The whole genome sequences of 10 extremely thermophilic Methanococci plus M. kandleri were analyzed for known hydrogenases (see Supplementary materials). All 11 of the Methanococci and Methanopyri in the genome survey have at least one of the following hydrogenase genes (see Greening et al., 2016 for a review): (1) eha and ehb operons, which encode for membrane-bound multimeric hydrogenases that couple H₂ oxidation to ferredoxin reduction and are H⁺/Na⁺ driven for anaplerotic (Eha) and anabolic (Ehb) purposes (Porat et al., 2006; Lie et al., 2012); (2) an frh operon, which encodes for a soluble complex that directly couples H₂ oxidation to coenzyme F₄₂₀ reduction (Hendrickson and Leigh, 2008); (3) an hmd gene, which encodes a soluble methylenetetrahydromethanopterin dehydrogenase that couples oxidation of H₂ to the reduction of methenyltetrahydromethanopterin in the archaeal Wood-Ljungdahl CO₂ fixation pathway (Hendrickson and Leigh, 2008); and (4) a vhu operon, which encodes for soluble heterodisulfide reductase-linked complexes that bifurcate electrons from H₂ to heterodisulfide (coenzyme M-coenzyme B) and ferredoxin (Kaster et al., 2011). These hydrogenases are described and listed in Figure 2, Table 2, and Supplementary Table 1. Coenzyme F₄₂₀, ferredoxin, coenzyme M, and coenzyme B are soluble electron carriers in methanogens (Thauer et al., 2008). Extremely thermophilic Methanococci and Methanopyri will often have two or three copies of the genes encoding these enzymes (Table 2 and Supplementary Table 1).

The growth of natural assemblages of extremely thermophilic Methanococci in hydrothermal vent fluids from Axial Seamount is largely dependent on H₂ availability and temperature (Topçuoğlu et al., 2016). The Monod kinetic half-saturation value (K_s) for growth of extremely thermophilic methanogens was 27–66 μM with maximum methane production rates of 24–43 fmol CH₄ produced cell^–1 h^–1 (Ver Eecke et al., 2012; Stewart et al., 2019). Methanocaldococcus jannaschii and Methanothermococcus thermolithotrophicum were shown to grow by interspecies H₂ transfer when grown in co-culture with Thermococcus celer, Thermococcus stetteri, and Pyrococcus furiosus (Bonch-Osmolovskaya and Stetter, 1991). When M. jannaschii was grown in monoculture at high (80–83 μM) and low (15–27 μM) H₂ concentrations and in co-culture with the hyperthermophilic H₂ producer Thermococcus paralvinellae (representing very low H₂ flux), growth and cell-specific CH₄ production rates decreased with decreasing H₂ availability (Topçuoğlu et al., 2019). However, the number of cells produced per mole of CH₄ produced (i.e., cell yield) increased six-fold with decreasing H₂ indicating increased growth efficiency when growth was limited by H₂ (Topçuoğlu et al., 2019). Relative to high H₂ concentrations, isotopic fractionation of CO₂ to CH₄ was 16‰ larger for cultures grown at low H₂ concentrations and 45–56‰ larger in co-culture suggesting reversal of the Wood-Ljungdahl pathway during methanogenesis with low H₂ flux (Valentine et al., 2004; Topçuoğlu et al., 2019). While all four types of hydrogenases were synthesized by M. jannaschii with high and low H₂ flux, transcript levels of hmd and eha decreased with decreasing H₂ availability (Topçuoğlu et al., 2019).

The whole genome sequences of 30 Thermococci were analyzed for known hydrogenases (see Supplementary materials). All 30 Thermococci analyzed have at least one of the following hydrogenase operons: (1) An mbh operon, which encodes for a membrane-bound hydrogenase that couples oxidation of ferredoxin to H₂ evolution with concomitant H⁺/Na⁺ translocation across the membrane using antiporters (Sapra et al., 2003); (2) an sh operon, which encodes for a soluble sulfhydrogenase that couples oxidation of H₂ oxidation to the reduction of NAD(P)⁺ (Van Haaster et al., 2008); (3) an frh operon, which encodes for cytoplasmic coenzyme F₄₂₀ reducing-type hydrogenase that oxidizes H₂ and passes electrons to a thioredoxin reductase (Jung et al., 2020); and (4) a codh operon, which encodes for a membrane-bound hydrogenase that couples oxidation of CO to H₂ evolution with concomitant H⁺/Na⁺ translocation across the membrane using antiporters (Bae et al., 2012; Moon et al., 2012). These hydrogenases are described and listed in Figure 3, Table 3, and Supplementary Table 2.

