Evidence for a Hexaheteromeric Methylenetetrahydrofolate Reductase in Moorella thermoacetica

Abstract

Moorella thermoacetica can grow with H₂ and CO₂, forming acetic acid from 2 CO₂ via the Wood-Ljungdahl pathway. All enzymes involved in this pathway have been characterized to date, except for methylenetetrahydrofolate reductase (MetF). We report here that the M. thermoacetica gene that putatively encodes this enzyme, metF, is part of a transcription unit also containing the genes hdrCBA, mvhD, and metV. MetF copurified with the other five proteins encoded in the unit in a hexaheteromeric complex with an apparent molecular mass in the 320-kDa range. The 40-fold-enriched preparation contained per mg protein 3.1 nmol flavin adenine dinucleotide (FAD), 3.4 nmol flavin mononucleotide (FMN), and 110 nmol iron, almost as predicted from the primary structure of the six subunits. It catalyzed the reduction of methylenetetrahydrofolate with reduced benzyl viologen but not with NAD(P)H in either the absence or presence of oxidized ferredoxin. It also catalyzed the reversible reduction of benzyl viologen with NADH (diaphorase activity). Heterologous expression of the metF gene in Escherichia coli revealed that the subunit MetF contains one FMN rather than FAD. MetF exhibited 70-fold-higher methylenetetrahydrofolate reductase activity with benzyl viologen when produced together with MetV, which in part shows sequence similarity to MetF. Heterologously produced HdrA contained 2 FADs and had NAD-specific diaphorase activity. Our results suggested that the physiological electron donor for methylenetetrahydrofolate reduction in M. thermoacetica is NADH and that the exergonic reduction of methylenetetrahydrofolate with NADH is coupled via flavin-based electron bifurcation with the endergonic reduction of an electron acceptor, whose identity remains unknown.

INTRODUCTION

Acetogens are anaerobic, mostly Gram-positive bacteria that reduce 2 CO₂s to acetic acid via the Wood-Ljungdahl pathway in their energy metabolism (1). They are of biotechnological interest, since many of them can grow chemolithoautotrophically on H₂ and CO₂ or on CO or mixtures thereof (syngas) as the sole energy sources. When CO is used, ethanol and 2.3-butanediol can be major fermentation end products (2, 3). Genetically engineered acetogens can also produce acetone, butanol, or other products of industrial interest (4–6).

The model organism for the study of the biochemistry of the Wood-Ljungdahl pathway has been Moorella thermoacetica (formerly Clostridium thermoaceticum) (7–10). This Gram-positive anaerobic bacterium ferments sugars to three acetic acids, H₂ and CO₂ to acetic acid, and CO to acetic acid and ethanol (11, 12). A scheme of the energy metabolism of M. thermoacetica during growth with H₂ and CO₂ is given in Fig. 1.

FIG 1.

Proposed scheme of the energy metabolism of M. thermoacetica growing on H₂ and CO₂. Reactions in black are catalyzed by enzymes that have been characterized. Reactions in red are still under investigation. The numbers in the scheme correspond to Moth_ORF genome identification numbers (20). The yellow dot represents the site of flavin-based electron bifurcation. Two [FeFe]-hydrogenases (HydABC and HytA to -F) and probably one energy-converting [NiFe]-hydrogenase (EchABCE to -I) are involved in H₂ activation (19). The cytoplasmic formate dehydrogenase (FdhAB) has been shown to be a tungsten seleno iron-sulfur protein (23). FdhA2 shows 50% sequence identity to subunit FdhA of the cytoplasmic formate dehydrogenase FdhAB. H₄F, tetrahydrofolate.

The Wood-Ljungdahl pathway starts with CO₂ reduction to formate, which is subsequently activated with ATP to form N¹⁰-formyltetrahydrofolate (formyl-H₄F). The latter is cyclized to methenyl-H₄F, which is then reduced to methylene-H₄F and further to methyl-H₄F. Finally, from methyl-H₄F and CO, which is formed from CO₂ via reduction, acetyl coenzyme A (acetyl-CoA) is generated. The pathway ends with the formation of acetic acid from acetyl-CoA via acetyl phosphate in an ATP-generating reaction. In the acetokinase reaction, the same amount of ATP is formed as is consumed in the formyl-H₄F synthetase reaction (Fig. 1). Therefore, growth with H₂ and CO₂ is only possible if the pathway is coupled to electron transport phosphorylation. Of the redox reactions involved, only the reduction of methylene-H₄F to methyl-H₄F (E₀′, −200 mV) (13) with H₂ (E₀′, −414 mV) is exergonic enough to be coupled with energy conservation (14–16).

M. thermoacetica contains three hydrogenases for H₂ activation: (i) a cytoplasmic electron-bifurcating [FeFe]-hydrogenase (HydABC), which uses NAD⁺ plus ferredoxin as electron acceptors (17); (ii) a cytoplasmic [FeFe]-hydrogenase (HytA to -F) (18), which uses NADP as an electron acceptor (19); and (iii) a membrane-associated, energy-converting [NiFe]-hydrogenase (EchABCE to -I) connected via an iron-sulfur protein to a formate dehydrogenase (FdhA2) (20) (Fig. 1).

In M. thermoacetica, CO₂ reduction to CO proceeds with reduced ferredoxin as an electron donor (21, 22). The cytoplasmic formate dehydrogenase (FdhAB) (23) and cytoplasmic methylene-H₄F dehydrogenase (24) are NADP specific. The physiological electron donor for the cytoplasmic methylene-H₄F is not known. Cell extracts of M. thermoacetica do not catalyze the reduction of methylene-H₄F to methyl-H₄F with NADH or with NADPH. Methylene-H₄F reductase activity is only found with reduced benzyl viologen (19, 25).

M. thermoacetica contains cytochromes and menaquinone (26, 27). The two membrane-associated electron carriers have been proposed to be involved in electron transport during acetogenesis (14, 28, 29). This proposal was substantiated by the finding that during growth of M. thermoacetica on glucose and nitrate, acetogenesis and cytochrome b biosynthesis are repressed (30, 31). However, it was recently shown that despite its classification as a strict anaerobe, M. thermoacetica contains a membrane-bound cytochrome bd oxidase that can catalyze the reduction of low levels of dioxygen involving menaquinone and a high-potential cytochrome b (E₀′ ≈ −50 mV) in the electron transport chain to O₂ (25, 27, 32). A second, low-potential (E₀′ ≈ −210 mV) b-type cytochrome in M. thermoacetica appears to be associated with a putative “periplasmic” formate dehydrogenase, FdhA3 (20, 25, 27), which could also be a dimethyl sulfoxide reductase (32). These findings and the relatively high redox potentials of menaquinone (E₀′, −74 mV) and the two b-type cytochromes make it unlikely that these membrane-associated electron transport components are involved in CO₂ reduction with H₂ to acetic acid. The highest redox potential in this pathway is that of the methylene-H₄F/methyl-H₄F couple (E₀′, −200 mV) (33).

All enzymes involved in the Wood-Ljungdahl pathway in M. thermoacetica have been characterized to date, except for methylene-H₄F reductase (MetF). The MetF proteins from Escherichia coli (34), Thermus thermophilus (35), and Peptostreptococcus productus (now Blautia producta) (36) are composed of only one type of subunit, are NAD specific, and contain only flavin adenine dinucleotide (FAD) as the prosthetic group. The homologous reductases from mammals differ mainly in that the homomeric enzyme is allosterically regulated by S-adenosylmethionine and is NADP specific (37). The plant reductases are NADH dependent and S-adenosylmethionine insensitive (38). The E. coli type of methylene-H₄F reductase has been structurally well studied and has an NAD(P) and FAD binding site (39).

