MFe444 produced a greater negative current than MFe408 (Figure 45a). Similar to the anodic electro-fermentation, the two strains displayed similar OD600 measurements throughout the course of the cultivation, indicating that the difference in current between the strains is due to Mtr pathway activity and not a difference in cell density (Figure 46). MFe444 production of the reduced fermentation product succinate increased 5.2% during cathodic electro-fermentation, indicating that the electrons taken up by the Mtr pathway could be used toward production of more reduced fermentation products under these conditions (Figure 45b, Table 17). 142 Figure 44. Fe2+ and metabolite concentrations during anaerobic cultivation of MFe408 and MFe444 with 20 mM D- glucose as the electron donor and 50 mM Fe(III)-NTA as the electron acceptor. 143 Figure 45. Current (A) and metabolite concentrations (B) of MFe408 and MFe444 during cathodic electro- fermentation. Strains cultivated in a bioelectrochemical system with 50 mM D-glucose, 5 μM riboflavin, and a cathode poised at -500 mVAg/AgCl. Inoculation occurred at time 0. Metabolite concentrations are reported starting at day 1 to promote ease of trend visualization. All fermentation product concentrations were measured and found to be 0 mM at the time of inoculation. P-values were calculated at each timepoint between the MFe408 and MFe444 conditions. * 0.05 < p < 0.1, ** 0.01 < p < 0.05, *** 0.001 < p < 0.01, **** p < 0.001. P values for glucose not calculated. 144 Figure 46. OD600 measurements of MFe408 and MFe444 during cathodic electro- fermentation depicted in Figure 45. Strains cultivated in a bioelectrochemical system with 50 mM D-glucose, 5 μM riboflavin, and a cathode poised at -500 mVAg/AgCl. Table 17. Final concentrations and percent differences of metabolites from MFe408 and MFe444 cathodic electro-fermentation depicted in Figure 45. Lactate (mM) Succinate (mM) Formate (mM) Acetate (mM) Ethanol (mM) MFe408 12.92 ± 0.37 1.16 ± 0.06 17.78 ± 0.26 7.57 ± 0.04 15.01 ± 0.57 MFe444 12.09 ± 0.01 1.22 ± 0.06 18.21 ± 0.02 7.94 ± 0.04 15.30 ± 0.43 Percent difference 6.4% 5.2% 2.4% 4.9% 1.9% DISCUSSION In this work, we demonstrate proof-of-concept that electro-fermentation can be used to influence E. coli fermentation product ratios, providing a potential mechanism for overcoming the redox balance restrictions of traditional fermentation. We show that an engineered E. coli strain MFe444 can use the fermentable sugar glucose for electro- fermentation and the electrode potential influences fermentation product ratios. The shift 145 in fermentation product ratios during anodic electro-fermentation toward oxidized products and away from reduced products is justified by the changes in electron transfer capabilities between the two strains. MFe444 increases electron transfer to the electrode by 104% over MFe408 with the Mtr pathway. This is accompanied by a 2.5% decrease in electrons routed towards fermentation products, but only a 0.9% decrease in carbon to fermentation products. The greater difference in electrons than carbon being directed to fermentation is directly observable in the fermentation product ratio, accounting for the decrease in lactate and succinate and concurrent increase in formate and acetate. While the electrode clearly has influence over the fermentation product ratios, the difference of “lost” electrons in the fermentation pathway of MFe444 cannot be completely explained by extracellular electron transfer. MFe444 produces 2.7 mmol fewer electrons in the form of fermentation products’ chemical bonds than MFe408. The difference in electrons deposited on the anode is only 0.101 mmol, which accounts for only 3.7% of the electrons removed from the fermentation pathway. We hypothesize the anode may also be responsible for regulatory changes that provide additional mechanisms of altering fermentation pathway flux beyond the redirection of electrons to the anode. Changes in regulation were also observed in MFe444 when exposed to a cathode electrode previously.36 We also show that influence over the fermentation product ratios can be achieved by cultivating MFe444 anaerobically in culture tubes with solid iron as the terminal electron acceptor. This set up may be more accessible and reasonable to scale up than bioelectrochemical systems, making it a viable alternative method. Cultivating MFe444 with iron as the terminal electron acceptor generated lower amounts of the fermentation 146 products than during anodic electro-fermentation, except in the case of succinate. The fermentation product ratios in the iron reducing cultures resemble that of the anodic electro-fermentation for changes in lactate and acetate production, but not in succinate and formate production. In iron reduction cultures, MFe444 produces 27.6% less lactate and 2.6% more acetate, whereas during anodic electro-fermentation MFe444 produces 19.0% less lactate and 16.4% more acetate. These changes reflect the expected observation that less reduced products and more oxidized products would be produced more in a system where electrons are transported out of the cell. Cultivation with iron as the terminal electron acceptor deviates from expectations for other metabolites, with