Formate-Dependent H₂ Production by the Mesophilic Methanogen Methanococcus maripaludis

Abstract

Methanococcus maripaludis, an H₂- and formate-utilizing methanogen, produced H₂ at high rates from formate. The rates and kinetics of H₂ production depended upon the growth conditions, and H₂ availability during growth was a major factor. Specific activities of resting cells grown with formate or H₂ were 0.4 to 1.4 U·mg⁻¹ (dry weight). H₂ production in formate-grown cells followed Michaelis-Menten kinetics, and the concentration of formate required for half-maximal activity (K_f) was 3.6 mM. In contrast, in H₂-grown cells this process followed sigmoidal kinetics, and the K_f was 9 mM. A key enzyme for formate-dependent H₂ production was formate dehydrogenase, Fdh. H₂ production and growth were severely reduced in a mutant containing a deletion of the gene encoding the Fdh1 isozyme, indicating that it was the primary Fdh. In contrast, a mutant containing a deletion of the gene encoding the Fdh2 isozyme possessed near-wild-type activities, indicating that this isozyme did not play a major role. H₂ production by a mutant containing a deletion of the coenzyme F₄₂₀-reducing hydrogenase Fru was also severely reduced, suggesting that the major pathway of H₂ production comprised Fdh1 and Fru. Because a Δfru-Δfrc mutant retained 10% of the wild-type activity, an additional pathway is present. Mutants possessing deletions of the gene encoding the F₄₂₀-dependent methylene-H₄MTP dehydrogenase (Mtd) or the H₂-forming methylene-H₄MTP dehydrogenase (Hmd) also possessed reduced activity, which suggested that this second pathway was comprised of Fdh1-Mtd-Hmd. In contrast to H₂ production, the cellular rates of methanogenesis were unaffected in these mutants, which suggested that the observed H₂ production was not a direct intermediate of methanogenesis. In conclusion, high rates of formate-dependent H₂ production demonstrated the potential of M. maripaludis for the microbial production of H₂ from formate.

Many hydrogenotrophic methanogens use H₂ or formate for the reduction of CO₂ to obtain energy for growth. Methanococcus maripaludis, the model microorganism in this study, is a hydrogenotrophic, formate-utilizing, mesophilic methanogen. It is common in salt marsh sediments, from which it was isolated (12). An extraordinarily active H₂ consumer, M. maripaludis is exceptionally well equipped with enzymes responsible for H₂ metabolism. M. maripaludis contains genes for seven different hydrogenases, whose expression depends upon the growth conditions (18). It possesses two membrane-bound, energy-converting [Ni-Fe] hydrogenases, designated Eha and Ehb, that are involved in the reduction of low-potential ferredoxins for anabolism (16, 26). There are also four cytoplasmic [Ni-Fe] hydrogenases, including two coenzyme F₄₂₀-reducing (Fru and Frc) and two coenzyme F₄₂₀-nonreducing (Vhu and Vhc) hydrogenases. One hydrogenase of each type (Fru and Vhu) contains a selenocysteinyl residue. The other hydrogenases (Frc and Vhc) contain cysteinyl residues at homologous positions (18). The selenocysteine-containing isozymes are abundant during cultivation in medium containing selenium, and the cysteine-containing isozymes (Frc and Vhc) are produced only upon selenium limitation (27). Lastly, the cells contain a cytoplasmic [Fe-S] cluster-free hydrogenase, the H₂-forming methylenetetrahydromethanopterin (methylene-H₄MPT) dehydrogenase (Hmd), which in other species has been shown to play an important role at high levels of H₂ or under nickel limitation (1, 2, 29).

When formate is the substrate, it is oxidized for the reduction of CO₂ to methane. The key enzyme for formate utilization is formate dehydrogenase, Fdh. The genome of M. maripaludis harbors two sets of genes encoding Fdh, fdhA1B1 and fdhA2B2 (33). Both Fdhs contain selenocysteinyl residues. While fdhA1B1 are found in an apparent operon with genes for a putative formate transporter and carbonic anhydrase, fdhA2B2 are not linked with other genes in formate utilization. In methanococci as well as in methanobacteria, the deazaflavin coenzyme F₄₂₀ is the electron acceptor of the Fdhs (5, 20).

