A novel hexameric NADP+‐reducing [FeFe] hydrogenase from Moorella thermoacetica

Acetogenic bacteria such as the thermophilic anaerobic model organism Moorella thermoacetica reduce CO2 with H2 as a reductant via the Wood‐Ljungdahl pathway (WLP). The enzymes of the WLP of M. thermoacetica require NADH, NADPH, and reduced ferredoxin as reductants. Whereas an electron‐bifurcating ferredoxin‐ and NAD+‐reducing hydrogenase HydABC had been described, the enzyme that reduces NADP+ remained to be identified. A likely candidate is the HydABCDEF hydrogenase from M. thermoacetica. Genes encoding for the HydABCDEF hydrogenase are expressed during growth on glucose and dimethyl sulfoxide (DMSO), an alternative electron acceptor in M. thermoacetica, whereas expression of the genes hydABC encoding for the electron‐bifurcating hydrogenase is downregulated. Therefore, we have purified the hydrogenase from cells grown on glucose and DMSO to apparent homogeneity. The enzyme had six subunits encoded by hydABCDEF and contained 58 mol of iron and 1 mol of FMN. The enzyme reduced methyl viologen with H2 as reductant and of the physiological acceptors tested, only NADP+ was reduced. Electron bifurcation with pyridine nucleotides and ferredoxin was not observed. H2‐dependent NADP+ reduction was optimal at pH 8 and 60 °C; the specific activity was 8.5 U·mg−1 and the Km for NADP+ was 0.086 mm. Cell suspensions catalyzed H2‐dependent DMSO reduction, which is in line with the hypothesis that the NADP+‐reducing hydrogenase HydABCDEF is involved in electron transfer from H2 to DMSO.


Introduction
Acetogenic bacteria are a specialized group of strictly anaerobic bacteria that grow by oxidation of organic or inorganic electron donors coupled to the reduction of CO 2 to acetate in the Wood-Ljungdahl pathway (WLP) [1][2][3].The WLP has two branches (Fig. 1).In the methyl branch, carbon dioxide is first reduced to formate, which is then bound to the C1 carrier tetrahydrofolate (THF) [4][5][6][7].Water is split off formyl-THF and methenyl-THF is reduced via methylene-to methyl-THF, the precursor of the methyl group of acetate [8,9].In the carbonyl branch, carbon dioxide is reduced to enzyme-bound CO, which is the precursor of the carbonyl group of acetate [10][11][12].Condensation of the two C1 precursors is catalyzed by the key enzyme of the pathway, the CO dehydrogenase/acetyl-CoA synthase (CODH/ACS) [13,14].The electron donor for the CODH reaction is reduced ferredoxin [15].
Moorella thermoacetica (formerly Clostridium thermoaceticum) has been the model bacterium to unravel the WLP and the enzymes involved [1,2,16,17].The formate dehydrogenase and methylene-THF dehydrogenase use NADPH as reductant, whereas the putative electron-bifurcating methylene-THF reductase uses NADH as electron donor and an unknown second electron acceptor and the CODH uses reduced ferredoxin [5,8,9,15].On the other hand, oxidation of glucose via the Emden-Meyerhof-Parnas pathway and pyruvate oxidation yields 2 mol NADH and 2 mol reduced ferredoxin [18].The missing NADPH is generated from NADH and reduced ferredoxin by the NADH-dependent ferredoxin:NADP + oxidoreductase (Nfn), which generates 2 mol of NADPH from 1 mol of reduced ferredoxin and NADH [3].Glucose oxidation also leads to the production of minor amounts of H 2 , indicating the presence of a hydrogenase in glucose-grown cells [3].An electron-bifurcating FeFehydrogenase is also key to acetogenesis from H 2 and CO 2 , since the oxidation of H 2 must ultimately lead to the reduction of NADP + , NAD + , and ferredoxin in the right stoichiometry [4].Genome analyses revealed the presence of an electron bifurcating hydrogenase HydABC that is similar to the electron-bifurcating HydABC hydrogenase from Thermotoga maritima and Acetobacterium woodii; these enzymes catalyze the H 2 -dependent reduction of NAD + and ferredoxin in stoichiometric amounts [19][20][21][22][23]. HydABC activity was present in glucose-and H 2 and CO 2 -grown cells with activities of 0.3 and 0.1 UÁmg À1 , respectively.Interestingly, cell-free extracts of H 2 and CO 2 -grown cells also catalyzed H 2 -dependent NADP + reduction, whereas this activity was not observed in glucose-grown cells [3,22].Genome analysis revealed the presence of a gene cluster (Mothe_c19180 to Mothe_c19230) that is annotated as NADP + -reducing hydrogenase and that may catalyze the observed H 2 -dependent NADP + reduction [3,19].
In order to clarify the electric connectivity between the oxidation of H 2 and reduction of CO 2 in acetogenesis by M. thermoacetica, we attempted to isolate the NADP + -reducing hydrogenase (Fig. 1).Since cells grow poorly on H 2 and CO 2 and the activity is not present in glucose-grown cells, we first identified conditions under which the enzyme is produced and cell mass can be easily obtained.

