The genomes of 10 Cloacimonadota MAGs were interrogated (58–97% completeness; 0–4.4% contamination), which represent four novel genus-level and eight novel species-level taxa (Table 1 and Supplementary Table S2). Based on phylogenetic analysis and GTDB taxonomy, the four novel genera are deeply nested within the phylum Cloacimonadota, within the class Candidatus Cloacimonadia. None of the four genera were found to be closely related to Ca. Cloacimonas or Ca. Syntrophosphaera, both of which belong to the family Candidatus Cloacimonadaceae (Dyksma and Gallert, 2019; Figure 1 and Supplementary Figure S1).

The genomes of 14 Omnitrophota MAGs were interrogated (61–93% completeness; 0–9.1% contamination) that represent 11 novel genus-level and 13 novel species-level taxa (Table 2 and Supplementary Table S2). Phylogenetic analysis and GTDB taxonomy revealed that the 11 genera represent two class-level and eight order-level clades. None of the Ace Lake MAGs were closely related to Ca. Omnitrophus (class Candidatus Omnitrophia, order Candidatus Omnitrophales), which the phylogenetic analysis recovered in a relatively basal position within the phylum (Figure 2 and Supplementary Figure S2).

For both Cloacimonadota and Omnitrophota, individual genera and species are provided along with etymologies of all proposed names (Tables 1, 2 and Supplementary Table S2). Proteins and pathways discussed for all MAGs assigned to each genus are provided for Cloacimonadota (Supplementary Tables S3, S4) and Omnitrophota (Supplementary Tables S5, S6).

Cloacimonadota are here inferred to be capable of degrading recalcitrant organic matter under anoxic conditions in Ace Lake. The MAGs of the four Ace Lake genera encode multiple glycoside hydrolase (GH) enzymes with signal peptides (indicating an extracytoplasmic location) and include enzymes that degrade polysaccharides and certain glycoconjugates, which would release oligosaccharides and simple sugars such as glucose (Table 1). These hydrolytic enzymes belong to various GH families, indicating a range of potential substrates, such as starch, β-glucans, β-glucosides, chitin, and glucosylceramides (Table 1 and Supplementary Tables S3, S4). The Ace Lake Cloacimonadota also encode diverse proteases and peptidases, including both secreted and cytoplasmic, indicating that polypeptides could be used as amino acid sources (Supplementary Tables S3, S4). Histidine degradation pathways are encoded in MAGs of all four genera, and Ca. Celaenobacter gen. nov. encodes proteins for tryptophan degradation (Supplementary Table S3). The abilities of Cloacimonadota to use complex sugars and proteins as organic substrates have been previously reported for this clade in both anaerobic digestors and lakes (Pelletier et al., 2008; Limam et al., 2014; Suominen et al., 2021).

However, in MAGs of the Ace Lake genus Ca. Stygibacter gen. nov. we identified genes for components of a putative extracellular, multienzyme complex for the binding and degradation of biopolymers (Table 1 and Figure 3A), not previously reported for Cloacimonadota. In Ca. Stygibacter, certain signal-peptide-bearing enzymes contain C-terminal dockerin domains: chitinase homolog (GH18), α-amylase/α-mannosidase homolog (GH57), serine peptidase (Peptidase S8/S53 domain), and gingipain-like peptidase (Peptidase C25). Ca. Stygibacter also encodes a large (3,755 amino acids) non-catalytic scaffoldin-like protein (Artzi et al., 2017) that contains tandemly repeated cohesin and carbohydrate-binding (CBM2/CBM3) domains. We infer that these dockerin- and cohesin-domain proteins combine to produce a cellulosome-like structure, with biopolymer-degrading enzymes integrated into this scaffoldin-like protein via complementary cohesin–dockerin interactions (Artzi et al., 2017).