Formate-hydrogen lyase activity (reaction 1) is common in methanococci and other methanogens (5, 6, 11, 20, 34). Reaction 1: HCO₂⁻ + H₂O → HCO₃⁻ + H₂ (+1.3 kJ/reaction).

Two pathways are likely to contribute to this activity in whole cells of M. maripaludis. In the first pathway, reduced coenzyme F₄₂₀ (F₄₂₀H₂) generated by Fdh is oxidized by the reversible F₄₂₀-dependent hydrogenase Fru (Fig. 1). In the second pathway, F₄₂₀H₂ is oxidized by the F₄₂₀-dependent methylene-H₄MPT dehydrogenase (Mtd) and Hmd (1, 2, 29) (Fig. 1). While previous studies demonstrated H₂ production from formate for some methanogens (7), the rates were very low and the pathways were not determined.

FIG. 1.

Potential pathways of H₂ production in the hydrogenotrophic methanogens. The first pathway includes Fdh and Fru. In this pathway, Fdh reduces F₄₂₀. In selenium-grown cells, F₄₂₀H₂ is oxidized by the [Ni-Fe] hydrogenase Fru. In the second pathway, F₄₂₀H₂ is oxidized by Mtd to reduce methenyl-H₄MPT to methylene-H₄MPT, which is then reoxidized by Hmd, a Ni-free hydrogenase, to produce H₂. Abbreviations: Fdh, formate dehydrogenase; Fru, F₄₂₀-reducing hydrogenase; F_{420 ox}, oxidized coenzyme F₄₂₀; F_{420 red}, reduced coenzyme F₄₂₀; Mtd, F₄₂₀-dependent methylene-H₄MPT dehydrogenase; Hmd, H₂-forming methylene-H₄MPT dehydrogenase.

Because of the low energy yield of CH₄ production, these microorganisms possess extremely high specific activities for enzymes involved in methanogenesis, including the hydrogenases used for H₂ consumption. If H₂ utilization were a reversible process, high rates of H₂ production would be possible. To test this hypothesis, the rates of H₂ production from formate were tested in M. maripaludis, a representative of a diversified and important group of methanogens.

MATERIALS AND METHODS

Strains and growth conditions.

M. maripaludis wild-type strain S2 and the mutant strains are listed in Table 1. For growth experiments, the strains were grown in 28-ml Balch tubes (4) filled with 5 ml of either McNA (McN minimal medium [32] supplemented with 10 mM acetate) or McCV medium (McNA medium supplemented with 0.2% yeast extract and Casamino Acids as well as vitamin solution [32]). The tubes were pressurized to 276 kPa with H₂-CO₂ gas (80:20 [vol/vol]) and grown at 37°C. When sodium formate was used (100 mM), 100 mM Tris-HCl solution, pH 7, was added to reduce the increase in pH. The tubes were also pressurized to 138 kPa with N₂-CO₂ (80:20 [vol/vol]).

TABLE 1.

Strains used in this study

Strain	Genotype	Reference
S2	Wild type	32
MM707	ΔfdhA1	33
MM708	ΔfdhA2	33
MM709	ΔfdhA1-ΔfdhA2	33
MM1145	ΔfruA	19
MM1183	ΔfrcA	19
MM1184	ΔfruA-ΔfrcA	19
MM1097	Δhmd	19
MM1020	Δmtd	19

For obtaining cell mass, M. maripaludis cultures were grown in 1-liter Wheaton bottles (3) with 100 ml of McNA or McCV medium and 138 kPa of H₂-CO₂ (80:20 [vol/vol]) at 37°C. When sodium formate was the substrate (0.4 M), the bottles were flushed with N₂-CO₂ (80:20 [vol/vol]) and pressurized to 35 kPa.