Genetic organization of the NADP + -reducing hydrogenase and properties of the deduced proteins
The genes encoding the potential NADP + -reducing hydrogenase are clustered on the chromosome.hydA is 1383 bp long and the first gene of the cluster; it is preceded by a gene encoding an aconitate hydratase (Fig. 2).hydA overlaps by 8 bp with the next gene of the cluster, hydD, which is 531 bp long and separated by 21 bp from the downstream gene hydE.hydE is 402 bp long and separated from the following gene, hydC by 48 bp.hydC is 453 bp long and separated by 11 bp from hydB, the following gene.hydB is 1779 bp long and separated by 14 bp from the last gene of the cluster, which is the 924 bp long hydF.The gene hydF is followed by a gene encoding for a sigma54-specific transcriptional regulator that is apparently not part of the hydrogenase gene cluster.
The predicted HydA has a molecular mass of 50.5 kDa.It is 44%, 37%, 33%, and 42% identical to HydA from the electron-bifurcating hydrogenases of M. thermoacetica, A. woodii, T. maritima, and Thermoanaerobacter kivui, and 40% and 45% identical to HydA from the non-electron-bifurcating NAD + - dependent hydrogenases from Syntrophomonas wolfei and Caldanaerobacter tengcongensis, respectively.HydA is predicted to harbor the H-cluster and two [4Fe-4S] cluster (Fig. 3).HydA from the before-mentioned HydABC complexes is much bigger (61-72 kDa).A protein similar to HydD is not present in the electronbifurcating hydrogenases from M. thermoacetica, A. woodii, T. maritima, and T. kivui or the NAD + -dependent hydrogenases from S. wolfei and C. tengcongensis.HydD has an identity of 41% and 40% with the postulated electron transport protein HydN from Symbiobacterium thermophilum and Clostridium magnum, respectively.HydD has a predicted molecular mass of 19.0 kDa and is predicted to harbor two [4Fe-4S] cluster.A HydE homolog is also not found in electron-bifurcating or NAD +dependent hydrogenases.It is similar to a ferredoxin family protein from Selenomonas sp CM52 and the [4Fe-4S] binding domain protein Selsp_2180 from Selenomonas sputigena, with 40% and 39% identical residues, respectively.Similar to HydD, HydE is a rather small protein with only 14.3 kDa.Nonetheless, HydE is predicted to harbor two [4Fe-4S] cluster as well.HydC has a molecular mass of 16.4 kDa and is predicted to harbor one [2Fe-2S] cluster.HydC is 44%, 46%, 44%, and 49% identical to HydC from the electron-bifurcating hydrogenase from M. thermoacetica, A. woodii, T. maritima, and T. kivui, and 33% and 46% identical to HydC from the NAD + -dependent hydrogenase from S. wolfei and C. tengcongensis, respectively.All HydC subunits harbor one [2Fe-2S] cluster.The 65-kDa HydB is the largest protein in the complex; it is 54-62% identical to HydB from electron-bifurcating hydrogenases and NAD + -dependent hydrogenases.All HydBs have similar molecular masses (64-68 kDa); only HydB of S. wolfei is much smaller, with only 43.9 kDa (Table 1).HydB is predicted to harbor 3 [4Fe-4S], one [2Fe-2S] cluster as well as FMN.The last subunit of the NADP + -reducing hydrogenase of M. thermoacetica is HydF, which has a molecular mass of 33.6 kDa and is predicted to harbor three [4Fe-4S] and one [2Fe-2S] cluster.HydF is 46-53% and 43-46%, identical to the HydA subunit of the electron-bifurcatingand NAD + -dependent hydrogenases, but lacks the Hcluster.
Conditions that promote the production of the NADP + -reducing hydrogenase HydABCDEF Transcriptome analyses that will be published elsewhere revealed that transcript levels of hydABCDEF genes were about 8-14-fold (log2 fold) increased in cells grown on H 2 and CO 2 compared to glucosegrown cells, which is in agreement with the enzyme activities mentioned above.M. thermoacetica can use DMSO instead of CO 2 as an electron acceptor [27], and the transcriptome data also revealed that, to our surprise, the hydABCDEF genes were also highly upregulated to the same level as in H 2 and CO 2 -grown cells during growth on glucose and DMSO in HCO 3 À / CO 2 -free complex medium.Under the same conditions, the genes encoding for the electron-bifurcating hydrogenase HydABC were downregulated.Since growth on glucose and DMSO leads to high cell titers, we grew cells on glucose and DMSO to search for the NADP + -reducing hydrogenase.

