Potassium transport systems are found in many bacteria as potassium is the major intracellular cation responsible for osmotic regulation, activating intracellular enzymes, regulating internal pH and as a secondary messenger (Epstein, 2003; Korolev, 2021; Raven, 2021). It is widely accepted that potassium transporters are regulated by ion concentrations and turgor pressure (Epstein, 2003). In terms of acid tolerance in extreme acidophiles, potassium transporters are hypothesized to be responsible for the reversed membrane potential, which is the primary mechanism reducing proton influx (Baker-Austin and Dopson, 2007). In halophiles, it has been shown that potassium transport systems aid in osmotic stress and typically halophiles have multiple copies of potassium transport genes (Epstein, 2003; Kraegeloh et al., 2005; Mongodin et al., 2005).

The Trk potassium system symports K⁺ and H⁺ into the cell and has been identified in plants, Bacteria and Archaea (reviewed in Epstein, 2003; Pandey and Mahiwal, 2020). In prokaryotes the most studied Trk system is in Escherichia coli and consists of TrkH and TrkG, the multimeric functional transporters; TrkA, the cytoplasmic regulating subunit; and TrkE (also known as sapD), the ATP-binding component of the Trk system (Epstein, 2003; reviewed in Stautz et al., 2021). In E. coli, the genes involved in the Trk system are distributed throughout the genome, however in halophilic organisms trkA and trkH have been found clustered together (Nakamura et al., 1998; Kraegeloh et al., 2005). Many acidophile genomes lack trkHG, and only contain the trkA homolog (exception Alicyclobacillus ferrooxidans) (Rivera-Araya et al., 2020). In Acidihalobacter spp. and H. neapolitanus c2 the TrkAH potassium transporter system was present. The system is in a highly conserved syntenic region in all genomes and the genes are arranged contiguously (Supplementary Figure 1). A second copy of trkH was also found in the Acidihalobacter genomes directly downstream of the first trkH copy. The genomic arrangement of the Trk system in Acidihalobacter and H. neapolitanus c2 is very similar to that found in Vibrio alinolyticus (a halophile) with fmt (methionyl-tRNA formyltransferase) and rsmB (a ribosomal RNA small subunit methyltransferase B, also identified as fmu in other bacteria) being present upstream of trkA (Supplementary Figure 1) (Nakamura et al., 1998). Three extra genes were found between rsmB and trkA in Acidihalobacter and H. neapolitanus, a predicted gene that produces a protein with an unknown function, a histidine kinase, and a sigma-54 RNA polymerase holoenzyme. Both the histidine kinase and σ-54 proteins are predicted to be nitrogen regulated and involved in nitrogen assimilation (ntrY and atoC also known as ntrX). Trk proteins have been identified in the genomes of other acidophiles such as Leptospirillum (Vergara et al., 2020), however, information about how the Trk system is arranged in their genome is lacking. No other case of a Trk system with two adjacent TrkH proteins has been published. Multiple copies of trkA and trkH have been found in the same region in Salinibacter ruber, however, they are not juxtaposed (López-Pérez et al., 2013). Whether this alteration results in a more efficient adaptation to acidity or salinity remains to be explored. Evidence of HGT in the genome surrounding the Trk system in Acidihalobacter was not found, and phylogenetic analysis of the protein sequences and their best BlastP hits show the Acidihalobacter TrkH proteins clustering with other TrkH proteins found in other organisms in the same Chromatiales order (Supplementary Figure 2). In light of the Trk system being in a highly conserved region of the genome in both Acidihalobacter and H. neapolitanus c2 and the Trk proteins all clustering with Trk proteins in the Chromatiales order, vertical descent of the Trk system in Acidihalobacter is likely. Whether the alteration of the gene structure in Acidihalobacter and H. neapolitanus when compared to Trk systems in other halophiles results in differential control or efficiency of Trk in these bacteria remains to be investigated.

