Understanding the Redox Metabolism of Pyrococcus Furiosus For

UNDERSTANDING THE REDOX METABOLISM OF PYROCOCCUS FURIOSUS FOR

EFFICIENT METABOLIC ENGINEERING

DIEP MINH NGOC NGUYEN

(Under the Direction of Michael W. W. Adams)

ABSTRACT

The hyperthermophilic archaeon Pyrococcus furiosus has been a valuable source for thermostable and thermoactive enzymes for many biotechnological applications. With the availability of its genome sequence and a genetic system, P. furiosus became the subject for metabolic engineering projects to produce various alcohols and industrially relevant chemicals.

Proofs of concepts have been established to show that P. furiosus could be engineered to produce ethanol, butanol and 3-hydroxypropionate. These fuel synthesis pathways were completed by insertion of one or more genes from foreign donors into the P. furiosus genome. Product synthesis was controlled in a temperature-dependent manner, in which the host was subjected to suboptimal growth temperatures to induce product formation. However, improvements in yield are necessary to advance P. furiosus as a platform relevant for industrial applications. To optimize yield in the current engineered P. furiosus strains, a deeper understanding of the electron metabolism at the suboptimal temperatures and knowledge of the host’s redox flux will be beneficial to redirect carbon and electron flow toward desired products. In this study, at temperature ranging from 70 to 80C, P. furiosus was shown to produce acetoin as a side-product of the biosynthesis of the branched chain amino acids. This work established that another industrially relevant compound could be produced from this organism. Additionally, we also were able to redirect carbon from acetoin toward ethanol production by deleting the gene encoding for an acetolactate synthase

(ALS). Deletion of ALS increased the ethanol yield by 50% in the ethanol producing strain. This result implied that the glycolytic enzyme pyruvate ferredoxin oxidoreductase (POR) was most likely to be the bottleneck step in the current pathways for ethanol production. The second aspect of this study focuses on how electron flow is regulated in P. furiosus. We investigated the physiological functions of the two ferredoxin:NADP+ oxidoreductases, NfnI and Xfn. The bifurcating NfnI is responsible for NADPH generation by utilizing both reduced ferredoxin and

NADH as electron donors. Xfn harbors a similar physiological function to NfnI, however, via a different mechanism. Xfn does not catalyze the same reaction as NfnI and whether it is another type of electron bifurcation is to be determined. The mechanism by which NfnI utilizes reduced ferredoxin and NADH to generate NADPH involves novel flavin and iron sulfur cluster biochemistry and understanding it in detail will not only be beneficial to control NfnI catalysis in metabolically engineered strains, but also provide insights into designing synthetic enzymes or bioinspired catalysts.

INDEX WORDS: Pyrococcus furiosus, archaea, anaerobe, hyperthermophile, biotechnology,

metabolic engineering, electron bifurcation, biofuels

UNDERSTANDING THE REDOX METABOLISM OF PYROCOCCUS FURIOSUS FOR

EFFICIENT METABOLIC ENGINEERING

DIEP MINH NGOC NGUYEN

BS, Virginia Polytechnic Institute and State University, 2013

A Dissertation Submitted to the Graduate Faculty of The University of Georgia in Partial

Fulfillment of the Requirements for the Degree

DOCTOR OF PHILOSOPHY

ATHENS, GEORGIA

2018

Diep Minh Ngoc Nguyen

UNDERSTANDING THE REDOX METABOLISM OF PYROCOCCUS FURIOSUS FOR

EFFICIENT METABOLIC ENGINEERING

DIEP MINH NGOC NGUYEN

Major Professor: Michael W. W. Adams Committee: William Lanzilotta Yajun Yan

Electronic Version Approved:

Suzanne Barbour Dean of the Graduate School The University of Georgia August 2018

DEDICATION

To my mother who loves and supports me unconditionally, who always believes in the choice I make in this life. There are things she wished she could have done differently, I hope that raising me to the person I am today is not one of those.

ACKNOWLEDGEMENTS

There are many who have helped me through my development from a mere student into a scientist. Throughout the 24 years of being students, I have encountered the right persons that helped and showed me the way. Special thanks go to Dr. Michael W. W. Adams for his support, oversight, and insight throughout my graduate school career. I also would like to thank my committee members, Dr. William Lanzilotta and Dr. Yajun Yan for their help and advice during my committee meetings. Big thanks and appreciation go to Dr. Gina Lipscomb and Dr. Gerrit

Schut, my two special mentors. They have worked with me closely on my research projects, taught me the knowledge I need and above all, became friends who also support me emotionally. It was my luck to be able to work under their wings. Moreover, I have to thank the entire Adams lab, especially my manager Farris Poole, for their individual help, technical and emotional support.

They have become friends that I can count on and enjoy working together. I would like to express my gratitude to Dr. Timothy Larson, my undergraduate advisor, who first taught me and showed me that I could be a scientist.

For me to be who I am today, my deepest thanks and appreciation to my mother-Tuyet Thi

Nguyen, my father-Hanh Minh Nguyen, my step-father-Tuan Anh Nguyen, my sister-Tram Minh

Bao Nguyen and my fiancée-Arthur Boon Teck Ong for their unconditional love and support.

Without them, I could have lost my way. And lastly, to my pet Daiki, he has worked very hard to keep me sane during my graduate school career.

“The only way of finding the limits of the possible is going beyond them into the impossible”

-Arthur C. Clarke-

TABLE OF CONTENTS

Page

ACKNOWLEDGEMENTS ...... v

LIST OF TABLES ...... x

LIST OF FIGURES ...... xi

CHAPTER

1 INTRODUCTION AND LITERATURE REVIEW ...... 1

1.1 Introduction to archaea ...... 1

1.2 Pyrococcus furiosus and its metabolism ...... 3

1.3 Electron bifurcation ...... 12

1.4 Research objectives ...... 17

1.5 Tables and figures ...... 19

2 TEMPERATURE-DEPENDENT ACETOIN PRODUCTION BY PYROCOCCUS

FURIOSUS IS CATALYZED BY A BIOSYNTHETIC ACETOLACTATE

SYNTHASE AND ITS DELETION IMPROVES ETHANOL PRODUCTION ...... 37

2.1 Introduction ...... 39

2.2 Materials and methods ...... 41

2.3 Results and discussion ...... 46

2.4 Conclusion ...... 55

2.5 Figures...... 56

vii

3 TWO FUNCTIONALLY DISTINCT NADP+-DEPENDENT FERREDOXIN

OXIDOREDUCTASES MAINTAIN THE PRIMARY REDOX BALANCE

OF PYROCOCCUS FURIOSUS ...... 76

3.1 Introduction ...... 78

3.2 Experimental procedures ...... 81

3.3 Results and discussion ...... 91

3.4 Tables and figures ...... 106

4 THE BIFURCATING NFN OF PYROCOCCUS FURIOSUS: A SIMPLE MODEL

TO DETERMINE THE MECHANISM OF FLAVIN-BASED ELECTRON

BIFURCATION ...... 150

4.1 The bifurcating center of P. furiosus NfnI: L-FAD ...... 151

4.2 The high potential electron transfer pathway...... 153

4.3 The low potential electron transfer pathway ...... 154

4.4 Protein dynamics and the effect of cluster-ligands on bifurcating activity ...155

4.5 Figures...... 160

5 DISCUSSION AND CONCLUSION...... 178

5.1 Optimizing P. furiosus for biofuel and chemical production ...... 180

5.2 NfnI as model for an efficient bioinspired catalyst ...... 183

5.3 Conclusion and outlook ...... 189

5.4 Figures...... 191

REFERENCES ...... 197

APPENDICES

viii

A SINGLE GENE INSERTION DRIVES BIOALCOHOL PRODUCTION BY A

THERMOPHILIC ARCHAEON...... 220

A1 Introduction ...... 223

A2 Experimental procedures...... 226

A3 Results and discussion ...... 232

A4 Tables and figures ...... 240

References ...... 271

LIST OF TABLES

Page

Table 1.1: Summary of characterized electron-bifurcating enzymes ...... 19

Table 3.1: Strains used and constructed in this study ...... 106

Table 3.2: Purification of NfnII ...... 107

Table 3.3: Data collection and refinement statistics ...... 108

Table 3.4: Activities of NfnI and NfnII ...... 110

Table S3.1: Primers designed and used recombinant strains in this study ...... 111

Table S3.2: Doubling times for growth of recombinant strains...... 113

Table S3.3: Nfn bifurcating activity detected in whole cell extracts ...... 114

Table S3.4: NfnII catalytic bias of FNOR activities ...... 115

Table A.S1: Strains used in this study ...... 248

Table A.S2: Primers used in this study ...... 249

LIST OF FIGURES

Page

Figure 1.1: Sugar metabolism of P. furiosus ...... 22

Figure 1.2: Peptide metabolism of P. furiosus ...... 24

Figure 1.3: Hydrogen metabolism of P. furiosus ...... 26

Figure 1.4: Sulfur metabolism of P. furiosus ...... 28

Figure 1.5: The alcohol producing AOR/AdhA pathway in P. furiosus ...... 30

Figure 1.6: Schematic of electron bifurcation and electron confurcation ...... 32

Figure 1.7: The Q-cycle in the respiratory Complex III ...... 34

Figure 1.8: Redox substrates and their relative reductive potentials in flavin-based electron

bifurcation ...... 36

Figure 2.1: P. furiosus produces acetoin at sub-optimal growth temperatures ...... 57

Figure 2.2: Acetoin formation via ALS and its genomic context ...... 59

Figure 2.3: ALS activity is temperature-dependent ...... 61

Figure 2.4: ALS deletion abolishes acetoin formation and improves ethanol production in the

ADHA strain ...... 63

Figure 2.5: ALS deletion improves ethanol yield in the ADHA strain ...... 65

Figure 2.6: Expression of ALS restores acetoin production in the ΔALS strain ...... 67

Figure 2.7: Acetoin formation is increased in the presence of excess pyruvate ...... 69

Figure 2.8: ALS is required for branched chain amino acid biosynthesis ...... 71

Figure S2.1: Phylogenetic tree based on the amino acid sequences of ALS found in different

bacterial and archaeal species ...... 73

Figure S2.2: α-acetolactate decarboxylase activity was not detected in P. furiosus ...... 75

Figure 3.1: Phylogenetic reconstruction of a subset of Nfn homologs encoded by Archaea ...... 117

Figure 3.2: Growth analysis ...... 119

Figure 3.3: Redox nucleotide pool analysis ...... 121

Figure 3.4: Native and chemical cross-linking mass spectrometry of NfnI and NfnII ...... 123

Figure 3.5: Structure of NfnII ...... 125

Figure 3.6: Cartoon models of NfnI and Xfn ...... 127

Figure S3.1: Phylogenetic reconstruction of 467 archaeal and bacterial Nfn ...... 129

Figure S3.2: Network analysis of proteins encoded by gene flanking NfnS ...... 131

Figure S3.3: Network analysis of multiple isoforms of Nfn ...... 133

Figure S3.4: Expression analysis of the up- and downstream genes of nfnI ...... 135

Figure S3.5: Expression analysis of NfnI and NfnII...... 137

Figure S3.6: Redox nucleotide pool analysis ...... 139

Figure S3.7: Expression analysis of genes encoding for enzymes involved in the redox

metabolism ...... 141

Figure S3.8: Identification of protein components in Nfn complexes ...... 143

Figure S3.9: The omit electron density map of NfnII ...... 145

Figure S3.10: The simple composite omit electron density map of NfnII...... 147

Figure S3.11: Multiple sequence alignment of NfnI, NfnII and NfnIII ...... 149

Figure 4.1: Structure of NfnI and cofactors orientation ...... 161

Figure 4.2: The isoalloxazine ring of FAD ...... 163

xii

Figure 4.3: Square wave voltammograms from NfnI ...... 165

Figure 4.4: Evidence for a short-lived anionic semiquinone of L-FAD in NfnI ...... 167

Figure 4.5: Evidence for the formation of a long-lived neutral semiquinone at S-FAD ...... 169

Figure 4.6: The energy landscape of NfnI ...... 171

Figure 4.7: Putative co-evolving residues that connect the NAD(P)(H) and ferredoxin binding

sites through allostery ...... 173

Figure 4.8: The iron-sulfur clusters of NfnI contains non-cysteine ligands ...... 175

Figure 4.9: Activity of NfnI mutants ...... 177

Figure 5.1: The proposed pathway for 2.3-butanediol production in P. furiosus ...... 192

Figure 5.2: Proposed engineering of P. furiosus for improving biofuel and biochemical

production ...... 194

Figure 5.3: Hydrogen-bonding network around N5 of L-FAD in NfnI ...... 196

Figure A1: Sugar fermentation coupled to alcohol production by P. furiosus strain A ...... 241

Figure A2: Formation of ethanol from sugars by engineered P. furiosus strains ...... 243

Figure A3: Reduction of organic acids to alcohols by P. furiosus strain A ...... 245

Figure A4: CO as source of reductant for conversion of organic acids to alcohols by P. furiosus

strain A/Codh ...... 247

Figure A.S1: Maps of plasmids used in this study...... 252

Figure A.S2: Formation of 13C-ethanol from 13C-acetate by P. furiosus strain A ...... 254

Figure A.S3: Effect of aor deletion on ethanol formation from maltose ...... 256

Figure A.S4: Conversion of sugars to ethanol and butyrate to butanol by P. furiosus ...... 258

Figure A.S5: The T. onnurineus Codh complex and vector for insertion of the Codh expression

construct into the P. furiosus chromosome ...... 260

xiii

Figure A.S6: Reduction of organic acids to alcohols by Pyrococcus furiosus strain A/Codh in the

absence of CO ...... 262

Figure A.S7: Reduction of organic acids to alcohols by Pyrococcus furiosus strain A/Codh cell

suspensions...... 264

Figure A.S8: Substrate specificity and temperature optimum of T. X514 AdhA ...... 266

Figure A.S9: Michaelis-Menten kinetics of T. X514 AdhA ...... 268

Figure A.S10: Temperature dependence of the AOR/AdhA synthetic pathway ...... 270

xiv

CHAPTER 1

INTRODUCTION AND LITERATURE REVIEW

1.1 Introduction to archaea

There are three kingdoms of life, known as Eukarya, Bacteria, and Archaea (1-3). Their classification was based on sequence comparisons of the small-subunit ribosomal RNA (2). From the phylogenetic tree, the domain Archaea was found to be more closely related to eukaryotes than bacteria. Both archaea and eukaryotes have similar mechanisms of information processing, such as DNA replication, RNA transcription, and protein translation, although these processes are generally more complex in eukaryotes (4). On the other hand, archaea are similar to bacteria in that they lack the nuclei and membrane-bound organelles and both domains are exclusively single- cell. Even though the central information processes are different, archaea and bacteria are very similar in term of their metabolic processes (5-7).

Archaea are often found in extreme environments, such as hydrothermal vents, solfataric fields, hot springs, saturated salt seas and even hot oil wells (8-10). They are believed to be the

“living fossils” of the most ancient form of life on earth (3). The kingdom of archaea is further divided into five phyla, the Crenarchaeota, Euryarchaeota, Nanoarchaeota, Thaumarcheaota, and Korarchaeota with the majority of archaeal organism belong to the first two phyla (2,11-15).

The Euryarchaeota phylum is among the best-studied within the Archaea kingdom. It consists of the order Archaeoglobales (sulfate-reducers), the Methanococcales (methanogens) and the

Thermococcales (sulfur-reducer) (16-18). The Thermococcales has the highest number of

characterized isolates and is made up of three genera, namely, Pyrococcus, Thermococcus and

Paleococcus. Organisms belonging to the phylum Crenarchaeota are mostly thriving at temperature above 80C, thus they are considered as hyperthermophiles. The Euryarchaerota, however, has a wider range of growth temperature, containing organisms ranging from mesophiles to hyperthermophiles. There are only two genera of bacteria that contain hyperthemophilic species:

Aquifex and Thermotoga, yet they cannot thrive at temperature above 90C. Only the archaeal species within the order Thermococcales are known to live at such extreme temperatures. (19).

From the evolutionary perspective, archaea are thought to be an ancient life-form from several billion years ago when the earth was still “hot” and anoxic (3). From the biotechnology perspective, hyperthermophilic enzymes are of keen interest due to their intrinsic thermostability and robust activities that able to tolerate in industrial settings (20-22). Some of their most famous applications are in copper biomining (23) and the use of the high-fidelity DNA polymerases in

Polymerase Chain Reaction (PCR), which is arguably the most important innovation in molecular biology (24). Engineering hyperthermophilic organisms to be hosts for the production of commodity chemicals has recently become possible when genetic systems were developed in some species. Metabolic engineering is an essential tool to produce biofuels and chemicals in well- studied organisms such as Escherichia coli and Saccharomyces cerevisiae. However, heterologous expression of hyperthermophilic enzymes in these popular hosts poses challenges because of the temperature differences between the gene donors and the hosts. To overcome this problem, the hyperthermophilic Pyrococcus furiosus has been engineered to produce several useful products, including ethanol, butanol, and 3-hydroxypropionate (3-HP) (25-28). It is now considered as the

E. coli version of archaeal hyperthermophiles. However, the yields of these products on the small scales in the laboratory do not yet meet industrial standards and at present they are not

economically profitable. This is understandable because research with both archaeal and hyperthermophilic organisms is lagging decades behind their bacterial counterparts. Therefore, establishing hyperthermophiles as a platform for metabolic engineering will only be possible with a better understanding of their metabolic processes, redox metabolism and improved genetic tools.

P. furiosus is the most promising of all the hyperthermophiles in terms of available genetic tools and unique biological characteristics. The metabolism and current status of engineering P. furiosus will be discussed in following section.

1.2 Pyrococcus furiosus and its metabolism

Pyrococcus furiosus is a strict anaerobic archaeon that was isolated from geothermally- heated marine vents off the coast of Vulcano Island, Italy by Karl Stetter. P. furiosus grows between 70 C and 103 C and thrives optimally near 100C. P. furiosus can convert a wide range of substrates, such as starch, maltose, peptides and yeast extract to organic acids, CO2, and H2 in

0 0 the absence of S or H2S in the presence of S , which serves as the terminal electron acceptor (29).

Sugar metabolism in P. furiosus

P. furiosus metabolizes sugars via a modified Embden-Meyerhof glycolytic pathway

(Figure 1.1) which converts one molecule of glucose into two molecules of pyruvate, which are further converted into two molecules of acetate, two molecules of CO2, and produce two ATP (30).

The fundamental differences between this and the conventional glycolysis pathway are the electron acceptor and the net ATP yield from substrate level phosphorylation. In P. furiosus, the first step involves the phosphorylation of glucose to glucose-6-phosphate by glucokinase (GK), but this is an ADP-dependent enzyme rather than an ATP-dependent one (31). The oxidation of

glyceraldehyde-3-phosphate (GAP) to 3-phosphoglycerate is catalyzed by a tungsten-containing enzyme glyceraldehyde-3-phosphate oxidoreductase (GAPOR), a ferredoxin (Fd) utilizing enzyme. This is in contrast to the conventional pathway which has glyceraldehyde-3-phosphate dehydrogenase (GAPDH), a NAD+-dependent enzyme (30,32). GAPOR replaces the functions of both GAPDH and phosphoglycerate kinase (PGK) in the conventional glycolysis (32). As a result, ferredoxin serves as the only electron acceptor, and no net ATP is generated as glucose is converted to pyruvate by P. furiosus. Pyruvate is further oxidized by pyruvate oxidoreductase (POR), another ferredoxin-dependent enzyme in P. furiosus, which catalyzes the oxidative decarboxylation of pyruvate to acetyl-CoA (33). Therefore, the conversion of one mole of glucose to 2 moles of acetyl-CoA also yields two moles of CO2 and eight equivalents of electrons in the form of reduced ferredoxin. ATP is generated via acetyl-CoA synthase I and II (ACSI/II), which convert acetyl-CoA to acetate coupled with substrate level phosphorylation of ADP to ATP (34).

These ACS type enzymes are unique to the archaea. The reaction ACS catalyzes is accomplished by two different enzymes in bacteria, a phosphate acetyltransferase and an acetate kinase (35).

Overall, Fd is the only reductant generated from glucose oxidation in P. furiosus.

Peptide metabolism in P. furiosus

When peptides are used as the carbon source for growth, the metabolism of P. furiosus is shifted to produce different end products compared to sugar metabolism and elemental sulfur (S0) is required as the terminal electron acceptor (29,36). Amino acids derived from peptides are deaminated by the appropriate transaminase to generate 2-keto acids (Figure 1.2). There are four different types of 2-keto acid oxidoreductases to further metabolize these 2-keto acids into their corresponding acyl CoA derivatives, specifically, pyruvate by POR, aromatic 2-keto acids by

indolepyruvate oxidoreductase (IOR), branched-chain 2-keto acids by 2-ketoisovalerate oxidoreductase (VOR) and 2-ketoglutarate by 2-ketoglutarate oxidoreductase (KGOR) (37-40).

All of these reactions are coupled to reduction of Fd. Generation of 2-ketoglutarate from glutamate is catalyzed by glutamate dehydrogenase (GDH), one of the most abundant enzymes in P. furiosus making up to approximately 20% of the cytoplasmic proteins (41). However, NADPH is produced by the GDH reaction, therefore, Fd and NADPH are both generated during peptide metabolism.

This is in contrast to sugar metabolism where Fd is the only major redox carrier.

Ferredoxin and P. furiosus redox metabolism

Ferredoxins are small redox protein (6-12 kDa), and contain one [2Fe-2S] cluster or one or more [4Fe-4S] clusters (42,43) with reduction potential (E0) typically below -420 mV (44). Fd from many anaerobic microorganisms can have reduction potentials as low as -500 mV (45). Fd from P. furiosus (Pf) was first characterized in 1989 and still remains one of the most thermostable

Fd known. Pf-Fd was purified as a homodimer, where the monomer contains one [4Fe-4S] cluster that has one non-cysteine ligand. Pf-Fd has been purified to contain various type of FeS cluster, namely, [4Fe-4S], [3Fe-4S] and even [2Fe-2S] with the latter two as the results of aerobic isolation

(46,47). Under anaerobic purification, Pf-Fd retains its [4Fe-4S] cluster with the mid-point potential determined by EPR-monitored redox titration at pH 8 to be approximately -345 mV (46).

However, under physiological condition, the potential of Fd is estimated to be near -480 mV (48).

In P. furiosus, Fd is the primary electron carrier under standard sugar fermentation condition (Figure 1.3), where all of the reductant is used to generate reduced Fd and it is oxidized by the membrane-bound [NiFe] hydrogenase (MBH) to evolve hydrogen gas (H2) and form a

Sodium ion (Na+) gradient. This is coupled to a Na+-dependent ATP synthase that overall generates

~1.2 ATP per glucose (49). Deletion of the catalytic subunit of MBH rendered P. furiosus unable

0 to grow in the absence of S as a terminal electron acceptor and H2 production was abolished (50).

This result reinforces the idea that the physiological function of MBH is to get rid of excess reductant in the form of H2 and to play a key role in energy conservation (49). Since only reduced

Fd is produced during sugar fermentation, it is not clear how NADPH was generated for biosynthesis. P. furiosus also harbors two soluble [NiFe] hydrogenases (SHI and SHII), one of

+ which is able to oxidize H2 and reduce NADP in vitro (51). SHI has an order of magnitude higher specific activity than SHII and originally, it was hypothesized to be the main mechanism to produce NADPH for biosynthesis (51). However, deletion of SHI did not produce any growth phenotype suggesting that this is not the role of this enzyme under standard growth conditions

(50). Reduced Fd can also be utilized by the cytosolic enzymes ferredoxin NADP+ oxidoreductase

(FNOR) I and II to generate NADPH for biosynthesis (52,53). This class of enzyme has now been renamed as NADH dependent ferredoxin:NADP+ oxidoreductase I and II (NfnI and NfnII) (54) and they will be the subject of Chapter 3 and Chapter 4 of this study.

When S0 is present in the growth medium of P. furiosus, the sulfur response regulator,

SurR, acts as both activator and repressor to turn on and off the expression of genes relevant for the growth of P. furiosus (55-57). For example, the expression of MBH is turned off, and the expression of an MBH homolog, the membrane-bound oxidoreductase complex MBX is switched on. The production of H2 therefore is shut down and hydrogen sulfide (H2S) is produced instead by presumably the MBX complex as an alternative mechanism to dispose of excess reductant (58).

Deletion of the gene encoding for MBX catalytic subunit rendered P. furiosus unable to grow in

0 the presence of S and very little H2 was produced, consistent with the proposed role of MBX in energy conservation under S0 reducing conditions (59). The expression of SHI and SHII are also

down-regulated in the presence of S0, hence they cannot be involved in redox metabolism. SurR acts as the repressor of genes encoding for NfnI but activates the expression of genes encoding for

NfnII (57). However, the expression of NfnI and NfnII is also dependent on the carbon source used for growth (sugars versus peptides), and this is further discussed in Chapter 3. SurR also activates another important NADPH utilizing enzyme, NADPH sulfur reductase (NSR) (57). NSR was thought to be the major H2S-evolving enzyme together with MBX. NSR has a preference for

NADPH over NADH for the reduction of S0 (60). However, a knockout of genes encoding for

NSR resulted in a mutant strain with a growth profile similar to that of the parent strain (59). The proposed relationship between these sulfur-activating enzymes is shown in Figure 1.4. Reduced

Fd produced via the oxidation of keto acids to their respective Co-A derivative is used to reduce

0 + S to H2S principally by the MBX complex. Fd and NADH (generated via the NAD salvage pathway) serve as electron donors to generate NADPH by the bifurcating complex NfnI and II, which in turn can be used for biosynthesis. The functions of NfnI and II are discussed in more detail in chapter 3.

P. furiosus as host for production of biofuels and biochemical

With the availability of a genetic system, P. furiosus has become an attractive host for metabolic engineering in which carbon and electron flow are redirected into pathways for biofuel and chemical production. To date, most efforts have been involved heterologous expression of enzymes from other thermophilic organisms in P. furiosus to form pathways for biofuel production, such as ethanol and butanol, or to biologically synthesize industrially relevant chemicals, such as 3-hydroxypropionate (3-HP). For metabolic engineering purposes, there are not many organisms that grow near 100C that could be used as gene donors for P. furiosus. Therefore,

less thermophilic organisms were used to provide a wider array of genes, but this requires the host to be grown at suboptimal temperatures (~70C) for the synthetic pathways to function properly.

For example, insertion of the genes encoding for a lactate dehydrogenase from

Caldicellulosiruptor bescii (optimal growth temperature of 75C), using a P. furiosus cold shock response promoter PcipA, allowed P. furiosus to produce lactate at 72C (61). Thus, temperature dependent product formation has been a core strategy for metabolic engineering work in P. furiosus. Also utilizing this strategy, P. furiosus has been engineered to produce ethanol (2 g L-1) at 72C using the end product of the host’s central metabolism – acetate (25). Acetate was reduced to acetaldehyde via the unusual ferredoxin-dependent aldehyde oxidoreductase (AOR) of P. furiosus. The single insertion of an alcohol dehydrogenase A (AdhA) from Thermoanaerobacter sp X514, which grows optimally at 70C, into the P. furiosus genome completed a simple but effective pathway to further convert acetaldehyde to ethanol. This pathway is independent of ATP consumption and solely dependent on reduced Fd (from AOR) and NADPH (from AdhA) (Figure

1.5). Furthermore, the AOR/AdhA pathway can reduced a wide range of organic acids to their corresponding alcohols (Appendix A, Figure A3) (25).

P. furiosus natively cannot metabolize carbon monoxide (CO) as an energy source, yet the insertion of a 16 gene cluster from Thermococcus onnurineus, which encodes for a membrane- bound carbon monoxide dehydrogenase (CODH) into P. furiosus allowed for CO oxidation to be coupled to Fd reduction, generating CO2 and H2 (62). T. onnurineus grows optimally at 80C and so the recombinant P. furiosus strain was also grown at the same temperature. This strain readily oxidized CO and produced H2 at 80C in growth medium that was limiting in both maltose and yeast extract (0.5 and 1 g L-1, respectively) (62). Expression of this CODH complex into P. furiosus

strain containing AOR/AdhA pathway allowed for production of ethanol, driven by CO as carbon and energy source (Appendix A, Figure A4) (25).

A more complex pathway has also been inserted into the genome of P. furiosus with the goal of utilizing carbon dioxide (CO2), a greenhouse gas, to generate butanol, a biofuel (27), and

3-HP (26), an industrial plastic precursor. The 3-HP pathway branches from the central metabolite acetyl-CoA and consists of other enzymes, utilizing an ATP dependent acetyl-CoA carboxylase and two NADPH-dependent reductases to produce 3-HP (26,63,64). The recombinant P. furiosus strain contained five genes encoding for three enzymes, taken from Metallosphaera sedula, which grows optimally at 73C. P. furiosus was able to produce 0.05 g L-1 3-HP at 72C in a small scale fermentation in the laboratory (26). The 3-HP yield was doubled by deletion of ACSI and ACSII to remove the competing pathway as these enzymes convert acetyl-CoA to acetate (65). Expressing additional accessory genes for the carboxylase function along with improved bioreactor conditions increased the yield of 3-HP to 0.5 g L-1 (63,64). Another artificial operon containing genes encoding six enzymes taken from three different donors: Thermoanaerobacter tengcongensis,

Spirochaeta thermophila and Thermoanaerobacter sp. X514, was inserted into P. furiosus for the generation of 1-butanol (27). These three bacteria grow optimally in the range 60-72C. This pathway, however, is dependent on NADH and it requires 4 NADH per butanol equivalent. At

60C, this strain produces ~ 0.07 g L-1 butanol after 48h of growth, which is the highest reported temperature for butanol production to date (27).

The future of P. furiosus as a platform for chemical production is promising, however, it is still at early stage. For these engineered pathways to be industrially relevant, the yield of product must be at the grams per litter scale and there is much room for improvement. These synthetic pathways are highly temperature-dependent since the optimal growth temperatures for the

organisms used as gene donors are 20-40C below that of the host, which is near 100C. Hence, the optimal temperature for production of ethanol and 3-HP is 72C (25,26), and it is 60C for the production of butanol (27). Because all of these pathways operate at temperatures lower than that for the optimal growth of P. furiosus, it is possible that the metabolism of P. furiosus is also affected at the lower temperatures in term of carbon and electron flow. It has been shown that ethanol, butanol and 3-HP production can be controlled by a temperature shift strategy (26,27,61).

This strategy entails growing the host near its optimal temperature until the end of log phase and then the temperature is rapidly shifted down to the pathway operation temperature (60 - 75C) to allow for heterologous protein production and product formation while significantly slowing down the host metabolism (61). However, it has been shown in vitro that the activities of the enzymes

POR, ACSI, ACSII and AOR are decreased at suboptimal temperatures and the effect of this on overall metabolism is not known (32,33,40). Another point worth noting is that the formation of biofuels and commodity chemicals by P. furiosus depends on reducing equivalents in the form of reduced Fd, NADH and NADPH, which could compete with the host’s metabolism and the biosynthesis necessary to sustain growth, albeit at a greatly decreased level at the lower temperatures. Therefore, one must gain more insights into how the host functions at suboptimal temperatures and how the redox carriers are regulated internally. Such information could be exploited for redirecting carbon and electron flow into the desired production pathway.

Current status of acetoin production in microbes

Acetoin or 3-hydroxybutanone is a four-carbon compound that is used as a building block for many bio-based chemicals (66). Acetoin is utilized as a flavor additive and preservative in the food industry. Fossil feedstocks are the common source used to chemically synthesize acetoin.

However, in the food and cosmetic industries, ecofriendly and biologically-produced acetoin from microbes is preferred despite its higher price tag (67). Thus, efforts have been made to increase acetoin yield by genetic engineering. In microbes, acetoin production depends on a two-enzyme pathway. Acetolactate synthase (ALS) catalyzes the reduction of pyruvate to acetolactate, which is then further converted to acetoin by acetolactate decarboxylase (ALDC). In lactic acid producing bacteria, acetoin serves not only as an internal pH neutralizer, but also an additional energy storage compound (68). The common strategies to overproduce acetoin is to overexpress ALS and ALDC and eliminate by-products such acetate, ethanol and lactate. Greatest effort has been made with natural acetoin producers (Bacillus, Klebsiella and Clostridium strains), or in E. coli and S. cerevisiae to co-produce acetoin and 2,3-butanediol or butanol and acetoin (67,69,70). 2,3- butanediol is produced via decarboxylation of acetoin by 2,3-butanediol dehydrogenase (Bdh), a reaction that is dependent on NADH. Typically, the genes encoding for ALS, ALDC and Bdh are located in the same operon and under the regulation of the LysR family transcriptional activators

(71).

The majority of work to produce acetoin has been done using mesophiles. The highest production yield was reported to be 100 g L-1 in fed-batch fermentation using S. cerevisiae strain

JHY605 as the host (72). This acetoin producing strain was extensively engineered, including deletion of four adh and a bdh gene to eliminate alcohol and 2,3-butanediol production, overexpression of als genes to improve acetoin production and insertion of an NADH oxidase to relieve the cofactor imbalance during sugar fermentation (72). However, in general, production of chemicals in mesophilic microbes is prone to batch contamination and some efforts have been made to isolate or engineer thermophilic strains to produce acetoin and/or 2,3-butanediol at elevated temperatures. For example, a Geobacillus strain, growing optimally at around 50C, has

been reported to produce up to 8 g L-1 acetoin naturally (73). Additionally, a thermophilic Baccilus strain is able to produce 50 g L-1 acetoin and 32 g L-1 2,3-butanediol, reaching 94% theoretical yield, in an optimized and pH controlled fermentation at 50C (74). Clearly, a much deeper understanding of the metabolism is needed in order to reach comparable level of product formation. At present, P. furiosus naturally produced 0.5 g L-1 at 75C (75).

1.3 Electron bifurcation

Electron bifurcation, in a nutshell, is a biochemical term to describe the splitting of a hydride into two electrons, one at a more positive and one at a more negative reduction potential than that of the electron pair (Figure 1.6). The reverse reaction is called electron confurcation, in which the bifurcating cofactor is reduced by two separate electrons from two separate cofactors.

This process is also referred to as the third mechanism of energy conservation, besides generation of a proton motive force and by substrate level phosphorylation.

The principle of electron bifurcation was first used by Peter Michell in 1975 to explain the so-called Q cycle, in which the respiratory complex III coupled quinol (QH2) oxidation with the reduction of quinone (Q) in the respiratory transport chain to achieve an overall thermodynamically favorable electron transfer (ET) reaction (76). This work was the foundation of quinone-based electron bifurcation. During the Q cycle (Figure 1.7), the first step occurs at the outer membrane where ubihydroquinone Q0 (QH2) is oxidized (Em= +90 mV) and one electron is transferred to high potential cytochrome c (Em= +250 mV) and the other electron goes to the low potential cytochrome bL (Em= -60 mV) and subsequently to the high potential cytochrome bH (Em=

+82 mV) and finally to UQ at the inner membrane, to yield the one electron reduced QH• state. A

• second round of electron flow is necessary to further reduce QH to QH2, which then diffuses

outside the membrane and the Q cycle repeats. One full Q cycle results in an additional gradient of 2H+/2e- and doubles the amount of energy conserved (77).

Until 2008, electron bifurcation was thought to be a unique phenomenon only applicable to the Q cycle. However, Wolfgang Buckel and Rolf Thauer discovered that electron bifurcation was, in fact, widespread in anaerobic metabolism (78). Compared to energy conservation in aerobes, the conversion of glucose to CO2 and H2O via the Embden-Meyerhof pathway, the Krebs cycle and the mitochondrial respiratory chain results in 38 equivalents of ATP, but glucose catabolism by anaerobic microorganisms can only yield approximately 2.5 ATP (77). Anaerobes

+ must have utilized very low potential reductant for the reduction of H , CO2 and N2 in the form of reduced Fd rather than NAD(P)H. How Fd is reduced by NAD(P)H remained an open question because the reduction potential of ferredoxin (Em= -480 mV) is lower than that of NAD(P)H (Em=

-320 mV (79), thus making it an endergonic reaction. It was proposed by Buckel’s group in 2008 that such a reaction was possible by coupling the reduction of Fd by NADH to a second and exergonic reaction, with a third substrate all catalyzed by a single enzyme. They proposed that the reduction of crotonyl-CoA (Em= -10 mV) by NADH (an exergonic reaction) is coupled to the reduction of Fd by NADH by a cytoplasmic electron-transfer flavoprotein, EftAB-butyryl-CoA dehydrogenase (78). Energetically, it was proposed that the energy released from fatty acid oxidation was harnessed to drive the endergonic half-reaction. The enzyme that catalyzes this reaction contains only flavin redox cofactors, hence, it was concluded that the splitting of the electron pair must happen at the flavin site, thus the name flavin-based electron bifurcation (FBEB) was proposed (78).

