Genomic Analysis of a Newly Isolated Acidithiobacillus ferridurans JAGS Strain Reveals Its Adaptation to Acid Mine Drainage

: Acidithiobacillus ferridurans JAGS is a newly isolated acidophile from an acid mine drainage (AMD). The genome of isolate JAGS was sequenced and compared with eight other published genomes of Acidithiobacillus . The pairwise mutation distance (Mash) and average nucleotide identity (ANI) revealed that isolate JAGS had a close evolutionary relationship with A. ferridurans JCM18981, but whole-genome alignment showed that it had higher similarity in genomic structure with A. ferrooxidans species. Pan-genome analysis revealed that nine genomes were comprised of 4601 protein coding sequences, of which 43% were core genes (1982) and 23% were unique genes (1064). A. ferridurans species had more unique genes (205–246) than A. ferrooxidans species (21–234). Functional gene categorizations showed that A. ferridurans strains had a higher portion of genes involved in energy production and conversion while A. ferrooxidans had more for inorganic ion transport and metabolism. A high abundance of kdp , mer and ars genes, as well as mobile genetic elements, was found in isolate JAGS, which might contribute to its resistance to harsh environments. These ﬁndings expand our understanding of the evolutionary adaptation of Acidithiobacillus and indicate that A. ferridurans JAGS is a promising candidate for biomining and AMD biotreatment applications.


Introduction
With continually increasing concerns about acid mine drainage (AMD) contamination and the depletion of high-grade ores, innovative and sustainable methods to recover heavy metals from tailings and AMD as well as to treat AMD pollution are urgently needed [1]. Pyrometallurgical and hydrometallurgical routes are the conventional methods for metal recovery, but they are environmentally unsustainable, with a high cost in terms of operating on low-grade ores [2,3]. Though many techniques have been applied for AMD management, such as neutralization, adsorption, oxygen barriers, bactericides and so on, most of those options are unsustainable and unaffordable [4]. Compared with conventional and other emerging reprocessing techniques, bioleaching is considered as a simple, highly efficient, safe, low-cost, more easily managed and eco-friendly technique to facilitate sustainable mining and prevent AMD [3]. Bioleaching facilitates metal mobilization from solid metal sulfides into their water-soluble forms by different microorganisms via direct and indirect bioleaching [5]. Direct bioleaching can be summarized as: MeS + 2O 2 → MeSO 4 (1) and low-pH tolerance between A. ferridurans JAGS and A. ferrooxidans ATCC23270 was conducted. To reveal the genetic traits associated with heavy metal and acid resistance in isolate JAGS, its genomic data were compared with the reported genomes of Acidithiobacillus strains. A pan-genome analysis was further conducted on these genomes to explore the metabolic features leading to the diversity of physico-biochemical traits. Functional genes and pathways responsible for heavy metal and acid resistance were analyzed and compared. A mobile genetic element analysis further suggested that gene transfers among these strains likely enabled adaptation to challenging environments. The insights gained in this study enhanced our understanding of the mechanism and evolutionary history of heavy metal and acid resistance in A. ferridurans and we suggest possible approaches for engineering A. ferridurans as a microbial chassis for biomining processes.