Canonical cellulosomes are extracellular complexes equipped with cellulose-, hemicellulose-, and polypeptide-targeting enzymes that are used by cellulolytic Clostridia (Firmicutes) to bind and degrade plant cell walls (Schwarz and Zverlov, 2006; Peer et al., 2009; Fontes and Gilbert, 2010). As in clostridial cellulosomes, the Ca. Stygibacter complex has a scaffoldin-like protein with tandem cohesin domains for integration of multiple dockerin-containing enzymes, and a conserved C-terminal domain (CTD) for direct attachment to its own cell surface (Lasica et al., 2017). In common with the clostridial cellulosome, we infer that the cellulosome-like structure in Ca. Stygibacter can mediate attachment of cells to insoluble substrates and promote degradation to soluble products that are taken up by the cell (Lamed et al., 1983). Unlike cellulosomes, we propose that, based on the annotated GHs and peptidases, this complex in Ca. Stygibacter is utilized for the attachment to and degradation of starch (via endohydrolysis) and chitinous and proteinaceous material rather than cellulose-rich plant-derived material. A non-canonical cellulosome-like structure (“planctosome”) has also been reported in certain freshwater Planctomycetes (Nemodlikiaceae) for polypeptide degradation (Andrei et al., 2019), also distinct from the Ca. Stygibacter structure described here; our finding for the Ace Lake Ca. Stygibacter therefore adds to the repertoire of cellulosome-like complexes represented across the domain Bacteria.

The Ace Lake Cloacimonadota MAGs also encode other putative extracellular structures not previously reported for this phylum, all of which indicate an elaborate cell envelope (Figure 3A). Three genera (Ca. Stygibacter, Ca. Tenebribacter gen. nov., Ca. Zophobacter gen. nov.) encode poly-γ-glutamate synthetase (CapBC) and other proteins required for synthesis and transport of poly-γ-glutamate (PGA), a biopolymer involved in capsule formation or released extracellularly as a water-binding component of a biofilm matrix (Rehm, 2010). The water-binding properties of PGA allow it to locally decrease high salt concentrations, allowing survival of the cell in high-salt environments (Kandler et al., 1983; Rehm, 2010). Additionally, the Ca. Stygibacter MAGs encode a glycine-rich protein (944 amino acids), with the N-terminal half containing a ∼440 amino acid region that has 27–30% identity to a ∼400–500 amino acid repeat sequence in halomucin (a very large protein in the halophilic archaeon Haloquadratum walsbyi) (Bolhuis et al., 2006), and the C-terminal half includes a dockerin domain. As proposed for H. walsbyi, it is possible that both PGA and the halomucin-like protein form a water-enriched capsule around the cell that facilitates growth in high concentrations of salt (Bolhuis et al., 2006). These cell envelope features combined with Na⁺-translocating homeostasis abilities (see section “Cloacimonadota: Fermentation”) may assist growth of Cloacimonadota at the bottom of the Ace Lake water column where salinity is at its highest (43 g L^–1) (Rankin et al., 1999; Lauro et al., 2011).

A fermentative metabolism is inferred for the Ace Lake Cloacimonadota, by which energy is derived from the oxidation of organic substrates (sugars, amino acids, 2-oxoacids, aldehydes, alcohols, and formate), and protons are used as the electron acceptor (Calusinska et al., 2010; Table 1 and Figures 3A,B). Simple sugars and amino acids, including those generated by extracellular degradation of biopolymers, appear to be principally imported by secondary transport, although a complete primary transporter for branched-chain amino acids (BCAAs) was identified in Ca. Celaenobacter MAGs. For simple sugars, the Embden–Meyerhof–Parnas (EMP) pathway for glycolysis generates NADH and reduced ferredoxin, as well as anabolic precursors (such as for the pentose phosphate pathway). There is no evidence in any of the 10 Ace Lake Cloacimonadota MAGs of a capacity for propionate oxidation, unlike members of Ca. Cloacimonadaceae (Pelletier et al., 2008; Dyksma and Gallert, 2019).

Ca. Tenebribacter, Ca. Stygibacter, and Ca. Zophobacter MAGs encode a membrane-bound [NiFe] hydrogenase (Mbh) (Group 4g) (Søndergaard et al., 2016). It has been proposed that Mbh transfers electrons from reduced ferredoxin to protons, thereby producing H₂ gas; this would generate a Na⁺ gradient across the cell membrane via a Mrp-type Na⁺/H⁺ antiporter module (Mayer and Müller, 2014; Søndergaard et al., 2016; Yu et al., 2018). All four genera encode the Rnf complex, which couples electron transfer from reduced ferredoxin to NAD⁺ to generate NADH, with concomitant translocation of Na⁺ ions across the membrane (Biegel et al., 2011). The Na⁺ gradient drives ATP synthesis via a Na⁺-dependent F-type ATP synthase (Meier et al., 2009). This Na⁺ gradient can also be used for other purposes, such as phosphate uptake via a Na⁺/phosphate cotransporter (NptA-like). NADH can also be used for anabolic purposes. Additionally, Ca. Tenebribacter, Ca. Stygibacter, and Ca. Zophobacter encode a Na⁺-translocating NADH:quinone oxidoreductase complex (NQR), which couples NADH re-oxidation to Na⁺-extrusion, as well as maintaining ionic balance inside the cell (Verkhovsky and Bogachev, 2010).