For electron bifurcation to happen, a protein complex must have the following three properties: (1) a bifurcating center or cofactor that is capable of both 1 and 2 electron transfers, (2)

two electron conduit or electron transfer pathways, and (3) the hydride donor or acceptor must be the at an intermediate potential (Figure 1.8). To date, twelves different types of bifurcating flavoenzymes have been characterized (Table 1.1). These enzymes differ in the number of subunits, their cofactor contents and the redox potential ranges of their substrates (Table 1.1 and

Figure 1.8). Sequence analysis of these bifurcating complexes reveals two common themes. They contain at least one FAD bound subunit and these enzymes are all involve in ferredoxin-dependent reactions (80). The cofactor composition ranges from a flavin only complex (the butyryl-CoA dehydrogenase-electron-transferring flavoprotein, Bcd-EtfAB), to one that harbors flavin and contains strings of FeS clusters and also another metal center, which can be the Fe only H cluster or [NiFe]-cofactor of the bifurcating hydrogenases, or tungsten- or molybdenum-pterins. In term of subunit composition, bifurcating complexes consist of at least two subunits (Nfn) and up to six subunits (the hydrogenase-heterodisulfide reductases).

Besides the butyryl-CoA dehydrogenase complex described above, there are other Eft- related enzymes that were found to bifurcate. These include the caffeyl-CoA dehydrogenase

(CarCDE) and lactate dehydrogenase (LctBCD). They all consist of three subunits, with CarDE and LctBC homologous to EtfAB (81,82). EftAB is the core of bifurcation, in which the FAD of

EtfB is proposed to be the bifurcating center, and the third subunit Bcd, CarD and LctD, enable the reduction of crotonyl-CoA, caffeyl-CoA, and lactate, respectively. In methanogens, low potential reduced Fd is generated via a hydrogenase-heterodisulfide reductase (MvhADG-

HdrABC). This enzyme bifurcates electrons from H2 (Em = -414 mV) to reduce heterodisulfide

(Em = -414 mV) and Fd (83). The HdrABC module that serves as the core for bifurcation, similar to the role played by the EftAB. Most characterized bifurcating complexes are cytoplasmic enzymes with the exception of the Etf (or Etf-like) complex that harbor an additional subunit C to

serve as the membrane anchor as seen in the FixABCX in Azotobacter vinelandii. This EftC subunit contains the binding site for a quinone-type compound. The physiological role of

FixABCX is thought to generate the low potential reduced Fd, from NADH and quinone, for nitrogenase (84). Demonstration of FixABCX bifurcating activity in vitro shows that bifurcating complexes are not restricted to only anaerobic organisms. To date, crystal structures of four types of bifurcating enzyme have been solved: the butyryl-CoA dehydrogenase-electron-transferring flavoprotein (Bcd-EtfAB) from Clostridium difficile (85), caffeyl-CoA reductase (CarCDE) from

Acetobacterium woodii (86), NADH-dependent ferredoxin:NADP oxidoreductase (NfnSL) from

Thermotoga maritima (54) and Pyrococcus furiosus (87), and the [NiFe]-hydrogenase- heterodisulfide reductase (MvhADG-HdrABC) complex from Methanothermococcus thermolitotrophicus (88). In this study, we focus on the Nfn enzyme from Pyrococcus furiosus, describing their roles in anaerobic metabolism (Chapter 3) and their bifurcating mechanisms

(Chapter 4).

NADH-dependent ferredoxin:NADP oxidoreductase

In the early 1970s, Thauer’s group reported that the NADPH-dependent reduction of Fd required the presence of NAD+ (Equation 1.1) and NADH-dependent reduction of NADP+ in the presence of reduced Fd (Equation 1.2) in cell extracts of Clostridium kluyveri (89). This indicated the presence of a ferredoxin:NADP reductase type enzyme with transhydrogenase activity in cell extracts (90). With the availability of the C. klyuveri genome in 2008 (91), a homolog of a plant ferredoxin:NADP reductase (FNR) was identified and later purified and found to catalyze the reversible reduction of Fd and NAD+ by 2 NADPH (Equation 1.3), hence the name NADH- dependent ferredoxin:NADP oxidoreductase or Nfn (92).

+ + + NADPH + 2 Fdox → NADP + 2Fdred + H (dependent of NAD ) (1.1) + + NADH + NADP → NAD + NADPH (dependent of Fdred) (1.2)

+ + 2 Fd(ox) + NAD + 2 NADPH  Fd(red) + NADH + 2 NADP (1.3)

The first structurally characterized Nfn was from the hyperthermophilic T. maritima

(54,93) and the structure of NfnI from P. furiosus was reported soon afterward (87,94). Nfn consists of two subunits, the large subunit (L) of ~50 kDa and the small subunit (S) of ~30 kDa.

This enzyme contains two FAD cofactors, one in each subunit (termed S-FAD and L-FAD), and three iron-sulfur (FeS) clusters, the [2Fe-2S] in the small subunit and the two [4Fe-4S] in the large subunit. The NAD(H) binding pocket is located near the S-FAD in the small subunit. The

NADP(H) binding site is near the L-FAD and ferredoxin binds near the distal [4Fe-4S] clusters in

NfnL. Based on this structure, it was proposed that the L-FAD is the site of bifurcation and the rest of cofactors form two electron transfer pathways to reduce the high and low potential substrates: NAD+ and Fd, respectively (54). The high potential pathway is formed by the [2Fe-2S] cluster and S-FAD, and the low potential pathway consists of the proximal and distal [4Fe-4S] clusters relative to L-FAD. Kinetically, it was demonstrated that NADH could not replace NADPH to serve as the hydride donor/acceptor and NADP+ could not replace NAD+ (54,93). Under physiological conditions where the NADPH/NADP+ ratio is greater than 1 and the NADH/NAD+ ratio is less than 1 (95), the reduction potential of NADP(H) and NAD(H) are close to -380 mV and -280 mV, respectively, compared to standard midpoint potential of -320 mV. Therefore, the

NADP(H) and NAD(H) binding pockets must have high specificities to ensure bifurcating or confurcating activity.

Nfn homologs are found in many anaerobic organisms where their subunit, and cofactor compositions are highly conserved (54,92,96). The P. furiosus genome contains two sets of genes

(PF1327-28 and PF1910-11) encoding for two NfnSL enzymes, termed NfnI and NfnII. NfnI was

previously purified and characterized as sulfide dehydrogenase (SuDH) in 1994 (97), in which it was reported to catalyze the reduction to polysulfide to H2S by using NADPH, and not NADH, as electron donor. Later in 2001, it was renamed ferredoxin:NADP oxidoreductase (FNOR) (53). In

2017, it was reclassified as a bifurcating Nfn, which catalyzed the reaction shown in equation 1.3

(87). Both NfnI and NfnII are under the control of SurR regulator, but while expression of NfnI is repressed in the presence of sulfur, expression of NfnII is activated (57). However, the functions of these two enzymes under different growth condition was not at all clear and our current understanding is discussed in Chapter 3.

1.4 Research objectives

The goal of this research is to gain insights into how carbon and electron flux in P. furiosus are affected by suboptimal growth temperatures and how redox metabolism is balanced. Efforts, to date, on engineering P. furiosus as a platform for biofuel and chemical production have relied on heterologous expression of genes to form hybrid or synthetic pathways in a temperature- controlled manner, which requires the host to be grown at suboptimal temperatures. Herein, we report that there is a bottleneck in sugar fermentation at suboptimal temperatures that causes accumulation of pyruvate, which leads to the formation of acetoin, and that this significantly impacts the AOR/AdhA strain in term of ethanol yield. Such findings are described in detail in

Chapter 2. Chapter 3 showcases the importance of the electron bifurcating Nfn complex in maintaining the redox metabolism in P. furiosus, as well as providing the first characterization of the second type of Nfn in this organism. With the recognition of the importance of electron bifurcation in anaerobic metabolism as a new mechanism of energy conservation, we describe in detail in Chapter 4 the mechanistic insights into flavin-based electron bifurcating provided by our

studies of NfnI. This is the simplest bifurcating model enzyme to date and provides an elegant means of controlling electron to overcome thermodynamically unfavorable reactions.

Understanding the physiological function and the bifurcating mechanism of Nfn will increase our understanding of redox control in general and how this can benefit metabolic engineering P. furiosus to produce chemicals. It will also provide principles for the design of biology-inspired redox catalysts for energy production.

1.5 Tables and Figures

Table 1.1 Summary of characterized electron-bifurcating enzymes

Enzyme Electron donor Electron acceptor Cofactor content

Butyryl-CoA dehydrogenase- Crotonyl-CoA, electron-transfering 2 NADH 3 FAD Fld/Fd flavoprotein Caffeyl-CoA reductase- 2 NADH Caffeyl-CoA, Fd 3 FAD, 2 [4Fe-4S] EtfAB Lactate dehydrogenase- 2 NADH Pyruvate, Fd 3 FAD, 2 [4Fe-4S EtfAB 3 FAD, 2 [4Fe-4S], UQ reductase-EftAB 2 NADH Ubiquinone, Fld riboflavin NADH-dependent 2 FAD, 1 [2Fe-2S], ferredoxin:NADP 2 NADPH NAD, Fd 2[4Fe-4S] oxidoreductase 1 FMN, 3 [2Fe-2S], [Fe-Fe]-hydrogenase 2 H2 NAD, Fd 6 [4Fe-4S], H- cluster1 1 FMN, 3 [2Fe-2S], [Fe-Fe]-hydrogenase-CO2 2 H2 NADP, Fd 6 [4Fe-4S], H- reductase cluster1, W-pterin 1 FMN, 4 [2Fe-2S], Formate dehydrogenase 2 HCOO- NAD, Fd 11 [4Fe-4S], Mo- pterin [NiFe]-hydrogenase- 1 FAD, 1 [2Fe-2S], 2H2 CoM-S-S-CoB, Fd heterodisulfide reductase 13 [4Fe-4S], [NiFe]2 2 FAD, 1 [2Fe-2S], Formate dehydrogenase- 2 HCOO- CoM-S-S-CoB, Fd 12 [4Fe-4S], Mo/W- hetero-disulfide reductase pterin

F420H2-dependent 1FAD, 1 [2Fe-2S], 7 2 F420H2 CoM-S-S-CoB, Fd heterodisulfide reductase [4Fe-4S] 2 FAD, 1 [2Fe-2S], 9 Methylene-H4 F reductase 2 NADH Methylene-H4F, ? [4 Fe-4S], 2 FMN

Modified from (80). 1 [FeFe] center containing six iron atoms 2 [NiFe] center containing one iron and one nickel atom

Figure 1.1 Sugar metabolism of P. furiosus

P. furiosus converts one molecule of glucose into 2 molecules of acetate, 2 molecule of CO2, 2 molecules of ATP, and 4 equivalents of reduced ferredoxin. The only ATP synthesis step in this pathway is by substrate level phosphorylation in the conversion of acetyl-CoA to acetate via acetyl-CoA synthetase (ACS) (34). Reduced ferredoxin is generated via glyceraldehyde-3- phosphate ferredoxin oxidoreductase (GAPOR) (32) and pyruvate ferredoxin oxidoreductase

(POR) (40).

Glucose ADP Glucokinase (GLK) AMP Glucose-6-phosphate Glucose-6-P isomerase (PGI)

Fructose-6-phosphate ADP Phosphofructokinase (PFK) AMP Fructose-1, 6-bisphosphatase Fructose-1, 6-bisphosphate aldolase (FBA)

Glyceraldehyde-3-phosphate Dihydroxyacetone phosphate 2e- Glyceraldehyde-3-P Fd oxidoreductase (GAPOR) 2 Fd 3-Phosphoglycerate Phosphoglycerate mutase

2-Phosphoglycerate Enolase H2O Phosphoenolpyruvate ADP Pyruvate kinase ATP Pyruvate CoASH Pyruvate Fd oxidoreductase (POR) 2e- 2 Fd Acetyl CoA + CO2 Pi + ADP (GDP) Acetyl-CoA synthase (ACS) CoASH + ATP (GTP) Acetate

Figure 1.2 Peptide metabolism of P. furiosus

Peptide metabolism by P. furisosus involves a series of peptidase and transaminase reactions convert peptides into individual amino acids and subsequently to 2-keto acids. Reduced ferredoxin is generated from the conversion of 2-keto acids to acyl-CoA, aldehydes, and CO2 via the four oxidoreductases, pyruvate ferredoxin oxidoreductase (POR), indolepyruvate oxidoreductase

(IOR), 2-ketoisovalerate oxidoreductase (VOR) and 2-ketoglutarate oxidoreductase (KGOR) (37-

40). Acyl-CoA and aldehydes are further reduced to their corresponding acids by acetyl-CoA synthase (ACS) (34) and aldehyde ferredoxin oxidoreductase (AOR) (100), produces ATP and forms reduced Fd. Glutamate is converted to 2-ketoglutarate by the NADPH dependent enzyme glutamate dehydrogenase (GDH) (101).

Peptides Peptidase Amino acids 2-ketoglutarate NADP Transaminases GDH Glutamate NADPH 2-keto acids Fd + CoASH CO ox POR 2 IOR Fdred VOR KGOR Acyl-CoAs Aldehydes ADP +Pi Fdox ACS AOR ATP + CoASH Fdred Acids Acids

Figure 1.3 Hydrogen metabolism of P. furiosus

During sugar fermentation, reduced ferredoxin generated from the glycolysis is oxidized by the membrane-bound hydrogenase (MBH) to evolve hydrogen gas (H2) (49). A sodium ion gradient is formed across the membrane and this is used by ATP synthase resulting in an additional 1.2

ATP produced per glucose molecule. Some of the H2 may be taken up by soluble hydrogenase I and II (SHI and SHII) to generate NAD(P)H (51). The bifurcating NADH-dependent ferredoxin:NADP oxidoreductases I also utilize reduced Fd and NADH to form NADPH for biosynthesis (see Chapter 3 for additional information).

Na+ Na+ outside MBH ATP Complex Synthase inside 2H+ Na+ + NAD(P)H H H2 Na SHI & 2 ADP ATP NAD(P) SHII 2H+ Fdox Fdred GAPOR POR Glucose Acetate NADPH PK + CO2

Fdred ACS NADP NfnI ADP ATP

Fdox NADH NAD

Figure 1.4 Sulfur metabolism of P. furiosus

During growth in the presence of elemental sulfur (S0), P. furiosus produces hydrogen sulfide

(H2S) instead of hydrogen gas (H2). A second membrane-bound oxidoreductase (MBX, with X predicted to be a sulfur contained compound) becomes active and replaces MBH to oxidize reduced ferredoxin (generated from sugar or peptide metabolism) and ultimately to reduce S0 to

H2S (59). Sulfide is also produced by a NADPH depended sulfur reductase (NSR) (60). NfnI has also been shown to contain S0 reduction activity in vitro (52,53). However, such activity is unlikely to be its physiological function, which is to produce NADPH from reduced Fd and NADH. The function of NfnII is not clear at present.

0 S Na+ Na+ outside MBX ATP Complex Synthase inside S0 + 0 Na S + H2S Na ADP ATP Fd NADPH red Fd Fdred NfnI ox GAPOR NADP Fdox POR Acetate NADH NAD Glucose PK + CO2 ACS

ADP ATP NAD(P)H H2S NSR NAD(P) S0

Figure 1.5 The alcohol producing AOR/AdhA pathway in P. furiosus

The novel synthetic AOR/AdhA pathway (25) takes advantage of acetate produced from glucose fermentation. Acetate is converted into acetaldehyde via aldehyde ferredoxin oxidoreductase

(AOR) using reduced ferredoxin as reductant. Acetaldehyde is further reduced to ethanol by

NADPH-dependent alcohol dehydrogenase (AdhA) from Thermoanaerobacter sp. X514, utilizing

NADPH as electron donor. This pathway has also been shown to convert various organic acids to their corresponding alcohol (25). This pathway is redox balanced because the reduced Fd and

NADPH are recycled by the GAPOR, POR and SHI reactions.

Glucose

MBH GAP 2H+ 2Fdox GAPOR 2Fdred H2

PEP ADP + Pi PK ATP Pyruvate 2Fd POR ox 2Fdred Acetyl CoA ADP + Pi ACS ATP Organic Acids Acetate 2Fd AOR red 2Fdox Aldehyde Acetaldehyde SHI NADPH AdhA H2 NADP 2H+ Alcohol Ethanol

Figure 1.6 Schematic of electron bifurcation and electron confurcation

Electron bifurcation refers to the reversible splitting of two electrons from a donor (D) into two separate pathways to each reduce a low potential acceptor (A1) and a high potential acceptor (A2) by a bifurcating complex (grey box). The combination of two electrons, one each from a low potential donor (D1) and a high potential donor (D2), to reduce a single two electron acceptor (A) is called electron confurcation.

A D1 1 -ve e- -ve e- 2e- 2e- D A - e e- D2 A2 +ve +ve Bifurcation Confurcation

Figure 1.7 The Q-cycle in the respiratory Complex III

Complex III was the first characterized electron bifurcating complex (76). The term electron bifurcation was used to describe the quinone (Q) cycle. When the Q0 site is reduced by QH2 in the membrane, the electron pair is split into two separate electron transfer pathways, the first high potential electron reduces the FeS cluster which in turn reduces cytochrome C1. The second low potential electron reduces cytochrome bL, then subsequently cytochrome bH and the Q1 site to yield

- a one electron reduced state Q• . A second round is necessary to fully reduce the Q1. QH2 is then released to the membrane matrix to complete one full electron bifurcating reaction. During the Q- cycle, an additional 2 protons and 2 electrons are released to the inner membrane to generate a proton gradient. This figure is modified from (102).

4 H+ 2 Cyt cox

2 e- IM FeS c1 2 Cyt cred 2 e-

2 QH2 Q 2 e- 2 Q 0 bL 2 e-

- Q 2 e bH Q1 QH2 M Complex III

(Cytochrome bc1) 2 H+

Figure 1.8 Redox substrates and their relative reductive potentials in flavin-based electron bifurcation

Flavin-based electron bifurcation (FBEB) utilizes flavin (FAD or FMN) site as the center of bifurcation. The bifurcating center is characterized by its ability to accept a hydride (a proton and two electrons) from a donor and facilitate electron transfer one each to the high and low potential acceptors. The number of electrons shown is the total electron involved in one full cycle of bifurcation. The midpoint potentials of donors and accepters are shown as numbers in brackets:

+ NAD(P)(H), H /H2, CO2/formate, pyruvate/lactate (103); caffeyl-CoA/dihydrocaffeyl-CoA, crotonyl-CoA/butyryl-CoA (104); F420/F420H2 (105); heterodisulfide/(CoMSH + CoBSH) (106); flavodoxin (107) and ferredoxin (108). Figure modified from (77).

Em, mV - Low potential acceptor Ferredoxin(-420) Flavodoxin (-420)

Electron donor

a Bifurcating center - i e t x

n Formate (-430) 2 e

t H2 (-414) o

P F420H2 (-360) Flavin -

x NADH (-320) 2 x 2e o

d NADPH (-320) 2 x e

e - High potential acceptor R

NAD+ (-320) Pyruvate (-190) CoM-S-S-CoB (-140) Menaquinone (-70) Caffeyl-CoA (-30) Crotonyl-CoA (-10) Ubiquinone (+90) +

CHAPTER 2

TEMPERATURE-DEPENDENT ACETOIN PRODUCTION BY PYROCOCCUS FURIOSUS

IS CATALYZED BY A BIOSYNTHETIC ACETOLACTATE SYNTHASE AND ITS

DELETION IMPROVES ETHANOL PRODUCTION

Nguyen, D. M. N., Lipscomb, G. L., Schut, G. J., Vaccaro, B. J., Basen, M., Kelly, R. M., and Adams, M. W. W. (2016) Temperature-dependent acetoin production by Pyrococcus furiosus is catalyzed by a biosynthetic acetolactate synthase and its deletion improves ethanol production. Metab Eng 34, 71-79 Reprinted here with permission of the publisher

ABSTRACT

The hyperthermophilic archaeon, Pyrococcus furiosus, grows optimally near 100°C by fermenting sugars to acetate, carbon dioxide and molecular hydrogen as the major end products.

The organism has recently been exploited to produce biofuels using a temperature-dependent metabolic switch using genes from microorganisms that grow near 70°C. However, little is known about its metabolism at the lower temperatures. We show here that P. furiosus produces acetoin

(3-hydroxybutanone) as a major product at temperatures below 80°C. A novel type of acetolactate synthase (ALS), which is involved in branched-chain amino acid biosynthesis, is responsible and deletion of the als gene abolishes acetoin production. Accordingly, deletion of als in a strain of P. furiosus containing a novel pathway for ethanol production significantly improved the yield of ethanol. These results also demonstrate that P. furiosus is a potential platform for the biological production of acetoin at temperatures in the 70 to 80°C range.

2.1 Introduction

Acetoin, or 3-hydroxybutanone, is an important four-carbon compound that serves as a building block for valuable bio-based chemical compounds and is a common flavor additive and preservative in the food industry (66). Acetoin can be chemically synthesized from fossil feedstocks, but biologically produced acetoin is preferable for use in the food and cosmetics industries despite its higher cost (67). The most efficient acetoin producers are species from the bacterial genera Bacillus, Paenibacillus, Serratia, Enterobacter and Klebsiella. These microorganisms produce acetoin via a well-characterized pathway that involves two key enzymes, acetolactate synthase (ALS)1 and acetolactate decarboxylase (ALDC), versions of which have been extensively characterized in Lactococcus lactis (109). Naturally produced acetoin has been in demand throughout the past decade, leading to the metabolic engineering of many Bacillus and

Clostridium strains for producing acetoin and butanediol or co-producing butanol and acetoin

(67,69). To date, most reports on acetoin and butanediol production involve mesophiles, with optimal growth temperatures ranging from 30-40°C, making the processes more prone to contamination (110). To reduce the risks of contamination during fermentation, thermophilic organisms have been considered. For example, a recently isolated strain of Geobacillus that grows optimally at 45-55°C produced up to 90 mM acetoin (73).

Herein, we report that the hyperthermophilic archaeon Pyrococcus furiosus produces acetoin in a temperature-dependent manner. P. furiosus is an obligate heterotroph that grows optimally near 100°C by fermenting simple and complex sugars to acetate, carbon dioxide and molecular hydrogen (29). With the recent availability of a genetic system (111), P. furiosus has been exploited and engineered to produce biofuels and desired products via a strategy based on a

temperature-dependent metabolic switch (25-27,61). This approach involves the expression of genes and pathways from organisms with growth optima almost 30°C lower than that of P. furiosus. As such, cultivation at these lower temperatures drives product formation from a heterologous pathway, while minimizing host metabolism, but not to the extent that cell function is completely compromised.

Using metabolic engineering hosts at temperatures substantially below their normal growth temperatures represents a new strategy for producing bio-based chemicals and fuels (10, 12).

Although P. furiosus is one of the best-studied hyperthermophiles, very little is known about its metabolic function at suboptimal growth temperatures. Genome-wide transcriptional analyses of cold-shocked and cold-adapted P. furiosus cells demonstrated that the expression of a surprising number of genes were either up-regulated or down-regulated upon a temperature shift from 98° to

72°C (112). However, no analyses of metabolites were performed during growth at the lower temperature. Herein, we demonstrate that P. furiosus produces acetoin as a major end product at growth temperatures below 85°C in a temperature-dependent manner, and that this is accomplished via a unique type of ALS whose gene expression and biochemical function are also temperature- dependent. Growth studies on deletion strains lacking ALS showed that it is necessary for branched-chain amino acid biosynthesis and, furthermore, that its absence can improve biofuel production in P. furiosus by redirecting the carbon flow toward ethanol.

2.2 Materials and methods

PCR product and vector construction-

A linear PCR construct was designed to delete the als gene (PF0935) from the P. furiosus genome. The pyrF pop-out marker cassette containing 65 bp identical flanking regions for marker removal via homologous recombination (26) was sandwiched between ~0.5 kb homologous flanking regions of the als gene. Each PCR product was obtained separately, and the full linear ~2 kb PCR construct was assembled via splicing by overlap extension and PCR (113).

The pDN005 vector was designed to insert an ALS overexpression cassette into the P. furiosus genome. The als expression cassette, consisting of the 182 bp slp promoter (26), the als gene (PF0935) and a 19 bp terminator sequence of the hpyA1 gene (5'-aatcttttttagcactttt), was assembled with the backbone vector pGL009 containing the pyrF popout cassette and ~0.5 kb homologous flanking regions targeting the intergenic space between convergent genes PF1232 and

PF1233 (termed genome region 5) (25). The pDN005 vector was assembled using Gibson

Assembly (NEB) (114) and was sequence-verified.

Strain construction

All strains used and constructed are listed in Table 2.1. The linear als deletion construct was transformed into P. furiosus COM1, as previously described (111). Transformants were cultured and purified twice on liquid and solid defined cellobiose medium (111). Genomic DNA was isolated using the ZymoBeadTM Genomic DNA Kit (Zymo Research) and isolates were screened by PCR, using primers that target outside of the homologous flanking regions. A strain

containing the als deletion was sequence-verified and designated MW250 (herein referred to as

ΔALS).

Markerless versions of the ADHA strain (25) and the markerless ΔALS strain were constructed by counter-selection for the loss of the pyrF marker cassette, as previously described

(115,116) to allow further genetic manipulation. The markerless strains were designated as

MW610 and MW260, respectively. MW610 subsequently was transformed with the linear als deletion construct in the same manner described above to create an als deletion in the ADHA strain background. A sequence-verified isolate was saved, designated MW618, and is referred herein as

ADHA-ΔALS. MW260 was transformed with the linearized pDN005 plasmid in order to complement the als deletion with a highly expressed als gene. Transformants were screened and purified, as detailed above. This strain was designated MW268 and is referred herein as GR5:ALS.

Growth and cell protein quantification

Strains were cultured in artificial seawater medium containing the following components per litter: 1x base salt (51), 1x trace minerals (51), 0.26 μM sodium tungstate, 0.25 μg resazurin,

0.5 g cysteine, 1g sodium bicarbonate, 0.5 g sodium sulfide, 5 g maltose, and either 0.5 or 2 g yeast extract for minimal (YM) and rich (YMR) complex medium, respectively. For growth curves using medium lacking branched chain amino acids, yeast extract was replaced with 1x amino acids mix

(either including or lacking the branched chain amino acids Val, Leu, and Ile) and 1x vitamin mix

(111). After all components were added, media were adjusted to pH 6.8 prior to bottling and the bottle head space was replaced with argon, with cycling three times between vacuum and argon to ensure no trace of oxygen was present. In this study, growth was performed in 100 mL serum bottles containing 50 mL of medium, unless otherwise indicated. Fresh cultures to be used as

inocula were prepared a day in advance of experiments, and inocula were standardized by cell count using a Petroff-Hausser counting chamber. During growth, 1 mL samples were collected periodically. Cell pellets were obtained by centrifugation at 10,000 x g for 10 min and used for protein quantification. Spent media were saved for further metabolite identification and quantification (see below). Protein was quantified as follows: cells were lysed by adding 1 mL water with vortexing and a freeze-thaw cycle, insoluble cell debris was removed by 1 min centrifugation at 10,000 x g, and soluble protein was quantified with a BioRad protein assay kit.

Cell Suspension Experiments

P. furiosus ADHA and ADHA-ΔALS strains were grown at 72 °C for 46 h to reach late log phase (<1 × 108 cells per milliliter), pelleted by centrifugation (6,000 × g) for 10 min, and then resuspended in 1x base salt (51) and 100 mM MOPS buffer pH 7.5 (1/10 of the original culture volume to achieve a 10× concentration). Maltose and/or pyruvate were added to achieve 10 mM and 100 mM concentration, respectively. Then the headspace was replaced with argon and the assay tubes were incubated at 72 °C, 50 rpm.

Metabolites quantification

Ethanol, acetate, and acetoin in spent media were measured and quantified using an Agilent

7890A Gas-Chromatography instrument, equipped with a Carbowax/20m column and an FID detector. The spent media was acidified with formic acid to 100 mM concentration prior to injecting.

Cytoplasmic extract preparation

To obtain whole cell extract for ALS activity assays, strains were cultured in YMR medium to stationary phase. Cells were collected and suspended in 50 mM MOPS pH 7.5 under strict anaerobic conditions and then lysed by three rounds of 45 sec sonication (Qsonica Q55, amplitude

30) and 15 sec rest. To remove the membrane fraction, cell lysates were centrifuged at 100,000 x g for 30 min. The cytoplasmic extract was stored in stoppered glass vials at -20 °C until used.

ALS assay

The ALS activity assay was modified from a previously published method which allows the indirect quantitation of α-acetolactate by measuring acetoin formation (117). The assay was conducted anaerobically in sealed vials containing 10 μM of flavin adenine dinucleotide (FAD), 1 mM of thiamine pyrophosphate (TPP) and 10 mM of pyruvate in 50 mM EPPS buffer pH 8.4, supplemented with 2 mM MgCl2. Whole cell extracts were added, and the assay vials were incubated at temperature ranging from 60 to 95°C to activate ALS, thereby catalyzing the condensation and decarboxylation of two pyruvates to one α-acetolactate . Aliquots (100 μL) were taken from the assay vials at intervals of 0, 5, 10, 30 and 60 min. The assay samples were acidified with 1% (vol/vol) H2SO4 at 60°C to chemically convert α-acetolactate to acetoin. Acetoin was detected by adding equal volumes of creatine (1.5%, wt/vol) and α-napthol (5%, wt/vol) in 5 N of

NaOH to the assay samples. The amount of acetoin produced was quantitated by the absorbance at 525 nm (117).

ALDC assay

The α-acetolactate decarboxylase activity assay was modified from a previously published method (118). Ethyl-2-acetoxy-2-methylacetoacetate (Sigma-Aldrich) was treated with 0.5 M

NaOH for approximately 60 min at room temperature to produce α-acetolactate, the direct substrate of ALDC. The assay vials contained the substrate solution (approximately 5 mM final concentration) in 50 mM EPPS buffer, pH 8.4, supplemented with 2 mM MgCl2. Whole cell extract was added and vials were incubated at 4, 25, 60, 70, 80 and 95 °C for 30 minutes. A control vial lacking cell extract was also included. Acetoin was detected as described in section 2.7.

2.3 Results and discussion

Acetoin is produced by wild-type P. furiosus during growth at sub-optimal temperatures

A strain of P. furiosus, termed ADHA, was previously engineered to heterologously express an alcohol dehydrogenase gene (adhA) from Thermoanaerobacter sp. X514 that grows optimally near 70°C. The ADHA strain was shown to produce ethanol optimally between 70 and

80°C by way of the heterologous AdhA functioning in concert with the P. furiosus enzyme aldehyde ferredoxin oxidoreductase (AOR), a pathway termed the AOR/AdhA pathway (25).

However, metabolite analyses showed that strain ADHA generated acetoin in addition to ethanol.

To determine if this was a property inherent to the recombinant strain, acetoin production was compared in strain ADHA, the wild type (DSM3638) and in the genetic background control strain

COM1c (116). COM1c is a uracil prototrophic version of the COM1 uracil auxotrophic genetic background strain and is used as a control in place of COM1 since there is a slight phenotype associated with uracil auxotrophy even when uracil is supplied in the growth medium.

All strains were cultured at 78C for 36 h as well as at 98C for 12 h (to late stationary phase) in minimal maltose medium (YM), and the spent medium was analyzed for metabolite identification and quantification. As shown in Figure 2.1A, acetate is produced at 78 and 98°C in all three strains but, surprisingly, acetoin is produced at significant concentrations only at 78°C, and the ADHA has the highest acetoin formation at 78°C. To determine the temperature range of acetoin production, the ADHA strain was cultured in YM medium at temperatures ranging from

70°C to 93°C for six days, and acetoin production was analyzed at the end point of growth (Figure

2.1B). The highest amount of acetoin (6 mM) was produced between 70°C to 78°C. In contrast, acetoin accumulation dropped rapidly above 80°C, and only 0.3 mM acetoin was formed at 98°C.

Hence, P. furiosus produces acetoin as one of its major metabolites in a temperature-dependent manner.

Acetoin is produced via spontaneous breakdown of α-acetolactate

P. furiosus contains no homolog for ALDC, the enzyme that catalyzes the decarboxylation of α-acetolactate to acetoin as a mechanism bacteria use to produce acetoin (109). Hence, the formation of acetoin in P. furiosus is likely due to the spontaneous breakdown of α-acetolactate during growth at high temperatures (119) (Figure 2.2A). To validate this hypothesis, approximately 5 mM α-acetolactate substrate solution was incubated at different temperatures with and without P. furiosus, ADHA strain, whole cell extract (Figure S2.2). When the assay solutions were incubated at 4 °C and 25 °C, there was virtually no acetoin formation, while at 60 °C and above, α-acetolactate breakdown to acetoin was observed equally in samples with or without added cell extract. At 70 °C and above, about 80% of the α-acetolactate was converted to acetoin after

30 min (Figure S2.2). This result leads us to believe that P. furiosus does not contain an ALDC or a non-specific decarboxylase that could catalyze the conversion of α-acetolactate to acetoin, and that acetoin formation is merely due to the spontaneous breakdown of α-acetolactate at temperatures above 70 °C.

ALS activity is temperature-dependent

There are two enzymes in nature that are known to produce α-acetolactate from pyruvate, a catabolic ALS encoded by the alsSD operon, which is involved in the production of acetoin and butanediol, and an anabolic α-acetohydroxy acid synthase (AHAS) that participates in biosynthesis of the branched chain amino acids isoleucine, leucine and valine (ILV) (119). Some species contain

both enzymes and/or multiple copies of AHAS (Figure S2.1). The flux of α-acetolactate toward amino acid biosynthesis is regulated via allosteric activation of ALDC by either valine or leucine

(119). P. furiosus contains one ALS enzyme, encoded by PF0935, and it is presumed to be involved in branched chain amino acid biosynthesis, since its gene is clustered with those involved in ILV biosynthesis (PF0935-PF0942; Figure 2.2B).

To determine if P. furiosus contained ALS activity, we assayed for the conversion of pyruvate to α-acetolactate in the ADHA strain which also contains a pathway for ethanol production (see diagram in Figure 2.2C). The ADHA strain was grown at 77.5°C, within the optimal temperature range for acetoin production (see Figure 1B), and the cytoplasmic extract was assayed from 60 to 95°C. As shown in Figure 2.3A, ALS specific activity peaked between 80 and

85°C, indicating that the enzyme is optimally active at 15 to 20ºC below the optimal growth temperature of P. furiosus. To determine how growth temperature affects ALS activity, ALS specific activity was assayed at 80°C in cytoplasmic extracts prepared from cells grown at temperatures ranging from 70 to 93°C. ALS activity, normalized to total protein detected in whole cell extract, was maximal between 75 and 77.5°C, and decreased as the temperature increased, with very low activity measured in extract from cells grown above 85°C (Figure 2.3B). These results suggest that the enzyme responsible for ALS activity in P. furiosus is maximally expressed between 72 and 80°C. The observed temperature-dependent activity profile of ALS correlates with the increased expression of this gene at the lower growth temperatures since previous transcriptomic analyses showed that expression of als increased by ~10 fold when P. furiosus was grown at 72 versus 98°C (112). All of these data correlate well with the temperature-dependent formation of acetoin in vivo (see Figure 2.1B) and suggest that ALS is the enzyme responsible.

ALS is responsible for acetoin formation in P. furiosus

To determine whether the ALS enzyme encoded by the gene PF0935 is solely responsible for acetoin production in P. furiosus, deletions of als were constructed in the P. furiosus COM1 strain as well as in the ADHA strain (Table 2.1). These strains will be referred to herein as ΔALS and ADHA-ΔALS, respectively. To analyze growth and metabolite formation (ethanol, acetate and acetoin), the ΔALS and ADHA-ΔALS strains were grown at 75°C, along with their parental

COM1c and ADHA strains using production of soluble cell protein as a measure of growth. While the growth of ADHA-ΔALS was similar to COM1c (Figure 2.4A), the ADHA strain grew slightly better than COM1c while the ΔALS strain exhibited a lag phase (Figure 2.4A).

The ΔALS and ADHA-ΔALS strains produced no acetoin, even after 57 h of growth, while specific acetoin production by the parental COM1c and ADHA strains was comparable (105.1 ±

24.6 and 102.0 ± 27.3 μmol mg-1 protein, respectively; Figure 2.4B). These results confirm that

ALS is responsible for acetoin production in P. furiosus. Furthermore, they suggest that acetoin production is beneficial to the cell, since both strains lacking ALS did not grow as well as their respective parental strains. The ADHA strain produced the highest amount of acetoin (6.1 ± 1.0 mM) (Figure 2.4C). These data also suggest that P. furiosus is unable to take up acetoin, since the concentration of acetoin continues to rise throughout the growth phase for the COM1c and AHDA strains. Only one archaeon, Sulfolobus solfataricus, has been reported to utilize acetoin as an energy source and, although it contains an als homolog similar to P. furiosus als, it is not known to produce acetoin (120).

