P. thermoglucosidasius 23.6 genome sequences are available from Genbank in Bioproject PRJNA668107. The complete genome sequence of this strain is available under the accession numbers CP063414–CP063417 and was inspected for the presence and organization of putative [NiFe]-hydrogenases-encoding genes. HydDB, a webtool for the structural and functional classification of hydrogenases (Søndergaard et al., 2016) was used to identify the class of [NiFe]-hydrogenases potentially encoded by P. thermoglucosidasius strain 23.6.

P. thermoglucosidasius 23.6 was plated from a glycerol stock on non-selective agar medium (Santana et al., 2020) and grown at 50 °C. A single colony was selected and grown to saturation in liquid medium overnight at 50 °C. Fresh non-selective medium (Santana et al., 2020) was inoculated with 1% (v/v) from the overnight culture in an Erlenmeyer with five times volume air/volume culture. The culture was incubated at 50 °C, 110 rpm and grown until early stationary phase and its cfu mL⁻¹ value was then determined following serial dilution and plating. The culture was divided into 2 mL aliquots, which were centrifuged at ambient temperature at 4,000 × g (Spectrafuge 16 M Microcentrifuge; Labnet, Edison, NJ, United States). The pellet of each aliquot was resuspended in 50 μL of Tris 20 mM pH 7.5 after determination of the fresh weight (FW). The cell suspensions were kept at −20 °C and defrost for the hydrogenase activity assay, which consisted of a protocol adapted from Lacasse et al. (2019). Briefly, dilutions of the cell suspensions were performed in 200 μL Tris 20 mM pH7.5 to determine the best cell density for the assay. Triplicates of the dilutions and of negative control (200 μL Tris 20 mM pH7.5) were placed in the wells of a Deltalab 96-well plate (code 900011) and the OD at 630 nm was recorded (MicroPlate reader Synergy HTX, BioTek). The plate was kept briefly in the dark and 20 μL of a freshly prepared developing solution −10 mg/mL benzyl viologen dichloride (Sigma-Aldrich, Massachusetts, United States) and 250 mM sodium formate (Riedel-de Häen) in 20 mM Tris buffer pH 7.5 - were added to each well. The change in absorbance at 630 nm, caused by the reduction of benzyl viologen, was recorded every 42 s for 7 min. Readings continued every 10 min for 1 h. Data were converted to ΔAbs·min⁻¹ cfu⁻¹and ΔAbs·min⁻¹ mg⁻¹ (FW). The procedure described was repeated a total of three times using independently grown cell suspensions.

Parageobacillus thermoglucosidasius strain 23.6, isolated from an upper soil sample from Southwestern Andalusia (Coria del Rio, Sevilla, Spain) (Gomez et al., 2020, 2021), was grown in Nutrient Broth (BD Difco, Sparks, MD) at 60 °C at different growth rates. Exponential growth under optimum conditions (2.2 h⁻¹) was obtained in a bioreactor and this condition was used as a reference for comparison to slow growing cells. Growth rates down to 0.025 h⁻¹ were achieved in a chemostat (Gonzalez and Aranda, 2023) under the same conditions. Lower growth rates and near-zero growth rates were obtained in a retentostat culturing system following previous procedures (Boender et al., 2011; Overkamp et al., 2015; Gonzalez and Aranda, 2023). A retentostat differs from a chemostat in that the retentostat is a closed system for the cells and open to medium (Gonzalez and Aranda, 2023). Cells were filtered back to the vessel by tangential flow filtration (Boender et al., 2011; Overkamp et al., 2015) and, thus, cells were slowly accumulated, forced to share the same nutrient resource, which progressively led to slower growth rates (down to ca. 10,000-fold lower rates than optimum growth). These nutrient-limiting conditions mimic the expected limitation of bacterial growth in the natural environment. Three different independent cultures for each growth rate were selected for RNA extraction and sequencing. Cells were collected at optimum rate (18 min doubling time), slow rate (chemostat; 1.1 days doubling time) and very slow and near-zero growth rates (retentostat; 14 d and 141 days doubling times) (Gonzalez and Aranda, 2023). Cells were collected by centrifugation at 4 °C for 5 min, RNA extracted (ZymoBIOMICS RNA extraction kit, ZymoResearch Corp., Irvine, CA). The rRNA was removed from the total RNA and sequenced (RNA-seq) using Illumina HiSeq system (BGI Genomics., Shenzhen, China). RNA-seq based comparative transcriptomic analysis between the gene expression at optimum and decreasing growth rates (down to near-zero growth rates) was carried out using the software suite and the bioinformatic protocol described by Pertea et al. (2016). This procedure follows a statistical test that uses a cumulative upper quartile normalization of FPKM abundance estimates according to Paulson et al. (2013). RNA sequencing data are accessible under DDBJ BioProject PRJDB39277, runs DRR797640-DRR797651.