MFe444 producing 7.0% more succinate and 17.4% less formate than MFe408. This is contrasted with the expected effect observed in the anodic electro-fermentation, where MFe444 produced 28.5% less succinate and 10.4% more formate. Another explanation for the differences between anodic electro-fermentation and iron reduction cultures could be the atmosphere in which they were cultivated. Bioelectrochemical systems were continually sparged with N2 whereas the iron reduction cultures were incubated in an anaerobic chamber containing 95% N2 and 5% H2. H2 and formate metabolism are linked through the quinone pool and through formate hydrogen lyase.46,47 The presence of H2 could therefore be a cause of altered formate metabolism. Additionally, the initial glucose concentrations were different between the conditions. The iron reduction cultures contained 20 mM D-glucose to create an electron-acceptor limited environment to encourage extracellular electron transfer to Fe2O3. D-glucose was included at 50 mM in the bioelectrochemical reactors because the anode acts as a limitless electron acceptor. The greater availability of glucose in the bioelectrochemical reactors provided 147 more carbon to the system and it is therefore reasonable that MFe444 produces a greater amount of fermentation products than in the iron-reducing cultures. The fine-tuning of aspects such as glucose concentrations and atmospheric composition could be useful tools in refining MFe444 fermentation. MFe444 uses inward electron transfer to take up 90.3% more electrons from the cathode than MFe408. This is accompanied by a 0.3% increase in electrons and a 1.1% increase in carbon routed to fermentation products. While cathodic electro-fermentation displays operational inward electron transfer via the Mtr pathway, the effect on fermentation products is less pronounced. Overall, the maximum current, total charge transfer, and shift in metabolites was smaller for the cathodic electro-fermentation than for the anodic electro-fermentation. This is consistent with studies on the Mtr pathway in S. oneidensis, which generally indicate lower cathodic current than anodic current.30,48–51 Inward electron transfer during E. coli cathodic electro-fermentation may be hindered by a limitation of proton-motive force. We have previously shown that in S. oneidensis, inward electron transfer to NAD+ can occur but is dependent on proton-motive force because it requires reverse activity of respiratory NADH dehydrogenases.29,52 Without a robust source of ATP or proton motive force generation, electron transfer from respiratory quinones to NAD+ is limited. Taken together, these data show that the fermentation products of electroactive E. coli can be influenced by the availability of reducing equivalents, and that this can be exploited for use in unbalanced fermentation. The Mtr pathway can operate in both the inward and outward electron transfer directions, which can allow the user to control production of more reduced or more oxidized products. Further strain and methodological 148 optimization could enhance control over fermentation product ratios. For example, expression of a periplasmic electron carrier such as S. oneidensis FccA could improve electron transfer rates between CymA and MtrCAB; and greater electrode potentials could create a stronger electron source or sink thus more greatly favoring reduced or oxidized fermentation products. Control over the availability of reducing equivalents during E. coli electro-fermentation could be a useful tool for optimizing biosynthesis of various products that are currently limited by the redox balance constraints of traditional fermentations. FUNDING SOURCES This work was supported by a 2018 Beckman Young Investigator Award and a National Science Foundation CAREER Award (1750785) to MT. CG was partially supported by the MSU Building Bridges Program. 149 REFERENCES 1. Müller, eLS doi:https://doi.org/10.1002/9780470015902.a0001415.pub2. Fermentation. Bacterial V. in (Wiley, 2008). 2. 3. 4. 5. 6. Folch, P. L., Bisschops, M. M. M. & Weusthuis, R. A. Metabolic energy conservation for fermentative product formation. Microb Biotechnol 14, 829–858 (2021). Clark, D. The fermentation pathways of Escherichia coli. FEMS Microbiol Rev 63, 223–234 (1989). Diaz, A. B., Blandino, A. & Caro, I. Value added products from fermentation of sugars derived from agro-food residues. Trends Food Sci Technol 71, 52–64 (2018). Iram, A., Ozcan, A., Turhan, I. & Demirci, A. Production of Value-Added Products as Food Ingredients via Microbial Fermentation. Processes 11, 1715 (2023). Tse, T. J., Wiens, D. J., Chicilo, F., Purdy, S. K. & Reaney, M. J. T. Value-Added Products from Ethanol Fermentation—A Review. Fermentation 7, 267 (2021). 7. Martínez, J. L., Liu, L., Petranovic, D. & Nielsen, J. Pharmaceutical protein production by yeast: towards production of human blood proteins by microbial fermentation. Curr Opin Biotechnol 23, 965–971 (2012). 8. 9. Elander, R. P. Industrial production of β-lactam antibiotics. Appl Microbiol Biotechnol 61, 385–392 (2003). Seo, S.-O. & Jin, Y.-S. Next-Generation Genetic and Fermentation Technologies for Safe and Sustainable Production of Food Ingredients: Colors and Flavorings. Annu Rev Food Sci Technol 13, 463–488 (2022). 