Subsurface Archaea associated with rapid geobiological change in a model Yellowstone hot spring

Acidification of CP

Historical geochemical data suggest that the water chemistry of Cinder Pool (CP) has been relatively stable from the time of first reported geochemical data in 1947 until autumn 2018, followed by pronounced acidification between winter and spring 2019 (Supplementary Data 1, Fig. 1a, b). Images and documentation dating to even earlier (1927) reveal the presence of cinders covering ~50% of the spring surface at that time, a temperature near boiling (91.5 °C), and a description of having high sulfate and chloride levels (although data was not provided), suggesting that its chemistry has been generally stable since its discovery¹. Spring pH ranged between ~3.6 and 4.5 in 22 yearly measurements spanning 71 years (1947–2018; multiple measurements in the same year were averaged to represent each year) (Fig. 1b), while the pH has been subsequently measured after 2018 as low as 2.5 (Fig. 1b). A single pH measurement of 2.5 was also recorded in a 2003 publication²⁷, although other measurements in 2003, 2000, and 2001 were more consistent with the long-term average (i.e., pH 4.2–4.3; Supplementary Data 1). Scrutiny of chemical data accompanying the pH 2.5 measurement in 2003 indicates a SO₄²⁻ concentration (~48 mg L⁻¹) that is considerably lower than would be expected for CP, even when the pH is much higher (SO₄²⁻ = 80 mg L⁻¹; pH = 4.2–4.3). Considering that sulfuric acid is the predominant buffer of pH in these systems^7,28, the pH 2.5 reading in 2003 is considered questionable. Nevertheless, the 2018 shift in pH towards more acidic conditions was accompanied by a notable change in the appearance of CP. Prior to autumn 2018, the spring waters were cloudy gray with the considerable suspension of kaolinite clay particles²⁰ and black cinders¹⁰. However, between autumn 2018 and spring 2019, the spring waters visibly turned blue-green and contained colloidal S° particles that were also deposited along the pool shelves, while the pool also lacked its characteristic black cinders (Fig. 1a). The spring has maintained this appearance since spring 2019 until at least July 2022.

The source of fluids in YNP hot springs can be broadly defined by concentrations of sulfate (SO₄²⁻) and chloride (Cl⁻)^2,7. These indicators have been previously used to define the source of YNP springs as either (1) hydrothermal only (HO) waters that have moderate concentrations of SO₄²⁻ (~30 mg L⁻¹ depending on the depth of boiling; described below) but high concentrations of Cl⁻ (~300 mg L⁻¹), (2) meteoric-only (MO) waters containing lower concentrations of both solutes, or (3) MO waters infused with gas (MG) that have lower Cl⁻ concentrations and higher SO₄²⁻ concentrations (Fig. 1c). Subsequent boiling and/or evaporation of HO waters can concentrate Cl⁻ and SO₄²⁻ to higher concentrations (termed hydrothermal plus boiling; HB), while additional gas input into HO or HB waters can lead to particularly high concentrations of both Cl⁻ and SO₄²⁻ (hydrothermal + boiling + gas; HBG)⁷ (Fig. 1c). Geochemical data from surveys spanning 1947 to 2018 suggest that CP was largely sourced by hydrothermal (HO) waters that have undergone boiling and/or evaporation (HB) during this time frame (Fig. 1c).