During growth on carbohydrates, P. furiosus metabolizes glucose via a modified Embden-

Meyerhof pathway (31,32), and the resulting pyruvate is converted to acetyl-CoA via pyruvate ferredoxin oxidoreductase (POR; (33)). Acetyl-CoA is then converted to acetate and CO2 via

acetyl-CoA synthase I (ACSI) in an energy-conserving manner (121,122). We, therefore, hypothesized that deletion of als would increase the flux of pyruvate through ACSI (Figure 2.2C).

The amount of acetate produced in the spent media was analyzed and, as expected, acetate formation in ΔALS and ADHA-ΔALS was higher than in the parental strains COM1c and ADHA, respectively (Figure 2.4D). Specific acetate formation was ~30% higher in ΔALS compared to

COM1c and ~56% higher in ADHA-ΔALS compared to ADHA (Figure 2.4B).

Deletion of ALS improves ethanol production

While P. furiosus is not known to natively produce ethanol in significant amounts, the engineered ADHA strain can convert acetate to ethanol through the AOR/AdhA pathway, while conserving energy via the recycling of ferredoxin (25). Therefore, we hypothesized that ethanol production would increase in the ADHA strain when ALS was deleted. While insignificant amount of ethanol were produced by COM1c and ΔALS, ethanol production in ADHA-ΔALS reached 4.0

± 0.2 mM after 72 h of growth, which is about 25% higher than that produced in the ADHA strain

(Figure 2.4E). In addition, specific ethanol production in ADHA-ΔALS was about 50% higher than that in ADHA (107.5 ± 15.2 vs. 68.1 ± 10.0 μmol mg-1 protein, respectively; Figure 2.4B).

To further investigate the impact of ALS deletion on the ADHA strain, a 10x concentrated whole cell suspension assay was employed to examine the distribution of major metabolites in the

ADHA and ADHA-ΔALS strains using an assay temperature of 72 °C, with the addition 10 mM maltose as the carbon source. Both strains produced ethanol linearly for 18 h. The ethanol production rate of ADHA-ΔALS is at 1.9 mM per hour, 73% faster than that of the ADHA strain,

1.1 mM per hour. The AOR/AdhA synthetic pathway theoretically converts 0.5 mol glucose to 1 mol ethanol as shown in the equation below:

0.5 glucose + ADP + Pi → ethanol + CO2 + ATP (25)

Therefore, the maximal theoretical yield of ethanol from 10 mM maltose is 40 mM. After 18 h,

The ADHA-ΔALS strain produced 35.0 mM ± 2.0 mM ethanol, which is near 90% of the maximal yield, higher than the amount of ethanol produced by the ADHA strain, at 20.2 ± 2.8 mM or about

50% of the maximal yield (Figure 2.5). Since the flux from pyruvate to acetate is increased in the absence of ALS (Figure 2.2C), more acetate can be converted to ethanol in the ADHA-ΔALS strain via the AOR/AdhA pathway. Taken together, these results demonstrate that ALS is responsible for acetoin production in P. furiosus, and that deletion of the corresponding gene can be used to effectively increase product formation, depending on the growth condition, to either acetate or ethanol, at temperatures in the range of 72 to 80°C. It should be noted that the concentration of ethanol produced by the ADHA strain shown in Figure 2.4E is lower than that previously reported (11). This is because in the present study, all cultures were incubated with shaking, causing an increase in hydrogen production and a corresponding decrease in ethanol production.

High level expression of ALS uncouples temperature-dependence of acetoin formation

To determine if addition of als would restore acetoin production in the ΔALS strain, an als expression cassette utilizing the promoter (Pslp) of the highly expressed gene encoding the S-layer protein was inserted at genome region 5, a region that has been determined to have little to no transcriptional activity according to tiling array data (123). This strain will be referred to as

GR5:ALS. To compare growth and metabolite production at 78°C versus 98°C, the GR5:ALS,

ΔALS and COM1c strains were grown in YM medium at both 78 and 98°C. At both temperatures,

ΔALS and COM1c strains have similar growth characteristics. However, at 78°C, GR5-ALS has

a slower growth rate than the control, but not at 98ºC (Figure 2.6A). Acetoin formation was absent in the ΔALS strain, as expected, at both 78 and 98°C. However, GR5:ALS produced acetoin during growth at 78°C, and surpassed COM1c in acetoin production in late stationary phase (4.7 ± 0.7 vs.

3.4 ± 0.2 mM, respectively; Figure 2.6B). The GR5:ALS strain also produced acetoin even at

98°C, up to 1.6 ± 0.1 mM after 16 h of growth, which is three times higher than the acetoin formation of COM1c (0.50 ± 0.05 mM), although this concentration is still far lower than that produced at 78°C. Interestingly, even when the als gene is highly expressed at 98C, not as much acetoin is produced compared to growth at 78°C. This could be due either to the lower specific activity of the enzyme ALS at the higher temperatures (Figure 2.3A) or to an increase in activity of ketol-acid reductoisomerase (IlvC), the ILV pathway enzyme utilizing α-acetolactate , either of which would decrease the pool of α-acetolactate that can be spontaneously converted to acetoin.

We hypothesized that als overexpression would effectively reduce pyruvate flux toward acetate.

However, the GR5:ALS strain produced similar concentrations of acetate as the COM1c strain, while ΔALS, as expected, produced the highest amount of acetate at 78°C (Figure 2.6C).

Presumably, even though als is overexpressed, sufficient amounts of pyruvate must be oxidized and acetate produced to maintain the redox balance (via reduced ferredoxin production by POR) and ATP supply (via ACSI).

In addition, ALS activity was measured in cytoplasmic extracts of strain GR5:ALS along with those of COM1c and ΔALS, grown at 78°C and 98°C. The ΔALS extract did not have any detectable ALS activity, confirming a clean genetic background for this strain. The extract of

COM1c cells grown 98°C exhibited very low activity, and very low activity was detected in

COM1c cells grown at 78°C growth (0.020 ± 0.001 μmol min-1 mg-1). Indeed, the activity of ALS in cells grown at 98ºC is below the detection limit, even though a low concentration of acetoin is

produced by COM1c at 98°C. In contrast, the specific activity of GR5:ALS cells grown at 78°C was approximately 20-fold higher (0.34 ± 0.03 μmol min-1 mg-1) than that of COM1c grown at

78°C. The dramatic increase in ALS activity in the GR5-ALS strain does not correspond with the marginal increase in acetoin production in vivo compared to the control strain COM1c. To rule out the possibility that the availability of pyruvate is a major factor that limits acetoin production, acetoin formation in the ADHA strain was monitored in a 10x concentrated cell suspension assay with or without addition of 100 mM pyruvate. In medium with excess pyruvate (100 mM), only

8.5 ± 0.8 mM acetoin was detected after 18 h. However, it is still about five-fold higher than that detected in medium without additional pyruvate (1.6 ± 0.3 mM) (Figure 2.7). This result suggests that the availability of pyruvate is not what limits ALS activity. It is possible that the flux of pyruvate to ALS may be controlled by the redox balance (via POR activity) and/or energy charge of the cell (via ACSI activity), although the mechanisms involved are not known.

It is unclear why P. furiosus produces acetoin rather than acetate (or ethanol with the

ADHA strain) at sub-optimal growth temperatures, as there appears to be no energetic advantage.

It is possible that cold-adapted growth causes a decrease in POR activity, relative to the glycolytic flux. The expression of the genes encoding POR do not change significantly between 72 and 98°C

(112); however, its specific activity is ~20 fold lower at 70ºC compared to 98°C (40). Expression of als may, therefore, be up-regulated at the lower temperature in response to pyruvate accumulation, either directly or indirectly. In addition, ALS activity could be regulated via branched-chain amino acids via feed-back inhibition as it is noted that as the amount of yeast extract used in media increases, the amount of acetoin produced decreases (data not shown). The third possibility is that the second enzyme in the ILV biosynthesis pathway, ketol-acid reductoisomerase (encoded by PF0936), may be optimally active at 98°C and have decreased

activity at lower temperatures where ALS is optimally active, thereby causing α-acetolactate to accumulate and be non-enzymatically converted to acetoin. In bacteria, acetoin formation is also driven by the production of acidic metabolites, such as acetate, that cause culture pH to drop.

However, this is not the case for P. furiosus, since the medium pH during growth of the als strains

(ΔALS and ADHA-ΔALS) and their parental strains (COM1c and ADHA) were similar (data not shown).

ALS is required for branched-chain amino acid biosynthesis in P. furiosus

ALS is involved in the biosynthesis of the branched-chain amino acids via the ILV biosynthesis pathway. In P. furiosus, this pathway is encoded by a gene cluster that contains ALS

(PF0935-PF0942). ALS catalyzes the condensation and decarboxylation of two molecules of pyruvate to α-acetolactate (Figure 2.2A). α-acetolactate is the substrate for ketol-acid reductoisomerase, which is encoded by a gene (PF0936 or ilvC) located immediately downstream of als (Figure 2.2B). To determine whether ALS is required for ILV biosynthesis, ΔALS, COM1c and GR5:ALS strains were cultivated at 78°C in a defined medium using maltose as the carbon source and that either contained (DM) or lacked isoleucine, leucine and valine (DM-ILV) (Figure

2.8). The GR5:ALS strain displayed a long lag phase in both types of media. In the DM medium, both COM1c and ΔALS strains grew similarly (TD of 3.3 and 3.2 h, respectively), while the doubling time for GR5:ALS was twice as long (5.8 h). The slower growth of GR5:ALS is possibly due to the burden of overexpressing als in addition to the unfavorable growth temperature.

However, in the absence of branched-chain amino acids (DM-ILV medium), the ΔALS strain was unable to grow. In addition, the length of the lag phases of COM1c and GR5:ALS increased by 24 h in the DM-ILV medium. These data show that ALS is necessary for ILV biosynthesis, and that

insertion of als in to the ΔALS background successfully rescues the growth phenotype on a medium lacking ILV.

2.4 Conclusion

We show herein that the archaeon P. furiosus contains a functional ALS. This enzyme is responsible for the first step in branched chain amino acid biosynthesis from pyruvate to generate

α-acetolactate. It also leads to acetoin production via spontaneous breakdown of α-acetolactate.

These findings present several options to terms of engineering P. furiosus strains to produce desired products. First, current temperature-dependent biofuel producing strains can be improved by deleting als from their genomes, thereby directing carbon flow away from acetoin and towards the desired end product. Hence, as was shown here, deleting als leads to improved ethanol production in the ADHA strain, and the same should be true with a butanol-producing strain of P. furiosus (27). Second, P. furiosus is a potential platform for the biological production of acetoin at temperatures in the 70 to 80°C range. Studies are underway to exploit both of these options.

Acknowledgements

This work was supported by the ARPA-E Electrofuels Program of the US Department of Energy

(DE-AR0000081) and by the U.S. National Science Foundation (CBET-1264052 and CBET-

1264053)

2.5 List of figures

Figure 2.1 P. furiosus produces acetoin at sub-optimal growth temperatures

(A) Acetate (left) and acetoin (right) accumulation after 36 h of growth at 78 °C (open bars) or 12 h of growth at 98 °C (closed bars) for the indicated strains (see Table 1). (B) Temperature dependence of acetoin formation in the ADHA strain. Samples were collected from cultures grown at temperatures ranging from 70°C to 93°C. Ethanol was produced by the ADHA strain at concentrations up to 4 mM. Each strain was grown in biological triplicate.

Figure 2.2 Acetoin formation via ALS and its genomic context

(A) Schematic of proposed pathway for acetoin formation. (B) Diagram of ALS-containing gene cluster including ILV biosynthetic genes. (als: α-acetolactate synthase, PF0935; ilvC: ketol-acid reductoisomerase, PF0936; leuA1: 2-isopropylmalate synthase, PF0937; leuC: 3-isopropylmalate dehydratase large subunit, PF0938; leuD: 3-isopropylmalate dehydratase small subunit, PF0939; leuB: 3-isopropylmalate dehydrogenase, PF0940; leuA2: α-isopropylmalate transferase, PF0941; ilvD: dihydroxy-acid dehydratase; PF0942). (C) Schematic of carbon flow from glucose to pyruvate, acetoin and ethanol in the recombinant ADHA strain (POR: pyruvate ferredoxin oxidoreductase, ACS: acetyl-CoA synthase; AOR: aldehyde oxidoreductase; AdhA: alcohol dehydrogenase A; ALS: α-acetolactate synthase). Glucose fermentation branches to acetoin production pathway.

Figure 2.3 ALS activity is temperature-dependent

(A) Temperature profile for ALS activity in the cytoplasmic fraction of strain ADHA grown at

77.5°C. (B) ALS activity profile in cytoplasmic extracts of strain ADHA grown at various temperatures. Assays were performed at 80°C. The assay substrate, pyruvate, becomes unstable at temperatures above 85°C. Cultures were grown in triplicate and assays were performed in triplicate.

Figure 2.4 ALS deletion abolishes acetoin formation and improves ethanol production in the

ADHA strain

The strains are designated as follows: COM1c (blue, diamond), ΔALS (red, square), ADHA

(green, triangle) and ADHA-ΔALS (purple, circle). Growth curves based on soluble protein content of samples from cultures grown in YM medium at 75°C (A). Specific metabolite accumulation at 57 h of growth was measured (B), and the concentrations of acetoin (C), ethanol

(D) and acetate (E) were determined. All growths and analyses were performed in biological triplicate.

Figure 2.5 ALS deletion improves ethanol yield in the ADHA strain

Formation of ethanol over time by the ADHA (opened bars) and the ADHA-ΔALS (closed bars) strains in the 10x cell-suspension assays. Both strains were supplemented with 10 mM maltose and 5 g L-1 yeast extract in 100 mM MOPS buffer, pH 7.5, and incubated at 72 °C with 50 rpm.

All assays and analyses were performed in triplicate.

Figure 2.6 Expression of ALS restores acetoin production in the ΔALS strain

Growth comparison between strains COM1c (blue, diamond), ΔALS (red, square) and GR5:ALS

(black, circle) at 78 °C versus 98 °C in YM medium, showing growth (A), acetoin production (B), and acetate production (C). Growths were performed in biological triplicate.

Figure 2.7 Acetoin formation is increased in the presence of excess pyruvate

Acetoin production over time by the ADHA strain in a 10x cell suspension assay without (opened bars) and with 100 mM pyruvate (closed bars). In addition, 10 mM maltose and 5 g L-1 yeast extract were also added. All assays were performed at 72 °C and 50 rpm, in triplicate.

Figure 2.8 ALS is required for branched chain amino acid biosynthesis

Growth comparison between COM1c (blue, diamond), ΔALS (red, square) and GR5:ALS (black, circle) in media with (DM) and without (DM-ILV) the branched-chain amino acids (leucine, isoleucine and valine). Cultures were grown at 78 °C in biological triplicate.

Figure S2.1 Phylogenetic tree based on the amino acid sequences of ALS found in different bacterial and archaeal species

The sequences were extracted from the NCBI protein database by a BLAST using the

Lactobacillius lactis AlsC sequence. The selected sequences were aligned, and the tree was created using a comprehensive PhyML program, ONE-CLICK mode (30-35).

Biosynthetic

Catabolic

Figure S2.2 α-acetolactate decarboxylase activity was not detected in P. furiosus whole cell extract

In the absence of cell-extract (opened bars), there is no significant difference in α-acetolactate breakdown to acetoin when compared to that in the presence of whole cell extract (closed bars).

CHAPTER 3

TWO FUNCTIONALLY DISTINCT NADP+-DEPENDENT FERREDOXIN

OXIDOREDUCTASES MAINTAIN THE PRIMARY REDOX BALANCE OF PYROCOCCUS

FURIOSUS

Nguyen, D. M. N., Schut, G. J., Zadvornyy, O. A., Tokmina-Lukaszewska, M., Poudel, S., Lipscomb, G. L., Adams, L. A., Dinsmore, J. T., Nixon, W. J., Boyd, E. S., Bothner, B., Peters, J. W., and Adams, M. W. W. (2017) Two functionally distinct NADP(+)-dependent ferredoxin oxidoreductases maintain the primary redox balance of Pyrococcus furiosus. J Biol Chem 292, 14603-14616 Reprinted here with permission of the publisher

ABSTRACT

Electron bifurcation has recently gained acceptance as the third mechanism of energy conservation in which energy is conserved through the coupling of exergonic and endergonic reactions. A structure-based mechanism of bifurcation has been elucidated recently for the flavin- based enzyme NADH-dependent ferredoxin NADP+ oxidoreductase I (NfnI) from the hyperthermophilic archaeon Pyrococcus furiosus. NfnI is thought to be involved in maintaining the cellular redox balance, producing NADPH for biosynthesis by recycling the two other primary redox carriers, NADH and ferredoxin. The P. furiosus genome encodes an NfnI paralog termed

NfnII, and the two are differentially expressed depending on the growth conditions. In this study, we show that deletion of the genes encoding either NfnI or NfnII affects the cellular concentrations of NAD(P)H and particularly NADPH. This results in a moderate to severe growth phenotype in deletion mutants, demonstrating a key role for each enzyme in maintaining redox homeostasis.

Despite their similarity in primary sequence and cofactor content, crystallographic, kinetic, and mass spectrometry analyses reveal that there are fundamental structural differences between the two enzymes and NfnII does not catalyze the NfnI bifurcating reaction. Instead it exhibits non- bifurcating ferredoxin NADP oxidoreductase-type activity. NfnII is therefore proposed to be a bifunctional enzyme and to also catalyze a bifurcating reaction, although its third substrate, in addition to ferredoxin and NADP(H), is as yet unknown.

3.1 Introduction

Pyrococcus furiosus is a hyperthermophilic archaeon that grows optimally near 100 °C and is able to utilize a wide range of simple and complex carbohydrates and peptides as carbon sources to produce acetate, CO2 and H2 or, in the presence of elemental sulfur, H2S. The carbohydrate metabolism of P. furiosus proceeds through a modified Embden-Meyerhof pathway that only utilizes ferredoxin (Fd) as an electron acceptor and NAD+ is not required (30). NADPH for biosynthesis is thought to be generated from evolved H2 recycled by soluble hydrogenases (SHI and SHII) (50), as well as by a soluble ferredoxin (Fd) NADP+ oxidoreductase (FNOR) (53,97).

Fd generated from glycolysis is in turn oxidized by an energy conserving membrane-bound hydrogenase (MBH) resulting in the production of H2 (49). Upon addition of elemental sulfur, P. furiosus metabolism is shifted rapidly to shut down the hydrogenases and H2 production, while initiating the production of H2S likely via the MBX complex, together with the NADPH-dependent sulfur reductase (NSR) (59). The presence of sulfur causes a dramatic decrease in the expression of genes encoding the three hydrogenases and an increase in the expression of MBX and other so- called sulfur-response genes (60). Specifically, the regulatory transcription factor SurR orchestrates the metabolic switch from H2 to H2S formation and also functions as a global regulator of electron flow pathways in P. furiosus (55-57). P. furiosus metabolism also changes dramatically depending on whether the carbon source is carbohydrates or peptides (36). During growth on sugar

(maltose), Fd serves as the main cellular redox carrier whereas during growth on peptides (in the presence of sulfur), both Fd (generated from keto-acid oxidoreductases) and NADPH (generated from amino acid deamination) are thought to serve as primary reductants (36).

Recently a new type of enzyme system was described to regulate the redox pools of Fd,

NAD(H) and NADP(H) in anaerobes, which lack the canonical transhydrogenases. The redox pools in some anaerobes are balanced by a bifurcating enzyme called NADH-dependent ferredoxin

NADP+ oxidoreductase (Nfn). This was first reported in Clostridium kluyveri and catalyzes the production of NADPH coupled to the simultaneous oxidation of NADH and reduced Fd. This enzyme carried out flavin-based electron bifurcation, a fundamental mechanism of biological energy conservation (48,92). In anaerobes, Nfn is directly involved in shuttling electrons between the three main redox carriers in anaerobes, Fd, NADH and NADPH (equation 3.1):

+ + NADH + 2 NADP + 2 Fdred  NAD + 2 NADPH + 2 Fdox (3.1)

Nfn effectively couples the endergonic reduction of NADP by NADH and the exergonic reduction

+ of NADP by reduced Fd (Fdred), thereby maintaining a high ratio of NADPH/NADP to drive biosynthesis (48,87). The genome of P. furiosus contains two Nfn homologs. One (PF1327-28) was initially purified as a sulfide dehydrogenase (97) and was subsequently found to have ferredoxin NADP+ oxidoreductase (FNOR) (53) activity (equation 3.2). This enzyme was recently renamed NfnI as it was shown to

+ NAD(P) + 2 Fdred  2 NAD(P)H + 2 Fdox (3.2) catalyze equation 1. NfnI was functionally and structurally characterized and the mechanism of flavin-based electron bifurcation was elucidated (87). NfnI contains a large (L) subunit of 53 kDa harboring two [4Fe4S] clusters and one FAD, and a small (S) subunit of 31 kDa containing one

[2Fe2S] cluster and one FAD. Nfn homologs are found mainly in anaerobic microorganisms, both bacteria and archaea, and the subunit and cofactor composition are highly conserved (87).

The second Nfn homolog of P. furiosus, termed NfnII (PF1910-11), has yet to be structurally and functionally characterized and is the subject of this study. Interestingly, the

expression of NfnI and NfnII are dependent upon sulfur availability and carbon source in a reciprocal fashion (36,57). Specifically, expression of nfnI is up-regulated under H2 producing conditions (sugar fermentation) while nfnII is up-regulated under S0-reducing conditions (with sugars or peptides as the carbon source). It was assumed that NfnI and NfnII have similar functions in shuttling electrons between NAD(P)H and Fd, adapting to the cells needs during growth, but why the cell needs two such enzymes and how they differ from each other was not at all clear, especially given their high sequence similarity.

Our current understanding of the physiological functions of Nfn in redox metabolism is very limited. Deletion of the two genes encoding Nfn in the fermentative bacterium Clostridium thermocellum was reported to have no significant effect on its metabolism (124). In

Thermoanaerobacterium saccharolyticum, Nfn was proposed to play a role in the NADPH- dependent ethanol production pathway since its deletion decreased ethanol yield in the NADPH- dependent strain (125). In this study, we assess the distribution and evolution of Nfn in the order

Thermococcales and determine the importance of the Nfn enzymes in P. furiosus through characterization of Nfn deletion mutants. In contrast to what was reported for C. thermocellum, both NfnI and NfnII play key roles in the metabolism of P. furiosus. In addition, we provide the first biochemical and structural characterization of P. furiosus NfnII.

3.2 Experimental Procedures

Characterization of nfnS Gene Neighborhood

Homologs of NfnS and NfnL were identified in publicly available complete archaeal genomes (n=230) using previously characterized NfnS (TM1639) and NfnL (TM1640) from

Thermotoga maritima (54) as BLASTp queries. Only NfnSL that were co-localized in the genome were retained for further phylogenetic and gene neighborhood analyses. A custom python (ver.

2.7.6) script was used to extract gene sequences (10 upstream and 10 downstream) that flanked nfnS. The twenty inferred protein sequences were subjected to pairwise alignment and were clustered using the cd-hit program (126). Inferred protein sequences were clustered using identity thresholds of 90 %, 60 % and 30 % while holding the pairwise sequence coverage threshold constant at >60 %. The clusters generated by the three-step clustering method were later combined to get a final ‘averaged’ cluster identity. Protein sequence clusters were then used to generate a binary matrix describing the presence or absence of cluster for use in statistical analyses.

Network analysis

The binary matrix describing the distribution of protein bins in the gene neighborhood of nfnS was organized based on the taxonomic rank of the host genome. To predict the potential functional role of Nfn without bias, only the archaeal phyla that contained greater than 10 nfnS encoding genomes were considered further and these were ultimately analyzed at the class taxonomic level, or at the level of genes that flank individual Nfn isoforms. Proteins that were in high abundance in the flanking region of nfnS (>20%) were visualized using Cytoscape specifying the force directed organic layout (127).

Phylogenetic analysis

Identified NfnS and NfnL proteins were aligned individually with ClustalW specifying default settings (128). NfnS and NfnL alignment blocks were concatenated using a custom python

(version 2.7.6) script. Phylogenetic reconstruction was performed with PhyML (version 3.1) (129) specifying the LG substitution matrix and Chi2 to approximate the likelihood ratios. Trees were projected using Itol (130).

PCR product construction

For the over-expressed (OE-) strain, a linear PCR product was assembled to overexpress

NfnII by targeted replacement of the native promoter with the promoter (slp) of the gene encoding the highly-expressed S-layer protein, similar to the overexpression cassette for NfnI (15). The full linear ~2.5 kb PCR construct was assembled via splicing by overlap extension and PCR (113). The

NfnI and NfnII deletion cassettes were also designed to knock out the NfnI (PF1327-1328) and

NfnII (PF1910-1911) genes in the P. furiosus genome. These cassettes contained the pyrF pop-out marker cassette, including the 65 bp identical flanking region were inserted between the homologous up and downstream flanking regions of the NfnI and NfnII genes. The full ~2.3 kb

PCR constructs were obtained using similar PCR technique as described above. All primers designed and used in this study were reported in Supplemental Table S3.1.

Strain constructions

Strains constructed for this study are listed in Table 3.1. The linear PCR constructs were designed to insert an NfnII overexpression cassette into the P. furiosus genome at the native NfnII

gene location, as described above. The overexpression PCR construct was transformed into P. furiosus COM1, as previously described (131). The transformants were cultured and purified three times on liquid and solid defined cellobiose-containing medium (131). Genomic DNA was isolated using the ZymoBeadTM Genomic DNA Kit (Zymo Research) and isolates were screened by PCR, using primers that target outside of the homologous flanking regions. A strain containing the NfnII overexpression constructs were sequence-verified and designated as MW333, or OE-NfnII, respectively.

The single deletion mutant of NfnI (PF1327-28) was also constructed using similar methods to those described above, and the sequence-verified strain was termed MW190 and will be referred to herein as the NfnI strain. The single deletion mutant of NfnII mutant was also constructed but the genes encoding it (PF1910-11) are located downstream and transcribed divergently from the gene encoding ferredoxin, PF1909. Therefore, to avoid any polar effect, the genetic marker pyrF was immediately spliced out by counter-selection for the loss of the marker cassette, as previously described (116,132). The markerless strain, after sequence verification, was termed MW187. This strain was subsequently transformed with linear DNA that targeted restoration of the pyrF gene at its native location. This new strain was named MW379 and is referred to as the NfnII strain

To generate the double deletion mutant lacking both NfnI and NfnII, the NfnI deletion construct was transformed into the markerless MW187 strain, as described above. This strain was named MW383 and is referred to herein as the NfnI-NfnII strain. It should be noted that attempts to construct this double deletion strain were unsuccessful as judged by DNA-gel electrophoresis (data not shown), with roughly 50% of the cells still containing genes encoding

NfnI.

Growth of P. furiosus

Strains (Table 3.1) were cultured in artificial seawater medium containing the following per liter: 1x base salts (27), 1x trace minerals (27), 10 µM sodium tungstate (27), 0.25 µg resazurin,

10 µM riboflavin, 10 µM cobalamin, 0.5 g cysteine, 1 g sodium bicarbonate and 1 mM potassium phosphate buffer, with the pH adjusted to 6.8. 50 or 75 mL medium were aliquoted into 100 or

150 mL serum bottles. The medium bottles were then capped and the headspace was replaced with argon after three cycles of vacuum. For the growth experiments, three types of medium were used: the minimal maltose medium (M), which was supplemented with 5 g maltose and 0.5 g yeast extract per liter (133), the maltose sulfur (MS) medium, which is the M medium containing 2 g elemental sulfur per liter, and the minimal peptide sulfur medium (PS), which contained 5 g casein,

0.5 g yeast extract and 2 g elemental sulfur per liter (36). P. furiosus cells were inoculated to ~3 x

106 cells ml-1, and cultures were incubated at 90 °C with shaking at ~200 rpm. Growth was monitored by cell counting using a Petroff-Hausser counting chamber. Cell protein was also quantified from 1 ml culture samples using the Bradford protein assay kit (Bio-Rad). Cells were harvested by centrifugation and lysed by osmotic shock in an equal volume of water with vortexing. Lysate was centrifuged at 10,000 g for 1 min to pellet insoluble cell debris prior to quantitation of soluble cell protein. The over-expressing recombinant strains, OE-NfnI and OE-

NfnII, were each grown on a 20-liter scale as previously described (87). The growth medium was supplemented with 5 g maltose and 5 g yeast extract per liter. After reaching the optimal cell density, cells were collected by centrifuging at 6000 g for 10 min and were stored at -80 °C until use.

Measurements of NAD(H) and NADP(H) in P. furiosus

The COM1c, NfnI, and NfnI strains were grown in 50 mL medium bottles of M, MS and PS media until mid-log phase at 90 °C with shaking at ~200 rpm. Cells were collected via centrifugation at 6000 g for 10 min at 4 °C and lysed with 100 L of 50 mM ammonium acetate anaerobically inside a Coy chamber. Lysates were filtered through a 10 kDa cutoff-filters (Merck

Millipore) to remove proteins. The flow-through samples were immediately used for NAD(H) and

NADP(H) analysis by high performance liquid chromatography (HPLC). HPLC measurements were performed according to the previously described method (134) with some modification. A

Hydrosphere C18 column (5 m, 150 x 4.6 mm I.D., 12 nm (YMC Co., Ltd., Kyoto, Japan) was used connected to a YMC Guard Cartridge column and run on an Agilent 1260 HPLC (Hewlett-

Packard Wilmington, DE, USA). Filtered lysates were kept anaerobically and 20 L aliquots were injected for analysis. NAD(P)(H) (5 M) was added to samples as internal controls. NADPH,

NADH, NADP+ and NAD+ were quantified based on absorption at 340 nm and 260 nm, respectively. Calculated concentrations of NAD(P)(H) were normalized based on the protein concentrations of the lysates. One unit represents one mol of nicotinamide nucleotide per g protein.

RNA extraction

RNA was extracted using a phenol:chloroform method as previously described (112).

Contaminating genomic DNA was digested using TURBO DNase (Ambion). RNA quality was assessed by A260/A280 ratios and qPCR.

Quantitative RT-PCR

Synthesis of cDNA was performed with 1 µg purified RNA using the Affinity Script QPCR cDNA synthesis kit (Agilent). The Brilliant III SYBR® Green QPCR Master Mix (Agilent) was used for quantitative RT-PCR experiments with primers designed to amplify a ~150 bp product within the target gene. The constitutively expressed PF0983 gene encoding the sliding clamp subunit of the DNA polymerase was used as a reference.

Purification of the recombinant His-tagged NfnI and NfnII

Approximately 60 g of frozen cells were lysed by resuspending them in 300 mL of 50 mM phosphate buffer, pH 7.0, under strict anaerobic condition inside a Coy chamber (CoyLab, MI,

USA). The S100 cell-free extract fraction was obtained by ultra-centrifugation at 100,000 g for 1 h at 12 C to remove the membrane remnants and cell debris. Anaerobic conditions were maintained during Nfn (I and II) purification by adding 1 mM cysteine to all buffers. The S100 was loaded onto a 5 ml HisTrapTM FF crude (GE Healthcare Bio-Science) and the His-tagged protein was eluted by applying a gradient of imidazole in 50 mM phosphate buffer, pH 7.0, per the manufacturer’s instructions. All fractions that contained Nfn activity, measured by the NADPH- dependent reduction of benzyl viologen, were pooled and concentrated using an Amicon Ultra-4 ultra-filtration centrifugal filter (10 kDa cutoff; Merck Millipore). The concentrated fractions were loaded onto a 5 ml HiTrapTM QHP column (GE Healthcare Bio-Science) equilibrated with 25 mM

Tris/HCl, pH 8.0, and Nfn was eluted with a gradient from 0 to 500 mM NaCl. The purity of active fractions was judged by gel-electrophoresis before pooling and storing at -80 °C until use.

Approximately 0.5 mg of protein was purified from 1 g wet cell paste for both NfnI and NfnII.

Protein identification

Digestion of gel bands and proteins in solution was performed according to standard protocols recommended by the manufacturers using a trypsin (Promega) protease:complex ratio of 1:50-1:100 overnight and pepsin (Sigma) protease:complex ratio of 1:10 for 60 sec. Proteins were identified as described (135) using a maXis Impact UHR-QTOF instrument (Bruker

Daltonics, Billerica, MA) interfaced with a Dionex 3000 nano-uHPLC (Thermo-Fisher, Waltham,

MA) followed by data analysis in Peptide Shaker (136). Intact protein analysis was performed as described previously using a Bruker Micro-TOF mass spectrometer (Bruker Daltonics) (137).

Chemical cross-linking

Protein cross-linking was performed using 10 mM glutaraldehyde (Sigma) and 20 μg of the NfnI and NfnII at 14 µM, complexed with Fd 1:1 ratio (Fd 14 µM). The reaction was carried out in 50 mM HEPES, pH 7.2, 150 mM NaCl at room temperature. The reaction was quenched after 10 min by addition of 1.7 M Tris buffer, pH 8 to a final concentration of 100mM. The resulting mixtures were separated by SDS-PAGE (4-20 % linear gradient gel, Bio-Rad) and stained with Coomassie Brilliant Blue (Thermo-Fisher). Protein bands of interest were excised from the gel, digested with trypsin and analyzed by LCMS as described above.

Native Mass Spectrometry

Non-covalent mass spectrometry under native conditions was conducted on a SYNAPT

G2-Si instrument (Waters) in a similar fashion to that previously described (138). Briefly, the NfnI and NfnII complex samples were buffer exchanged with 200 mM ammonium acetate, pH 7

(Sigma) using 3 kDa molecular weight cutoff spin filters (Pall Corporation) and infused from in-

house prepared gold-coated borosilicate glass capillaries to electrospray source at protein concentration of 5 μM and the rate around 90 nL min-1. The instrument was tuned to enhance performance in the high mass-to-charge range. Settings were as follows: source temperature 30

°C, capillary voltage 1.7 kV, trap bias voltage 16 V and argon flow in collision cell (trap) 7 mL min-1. Transfer collision energy was held at 10 V while trap energy varied between 10–200 V. To determine accurate mass of individual protein components, complexes were denatured by dilution in a 50:50 solution of 1 % formic acid (Sigma) and acetonitrile (Thermo-Fisher). Data analysis was performed in MassLynx software version 4.1 (Waters). Molecular graphics were created using the UCSF Chimera package (139). Intact protein analysis showed that all proteins, except NfnII-

S, have missing N-terminal Met residues. NfnII-S 34,029.1 Da (calculated 34,028.71 Da), NfnII-

L 52,368.37 Da (calculated 52,497.8 Da, missing N-terminal Met residue); NfnI-S 31,376.7 Da

(calculated: 31,507.9 Da, missing N-terminal Met residue), NfnI-L 52,597.8 Da (calculated 52,729

Da, missing N-terminal Met residue); Fd 7,167.0 Da (calculated: 7,298.2 Da, missing N-terminal

Met residue).

Structure determination and refinement

NfnII crystals were obtained by the vapor diffusion method under anaerobic conditions in a Coy anaerobic chamber using 0.22 M magnesium sulfate, 27 % (w/v) polyethylene glycol 3,350, and 0.4 % (v/v) ethyl acetate in presence of 1 mM sodium dithionite. In addition, before flash freezing in liquid nitrogen, NfnII crystals under argon flow were dragged through crystallization solution containing 10 % (v/v) glycerol as a cryo protectant. The data were collected from flash- cooled crystals with a continuous flow of liquid nitrogen at 100 K on BL12 -2 (SLAC National

Accelerator Laboratory). The diffraction images were indexed, integrated and scaled using

HKL2000 (140). The structure was solved to 2.6 Å by molecular replacement using the structure of NfnI (sequence identity 48.3 %, PDB code 5JFC), with phenix.phaser (141). The solutions were refined and improved by phenix.refine (142) with final R/Rfree to 20.9%/27.4% (Table 3). Model building was subsequently completed manually using COOT (143). Figures were prepared using

PyMol (144) (http://www.pymol.org). The root-mean-square deviation (RMSD) was calculated using SUPERPOSE (145). Composition of the crystal was confirmed by dissolving the protein and running on SDS-PAGE. Protein bands were digested in gel with trypsin and identified using

LCMS as described in Chemical cross-linking section. The NfnII structure was deposited in the

PBD databank with code 5VJ7.

Nfn and FNOR dye-linked activity assay

NADH and NADPH-dependent dye-linked activity assays were performed in 50 mM

MOPS, pH 7.5 and 2 mM benzyl viologen (BV) was used as the electron acceptor. BV reduction was monitoring at 600 nm, ε = 7.4 mM-1cm-1. The assay contained 1 mM NAD(P)H and 2 mM

BV. One unit is defined as 1 µmol BV reduced per min.

