Tungsten and Molybdenum Regulation of Formate Dehydrogenase Expression in Desulfovibrio vulgaris Hildenborough

Abstract

Formate is an important energy substrate for sulfate-reducing bacteria in natural environments, and both molybdenum- and tungsten-containing formate dehydrogenases have been reported in these organisms. In this work, we studied the effect of both metals on the levels of the three formate dehydrogenases encoded in the genome of Desulfovibrio vulgaris Hildenborough, with lactate, formate, or hydrogen as electron donors. Using Western blot analysis, quantitative real-time PCR, activity-stained gels, and protein purification, we show that a metal-dependent regulatory mechanism is present, resulting in the dimeric FdhAB protein being the main enzyme present in cells grown in the presence of tungsten and the trimeric FdhABC₃ protein being the main enzyme in cells grown in the presence of molybdenum. The putatively membrane-associated formate dehydrogenase is detected only at low levels after growth with tungsten. Purification of the three enzymes and metal analysis shows that FdhABC₃ specifically incorporates Mo, whereas FdhAB can incorporate both metals. The FdhAB enzyme has a much higher catalytic efficiency than the other two. Since sulfate reducers are likely to experience high sulfide concentrations that may result in low Mo bioavailability, the ability to use W is likely to constitute a selective advantage.

INTRODUCTION

Formate is a key metabolite in anaerobic habitats, arising as a metabolic product of bacterial fermentations and functioning as a growth substrate for many microorganisms (for example, methanogens and sulfate-reducing bacteria [SRB]). Formate is also an intermediate in the energy metabolism of several prokaryotes and a crucial compound in many syntrophic associations, whereby organisms live close to the thermodynamic limit (30, 45). Recent reports indicate that formate plays an even more important role in anaerobic microbial metabolism than previously considered (14, 24, 27). The key enzyme in formate metabolism is formate dehydrogenase (FDH) (50), a member of the dimethyl sulfoxide (DMSO) reductase family. It catalyzes the reversible two-electron oxidation of formate or reduction of CO₂ and plays a role in energy metabolism and carbon fixation. In anaerobic microorganisms, FDH includes a molybdenum or tungsten bis-(pyranopterin guanidine dinucleotide) cofactor and iron-sulfur clusters (20, 41) and shows great variability in quaternary structure, physiological redox partner, and cellular location (7, 23, 38, 50).

FDH was the first enzyme shown to naturally incorporate tungsten, at a time when this element was considered to be mostly an antagonist to molybdenum (2, 52). Since then, several tungstoenzymes have been isolated and characterized, mainly but not exclusively from archaeal organisms, including FDHs, formylmethanofuran dehydrogenases (FMDH), aldehyde oxidoreductases (AOR) (not belonging to the xanthine oxidase family), and acetylene hydratase (3, 4, 20, 25, 31, 41). FDHs and FMDHs can naturally incorporate either tungsten or molybdenum. Since these two elements have very similar chemical and catalytic properties, several studies have addressed the effect of substituting molybdenum for tungsten (16). Some molybdoenzymes are able to incorporate tungsten and retain activity (e.g., Escherichia coli trimethylamine N-oxide [TMAO] reductase [9] or Rhodobacter capsulatus DMSO reductase [46]), whereas this substitution results in production of inactive enzymes in other cases, such as bacterial nitrate reductases (17, 36) or Methanobacterium formicicum FDH (29). In contrast, a fully active W-nitrate reductase was recently reported in the archaeon Pyrobaculum aerophilum, which lives in a high-tungsten environment (13). Substitution of molybdenum for tungsten has been reported for an acetylene hydratase, in which the Mo-substituted enzyme shows 60% activity of the natural tungsten protein (6). The Pyrococcus furiosus AOR tungstoenzymes were recently shown to be able to incorporate Mo, albeit with no activity (43). Given the high similarity of the two elements, it is interesting to understand how biological systems have developed solutions to discriminate between the two. Significant advances have been made in the study of tungsten uptake by the cell through the identification of selective transporters, the TupABC (28) and WtpABC (5) systems, but comparatively little is known about the intracellular regulation of protein expression in response to the two metals (4, 31).

