Syntrophomonas wolfei Uses an NADH-Dependent, Ferredoxin-Independent [FeFe]-Hydrogenase To Reoxidize NADH

ABSTRACT Syntrophomonas wolfei syntrophically oxidizes short-chain fatty acids (four to eight carbons in length) when grown in coculture with a hydrogen- and/or formate-using methanogen. The oxidation of 3-hydroxybutyryl-coenzyme A (CoA), formed during butyrate metabolism, results in the production of NADH. The enzyme systems involved in NADH reoxidation in S. wolfei are not well understood. The genome of S. wolfei contains a multimeric [FeFe]-hydrogenase that may be a mechanism for NADH reoxidation. The S. wolfei genes for the multimeric [FeFe]-hydrogenase (hyd1ABC; SWOL_RS05165, SWOL_RS05170, SWOL_RS05175) and [FeFe]-hydrogenase maturation proteins (SWOL_RS05180, SWOL_RS05190, SWOL_RS01625) were coexpressed in Escherichia coli, and the recombinant Hyd1ABC was purified and characterized. The purified recombinant Hyd1ABC was a heterotrimer with an αβγ configuration and a molecular mass of 115 kDa. Hyd1ABC contained 29.2 ± 1.49 mol of Fe and 0.7 mol of flavin mononucleotide (FMN) per mole enzyme. The purified, recombinant Hyd1ABC reduced NAD+ and oxidized NADH without the presence of ferredoxin. The HydB subunit of the S. wolfei multimeric [FeFe]-hydrogenase lacks two iron-sulfur centers that are present in known confurcating NADH- and ferredoxin-dependent [FeFe]-hydrogenases. Hyd1ABC is a NADH-dependent hydrogenase that produces hydrogen from NADH without the need of reduced ferredoxin, which differs from confurcating [FeFe]-hydrogenases. Hyd1ABC provides a mechanism by which S. wolfei can reoxidize NADH produced during syntrophic butyrate oxidation when low hydrogen partial pressures are maintained by a hydrogen-consuming microorganism. IMPORTANCE Our work provides mechanistic understanding of the obligate metabolic coupling that occurs between hydrogen-producing fatty and aromatic acid-degrading microorganisms and their hydrogen-consuming partners in the process called syntrophy (feeding together). The multimeric [FeFe]-hydrogenase used NADH without the involvement of reduced ferredoxin. The multimeric [FeFe]-hydrogenase would produce hydrogen from NADH only when hydrogen concentrations were low. Hydrogen production from NADH by Syntrophomonas wolfei would likely cease before any detectable amount of cell growth occurred. Thus, continual hydrogen production requires the presence of a hydrogen-consuming partner to keep hydrogen concentrations low and explains, in part, the obligate requirement that S. wolfei has for a hydrogen-consuming partner organism during growth on butyrate. We have successfully expressed genes encoding a multimeric [FeFe]-hydrogenase in E. coli, demonstrating that such an approach can be advantageous to characterize complex redox proteins from difficult-to-culture microorganisms.

yield of 12.7% (Table 1). Approximately 0.55 mg of purified recombinant enzyme was obtained from 10.8 g (wet weight) of cells. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) analysis showed that the purified recombinant enzyme consisted of three subunits with molecular weights that were consistent with molecular weights predicted from the sequence of encoding genes, namely, 63 kDa compared to the predicted 63.0 kDa (HydA1), 43 kDa compared to the predicted 43.9 kDa (HydB1), and 13 kDa compared to the predicted 17.5 kDa (HydC1) with the included molecular weight of the His tag (MGSSHHHHHHSQDP, 1.6 kDa) (Fig. 1). Native PAGE analysis showed that the purified enzyme migrated as a single band (Fig. 2), and size exclusion chromatography gave a molecular mass of 115 kDa. The single visible band from native gel electrophoresis when subjected to peptide analysis showed high coverage scores for the intended peptides (Table S1). Based on the subunit analysis and the native molecular weight, Hyd1ABC is a heterotrimer with an ␣␤␥ configuration and a molecular mass of 124.5 kDa as predicted by the sequence of the encoding genes. Iron content was determined to be 29. Activity. The purified recombinant Hyd1ABC had high hydrogen-dependent methyl viologen-reducing activity and reduced methyl viologen-oxidizing activity (Table 2), typical for many [FeFe]-hydrogenases. The purified recombinant Hyd1ABC also had high NAD ϩ -reducing activity with hydrogen as the electron donor (specific activity of a Activity was determined at 37°C, pH 7.5; 1 unit of activity (U) equals 2 mol of electrons transferred per minute. MV ox , oxidized methyl viologen. 