Student’s t-test (IBM SPSS Statistics 30.0.0.0) was used to compare significant differences (p < 0.05) between hydrogenase activity measurements performed at 25 °C and 50 °C.

The soil Actinomycetota Mycobacterium smegmatis has been extensively studied regarding its H₂-scavenging activity and associated hydrogenases. This bacterium has three hydrogenases that belong to distinct phylogenetic groups, −1 h, 2a, and 3b—which have homologs in other actinomycetes (Greening et al., 2014a). Both 1 h hydrogenase (Hhy), considered a typical high-affinity hydrogenase (Km < 100 nM) and 2a, named Huc, are capable to mediate high-affinity H₂ oxidation, while Huc is also efficient at a wide range of H₂ concentrations (Greening et al., 2014a). We have clearly identified in P. thermoglucosidasius 23.6 genome sequence the genes encoding a Huc-type enzyme among three putative [NiFe]-hydrogenases-encoding genes; the hydrogenase primary sequences were assigned to hydrogenase groups 1d (Hup-type), 2a (Huc-type) and 4a (Hyf-type).

Although not being classed as H₂-scavenging, having low affinity for hydrogen (Mohammadi et al., 2017), hydrogenases from group 1d are membrane-bound H₂-uptake enzymes known to be relatively oxygen-tolerant. Hydrogenases of this group mediate the electron input for respiratory reduction of various electron acceptors such as nitrate, sulfate, and fumarate (Mohr et al., 2018) but are in general associated with aerobic respiration and oxygen-tolerant anaerobic respiration in facultative anaerobes (Greening et al., 2016) and may support autotrophic growth (Mohammadi et al., 2017).

Contrary to the above groups, group 4a comprehends oxygen-sensitive, membrane-bound H₂ evolving enzymes (Mohr et al., 2018). The group 4 [NiFe]-hydrogenases is traditionally associated with a fermentative role, namely with 4a formate hydrogenlyase, which couples the oxidation of formate to CO₂ and to fermentative evolution of H₂ (McDowall et al., 2014).

In Mycobacterium smegmatis, each of three hydrogenases 1 h (Hhy), 2a (Huc), and 3b, is encoded by a distinct structural operon, and numerous genes encoding maturation factors are clustered with hyd1 and hyd2 operons encoding, respectively, Huc and Hhy (Greening, 2013). Those maturation factors are required for production of a functional hydrogenase and include, among others, the nickel insertase HypA and the nickel chelator HypB (Greening et al., 2015), required for the nickel insertion into the enzyme (Chan et al., 2012). Like M. smegmatis, P. thermoglucosidasius 23.6 1d (Hup-type), 2a (Huc-type) and 4a (Hyf-type) hydrogenases are encoded by distinct putative operons (operons 1, 2, and 5; Table 1; Figure 1) and the high similarity of P. thermoglusosidasius 23.6 Huc-type hydrogenase to M. smegmatis Huc was determined by BlastP analysis of WP_003250668.1 nickel-dependent hydrogenase large subunit (E value = 0). Two additional clusters (operon 3 and 4; Table 1; Figure 1), located in proximity to operon 2, encode maturation proteins, probably used in the assembly of 2a hydrogenase.