Nfn bifurcating activity assay

The bifurcating activity was measured based on the following reaction: 2 NADPH + 2 Fdox

+ + + + NAD →2 NADP + H + NADH + 2 Fdred. Fd was purified based on the previously described protocol (146). Fd reduction was monitored at 425 nm (ε = 13 mM-1 cm-1) using 1 mM NADPH,

2 mM NAD and 25 µM Fd in 50 mM MOPS, pH 7.5 at 80 C. One unit of bifurcating activity is defined as 1 µmol Fd reduced per min.

FNOR activity assay

The FNOR activity assay was measured according to the following equation: 2 NAD(P)+

+ 2 Fdred  2 NAD(P)H + 2 Fdox. Fd reduction or oxidation was monitored at 425 nm using 1 mM NADP(H) and 50 µM Fd (to an absorbance of approximately 0.9) in 50 mM MOPS, pH 7.5 at 80 C. Where reduced Fd was used, the protein was chemically reduced with freshly-prepared

7 mM Ti-citrate at the beginning of each assay. One unit of FNOR activity is defined as 1 µmol

Fd reduced or oxidized per min.

3.3 Results and Discussion

NfnSL taxonomic distribution

Genes encoding homologs of the S and L subunits of P. furiosus NfnI were found in 72 archaeal genomes (or 31 % of the total). Within these genomes, a total of 92 archaeal NfnSL homologs were identified, with 17 genomes encoding multiple copies. (Supplemental Figure

S3.1). Of the total 72 archaeal genomes, 66 were from the phylum Euryarchaeota, one from the phylum Korarchaeota and five from the phylum Crenarchaeota. Of the 66 Euryarchaeota genomes that contained NfnSL, 41 genomes belonged to the order Methanomicrobia, 23 to the

Thermococci, five to the Thermoprotei, and two belonged to an unclassified order. Furthermore, all 17 genomes that encoded two or three isoforms of NfnSL were in the order Thermococcales

(Figure 3.1). Interestingly, genes encoding for homologs of P. furiosus NfnII are always present in the genomes of the order Thermococcales together with either NfnI or a third type of isoform,

NfnIII. The Thermococcales strains Thermococcus onnurineus, Pyrococcus sp. ST04, and

Pyrococcus sp. NCB100 encode for homologs of all three isoforms of Nfn

Phylogenetic analysis of a concatenation of NfnSL homologs reveals monophyly of Nfn isoforms at the taxonomic level (e.g., evidence for vertical inheritance). To provide insight into the physiological role of Nfn in archaeal genomes, in particular enzymatic processes or pathways that are NADP(H), NAD(H) or Fd specific, we investigated the proteins encoded in the gene neighborhood of those encoding for NfnSL in the Thermococci and Methanomicrobia. In

Thermococci, neighboring genes included those encoding [NiFe]-hydrogenases and transferases

(e.g. nicotinate phosphoribosyltranserase-like and carboxyl methyltransferase-like proteins.). In contrast in Methanomicrobia, genes encoding isomerases (e.g. ketol-acid reductoisomerase-like

and mannose phosphate isomerase-like proteins) and nitroreductases, which are NADP(H)- dependent enzymes (147-149), were enriched in the neighborhood of nfn (Supplemental Figure

S3.2).

Phylogenetic reconstruction of archaeal NfnSL also revealed that the multiple isoforms were the result of at least two independent gene duplication events, yielding monophyletic lineages that we have termed NfnI, II and III. To provide more insight into specific NADP(H), NAD(H) or Fd requiring processes that might lead to selective pressure to retain multiple copies of Nfn, we analyzed the proteins encoded in the gene neighborhood of all three Nfn groups. This analysis revealed unique genes that were abundant in the flanking regions of each enzyme complex

(Supplemental Figure S3.3). Genes encoded for [NiFe]-hydrogenases were abundant in the neighborhoods of the NfnI group. Likewise, genes encoded for enzymes involved in glycine cleavage as well as biotin synthase-like enzymes that are NADP(H) or NAD(H) dependent

(150,151) were abundant in the flanking region in the NfnII group. Genes encoding for nicotinate phosphoribosyltransferase, which is one of the primary enzymes involved in NAD+ synthesis, was clearly enriched in the flanking regions of NfnIII group (152).

Effects of deleting genes encoding NfnI and NfnII in P. furiosus

To analyze the roles of NfnI and NfnII under different carbon (sugar versus peptides) and redox (H+ versus S0 reduction) metabolisms, strains harboring deletions of the nfnI and nfnII genes were constructed for phenotypic characterization. Multiple attempts at constructing a strain containing deletions of both NfnI and NfnII were made; however, this strain could not be completely purified to prevent the reemergence of the wild-type nfnI allele during subsequent growth. All cultures used as inocula in this comparative growth study were sub-cultured once from

the revived glycerol stocks, in order to maintain a higher fraction of the NfnI and NfnII double deletions in the NfnI-NfnII strain. The single and double deletions of NfnI and NfnII were directly compared to the control strain COM1c under three different growth conditions: maltose

(M), maltose and sulfur (MS), or peptides and sulfur (PS). Growth yields, represented by the total protein concentration, were compared and the doubling time (TD) was calculated for each strain.

During growth with maltose (M), the single deletion strains lacking either NfnI or NfnII exhibited moderate to severe growth phenotypes (Figure 3.2A, Supplemental Table S3.2). The growth phenotype was less prominent in the NfnII strain, with similar TD and final protein concentration of approximately 60 % that of the control strain (42  2 versus 71  3 g mL-1).

Deletion of NfnI caused a more severe growth phenotype, with a two-fold increase in TD and a final protein concentration approximately four times lower than the control strain (17  1 versus

71  3 g mL-1). However, the phenotype observed with the NfnI strain under sulfur-reducing conditions was less severe, with the final protein concentration similar to that of NfnII and having about 75 % of the growth of the control (34  2 and 32  2 versus 44  1 g mL-1) (Figure 3.2B).

Furthermore, when peptides were used as the carbon source (PS), deletion of NfnII caused a more pronounced phenotype than NfnI (Figure 3.2C). The NfnI strain grew more slowly but reached a final protein concentration close to the control strain (30  1 versus 37  1 g mL-1). There was no significant difference in growth of the NfnII strain compared to that grown in MS medium.

To rule out any polar effects caused by deleting nfnI and nfnII, RT-qPCR analysis were performed to determine if there were any changes in expression of the neighboring genes, PF1326, PF1329,

PF1909 and PF1912. Although there was a small increase in the expression of PF1326, PF1329 and PF1912, these changes are not thought to be significant (Supplemental Figure S3.4).

Deletion of both NfnI and NfnII caused the most severe growth phenotype in all three growth conditions, suggesting that NfnI and NfnII are important and required for robust and healthy growth of P. furiosus. Taken altogether, the phenotype caused by deletion of NfnI was the most severe in M medium lacking S0, and less severe in MS and PS, whereas the NfnII strain maintained a similar phenotype in the different media used for growth. The result of this phenotypic study was in agreement with the expression analysis of NfnI and NfnII genes in all three growth media. Specifically, the expression of NfnI was decreased in the presence of S0, and more so when peptides were used as a carbon source, whereas, the expression of NfnII followed an opposite trend, increasing in the presence of S0 and peptides (Supplemental Figure S3.5). This suggests that NfnI and NfnII are differentially expressed depending on the availability of carbon sources and whether or not S0 is present as the terminal electron acceptor. NfnI appears to be more necessary for P. furiosus during carbohydrate metabolism whereas NfnII is utilized more during sulfur and peptide metabolism. Also, at least one of the two Nfn paralogs appear to be required to sustain observable growth since attempts at deleting both NfnI and NfnII resulted in generation of an impure strain retaining ~50 % of the wild type NfnI allele. Furthermore, growth of this strain in all three media formulations tested was minimal (Figure 3.2).

To determine if the growth defects observed in the ΔNfnI and ΔNfnII strains were linked to changes in the internal redox pools, total NADP(H) and NAD(H) concentrations and their reduced:oxidized ratios were determined under the various growth condition. In general, deletion of either NfnI or NfnII caused an increase in total NADP(H) and NAD(H) pools and decreased the ratio of NADPH/NADP+ (Figure 3.3, Supplemental Figure S3.6). During growth with maltose, deletion of NfnI caused a 30-fold decrease in the NADPH/NADP+ ratio compared to the control strain COM1c (0.10  0.01 versus 3.2  0.4) and deletion of NfnII also affected this ratio but only

by 3-fold (1.0  0.1 versus 3.2  0.4). Importantly, no significant changes were observed in the

NADH/NAD+ ratio in these Nfn deletion strains (Figure 3.3A). These results imply that the physiological functions of both NfnI and NfnII are indeed to generate NADPH, presumably for biosynthesis. In the presence of S0 (MS medium), the effects on the total amounts and the ratio of reduced to oxidized nicotinamide nucleotides were not as dramatic as those seen in the absence of

S0 (Figure 3.3B). The total NADP(H) and NAD(H) only doubled in the Nfn deletion strains and the redox ratios was only reduced by half in the ΔNfnI strain, and even less than that in the ΔNfnII strain. Another noticeable increase was for the NADH/NAD+ ratio in the ΔNfnII strain (0.30 

0.02 versus 0.10  0.02) while this ratio measured in the ΔNfnI strain was comparable to the control COM1c strain. Hence, under peptide growth conditions, deletion of NfnI and NfnII mainly affects NADPH formation, reflected in an increase in total NADP(H) and the redox ratio, but

NADH/NAD+ values remain unchanged even though the total amount of NAD(H) slightly increased (Supplemental Figure S3.6). Contrary to the differential expression of NfnI and NfnII regulated by S0 (Supplemental Figure S3.5), our redox pool analysis does not support the hypothesis that NfnII is more important for growth in the presence of S0. However, these results show that the growth phenotypes observed when either NfnI or NfnII is deleted are due at least in part to internal changes of the nucleotide ratios, and suggest that NfnI and NfnII are the two major enzymes responsible for NADPH production in P. furiosus.

In order to determine if deletion of either NfnI or NfnII had any effect on expression of other related genes, RT-qPCR analysis was performed. There was no significant change of expression in either NfnI or NfnII genes in the respective ΔNfnII or ΔNfnI strains, under sulfur or non-sulfur reducing conditions, indicating there is no transcriptional compensation between the paralogs when the other Nfn is deleted (Supplemental Figure S3.7). Expression of key genes in

the NAD+ salvage pathway was also tested because of the roles the Nfn paralogs are expected to

+ play in balancing the redox pools. The genes encoding NAD kinase (NADK, PF1103), (NH3)- dependent NAD+ synthase (NadE, PF0098), and NAD+ diphosphorylase (NMNAT, PF0458), did not change significantly in expression; however, a significant change occurred in the gene encoding L-aspartate oxidase (NadB, PF1976) in the NfnI strain. NadB catalyzes the conversion of L-aspartate to iminoaspartate, the very first step in the NAD+ salvage pathway. When NfnI was deleted, NadB expression decreased by nearly eight-fold in the presence of sulfur compared to the control strain (Supplemental Figure S3.7). If NfnI is the major enzyme producing NADPH for biosynthesis from NADH and reduced ferredoxin, and is responsible for regulating the favorable concentration of NADH and NADPH in P. furiosus, deletion of NfnI would most likely cause an accumulation of NADH internally. Thus, NadB expression was decreased as a result of feedback inhibition, thereby turning off or slowing down the NAD+ salvage pathway to avoid excess NADH from accumulating.

Unlike NfnI, deletion of NfnII had little to no effect on NadB and other NAD+ synthesis enzymes, despite the growth phenotype observed under sulfur-reducing conditions. In addition, deletion of NfnI or NfnII did not change the expression of Fd significantly compared to that of the control strain. However, it is possible that the reduced:oxidized Fd ratio is altered in the Nfn deletion strains. Because there are many other Fd-utilizing enzymes that have a relatively high transcription level (e.g. pyruvate oxidoreductase (POR), MBH, etc.), it is challenging to determine the effect of an Nfn deletion on the ratio of reduced:oxidized Fd in the cell.

Expression of other major NAD(P)(H) utilizing enzymes was also analyzed in the Nfn deletion strains under sulfur and non-sulfur growth conditions, including genes encoding subunits of the soluble hydrogenases I and II (SHI and SHII) and the gene encoding the NAD(P)H sulfur

oxidoreductase (NSR). Overall, deletion of NfnI or NfnII did not affect expression of the SHI and

SHII operons (Supplemental Figure S3.7), as the transcription levels are comparable to the control strain. However, expression of NSR decreased significantly in the ΔNfnI strain but not in the ΔNfnII strain. It is not clear whether there is a strong connection between NfnI and NSR since these two enzymes are differentially expressed under sulfur-reducing conditions. NSR has a preference for NADPH over NADH; therefore, it is possible that the expression of NSR decreased in response to the drop of NADPH level in the NfnI strain.

Effect of NfnI and NfnII overexpression in P. furiosus

In order to analyze the phenotypic effects of overexpressing NfnI and NfnII, strains that contained either NfnI or NfnII placed under control of the strong, constitutive slp promoter were constructed in P. furiosus. These are referred to as OE-NfnI and OE-NfnII, respectively. The gene encoding the S subunit for each also contained a his-tag at the N-terminus for purification purposes

(see below). To screen for possible growth phenotypes due to expressing these redox enzymes at high levels, the OE-NfnI and OE-NfnII strains were grown in M, MS and PS media. When grown using different carbon sources (maltose or peptides) and terminal electron acceptors (H+ or S0) both OE-NfnI and OE-NfnII strains showed growth defects at different levels of severity. OE-NfnI had most severe phenotype in M medium, similar to that observed with the NfnI strain (Figure

3.2A). The phenotype is less prominent in the MS medium and insignificant in PS medium. This observation, together with the decrease in transcription of NfnI during sulfur and peptide metabolism, suggests that NfnI function is more important in carbohydrate metabolism. On the other hand, the OE-NfnII strain displayed similar phenotypes when grown in all three media, reaching approximately 50 % of the growth compared to the control strain. It appears that the NfnI

and NfnII overexpression strains have similar and sometimes more severe growth phenotypes compared to the ΔNfnI and ΔNfnII strains. It is likely that these growth defects are due to the increased NfnI and NfnII activities changing the composition of the three main redox pools (Fd,

NAD(H) and NADP(H)) thereby affecting the activities of other redox enzymes. Taken together, these data suggest that the native expression levels of the NfnI and NfnII operons are finely tuned to regulate redox balance in P. furiosus metabolism.

The differential expression of NfnI and NfnII together with the different phenotypes observed in the overexpression and deletion strains suggest NfnI and NfnII play pivotal role in carbohydrate and peptides metabolism while S0 metabolism adds an additional layer of complexity.

In short, NfnI is mainly responsible for the transfer of electrons from NADH and reduced Fd to

NADP+, which is used for biosynthesis (87) under normal carbohydrate metabolism in P. furiosus.

However, whether NfnII has the same function as NfnI is still unclear since during growth with peptides, where NfnII expression is highest, expression of GDH, another NAD(P)H-forming enzyme is also up-regulated (101). Judging from sequence homology, NfnI and NfnII could be catalyzing very similar reactions but with different kinetics. However, to better understand the roles of each Nfn, characterization of NfnII was undertaken.

Purification of NfnII in P. furiosus

To obtain protein for structural and biochemical characterization, his-tagged NfnII was purified from the P. furiosus OE-NfnII strain. Approximately 43 mg purified NfnII was obtained from 60 g of wet P. furiosus cells. All peak fractions from the Ni:NTA column that were brownish to yellow in color (from flavin and iron-sulfur clusters) and had NADPH-dependent BV reduction activity were pooled, concentrated and further purified via anion-exchange chromatography. A

similar two-step purification protocol was used previously for the purification of NfnI (87). The total protein recovery after the two-column purification was 62 % (Table 3.2). The NfnII protein was purified as the expected heterodimer containing large (L) and small (S) subunits as shown by gel-electrophoresis (Supplemental Figure S3.8).

In-gel digestion and LCMS analysis confirmed the identity of the L and S subunits of both

NfnI and NfnII. We suspected that the lower molecular weight protein band seen on the electrophoresis gel was Fd, as described for NfnI (87). LCMS analysis of both samples after digestion with the low specificity protease pepsin confirmed the presence of Fd in both samples.

Note that Fd is a small protein (7.5 kDa) and does not stain well on an electrophoresis gel

(Supplementary Figure S3.8). These results inspired us to pursue a more in-depth analysis of the as purified complexes using native mass spectrometry. Analysis of the purified complexes confirmed the heterodimeric state and were consistent with the presence of two FADs, two [4Fe4S] clusters and one [2Fe-2S] cluster in each NfnI and II dimer in solution (Figure 4), based on the crystallography data presented below and by Lubner et al (14). By activating the complex with low collision energy, the dimer was dissociated revealing a small subunit with and without one

FAD. At this point the two complexes appeared highly similar. However, further collision induced disassociation at higher energies showed that the NfnI complex was much more tightly associated than NfnII. Even at high collision energy (200 V) or in the presence of acetonitrile, a fraction of

NfnI remained dimeric. Fd appears to have a lower affinity for NfnII, as the trimeric complex was not observed in the gas-phase.

To capture the trimeric complex, chemical cross-linking using glutaraldehyde was employed. Cross-linked complexes where separated on SDS-PAGE and then cross-linked species where subjected to in-gel proteolysis and LCMS analysis (Figure 3.4C). NfnI (K371-F392)

interacted with the N-terminus of Fd, whereas NfnII (residues G55-K68) cross-linked around K32 on Fd (Figure 3.4D and E). Thus, while the NfnI and NfnII have the same stoichiometry and cofactor composition, the complexes have specific differences in stability and potentially in the Fd docking site.

Structure determination of NfnII

The NfnII enzyme was crystallized as a heterodimer with a large (NfnII-L) and a small (NfnII-S) subunit. The structure was determined by molecular replacement using the known structure of

NfnI as a search model and refined to 2.6 Å resolution. NfnII-L contains one FAD (L-FAD) and two [4Fe-4S] clusters, whereas NfnII-S contains one FAD (S-FAD) and one [2Fe-2S] cluster

(Figure 3.5A, Supplemental Figure S3.9, and S3.10). The [2Fe-2S] and the [4Fe-4S] clusters proximal to L-FAD are coordinated by aspartate and glutamate ligands, respectively (Figure 3.5C and D). The NfnII-L is very similar to NfnI-L with a root-mean-square deviation (RMSD) of 0.70

Å. The main differences between the two enzymes are in the small subunit, having an RMSD of

1.06 Å. The Fe-S clusters in NfnII have similar coordination environments as in NfnI (Figure

3.5C, D, and F). The most notable difference in the structure is the presence of an extended loop in the S subunit, composed of residues 164-178 which appear to occlude the assumed NAD(H) binding site, as determined by comparison with the structure of NfnI. In addition, within this FAD binding site in the S subunit, there are substitutions in NfnII of key residues that help coordinate

NADH binding in the NfnI structure. Specifically, Asn218 (in NfnI) is replaced by His226 (in

NfnII), Gly113 replaces Tyr114, and Val72 replaces Arg72 (numbering according to NfnI structure; PDB ID 5JFC). Hence, given that the NAD(H) binding site is blocked and key binding residues are absent, the primary conclusion from the crystal structure of NfnII is that this enzyme

100

is unlikely to use NAD(H) as a substrate. Accordingly, as described below, NfnII did not exhibit

NADH-dependent dye reduction, in contrast to NfnI.

Comparison of NfnII and NfnI in vitro activities

A comprehensive list of the catalytic activities of NfnI and NfnII measured in this study are given in Table 3.4. Consistent with the structural data, NfnII did not exhibit significant NADH dependent reduction of the dye BV, which was approximately 60 times lower than that observed with NfnI (1.1  0.3 versus 61  7 units mg-1). However, the NADPH-dependent reduction of BV activities were similar between NfnI and NfnII, (1277  165 and 408  60 units mg-1). In light of the NADH-dependent data, it was not surprising to find that the Nfn bifurcating activity of NfnII

+ -1 (using NADH, NADP and Fdox) was insignificant (<0.05 units mg ), compared with that measured using NfnI (21  3 units mg-1, measured by adding NfnI into the same assay cuvette used to measure NfnII activity). To rule out the possibility that NfnII was inactive after purification and/or lacked some active cofactor, the bifurcating Nfn activity was measured using the cytosolic extracts of the following strains under non-sulfur and sulfur-reducing conditions: COM1c (as the positive control), NfnI (to separate NfnII activity from NfnI) and NfnII (as the negative control).

There was no bifurcating activity detected in the cytosolic extract of the NfnI strain under both conditions, whereas comparable activities were measured in the extract of the COM1c and NfnII strains (Supplemental Table S3). These data suggest that NfnI is the sole enzyme responsible for the Nfn bifurcating activity in P. furiosus and that “NfnII” has FNOR-type (measured by the

NADPH-dependent reduction of BV) rather than true Nfn bifurcating activity.

The apparent FNOR-like activity of NfnII was confirmed using the physiological substrate

Fd in place of the artificial electron carrier BV. By using Fd both the oxidative and reductive

101

directions of the FNOR reaction could be measured. No significant NADPH-dependent Fd reduction activity was detected with NfnII, although NfnI had measurable activity (0.70  0.01

-1 + units mg , Table 3.4). To measure the NAD(P) reduction activity, Ti-citrate was used to obtain approximately 70 % reduced Fd to serve as the electron donor (in order to mimic the cellular redox

+ + ratio of Fd). The Fdred dependent NAD and NADP reduction activities of NfnI were low because electron transfer was gated in this “tight” bifurcating enzyme. There were little activity unless all three substrates (NAD(H), NADP(H) and Fdred/ox) were present, which prevented Fd from directly reducing NAD+ or NADP+ in the absence of the other cofactor (87). In contrast, this was not the case with NfnII, which exhibited significant Fd-dependent NADP+ reduction activity compared to

NfnI (5.0 ± 0.8 vs 0.4 ± 0.2 units mg-1, Table 3.4).

These results show that NfnII does not have Nfn bifurcating activity and does not utilize

NAD(H) as a substrate. It exhibits one related activity, the Fd-dependent reduction of NADP+

(FNOR activity). The lack of NAD(H)-linked activity is consistent with the structure of the enzyme discussed above (Figure 3.5B). Interestingly, this additional loop in the NfnII structure that blocks

NAD(H) binding is found in the sequences of all the NfnII homologs in the order Thermoccocales

(Supplemental Figure S3.11) but is not found in the sequences of NfnI or NfnIII. Additionally,

Fd was found to bind differently in NfnI and NfnII (Figure 3.4D and E). Investigation is underway to determine if this considerable difference in Fd mode of binding affects NfnII activity. However, as expected from thermodynamic considerations, the FNOR activity exhibited by NfnII is catalytically-biased towards NADP+ reduction (Supplemental Table S3.4), in agreement with the proposed physiological role of NfnII, which is to generate NADPH from reduced Fd, independent of NAD(H).

102

Hence, unlike NfnI, NfnII is not a bifurcating Nfn. However, whether it is another type of bifurcating enzyme utilizing another substrate (in place of NAD(H)) or a new type of non- bifurcating FNOR enzyme is unclear at this point. There are two scenarios to be considered

(Figure 3.6). First, if NfnII functions as a non-bifurcating FNOR, then structurally, only the NfnII-

L subunit contributes to the catalytic activity of the holoenzyme, begging the question of the role of its S-subunit, and why NfnII retains a structure and cofactor composition similar to NfnI?

Alternatively, given that the NfnII L subunit contains a bifurcating flavin (by analogy with NfnI),

NfnII may be a “bifunctional” enzyme in that, as well as FNOR activity, it also bifurcates using a third substrate that is not NAD(H). We hypothesize that, if this is the case, then the third substrate of NfnII is somehow involved in peptide and sulfur metabolism by P. furiosus. However, NfnII is clearly not a bifurcating NADH-dependent ferredoxin NADP+ oxidoreductase and the term Nfn is not appropriate. At presence, we favor the notion that NfnII is a bifunctional enzyme with both non-bifurcating (FNOR) and bifurcating activity. Henceforth it will be referred to as Xfn, with X representing the unknown third substrate used in bifurcation. We are currently using spectroscopic tools to determine if the FAD in the L-subunit of Xfn has the expected and characteristic signature of a bifurcating flavin, and are attempting to identify substrate X. Substrate X could be another redox protein, an oxidoreductase or a small molecule. Understanding the properties of Xfn will add another layer to our limited knowledge of both bifurcating enzymes and the primary redox metabolism of P. furiosus.

In conclusion, in the order Thermococcales, NfnI is an important bifurcating enzyme that functions in primary redox metabolism by balancing three separate pools of redox cofactors, Fd,

NADPH and NADH during carbohydrate metabolism. NfnII, a homolog of NfnI, which we now term Xfn, appears to be a bifunctional enzyme and plays an important role in maintaining the

103

primary redox pool during sulfur and peptide metabolism. In P. furiosus, the physiological functions of NfnI and Xfn are overlapping and crucial and no other enzyme can compensate for them.

Acknowledgments

This work was supported as part of the Biological Electron Transfer and Catalysis Energy

Frontier Research Center funded by the U.S. Department of Energy, Office of Science, Basic

Energy Sciences under Award # DE-SC0012518.

We would like to thank the Montana State University Microfabrication Facility for help in preparation of gold-coated borosilica capillaries for non-covalent mass spectrometry and Dr. Ravi

Kant for assistance with protein identification. The Mass Spectrometry Facility at MSU is supported in part by the Murdock Charitable Trust and an NIH IDEA program grant

P20GM103474

The use of the Stanford Synchrotron Radiation Lightsource, SLAC National Accelerator

Laboratory, is supported by the U.S. Department of Energy, Office of Science, Office of Basic

Energy Sciences under Contract No. DE-AC02-76SF00515. The SSRL Structural Molecular

Biology Program is supported by the DOE Office of Biological and Environmental Research, and by the National Institutes of Health, National Institute of General Medical Sciences

(including P41GM103393) . The contents of this publication are solely the responsibility of the authors and do not necessarily represent the official views of NIGMS or NIH

104

Author contributions

DMNN, GJS, GLL and MWWA conceived and coordinated the study and wrote the paper. SP and

EB carried out the phylogenetic analyses, OAZ and JWP performed the crystallographic study, and MTL and BB carried out the MS analyses. DMNN, GLL and GJS designed and performed all of the genetic, enzymatic and physiological studies with technical assistance from LAA, JTD and

WJN.

105

3.4 Tables and Figures

Table 3.1. Strains used and constructed in this study.

Strain number Alias Genotype References

MW002 COM1/parent ΔpyrF (131)

MW004 COM1c/control ΔpyrF::pyrF (131)

MW187 ΔNfnII ΔpyrF ∆nfnII This study

MW190 ΔNfnI ΔpyrF ∆nfnI::Ppep pyrF This study

MW379 ΔNfnII ΔpyrF::pyrF ∆nfnII This study

MW367 OE-NfnI ΔpyrF Pslp His9-Gly-nfnI::Ppep pyrF (87)

MW333 OE-NfnII ΔpyrF Pslp His9-Gly-nfnII::Ppep pyrF This study

MW383 ΔNfnI-ΔNfnII ΔpyrF ∆nfnII ∆nfnI::Ppep pyrF This study

OE, overexpression

106

Table 3.2. Purification of NfnII

Steps Volume Total Total activity Specific Activity Purification Yield

-1 (mL) protein (units) (units mg ) fold (%) (mg)

S100 450 6480 96 17 1.0 100

His-trap 28 165 62 74 4.1 65

QHP 17 42 59 275 15 62

107

Table 3.3. Data collection and refinement statistics

Data collection NfnII

Wavelength (Å) 0.9369

Unit cell parameters

a, b, c (Å) 55.74, 73.14, 99.96

α, β, γ (°) 90.0, 96.7, 90.0

Space group P 21

Resolution range (Å) 39-2.6

Total reflections 108586

Unique reflections 25877

Rmerge (%) 10.0 (36.2)

Rpim (%) 6.1 (21.5)

CC1/2 0.991 (0.934)

I/σ(I) 7.7 (2.6)

Completeness (%) 98.8 (99.1)

Redundancy 4.2 (4.2)

Refinement

Resolution limits (Å) 39-2.6

# reflections 25825

# atoms 6196

Rfactor (%) 20.9

Rfree (%) 27.4

108

Wilson B-factor (Å2) 35.3

Ramachandran plot

Favorable region (%) 95.0

Allowed regions (%) 5.0

Disallowed regions (%) 0

RMSD from ideality

Bond distance (Å) 0.010

Angles (°) 1.19

* Values in parentheses are for the highest resolution shell.

109

Table 3.4. Activities of NfnI and NfnII

NfnI activity NfnII activity Electron donors Electron acceptors (units mg -1) (units mg -1)

+ NADPH NAD + Fdox 21 ± 3 <0.01**

NADPH BV 1277 ± 165 408 ± 60

NADH BV 61 ± 7 1.1 ± 0.3

NADPH Fdox 0.7± 0.01 <0.01**

+ Fd70red* NADP 0.4 ± 0.2 5.0 ± 0.8

+ Fd70red* NAD 0.7 ± 0.2 0.6 ± 0.2

*Chemically reduced by ~ 70% with Ti-citrate

**Below detection limit of 0.01 units mg-1

Standard deviations were derived from three technical measurements

110

Table S3.1 Primers designed and used recombinant strains in this study

Primers Sequence

DN.039 CTAGGTCTATCTTCCTCC CTT C

DN.040 GGTGTTCCTCAAACATTTTCAAGTATGCACATCACCCTACAAG

DN.041 ATCCATCGGGCAATTCATGG

DN.042 GATTATTGGGAGGTGGAGAAAAATGCATCACCACCATCACCA

CCACCATCACGGTTATAAAATCCTCGAGAAAAAGGAAATCG

DN.043 CTCTTTTACAACTTCAAATACCTG

DN.044 GTTCCTCAAACATTTTCAAGGATGAACACCTCCGATCACG

DN.045 CTTTACCCATTCAACAATCTTCTCTG

DN.053 CTAGGTCTATCTTCCTCCCTTC

DN.060 GAT TAT TGG GAG GTG GAG AAA AAT GCA TCA CCA CCA TCA CCA

CCA CCA TCA CGG TTT CAA AAT TTT AAG AAA AGA GAG GC

DN.073 GGGAAGCCGCTAAGAAGATTTTC

DN.074 CGAAAATCTTCTTAGCGGCTTCCCTTAAATTAACATCTTTATTTTTTCA

AG G

DN.075 CGAAAATCTTCTTAGCGGCTTCCCTATGCACATCACCCTACAAG

DN.076 GCTCTGCCCAATATGTCCACGCGGCCGCGTTTAAACGG

WN.017 CAGAGGCAAGTAACGAGAG

WN.018 GTGGACATATTGGGCAGAGCTGTTAGAACTAAACCTATTGAAATCGT

WN.019 GCTCTGCCCAATATGTCCACTTATCTTGAGCTCCATTCTTTCAC

WN.020 TGTTAGAACTAAACCTATTGAAATCGTTGGTCAAATGCTCATCATTTA

GTTTTATG

111

WN.021 CAATAGGTTTAGTTCTAACAGCTCTGCCCAATATGTCCACTATGCACA

TCACCCTACAAG

WN.022 AAA TCT GTC AAG CCT CGT GG

WN.023 GGTCTACTGGATTGGAACAG

WN.024 CACCTCTTTCTTATAACCTTTTTAGGAC

WN. 025 AAGGTTATAAGAAAGAGGTGTTATCTTGAGCTCCATTCTTTCAC

WN.026 CTTTTTAGGACGAAAGGTTTATATCTCCAGGATGTCAAATGCTCATCA

TTTAGTTTTATG

WN.027 ATATAAACCTTTCGTCCTAAAAAGGTTATAAGAAAGAGGTGGATGAA

CACCTCCGATCAC

WN.028 CACTTTCTCCTGGGAAACCAAC

SP2.088 CTTGAAAATGTTTGAGGAACACC

SP2.055 TTTTCTCCACCTCCCAATAATC

112

Table S3.2 Doubling times for growth of recombinant strains

Doubling time (TD) Strains M MS PS

COM1c 1.06 1.35 1.7

ΔNfnI 2.07 1.14 2.1

ΔNfnII 1.28 1.14 2.7

ΔNfnI- ΔNfnII ND ND ND

OE-NfnI 1.5 1.1 3.4

OE-NfnII 3.1 1.3 3.9

ND, not determined

OE, overexpression

113

Table S3.3 Nfn bifurcating activity detected in whole cell extracts

Strains Nfn bifurcating activity (units mg-1)

M MS

COM1c 0.13 ± 0.02 0.06 ± 0.02

ΔNfnI <0.01** <0.01**

ΔNfnII 0.11 ± 0.01 0.02 + 0.0

**Below detection limit

Standard deviations were derived from three technical measurements

114

Table S3.4 NfnII catalytic bias of FNOR activities

Electron donors Electron acceptors NfnII activity (units mg -1)

NADPH Fdox <0.01**

+ Fdred* NADP 12.2 ± 1.5

*Chemically reduced with Ti-citrate

**Below detection limit

Standard deviations were derived from three technical measurements

115

Figure 3.1 Phylogenetic reconstruction of a subset of Nfn homologs encoded by Archaea within the Thermococci and Thermoprotei classes (n=45). NfnSL were concatenated prior to phylogenetic reconstruction. Asterisks (*) denote genomes that comprise multiple copies of Nfn.

116

117

Figure 3.2 Growth analysis of COM1c (black, control strain), ΔNfnI (blue), ΔNfnII (red), the

ΔNfnI –ΔNfnII (green), OE-NfnI (orange) and OE-NfnII (purple) strains in M (A), MS (B) and

PS media (C) at 90 ℃. The standard deviation analyses were taken from three independent biological samples. Abbreviation used: OE, overexpression.

118

119

Figure 3.3 Redox nucleotide pool analysis: the total concentration of NADP(H), NAD(H),

NADPH/NADP+ and NADH/NAD+ ratios of COM1c (black, control), ΔNfnI (blue), ΔNfnII (red) strains in M (A) and MS media (B) at 90 ℃. The standard deviation analyses were taken from three independent biological samples.

120

121

Figure 3.4 Native and chemical cross-linking mass spectrometry of NfnI and NfnII. Native mass spectrum of NfnI complex (A) and NfnII complex (B) in the gas phase. Yellow diamonds denote charge states of the intact complex centered around charge 16+. After deconvolution the intact

NfnI complex is 86,407 Da (expected: 86,425 Da) and NfnII is 88,951 Da (expected 88,901 Da), consistent with complexes containing one small and large subunit, two FAD molecules, two [4Fe-

4S] and one [2Fe-2S] cluster. Under conditions of low collision energy (80 V and 60 V) NfnI and

NfnII, respectively, the subunits dissociated (insets in panels A and B). Blue and magenta diamonds represent charge envelopes of small subunits with and without 1 FAD cofactor. (C) SDS

PAGE of glutaraldehyde (GA cross-linked NfnI-Fd and NfnII-Fd complexes. LCMS analysis revealed that Ferredoxin (brown) interacts with the large subunits (green) of NfnI (D) and NfnII

(E). Red space-filling regions show cross-linked peptides of NfnI and NfnII. Small subunits colored in cyan and red, respectively. In all experiments Nfn enzymes were used “as-purified”.

NfnI structure PDB ID: 5JFC; NfnII structure PDB ID: 5VJ7

122

123

Figure 3.5 Structure of NfnII. (A) Superposition of NfnII and NfnI (PDB ID:5JFC) structures shows the similarity in the large NfnII and NfnI (green) subunits, and differences in the small NfnII

(red) and NfnI (cyan) subunits. Protein structures are presented as cartoons, FeS clusters as balls, and FAD molecules as sticks. (B) Difference between the NfnII (red) and NfnI (cyan) small subunits at the FAD-S binding site. The pairs of amino acids, which potentially result in the blockage of NAD(H) binding to the active site are shown in sticks and are colored according to the color of the subunits. Similarity between NfnII and NfnI in the coordination of the 2Fe-2S cluster (C), proximal 4Fe-4S cluster (D), and distal 4Fe-4S cluster (F) is shown in sticks. Fe atoms are shown in brown, S atoms are shown in yellow. Amino acids are colored according to the color of the subunits.

124

125

Figure 3.6 Cartoon models depict the cofactor content and reaction catalyze by NfnI (A) and Xfn in two possible scenarios: (B) Xfn is a bifurcating enzyme with substrate X that replaces NAD(H) and (C) Xfn is a regular FNOR type enzyme with only the large subunit Xfn-L contributes to the catalytic activity.

126

A B C

NfnI-S NADH Xfn-S XH Xfn-S S-FAD S-FAD S-FAD E’= -280 mV E’< -380 mV - e- NAD e X [2Fe-2S] [2Fe-2S] [2Fe-2S]

- NfnI-L e- NADP Xfn-L e- NADP Xfn-L NADP L-FAD L-FAD L-FAD E’= -380 mV E’= -380 mV E’= -380 mV e- e- e- NADPH NADPH NADPH [4Fe-4S] [4Fe-4S] [4Fe-4S] [4Fe-4S] [4Fe-4S] [4Fe-4S]

Fdred Fdox Fdred Fdox Fdred Fdox E’= -500 mV E’= -500 mV E’= -500 mV

127

Figure S3.1 Phylogenetic reconstruction of a concatenation of 467 archaeal and bacterial NfnS and NfnL subunits. Phylum level taxonomic ranks were mapped onto the terminals of the tree.

