Unique H2-utilizing lithotrophy in serpentinite-hosted systems

Serpentinization of ultramafic rocks provides molecular hydrogen (H2) that can support lithotrophic metabolism of microorganisms, but also poses extremely challenging conditions, including hyperalkalinity and limited electron acceptor availability. Investigation of two serpentinization-active systems reveals that conventional H2-/CO2-dependent homoacetogenesis is thermodynamically unfavorable in situ due to picomolar CO2 levels. Through metagenomics and thermodynamics, we discover unique taxa capable of metabolism adapted to the habitat. This included a novel deep-branching phylum, “Ca. Lithacetigenota”, that exclusively inhabits serpentinite-hosted systems and harbors genes encoding alternative modes of H2-utilizing lithotrophy. Rather than CO2, these putative metabolisms utilize reduced carbon compounds detected in situ presumably serpentinization-derived: formate and glycine. The former employs a partial homoacetogenesis pathway and the latter a distinct pathway mediated by a rare selenoprotein—the glycine reductase. A survey of microbiomes shows that glycine reductases are diverse and nearly ubiquitous in serpentinite-hosted environments. “Ca. Lithacetigenota” glycine reductases represent a basal lineage, suggesting that catabolic glycine reduction is an ancient bacterial innovation by Terrabacteria for gaining energy from geogenic H2 even under hyperalkaline, CO2-poor conditions. Unique non-CO2-reducing metabolisms presented here shed light on potential strategies that extremophiles may employ for overcoming a crucial obstacle in serpentinization-associated environments, features potentially relevant to primordial lithotrophy in early Earth.

The Candidatus Lithacetigenota is defined by five metagenome bins (HKB111, HKB210, BS525, BS5B28, and GPS1B18) recovered using culture-independent metagenomics from two serpentinite-hosted systems (Hakuba Happo hot springs in Hakuba, Japan, and The Cedars springs in California, USA). The metagenome bins are deposited in the National Center for Biotechnology Information (NCBI) under BioProject number PRJNA453100 and WGS numbers QLUP00000000, QLUQ00000000, QLTW00000000, QLTX00000000, and QLTY00000000. See also Table S3 for the quality and completeness of each bin-genome. Genomic analysis predicts that Ca. Lithacetigenota bacteria have the capacity to produce acetate from inorganic substrates.
Description of Candidatus Lithacetigena gen. nov.
The type species is Candidatus Lithacetigena glycinireducens with two metagenome bins (HKB111 and HKB210) recovered using metagenomics from Hakuba Happo hot springs in Hakuba, Japan.
The genome of this candidatus species was discovered in Hakuba Happo hot springs in Hakuba, Japan.
The type species is Candidatus Psychracetigena formicireducens with three metagenome bins (BS525, BS5B28, and GPS1B18) recovered using metagenomics from The Cedars springs in California, USA.
The genome of this candidatus species was discovered in The Cedars springs in California, USA. Figure S1. Thermodynamics of H2-oxidizing reduction of CO2 (black), formate (pink), and glycine (blue), and formate disproportionation (orange) with changing H2 and total inorganic carbon (TIC) concentrations. The maximum observed H2 concentration in Hakuba Happo hot springs (664 µM) is marked (triangles) in the H2-based plots and the TIC concentrations observed in Hakuba (<0.1 µM) and The Cedars GPS1 (35 µM) are indicated correspondingly. See also Tables S1 and S2 and Supplementary Equations. Figure S2. Thermodynamics of H2-oxidizing CO2-reducing homoacetogenesis under differing pH, temperature, H2, and total inorganic carbon (TIC) concentrations. (A) For temperatures 18 (blue), 45 (green), 85 (orange), and 115 (red) °C, the H2 and TIC concentrations at which H2/CO2 homoacetogenesis has a Gibbs free energy yield (∆G) of -10 kJ mol -1 is shown for pH of 9 (dotted line) and 11 (solid line) (atmospheric pressure of 1 atm). For reference, the same is shown for 25 °C at pH 7 (gray solid line). (B) The ∆G of H2/CO2 homoacetogenesis in various serpentinite-hosted systems (Hakuba Happo hot springs -HS; The Cedars springs -CS; Lost City -LC; Voltri Massif -VM; Coast Range Ophiolite Microbiological Observatory -CR; Santa Elena ophiolite -SEO; Table Lands -TLE) are shown based on reported environmental conditions for individual sampling locations. Each data point is colored based on temperature: psychrophilic (blue), mesophilic (purple), thermophilic (green), and hyperthermophilic (orange). Samples with associated acetate measurements are labelled black, and those that have >2 µM acetate are circled. For samples with no reported acetate concentrations (gray), the average of reported concentrations was used (8.57 µM Acetate). For The Cedars spring sample, no H2 concentration has been reported, so the highest on-land serpentinite-hosted system H2 concentration was used (664 µM H2 from Hakuba Happo #1). Thermodynamic calculations were performed using ∆G°f and ∆H°f values at 298 K values and temperature adjustment through the Gibbs-Helmholtz equation. The effect of pressure was approximated as described by Wang et al. (Wang et al 2010). See also Tables S1 and S2 and Supplementary Equations. Figure S3. Ribosomal protein tree for high-quality MAGs. Universally conserved ribosomal proteins were collected from each genome, aligned with MAFFT v7.394 (Katoh et al 2005), trimmed with trimAl 1.2rev59 (-gt 0.70) (Capella-Gutiérrez et al 2009), and concatenated. A maximum likelihood tree was calculated using phyML 3.3.20190321 with the LG model and 100 bootstrap iterations (Guindon and Gascuel 2003). GTDBtk-based phylogeny is shown. Figure S4. Ribosomal protein tree including high-quality MAGs from 74 GTDB-defined phylum-level lineages. See Fig. 1 for details. GTDBtk-based phylogenetic nomenclature is shown. Figure S5. Schemes for energy metabolism. Arrow colors indicate oxidative (pink), reductive (blue), ATP-yielding (orange), and ATP-consuming (green) steps. The corresponding electron carrier (reduced form) is shown for each redox step. Redox reactions shown in gray are necessary for generating reducing equivalents for biosynthesis (e.g., if the putative catabolic pathway does not produce NADPH, an NADPH-generating pathway is shown in gray). Abbreviations: NAD -nicotinamide adenine dinucleotide, NADP -nicotinamide adenine dinucleotide phosphate, Fd -ferredoxin, THFtetrahydrofolate, CoA -Coenzyme A, Rnf -Ion-translocating NADH:ferredoxin oxidoreductase, Nfn -NADH-dependent ferredoxin:NADP + oxidoreductase, *OR -putative NADH-dependent NADPH:ferredoxin oxidoreductase (see supplemental results). † For the 5,10-methylenetetrahydrofolate dehydrogenase reaction, NAD + /NADH is shown as the electron carrier as has been observed for Acetobacterium woodii. Figure S6. Complete phylogenetic tree of glycine reductase subunits GrdBE and homologs from Hakuba Happo hot spring*, The Cedars springs † , and other serpentinite-hosted system metagenomes # . See Fig. 3 for details. Phylogenetic nomenclature indicated with an asterisk are GTDB-defined phylumlevel nomenclature. For those with underscores, the name is abbreviated to the last capital letter (e.g., Firmicutes_A is shown as (A) and Desulfobacterota_E is shown as (E)).  Culture medium based on Widdel medium (pH 10) with an N2-CO2 (80:20, vol/vol) headspace was supplemented with 0.01 g l -1 yeast extract and 25 g l -1 elemental iron granules. Hakuba hot spring water 100 mL was passed though a membrane filter and the filter was submerged in the culture medium. After a 4 month incubation at 40 °C, 1 mL of the culture was used for DNA extraction, PCR amplification, and clone library construction (about 600 bp of 16S rRNA gene). UBA1414-derived 16S rRNA gene fragments, which shared high sequence identity (>99%) to bin HKB206, comprised 7 out of 20 clones. The remaining 13 clones consisted of obligately aerobic Methylobacterium-and Pseudomonas-related sequences that may be contaminants, but further investigation is required. (B) Micrographs: phase-contrast (left) and SYBR-Green-Istained microbial cells (green, right); scale bar, 2 μm.