Cells were collected by centrifugation in sealed plastic centrifuge bottles that had been equilibrated in an anaerobic glove box (Coy Laboratories, Ann Arbor, MI) for at least 24 h to remove O₂ adsorbed to the plastic. The cells were centrifuged at 6,000 × g for 20 min at 4°C using a Beckman model J2-21 centrifuge (Beckman Coulter, Inc., Fullerton, CA) fitted with a Beckman JA-14 rotor. The cells were resuspended in 1/100 of the initial volume of an anaerobic buffer containing 50 mM piperazine-1,4-bis(2-ethanesulfonic acid) (PIPES), 400 mM NaCl, 20 mM KCl, 20 mM MgCl₂, 1 mM CaCl₂, and 5 mM dithiothreitol, pH 6.9.

H₂ measurements.

The H₂ measurements were performed at 37^οC in a custom-made anaerobic, water-jacketed cuvette (2.8 ml) fitted with a rubber stopper and gassed with O₂-free N₂ (31). The concentration of the dissolved H₂ in the anaerobic buffer was measured using a modified amperometric O₂ Clark-type electrode (15, 31) (Yellow Springs Instrument, Yellow Springs, OH) connected to a picoammeter PA2000 (Unisense, Aarhus, Denmark). The connections for the electrode (6.3-mm TRS connector) and picoammeter (BNC connector) were incompatible, and so the appropriate connection was manufactured. Standard curves were prepared with H₂-saturated distilled water. Cell suspensions of 0.1 mg (dry weight) were added via microsyringes to 1 ml of the same buffer used to suspend the cells. The assay was started by adding sodium formate. One unit was defined as 1 μmol of H₂ produced per minute. The cell dry weight was calculated from the slope of a standard curve relating absorbance at 600 nm to dry weight. From this curve, a suspension with an A₆₀₀ of 1 corresponded to 0.34 mg (dry weight)·ml⁻¹.

CH₄ detection.

Resting cells (0.1 mg [dry weight]) suspensions in 0.5 ml of buffer were transferred to 3.5-ml vials under an atmosphere of N₂. The assay was initiated by adding formate to a final concentration of 20 mM or flushing the vials with H₂-CO₂ (80:20 [vol/vol]) for 1 min. The samples were incubated at 37°C for 10 min. CH₄ was detected with an SRI 8610-C gas chromatograph (SRI Instruments, Torrance, CA) fitted with on-column injection, a Porpak Q teflon column at 90°C, and a flame ionization detector operating at 150°C. The carrier gas was N₂. One unit was defined as 1 μmol of CH₄ produced per minute.

Preparation of cell extracts.

All procedures were performed anaerobically. Cells were collected by centrifugation as described above and loaded into a chilled French pressure cell in the anaerobic glove box. Cell extracts were prepared by passing 10 ml of cell suspension (about 10 mg [dry weight] per ml) through the French pressure cell operated at 65 MPa equipped with a 22-gauge needle. Cell extracts were collected in the sealed tubes or serum bottles previously equilibrated in the anaerobic glove box. Subsequently, the extracts were centrifuged at 10,000 × g for 20 min at 4°C. Protein concentrations were determined using the Bradford protein kit (Bio-Rad, Hercules, CA).

Enzymatic assays.

The Fdh and F₄₂₀-reducing hydrogenase activities were measured spectrophotometrically under anaerobic conditions. H₂-dependent F₄₂₀ reduction was assayed in 1 ml of anaerobic buffer containing 100 mM PIPES, pH 6.9, 20 mM NaCl, 10 mM KCl, 10 mM MgCl₂, 1 mM CaCl₂, and 5 mM dithiothreitol. The final concentration of F₄₂₀ was 10 μM. The 1.6-ml glass cuvettes were sealed with rubber stoppers and flushed with O₂-free H₂ for 5 min before each assay. Changes in absorbance at a λ of 420 nm were measured using a Beckman DU-640B spectrophotometer. An extinction coefficient (ɛ₄₂₀) of 40 mM·cm⁻¹ was used for calculations. One unit was defined as 1 μmol of coenzyme F₄₂₀ reduced per minute. F₄₂₀ was purified from the M. maripaludis cell paste according to a modification of the method of Eirich et al. (9).