94.5 U · mg Ϫ1 at 37°C) and catalyzed hydrogen production from NADH with a specific activity of 6.6 U · mg Ϫ1 at 37°C ( Table 2). The enzyme neither oxidized NADPH nor reduced NADP ϩ under any condition tested. Neither the rate of NAD ϩ reduction nor the rate of NADH-dependent hydrogen production changed when oxidized or reduced clostridial ferredoxin, respectively, was added ( Fig. 3 and 4). In addition, the purified recombinant Hyd1ABC did not reduce clostridial ferredoxin either alone or in the presence of NAD ϩ with hydrogen as the electron donor. Hyd1ABC did not produce hydrogen from reduced clostridial ferredoxin in the absence of NADH, and the amount of hydrogen produced when both NADH and reduced ferredoxin were present was similar to that produced when only NADH was present (Fig. 4). The above-described  nases; however, the K m for reduction of NAD ϩ by Hyd1ABC was higher than that of other multimeric [FeFe]-hydrogenases (Table S2). In the above-described assays, the purified Hyd1ABC produced hydrogen from NADH under conditions where only a small amount of NADH would be oxidized to NAD ϩ during the time course of the assay and thus the ratio of NADH to NAD ϩ would remain high. To determine whether Hyd1ABC could produce hydrogen from NADH under more physiological conditions, we varied the ratio of NADH to NAD ϩ and the pH but kept the total pyridine nucleotide concentration constant at 5 mM (Table 3). In each case, more hydrogen was produced with a NADH/NAD ϩ ratio of 1.5 than with a ratio of 1.0. More hydrogen was also produced at pH 6.5 than at pH 7.0. Hydrogen partial pressures varied from 17.0 Pa, with a range of 11.3 Pa, to 40.0 Pa with a range of 6.9 Pa, with higher partial pressures observed at lower pH values and higher NADH/NAD ϩ ratios. Final hydrogen partial pressures detected in these experiments were below those predicted at thermodynamic equilibrium for the given NADH/NAD ϩ ratios. With the given liquid and gas volumes, a NADH/NAD ϩ ratio of 1.5, and a pH of 7.0, the redox potential (E=) of the NADH/NAD ϩ pair should be greater than Ϫ320 mV, even with the shift in NADH/NAD ϩ ratio as NADH is oxidized. The predicted equilibrium hydrogen partial pressure is in excess of 70 Pa for these conditions; however, the measured hydrogen partial pressure was lower, at 25.9 Pa, with a range of 0.8, suggesting that thermodynamic equilibrium was not reached.
Hydrogen levels in growing coculture. Hydrogen partial pressures measured during growth of S. wolfei in coculture with Methanospirillum hungatei on either butyrate or crotonate were between 4.6 and 18.0 Pa (Fig. S2), with the highest concentrations observed at day four and lower concentrations observed during the later stages of growth. Hydrogen partial pressures measured for pure culture S. wolfei grown on crotonate were between 27.4 and 40.0 Pa.

DISCUSSION
The identification of a NADH-dependent [FeFe]-hydrogenase in S. wolfei provides an explanation for how electrons derived from 3-hydroxybutyryl-CoA oxidation are used to make hydrogen during syntrophic growth. Stoichiometrically, all electrons derived from the oxidation of butyrate are converted to either hydrogen or formate, as S. wolfei does not produce any other reduced end products during syntrophic butyrate metabolism (20,21). The electrons generated by the oxidation of 3-hydroxybutyryl-CoA to acetoacetyl-CoA are used to reduce NAD ϩ to NADH (6). The presence of genes encoding a multimeric [FeFe]-hydrogenase and NADH-linked formate dehydrogenase in the genome of S. wolfei (4) suggested that NADH could be reoxidized using these enzyme systems. Hyd1 represents the only hydrogenase encoded within the genome of S. wolfei predicted to catalyze NADH oxidation. Of the two other hydrogenases present within the S. wolfei genome, Hyd2, encoded by SWOL_RS09950, is a membrane-bound [FeFe]-hydrogenase believed to interact with the menaquinone pool (4,7,8), while Hyd3, encoded by SWOL_RS12620, is a monomeric type [FeFe]hydrogenase (4), which has not been reported to interact with NADH (22). The hyd1A gene was upregulated under coculture conditions on butyrate compared to coculture conditions on crotonate (9), and Hyd1ABC was confirmed to be present by proteomic techniques (5, 7) during syntrophic coculture growth on butyrate with M. hungatei. a Assays were performed in 6.5-ml serum bottles with 0.5 ml of liquid volume containing 0.86 g of protein and 6 ml of headspace.
