Hydrogen is a preferred intermediate in the energy-conserving electron transport chain of Methanosarcina barkeri

Abstract

Methanogens use an unusual energy-conserving electron transport chain that involves reduction of a limited number of electron acceptors to methane gas. Previous biochemical studies suggested that the proton-pumping F₄₂₀H₂ dehydrogenase (Fpo) plays a crucial role in this process during growth on methanol. However, Methanosarcina barkeri Δfpo mutants constructed in this study display no measurable phenotype on this substrate, indicating that Fpo plays a minor role, if any. In contrast, Δfrh mutants lacking the cytoplasmic F₄₂₀-reducing hydrogenase (Frh) are severely affected in their ability to grow and make methane from methanol, and double Δfpo/Δfrh mutants are completely unable to use this substrate. These data suggest that the preferred electron transport chain involves production of hydrogen gas in the cytoplasm, which then diffuses out of the cell, where it is reoxidized with transfer of electrons into the energy-conserving electron transport chain. This hydrogen-cycling metabolism leads directly to production of a proton motive force that can be used by the cell for ATP synthesis. Nevertheless, M. barkeri does have the flexibility to use the Fpo-dependent electron transport chain when needed, as shown by the poor growth of the Δfrh mutant. Our data suggest that the rapid enzymatic turnover of hydrogenases may allow a competitive advantage via faster growth rates in this freshwater organism. The mutant analysis also confirms the proposed role of Frh in growth on hydrogen/carbon dioxide and suggests that either Frh or Fpo is needed for aceticlastic growth of M. barkeri.

Methanogenesis is the terminal step in biomass degradation in many anaerobic environments and plays a central role in the global carbon cycle. Although most of the methane (CH₄) produced is oxidized to carbon dioxide (CO₂) by methane-consuming organisms, substantial quantities (ca. 10¹⁴ g/year) escape into the atmosphere where it acts as a potent greenhouse gas (1). Most methanogens produce CH₄ by reducing CO₂ with hydrogen gas (H₂) (2). However, some Methanosarcina species such as M. barkeri and M. mazei also are capable of using a variety of other substrates, including acetate, which accounts for ca. 2/3 of global CH₄ production (3), and C1 compounds such as methanol, methylsulfides, and methylamines (4).

Methanogenic organisms produce CH₄ as a byproduct of anaerobic respiration involving a unique energy-conserving electron transport chain found only in Archaea. At least 2 distinct types of methanogenic respiration exist: 1 found in methanogens, including Methanosarcina species that synthesize cytochromes, and the other in those that lack cytochromes (1). The penultimate step of both respiratory pathways involves the reduction of methyl-coenzyme M (CoM-SH, mercaptoethanesulfonic acid) to CH₄ using coenzyme B (CoB-SH, N-7-mercaptoheptanoyl-O-phospho-L-threonine) and CoM-SH as electron donors. The other product of this reaction is the heterodisulfide of CoM-SH and CoB-SH (CoM-S-S-CoB), which serves as the terminal electron acceptor in the energy-conserving electron transport chain. In methanogenic Archaea that lack cytochromes, the means by which energy is conserved is poorly understood but probably involves a cytoplasmic electron bifurcation pathway similar to that recently characterized in Clostridium (1, 5). In contrast, the energy-conserving electron transport chain of cytochrome-containing methanogens, exemplified by Methanosarcina species, has been studied in detail, including the reconstitution of 2 distinct proton-translocating electron transport systems in vitro (6–9).

In Methanosarcina species, a membrane-bound electron transport chain that terminates with the reduction of the CoM-S-S-CoB heterodisulfide generates ion-motive force that can be used by ATP synthase to form ATP (Fig. 1). Either H₂ or reduced coenzyme F₄₂₀ (F₄₂₀H₂) can donate electrons for reduction of CoM-S-S-CoB (2, 4). In the H₂:heterodisulfide oxidoreductase system, a methanophenazine-dependent hydrogenase (Vht or Vhx) (10) oxidizes H₂ in the periplasm with transfer of 2 electrons to the membrane-soluble electron carrier methanophenazine. The reduced methanophenazine (MPH₂) subsequently is oxidized by the enzyme heterodisulfide reductase (Hdr) with concomitant reduction of CoM-S-S-CoB. In the F₄₂₀H₂:heterodisulfide oxidoreductase system, the F₄₂₀H₂ dehydrogenase (Fpo) catalyzes electron transfer from F₄₂₀H₂ to methanophenazine, concomitantly pumping 2 protons out of the cell. As in the H₂:heterodisulfide oxidoreductase system, MPH₂ passes electrons to CoM-S-S-CoB, leading to translocation of 2 additional protons. In vitro measurements suggest that the magnitude of proton motive force generated is the same for both oxidoreductase systems, probably 4H⁺/2e⁻ (9, 11).