All 10 Ace Lake Cloacimonadota MAGs lack genes for two essential enzymes of the oxidative tricarboxylic acid (TCA) cycle: citrate synthase and succinate dehydrogenase. Thus, we infer that they cannot operate a complete oxidative TCA cycle. All four genera encode phosphoenolpyruvate (PEP) carboxykinase, which converts oxaloacetate to PEP. The Ace Lake Cloacimonadota MAGs encode a pyrophosphate-dependent phosphofructokinase (PP_i-PFK) as well as the more widely distributed ATP-dependent 6-phosphofructokinase (ATP-PFK), the former of which can reversibly function in both glycolysis and gluconeogenesis (Mertens, 1991; Kemp and Tripathi, 1993). Because PP_i is a byproduct of biosynthetic reactions, the use of PPi-PFK rather than ATP-PFK increases the energetic efficiency of glycolysis, especially during fermentation (Mertens, 1991; Reshetnikov et al., 2008). The presence of the reversible enzyme PP_i-PFK is consistent with the absence of the gluconeogenesis-specific enzyme fructose-1,6-bisphosphatase from the Ace Lake Cloacimonadota MAGs. Having dual enzymes for the conversion of fructose-6-phosphate to fructose 1,6-bisphosphate might allow the Ace Lake Cloacimonadota to respond to the flux of high-energy phosphoryl donors in the cell (ATP vs. PP_i). Furthermore, PP_i may also be diverted directly to energy conservation using a PP_i-dependent Na⁺ pump (HppA) that utilizes the energy of PP_i hydrolysis as the driving force for Na⁺ translocation.

In addition to the abilities to ferment sugars, the Ace Lake Cloacimonadota MAGs encode multiple ferredoxin oxidoreductases that oxidize 2-oxoacids (including products of amino acid degradation), as inferred for Ca. Cloacimonas (Pelletier et al., 2008). These include pyruvate:ferredoxin oxidoreductase (POR); 2-oxoglutarate:ferredoxin oxidoreductase (OGOR); branched-chain 2-oxoacid (2-oxoisovalerate):ferredoxin oxidoreductase (VOR); indolepyruvate:ferredoxin oxidoreductase; and aldehyde:ferredoxin oxidoreductase (Supplementary Tables S3, S4). In addition to VOR, the Ace Lake Cloacimonadota encode phosphate butyryltransferase and butyrate kinase, suggesting the potential for further catabolism of branched-chain 2-oxoacids derived from degradation of BCAAs.

Ca. Stygibacter and Ca. Zophobacter MAGs encode phosphate acetyltransferase and acetate kinase for the conversion of acetyl-CoA to acetate via acetyl phosphate with concomitant production of ATP via substrate-level phosphorylation (Sapra et al., 2003), also inferred for Cloacimonadota MAG TCS47 (Zhang et al., 2016). In addition to Mbh, two reversible [FeFe] cytoplasmic hydrogenases were identified in certain Ace Lake Cloacimonadota: a tetrameric Group A3 hydrogenase (Ca. Stygibacter) and a monomeric Group C1 hydrogenase (Ca. Tenebribacter and Ca. Stygibacter) (Søndergaard et al., 2016). As a bidirectional hydrogenase, the Group A3 hydrogenase could use H₂ as an energy source through the bifurcation of electrons from H₂ to ferredoxin and NAD⁺ (Poudel et al., 2016; Søndergaard et al., 2016; Kpebe et al., 2018), or it could serve as a confurcating hydrogenase to dissipate surplus reductant (from both NADH and reduced ferredoxin) that is generated during fermentation (Schut and Adams, 2009; Poudel et al., 2016). For the latter, substrate-level phosphorylation in the conversion of glucose to acetate would be facilitated by the dissipation of both reducing equivalents (NADH and reduced ferredoxin) as H₂ (Herrmann et al., 2008). These findings indicate that the Ace Lake Cloacimonadota would generate H₂ and acetate as byproducts of carbohydrate fermentation.