Culture Media, Phenotypic and Growth Observations
The strain A. ferridurans JAGS was isolated from acidic mine drainage in our previous study [22]. It was the dominant species and made up 92.6% of the enriched culture, based on 16S rRNA gene sequence analysis. The strain A. ferrooxidans ATCC23270 was purchased from American Type Culture Collection (ATCC). The phenotypic features of isolated JAGS were observed on a light microscope (Nikon Eclipse E400, Nikon, Melville, NY, USA) and a scanning electron microscope (FEI XL30 SEM, Philips, Eindhoven, Holland), separately. The strain JAGS was cultured with either 9K-Fe 2+ (160 mM ferrous iron, pH 2.0) or 9K-S 0 (0.5% elemental sulfur, pH 3.0) at 30 • C with shaking at 180 rpm or with 2:2 solid medium in an incubator at 30 • C [24].
The abilities to tolerate elevated concentrations of ferrous iron and nickel and low pH were tested in a side-by-side comparison between A. ferridurans JAGS and A. ferrooxidans ATCC23270. Cultures grown in the 9K-Fe 2+ medium were inoculated at a ratio of 20% into the same medium with Fe 2+ (200 or 320 mM), Ni 2+ (100 or 200 mM) or at pH 1.5 and then incubated at 30 • C for 22 h. The Fe 2+ concentration was tested by the colorimetric ferrozine-based assay [25] and the ferrous iron oxidation rate was calculated by using consumed Fe 2+ divided by its initial concentration, as in a previous study [26].
The growth features of A. ferridurans JAGS in 9K-Fe 2+ or 9K-S 0 medium were further monitored by detecting pH, iron or sulfate concentrations and cell numbers during incubation by removing samples at intervals. The pH value was detected using a pH meter (Thermo Scientific ® , Orion Star A211, Waltham, MA, USA). Ferrous and ferric iron concentrations were examined by the colorimetric ferrozine-based assay [25]. Sulfate was detected using a turbidimetric method [27]. Three cell-counting methods were tested: direct cell counting, optical density (OD 600 ) measurements and plate counting. For the direct cell-counting method, samples were taken from media and cell numbers were estimated using a hemocytometer (Hausser Scientific, Horsham, PA, USA). For the OD 600 method, cells were harvested, washed twice with a basal salt buffer (4.5 g/L (NH 4 ) 2 SO 4 , 0.15 g/L KCl and 0.75 g/L MgSO 4 ·7H 2 O) and resuspended in 6% betaine prior to measurements. The plate counting was carried out by spreading proper diluted samples on 2:2 solid plates and colonies were counted after 7-10 days.

Comparative Genomics
Details of A. ferridurans JAGS genomic DNA extraction, sequencing, assembly and annotation are described in our previous paper [23]. The complete genome sequence of the isolate JAGS contains 2,933,811 bp with a GC (guanine-cytosine) content of 58.6%. The Similar Genome Finder service on the PATRIC website was used with the default parameters to find the other similar Acidithiobacillus genomes published and to calculate their Mash/MinHash distances with isolate JAGS [28]. For these genomes, average nucleotide identities based on BLAST (ANIb) and MUMmer (ANIm) were calculated in JSpeciesWS [29]. Genome alignment among four whole-genome sequences was achieved using progressiveMauve within PATRIC [30]. Genes related to acid stress and metal resistance were analyzed using PATRIC and created with BioRender (https://biorender.com).

Pan-Genome Analysis
NCBI PGAP [31] was used to predict coding sequences for A. ferridurans JAGS and 8 other Acidithiobacillus genomes, and these amino acid sequences were used as the input for the Bacterial Pan-genome Analysis tool (BPGA ver. 1.2) to estimate core and pan genomes using the USEARCH program (ver. 9.0) available in BPGA, with a 50% cut-off of sequence identity [32]. The empirical power law equation f(n) = a × n α and the exponential equation f1(n) = c × e (d.n) were used for extrapolation of the pan and core genome curves, respectively. Core, accessory and unique genes defined in USEARCH were mapped into various cluster of orthologous group (COG) categories and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways. The EggNOG (ver. 5.0.0) program [33] with default parameters was further used to cluster genes into functionally related groups and to analyze metabolic pathways.

Prediction of Mobile Genetic Elements
Insertion sequences (ISs) and transposases (Tn) distributed over the 9 Acidithiobacillus genomes were predicted and classified using the ISFinder platform with manual inspection of search hits (E-value ≤ 10 −5 ) [34]. IslandViewer (ver. 4), which has integrated the three most accurate and complementary genomic islands (GIs) prediction tools, IslandPath-DIMOB, SIGI-HMM and IslandPick [35], was applied for the computational identification of putative GIs. In addition, the web tool CRISPRFinder was mainly used to identify the Clustered Regularly Interspaced Short Palindromic Repeats-Cas protein (CRISPR-Cas) array.