Thus, Hyd1ABC along with NADH-linked formate dehydrogenases are the likely routes of NADH reoxidation in S. wolfei. The possibility that S. wolfei utilizes a NADH-dependent hydrogenase was suggested previously due to the closeness of the redox potentials of 3-hydroxybutyryl-CoA/ acetoacetyl-CoA (E 0= ϭ Ϫ250 mV), NADH/NAD ϩ (E 0= ϭ Ϫ320 mV), and H 2 /H ϩ at 1 Pa H 2 (E= ϭ Ϫ261 mV) (3). The amino acid sequence comparisons suggested that Hyd1ABC might be a confurcating hydrogenase (4,10). Confurcating/bifurcating [FeFe]hydrogenses produce hydrogen from NADH (E 0= ϭ Ϫ320 mV) and reduced ferredoxin (E 0= ϭ Ϫ398 mV or less) (11,23). Coupling the unfavorable hydrogen production from NADH with the favorable hydrogen production from reduced ferredoxin would allow hydrogen production at partial pressures greater than 1,000 Pa at equilibrium (equivalent to E= ϭ Ϫ367 mV). The NADH/NAD ϩ midpoint potential (E 0= ϭ Ϫ320 mV) would allow for a much lower hydrogen partial pressure of close to 70 Pa at equilibrium. Using fixed NADH/NAD ϩ ratios of 1.0 and 1.5, Hyd1ABC produced partial pressures of hydrogen of 17 to 40 Pa (Table 3). NADH/NAD ϩ ratios reported for organisms grown under anaerobic conditions can vary, with some values lower than those tested here, for example, 0.27 for Clostridium kluyveri, 0.29 for Clostridium welchii, and 0.4 for E. coli (24), while other studies have reported higher ratios of 0.75 for E. coli (25) and 1.16 for Enterococcus faecalis (26). Thus, the production of hydrogen by Hyd1ABC from NADH/ NAD ϩ ratios of 1.0 or higher resembles physiological conditions for S. wolfei cells, although the actual NADH/NAD ϩ ratios during syntrophic growth conditions have not yet been determined.
The partial pressure of hydrogen produced by Hyd1ABC was within the range of the hydrogen partial pressure (25 Pa) produced by suspensions of butyrate-grown S. wolfei cells when conditions allowed only hydrogen production from NADH (27). Butyrateoxidizing cocultures of S. wolfei and Methanobacterium formicicum maintained a partial pressure of hydrogen of 10.6 Pa (28), while hydrogen concentrations for cocultures of S. wolfei and M. hungatei ranged from 4.6 to 18.0 Pa (Fig. S2). Measurements of partial pressures of hydrogen from methanogenic butyrate-degrading environments, such as lake sediment and sewage sludge, ranged from 3.7 to 26.9 Pa (29). Hydrogen partial pressures from syntrophic, butyrate-oxidizing biogas reactors ranged from 1 to 4.5 Pa (30). Methanogenic syntrophic microcosms established from peat soil implicated members of the Syntrophomonas genus in butyrate oxidation and had hydrogen partial pressures that ranged from 3 to 12.5 Pa (31). Thus, the hydrogen partial pressure that would allow for reoxidation of NADH during syntrophic butyrate metabolism is considerably less than that achieved by confurcating hydrogenases, but it is within the range of the hydrogen partial pressures produced by Hyd1ABC from NADH alone (Table  3). Furthermore, such hydrogen partial pressures are observed in coculture, microcosms, and environments where members of the Syntrophomonas genus would perform syntrophic butyrate oxidation. Thus, the hydrogen partial pressure achieved by Hyd1ABC from NADH would appear to be sufficient to allow syntrophic growth of S. wolfei on butyrate in coculture with a hydrogenotrophic methanogen.