HO and HB waters are typically circumneutral⁷, while CP (which is also sourced by HB waters) has maintained a moderately acidic pH of ~4 until autumn 2018 (Fig. 1b). Several other low pH HB waters have been previously observed within the NGB⁷. The moderately acidic pH in CP (prior to 2018) has been attributed to the hydrolysis of molten S° that occurs at depths of >18 m that leads to the formation of S₂O₃^2– 11. Oxygen (O₂)-dependent oxidation of S₂O₃²⁻, catalyzed by trace iron sulfide in the cinders, forms S_xO₆²⁻ that can then react with sulfide to yield S₂O₃²⁻ and S° ¹¹. Alternatively, S_xO₆²⁻ can be disproportionated to form S₂O₃²⁻ and SO₄^2− 11. The relative rates of these reactions in CP prior to 2018 are not known although similar concentrations of S₂O₃²⁻ measured between 1995 and 1997 suggest that rates of S° hydrolysis and rates of S₂O₃²⁻ formation have been relatively constant over yearly time scales¹¹. The consumption of O₂ by reaction with S₂O₃²⁻ and the consumption of sulfide involving reactions with S_xO₆²⁻ would limit the amount of sulfuric acid that could be formed, thereby maintaining a less acidic pH than other sulfuric acid buffered acidic springs in YNP⁷.

Between November 2018 and March 2019, the pH of CP markedly decreased to 2.8 in 2019, 2.7 in 2020, and 2.6 in 2021. This coincided with a marked increase in SO₄²⁻ concentrations of ~3–5 fold above historical ranges (Fig. 1b), while Cl⁻ concentrations fluctuated without clear trends during this time (Supplementary Fig. 1c). Thus, CP transitioned from an HB water type to an HBG water type between autumn 2018 and spring 2019 and has remained this way since (Fig. 1c). This is interpreted to reflect a substantial increase in H₂S/S° oxidation that results in the formation of SO₄²⁻ and H⁺ (sulfuric acid). Several observations suggest a fundamental restructuring of CP’s unique sulfur cycling due to dramatic physical and chemical changes at this time. As described in more detail below, the molten S° layer was detected at a depth of 18 m in 2016. However, in 2020 and 2021 there was no evidence of molten S° at ~18 to 20 m depth as previously documented, and sampling equipment could be freely dropped to a depth of 22 m (length of the cable) without interruption. In the absence of the molten S° at depth, the S° hydrolysis product S₂O₃²⁻, and the cinders that catalyze S_xO₆²⁻ formation from S₂O₃²⁻ and H₂S, it is possible that such reactions that previously competed for H₂S or O₂ (i.e., those involving S₂O₃²⁻ and S_xO₆²⁻) are no longer taking place in CP. This in turn would allow for sulfur compounds (H₂S and S°) to now be oxidized, thereby contributing to spring acidification.

Alternative scenarios underlying the dramatic changes in CP waters also warrant consideration, and the three most logical are presented below. First, it is possible that the waters sourcing CP may have shifted either via replacement of the primary source or by altered mixing of multiple water sources. Water isotope values (δ²H and δ¹⁸O) can be used to further deconvolute the sources of hydrothermal waters because distinctive isotope values are associated with distinct water sources and the various influences upon them including meteoric water recharge, boiling (and/or evaporation), and water–rock interactions^7,29. The water isotope values measured among the measured depths in CP in 2020 were near the range of water isotope values observed in CP across multiple months in 2016¹³ (depth-resolved water isotope measurements were not made in 2016). The 2020 CP water isotope values were slightly right-shifted relative to those of 2016, suggesting a minor increase in the evaporation and concentration of CP water isotopes between 2016 and 2020⁷ (Supplementary Fig. 2). These data thus do not support the hypothesis that the source of waters in CP dramatically shifted between 2016 and 2020, consistent with the SO₄²⁻ and Cl⁻ measurements indicating that the primary change to CP waters was increased input or availability of H₂S for oxidation.

A second alternative explanation is that a change in the water level of CP could potentially alter residence times which could allow for more oxidation of sulfur compounds in the spring and increased acidification. Such a scenario would also likely result in increased evaporation and concentration of solutes. However, the minimal increase in water isotope values (Supplementary Fig. 2) and similar Cl⁻ concentrations (Supplementary Fig. 1c) accompanying a ~3–5 fold increase in SO₄²⁻ concentration pre- and post-acidification (Fig. 1b) argue that increased residence time was of minimal importance in acidification.