Both molybdenum and tungsten FDHs have been reported in sulfate-reducing bacteria. These ancient organisms live in sulfide-rich environments, where molybdenum availability may be lower than that of tungsten due to the very low solubility of molybdenum sulfides (47). Nevertheless, two Mo-FDHs have been reported in SRB, from Desulfovibrio vulgaris Hildenborough (42) and Desulfovibrio desulfuricans ATCC 27774 (10). Both these enzymes are trimeric proteins that include the catalytic molybdopterin α subunit, the iron-sulfur electron transfer β subunit, and a tetraheme cytochrome c. In contrast, a dimeric αβ W-containing FDH was isolated from Desulfovibrio gigas (1). This W-FDH was the first tungstoprotein from a mesophile to have its structure determined (38). Interestingly, D. gigas W-FDH was purified from cells not depleted of molybdenum and from which a Mo-containing AOR was also isolated (1). In addition, a dimeric FDH, homologous to the D. gigas W-FDH, was isolated from Desulfovibrio alaskensis cells grown in a rich medium (Postgate medium C) without supplementation of either metal and was shown to incorporate both Mo and W (8). The first SRB genome to be sequenced, from D. vulgaris Hildenborough (19), revealed that this organism has three selenocysteine-containing FDHs. Analysis of gene organization indicates that FDH-1 (DVU0587-DVU0588) is a periplasmic dimeric protein (here referred to as FdhAB) homologous to the D. gigas W-FDH; FDH-2 is a periplasmic-facing oligomeric protein in which the αβ subunits associate with two c cytochromes and a membrane protein (DVU2481-DVU2482; here referred to as FdhM, for membrane associated); and FDH-3 (DVU2811-DVU2812-DVU2809) is the trimeric periplasmic protein in which the αβ subunits associate with a tetraheme cytochrome c₃ (here referred to as FdhABC₃), which was reported to be a Mo protein (42). In this work, we addressed the role of the molybdenum and tungsten metals on the relative expression of the three FDHs in D. vulgaris and show that different isoenzymes are expressed in the presence of either metal.

MATERIALS AND METHODS

Culture media, growth conditions, and preparation of cellular extracts.

D. vulgaris Hildenborough was grown in Postgate medium C (37) or Widdel-Pfenning (WP) defined medium (51). Postgate medium C contains 1 g/liter of yeast extract and is supplemented only with iron at a concentration of 25 μM. Molybdate or tungstate was added to a final concentration of 0.1 μM. In WP medium, molybdate or tungstate was added separately from the other trace elements to a final concentration of 0.4 μM. In each case, different electron donors were used (formate, lactate, or hydrogen) at a final concentration of 40 mM, with sulfate as the electron acceptor (38 mM in Postgate medium C and 28 mM in WP medium). When hydrogen or formate was used as the electron donor, acetate (20 mM) was also included. Growth with hydrogen in WP medium with tungstate required the presence of trace amounts of molybdate (6 nM added).

Growth with formate or lactate was performed in 1-liter closed flasks containing half the volume of medium and a gas phase of 100% N₂ for Postgate medium C and 80% N₂-20% CO₂ for WP medium. Growth with hydrogen was performed in a 3-liter bioreactor with a continuous flow of 80% H₂-20% CO₂ at 500 ml/min and stirring at 250 rpm at constant pH of 7. In all cases, the cells were grown at 37°C. Cells were collected at mid-log phase, and the different cellular extracts were obtained as described elsewhere (11).

Activity and kinetic assays.

Formate dehydrogenase activities of cell extracts were performed as described before (11). The assays were performed in a stirred cell in 50 mM Tris-HCl (pH 7.6), inside the anaerobic chamber, with the enzymes prereduced with formate as described in reference 40. For the kinetic assays with the pure enzymes, the concentrations of each FDH used were 0.0124 nM (FdhAB), 0.48 nM (FdhM), and 0.25 nM (FdhABC₃). The formate concentrations used were 2 μM, 5 μM, 25 μM, 50 μM, 250 μM, and 500 μM. Formate oxidation was measured following benzyl viologen (Sigma) reduction at 555 nm. CO₂ reductase activity with the pure enzymes was determined as described in reference 12. Each experiment was repeated at least three times.

Sulfate reduction rates.

Cells were grown in medium C without addition of Mo or W, with formate, lactate, or hydrogen as the electron donor. At mid-log phase, cells were collected and transferred to fresh culture medium, and samples were taken for sulfate quantification over the course of 3 h. Sulfate quantification was performed by HPLC analysis with indirect UV detection at 310 nm, with a PRP-×100 column (Hamilton) and a mobile phase of 3% (vol/vol) methanol and 97% (vol/vol) 6 mM hydroxybenzoic acid (pH 10).

Gel electrophoresis.

