Sulfurospirillum multivorans (DSMZ 12446) was grown anaerobically at 28°C in rubber stoppered glass bottles shaken at 150 rpm with an aqueous to gas phase ratio of 1:1. The medium was described previously containing cysteine as reductant but did not contain vitamin B₁₂ (cyanocobalamin) here (Scholz-Muramatsu et al., 1995). Before autoclaving, the gas phase contained 150 kPa N₂. Pyruvate (40 mM) was used as electron donor and fumarate (40 mM), PCE, nitrate (10 mM), or 5% O₂ as electron acceptor. In the latter case, O₂ concentration in the gas phase was measured with the Microx 4 oxygen meter (PreSens Precision Sensing GmbH, Regensburg, Germany). The amount of 5% O₂ in the gas phase corresponded to approximately 0.5 mg/L medium in the liquid phase. Cells growing with pyruvate as electron donor were cultivated at least 50 transfers without H₂ in the gas phase to test for possible long-term regulation effects. PCE was added to the medium in a nominal concentration of 10 mM from a 0.5 M PCE hexadecane stock solution. When cells were cultivated with H₂ as electron donor, a gas phase of H₂ (150 kPa) was applied and acetate (5 mM) was added as carbon source. For experiments with N₂ as sole nitrogen source, ammonium chloride was omitted from the basal medium and N₂ (150 kPa) was used as gas phase. The following energy substrate combinations were used: pyruvate/fumarate, pyruvate/PCE, pyruvate/O₂, pyruvate alone, H₂/PCE, and H₂/nitrate. Growth was monitored photometrically at 578 nm.

Sulfurospirillum multivorans cells were harvested oxically in the mid-exponential growth phase by centrifugation (12,000 × g, 10 min at 10°C). Resulting cell pellets were washed twice in 50 mM Tris-HCl (pH 7.5) and resuspended in two volumes (2 ml per g cells) of the same buffer with additional DNase I (AppliChem, Darmstadt, Germany) and protease inhibitor (one tablet for 10 ml buffer; cOmplete Mini, EDTA-free; Roche, Mannheim, Germany). Cell disruption was carried out semi-anoxically using a French Press cell (1500 psi), which was made anoxic in a glovebox (Coy Laboratory Products Inc., Grass Lake, MI, USA), while the disruption itself was performed outside the tent. The obtained crude extracts were used for biochemical experiments. Cells were fractionated by ultracentrifugation (36,000 × g, 45 min at 4°C), supernatants were designated soluble fractions (SF). The pellets were resuspended in two volumes (2 ml per g membrane pellet) 50 mM Tris-HCl (pH 7.5); a part of this suspension was used further as membrane fraction (MF), while the rest was mixed with 0.1% Triton X-100 and stirred at 4°C overnight. Solubilized membrane extract (ME) was obtained from the resulting supernatant after ultracentrifugation (36,000 × g, 45 min at 4°C).

Enrichment was carried out at room temperature in an anoxic chamber with an atmosphere of 98% N₂ and 2% H₂ (Coy Laboratory Products Inc., Grass Lake, MI, USA) using an ÄKTA-FPLC system (GE Healthcare Europe GmbH, Freiburg, Germany). Cells were grown in 2 L Schott bottles (aqueous to gas phase ratio 1:1) with pyruvate and fumarate. Harvesting, disruption and fractionation was done as described in the previous chapter, except that 1% digitonin was used as detergent. The filtered solubilized ME was fractionated via an anion exchange column (Q-Sepharose HP column 10/10, GE Healthcare Europe GmbH, Freiburg, Germany). The Q-Sepharose column was pre-equilibrated with 50 mM Tris-HCl (pH 8.0), 0.5 mM DTT and 0.1% (v/v) Triton X-100. Subsequently, the column was washed with 200 ml of the same buffer. Elution of proteins was achieved with a linear salt gradient from 0 to 0.5 M KCl at a flow rate of 2 ml/min. Fractions containing hydrogenase activity eluted at approximately 0.2 M KCl. H₂-oxidizing activity was checked photometrically and by activity stained blue native (BN) polyacrylamide gel electrophoresis (PAGE), while purity was checked by SDS PAGE and silver-stained BN PAGE.