Double plus symbols (++) inside of the color ring denoting taxonomic ranks indicates that the subunits are fused, whereas an asterisk (*) denotes the presence of multiple copies of Nfn.

128

129

Figure S3.2 Network analysis of proteins encoded by genes flanking (+/-10) nfnS, as organized by the taxonomic rank of the genome where NfnS was recovered. Only proteins (n=29) that were identified in >20 % of the NfnSL encoding archaeal genomes were considered. Node color represents the betweenness centrality (a measure of the ‘connectedness’ of each gene) while the edge color represents the percent abundance.

130

131

Figure S3.3 Network analysis of multiple isoforms of Nfn (i.e., NfnI, NfnII, and NfnIII) identified in archaeal genomes. Only proteins encoded in the flanking regions of >50 % of the each Nfn group (i.e. relative frequency of >50 %) for each individual Nfn were considered in this analysis.

Here, the edge color represents the abundance of the protein (depicted as a node) in each group.

The force directed organic layout was used to visualize the network.

132

133

Figure S3.4 Expression change analysis of the up- and downstream genes of nfnI (PF1327-28) and nfnII (PF1910-11) genes in the ∆NfnI and ∆NfnII when using S0 as terminal electron acceptor versus H+ by RT-qPCR. The constitutively expressed DNA polymerase subunit gene PF0983,

DNAp, was used as a control. PF1326, and PF1912 genes encoded for hypothetical genes. PF1329 encoded for the β-subunit of soluble hydrogenase II and PF1909 encoded for ferredoxin. The propagation errors were calculated from Ct values measured from cDNA that were derived from two biological samples. The top graph represents the fold changes in log based 10 scale and the bottom graph displays the observed ∆∆Ct values.

134

135

Figure S3.5 Expression change analysis of NfnI and NfnII when using S0 as terminal electron acceptor versus H+ (A) and when using peptides as the sole carbon source versus carbohydrates

(B) by RT-qPCR. The constitutively expressed DNA polymerase subunit gene PF0983, DNAp, was used as a control. PF1328 and PF1911 genes encoded for NADH dependent ferredoxin NADP oxidoreductase I and II- small subunits (NfnI-S and NfnII-S, respectively). The propagation errors were calculated from Ct values measured from cDNA that were derived from two biological samples. The top graphs of (A) and (B) represent the fold changes in log based 10 scale and the bottom graphs of (A) and (B) display the observed ∆∆Ct values.

136

137

Figure S3.6 Redox nucleotide pool analysis: the total concentration of NADP(H), NAD(H),

NADPH/NADP and NADH/NAD ratios of COM1c (black), ΔNfnI (blue), ΔNfnII (red) strains in

PS medium at 90 ℃, 200 rpm. The standard deviations were derived from measurement taken from three independent biological samples.

138

139

Figure S3.7 Expression change analysis of the control strain (black), ΔNfnI (blue) and ΔNfnII

(red) trains when using S0 as terminal electron acceptor versus H+ (M versus MS media) by RT- qPCR. The constitutively expressed DNA polymerase subunit gene PF0983, DNAp, was used as a control. The following genes encoding proteins were analyzed: NADH dependent ferredoxin:NADP oxidoreductase I and II- small subunits (PF1328, NfnI-S, and PF1911, NfnII-S, respectively), NAD kinase (PF1103), L-aspartate oxidase (PF1976), NAD diphosphorylase

(PF0458), NH3-dependent NAD synthetase (PF0098), ferredoxin (PF1909), β-subunit of soluble hydrogenase I (PF0891) and II (PF1329) and NAD(P)H elemental sulfur oxidoreductase (PF1186,

NSR). The propagation errors were calculated from Ct values measured from cDNA that were derived from two biological samples. The top graph represents the fold changes in log based 10 scale and the bottom graph displays the observed ∆∆Ct values.

140

141

Figure S3.8 Identification of protein components in Nfn complexes. (A) SDS PAGE separation of purified complexes NfnI and NfnII. Protein bands indicated by the arrows were identified as

NfnI or NfnII small and large subunits in the course of tryptic digestion. (B) Three overlapping peptides from pepsin digestion used to identify ferredoxin (identified peptide is highlighted in red).

Below the sequence is the MS/MS fragmentation matches, where red and blue lines represent b and y ions, respectively.

142

143

Figure S3.9 Structure of NfnII. The omit electron density map of NfnII contoured at 1.5 σ shown in blue mesh calculated with omitted L-FAD (A), S-FAD and [2Fe-2S] cluster (B), and two [4Fe-

4S] clusters (C). The NfnII-L subunit is colored in green, and NfnII-S subunit colored in red. L-

FAD and S-FAD are shown in sticks, FeS clusters in balls and sticks.

144

145

Figure S3.10 Structure of NfnII. The simple composite omit electron density map of NfnII contoured at 1.5 σ shown in blue mesh around L-FAD (A), S-FAD and [2Fe-2S] cluster (B), and two [4Fe-4S] clusters (C). The NfnII-L subunit is colored in green, and NfnII-S subunit colored in red. L-FAD and S-FAD are shown in sticks, FeS clusters in balls and sticks.

146

147

Figure S3.11 Multiple sequence alignment of a fragment of representative NfnI, NfnII and NfnIII protein sequences. Positions 163 to 178 are highlighted with a red box. The asterisk represents

NfnI and NfnII from P. furiosus reported in this paper.

148

149

CHAPTER 4

THE BIFURCATING NFN OF PYROCOCCUS FURIOSUS: A SIMPLE MODEL TO

DETERMINE THE MECHANISM OF FLAVIN-BASED ELECTRON BIFURCATION

Energy conservation by the NADH-dependent ferredoxin:NADP+ oxidoreductase (Nfn) is accomplished via the coupling of the reduction of ferredoxin by NADPH (an endergonic reaction) to the reduction of NAD+ by NADPH (an exergonic reaction) through a process called flavin-based electron bifurcation (FBEB) (Equation 4.1).

+ + 2 Fd(ox) + NAD + 2 NADPH  2 Fd(red) + NADH + 2 NADP (4.1)

In a nutshell, electron bifurcation refers to the process by which the electron pair is split into two electrons with different reduction potentials, one high and one lower than that of the potential of the donor of the electron pair. Many bifurcating enzymes have recently been found to be involved in various aspects of anaerobic metabolism. There are common themes between these enzymes.

They all contain a bifurcating center that is able to perform both one and two electron chemistry, and two electron conduits that separate the high potential electron transfer pathway from the low potential electron transfer pathway.

The NfnI of Pyrococcus furiosus is a heterodimeric protein consisting of a small (NfnS,

~29 kDa) and a large (NfnL, ~53 kDa) subunit. In term of cofactor content, NfnS harbors a flavin site (S-FAD) and a [2Fe-2S] cluster. NfnL contains one flavin (L-FAD), and proximal and distal

FeS clusters (p[4Fe-4S]; d[4Fe-4S]). The crystal structure of NfnI from P. furiosus was obtained

(Figure 4.1) (87).

150

NfnI has high sequence and structural similarity to that of the Nfn from Thermotoga maritima. The NAD(H) binding site is located near the S-FAD in NfnS while the NADP(H) binding site is close to the L-FAD at NfnL. We also showed that ferredoxin, in one of our unpublished crystal structures of NfnI, binds near the distal FeS cluster of NfnL. Nfn is the simplest bifurcating enzyme to be characterized to date and has become the subject of extensive studies to elucidate the mechanism of electron bifurcation. This chapter summarizes the structural, spectroscopic and electrochemical studies NfnI and includes new data that provides additional insight into the mechanism of electron bifurcation (87,94,153,154).

4.1. The bifurcating center of P. furiosus NfnI: L-FAD

In NfnI, the L-FAD is located near the binding site the intermediate potential substrate

(NADPH) and it was proposed to be the site of bifurcation (54). FAD (and FMN) has a unique property to be the heart of bifurcation and this is the ability to undergo both one and two electron transfers. Free flavin has three oxidation states: fully oxidized (OX), semiquinone state (SQ) and fully reduced hydroquinone state (HQ) (Figure 4.2). However, its unique properties lie in the crossover of the one electron reduction potentials of flavin, where the OX/SQ couple has much lower reductive power than that of the SQ/HQ couple. The one electron OX/SQ does not accumulate and is often treated as a transition state because of its reactive nature (155,156). The reduction potentials of these species in free flavin can be measured by pulse radiolysis and by

’ redox titration and were determined to be: E0 = -314 mV for the OX/SQ couple, -219 mV for

OX/HQ and -124 mV for SQ/HQ (157). Notably, the reduction potential of the OX/HQ couple is the intermediate of the three. Thus, when a fully reduced FAD (HQ) transfers the first electron

(HQ/SQ couple), this electron is at higher potential than that of OX/HQ, and what is left behind is

151

the high energy, low potential SQ/OX. This was described as “hot”, redox active and able to reduce low potential substrates such as ferredoxin (Fd) and flavodoxin (Fld) (155). The difference in potentials of the HQ/SQ and SQ/OX in free flavin is 200 mV. In the case of NfnI, the low potential electron acceptor is P. furiosus ferredoxin, whose potential was determined to be around -345 mV at pH 8 and at 25C (46). Under physiological conditions (above 80C), the reduction potential of

Fd is estimated to be near -480 mV (48). For protein-bound FAD, the reduction potential of the one electron cross-over SQ/OX and HQ/SQ couples are expected to be much further apart so that the SQ/OX is low enough to reduce Fd.

To establish the energy landscape of NfnI, it is necessary to determine the redox potential of each cofactor within the electron transfer pathway. Square wave voltammetry (SWV) was employed to directly measure the redox potential of cofactors in NfnI at 60C, and a peak at -276 mV was assigned to FAD (Figure 4.3). This number represents the average potential (HQ/OX state) of L-FAD and S-FAD. Transient absorption spectroscopy (TAS) was used to probe for the

SQ state of NfnI L-FAD. In this experiment, a pump laser was used to induce electron transfer to the L-FAD, thus a one-electron semiquinone state of this flavin site was generated. The oxidation of the L-FAD was followed as a function of time and electron transfer rate could be determined.

It should be noted that the L-FAD and S-FAD could be distinguished by the presence of a short- lived anionic semiquinone (ASQ, peak at 366 nm) (Figure 4.4A), which has an optical signature different than the neutral semiquinone (non-crossed flavin). A short-lived ASQ with a half-life of

10ps is assigned to the L-FAD (Figure 4.4B). By using the rate of electron transfer, the half-life of the ASQ species and the distance the electron has to travel to the [2Fe-2S] and the proximal

[4Fe-4S], an estimation of the reduction potential of each electron can be made. The values were

E(HQ/ASQ) = +359 mV and E(ASQ/OX) = -911 mV (87). Our initial measurements with NfnI therefore

152

indicates that the potentials of two electrons are about 1.3 V apart. To confirm these numbers, heterologous expression of just NfnIL subunit in E. coli was carried out and a similar TAS experiment was performed on this recombinant L-subunit to re-define the energy landscape of L-

FAD in the absence of the S-FAD. L-FAD in its native protein environment had a lower reduction potential than what was determined by SWV previously (Em= -411 mV compared to Em = -276 mV). The revised reduction potentials of the HQ/ASQ and ASQ/OX couples were -911 mV and

+21 mV, respectively (Lubner and King, unpublished data). In term of mechanism, it seems that a separation of redox potentials of the electron pair by almost 1 V is necessary for electron bifurcation to occur.

4.2. The high potential electron transfer pathway

The small subunit of NfnI harbors a [2Fe-2S] cluster close to S-FAD near the NAD(H) binding site. (Figure 4.1) Structurally, this FeS cluster and S-FAD form the high potential electron transfer pathway to reduce NAD+. Previously, the redox potential of the [2Fe-2S] cluster was measured using electron paramagnetic resonance (EPR) to be +80 mV, which is unusually positive for an FeS cluster (53). This redox potential signal was unfortunately not observed in our square wave voltammetry (SWV) experiment (87). The interaction between this [2Fe-2S] cluster and the

S-FAD can be seen by EPR analysis, in which the observed signal indicates a spin-interaction between the two cofactors (87). During UV/vis redox titrations, an absorbance at 620 nm was seen as an indication of a neutral semiquinone (NSQ). The intensity of this NSQ signal increases as the concentration of NADPH (as a reductant) increases suggesting that there is an accumulation of

NSQ at the S-FAD site, in the absence of NAD+. (Figure 4.5). This NSQ signal is also seen in

NADPH-reduced NfnI by EPR, confirming that the S-FAD serves as part of electron conduit,

153

together with the [2Fe-2S] to form the high potential electron transfer pathway. The redox potential of the S-FAD measured by SWV was -276 mV (OX/HQ couple). However, this redox potential is the average potential of the two FAD in NfnI. Efforts to express and to isolate just the small subunit in P. furiosus or heterologously produce it in E. coli have so far been unsuccessful

(Nguyen, unpublished data). Since the average potential of FAD in NfnI is -276 mV, and the revised potential of the L-FAD is -411 mV (see Section 4.1), the potential of S-FAD can be estimated to be approximately -140 mV. Additionally, it should be noted that the potential of

NAD(H) under physiological condition is approximately -280 mV (48,95). Taken together, electron transfer from L-FAD to the high potential pathway proceeds from HQ/ASQ (+21 mV) to the [2Fe-2S] cluster (+80 mV), to the S-FAD (-140mV) and to the final destination, NAD+ (-280 mV). Overall this is an exergonic reaction, where NADPH (-380 mV) reduces NAD+ (-280 mV)

(Figure 4.6). However, reduction of NAD+ requires two electrons, therefore oxidation of a second

NADPH oxidation is necessary to fully reduce S-FAD and subsequently reduce NAD+.

4.3. The low potential electron transfer pathway

Square wave voltammetry was utilized to determine the reduction potentials of the two

[4Fe-4S] clusters in NfnI at 60C (Figure 4.3). The result revealed oxidation-reduction peaks at

-513 mV and -718 mV at that could be assigned to the distal and the proximal clusters, respectively.

This result also indicated that this pathway branch was indeed poised toward the low potential pathway to reduce Fd, whose redox potential was estimated to be -480 mV under physiological conditions (48). To confirm that these two [4Fe-4S] clusters formed the low potential pathway,

EPR analysis was conducted and this revealed a spin interaction between the two FeS clusters.

Hence, although the reduction of Fd by NADPH is an unfavorable reaction, NfnI is able to

154

overcome the thermodynamic barrier without any additional energy input. As discussed in Section

4.1, after the fully reduced FAD facilitates the transfer of the first high potential electron (E(HQ/ASQ)

= +21 mV), the remaining one-electron ASQ/OX couple is short-lived and of high energy as its reduction potential was calculated to be approximately -911 mV. This electron has enough reductive power to reduce the proximal FeS cluster, at -718 mV, which in turn reduces the distal cluster at -513 mV and subsequently reduces Fd (Figure 4.6).

4.4. Protein dynamics and the effect of cluster-ligands on bifurcating activity

Electron bifurcation is an elegant means to enable Nfn-type enzymes to couple the reduction of the low potential substrate (Fd) and reduction of a high potential substrate (NAD+) by an intermediate electron donor (NADPH). The one-electron crossed-over state is a unique feature of the bifurcating L-FAD, in which it allows the separation of electron pair into two independent electrons with reduction potentials nearly 1 V apart. The reduction potentials of the [2Fe-2S] (+80 mV) and the proximal [4Fe-4S] (-718 mV) cluster gate the electron transfer to ensure that there are two distinct pathways within this Nfn complex.

Hydrogen/deuterium exchange and mass spectrometry (HDX-MS) analysis of NfnI revealed that catalysis by NfnI also involved a coordinated movement of a network of amino acids

(94). The distance between the [2Fe-2S] cluster and the L-FAD is >14 Å, which is not an ideal distance for rapid electron transfer (158). Our results showed that when NADPH bound, NfnI underwent a structural rearrangement to bring the [2Fe-2S] cluster closer to the L-FAD, and the distance between them become < 13 Å, effectively increasing the electron transfer efficiency

(93,94). In addition, our HDX-MS analysis suggests that alongside the cofactors that facilitate electron transfer in NfnI, there is a physical network of amino acids that allowed allosteric

155

communication upon substrate binding. When NADPH and NAD+ simultaneously bind to NfnI, loop 210-226 (involves in the S-FAD and [2Fe-2S] coordination region) exhibits a 30% decrease in deuterium exchange compared to as purified NfnI, suggesting that this region is highly protected. This is consistent with the hypothesis that there is a structural rearrangement around the

S-FAD and [2Fe-2S] to bring these cofactors closer together (94). Similar trend was also observed with residues 245-262 that are involved in L-FAD binding. This, therefore, explains how the L-

FAD and [2Fe-2S] cluster were brought closer together (<13 Å) during catalysis. Statistical coupling analysis (SCA) was used to identify co-evolving amino acids in NfnI that participate in the communication network. A structural-based alignment of 467 Nfn-like sequences revealed a total of 137 co-evolving residues (94). By overlaying peptides that exhibited the same deuterium exchange profile in either bifurcating or confurcating direction with the SCA results, 14 residues were identify and they were within 6 Å from one another or from the cofactors (Figure 4.7). This result implies that these 14 amino acids are potentially participating in the communication pathway within NfnI. In the future they could be the subject of mutagenesis studies to elucidate the role of each amino acid in allosteric control and proton/electron transfers.

Another piece of the puzzle to help understand electron transfer in the bifurcating NfnI enzyme is how the protein prevents the second high-energy electron (ASQ/OX couple) from undergoing transfer to the high potential chain to reduce NAD+. Theoretically, Marcus theory can explain why it is not energy efficient for electron transfer from ASQ/OX (-911 mV) to the [2Fe-

2S] cluster (+80 mM). The difference in potentials places this transfer event under the Marcus inverted region and the long distance between the L-FAD and the [2Fe-2S] cluster (> 14 Å) makes it slow to transfer this electron (154). One question that arises is what determines the redox potentials of the [2Fe-2S] and the proximal [4Fe-4S] clusters? Both of these clusters contain a non-

156

cysteine ligand, an aspartate ligand (D223) to the [2Fe-2S] cluster and a glutamate (E126) to the proximal [4Fe-4S] cluster (Figure 4.8). The aspartate ligand is highly conserved in the Nfn enzymes whereas it is more difficult to determine how conserved the glutamate ligand is by sequence analysis because of the large number of cysteine residues present within NfnL. We hypothesize that these two non-cysteine ligands are a major factor in dictating the unusual redox potentials of the FeS clusters and aid in electron transfer events. To determine the importance of these non-cysteine ligands, we constructed P. furiosus strains to overexpress three different NfnI mutants that harbored the point-mutations, D223C to the [2Fe-2S], E126 to the proximal [4Fe-4S], and a combination of both D223C-E126C. The His-tagged proteins were purified from P. furiosus for kinetic and spectroscopic studies to elucidate the effect of replacing the non-cysteine ligands on (1) the bifurcating activity, (2) the reduction potential of the FeS clusters and (3) their interaction with the bifurcating L-FAD and other FeS clusters.

The D223C, E126C, and D223C-E126C proteins were purified by the same method used for the wild-type NfnI indicating that they were all intact proteins. Their bifurcating activities were measured and are shown in Figure 4.9A. Under standard assay conditions (159), the D223C and

E126C mutations caused a 60% and 85% decrease in bifurcating activity at pH 8, respectively.

The double D223C-E126C rendered NfnI almost inactive in the bifurcating assay. Interestingly, the mutations did not change the dye-linked activities, namely, reduction of benzyl viologen by

NADH or NADPH (Nguyen, unpublished data). The D223C mutant also had very high reduction activity of Fd by NADPH, independent of NAD+ activity (the so-called FNOR activity), which was 30% of the measured bifurcating activity (Figure 4.9B). This compares with < 2% of the bifurcating activity in the wild type NfnI. The E126C mutant also exhibited higher FNOR activity than the wild type (6%) but not to the extent as seen in the D223C mutant (Nguyen, unpublished

157

data). These results imply that the non-cysteine ligands are important for the bifurcating activity of NfnI. In the absence of the D223 ligand, there is non-bifurcating activity suggesting that the

[2Fe-2S] plays a key role in electron gating. The change from the aspartate to the cysteine ligand may also cause changes in the potential of this [2Fe-2S] cluster, as well as conformational changes and allosteric communication within the enzyme. Further analyses will be required to determine the effects on cluster potential.

SWV analysis on the D220C, E126C, and D220C-E126C mutants was carried out in collaboration with Dr. Anne Jones (Arizona State University) to determine the redox potentials of the [4Fe-4S] clusters, as described in (87). The SWV spectrum of D223C showed similar features to those of the wild type spectrum, in that a signal of the [2Fe-2S] was not observed. However, the very negative potential peak at -718 mV in the wild type spectrum, was not seen in SWV spectrum of the E126C mutant, suggesting that the redox potential of this cluster was shifted. The disappearance of this low potential peak was also confirmed in the double D223C-E126C mutant, indicating that our assignment of FeS clusters in (87) was correct (Jennings and Jones, unpublished data). This result implies that the loss of bifurcating activity in the E126C mutant is likely due to the changes in potential of the proximal [4Fe-4S] cluster. The potential peak -513 mV that was assigned to the distal [4Fe-4S] was seen in all of SWV spectra of the mutants. This means that the potential of the distal cluster is not affected. These results therefore suggest that electron gating in

NfnI and its bifurcating activity are strongly influenced by the non-cysteine ligands to the [2Fe-

2S] and the proximal [4Fe-4S] clusters. EPR analyses are underway to interrogate the changes in potentials of these clusters as well as the effect of non-cysteine ligands on controlling co-factor interactions in the high and low potential electron transfer pathways. UV/vis titration and transient absorption spectroscopy will also provide insight into how the NfnI cluster ligand mutations affect

158

electron accessibility and FAD optical signatures, thus influencing the energy landscape of L- and

S-FAD. HDX-MS should also be employed to monitor the changes in communication pathway within each electron transfer pathway. All this information is necessary to advance our knowledge on the structural requirements that enable Nfn to bifurcate.

159

4.5 List of figures

Figure 4.1 Structure of NfnI and cofactors orientation

(A) NfnI is a heterodimer consisting of a small subunit (NfnI-S, green) and a large subunit (NfnI-

L, cyan). Ferredoxin (Fd, brown) was computationally docked in to its binding site near the distal

[4Fe-4S] cluster of NfnI-L. (B) the distances between each co-factor. NAD binds to NfnI-S, near

S-FAD. NADPH binds to NfnI-L, close to L-FAD. Figure taken from (102)

160

NfnI-S S-FAD

NfnI-L

161

Figure 4.2 The isoalloxazine ring of FAD

Three oxidation states of flavin: the fully reduced hydroquinone (HQ or FADH), the one electron reduced semiquinone (SQ) and the fully oxidized (Q or OX, or FAD).

162

Hydroquinone (HQ) Semiquinone (SQ) Oxidized (OX) FAD FADH 1 e- reduced

e- e- (H+) ( )

163

Figure 4.3 Square wave voltammograms from NfnI

SWV of NfnI absorbed on a graphite edge electrode (top trace) and a protein-free electrode. Three oxidation-reduction peaks that could be assigned to (from left to right): the proximal [4Fe-4S]

(-718 mV), the distal [4Fe-4S] (-513 mV) and FAD (average potential of S-FAD and L-FAD,

-276 mV). The measured potentials were the average of four scans. Figure taken with permission from (87).

164

165

Figure 4.4 Evidence for a short-lived anionic semiquinone of L-FAD in NfnI

(A) TAS difference spectra of NfnI poised in reductive direction with NADPH. ASQ absorption

(366 nm) and OX bleach (455 nm) are shown. (B) TAS kinetic traces of ASQ decay and OX bleach recovery are correlated with time. The half-life of ASQ is 10.2 ± 0.2 ps, and that of OX is 10.0 ±

0.7 ps. n = 2 experiment. Figure taken with permission from (87).

166

167

Figure 4.5 Evidence for the formation of a long-lived neutral semiquinone at S-FAD

Reduction of NfnI by NADPH under steady-state conditions formed a long-lived neutral semiquinone of S-FAD that could be visualized at 620 nm. The increased of absorption at 620 nm correlated with the decrease in the concentration of the oxidized species (452 nm). Figure taken with permission from (87).

168

169

Figure 4.6 The energy landscape of NfnI

Electron transfer events in NfnI during catalysis involves the reduction of L-FAD by a hydride transfer from NADPH, the medium electron donor. The first high potential crossed-one electron

HQ/ASQ (+21 mV) reduced the positive [2Fe-2S] cluster (+80 mV), and then the S-FAD (-276 mV) in the high potential branch. The short-lived low potential ASQ/OX (-911 mV) reduces the proximal and distal [4Fe-4S] clusters (-718 mV and -513 mV, respectively) and eventually Fd

(-480 mV) in the low potential pathway. Reduction of NAD+ requires two electrons, thus, a second round of NADPH oxidation is required. The S-FAD is in its stable one electron state neutral semiquinone that can be observed optically. The reduction potential of [2Fe-2S] is taken from

(53), NADPH and NADH: (48,95), Fd: (46,48)

1Reduction potentials were calculated from TAS (87).

2Reduction potentials were taken directly from SWV (87).

170

E’ (mV) More Oxidized Lower Energy [2Fe-2S] HQ/ASQ (+80 mV) +21 mV1 S-FAD NAD (-276 mV2) NAD 2 e- (-280 mV)

2 NADPH 4 e- +ΔG e- L-FAD High potential branch (-380 mV) (HQ/OX) Low potential branch -ΔG -411 mV1 2 Fd 2 e- [4Fe–4S]-d (-480 mV) (-513 mV2) [4Fe-4S]-p 1 2 More Reduced -911 mV (-718 mV ) Higher Energy ASQ/OX

171

Figure 4.7 Putative co-evolving residues that connect the NAD(P)(H) and ferredoxin binding sites through allostery.

Structure of NfnI is shown with NfnI-S in green and NfnI-L in blue (PDB ID: 5JCA). Co-evolving residues (shown in red stick) are potentially involving in allosteric communication and are within

6 Å of each other. The number in parentheses indicate the percent conservation of the residue in the statistical coupling analysis comprising of 467 Nfn homologous sequences. Figure taken with permission from (94).

172

173

Figure 4.8 The iron-sulfur clusters of NfnI contains non-cysteine ligands

[2Fe-2S] cluster located at the S-subunit is coordinated by an aspartate ligand at position 223 (A).

The proximal [4Fe-4S] cluster (relative to the L-FAD) is coordinated by a glutamate at position

126 (B). These non-cysteine ligands are hypothesized to be the major factor in tuning the cluster potential and controlling bifurcating activity.

174

175

Figure 4.9 Activity of NfnI mutants

(A) The Nfn bifurcating activities and (B) the non-bifurcating FNOR activity were measured in the wild-type and in the D223C, E126C and D223C/E126C mutant proteins. The unit of activity in panel (A) is in moles Fd reduced per minute per mg protein (Unit mg-1). In panel (B), activity is represented as a percentage (%) of the FNOR activity compared to the bifurcating activity. (A) n = 3, (B) n = 1. (Nguyen, unpublished data)

176

A Bifurcating activity: + + Fdox + NAD + NADPH à Fdred + NADH + NADP

140

y t

i 120

v i

t 100

)

g a

/ i

f 60

( c

e 40

p 20 S 0 Pf-NNfnfnI I D[22F2e3-2CS] E[41F2e6-C4S] DD2o2u3bCl-e E1m26uCt B Non-bifurcating FNOR activity: + Fdox + NADPH à Fdred + NADP

. t

c 30 a

25 R

O 20

N 15

f 10 o 5 % 0 PfN-NfnfInI [D2F2e2-32CS] [E41F2e6-4CS] DD2o2u3bCle- E1m2u6tC

177

CHAPTER 5

DISCUSSION AND CONCLUSION

Very recently Pyrococcus furiosus has been engineered to produce biofuels and useful chemicals in a temperature-dependent manner. The successful examples of this approach include alcohol production by the AOR/AdhA pathway, ethanol production via the expression of a bifunctional AdhE, butanol formation via a 1-BuOH hybrid pathway, and 3-hydroxypropionate

(3-HP) via the 3-HP cycle (25-28,63,64). However, the yields of these products are not yet industrial relevant. With the groundwork established and proof of concept demonstrated, efforts to improve the yield can be anticipated. When engineering an organism to produce commodity chemicals, understanding the carbon and electron flux are essential since the goal is to redirect carbon and electrons toward the engineered pathway with minimal unwanted side products, such as lactate and acetate. P. furiosus is considered the E. coli of the hyperthermophile world with the availability of a robust genetic system, and the metabolic engineering of this organism has made great progress in just the last few years. However, it is still very much an emerging field compared to what has been achieved with Escherichia coli and Saccharomyces cerevisiae (yeast).

Significant effort is therefore still needed to optimize P. furiosus as a host for chemical and biofuel production. And to do this, an understanding of its central metabolism at suboptimal temperatures is essential. In particular, we must be able to fine-tune the redox flux since most of the current hybrid pathways in P. furiosus utilize reduced ferredoxin (Fd) and NAD(P)H as electron donors. Previously, it was shown that “cold-shock” stress of P. furiosus, where it was

178

grown at 72C rather than at the optimal temperature (100C), affected the expression of genes that were involved in central metabolism and amino acid biosynthesis (112). Herein, we reported that unexpectedly, at suboptimal growth temperatures, P. furiosus produced acetoin as a major metabolite via acetolactate synthase (ALS). Deletion of the gene encoding ALS increased the concentration of acetate that was produced, and its deletion in the AOR/AdhA strain significantly increased the ethanol yield to approximately 90% of the theoretical value (Chapter 2) (75). This is an excellent demonstration of how fundamental insights into metabolism can be used to impact product yields.

Moreover, we provided evidence that NfnI and Xfn were the key enzymes responsible for controlling redox flux, especially the NADPH/NADP ratio. NfnI played an essential role in sugar metabolism while Xfn was more important for P. furious during growth on peptides. The physiological function of these two enzymes is to generate NADPH for biosynthesis as deletion of either one altered the internal NADPH/NADP ratio significantly and caused growth phenotypes

(Chapter 3) (159). Additionally, the first biochemical characterization and 3D-structure of Xfn were reported. Despite sharing high sequence and structural similarity with NfnI, Xfn did not catalyze the Nfn bifurcating reaction although from mutational studies it has a key role in maintaining the redox pool (159). With growing interest in electron bifurcation as the third mechanism of energy conservation used by many anaerobic organisms (48,80,102), understanding how bifurcating enzymes mechanistically overcome the thermodynamic barrier to reduce low potential substrates using medium potential donors is especially important, not only for engineering metabolic pathways for chemical production but also for the design of bioinspired catalysts. Nfn is the simplest flavin-based bifurcating system to have been characterized so far. In

179

chapter 4, mechanistic insights into flavin-based bifurcation were discussed using NfnI from P. furiosus as the model enzyme (87,94).

5.1 Optimizing P. furiosus for biofuel and chemical production

As discussed in Chapter 2, formation of acetoin was due to the accumulation of pyruvate during growth at sub-optimal temperatures. POR is responsible for the conversion of pyruvate to acetyl-CoA, a reaction that is dependent on ferredoxin. At 75ºC, the specific activity of POR decreased by approximately 50% compared to its activity at 90ºC (160). Hence, expression of the gene encoding for ALS increased in response to pyruvate accumulation. ALS catalyzes the conversion of pyruvate to acetolactate, which spontaneously breaks down to acetoin in the absence of an acetolactate dehydrogenase (ALDH) (Figure S2.2) (75). The optimal temperature for production of acetoin was shown to be 75ºC (Figure 2.1B), the highest temperature for microbial acetoin production reported to date (73,75).

Additionally, it is possible that P. furiosus could be engineered to reduce acetoin to 2,3- butanediol, which is another chemical of industrial interest, by insertion of genes encoding for alcohol dehydrogenase B (AdhB) into the P. furiosus genome (Figure 5.1A). A candidate for the adhB gene donor is Thermoanerobacter sp. X514 since it contains AdhB and its AdhA was well- expressed in P. furiosus (25). AdhB uses NADH as the electron donor, which is likely one of the limiting factors in 2.3-butanediol production. A bacterial membrane-bound complex known as Rnf could be expressed in P. furiosus to generate NADH. Rnf consist of six subunits and is an ion translocating ferredoxin:NAD+ oxidoreductase. It has been shown in acetogens that the Rnf complex is responsible for generating Na+ gradient for energy conservation and its deletion in

Clostridium ljungdahlii hindered both autotrophic and heterotrophic growths of this organism

180

(161). Moreover, the Rnf complex had been shown to increase the ethanol yield in C. thermocellum

(124). Insertion of the genes encoding for Rnf into the genome of P. furiosus might therefore be very useful in generating additional NADH. The sodium ion gradient formed by Rnf would also be beneficial as P. furiosus has a sodium dependent ATP synthase (Figure 5.1B) (49).

Also reported in Chapter 3 was that the deletion of ALS abolishes acetoin production. As a result, carbon was redirected toward acetate and subsequently to ethanol via the AOR/AdhA pathway. In particular, deletion of als in the AOR/AdhA strain increased the ethanol yield by 40% reaching 90% of the theoretically yield, compared to 50% in the parent AOR/AdhA strain (Figure

2.5). There is no doubt that deletion of als can improve ethanol production and will likely increase the product yield for any engineered pathway that utilizes metabolites from pyruvate. These include acetyl-CoA, which is the precursor of butanol and 3-HP (26,27) and as well as ethanol

(28). However, deletion of als may not be the best option to improve product yields because ALS is also responsible for the biosynthesis of branch chain amino acids, hence strains with the als deletion typically showed a slight lag phase during growth (Figure 2.4A). An alternative approach is to cultivate P. furiosus recombinant strains under conditions in which als expression is suppressed via product inhibition. Since ALS in P. furiosus is responsible for synthesis of valine, leucine and isoleucine, inhibition of ALS can be achieved by using peptide rich medium. When growth media containing 2 or 5 g L-1 yeast extract were used, the control COM1c and AOR/AdhA strain produced less acetoin by approximately 50% (Nguyen, unpublished data). Another approach that could be beneficial that does not rely on inhibition or deleting als is to increase the pyruvate flux toward acetyl-CoA and acetate production by heterologous expression of a POR that has a maximal specific activity at 70-75ºC. Bacteria or archaea with optimal growth temperatures in the range 60-80 ºC could be investigated for gene donors of such POR. With targeted improvements

181

by a more temperature-appropriate POR, pyruvate will be less likely to accumulate and cause side- reactions such as the formation of acetoin (Figure 5.2).

Overall, in vivo ethanol, butanol and 3-HP formation is a stepping stone to eventually establish P. furiosus as a platform organism for the production of various fuels and chemicals. At present, product yields in small scale cultures were low but provide proof of concept. There is only one occasion that we observed a 90% theoretical yield of ethanol formation, but this used a concentrated cell suspension assay (Figure 2.5) (75). Perhaps a better approach to significantly improve biofuel and biochemical yield is to use in vitro cell-suspension conversion of sugars

(maltose) to bioproducts. In in vitro cell-suspension assay reported in Chapter 2, 1.7 g L-1 (37 mM) ethanol was formed from 3.4 g L-1 (10 mM) supplemented maltose after 18 hours. In a continuous assay with maltose supplemented every 12 hours, the ethanol yield should be even higher. For example, fed-batch fermenter of an ethanol producing Thermoanaerobacterium sacchrolyticum strain (a thermophilic bacterium that grows optimally at 50ºC) produced up to 37 g L-1 ethanol, using a sugar mixture (glucose, xylose, galactose, and mannose) of 330 g L-1 that was supplemented at 3 g h-1 for each sugar (162). This approach is very feasible for engineered pathways that are redox balanced, such as the AOR/AdhA pathway in P. furiosus (25). However, in more complex pathways involved in producing butanol and 3-HP, the supply of reducing equivalents such as Fd, NADPH or NADH is not balanced and this is still an obstacle to be overcome.