A third possible explanation is that a change in the plumbing system of CP is now delivering more vapor phase gas that contributes H₂S and acidity when oxidized. Such a scenario could be consistent with increased surface deformation, subsurface gas accumulation, and seismic activity that has been taking place near NGB just prior to these changes²¹, and the transition from HB-type to HBG-type waters in CP. Sulfur species isotope analyses would help deconvolute the sources of SO₄²⁻ in CP, but samples for sulfur isotopic analyses were not collected prior to acidification. Thus, it is unclear if this process may also be contributing to the acidification of CP. Regardless, the disappearance of the molten S° cap either by consumption or displacement would in effect make H₂S more available for oxidation, similar to increased vapor phase input. The acidification of hot springs involves the oxidation of H₂S by O₂³⁰. More specifically, partial oxidation of H₂S at acidic pH (<4.0) leads to the formation of S₂O₃²⁻ which is unstable and disproportionates to form sulfite and S^0 28. In the absence of microbial S⁰ oxidation or reduction activities and at temperatures <100 °C, S° is stable and often accumulates in acidic and sulfidic springs²⁸. Aerobic S° oxidizing thermoacidophiles are thought to be the primary catalysts of S° oxidation and this involves reactions that contribute acid^26,28. At temperatures >80 °C, such as in CP, S° oxidation is primarily catalyzed by members of the archaeal order, Sulfolobales^7,28. Taken together, these observations imply that the near-surface acidification of CP waters was enabled by the disappearance of molten S° at depth and the cessation of reactions that S° hydrolysis made possible, including those that consumed H₂S and O₂. It is also possible that the increased input of vapor phase H₂S into CP contributed to its acidification and increased overall sulfur content. In either case, H₂S and O₂ would have been made available for the oxidation of H₂S or S° by thermoacidophiles, possibly related to members of the Sulfolobales^{7,10,15,16,31}.

Subsurface geochemistry and microbial communities in CP

To better understand the existence of a subsurface biosphere in CP, depth-resolved sampling of the water column was conducted in July 2016. The in situ measured temperature of CP water at the surface in 2016 was 85.8 °C and gradually increased to 88.3 °C at 9 m and then abruptly to 93.1 and 95.6 °C at 12 and 15 m, respectively (Fig. 2b; Supplementary Data 2). At ~18 m, the temperature dramatically increased to ~120 °C. Upon retrieval, a metallic gray solidified material covered the sampling apparatus thermistor, weight, and support cable (Supplementary Fig. 3), consistent with previous reports of encountering a similarly viscous material at ~18 to 20 m with a temperature of ~120 °C and that had a metallic appearance upon solidification¹⁰. The marked increase in temperature over this short distance is attributed to the molten S⁰ acting as an insulating lid on the hydrothermal system¹⁰. Sampling was consequently restricted to the 0–15 m depth interval. pH did not vary considerably over the entire depth profile (Fig. 2a, Supplementary Data 2). Like temperature, ORP did not markedly vary over the 0–9 m intervals (−78 to −81 mV) but decreased to ~−117 mV at 12 and 15 m (Fig. 2c; Supplementary Data 2). A minimal correlation was observed between SO₄²⁻ and Fe(II) concentrations with depth, but total sulfide concentrations notably increased with depth (Fig. 2d–f; Supplementary Data 2). These data suggested the presence of two thermally and chemically stratified bodies of water in the water column of CP: one that was more reduced, likely due to the influence of molten S° at depth, and the other that was more oxidized and influenced by contact with the atmosphere.