Kdp is a high-affinity potassium transport system known to be involved in potassium homeostasis and the osmotic stress response in many bacteria, encompassing halophiles and non-halophiles, and is responsible for the active influx of potassium ions in environments with low potassium concentrations (Epstein, 2003; Strahl and Greie, 2008; Kixmüller et al., 2011; Price-Whelan et al., 2013). The Kdp system is also present in many acidophiles and has been found to be actively transcribed in acidic environments (Bakker et al., 1987; Acuña et al., 2013; Cholo et al., 2015; Rivera-Araya et al., 2019, 2020; Vergara et al., 2020). In E. coli the Kdp system consists of two operons, one encompassing the structural genes kdpFABC, and the other encompassing the regulatory genes kdpDE (Walderhaug et al., 1992). These two operons overlap in E. coli, however, different genetic arrangements of the system are found in other bacteria (Walderhaug et al., 1992; Epstein, 2003). KdpF is involved in stabilizing the Kdp complex but is not essential. KdpA is responsible for binding and translocating the K⁺ and KdpB and KdpC are responsible for the ATP hydrolysis (Hesse et al., 1984; Gaßel et al., 1998, 1999; Dorus et al., 2001; Reviewed in Epstein, 2003). The kdpDE operon is a two-component response regulator responsible for regulating the Kdp complex as KdpD is a histidine kinase and KdpE is a response regulator (Walderhaug et al., 1992; Reviewed in Epstein, 2003). The Kdp system was only identified in the genome of A. aeolianus. The gene arrangement of the system was similar to that of E. coli (Supplementary Figure 3A) (Walderhaug et al., 1992). Both carbonic anhydrase (can) and the sigma-54 regulator (σ-54) have been identified in the same genomic context and previously have been indicated to be involved in acid tolerance of other microorganisms (Sachs et al., 2006; Bury-Moneì et al., 2008; Mitra et al., 2012, 2014). HGTector analysis predicts KdpA, KdpB, KdpC, KdpD, KdpE, and the sigma-54 transcription regulator to have been gained through HGT from a Bacteria, and the can from a Proteobacteria. The sigma-54 regulator, a hypothetical zinc-containing protein and the can upstream of the Kdp system all have significant similarity to their corresponding proteins in multiple Acidithiobacillus spp. (Acidithiobacillus ferriphilus and Acidithiobacillus ferrivorans) as well as the closely related Ambacidithiobacillus sulfuriphilus and Fervidacidithiobacillus caldus acidophiles. The lack of kdp genes in all other Acidihalobacter genomes and the outgroup suggests it has been acquired via HGT, as well as the genes and gene contexts having significant similarity to species in the Burkholderiales order. We propose that the Kdp system in A. aeolianus was a result of HGT and has the potential to aid acid tolerance in low potassium environments.

Kch is a voltage-gated membrane potassium transporter commonly found in plants but has also been identified in E. coli, although its biological role is currently not understood (Reviewed in Epstein, 2003; Lundbäck et al., 2009). In eukaryotes, it is recognized as a voltage-gated ion channel that is associated with membrane potential regulation (Capera et al., 2019). The Kch protein was identified in all Acidihalobacter genomes and was not identified in H. neapolitanus. Best-hit analysis of the Kch proteins to the NCBI BlastP non-redundant database revealed that the Kch proteins in Acidihalobacter have significant similarity (46–53%) to Kch proteins from the acidophiles, F. caldus and Am. sulfuriphilus. putA (a bifunctional proline dehydrogenase/L-glutamate gamma-semialdehyde dehydrogenase) was found closely associated with kch in A. yilgarnensis and A. prosperus and is known to be involved in osmotolerance. Phylogenetic analysis, however, shows the Kch proteins from Acidihalobacter clustering closely to Kch proteins from other Chromatiales, and not the acidophiles Kch proteins (see Supplementary Figure 4). As Kch proteins are present, and cluster closely to the Chromatiales proteins, it is likely the Kch protein is a result of vertical decent in the Acidihalobacter genomes. The proximity of the kch to other osmotic tolerance genes may point to this area of the genome being important to both acid and osmotolerance in Acidihalobacter.

