Roles of Coenzyme F₄₂₀-Reducing Hydrogenases and Hydrogen- and F₄₂₀-Dependent Methylenetetrahydromethanopterin Dehydrogenases in Reduction of F₄₂₀ and Production of Hydrogen during Methanogenesis

Abstract

Reduced coenzyme F₄₂₀ (F₄₂₀H₂) is an essential intermediate in methanogenesis from CO₂. During methanogenesis from H₂ and CO₂, F₄₂₀H₂ is provided by the action of F₄₂₀-reducing hydrogenases. However, an alternative pathway has been proposed, where H₂-dependent methylenetetrahydromethanopterin dehydrogenase (Hmd) and F₄₂₀H₂-dependent methylenetetrahydromethanopterin dehydrogenase (Mtd) together reduce F₄₂₀ with H₂. Here we report the construction of mutants of Methanococcus maripaludis that are defective in each putative pathway. Their analysis demonstrates that either pathway supports growth on H₂ and CO₂. Furthermore, we show that during growth on formate instead of H₂, where F₄₂₀H₂ is a direct product of formate oxidation, H₂ production occurs. H₂ presumably arises from the oxidation of F₄₂₀H₂, and the analysis of the mutants during growth on formate suggests that this too can occur by either pathway. We designate the alternative pathway for the interconversion of H₂ and F₄₂₀H₂ the Hmd-Mtd cycle.

The methanogenic Archaea (methanogens) occupy a variety of anaerobic habitats, where they play essential roles in the conversion of hydrogen and other intermediates to methane (10). The hydrogenotrophic methanogens use hydrogen to reduce CO₂ to methane. In addition, some hydrogenotrophs use formate, and a few substitute certain low-molecular-weight alcohols for hydrogen.

The deazaflavin F₄₂₀ is an essential coenzyme of methanogenesis. The reduction of CO₂ to methane requires reduced F₄₂₀ (F₄₂₀H₂), since it is the sole electron donor for the step that reduces methylenetetrahydromethanopterin (methylene-H₄MPT) (Mer in Fig. 1). In addition, F₄₂₀H₂ is the electron donor for F₄₂₀H₂-dependent methylenetetrahydromethanopterin dehydrogenase (Mtd), one of two enzymes that reduce methenyl-H₄MPT. The other enzyme, H₂-dependent methylenetetrahydromethanopterin dehydrogenase (Hmd), uses H₂ directly. mRNA abundance for mtd increased markedly under hydrogen-limited growth conditions (4), suggesting that Mtd may be more important when H₂ is limiting.

FIG. 1.

The hydrogenotrophic methanogenic pathway. See reference 3 for a full description of methanogenesis. The Hmd-Mtd cycle is boxed. Abbreviations: CoB, coenzyme B; CoM, coenzyme M; F₄₂₀, coenzyme F₄₂₀; Fd, ferredoxin; Frc, cysteine-containing F₄₂₀-reducing hydrogenase; Fru, selenocysteine-containing F₄₂₀-reducing hydrogenase; Mer, methylenetetrahydromethanopterin reductase; MFR, methanofuran.

The F₄₂₀-reducing hydrogenases (Fru and Frc) reduce F₄₂₀ with H₂. However, an alternative route for this process has been proposed. In Methanothermobacter marburgensis the specific activity of F₄₂₀-reducing hydrogenase, a Ni-Fe hydrogenase, decreased 20-fold under nickel-limited growth conditions. In contrast, the specific activities of Hmd and Mtd, neither of which requires nickel for activity, increased six- and fourfold, respectively (1). These observations led to the proposal that under nickel-limited conditions, F₄₂₀ may be reduced by the concerted action of Hmd and Mtd, the former working in the forward direction (with respect to the methanogenic pathway) and the latter in the reverse direction (1, 2). This pathway is boxed in Fig. 1.

Here we report on the properties of mutants of Methanococcus maripaludis that are deficient in Hmd, Mtd, or the F₄₂₀-reducing hydrogenases. The results demonstrate that neither Hmd nor Mtd is essential, confirming that either enzyme is sufficient for methenyl-H₄MPT reduction. The results also indicate that, in vivo, Hmd and Mtd do indeed constitute an alternate pathway for the reduction of F₄₂₀ with H₂, which we designate the Hmd-Mtd cycle. Furthermore, we show that during growth on formate, H₂ production occurs, evidently by reversal of either the F₄₂₀-reducing hydrogenase or the Hmd-Mtd cycle.