Hyd1ABC was capable of both the reduction of NAD ϩ with hydrogen and the oxidation of NADH to produce hydrogen in the absence of ferredoxin ( Fig. 3 and 4; Table 2), indicating the lack of the tight coupling that is reported for bifurcating/ confurcating [FeFe]-hydrogenases (11,14,15). In addition, Hyd1ABC did not reduce ferredoxin even in the presence of NAD ϩ . Thus, if Hyd1ABC is capable of producing hydrogen using electrons from reduced ferredoxin under as-yet-unidentified conditions, the concomitant reduction of NAD ϩ by hydrogen would result in hydrogen being equilibrated with the redox potential of the NADH/NAD ϩ pool, preventing higher hydrogen partial pressures that would normally be achievable by confurcating hydrogenases. It is not surprising that Hyd1 does not interact with ferredoxin. Ferredoxin levels are very low in the proteome of S. wolfei (7), and we did not detect ferredoxin using the traditional approach to purify ferredoxin (N. Losey and M. J. McInerney, unpublished data), e.g., elution from an ion exchange column after high salt treatment and monitoring of the A 390 / 280 ratio (32,33). Unlike other organisms known to possess confurcating hydrogenases (11,(14)(15)(16), S. wolfei does not have a clear mechanism or need to produce reduced ferredoxin in amounts equivalent to the amount of NADH made from butyrate (4,5). Microorganisms known to possess confurcating [FeFe]hydrogenases, such as T. maritima (11), M. thermoacetica (15), and R. albus (16), have fermentative metabolisms with ferredoxin-dependent enzymes such as pyruvate:ferredoxin oxidoreductase, which can produce reduced ferredoxin in amounts needed for the coupled hydrogen production from NADH. However, S. wolfei uses fatty acids exclusively as the substrates for growth (1,21,34). Beta-oxidation of fatty acids by S. wolfei results in the formation of reduced electron transfer flavoprotein and NADH but not directly in reduced ferredoxin (6).
The presence of both HydB/HydC subunits has been suggested as being more determining for identifying bifurcating [FeFe]-hydrogenases than the composition of the HydA subunit (22). Compared to the iron-sulfur centers of known bifurcating/ confurcating [FeFe]-hydrogenases, the iron-sulfur centers of S. wolfei Hyd1ABC have two differences as predicted by NCBI's Conserved Domain Database (35) that may relate to its inability to use ferredoxin as an electron donor (Fig. 5) (Fig. 5) (12). The number of iron-sulfur centers also seems to match with the ability to bifurcate in multimeric molybdopterin formate dehydrogenases. The multimeric, molybdopterin formate dehydrogenase from Clostridium acidurici has been shown to catalyze electron bifurcation and has the same iron-sulfur centers in its HydB and HydC equivalent subunits that known bifurcating [FeFe]-hydrogenases have while a nonbifurcating, NADH-dependent formate dehydrogenase from Rhodobacter capsulatus lacks a C-terminal [4Fe-4S] cluster in its HydB subunit (13,36). The NADHdependent [FeFe]-hydrogenase from Caldanaerobacter (formerly Thermoanaerobacter) tengcongensis may be an example of an enzyme whose number of iron-sulfur centers is equivalent to those of bifurcating [FeFe]-hydrogenases but with the ability to reduce NAD ϩ and oxidize NADH without ferredoxin (37). However, the C. tengcongensis [FeFe]-hydrogenase was not tested under bifurcating/confurcating conditions by supplying ferredoxin in the presence of NAD ϩ or NADH. While the number of examples of known bifurcating/confurcating and nonbifurcating [FeFe]-hydrogenases and formate dehydrogenases is small, it appears that the type and number of conserved iron-sulfur domains in the HydB/HydC subunits may be important for electron bifurcation activity. Electron bifurcation involves two independent paths of electron flow, either away from the bifurcation site for bifurcating enzymes or toward the bifurcation site for confurcating enzymes (38,39). A change in the presence of an iron-sulfur cluster or its ability to facilitate the directionality of electron flow due to a change in redox potential or its distance from other redox centers may affect the bifurcation reaction. It is also possible that the loss of the ability to bifurcate was independent of the change in the number of iron-sulfur clusters and such centers were eliminated over time as they no longer served a critical function.