Spermidine is a positively charged membrane associated aliphatic, polycation polyamine that is possibly involved in acid and osmotic stress responses in E. coli and has been shown to reduce the outer membrane permeability caused by porins (Samartzidou et al., 2003). Spermidine biosynthesis pathways vary between bacterial species. Arginine is the precursor which can either be converted to citrulline or ornithine (via ArcB and RocF, respectively) which is then converted to putrescine (via SpeC) and subsequently converted to Spermidine via SpeE. An alternative pathway also includes arginine conversion to agmatine (AguA), which can be either converted directly to putrescine via SpeB, or to an intermediate CPT by AguA. The full spermidine biosynthesis pathway identified in Acidihalobacter is summarized in Figure 2. Both speE and speH (involved in the production of dSAM important in the conversion of putrescine to spermidine) were identified in all Acidihalobacter genomes and had significant similarity (66–68%) to SpeE and (73–79%) SpeH proteins in the closely related sister clade Thioalkalivibrio genus and other bacteria in the Chromatiales order. speE is found in the same genomic context as the urease accessory genes. speH is in a separate part of the genome that is also conserved between the Acidihalobacter genomes. Genes involved in osmotic tolerance are found upstream of speH including yhfA (a protein belonging to the osmotically inducible OsmC superfamily). As spermidine has been shown to be involved in both acid and osmotic stress responses, it is likely that spermidine synthesis in Acidihalobacter has been acquired via vertical descent from a common ancestor considering the presence of spermidine genes in other organisms from Chromatiales order and has possibly been lost in the H. neapolitanus outgroup genome.

Hopanoid biosynthesis is hypothesized to be involved in regulating the fluidity and permeability of bacterial membranes, aiding in stress responses involved in acidity, temperature and salinity (Belin et al., 2018). Squalene is the precursor to hopanoid biosynthesis and the genes involved in it (hpnDE and hpnC) were predicted in all Acidihalobacter genomes, and were not identified in the H. neapolitanus genome. hpnDEC all had high similarity to orthologous proteins in the Thioalkalivibrio genus (>90%). A. ferrooxydans was the only genome that had the complete set of genes required to produce bacteriohopanetetrol (BHT) (hpnFHG) (the full hopanoid biosynthesis pathway can be seen in Figure 3). HpnF and HpnH both had high similarity to proteins from the Nitrosomonas genus (61.5 and 74–68%, respectively) and HpnH also had high similarity to proteins from microbes in the Methylococcaceae family (70–67%). hpnH, ispH, hpnF, hpG, and hpnA are all in a region of the genome with no synteny with the other Acidihalobacter genomes. An assembly gap was detected in the genome adjacent to hpnA, thus the full genomic context and evidence of HGT could not be determined.

The outer membrane “starvation lipoprotein” gene slp is associated with the protection of E. coli during carbon starvation and metabolites that are toxic in low pH environments (namely organic acids) (Alexander and St John, 1994; Hommais et al., 2004; Mates et al., 2007). slp is located in the “acid fitness island” of E. coli and is co-expressed with other acid resistance genes. Two copies of slp were identified in all species of Acidihalobacter, excluding A. ferrooxydans, which only had one copy, and was absent in the outgroup. Slp copies contained a characteristic lipobox motif (Zückert, 2014; Vergara et al., 2020) with an Alanine amino acid residue in the position +2, suggesting an exportation of Slp (Supplementary Figure 5). Both copies, named slp-1 and slp-2, are adjacent to each other in A. yilgarnensis, A. prosperus and A. aeolianus, and the area of the genome is conserved between the three species (see Supplementary Figure 6). A cardiolipin synthase (cls) gene was found upstream of slp in all Acidihalobacter genomes and is involved in the structural organization of membranes and is associated with osmoregulation and acidic resistance (Romantsov et al., 2009; MacGilvray et al., 2012). The cardiolipin synthase presented higher similarity with Planctomycetes microorganisms according BlastP results. engB, a gene associated with the slp and glutamate decarboxylase genes in E. coli, was also identified in the same genomic context as Acidihalobacters slp (Mates et al., 2007). NhaP, a K⁺(Na⁺)/H⁺ antiporter, and lysR, a transcriptional regulator, were identified downstream of slp in A. ferrooxydans and both are associated with acid stress response of Vibrio cholerae (Kovacikova et al., 2010; Ante et al., 2015; Mourin et al., 2019). Phylogenetic analysis of the proteins shows the Slp-1 protein clustering with Slp proteins of Acidithiobacillus, and Slp-2 clustered with Chromatiales Slp proteins (see Figure 4). HGT events signals were identified for two genes upstream of slp-1 (AOU99095.1–AOU99096.1) by HGTector, suggesting Proteobacteria donor. This information, in addition to the genomic arrangement and phylogenetic analysis, allow us to hypothesize that genomic segment of Acidihalobacter including slp-1, cardiolipin synthase and genes upstream were gained by HGT events from a Proteobacteria donor, such as the extreme acidophile Acidithiobacillaceae; meanwhile slp-2 copies correspond to an ancestral gene from Chromatiales.