All Thermococci have at least one mbh operon and all but one have at least one sh operon (Table 3). These enzymes are the core hydrogenases for Thermococci (Schut et al., 2012; Boyd et al., 2014). Twelve of the 30 Thermococci in the survey have frh operons. Five of the 30 Thermococci have codh operons. It was shown that the growth of Thermococcus sp. strain AM4 and Thermococcus onnurineus can be supported by CO with concomitant H₂ production (Sokolova et al., 2004; Bae et al., 2012; Moon et al., 2012), although the physiological role of this enzyme in Thermococcus is yet to be determined for growth in its natural environment.

In Thermococci, the reduction of S⁰ is the preferred route for electron disposal over the reduction of H⁺ to H₂. In P. furiosus, the presence of S⁰ in growth media resulted in decreases in Mbh and Sh hydrogenase specific activities, each by an order of magnitude (Adams et al., 2001). There was an immediate downregulation of mbh and an upregulation of mbs (membrane-bound sulfane reductase) (Wu et al., 2018) and nsr (NAD(P)H:S⁰ reductase) in P. furiosus when S⁰ was added to growth medium (Schut et al., 2001, 2007). A sulfur response regulator protein (SurR) was identified as the transcription factor regulating hydrogenase and sulfur responsive genes (Lipscomb et al., 2009, 2017). The proposed model suggests that SurR contains a redox-active cysteine disulfide that can reduce S⁰ to H₂S (Yang et al., 2010). SurR is reduced in a redox cascade involving NAD(P)H-dependent thioredoxin reductase (TrxR) and protein disulfide oxidoreductase (Pdo) as the electron donors (Lim et al., 2017). In the absence of S⁰, SurR remains reduced, binds to GTTn₃AAC(n₅GTT), promotes the transcription of mbh and sh genes, and represses the expression of mbs and nsr genes (Lipscomb et al., 2009). Thermococci with the Frh hydrogenase also can reduce TrxR using H₂ as the electron donor (Jung et al., 2020).

Abiotic formation of formate, carbon monoxide, methane, and hydrocarbons in hydrothermal vents is of interest as potential growth substrates for microbes. Methane and hydrocarbons in vents were suggested to form through Fischer-Tropsch type reactions [(2n + 1)H₂ + nCO → C_nH_2n+2 + nH₂O] or leach from fluid inclusions in plutonic rocks (Berndt et al., 1996; Horita and Berndt, 1999; McCollom and Seewald, 2001; McDermott et al., 2015). In contrast to hydrocarbons, there is a strong thermodynamic drive toward rapid C-H-O equilibrium in hydrothermal fluids within hours to days. Kinetic barriers preclude the formation of CH₄ in this equilibrium (Shock, 1990). This permits the creation of metastable formate species (H₂ + CO₂ ↔ HCOOH), CO (HCOOH ↔ CO + H₂O), formaldehyde (HCOOH + H₂ ↔ CH₂O + H₂O), and methanol (CH₂O + H₂ ↔ CH₃OH) through the sequential reduction of CO₂ using H₂ as the reductant (Seewald et al., 2006).

The abundance of formate in chemical equilibrium with dissolved inorganic carbon is strongly dependent on H₂ concentration, pH, and temperature (McCollom and Seewald, 2003; Seewald et al., 2006). In a gold-titanium reaction cell, HCOO^– was formed from CO₂ at 300°C and 350 bar in less than 24 h from H₂ generated from hydrothermal alteration of olivine serving as the reductant (McCollom and Seewald, 2001). In a separate study, incubation of a 175 mmol/kg HCOOH solution at 300°C and 350 bar in the gold reaction cell led to near complete conversion to H₂ and CO₂ within 48 h, CO reached 0.83 mmol/kg, and HCOO^– + HCOOH (or ΣHCOOH) decreased to 0.38 mmol/kg (Seewald et al., 2006). Reducing the temperature to 200°C and then to 150°C each led to an increase in ΣHCOOH, a decrease in CO, and C-H-O equilibrium within 115 h and 71 h, respectively. Injection of 172 mmol/kg CO led to production of H₂, ΣCO₂, and ΣHCOOH, and decreasing CO. Alkaline conditions favored the formation of HCOOH, HCO₃^–, and CO₃^2– (Seewald et al., 2006). Therefore, the abundance of formate, CO, and CH₃OH in seafloor hydrothermal systems will be regulated by the residence times of fluids in reactions zones, and physical and chemical conditions in the subsurface environments.