Nondenaturing polyacrylamide gels were stained for FDH activity with 2,3,5-triphenyltetrazolium chloride as described elsewhere (11). Purified proteins were analyzed with 12% SDS-PAGE stained with a Coomassie blue solution (0.1%). For heme visualization, gels were treated with a 10% trichloroacetic acid (TCA) solution and incubated in a DMB (3,3′-dimethoxybenzidene dihydrochloride) solution (15).

Protein and metal quantification.

The protein content of cell extracts was determined by the Bradford method (Sigma) with bovine gamma globulin as the standard (Bio-Rad). The concentration of pure proteins was determined by bicinchoninic acid (BCA) using the BCA protein assay kit (Novagen) with bovine serum albumin as the standard. Molybdenum and tungsten were quantified by inductively coupled plasma mass spectrometry (ICP-MS).

Western blot analysis.

Samples were run in a 10% SDS-polyacrylamide gel and transferred to 0.45-μm polyvinylidene difluoride (PVDF) membranes (Roche) for 1 h at 100 mV and 4°C in a Mini Trans-Blot electrophoretic transfer cell (Bio-Rad). The membranes were equilibrated with a Tris-buffered saline solution (10 mM Tris-HCl [pH 7.6], 150 mM NaCl) and then treated with antiserum raised against artificial peptides of each D. vulgaris FDH (Davids Biotechnologie GmbH). Immunodetection of bound antibodies was done by treatment with anti-rabbit immunoglobulin G alkaline phosphatase conjugate (Promega) diluted 1:10,000, followed by treatment with a solution of nitroblue tetrazolium salt (NBT) and 5-bromo-4-chloro-3-indolyl phosphate toluidine salt (BCIP) dissolved in 100 mM Tris (pH 9.5), 100 mM NaCl, and 5 mM MgCl₂ buffer.

Quantitative real-time PCR.

FDH gene expression changes were monitored by quantitative real-time PCR (qRT-PCR) through the amplification of the fdhB gene that encodes the small subunit of each protein. D. vulgaris was grown in WP or medium C with 0.2 g/liter of yeast extract, supplemented with Mo or W as described above, and with formate, lactate, or hydrogen as electron donors. The cells from three independent experiments were collected anaerobically at mid-log phase, and after centrifugation the pellet was immediately frozen in liquid nitrogen and later used for RNA extraction. Total RNA was obtained using the RNeasy minikit (Qiagen) and was treated with Turbo DNAfree (Ambion). cDNA synthesis from each RNA sample (1 μg) was performed using the First-Strand cDNA synthesis kit for qRT-PCR (Roche).

Primers were designed to generate amplicons of 150 bp for the β subunit genes of each FDH and the 16S rRNA gene (Table 1). The qRT-PCR was conducted in a LightCycler 1.5 real-time PCR system (Roche) using LightCycler FastStart DNA master SYBR green I (Roche). Relative standard curves were constructed for each fdhB gene and for the 16S rRNA gene by using triplicate serial dilutions of cDNA. The relative expression of the fdhB gene of FdhAB and FdhABC₃ in different growth conditions was calculated by the relative standard curve method as described in reference 26, using the 16S rRNA gene as a reference.

Table 1.

Primers used in qRT-PCR to determine the relative expression of each D. vulgaris FDH

Target	Primer sequence (5′→3′)
FdhAB β subunit	Forward: GCGGAAGTGAAGAAGACATA
	Reverse: AACCGTGCGAACAGTTGCT
Membrane Fdh β subunit	Forward: CAGGGCCGACATGCTTCAA
	Reverse: TTCCGCCACCATGTGCGAA
FdhABC3 β subunit	Forward: CTTCGACCGTGTCTCCTCT
	Reverse: CATCGACCAGATGCGCCAT
16S RNA	Forward: CCCTATTGCCAGTTGCTACC
	Reverse: AAGGGCCATGATGACTTGAC

Protein purification.

Cells grown in hydrogen-sulfate medium with either Mo or W were collected at exponential phase and broken in a French press in anaerobic conditions. The soluble fractions obtained after centrifugation were used for further purification. All chromatographic steps were performed at 4°C inside a Coy anaerobic chamber with an argon-2% hydrogen atmosphere.

(i) Step 1.

The soluble fraction from Mo- or W-supplemented cells was first purified in a Q-Sepharose high-performance 26/10 (Pharmacia) column equilibrated with 10 mM Tris-HCl (pH 7.6). A step gradient with 10 mM Tris-HCl and 1 M sodium chloride was performed.

(ii) Step 2.

The fractions containing the FDH activity (one for Mo- and two for W-soluble extracts) were concentrated and further purified on the same column (Mo conditions) or in a Resource Q column (W conditions), using the same buffers as those described in step 1.