Bioinformatic analyses indicate that in some bacteria (e.g., M. thermoacetica and Acetobacterium woodii) and archaea (e.g., Pyrococcus furiosus) (40), the methylene-H₄F reductase differs greatly from that found in E. coli. In the genome of these microorganisms, a metF gene related to that in E. coli (about 20% identity at the protein level) is present, but this metF gene is tightly associated with a gene designated metV (41), which at the protein level in part shows sequence similarity to the N terminus of MetF. Otherwise, metV is predicted to code for an iron-sulfur zinc protein.

MetFV-type methylene-H₄F reductases have not yet been characterized. The only probable reductases of this type that have been purified are the enzymes from Clostridium formicaceticum (25, 42) and M. thermoacetica (25, 42). The purified methylene-H₄F reductase from C. formicoaceticum had an α₄β₄ structure, contains 2 FADs, 15 Fe's, 20 acid-labile sulfurs, and 2.3 zincs, and exhibited methylene-H₄F reductase activity with artificial dyes rather than with NAD or NADP. The apparent molecular masses of the α-subunit (35 kDa) and β-subunit (25 kDa) are in the range predicted for proteins encoded by metF and metV, respectively. Since neither the genome sequence of C. formicoaceticum nor the amino acid sequence of the α- and β-subunits of the purified reductase is presently available, the assignment remains tentative.

The methylene-H₄F reductase from M. thermoacetica was reported in 1991 to be composed of only one type of subunit with an apparent molecular mass 42.6 kDa, to have an α₈ structure, and to contain 34 Fe's and 2.3 FADs per homooctamer when purified in the presence of FAD. When purified in the absence of FAD, the enzyme was still active but did not contain FAD. The enzyme exhibited methylene-H₄F reductase activity only with artificial dyes (25, 42). The gene coding for the α-subunit was cloned and sequenced, and from the derived amino acid sequence (not published) a molecular mass of 36.143 kDa was calculated (25, 42), which is larger by more than 2 kDa than the mass of 33.924 kDa predicted for the metF gene product of M. thermoacetica (20).

We show here that the metVF genes in M. thermoacetica form a transcription unit together with the hdrCBA-mvhD genes. The six encoded proteins copurified in a complex with an apparent molecular mass in the 320-kDa range. The subunit MetF contained flavin mononucleotide (FMN) rather than FAD and catalyzed the reduction of methylene-H₄F with reduced benzyl viologen. The subunit MetV is an iron-sulfur zinc protein that is required by MetF for full activity. The subunit HdrA contains iron-sulfur clusters and 2 FADs and catalyzes the reduction of benzyl viologen with NADH.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

M. thermoacetica DSM 521 and Clostridium pasteurianum DSM 525 were obtained from the Deutsche Sammlung für Mikroorganismen und Zellkulturen, Braunschweig, Germany. Escherichia coli C41(DE3) harboring the plasmids pCodonPlus and pRKISC for optimal codon usage and iron-sulfur cluster production, respectively, was obtained from Nakamura and coworkers (Osaka University, Japan) (43).

M. thermoacetica was cultivated anaerobically at 55°C on glucose with 100% CO₂ as the gas phase as previously described (19). When the optical density at 660 nm (OD₆₆₀) was about 5, the cells were harvested by centrifugation under N₂ and were stored at −80°C until used. C. pasteurianum was grown anaerobically at 37°C on a glucose-ammonium medium (44), and the cells were harvested in the mid-exponential phase at an OD₆₆₀ near 5.

Biochemicals.

Tetrahydrofolate (H₄F), flavin adenine dinucleotide (FAD), flavin adenine mononucleotide (FMN), and benzyl viologen were from Sigma-Aldrich Chemie GmbH (Taufkirchen, Germany). Methyl-H₄F was from Schircks Laboratories (Jona, Switzerland). Methylene-H₄F was prepared from H₄F and formaldehyde in a spontaneous reaction. Ferredoxin (Fd), in the oxidized form, was purified from C. pasteurianum (45).

RNA isolation and cotranscription analysis by RT-PCR.

Total RNA was isolated from M. thermoacetica cells grown exponentially on glucose plus CO₂ and harvested at an OD₆₆₀ of 2.0. RNA was extracted with TRIzol reagent (Invitrogen, Darmstadt, Germany) following the manufacturer's protocol. After treatment with RNase-free DNase I (Fermentas, St. Leon-Rot, Germany), 3-μg samples of RNA were analyzed by electrophoresis in a 1.0% agarose gel to assess the integrity of the RNA. Cotranscription of the hdrCBA-mvhD-metVF genes and their neighboring genes was analyzed by amplification of the intergenic regions by reverse transcription-PCR (RT-PCR) and with the ProtoScript Moloney murine leukemia virus (M-MuLV) Taq RT-PCR kit (New England BioLabs, Frankfurt, Germany) as recommended by the manufacturer. The random primer provided in the kit was used for reverse transcription. The resulting cDNA was used as the template to amplify the intergenic region of every pair of neighboring genes. The specific primers used for amplification are listed in Table 1. Genomic DNA and total RNA were used as positive and negative controls, respectively (17).

TABLE 1.

Primers used for RT-PCR to check cotranscription of the methylene-H₄F reductase gene cluster hdrCBA-mvhD-metVF in M. thermoacetica

Primer	Sequence (5′→3′)	Application
9796-503bp-sense	AGTAATGTCTCCCAGAACTGCC	Amplification of intergenic region between methyltransferase gene and hdrC
9796-503bp-antisense	ATGACCGCAGTAACGGGACAA
9695-401bp-sense	AAGTTGCTGCCTTCCACCAG	Amplification of intergenic region between hdrC and hdrB
9695-401bp-antisense	GCAGGGCATGGCTCAAGAA
9594-448bp-sense	CCAACCGCAAATCGTCGAAG	Amplification of intergenic region between hdrB and hdrA
9594-448bp-antisense	TTATCCAGCTGGGCCATGTG
9493-438bp-sense	GGTGGTAGCCGTTGTCGATGA	Amplification of intergenic region between hdrA and mvhD
9493-438bp-antisense	ATGCAACCCAGGACGTAGACC
9392-432bp-sense	GTGTACGGCAGGCCAAGAAA	Amplification of intergenic region between mvhD and metV
9392-432bp-antisense	GGTCAGGGTGACTTCAACTGTT
9291-459bp-sense	GGTGGCAAATGCGAGGTCAA	Amplification of intergenic region between metV and metF
9291-459bp-antisense	TAGGCACCCAGGAGATCACT
9190-410bp-sense	GTCCCCGACGAAATCGTCAA	Amplification of intergenic region between metF gene and Moth_1190
9190-410bp-antisense	AATTGTTCTTCCCGGGCCGA

Enzyme activity assays.

Except where indicated, enzymes were assayed at 45°C in 1.5-ml anoxic cuvettes closed with a rubber stopper and filled with 0.8 ml reaction mixture and 0.7 ml N₂ or H₂ at 1.2 × 10⁵ Pa. The basal reaction mixture contained 100 mM Tris-HCl (pH 7.5), 20 mM ascorbate (25), 2 mM dithiothreitol (DTT), and 10 μM FAD. The reactions were monitored photometrically at the wavelength indicated. One unit was defined as the transfer of 2 μmol electrons per minute. Protein was determined with the Bio-Rad reagent using bovine serum albumin as standard (19).

The optimum growth temperature of M. thermoacetica is 55°C. For technical reasons, the enzyme activities were assayed at 45°C. At 55°C, the specific activities were twice as high as at 45°C (19).

Methylene-H₄F reduction with reduced benzyl viologen.