10. Kracke, F., Lai, B., Yu, S. & Krömer, J. O. Balancing cellular redox metabolism in microbial electrosynthesis and electro fermentation – A chance for metabolic engineering. Metab Eng 45, 109–120 (2018). 11. Yamada, S., Takamatsu, Y., Ikeda, S., Kouzuma, A. & Watanabe, K. Towards Application of Electro-Fermentation for the Production of Value-Added Chemicals From Biomass Feedstocks. Front Chem 9, (2022). 12. Sravan, J. S., Butti, S. K., Sarkar, O. & Mohan, S. V. Electrofermentation. in Microbial Electrochemical Technology 723–737 (Elsevier, 2019). doi:10.1016/B978-0-444- 64052-9.00029-7. 13. Xu, S., Jangir, Y. & El-Naggar, M. Y. Disentangling the roles of free and cytochrome- bound flavins in extracellular electron transport from Shewanella oneidensis MR-1. Electrochim Acta 198, 49–55 (2016). 150 14. Marsili, E. et al. Shewanella secretes flavins that mediate extracellular electron transfer. PNAS 105, 3968–3973 (2008). 15. Brutinel, E. D. & Gralnick, J. A. Shuttling happens: soluble flavin mediators of extracellular electron transfer in Shewanella. Appl Microbiol Biotechnol 93, 41–48 (2012). 16. Okamoto, A., Nakamura, R. & Hashimoto, K. In-vivo identification of direct electron transfer from Shewanella oneidensis MR-1 to electrodes via outer-membrane OmcA–MtrCAB protein complexes. Electrochim Acta 56, 5526–5531 (2011). 17. Hartshorne, R. S. et al. Characterization of Shewanella oneidensis MtrC: a cell- surface decaheme cytochrome involved in respiratory electron transport to extracellular electron acceptors. J Biol Inorg Chem 12, 1083–1094 (2007). 18. Firer-Sherwood, M., Pulcu, G. S. & Elliott, S. J. Electrochemical interrogations of the Mtr cytochromes from Shewanella: opening a potential window. J Biol Inorg Chem 13, 849–854 (2008). 19. Idalia, V.-M. N. & Bernardo, F. Escherichia coli as a Model Organism and Its Application in Biotechnology. in Escherichia coli - Recent Advances on Physiology, Pathogenesis and Biotechnological Applications (ed. Samie, A.) 253–274 (IntechOpen, Rijeka, 2017). doi:10.5772/67306. 20. Lee, Y., Lee, J. & Kim, S. Direct extracellular electron transfer from Escherichia coli through modified carbon nanoparticles. Electrochim Acta 437, 141497 (2023). 21. Pitts, K. E. et al. Characterization of the Shewanella oneidensis MR-1 Decaheme Cytochrome MtrA. J Biol Chem 278, 27758–27765 (2003). 22. Gescher, J. S., Cordova, C. D. & Spormann, A. M. Dissimilatory iron reduction in Escherichia coli : identification of CymA of Shewanella oneidensis and NapC of E. coli as ferric reductases. Mol Microbiol 68, 706–719 (2008). 23. Schuetz, B., Schicklberger, M., Kuermann, J., Spormann, A. M. & Gescher, J. Periplasmic Electron Transfer via the c -Type Cytochromes MtrA and FccA of Shewanella oneidensis MR-1. Appl Environ Microbiol 75, 7789–7796 (2009). 24. Jensen, H. M., TerAvest, M. A., Kokish, M. G. & Ajo-Franklin, C. M. CymA and Exogenous Flavins Improve Extracellular Electron Transfer and Couple It to Cell Growth in Mtr-Expressing Escherichia coli. ACS Synth Biol 5, 679–688 (2016). 25. Marritt, S. J. et al. A functional description of CymA, an electron-transfer hub supporting anaerobic respiratory flexibility in Shewanella. Biochem J 444, 465–474 (2012). 151 26. Myers, C. R. & Myers, J. M. Cloning and sequence of cymA, a gene encoding a tetraheme cytochrome c required for reduction of iron(III), fumarate, and nitrate by Shewanella putrefaciens MR-1. J Bacteriol 179, 1143–1152 (1997). 27. Coursolle, D. & Gralnick, J. A. Reconstruction of Extracellular Respiratory Pathways for Iron(III) Reduction in Shewanella oneidensis Strain MR-1. Front Microbiol 3, (2012). 28. Coursolle, D., Baron, D. B., Bond, D. R. & Gralnick, J. A. The Mtr Respiratory Pathway Is Essential for Reducing Flavins and Electrodes in Shewanella oneidensis. J Bacteriol 192, 467–474 (2010). 29. Tefft, N. M. & Teravest, M. A. Reversing an Extracellular Electron Transfer Pathway for Electrode-Driven Acetoin Reduction. ACS Synth Biol 8, 1590–1600 (2019). 30. Ross, D. E., Flynn, J. M., Baron, D. B., Gralnick, J. A. & Bond, D. R. Towards Electrosynthesis in Shewanella: Energetics of Reversing the Mtr Pathway for Reductive Metabolism. PLoS One 6, e16649 (2011). 31. Rowe, A. R. et al. Identification of a pathway for electron uptake in Shewanella oneidensis. Commun Biol 4, 957 (2021). 32. Rowe, A. R. et al. Tracking Electron Uptake from a Cathode into Shewanella Cells: Implications for Energy Acquisition from Solid-Substrate Electron Donors. mBio 9, (2018). 33. TerAvest, M. A. & Ajo‐Franklin, C. M. Transforming exoelectrogens biotechnology using synthetic biology. Biotechnol Bioeng 113, 687–697 (2016). for 34. Lienemann, M. et al. Towards patterned bioelectronics: facilitated immobilization of exoelectrogenic Escherichia coli with heterologous pili. Microb Biotechnol 11, 1184– 1194 (2018). 35. TerAvest, M. A., Zajdel, T. J. & Ajo‐Franklin, C. M. The Mtr Pathway of Shewanella oneidensis MR‐1 Couples Substrate Utilization to Current Production in Escherichia coli. ChemElectroChem 1, 1874–1879 (2014). 36. Baruch, M., Tejedor-Sanz, S., Su, L. & Ajo-Franklin, C. M. Electronic control of redox reactions inside Escherichia coli using a genetic module. PLoS One 16, e0258380 (2021). 