Formate is also generated across a pH gradient of more than three pH units using a mineral precipitate bridge at the interface of two fluids (Hudson et al., 2020). This may be relevant to the formation of formate on the early Earth or possibly in extraterrestrial oceans where high pH serpentinized fluids are emitted into an acid ocean. Under standard conditions, the generation of formate from H₂ and CO₂ is not thermodynamically favorable. However, H₂ in synthetic alkaline vent fluid (pH 12.3) passed electrons to dissolved CO₂ in a synthetic acid ocean (pH 3.9) at 25°C through a Fe(Ni)S mineral interface generating 1.5 μM HCOO^– in the ocean fluid (Hudson et al., 2020). Isotopic labeling showed that protonation occurred using H₂O on the ocean side of the interface, not H₂ on the vent side. Weakening the pH gradient led to decreased concentrations of HCOO^– produced. Nickel in the precipitate is a crucial part of the reduction mechanism as HCOO^– yield dropped below detection without Ni in the ocean precipitation fluid.

There have been very few measurements of formate in natural hydrothermal fluids due in part to the analytical difficulty of measuring formate at low concentrations (Schink et al., 2017). Formate has been measured mostly at sites with high H₂ concentrations such as at the Lost City, Von Damm, and Piccard hydrothermal vent sites and were 36–669 μM (Table 1). Formate and H₂ were also measured at Snake Pit and TAG hydrothermal vents, which are mafic hydrothermal vents on the Mid-Atlantic Ridge, where formate concentrations were 1–2 nM and H₂ concentrations were 0.08–2.4 μM (Konn et al., 2022). At ultramafic sites, formate concentrations are generally 10–100 fold lower than that of H₂ at the same site (Lang et al., 2010; McDermott et al., 2015) while at mafic sites the formate concentration is often more than 1,000 fold lower than the H₂ concentration (McDermott et al., 2018; Konn et al., 2022).

Methanocaldococcus jannaschii was shown to incorporate ¹⁴C-formate into biomass during growth (Sprott et al., 1993), which may be used in part for de novo purine biosynthesis. Inosine monophosphate (IMP) is a precursor for adenine and guanine synthesis for purine biosynthesis and is made from ribose-5-phosphate (Figure 4). In most organisms, the pathway intermediates glycinamide-ribose-5-phosphate (GAR) and aminoimidazole carboxamide-ribose-5-phosphate (AICAR) are formylated using N¹⁰-formyl-tetrahydrofolate as the formyl donor. However, the genes for these enzymes are absent in Methanococci and Methanopyri and are replaced with genes that encode for enzymes that use free formate and energy from ATP to formylate their substrates (White, 1997; Brown et al., 2011). These enzymes are formylglycinamide-ribose-5-phosphate synthetase (PurT) and forminido imidazole carboxamide-ribose-5-phosphate synthetase (PurP) (Figure 4). M. jannaschii was shown to have PurP activity and that it produced free ¹³C-formate in the cell when incubated with H₂ and H¹³CO₃ (Ownby et al., 2005). Herein, a genome survey of the eleven extremely thermophilic methanogens showed that all the organisms have homologs for purP and all but M. kandleri have homologs for purT (Table 2 and Supplementary Table 1). This suggests that these organisms have a mechanism for formate synthesis.