The basal reaction mixture was supplemented with 0.75 mM tetrahydrofolate, 10 mM formaldehyde, and 1 mM benzyl viologen. Methylene-H₄F is formed from tetrahydrofolate and formaldehyde in a spontaneous reaction (46). The gas phase was 100% N₂. Just before starting the reaction with enzyme, benzyl viologen was reduced to a ΔA₅₅₅ of ∼1.5 (the limit of the spectrophotometer used) with sodium dithionite. Reduced benzyl viologen oxidation was monitored at 555 nm (ε, 12 mM⁻¹ cm⁻¹) (47, 48).

Benzyl viologen reduction with methyl-H₄F.

The assay described by Clark and Ljungdahl (42) was modified as follows. The basal reaction mixture was supplemented with 20 mM benzyl viologen and 1 mM methyl-H₄F. The gas phase was 100% N₂. After prereduction of benzyl viologen to an ΔA₅₅₅ of 0.3 with sodium dithionite to secure completely anoxic conditions, the reactions were started by the addition of enzyme. Benzyl viologen reduction was monitored at 555 nm (ε, 12 mM⁻¹ cm⁻¹) (47, 48).

Methylene-H₄F reduction with NAD(P)H or reduced ferredoxin.

The basal reaction mixture was supplemented with 0.5 mM tetrahydrofolate, 10 mM formaldehyde, and 0.1 mM NADH or NADPH or 30 μM reduced ferredoxin (reduced to 80% with sodium dithionite). Methylene-H₄F is formed from tetrahydrofolate and formaldehyde in a spontaneous reaction (46). The gas phase was 100% N₂. After starting the reaction with enzyme, the reduction of NAD(P)⁺ was monitored at 350 nm (ε, 5.6 mM⁻¹ cm⁻¹) (49), or the oxidation of reduced ferredoxin was monitored at 430 nm (ε_Δox-red, ≈13.1 mM⁻¹ cm⁻¹) (17, 50).

Benzyl viologen reduction with NADH.

The basal reaction mixture was supplemented with 20 mM benzyl viologen and 2 mM NADH. After starting the reaction with enzyme, benzyl viologen reduction was monitored at 555 nm (ε, 12 mM⁻¹ cm⁻¹) (47, 48).

Reduced benzyl viologen oxidation with NAD⁺.

The basal reaction mixture was supplemented with 20 mM benzyl viologen, which was subsequently reduced with sodium dithionite to an ΔA₅₅₅ of 1.5. After the addition of 1 mM NAD, the reaction was started with enzyme, and reduced benzyl viologen oxidation was monitored at 555 nm (ε, 12 mM⁻¹ cm⁻¹) (25, 47).

H₂ formation in cell extracts.

The assay was performed in 6.5-ml serum bottles sealed with rubber stoppers and containing 1 ml reaction mixture and 5.5 ml N₂ at 1.2 × 10⁵ Pa. The basal reaction mixture contained 1 mM H₄F, 10 mM formaldehyde, 1 mM NADH, 4 μM ferredoxin from C. pasteurianum, and up to 6 mg cell extract protein. In control experiments, the basal reaction mixture contained 5 mM sodium pyruvate, 4 mM CoA, 1 mM NADH, 4 μM ferredoxin, and up to 6 mg cell extract protein. The reaction was started with enzyme, and the serum bottles were then continuously shaken at 200 rpm to ensure H₂ transfer from the liquid phase into the gas phase. A gas sample (0.2 ml) was withdrawn only once from each bottle 1, 2, 3, or 4 min after the start, and H₂ was quantified by gas chromatography. The gas chromatograph was equipped with a thermal conductivity detector (Carlo Erba GC series 6000) and a ShinCarbon ST micropacked column (82/100 mesh, 2.0 mm by 2 m) (Restek GmbH, Germany). The oven and injection port temperatures were set at 100°C; the detector was set at 150°C. N₂ was used as the carrier gas, and the flow rate was 34 ml/min. The amount of H₂ was calculated according to the standard curve correlated with peak areas (17). The detection limit was 50 nmol H₂ in the 5.5-ml gas phase.

Purification of methylene-H₄F reductase.

All purification steps were performed under strictly anoxic conditions at room temperature in a type B vinyl anaerobic chamber (Coy, Grass Lake, MI) filled with 95% N₂–5% H₂ and containing a palladium catalyst for O₂ reduction with H₂. All buffers were flushed with N₂, supplemented with 2 mM DTT, and maintained under a slight overpressure of N₂. During purification, reduction of benzyl viologen with methyl-H₄F and reduction of benzyl viologen with NADH were monitored.

Frozen wet cells of M. thermoacetica (18 g) were suspended at room temperature in 20 ml 50 mM Tris-HCl (pH 7.5) containing 2 mM DTT and 5 μM FAD (buffer A). Traces of DNase were added to the cell suspension before it was passed through a prechilled French pressure cell three times at 120 MPa. Unbroken cells, cell debris, and membranes were removed by ultracentrifugation at 115,000 × g at 4°C for 60 min. The 20-ml supernatant containing methylene-H₄F reductase activity and about 60 mg protein per ml was divided into two parts to avoid overloading the columns during the following steps.

The first half of the 115,000 × g supernatant (approximately 10 ml) was loaded onto a MonoQ column (1 × 10 cm) equilibrated with buffer A. MonoQ and the other materials for protein purification were from GE Healthcare, Freiburg, Germany. The column was washed with 2 column volumes of buffer A, and then protein was eluted with 20 column volumes of a 0 to 500 mM KCl linear gradient at a flow rate of 2 ml min⁻¹. The reductase activity eluted at 150 to 250 mM KCl. The fractions containing the activity were pooled and concentrated using an Amicon cell with a 50-kDa-cutoff membrane. The concentrate was loaded onto a Sephacryl S-300 gel filtration column (2.6 by 60 cm), which was equilibrated with 50 mM Tris-HCl (pH 7.5) containing 150 mM KCl, 2 mM DTT, and 5 μM FAD (buffer A2). The reductase activity eluted at 100 to 150 ml at a flow rate of 1.2 ml min⁻¹. The second half of the 150,000 × g supernatant was similarly subjected to MonoQ and Sephacryl S-300 chromatography.

The active Sephacryl S-300 fractions obtained with both halves of the 150,000 × g supernatant were pooled, concentrated to 32 ml using an Amicon cell with a 50-kDa-cutoff membrane, and supplemented with KH₂PO₄-K₂HPO₄ (pH 7.5) to a final concentration of 5 mM. The concentrated protein was loaded onto a hydroxyapatite type I 40-μm column (1.6 by 8 cm), which was equilibrated with 50 mM Tris-HCl (pH 7.5) containing 5 mM KH₂PO₄-K₂HPO₄ (pH 7.5), 150 mM KCl, 2 mM DTT, and 5 μM FAD (buffer A3). The column was washed with two column volumes of buffer A3, and then the protein was eluted with 20 column volumes of a linear gradient of 5 to 500 mM KH₂PO₄-K₂HPO₄ (pH 7.5) at a flow rate of 3 ml min⁻¹. The reductase activity eluted at about 75 mM KH₂PO₄-K₂HPO₄ (pH 7.5). Active fractions were pooled, concentrated using an Amicon cell with a 50-kDa-cutoff membrane, and divided into three parts. Each part was loaded onto a Superdex 200 Prep Grade (PG) column (1 by 28.5 cm), which was equilibrated with buffer A2. The reductase activity eluted at around 9 ml at a flow rate of 0.4 ml min⁻¹. Active fractions were pooled, glycerol was added to 10%, and samples were stored at −80°C.