To continue to assess the possibility of a subsurface biosphere in CP, and to assess its potential role in spring acidification between autumn 2018 and spring 2019, depth-resolved sampling was conducted again in August 2020. The same sampling strategy was used to sample CP as in 2016. As observed in 2016, the temperature varied minimally between the surface (86.4 °C) to 9 m (87.3 °C) and slightly increased to ~90 to 91 °C at 12–15 m (Fig. 2b; Supplementary Data 2). As mentioned above, a molten S° layer was not encountered while lowering the sampling equipment to 22 m in 2020 and 2021 (Supplementary Fig. 4) and the temperature increased only to 93 °C, consistent with the absence of an insulating S° lid. Spring pH did not vary substantively across the depth profile in 2020. In contrast to the 2016 sampling profile, ORP did not markedly vary with depth, nor did SO₄²⁻ and Fe(II) concentrations, although sulfide again increased with depth (Fig. 2c–f; Supplementary Data 2). Notably, measurements of ORP in the 2020 CP waters were substantially higher by ~400 mV (~+270 to +280 mv) than those of the 2016 CP waters (−78 to −117 mV), indicative of a much more oxidized and acidic spring. This is consistent with sulfuric acid buffering of this spring to a pH of 2.5–2.6⁷.

To investigate the abundance, composition, and functional potentials of CP communities in 2016 and 2020, biomass was filtered from 0, 9, and 15 m depths and subjected to DNA extraction and shotgun metagenomic sequencing. Total DNA concentrations, normalized to L of spring water filtered, were much lower in the 2016 water samples than in the 2020 samples (Fig. 3a; Supplementary Data 2), suggesting lower overall planktonic biomass in 2016. The decreased biomass detected in samples from 2016 may be due to increased evidence for viral infections in those cells at that time (discussed below) that, if lytic, would be expected to decrease the number of cells captured on filters. Interestingly, while DNA concentrations considerably declined with depth in the 2016 waters, DNA concentrations increased with depth in the 2020 waters (Fig. 3a; Supplementary Data 2). This may point to the presence of populations in the 2020 samples that were better adapted to growing under subsurface conditions (e.g., lower O₂ and increased H₂S/S° concentrations relative to the surface). Measurements of the abundance of planktonic cells in waters collected from 2021 at 0, 9, and 21 m confirmed the increase of biomass with depth (Fig. 3b). Moreover, the presence of abundant cells at 21 m in samples from 2021 suggests that cells could be present at even greater depths within the CP subsurface.

Shotgun metagenomic sequencing, assembly, and binning were used to reconstruct population-level genome bins from samples collected in 2016 and 2020. Taxonomically, both 2016 and 2020 CP communities comprised only Archaea and were dominated by members of the class Thermoprotei (Table 1 and Fig. 3c), consistent with previous characterizations of acidic, high-temperature hot spring communities in YNP^16,32, including CP^13,20. Taxonomically, the community structure did not considerably differ with depth in either the 2016 or 2020 water samples (Fig. 3c), suggesting the presence of generally homogenous communities along depths. Metagenome-assembled-genomes (MAGs) belonging to a group of uncharacterized Sulfolobales were dominant members of all communities in 2016 and 2020, followed to a lesser extent by those of uncultured Acidilobus sp. and Vulcanisaeta sp. (Fig. 3c). These compositions are also consistent with 16S rRNA gene community profiles of CP surface waters from across ~5 months of 2016¹³. In addition, a second Vulcanisaeta sp. MAG was also present in low abundance in all three 2020 water communities along with several low abundance Candidatus Nanopusillus-affiliated MAGs (Table 1). Lastly, a Stygiolobus-like population was also present in the 0 m 2020 water communities (Table 1).