37. Stookey, L. L. Ferrozine---a new spectrophotometric reagent for iron. Anal Chem 42, 779–781 (1970). 38. R Core Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing Preprint at https://www.R-project.org/ (2021). 39. H. Wickham. ggplot2: Elegant Graphics for Data Analysis. Preprint at (2016). 152 40. Hadley Wickham, Romain François, Lionel Henry & Kirill Müller. dplyr: A Grammar of Data Manipulation. R package version 1.0.9 Preprint at https://CRAN.R- project.org/package=dplyr (2022). 41. Garnier, S. et al. viridis(Lite) - Colorblind-Friendly Color Maps for R. (2023). 42. Ulrich, J. TTR: Technical Trading Rules. (2021). 43. Auguie, B. gridExtra: Miscellaneous Functions for ‘Grid’ Graphics. Preprint at (2017). 44. Förster, A. H. & Gescher, J. Metabolic Engineering of Escherichia coli for Production of Mixed-Acid Fermentation End Products. Front Bioeng Biotechnol 2, (2014). 45. Song, C. W., Kim, D. I., Choi, S., Jang, J. W. & Lee, S. Y. Metabolic engineering of Escherichia coli for the production of fumaric acid. Biotechnol Bioeng 110, 2025– 2034 (2013). 46. Meshulam-Simon, G., Behrens, S., Choo, A. D. & Spormann, A. M. Hydrogen Metabolism in Shewanella oneidensis MR-1. Appl Environ Microbiol 73, 1153–1165 (2007). 47. Kane, A. L. et al. Formate Metabolism in Shewanella oneidensis Generates Proton Motive Force and Prevents Growth without an Electron Acceptor. J Bacteriol 198, 1337–1346 (2016). 48. Yang, Y. et al. Enhancing Bidirectional Electron Transfer of Shewanella oneidensis by a Synthetic Flavin Pathway. ACS Synth Biol 4, 815–823 (2015). 49. Coursolle, D., Baron, D. B., Bond, D. R. & Gralnick, J. A. The Mtr Respiratory Pathway Is Essential for Reducing Flavins and Electrodes in Shewanella oneidensis. J Bacteriol 192, 467–474 (2010). 50. Ford, K. C. & TerAvest, M. A. The electron transport chain of Shewanella oneidensis MR-1 can operate bidirectionally to enable microbial electrosynthesis. Appl Environ Microbiol 90, (2024). 51. Madsen, C. S. & TerAvest, M. A. NADH dehydrogenases Nuo and Nqr1 contribute in transfer by Shewanella oneidensis MR-1 to extracellular electron bioelectrochemical systems. Sci Rep 9, 14959 (2019). 52. Ford, K. C. & TerAvest, M. A. The electron transport chain of Shewanella oneidensis MR-1 can operate bidirectionally to enable microbial electrosynthesis. Appl Environ Microbiol 90, (2024). 53. Lee, S. H. S. Importance of redox balance on the production of succinic acid by metabolically engineered Escherichia coli. Appl Microbiol Biotechnol 58, 286–290 (2002). 153 54. Goldbeck, C. P. et al. Tuning Promoter Strengths for Improved Synthesis and Function of Electron Conduits in Escherichia Coli. ACS Synth Biol 2 (3), 150–159 (2013). 55. Arslan, E. et al. Overproduction of The Bradyrhizobium Japonicum C-Type Cytochrome Subunits of Thecbb3Oxidase In Escherichia coli. Biochem Biophys Res Commun 251 (3), 744–747 (1998). 56. Jensen, H. M. et al. Engineering of a Synthetic Electron Conduit in Living Cells. PNAS 107 (45), 19213-19218, (2010). 154 CHAPTER 6: CONCLUSION 155 OBJECTIVES This dissertation presents original work toward developing S. oneidensis and E. coli as improved hosts for MES. Both organisms have strengths that make them attractive options for creating improved MES strains (pages 9 and 20). Currently, MES strains comprised of acetogens or mixed cultures are limited in their product scope, off-target production, and small genetic toolbox.1–3 Acetate remains the main product for both biocatalysts, and off-target methane production plagues mixed-culture MES due to methanogen populations.1 Since these organisms are not as well understood as model organisms like S. oneidensis and E. coli, there remain open questions for how to divert production away from acetate and toward more valuable products. Additionally, their limited tools for genetic engineering cause roadblocks for implementing what little insights do exist for improving product production and scope.1–6 S. oneidensis is a promising MES host as it is a model organism for EET, having well characterized metabolome, proteome, transcriptome, and genome.7–12 The Mtr pathway which drives both OET and IET has been the subject of many characterization and practical studies.13–21 Researchers have a good understanding of how to manipulate EET in S. oneidensis to provide desired system modifications. Additionally, there is some evidence that S. oneidensis has reversible formate dehydrogenases which can catalyze CO2 reduction under cathodic potentials.22 This is important because it suggests that if formate assimilation can be engineered in this organism, it could use CO2 as a carbon and energy source. The objective of Chapter 2 was to engineer a formate-tolerant strain of S. oneidensis as a first step toward developing formatotrophy in this organism. Chapter 3 aimed to characterize Rnf, an important enzyme which connects the ferredoxin and NADH 156 pools that could be a source of low potential electrons for CO2 fixation. Finally, the influence of a common inducer molecule used in strain development was characterized in Chapter 4. E. coli is a model organism with a vast number of pathways engineered for fermentative biosynthesis of value-added product production,23–27 which could