Evolutionary constraints and genomic plasticity of a distinct thaumarchaeotal group 1.1c lineage from CO2-rich groundwater

Thaumarchaeotal Group 1.1c (family Gagatemarchaeaceae) comprises deeply branching, non-ammonia-oxidizing archaea detected in acidic soils and subsurface habitats. However, their roles in groundwater microbiomes and the evolutionary processes shaping their genomes remain poorly understood. At a site previously classified as CO 2 -rich, we examined Group 1.1c in acidic (pH 4.8), microoxic groundwater with a measured temperature of 7.6 °C and placed the recovered genomes within a lineage-wide comparative framework. Among 789 archaeal SSU rRNA gene assignments recovered from metagenomic reads, 67.2% were classified as Group 1.1c. We reconstructed three metagenome-assembled genomes (85.3–97.5% complete; 0.07–3.28% contamination) and compared them with 33 public Group 1.1c genomes. The groundwater metagenome-assembled genomes (MAGs) lacked amoABC and the canonical 3-hydroxypropionate/4-hydroxybutyrate carbon-fixation pathway but encoded genes associated with carbohydrate utilization, fatty-acid degradation, central heterotrophic metabolism, an aa 3 -type cytochrome c oxidase, and superoxide dismutase, consistent with heterotrophic metabolism, aerobic respiratory potential, and oxidative-stress management. Phylogenomics resolved five subgroups (G1–G5) that differed in accessory-gene repertoires and synonymous-site composition. The two G1 MAGs had lower genomic GC content and weaker codon-usage bias than G2–G5. Available source metadata showed that all four G2 MAGs were recovered from groundwater, including the three CO₂-rich groundwater MAGs reconstructed here; seven of nine G4 MAGs originated from temperate grasslands, whereas G1 and G5 were represented by soil- or peat-derived MAGs and G3 spanned peat, soil, sediment, and groundwater sources. Because environmental measurements were not harmonized across the public MAGs, these patterns were treated descriptively rather than as evidence of habitat-specific ecotypes. The fitted recombination model indicated a lower contribution of recombination than mutation to nucleotide change (median r/m = 0.264; central 95% range across 100 emsim draws, 0.253–0.275), with kilobase-scale inferred importation tracts. Fixed-coefficient tAI estimates were lowest in G1, but this pattern was not robust to coefficient optimization, and its interpretation was limited by the incompleteness, fragmentation, and sparse tRNA recovery of the two available G1 MAGs; it was therefore treated as descriptive rather than as evidence of translational adaptation. Of 208 one-to-one relaxed-core candidates, 135 passed alignment and taxon-coverage filters; all had gene-wide dN/dS estimates below 1 (median ω = 0.046), including the 14 genes meeting both the strict starting-value stability and no-saturation criteria. Group 1.1c accounted for a large proportion of archaeal SSU rRNA gene assignments in the sampled groundwater, and the recovered MAGs encoded metabolic capacities consistent with heterotrophic metabolism, aerobic respiratory potential, and oxidative-stress management. Gene-wide dN/dS estimates were consistent with broad purifying constraint across the model-estimable component of the relaxed core, whereas accessory-gene and synonymous-site variation accompanied deep lineage diversification. These data expand the genomic representation of groundwater Group 1.1c but do not demonstrate direct adaptation to acidic, microoxic, or CO 2 -influenced conditions.