The glutamate decarboxylase system has been identified as a major mechanism in the acid stress response of E. coli (reviewed in Foster, 2004). Glutamate is transported into the cell by GadC (an amino acid permease), and GadA or GadB (the glutamate decarboxylase isozymes) replaces the a-carboxyl group of glutamate with H⁺, creating y-amino butyric acid (GABA) and CO₂ removing a proton from the cytoplasm in the process. GABA is subsequently removed from the cell by GadC. The Gad decarboxylase system has also been identified in extreme acidophiles such as Leptospirillum spp. and F. caldus (Mangold et al., 2013; Vergara et al., 2020). GadA and GadC were both identified in all Acidihalobacter genomes but was missing in the H. neapolitanus genome. gadA and gadC in A. aeolianus and A. ferrooxydans were separated by a protein with no predicted putative domains. This hypothetical protein was predicted to be 248 amino acids and 7 transmembrane domains by HMMER, and has matches to ArcD, an arginine/ornithine antiporter in the HMMER database. In all genomes, best-hit analysis of GadA and GadC proteins showed 100% coverage and high amino acid percentage identity (ranging from 75 to 85.12%) to orthologs in Salinisphaera sp. LB1 (a haloacidophile), Mangrovitalea sediminis (a halophile), and Acidithiobacillus spp. (including At. thiooxidans, At. Ferridurans, and At. ferrivorans). This is an indication of a possible gene gain event in Acidihalobacter (see Supplementary Figures 7A,B).

The Arginine decarboxylase system is another acid stress tolerance system identified in E. coli that works in the same way as the glutamate decarboxylase system (reviewed in Richard and Foster, 2004). Arginine decarboxylase, encoded by the adiA, replaces the a-carboxyl group of arginine with a proton, producing CO₂ and agmatine, which is removed from the cell via antiporter AdiC. All Acidihalobacter genomes had hits to adiA, however, no hit to adiC was identified (no arginine decarboxylase genes were found in the H. neapolitanus genome). The complete arginine synthesis pathway was identified (argABCDEFGHO), thus Acidihalobacter may be capable of producing its own arginine, to be used in the arginine decarboxylase system. Best hits of the Acidihalobacter AdiA proteins to the BlastP non-redundant database indicated significant similarity to AdiA proteins of other Chromatiales organisms, suggesting vertical descent.

One of the key mechanisms used by the enteric pathogen H. pylori to combat acid stress is the production of a highly expressed urease enzyme (Marcus and Scott, 2016). Urease converts urea to ammonia which is protonated to NH₄⁺, neutralizing its environment. Urea import in H. pylori is controlled by UreI, a proton gated inner membrane urea channel, found to be activated by low pH (Bury-Moné et al., 2001; McNulty et al., 2013). Further buffering capacity is achieved by carbonic anhydrase which converts the carbon dioxide produced as a by-product of the urease activity into carbonic acid, which has a higher buffering capacity compared to ammonium (Marcus et al., 2005). The urease enzyme is coded for by ureA, ureB, and ureC (urease subunit gamma, beta, and alpha, respectively) and ureD (an ortholog of ureH), ureE, ureF, and ureG all of which are accessory proteins necessary for the expression of a functional urease enzyme (Akada et al., 2000). In H. pylori the regulation in the expression of the urease complex is dependent on the cellular pH and environmental pH, conditions sensed by the ArsRS and FlgRS two-component system. Urease is also found widely in soil bacteria, cyanobacteria and the extreme acidophile Ferrovum (Ullrich et al., 2016a) and studied in halophilic alkaliphiles in the production of biocement (Burbank et al., 2012; Dhami et al., 2013; Stabnikov et al., 2013; Veaudor et al., 2019). In the previously mentioned cases the urease activity is largely related to nitrogen assimilation and microbially induced calcite precipitation (Burbank et al., 2012). In cyanobacteria urea transport is controlled by the urea transport genes urtABCDE (Veaudor et al., 2019).