be expressed in a MES-compatible strain. This would immediately greatly expand the product scope of MES. E. coli has previously been engineered to fix CO2, showing that this feat is achievable in this organism.28–30 Additionally, EET was introduced in strain MFe444.31 Together these two engineering strategies could be used to generate an MES-compatible E. coli host. Chapter 5 investigates the potential for MFe444 to be used in a BES using a fermentable carbon source. CONTRIBUTION OF PRESENTED WORK TOWARD DEVELOPMENT OF S. ONEIDENSIS AS A MES HOST This dissertation contributes toward the development of S. oneidensis as a MES host by describing the evolution of a formate-tolerant strain, characterization of electron transfer enzyme Rnf, and phenotype assessment of IPTG on the wild-type strain. In Chapter 2, a formate-tolerant strain of S. oneidensis is described. As formate is toxic to this organism, a formate-tolerant strain could improve future efforts at development of formatotrophy. The formate tolerant strains were generated through adaptive laboratory evolution in minimal lactate medium with increasing concentrations of formate. After just 30 days, eight strains were obtained which were tolerant to 100 mM formate. The sequencing of individual colonies from each evolution line revealed that two mutations appeared to be aiding in formate tolerance. The first was a substitution in sodium- 157 dependent bicarbonate transporter encoded by SO_3758. This transporter had a substitution in the binding pocket which brought the predicted binding affinity of formate much closer to the predicted binding affinity of the native substrate bicarbonate. This points to the role of this alteration being improved formate efflux, thus preventing toxic accumulation of formate in the cells. Another gene whose mutation improved formate tolerance was SO_1320. This gene had different mutations in different strains affecting protein residues V106I, V106F, and S52I. As a protein of unknown function, it is unclear what benefit these substitutions provide. The protein product is expected to localize to the cell membrane, which could indicate transporter activity or aid in membrane integrity. This first step toward developing a formate tolerant strain of S. oneidensis could pave the way for further development as a formatotroph. Development of an MES host could come from a formatotrophic strain through the CO2 reductase activity of formate dehydrogenase or through the electrochemical reduction of CO2 in the MES reactor. The S. oneidensis Rnf complex was characterized in Chapter 3, as the Rnf complex is a respiratory enzyme with potential ties to the EET pathways. The Rnf complex connects the ferredoxin and NADH pools and could be an important source of low-potential electrons for CO2 fixation. 32 This study found that in S. oneidensis growing on minimal lactate medium, Rnf operates in the reduced ferredoxin-forming direction and is powered by PMF. Importantly, a connection between Rnf and the iron-sulfur cluster formation and repair enzyme cysteine desulfurase, IscS, was established. IscS is a reduced ferredoxin- dependent enzyme which contributes to the synthesis and repair of cellular iron-sulfur clusters. This is activity which influences numerous enzymes, and the effects propagate throughout the cell. This work explored the metabolomic profile of Rnf and IscS CRISPRi 158 knockdown strains and found that they similarly accumulate MEP pathway metabolites. As the MEP pathway requires both ISC proteins (IspG and IspH) and reduced ferredoxins, it is a great example for how Rnf activity affects ISC biosynthesis and repair. We posit that the effect of Rnf is two-pronged across species. First is that Rnf generates reduced ferredoxins for enzymes requiring it as a cofactor. This includes both ISC proteins such as Nif, SoxR, and IspG and non-ISC proteins such as Fpr and IscS. The second prong is that Rnf provides reduced ferredoxins for ISC biosynthesis and repair. Disruption to this process is propagated throughout the cell, impacting activities of both ferredoxin- dependent ISC proteins such as Nif, SoxR, and IspG, as well as non-ferredoxin-dependent ISC proteins such as Nth, Nuo, and Acn. This connection of Rnf to multiple pathways can explain how its activity has been attributed to the function of many seemingly unrelated enzymes in the literature. These findings indicate that Rnf could be an important source of reduced ferredoxins for the EET pathway as well, as some evidence exists for reduction of the quinone pool by reduced ferredoxin-dependent enzymes in other species.33,34 The effect of IPTG on the growth of wild-type S. oneidensis on NAG was discussed in Chapter 4. S. oneidensis cannot use most six carbon sugars, including glucose, as a substrate because of missing genes for glucose permease, glucose kinase, and 6- phosphofructokinse.35,36 However, NAG is one of the few six carbon sugars that can be metabolized by this organism through conversion of the sugar to fructose 6-phosphate which enters metabolism at glycolysis. 