Fdh activity was assayed in 1 ml of the same buffer described above plus 2 mM methyl viologen (MV). The cuvettes were flushed with O₂-free N₂ for 5 min before each assay. The MV was reduced with a few microliters of 200 mM dithionite until the assay buffer turned slightly blue. The cell extract was added, and the reaction was initiated by adding formate to a final concentration of 10 mM. Changes in absorbance at a λ of 605 nm were recorded, and an extinction coefficient (ɛ₆₀₅) of 13.9 mM·cm⁻¹ was used for calculations. One unit was defined as 2 μmol of MV reduced per minute.

RESULTS AND DISCUSSION

Initial rates of H₂ production.

Previous reports of formate-dependent H₂ production measured formation of headspace H₂ during growth (7, 19, 28). However, preliminary measurements of the rate of production of dissolved H₂ with an H₂ probe and using resting cells far exceeded these reports. Therefore, the earlier measurements underestimated the production rate, possibly because H₂ uptake was occurring simultaneously or gas transfer to the headspace was rate limiting.

To examine these high rates in more detail, further experiments were performed to standardize the reaction conditions. Whole cells were grown either with H₂ or formate, washed in buffer, and resuspended in a reaction cuvette. In both cases, formate-dependent H₂ production proceeded linearly for the first minute (Fig. 2). In the subsequent 2 to 5 minutes, the apparent rate of H₂ generation declined until a plateau was reached. With 20 mM formate, the maximal H₂ concentrations ranged between 0.4 and 0.8 mM, depending upon the experiment. These concentrations approached 3 mM, the concentration expected at equilibrium. The observed rate would be lower than the true production rate if H₂ utilization were occurring simultaneously. However, H₂ utilization for methanogenesis was not likely to be a factor, because the initial concentration of CO₂ was very low. In addition, the inhibition of methanogenesis by bromoethanesulfonate (1 mM) had little effect on the initial rates or maximum values of H₂ production (data not shown). Therefore, the initial observed rate was not affected by simultaneous H₂ consumption for methanogenesis.

FIG. 2.

Initial rates of H₂ production from formate (solid line) by resting cells of M. maripaludis. Cells were grown with formate, washed in buffer, and resuspended in the reaction cuvette. Assays were initiated by adding 20 mM sodium formate (arrow). Results from incubation of the cell suspension in the absence of formate is shown by the broken line.

Both H₂- and formate-grown cells produced H₂ from formate at comparable rates. Although the rates depended greatly upon the experiment (see below), the rates varied from about 0.4 to 1.4 U·mg⁻¹ (dry weight) regardless of how the cells were grown (data not shown). These results suggested that the levels of Fdh and hydrogenase were high in both cell types. Since these enzymes are essential components of the methanogenesis system, the rates of methanogenesis were also compared. In one experiment, the rates of methanogenesis by H₂-grown cells were 0.37 ± 0.02 U·mg⁻¹ (dry weight) and 0.32 ± 0.04 U·mg⁻¹ (dry weight) with H₂ or formate, respectively (means ± standard deviations of triplicate measurements with two independently grown cultures). Similarly, the rates of methanogenesis by formate-grown cells were 0.25 ± 0.02 U·mg⁻¹ (dry weight) and 0.43 ± 0.04 U·mg⁻¹ (dry weight) with H₂ or formate, respectively. Since four molecules of H₂ or formate are consumed per molecule of CH₄ formed, the capacity of H₂ or formate consumption is comparable to the rate of formate-dependent H₂ production regardless of the growth substrate.

Influence of growth conditions on cellular rates of H₂ production.