In all the Acidihalobacter genomes, ureA, ureB, ureC, ureD, ureE, ureF, and ureG were detected (ureC in A. yilganensis has multiple frameshift mutations resulting in a truncated version of UreC). These genes were not detected in the outgroup H. neapolitanus c2. The gene cluster structure of the urease system was unique in the Acidihalobacter genomes with a cyanase (cynS) gene between ureA and ureB. Cyanase is an enzyme that converts cyanate to carbamate, which spontaneously converts to ammonia and carbon dioxide (OCN^– + HCO₃^– + 2H⁺ ↔ NH₃ + 2CO₂), directly consuming two protons as well as more protons consumed via the spontaneous conversion of ammonia to ammonium (NH₃ + H⁺ ↔ NH₄⁺) (Anderson and Little, 1986). Reporting of cynS incorporated into the urease operon in other genomes has not been found and could possibly be a novel alteration in Acidihalobacter allowing for further buffering capacity. As previously stated, H. pylori contains arsR, a pH activated transcriptional regulator, responsible for the transcription of the urease operon in low pH environments (Pflock et al., 2005). The arsR transcriptional regulator was found in A. yilgarnensis, A. prosperus, and A. aeolianus. This regulator was found upstream of the urease operon and also encompasses the genes adiA and speE (Figure 5). SpeE is a polyamine aminopropyltransferase in E. coli that is involved in the production of spermidine and is found to be involved in the acid tolerance of E. coli, possibly through the regulation of a wide range of acid tolerance genes including the glutamate decarboxylase system and the arginine decarboxylase system (Rhee et al., 2007). Whether or not the urease operon, adiA and speE are controlled by the arsR transcriptional regulator remains to be seen and must be tested further. The only urease transport system detected was urtABCDE in A. ferrooxydans and two copies in H. nepolitanus c2. In the case of H. pylori, gastrointestinal urea concentrations are very low (∼3 mM), and thus requires the active transport of urea into its cell for the urease system to be effective at combating acid stress (Scott et al., 2000). Environmental urea typically ranges from 0 to 13uM in concentration, over 2 orders of magnitude lower than the case of H. pylori; thus, it is unclear if the urease system found in A. yilgarnensis, A. prosperus and A. aeolianus would be effective enough in combating acid stress if no active urea transport system is found. A. ferrooxydans may be able to use the urease system as an acid resistance mechanism more effectively than the other member of Acidihalobacter due to the presence of the active urease transporter system in its genome. Further proteomic studies would be needed to validate this hypothesis. Due to the lack of the urease gene cluster in the outgroup and the similarity of the Acidihalobacter and Betaproteobacteria urease gene clusters, we speculate that the urease system was acquired via HGT in an ancestral Acidihalobacter.

Carbonic anhydrase (CA) is a proton-consuming enzyme that has been found to be upregulated in response to acid stress in H. pylori and the lack of the enzyme delays growth of H. pylori in acidic conditions (Sachs et al., 2006; Bury-Moneì et al., 2008). CAs are found in the cytoplasm and thus would directly contribute to minimizing the effect of protons leaking into the cytoplasm. CA is also known to catalyze the dissolution of carbonate minerals and in Aspergillus fumigatus is upregulated in potassium limited environments to increase the dissolution of potassium minerals to obtain potassium (Xiao et al., 2012; Sun et al., 2013). CA have also been found in acidophiles, but conclusions about their involvement in acid tolerance is not clear (Levicán et al., 2008; Li et al., 2012; Acuña et al., 2013). HGTector predicted the CA gene to have been acquired in the A. aeolianus and the A. yilgarnensis genomes from a Proteobacteria. The proximity of the CA to the Kdp system in A. aeolianus, as well as its clustering with CA proteins from the extreme acidophile class Acidithiobacillia may indicate an involvement in acid tolerance. Although, best BlastP hit analysis had many top hits to CAs of the extreme acidophile class Acidithiobacillia, phylogenetic analysis reveals clustering of the Acidihalobacter CAs to both Acidithiobacillia and Chromatiales CAs (see Supplementary Figure 8). This makes it unclear as to whether the horizontal gene transfer occurred from the acidophiles to the Acidihalobacter genomes, and subsequent loss in A. prosperus and A. ferrooxydans, or if the HGT occurred from an early Acidihalobacter common ancestor to the acidophiles.