Although the community compositions were generally similar in 2016 and 2020, their structures varied, with the uncultured Sulfolobales more abundant in the 2020 post-acidification waters, regardless of sample depth (Fig. 3c). Differences in estimated abundance were largely due to the presence of much higher proportions of putatively viral (and/or non-cellular) genomic DNA in the 2016 community metagenomes than those in 2020 (Fig. 3c). Although further characterization of these putative viral contigs was not conducted, the similarity of many of these contigs to those from previously characterized archaeal viruses from YNP (Supplementary Data 3) suggests that the viral load in CP waters in 2016 was overall higher than in waters from 2020. These results are potentially consistent with more stressful conditions for thermoacidophiles (i.e., from higher temperatures and more reducing conditions) in CP waters sampled in 2016 compared to those from 2020. The near homogenous composition of the communities sampled in 2016 and 2020 (Fig. 3c), albeit in decreasing absolute abundance with depth (based on total DNA concentrations) in 2016 and increasing in abundance with depth in 2020 (Fig. 3a, b), suggests that strain level functional variation among populations comprising each assemblage may enable higher fitness in surface waters in 2016 and deeper waters in 2020.

Strain-level shifts in CP populations coinciding with acidification

To further evaluate the involvement of CP populations in hot spring acidification and to identify potential adaptations that may enable higher fitness in surface versus subsurface waters, the MAGs were subjected to phylogenetic and metabolic reconstructions. MAGs for the three primary populations (Uncultured Sulfolobales, Acidilobus sp., and Vulcanisaeta 1) that were present among all three water depth communities and in both 2016 and 2020 communities comprised distinct monophyletic groups corresponding to each year (Fig. 4). The 2016 and 2020 Sulfolobales and Acidilobus populations were not most closely related to corresponding groups in CP waters from the other sampling year, but rather to genomes previously sampled from other YNP hot springs in other studies (Fig. 4). In contrast, the Vulcanisaeta 1 populations were also monophyletic within each year’s communities, but the 2016 and 2020 MAGs were most similar to one another, to the exclusion of Vulcanisaeta MAGs recovered from other springs in YNP. Thus, the altered geochemical conditions of CP that occurred between 2016 and 2020 selected for slightly different strain-level variants of the three pre-dominant populations, in addition to other low-abundance Archaea that were only present in 2020 (Vulcanisaeta 2, Stygiolobus, and Ca. Nanopusillus). The Sulfolobales and Acidilobus MAGs from CP in 2016 and 2020 (in addition to others sampled from CP sediments in 2018 and 2019) were phylogenetically separated by MAGs recovered from numerous other hot springs across YNP and from years ranging from 2008–2018, suggesting that the shifting dominance of strain-level variants in CP between 2016 and 2020 was due to the recruitment and/or proliferation of strains from a pool of metapopulation genotypes distributed across YNP.

Genomic plasticity of a subsurface archaeon

The uncultured Sulfolobales that dominated CP waters in 2016 and 2020 belonged to a family-level lineage comprising closely related genomes (all with >95% whole-genome amino acid identity to each other) recovered from numerous acidic or slightly acidic springs in YNP (Figs. 4a and 5)^{26,32,33,34,35}, and that is hereafter referred to as the uncultured YNP Sulfolobales (UYS) group. A total of 22 UYS genomes were subjected to further analyses, comprising six from CP waters in 2016 and 2020, three from CP sediments in 2018 and 2019, four previously published MAGs from other hot springs, and nine other MAGs generated from eight other acidic springs across YNP. While the UYS were present in hot springs with pH ranging from 1.6 to 5.5, the spring pH from where they were identified was highly associated with the phylogenetic placement of individual MAGs (Fig. 5). Specifically, the deepest branching MAGs were generally present in the least acidic springs (pH 4.0–5.5; with the exception of one MAG from Moose Pool [pH 1.6]; inclusive of the 2016 CP MAGs), while the most derived MAG populations were only present in highly acidic springs (pH 2.3–3.7; with the exception of one MAG from Cinder Pool sediments in 2018 [pH 4.0]; inclusive of the 2020 CP MAGs) (Fig. 5). These trends are consistent with the hypothesized trajectory of thermoacidophile evolution towards greater acid tolerance/preference¹⁶. Moreover, these trends may also be consistent with niche construction models, wherein H₂S/S° oxidation by thermoacidophiles helps construct increasingly acidic environments via acid production, which then concurrently drives adaptation to these increasingly acidic environments¹⁶.