Fig. 1.

The electron transport chain of M. barkeri has been proposed to comprise 2 energy-conserving systems, the F₄₂₀H₂:heterodisulfide oxidoreductase and the H₂:heterodisulfide oxidoreductase. In the former system, F₄₂₀H₂ is oxidized by FpoF releasing 2 electrons that are transferred through FpoBCDI and then FpoAHJKLMN to membrane-soluble methanophenazine. This reaction is coupled to the pumping of 2 protons outside the cell. Reduction of methanophenazine consumes 2 protons from the cytoplasm, which subsequently are released outside the cell upon oxidation of MPH₂. The electrons then are transferred through HdrED to reduce CoM-S-S-CoB with 2 protons from the cytoplasm. Alternatively, in the H₂:heterodisulfide oxidoreductase, H₂ is oxidized by Vht/Vhx to produce 2 protons outside the cell and 2 electrons that are transferred to MPH₂, which then is used to reduce CoM-S-S-CoB. The dashed arrow represents a third possible energy-conserving mechanism that is proposed in this study. In this pathway, F₄₂₀H₂ is oxidized by the cytoplasmic hydrogenase Frh to generate H₂. The H₂ then diffuses outside the cell to the active site of membrane-bound hydrogenase Vht/Vhx, where it is reoxidized, resulting in the translocation of 2 protons via a H₂-cycling mechanism. The electrons are passed through methanophenazine to CoM-S-S-CoB, as in the other 2 systems. In all 3 systems, the entire electron transport process leads to the net translocation of 4 protons (highlighted in red) outside the cell per 2 electrons transferred from F₄₂₀H₂ or H₂ to the CoM-S-S-CoB. The electrochemical gradient generated is coupled to ATP synthesis via an A-type ATPase. Abbreviations: CoB-SH, coenzyme B; CoM-SH, coenzyme M; CoM-S-S-CoB, mixed disulfide of CoM-SH and CoB-SH; Cytb₂, cytochrome b₂; F₄₂₀/F₄₂₀H₂, oxidized and reduced Factor 420; FAD, flavin adenine dinucleotide; [FeS], iron-sulfur cluster; Fpo, F₄₂₀H₂:phenazine oxidoreductase; Frh, F₄₂₀-reducing hydrogenase; Hdr, heterodisulfide reductase; MP/MPH₂, oxidized and reduced methanophenazine; [NiFe], bimetallic catalytic center; Vht/Vhx, methanophenazine-dependent hydrogenase.

The F₄₂₀H₂:heterodisulfide oxidoreductase system shares features with the aerobic electron transport chain found in many bacteria. Fpo was purified from M. mazei Gö1 as a 5-subunit complex (FpoBCDIF) capable of oxidizing F₄₂₀H₂ with a variety of artificial electron acceptors (12). Most of these proteins were shown later to be encoded by a 13-gene operon, fpoABCDHIJJKLMNO (fpo), with fpoF being located elsewhere on the genome (supporting information (SI) Fig. S1) (11). Fpo shares significant homology with NADH dehydrogenase I (Nuo) of Escherichia coli, and the nomenclature reflects that similarity (Table S1). Accordingly, both Fpo and Nuo are composed of similar membrane-integral modules (the A, H, J, K, L, M, and N subunits) and membrane-associated modules (the B, C, D, and I subunits). However, the enzymes differ in their substrates and hence in their input modules, so that NuoEFG oxidizes NADH, whereas FpoF oxidizes F₄₂₀H₂, which is a hydride carrier analogous to NADH (2).