(iii) Step 3.

The fraction with FDH activity (Mo conditions) was further purified in a Resource Q column. In W conditions, a third step was not required.

N-terminal determination.

The three purified FDHs were run in a 12% SDS-polyacrylamide gel and transferred to a PVDF membrane (Roche) as described above. The membrane was stained with a 0.1% Coomassie blue solution and destained with 40% methanol. The band corresponding to each FDH large subunit was used for N-terminal sequencing by the method of Edman and Begg, using an Applied Biosystems 491 HT sequencer.

RESULTS

Effect of Mo and W on D. vulgaris growth and FDH activities.

Initial experiments with extracts of D. vulgaris cells grown with either Mo or W in medium C (lactate as electron donor) showed that different bands were detected in native gels stained for FDH activity in either condition. This preliminary result indicated an effect of metal supplementation on FDH expression and prompted further studies.

The growth of D. vulgaris was tested in Postgate medium C, with lactate, formate, or hydrogen as the electron donor in the presence of either Mo or W, with no striking differences being observed. In lactate-sulfate medium, a small increase in growth rate and growth yield was observed in W versus Mo conditions. This effect was slightly more pronounced in formate-sulfate medium but was not detected upon growth with H₂ in a bioreactor (data not shown). Cells were collected at mid-log phase, and the FDH activities were measured in both soluble and membrane extracts (Fig. 1). For all growth conditions, the FDH activity in the membrane extracts is much lower than that of the soluble extracts. This suggests a very low level of the putatively membrane-associated FdhM and/or that its catalytic subunit does not remain associated with the membrane. Both these factors were observed in the subsequent experiments.

Fig. 1.

FDH activity in soluble fraction (SF) and membrane fraction (MF) of D. vulgaris cells grown with Mo (light gray), with W (dark gray), with both metals (stripes), or without either metal (black) added to the culture medium and with different electron donors: lactate (A), formate (B), or hydrogen (C).

With lactate as the electron donor, the FDH activities are quite low (Fig. 1A). Similar levels of activity were observed for the soluble extracts of Mo- and W-grown cells, but a higher activity for the W-grown cells was measured in the membrane extract. Notably, when cells were grown with formate or hydrogen, there was a very high increase in activity when Mo was replaced with W, and this effect was more pronounced in the soluble fraction than in the membrane fraction. The higher activities of cells grown with formate, and particularly with hydrogen, are in line with the reported higher expression of the D. vulgaris FDH genes during growth with these electron donors than that with lactate (34, 53). For these two conditions, we also tested the simultaneous addition of both metals or no addition. When the medium is not supplemented with either Mo or W, growth with formate or H₂ is still observed, but the cellular FDH activities are much lower. Medium C contains 0.1% (wt/vol) yeast extract, which supplies trace amounts of both metals (1.4 nM Mo and 0.3 nM W [25]). The low levels of Mo or W present in medium C are apparently enough to sustain growth with formate or H₂, but they do not allow full FDH activity. When both metals are present simultaneously during growth, the FDH activity is not significantly different from that with tungsten alone, indicating that an antagonistic effect is not present.

The sulfate reduction rates of mid-log cells grown in the absence of Mo or W were also measured for comparison with the FDH activities. The results were 13.0 ± 3.2 U g⁻¹ cells in lactate, 6.4 ± 0.8 U g⁻¹ cells in formate, and 11.4 ± 2.4 U g⁻¹ cells in hydrogen. Since four molecules of formate have to be oxidized to reduce one molecule of sulfate, these results indicate that in the absence of metals or in the presence of Mo, the measured FDH activity in the cell extracts (Fig. 1) is less than that expected from the observed sulfate reduction rates. In contrast, in the presence of W, this activity is higher than expected with formate or hydrogen as electron donors, indicating that under these conditions the growth is not limited by the FDH activity.

Analysis of FDHs by activity-stained native gels and Western blot analysis.