To assess the potential role of the UYS in sulfur biogeochemical cycling, the metabolic functional potentials of these populations were evaluated in greater detail based on their reconstructed genomes (Fig. 5, Supplementary Data 3). The UYS encoded the capacity for autotrophy via full complements of enzymes involved in the 3-hydroxypropionate/4-hydroxybutyrate cycle (3HP-4HB) (Supplementary Data 4), consistent with the general potential for autotrophy in most other Sulfolobales³⁶. Consistently, the SoxM subunit that has been suggested as a marker for (facultatively) heterotrophic growth of Sulfolobales³⁷ was absent in all UYS MAGs (Fig. 5, Supplementary Data 4). Given that all known Acidilobus and Vulcanisaeta spp. are characterized heterotrophs without known autotrophic capacity^38,39, the UYS are likely the sole primary producers in the CP surface and subsurface waters, consistent with their considerable dominance in CP water communities over time.

Also consistent with almost all other Sulfolobales³⁶, the UYS universally encode the ability to reduce O₂ via terminal cytochrome oxidases, although not via Sulfolobus oxidase (SoxABCD) complexes that are common among many Sulfolobales³⁶ but rather via Desulfurolobus oxidase complexes (DoxBCE) (Fig. 5, Supplementary Data 4). An additional terminal oxidase complex (CbsAB-SoxLN) was encoded in the 2020 CP MAGs along with several other UYS MAGs from other YNP springs, although homologs of CbsAB-SoxLN were not present in the 2016 CP MAGs or several others recovered from sediments of other hot springs (Fig. 5). Thus, a potentially important metabolic difference between the pre- and post-acidification (2016 and 2020, respectively) CP Sulfolobales was the ability to use different terminal cytochrome oxidase compliments for aerobic respiration. The capacity to use multiple terminal oxidases has been suggested as an adaptation to varying oxygen tensions/availabilities^37,40 that likely substantively differed between the low ORP 2016 CP waters and the high ORP 2020 CP waters (Fig. 2c). Consequently, these data point to the ecological succession of UYS strains within CP that are, at least in part, related to strain-level differences in aerobic respiration capacities.

A defining feature of most cultured Sulfolobales is the ability to grow chemolithoautotrophically by coupling the oxidation of sulfur compounds (e.g., S⁰) to aerobic respiration³⁷. The slow kinetics associated with abiotic oxidation of S⁰ with O₂ at temperatures <100 °C combined with the widespread distribution of Sulfolobales in moderately acidic to acidic high-temperature hot springs has been used to support their role as key mediators of sulfur compound oxidation and hot spring acidification^7,16,28. The membrane-bound tetrathionate hydrolase (TetH) and intracellular sulfur oxygenase reductase (SOR) proteins that are present among many Sulfolobales and that mediate the production of S° from tetrathionate (S₄O₆²⁻) and the disproportionation of S° to H₂S and SO₃²⁻/S₂O₃⁻, respectively^36,37, were absent in all UYS MAGs (Fig. 5). The absence of SOR has been suggested to preclude the ability of Sulfolobales to use S° as an electron donor³⁷, although Sulfolobales (i.e., Metallosphaera) are known to use S° as electron donor but do not encode SOR^41,42. Further, SOR is not necessary for the growth of acidophilic bacteria (e.g., Acidithiobacillus) via S° oxidation and is thus, likely a supplementary enzyme for S° oxidation⁴³.