37 Because of this, NAG is commonly used as a sugar substrate for S. oneidensis during phenotype assessments of engineered strains. The work described in Chapter 4 shows that the common inducer molecule IPTG enhances growth of wild-type S. oneidensis on NAG. This is an important effect for all 159 researchers developing engineered S. oneidensis strains because it could lead to false positive correlations between a heterologous protein and the desired outcome. This is especially important when developing an MES strain because one of the main goals for MES is to generate high titers of value-added products. Expression of heterologous genes from an IPTG-inducible promoter is an inefficient method for scaled-up MES processes because plasmid maintenance through the addition of antibiotics and induction of gene expression through the addition of IPTG can dramatically raise costs. For design and screening purposes, this is the most efficient way to test the utility of various genomic additions. However, a strain suitable for scale-up would require the transfer of the gene to the genome and preferably under the control of a constitutive promoter. If a false positive correlation between the preliminary engineered strain and improved titers is due to the inducer, both time and funding will be spent on the development of an inducer-free strain, which would not yield the same results. Therefore, it is imperative to understand the interactions of inducers on the native system. IPTG does impact growth of S. oneidensis on NAG under both aerobic and anaerobic conditions, although the effect is more pronounced under anaerobic conditions. However, IPTG does not appear to affect the strain’s growth on other common substrates such as lactate and acetate. FUTURE DIRECTIONS FOR DEVELOPING S. ONEIDENSIS AS A MES HOST The analysis of the formate tolerant strains indicates that short-term evolution in increasing formate concentrations may not be enough to evolve a formate-assimilating mechanism. Rather, because improved formate efflux grants the cells greater tolerance, it appears that these strains highlight the toxicity of formate to S. oneidensis. A longer 160 evolution with smaller steps in formate concentration may be required for the evolution of a strain with formate-consuming ability. Alternatively, creating an S. oneidensis strain which requires formate for growth could be a promising approach. A formate-dependent strain could be evolved to improve its formate utilization to the point of supporting biomass production form the once-toxic molecule. Replication of Davies’ results22 that S. oneidensis formate dehydrogenases can catalyze CO2 reduction under cathodic potentials is an important next step for developing a MES-compatible strain. If formatotrophy can be evolved or engineered in S. oneidensis, the conversion of the formatotrophic strain to an autotrophic strain relies upon the capability of reducing CO2 to formate. Anaerobic pre-culturing of the formatotrophic strain would be imperative as formate dehydrogenases have been shown to be oxygen sensitive. Formate production from CO2 would require cultivation under cathodic potentials, thus providing electrons for the reduction reaction. An increase in current upon flushing the BES with CO2 would indicate that CO2 reduction is occurring and could be confirmed by monitoring formate levels via HPLC during the course of the experiment. Should a formatotrophic or autotrophic strain of S. oneidensis be successfully created in the future, it is imperative to test for IPTG-correlated changes in metabolism once again. This is necessary to determine if IPTG is a suitable inducer to test strains engineered for value-added product production from formate and CO2. Additional studies on the Rnf complex and participation of reduced ferredoxins in EET could be beneficial for generating a more complete picture of electron transfer in S. oneidensis. Since the S. oneidensis electron transport chain is a branched pathway, it would be unsurprising to find a yet unidentified link connecting low-potential electrons from 161 ferredoxin to the EET pathway. In this work, Rnf activity was characterized during aerobic growth on minimal lactate medium. For a more comprehensive understanding of this complex, additional characterization using other minimal media substrates, rich media, and anoxic conditions can be done. Rnf activity is expected to be reversible, and this characteristic could be observed through the oxidation of ferredoxin under different conditions. Our examination of Rnf began with the hypothesis that it could be beneficial for engineering autotrophy in this organism. Therefore, additional experiments should be conducted which examines the feasibility of using reduced ferredoxins to drive CO2 reduction in S. oneidensis. Repetition of experiments described in this work for Rnf characterization should also be performed in E. coli. To date, the only function Rnf is known to have in E. coli is for its participation in SoxR reduction. In this work, we present a model for Rnf activity across species in that it is essential for ISC biosynthesis and repair. Testing of this model in other organisms is essential for confirmation of this model. CONTRIBUTAION OF PRESENTED WORK TOWARD DEVELOPMENT OF E. COLI AS A MES HOST As one of the most well-studied microbial production hosts, conversion of E. coli to a MES host would bring along