MATERIALS AND METHODS

Growth of strains and measurement of H₂.

M. maripaludis was grown on H₂ and CO₂ by standard anaerobic techniques in McCas medium as described elsewhere (6). For growth on formate, McCas medium was modified to contain 200 mM sodium formate and 200 mM MOPS (morpholinepropanesulfonic acid) buffer (pH 7.0), NaCl was decreased to 0.18 M, and the gas atmosphere was 80% N₂ and 20% CO₂ at a pressure of 15 lb/in². Cultures (5-ml volume) were inoculated with 0.25 to 0.5 ml of a culture actively growing on formate. Growth was monitored by optical density at 660 nm. The accumulation of H₂ in the headspace (20-ml volume) was measured using a Hach CARLE Series 100 AGC gas chromatograph equipped with a Supelco 60/80 mesh molecular sieve 5A column (6 ft by 1/8 in.) and a trace analytical RGD2 reduction gas detector.

Construction of plasmids and strains.

Primers are listed in Table 1, and strains and plasmids are listed in Table 2. PCR products containing the genes hmd, mtd, frcA, and fruA and their flanking regions were generated using the primer pairs hmdcln5for and hmdcln5rev, mtdcln5for and mtdcln5rev, frcAfor2 and frcArev2, and fruAfor and fruArev, respectively. The products were cloned into pCR2.1topo to generate phmdtopo, pmtdtopo, pfrcAtopo, and pfruAtopo. An in-frame deletion of hmd was produced by PCR of phmdtopo using primers hmddel1 and hmddel2, followed by digestion with AscI and ligation to produce phmddeltopo. pmtddeltopo, pfrcAdeltopo, and pfruAdeltopol were generated in the same way using pmtdtopo and the primers mtddel1 and mtddel3, pfrcAtopo and the primers frcdel1 and frcdel2, and pfruAtopo and the primers frudel1 and frudel2, respectively. The in-frame deletion of hmd was amplified from phmddeltopo using the primers hmddelamp1 and hmddelamp3; the resulting fragment was digested with BamHI and ligated into the vector pCRprtneo to produce pCRprtΔhmdneo. pCRprtΔmtdneo was produced in the same way from pmtddeltopo using the primers mtddelamp1 and mtddelamp2 and digesting with BamHI. pCRprtΔfrcneo was produced from pfrcAdeltopo using frcdelamp5 and frcdelamp6 and digesting with XbaI, and pCRprtΔfruneo was produced from pfruAdeltopo using frudelamp5 and frudelamp6 and digesting with XbaI.

TABLE 1.

Primers

Name	Sequence	Restriction site
hmdcln5for	GCTGTTGGAATAGACTGCTG
hmdcln5rev	GCCCTTATTACTTCTTTTCC
mtdcln5for	CGTTTCAGCAGGTTCGAAGG
mtdcln5rev	GGGTGTTGCATTAATTGGCG
frcAfor2	GCACCTCTTTAAAAGCTTT
frcArev2	AATGAAACAGCGCCATCTAC
fruAfor	CCAGTACTTCAATATCTTTCAC
fruArev	TACTTCTTCTGACAACCGAC
hmddel1	AGGCGCGCCACTTTCATATCATACACCTCA	AscI
hmddel2	AGGCGCGCCCAATAAAACCTTAAGTATTAC	AscI
mtddel1	AGGCGCGCCCATTATATCACCGAAAGATAT	AscI
mtddel2	GGGCGCGCCAGAATAAATTTGCATCAAAAT	AscI
frcdel1	GGCGCGCCTTACCCATCAGATCACCTATC	AscI
frcdel2	GGCGCGCCAATAAATACTGGTGAATCATGC	AscI
frudel1	GGCGCGCCACTTTATTCACCTCC	AscI
frudel2	GGCGCGCCAATTCTAAATTCCTGAAAAGG	AscI
hmddelamp1	ATGGATCCGGCTTGCTGTTGGAATAGAC	BamHI
hmddelamp3	TTGGATCCGCCCTTATTACTTCTTTTCC	BamHI
mtddelamp1	GAGCTCGGATCCACTAGTAACGGCCGCCAAGTGT	BamHI
mtddelamp2	AGAATTGGATCCCGTTTCAGCAGGTTCGAAGGA	BamHI
frcdelamp5	ATCTAGAGCACCTTCTTTAAAAGCTTT	XbaI
frcdelamp6	CTCTAGAAATGAAACAGCGCCATCTAC	XbaI
frudelamp5	TTCTAGACCAGTACTTCAATATCTTTCAC	XbaI
frudelamp6	CTCTAGATACTTCTTCTGACAACCGAC	XbaI

TABLE 2.