Purification of the NADP + -reducing hydrogenase
The cytoplasm of glucose-and DMSO-grown cells had a hydrogenase activity of 1 UÁmg À1 of protein, measured with H 2 as an electron donor and methylviologen as an electron acceptor.The proteins of the cytoplasm were separated by anion exchange chromatography on Q-sepharose, thereby eluting the hydrogenase at a NaCl concentration of 250 to 570 mM (Fig. 4A).Fractions showing the highest activity were pooled, precipitated with 20% ammonium sulfate, and further separated by hydrophobic interaction chromatography on phenylsepharose (Fig. 4B).The hydrogenase was eluted in the absence of ammonium sulfate.Fractions showing the highest activity were again pooled, concentrated using a 100-kDa Vivaspin ultrafiltration tube, and separated by size exclusion chromatography; the hydrogenase eluted in a single peak at 15.2 mL (Fig. 4C).Using this procedure, the enzyme could be enriched 127-fold with a specific activity of 127.1 UÁmg À1 (Table 2).In order to check the progress of the purification and the purity of the protein, samples from each purification step were applied to a denaturing SDS gel.After the last purification step, size exclusion chromatography, six proteins with molecular masses corresponding to the subunits of HydB (65.0 kDa), HydA (50.5 kDa), HydF (33.6 kDa), HydD (19.0 kDa), HydC (16.4 kDa), and HydE (14.3 kDa) were present (Fig. 5).If a copy number of one is assumed for each subunit in the complex, the molecular mass would be 198.8kDa.The native size of the NADP-reducing hydrogenase was determined by size exclusion chromatography at around 389 kDa, which may suggest that the native complex is a dimer of a heterohexamer (Fig. 4D).Unfortunately, the complex disintegrated in a native PAGE, and only a complex of approximately 100 kDa had hydrogenase activity (Fig. S1).This complex was cut out of the gel, and all of the proteins encoded by hydABCDEF were found by peptide mass fingerprinting (Table S1), demonstrating that we have purified the HydABCDEF complex.

Basic biochemical properties of the NADP +reducing hydrogenase
First, we assessed key biochemical properties of the purified hydrogenase, including the use of artificial and physiological electron acceptors, pH-and temperature optima, as well as the substrate affinity.The purified hydrogenase reduced the artificial electron acceptors methylviologen and benzyl viologen with activities of 127.1 and 115.3 lmol viologen/min 9 mg, respectively; NADP + was the only physiological electron acceptor and was reduced with an activity of 8.5 UÁmg À1 ; NAD + or ferredoxin were not reduced (Fig. 6A).The NADP + -reducing activity was not stimulated by the presence of ferredoxin, and the addition of FMN or FAD did not lead to a reduction of ferredoxin.The enzyme did not bifurcate electrons from H 2 to NADP + and ferredoxin neither from H 2 to NAD + and ferredoxin.

Cofactor determination of the NADP + -reducing hydrogenase
The enzyme contained 58.03 AE 1.89 mol of iron per mol of protein.This result is consistent with the predicted 60 mol of iron per mol of protein distributed to the H-cluster (two Fe and one [4Fe-4S] and two [4Fe-4S] in HydA, two [4Fe-4S] in HydD, two [4Fe-4S] in HydE, one [2Fe-2S] in HydC, one [2Fe-2S], and three [4Fe-4S] in HydB and one [2Fe-2S] and three [4Fe-4S] in HydF).To determine the flavin content, the purified enzyme was precipitated with trichloroacetic acid, and after centrifugation, the flavin-containing supernatant was analyzed by thin layer chromatography (TLC), revealing FMN in the preparation (Fig. 7).The amount of enzyme-bound flavin was determined photometrically to be 0.6 AE 0.1 molÁmol À1 of protein.This agrees with a predicted single flavin-binding site in the HydB subunit.