a plethora of pathways for implementing CO2 conversion to value-added products.23–27 E. coli strain MFe444 was previously engineered to participate in EET via heterologous expression of the Mtr pathway.31 Testing of this strain has only been carried out with lactate, a non-fermentable carbon and electron source, in order to display proof-of-concept that the pathway is functional.31 Without electron acceptors, E. coli ferments glucose. Fermentation balances redox reactions to gain energy in an 162 environment which makes respiration impossible. This careful balance of fermentation products is a major constraint in E. coli fermentative biosynthesis reactions. Therefore, EET in E. coli serves two functions. The first is to allow for unbalanced fermentation, relieving redox balance constraints and opening avenues for greater product yields. The second is to connect the cells to an electrode, a necessary step for engineering this host for MES. Chapter 5 discusses work that explores the biotechnological potential of the strain by using the fermentable sugar glucose as a carbon and electron donor in both anodic and cathodic electro-fermentation in BESs. This work demonstrates that MFe444 can use an anode as an electron sink via OET, drawing electrons away from the production of reduced fermentation products lactate and succinate and toward production of oxidized fermentation products formate and acetate. Alternatively, MFe444 can use a cathode as an electron donor via IET, influencing the fermentation product ratios toward the reduced products and away from the oxidized products. This control of fermentation product ratios could be an important tool in optimizing E. coli as an MES host because it removes the redox balance constraint which is present in standard E. coli fermentative biosynthesis reactions. This could make MES attractive not only for its conversion of CO2 and clean electricity, but also for potentially higher titers than the same molecule production by traditional fermentation. FUTURE DIRECTIONS FOR DEVELOPING E. COLI AS A MES HOST MFe444 appears to be a promising host for electro-fermentation. However, it does not have CO2-fixing capabilities, making it ill-suited for MES at this stage. To create an MES-compatible E. coli strain, it is necessary to connect the EET capabilities of MFe444 to the CO2 fixing capabilities of other engineered strains. One possible way to accomplish 163 this is to add the Mtr and Ccm vectors from MFe444 to previously engineered CO2-fixing E. coli. This would be the quickest and most efficient way to test whether the combination of these activities allows for product production from CO2 in a BES. This approach would be more efficient than engineering CO2 production in E. coli, because as discussed in Chapter 5, the engineering of CO2 fixation in a heterotrophic organism is a challenging feat. On the contrary, EET can be transferred to E. coli by transforming two vectors into the host. Should the transfer of EET via expression of MtrCAB, CymA, and Ccm proteins create a strain capable of MES, the next step would be integration of these proteins in the genome. This would create a more stable strain with no antibiotic requirements, thus keeping production costs down. As the cytochrome maturation proteins used in MFe444 are those from E. coli, replacing the native promoter on the genome with a stronger constitutive promoter could eliminate the need for another copy of the genes elsewhere in the genome. CONCLUDING REMARKS MES could be a powerful technology for aiding in the turn away from fossil fuels. While current biocatalysts are limited in their product scope and potential for genetic engineering, development of additional strains could be the answer this technology needs to become industrially relevant. This dissertation outlines the progress toward development of more versatile and robust MES biocatalysts from S. oneidensis and E. coli. Future building upon this work could have great potential for aiding in the fight against climate change. 164 REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. Prévoteau, A., Carvajal-Arroyo, J. M., Ganigué, R. & Rabaey, K. Microbial electrosynthesis from CO2: forever a promise? Curr Opin Biotechnol 62, 48–57 (2020). Rosenbaum, M. A. & Henrich, A. W. Engineering microbial electrocatalysis for chemical and fuel production. Curr Opin Biotechnol 29, 93–98 (2014). Schiel-Bengelsdorf, B. & Dürre, P. Pathway engineering and synthetic biology using acetogens. FEBS Lett 586, 2191–2198 (2012). Arends, J. B. A., Patil, S. A., Roume, H. & Rabaey, K. Continuous long-term electricity-driven bioproduction of carboxylates and isopropanol from CO2 with a mixed microbial community. J CO2 Util 20, 141–149 (2017). Ganigué, R., Puig, S., Batlle-Vilanova, P., Balaguer, M. D. & Colprim, J. Microbial electrosynthesis of butyrate from carbon dioxide. Chem Comm 51, 3235–3238 (2015). Raes, S. M. T., Jourdin, L., Buisman, C. J. N. & Strik, D. P. B. T. B. Continuous Long‐Term Bioelectrochemical Chain Elongation to Butyrate. ChemElectroChem 4, 386–395 (2017). Pinchuk, G. E. et al. Pyruvate and Lactate Metabolism by Shewanella oneidensis MR-1 under Fermentation, Oxygen Limitation, and Fumarate Respiration Conditions. Appl Environ Microbiol 77, 8234–8240 (2011). Beliaev, A. S. et al. Global Transcriptome Analysis of Shewanella oneidensis MR-1 Exposed to Different Terminal Electron Acceptors. J Bacteriol 187, 7138–7145 (2005). Leaphart, A. B. et al. Transcriptome Profiling of Shewanella oneidensis Gene Expression following Exposure to Acidic and Alkaline pH. J Bacteriol 188, 1633– 1642 (2006). 