Analysis of cell extracts with native gels stained for FDH activity showed that one major band is detected for cells grown with Mo in the presence of any of the three electron donors (Fig. 2). In extracts of cells grown with W, two other bands are detected, of which band 2 is very intense in formate or hydrogen extracts. After purification of the three D. vulgaris FDHs (see below), it was possible to assign band 1 to FdhABC₃, band 2 to FdhAB, and band 3 to the αβ subunits of FdhM. Thus, growth in the presence of W leads to a strong increase in activity of FdhAB (and to a lesser extent FdhM). To test whether this increase is due to specific incorporation of W in FdhAB or to an effect in protein levels, we performed Western blot analysis of cell extracts using polyclonal antibodies raised against specific peptides of the D. vulgaris FDHs. The immunoblots of cell extracts with anti-FdhAB antibodies show clearly that there is a significant increase in FdhAB levels when Mo is replaced with W in formate-grown cells, and this effect is even more pronounced in hydrogen-grown cells (Fig. 3 A). Similar high levels of FdhAB are observed when both Mo and W are present during growth, indicating that this enzyme is not repressed by Mo. Immunoblots of hydrogen-grown cell extracts with anti-FdhM and anti-FdhABC₃ antibodies show a small increase in the levels of FdhM upon replacement of Mo for W, but no difference is detected in the levels of FdhABC₃ (Fig. 3B). In fact, the response level of FdhABC₃ antibodies is rather poor (low titer), as observed in Fig. 3B, where a low signal is observed in Mo conditions, even though the protein is known to be present from the activity-stained gels (Fig. 2) and protein purification (see below). Thus, the Western blotting results are not informative in the case of FdhABC₃.

Fig. 2.

Nondenaturing polyacrylamide gels stained for FDH activity of soluble extracts of D. vulgaris cells grown with Mo or W and different electron donors. FS, formate-sulfate (70 μg); HS, hydrogen-sulfate (60 μg); LS, lactate-sulfate (70 μg). The bands were identified after isolation of each FDH: 1, FdhABC₃; 2, FdhAB; 3, αβ subunits of FdhM.

Fig. 3.

Western blots of D. vulgaris soluble fractions using antibodies against FdhAB in cells grown with hydrogen-sulfate (HS) or formate-sulfate (FS) (A) or FdhAB (1), FdhM (2), and FdhABC₃ (3) in cells grown with hydrogen/sulfate (HS) (B). The culture medium used for cell growth was supplemented with only Mo, only W, both metals (Mo/W), or no metals (−). The amounts used for immunodetection were 75 μg of soluble fraction for FdhAB antibodies and 100 μg for membrane Fdh and FdhABC₃ antibodies.

FDH gene expression analysis by qRT-PCR.

The Western blot analysis shows a clear effect of tungsten on the protein levels of FdhAB and suggests that there is a metal regulation of FDH expression in D. vulgaris. To further test this, we studied the consequence of replacing Mo for W in the mRNA levels of the fdhB gene encoding the small subunit of the FDHs by quantitative real-time PCR. The results show clearly that the expression of FdhAB is higher in W conditions than in Mo conditions, with either of the three electron donors (Fig. 4). This effect is particularly dramatic in hydrogen-grown cells, where a very high level of expression is observed with W and a very low level with Mo, in full agreement with the observations by activity-stained gels and Western blot analyses. In lactate-Mo conditions, the expression level of FdhAB is vanishingly small relative to that of W conditions, whereas with formate as the electron donor, the difference in FdhAB expression between W and Mo conditions is less pronounced. In fact, the difference in FdhAB transcript levels in formate conditions with Mo relative to W seems less than that observed at the protein level in the activity gels (Fig. 2) and Western blots (Fig. 3A). However, the mRNA levels are not necessarily reflected in protein levels. Thus, formate induces higher expression of this enzyme in the presence of both Mo and W, whereas hydrogen leads to a strong increase in expression only in the presence of W, suggesting that the interplay between regulation by the metals and the growth substrates is different for formate relative to lactate or H₂. Previous microarray experiments have already reported that formate and hydrogen lead to increased expression of the three FDHs in D. vulgaris relative to lactate (34, 53). Overall, the qRT-PCR results for the FdhAB fdhB mRNA levels are in agreement with the protein levels and activity results and suggest that tungsten is an inducer of FdhAB gene expression.

Fig. 4.

Relative expression of the D. vulgaris FdhAB β subunit gene determined by qRT-PCR, in cells grown with either Mo or W and lactate-sulfate (LS), formate-sulfate (FS), or hydrogen-sulfate (HS). The log₁₀ relative transcription values are shown on the y axis. The 16S rRNA gene was used as the reference, and the values are normalized with the LS/W condition. Results are from three independent biological experiments (means ± standard errors).