The rate of formate-dependent H₂ production varied greatly between experiments. To determine if some of this variability resulted from the growth phase, H₂-grown cultures were monitored for formate-dependent H₂ production (Fig. 3). In batch cultures of methanococci, exponential growth is only observed at low cell densities (25). Above an absorbance of ∼0.4, growth becomes linear as the rate of H₂ transfer to the aqueous phase becomes rate limiting. Finally, the linear phase usually ends at absorbances of ∼0.8, as cells enter early stationary phase. When cultures were allowed to nearly exhaust H₂ late in growth, the specific activity of H₂ production increased 6- to 10-fold (Fig. 3). In addition to growth phase, the medium composition also affected these rates. At the end of growth, cells grown in the minimal medium possessed about 30% higher specific activities than cells grown in the rich medium (Fig. 3). Methanococci require higher levels of H₂ for anabolism during growth on minimal medium (26). Taken together, these results were consistent with a role for H₂ limitation in the regulation of H₂ production.

FIG. 3.

Changes in formate-dependent H₂ production by resting cells following growth under H₂ limitation. The wild-type S2 was grown in minimal or McNA (○) and rich or McCV (•) medium in 1-liter Wheaton bottles with 100 ml of broth. At the beginning of the experiment, the bottles were pressurized with 138 kPa of H₂-CO₂ (80:20 [vol/vol]). Cell samples (20 ml) were collected at different absorbances, as indicated by the points on the inset growth curves. After each sampling, the bottles were repressurized with N₂-CO₂ (80:20 [vol/vol]). This resulted in H₂ limitation but maintained the concentration of CO₂. Error bars represent 1 standard deviation from the four measurements.

To examine directly the role of H₂ partial pressure on the specific activity of H₂ production, cultures were grown in bottles with very large headspaces in order to minimize gas pressure fluctuations due to H₂ consumption. Parallel cultures were grown with H₂ partial pressures of 220 and 80 kPa to absorbances of 0.45 to 0.50. The specific activities of H₂ production were 0.12 ± 0.03 and 0.41 ± 0.04 U·mg⁻¹ (dry weight), respectively. Thus, the levels of H₂ during growth had a direct effect on the expression of the enzymes involved in H₂ production. These results are consistent with microarray observations, where the levels of mRNA for both the formate dehydrogenase (fdh1) and F₄₂₀-reducing hydrogenase (fru) genes are significantly higher during H₂ limitation (17).

Role of formate dehydrogenase.

To ascertain the importance of the two Fdh isozymes in H₂ production, the mutant strains ΔfdhA1, ΔfdhA2, and ΔfdhA1-ΔfdhA2 were tested for their ability to grow with formate, MV-dependent Fdh activity, and for H₂ production. The double mutant ΔfdhA1-ΔfdhA2, which was constructed by marker exchange of internal portions of both fdhA1 and fdhA2, was unable to utilize formate for either growth, which was followed for 140 h, or H₂ production (Fig. 4, Table 2, and data not shown). In addition, in extracts of H₂-grown cells, the MV-dependent Fdh activity was <0.02 U·mg⁻¹ (dry weight). Thus, H₂ production required Fdh. MV-dependent Fdh activities in cell extracts of the ΔfdhA2 mutant and the wild-type S2 strains were 1.0 ± 0.25 and 3.6 ± 1.2 U·mg⁻¹ (dry weight), respectively. In contrast, growth on formate of strain ΔfdhA2 was comparable to that of the wild type, and H₂ production was only slightly reduced. Therefore, the rate of formate oxidation did not appear to limit growth and H₂ production in this mutant. Notably, the ΔfdhA1 mutant failed to grow with formate without an extended incubation of 70 h and produced H₂ poorly (Fig. 4, Table 2, and data not shown). In addition, the MV-dependent Fdh activity was <0.02 U·mg⁻¹ (dry weight). Subsequent transfers of formate-grown cultures of the ΔfdhA1 mutant to the fresh medium decreased the lag phase, and the third subculture grew with formate after a lag of about 48 h. Presumably, mutations at other sites on the genome were responsible for the adaptation of the ΔfdhA1 mutant to growth on formate. For instance, increased expression of Fdh2 could account for this phenotype. Thus, Fdh2 played only a small role in H₂ production, and Fdh1 appeared to be the major isozyme in formate utilization under these conditions. In previous studies, the ΔfdhA1 mutant grew with formate at similar rates as the ΔfdhA2 strain (33). Presumably, this difference reflects the differences in the medium composition and the experimental design.