Multimeric [FeFe]-hydrogenases with the same number of iron-sulfur centers as S. wolfei Hyd1ABC are present in the genomes of other members of the family Syntrophomonadaceae (see Table S3 in the supplemental material). This suggests that Syntrophomonas zehnderi, Syntrophothermus lipocalidus, and Thermosyntropha lipolytica, which perform syntrophic oxidation of fatty acids in coculture with hydrogenotrophic methanogens (40)(41)(42), may also utilize nonconfurcating, NADH-dependent hydrogen production. Future studies of additional multimeric [FeFe]-hydrogenases may allow the determination of the relationship, if any, between certain iron-sulfur centers in the HydB and HydC subunits and the ability to perform electron bifurcation.  (Continued on next page) hydrogenases (Fig. 5). However, the metabolism of propionate by the methylmalonyl-CoA pathway by S. fumaroxidans leads to pyruvate formation and the ability to produce reduced ferredoxin, which is required as the low potential donor for electron bifurcation (43). The [FeFe]-hydrogenase in Syntrophus aciditrophicus lacks the HydC subunit, which contains a [2Fe-2S] domain (Fig. 5). S. aciditrophicus may also have the ability to generate reduced ferredoxin if the acyl-CoA dehydrogenases involved in cyclohex-1ene-1-carboxylate and cyclohexane-1-carboxylate formation can catalyze electron bifurcation (44). The syntrophic metabolism of ethanol and lactate are believed to generate reduced ferredoxin (3,10,45). The partial pressures observed during syntrophic ethanol oxidation by cocultures of Pelobacter species with M. hungatei exceeded 1,000 Pa (46), which would not be thermodynamically favorable if NADH was the only electron donor for hydrogen production. Given the higher hydrogen partial pressures and presence of reduced ferredoxin-generating systems, it seems likely that some syntrophic oxidizers may utilize ferredoxin, possibly in a confurcating mechanism, for hydrogen production. The S. wolfei Hyd1ABC was previously found to copurify with a multimeric formate dehydrogenase encoded by SWOL_RS03955, SWOL_RS03960, SWOL_RS03965, and SWOL_RS03970 (47). The copurification of Hyd1ABC with an apparent formate dehydrogenase suggested a possible formate dehydrogenase-hydrogenase complex, although an alternative would be that Hyd1 and the multimeric formate dehydrogenase are separate complexes that copurify, perhaps due to the similar nature of the HydB/ HydC subunits (Fig. 5). Either as separate complexes or as a single larger complex, Hyd1 and the multimeric formate dehydrogenase could function to convert NADH to hydrogen and formate, respectively, which would mean that both metabolites are linked to the NADH/NAD ϩ pool.
The identification of Hyd1ABC as a NADH-dependent hydrogenase suggests that NADH-dependent hydrogen production occurs during syntrophic growth on butyrate by S. wolfei without the need for reduced ferredoxin. Ferredoxin-independent hydrogen production from NADH would allow S. wolfei to avoid the energetically costly reaction to produce reduced ferredoxin from NADH. Based on the hydrogen partial pressures generated by Hyd1ABC (Table 3), continual NADH-dependent hydrogen production would require the presence of a hydrogen-consuming organism to maintain a low hydrogen partial pressure. NADH-dependent hydrogen production could explain in part the obligate requirement S. wolfei has for a hydrogen-consuming organism such as M. hungatei during growth on energetically poor substrates such as butyrate.

Plasmids and strains.
Genes encoding a putative multimeric [FeFe]-hydrogenase were amplified by PCR from S. wolfei genomic DNA with Phusion High-Fidelity DNA polymerase (Thermo Fisher Scientific, Waltham, MA) using the primers listed in Table 4. An amplicon with genes encoding HydA, HydB, and HydC (SWOL_RS05165, SWOL_RS05170, SWOL_RS05175) with the addition of BamHI and AscI restriction sites was inserted into the first multiple cloning site (MCS) of pETDuet-1 (Novagen, Merck KGaA, Darmstadt, Germany) (see Fig. S3 in the supplemental material) with the addition of an N-terminal (His) 6 residue on HydC to produce pETDuet-1 SwHydABC.
