Chemoautotrophy in subzero environments and the potential for cold-adapted Rubisco

Abstract

The act of fixing inorganic carbon into the biosphere is largely facilitated by one enzyme, Rubisco. Beyond well-studied plants and cyanobacteria, many bacteria use Rubisco for chemolithoautotrophy in extreme environments on Earth. Here, we characterized the diversity of autotrophic pathways and chemolithoautotrophic Rubiscos from two distinct subzero, hypersaline Arctic environments: 40-kyr relic marine brines encased within permafrost (cryopeg brines) and first-year sea ice. The Calvin-Benson-Bassham (CBB) cycle was widely found in both environments, although with different predominant Rubisco forms. From cryopeg brine, reconstructions of metagenome-assembled genomes (MAGs) uncovered four MAGs with the potential for chemolithoautotrophy, of which the CBB-containing genus Thiomicrorhabdus was most abundant. A broader survey of Thiomicrorhabdus genomes from diverse environments identified a core complement of three Rubisco forms (II, IAc, IAq) with a complex pattern of gain and loss, with form II constitutively present in genomes from subzero environments. Using representative kinetic data, we modeled carboxylation rates of Rubisco forms II, IAc, and IAq across CO₂, O₂, and temperature conditions. We found that form II outcompetes form I at low O₂, but cold temperatures minimize this advantage. Inspection of form II from genomes from cold environments identified signals of potential thermal adaptation due to key amino acid substitutions, which resulted in a more exposed active site. We argue that subzero form II from Thiomicrorhabdus warrants further study as it may have unique kinetics or thermal stability. This work can help address the limits of autotrophic functionality in extreme environments on Earth and other planetary bodies.IMPORTANCEAutotrophy, or the fixation of inorganic carbon to biomass, is a key factor in life's ability to thrive on Earth. Research on autotrophy has focused on plants and algae, but many bacteria are also autotrophic and can survive and thrive under more extreme conditions. These bacteria are a window to past autotrophy on Earth, as well as potential autotrophy in extreme environments elsewhere in the universe. Our study focused on dark, cold, saline environments, which are likely to be found on Enceladus and Europa, as well as in the Martian subsurface. We found evidence for potential cold adaptation in a key autotrophic enzyme, Rubisco, which could expand the known boundaries of autotrophy in rapidly disappearing icy environments on Earth. We also present a novel model framework that can be used to probe the limits of autotrophy not only on Earth but also on key astrobiological targets like Enceladus and Europa.

Keywords: Rubisco; astrobiology; autotrophs; chemoautotrophy; psychrophiles.

Fig 1

Rubisco forms in metagenomes from cryopeg brines and sea ice. (A) Relative abundance of key genes from carbon fixation pathways. Abundance calculated from mapped reads to gene-encoding scaffolds (see Table S1 for the list of genes used). CB metagenomes from cryopeg brine; SB metagenomes from sea-ice brine; SI metagenomes from melted sea ice. See Reference for metagenome details. Abbreviations: 4-HB, 4-hydroxybutyrate/dicarboxylic acid cycle and 3-hydroxyproprionate/4-hydroxybutyrate cycle; others shared with Table 1. (B) Proportion of bacterial Rubisco abundance attributed to each form within cryopeg brine and sea-ice metagenomes. Abundance calculated from mapped reads to Rubisco-encoding scaffolds from bacteria only. No form III was detected in either cryopeg brine or sea-ice samples. The number of sequences and the proportion of read depth attributed to form I subforms and form II in Fig. S4.

Fig 2

Phylogenetic tree of Thiomicrorhabdus genomes alongside Rubisco gene complement and results from amino acid thermal analysis. (A) Maximum-likelihood phylogenetic tree of 22 Thiomicrorhabdus genomes. H. marinus and H. thermophila designated as outgroup. Bootstrap values for each node are depicted on the branch. MAGs designated with (*). Genomes added in this study designated with (**). Highlighting in blue shows species found in stable cold environments or with reported growth at ≤5°C. Thiosulfativibrio (Tsv.) zosterae reclassified after inclusion. (B) Presence/absence of Rubisco forms and result of thermal adaptation analysis. Shape represents the presence of form: square, form II; circle, form IAc; triangle, form IAq. Color indicates the result of thermal analysis: blue, cold-inferred; red, hot-inferred; outline only, neutral; gray, used as reference for thermal analysis. See Fig. S6 for an expanded Clade 3 that includes two other genomes in Clade 3 that passed the quality control but did not contain a complete Rubisco gene.

Fig 3

Theoretical kinetics model of Rubisco net carboxylation. Modeled Rubisco carboxylation speed for forms II–IAc at varying CO₂ concentration, O₂ concentration, and temperature. Red indicates form II > IAc, blue indicates form IAc > II, with shading indicating the magnitude in carbons per second (C/s). Gray shading at low CO₂ indicates conditions where there was a net loss of carbon for both enzymes. (A) and (C) show gross carboxylation rate equation (R_C); (B) and (D) show carboxylation rate when the PGS pathway (C2 cycle) is included at 0°C and 25°C, respectively. All graphs share a color scale.

Fig 4

Partial amino acid alignments of Rubisco form II. Rubisco form II sequence excerpts from H. marinus, T. frisia Kp2, T. arctica, T. arctica UC1, and T. sp. ZW0627 around common, prominent substitutions/gaps found within the cold-inferred sequences (blue) compared to the neutral H. marinus and T. frisia Kp2. Amino acids highlighted in red show deviation from the H. marinus sequence.

Fig 5

Rubisco form II structural differences between T. arctica UC1 and H. marinus. (A) AlphaFold-generated structures for H. marinus (transparent white) and T. arctica UC1 (dark blue) Rubisco form II dimers. Active sites are colored red. Arrows point to the additional exposed active site residues in UC1 compared to H. marinus. The missing residues in UC1 at amino acid positions 78–82 (Fig. 4) are colored yellow on H. marinus structure (a close-up view in B). (C) UC1 form II dimer with residue differences to H. marinus that confer a semi-conservative or non-conservative change highlighted in orange.