H 2 -dependent DMSO reduction by resting cells
Since the hydrogenase-encoding genes were highly upregulated during growth on glucose and DMSO in HCO 3 À /CO 2 -free complex medium, we addressed the question of whether H 2 can be used by M. thermoacetica as an electron donor for the reduction of DMSO.Therefore, cells were grown on glucose in HCO 3 À / CO 2 -free complex medium with DMSO as an electron acceptor, and cell suspensions were prepared.Resting cells suspended in buffer are metabolically active but do not grow.These resting cells started immediately to reduce DMSO with a rate of 11.9 AE 0.3 mMÁh À1 in the presence of H 2 , and neither acetate nor lactate was produced.Resting cells did not reduce DMSO in the absence of H 2 (Fig. 8).

Discussion
The NADP + -dependent hydrogenase from M. thermoacetica as purified here has six subunits, HydABCDEF.
The HydABC subunits of the HydABCDEF complex are very similar to the subunits of the electronbifurcating hydrogenases from, for example, A. woodii and T. kivui and the non-bifurcating hydrogenase from  S. wolfei (Fig. 9A) [21,24,25].The structure of the enzymes from A. woodii and T. kivui was solved recently to 3.4 and 3.1 A and revealed the electron transport pathway in the complex; the functional analysis also allowed to propose a new mechanism for electron bifurcation in an enzyme missing the electronbifurcating flavin [28].HydA harbors the H-cluster that catalyzes the oxidation of H 2 , and the electron-transferring [4Fe-4S] center A1, A2, A3, and A4.From A4, the electrons are passed to B1 in HydB, which is in electronic connection with a flavin that transfers the electrons to NADP + (T.kivui) or NAD + (A.woodii) in HydB (Fig. 9B) [28].From the flavin, the electron can also be transferred to the [2Fe-2S] cluster B2 and the [2Fe-2S] cluster C1 of HydC.C1 transfers the electrons to the [4Fe-4S] cluster B3, which is electronically  connected with the 4Fe-4S cluster B4 [28].From the [4Fe-4S] cluster B4, electrons are transferred to ferredoxin [28].In addition to the FeS cluster present in HydB, a Zn 2+ was experimentally detected [28].Its role in electron bifurcation is currently not fully understood [28].Evaluation of the structure of the electron-bifurcating hydrogenase from T. maritima also shows the incorporation of a Zn 2+ into the HydB subunit [29].The authors suggested that Zn 2+ is needed to allow structural movements that reduce the distance between the FeS centers in HydB and HydC [29].Furthermore, it is postulated that Zn 2+ could replace an FeS center.This would suggest Zn 2+ as part of the electron transport chain in the enzyme.
The non-bifurcating hydrogenase of S. wolfei only has the B1 cluster but misses B2, B3, and B4, thus is unable to bifurcate electrons to ferredoxin (Fig. 9C) [24].HydA of the HydABCDEF from M. thermoacetica has a different relay of iron and sulfur cluster.Apparently, the H-cluster is only followed by A1 and A2, and the [4Fe-4S] cluster A3 and A4 are missing.This may be due to the shorter polypeptide chain of HydA.According to sequence alignment, the HydB of the HydABCDEF complex contains 3 [4Fe-4S] and one [2Fe-2S] cluster, similar to the electron-bifurcating hydrogenases from T. kivui and A. woodii (Fig. S2).However, the distances between A2 and the NADP +binding site are too long, and that may be the reason for the presence of HydD and HydE, two subunits absent in the other hydrogenases that could complement the electron path from the H-cluster to NADP + .
We speculate that these additional subunits may cause sterical hindrance and abolish the ability of HydB and HydC to perform the conformational change that is required to lower the distance between the FeS cluster C1 and B3 in electron bifurcation.In electron-bifurcating hydrogenases, the long distance between C1 and B3 is shortened by a large conformational movement of HydC that allows electronic connection between C1 and B3 [28].This movement may be blocked by the additional subunits in HydDE.Next, electrons are transferred from the [4Fe-4S] cluster B3 to B4.The [4Fe-4S] cluster B4 is responsible for the electron transfer to ferredoxin.A similar electron flow can also be suggested for the hydrogenases from C. tengcongensis (Fig. 9D), but the hydrogenase HydABC from C. tengcongensis is able to reduce NAD + in the absence of ferredoxin, and electron bifurcation has not been observed [26].The basis of the inability of the enzyme to bifurcate electrons remains elusive.The NADP + -reducing hydrogenase has a third additional subunit, HydF, which is not part of any other electron-bifurcating or non-bifurcating hydrogenase known so far.HydF contains 3 [4Fe-4S] and one [2Fe-2S] cluster.The amount of FeS cluster in HydF is identical to that detected in the HydA subunit of  electron-bifurcating hydrogenases but misses the Hcluster.A structure prediction suggests that HydB and HydF physically interact, which might allow, on the one hand, the electron transfer from HydF to HydB and vice versa, and furthermore, this complex might interfere with the electron transfer from B4 to ferredoxin due to sterical hindrance (Fig. 10).The determined Fe content argues for the presence of HydDEF in the hydrogenase complex, but their roles must be identified by structural and genetic analyses.In addition, further spectroscopical analyses have to be conducted to verify the FeS cluster composition of the single subunits of the HydABCDEF complex.