Recent experimental studies of acidophilic S°-oxidizing bacteria have indicated that the heterodisulfide reductase complex (HdrABCB₂C₂)/sulfane-sulfur carrier protein TusA/DsrE3 sulfurtransferase/lipoate-binding protein (LbpA) intracellular system is necessary for growth via S° oxidation^44,45. These proteins are highly conserved across Sulfolobales³⁶ and are proposed to operate similarly as in bacterial S° oxidation^42,44, although the exact mechanism by which S° is transferred to the sulfur-carrier proteins is currently unclear⁴⁴. These proteins are nevertheless conserved in nearly all UYS MAGs (and in nearly all Sulfolobales³⁶), with the notable exception of the 2016 CP UYS (Fig. 5, Supplementary Data 4). Given the high level of estimated genome completeness for the 2016 CP UYS (95–100%; Supplementary Data 4), their high level of genomic sequencing coverage (~2200–16,700×; Supplementary Data 4), the presence of most intracellular S° oxidation proteins in most UYS genomes, and the presence of at least some of these genes in MAGs or the unbinned contigs of their corresponding metagenomes suggests that the 2016 CP UYS indeed lacked the genes encoding these proteins and intracellular sulfur oxidation capacity in general. The co-localized HdrABCB₂C₂ gene structure, along with associated lipoate binding/ligase proteins and sulfur-carrier proteins, that are all syntenous within a 30–40 gene locus present among all UYS MAGs with available data for this genomic region, supports this assertion (Supplementary Fig. 5). Moreover, the core syntenic gene structures share considerable similarity to model Sulfolobales shown to upregulate Hdr expression during autotrophic S° oxidizing growth (e.g., Metallosphaera cuprina, Supplementary Fig. 5)⁴⁶, along with bacterial S° oxidizers, where this complex has been validated⁴⁴.

All UYS MAGs also encoded the ability to oxidize H₂S via sulfide quinone oxidoreductases (SQO), with many MAGs even encoding multiple copies of sqo (Supplementary Data 4). Further, the coupling of H₂S oxidation to O₂ reduction could be mediated by thiosulfate–quinone oxidoreductase (TQOab) complexes that are encoded by nearly all UYS. Thus, the UYS generally appear to be able to couple the oxidation of S₂O₃⁻, S° (via sulfur carrier proteins), and H₂S via TQO, Hdr/TusA/DsrE3/lipoate-metabolizing proteins, and SQO, respectively, to the generation of a reduced quinone pool that can be used to reduce O₂ via cytochrome oxidases (Fig. 5, Supplementary Data 4).

Aerobic oxidation of S₂O₃⁻, S°, and H₂S can contribute to the production of H⁺ (either through direct oxidation to SO₄²⁻ and H⁺ or through the production of intermediates that could then be further oxidized to SO₄²⁻ and H⁺), and all of these sulfur compounds are critical focal points in the acidification of hot spring waters⁴⁷. Thus, aerobic oxidation of these sulfur compounds is likely also occurring in the near subsurface of other regions of YNP where similar geochemical conditions exist. The lack of proteins allowing for intracellular S° oxidation by the 2016 CP UYS suggests that they may have lacked the ability to oxidize sulfur species and were limited to the use of H₂S via SQO and S₂O₃⁻ via TQO, potentially generating S° as a by-product. As such, they may have had a decreased potential to contribute to sulfur oxidation and H⁺ production compared to 2020 CP UYS. Notably, CP was previously considered anomalous due to its high concentrations of S₂O₃⁻ which were among the highest in YNP up until 2018¹¹. This is due to S₂O₃⁻ being unstable at pH < 4²⁸ and thus not generally being found at high concentrations in acidic hot springs, with the exception of CP¹¹. Consequently, the 2016 CP UYS may have been uniquely adapted to the S₂O₃²⁻ replete historical conditions of CP, but this population may have been quickly outcompeted by the 2020 CP UYS strain upon the disappearance of the molten S° and subsequent (or concomitant) acidification of CP (and thus, the disappearance of abundant S₂O₃⁻). Consequently, the 2020 CP UYS strain is uniquely poised to utilize H₂S and S° compounds that would have been more available in CP following its acidification, and thus the strain would have outcompeted the 2016 strain based on the availability of sulfur compounds as electron donors.

Source: Ecology - nature.com