Heterologous expression of metF, metV, metVF, and hdrA.

M. thermoacetica genomic DNA was extracted following the method described by Murray and Thompson (51). The genes were amplified by PCR with KOD Hot Start DNA polymerase (Merck, Darmstadt, Germany) using M. thermoacetica genomic DNA as a template. The following primers were used (with the inserted sequence specific for ligation-independent cloning underlined): hdrA, 5′-GACGACGACAAGATGAGTGCAAAGAAGGAAACTAC-3′ (forward) and 5′-GAGGAGAAGCCCGGTAACACCTCCCCGAAGAGG-3′ (reverse); metF, 5′-GACGACGACAAGATGGTTGAAAGCAAGATGGCG-3′ (forward) and 5′-GAGGAGAAGCCCGGTTACTCTACCTTGGGACGG-3′ (reverse); and metV, 5′-GACGACGACAATATGATTATCGCCGAAGGCAAAC-3′ (forward) and 5′-GAGGAGAAGCCCGGTCAACCATCGATTGTCACATCC-3′ (reverse). For the construction of a gene cassette containing both metF and metV, the forward metV primer and the reverse metF primer were used for amplification. After purification with the MinElute PCR purification kit (Qiagen, Hilden, Germany), the blunt-ended PCR product was treated with T4 DNA polymerase in the presence of dATP to generate specific vector-compatible overhangs. Then the PCR product was annealed to the linear pET-51b(+) Ek/LIC vector carrying a Strep-tag cassette with single-stranded complementary overhangs, which was subsequently introduced into NovaBlue GigaSingles (Merck, Darmstadt, Germany) competent cells by transformation. After amplification, the construct was verified by DNA sequencing. It was then introduced by transformation into E. coli C41(DE3) for expression, which already harbored pCodonPlus and pRKISC. pRKISC contains the E. coli isc locus (52) and has been successfully used for production of iron-sulfur proteins (43).

For expression, the cells were grown aerobically in 2 liters of Terrific Broth (TB) medium (24 g yeast extract, 12 g tryptone, and 4 ml glycerol per liter, with 0.17 M KH₂PO₄ and 0.72 M K₂HPO₄ as buffer) at 37°C and a high stirring speed (750 rpm). Before inoculation, the medium was supplemented with carbenicillin (50 mg/liter), chloramphenicol (25 mg/liter), and tetracycline (10 mg/liter) to maintain the plasmids, and with cysteine (0.12 g/liter), ferrous sulfate (0.1 g/liter), ferric citrate (0.1 g/liter), and ferric ammonium citrate (0.1 g/liter) to enhance the synthesis of iron-sulfur clusters. When an OD₆₀₀ of about 0.6 was reached, the stirring speed was lowered to 250 rpm. Concomitantly, the culture was supplemented with isopropyl-β-d-thiogalactoside (IPTG [0.5 mM]) to induce gene expression. After induction of the cells with IPTG, the cells were incubated at room temperature for 20 h with stirring at 250 rpm and then incubated for another 20 h at 4°C with stirring at 250 rpm before harvesting by centrifugation at 115,000 × g at 4°C. Subsequently, the recombinant E. coli cells (2 to 4 g wet mass per liter) were washed with anoxic 100 mM Tris-HCl (pH 7.5) and stored at −80°C under N₂ until used (19).

Purification of Strep-tagged MetF, MetV, MetFV, and HdrA.

The purification steps were performed at room temperature in an anaerobic chamber (Coy, Ann Arbor, MI) filled with 95% N₂–5% H₂ and containing a palladium catalyst for O₂ reduction with H₂. The E. coli cells (2 to 4 g wet mass) were resuspended in 20 ml 100 mM Tris-HCl (pH 7.5) containing 2 mM DTT and 10 μM FAD and disrupted by sonication (10 × 2 min at 32 W). Cell debris and membranes were removed by centrifugation at 115,000 × g and 4°C for 60 min. The supernatant was then loaded onto a 5 ml StrepTrap column, which was equilibrated with 20 ml buffer W (100 mM Tris-HCl [pH 7.5], 150 mM NaCl, 2 mM DTT, 10 μM FAD). Then the column was washed with 35 ml buffer W. The recombinant protein was eluted with 15 ml buffer E (buffer W containing 2.5 mM desthiobiotin). The fractions containing the target protein were pooled and concentrated by ultrafiltration with Amicon filters (50-kDa-cutoff; Millipore). The purified protein was washed with 100 mM Tris-HCl (pH 7.5) containing 2 mM DTT and 10 μM FAD and then stored at −20°C under N₂ until used.

Analytical methods.

Proteins were separated by SDS-PAGE on Mini-Protean TGX precast gels (any kDa) according to the manufacturer's manual (Bio-Rad, Munich, Germany), calibrated with PageRuler prestained protein ladder (Thermo Scientific, Rockford, IL). The proteins separated by SDS-PAGE were digested with sequencing-grade modified trypsin (Promega), and the resulting peptide mixtures were analyzed by peptide mass fingerprinting using a nano-high-pressure liquid chromatography (nano-HPLC) matrix-assisted laser desorption ionization–time of flight tandem mass-spectrometry (MALDI-TOF MS/MS) unit (4800 Proteomics analyzer; MDS Sciex). The mass spectrometry (MS) and tandem MS (MS/MS) data were searched against an in-house protein database using Mascot embedded into GPS Explorer software (MDS Sciex).

The stoichiometry of recombinant MetFV was estimated from a scan of the Coomassie brilliant-blue-stained SDS gel using ImageJ 1.46r software (Bethesda, MD).

The apparent molecular mass of the methylene-H₄F reductase preparation was determined by gel filtration on a Superose 6 column (1.0 by 30 cm), calibrated with a GE Healthcare high-molecular-weight (HMW) gel filtration calibration kit (GE Healthcare, Freiburg, Germany). The column was equilibrated with buffer A containing 150 mM NaCl and 10% glycerol and run at a flow rate of 0.5 ml/min.

The iron content of the purified enzymes was determined colorimetrically with 3-(2-pyridyl)-5,6-bis(5-sulfo-2-furyl)-1,2,4-triazinedisodium trihydrate (Ferene; Sigma) (53). The iron and zinc contents were determined by inductively coupled plasma mass spectrometry (ICP-MS) (Fachbereich Chemie, Philipps-Universität, Marburg, Germany). Buffer A2 or buffer E was used as the blank. For identification of flavin and measurement of the flavin content of the enzymes, the purified enzymes were first heat denatured in a thermocycler at 80°C for 10 min. The denatured protein was removed by centrifugation at 13,000 × g at 4°C for 10 min. The supernatant was subjected to high-performance liquid chromatography (HPLC) on a Kinetex 5-μm C₁₈ 100-Å (core shell) 250- by 4.6-mm column (Phenomenex, Aschaffenburg, Germany). FAD eluted with a retention time of 9.9 min, and FMN eluted with a retention time of 10.1 min: both were quantified with a diode array detector at a wavelength of 450 nm using commercially available FAD or FMN as the standard. Again, buffer A2 or buffer E was used as the blank.

Nucleotide sequence accession number.

The M. thermoacetica genome GenBank accession number is CP000232.1.