The arrangement of the [Ni-Fe] hydrogenase loci associated with the 2a hydrogenase in the 23.6 strain (including the putative operons 3 and 4 encoding maturation proteins) is similar to the one already described by Mohr et al. (2018) for P. thermoglucosidasius DSM 2542, mostly conserved among several P. thermoglucosidasius strains, and roughly retained by other more distant Bacillaceae (Figure 2).

Interestingly, located just upstream of the P. thermoglucosidasius DSM 2542 4a hydrogenase locus, three putative genes, cooC, cooS, and cooF encoding a CO dehydrogenase maturation factor, a CO dehydrogenase catalytic subunit and a CO dehydrogenase Fe–S protein, respectively, which are involved in the oxidation of CO to CO₂, were identified leading to the suggestion that 4a hydrogenase may form a complex with a CO dehydrogenase to produce CO₂ and H₂. Herein, we also identified putative CO dehydrogenase-encoding genes in the genome of strain 23.6 (see Figure 1; Table 1) indicating the potential of this strain for H₂ production.

We used benzyl viologen, an established hydrogenase alternative substrate and a dye previously assessed as a colorimetric indicator primarily reporting on [NiFe]-hydrogenase activity (Lacasse et al., 2019). We found hydrogenase activity for cells grown at 50 °C, which was measured at 25 °C and 50 °C. At 25 °C, the activity was (η = 3) 4.02 10⁻⁵ mg⁻¹ (FW) (sd = 9.48 10⁻⁶) and 5.42 10⁻¹¹ cfu⁻¹ (sd = 1.53 10⁻¹¹), while at 50 °C the corresponding values were 6.39 10⁻⁴ mg⁻¹ (FW) (sd = 9.86 10⁻⁵) and 7.49 10⁻¹⁰ cfu⁻¹ (sd = 1.53 10⁻¹¹). The activity is therefore over 15-fold higher at 50 °C (Figure 3) than at 25 °C although it remains measurable at this low temperature.

Table 1 shows the expression data at different growth rates for the genes within the putative [NiFe]-hydrogenases-encoding operons found in P. thermoglucosidasius 23.6. We observed significant overexpression of the genes encoding the subunits of putative [NiFe]-hydrogenases belonging to groups 1d and 2a during limiting growth conditions and located in putative hydrogenase-encoding operons 1 and 2, respectively. 1d (Hup-type) and 2a (Huc-type) nickel-dependent hydrogenase large subunit gene was found to be overexpressed 8-fold and over 900-fold, respectively, at near-zero growth rate compared with optimum growth rate. Decreasing the growth rate to about 89-fold lower than optimum rate indicated important differences, showing overexpression of hup and huc genes; Huc-type hydrogenase genes were increasingly and drastically overexpressed at decreasing growth rates, reaching maxima at near-zero growth rates, whereas Hup-type enzyme genes were overexpressed at slow growth and near-zero growth, but their expression was highest at a low growth rate (0.025 h⁻¹) (Table 1).

Unlike the expression of Hup and Huc [NiFe]-hydrogenase genes, we observed a modest increase in the gene expression of P. thermoglucosidasius 23.6 Hyf-type hydrogenase under limiting nutrient conditions versus optimal growth (a 1.6-fold increase of the hydrogenase large subunit gene expression at 0.025 h⁻¹).