The source of electrons for the 2 disparate Methanosarcina electron transport chains varies with the growth substrate (Fig. S2). During growth on H₂/CO₂ (the hydrogenotrophic pathway), or on H₂ in combination with a C1 electron acceptor (the methyl-respiration pathway), electrons are derived from H₂ oxidation and are used via the H₂:heterodisulfide oxidoreductase system (2). Genetic and biochemical experiments in M. barkeri suggest that use of acetate (the aceticlastic pathway) also involves the obligate intermediacy of H₂ and the H₂:heterodisulfide oxidoreductase system (13, 14). In contrast, metabolism of compounds such as methanol or methylamines (the methylotrophic pathway) produces 2 equivalents of F₄₂₀H₂ per C1 molecule oxidized. Thus, it has been suggested that F₄₂₀H₂:heterodisulfide oxidoreductase system plays a central, perhaps essential, role in the energy-conserving electron transport chain of the methylotrophic pathway in all Methanosarcina species (reviewed in refs. 2, 4, 15–18).

Surprisingly, the in vivo genetic data presented here tell a different story. Contrary to expectations, Fpo is not required for growth under any condition tested, including during growth on methanol. Instead, reducing equivalents from methanol oxidation seem to be preferentially passed to molecular H₂ by the cytoplasmic F₄₂₀-reducing hydrogenase (Frh). Subsequently, H₂ may enter the electron transport chain via the H₂:heterodisulfide oxidoreductase system (Fig. 1). These data are reminiscent of the hydrogen-cycling model for electron transport first proposed by Odom and Peck for sulfate-reducing bacteria (19) and suggest that a reanalysis of energy-conservation mechanisms in freshwater Methanosarcina species is warranted.

Results

The F₄₂₀H₂:Heterodisulfide Oxidoreductase Is Conserved in All Sequenced Methanosarcina Genomes.

The genomes of M. barkeri Fusaro (20) and M. acetivorans C2A (21) encode fpoABCDHIJJKLMNO and fpoF operons that are nearly identical to those found in M. mazei (Fig. S1 and Table S1) (22). All known structural and catalytic amino acid residues are conserved in the predicted protein sequences of the Fpo subunits from all 3 Methanosarcina species (Fig. S3) (2, 4, 11). Further, each sequenced genome carries homologues of the hdrDE operon, which encodes the methanophenazine-linked Hdr (Fig. S1) (20–23). Thus, although biochemical activity has been examined only in M. mazei, functional F₄₂₀H₂:heterodisulfide oxidoreductase systems probably are present in each of the Methanosarcina species examined to date.

F₄₂₀H₂ Dehydrogenase Is Not Required for Methanogenesis or Growth in M. barkeri.

As described in earlier sections, in vitro biochemical experiments led to the suggestion that the F₄₂₀H₂:heterodisulfide oxidoreductase system should be essential for growth on C1 compounds via the methylotrophic pathway. To test this hypothesis in vivo, we constructed M. barkeri mutants lacking F₄₂₀H₂ dehydrogenase by deleting the fpoA-O operon or the fpoF gene (Table S2).

The ΔfpoA-O and ΔfpoF mutants were tested for their ability to grow on various methanogenic substrates (Table 1). As expected, both mutants grow on H₂/CO₂, on methanol plus H₂/CO₂, and on acetate with growth rates and yields similar to the isogenic parental strain. However, we were surprised to discover that growth of the mutants on methanol also was unaffected. Further, CH₄ and CO₂ were produced by the ΔfpoA-O and ΔfpoF mutants in the expected 3:1 stoichiometry in amounts and at rates similar to the parent on all substrates (Tables 2 and 3). Thus, loss of F₄₂₀H₂ dehydrogenase, and therefore of the F₄₂₀H₂:heterodisulfide oxidoreductase system, does not measurably affect methanogenesis or growth in M. barkeri.

Table 1.

Generation time (h) and relative growth yield (%)^* of M. barkeri Fusaro strains^† in various media

Strain	Substrates
	CH3OH	CH3OH/H2/CO2	H2/CO2	CH3COOH
Δhpt‡	7.3 ± 0.3 (100%)	5.7 ± 0.1 (100%)	13.7 ± 2.5 (100%)	37 ± 3.4 (100%)
Δfpo	7.3 ± 0.2 (101%)	6.1 ± 0.5 (100%)	11.9 ± 0.9 (100%)	36 ± 2.6 (100%)
ΔfpoF	7.3 ± 0.2 (96%)	5.8 ± 0.5 (96%)	12.4 ± 1.2 (90%)	39 ± 1.9 (84%)
Δfre	7.7 ± 0.3 (116%)	5.5 ± 0.4 (125%)	8.8 ± 0.8 (95%)	41 ± 3.1 (84%)
Δfrh	13.7 ± 0.6 (52%)	6.5 ± 0.3 (80%)	NG (NA)	55 ± 7.0 (84%)
Δfpo Δfrh	NG (NA)	5.0 ± 0.3 (80%)	NG (NA)	NG (NA)
ΔfpoF Δfrh	NG (NA)	6.4 ± 0.4 (73%)	NG (NA)	76 ± 0.8 (66%)