Non-denaturing PAGE was performed with native gels from SERVA (SERVAGEL N, Heidelberg, Germany). The anaerobic running buffer system was composed of a cathode buffer (50 mM Tricine, 15 mM BisTris-HCl, 1 ml/L Serva Blue G; pH 7.0) and anode buffer (15 mM BisTris-HCl; pH 7.0). Samples were mixed with twofold loading dye (1 M 6-Aminocaproic acid; 100 mM BisTris-HCl pH 7.0; 100 mM NaCl; 20% glycerol; 0.1% Serva Blue G250). The protein marker used was a SERVA Native Marker Liquid Mix. The gels were run in a vertical polyacrylamide electrophoresis system (Minigel-Twin, Biometra GmbH, Göttingen, Germany) starting with 50 V for 10 min and subsequently 200 V for about 2 h. The gel electrophoresis was performed in an anoxic glovebox (Coy Laboratory Products Inc., Grass Lake, MI, USA). H₂-oxidizing activity was visualized as follows: The gel was transferred to an anoxic Schott bottle filled with 50 mM Tris-HCl (pH 8.0) buffer containing 1 mM benzyl viologen (BV) as primary electron acceptor and redox mediator and 1 mM triphenyl tetrazolium chloride (TTC) as terminal electron acceptor for permanent staining. The bottle was then flushed with pure H₂ for 30 min. The incubation was carried out in a 28°C water bath until first bands were visible and then stopped by removing the native gel from the bottle. Fractions obtained from enrichment procedure were subjected to SDS-PAGE, which was silver stained after the run.

H₂-oxidizing activity was measured spectrophotometrically in butyl rubber stoppered glass cuvettes filled with 1 ml H₂-saturated buffer (50 mM Tris-HCl, pH 8.0) made anoxic by flushing with H₂ for 5 min. Redox dyes used were 1 mM benzyl viologen (BV), 1 mM methyl viologen (MV), or 1 mM methylene blue (MB) at 30°C. The colorization caused by the reduction of BV or MV was recorded at 578 nm and decolorization of MB was monitored at 570 nm. Additionally, activity was measured with the menaquinone analogs 1,4-naphthoquinone and 2,3-dimethyl-1,4-naphthoquinone (DMN). Changes in absorption during reduction of these derivatives were recorded at 270 nm. The anoxic reaction buffer contained 0.2 mM DMN or 1,4-naphthoquinone in 50 mM glycylglycine (pH 8.0) and 0.5 mM DTT. The assay was performed in rubber-stoppered quartz cuvettes, flushed with H₂. Reduction of DMN was recorded using the absorbance difference at 270 and 290 nm. Determination of protein concentration was done according to the method of Bradford (1976). Activity values are given in nanokatal (oxidation of 1 nmol H₂ per second). Temperature and pH dependance was measured with enriched hydrogenase using a 50 mM Britton-Robinson buffer system (50 mM H₃BO₃, 50 mM H₃PO₄, 50 mM acetate) in a pH range from 5.5 to 10 and a temperature range between 10 and 50°C.

Isolation of total RNA from S. multivorans was done using the RNeasy minikit (Qiagen, Hilden, Germany). Remaining genomic DNA was removed with DNase I (RNase free; Roche, Mannheim, Germany). For quality check of the RNA, agarose gel electrophoresis was performed. For all quantitative real-time PCR (qPCR) experiments, the RNA was isolated in the mid-exponential growth phase of three independently grown cultures. For cDNA synthesis, 1 μg of RNA was used as starting material in the RevertAid First Strand cDNA Synthesis kit (Thermo Scientific, Schwerte, Germany). The reaction mixture contained 1 μg RNA, 2.5 μl reverse primer, 3.5 μl 5x reaction buffer, and 2 μl 10 mM dNTP mix. It was filled up to a final volume of 17.5 μl with PCR-grade water (Fermentas, St. Leon Rot, Germany). To 10.5 μl of the mix, 0.5 μl RevertAid Reverse transcriptase (RT; 200 U/μl) was added, the residual amount (mix without RT) was used as negative control. The mix was incubated for 1 h at 42°C in a thermo cycler (Mastercycler, Personal, Eppendorf, Hamburg, Germany) and the reaction was stopped at 72°C for 5 min. Transcript levels of the different genes were compared by qPCR with the Maxima SYBR green qPCR master mix (Fermentas, St. Leon Rot, Germany); the primer pairs used are listed in Supplementary Table 1. The assay was performed in triplicates in a CFX96 qPCR machine (Bio-Rad, Munich, Germany). The PCR reaction mixture included 2.5 μl cDNA, 0.5 μM of each primer, and 6 μl 2x Maxima SYBR green qPCR master mix and was filled up to a final volume of 12 μl with PCR-grade water. Two negative controls were used, one with water as qPCR template and a reverse transcriptase negative control for each tested gene. Melting curve analysis was performed to exclude the formation of primer dimers or unspecific byproducts. Before performing the qPCR experiments, primer efficiency was tested for all primer pairs used in this study to ensure accurate and comparable amplification (see Supplementary Data Sheet 2). Three different control genes for normalization were tested: 16S rRNA, recA, rpoB. Only the first was useful as control gene under all growth conditions applied. Obtained data were thus normalized to 16S cDNA (diluted 1:10,000) and the calculation of the relative gene expression level was done according to the 2^-ΔΔCT method (Schmittgen and Livak, 2008), except where stated otherwise. For statistical analysis, student’s t-test was performed on the ^ΔCT-values; p-values lower than 0.05 were regarded as significant.