In most Actinobacteria, enzyme kinetic parameters were determined in whole cells, thus, it is still not clear and not entirely predicted by the primary sequences (Greening et al., 2015), if high-affinity hydrogenases are intrinsically high affinity enzymes or if their affinities are modulated by interactions with the respiratory chain (Greening et al., 2014a). However, recent kinetic analysis of purified M. smegmatis Huc enzyme showed that this group 2a enzyme is per se capable of oxidizing atmospheric H₂ with a high affinity (Km = 129 nM) and is highly efficient at low H₂ concentrations (Grinter et al., 2023). Overall, in addition to the canonical group 1 h [NiFe]-hydrogenase, it has been shown that three other hydrogenase lineages, taxonomically widespread, groups 1f, 1l, and 2a [NiFe]-hydrogenases, can oxidize atmospheric H₂ (Greening et al., 2014a; Myers and King, 2016; Islam et al., 2020; Ortiz et al., 2020). Since its discovery in Actinomycetota, atmospheric H₂ oxidation has been shown to be performed by diverse soil organoheterotrophic isolates from phyla Acidobacteriota, Chloroflexota, and Bacteroidota, spanning from thermophiles to psychrophiles (Greening et al., 2022). Globally, bacteria from nine phyla and 17 genera have been experimentally proven to oxidize atmospheric H₂, and genomic inspections suggest that at least 12 additional phyla contain hydrogenase classes known to intervene in this process (Bay et al., 2021; Ortiz et al., 2020; Greening et al., 2022). One of these phyla is the Bacillota, as P. thermoglucosidasius 23.6 genome contains genes putatively encoding a Huc-type enzyme, herein identified through its sequence.

The strain also contains genes encoding putative [NiFe]-hydrogenases of groups 1d (Hup-type) and 4a (Hyf-type). Hup-type hydrogenases, with low H₂ affinity, are present in a broad range of obligatory aerobic and facultative anaerobic soil-borne, aquatic and host-associated taxa such as Ralstonia eutropha and Escherichia coli (Alvarez et al., 2010; Greening et al., 2016). In anoxia, 1d O₂-tolerant hydrogenases may function to keep intracellular O₂ levels low, thereby protecting O₂-sensitive enzymes during O₂ fluctuations (Volbeda et al., 2013). In P. thermoglucosidasius, a facultative anaerobe, Hup hydrogenase could have a similar function in anaerobic or microoxic environments. However, it could also be functioning under aerobiosis (see section 4.2.), where it might contribute to hydrogen uptake under elevated H₂ concentrations, which may be associated with fermentation under moderate carbon availability, favoring mixotrophic growth.

Formate hydrogenlyases, classified as a group 4a [NiFe]-hydrogenases, are widespread in enteric bacteria with a facultatively fermentative lifestyle (Greening et al., 2016). However, group 4 [NiFe]-hydrogenases have been shown to have a respiratory function, as they associate into complexes comprising primary dehydrogenases and terminal hydrogenases to build minimalistic respiratory chains and conserve the energy liberated during electron transfer as a proton or sodium motive force (Buckel and Thauer, 2013). Although minimalistic, these respiratory chains may be of major importance as a primary strategy for energy generation particularly within oligotrophic environments (Greening et al., 2016); these constitute P. thermoglucosidasius primary habitat, supporting the relevance of the 4a [NiFe]-hydrogenase in this STB.

Following BlastP analysis to identify [NiFe]-hydrogenases orthologous loci in other bacterial taxa, Mohr et al. (2018) referred to the unique combination of [Ni-Fe] group 1- 2a- 4 hydrogenases, which appeared to be restricted to P. thermoglucosidasius strains. Orthologous [NiFe] group 2a uptake hydrogenase loci were common among the Bacillota, but more restricted to Bacilli and within the family Bacillaceae. As Mohr et al. (2018), we also found orthologous [NiFe] group 2a loci among several P. thermoglucosidasius strains with similar organization (Figure 2).