A contrasting picture is observed for FdhABC₃. The fdhB mRNA levels of this enzyme indicate a reduced expression when Mo is replaced by W in the three growth conditions (Fig. 5). This effect is more pronounced in formate-grown cells, for which a high level of expression of FdhABC₃ in Mo conditions is observed. This effect was not apparent in the activity-stained native gels (Fig. 2), but the intensities of the bands cannot be compared among different gels, as they vary with time and this was not controlled. For this reason, the soluble extracts of cells grown in the presence of Mo were run simultaneously in the same gel (Fig. 6). This confirmed an increased activity of FdhABC₃ in formate-Mo conditions relative to the standard lactate-Mo condition. Thus, the qRT-PCR results show that the presence of tungsten/absence of molybdenum leads to a decrease of FdhABC₃ expression. Relative to lactate, formate induces expression of this enzyme but hydrogen does not, with low levels observed even in the presence of Mo. No reproducible results could be obtained with FdhM.

Fig. 5.

Relative expression of the D. vulgaris FdhABC₃ β subunit gene determined by qRT-PCR, in cells grown with either Mo or W and lactate-sulfate (LS), formate-sulfate (FS), or hydrogen-sulfate (HS). The log₁₀ relative transcription values are shown on the y axis. The 16S rRNA gene was used as the reference, and the values are normalized with the LS-Mo condition. Results are from three independent biological experiments (means ± standard errors).

Fig. 6.

Nondenaturing polyacrylamide gel stained for FDH activity of soluble extracts of D. vulgaris cells grown with Mo and different electron donors. FS, formate-sulfate (80 μg); LS, lactate-sulfate (80 μg); HS, hydrogen-sulfate (60 μg).

Purification of the three FDHs and metal analysis.

Since only one FDH has been reported from D. vulgaris Hildenborough (42), it was of interest to purify and characterize the other two FDHs present in this organism and determine their metal content. For this process, we used cells grown with hydrogen as the electron donor in medium C supplemented with either Mo or W, since this is the electron donor leading to higher FDH expression. The two purifications (Mo and W conditions) were carried out separately, in anaerobic conditions, following the FDH activity. From Mo-grown cells, only one peak of activity was detected after the first ion-exchange chromatography of the soluble fraction. Further purification led to the isolation of a single protein displaying three bands on an SDS gel (Fig. 7 A). The smaller band stained poorly with Coomassie blue but was clearly visible after heme staining. This indicated the presence of heme c (also visible in the UV-visible spectrum), suggesting that the protein isolated under these conditions was FdhABC₃, which was confirmed by N-terminal sequencing of the large subunit (YAVKL). From W-grown cells, two peaks of activity were detected after the first ion-exchange chromatography of the soluble fraction. Further purification of the major peak led to the isolation of a protein with only two subunits by SDS-PAGE (Fig. 7B), and N-terminal sequencing of the larger subunit (ELQKL) identified this protein as FdhAB. From the minor peak, a two-subunit protein was isolated, which from N-terminal sequencing of the larger subunit (AELKI) corresponds to the αβ subunits of FdhM. This indicates that these two subunits of FdhM do not form a strong complex and dissociate from the membrane-bound protein and the two c cytochromes encoded in the same gene locus. Thus, protein purification gave further confirmation that FdhABC₃ is the major FDH expressed in D. vulgaris Mo-grown cells, whereas FdhAB and FdhM are the main FDHs present in W-grown cells.

Fig. 7.

SDS-PAGE of FDHs purified from a D. vulgaris soluble fraction stained with both Coomassie blue and heme (A) or just Coomassie blue (B). M, molecular markers; 1, FdhABC₃ isolated from Mo-hydrogen-grown cells; 2, FdhAB; and 3, αβ subunits of FdhM, both isolated from W-hydrogen-grown cells.

Quantification of molybdenum and tungsten in the FDHs isolated from cells grown in medium C revealed that FdhABC₃ contained only Mo (Table 2), whereas FdhAB contained nearly equimolar amounts of Mo and W, as previously reported for D. alaskensis FdhAB (8). Metal analysis of FdhM was not possible due to the very low levels of protein obtained. Since medium C is a rich medium containing small amounts of Mo and W, we repeated the purification of FdhAB and FdhABC₃ from cells grown in the same conditions as above but in defined WP medium to which either Mo or W was added. However, growth with hydrogen in defined medium with W required the presence of a low concentration of Mo (6 nM added). Under these conditions, metal analysis by ICP-MS revealed that FdhABC₃ again incorporated only Mo (from a hydrogen-Mo medium) (Table 2), whereas FdhAB had a metal content of 80% W (from a hydrogen medium containing 0.4 μM W and 6 nM Mo). These results show that FdhABC₃ specifically incorporates Mo, whereas FdhAB can incorporate both metals.

Table 2.