FIG. 4.

Growth of the wild-type and fdh mutant strains with H₂ or formate. Solid symbols, growth with H₂ (276 kPa; H₂-CO₂ 80:20 [vol/vol]); open symbols, growth with formate (100 mM). Circles, S2; inverted triangles, ΔfdhA1; squares, ΔfdhA2; diamonds, ΔfdhA1-A2. Each point represents the mean value of two replicates. Similar curves were obtained in a replicate experiment.

TABLE 2.

Activities of formate-dependent H₂ production by fdh mutant strains^a

Genotype	Sp act (U·mg−1 [dry wt])
Wild type	0.32 ± 0.05
ΔfdhA1	0.04 ± 0.01
ΔfdhA2	0.29 ± 0.02
ΔfdhA1-ΔfdhA2	<0.02

^aCultures were grown with H₂ at 136 kPa of H₂-CO₂, (80:20 [vol/vol]) to an absorbance of about 0.5. The values are means ± standard deviations of triplicate measurements with two independently grown cultures.

Kinetics of formate-dependent H₂ production by whole cells.

The kinetics depended on the growth substrates of the cells. For incubations of formate-grown cells, H₂ production followed Michaelis-Menten kinetics. The K_f, or concentration of formate required for half-maximal activity, was 3.6 ± 0.5 mM (mean ± 1 standard deviation), and the V_f, or the maximal specific activity, was 1.5 ± 0.1 U·mg⁻¹ (dry weight) (Fig. 5). For H₂-grown cells, H₂ production followed sigmoidal kinetics, and the kinetic values were 9 ± 1 mM and 1.1 ± 0.2 U·mg⁻¹ (dry weight), respectively. The sigmoidal kinetics for the H₂-grown cells were not due to the presence of two isozymes in the wild-type cells, because the ΔfdhA2 mutant possessed similar kinetics (Fig. 5). Instead, the biphasic kinetics suggested that the formate transporter FdhC plays a significant role in the kinetics of H₂ production. Expression of fdhC is reduced in H₂-grown cells (33). Thus, at low formate concentrations, low levels of the transporter may limit the rate of formate uptake and H₂ production. High formate concentrations would then compensate for low levels of the transporter. In contrast, the levels of the transporter may be sufficient in the formate-grown cells so that uptake is no longer rate limiting at millimolar formate concentrations.

FIG. 5.

Kinetics of formate-dependent H₂ production by resting cells of the wild-type strain S2 previously grown either with H₂ (A) or formate (B). The kinetics of the ΔfdhA2 mutant are shown in the insets. The S2 and ΔfdhA2 mutant strains were grown to absorbances of about 0.8 and 0.6, respectively.

Pathways of H₂ production.

Given the requirement for Fdh, two pathways of H₂ production were likely (Fig. 1). In selenium-grown cells, the major pathway was expected to include the oxidation of Fdh-generated F₄₂₀H₂ by the F₄₂₀-dependent [Ni-Fe] hydrogenase Fru. The selenium-independent isozyme Frc was not expected to play a role, because it would not be expressed under these conditions (27). In extracts of cells grown with both selenium and formate, the specific activities of MV-dependent Fdh and Fru were 18 and 2.1 U·mg⁻¹ of the total protein. Assuming that 60% of the cell was protein, the corresponding cellular specific activities were 10.8 and 1.3 U·mg⁻¹ (dry weight), values that were consistent with a role in whole-cell H₂ production. To examine this point further, the H₂ production activities of a series of in-frame deletion mutants in the F₄₂₀-dependent hydrogenases, ΔfruA, ΔfrcA, and ΔfruA-ΔfrcA, were examined. The activity of the ΔfruA-ΔfrcA mutant was severely reduced (Table 3). The ΔfruA mutant possessed nearly identical activity as the double mutant, suggesting the FruA played an important role (data not shown). In contrast, the activity of the ΔfrcA mutant was identical to the wild type (data not shown), as expected if this enzyme is not produced in the presence of selenium. These observations clearly suggested that F₄₂₀ is an intermediate in the process and provided evidence for an F₄₂₀-dependent formate-hydrogen lyase system in M. maripaludis. Because the ΔfruA-ΔfrcA mutant retained about 10% of the wild-type H₂ production activity (Table 3), an additional pathway must be present. Deletions of either mtd or hmd reduced the activity by about 30 to 40%, which indicated that the second system utilized the H₄MPT-dependent pathway (Fig. 1).