Nine of the 11 Methanococci and Methanopyri genomes have genes that encode for a cytoplasmic formate dehydrogenase (Table 2 and Supplementary Table 1). Formate dehydrogenases catalyze the reversible oxidation of formate to CO₂ using various electron acceptors. The catalytic α subunit (FdhA) contains tungsten, selenocysteine, and a (Fe₄-S₄) cluster as cofactors while the β subunit (FdhB) contains three (Fe₄-S₄) clusters (Niks and Hille, 2019). FdhAB in Methanococci and Methanopyri is homologous to two formate dehydrogenases in the mesophilic methanogen Methanococcus maripaludis, also a Methanococci, that use coenzyme F₄₂₀ as their redox partner (Figure 2; Wood et al., 2003; Lupa et al., 2008). M. maripaludis grows hydrogenotrophically on H₂ and CO₂ but also grows on formate in their absence (Jones et al., 1983b). When fdhA1 was mutated in M. maripaludis, the organism was unable to grow on formate and formate dehydrogenase activity in cell extracts was undetectable (Lupa et al., 2008). Observations with hydrogenase mutants in M. maripaludis suggest that coenzyme F₄₂₀ is an intermediate in formate-to-H₂ conversion (Lupa et al., 2008). An M. maripaludisΔfdhA1ΔfdhA2 double mutant grown in purine-free defined medium grew as well as the wild-type strain suggesting that formate dehydrogenase is not essential for de novo purine biosynthesis (Wood et al., 2003). The absence of fdhAB genes in Methanocaldococcus infernus and Methanofervidicoccus abyssi (Table 2 and Supplementary Table 1) also supports the idea that formate dehydrogenase is not essential for purine biosynthesis. However, it is likely that H₂ and coenzyme F₄₂₀ are electron donors for formate production and can help meet the cellular demand for formate for purine biosynthesis.

The formate dehydrogenase (FdhA1B1) from M. maripaludis also forms an enzyme complex with heterodisulfide reductase, the soluble hydrogenase Vhu, and formylmethanofuran dehydrogenase (Figure 2; Costa et al., 2010). It was necessary for the organism’s growth on formate but not on H₂ (Costa et al., 2010). Therefore, in addition to coenzyme F₄₂₀ reduction, this formate dehydrogenase also oxidizes formate to reduce the heterodisulfide coenzyme M-coenzyme B and ferredoxin through electron bifurcation. Coenzyme M, coenzyme B, and ferredoxin are cytoplasmic electron carriers in these methanogens. Expression of the second formate dehydrogenase gene (fdhA2) in M. maripaludis increased when cells were grown under H₂ limited conditions but was unchanged under formate limited conditions (Costa et al., 2013) and was not required for growth on formate (Lupa et al., 2008) suggesting that this isoenzyme may have a separate physiological function.

For extremely thermophilic methanogens, it appears that a formate transporter is required for growth on formate. Formate transporters import or export formate across the cytoplasmic membrane and require co-translocation of a H⁺ (Figure 2). Three thermophilic methanogens in our survey (Methanotorris formicicus, Methanothermococcus okinawensis, and Methanothermococcus thermolithotrophicus) grew on formate in the absence of H₂ and CO₂ but not any of the other methanogens examined (Table 2). Each of these methanogens that grew on formate has a gene that encodes for a membrane-bound formate transporter (fdhC) in its genome, which is absent in all other methanogens examined, except for Methanocaldococcus fervens which was not tested for growth on formate (Table 2 and Supplementary Table 1). M. maripaludis has an fdhC gene in an operon with fdhA1B1 (Sattler et al., 2013). In M. fervens and M. okinawensis, the formate transporter gene fdhC appears to be in the same operon as fdhAB suggesting they are co-transcribed (Supplementary Table 1).

Like Methanococci, all Thermococci lack the enzymes that use N¹⁰-formyl-tetrahydrofolate as the formyl donor for de novo purine biosynthesis (Brown et al., 2011). Instead, most Thermococci use formate-dependent enzymes (PurT and PurP) for de novo purine biosynthesis (Figure 4, Table 3, and Supplementary Table 2). Therefore, they depend on a source of free formate in the cell for de novo synthesis. However, some Thermococcus species (T. paralvinellae, T. barophilus CH5, T. onnurineus, and T. gorgonarius) lack most or all the genes for the purine biosynthesis pathway (Brown et al., 2011) and likely rely on environmental sources of purines.