Mo and W quantification in FdhAB and FdhABC₃

Protein	Mol W/mol protein	Mol Mo/mol protein
FdhABa	1.0 ± 0.20	0.8 ± 0.16
FdhABC3a	0.03 ± 0.006	0.7 ± 0.14
FdhABb	0.8 ± 0.12	0.25 ± 0.06
FdhABC3b	NDc	0.8 ± 0.08

^aPurified from cells grown in medium C.

^bPurified from cells grown in defined medium (WP).

^cND, not detected.

Kinetic characterization of the three isolated FDHs (Table 3) showed that FdhABC₃ followed Michaelis-Menten kinetics, with a turnover number of 262 s⁻¹ and a K_m of 8 μM for formate. The FdhAB (isolated from medium C and containing ∼50% Mo and W) displayed a much higher turnover number (3,684 s⁻¹) and also a lower K_m (1 μM). Thus, D. vulgaris FdhAB is an enzyme with a significantly higher catalytic efficiency (k_cat/K_m of 3,684 μM⁻¹ s⁻¹) than FdhABC₃ (k_cat/K_m of 33 μM⁻¹ s⁻¹). FdhAB isolated from cells grown in defined medium exhibited a lower level of activity, but this could have been the result of a harsher purification protocol. Since its metal content was also not 100% W, we cannot derive any conclusions as to whether the metal has an effect on the activity of the enzyme. Further studies will have to be carried out to clarify this point. The αβ subunits of FdhM showed a low turnover number of 81 s⁻¹ and a K_m of 4 μM for formate. However, this low in vitro activity may not actually reflect the in vivo situation, as it may result from the improper quaternary arrangement.

Table 3.

Kinetic characterization of FDHs isolated from D. vulgaris

Protein	Specific activity (U mg−1)	Turnover no. (s−1)	Km (μM)	kcat/Km (μM−1 s−1)
FdhAB	903	3,684	1	3,684
FdhM	367	81	4	20
FdhABC3	77	262	8	33

Several W-FDHs have been reported to reduce CO₂ (12, 18), and given the lower redox potential of the W(IV)/W(VI) couple compared to Mo(IV)/Mo(VI), it was proposed that all CO₂ reductases are W-containing reversible FDHs, with Mo-containing enzymes probably operating only in the direction of formate oxidation (12). We tested the CO₂ reduction activity of the isolated FDHs, and contrary to the expectation, we could detect no activity for the W-induced FdhAB and FdhM enzymes, but a low level of activity was observed for the Mo-containing FdhABC₃ (1 U mg⁻¹). This reveals that CO₂ reduction is not limited to W-containing FDHs and that even a Mo-containing FDH can act reversibly, even if with considerably lower activity for CO₂ reduction than formate oxidation.

DISCUSSION

The D. vulgaris Hildenborough genome encodes three FDHs, which are all in the periplasm and thus are likely to operate as formate uptake enzymes contributing to the proton motive force. Two of them (FdhABC₃ and FdhAB) are believed to transfer electrons to the cytochrome c₃ pool (19, 33), similarly to the four periplasmic hydrogenases, whereas the third FDH is associated with a membrane subunit that may allow direct reduction of the menaquinone pool. The FdhABC₃ from D. vulgaris Hildenborough has been reported to be a Mo-containing protein (42), whereas the FdhAB enzyme from the closely related D. gigas was reported to be a W protein (1). In addition, the genome of D. vulgaris Hildenborough encodes also for a Mo/W ModABC transporter and a W-specific TupABC transporter.

In this work, we studied the effect of Mo and W on the relative expression of the three D. vulgaris FDHs. There were no significant differences in growth with either Mo or W when lactate, formate, or hydrogen was used as the electron donor. However, growth with W led to a much higher FDH activity in formate-grown cells and especially in hydrogen-grown cells, whereas only a small difference was observed in lactate-grown cells. Using native gels stained for FDH activity, we observed that different FDH isoenzymes are expressed during growth with either Mo or W, suggesting that a metal-dependent regulatory mechanism is present. This effect was confirmed by Western blot analysis, qRT-PCR, and protein isolation, which showed that the level of the FdhABC₃ and FdhAB enzymes depends on the metal present, besides being affected by the electron donor as previously reported (34, 53). The presence of tungsten and absence of molybdenum induce the increase of FdhAB and reduce the level of FdhABC₃ with either of the three electron donors. Activity-stained native gels and Western blot analysis suggest that this condition also leads to an increase of FdhM, although this could not be confirmed by qRT-PCR. In cells grown in the presence of Mo, the FdhABC₃ is the main FDH present, and FdhAB is expressed only significantly with formate. In the presence of W, FdhAB is the main FDH, and FdhM is present at a low level. Overall, the results indicate that FdhABC₃ is the main FDH in formate-Mo conditions, and FdhAB is the main FDH in H₂-W conditions.