TABLE 3.

Activities of formate-dependent H₂ and CH₄ production for hydrogenase and methylene-H₄MPT dehydrogenase mutant strains^a

Genotype	Sp act (U·mg−1 [dry wt])
	H2	CH4
Wild type	0.56 ± 0.10	0.36 ± 0.03
ΔfruA-ΔfrcA	0.06 ± 0.02	0.20 ± 0.04
Δhmd	0.40 ± 0.08	0.12 ± 0.02
Δmtd	0.35 ± 0.05	0.13 ± 0.02

^aCells were grown with formate to an absorbance of about 0.6, except for Δmtd, for which the absorbance was about 0.4. H₂ and CH₄ production levels were measured using cells from the same culture. The values are means ± 1 standard deviation of triplicate measurements from each of two independently grown cultures.

Is H₂ an intermediate of methanogenesis with formate?

The rates of H₂ generation were comparable to the rates of methanogenesis, which suggested that H₂ was produced in sufficient amounts to be an intermediate during methanogenesis with formate. During hydrogenotrophic growth, H₂ is proposed to be the electron donor for the Eha-dependent reduction of CO₂ to formylmethanofuran as well as the heterodisulfide reductase, the first and last steps of methanogenesis, respectively (for a review see reference 10). During growth with formate, F₄₂₀ is initially reduced. If methanogenesis proceeds in a fashion similar to hydrogenotrophic growth, H₂ could be generated from F₄₂₀H₂ to produce the reductant for the first and last steps of methanogenesis. To test this hypothesis, rates of formate-dependent H₂ and CH₄ production were measured in cells derived from the same cultures in order to reduce the variability caused by growth and handling (Table 3). In these experiments, there was little correlation between the rates of H₂ and CH₄ production in the wild type and ΔfruA-ΔfrcA, Δhmd, and Δmtd mutants. In fact, for the ΔfruA-ΔfrcA mutant, the rate of methanogenesis exceeded the rate of H₂ production, which seemed to preclude the possibility that H₂ could be an obligate intermediate of methanogenesis.

Methanogenesis is very O₂ sensitive, and it is possible that the activity may have been damaged during preparation of the cell suspensions used in this experiment. Therefore, the rates of methanogenesis were also determined during growth of the wild type and mutants without preparation of resting cells. For the wild type, the specific activity for methanogenesis was low except during the exponential and linear growth phases (Fig. 6). For the ΔfruA-ΔfrcA and Δhmd mutants, growth and methanogenesis were nearly the same as the wild type. Even though the growth and CH₄ production were delayed for the Δmtd mutant (Fig. 6), the maximum rate of methanogenesis during the exponential growth phase was nearly the same as that of the wild type. For this mutant, H₂ must first accumulate in the medium to allow for activity of the low-affinity Hmd before growth commences (19). In conclusion, while the rate of H₂ production measured with a H₂ probe is too low to be an obligatory intermediate for methanogenesis from formate, it is still formally possible that a H₂ cycle could still exist within the cell. In this case, the cellular H₂ levels would not equilibrate with the bulk H₂ dissolved in the medium. However, this possibility seems unlikely, given that inactivation of each of the major pathways of H₂ production, either in the ΔfruA-ΔfrcA or the Δhmd and Δmtd mutants, had little effect on the cellular rate of methanogenesis from formate.