RESULTS

Cell extracts of M. thermoacetica catalyzed the reduction of benzyl viologen at specific initial rates of 0.4 μmol per min and mg protein (Table 2), indicating the presence of methylene-H₄F reductase. The reduction of benzyl viologen in the assays leveled off due to the reoxidation of the dye via the [FeFe]-hydrogenase present (17). Reduced benzyl viologen was also oxidized in the absence of methyl-H₄F or methylene-H₄F. This is why it was not possible to determine the specific activity of the reverse reaction, the reduction of methylene-H₄F with reduced benzyl viologen. The cell extracts did not catalyze the oxidation of NADH or NADPH in either the absence or presence of methylene-H₄F. An oxidation of NADPH by methylene-H₄F could have been overlooked because the cell extracts contained an active NADP-specific methylene-H₄F dehydrogenase (1.4 U/mg) that catalyzed the reduction of NADP⁺ with methylene-H₄F. We tested whether ferredoxin from C. pasteurianum was reduced by methylene-H₄F in the presence of NADH, which was not the case. Since the cell extracts rapidly catalyzed the oxidation of reduced ferredoxin (regenerated via pyruvate:ferredoxin oxidoreductase) in the presence of NADH due to the activity of the reversible NAD- and ferredoxin-dependent electron-bifurcating [FeFe]-hydrogenase (0.1 U/mg) (18), we also tested whether the cell extracts catalyzed the formation of H₂ in the presence of methylene-H₄F, NADH, and ferredoxin from C. pasteurianum, but H₂ formation at significant rates was not observed. In the H₂ formation assays, very high cell extract concentrations (6 mg protein/ml) were employed so that most likely also ferredoxins I and II from M. thermoacetica were present in sufficient concentrations (54, 55).

TABLE 2.

Partial purification of methylene-H₄F reductase from M. thermoacetica

Purification step	Amt of protein (mg)	Activitya (U)	Sp acta (U/mg)	Yield (%)	Enrichment factor	Diaphorase activity (U/mg)b
Cell extract	1,170	413	0.4	100	1	1.5
MonoQ	84	115	1.4	28	3.5	7
Sephacryl S-300	39	83	2.1	20	5	23
Hydroxyapatite	10	79	7.9	19	20	60
Superdex 200 PG	4	64	16	15	40	48

^aMethylene-H₄F reductase activity was monitored by measuring the reduction of benzyl viologen with methyl-H₄F.

^bDiaphorase activity was monitored by measuring the reduction of benzyl viologen with NADH.

Genes encoding methylene-H₄F reductase.

The genome of M. thermoacetica harbors only one gene related to the gene metF from E. coli. This metF gene is directly downstream from a metV gene, and only one copy of this gene pair is present.

The metVF genes in M. thermoacetica are flanked upstream by the hdrCBA and mvhD genes, forming a cluster that could be a transcription unit (Fig. 2). The open reading frame (ORF) (Moth_1197) upstream of the cluster is predicted to encode a methyltetrahydrofolate:corrinoid methyltransferase, and the ORF (Moth_1190) downstream of the cluster is predicted to encode a protein of unknown function. Upstream of the hdrCBA-mvhD-metVF cluster, a putative promoter sequence and downstream of the cluster putative terminator and promoter sequences were identified using the software tools Sofberry (Sofberry, Inc., New York, NY) and Arnold (Institut de Génétique et Microbiologie—UMR CNRS 8621, Université Paris-Sud, France).

FIG 2.

Genome region of M. thermoacetica putatively coding for methylenetetrahydrofolate reductase. The numbers in the arrows correspond to Moth_ORF genome identification numbers (20). MetF shows 20% sequence identity to the homomonomeric methylene-H₄F reductase (MetF) from E. coli. MetV (224 amino acids) shows 33% sequence identity to about 40 amino acids of the N terminus of MetF (306 amino acids) (20). hdrCBA and mvhD code for proteins with sequence similarity to heterodisulfide reductase HdrABC-MvhD from methanogenic archaeae (HdrA,15%; HdrB, 28%; HdrC, 22%; MvhD, 46%). The predicted cofactor content was deduced from the amino acid sequences of the proteins. Bioinformatics indicated that hdrCBA, mvhD, and metVF form a transcription unit, which was confirmed by reverse transcription (RT)-PCR analysis. After reverse transcription, cDNA (third lane) was used as the template to amplify the intergenic regions. Genomic DNA (first lane) and total RNA (second lane) were used as positive and negative controls, respectively. CODH, carbon monoxide dehydrogenase; ACS, acetyl-CoA synthase; bp, base pairs.

Within the cluster, no such sequences are apparent. RT-PCR analysis indicated that the six genes indeed form a transcription unit (Fig. 2). There is a minor band between metF and Moth_1190 that is difficult to explain.

The hdrCBA and mvhD genes (20) are predicted to encode proteins with sequence similarity to the cytoplasmic heterodisulfide reductase HdrABC-MvhD from methanogenic archaea (56). The hdrA gene product (163 kDa) is, however, twice as large as HdrA in methanogens (72 kDa) and is predicted to contain two flavin binding sites rather than one. The overall sequence identity to HdrA from Methanothermobacter marburgensis is low (15% identity). In methanogens, HdrA is assumed to harbor the site of flavin-based electron bifurcation and a site for ferredoxin reduction. The hdrB, hdrC, and mvhD gene products (31, 21, and 17 kDa, respectively) are similar in size to HdrB (33 Da), HdrC (21 kDa), and MvhD (16 kDa) from methanogens, but the sequence identities differ for each subunit: 28% for HdrB, 22% for HdrC, and 46% for MvhD. In methanogens, HdrB is assumed to harbor the site in which the heterodisulfide coenzyme M (CoM)-S-S-coenzyme B (CoB) is reduced, and HdrC and MvhD are assumed to be iron-sulfur proteins conducting electrons from the [NiFe]-hydrogenase MvhAB via HdrA to HdrB (16, 57).

The genome of M. thermoacetica does not contain the genes comC, comD, and comE specific for the biosynthesis of coenzyme M (CoM) (56). The presence of genes for coenzyme B (CoB) biosynthesis cannot be excluded because the three enzymes known to be involved are homologs of isopropylmalate synthase (LeuA), isopropylmalate isomerase (LeuC/D), and isopropylmalate dehydrogenase (LeuB) present in M. thermoacetica (56). However, since coenzyme B has never been detected in bacteria, it is very unlikely that HdrB in the acetogen catalyzes the reduction of CoM-S-S-CoB. Still, an unknown disulfide could be involved in methylene-H₄F reduction.

The hdrCBA-mvhD-metVF cluster (Moth_1191 to -1196) is found in a genome region in which other genes involved in the Wood-Ljungdahl pathway are also located. Upstream are the genes coding for methyl-H₄F:corrinoid/iron-sulfur protein methyltransferase (Moth_1197), acetyl-CoA decarbonylase/synthase subunits α, β, γ, and δ (Moth_1198, -1201, and -1202), cobyrinic acid a,c-diamide synthase (Moth_1199, and -1204), an uncharacterized metal-binding protein (Moth_1200), and carbon-monoxide dehydrogenase (Moth_1203) (Fig. 2). The genes coding for formate dehydrogenase (Moth_2312 and -2314), formyl-H₄F synthetase (Moth_0109), and bifunctional methenyl-H₄F cyclohydrolase/methylene-H₄F dehydrogenase (Moth_1516) are located elsewhere.

Partial purification of MetFV in complex with HdrABC and MvhD.

Methylene-H₄F reductase activity was purified from cell extracts of M. thermoacetica grown on glucose and CO₂. The purification buffers all contained 5 μM FAD, which appeared to be required for stabilization of methylene-H₄F reductase activity. The 40-fold purification to a specific activity of 16 U/mg involved anion-exchange chromatography on MonoQ, size exclusion chromatography on Sephacryl, hydroxyapatite chromatography, and size exclusion chromatography on Superdex 200 PG (Table 2). The methylene-H₄F reductase activity eluted from the Superose 6 column with an apparent molecular mass of about 320 kDa (Fig. 3). SDS-PAGE and matrix-assisted laser desorption ionization–time of flight (MALDI-TOF) MS analysis of the single protein bands revealed that the proteins MetF, MetV, HdrA, HdrB, HdrC, and MvhD were present in the preparation (Fig. 4). The masses of these proteins add up to 280 kDa, consistent with the apparent molecular mass of 320 kDa. The results indicated that the six proteins form a hexaheteromeric complex.