A second plasmid was created to express genes encoding [FeFe]-hydrogenase maturation proteins. The Hyd maturases HydE (SWOL_RS05180) and HydG (SWOL_RS05190), as well as a 348-bp open reading frame with no identifiable conserved domains (SWOL_RS05185), were amplified from the S. wolfei genomic template with the addition of NcoI and BamHI restriction sites and were inserted into the MCS of pCDFDuet-1 (Novagen, Merck KGaA, Darmstadt, Germany) (Fig. S1). In addition, the gene encoding the

FIG 5 Legend (Continued)
The boxes highlight the heterogeneity in the number of conserved domains in HydB (␤) subunits of characterized nonbifurcating NADH-dependent [FeFe]-hydrogensases and formate dehydrogenases compared to the similarity of conserved domains of the HydB (␤) subunits of characterized bifurcating [FeFe]-hydrogenases and a bifurcating formate dehydrogenase. Heterogeneity in conserved domains is also present in two unstudied [FeFe]-hydrogenases from syntrophic metabolizers S. aciditrophicus and S. fumaroxidans and a formate dehydrogenase from S. wolfei previously found to copurify with Hyd1ABC. The figure is a visual representation of the conserved domains as predicted by NCBI's Conserved Domain Database (40). Greek symbols at the top are the subunit designations. Accession numbers for the sequences of the subunits are given in Table S3. Reported functions of enzymes were obtained from the following references: S. wolfei hydrogenase (this publication), C. tengcongensis (37), D. fructosivorans (12), R. capsulatus (13), T. maritima (11), A. woodii (14), M. thermoacetica (15), R. albus (16), and C. acidurici (36).
Hyd maturase HydF (SWOL_RS01625) was amplified with restriction sites for NdeI and XhoI and inserted into the second MCS of pCDFDuet-1 to generate pCDFDuet-1 SwHydEFG.
In addition, a gene encoding a putative ferredoxin, SWOL_RS010890, was amplified by PCR with the addition of BamHI and HindIII restriction sites and inserted into the first MCS of pCDFDuet-1 with the addition of an N-terminal (His) 6 residue to generate pCDFDuet-1 SwFd. SWOL_RS010890 is predicted to encode a 6-kDa peptide containing a conserved domain for two [4Fe-4S] centers. The SWOL_RS010890 gene product shares 57% amino acid sequence identity with the ferredoxin from Clostridium pasteurianum (GenBank accession number M11214).
Due to changes in the locus tags for S. wolfei genes, a table listing the newly assigned locus tags for S. wolfei used in this publication, older S. wolfei locus tags, and GenBank protein accession numbers is included as Table S4. S. wolfei gene sequences included in this article are available from the S. wolfei genome (GenBank accession number NC_008346), which was published previously (4).
Expression conditions. E. coli BL21(DE3) (Thermo Fisher Scientific, Waltham, MA) cells for expression of the genes for His-tagged hydrogenase and His-tagged ferredoxin were grown in LB medium with 50 mM potassium phosphate (pH 7.5) and 10 g · liter Ϫ1 glucose with appropriate antibiotics at 37°C. Cultures were grown aerobically with shaking at 200 rpm to an optical density (OD) 600 of 0.4 to 0.6 before induction by the addition of 0.5 mM IPTG (isopropyl-␤-D-thiogalactopyranoside), followed by the addition of 2 mM ferric ammonium citrate, 10 mM sodium fumarate, and 2 mM cysteine. After induction, the cultures were sealed with rubber stoppers and sparged with nitrogen gas for 6 h until harvesting. Cells were harvested anaerobically by centrifugation at 6,000 ϫ g for 20 min at 4°C, washed by resuspension in 50 mM potassium phosphate (pH 7.5) and then pelleted again by centrifuging at 6,000 ϫ g for 10 min at 4°C. The final cell pellets were frozen in liquid N 2 until further use.