The [Ni-Fe] group 4a H₂-evolving hydrogenase (Hyf-type) locus showed the most restricted distribution of the three hydrogenase-encoding loci among the Bacillota, with orthologous loci only present in P. thermoglucosidasius strains and members of the Thermoanaerobacteraceae family (Class Clostridia) (Mohr et al., 2018). In this cases, Hyf-type loci organization is similar to that reported for Caldanaerobacter subterraneus (Thermoanaerobacteraceae) (Sant’Anna et al., 2015) where 4a H₂-evolving hydrogenase genes are flanked by three genes, cooCSF, coding for a carbon monoxide (CO) dehydrogenase. The coo gene cluster was colocalized with the 4a hydrogenase locus in nine P. thermoglucosidasius genomes inspected by Mohr et al. (2018), but orthologues were not found on the genomes of any other Parageobacillus or Geobacillus spp. P. thermoglucosidasius DSM 2542 was indeed shown to be capable of producing H₂ when cultured in an initial gas atmosphere consisting of 50% CO and 50% air (Mohr et al., 2018). Herein, we have identified putative CO dehydrogenase-encoding genes in the genome of strain 23.6, adding this strain to the collection of P. thermoglucosidasius strains with the potential of CO-dependent H₂ production.

In M. smegmatis, the expression and activities of 1 h and 2a hydrogenases are significantly increased during carbon limitation; experiments showed that this two H₂-scavenging enzymes were mostly synthesized and active when the organism growth is severely limited by its preferred organic carbon sources and with highest activity in stationary phase, when organic carbon is scarce (Berney and Cook, 2010; Berney et al., 2014; Greening et al., 2014a; Greening et al., 2022). Genes encoding the structural components of the groups 1 h and 2a were induced 30-fold and 6-fold, respectively, during slow versus fast growth (Berney and Cook, 2010; Berney et al., 2014). Mutant strains lacking these hydrogenases showed a 40% reduction in long-term viability (Berney and Cook, 2010; Greening et al., 2014b). Also, group 1 h hydrogenases show higher activity during energy limitation in other actinomycetes and during sporulation in streptomycetes (Constant et al., 2010; Meredith et al., 2013). In such cases, H₂ scavenging may function as an electron input to the respiratory chain, H₂ being the primary electron donor used to maintain redox balance and a membrane potential. Although both Huc and Hhy are oxygen-tolerant, contribute to oxidize H₂ at sub-atmospheric concentrations, and enhance bacterial survival during carbon limitation, together they enable M. smegmatis to oxidize tropospheric H₂ at significantly faster rates than cultured streptomycetes harboring only group 1 h hydrogenases (Constant et al., 2010; Greening et al., 2014a). Huc and Hhy are indeed differentially expressed, localized, and integrated into the respiratory chain via the menaquinone pool; Hhy is most active during long-term persistence, yielding energy for maintenance processes ensuing carbon exhaustion, whereas Huc is active in late exponential and early stationary phases, “supporting energy conservation during mixotrophic growth and transition into dormancy” (Cordero et al., 2019b). Hence, besides a physiological role in the release of an electron flux to the aerobic respiratory chain that may create sufficient proton-motive force for cells to persist under starvation, hydrogen scavenging is used in M. smegmatis for mixotrophic growth with organic carbon sources. The wild-type bacterium appears to be unable to grow chemolithoautotrophically using H₂ as the sole electron donor (Berney et al., 2014), but the growth rate and yields of deletion strains of group 1 h and 2a [NiFe]-hydrogenases were significantly reduced compared to those of the wild type during growth on organic carbon sources, showing that M. smegmatis preferentially grows mixotrophically by co-oxidizing organic electron donors and H₂ (Berney and Cook, 2010; Berney et al., 2014; Greening et al., 2014b).

More recently, Islam et al. (2020) showed that the group 2a [NiFe]-hydrogenases, moderately to highly abundant in many soils, is largely distributed across several bacterial phyla and can display optimal expression and activity during growth, supporting a mixotrophic growth. In the tested species, belonging to different phylogenetic lineages and ecological niches, the 2a [NiFe]-hydrogenase expression significantly decreased during the transition from growth to persistence, and the enzyme could be used to oxidize H₂ at sub-atmospheric levels. Hence, such findings demonstrated that “atmospheric H₂ oxidation is not solely a persistence-linked trait”. Bacteria with 2a [NiFe]-hydrogenases may have selective advantages by co-oxidizing H₂ with other organic or inorganic energy sources in environments where nutrient availability is very low or variable. In fact, thermodynamic modeling indicates that atmospheric H₂ used as the sole substrate cannot sustain autotrophic growth, but probably supports mixotrophic growth, especially during growth limitation by organic resources (Conrad, 1999; Bay et al., 2021). In organoheterotrophs, mixotrophic growth may enhance carbon use efficiency by allowing more organic carbon input for anabolism rather than catabolism (Carini, 2021).