FIG 3.

Apparent molecular mass of methylene-H₄F reductase determined by gel filtration on Superose 6. The flow rate was 1 ml/min. mAU, milliabsorbance units. The inset shows the calibration curve. The solid circles show size standards: thyroglobulin, 669 kDa; ferritin, 440 kDa; aldolase, 158 kDa; and ovalbumin, 43 kDa. The open circle shows the position of the methylene-H₄F reductase.

FIG 4.

SDS-PAGE of methylene-H₄F reductase partially purified from M. thermoacetica. Lane 1, molecular mass markers; lane 2, partially purified methylen-H₄F reductase complex. The proteins were identified by MALDI-TOF MS analysis of the protein bands after digestion with trypsin. The molecular masses were deduced from the amino acid sequences. The five other proteins identified after digestion were as follows: pyruvate:ferredoxin oxidoreductase (POR) (Moth_0064), copper amine-oxidase-like (amine oxidase) (Moth_2343), GroEL (a molecular chaperonin) (Moth_0546), ATP-dependent protease (Moth_1030), and LacI/GalR, transcriptional regulator (Moth_2022). (See the text for details.)

The hexaheteromeric complex was, however, still contaminated with at least five other proteins, which were identified after digestion with trypsin by mass spectrometry via their amino acid sequence as pyruvate:ferredoxin oxidoreductase (number of peptides, 53; protein score, 795; no other hit), copper amine oxidase-like (number of peptides, 45; protein score, 956; no other hit), the molecular chaperonin GroEL (number of peptides, 44; protein score, 1,060; no other hit), ATP-dependent protease (number of peptides, 31; protein score, 566; best hit), and a LacI/GalR-transcriptional regulator (number of peptides, 20; protein score, 550; best hit) (Fig. 4). The contaminating enzyme complexes were not separated from the MetFV-HdrABC-MvhD complex by size exclusion chromatography, indicating that they had similar sizes. For example, pyruvate:ferredoxin oxidoreductase from M. thermoacetica is known to be a homodimer with an apparent molecular mass of 255 kDa (58). Unfortunately, all attempts to separate these from the complex failed either because the yields were too low or because the steps employed (hydrophobic chromatography on Source 15PHE [GE Healthcare Life Sciences], butyl sepharose, and octyl sepharose) led to a dissociation of the hexaheteromeric complex.

In Table 2, the copurification of a diaphorase activity catalyzing the reduction of benzyl viologen with NADH is also documented because, as will be shown below, one of the subunits, namely, HdrA, exhibited diaphorase activity. The diaphorase activity was purified 32-fold with 9.5% yield. The lower purification factor relative to that of methylene-H₄F reductase activity (40-fold) can be explained by the fact that cell extracts of M. thermoacetica contained at least one enzyme, namely, the ferredoxin- and NAD-dependent [FeFe]-hydrogenase, with diaphorase activity (17), which was separated from the MetFV-HdrABC-MvhD complex during purification.

The enzyme preparation catalyzed the oxidation of reduced benzyl viologen in the absence of methylene-H₄F, probably using dehydroascorbate as the electron acceptor. The assay mixture contained 20 mM ascorbate for O₂ removal. When ascorbate was omitted, reduced benzyl viologen was not oxidized. Reduced benzyl viologen oxidation could be completely inhibited by 2 mM cyanide, which indicated that the activity could be due to the copper amine oxidase (ascorbate oxidase) (59) present as contamination in the preparation (Fig. 4).

Catalytic properties of MetFV in complex with HdrABC and MvhD.

The hexaheteromeric enzyme exhibited two activities: the reversible reduction of methylene-H₄F to methyl-H₄F with reduced benzyl viologen (16 U/mg) (Tables 2 and 3) and the reduction of benzyl viologen with NADH (48 U/mg) (Tables 2 and 3). The apparent K_ms for methylene-H₄F (1 mM) reduction with reduced benzyl viologen (0.15 mM) were 1.8 mM for methylene-H₄F and 0.1 mM for reduced benzyl viologen. The apparent K_ms for benzyl viologen (20 mM) reduction with NADH (1 mM) were 0.2 mM for NADH and 14 mM for benzyl viologen. Because of the high apparent K_m for benzyl viologen of the diaphorase activity, the addition of small amounts of benzyl viologen to the enzyme assay mixtures did not promote methylene-H₄F reduction with NADH. Interestingly, the reduction of benzyl viologen with NADH was competitively inhibited by NADP⁺. An apparent K_i for NADP⁺ of 0.3 mM was determined (not shown).

TABLE 3.

Properties of the methylene-H₄F reductase preparation from M. thermoacetica and of heterologously produced MetFV, MetF, MetV, and HdrA from M. thermoacetica in E. coli^a

Enzyme	Methylene-H4F reduction with BVb (U mg−1)	Methyl-H4F oxidation with BV (U mg−1)	BV reduction with NADH (U mg−1)	Content [(nmol mg−1)/(mol/mol protein)] of:
				FAD	FMN	Iron
Methylene-H4F reductase preparation	18c	16.3	48	3.1/0.9	3.4/1.0	110/30
MetFV	50	50	NDd	ND	18/1.0	165/9
MetF	0.7	0.6	ND	ND	33/1.1	ND
MetV	ND	ND	ND	ND	ND	160/3.8
HdrA	ND	ND	190	18/2.9	ND	125/20

^aThe assays were performed at least twice, and results deviated by less than 10%.

^bBV, benzyl viologen.

^cMeasured in the presence of 2 mM cyanide for inhibition of contaminating copper amine oxidase activity (see the text).

^dND, not detectable.

The enzyme complex did not catalyze the reduction of methylene-H₄F with NADH or NADPH in either the absence or presence of oxidized ferredoxin. The activity in the presence of oxidized ferredoxin was tested because it has been proposed that the exergonic reduction of methylene-H₄F with NADH is coupled with the endergonic reduction of ferredoxin with NADH via flavin-based electron bifurcation (15). The enzyme complex also did not catalyze the reduction of CoM-S-S-CoB, cystine, or oxidized lipoic acid with reduced benzyl viologen or the reduction of methylene blue with coenzyme M plus coenzyme B or cysteine or glutathione. The reduction of CoM-S-S-CoB with reduced benzyl viologen and the reduction of methylene blue with coenzyme M plus coenzyme B are reactions catalyzed by the HdrABC complex from methanogenic archaea (60).

Iron and flavin contents of the methylene-H₄F reductase preparation.

From the primary structures, we predicted that MetV harbors one [4Fe4S] cluster and one zinc, HdrA harbors four [4Fe4S] clusters, HdrB harbors one [4Fe4S] cluster and one zinc, HdrC harbors two [4Fe4S] clusters, and MvhD harbors one [2Fe2S] cluster (Fig. 2). The complex should therefore contain 34 irons per mol enzyme with a mass of 280 kDa and 121 nmol iron per mg protein. We found the preparation to contain 110 nmol iron per mg protein, almost the same amount of iron reported by Park et al. in 1991 (25) for the homooctameric methylene-H₄F reductase from M. thermoacetica (see the introduction) As indicated above, the preparation obtained after four chromatographic steps (Table 2) was still contaminated with several proteins, one being pyruvate:ferredoxin oxidoreductase (Fig. 4), which is a 120-kDa iron-sulfur protein containing three [4Fe4S] clusters and with an iron content of 76 nmol per mg protein (61). The other contaminating proteins in the preparation (GroEL, copper amine oxidase, and the LacI/GalR-transcriptional regulator) are not known to contain iron.