Strains and plasmids

Name	Feature(s)	Source or reference
Plasmids
pCR2.1topo	Ampr Kanr cloning vector	Invitrogen
pCRprtneo	hmv-promoter-hpt fusion + Neor cassette in pCR2.1topo	6
phmdtopo	hmd plus flanking DNA in pCR2.1topo	This study
pmtdtopo	mtd plus flanking DNA in pCR2.1topo	This study
pfrcAtopo	frcA plus flanking DNA in pCR2.1topo	This study
pfruAtopo	fruA plus flanking DNA in pCR2.1topo	This study
phmddeltopo	In-frame deletion of hmd in pCR2.1topo	This study
pmtddeltopo	In-frame deletion of mtd in pCR2.1topo	This study
pfrcAdeltopo	In-frame deletion of frcA in pCR2.1topo	This study
pfruAdeltopo	In-frame deletion of fruA in pCR2.1topo	This study
pCRprtΔhmdneo	In-frame deletion of hmd in pCRprtneo	This study
pCRprtΔmtdneo	In-frame deletion of mtd in pCRprtneo	This study
pCRprtΔfrcneo	In-frame deletion of frcA in pCRprtneo	This study
pCRprtΔfruneo	In-frame deletion of fruA in pCRprtneo	This study
Strains
Mm900	M. maripaludis Δhpt	6
Mm1097	Mm900 Δhmd	This study
Mm1020	Mm900 Δmtd	This study
Mm1183	Mm900 ΔfrcA	This study
Mm1145	Mm900 ΔfruA	This study
Mm1184	Mm900 ΔfrcA ΔfruA	This study

Strains containing markerless in-frame deletions of hmd, mtd, frcA, and fruA were constructed in strain Mm900 as described elsewhere (6) using the plasmids pCRprtΔhmdneo, pCRprtΔmtdneo, pCRprtΔfrcneo, and pCRprtΔfruneo, respectively, to produce strains Mm1097, Mm1020, Mm1183, and Mm1145, respectively. A double mutant of frcA and fruA was constructed by the same procedure from the frcA mutant strain Mm1183 by using pCRprtΔfruneo to produce Mm1184. Deletions were confirmed by Southern analysis. For experiments testing whether hmd deletion mutations could be made, pCRprtΔhmdneo was transformed into recipient strains. The resultant merodiploids were streak purified, allowed to grow overnight without antibiotic selection, and plated on counterselection plates containing 8-azahypoxanthine. Colonies were analyzed by Southern blotting to distinguish strains containing deletions of the hmd gene from those containing the wild-type hmd gene.

RESULTS AND DISCUSSION

F₄₂₀ reduction during growth on H₂.

We used a genetic approach in M. maripaludis to test whether F₄₂₀-reducing hydrogenase and the Hmd-Mtd cycle constitute two alternative pathways for the reduction of F₄₂₀ in vivo. M. maripaludis contains genes for Hmd and Mtd and two sets of genes for F₄₂₀-reducing hydrogenases, fruADGB and frcADGB (5). FruA contains selenocysteine residues, while FrcA contains cysteine residues in corresponding positions, and in the closely related Methanococcus voltae frc expression is repressed in the presence of selenium in the medium (7, 8). We hypothesized that if Hmd and Mtd can provide an alternative pathway for the reduction of F₄₂₀, then mutants with deletions in fru and frc should be viable in the presence of wild-type hmd and mtd. Conversely, mutants with mutations in either hmd or mtd should be viable in a fru⁺ frc⁺ background.

Using H₂ and CO₂ as growth substrates, we made the following mutants, all containing markerless in-frame deletions: ΔfruA, ΔfrcA, double mutant ΔfruA ΔfrcA, Δmtd, and Δhmd strains. ΔfruA ΔfrcA, Δmtd, and Δhmd strains each grew normally on H₂ and CO₂ (Fig. 2A). Since F₄₂₀H₂ is essential for methanogenesis, each mutant must retain a pathway for F₄₂₀ reduction using H₂. Hence, the results imply that F₄₂₀-reducing hydrogenase and the Mtd-Hmd cycle are each sufficient for this function.

FIG. 2.