Herein, as P. thermoglucosidasius 23.6 growth rate decreases down to near-zero growth, cells triggered the increased gene expression of the putative high-affinity type 2a [NiFe]-hydrogenase (Huc-type), which might have a role in a strategy of mixotrophy to compensate for the scarcity of the organic electron donor, hence supporting energy conservation during slow growth. This enzyme also has a potential role in cellular long-term persistence, at near-zero growth of the bacterial population, when it would provide the energy for maintenance processes that follow carbon depletion. Hence, the enzyme could combine the characteristics of M. smegmatis Hhy and Huc hydrogenases, confirming the dual role of 2a [NiFe]-hydrogenases indicated by Islam et al. (2020). P. thermoglucosidasius 23.6 will gradually increase the expression of the Huc-type enzyme when decreasing cell growth rate, a strategy that would spare cell resources in stressful nutrient-deprived environments and this is certainly related to the extraordinarily adaptability of P. thermoglucosidasius to diverse environmental conditions.

In the methanotroph Methylacidiphilum fumariolicum SoIV, a hup-type hydrogenase contributed to M. fumariolicum autotrophic growth on hydrogen and carbon dioxide, without addition of methane (Mohammadi et al., 2017). In E. coli, the 1d hydrogenase shows maximal expression during fermentation when electron acceptors are sparse, but also under several types of stress such as carbon and phosphate starvation, and under stationary phase conditions (Atlung et al., 1997). Likely, the Hup-type hydrogenase of P. thermoglucosidasius has a role in energy provision under stress occurring in our experimental case during nutrient-limiting growth (i.e., at slow growth rates), suggesting this enzyme provides the bacterium with an additional mechanism to compete in oligotrophic environments, possibly when local hydrogen levels are high enough due to biological processes that are nevertheless limited by nutrient availability.

In contrast to the expression of Hup and Huc [NiFe]-hydrogenase genes, the increase in P. thermoglucosidasius 23.6 Hyf-type hydrogenase gene expression under limiting nutrient conditions is small. P. thermoglucosidasius 23.6 could use atmospheric CO as a supplement during carbon starvation, supporting microbial growth and survival, as already demonstrated for M. smegmatis (Cordero et al., 2019a) and other Bacillota (Gonzalez and Robb, 2000; Wu et al., 2005; Sant’Anna et al., 2015). Hypothetically, the three P. thermoglucosidasius 23.6 [NiFe]-hydrogenases may act cooperatively, with locally evolved H₂ produced by Hyf being recycled by Hup and Huc-type hydrogenases.