In addition to iron, we determined the FAD and FMN contents to be 3.1 nmol FAD and 3.4 nmol FMN per mg protein, which indicated that the enzyme complex of 280 kDa contains both FAD and FMN, each at about 1 mol per mol (Table 3). The content is underestimated because of the contaminating proteins in the preparation (Fig. 4) that are not flavoproteins. The presence of FAD can be explained by HdrA from methanogens having one FAD. However, the finding of FMN was unexpected. As will be shown below, heterologously produced MetF and MetFV from M. thermoacetica contain FMN rather than FAD.

Purification of heterologously produced MetF, MetV, MetFV, and HdrA.

The subunits MetF, MetV, MetFV, and HdrA were heterologously produced in E. coli C41(DE3) harboring the plasmids pCodonPlus for optimal rare codon usage and pRKISC for the production of iron-sulfur proteins. The N-terminally Strep-tagged fusion proteins were purified on a StrepTrap column. The protein yields from a 1-liter E. coli culture were 1.4 mg MetF, 42 μg MetV, 50 μg MetFV, and 300 μg HdrA. SDS-PAGE revealed only one band of MetF with an apparent molecular mass of 34 kDa, only one band of MetV with an apparent molecular mass of 24 kDa, and two bands of MetFV with apparent molecular masses of 34 and 24 kDa (Fig. 5). Scans of the two Coomassie brilliant-blue-stained proteins showed that MetF and MetV were present in an almost 1:1 molar ratio. HdrA migrated in three bands: the largest at 180 kDa and the others at 100 and 70 kDa (Fig. 5). MALDI-TOF MS analysis indicated that the two bands of lower molecular mass were proteolysis products of HdrA. These products were not removed before being assayed for activity.

FIG 5.

SDS-PAGE of Strep-tagged MetF, MetV, MetFV, and HdrA from M. thermoacetica heterologously produced in E. coli. Lane 1, molecular mass markers. Lanes MetF, MetV, MetFV and HdrA contained about 3, 2, 3 and 1 μg protein, respectively. The molecular masses of the proteins were calculated from amino acid sequences deduced from the encoding genes. The 70-kDa protein in lane MetV is the Hsp70 heat shock protein from E. coli, as identified by MALDI-TOF MS analysis of the protein band after digestion with trypsin.

Catalytic properties and cofactor content of MetF, MetV, and MetFV.

Recombinant MetF from M. thermoacetica catalyzed the reversible reduction of methylene-H₄F with benzyl viologen. For methyl-H₄F oxidation with benzyl viologen, a specific activity of 0.6 U/mg was measured (Table 3). When tested alone, MetV showed no activity in the methylene-H₄F reductase assay, but MetV stimulated the reductase activity of MetF up to 2-fold when added to the assay. Stimulation was highest when 1 mol MetV was added per mol MetF (result not shown).

Much higher specific activities were observed when the metF and metV genes were expressed together. The purified MetFV complex catalyzed the reduction of benzyl viologen with methyl-H₄F at a specific activity of 50 U/mg and the reduction of methylene-H₄F with reduced benzyl viologen also at a specific activity of 50 U/mg (Table 3). Methylene-H₄F reduction followed Michaelis-Menten kinetics, with an apparent K_m for methylene-H₄F of 1 mM and one for reduced benzyl viologen of about 0.1 mM. The apparent K_m values were on the same order as those determined for the complex of MetFV with HdrABC and MvhD purified from M. thermoacetica. The specific activity of recombinant MetFV was about 3-fold higher than that of the partially purified complex, in agreement with the complex containing the additional proteins HdrABC and MvhD and the preparation of the complex containing some contaminating proteins (Fig. 4). One difference between recombinant MetFV and MetFV in complex with HdrABC and MvhD appeared to be that the activity of recombinant MetFV was lost much more rapidly upon storage at 4°C and when in contact with traces of molecular oxygen.

Just like MetFV in complex with HdrABC and MvhD, recombinant MetF and MetFV did not catalyze the reduction of methylene-H₄F with NADH or NADPH. Sequence comparisons of MetF from M. thermoacetica and MetF from E. coli revealed that two of the amino acids, F28 and T57, important for NAD binding in MetF of E. coli (34) are exchanged with aspartate and arginine, respectively, in MetF of M. thermoacetica. The exchange of these amino acids appears to be conserved and to be a general property of the MetFV type of methylene-H₄F reductases.

Surprisingly, recombinant MetF and MetFV contained FMN rather than FAD (Table 3), at 1.1 mol per mol MetF and 1.0 mol per mol MetFV. No flavin was detected in the recombinant MetV preparation. Sequence comparison of MetF from E. coli and MetF from M. thermoacetica revealed that the two proteins differ in the flavin binding site by only one amino acid. The E. coli protein contains an arginine at position 118 (R118), whereas the M. thermoacetica protein contains a serine. In MetF of E. coli, amino acid R118 interacts with the AMP moiety of FAD (34). Exchange of R118 with serine is predicted to favor FMN binding, as indicated by in silico mutagenesis with PyMOL software (PyMOL Molecular Graphics System, version 1.3; Schrödinger, LLC) on PDB accession no. 1ZP3.

We routinely purified the proteins in the presence of FAD. Since FMN can be formed from FAD via hydrolysis, we repeated the purification of recombinant MetF and MetFV in the absence of FAD. Even in these preparations, only FMN was found. The FMN appears to be very tightly bound, since addition of FMN had no effect on the reductase activity. Interesting in this respect is the finding by Park et al. in 1991 (25) that their homooctameric methylene-H₄F reductase preparation from M. thermoacetica did not contain FAD when purified in the absence of FAD. However, these authors did not report whether they looked for the presence of FMN (25).

Recombinant MetV contained 3.8 mol iron per mol protein, while recombinant MetFV contained 9 mol iron per mol protein (Table 3), which cannot be explained since, in agreement with structural predictions, the MetF preparations contained only background amounts of iron.

The sequence of the middle of MetV from M. thermoacetica shows similarity (33% identity) to a sequence of 40 amino acids in the N-terminal region of MetF from M. thermoacetica. In MetF from E. coli, this sequence is involved in both FAD and NAD binding, but the sequence does not contain all of the residues required to bind the two cofactors, as revealed from the crystal structure (34).

Catalytic properties and cofactor content of HdrA.

Recombinant HdrA from M. thermoacetica catalyzed the reduction of benzyl viologen with NADH at a specific activity of 190 U/mg (Table 3). Apparent K_m values of 0.4 mM for NADH and 20 mM for oxidized benzyl viologen were measured. NADP⁺ was an inhibitor that competed with NADH (K_i, 0.6 mM). The catalytic properties were therefore very similar to those of the HdrABC-MvhD-MetFV complex purified from M. thermoacetica.

Recombinant HdrA contained 2.9 mol FAD per mol protein, no FMN, and 20 mol nonheme iron per mol protein (Table 3). The cofactor content is somewhat higher than structurally predicted (Fig. 2).

DISCUSSION

The genome of M. thermoacetica encodes only one gene product with sequence identity (about 20%) to the methylene-H₄F reductase MetF from E. coli. MetF from M. thermoacetica differed from the E. coli enzyme in that it contained FMN rather than FAD, and it did not catalyze the reduction of methylene-H₄F with NADH. It also differed from the E. coli enzyme in that for full reductase activity, the iron-sulfur protein MetV was required. The genome of E. coli is devoid of a metV homologous gene. The MetF proteins from M. thermoacetica (25) and from E. coli (62) are similar in that both enzymes can catalyze the oxidation of methyl-H₄F with one electron acceptor, such as menadione.