Growth and H₂ production by wild-type and mutant strains of M. maripaludis. (A) Growth on H₂; (B) growth and H₂ production on formate. For the Δmtd mutant on formate, data from three separate growth experiments are plotted and are represented by a single line. OD₆₆₀, optical density at 660 nm.

As a formal possibility, a third, unknown pathway for the reduction of F₄₂₀, different from the F₄₂₀-reducing hydrogenase and the Mtd-Hmd cycle, could exist. To test this possibility, we attempted to construct an Δhmd mutation in a ΔfruA ΔfrcA background. Following our regular procedure for generation of markerless mutations (6), we introduced Δhmd (containing the N- and C-terminal flanking regions of hmd) on an integrative vector to produce merodiploids of Δhmd and hmd⁺. We made such merodiploids in the ΔfruA ΔfrcA, ΔfrcA, and fru⁺ frc⁺ backgrounds. We then selected for resolution of the merodiploids via a second recombination event and analyzed the resulting strains by Southern blotting. In principle a mixture of wild-type and deletion strains should result, depending on where the second recombination event occurs. We counted the numbers of resulting Δhmd and hmd⁺ strains in each background. In the fru⁺ frc⁺ background six out of eight strains tested contained Δhmd and the remaining two contained hmd⁺. In the ΔfrcA background, which should express fru and therefore retain active F₄₂₀-reducing hydrogenase, three strains contained Δhmd and five contained hmd⁺. In contrast, in the ΔfruA ΔfrcA background all 40 strains tested contained only hmd⁺. The results indicate that while Hmd can be eliminated in a strain with active F₄₂₀-reducing hydrogenase, it is essential in a strain lacking F₄₂₀-reducing hydrogenase. Therefore, no evidence could be found for the existence of a third pathway that would produce F₄₂₀H₂ from H₂.

H₂ production during growth on formate.

Growth on formate differs from growth on H₂ and CO₂ because F₄₂₀H₂ is a direct product of formate oxidation (Fig. 1). Neither the F₄₂₀-reducing hydrogenase nor the Mtd-Hmd cycle should be necessary for the production of F₄₂₀H₂. However, the reversal of either pathway might result in H₂ production. We characterized the growth of the ΔfruA ΔfrcA, Δmtd, and Δhmd mutants on formate. The ΔfruA ΔfrcA and Δhmd mutants grew normally, while the Δmtd mutant grew after a lag. For each strain, H₂ accumulated in the headspace of the tubes as growth commenced and disappeared when growth ended (Fig. 2B). This observation suggests that H₂ is produced from F₄₂₀H₂ and that either the F₄₂₀-reducing hydrogenase or the Mtd-Hmd cycle can mediate this conversion. H₂ accumulated to a substantially higher level in tubes containing cultures of the Δmtd mutant than in tubes containing any of the other strains. In the Δmtd strain, Hmd is the only enzyme for the reduction of methenyl-H₄MPT. Therefore, H₂ production, which would occur by the action of the F₄₂₀-reducing hydrogenase, should be essential. Due to the relatively low affinity of Hmd for H₂ (9), substantially higher H₂ levels accumulate. In contrast, in the other strains Mtd is present and can use F₄₂₀H₂ for the reduction of methenyl-H₄MPT. These results indicate that H₂ production from F₄₂₀H₂ occurs during growth on formate and that either the F₄₂₀-reducing hydrogenase or the Mtd-Hmd cycle can carry out this process.

Whether H₂ is a necessary intermediate during growth on formate cannot be determined from the present data. The generation of a ΔfruA ΔfrcA Δhmd triple mutant, which is expected to grow in the presence of formate, could resolve this question. Growth of the mutant on formate alone without the addition or generation of H₂ would indicate that H₂ is not a required intermediate. A requirement for added H₂ would indicate that H₂ production is required during growth on formate. Efforts to construct such a mutant are under way.

Concluding remarks.

The genetic approach taken here has shown that two alternative pathways, the F₄₂₀-reducing hydrogenase and the Hmd-Mtd cycle, can function in vivo for the reduction of F₄₂₀ with H₂. Furthermore, during growth on formate the same pathways function in reverse to produce H₂ from F₄₂₀H₂. The lack of growth differences between the wild-type and mutant strains on H₂ and CO₂ (Fig. 2A) suggests that neither pathway for F₄₂₀ reduction was rate limiting. However, in nature the F₄₂₀-reducing hydrogenase may constitute the major pathway when sufficient nickel is present, while the Hmd-Mtd cycle may be important when nickel is limiting (1, 2).