These results provide the first direct evidence of transcriptional/posttranscriptional control of FDH isoenzymes performed by Mo and W. Previous reports have indicated that similar mechanisms are present in other organisms and with another Mo/W enzyme. In Methanococcus vannielii, a single FDH was reported when cells were grown in the presence of tungsten, whereas two enzymes were observed in its absence (22). In Syntrophobacter fumaroxidans, a syntrophic acetogenic bacterium that can also grow by reduction of sulfate or fumarate, two W-containing FDHs have been isolated (12), and several others are encoded in the genome (32). In this organism, growth in the presence of W also leads to a strong increase in total FDH activity relative to growth with Mo, either during syntrophic growth with a methanogen or during growth with propionate and fumarate (35). In the presence of both metals, a decrease in FDH activity is observed for the pure culture relative to the W-supplemented culture, suggesting an antagonist effect of Mo in a W-FDH, whereas such effect is not observed for the cocultured cells, which indicates the involvement of different FDHs. In the pathogen Campylobacter jejuni, which also has both ModABC and TupABC transporters, its single FDH was shown to be active with either metal but to display a preference for tungsten, with the TupABC system acting as a specific transporter for W to be incorporated in FDH (44, 48). In this organism, a ModE-like regulatory protein represses the mod operon in the presence of both Mo or W and the tup operon only in the presence of W.

A previous example of two isoenzymes incorporating either Mo or W with differential metal regulation has been reported for FMDHs from Methanobacterium wolfei and Methanobacterium thermoautotrophicum (21). The two isoenzymes have different subunit compositions, and the tungsten protein (Fwd) is constitutively expressed, whereas the one containing molybdenum (Fmd) is induced by molybdate. Like D. vulgaris FdhAB, the Fwd protein can also incorporate Mo. Curiously, the two proteins share one of the subunits, FwdA, encoded in the fwd operon.

The understanding of Mo/W-dependent gene regulation is still rather limited, and studies have focused mainly on regulation of molybdate transport by Mo and not on intracellular mechanisms (4, 54). Nevertheless, a widespread bacterial riboswitch was recently identified that can sense Moco (molybdopterin cofactor) in the cell and control gene expression of molybdate transporters, Moco biosynthesis enzymes, and some proteins that use Moco as a cofactor (39). Interestingly, this RNA can distinguish between Moco and Tuco (tungstopterin cofactor), and a second group of closely related RNAs was identified in bacteria that use tungsten and were proposed to be Tuco riboswitches (39). We could not identify either of the riboswitch sequences close to the genes for D. vulgaris FDHs. So, at this stage, it is difficult to speculate on the mechanism of regulation that can discriminate between Mo and W. The recent report of low Mo incorporation in the P. furiosus tungsten AORs, when the organism is grown with very low W and high Mo concentrations, reveals that also in this organism a selective intracellular mechanism is present to discriminate between the two metals, even when high intracellular concentrations of Mo are present in this tungsten-dependent organism (43).

Our results show that in D. vulgaris, the incorporation of Mo or W at the active site of FDHs is regulated not only at the level of gene expression of the different isoenzymes but also by their different selectivities in metal incorporation. For FdhABC₃, a high selectivity for Mo incorporation is observed, which is paralleled by a strong decrease in protein level in the absence of Mo/presence of W. In contrast, the FdhAB enzyme can incorporate either metal, even when grown with much higher concentrations of W, indicating that the process of metal incorporation is not so specific or tightly regulated. This enzyme has a much higher catalytic efficiency than FdhABC₃ but is detected only if W is available, suggesting an overall preference of the organism for W. The ability of D. vulgaris (and probably other SRB) to use both tungsten and molybdenum, and to regulate FDH expression according to their levels, has significant environmental implications. SRB are important players in the degradation of organic matter in anaerobic habitats, namely involving syntrophic associations where formate and hydrogen are key intermediates (30, 32, 45). If sulfate is available, the activity of SRB will lead to sulfide formation and metal sulfide precipitation. The higher solubility of tungsten sulfides than molybdenum sulfides may change the relative abundance and bioavailability of these metals, and thus the capacity to use both of them increases the metabolic versatility and is likely to confer a selective advantage to the organism. A similar versatility is observed in D. vulgaris regarding the periplasmic uptake hydrogenases, where three different enzymes are expressed in response to Ni and Se availability (49). Thus, adaptation to trace element availability seems to be a fitness factor in D. vulgaris.