MetFV from M. thermoacetica formed a genetic and functional complex with the proteins HdrABC and MvhD (Fig. 2). Homologues of these proteins in methanogenic archaea are involved in heterodisulfide reduction with H₂ or formate and in coupling these exergonic reactions with the endergonic reduction of ferredoxin with H₂ or formate, presumably via flavin-based electron bifurcation (16, 56, 57, 63–65). HdrA from M. thermoacetica differed from HdrA in methanogens in that it contained two FADs rather than one and exhibited NADH diaphorase activity (Table 3). Nevertheless, the complex of MetFV with HdrABC and MvhD did not catalyze the reduction of methylene-H₄F with NAD(P)H in either the absence or presence of ferredoxin from C. pasteurianum. It is very unlikely that activity would have been found with ferredoxin I or II from M. thermoacetica (54, 55), because there is no case known to date in which ferredoxins from different organisms could not substitute for another, albeit sometimes with lower activity. Consistent with this interpretation is our finding that cell extracts, even at high protein concentrations, did not catalyze the formation of H₂ from methylene-H₄F in the presence of NADH, although under these conditions the cell extracts catalyzed the formation of H₂ from pyruvate in the presence of NADH and CoA (see the first paragraph of the Results section).

What could the physiological electron donor or second electron acceptor of methylene-H₄F reduction in M. thermoacetica then be? It is predicted that the gene cluster Moth_2184 to Moth_2194 in the genome of M. thermoacetica encodes an energy-conserving [NiFe]-hydrogenase (EchABCE to -I) (Moth_2184 to -2191) in complex with a formate dehydrogenase FdhA2 (Moth_2193) (20) (Fig. 1), similar to the formate-hydrogen lyase complex from E. coli (66–68). As in the E. coli complex, the different subunits of the Ech hydrogenase appear to be electrically connected with the formate dehydrogenase (FdhA2) via a membrane-associated electron transfer protein containing two [4FeS] clusters: i.e., HycB in E. coli and the protein encoded by Moth_2192 in M. thermoacetica (FeS protein in Fig. 1). These two electron transfer proteins have 49% sequence identity. It is therefore conceivable that in M. thermoacetica, the cytoplasmic methylene-H₄F reductase complex is somehow directly docked to the membrane-associated formate-hydrogen lyase complex. This could allow coupling of the exergonic reduction of methylene-H₄F with NADH with the endergonic reduction of protons to H₂. Since proton reduction via the Ech hydrogenase complex is associated with the buildup of a proton-motive force, the latter would drive the synthesis of ATP via the membrane-associated ATP synthetase. The H₂ formed could subsequently be recycled by one of the other hydrogenases (Fig. 1), as has been proposed for H₂ in the energy metabolism of Methanosphaera stadtmanae, which grows with methanol and H₂ as the sole energy sources (63). In this respect, it is of interest that methylene-H₄F reductase activity in M. thermoautotrophica appears to be membrane associated (28) and that the membrane fraction from this acetogen can be energized by reduced ferredoxin regenerated via CO oxidation (14, 29).

Genes encoding HdrABC and MvhD homologues are found not only in methanogens and M. thermoacetica but also in many other anaerobic bacteria and archaea, notably in sulfate-reducing bacteria and in Archaeoglobus sp. In dissimilatory sulfate reduction, a function in coupling of cytoplasmic redox reactions to membrane-associated redox reactions has been proposed (16, 69). The HdrABC complex appears to be an electron-bifurcating module used over and over again in different contexts, as is the EtfAB complex (70), which is used in many anaerobes as an electron-bifurcating module in crotonyl-CoA reduction (71), caffeyl-CoA reduction (72), and lactate oxidation (73). This hypothesis still remains to be experimentally verified.

Besides M. thermoacetica, A. woodii is used as model organism to study energy conservation in acetogenic bacteria. Both Gram-positive bacteria can grow with sugars or with H₂ and CO₂, but they differ considerably in many properties, such as the presence of cytochromes and menaquinone (M. thermoacetica) (26, 27), the presence of the transhydrogenase NfnAB (M. thermoacetica) (19), and the dependence on sodium ions (A. woodii) (74). The two organisms have different membrane-associated enzyme complexes required for conserving energy during growth with H₂ and CO₂, as revealed from genome sequences (20, 75). A. woodii involves a membrane-associated sodium ion-translocating reduced ferredoxin:NAD oxidoreductase (RnfA to -G) (75, 76), which is lacking in M. thermoacetica. On the other hand, it is predicted that M. thermoacetica contains a membrane-associated, energy-converting [NiFe]-hydrogenase (EchABCE to -I) in complex with formate dehydrogenase (FdhA2) (20) (Fig. 1), which is not found in A. woodii. CO₂ is reduced to formate with NADPH in M. thermoacetica (23) and with H₂ or reduced ferredoxin in A. woodii (77).

In the genome of A. woodii, the metFV genes for methylene-H₄F reductase are clustered with a gene, rnfC2, coding for an NADH dehydrogenase, which points to methylene-H₄F reduction with NADH (75). Consistently, cell extracts of A. woodii show methylene-H₄F reductase activity with NADH (78). In the genome of M. thermoacetica (20), which lacks rnf genes, the metVF genes are located together with the genes hdrCBA-mvhD in a transcription unit (Fig. 2). In the genome of A. woodii, the genes hdrBC and mvhD are lacking, and an hdrA gene homologue is located far away from the metVF genes.

In the genome of the acetogens Clostridium ljungdahlii (15) and Clostridium autoethanogenum (79), which like A. woodii lack cytochromes and menaquinone but are not sodium ion dependent, the metVF genes are associated with neither an rnfC2 gene nor with hdrABC or mvhD genes. Cell extracts of C. autoethanogenum do not catalyze the reduction of methylene-H₄F with NADH or NADPH (18). Methylene-H₄F reductase activity is found, as in cell extracts of M. thermoacetica, only with benzyl viologen. In A. woodii (78) and Clostridium formicaceticum (80), the methylene-H₄F dehydrogenase is NAD specific, and in M. thermoacetica (24, 81) and Peptostreptococcus productus (24, 81), the dehydrogenase is NADP specific. Since in vivo the redox potential E′ of the NAD⁺/NADH couple is −280 mV and that of the NADP⁺/NADPH couple is −370 mV (82) and the E₀′ of the methenyl-H₄F/methylene-H₄F couple is −300 mV (33), it is predicted that the NAD specificity of the methylene-H₄F dehydrogenase affects the thermodynamics of the methylene-H₄F reductase reaction via a lower in vivo methylene-H₄F concentration. All of these findings indicate that although all of these different acetogens use the Wood-Ljungdahl pathway of CO₂ reduction to acetic acid, they have evolved different ways to directly or indirectly conserve the energy associated with methylene-H₄F reduction.

Where do we go from here? In our opinion, all that could be done by biochemical analysis of M. thermoacetica has been done. A genetic analysis is now required. Thus, the question of whether cytochromes are involved in CO₂ reduction to acetic acid or whether they are only required to protect the organism from oxygen stress (32) will only be clarified by deleting genes essential for heme biosynthesis or genes for individual cytochromes. Fortunately, M. thermoacetica can grow on a broad spectrum of substrates with and without producing acetic acid from CO₂ (30, 31). Once a genetic system for M. thermoacetica is in hand, this and other open questions can successfully be addressed. In several laboratories, the genetic tools are already being developed (83, 84).