*Growth rate and yield were measured as described in Materials and Methods; growth yield is relative to the parental strain on the same substrate. Values represent the average and standard deviation of at least triplicate measurements.

^†Strains used were WWM85 (Δhpt), WWM71 (Δfpo), WWM123 (ΔfpoF), WWM122 (Δfrh), WWM108 (Δfpo Δfrh), and WWM145 (ΔfpoF Δfrh).

^‡M. barkeri Fusaro parent strain in which all deletions were constructed.

NG, no growth for at least 6 months of incubation; NA, not applicable.

Table 2.

Methane (μmoL) and carbon dioxide (μmoL) production^* from resting cell suspensions of M. barkeri Fusaro strains^†

Strain	Substrates‡
	N2		CH3OH		CH3OH/H2		H2/CO2
	CH4	CO2	CH4	CO2	CH4	CO2	CH4	CO2
Δhpt§	<1	2 ± 0.1	343 ± 4	108 ± 2	465 ± 17	<1	298 ± 21	NA
Δfpo	<1	1 ± 0.1	365 ± 3	115 ± 1	494 ± 11	<1	356 ± 6	NA
ΔfpoF	<1	1 ± 0.1	332 ± 7	106 ± 2	439 ± 27	<1	359 ± 9	NA
Δfre	<1	1 ± 0.1	399 ± 14	133 ± 4	508 ± 66	<1	381 ± 21	NA
Δfrh	<1	2 ± 0.9	149 ± 3	42 ± 1	481 ± 14	<1	7 ± 2	NA
Δfpo Δfrh	<1	2 ± 0.7	47 ± 2	9 ± 1	480 ± 59	<1	10 ± 1	NA
ΔfpoF Δfrh	<1	4 ± 0.1	54 ± 3	16 ± 1	465 ± 9	<1	11 ± 5	NA

*Values are the average and standard deviation of at least 3 trials.

^†Strains used were WWM85 (Δhpt), WWM71 (Δfpo), WWM123 (ΔfpoF), WWM122 (Δfrh), WWM108 (Δfpo Δfrh), and WWM145 (ΔfpoF Δfrh).

^‡ Assays were conducted as described in Supplementary Information.

^§M. barkeri Fusaro parent strain in which all deletions were constructed.

NA, not applicable (CO₂ produced could not be measured because it was added to headspace).

Table 3.

Rate (nmol min⁻¹ mg⁻¹) of methane production^* from resting cell suspensions of M. barkeri Fusaro strains^†

Strain	Substrate‡
	N2	CH3OH	CH3OH/H2
Δhpt§	<1	70 ± 6	121 ± 35
Δfpo	<1	93 ± 13	127 ± 29
ΔfpoF	<1	82 ± 2	143 ± 7
Δfre	<1	95 ± 13	134 ± 32
Δfrh	<1	14 ± 0.5	97 ± 3
Δfpo Δfrh	<1	4 ± .5	151 ± 3
ΔfpoF Δfrh	<1	4 ± 0.1	135 ± 15

*Values are the average and standard deviation of at least three trials.

^†Strains used were WWM85 (Δhpt), WWM71 (Δfpo), WWM123 (ΔfpoF), WWM122 (Δfrh), WWM108 (Δfpo Δfrh), and WWM145 (ΔfpoF Δfrh).

^‡Assays were conducted as described in SI Text.

^§M. barkeri Fusaro parent strain in which all deletions were constructed.

Frh Is Essential for Growth of M. barkeri on H₂/CO₂ and Plays an Important Role in the Methylotrophic Pathway.

The dispensability of Fpo raises the question of which enzyme(s) transfers electrons from F₄₂₀H₂ to methanophenazine in the methylotrophic pathway. Because M. barkeri possesses a highly active F₄₂₀-reducing hydrogenase (24), we considered the possibility of electrons being channeled from F₄₂₀H₂ into the H₂:heterodisulfide oxidoreductase system via production of H₂ by Frh (Fig. 1).