In Chapter 3, we discussed the physiological function of NfnI and Xfn in P. furiosus, which are in general to generate NADPH for biosynthesis (159). More precisely, NfnI was shown to have a major effect in mediating redox balance for hydrogen production by P. furiosus. It is not clear how changing NfnI expression would benefit the current engineered P. furiosus strains because

182

both deletion and overexpression of NfnI led to very noticeable phenotypes (Figure 3.2). The logical explanation would be that NfnI when overexpressed might out compete the energy conserving MBH for reduced Fd, which in turn affects generation of the sodium ion gradient used for ATP synthesis. However, in strains producing biofuels and chemicals, NADPH will quickly be consumed, for example, by AdhA, and it is possible that NfnI overexpression would cause much less or no growth defect. There is little doubt that the NADPH-dependent product formation pathways are primarily dependent on NfnI for NADP+ recycling. It was shown previously in T. sacchrolyticum that deletion of its Nfn decreased ethanol yield in the NADPH-dependent ethanol- producing strain but not in the NADH-dependent strain (125). With our current understanding of

P. furiosus redox metabolism, in which NfnI plays a critical role, we argue that perhaps, the first step would be to increase NfnI expression in the biofuel and chemical producing strains requiring

NADPH. We should also consider designing a synthetic enzyme and/or catalyst based on NfnI to generate NADPH from NADH and a low redox potential artificial electron donor (such as benzyl viologen), independent of Fd. In that case, a much more complete knowledge of the catalytic, biophysical and electronic properties of NfnI are essential. Another issue is utilizing Xfn in the engineered P. furiosus strains. However, more information on its catalytic properties are needed before a suitable strategy can be designed.

5.2. NfnI as model for an efficient bioinspired catalyst

Bioinspired catalysts are “materials” designed to mimic the structure and catalytic function of enzymes. They often contain the minimum structural requirement while having similar or better catalytic activities than the natural enzyme. These catalysts are also more stable in harsh reaction conditions (high temperature, high pressure, and extreme pH, etc.) and their catalytic activity is

183

tunable to favor a certain reaction direction (163). Electron bifurcation is a fascinating mechanism that some enzymes use to achieve difficult reactions that involve reduction of a very low redox potential substrate, for example, ferredoxin, by a medium potential electron donor, for instance,

NADPH. This unfavorable reaction is coupled to a favorable one, so that the same electron donor reduces a second substrate but of high potential. The energy that is required to drive the endergonic half-reaction is overcome by fine-tuning the energy landscape of the enzyme’s cofactors so that the process is favorable for each electron transfer step. There is no doubt that a bifurcating catalyst could be an important innovation in biotechnology in general. A bifurcation complex contains three main components: a bifurcating center and two electron transfer pathways (one of high and one of low potential). The bifurcating center is characterized by its capacity to facilitate both one and two electron chemistry. Quinone and flavin are two well-known bifurcating centers that have been characterized in detail. Recently, metal centers were proposed to function as bifurcating site as well (102). Such a concept stemmed from our work on the hydrogen-evolving hydrogenase complex from the hyperthermophilic bacterium Thermotoga maritima, Tm-HydABC. This enzyme reversibly oxidizes the lower potential electron donor Fd (-480 mV) and the higher potential donor NADH (-280 mV) and evolves H2 gas (164). In this the complex, the catalytic H- cluster is proposed to be the site of bifurcation, as this reacts with the medium potential donor, H2 gas. The FMN in this enzyme is thought to interact with NAD+ and is not the bifurcating site

(102,164).

Destabilization of the L-FAD anionic semiquinone (ASQ) state

Out of all flavin-based bifurcating complexes that have been characterized, NfnI is the simplest in terms of cofactor content and subunit composition. Its structural, biochemical,

184

biophysical, and electrochemical properties have been characterized in great details (87,94).

Although the fundamental mechanism by which the electron pair is split into high and low potential electron transfer pathways and the energy landscape of NfnI are well-understood, there are still fundamental questions that need to be further answered. For example, how does the protein environment control the functions of the bifurcating and the non-bifurcating flavins in a multi- flavin enzyme complex? How far apart can the redox potentials of the one-electron cross-over species be? How does the environment of the bifurcating flavin aid in generating the short-lived low potential anionic semiquinone (ASQ)? Which amino acids control the redox properties of the

[2Fe-2S] and the [4Fe-4S] clusters?

In terms of the ASQ, upon close inspection of the L-FAD near the N5 of the isoalloxazine ring, there are two residues that are directly involved in a hydrogen network, the arginine at position 201 (R201) and the aspartate at position 307 (D307) (Figure 5.3). The R201 with its positive charge is expected to stabilize the ASQ state. However, in NfnI, the positive charge of

R201 is neutralized by D307. This should destabilize the ASQ, resulting in the short-lived species that was observed by the transient absorption spectroscopy (TAS) experiment. If this is true, disruption of the R201 and D307 interaction by mutating either residue to the opposite charge amino acid should stabilize this ASQ. The expected outcome would be a longer-lived ASQ and potentially loss of bifurcating activity. Mutating R201 to glutamate (R201Q) should disrupt this interaction but retain the bidentate H-bond with N5 due to the side chain of glutamate while the

D307N mutation should directly disrupt the R201-D307 charge neutralization. However, our attempt at constructing the R201Q or D307N of NfnI in P. furiosus was unsuccessful as the point mutations reverted back to the wild-type (R201 and D307) after three consecutive colony purifications (Nguyen, unpublished data). This result suggests that the mutations might be lethal

185

due to loss of bifurcating activity. Additionally, a study of the L-FAD environment by hydrogen- deuterium exchange mass spectrometry (HDX-MS) revealed an extensive hydrogen bonding network around the N5 of the flavin ring that the R201 is a part of (87). This extends toward the surface of the protein and a substantial confirmation change around this region was also observed upon NADP(H) binding (87,94). Hence, the hydrogen bonding network of L-FAD N5 is highly controlled by substrate binding and the R201-D307 interaction might be the key factor that favors the electron bifurcation at L-FAD site. To investigate this, it might be possible to generate NfnI mutant in a wild-type NfnI background strain so that the mutants do not dramatically influence growth.

The roles of the non-cysteine ligands on the catalytic and biophysical characteristics of NfnI

It appears that the bifurcating L-FAD is not the only factor that contributes to bifurcation in NfnI. We provided evidence in Chapter 4 that the ligands to the iron-sulfur clusters might tune their redox potentials. In addition, we showed that the non-cysteine ligands to each of the [2Fe-

2S] and the proximal [4Fe-4S] clusters are necessary for bifurcating activity. The Asp ligand to the [2Fe-2S] cluster is likely to be the key in controlling electron gating since mutating this Asp ligand to the canonical cysteine ligand allowed uncoupling of the endergonic and the exergonic reaction (Figure 4.9). The D223C (to the [2Fe-2S] cluster) and the E126C (to the proximal [4Fe-

4S] cluster) ligand changes have been shown to prevent bifurcating activity in vitro, but do not abolish NAD(P)H-dye linked non-bifurcating activity (Nguyen, unpublished data). These results suggest that the electron transfer pathways in the NfnI mutants are still intact but they are uncoupled. Using square wave voltammetry (SWV), we were able to conclude that the Glu ligand to the [4Fe-4S] cluster contributes to tuning the redox potential of this cluster based on the

186

disappearance of the peak corresponding the -718 mV peak in the voltammogram (Jennings and

Jones, unpublished data). We speculate that the reduction potential of this cluster was shifted to that of the distal [4Fe-4S] cluster (-513 mV) and therefore was hidden. Even so, the overall reduction potential of the electron pair of L-FAD must be affected because thermodynamically the

-911 mV ASQ/OX couple should be able to reduce both the [2Fe-2S] and the proximal [4Fe-4S] cluster. Hence, it is likely that the E126C mutation changes the cluster environment and disrupts the communication pathway with the L-FAD. A similar outcome could be expected for the D223C mutation at the [2Fe-2S] cluster and we expect a change in redox potential of this cluster. Since the bifurcating activity decreased by 60% compared to the wild-type but not to the extent observed in the E126C mutant (Figure 4.9), it is possible that the D223 residue is essential for controlling catalytic turnover in NfnI. In previous studies, it was hypothesized that upon substrate binding, the small subunit, Nfn-S underwent a rigid body movement to bring the [2Fe-2S] cluster closer to the

L-FAD (93), and a network of amino acids acted as a signaling pathway (94). Together with our preliminary data obtained from kinetic analysis with the D223C mutant, there is a likelihood that this D223 ligand plays a critical role in promoting the conformation changes and signaling network in the high potential pathway of NfnI. Clearly, experiments are needed to determine if the D223C and the E126C cluster mutation affect the reduction potentials of the electron pair via TAS, and the [2Fe-2S] and the proximal [4Fe-4S] cluster via electron paramagnetic resonance (EPR) titration to remap the energy landscape of NfnI. In addition, HDX-MS should be employed to monitor changes in structural confirmations and the communication network surrounding the L-

FAD, as well as the electron communication pathways within each electron transfer branch.

187

Comparative studies enable understanding of key features in bifurcating enzymes: Xfn versus NfnI

Our current perspective is that Xfn is not a bifurcating type of Nfn despite the striking similarity in term of sequence, cofactor content and overall structure between it and NfnI (159). It is unclear at this time whether Xfn bifurcates. The L-FAD of Xfn had a short-lived ASQ, with a half-life of ~6 ps by TAS analysis (Lubner and King, unpublished data) and SWV produced a voltammogram that contained similar peaks to that seen in NfnI (Jennings and Jones, unpublished data). These results suggest that the lower potential branch (to reduce Fd) has a similar energy landscape to that of NfnI. The short-lived ASQ is a necessary feature of bifurcating enzymes (to generate the low potential electron) but is not enough to promote bifurcation alone (153). Further comparative studies of Xfn and NfnI revealed that while the L-subunits are highly similar (97 % identity), the major differences lie within the small subunit XfnS (approximately 38% sequence identity to NfnI-S). First, there is an amino acid loop (164-178) that prevents NAD(H) from binding to the S-FAD site. This is consistent with the absence of any NADH-dependent activity for Xfn. Hence, it was proposed that Xfn is a non-bifurcating ferredoxin NADP oxidoreductase enzyme (FNOR) (159).

However, biophysical studies of Xfn have provided mixed results. In a redox titration monitored with visible spectroscopy, reduction of Xfn with titanium citrate (Ti-citrate) or NADPH did not yield a stable neutral semiquinone (NSQ) (~620 nm) that could be assigned to the S-FAD as seen in NfnI (Schut and Nguyen, unpublished data). However, EPR spectra of NAD(P)H- reduced Xfn revealed an NSQ radical signal, similar to NfnI although with weaker intensity

(Mulder and King, unpublished data). To further investigate Xfn, HDX-MS is necessary to provide information on the communication pathway and allosteric effects upon substrate binding, with a focus on the loop 164-178 near the S-FAD. Perhaps under the right physiological conditions, the

188

loop becomes mobile to allow access to NAD(H). For example, it was shown that the expression of genes encoding for Xfn increased by 50-fold during growth on peptides (Figure S3.5B).

However, it is not clear that if this loop regulates NAD(H) access to the S-subunit, or how this provides any benefit in maintaining the redox pools in P. furiosus under peptide growth conditions.

Alternatively, the third substrate of Xfn may not be NAD(H) but a small sulfur-containing compound that is only produced when P. furiosus is grown with elemental sulfur (S0), which is required for growth on peptides. In this case, a sulfur-based metabolomics study might reveal key compounds that could potentially be third substrate of Xfn, in addition to Fd and NADP(H).

5.3 Conclusion and outlook

P. furiosus has been metabolically engineered to produce biofuels and chemicals in a temperature-dependent manner. The host was cultivated near 100 ºC, the optimal temperature for cell-mass generation, and was then shifted to lower temperature near 70ºC for production of biofuels and chemicals via specific engineered pathways. Understanding the host redox metabolism at suboptimal growth temperature is therefore necessary to improve the product yields in P. furiosus. At suboptimal growth temperatures, P. furiosus produced acetoin as a major metabolite by a biosynthetic acetolactate synthase (ALS). This is a way to produce acetoin for industrial uses in a temperature-dependent manner. Acetoin production was the result of a bottleneck step after the glycolysis pathway in which pyruvate is accumulate due to a decrease in pyruvate ferredoxin oxidoreductase (POR) activity at temperatures below 80ºC. Deletion of the genes encoding for ALS abolished acetoin production and improved the ethanol yield in the alcohol producing P. furiosus strain that contained the AOR/AdhA pathway. Deletion of als could improve product yield significantly for any engineered pathway in P. furiosus that required acetyl-

189

CoA or acetate as a feedstock. Moreover, redox balance is an important aspect when engineered microbes are used for biofuel and chemical production.

In P. furiosus, the redox pools of ferredoxin and NAD(P)(H) are thought to be maintained primarily by two ferredoxin:NADP oxidoreductases: NfnI and Xfn. Their physiological function is thought to be generating NADPH, which is used as electron donor for fuel and chemical synthesis. NfnI is more relevant in a metabolic engineering context because its expression and functions are better understood. It is involved in sugar metabolism, which is the primary growth condition for cultivating P. furiosus engineered strains for generation of reduced products.

Debottlenecking and redox balance are clearly two crucial aspects to consider when metabolic engineering P. furiosus. In terms of carbon flow, we propose that replacement of the host’s POR for a more temperature-appropriate enzyme will promote both electron and carbon flow toward the engineered pathways. In terms of electron flow, the insertion and expression of a membrane- bound Rnf complex could not only aid in recycling Fd and NAD+ but also provide a sodium ion gradient, which may be bioenergetic advantage. Herein, we also summarize in detailed bifurcating mechanism for NfnI. This can be used to advance our understanding of bifurcating enzymes in general and guide the design of novel bioinspired catalysts.

190

5.4 List of figures

Figure 5.1 The proposed pathway for 2.3-butanediol production in P. furiosus

Schematic of proposed pathway from pyruvate to 2,3-butandiol formation (A). At temperatures ranging from 70 to 80ºC, P. furiosus can be engineered to produce 2,3- butanediol by insertion and expression of genes encoding for alcohol dehydrogenase B (AdhB), potentially taken from

Thermoanaerobacter sp. X514. AdhB catalyzes the conversion of acetoin to 2.3-butanediol and consumes one NADH as electron donor. It is also proposed that expression of a sodium translocating ferredoxin:NAD+ oxidoreductase (Rnf) into P. furiosus will be beneficial for this pathway in term of providing the necessary NADH for the AdhB reaction (B).

191

A Non- ALS enzymatic AdhB

NADH NAD+ CO2 CO2 2 Pyruvate Acetolactate Acetoin 2,3-Butanediol

B Glucose

2Fdred 2Fdox GAP 2Fdox Rnf GAPOR 2Fdred

PEP NADH NAD ADP + Pi PK ATP ALS AdhB Pyruvate Acetolactate Acetoin 2,3-Butanediol POR 2Fdox 2Fdred Acetyl CoA

ADP + Pi ACS ATP Acetate

192

Figure 5.2 Proposed engineering of P. furiosus for improving biofuel and biochemical production

Heterologous expression of temperature-appropriated pyruvate ferredoxin oxidoreductase (POR, yellow) and proton translocating ferredoxin:NAD+ oxidoreductase (Rnf, yellow) into P. furiosus to increase the carbon flux from pyruvate to acetyl-CoA and acetate and to provide additional

NADH that can be utilized by NfnI or any NADH-dependent biofuel formation pathway. Ethanol production via the AOR/AdhA pathway (shown in grey) requires one reduced Fd and one NADPH as electron donors (25). The 3-HP cycle (highlighted in green) requires two NADPH and one ATP

(26), whereas two NADH are needed to power the 1-BuOH pathway for butanol production (27).

This proposed model focuses on maximizing ATP synthesis via pyruvate kinase (PK), acetyl-CoA synthase (ACS) and ATP synthase. It is noted that in the 3-HP cycle and 1-BuOH pathway, the

ATP generation step by ACS is omitted.

193

2Fdred 2Fdox Glucose

H+/Na+ MBH H+/Na+ GAP H 2Fdox 2 GAPOR NADPH + 2Fdred SHI 2H 2NADPH 2NADP NADP + ATP + ADP PEP ATP ADP + Pi PK Na+ ATPase Na+ ATP 2Fdox Pyruvate 2Fdred ADP + Pi 3-HPcycle 2Fdox 3-HP POR 2Fdred NADH Acetyl CoA NfnI 2NADPH Butanol ADP + Pi NAD 1-BuOH ACS ATP 2NADP Acetate Na+ Rnf Na+ Fd + NADPH 2NAD 2NADH red 2Fdox AOR/AdhA 2Fdred

Fdred + NADPH Ethanol

Inside Outside

194

Figure 5.3 Hydrogen-bonding network around N5 of L-FAD in NfnI

D307 and R201 are proposed to be the critical residues within the hydrogen-bonding network at the L-FAD site that control flavin reactivity. The R201 positive charge is neutralized by D307 and is thought to destabilize the semiquinone state.

195

D307

R201 N5

V452

V100

196

REFERENCES

1. Leigh, J. A., Albers, S. V., Atomi, H., and Allers, T. (2011) Model organisms for genetics

in the domain Archaea: methanogens, halophiles, Thermococcales and Sulfolobales. FEMS

Microbiol Rev 35, 577-608

2. Winker, S., and Woese, C. R. (1991) A definition of the domains Archaea, Bacteria and

Eucarya in terms of small subunit ribosomal RNA characteristics. Syst Appl Microbiol 14,

305-310

3. Blochl, E., Burggraf, S., Fiala, G., Lauerer, G., Huber, G., Huber, R., Rachel, R., Segerer,

A., Stetter, K. O., and Volkl, P. (1995) Isolation, taxonomy and phylogeny of

hyperthermophilic microorganisms. World J Microbiol Biotechnol 11, 9-16

4. Bell, S. D., and Jackson, S. P. (1998) Transcription and translation in Archaea: a mosaic

of eukaryal and bacterial features. Trends Microbiol 6, 222-228

5. Olsen, G. J., and Woese, C. R. (1996) Lessons from an Archaeal genome: what are we

learning from Methanococcus jannaschii? Trends Genet 12, 377-379

6. Olsen, G. J., and Woese, C. R. (1997) Archaeal genomics: an overview. Cell 89, 991-994

7. Koonin, E. V., Mushegian, A. R., Galperin, M. Y., and Walker, D. R. (1997) Comparison

of archaeal and bacterial genomes: computer analysis of protein sequences predicts novel

functions and suggests a chimeric origin for the archaea. Mol Microbiol 25, 619-637

8. Chaban, B., Ng, S. Y., and Jarrell, K. F. (2006) Archaeal habitats--from the extreme to the

ordinary. Can J Microbiol 52, 73-116

197

9. Robertson, C. E., Harris, J. K., Spear, J. R., and Pace, N. R. (2005) Phylogenetic diversity

and ecology of environmental Archaea. Curr Opin Microbiol 8, 638-642

10. Stein, J. L., and Simon, M. I. (1996) Archaeal ubiquity. Proc Natl Acad Sci U S A 93, 6228-

6230

11. Huber, R., Dyba, D., Huber, H., Burggraf, S., and Rachel, R. (1998) Sulfur-inhibited

Thermosphaera aggregans sp. nov., a new genus of hyperthermophilic archaea isolated

after its prediction from environmentally derived 16S rRNA sequences. Int J Syst Bacteriol

48 Pt 1, 31-38

12. Brochier, C., Gribaldo, S., Zivanovic, Y., Confalonieri, F., and Forterre, P. (2005)

Nanoarchaea: representatives of a novel archaeal phylum or a fast-evolving euryarchaeal

lineage related to Thermococcales? Genome Biol 6, R42

13. Brochier-Armanet, C., Boussau, B., Gribaldo, S., and Forterre, P. (2008) Mesophilic

Crenarchaeota: proposal for a third archaeal phylum, the Thaumarchaeota. Nat Rev

Microbiol 6, 245-252

14. Auchtung, T. A., Takacs-Vesbach, C. D., and Cavanaugh, C. M. (2006) 16S rRNA

phylogenetic investigation of the candidate division "Korarchaeota". Appl Environ

Microbiol 72, 5077-5082

15. Barns, S. M., Delwiche, C. F., Palmer, J. D., and Pace, N. R. (1996) Perspectives on

archaeal diversity, thermophily and monophyly from environmental rRNA sequences.

Proc Natl Acad Sci U S A 93, 9188-9193

16. Klenk, H. P., Renner, O., Schwass, V., and Zillig, W. (1992) Nucleotide sequence of the

genes encoding the subunits H, B, A' and A'' of the DNA-dependent RNA polymerase and

the initiator tRNA from Thermoplasma acidophilum. Nucleic Acids Res 20, 5226

198

17. Jones, W. J., Nagle, D. P., Jr., and Whitman, W. B. (1987) Methanogens and the diversity

of archaebacteria. Microbiol Rev 51, 135-177

18. Zillig, W., and Reysenbach, A. L. (1992) The order Thermococcales. The Prokaryotes,

702-706

19. Segerer, A. H., Burggraf, S., Fiala, G., Huber, G., Huber, R., Pley, U., and Stetter, K. O.

(1993) Life in hot springs and hydrothermal vents. Orig Life Evol Biosph 23, 77-90

20. Adams, M. W. W. (1994) Biochemical diversity among sulfur-dependent,

hyperthermophilic microorganisms. FEMS Microbiol Rev 15, 261-277

21. Adams, W. W. M., and Kelly, R. M. (1995) Enzymes from microorganisms in extreme

environments. Chem Eng News, 37-39

22. Adams, M. W. W., and Kelly, R. M. (1998) Finding and using hyperthermophilic enzymes.

Trends Biotechnol 16, 329-332

23. Wheaton, G., Counts, J., Mukherjee, A., Kruh, J., and Kelly, R. M. (2015) The confluence

of heavy metal biooxidation and heavy metal resistance: Implications for bioleaching by

extreme thermoacidophiles. Minerals 5, 397-451

24. Pavlov, A. R., Pavlova, N. V., Kozyavkin, S. A., and Slesarev, A. I. (2004) Recent

developments in the optimization of thermostable DNA polymerases for efficient

applications. Trends Biotechnol 22, 253-260

25. Basen, M., Schut, G. J., Nguyen, D. M., Lipscomb, G. L., Benn, R. A., Prybol, C. J.,

Vaccaro, B. J., Poole, F. L., 2nd, Kelly, R. M., and Adams, M. W. (2014) Single gene

insertion drives bioalcohol production by a thermophilic archaeon. Proc Natl Acad Sci U

S A 111, 17618-17623

199

26. Keller, M. W., Schut, G. J., Lipscomb, G. L., Menon, A. L., Iwuchukwu, I. J., Leuko, T.

T., Thorgersen, M. P., Nixon, W. J., Hawkins, A. S., Kelly, R. M., and Adams, M. W.

(2013) Exploiting microbial hyperthermophilicity to produce an industrial chemical, using

hydrogen and carbon dioxide. Proc Natl Acad Sci U S A 110, 5840-5845

27. Keller, M. W., Lipscomb, G. L., Loder, A. J., Schut, G. J., Kelly, R. M., and Adams, M.

W. (2015) A hybrid synthetic pathway for butanol production by a hyperthermophilic

microbe. Metab Eng 27, 101-106

28. Keller, M. W., Lipscomb, G. L., Nguyen, D. M., Crowley, A. T., Schut, G. J., Scott, I.,

Kelly, R. M., and Adams, M. W. W. (2017) Ethanol production by the hyperthermophilic

archaeon Pyrococcus furiosus by expression of bacterial bifunctional alcohol

dehydrogenases. Microb Biotechnol 10, 1535-1545

29. Fiala, G., and Stetter, K. (1986) Pyrococcus furiosus sp. nov. represents a novel genus of

marine heterotrophic archaebacteria growing optimally at 100°C. Arch. Microbiol. 145,

56-61

30. Verhees, C. H., Kengen, S. W., Tuininga, J. E., Schut, G. J., Adams, M. W., De Vos, W.

M., and Van Der Oost, J. (2003) The unique features of glycolytic pathways in Archaea.

Biochem J 375, 231-246

31. Kengen, S. W., de Bok, F. A., van Loo, N. D., Dijkema, C., Stams, A. J., and de Vos, W.

M. (1994) Evidence for the operation of a novel Embden-Meyerhof pathway that involves

ADP-dependent kinases during sugar fermentation by Pyrococcus furiosus. J Biol Chem

269, 17537-17541

200

32. Mukund, S., and Adams, M. W. W. (1995) Glyceraldehyde-3-phosphate ferredoxin

oxidoreductase, a novel tungsten-containing enzyme with a potential glycolytic role in the

hyperthermophilic archaeon Pyrococcus furiosus. J Biol Chem 270, 8389-8392

33. Blamey, J. M., and Adams, M. W. (1993) Purification and characterization of pyruvate

ferredoxin oxidoreductase from the hyperthermophilic archaeon Pyrococcus furiosus.

Biochim Biophys Acta 1161, 19-27

34. Mai, X., and Adams, M. W. W. (1996) Purification and characterization of two reversible

and ADP-dependent acetyl coenzyme A synthetases from the hyperthermophilic archaeon

Pyrococcus furiosus. J Bacteriol 178, 5897-5903

35. Schonheit, P., and Schafer, T. (1995) Metabolism of hyperthermophiles. World J Microbiol

Biotechnol 11, 26-57

36. Schut, G. J., Brehm, S. D., Datta, S. F., and Adams, M. W. W. (2003) Whole-genome DNA

microarray analysis of a hyperthermophile and an archaeon: Pyrococcus furiosus grown

on carbohydrates or peptides. J Bacteriol 185, 3935-3947

37. Mai, X., and Adams, M. W. (1996) Characterization of a fourth type of 2-keto acid-

oxidizing enzyme from a hyperthermophilic archaeon: 2-ketoglutarate ferredoxin

oxidoreductase from Thermococcus litoralis. J Bacteriol 178, 5890-5896

38. Heider, J., Mai, X., and Adams, M. W. (1996) Characterization of 2-ketoisovalerate

ferredoxin oxidoreductase, a new and reversible coenzyme A-dependent enzyme involved

in peptide fermentation by hyperthermophilic archaea. J Bacteriol 178, 780-787

39. Kletzin, A., and Adams, W. W. M. (1996) Molecular and phylogenetic characterization of

pyruvate and 2-ketoisovalerate ferredoxin oxidoreductases from Pyrococcus furiosus and

pyruvate ferredoxin oxidoreductase from Thermotoga maritima. J Bacteriol 178, 248-257

201

40. Ma, K., Hutchins, A., Sung, S. J., and Adams, M. W. W. (1997) Pyruvate ferredoxin

oxidoreductase from the hyperthermophilic archaeon, Pyrococcus furiosus, functions as a

CoA-dependent pyruvate decarboxylase. Proc Natl Acad Sci U S A 94, 9608-9613

41. Klump, H., Di Ruggiero, J., Kessel, M., Park, J. B., Adams, M. W., and Robb, F. T. (1992)

Glutamate dehydrogenase from the hyperthermophile Pyrococcus furiosus. Thermal

denaturation and activation. J Biol Chem 267, 22681-22685

42. Johnson, M. K., Spiro, T. G., and Mortenson, L. E. (1982) Resonance Raman and electron

paramagnetic resonance studies on oxidized and ferricyanide-treated Clostridium

pasteurianum ferredoxin. Vibrational assignments from 34S shifts and evidence for

conversion of 4 to 3 iron-sulfur clusters via oxidative damage. Vibrational assignments

from 34S shifts and evidence for conversion of 4 to 3 iron-sulfur clusters via oxidative

damage. J Biol Chem 257, 2447-2452

43. Fukuyama, K. (2004) Structure and function of plant-type ferredoxins. Photosynth Res 81,

289-301

44. Fitzgerald, M. P., Rogers, L. J., Rao, K. K., and Hall, D. O. (1980) Efficiency of ferredoxins

and flavodoxins as mediators in systems for hydrogen evolution. Biochem J 192, 665-672

45. Bengelsdorf, F. R., Gerischer, U., Langer, S., Zak, M., and Kazda, M. (2013) Stability of

a biogas-producing bacterial, archaeal and fungal community degrading food residues.

FEMS Microbiol Ecol 84, 201-212

46. Aono, S., Bryant, F. O., and Adams, M. W. (1989) A novel and remarkably thermostable

ferredoxin from the hyperthermophilic archaebacterium Pyrococcus furiosus. J Bacteriol

171, 3433-3439

202

47. Conover, R. C., Kowal, A. T., Fu, W. G., Park, J. B., Aono, S., Adams, M. W., and Johnson,

M. K. (1990) Spectroscopic characterization of the novel iron-sulfur cluster in Pyrococcus

furiosus ferredoxin. J Biol Chem 265, 8533-8541

48. Buckel, W., and Thauer, R. K. (2013) Energy conservation via electron bifurcating

ferredoxin reduction and proton/Na(+) translocating ferredoxin oxidation. Biochim

Biophys Acta 1827, 94-113

49. Sapra, R., Bagramyan, K., and Adams, M. W. W. (2003) A simple energy-conserving

system: proton reduction coupled to proton translocation. Proc Natl Acad Sci U S A 100,

7545-7550

50. Schut, G. J., Nixon, W. J., Lipscomb, G. L., Scott, R. A., and Adams, M. W. W. (2012)

Mutational analyses of the enzymes involved in the metabolism of hydrogen by the

hyperthermophilic archaeon Pyrococcus furiosus. Front Microbiol 163,

10.3389/fmicb.2012.00163

51. Ma, K., and Adams, M. W. (2001) Hydrogenases I and II from Pyrococcus furiosus.

Methods Enzymol 331, 208-216

52. Ma, K., and Adams, M. W. (1994) Sulfide dehydrogenase from the hyperthermophilic

archaeon Pyrococcus furiosus: a new multifunctional enzyme involved in the reduction of

elemental sulfur. J Bacteriol 176, 6509-6517

53. Ma, K., and Adams, M. W. W. (2001) Ferredoxin:NADP oxidoreductase from Pyrococcus

furiosus. Methods Enzymol 334, 40-45

54. Demmer, J. K., Huang, H., Wang, S., Demmer, U., Thauer, R. K., and Ermler, U. (2015)

Insights into flavin-based electron bifurcation via the NADH-dependent reduced

ferredoxin:NADP oxidoreductase structure. J Biol Chem 290, 21985-21995

203

55. Lipscomb, G. L., Keese, A. M., Cowart, D. M., Schut, G. J., Thomm, M., Adams, M. W.

W., and Scott, R. A. (2009) SurR: A transcriptional activator and repressor controlling

hydrogen and elemental sulfur metabolism in Pyrococcus furiosus. Mol Microbiol 71, 332-

349

56. Yang, H., Lipscomb, G. L., Schut, G. J., Thomm, M., Adams, M. W. W., Wang, B. C., and

Scott, R. A. (2010) SurR regulates hydrogen production in Pyrococcus furiosus by a sulfur-

dependent redox switch. Mol Microbiol 77, 1111-1122

57. Lipscomb, G. L., Schut, G. J., Scott, R. A., and Adams, M. W. W. (2017) SurR is a master

regulator of the primary electron flow pathways in the order Thermococcales. Mol

Microbiol, 10.1111/mmi.13668

58. Schicho, R. N., Ma, K., Adams, M. W., and Kelly, R. M. (1993) Bioenergetics of sulfur

reduction in the hyperthermophilic archaeon Pyrococcus furiosus. J Bacteriol 175, 1823-

1830

59. Bridger, S. L., Clarkson, S. M., Stirrett, K., DeBarry, M. B., Lipscomb, G. L., Schut, G. J.,

Westpheling, J., Scott, R. A., and Adams, M. W. W. (2011) Deletion strains reveal

metabolic roles for key elemental sulfur-responsive proteins in Pyrococcus furiosus. J

Bacteriol 193, 6498-6504

60. Schut, G. J., Bridger, S. L. l., and Adams, M. W. W. (2007) Insights into the metabolism

of elemental sulfur by the hyperthermophilic archaeon Pyrococcus furiosus:

characterization of a coenzyme A- dependent NAD(P)H sulfur oxidoreductase. J Bacteriol

189, 4431-4441

61. Basen, M., Sun, J., and Adams, M. W. (2012) Engineering a hyperthermophilic archaeon

for temperature-dependent product formation. mBio 3, e00053-00012

204

62. Schut, G. J., Lipscomb, G. L., Nguyen, D. M., Kelly, R. M., and Adams, M. W. (2016)

Heterologous production of an energy-conserving carbon monoxide dehydrogenase

complex in the hyperthermophile Pyrococcus furiosus. Front Microbiol 7, 29

63. Hawkins, A. B., Lian, H., Zeldes, B. M., Loder, A. J., Lipscomb, G. L., Schut, G. J., Keller,

M. W., Adams, M. W., and Kelly, R. M. (2015) Bioprocessing analysis of Pyrococcus

furiosus strains engineered for CO(2)-based 3-hydroxypropionate production. Biotechnol

Bioeng 112, 1533-1543

64. Lian, H., Zeldes, B. M., Lipscomb, G. L., Hawkins, A. B., Han, Y., Loder, A. J.,

Nishiyama, D., Adams, M. W., and Kelly, R. M. (2016) Ancillary contributions of

heterologous biotin protein ligase and carbonic anhydrase for CO2 incorporation into 3-

hydroxypropionate by metabolically engineered Pyrococcus furiosus. Biotechnol Bioeng

113, 2652-2660

65. Thorgersen, M. P., Lipscomb, G. L., Schut, G. J., Kelly, R. M., and Adams, M. W. (2014)

Deletion of acetyl-CoA synthetases I and II increases production of 3-hydroxypropionate

by the metabolically-engineered hyperthermophile Pyrococcus furiosus. Metab Eng 22,

83-88

66. Werpy, T., and Petersen, G. (2004) Top value added chemicals from biomass: Volume I -

- Results of screening for potential candidates from sugars and synthesis gas. in Other

Information: PBD: 1 Aug 2004

67. Xiao, Z., and Lu, J. R. (2014) Strategies for enhancing fermentative production of acetoin:

a review. Biotechnology advances 32, 492-503

68. Snoep, J. L., de Graef, M. R., Teixeira de Mattos, M. J., and Neijssel, O. M. (1992)

Pyruvate catabolism during transient state conditions in chemostat cultures of

205

Enterococcus faecalis NCTC 775: importance of internal pyruvate concentrations and

NADH/NAD+ ratios. Journal of general microbiology 138, 2015-2020

69. Liu, D., Chen, Y., Ding, F., Guo, T., Xie, J., Zhuang, W., Niu, H., Shi, X., Zhu, C., and

Ying, H. (2015) Simultaneous production of butanol and acetoin by metabolically

engineered Clostridium acetobutylicum. Metab Eng 27, 107-114

70. Yang, T., Rao, Z., Zhang, X., Xu, M., Xu, Z., and Yang, S. T. (2017) Metabolic engineering

strategies for acetoin and 2,3-butanediol production: advances and prospects. Crit Rev

Biotechnol 37, 990-1005

71. Mayer, D., Schlensog, V., and Bock, A. (1995) Identification of the transcriptional

activator controlling the butanediol fermentation pathway in Klebsiella terrigena. J

Bacteriol 177, 5261-5269

72. Bae, S. J., Kim, S., and Hahn, J. S. (2016) Efficient production of acetoin in Saccharomyces

cerevisiae by disruption of 2,3-butanediol dehydrogenase and expression of NADH

oxidase. Sci Rep 6, 27667

73. Xiao, Z., Wang, X., Huang, Y., Huo, F., Zhu, X., Xi, L., and Lu, J. R. (2012) Thermophilic

fermentation of acetoin and 2,3-butanediol by a novel Geobacillus strain. Biotechnol

Biofuels 5, 88

74. Xiao, Z., Gu, R., Hou, X., Zhao, J., Zhu, H., and Lu, J. R. (2017) Non‐sterilized

fermentative production of acetoin with 2,3‐butanediol as a main byproduct from maize

hydrolysate by a newly isolated thermophilic Baccilus strain. J. Chem. Technol. Biotechnol

92, 2845-2852

75. Nguyen, D. M. N., Lipscomb, G. L., Schut, G. J., Vaccaro, B. J., Basen, M., Kelly, R. M.,

and Adams, M. W. W. (2016) Temperature-dependent acetoin production by Pyrococcus

206

furiosus is catalyzed by a biosynthetic acetolactate synthase and its deletion improves

ethanol production. Metab Eng 34, 71-79

76. Mitchell, P. (1975) The protonmotive Q cycle: a general formulation. FEBS letters 59, 137-

139

77. Buckel, W., and Thauer, R. K. (2018) Flavin-based electron bifurcation, a new mechanism

of biological energy coupling. Chem Rev 118, 3862-3886

78. Herrmann, G., Jayamani, E., Mai, G., and Buckel, W. (2008) Energy conservation via

electron-transferring flavoprotein in anaerobic bacteria. J Bacteriol 190, 784-791

79. Decker, K., and Pfitzer, S. (1972) Determination of steady-state concentrations of adenine

nucleotides in growing C. kluyveri cells by biosynthetic labeling. Anal Biochem 50, 529-

539

80. Buckel, W., and Thauer, R. K. (2018) Flavin-based electron bifurcation, ferredoxin,

flavodoxin, and anaerobic respiration with protons (Ech) or NAD(+) (Rnf) as electron

acceptors: a historical review. Front Microbiol 9, 401

81. Weghoff, M. C., Bertsch, J., and Muller, V. (2015) A novel mode of lactate metabolism in

strictly anaerobic bacteria. Environ Microbiol 17, 670-677

82. Bertsch, J., Parthasarathy, A., Buckel, W., and Muller, V. (2013) An electron-bifurcating

caffeyl-CoA reductase. J Biol Chem 288, 11304-11311

83. Kaster, A. K., Moll, J., Parey, K., and Thauer, R. K. (2011) Coupling of ferredoxin and

heterodisulfide reduction via electron bifurcation in hydrogenotrophic methanogenic

archaea. Proc Natl Acad Sci U S A 108, 2981-2986

84. Ledbetter, R. N., Garcia Costas, A. M., Lubner, C. E., Mulder, D. W., Tokmina-

Lukaszewska, M., Artz, J. H., Patterson, A., Magnuson, T. S., Jay, Z. J., Duan, H. D.,

207

Miller, J., Plunkett, M. H., Hoben, J. P., Barney, B. M., Carlson, R. P., Miller, A. F.,

Bothner, B., King, P. W., Peters, J. W., and Seefeldt, L. C. (2017) The electron eifurcating

FixABCX protein complex from Azotobacter vinelandii: generation of low-potential

peducing equivalents for nitrogenase catalysis. Biochemistry 56, 4177-4190

85. Demmer, J. K., Pal Chowdhury, N., Selmer, T., Ermler, U., and Buckel, W. (2017) The

semiquinone swing in the bifurcating electron transferring flavoprotein/butyryl-CoA

dehydrogenase complex from Clostridium difficile. Nat Commun 8, 1577

86. Demmer, J. K., Bertsch, J., Oppinger, C., Wohlers, H., Kayastha, K., Demmer, U., Ermler,

U., and Muller, V. (2018) Molecular basis of the flavin-based electron-bifurcating caffeyl-

CoA reductase reaction. FEBS letters 592, 332-342

87. Lubner, C. E., Jennings, D. P., Mulder, D. W., Schut, G. J., Zadvornyy, O. A., Hoben, J.

P., Tokmina-Lukaszewska, M., Berry, L., Nguyen, D. M., Lipscomb, G. L., Bothner, B.,

Jones, A. K., Miller, A. F., King, P. W., Adams, M. W. W., and Peters, J. W. (2017)

Mechanistic insights into energy conservation by flavin-based electron bifurcation. Nature

Chemical Biology, 10.1038/nchembio.2348

88. Wagner, T., Koch, J., Ermler, U., and Shima, S. (2017) Methanogenic heterodisulfide

reductase (HdrABC-MvhAGD) uses two noncubane [4Fe-4S] clusters for reduction.