Supplementary Equations
Gibbs-Helmholtz equation: ./ van't Hoff equation: Pressure adjustment equation (optional): ./ ΔG1 : Gibbs free energy of the reaction at in situ temperature ΔG2 : Gibbs free energy of the reaction at the reference temperature R : Universal gas constant T1 : In situ temperature (Kelvin) T2 : Reference temperature (Kelvin) ΔH : Molar enthalpy change of the reaction at the reference temperature Keq1 : Equilibrium constant at in situ temperature Keq2 : Equilibrium constant at the reference temperature ΔV : Molar volume change in the reaction at the reference temperature and pressure P1 : In situ pressure (atm) P2 : Reference pressure (atm)  Table S3. Phylogeny and quality of bins from Hakuba Happo hot springs and The Cedars springs (GPS1 and BS5). Phylogeny was defined by GTDBtk (g) or comparing GTDBtk-defined phylogeny with EMBL (e) or SILVA (s). GTDBtk annotations with RED values less than 0.5 were not considered and phylogeny was checked by constructing a concatenated ribosomal protein tree (see Supplement). * Low quality bins that were only used for comparative purposes (e.g., whether a function found in a high-quality bin from HKB1 is present in HKB2 with >99% similarity)   SRB2  SRB2  HKB210  BS5B28  HKB206  GPS1B11  HKB214  GPS1B09  HKB109  BS5B11  HKB212  GPS119  GrdE-TrxA DDT22_00297  DDT41_00874  -----

ATP synthase (A-type)
ATP synthase (F-type) " "E.M.C..GC..G" (only retained fully conserved residues and positions with low variation [≤3 residue types]). † though not all components of the putative NADPH-dependent oxidoreductase found in Lithoacetigenota DDT41 is present in Lithoacetigenota genome DDT22, one gene with high homology (84.15% sequence identity) is detected at the edge of a contig, suggesting that the remaining two genes may be present in the DDT22 genome but was not assembled/recovered. --<dl <dl, below our quantification limit; -, no data. Table S6. Environmental paramers and chemical coposition of Hakuba Happo spring water (artificially pumped from a drilling well named Happo #3) used for microbiological analysis.