The F₄₂₀-reducing hydrogenase couples oxidation of H₂ to F₄₂₀ reduction in vitro and is fully reversible (24–26). M. barkeri has 2 operons, frhADGB and freAEGB, with the potential to encode F₄₂₀-reducing hydrogenases (Fig. S1) (27). It is unclear whether freAEGB encodes an active hydrogenase, because it lacks the required maturation protease encoded by frhD in the homologous operon. Nevertheless, the putative fre-encoded hydrogenase shares all important catalytic and structural amino acid residues with the frh-encoded enzyme. Thus, freAEGB could encode a functional F₄₂₀-reducing hydrogenase if the FrhD protease can function in trans (10). The F₄₂₀-reducing hydrogenase is proposed to provide F₄₂₀H₂, which is needed for CO₂ reduction in the hydrogenotrophic pathway (Fig. S2) (24); however, frhADGB and freAEGB are expressed during growth on both H₂/CO₂ and methanol (10, 27). Thus, it seems possible that these genes play a role in both the hydrogenotrophic and methylotrophic pathways.

To test this possibility, the freAEGB or frhADGB operons were deleted from the chromosome of M. barkeri (Table S2), and the resulting mutants were characterized. The Δfre mutant is indistinguishable from its parent with respect to growth rate and yield on all substrates tested (Table 1). Further, deletion of fre from M. barkeri does not affect the amount and rate of CH₄ produced in resting cell suspensions from the various substrates, nor does the mutation change the expected 3:1 ratio of CH₄ to CO₂ on methanol (Tables 2 and 3). These data indicate that Fre is not required for growth of M. barkeri on any of the substrates tested.

In contrast, the Δfrh mutation has severe phenotypic consequences on several of the growth media examined (Tables 1–3). Although the Δfrh mutant is indistinguishable from its parent when grown on methanol plus H₂/CO₂ and acetate, it is unable to grow on H₂/CO₂. Moreover, the Δfrh mutant has a 2-fold slower growth rate and a 50% reduction in growth yield when methanol alone is used as a substrate. The Δfrh mutant exhibits a very long lag phase of 911 ± 21 h, as compared with 31 ± 0.4 h for wild-type M. barkeri on methanol. The growth defects of the Δfrh mutant are reflected further in CH₄ production. Resting cell suspensions of the Δfrh mutant produce negligible amounts of CH₄ from H₂/CO₂ and although they disproportionate methanol in the expected 3:1 ratio, they consistently produce only half as much CH₄ and CO₂ as the parent. Finally, the Δfrh mutation lowers the rate of CH₄ production from methanol in resting cells by ca. 4-fold relative to the parent.

Taken together, these data indicate that Frh is essential for growth by the hydrogenotrophic pathway and plays an important but dispensable role in the methylotrophic pathway, the latter potentially being the delivery of electrons from F₄₂₀H₂ into the H₂:heterodisulfide oxidoreductase system. Importantly, Fre is not able to substitute for the role of Frh under the conditions tested, suggesting that it does not encode a functional F₄₂₀-reducing hydrogenase in the absence of frhADGB.

M. barkeri Possesses Two Functional Pathways for Electron Transfer from F₄₂₀H₂ to Methanophenazine.

The fact that the Δfrh mutant retains the ability to grow and produce CH₄ from methanol, although at reduced rates, suggests that the cell has an alternative, less efficient route to deliver electrons from F₄₂₀H₂ to the CoM-S-S-CoB heterodisulfide. The most obvious candidate for this electron transfer pathway is the F₄₂₀H₂:heterodisulfide oxidoreductase system. To test this hypothesis, we constructed and characterized double mutants lacking frhADGB and either fpoA-O or fpoF (Tables S2 and Tables 1, 2, and 3).