In soil ecosystems, H₂ partial pressure (pH₂) can vary by several orders of magnitude across time and space. For instance, H₂ partial pressure decreases with soil depth from ambient atmospheric concentrations (530 ppbv) at the surface to threshold levels (<50 ppbv) at about 10 cm depth, the gradient depends on the microbial H₂ consumption by microbiota (e.g., Smith-Downey et al., 2008). Greening et al. (2015) hypothesized that in zones of high partial pressure of H₂ (pH₂), such as those in the vicinity of root nodules due to the H₂-evolving activity of nitrogenase from N₂-fixing rhizobacteria, the growth of low-affinity H₂-oxidizing bacteria (i.e., Alpha-, Beta-, and Gammaproteobacteria harboring group 1 [NiFe]-hydrogenases) is selected (Vignais and Billoud, 2007). In opposite, atmospheric and sub-atmospheric concentrations would support the survival of high-affinity H₂-oxidizing bacteria, such as the sporulating streptomycetes and persistent mycobacteria. Hence, pH₂ may be a selective factor for the abundance of copiotrophs versus oligotrophs, since atmospheric H₂ may confer a dependable lifeline for the latter, and thus influence microbial community structure. As an example, Osborne et al. (2010) demonstrated that exposure of soils to a moderate pH₂ produced a shift in the soil bacterial community as the relative abundance of ribotypes corresponding to soil actinomycetes Pseudonocardia, Mycobacterium, and Streptomyces species increased. Most importantly, Greening et al. (2015) proposed that atmospheric H₂ scavenging is important for sustaining the survival of microbes in energy-starved soils by assuring the energy input for basic cell maintenance (i.e., for macromolecular repair, cell wall integrity, membrane potential conservation), environmental sensing and structural changes (e.g., in sporulating cells). Thus, H₂ scavenging could provide a way to increase cell persistence in the environment under growth limiting conditions, in sum, by generating energy for persisters. In addition, the potential for stress tolerance of high-affinity hydrogenases such as Huc hydrogenase, shown to be thermostable and largely insensitive to inhibition by O₂ in M. smegmatis (Grinter et al., 2023), combined with the ubiquity of atmospheric H₂ and the abundance of O₂, would enable scavenging to occur under changing environments and severely growth-limiting conditions, thus contributing to a relative stability of the microbial community structures in soils (Lennon and Jones, 2011) and to the maintenance of the viability of growth-limited microbial species.

Even if STB comprise bacteria of closely related genera (e.g., Parageobacillus, Geobacillus, Bacillus, Brevibacillus, Ureibacillus), it is compromising to generalize H₂ scavenging as a strategy to keep cell viability for all, or most STB species. In fact, the above-mentioned studies by Islam et al. (2020) and Mohr et al. (2018) on the distribution of [NiFe]-hydrogenases did not identify the group 2a in the Geobacillus genus, although our BlastP search (results not shown) revealed over 90% identity between 2a [NiFe]-hydrogenase of P. thermoglucosidasius 23.6 and hydrogenases of other soil thermophilic members (e.g., Aeribacillus pallidus previously known as Geobacillus pallidus, Brevibacillus, and Bacillus).

Overall, our work indicates that, besides sporulating streptomycetes and persistent mycobacteria, STB members are part of atmospheric H₂-scavenging soil microbiota. P. thermoglucosidasius populations probably use a putative high-affinity hydrogenase of group 2a in nutrient-limited soils either for slow mixotrophic growth or to persist at near-zero growth. This is in accordance with the source of these thermophiles, transported to cool soils from dust of hot deserts (Marchant et al., 2011). There, they may survive thanks to the activity of high-affinity [NiFe]-hydrogenases for the energy input required for maintenance and to release cytosolic metabolic water, which has been recently theorized to be sufficient to meet hydration needs (Ortiz et al., 2020, 2021). In fact, soil thermophiles such as P. thermoglucosidasius 23.6 are some of the best adapted bacteria to thrive under dry conditions (Gonzalez et al., 2023). Some soil thermophiles, including P. thermoglucosidasius 23.6, show extracellular enzymes with optimum activity at very dried soil conditions (at water activities below 0.6) (Gomez et al., 2020, 2021). Also, P. thermoglucosidasius 23.6 can decompose halogenated pollutants at optimum rates under very dried conditions (water activity around 0.5) in soils (Moxley et al., 2019). It is worth highlighting that STB survive under conditions where the temperature is below their defined temperature growth range (Santana et al., 2020; Milojevic et al., 2022; Gonzalez et al., 2023), namely in the rhizosphere milieu (Rosa et al., 2025), where [NiFe]-hydrogenases could play a fundamental role when these cells drastically reduce their growth rate to near-zero levels. [NiFe]-hydrogenase activity (herein shown to be measurable at 25 °C) and root-exuded metabolites (Rosa et al., 2025) might be sufficient for cell maintenance to assure the persistence of STB in temperate soils.