FIG. 6.

Growth (A) and specific activity of CH₄ production (B) by the wild-type S2 (•), ΔfruA-ΔfrcA (▿), Δhmd, (▪), and Δmtd (⋄) mutant strains grown with formate (100 mM). Each point represents the mean value of three replicates, and error bars in panel B represent 1 standard deviation. Similar curves were obtained in a replicate experiment.

Recently, a novel hypothesis for energy conservation by the hydrogenotrophic methanogens was proposed (30). This model predicts that the exergonic H₂-dependent reduction of the heterodisulfide is not membrane associated and does not generate a proton or sodium motive force. Instead, this exergonic reaction is directly coupled to the endergonic reduction of the low-potential ferredoxin required for CO₂ reduction by flavin-mediated electron bifurcation. According to this model, one low-potential electron from H₂ oxidation partially reduces the ferredoxin, while a high-potential electron partially reduces the heterodisulfide. An additional cycle is required to fully reduce both the ferredoxin and heterodisulfide (30). This model avoids the necessity of an energy-conserving hydrogenase, such as Eha, to reduce the low-potential ferredoxin. While the observation that H₂ is not an obligate intermediate during growth with formate supports this model, it remains unclear how F₄₂₀H₂ generated from formate donates electrons for the flavin-mediated electron bifurcation. If H₂ is not an intermediate, an enzymatic complex with the function of F₄₂₀H₂:flavin oxidoreductase would be required. At present, a candidate for this enzyme has not been identified, either from biochemical or genome analyses.

Possible applications.

The growing energy demand, environmental concerns, and limited resources of fossil fuels draw attention to H₂, which is a clean and efficient energy source. Microbial H₂ production has involved a range of approaches (8, 13, 14, 24). Methanogens are very active H₂ consumers; if these systems could be used for H₂ production, high rates would be possible. To test this concept, H₂ production from formate by resting cells of methanococci was evaluated. Rates of formate-dependent H₂ production have been previously examined in a few bacteria. The methylotrophs Methylomonas albus and Methylosinus trichosporium produced H₂ at rates of 1.6 and 0.4 mU·mg⁻¹ (dry weight), respectively, after 5 hours of incubation under anaerobic conditions (21). Similar low rates were observed for Shewanella oneidensis MR-1 (23) and Alcaligenes eutrophus (22). The highest rates of 1.7 to 4.2 U·mg⁻¹ (dry weight) were obtained with genetically engineered Escherichia coli strains (35). The rates obtained from wild-type methanococci, up to 1.4 U·mg⁻¹ (dry weight), were comparable. However, because of the equilibrium constant, high concentrations of H₂ cannot be produced from formate regardless of the catalyst employed. Therefore, biotechnological applications would require an efficient means of harvesting H₂ at low levels. Because F₄₂₀H₂ is a key intermediate in methanococcal H₂ production, one might speculate that the substrate range could be increased by genetically engineering methanococci to couple the reduction of F₄₂₀ to the oxidation of substrates other than formate.

While our studies were focused on M. maripaludis, other methanogens also possess many of the activities required for formate-dependent H₂ production and could be candidates for biotechnological applications. Thus, use of thermophilic or freshwater species might extend this application to a much broader range of conditions. The use of formate-utilizing methanogens appears to be the optimal solution. In this approach the growth and H₂ production are decoupled. In such bioreactors, formate first is used to obtain the cell mass. Then, by the continuous supply of this substrate, the cell mass could be an efficient catalyst for H₂ production. For wild-type cells, the rates of formate-dependent H₂ production are among the highest reported for prokaryotes. Further optimization of the growth and reaction conditions as well as genetic engineering could potentially increase these rates yet again. These observations open the possibilities for the use of methanococci in bioreactors for the generation of H₂ from the relatively inexpensive chemical, which can be derived from biomass (35).