Like the single Δfrh mutant, the ΔfpoA-O/Δfrh and ΔfpoF/Δfrh double mutants are unable to grow on H₂/CO₂ and produce negligible amounts of CH₄ in resting cell suspensions from this substrate. Both mutants are able to grow and produce CH₄ at levels and rates comparable to the parent on methanol plus H₂/CO₂. However, in contrast to the Δfrh, ΔfpoA-O, and ΔfpoF single mutants, the double mutants are incapable of growth on methanol. Thus, Frh and Fpo fulfill a similar role during growth on this substrate that is lost in the absence of both enzymes. Quantitative RT-PCR experiments show that although fpoA-O transcripts levels are slightly higher in the Δfrh mutant (3.9 ± 1.5-fold), they are easily detectable in the parental strain. Therefore, Fpo is expressed and has the potential to contribute to the methylotrophic electron transport chain in wild-type M. barkeri. Interestingly, the ΔfpoA-O/Δfrh and ΔfpoF/Δfrh mutants still produce small amounts of CH₄ and CO₂ from methanol despite their inability to grow on this substrate. Hence, an additional minor electron transport pathway(s) exists in M. barkeri that is incapable of supporting growth.

Surprisingly, the double mutants display different phenotypes when grown on acetate. The ΔfpoF/Δfrh mutant grows slowly on acetate, whereas the ΔfpoA-O/Δfrh mutant is unable to use acetate. It is unclear how the absence of both Fpo and Frh affects growth on acetate, because neither the ΔfpoA-O nor the Δfrh single mutant exhibits a growth defect on this substrate. These data seem to suggest that the input module of Fpo (FpoF) can be dispensable under conditions in which the proton-pumping methanophenazine oxidoreductase (FpoABCDHIJKLMNO) is not.

Discussion

Methanosarcina species have proven to be exceptional model organisms for genetic analysis of methanogenesis (14, 28–30), an approach that modified our concept of energy conservation in M. barkeri. The data presented here indicate that M. barkeri has 2 distinct energy-conserving electron transport pathways during growth via the methylotrophic methanogenic pathway (Fig. 1). Contrary to expectations, M. barkeri apparently prefers to transfer electrons obtained from C1 compound oxidation to the H₂:heterodisulfide oxidoreductase system via H₂ rather than to transfer them directly into the F₄₂₀H₂:heterodisulfide oxidoreductase system. Our results suggest that the cytoplasmic F₄₂₀-reducing hydrogenase mediates electron transfer to H₂ via oxidation of F₄₂₀H₂, which, along with reduced ferredoxin (Fd_red), is the direct product of C1 compound oxidation (24, 26). When the cells lose the ability to produce H₂ via this route, growth on methanol is severely affected, with reduction in the rates of growth and CH₄ production, in total growth yield, and in the amount of CH₄ produced. We suggest that the H₂ diffuses out of the cell and enters the H₂:heterodisulfide oxidoreductase system via methanophenazine-dependent hydrogenase (Vht or Vhx) (10), whose active site is known to be in the periplasm (4). Because production of H₂ by Frh consumes protons within the cytoplasm, whereas oxidation of H₂ by Vht/Vhx releases protons outside the cell, this electron transport chain is capable of establishing a proton gradient across the membrane that can be used to generate ATP by the ATP synthase (31), thus conserving energy via a H₂-cycling mechanism (19). M. barkeri genome harbors 2 operons, vhtGACD and vhxGAC, that potentially can encode methanophenazine-dependent hydrogenases. Analogous to the fre operon, the vhx operon lacks gene D that encodes the hydrogenase maturation protease and therefore may not encode a functional hydrogenase. However, Vhx shares all important catalytic and structural amino acid residues with Vht and could encode an active methanophenazine-dependent hydrogenase if the VhtD protease can function in trans (10).

H₂ cycling, as a mechanism of energy conservation, was first proposed in sulfate-reducing bacteria (19). Later, it also was proposed in other anaerobic organisms such as Acetobacterium woodii (32) and Geobacter sulfurreducens (33). Based on the production of H₂ during growth of Methanosarcina species on methylated substrates such as methanol, trimethylamine, and acetate, H₂ cycling also was suggested to occur in Methanosarcina (34). Experimental support for H₂ cycling has been provided in Desulfovibrio vulgaris, wherein simultaneous production and consumption of H₂ were detected during metabolism of pyruvate and sulfate (35). Also, suppression of a D. vulgaris H₂-evolving hydrogenase led to reduced growth rates on lactate and sulfate, suggesting the importance of H₂ production in growth (36). Nevertheless, the H₂-cycling theory is not accepted universally, and several lines of evidence have been used to argue that this mechanism is unlikely (reviewed in ref. 37). For example, high concentrations of H₂ do not inhibit lactate oxidation in sulfate-reducing bacteria (38), nor does a mutation that blocks use of H₂ prevent growth on lactate (39). Further, the idea that the cell would transfer electrons preferentially to a molecule that can diffuse freely away from the cell seems highly problematic.