Purification of His-tagged NADH-dependent hydrogenase. All manipulations, including cell harvesting and protein purification, were performed inside a Coy chamber with an atmosphere of 95% nitrogen and 1 to 5% hydrogen. All purification buffers were prepared anoxically and contained 2 mM dithioerythritol (DTE) and 5 M both flavin adenine dinucleotide (FAD) and FMN. Cell pellets (10.8 g) were thawed and suspended in 20 ml of 50 mM Tris-HCl (pH 7.5) with 0.5 M NaCl and 30 mM imidazole, and the cells broken by passage through a French pressure cell operating at an internal pressure of 140 megapascals (MPa). Cell debris was removed by centrifugation at 13,000 ϫ g for 20 min at 4°C and passage through a 0.22-m filter. Resulting cell lysate was applied to a HisTrap HP 5-ml column (GE Healthcare Life Sciences, Pittsburgh, PA) equilibrated with the buffer used for cell breakage, followed by a wash with 50 mM imidazole in the breakage buffer, and His-tagged, NADH-dependent hydrogenase was eluted with 250 mM imidazole in the breakage buffer. Eluted fractions were then concentrated and desalted using an Amicon Ultra 0.5-ml centrifugal filter (EMD Millipore) with a 100-kDa nominal molecular mass limit filter. The desalted fractions were applied to a UNO-Q1 (Bio-Rad, Hercules, CA) column equilibrated with 50 mM Tris-HCl (pH 7.5) and eluted with a gradient of 0 to 0.6 M NaCl over a 17-min period at a flow rate of 2 ml/min. The purified protein was then aliquoted into several vials and flash frozen in liquid nitrogen until used for further analyses.
His-tagged S. wolfei ferredoxin was prepared as described for the His-tagged NADH-dependent hydrogenase, but the purification procedure ended after the HisTrap HP chromatography step.
Enzymatic assays. All assays were performed at 37°C unless noted otherwise. Enzyme activity measurements were determined in triplicate with various concentrations of protein to ensure activity was proportional to the amount of protein added. Hydrogenase-oxidizing activity was measured in rubber stopper-sealed, 1.4-ml quartz cuvettes (Nova Biotech, El Cajon, CA) with 600 l of reaction mix and with hydrogen at a pressure of 1.2 · 10 5 Pa. The reaction mixture for hydrogen-oxidizing assays consisted of 50 mM Tris-HCl (pH 7.5) or potassium phosphate (pH 7.5), 2 mM DTE, 5 M FAD, and 5 M FMN. NAD(P) ϩ reduction was tested with 1 mM NAD(P) ϩ with and without 10 M C. pasteurianum ferredoxin. Methyl viologen reduction assays were performed using a concentration of 10 mM methyl viologen. Reactions were initiated by either the addition of enzyme or by the addition of 1.2 · 10 5 Pa of hydrogen to a 100% nitrogen headspace. The kinetic constants for NAD ϩ reduction by hydrogen were determined with various NAD ϩ concentrations (0 to 2.0 mM). Spectrophotometric measurements for the reduction of NAD ϩ and NADP ϩ were performed at 340 nm ( ϭ 6.2 mM Ϫ1 cm Ϫ1 ). Clostridial ferredoxin reduction was measured at 430 nm ( Δox-red Ϸ 13.1 mM Ϫ1 cm Ϫ1 ) and methyl viologen reduction at 600 nm ( ϭ 10.0 mM Ϫ1 cm Ϫ1 ).
Hydrogen-producing activities were measured in serum bottles (6.5 ml) sealed with butyl rubber stoppers with a 0.5-ml reaction mixture and shaking at 200 rpm. The reaction mix consisted of 100 mM potassium phosphate at pH 7.5, unless otherwise indicated, as well as 2 mM DTE, 5 M FAD, and 5 M FMN. Reactions for hydrogen production were initiated by either the addition of enzyme or the addition of NADH. The headspace was sampled (0.4 ml) every 3.5 min, and percent hydrogen in the gas phase was determined by a gas chromatograph equipped with a reductive gas analyzer. The concentration of NADH and NADPH was 1 mM. A reduced ferredoxin-generating system was used, consisting of 10 mM pyruvate, 1 mM CoA, 0.1 mM thiamine pyrophosphate, 10 M clostridial or S. wolfei ferredoxin, and 0.1 U of clostridial pyruvate:ferredoxin oxidoreductase. Assays were set up to determine the final hydrogen concentrations obtained at different NADH/ NAD ϩ ratios and pH values. The conditions tested included a NADH/NAD ϩ ratio of either 1.0 or 1.5 with a total pyridine nucleotide pool size of 5 mM with a buffer of 100 mM potassium phosphate at pH 6.5 or 7.0. The ratios of NADH to NAD ϩ used in the above-described experiments were determined from the measured hydrogen partial pressures of growing S. wolfei cultures. Final hydrogen concentrations were determined in duplicate from sealed serum bottles after the reaction mixture was incubated for 24 h at room temperature.