In both the Ca. Tenebribacter and Ca. Stygibacter MAGs, the Group C1 [FeFe] hydrogenase gene is immediately downstream of a gene for a histidine kinase domain protein, providing support for a putative sensory function (Greening et al., 2016). However, in one Ca. Tenebribacter MAG (3300035698_1346) the same gene cluster also encodes homologs of hydrogenase subunits associated with electron bifurcation (Poudel et al., 2016), which raises the possibility of a metabolic role for the Group C1 [FeFe] hydrogenase.

The gene inventories of certain Ace Lake Cloacimonadota suggest that they are capable of operating the reverse tricarboxylic acid (rTCA) cycle for carbon fixation, driven by sulfur oxidation (Ca. Tenebribacter and Ca. Stygibacter) (Figure 3A) or H₂ oxidation (Ca. Celaenobacter) (Figure 3B). The MAGs of these three genera encode ATP citrate lyase (ACL), thiol:fumarate reductase (TFR), and OGOR; these three enzymes allow the TCA cycle to proceed in the reductive direction (Rubin-Blum et al., 2019). PEP carboxykinase would connect the rTCA cycle to gluconeogenesis (Marietou et al., 2020) (see section “Cloacimonadota: Fermentation”). In the Ca. Celaenobacter MAGs, the genes for TFR and OGOR are part of a gene cluster that also includes fumarate hydratase, succinyl-CoA synthetase, [NiFe] hydrogenase (Mvh) (Group 3c), and heterodisulfide reductase (Hdr), which is consistent with a functional link between all these proteins. ACL is encoded elsewhere in the Ca. Celaenobacter genome, in a gene cluster that also includes the TCA cycle enzymes aconitase and isocitrate dehydrogenase. In general, the cytoplasmic Mvh hydrogenase forms a complex with Hdr, and bifurcates electrons from H₂ to heterodisulfide (CoM-S-S-CoB) and ferredoxin; the Mvh–Hdr complex couples the exergonic reduction of heterodisulfide with the endergonic reduction of ferredoxin with H₂ (Heim et al., 1998; Kaster et al., 2011; Greening et al., 2016). Thus, in Ca. Celaenobacter, carbon fixation using the rTCA cycle would be driven by H₂ oxidation (Figure 3B), as in Aquifex aeolicus (Brugna-Guiral et al., 2003; Guiral et al., 2005), although the latter have been inferred to use a Group 2d cytoplasmic hydrogenase for carbon fixation (Greening et al., 2016). ACL, OGOR, TFR, and Hdr genes were also identified in MAGs assigned to Ca. Tenebribacter and Ca. Stygibacter. Hdr genes in one Ca. Stygibacter MAG (3300025642_13) are within a gene cluster that also contains genes implicated in sulfur metabolism, including thiosulfate:sulfurtransferase (TST) (with a predicted signal peptide), a cytoplasmic sulfur relay protein TusA, and a sulfur compound transporter (Gristwood et al., 2011; Tanaka et al., 2020). Homologs of these four proteins are also encoded in MAGs assigned to Ca. Tenebribacter. Thus, energy required for carbon fixation in Ca. Tenebribacter and Ca. Stygibacter may be derived from sulfur oxidation catalyzed by TST and Hdr, as in certain other autotrophic bacteria (Boughanemi et al., 2016; Koch and Dahl, 2018; Wang et al., 2019). Based on the presence of a TST homolog in these Ace Lake Cloacimonadota MAGs, the electron donor may be thiosulfate, with the initial reaction occurring in the periplasm (Figure 3A); however, elemental sulfur might also be utilized, as in A. aeolicus (Boughanemi et al., 2016).

Based on interrogation of 14 Ace Lake Omnitrophota MAGs, this candidate phylum possesses a heterotrophic and fermentative metabolism. None of the MAGs possess any genes necessary for motility or magnetotaxis. We infer the Ace Lake Omnitrophota to be heterotrophs that are capable of fermenting a narrow range of substrates for energy conservation. All MAGs encode proteases and peptidases (some with signal peptides) to degrade proteins to amino acids. However, there are very few enzymes encoded for the catabolism of amino acids, and there is no evidence in any of the 14 MAGs of genes required for amino acid fermentation, unlike Clostridium spp. (Herrmann et al., 2008; Perret et al., 2011). Thus, we posit that these Ace Lake Omnitrophota use amino acids derived from peptide hydrolysis as nitrogen sources (especially by deamination or transamination) or for protein synthesis.