10. Pinchuk, G. E. et al. Constraint-Based Model of Shewanella oneidensis MR-1 Metabolism: A Tool for Data Analysis and Hypothesis Generation. PLoS Comput Biol 6, e1000822 (2010). 11. Cao, X., Qi, Y., Xu, C., Yang, Y. & Wang, J. Transcriptome and metabolome responses of Shewanella oneidensis MR-1 to methyl orange under microaerophilic and aerobic conditions. Appl Microbiol Biotechnol 101, 3463–3472 (2017). 12. Heidelberg, J. F. et al. Genome sequence of the dissimilatory metal ion–reducing bacterium Shewanella oneidensis. Nat Biotechnol 20, 1118–1123 (2002). 13. Brutinel, E. D. & Gralnick, J. A. Shuttling happens: soluble flavin mediators of 165 extracellular electron transfer in Shewanella. Appl Microbiol Biotechnol 93, 41–48 (2012). 14. Kouzuma, A. Molecular mechanisms regulating the catabolic and electrochemical activities of Shewanella oneidensis MR-1. Biosci Biotechnol Biochem 85, 1572– 1581 (2021). 15. Richardson, D. J. et al. The ‘porin–cytochrome’ model for microbe‐to‐mineral electron transfer. Mol Microbiol 85, 201–212 (2012). 16. Beblawy, S. et al. Extracellular reduction of solid electron acceptors by Shewanella oneidensis. Mol Microbiol 109, 571–583 (2018). 17. Kane, A. L. et al. Formate Metabolism in Shewanella oneidensis Generates Proton Motive Force and Prevents Growth without an Electron Acceptor. J Bacteriol 198, 1337–1346 (2016). 18. Madsen, C. S. & TerAvest, M. A. NADH dehydrogenases Nuo and Nqr1 contribute to extracellular electron transfer by Shewanella oneidensis MR-1 in bioelectrochemical systems. Sci Rep 9, 14959 (2019). 19. Ford, K. C. & TerAvest, M. A. The electron transport chain of Shewanella oneidensis MR-1 can operate bidirectionally to enable microbial electrosynthesis. Appl Environ Microbiol 90, (2024). 20. Tefft, N. M. & Teravest, M. A. Reversing an Extracellular Electron Transfer Pathway for Electrode-Driven Acetoin Reduction. ACS Synth Biol 8, 1590–1600 (2019). 21. Tefft, N. M., Ford, K. & TerAvest, M. A. NADH dehydrogenases drive inward electron transfer in Shewanella oneidensis MR‐1. Microb Biotechnol 16, 560–568 (2023). 22. Davies, J. Characterisation of the reversible formate dehydrogenases of Shewanella. (University of East Anglia, 2017). 23. Pontrelli, S. et al. Escherichia coli as a host for metabolic engineering. Metab Eng 50, 16–46 (2018). 24. Hoesl, M. G. & Budisa, N. Recent advances in genetic code engineering in Escherichia coli. Curr Opin Biotechnol 23, 751–757 (2012). 25. Idalia, V.-M. N. & Bernardo, F. Escherichia coli as a Model Organism and Its Application in Biotechnology. in Escherichia coli - Recent Advances on Physiology, Pathogenesis and Biotechnological Applications (ed. Samie, A.) 253–274 (IntechOpen, Rijeka, 2017). doi:10.5772/67306. 26. Huang, T. & Ma, Y. Advances in biosynthesis of higher alcohols in Escherichia coli. 166 World J Microbiol Biotechnol 39, 125 (2023). 27. Yang, D., Park, S. Y., Park, Y. S., Eun, H. & Lee, S. Y. Metabolic Engineering of Escherichia coli for Natural Product Biosynthesis. Trends Biotechnol 38, 745–765 (2020). 28. Antonovsky, N. et al. Sugar Synthesis from CO2 in Escherichia coli. Cell 166, 115– 125 (2016). 29. Herz, E. et al. The genetic basis for the adaptation of E. coli to sugar synthesis from CO2. Nat Commun 8, 1705 (2017). 30. Gleizer, S. et al. Conversion of Escherichia coli to Generate All Biomass Carbon from CO2. Cell 179, 1255-1263.e12 (2019). 31. Jensen, H. M., TerAvest, M. A., Kokish, M. G. & Ajo-Franklin, C. M. CymA and Exogenous Flavins Improve Extracellular Electron Transfer and Couple It to Cell Growth in Mtr-Expressing Escherichia coli. ACS Synth Biol 5, 679–688 (2016). 32. Biegel, E., Schmidt, S., González, J. M. & Müller, V. Biochemistry, evolution and physiological function of the Rnf complex, a novel ion-motive electron transport complex in prokaryotes. CMLS 68, 613–634 (2011). 33. St. Maurice, M. et al. Flavodoxin:Quinone Reductase (FqrB): a Redox Partner of Pyruvate:Ferredoxin Oxidoreductase That Reversibly Couples Pyruvate Oxidation to NADPH Production in Helicobacter pylori and Campylobacter jejuni. J Bacteriol 189, 4764–4773 (2007). 34. Bojko, M. & Wieckowski, S. NADPH and ferredoxin:NADP+ oxidoreductase- dependent reduction of quinones and their reoxidation11Preliminary results were presented at the 10th FESPP Congress, Florence, Italy (Plant Physiol. Biochem., Special issue, 1996, p. 99). Phytochemistry 50, 203–208 (1999). 35. Chubiz, L. M. & Marx, C. J. Growth Trade-Offs Accompany the Emergence of Glycolytic Metabolism in Shewanella oneidensis MR-1. J Bacteriol 199, (2017). 36. Rodionov, D. A. et al. Genomic encyclopedia of sugar utilization pathways in the Shewanella genus. BMC Genomics 11, 494 (2010). 37. Yang, C. et al. Comparative Genomics and Experimental Characterization of N- Acetylglucosamine Utilization Pathway of Shewanella oneidensis. J Biol Chem 281, 29872–29885 (2006). 167