Voltage gated ClC-type Cl^–/H⁺ transporters found in E. coli are thought to be important in acid resistance systems. These proteins are hypothesized to function as a “electrical shunt for an outwardly directed virtual proton pump, linked to amino acid decarboxylation” (Iyer et al., 2002). Although chloride regulation is important in halophiles, no studies have indicated ClcA proteins to be involved in osmotic tolerance. Multiple predicted ClcA proteins were identified in the Acidihalobacter genomes, two in A. yilgarnensis, three in A. prosperus and A. aeolianus, four in A. ferrooxydans and one in the outgroup. Phylogenetic analysis of the ClcA proteins reveal five distinct clades of ClcA proteins which were named ClcA-1, ClcA-2, ClcA-3, ClcA-4, and ClcA-5 (see Supplementary Figure 9). clcA-1 and clcA-3 are located in the same area of the genome in A. yilgarnensis, A. prosperus, and A. aeolianus. Evidence of the loss of clcA-3 in A. yilgarnensis is evident as genomic synteny is conserved between these genomes (see Figure 6A). Both clcA-1 and clcA-3 are hypothesized to have been inherited via HGT, as phylogenetic analysis reveals clcA-1 clustering closely to ClcA proteins from Acidithiobacillaceae and clcA-3 clustering with Sneathiella spp., Symmachiella spp., Desulfobacteraceae spp., and Thiomicrospira spp. The genes surrounding clcA-1 and clcA-3 may also be involved in both acid and osmotic tolerance; pstACS (proton buffering), yggTUS (osmotic stress signaling), proH (associated with the production of proline as a compatible solute), and pilU and pilT (involved in cell adhesion, colonization, biofilm maturation and twitching) (Ito et al., 2009; Chen et al., 2015). pilG was identified to be upregulated in Acidihalobacter during increasing osmotic stress, thus these proteins may also play a role in osmotic tolerance, although no direct evidence is available of this mechanism in other bacteria (Dopson et al., 2016). This area of the genome may be important to acidic and osmotic tolerance in the Acidihalobacter, given that many of the genes in the same context are possibly involved in both osmotic and acidic tolerance. A. ferrooxydans is the only Acidihalobacter that had a second copy of the clcA-1 and is hypothesized to have been acquired via gene duplication of the original clcA-1 gene and has undergone further diversification. The second copy of clcA-1 in A. ferrooxydans is directly upstream of the hypothesized original clcA-1 gene and clusters closely with the original clcA-1 proteins (Figure 6B). clcA-2 was only identified in A. ferrooxydans and was in an area of the genome with no synteny between any other Acidihalobacter genomes (Figure 6A). Phylogenetic analysis indicates the clcA-2 protein also clustered closely with ClcA proteins in Acidithiobacillaceae, and is hypothesized to have been acquired via HGT from said Acidithiobacillaceae ClcA-4 clusters closely to ClcA proteins from other Chromatiales ClcA proteins, and is in a conserved region of the genome across all the Acidihalobacter genomes (see Supplementary Figure 10), thus we suspect that clcA-4 was acquired via vertical descent, and subsequently lost in H. neapolitanus. ClcA-5 was the only ClcA protein found in the outgroup H. neapolitanus c2 genome, however, is hypothesized to not be orthologous to any of the ClcA proteins found in the Acidihalobacter genomes as it was not phylogenetically related to any of the Acidihalobacter ClcA proteins. We hypothesize that the multiple copies of clcA in Acidihalobacter may point to its usefulness in the acidic response of Acidihalobacter. The abundance of chloride ions in Acidihalobacter environments could possibly allow for the ClcA mechanism of exporting H⁺ in exchange for Cl^– is an effective and energetically favored mechanism, due to the abundance of chloride in Acidihalobacter environment. Although this mechanism is found in other acidophiles, it may be more useful for a haloacidophile like Acidihalobacter as it is adapted to a high chloride environment.