It was previously reported that the presence of trace amounts of viologen dyes in cuvettes and stoppers can decouple bifurcating reactions (48). Hydrogen production assays involving NADH were performed using serum bottles and rubber stoppers that were not previously exposed to viologen dyes. Hydrogen-oxidizing assays were performed with stoppers that were not previously exposed to viologen dyes and with extensively washed cuvettes. Methyl viologen reduction assays were done after all other assays were completed.
Analytical techniques. Protein concentration was determined by using the Bradford protein assay (Thermo Fisher Scientific, Waltham, MA) with bovine serum albumin as the standard. SDS-PAGE and native PAGE analyses were performed using precast Novex 8 to 16% Tris-glycine (Life Technologies Co., Carlsbad, CA) according to the manufacturer's instructions. Gels were stained using Coomassie brilliant blue G-250 (Thermo Fisher Scientific, Waltham, MA). Bands from the native gel were excised for peptide identification. The protein in the excised band was digested with trypsin and subjected to high-performance liquid chromatography-tandem mass spectrometry (HPLC-MS/ MS) performed by the Laboratory for Molecular Biology and Cytometry Research at OUHSC (Oklahoma City, OK). Resulting peptide fragments were identified by Mascot search and compared to the NCBI nonredundant (nr) database.
Molecular mass of recombinant Hyd1ABC was determined by size exclusion chromatography. Size exclusion chromatography was performed using a Superdex 200 10/300 GL (GE Healthcare Life Sciences) calibrated with gel filtration standards (Bio-Rad Laboratories, Hercules, CA) using a buffer of 50 mM potassium phosphate (pH 7.5) with 0.5 M NaCl at a flow rate of 0.45 ml/min. Iron content of the purified recombinant Hyd1ABC was determined using a ferrozine assay (49). Identification and quantification of bound flavin were performed by HPLC. The purified recombinant Hyd1ABC was washed 20-fold by ultrafiltration using flavin-free 50 mM Tris-HCl (pH 7.5) and boiled for 10 min, and denatured protein was removed by centrifugation at 13,000 ϫ g for 10 min. The supernatant was then analyzed using a Kromasil 100-10-C 18 column (250 by 4.6 mm) using a buffer of 25% methanol in 50 mM ammonium formate similar to a previous HPLC technique (50) with a UV detector set to 275 nm. Retention times and peak areas were compared to FAD and FMN standards.
Hydrogen concentration was determined by a gas chromatograph (30,51) equipped with a reductive gas detector Peak Performer RCP-910 (Peak Laboratories, Mountain View, CA). Hydrogen concentrations were determined by comparing peak areas to standards containing hydrogen concentrations of 0.01 to 1%.
Culture growth conditions. Pure cultures of S. wolfei (strain DSM 2245B) and cocultures with M. hungatei strain JF-1 (strain ATCC 27890) were grown anaerobically with resazurin as the redox indicator and cysteine sulfide (0.05%) as the reducing agent. The culture medium consisted of 10 ml · liter Ϫ1 each of mineral, trace metal, and vitamin solutions (53) prepared anaerobically under a N 2 /CO 2 (80:20) atmosphere with 3.5 g · liter Ϫ1 NaHCO 3 as the buffering agent and 20 mM butyrate or crotonate as the substrate. Cultures were established in triplicate with 160-ml culture medium and 90-ml headspace volumes and incubated at 37°C with shaking at 80 rpm. Growth was monitored at OD 600 with measurements for fatty acid, hydrogen, and methane concentrations made at 4-day intervals. Determinations of butyrate, crotonate, and methane concentrations were performed by HPLC and gas chromatographyflame ionization detector (GC-FID) as described previously (9) while hydrogen concentrations were determined as described for hydrogen production assays.

SUPPLEMENTAL MATERIAL
Supplemental material for this article may be found at https://doi.org/10.1128/AEM .01335-17.

ACKNOWLEDGMENTS
We thank N. Q. Wofford for technical assistance. The initial cloning was developed and accomplished as part of the Biological and Electron Transfer and Catalysis (BETCy) EFRC, an Energy Frontier Research Center funded by the U.S. Department of Energy, Office of Science, Basic Energy Sciences under Award DE-SC0012518 to J.W.P. Additional production and characterization of Hyd1ABC were supported by National Science Foundation grant 1515843 to M.J.M. The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.