All 30 Thermococci genomes have at least one copy of the gene that encodes for the catalytic α subunit of formate dehydrogenase (FdhA) either in the form of formate hydrogenlyase, NAD(P)⁺-dependent formate dehydrogenase, or the catalytic subunit alone (Table 3 and Supplementary Table 2). The phylogeny of FdhA in extremely thermophilic Methanococci, Methanopyri, and Thermococci showed one clade for Methanococci and Methanopyri and five clades among the Thermococci (Figure 5). In Thermococci, hydrogenase operons often flank fdhA-containing operons in the genome (Figure 6 and Supplementary Table 2) suggesting a close association between formate and H₂ in these organisms. In Groups 1 and 2 in Figure 5, fdhA was encoded in an operon with a formate transporter gene. For Group 1, in nearly all instances, the fdhA-containing operon was immediately downstream from an frh operon and immediately upstream from one or two mbh operons on the same DNA strand suggesting that they may be co-transcribed (Figure 6). In Group 1A, fdhA was encoded in a formate hydrogenlyase (fhl) operon (Kim et al., 2010; Topçuoğlu et al., 2018; Le Guellec et al., 2021; Table 3; Supplementary Table 2). This enzyme reversibly couples formate oxidation to H₂ evolution on the cytoplasmic membrane with concomitant H⁺/Na⁺ translocation across the membrane via antiporter modules (Figure 4; Kim et al., 2010; Lim et al., 2014). In Group 1B, fdhA was encoded in a NAD(P)⁺-dependent formate dehydrogenase (nfd) operon (Figure 6). This soluble enzyme catalyzes the reversible oxidation of formate using NAD(P)⁺ or ferredoxin as its redox partner (Le Guellec et al., 2021; Yang et al., 2022; Figure 4). In Group 2, fdhA was encoded in an nfd operon but neighbored an sh operon in the genome instead of frh and mbh operons (Figure 6). These nfd and sh operons are transcribed in opposite directions from the same intergenic spacer region.

The fdhA from Groups 3 and 4 are in fhl operons that lack a formate transporter gene. In Group 3, the fhl operon was next to an sh operon (Figure 6). These fhl and sh operons are transcribed in opposite directions from the same intergenic spacer region. In Group 4, the fhl operon did not neighbor any hydrogenase operons in the genome, and in Group 5 the fdhA gene was the only formate dehydrogenase-related gene present in the genome (Figure 6). Often these solo genes in Group 5 are near the purine biosynthesis genes in genome sequences (Supplementary Table 2). In T. sibiricus, nearly all the genes for de novo purine biosynthesis (purFCMTEDPSQL) and fdhA are next to each other in the genome, although they are not all on the same DNA strand (Figure 6). In these organisms, it is unknown if fdhA alone encodes for a functional formate dehydrogenase or what the redox partner is for this putative enzyme. However, it is plausible that it might be used to produce formate for purine biosynthesis when other formate dehydrogenases and formate transport proteins are absent.

Under defined growth conditions, 11 of the 30 Thermococci strains analyzed either oxidized added formate as an energy source (plus trace levels of organic compounds as a carbon source) and produced H₂ (Kim et al., 2010; Topçuoğlu et al., 2018) or secreted formate when grown on organic compounds in the presence of high background H₂ and the absence of added formate (Hensley et al., 2016; Topçuoğlu et al., 2018; Le Guellec et al., 2021). These 11 strains are the only Thermococci in the survey that have a formate transporter gene (Table 3). The other 19 Thermococci lack this gene and were unable to grow on formate or secrete formate (Kim et al., 2010; Le Guellec et al., 2021). Therefore, it appears that a formate transporter is required for Thermococci to secrete formate or, like Methanococci, for growth of Thermococci on formate. The presence of a formate transporter gene or transcript should be a criterion when determining if Methanococci or Thermococci are potentially using or producing formate in their natural habitat.

The standard Gibbs energy for interconversion between formate and H₂ + CO₂ is small; therefore, the direction of the reaction is highly dependent upon the relative concentrations of formate and H₂ in the environment (Schink et al., 2017; Le Guellec et al., 2021). Le Guellec et al. (2021) calculated that CO₂ reduction to formate using H₂ is thermodynamically more favorable than formate oxidation to H₂ and CO₂ at Lost City, Von Damm, Rainbow, Lucky Strike, Snake Pit, and Ashadze 1 hydrothermal vent sites based on relative formate and H₂ concentrations in hydrothermal fluids. The physiological response of Thermococcus is in keeping with this idea. Growth of T. paralvinellae on a sugar or peptides when sparged with H₂ led to higher levels of fhl1 expression and higher formate secretion relative to cultures sparged with N₂ (Topçuoğlu et al., 2018). It was concluded that fhl and nfd expression in Thermococci is primarily for the purpose of ameliorating H₂ inhibition rather than for growth on formate (Topçuoğlu et al., 2019; Le Guellec et al., 2021). Thermococci would require an environment where formate concentrations exceed H₂ concentrations to grow on formate. The formate produced by Thermococci may supplement the growth of Methanococci even when Thermococci produce H₂, as is observed with fermenter-methanogen relationships in mesophilic environments (Schink et al., 2017).