Science 357, 699-703

89. Jungermann, K., Thauer, R. K., Rupprecht, E., Ohrloff, C., and Decker, K. (1969)

Ferredoxin mediated hydrogen formation from NADPH in a cell-free system of

Clostridium kluyveri. FEBS letters 3, 144-146

208

90. Thauer, R. K., Rupprecht, E., Ohrloff, C., Jungermann, K., and Decker, K. (1971)

Regulation of the reduced nicotinamide adenine dinucleotide phosphate-ferredoxin

reductase system in Clostridium kluyveri. J Biol Chem 246, 954-959

91. Seedorf, H., Fricke, W. F., Veith, B., Bruggemann, H., Liesegang, H., Strittmatter, A.,

Miethke, M., Buckel, W., Hinderberger, J., Li, F., Hagemeier, C., Thauer, R. K., and

Gottschalk, G. (2008) The genome of Clostridium kluyveri, a strict anaerobe with unique

metabolic features. Proc Natl Acad Sci U S A 105, 2128-2133

92. Wang, S., Huang, H., Moll, J., and Thauer, R. K. (2010) NADP+ reduction with reduced

ferredoxin and NADP+ reduction with NADH are coupled via an electron-bifurcating

enzyme complex in Clostridium kluyveri. J Bacteriol 192, 5115-5123

93. Demmer, J. K., Rupprecht, F. A., Eisinger, M. L., Ermler, U., and Langer, J. D. (2016)

Ligand binding and conformational dynamics in a flavin-based electron-bifurcating

enzyme complex revealed by Hydrogen-Deuterium Exchange Mass Spectrometry. FEBS

letters 590, 4472-4479

94. Berry, L., Poudel, S., Tokmina-Lukaszewska, M., Colman, D. R., Nguyen, D. M. N., Schut,

G. J., Adams, M. W. W., Peters, J. W., Boyd, E. S., and Bothner, B. (2018) H/D exchange

mass spectrometry and statistical coupling analysis reveal a role for allostery in a

ferredoxin-dependent bifurcating transhydrogenase catalytic cycle. Biochim Biophys Acta

1862, 9-17

95. Bennett, B. D., Kimball, E. H., Gao, M., Osterhout, R., Van Dien, S. J., and Rabinowitz, J.

D. (2009) Absolute metabolite concentrations and implied enzyme active site occupancy

in Escherichia coli. Nat Chem Biol 5, 593-599

209

96. Huang, H., Wang, S., Moll, J., and Thauer, R. K. (2012) Electron bifurcation involved in

the energy metabolism of the acetogenic bacterium Moorella thermoacetica growing on

glucose or H2 plus CO2. J Bacteriol 194, 3689-3699

97. Ma, K., and Adams, M. W. W. (1994) Sulfide dehydrogenase from the hyperthermophilic

archaeon Pyrococcus furiosus: a new multifunctional enzyme involved in the reduction of

elemental sulfur. J Bacteriol 179, 6509-6517

98. Forterre, P., Confalonieri, F., Charbonnier, F., and Duguet, M. (1995) Speculations on the

origin of life and thermophily: review of available information on reverse gyrase suggests

that hyperthermophilic procaryotes are not so primitive. Orig Life Evol Biosph 25, 235-249

99. Stetter, K. O. (1996) Hyperthermophilic procaryotes. FEMS Microbiol Rev 18, 149-158

100. Mukund, S., and Adams, M. W. W. (1991) The novel tungsten-iron-sulfur protein of the

hyperthermophilic archaebacterium, Pyrococcus furiosus, is an aldehyde ferredoxin

oxidoreductase. Evidence for its participation in a unique glycolytic pathway. J Biol Chem

266, 14208-14216

101. Robb, F. T., Park, J. B., and Adams, M. W. W. (1992) Characterization of an extremely

thermostable glutamate dehydrogenase: a key enzyme in the primary metabolism of the

hyperthermophilic archaebacterium, Pyrococcus furiosus. Biochim Biophys Acta 1120,

267-272

102. Peters, J. W., Beratan, D. N., Schut, G. J., and Adams, M. W. W. (2018) On the nature of

organic and inorganic centers that bifurcate electrons, coupling exergonic and endergonic

oxidation-reduction reactions. Chem Commun (Camb) 54, 4091-4099

103. Thauer, R. K., Jungermann, K., and Decker, K. (1977) Energy conservation in

chemotrophic anaerobic bacteria. Bacteriol Rev 41, 100-180

210

104. Sato, K., Nishina, Y., Setoyama, C., Miura, R., and Shiga, K. (1999) Unusually high

standard redox potential of acrylyl-CoA/propionyl-CoA couple among enoyl-CoA/acyl-

CoA couples: a reason for the distinct metabolic pathway of propionyl-CoA from longer

acyl-CoAs. J Biochem 126, 668-675

105. Walsh, C. (1986) Naturally occurring 5-deazaflavin coenzymes: biological redox roles.

Accounts of Chemical Research 19, 216-221

106. Tietze, M., Beuchle, A., Lamla, I., Orth, N., Dehler, M., Greiner, G., and Beifuss, U. (2003)

Redox potentials of methanophenazine and CoB-S-S-CoM, factors involved in electron

transport in Methanogenic archaea. Chembiochem 4, 333-335

107. Hans, M., Bill, E., Cirpus, I., Pierik, A. J., Hetzel, M., Alber, D., and Buckel, W. (2002)

Adenosine triphosphate-induced electron transfer in 2-hydroxyglutaryl-CoA dehydratase

from Acidaminococcus fermentans. Biochemistry 41, 5873-5882

108. Li, B., and Elliott, S. J. (2016) The Catalytic Bias of 2-Oxoacid:ferredoxin Oxidoreductase

in CO2: evolution and reduction through a ferredoxin-mediated electrocatalytic assay.

Electrochimica Acta 199, 349-356

109. Snoep, J. L., Teixeira de Mattos, M. J., Starrenburg, M. J., and Hugenholtz, J. (1992)

Isolation, characterization, and physiological role of the pyruvate dehydrogenase complex

and alpha-acetolactate synthase of Lactococcus lactis subsp. lactis bv. diacetylactis. J

Bacteriol 174, 4838-4841

110. Ji, X. J., Huang, H., and Ouyang, P. K. (2011) Microbial 2,3-butanediol production: a state-

of-the-art review. Biotechnology advances 29, 351-364

111. Lipscomb, G. L., Stirrett, K., Schut, G. J., Yang, F., Jenney, F. E., Jr., Scott, R. A., Adams,

M. W. W., and Westpheling, J. (2011) Natural competence in the hyperthermophilic

211

archaeon Pyrococcus furiosus facilitates genetic manipulation: construction of markerless

deletions of genes encoding the two cytoplasmic hydrogenases. Appl Environ Microbiol

77, 2232-2238

112. Weinberg, M. V., Schut, G. J., Brehm, S., Datta, S., and Adams, M. W. W. (2005) Cold

shock of a hyperthermophilic archaeon: Pyrococcus furiosus exhibits multiple responses

to a suboptimal growth temperature with a key role for membrane-bound glycoproteins. J

Bacteriol 187, 336-348

113. Horton, R. M., Hunt, H. D., Ho, S. N., Pullen, J. K., and Pease, L. R. (1989) Engineering

hybrid genes without the use of restriction enzymes: gene splicing by overlap extension.

Gene 77, 61-68

114. Merryman, C., and Gibson, D. G. (2012) Methods and applications for assembling large

DNA constructs. Metab Eng 14, 196-204

115. Farkas, J., Stirrett, K., Lipscomb, G. L., Nixon, W., Scott, R. A., Adams, M. W., and

Westpheling, J. (2012) Recombinogenic properties of Pyrococcus furiosus strain COM1

enable rapid selection of targeted mutants. Appl Environ Microbiol 78, 4669-4676

116. Thorgersen, M. P., Stirrett, K., Scott, R. A., and Adams, M. W. W. (2012) Mechanism of

oxygen detoxification by the surprisingly oxygen-tolerant hyperthermophilic archaeon,

Pyrococcus furiosus. Proc Natl Acad Sci U S A 109, 18547-18552

117. Xing, R. Y., and Whitman, W. B. (1987) Sulfometuron methyl-sensitive and -resistant

acetolactate synthases of the archaebacteria Methanococcus spp. J Bacteriol 169, 4486-

4492

212

118. Xing, R. Y., and Whitman, W. B. (1991) Characterization of enzymes of the branched-

chain amino acid biosynthetic pathway in Methanococcus spp. Journal of bacteriology

173, 2086-2092

119. Xiao, Z., and Xu, P. (2007) Acetoin metabolism in bacteria. Critical reviews in

microbiology 33, 127-140

120. Payne, K. A., Hough, D. W., and Danson, M. J. (2010) Discovery of a putative acetoin

dehydrogenase complex in the hyperthermophilic archaeon Sulfolobus solfataricus. FEBS

letters 584, 1231-1234

121. Scott, J. W., Poole, F. L., and Adams, M. W. W. (2014) Characterization of ten

heterotetrameric NDP-dependent acyl-CoA synthetases of the hyperthermophilic archaeon

Pyrococcus furiosus. Archaea (Vancouver, B.C.) 2014, 176863

122. Glasemacher, J., Bock, A. K., Schmid, R., and Schonheit, P. (1997) Purification and

properties of acetyl-CoA synthetase (ADP-forming), an archaeal enzyme of acetate

formation and ATP synthesis, from the hyperthermophile Pyrococcus furiosus. European

journal of biochemistry / FEBS 244, 561-567

123. Yoon, S. H., Reiss, D. J., Bare, J. C., Tenenbaum, D., Pan, M., Slagel, J., Moritz, R. L.,

Lim, S., Hackett, M., Menon, A. L., Adams, M. W., Barnebey, A., Yannone, S. M., Leigh,

J. A., and Baliga, N. S. (2011) Parallel evolution of transcriptome architecture during

genome reorganization. Genome research 21, 1892-1904

124. Lo, J., Olson, D. G., Murphy, S. J., Tian, L., Hon, S., Lanahan, A., Guss, A. M., and Lynd,

L. R. (2017) Engineering electron metabolism to increase ethanol production in

Clostridium thermocellum. Metab Eng 39, 71-79

213

125. Lo, J., Zheng, T., Olson, D. G., Ruppertsberger, N., Tripathi, S. A., Tian, L., Guss, A. M.,

and Lynd, L. R. (2015) Deletion of nfnAB in Thermoanaerobacterium saccharolyticum

and Its Effect on Metabolism. J Bacteriol. 197, 2920-2929

126. Li, W., and Godzik, A. (2006) Cd-hit: a fast program for clustering and comparing large

sets of protein or nucleotide sequences. Bioinformatics 22, 1658-1659

127. Smoot, M. E., Ono, K., Ruscheinski, J., Wang, P. L., and Ideker, T. (2011) Cytoscape 2.8:

new features for data integration and network visualization. Bioinformatics 27, 431-432

128. Thompson, J. D., Gibson, T. J., and Higgins, D. G. (2002) Multiple sequence alignment

using ClustalW and ClustalX. Curr Protoc Bioinformatics Chapter 2:Unit 2.3,

10.1002/0471250953.bi0471250203s0471250900

129. Guindon, S., Dufayard, J. F., Lefort, V., Anisimova, M., Hordijk, W., and Gascuel, O.

(2010) New algorithms and methods to estimate maximum-likelihood phylogenies:

assessing the performance of PhyML 3.0. Syst Biol 59, 307-321

130. Letunic, I., and Bork, P. (2007) Interactive Tree Of Life (iTOL): an online tool for

phylogenetic tree display and annotation. Bioinformatics 23, 127-128

131. Lipscomb, G. L., Stirrett, K., Schut, G. J., Yang, F., Jenney, F. E. J., Scott, R. A., Adams,

M. W. W., and Westpheling, J. (2011) Natural competence in the hyperthermophilic

archaeon Pyrococcus furiosus facilitates genetic manipulation: construction of markerless

deletions of genes encoding the two cytoplasmic hydrogenases. Appl Environ Microbiol

77, 2232-2238

132. Farkas, J., Stirrett, K. F., Lipscomb, G. L., Nixon, W., Scott, R. A., Adams, M. W. W., and

Westpheling, J. (2012) Recombinogenic properties of Pyrococcus furiosus strain COM1

enable rapid selection of targeted mutants. Appl Environ Microbiol 78, 4669-4676

214

133. Keller, M. W., Schut, G. J., Lipscomb, G. L., Menon, A. L., Iwuchukwu, I. J., Leuko, T.

T., Thorgersen, M. P., Nixon, W. J., Hawkins, A. S., Kelly, R. M., and Adams, M. W. W.

(2013) Exploiting microbial hyperthermophilicity to produce an industrial chemical, using

hydrogen and carbon dioxide. Proc Natl Acad Sci U S A 110, 5840-5845

134. Sporty, J. L., Kabir, M. M., Turteltaub, K. W., Ognibene, T., Lin, S. J., and Bench, G.

(2008) Single sample extraction protocol for the quantification of NAD and NADH redox

states in Saccharomyces cerevisiae. J Sep Sci 32, 3202-3211

135. Yang, Z. Y., Ledbetter, R., Shaw, S., Pence, N., Tokmina-Lukaszewska, M., Eilers, B.,

Guo, Q., Pokhrel, N., Cash, V. L., Dean, D. R., Antony, E., Bothner, B., Peters, J. W., and

Seefeldt, L. C. (2016) Evidence that the Pi release event is the rate-limiting step in the

nitrogenase catalytic cycle. Biochemistry 55, 3625-3635

136. Vaudel, M., Burkhart, J. M., Zahedi, R. P., Oveland, E., Berven, F. S., Sickmann, A.,

Martens, L., and Barsnes, H. (2015) PeptideShaker enables reanalysis of MS-derived

proteomics data sets. Nat Biotechnol 33, 22-24

137. Poudel, S., Tokmina-Lukaszewska, M., Colman, D., Refai, M., Schut, G., King, P.,

Maness, P.-C., Adams, M. W. W., Peters, J., Bothner, B., and Boyd, E. (2016) Unification

of [FeFe]-hydrogenases into three structural and functional groups. Biochim Biophys Acta

1860, 1910-1921

138. Luo, M. L., Jackson, R., Denny, S., Tokmina-Lukaszewska, M., Maksimchuk, K., Lin, W.,

Bothner, B., Wiedenheft, B., and Beisel, C. (2016) The CRISPR RNA-guided surveillance

complex in Escherichia coli accommodates extended RNA spacers. Nucleic Acids Res 44,

7385-7394

215

139. Pettersen, E. F., Goddard, T. D., Huang, C. C., Couch, G. S., Greenblatt, D. M., Meng, E.

C., and Ferrin, T. E. (2004) UCSF Chimera--a visualization system for exploratory research

and analysis. J Comput Chem 25, 1605-1612

140. Otwinowski, Z., and Minor, W. (1997) Processing of X-ray diffraction data collected in

oscillation mode. Methods Enzymol 276, 307-326

141. McCoy, A. J., Grosse-Kunstleve, R. W., D., A. P., Winn, M. D., and Read, R. J. (2007)

Phaser crystallographic software. J Appl Crystallogr 40, 658-674

142. Adams, P. D., Afonine, P. V., Bunkoczi, G., Chen, V. B., Davis, I. W., Echols, N., Headd,

J. J., Hung, L. W., Kapral, G. J., Grosse-Kunstleve, R. W., McCoy, A. J., Moriarty, N. W.,

Oeffner, R., Read, R. J., Richardson, D. C., Richardson, J. S., Terwilliger, T. C., and Zwart,

P. H. (2010) PHENIX: a comprehensive Python-based system for macromolecular

structure solution. Acta Crystallogr D Biol Crystallogr 66, 213-221

143. Emsley, P., and Cowtan, K. (2004) Coot: model-building tools for molecular graphics. Acta

Crystallogr D Biol Crystallogr. 60, 2126-2132

144. DeLano, W. L. (2002) The PyMOL Molecular Graphics System. Schrödinger, LLC

145. Krissinel, E., and Henrick, K. (2004) Secondary-structure matching (SSM), a new tool for

fast protein structure alignment in three dimensions. Acta Crystallogr D Biol Crystallogr

60, 2256-2268

146. Aono, S., Bryant, F. O., and Adams, M. W. W. (1989) A novel and remarkably

thermostable ferredoxin from the hyperthermophilic archaebacterium Pyrococcus furiosus.

J Bacteriol 171, 3433-3439

216

147. Kanai, T., Ito, S., and Imanaka, T. (2003) Characterization of a cytosolic NiFe-hydrogenase

from the hyperthermophilic archaeon Thermococcus kodakaraensis KOD1. J Bacteriol

185, 1705-1711

148. Nixon, J. E., Wang, A., Field, J., Morrison, H. G., McArthur, A. G., Sogin, M. L., Loftus,

B. J., and Samuelson, J. (2002) Evidence for lateral transfer of genes encoding ferredoxins,

nitroreductases, NADH oxidase, and alcohol dehydrogenase 3 from anaerobic prokaryotes

to Giardia lamblia and Entamoeba histolytica. Eukaryot Cell 1, 181-190

149. Bastian, S., X., L., Meyerowitz, J. T., Snow, C. D., Chen, M., M., and Arnold, F. H. (2011)

Engineered ketol-acid reductoisomerase and alcohol dehydrogenase enable anaerobic 2-

methylpropan-1-ol production at theoretical yield in Escherichia coli. Metab Eng 13, 345-

353

150. Maddocks, O. D., Labuschagne, C. F., and Vousden, K. H. (2014) Localization of NADPH

production: a wheel within a wheel. Mol Cell Biochem 55, 158-160

151. Kikuchi, G. (1973) The glycine cleavage system: composition, reaction mechanism, and

physiological significance. Mol Cell Biochem 1, 169-187

152. Berger, F., Ramirez-Hernandez, H. H., and Ziegler, M. (2004) The new life of a

centenarian: signalling functions of NAD(P). Trends Biochem Sci 29, 111-118

153. Hoben, J. P., Lubner, C. E., Ratzloff, M. W., Schut, G. J., Nguyen, D. M. N., Hempel, K.

W., Adams, M. W. W., King, P. W., and Miller, A. F. (2017) Equilibrium and ultrafast

kinetic studies manipulating electron transfer: A short-lived flavin semiquinone is not

sufficient for electron bifurcation. J Biol Chem 292, 14039-14049

217

154. Zhang, P., Yuly, J. L., Lubner, C. E., Mulder, D. W., King, P. W., Peters, J. W., and

Beratan, D. N. (2017) Electron bifurcation: thermodynamics and kinetics of two-electron

brokering in biological redox chemistry. Acc Chem Res 50, 2410-2417

155. Nitschke, W., and Russell, M. J. (2012) Redox bifurcations: mechanisms and importance

to life now, and at its origin: a widespread means of energy conversion in biology unfolds.

Bioessays 34, 106-109

156. Schoepp-Cothenet, B., van Lis, R., Atteia, A., Baymann, F., Capowiez, L., Ducluzeau, A.

L., Duval, S., ten Brink, F., Russell, M. J., and Nitschke, W. (2013) On the universal core

of bioenergetics. Biochim Biophys Acta 1827, 79-93

157. Anderson, R. F. (1983) Energetics of the one-electron reduction steps of riboflavin, FMN

and FAD to their fully reduced forms. Biochim Biophys Acta 722, 158-162

158. Page, C. C., Moser, C. C., and Dutton, P. L. (2003) Mechanism for electron transfer within

and between proteins. Curr Opin Chem Biol 7, 551-556

159. Nguyen, D. M. N., Schut, G. J., Zadvornyy, O. A., Tokmina-Lukaszewska, M., Poudel, S.,

Lipscomb, G. L., Adams, L. A., Dinsmore, J. T., Nixon, W. J., Boyd, E. S., Bothner, B.,

Peters, J. W., and Adams, M. W. W. (2017) Two functionally distinct NADP(+)-dependent

ferredoxin oxidoreductases maintain the primary redox balance of Pyrococcus furiosus. J

Biol Chem 292, 14603-14616

160. Mukund, S., and Adams, M. W. W. (1990) Characterization of a tungsten-iron-sulfur

protein exhibiting novel spectroscopic and redox properties from the hyperthermophilic

archaebacterium Pyrococcus furiosus. J Biol Chem 265, 11508-11516

218

161. Tremblay, P. L., Zhang, T., Dar, S. A., Leang, C., and Lovley, D. R. (2012) The Rnf

complex of Clostridium ljungdahlii is a proton-translocating ferredoxin:NAD+

oxidoreductase essential for autotrophic growth. mBio 4, e00406-00412

162. Shaw, A. J., Podkaminer, K. K., Desai, S. G., Bardsley, J. S., Rogers, S. R., Thorne, P. G.,

Hogsett, D. A., and Lynd, L. R. (2008) Metabolic engineering of a thermophilic bacterium

to produce ethanol at high yield. Proc Natl Acad Sci U S A 105, 13769-13774

163. Fontecave, M., and Artero, V. (2011) Bioinspired catalysis at the crossroads between

biology and chemistry: A remarkable example of an electrocatalytic material mimicking

hydrogenases. Comptes Rendus Chimie 14, 362-371

164. Schut, G. J., and Adams, M. W. (2009) The iron-hydrogenase of Thermotoga maritima

utilizes ferredoxin and NADH synergistically: a new perspective on anaerobic hydrogen

production. J Bacteriol 191, 4451-4457

219

APPENDIX A

SINGLE GENE INSERTION DRIVES BIOALCOHOL PRODUCTION BY A

THERMOPHILIC ARCHAEON

Basen, M.†, Schut, G. J†., Nguyen, D. M., Lipscomb, G. L., Benn, R. A., Prybol, C. J., Vaccaro, B. J., Poole, F. L., Kelly, R. M., and Adams, M. W. (2014) Single gene insertion drives bioalcohol production by a thermophilic archaeon. Proc Natl Acad Sci U S A 111, 17618-17623 †Authors contributed equally to this work Reprinted here with permission of the publisher

220

ABSTRACT

Bioethanol production is achieved by only two metabolic pathways and only at moderate temperatures. Herein a fundamentally different synthetic pathway for bioalcohol production at 70

°C was constructed by insertion of the gene for bacterial alcohol dehydrogenase (AdhA) into the archaeon Pyrococcus furiosus. The engineered strain converted glucose to ethanol via acetate and acetaldehyde, catalyzed by the host-encoded aldehyde ferredoxin oxidoreductase (AOR) and heterologously expressed AdhA, in an energy-conserving, redox-balanced pathway. Furthermore, the AOR/AdhA pathway also converted exogenously added aliphatic and aromatic carboxylic acids to the corresponding alcohol using glucose, pyruvate, and/or hydrogen as the source of reductant. By heterologous coexpression of a membrane-bound carbon monoxide dehydrogenase,

CO was used as a reductant for converting carboxylic acids to alcohols. Redirecting the fermentative metabolism of P. furiosus through strategic insertion of foreign genes creates unprecedented opportunities for thermophilic bioalcohol production. Moreover, the AOR/AdhA pathway is a potentially game-changing strategy for syngas fermentation, especially in combination with carbon chain elongation pathways.

SIGNIFICANCE

The microbial production of ethanol (bioethanol) is a massive commercialized technology.

While alcohols with longer carbon chains are chemically much better suited for current transportation needs, their biotechnological production remains challenging. Here we have engineered the model hyperthermophile Pyrococcus furiosus to produce various alcohols from their corresponding organic acids by constructing a novel synthetic route termed the AOR/AdhA pathway. Our study is also the first example of significant alcohol formation in an archaeon,

221

emphasizing the biotechnological potential of novel microorganisms. Moreover, we show that carbon monoxide and hydrogen (syngas) can be used as the driving forces for alcohol production.

The application of the AOR/AdhA pathway in syngas-fermenting microorganisms is potentially a game-changing platform technology for the production of longer bioalcohols.

222

A1. Introduction

Production of alcohol-based biofuels from renewable feedstocks is currently achieved by only a very limited number of metabolic pathways (1, 2). The US bioethanol industry depends upon glucose conversion by yeast wherein pyruvate (C3) is decarboxylated to acetaldehyde and then reduced to ethanol (C2) by a monofunctional alcohol dehydrogenase. The other major pathway is found in some anaerobic bacteria, wherein glucose-derived pyruvate is oxidized to acetyl-coenzyme A, and this is further reduced to ethanol by a bifunctional alcohol dehydrogenase

(AdhE) (3, 4). Recently, there has been increasing interest in microorganisms that produce longer chain alcohols (>C2), which have superior characteristics as fuel molecules compared to ethanol, to replace fossil fuels (1, 2). In this case glucose conversion requires microbial strains engineered to produce one specific alcohol at a time. For example, the acetone-butanol-ethanol fermentation pathway, found in some Clostridia, has been adapted in yeast, Escherichia coli and a few other bacteria (2, 5) to produce isopropanol or n-butanol. Similarly, n-propanol, 2-methyl-1-butanol, 3- methyl-1-butanol and 1-butanol are side products of amino acid fermentation by yeast (2) and modified pathways have been expressed in E. coli to produce a specific alcohol (6). In addition, isopentanol can be produced by a variation of the isoprenoid biosynthesis pathway in engineered

E. coli (2).

Production of bioalcohols at temperatures above 70°C has several advantages over ambient-temperature processes, including lower risk of microbial contamination, higher diffusion rates and lower cooling and distillation costs (7). However, very few microorganisms that are able to grow at such temperatures are able to generate ethanol from sugar (8-10), and no bacterium growing above 70°C produces an alcohol other than ethanol. In addition, no member of the domain

Archaea is known to produce any alcohol as a major product, regardless of growth temperature.

223

Herein, we describe the metabolic engineering of an archaeon to produce not only ethanol but a range of alcohols at 70−80°C via a synthetic pathway not known in Nature and fundamentally different from those previously known.

The archaeon Pyrococcus furiosus grows optimally near 100°C (11) by fermenting simple and complex sugars to acetate, carbon dioxide and hydrogen gas (12). P. furiosus has an unusual

Emden-Meyerhof pathway for the conversion of glucose to pyruvate since reductant is channeled not to NADH but to the redox protein ferredoxin (Fd, Figure A1) by glyceraldehyde-3-phosphate

(GAP) Fd oxidoreductase (GAPOR). Reduced Fd is re-oxidized by a membrane-bound, energy- conserving H2-evolving hydrogenase (12). Pyruvate produced by glycolysis is subsequently oxidized to acetyl-CoA by pyruvate Fd oxidoreductase (POR), and acetyl-CoA is converted by

ATP-forming acetyl-CoA synthetase (ACS) to acetate. P. furiosus was recently metabolically- engineered to generate end products other than acetate in a temperature-controlled manner without the need for chemical inducers. Lactate was produced from glucose and 3-hydroxypropionate was produced from carbon dioxide and glucose using heterologously-expressed enzymes encoded by foreign genes obtained from microbes that grow near 75°C (13, 14). At 98°C, the foreign enzymes were inactive and the engineered P. furiosus strains generated acetate, but near 70°C, the engineered strains produced either lactate or 3-hydroxypropionate instead.

The primary goal here was to engineer P. furiosus to produce ethanol near 70°C, applying a similar approach. We found that unexpectedly the insertion of a primary alcohol dehydrogenase,

AdhA, lead to the production of not only ethanol but also of a variety of bio-alcohols from their corresponding organic acids. We used gene deletion analysis and 13C-labeling to elucidate the biochemical pathway and hypothesize it might be a remnant of an ancient energy-conservation mechanism. Furthermore, a P. furiosus strain A/Codh was developed that expresses a multi-

224

subunit carbon monoxide dehydrogenase. That strain used carbon monoxide as the electron donor for organic acid reduction to bio-alcohols, emphasizing the biotechnological versatility and potential of the new synthetic pathway.

225

A2. Materials and Methods

Transformation of P. furiosus.

Escherichia coli XL1 Blue-MRF´ (Agilent Technologies, Santa Clara, CA) was used to amplify plasmid DNA. Plasmid DNA purification was performed using the StrataPrep Plasmid

Miniprep Kit (Agilent). Extraction of DNA from P. furiosus, transformation of P. furiosus, screening of transformants and strain purification were performed as previously described, except that the defined medium contained maltose (5 g L−1) instead of cellobiose as sole growth substrate.

The DNA sequence modification of isolated P. furiosus strains was verified by sequencing as previously described (17). Primers used to construct the strains, all plasmids and all strains are listed in the SI Appendix (Figure A.S1 and Tables A.S1 and A.S2).

Construction of adhA-containing strains A, E, EA and A/Codh.

P. furiosus strain COM117 served as the parent strain for genetic manipulations for the heterologous expression of the bifunctional aldehyde/alcohol dehydrogenase AdhE (Teth514

0627; GeneID: 5876124) and the primary alcohol dehydrogenase AdhA (Teth514 0564; GeneID:

5877753) from Thermoanaerobacter strain X514 (16). Genomic DNA was isolated according to

Zhou et al. (32). adhE was amplified from genomic DNA by PCR using the primer pairs AdhE-

Pslp-F/AdhE-R2 (for construction of plasmid pMB303SLP) or AdhE-Pslp-F/AdhE-SphI-R (for construction of plasmid pMB304SLP). AdhA (for construction of plasmid pMB303SLP) was amplified using AdhA-F2/AdhA-SphI-R. The constitutive promoter Pslpwas amplified from genomic DNA of P. furiosus with the primer set Pslp-SacII-F / Pslp-adhE-R. Fusion products of

Pslp and adhE or adhE and adhA were obtained by overlap PCR. Products from overlap PCR were

226

digested with the restriction enzymes SacII and SphI and ligated into plasmid vector pSPF300 as described previously (13) to make plasmids pMB303SLP (containing adhE and adhA under control of Pslp) and pMB304SLP (containing only adhE under control of Pslp). pMB303SLP contained a ribosomal binding site of the cold induced protein CipA (PcipA RBS, 16 bases) between adhE and adhA. pMB303SLP and pMB304SLP were used for transformation of P. furiosus strain

ΔpdaD.

For transformation of P. furiosus strain COM1, the Pslp-adhE-adhA or Pslp-adhE fusions were amplified from pMB303SLP and pMB304SLP using the primer pairs Pslp-SphI-F/AdhA-AscI-R or Pslp-SphI-F/AdhE-AscI-R, respectively; additionally introducing the hpyA1 terminator T hpyA1 (17). The resulting PCR products were digested with AscI and SphI, and then ligated into plasmid pGL007 (14) to make plasmids pMB403SLP and pMB404SLP. Plasmid pMB407SLP for construction of strain A (Figure A.2A) is derived from plasmid pMB403SLP. Using the primers AdhA-Pslp-F/SP2.055 everything but the adhE gene was amplified from plasmid pMB403SLP, and the PCR product was assembled to yield plasmid pMB407SLP using a Gibson

Assembly Master Mix (NEB, Ispwich, MA, USA). All plasmids were digested with the restriction enzyme NdeI, and the resulting linear DNA was used to transform strain COM1 to yield strains

EA, E and A (Figure A.2A).

A linear DNA construct was used for a knockout of the aor gene (PF0346) in strain A to make strain AΔaor. First, the marker PgdhpyrF was removed from strain A by selection on 5-fluoroacetic acid (17) to yield strain MW610. Then, primer pairs AOR1/AOR2 and AOR3/AOR4 were used to amplify 500 bp regions upstream and downstream of PF0346. AOR5/SP2.037 and SP2.088/AOR6 were used to amplify 800 bp the marker PgdhpyrF from pGL007 (14). The PCR products were combined by overlap PCR; and the resulting DNA fragment containing the marker PgdhpyrF

227

flanked by 500 bp regions upstream and downstream of PF0346 was used to transform strain A.

The deletion was verified by PCR and by sequence analysis.

The pGL058 plasmid containing the T. onnurineus Codh was constructed via Gibson

Assembly (NEB, Ispwich, MA, USA) of the following fragments: the 8.8 kb backbone BAC vector containing the pyrF genetic marker and flanking homologous recombination regions targeting the intergenic space between convergent genes PF1232-PF1233, amplified from pGL054(33) with primers SP.238 and SP.237; the 200 bp mbh1 (PF1423) promoter region of the membrane-bound hydrogenase gene cluster, amplified from P. furiosus genomic DNA with primers SP.239 and

SP.243; and the 13.3 kb Codh gene cluster (TON_1017-TON_1031), amplified from T. onnurineus genomic DNA using primers SP.244 and SP.245 (see SI Appendix (Figure A.S5)). The pGL058 plasmid was linearized using the unique PvuI restriction site on the BAC vector backbone prior to transformation of P. furiosus strain MW610.

Cultivation of the strains and alcohol production experiments.

Thermoanaerobacter strain X514 was cultivated at 65°C on complex medium used for cultivation of thermophilic heterotrophic anaerobes (modified DSMZ 516 medium) with 5 g L−1 cellobiose as electron donor (34). P. furiosus (DSM 3638) was routinely grown at the indicated temperatures with 5 g L−1 maltose and 2 g L−1 yeast extract as described previously (13). 20 µM uracil (Sigma Chemical, St. Louis, MO, USA) was added as needed (17). In temperature switch experiments, cells were grown at 95°C until mid to late exponential growth phase (0.5−1 x 108 cells), then cooled to 72°C and kept at this temperature for another 20−48 h as described previously(13, 14). Growth was followed by cell counting and by determination of cell-protein concentration.

228

Cell suspension experiments.

P. furiosus strain A and strain A/Codh were grown at 72°C for 4 days to reach high cell density (>1 x 108 cells ml−1), pelleted by centrifugation (6000 × g) for 10 minutes, and then resuspended in media (1/10 of the original culture volume to achieve a 10 x concentration).

Maltose, pyruvate (as the electron donor) and organic acids (as electron acceptors) were added in excess (≥40−100 mM). To test the effect of hydrogen or CO as the electron donor, the argon headspace was replaced by either gas (2 bar).

13C-acetate conversion experiment.

10 mL cultures of P. furiosus strain A were supplied with 8 mM double 13C-labeled sodium

13 −1 acetate (sodium acetate- C2; Sigma-Aldrich, St. Louis, MO, USA) in addition to 5 g L

(unlabeled) maltose, and incubated at 72°C. Samples were taken from the cultures over a 4-day time course to study the change in the carbon isotope signature of acetate and ethanol. Samples

(100 µL) of spent media in 2 mL glass vials were acidified by addition of 10 µL 2 M H2SO4.