The data presented here suggest a way to reconcile these conflicting views. Although M. barkeri apparently prefers to transfer electrons via H₂, they remain capable of using the F₄₂₀H₂:heterodisulfide oxidoreductase system when the ability to produce cytoplasmic H₂ is lost. Thus, M. barkeri employs a branched electron transport chain with most electrons flowing from F₄₂₀H₂ into the H₂:heterodisulfide oxidoreductase system via H₂ but with some fraction flowing directly into the F₄₂₀H₂:heterodisulfide oxidoreductase system. A similar branched electron transport chain in sulfate reducers would explain the results cited as arguing against H₂ cycling without invalidating the model. It should be noted that a recently proposed metabolic model for D. vulgaris suggests just such a branched electron transport chain (37).

Why, then, is the H₂-cycling pathway the preferred electron transport chain in M. barkeri? Because the initial electron donor (F₄₂₀H₂) and final electron acceptor (methanophenazine) are the same, the amount of energy available from the 2 electron transport schemes must be identical. Experimental measurements suggest that electron transfer from F₄₂₀H₂ to CoM-S-S-CoB via Fpo and Hdr is accompanied by translocation of 4 protons across the membrane (11). Similarly, the flow of electrons from F₄₂₀H₂ to methanophenazine via Frh and Vht/Vhx leads to the translocation of 2 protons across the membrane by virtue of the H₂-cycling mechanism, whereas electron transfer from MPH₂ to CoM-S-S-CoB translocates another 2 protons (9). Thus, the magnitude of proton motive force generated should be identical via either route (Fig. 1). In contrast, the rate of CH₄ production from methanol in resting cell suspensions of Δfrh mutant is ca. 4 times slower than that of ΔfpoA-O and ΔfpoF mutants, indicating that H₂ cycling is a much faster mechanism of energy conservation and, as observed in our mutant strains, allows correspondingly faster growth rates. The biochemical properties of the enzymes involved are remarkably consistent with our in vivo data. Thus, FpoBCDIF [molecular weight (MW) = 135.1 kDa] (11) purified from M. mazei catalyzes F₄₂₀H₂ oxidation with a turnover number (K_cat) of 38 s⁻¹ and catalytic efficiency [K_cat/Michaelis constant (K_m)] of 5.4 × 10⁶ M⁻¹s⁻¹ (12); FrhAGB (MW = 194.4 kDa) (27) purified from M. barkeri catalyzes H₂ oxidation using F₄₂₀ as the electron acceptor with a K_cat of 134 s⁻¹ and K_cat/K_m of 4.4 × 10⁷ M⁻¹s⁻¹ (24). Because methanol-metabolizing M. barkeri cells have been shown to maintain ratios of F₄₂₀H₂ and F₄₂₀ in thermodynamic equilibrium with the H₂ partial pressure (25), it is reasonable to assume that Frh catalyzes the reverse reaction (F₄₂₀H₂ oxidation) with similar efficiency as the forward reaction (H₂ oxidation). Therefore, the K_cat of Frh is ca. 4 times higher than that of Fpo for F₄₂₀H₂ oxidation, suggesting that Frh is faster than Fpo in catalyzing this reaction.

A variety of data indicate that H₂ cycling is important to many, but not all, Methanosarcina species. Growth of M. barkeri on acetate involves the obligate production of H₂ (14). Thus, in combination with the data presented here, it seems likely that this species prefers to use H₂ cycling for all soluble substrates. The situation probably is similar in M. mazei because it has functional Frh, Vht, and Fpo enzymes as well (11). However, methylotrophic species such as M. acetivorans (10, 21, 28), Methanolobus tindarius (40, 41), and Methanococcoides burtonii (42) do not encode functional hydro-genases (22). Hence, these organisms probably rely exclusively on the F₄₂₀H₂:heterodisulfide oxidoreductase system for energy conservation. Moreover, we made an intensive effort to delete the fpo genes from M. acetivorans without success, suggesting that F₄₂₀H₂ dehydrogenase is essential in this organism. It has been suggested previously that these organisms forego H₂-dependent electron transport pathways because of their high-salt marine habitat, wherein they exist as disaggregated single cells and would be prone to lose the freely diffusible H₂ gas to competing organisms. In contrast, freshwater organisms such as M. barkeri that exist as large multicellular aggregates have a higher chance of retaining H₂ gas within the aggregates, enabling them to use it as an electron carrier (28).