The Ace Lake Omnitrophota MAGs encode ABC transporter systems for sugars (disaccharides and/or oligosaccharides) and a number of GHs (including β-glucosidases and sugar phosphorolytic enzymes) to break down di- and oligosaccharides into simpler sugars such as glucose and/or glucose-1-phosphate (Table 2 and Supplementary Tables S5, S6). Enzymes necessary for the initial depolymerization of polysaccharides are absent from all MAGs, which suggests that the Ace Lake Omnitrophota are dependent on other microorganisms for initial degradation of biopolymers, as inferred for the Black Sea Omnitrophota (Suominen et al., 2021). Simple sugars imported into the cell could also be utilized by the Ace Lake Omnitrophota for the synthesis of the compatible solute trehalose, or for the synthesis of glycogen for carbon and energy storage; enzymes for both processes were encoded across the Omnitrophota MAGs (Table 2 and Supplementary Table S5).

The Ace Lake Omnitrophota MAGs also encode the capacity for fermentation of glucose to acetyl-CoA via the EMP pathway. The Rnf complex is predicted to couple the reduction of NAD⁺ to the oxidation of reduced ferredoxin and translocation of protons across the cell membrane, allowing ATP generation via a V-type ATP synthase. The majority of MAGs also encode the enzymes phosphate acetyltransferase and acetate kinase for the subsequent conversion of acetyl-CoA to acetate with concomitant production of ATP via substrate-level phosphorylation (Sapra et al., 2003).

Excess reductant generated during fermentation could be dissipated as H₂ using a cytoplasmic hydrogenase (Dombrowski et al., 2017; Suominen et al., 2021). The Ace Lake Omnitrophota MAGs encode various hydrogenases, although it is noteworthy that each MAG has only one identifiable hydrogenase: Group 3d [NiFe] hydrogenase (Ca. Aceula gen. nov. and Ca. Zapsychrus gen. nov.); Group 3b [NiFe] hydrogenase (Ca. Tantalella gen. nov., Ca. Aadella gen. nov., and Ca. Orphnella gen. nov.); Group 4g [NiFe] hydrogenase (Ca. Saelkia gen. nov., and Ca. Gorgyraea gen. nov.), and Group A3 [FeFe] hydrogenase (Ca. Gygaella. gen. nov, Ca. Susulua gen. nov., Ca. Kaelpia gen. nov., and Ca. Kappaea gen. nov.). We infer that these hydrogenases are used for redox balance, associated with the need to dispose of surplus reductant. MAGs of both genera that encode Group 4g Mbh also encode Mrp (Ca. Saelkia and Ca. Gorgyraea) (Figure 4A); this hydrogenase may therefore function in a complex with the Mrp antiporter to generate an ionic gradient across the cell membrane, as in the other Mbh (Mayer and Müller, 2014; Søndergaard et al., 2016; Yu et al., 2018). The Group 3b hydrogenases of Ca. Tantalella, Ca. Aadella, and Ca. Orphnella were annotated as a bifunctional sulfhydrogenase (Shy) with dual hydrogenase and sulfur reductase activity (Figure 4B), meaning that excess reductant generated during fermentation can be disposed of as H₂ and sulfide, respectively (Ma et al., 1993; Silva et al., 1999; Ma et al., 2000).

None of the 14 Ace Lake Omnitrophota MAGs encode a complete TCA cycle (Figures 4A,B), either in the oxidative or reverse directions, with OGOR, succinyl-CoA synthetase, succinate dehydrogenase, ACL, and fumarate reductase absent from all MAGs (Table 2). We infer that the Ace Lake Omnitrophota possess an incomplete, “horse-shoe”-type TCA cycle as found in certain other anaerobic bacteria (e.g., Herlemann et al., 2009; Marco-Urrea et al., 2011). The type of citrate synthase varies, with either (but never both) citrate (Si)-synthase or citrate (Re)-synthase (Li et al., 2007; Marco-Urrea et al., 2011) encoded in individual MAGs, with the distribution of the functional analogs mostly conforming to separate Omnitrophota clades (Supplementary Table S6). The right branch of the incomplete TCA pathway of Omnitrophota is inferred to occur in the oxidative direction and commence at citrate synthase and terminate at 2-oxoglutarate. The left branch allows the interconversion of oxaloacetate, malate, and fumarate (Herlemann et al., 2009). This could proceed in the oxidative direction, with fumarate (such as generated as a byproduct of arginine synthesis) converted to oxaloacetate and used for gluconeogenesis (van Vugt-Lussenburg et al., 2009). Alternatively, this left branch may operate in the reductive direction, and be initiated by PEP carboxykinase (Herlemann et al., 2009); the subsequent reduction of oxaloacetate to fumarate would provide redox balance to the oxidative branch (Meléndez-Hevia et al., 1996). The “horseshoe-type” TCA cycle has no energy conservation function but serves solely for biosynthesis (Wood et al., 2004). The carbon skeleton 2-oxoglutarate is required for ammonia assimilation, and all 14 MAGs encode enzymes for this process (Supplementary Tables S5, S6). However, the fate of fumarate in Omnitrophota is unclear; there is no identifiable fumarate reductase (for anaerobic respiration), fumarate-adding enzymes (for hydrocarbon degradation), or aspartase (for synthesis of aspartate directly from fumarate) in any of the 14 Omnitrophota MAGs.