Acidified samples were heated on a hot plate until boiling and 1-4 µL samples were removed and analyzed by gas chromatography (GC) - mass spectrometry (MS) to obtain 13C/12C ratios for acetic acid and ethanol. This ratio was taken to be equal to the ratio of the measured abundances for masses 62 and 60 for acetic acid and masses 45 and 47 for ethanol. Significant amounts of mixed compounds (containing both 13C and 12C) were not detected. A helium mobile phase was used at a head pressure of 12 psi on an Alltech Econo-Cap 30m x 0.25 mm EC-WAX column (0.25 µM film) using a Hewlett Packard HP5890A GC with a Hewlett Packard 5971A electron ionization

MS. The temperature was held at 40°C for a 3 min solvent delay and then increased to 220°C at a rate of 15°C/min where it was held for an additional minute. For acetic acid measurements in

229

samples with high amounts of ethanol, the method was modified to begin at 100°C to avoid overloading the MS detector with ethanol while still obtaining sufficient signal for acetic acid.

Mass spectra were collected at M/Z of 40 to 200 at a scan rate of 4 scans per second.

Preparation of cell extracts and enzyme assays.

P. furiosus cells were harvested by centrifugation for 10 min at 6,000 × g. Cells were lysed under anoxic conditions by osmotic shock in 50 mM Tris HCl, pH 8.0 and 2 mM sodium dithionite; and additionally by a short sonication treatment (30 s, max. 36 W). The lysis buffer contained 0.5

µg/ml deoxyribonuclease I (Sigma) to decrease the viscosity of the protein extract. Particles were removed by centrifugation at 30,000 × g for 10 minutes to yield the whole cell extracts (S30). The supernatant of the whole cell extract subjected to ultracentrifugation at 100,000 × g for 1 h yielded the cytoplasmatic protein fraction (S100). The protein content was determined using a standard

Bradford assay. Whole cell extracts and S100 were kept anoxic at all times; and enzyme activity assays were performed under reducing conditions in anoxic 50 mM MOPS, pH 7.5 plus 2 mM

DTT at 70°C. Unless noted otherwise, aldehyde dehydrogenase (ALDH, E.C. 1.2.1.3) was determined by oxidation of NADH (0.2 mM) with acetaldehyde (1 mM) as the substrate, and alcohol dehydrogenase was determined by the oxidation of NADPH (0.2 mM) with butyraldehyde

(1 mM) as the substrate. Absorption of both NADH and NADPH was measured at 340 nm (ԑ =

6.22 M−1 cm−1), and NAD(P)H oxidation activities are given in µmol min−1 mg−1. Aldehyde ferredoxin oxidoreductase was measured by the oxidation of butyraldehyde (1 mM) with benzyl viologen (1 mM) as electron acceptor as described previously (35). Vmax and Km values of AdhA were calculated using non-linear regression (nls function) in R (36). Standard Gibbs free energies

230

ΔG°ˊ were calculated from the free energies of formation ΔGf°, which were taken from Thauer et al. 1977 (37).

Chemical Analyses.

Alcohols and organic acids were measured using an Agilent 7890A GC equipped with a carbowax/20m column and a FID detector. Ethanol and organic acids were also determined using the Megazyme Ethanol assay kit (Megazyme, Wicklow, Ireland) and using a Waters high- performance liquid chromatography model 2690 equipped with an Aminex HPX-87H column (300 mm by 7.8 mm; Bio-Rad) and a photodiode array detector (Model 996, Waters). Hydrogen and

CO were determined on a GC-8A gas chromatograph (Shimadzu) equipped with a thermal conductivity detector and a molecular sieve column (model 5A 80/100, Alltech) with argon as the carrier gas.

231

A3. Results and Discussion

For ethanol production, the foreign genes to be inserted into P. furiosus encoded the bifunctional AdhE and the mono-functional AdhA enzymes, which generate ethanol from acetyl-

CoA and acetaldehyde, respectively (9, 15). The genes were obtained from the thermophilic bacterium Thermoanaerobacter strain X514, which grows near 70°C (16). These genes were inserted, individually and in combination, into the P. furiosus genome (17), yielding strain E

(containing adhE), strain A (containing adhA), and strain EA (containing adhE and adhA; see

Figure A2A and SI Appendix (Figure A.S1 and Table A.S1)). As expected, when grown at 98°C, no AdhA or AdhE activity could be detected in cell extracts of any strain, although both activities were measured when the strains were grown at 72°C (Figure A2B). The activity of AdhA was lower in cell extracts of strain EA compared to those of strain A, possibly due to a lower expression level of adhA as it is the second gene in the synthetic operon inserted into strain EA.

Surprisingly, however, at 72°C, strain E produced very little ethanol, only slightly more than the trace amounts produced by the parent strain (Figure A2C). This might be explained by high activity of the P. furiosus enzyme ATP-forming acetyl-CoA synthase (ACS), which competes with AdhE for the substrate acetyl-CoA (18). Even more unexpected was that strain A generated very high amounts of ethanol (>20 mM), even more than that produced by strain EA ((Figure

A2C) with very little acetate (<2 mM; (Figure A2D).

Since AdhA can generate ethanol only from acetaldehyde, and P. furiosus strain A does not contain bifunctional AdhE activity (Figure A2B), acetyl-CoA is not likely to be the source of this acetaldehyde for ethanol production. Acetaldehyde could arise in P. furiosus from the decarboxylation of pyruvate, which was previously shown to be a significant side reaction of POR

232

(19). Alternatively, it could arise by the reduction of acetate by P. furiosus aldehyde Fd oxidoreductase (AOR). This enzyme is highly expressed in P. furiosus when grown on sugars or peptides, and it has been shown to catalyze the reverse reaction in vitro, the oxidation of various aldehydes to their corresponding acid. It is thought that the in vivo function of AOR is to oxidize toxic aldehydes generated from the 2-ketoacids that are produced during sugar and peptide fermentation. However, this hypothesis has not been experimentally verified (20).

To distinguish between pyruvate or acetate as the source of acetaldehyde, 13C-labeled acetate was added to P. furiosus strain A growing at 72°C on sugar (the disaccharide maltose), and the isotopic composition of the ethanol produced was analyzed. Approximately 50% of the ethanol formed after 40 h incubation contained the 13C label (see SI Appendix (Figure A.S2)). This can only occur if the acetaldehyde for ethanol production was derived from the added labeled acetate, which was subsequently diluted by unlabeled acetate produced from maltose degradation. To prove that AOR was responsible for reducing the acetate to acetaldehyde, the gene encoding AOR

(PF0346) was deleted in strain A. As expected, the new P. furiosus strain A/Δaor, containing

Thermoanaerobacter strain X514 AdhA but lacking the host’s AOR, generated only trace amounts of ethanol from maltose, similar to that of the original parent strain (see SI Appendix (Figure

A.S3)). It is not clear if AOR normally generates any acetaldehyde from acetate in wild type P. furiosus (lacking AdhA) and, if so, how that the acetaldehyde is further metabolized. In any event, the properties of strain A call into question the previously proposed role of AOR (20).

The proposed synthetic pathway for ethanol production in P. furiosus strain A is shown in

Figure A1. Acetate generated from glucose oxidation is reduced by AOR, and the acetaldehyde produced is reduced to ethanol by heterologously-expressed AdhA. As indicated in Figure A1, ethanol production from glucose is redox balanced. Reduced Fd for acetate reduction by AOR is

233

supplied by POR and GAPOR (Figure A1), while the NADPH for ethanol production by AdhA must also be generated from reduced Fd. This could occur either by ferredoxin NAD(P) oxidoreductase or from H2 via the cytoplasmic hydrogenase (SHI) of P. furiosus. In addition, energy is conserved in the form of ATP by the ACS reaction. Consequently, this synthetic pathway theoretically converts 0.5 moles of glucose to one mole of ethanol and one mole of CO2, according to Equation. 1:

0.5 Glucose + ADP + Pi → Ethanol + CO2 + ATP (1)

Since the synthetic pathway converted the added 13C-labeled acetate to ethanol, we investigated whether other exogenously supplied organic acids would similarly be converted to their corresponding alcohol by the AOR-AdhA pathway. This would seem likely since AOR has a very broad substrate specificity; it oxidizes the decarboxylated forms of keto-acids derived from the transaminated derivatives of virtually all twenty amino acids (21). Hence, when 40 mM butyrate was added to a culture of P. furiosus strain A at 72°C, almost 30 mM butanol was generated (Figure A3A; SI Appendix (Figure A.S4)) with the reductant supplied by glucose, according to Equation. 2:

0.5 Glucose + Butyrate + ADP + Pi → Acetate + Butanol + CO2 + ATP (2)

Similar results were obtained when propionate, isobutyrate, valerate, isovalerate, caproate or phenylacetate were added to P. furious strain A, generating propanol, isobutanol, 1-pentanol, isoamylalcohol, 1-hexanol and phenylethanol, respectively (Figure A.3A). When butyrate was added to strain A/Δaor, insignificant amounts of butanol were formed (Figure A.3C), once more demonstrating the essential role of AOR in alcohol formation. P. furiosus strain A must, therefore, metabolize the sugar (maltose) to provide reductant for the conversion of the added acid to the corresponding alcohol (Figure A1, Equation 2). Consequently, one would expect acetate to be

234

also generated as the oxidized end product, and this was the case (Figure A3A). Furthermore, ethanol was produced by the reduction of the acetate generated from sugar oxidation according to

Equation 1 (Figure A3A). As shown in Figure A3B, with butyrate as the added acid, acetate and butanol were produced in a 1:1 ratio (Eqn. 2). As butyrate was provided in great excess (100 mM), only minimal amounts of ethanol were produced. Under growth conditions almost 40 mM (3 g

L−1) butanol was generated at a rate of 0.34 mmol h−1 g protein−1.

P. furiosus can also grow with pyruvate as a carbon and energy source and so pyruvate should be able replace maltose and supply reductant via POR for butyrate reduction, and this should also result in the formation of ATP (Figure A1; Equation 3). As shown in Figure A3D, this proved to be the case. The acetate:butanol ratio is predicted to be 2:1 for redox balance

(Equation 3) and this was confirmed experimentally (Figure A3D). Hydrogen gas (H2) could also be used as a source of reductant in addition to pyruvate (Figure A1). Use of H2 is predicted to result in the production of equimolar amounts of butanol and acetate (Eqn. 4) and this was also demonstrated in vivo (Figure A3D).

2 Pyruvate + Butyrate + 2 ADP + 2 Pi → 2 Acetate + Butanol + 2 CO2 + 2 ATP (3)

Pyruvate + Butyrate + H2 + ADP + Pi → Acetate + Butanol + CO2 + ATP (4)

Hence, exogenous acid to alcohol conversion by P. furiosus strain A can be driven by the oxidation of glucose, pyruvate or pyruvate plus H2. Hydrogen gas cannot be used as the sole

+ source of reductant for alcohol production, however, since its redox potential (H2/H , E0ˊ = −414 mV, pH 7.0) is low enough to reduce NADP (E0ˊ = −320 mV) but not low enough to drive the reduction of P. furiosus Fd (E0ˊ = − 480 mV (22)) for the AOR reaction. In contrast, carbon monoxide (CO) oxidation is a very low potential reaction (CO/CO2, E0ˊ = −558mV) that could potentially be coupled to Fd reduction for the AOR reaction, but P. furiosus does not metabolize

235

CO (or any other C-1 compound). Using a bacterial artificial chromosome, we genetically inserted into the chromosome of P. furiosus strain A the 16-gene operon encoding the complete carbon monoxide dehydrogenase/membrane-bound hydrogenase complex (CODH, see SI Appendix

(Figure A.S5)) of the carboxydotrophic thermophile Thermococcus onnurineus, which oxidizes

CO to H2 and CO2 at 80°C (23, 24). Remarkably, engineered P. furiosus strain A/Codh was able to utilize the strong reducing power of CO to produce high concentrations of the alcohol from the corresponding acid at 72°C. For example, the CO-dependent production of isobutanol (70 mM) from isobutyrate (105 mM) by strain A/Codh is shown in Figure A4A.

CO oxidation is not coupled to the production of acetate or any other organic compound

(Figure A4B), in contrast to the use of glucose or pyruvate to drive alcohol production by P. furiosus (see SI Appendix (Figure A.S6)). A cell suspension of strain A/Codh used CO as the only electron source to drive the AOR/AdhA pathway and convert isobutyrate to isobutanol, with no other products (except for CO2 from CO oxidation: see SI Appendix (Figure A.S7)). Organic acids are therefore converted to the corresponding alcohol with minimal input of the host’s energy metabolism (Figure A4B). The CODH complex is thought to convert CO to H2 and CO2 without the involvement of intermediate electron carriers like ferredoxin in T. onnurineus, but this cannot be the case in P. furiosus. Since H2 cannot be the sole source of reductant for organic acid production, T. onnurineus CODH expressed in P. furiosus must also reduce Fd directly thereby allowed the resulting P. furiosus strain A/Codh to utilize CO as reductant for the reduction of organic acids by the AOR-AdhA pathway (Figure A4B). CO-dependent conversion of acids to alcohols also results in H2 production (see SI Appendix (Figure A.S7)) hence while CODH reduces P. furiosus Fd directly for the AOR reaction, the NADPH for the AdhA reaction is supplied at least in part via H2 and SHI (Figure A4B). However, no net H2 production was observed until

236

isobutanol production slowed down (Figure A4A). The use of CO to provide the reducing equivalents to convert organic acids to their corresponding alcohols has great potential for using industrial syngas (CO and H2), as both an energy source and a carbon source in microbial fermentations to convert organic acids generated from H2 and CO2 (25) to the corresponding alcohol.

Hence, remarkably, with the introduction of a single foreign enzyme, encoded by adhA from Thermoanaerobacter strain X514, P. furiosus can convert glucose to ethanol as well as various organic acids to the corresponding alcohol. Moreover, with the introduction of second enzyme, CODH, CO can serve as the sole source of reductant for the reduction of structurally diverse acids, including aliphatic (C2-C6) and aromatic (phenyl acetate) derivatives. Accordingly, we found that recombinant AdhA (produced in P. furiosus) was able to reduce C2-C6 aldehydes and phenyl acetaldehyde to the corresponding alcohol and thus it has a broad substrate spectrum, matching that of AOR (21) (see SI Appendix (Figure A.S8)). AdhA also has high affinities for acetaldehyde, butyraldehyde and NADPH, with apparent Km values of 63, 166 and 31 µM, respectively (see SI Appendix (Figure A.S9)), and so it is able to efficiently reduce the aldehydes generated by AOR. Maximal activity of AdhA was produced in P. furiosus when cells were grown in the 70−77°C range, representing the optimal temperature for production and folding of the

AdhA polypeptide (see SI Appendix (Figure A.S10A)). This correlates well with the optimum temperature for in vivo production of butanol from butyrate (70−80°C; see SI Appendix (Figure

A.S10B)). In fact, some butanol was still produced at 94°C, which corresponds to the upper limit for AdhA activity (see SI Appendix (Figure A.S8)).

Conversion of organic acids to the corresponding alcohols has been reported using the mesophilic anaerobic bacterium Clostridium ljungdahlii (26) grown with CO as the energy source.

237

Although the mechanism and pathway of carbon and electron flow has not been demonstrated, this likely proceeds via activation of organic acids to their coenzyme A ester and CO-derived reducing equivalents are used to form ethanol from the acyl-CoA. In contrast, the synthetic AOR-AdhA pathway of P. furiosus for CO-dependent acid to alcohol conversion does not involve CoA- derivatives (Figure A4B). There are also reports suggesting that cell-free extracts and/or cell suspensions of the anaerobic bacteria Moorella thermoacetica and Clostridium formicaceticum catalyze a “through-reduction” of acids to alcohols (27-29). These reactions were performed using

CO, formate or H2 as electron donors in the presence of artificial viologen dyes as electron carriers.

However, the fermentation of glucose to ethanol via an AOR-AdhA type pathway has not been shown previously. Moreover, direct involvement of AOR in microbial alcohol production from organic acids has not been previously demonstrated. It is the low potential fermentative pathway of P. furiosus, where ferredoxin is the sole electron acceptor of sugar oxidation, that fuels the AOR reaction and alcohol production.

Conversion of organic acids to alcohols might have a primordial origin, as the synthesis of organic acids from CO2 has been shown experimentally using metal catalysts (30). Furthermore,

C2 – C6 carboxylic acids have been postulated to be the dominant carbon species in early earth hydrothermal vents based on thermodynamic considerations (31). P. furiosus was isolated from a hot marine vent system (11), and archaea in general are considered by some as the most primitive of all life forms. AOR might, therefore, be a remnant of an ancient pathway for energy conservation in a reducing early earth environment, where geochemically-formed organic acids could have served as electron acceptors with carbon monoxide as the potential electron donor.

238

Acknowledgements

We are indebted to Amanda M. Rhaesa for technical assistance and Sanjeev K. Chandrayan,

Matthew Keller, R. Chris Hopkins, Irina Kataeva, and Angeli L. Menon for helpful discussions.

This work was supported by Bioenergy Science Center, a US Department of Energy (DOE)

Bioenergy Research Center supported by the Office of Biological and Environmental Research in the DOE Office of Science, Grant DE-PS02-06ER64304; ARPA-E Electrofuels Program of DOE

Grant DE-AR0000081; and National Science Foundation Grant CBET-1264052/CBET-1264053.

Author contributions

R.M.K. and M.W.W.A. designed research; M.B., G.J.S., D.M.N., G.L.L., R.A.B., C.J.P., B.J.V., and F.L.P. performed research; M.B., G.J.S., D.M.N., G.L.L., F.L.P., and M.W.W.A. analyzed data; and M.B., G.J.S., R.M.K., and M.W.W.A. wrote the paper.

239

A4. Tables and Figures

Figure A1. Sugar fermentation coupled to alcohol production by P. furiosus strain A

Glucose from sugars is oxidized to acetate and CO2 and ferredoxin is reduced by GAPOR and

POR. In the engineered strain A, acetate is reduced to acetaldehyde by AOR and then reduced to ethanol by the heterologously expressed AdhA from Thermoanaerobacter strain X514. The redox balance is maintained by the production of H2 by the energy-conserving, membrane-bound hydrogenase (MBH) and H2 oxidation by SHI. Organic acids added exogenously are reduced to the corresponding aldehyde and alcohol by AOR and AdhA, respectively, using reductant generated by glucose oxidation.

240

241

Figure A2. Formation of ethanol from sugars by engineered P. furiosus strains

(A) Genetic constructs with Thermoanaerobacter strain X514 adhE and/or adhA for genome insertion into P. furiosus strain COM1. (B) Specific activities of AdhE (open bars) and AdhA

(solid bars) in cell-free extracts of P. furiosus strains EA, E, and A, and parent strain COM1 grown at 72 °C. (C) Ethanol (blue bars) and acetate (red bars) produced after 4 d incubation at 72 °C with maltose (5 g⋅L−1) as the carbon source. (D) Time course of ethanol (blue) and acetate (red) production in strain A (▲) and COM1 (●) at 72 °C with cellobiose (5 g⋅L−1) as the carbon source.

After 4 d, ∼35% of the cellobiose was converted to ethanol. Experimental data represent the average of three independently prepared cell extracts or cultures (n = 3; ±SD).

242

243

Figure A3. Reduction of organic acids to alcohols by P. furiosus strain A

(A) Various organic acids were reduced to the corresponding alcohols (black bars) with concomitant production of acetate (red bars) and ethanol (blue bars) during incubation of strain A with maltose for 5 d at 72 °C. (B) Equimolar formation of butanol (green squares) from butyrate and acetate (red circles) from maltose by a 10-fold concentrated cell suspension, with only minor amounts of ethanol (blue triangles) formed. (C) Effect of aor deletion on butanol formation in strain A. (D) Effect of hydrogen on the oxidation of pyruvate to acetate and reduction of butyrate to butanol by a 10-fold concentrated cell suspension of strain A. All experimental data represent the average of three independent cultures (n = 3; ±SD).

244

245

Figure A4. CO as source of reductant for conversion of organic acids to alcohols by P. furiosus strain A/Codh

(A) Isobutanol formation (black squares) from isobutyrate (orange diamonds) in the presence of

CO (purple diamonds) by cultures of strain A/Codh grown with maltose as the carbon source at

72 °C (n = 2; ±SD). (B) CO oxidation linked to organic acid reduction by P. furiosus strain

A/Codh. The CODH complex oxidizes CO with the production of H2 and also reduces ferredoxin to provide low potential electrons to the AOR reaction. NADPH for the AdhA reaction is supplied by H2 oxidation by SHI.

246

247

Table A.S1

Strains used in this study.

Strain Strain Relevant genotype Parent Source Name ID

DSM 3638 MW001 Wild Type n/a (1)

COM1 MW002 ΔpyrF DSM 3638 (2)

COM1c MW004 ΔpyrF PgdhpyrF COM1 (3)

ΔpyrPgdhpyrF EA MW606 COM1 This study Pslp(Teth514_0627;Teth514_0564)

E MW607 ΔpyrF PgdhpyrF PslpTeth514_0627 COM1 This study

A MW608 ΔpyrF PgdhpyrF PslpTeth514_0564 COM1 This study

AΔpyrF MW610 ΔpyrF PslpTeth514_0564 MW608 This study

AΔaor MW611 Δaor ΔpyrF PgdhpyrF PslpTeth514_0564 MW610 This study

ΔpyrF PgdhpyrF PslpTeth514_0564;Pmbh1 A/Codh MW258 MW610 This study TON_1017-1031

248

Table A.S2

Primers used in this study.

Primer Sequence (5’ – 3’) Source

Pslp-SacII-F gaatccccgcggaaatagatattatcggcaaacac This study

Pslp-AdhE-R cttgtaataaggtaggcatttttctccacctcccaataatc This study

AdhE-Pslp-F gattattgggaggtggagaaaaatgcctaccttattacaag This study

AdhE-R2 gtttcccacactgcatatcacctgccattattctccataggc This study

AdhE-SphI-R gcatgcggtaccagcctcctattattctccataggcttttcta This study

AdhA-F2 ctatggagaataatggcaggtgatatgcagtgtgggaaacaaaaataaatc This study

AdhA-SphI-R tacatgcatgcggtaccagcctcctattagaaagattcttcataaatc This study

Pslp-SphI-F tacatgcatgcaaatagatattatcggcaaacac This study

AdhA-AscI-R aggcgcgcctaaaaaagattttagaaagattcttcataaatcttg This study

AdhE-AscI-R aggcgcgcctaaaaaagattttattctccataggcttttc This study

åAdhA-Pslp-F gattattgggaggtggagaaaagtgtgggaaacaaaaataaatcc This study

SP2.055 ttttctccacctcccaataatc This study

AOR1 gatagctagcgaaacttctctgcatcgtcaaga This study

AOR2 actcttcttttcaattaac This study

AOR3 agaggtcaccaacatatttattg This study

AOR4 tctacatatgatcgatctagaactttcagtattctcg This study

AOR5 ggaaataaaaagttaattgaaaagaagagtcccgggaagccgctaag This study

AOR6 caataaatatgttggtgacctctgcggccgcgtttaaacggc This study

SP2.037 gcctttcagcattgtatatgg This study

249

SP2.088 cttgaaaatgtttgaggaacacc This study

SP.237 ctgagggagatatggttaatatg This study

SP.238 ggaattactcacaaatgttccaacggccgcgtttaaacggc This study

SP.239 ttggaacatttgtgagtaattcc This study

SP.243 gaaccggaaaaagctggcatcgccaaacctccttaacatttg This study

SP.244 atgccagctttttccggttc This study

SP.245 catattaaccatatctccctcagacaacccattgatagtcatgtgc This study

250

Figure A.S1 Maps of plasmids used in this study

(A) pMB403SLP, (B) pMB404SLP and (C) pMB407SLP were used to transform Pyrococcus

furiosus strain COM1, using pyrF as selective marker (2). They contained the gene(s) adhE and/or adhA from Thermoanaerobacter strain X514 under the control of the strong constitutive

promoter Pslp. They were inserted between the converging genes PF0574 and PF0575 (4).

SeeMaterials and Methods for details.

251

252

Figure A.S2 Formation of 13C-ethanol from 13C-acetate by P. furiosus strain A.

Percentage ratio of 13C vs. total C (12C+13C) determined in ethanol (blue triangles) and acetate (red circles) in cultures incubated at 72°C on unlabeled maltose (5 g L−1, 15 mM) supplied with 8 mM of double-labeled 13C-acetate. Average of three independent cultures (n=3; ± SD).

253

254

Figure A.S3 Effect of aor deletion on ethanol formation from maltose

Formation of ethanol (blue bars) and acetate (red bars) in Strains COM1, A (harboring adhA) and

AΔaor (strain A with aor gene deleted) when incubated with maltose (5 g L−1, 15 mM) at 72°C for 3 days (n=3; ± SD).

255

256

Figure A.S4 Conversion of sugars to ethanol and butyrate to butanol by P. furiosus.

(A) Product formation from maltose in strain COM1c. (B) Formation of ethanol from maltose by strain A. (C) Formation of butanol from butyrate by strain A. Concentration of metabolites are represented as follows: red circles, acetate; blue triangles, ethanol; purple circles, H2 (represented as mmol per L medium); open black circles, total cell protein (μg cell protein per mL medium); orange diamonds, butyrate; green squares, butanol.

257

258

Figure A.S5 The T. onnurineus Codh complex and vector for insertion of the Codh expression construct into the P. furiosus chromosome

(A) A schematic representation of the Ton CODH complex encoded by TON_1017-1031. This shows subunits of carbon monoxide dehydrogenase (Codh) in red, membrane-bound hydrogenase

(Mbh) in green, and the Na+/H+ transporter (Mrp) in yellow. CooC (TON_1019) encodes the maturation protein for CooS, and the function of CooX encoded by TON_1020 is not known. (B)

The BAC-based vector pGL058 containing the Codh operon TON_1017-1031 gene cluster (Codh,

TON_1017-20, red; Mbh, TON_1021-25, green; Mrp, TON_1025-31, yellow) under transcriptional control of the promoter for P. furiosus membrane-bound hydrogenase Pmbh1; the pyrF marker cassette (indigo) and 5' and 3' homologous recombination regions (purple) used for targeted insertion into the P. furiosus chromosome. The BAC vector backbone features are shown in grey (from left to right: cat marker, oriS, repE, sopA, sopB, sopC and cos).

259

260

Figure A.S6 Reduction of organic acids to alcohols by Pyrococcus furiosus strain A/Codh in the absence of CO

Strain A/Codh was incubated at 75 °C under argon and the production and utilization of alcohols and organic acids were determined. Less than 1 mM ethanol was produced under these conditions.

261

262

Figure A.S7 Reduction of organic acids to alcohols by Pyrococcus furiosus strain A/Codh cell suspensions.

Formation of hydrogen (circles) and of isobutanol (squares) from isobutyrate in the presence (filled symbols) or absence (open symbols, dotted lines) of CO as the only electron donor by a 10-fold concentrated cell suspension (~3x109 cells/ml).

263

264

Figure A.S8 Substrate specificity and temperature optimum of T. X514 AdhA.

(A) Relative specific activity of AdhA in the cell extracts of strain A with various aldehydes. (B)

Specific activity of AdhA at different temperatures (n=3; ± SD).

265

266

Figure A.S9 Michaelis-Menten kinetics of T. X514 AdhA.

Specific activity of AdhA in the cell extracts of P. furiosus strain A at different substrate concentrations of (A) butyraldehyde, (B) acetaldehyde and (C) NADPH.

267

268

Figure A.S10 Temperature dependence of the AOR/AdhA synthetic pathway.

(A) Activity of AdhA in cell extracts of strain A grown at different temperatures. (B) Conversion of butyrate (40 mM; orange diamonds) to butanol (green squares) and acetate (red circles) by strain

A incubated at different temperatures in the presence of maltose (15 mM). Only minor amounts of ethanol (blue triangles) were formed. All experimental data represent the average of three independent cultures (n=3; ± SD).

269

270

REFERENCES

1. Nielsen J, Larsson C, van Maris A, & Pronk J (2013) Metabolic engineering of yeast for

production of fuels and chemicals. Curr Opin Biotechnol 24: 398-404.

2. Peralta-Yahya PP & Keasling JD (2010) Advanced biofuel production in microbes.

Biotechnology Journal 5: 147-162.

3. Carere CR, Rydzak T, Verbeke TJ, Cicek N, Levin DB, & Sparling R (2012) Linking

genome content to biofuel production yields: a meta-analysis of major catabolic pathways

among select H2 and ethanol-producing bacteria. Bmc Microbiol 12: 295.

4. Chang TH & Yao S (2011) Thermophilic, lignocellulolytic bacteria for ethanol production:

current state and perspectives. Appl Microbiol Biotechnol 92: 13-27.

5. Branduardi P, de Ferra F, Longo V, & Porro D (2014) Microbial n-butanol production from

Clostridia to non-Clostridial hosts. Eng Life Sci 14: 16-26.

6. Atsumi S, Hanai T, & Liao JC (2008) Non-fermentative pathways for synthesis of

branched-chain higher alcohols as biofuels. Nature 451: 86-U13.

7. Taylor MP, Eley KL, Martin S, Tuffin MI, Burton SG, & Cowan DA (2009) Thermophilic

ethanologenesis: future prospects for second-generation bioethanol production. Trends

Biotechnol 27: 398-405.

8. Wiegel J & Ljungdahl LG (1981) Thermoanaerobacter ethanolicus gen. nov., spec. nov.,

a new extreme thermophilic anaerobic bacterium Arch Microbiol 128: 343-348.

271

9. Yao S & Mikkelsen MJ (2010) Identification and overexpression of a bifunctional

aldehyde/alcohol dehydrogenase responsible for ethanol production in

Thermoanaerobacter mathranii. J Mol Microbiol Biotechnol 19: 123-133.

10. Svetlitchnyi V, Kensch O, Falkenhan D, Korseska S, Lippert N, Prinz M, Sassi J, Schickor

A, & Curvers S (2013) Single-step ethanol production from lignocellulose using novel

extremely thermophilic bacteria. Biotechnol Biofuels 6: 31.

11. Fiala G & Stetter KO (1986) Pyrococcus furiosus sp.nov. represents a novel genus of

marine heterotrophic archaebacteria growing optimally at 100°C. Arch Microbiol 145: 56-

61.

12. Sapra R, Bagramyan K, & Adams MWW (2003) A simple energy-conserving system:

Proton reduction coupled to proton translocation. Proc Natl Acad Sci USA 100: 7545-7550.

13. Basen M, Sun JS, & Adams MWW (2012) Engineering a hyperthermophilic archaeon for

temperature-dependent product formation. mBio 3: e00053-00012.

14. Keller MW, Schut GJ, Lipscomb GL, Menon AL, Iwuchukwu IJ, Leuko TT, Thorgersen

MP, Nixon WJ, Hawkins AS, Kelly RM, et al. (2013) Exploiting microbial

hyperthermophilicity to produce an industrial chemical, using hydrogen and carbon

dioxide. Proceedings of the National Academy of Sciences of the United States of America

110: 5840-5845.

15. Pei JJ, Zhou Q, Jiang Y, Le YL, Li HZ, Shao WL, & Wiegel J (2010) Thermoanaerobacter

spp. control ethanol pathway via transcriptional regulation and versatility of key enzymes.

Metab Eng 12: 420-428.

272

16. Roh Y, Liu SV, Li GS, Huang HS, Phelps TJ, & Zhou JZ (2002) Isolation and

characterization of metal-reducing Thermoanaerobacter strains from deep subsurface

environments of the Piceance Basin, Colorado. Appl Environ Microbiol 68: 6013-6020.

17. Lipscomb GL, Stirrett K, Schut GJ, Yang F, Jenney FE, Scott RA, Adams MWW, &

Westpheling J (2011) Natural competence in the hyperthermophilic archaeon Pyrococcus

furiosus facilitates genetic manipulation: construction of markerless deletions of genes

encoding the two cytoplasmic hydrogenases. Appl Environ Microbiol 77: 2232-2238.

18. Thorgersen MP, Lipscomb GL, Schut GJ, Kelly RM, & Adams MWW (2014) Deletion of

acetyl-CoA synthetases I and II increases production of 3-hydroxypropionate by the

metabolically-engineered hyperthermophile Pyrococcus furiosus. Metab Eng 22: 83-88.

19. Ma K, Hutchins A, Sung SJS, & Adams MWW (1997) Pyruvate ferredoxin oxidoreductase

from the hyperthermophilic archaeon, Pyrococcus furiosus, functions as a CoA-dependent

pyruvate decarboxylase. Proc Natl Acad Sci USA 94: 9608-9613.

20. Heider J, Ma K, & Adams MWW (1995) Purification, characterization, and metabolic

function of tungsten-containing aldehyde ferredoxin oxidoreductase from the

hyperthermophilic and proteolytic archaeon Thermococcus Strain ES-1. J Bacteriol 177:

4757-4764.

21. Roy R & Adams MWW (2002) Tungsten-dependent aldehyde oxidoreductase: A new

family of enzymes containing the pterin cofactor. Met Ions Biol Syst 39: 673-697.

22. Park JB, Fan CL, Hoffman BM, & Adams MWW (1991) Potentiometric and electron

nuclear double resonance properties of the two spin forms of the [4Fe-4S]+ cluster in the

novel ferredoxin from the hyperthermophilic archaebacterium Pyrococcus furiosus. J Biol

Chem 266: 19351-19356.

273

23. Schut GJ, Boyd ES, Peters JW, & Adams MWW (2013) The modular respiratory

complexes involved in hydrogen and sulfur metabolism by heterotrophic

hyperthermophilic archaea and their evolutionary implications. FEMS Microbiol Rev 37:

182-203.

24. Yun SH, Kwon SO, Park GW, Kim JY, Kang SG, Lee JH, Chung YH, Kim S, Choi JS, &

Il Kim S (2011) Proteome analysis of Thermococcus onnurineus NA1 reveals the

expression of hydrogen gene cluster under carboxydotrophic growth. Journal of

Proteomics 74: 1926-1933.

25. Bengelsdorf FR, Straub M, & Dürre P (2013) Bacterial synthesis gas (syngas)

fermentation. Environ Technol 34: 1639-1651.

26. Perez JM, Richter H, Loftus SE, & Angenent LT (2013) Biocatalytic reduction of short-

chain carboxylic acids into their corresponding alcohols with syngas fermentation.

Biotechnol Bioeng 110: 1066-1077.

27. Fraisse L & Simon H (1988) Observations on the reduction of non-activated carboxylates

by Clostridium formicoaceticum with carbon monoxide or formate and the influence of

various viologens. Arch Microbiol 150: 381-386.

28. Simon H, White H, Lebertz H, & Thanos I (1987) Reduction of 2-enoates and alkanoates

with carbon monoxide or formate, viologens, and Clostridium thermoaceticum to saturated

acids and unsaturated and saturated alcohols. Angew Chem Int Edit 26: 785-787.

29. White H, Strobl G, Feicht R, & Simon H (1989) Carboxylic acid reductase - A new

tungsten enzyme catalyzes the reduction of non-activated carboxylic acids to aldehydes.

Europ J Biochem 184: 89-96.

274

30. Huber C & Wächtershauser G (1997) Activated acetic acid by carbon fixation on (Fe,Ni)S

under primordial conditions. Science 276: 245-247.

31. Amend JP, LaRowe DE, McCollom TM, & Shock EL (2013) The energetics of organic

synthesis inside and outside the cell. Phil Trans R Soc B 368.

32. Zhou JZ, Fries MR, Cheesanford JC, & Tiedje JM (1995) Phylogenetic analyses of a new

group of denitrifiers capable of anaerobic growth on toluene and description of Azoarcus

tolulyticus sp. nov. Int J Syst Bacteriol 45: 500-506.

33. Lipscomb GL, Schut GJ, Thorgersen MP, Nixon WJ, Kelly RM, & Adams MWW (2014)

Engineering hydrogen gas production from formate in a hyperthermophile by heterologous

production of an 18-subunit membrane-bound complex. J Biol Chem 289: 2873-2879.

34. Yang SJ, Kataeva I, Hamilton-Brehm SD, Engle NL, Tschaplinski TJ, Doeppke C, Davis

M, Westpheling J, & Adams MWW (2009) Efficient degradation of lignocellulosic plant

biomass, without pretreatment, by the thermophilic anaerobe "Anaerocellum

thermophilum" DSM 6725. Appl Environ Microbiol 75: 4762-4769.

35. Mukund S & Adams MWW (1991) The novel tungsten-iron-sulfur protein of the

hyperthermophilic archaebacterium, Pyrococcus furiosus, is an aldehyde ferredoxin

oxidoreductase - evidence for its participation in a unique glycolytic pathway. J Biol Chem

266: 14208-14216.

36. RCoreTeam (2013) R: A language and environment for statistical computing. R

Foundation for Statistical Computing, Vienna, Austria. URL http://www.R-project.org/.

37. Thauer RK, Jungermann K, & Decker K (1977) Energy conservation in chemotropic

anaerobic bacteria. Bacteriol Rev 41: 100-180.

275