To determine and compare the transcript levels of the hydrogenase genes under different growth conditions, S. multivorans was grown with the following substrate combinations: pyruvate plus fumarate (standard condition), PCE, 5% O₂, or without external electron acceptor, or H₂ plus PCE or nitrate. In addition, N₂-fixing conditions were achieved with N₂ as sole nitrogen source. Prior to the main qPCR experiments, three housekeeping genes were tested for their suitability as reference genes: 16S rRNA, recA and rpoB. The latter was not considered further due to unspecific products and primer specificities out of an acceptable range (more than 10% deviation from primer specificities for the other primer pairs). The recA gene was unsuitable, since the transcript levels varied too much among replicates and under the different growth conditions (see Supplementary Data Sheet 2). Therefore, hydrogenase qPCR data were normalized to the 16S rRNA gene when comparing transcript levels under different growth conditions. For hydrogenase qPCR, the following genes encoding the catalytic subunits of each hydrogenase gene cluster were chosen: hydB (SMUL_1424) of the MBH gene cluster, hupL (SMUL_1422) from the cytoplasmic uptake hydrogenase (hupSL), echE (SMUL_1307) from the CooH-like hydrogenase and hyfG (SMUL_2388) from the Hyf hydrogenase. Primers used in qPCR are listed in Supplementary Table 1.

Transcripts of only two hydrogenase genes, hydB and hyfG (large subunits of the MBH and of Hyf) were found at levels of 0.03 to 0.08 (after normalization to 16S rRNA transcript level) under nearly all tested growth conditions. This was even the case when pyruvate rather than H₂ was used as electron donor in cells cultivated for at least 50 transfers with pyruvate as sole energy source. A long-term transcriptional regulatory effect, as was seen for pceA (John et al., 2009), can therefore be excluded. Opposed to hydB and hyfG, hupL and echE transcripts were not detected under most conditions, except very low transcript levels under pyruvate/fumarate (below 0.001, see Figure 2 and Supplementary Data Sheet 2). To establish the hydrogenases that are involved in H₂-dependent PCE respiration, S. multivorans was grown in the presence of 100% H₂ in the gas phase as electron donor and 10 mM PCE in a hexadecane phase as electron acceptor. Only an insignificant increase in hydB transcript level (about twofold, p-value 0.224) and a decrease of hyfG (about threefold, p-value 0.562) was observed (Figure 2). To compare these transcript levels to those of another anaerobic H₂-dependent respiration, cells were grown with nitrate as electron acceptor and H₂ as electron donor. Nitrate rather than fumarate was chosen as electron acceptor, since S. multivorans was shown to be capable of growth with fumarate alone by disproportionation (Scholz-Muramatsu et al., 1995). The hydB gene transcript level dropped slightly, but insignificantly compared to H₂/PCE (twofold, p-value 0.384), while hyfG was reduced significantly in H₂/nitrate cells (about 30-fold lower; p-value 0.022) (Figure 2).