Although autotrophic pathways have been inferred in other Omnitrophota (Rinke et al., 2013; Kolinko et al., 2016) (see section “Cloacimonadota and Omnitrophota in the Ace Lake Ecosystem”), the Ace Lake Omnitrophota appear to be obligate heterotrophs. Ca. Gorgyraea encodes WL pathway genes (Table 2, Figure 4A, and Supplementary Table S5), but in the absence of rTCA cycle genes that link acetyl-CoA to central biosynthetic pathways (Youssef et al., 2019), we infer that the WL pathway does not function in autotrophic CO₂ fixation. Instead, we propose that the WL pathway, in combination with the Rnf complex, functions in the reductive direction as an electron sink during homoacetogenic glucose fermentation, to maintain redox balance (Schuchmann and Müller, 2016; Youssef et al., 2019).

Overall, we infer only minor differences in the metabolic strategies among the 14 Ace Lake Omnitrophota MAGs (e.g., ability to catabolize glycerol, alcohol, or certain sugars; mechanisms for redox balance) (Table 2). Thus, all 11 Omnitrophota genera described here appear to conform to the same metabolic template: fermentative heterotrophs capable of degradation of a narrow range of organic compounds (especially simple sugars), with a hydrogenase for redox balance. The biosynthetic potential of the Ace Lake Omnitrophota MAGs are impressive, with the genomic capacity to synthesize nucleosides, fatty acids, the majority of proteinogenic amino acids, and essential cofactors (Supplementary Table S5).

Although the ecophysiology of Ace Lake Cloacimonadota broadly agree with the fermentative, heterotrophic generalists inferred for members of this phylum from the Black Sea (Suominen et al., 2021), we infer a number of traits in certain Ace Lake Cloacimonadota that have not been previously reported for this candidate phylum. These include the presence of an extracellular cellulosome-like structure for the binding and degradation of biopolymers, PGA synthesis, a halomucin-like protein, and a chemolithoautotrophic pathway for carbon fixation via the rTCA cycle, fueled by oxidation of H₂ or sulfur compounds. These abilities attest to the physical and metabolic diversity of the Cloacimonadota, and emphasize the potential importance of this group in cycling of carbon, hydrogen, and sulfur in Ace Lake.

