Diversification of methanogens into hyperalkaline serpentinizing environments through adaptations to minimize oxidant limitation

Abstract

Metagenome assembled genomes (MAGs) and single amplified genomes (SAGs) affiliated with two distinct Methanobacterium lineages were recovered from subsurface fracture waters of the Samail Ophiolite, Sultanate of Oman. Lineage Type I was abundant in waters with circumneutral pH, whereas lineage Type II was abundant in hydrogen rich, hyperalkaline waters. Type I encoded proteins to couple hydrogen oxidation to CO₂ reduction, typical of hydrogenotrophic methanogens. Surprisingly, Type II, which branched from the Type I lineage, lacked homologs of two key oxidative [NiFe]-hydrogenases. These functions were presumably replaced by formate dehydrogenases that oxidize formate to yield reductant and cytoplasmic CO₂ via a pathway that was unique among characterized Methanobacteria, allowing cells to overcome CO₂/oxidant limitation in high pH waters. This prediction was supported by microcosm-based radiotracer experiments that showed significant biological methane generation from formate, but not bicarbonate, in waters where the Type II lineage was detected in highest relative abundance. Phylogenetic analyses and variability in gene content suggested that recent and ongoing diversification of the Type II lineage was enabled by gene transfer, loss, and transposition. These data indicate that selection imposed by CO₂/oxidant availability drove recent methanogen diversification into hyperalkaline waters that are heavily impacted by serpentinization.

Fig. 1. The estimated relative abundances of Methanobacterium MAGs in communities from subsurface fracture waters collected from wells intersecting the Samail Ophiolite in 2017 (A) and phylogenomic reconstruction of Oman methanogen metagenome assembled genomes (MAGs) in relation to representative Methanobacterium genomes (B).

In A, wells are listed in order of ascending pH from left to right with pH values indicated below the name of each well in parentheses. Contigs with ≥98% nucleotide identity to the Methanobacterium Type I MAG from WAB188 were classified as Type I and contigs with ≥98% nucleotide identity to the Methanobacterium Type II MAG from NSHQ14C were classified as Type II. Estimated relative abundances are shown as the number of reads affiliated with Methanobacterium Type I or II MAGs as a percentage of the total reads in each metagenome. In B, pink text depicts the Type I MAG and cyan text depicts Type II MAGs. All bootstrap values of displayed nodes are >980 out of 1000 bootstrap replicates. Clade-level triangles indicate the phylogenetic diversity within each group via side lengths that are proportional to the distances between the clade’s most closely related and furthest related taxa. Branch length is relative to the scale provided at the bottom of the figure indicating the expected number of substitutions per site. Filled boxes to the right of terminals indicate the presence of genes in reconstructed MAGs or genomes, whereas empty boxes indicate absence of genes in reconstructed MAGs or genomes. Frh coenzyme F₄₂₀-reducing (Group 3a) [NiFe]-hydrogenase, Mvh methyl viologen (Group 3c) [NiFe]-hydrogenase, Mrp-Mbh multiple resistance and pH adaptation module—membrane-bound (Group 4) [NiFe]-hydrogenase protein complex.

Fig. 2. Proposed hydrogenotrophic and formatotrophic methanogenesis pathways in Methanobacterium Type I (A) and Type II (B) populations, respectively, from the Samail Ophiolite, Oman.

Enzymes that differed between the two MAGs are denoted in the figures. Eha Group 4d energy-converting [NiFe]-hydrogenase A, Ehb Group 4d energy-converting [NiFe]-hydrogenase B, Fd ferredoxin, Fdh formate dehydrogenase, Frh coenzyme F₄₂₀-reducing (Group 3a) [NiFe]-hydrogenase, Hdr heterodisulfide reductase, MFR methanofuran, MPT methanopterin, Mrp-Mbh multiple resistance and pH adaptation module—membrane bound (Group 4) [NiFe]-hydrogenase protein complex, Mvh methyl viologen-reducing (Group 3c) [NiFe]-hydrogenase. Figure adapted from Boyd et al. [1].

Fig. 3. Genes inferred to be co-localized with those coding for methyl viologen-reducing (Group 3c) [NiFe]-hydrogenase (Mvh) and formate dehydrogenase (Fdh) in Type I and II MAGs (A) and the putative protein complexes they form to bifurcate electrons from H₂ (B) or formate (C), respectively, to simultaneously reduce ferredoxin and heterodisulfide.

Percentages below each gene indicate amino acid identities between homologs encoded by Type I and Type II MAGs. In panels B and C, squares represent [4Fe-4S] clusters and triangles represent [2Fe-2S] clusters. CoB coenzyme B, CoM coenzyme M, FAD flavin adenine dinucleotide, Fd ferredoxin, Fdh formate dehydrogenase, Hdr heterodisulfide reductase, Mvh methyl viologen-reducing (Group 3c) hydrogenase. Figure modified from Thauer et al. [71] and Costa et al. [77].

Fig. 4. Potential rates of biological methanogenesis from formate and bicarbonate by planktonic microbial communities in well water samples collected from the Samail Ophiolite in 2020.

Potential rates of biological substrate transformation were determined via microcosm assays using well waters collected from the Samail Ophiolite in 2020. Results are plotted on a logarithmic scale. The average rates of methane generation observed in four replicate abiological controls were subtracted from values in four replicate biological assays (Avg) and their combined standard deviations (SD) are presented at five timepoints over an 8-week incubation. P-values were determined between biological assays and abiological controls at each timepoint via Student’s t-test assuming unequal variance for each condition (*p < 0.05, **p < 0.01).

Fig. 5. Average nucleotide identities (ANIs) between Methanobacterium single amplified genomes (SAGs) and Type II metagenome assembled genomes (MAGs).

In A, ANIs relative to the NSHQ14C MAG are presented with blue circles representing individual SAGs and the red circle representing the NSHQ14B MAG. In B, pairwise ANIs among the SAGs and the NSHQ14C MAG were used to generate a dendrogram based on hierarchical clustering, with branch lengths (“Height”) representing Euclidean distances between ANIs calculated among the genomes. Red text represents the Type II MAGs and letters followed by two-digit numbers represent individual SAGs.

Fig. 6. ISNCY transposase protein phylogeny and genes co-localized with this transposase in the NSHQ14C metagenome assembled genome (MAG) and single cell genomes (SAGs).

Branch length is relative to the scale provided at the top of the figure indicating average substitutions per site. Bootstrap values are displayed at each node (out of 100 replicates). Contigs that encoded this transposase in each SAG or MAG are depicted to the right of the terminals. Red arrows represent genes encoding transposase orthologs, white arrows represent genes encoding hypothetical proteins, black arrows represent genes encoding CRISPR-associated (Cas) proteins, and the remaining colored arrows represent genes encoding other functional proteins. Gene lengths are relative to the scale provided at the top of the figure.