Since HupSL might be involved in recycling hydrogen produced during N₂ fixation, as was shown for similar enzymes in cyanobacteria (Tamagnini et al., 2007; Bothe et al., 2010) and S. multivorans was shown to fix N₂ (Ju et al., 2007; Goris et al., 2014), the transcript levels of hupL were investigated with N₂ as sole N-source. S. multivorans was grown under standard conditions (pyruvate/fumarate plus NH₄Cl as N-source) and with medium containing N₂ as sole N-source. The hupL transcript level did not increase with N₂ as sole N-source when compared to cells grown with NH₄⁺. Instead, hupL and hydB levels decreased about 7.5- and 3-fold, albeit insignificantly with a p-value of 0.592 and 0.892, respectively (Figure 3).

Aerobic growth of S. multivorans with pyruvate and an O₂-concentration of about 5% in the gas phase was described recently (Goris et al., 2014). Since NiFe hydrogenases are generally inactivated by O₂ (Stiebritz and Reiher, 2012), it was of interest whether O₂ could possibly cause a down-regulation of hydrogenase gene expression as found for other bacteria (Kovács et al., 2005). Therefore, qPCR of cells grown in the presence of 5% O₂ in the gas phase with pyruvate as electron donor was performed to test if hydrogenase transcription occurs in the presence of oxygen. Cells were harvested at an O₂ concentration in the medium of at least 0.2 mg/L. Surprisingly, the transcript levels of hydB and hyfG were similar or insignificantly higher (about twofold for hydB, p-value of 0.553) under microoxic conditions compared to cells grown with fumarate as electron acceptor (Figure 3), while hupL and echE transcripts were not detected. The cells were, however, not able to grow with H₂ as electron donor and 5% O₂ as electron acceptor and acetate as carbon source.

Sulfurospirillum multivorans was described to grow fermentatively with pyruvate as the sole energy source (Scholz-Muramatsu et al., 1995). To investigate whether one of the putative H₂-producing hydrogenases shows higher transcript levels under this growth condition, we compared qPCR results obtained from pyruvate-grown cells with those from pyruvate/fumarate-grown cells. The hydB transcript level was lowered insignificantly (p-value 0.229) about threefold, while the hyfG level was increased by a factor of 3 with a p-value of 0.056 (Figure 4).

To biochemically characterize the H₂-oxidizing enzyme of cells grown under PCE-respiring conditions, we measured the H₂-uptake activity of subcellular fractions from S. multivorans in different enzymatic assays. First, crude extracts of cells grown with pyruvate or H₂ as electron donor and PCE as electron acceptor were tested spectrophotometrically for H₂-oxidizing activity with several different artificial redox mediators as electron acceptors. The activity with BV was by far the highest and about five times higher than the activities with MV and approximately tenfold of the activity with MB (Figure 5A). The other redox mediators tested, NAD⁺, nitro blue tetrazolium chloride (NBT) or phenazine methosulfate (PMS), and negative controls without H₂ or without cell extracts showed no activity (Supplementary Table 2). In accordance with the quantitative PCR results, there was not much difference in the activity levels between pyruvate/PCE- and H₂/PCE-grown cells, with a specific H₂-oxidizing activity of 60 nkat/mg protein with BV as electron acceptor for pyruvate/PCE cells, 50 nkat/mg for pyruvate/fumarate cells and 70 nkat/mg for H₂/PCE cells (Figure 5A). Crude extracts obtained from cells grown with pyruvate as electron donor and 5% O₂ as electron acceptor showed H₂-dependent enzyme activity with BV (25 ± 2.3 nkat/mg). After subcellular fractionation, the highest specific activity (about 60 nkat/mg) was found in the MF while only around one tenth (7 and 5 nkat/mg for extracts obtained from pyruvate- and H₂-grown cells, respectively) was measured for the SF. The activities of crude extract, MF and solubilized MEs were about 50 to 70 nkat/mg (Figure 5B).

The different subcellular fractions were subjected to native PAGE followed by activity staining. This method allows to estimate the size and activity of catalytically active enzyme complexes and is often routinely used for the characterization of hydrogenases (Ballantine and Boxer, 1985; Pinske et al., 2012). For this purpose, S. multivorans cells were grown with either pyruvate or H₂ as electron donor and PCE as electron acceptor. Both, cell extracts and subcellular fractions (SF and ME solubilized with digitonin), were subjected to activity staining with BV as primary and TTC (triphenyl tetrazolium chloride) as secondary electron acceptor. After incubation of the gel with H₂ for 15–30 min at 28°C, a distinct band showing the typical reddish color of reduced TTC appeared at a size of about 270 kDa. This band was missing or only weakly visible in the SF (Figure 6). A prolonged incubation time up to 12 h did not lead to an additional band in the activity staining with membrane or SFs, suggesting the involvement of only one hydrogenase in H₂ oxidation, albeit a hydrogenase not performing H₂ oxidation under the given conditions might have been overlooked in this assay. The size of 270 kDa correlates to a dimeric form of the heterotrimeric S. multivorans MBH, which is predicted to be around 244 kDa (predicted sizes for the maturated MBH – large subunit HydB: 62 kDa, small subunit HydA: 34 kDa, and the membrane-integral cytochrome b, HydC: 26 kDa). This value, however, does not take into account the unknown contribution of the digitonin micelle size.