Growth conditions
Moorella thermoacetica (DSM 521) was cultivated under anaerobic conditions at 55 °C in phosphate-buffered complex medium under a 100% N 2 atmosphere.The medium was prepared according to [30] using the anaerobic techniques described previously [31,32].Cells were grown in 1-or 20-L flasks (Glasger€ atebau Ochs, Bovenden/Lenglern, Germany) supplemented with 50 mM glucose carbon and energy source and 20 mM DMSO as terminal electron acceptor.

Measurement of hydrogenase activity
All enzyme assays were carried out at 55 °C in 1.

Resting cell experiments
Resting cell experiments were performed as previously described [30] with cells pregrown on glucose and DMSO.

Analytical methods
The concentration of proteins was measured according to Bradford [33].Proteins were separated in a 12% polyacrylamide gel and stained with Coomassie brilliant blue G250 [34].The molecular mass of the purified hydrogenase was determined using a calibrated superose 6 column, buffer 4, and defined size standards (ovalbumin: 43 kDa; albumin: 158 kDa; catalase: 232 kDa; ferritin: 440 kDa; thyroglobulin: 669 kDa).The isolated hydrogenase was identified by MALDI-TOF analysis.Peptide mass fingerprinting was performed by the "Functional Genomics Center Z€ urich" at ETH Zurich, Switzerland, and results were analyzed using the SCAFFOLD-PROTEOME Software version 5.3.0 (Proteome Software Inc., Portland, OR, USA).The iron content of the purified enzyme was determined by colorimetric methods [35].Flavin determination was performed by TLC as described before [36].The amount of enzyme-bound flavin was determined photometrically, as described before [36].The DMSO concentration was monitored as previously described [27].Ferredoxin was isolated from Clostridium pasteurianum [37].Native PAGE and gel activity staining were performed according to Wittig et al. [38].The gel activity assay was performed by incubating the native PAGE in 50 mL buffer (50 mM Tris-HCl pH 8, 10 mM NaCl) with 0.25 mM tetrazolium chloride and 0.5 mM NADP + under an atmosphere of 100% H 2 .

Figure 1 .
Figure 1.Scheme of acetogenesis from H 2 and CO 2 in M. thermoacetica.Two moles of CO 2 are reduced to one mole of acetyl-CoA at the expense of eight electrons.An electronbifurcating hydrogenase oxidizes H 2 with a reduction of NAD + and ferredoxin.The missing link between H 2 oxidation and NADP +reducing hydrogenase (question mark).A Nfn-type transhydrogenase may also be involved in redox balancing.The electron flow is not balanced.It is speculated that the methylene-THF reductase is electron-bifurcating (question mark).

Figure 2 .
Figure 2. Genetic organization of the NADP + -reducing hydrogenase from M. thermoacetica.The potential hydrogenase genes are colored in blue.The hydrogenase gene cluster is surrounded by the genes Mothe_c19170 and Mothe_c19240, colored orange, encoding for a putative aconitate hydratase and a putative sigma54-specific transcriptional regulator.