Finally, the phenotypes of the mutants constructed here also provide insight into the metabolism of H₂/CO₂ and acetate by M. barkeri. Because of its ability to catalyze F₄₂₀ reduction with H₂, F₄₂₀-reducing hydrogenase was proposed to provide F₄₂₀H₂ for reduction of methenyl-tetrahydrosarcinapterin (H₄SPT) and methylene-H₄SPT in the hydrogenotrophic pathway (Fig. S2) (24, 43). The inability of the frh mutants to grow on H₂/CO₂ provides direct experimental support for this proposal. Interestingly, the putative fre-encoded hydrogenase cannot substitute for the frh-encoded hydrogenase. This inability may be caused by low expression of fre, absence of posttranslational processing, mutations in structural or catalytic residues, or some combination of these factors (10). Nevertheless, the dispensability of fre in M. barkeri is not surprising because M. mazei lacks the fre operon and is able to grow via all 4 methanogenic pathways (4, 10). Thus, the role of Fre in M. barkeri remains mysterious.

Methanogenesis from acetate does not require either Frh or Fpo; however, our results show that 1 of the 2 enzymes, but not both, is needed for growth on this substrate. We previously showed that mutations in the C1 oxidation pathway prevent growth on acetate, presumably by blocking the production of reducing equivalents needed for biosynthetic reactions (30). The lack of growth of the Δfrh/Δfpo double mutant on acetate medium suggests that these reducing equivalents must flow through either Frh or Fpo to allow growth. Interestingly, the Δfrh/ΔfpoF double mutant is able to grow on acetate, clearly suggesting that the membrane-bound proton-pumping module has the ability to accept electrons from input modules other than FpoF. In this regard, FpoF is homologous to the β-subunit (B) of F₄₂₀-reducing hydrogenases and shares a common substrate, coenzyme F₄₂₀ (11, 27). It is conceivable that in the absence of FpoF and FrhB, FreB serves as the input module, thus channeling electrons from F₄₂₀H₂ to Fpo and allowing growth of Δfrh/ΔfpoF double mutant on acetate; however, this conjecture remains to be tested.

Materials and Methods

Sequence Analysis.

All sequence data are from publicly available genomes (11, 16, 20, 21). The Integrated Microbial Genome (IMG) system was used to identify orthologs and assess the genomic context of genes (22).

Strains, Media, and Growth Conditions.

The construction and genotype of all Methanosarcina strains is presented in Table S2. Methanosarcina strains were grown as single cells at 37 °C in high-salt broth medium (44) or on agar-solidified medium as described (45). Standard conditions were used for growth of E. coli strains DH5α/λ-pir and DH10B (Stratagene) (46), which were used as hosts for plasmid constructions.

DNA Methods, Plasmid, and Strain Construction.

Standard methods were used for plasmid DNA isolation and manipulation in E. coli (47). Liposome-mediated transformation was used for Methanosarcina as described (48). Genomic DNA isolation and DNA hybridization were as described (44, 45, 49). DNA sequences were determined from double-stranded templates by the W.M. Keck Center for Comparative and Functional Genomics, University of Illinois. Plasmid constructions are described in the SI Text (Tables S3 and S4).

Characterization of Mutants in Terms of Growth Characteristics and CH₄ and CO₂ Production in Cell Suspensions.

Growth was quantified by measuring OD₆₀₀. Generation times were calculated during exponential growth phase, and growth yield was determined by measuring the maximal OD₆₀₀ of the culture. Growth curve and cell suspension experiments were performed as described in SI Text.

Quantitative RT-PCR.

Gene-specific primers (Table S3) were designed using Primer Express Software v2.0 (Applied Biosystems). A 1-step qRT-PCR was performed using SuperScript III Platinum SYBR Green One-Step qRT-PCR Kit with ROX (Invitrogen) in the ABI PRISM 7900HT Sequence Detection System (Applied Biosystems). The relative standard curve method was used to quantify expression of fpo (ABI PRISM 7700 Sequence Detection System User Bulletin #2) using rpoA1 as the reference gene (described in Supporting Information).