Formate consumption in Methanococci is closely associated with H₂ use in the cell. Therefore, a question that arises is whether formate or H₂ regulates fdhAB expression in these organisms. The thermophilic methanogen Methanobacterium thermoformicicum grows on H₂ and CO₂ as well as separately on formate. It has a formate transporter gene (fdhC) directly upstream of its formate dehydrogenase genes (fdhAB) (Nolling and Reeve, 1997). Transcripts of fdhCAB were present in M. thermoformicicum at all growth stages when grown on formate. When grown on H₂ and CO₂, fdhCAB transcripts were barely detectable in early exponential growth phase but increased dramatically as cells approached late exponential growth phase in a closed batch system when H₂ became more limiting. Similarly, fdh expression in M. maripaludis was controlled by the presence of H₂ and not formate (Wood et al., 2003). Using fdhC-lacZ gene fusions, β-galactosidase activity increased in M. maripaludis cells grown on H₂ and CO₂ as they approached late exponential growth phase, again when H₂ became limiting. When grown on formate, β-galactosidase activity was higher in cells with N₂ and CO₂ in the headspace relative to those with H₂ and CO₂ in the headspace. β-galactosidase activity increased in cells grown on formate plus H₂ and CO₂ after the H₂ and CO₂ was replaced mid-growth phase with N₂ and CO₂.

In M. maripaludis, genes for a putative response regulator and a histidine kinase are directly upstream of fdhC, which is three genes upstream of fdhA1B1 and part of a putative five-gene operon (Sattler et al., 2013). Random mutagenesis showed that disruption of this putative response regulator led to slower growth of M. maripaludis on formate relative to the wild type. It also led to increased fdhA1 transcriptional abundance regardless of whether H₂ and CO₂ or formate was the growth substrate. Impairment of derepression of the fdhC-fdhA1B1 operon is a plausible explanation (Sattler et al., 2013). Therefore, H₂ present at high concentrations may interact with the histidine kinase and activate the response regulator in a two-component regulatory system that represses fdhC-fdhA1B1 expression, which is derepressed when H₂ levels are low or absent.

Very little is known about transcriptional regulation of the fhl and nfd operons in Thermococci. Group 1 Thermococci genomes (Figure 6) encode syntenic frh, either fhl or nfd, and mbh operons with a formate transporter gene encoded in the fhl or nfd operon (Figure 6). These frh, fhl, nfd, and mbh operons each have GTTn₃AAC(n₅GTT) in their promoter region just upstream of BRE/TATA RNA polymerase binding sites suggesting they are also regulated and promoted by the sulfur response regulator protein SurR (see Section “4.2. Growth of Thermococci with and without S⁰”). Furthermore, Frh was shown to oxidize H₂ and reduce TrxR (Jung et al., 2020), which reduces SurR via Pdo, suggesting that it might serve as a regulatory hydrogenase that promotes frh, fhl, nfd, and mbh expression when H₂ concentrations increase in the cell. Therefore, like Methanococci, H₂ abundance appears to regulate formate use in Thermococci. A remaining question is whether formate also regulates gene expression in Thermococci. In T. paralvinellae, expression of the Group 1A fhl operon containing the formate transporter gene increased when cells were grown on formate relative to growth on maltose or peptides while expression of mbh either remained unchanged or decreased (Topçuoğlu et al., 2018). This suggests that in addition to SurR regulation, formate either directly or indirectly regulates gene expression in T. paralvinellae as well. Validation and the mechanism of this putative regulation is yet to be determined.

None of the promoter regions for the nfd, fhl, or sh operons in Groups 2–4 had a SurR nucleotide binding sequence. All but one of the Group 4 fhl operons have a gene encoding for a TetR/AcrR family transcriptional regulator that is ∼350 nucleotides upstream of and transcribed in the same direction as the fhl operon (Supplementary Table 2). TetR/AcrR family transcriptional regulators are one-component systems where a single protein contains both a sensory domain and a DNA-binding domain (Cuthbertson and Nodwell, 2013). They are widely associated with antibiotic resistance and the regulation of genes encoding small molecule exporters and are usually encoded alongside target operons (Colclough et al., 2019). In T. paralvinellae, expression of the Group 4 fhl operon decreased when cells were grown on formate relative to growth on maltose or peptides (Topçuoğlu et al., 2018). The mechanism for regulation of Group 2–5 formate dehydrogenase-related genes is unknown.