Figure 3 .
Figure 3.Comparison of the genetic organization, architecture, and cofactors of the NADP + -reducing hydrogenase from M. thermoacetica with described FeFe hydrogenases.The NADP + -reducing hydrogenase was compared to the electron-bifurcating hydrogenase from M. thermoacetica, A. woodii, T. kivui, and T. maritima and to the NAD + -dependent, non-electron-bifurcating hydrogenase from S. wolfei and C. tengcongensis.hydA is in red, hydB in blue, hydC in green, and subunits with no homologs are in white.The identity is given in percentages.The 2Fe-2S cluster in the electron-bifurcating hydrogenase subunit HydB from M. thermoacetica is a matter of debate since literature depicts both its presence and absence.[H], H-cluster; [F], FMN/FAD binding domain; [N], NAD(P) + binding domain ♦; [4Fe-4S] cluster; ◊, [2Fe-2S] cluster.

Figure 4 .
Figure 4. Purification of the NADP + -reducing hydrogenase from M. thermoacetica.Proteins of the cytoplasm were separated by anion exchange chromatography using a NaCl gradient from 0 to 1 M NaCl over 200 mL.Fractions containing hydrogenase activity are indicated by black shading (A).Pooled fractions after separation by anion exchange chromatography were precipitated and further separated by hydrophobic interaction chromatography.Fractions containing hydrogenase activity are indicated by black shading (B).Pooled fractions after hydrophobic interaction chromatography were concentrated by ultrafiltration and further separated by size exclusion chromatography.The hydrogenase eluted in a single peak (C, black shading).The absorbance during the chromatographic steps was monitored at 280 nm (blue line, A-C).The size exclusion chromatography column was calibrated using thyroglobulin (669 kDa, T), ferritin (440 kDa, F), catalase (232 kDa, C), albumin (158 kDa, A), and ovalbumin (43 kDa, O) as standard proteins (orange squares).The migration behavior of the NADP +reducing hydrogenase (H) is indicated by the purple circle (D).

Figure 5 .
Figure 5. SDS/PAGE monitoring the purification process of the NADP + -reducing hydrogenase from M. thermoacetica.Samples of the different purification steps were separated by SDS/PAGE (12%), and proteins were stained with Coomassie Brilliant Blue G250.Ten micrograms of protein were applied to each lane.M, prestained page ruler; CE, cell-free extract; CP, cytoplasm; Q, pooled fractions from Q-sepharose; P, pooled fractions from phenyl-sepharose; S, pooled fractions from size exclusion superose 6.

Figure 6 .
Figure 6. Biochemical characterization of the NADP + -reducing hydrogenase.All assays were performed in 1.8-mL anoxic cuvettes containing an overall liquid volume of 1 mL and 12.5 lg protein in a 100% H 2 atmosphere.(A) H 2 -dependent NADP + reduction.The assay contained 1 mM NADP + in buffer A. (B) pH optimum of the hydrogenase.The assay contained 1 mM NADP + in buffer B at 55 °C.(C) Temperature optimum of the hydrogenase.The assay contained 1 mM NADP + in buffer A at 40-80°C.(D) K m determination for NADP + .The assay contained 0-2 mM NADP + in buffer C at 60 °C (n = 3, SD).

Figure 7 .
Figure 7. Thin-layer chromatography of flavins extracted from the NADP + -reducing hydrogenase.Purified NADP + -reducing hydrogenase was precipitated, and after centrifugation, the flavincontaining supernatant was analyzed by TLC.FAD and FMN were used as standards.A mixture of 60% [v/v] n-butanol, 25% [v/v] H 2 O, and 15% acetic acid was used as the mobile phase.

Figure 8 . 3 À/
Figure 8. H 2 -dependent DMSO reduction by resting cells of M. thermoacetica.Cells of M. thermoacetica were grown on glucose in HCO 3 À /CO 2 -free complex medium with DMSO as an electron acceptor, harvested in the exponential growth phase, and resting cells prepared.The assays were performed in 120-mL serum bottles filled with HCO 3 À /CO 2 -free resting cell buffer (50 mM MOPS pH 7.0, 20 mM MgSO 4 , 20 mM KCl, 20 mM NaCl, 20% glycerol, 2 mM DTE, and 4 lM resazurin) and a 1 9 10 5 Pa H 2 (blue square) or N 2 atmosphere (green triangle).The final assay volume was 20 mL, and the assays were preincubated at 55 °C for 15 min before the reaction was started by the addition of DMSO (n = 3, SD).

Figure 10 .
Figure 10.ColabFold/AlphaFold2 predicted molecular architecture of the HydBF complex from the NADP + -reducing hydrogenase.Predicted structure without (A) and with atomic coloring (B) of the HydBF complex from M. thermoacetica.The predicted structure without (C) and with atomic coloring (D) of the HydBF complex rotated 90°counterclockwise.