The presumable physiological electron acceptor for the MBH is menaquinone, as was shown for the MBH from W. succinogenes (Dross et al., 1992). To test the H₂-dependent reactivity of the MBH of S. multivorans with quinones, H₂ oxidation activity was measured spectrophotometrically with two different quinone analogs. Similar to the MBH of W. succinogenes, crude and MEs of S. multivorans showed activity with 2,3-dimethyl-1,4-naphtoquinone (DMN) (Table 1). Extracts from cells grown with H₂ showed approximately 10% higher activities than those from cells grown with pyruvate (Table 1). The hydrophilic quinone analog, 1,4-naphthoquinone, was neither reduced by the MBH of S. multivorans nor by that of W. succinogenes (Dross et al., 1992).

To investigate the involvement of a cytochrome b, spectroscopic changes in UV/VIS absorption of MFs were recorded before and after the addition of H₂. The MF showed an absorption maximum at 414 nm in the absence of H₂. After flushing the sample with H₂, the Soret band at 414 nm shifted to 427 nm and two peaks at 530 and 561 nm appeared (Figure 7A). When calculating the difference spectra of H₂-reduced minus oxidized state, the spectrum showed α-, β-, and γ-absorption peaks at 561, 530, and 427 nm, respectively (Figure 7B). These spectral properties correspond to absorption spectra of b-type cytochromes (Yoon et al., 2008; Eguchi et al., 2012).

Attempts to express the MBH of S. multivorans heterologously did not lead to the formation of an active enzyme, therefore, an enrichment from MEs was carried out (Figure 8). After solubilization of the MF with digitonin, 18% of hydrogenase enzyme activity from the MF could be recovered. An enrichment of 221-fold was obtained using anion exchange chromatography with a Q-Sepharose column (Table 2). After elution, the fractions with the highest activities (fractions 19–23) were applied to SDS-PAGE, native PAGE and BN activity staining (Supplementary Figure 2). In the latter, all fractions showed the remarkable red band at 270 kDa, which was most intense in fractions 20 and 21. Similar to the cell extracts, no second band was visible after prolonged incubation (up to 12 h) in the reaction mixture. Silver stained native PAGE showed the same 270 kDa band in all applied fractions with the most intense band in fraction 20 (Figure 8A). Other distinct bands in fraction 20 appeared around 50 and 130 kDa. After SDS-PAGE, all tested fractions showed a band at about 65 kDa, which corresponds to the size of HydB (62 kDa). This band was strongest in fractions 20 and 21. Additionally, a band at nearly 35 kDa was showing up in fractions 20 to 23, correlating to the theoretical size of the mature small subunit of the MBH (34 kDa). Several bands between 25 and 30 kDa were visible in all fractions, but none of them could be attributed clearly to HydC (26 kDa). One predominant band at 25 kDa observed in fractions 20 and 21 might represent a cytochrome b, especially since both fractions showed activity with DMN as electron acceptor (Supplementary Figure 2). Fraction 20 appeared to exhibit less additional bands than fraction 21 after SDS-PAGE (Supplementary Figure 2), therefore, we chose this fraction for further purification. Hydrophobic interaction chromatography or gel filtration did not lead to further enrichment of active trimeric MBH, so that further characterization was performed with fraction 20. The presence of the membrane-integral cytochrome b (HydC) in this fraction was shown using UV/Vis spectroscopy. The same characteristic absorption spectra and difference spectrum were recorded as those obtained for the MF (Supplementary Figure 3). Compared to the membranes, the enriched MBH showed a higher absorption in the difference spectrum, suggesting a higher cytochrome b concentration. The optimal condition for H₂ oxidation by fraction 20 was at pH 8.0 and 40°C when BV was used as electron acceptor in the photometric assay.