Chlorobium, which is the dominant organism in Ace Lake, also employs the rTCA cycle; this anaerobic photoautotroph grows at the limits of the penetration of photosynthetically active radiation at the oxic–anoxic interface (Panwar et al., 2020; Figure 5). As a consequence, the abundance of Chlorobium in this lake system is dictated by the polar light cycle, with a peak relative abundance of 83% at the interface in summer, and a marked decline in winter (6%) to spring (1%) (Panwar et al., 2020). Chlorobium is not metabolically active in the perennially dark anoxic zone of Ace Lake, and sinks to the bottom as particulate matter (Rankin et al., 1999; Lauro et al., 2011; Panwar et al., 2020). By contrast, light-independent, facultative chemolithoautotrophs, including the Ace Lake Cloacimonadota, would not be directly impacted by the marked seasonal variation in light availability. Although Cloacimonadota were most abundant in the deepest part of the anoxic zone of Ace Lake, they were detected throughout the anoxic zone, as well as at the interface (Panwar et al., 2020). For those Cloacimonadota that we infer to use H₂ oxidation for carbon fixation (Ca. Celaenobacter), Chlorobium is potentially a major source of H₂, as a byproduct of nitrogen fixation by a membrane-bound nitrogenase (Lauro et al., 2011; Panwar et al., 2020). There are other bacteria present throughout the anoxic zone of Ace Lake that have been inferred to be H₂-evolving obligate anaerobes; aside from Omnitrophota, these include members of Bacteroidia, Firmicutes, and Atribacterota (Panwar et al., 2020). The cyanobacterium Synechococcus, the most abundant phototrophic bacterium in the oxic zone of Ace Lake, was also abundant in the interface and anoxic zone, and previously inferred to be capable of fermentation coupled to H₂ production (Panwar et al., 2020); thus, Synechococcus is also a potential source of H₂ for H₂-oxidizing Cloacimonadota. For those Ace Lake Cloacimonadota that we infer to rely on sulfur oxidation for carbon fixation (Ca. Tenebribacter and Ca. Stygibacter), Chlorobium cells could be a source of elemental sulfur and polysulfide, as intermediates in sulfide oxidation (Marnocha et al., 2016), whereas thiosulfate is likely generated from the chemical reaction of sulfide (such as those generated by sulfate-reducing Desulfobacterota) with dissolved oxygen (Kondo et al., 2000). However, unlike sulfate-reducing and sulfur-disproportionating Desulfobacterota in Ace Lake, which were previously inferred to be metabolically linked to Chlorobium via sulfur cycling (Lauro et al., 2011; Panwar et al., 2020; Figure 5), no seasonal variation was observed for Cloacimonadota in Ace Lake. Although Cloacimonadota and Omnitrophota in Ace Lake are inferred to be obligate anaerobes, both are found in the oxic–anoxic interface; the MAGs encode mechanisms to help protect against oxidative stress (e.g., desulfoferrodoxin in both phyla, catalase in Cloacimonadota), allowing survival at minimal oxygen concentrations (Pelletier et al., 2008). Thus, we infer that these anaerobes are aerotolerant, as previously inferred for Chlorobium in Ace Lake (Ng et al., 2010).

During heterotrophic growth, the anaerobic degradation of hexoses to acetate, CO₂, and H₂ does not yield sufficient energy to support growth unless H₂ levels are sufficiently low (Thauer et al., 1977; Schink, 1997; Morris et al., 2013). Acetate and H₂ would be beneficial to hydrogenotrophic and acetoclastic methanogens (Euryarchaeota), which, like Cloacimonadota, are most abundant in the deepest waters of Ace Lake (Panwar et al., 2020; Figure 5); as such, we posit that Cloacimonadota and methanogens may be metabolically linked. The extracellular GHs and peptidases, including the extracellular cellulosome-like structure inferred for the Cloacimonadota genus Ca. Stygibacter, suggest that Cloacimonadota are among the “first responders” in deconstructing and assimilating recalcitrant particulate organic matter, including microbial aggregates that sink to the bottom from higher in the water column (Rankin et al., 1999; Lauro et al., 2011).

All the Ace Lake Omnitrophota MAGs appear to be obligate heterotrophs, with no evidence of autotrophic capacity. Furthermore, these anaerobes appear to be dependent on hydrogenotrophic microbes to consume H₂ released via anaerobic glucose fermentation. In Ace Lake, Omnitrophota were most abundant at the oxic–anoxic interface and the water column immediately below the interface, the depths at which sulfate-reducing and sulfur-disproportionating Desulfobacterota were also most abundant (Panwar et al., 2020; Figure 5). These Desulfobacterota encode H₂-uptake hydrogenases (Panwar et al., 2020), and would therefore be capable of consuming H₂.

The ability to infer the ecophysiology of the members of Cloacimonadota and Omnitrophota emphasizes the importance of having metagenome data and accompanying analyses detailing the metabolisms of numerous other members of the microbial community (Ng et al., 2010; Lauro et al., 2011; Panwar et al., 2020). Here, we added to the understanding of the system by analyzing a total of 24 MAGs for these “microbial dark matter” lineages. Both of these candidate phyla are predicted to engage in metabolic associations with other Ace Lake microorganisms. The specific involvement of hydrogen is noteworthy in view of hydrogen cycling being previously identified as pivotal to multiple nutrient cycles in Ace Lake (Panwar et al., 2020), as well as being increasingly recognized as important to ecosystem function in global anoxic and oxic environments (Greening et al., 2016).