Phenotypic and Growth Features
The genus of Acidithiobacillus is widely distributed in natural environments such as acid mine drainage (AMD) settings. We isolated an Acidithiobacillus strain from an AMD sample collected from Sudbury, Canada, and named it A. ferridurans JAGS in our previous study [23]. Orange-brown colonies of isolate JAGS formed on solid media, taking the shape of dots after 10 days of incubation ( Figure 1A). The cells of isolate JAGS collected from 9K-Fe 2+ liquid media showed single and paired rods, approximately 0.5-1.5 µm long and 0.3 µm wide ( Figure 1B), which is slightly smaller than the reported A. ferrooxidans that is 1-2 µm long and 0.3-0.6 µm wide.  Figure S1). Figure S1A,B present a standard curve of OD600 versus the cell count obtained by plate counting. When OD600 = 1, we esti-  The species of A. ferridurans was reported to have a notably higher tolerance to many metals such as Fe 2+ , Ni 2+ and Mg 2+ and lower pH when compared with other Acidithiobacillus species [17,18]. In addition, there is 0.5-1% Ni existing in the pyrrhotite tailings of Sudbury [36]. Therefore, we compared the Fe 2+ oxidation rate (%) between A. ferridurans JAGS and A. ferrooxidans ATCC23270 under elevated concentrations of Fe 2+ and Ni 2+ and low pH pressures ( Figure 1C). The results showed that isolate JAGS had higher Fe 2+ oxidation rates compared with ATCC23270 under high concentrations of Fe 2+ and low pH values but similar levels of Fe 2+ oxidation rates under high concentrations of Ni 2+ . These results indicated the adaptation of A. ferridurans JAGS to the acid mine drainage in Sudbury and suggested that it might be a great candidate as the ferrous oxidizer in low-pH bioleaching.
The growth features of A. ferridurans JAGS in culture media with 9K-Fe 2+ and 9K-S 0 were investigated (Supplementary Material Figure S1). Figure S1A,B present a standard curve of OD 600 versus the cell count obtained by plate counting. When OD 600 = 1, we estimated that there were 8.8 × 10 9 cells/mL of A. ferridurans JAGS, which is slightly higher than the reported number of A. ferrooxidans (8.3 × 10 9 cells/mL) [37]. This may be due to the smaller cell size of isolate JAGS that causes it to absorb less light in the cuvette. We noted that counting cells of isolate JAGS was difficult due to its very low cell density in the lag phase and the interference of precipitate formation in the exponential phase. Therefore, monitoring growth required indirect tracking via changes in the pH and electron donor concentrations, but this was corroborated with data from OD 600 measurements and the plate count method. The growth behavior of isolate JAGS in 9K-Fe 2+ ( Figure S1C) and 9K-S0 culture media ( Figure S1D) shared similarities with what was reported for A. ferrooxidans ATCC 23270 [38,39], which indicates a close relationship between the two species. However, the cell numbers examined by OD 600 were much lower than those obtained by the plate count method, which is likely due to the cells lost during the process of precipitate removal prior to OD 600 measurements.

Genomic Features
To better understand the isolated strain A. ferridurans JAGS, its genome was sequenced (GenBank: CP044411) and analyzed. The genome of A. ferridurans JAGS is a single circular chromosome comprising 2,933,811 bases with a GC content of 58.56%, which contains 3001 protein-coding sequences (CDSs), 46 tRNAs and 6 rRNAs [23]. The genomic features of A. ferridurans JAGS are quite similar to those of A. ferridurans JCM18981, which are 2,921,399 bases with 58.4% GC content, containing 3026 CDSs, 47 tRNAs and 6 rRNAs.
To explore the relationship of A. ferridurans JAGS with other Acidithiobacillus species, the Similar Genome Finder service from PATRIC was used to find similar genomes with isolate JAGS as the reference. There were eight Acidithiobacillus genomes found (Table 1). These strains were collected from different environments but mainly from acid mine waters. Their genomes varied in size from 2.7 to 3.2 Mb, with total CDS numbers ranging from 2634 to 3179. For the Mash/MinHash distances, A. ferridurans JCM18981 and A. ferrooxidans IO-2C showed the closest distance with A. ferridurans JAGS compared to the other six strains, which suggests that the IO-2C strain might be incorrectly classified. Based on the average nucleotide identity (ANI) relatedness analysis, it appears that the strains JAGS, JCM18981 and IO-2C are all A. ferridurans species as they shared ANI values >98% with each other (Table 1 and Table S1), which is larger than the reported threshold of ≥96% for classification [14].
Genome alignment among the complete-genome sequences of A. ferridurans JAGS, A. ferridurans JCM18981, A. ferrooxidans ATCC53993 and A. ferrooxidans ATCC23270 was performed using progressiveMauve [24] and is shown in Figure 2. Surprisingly, the genomic arrangement of A. ferridurans JAGS had better co-linearity with A. ferrooxidans ATCC53993 and A. ferrooxidans ATCC23270 than with A. ferridurans JCM18981. The samecolor blocks suggest high conservation of gene orders among multiple genomes that are likely inherited through vertical transfer, while A. ferridurans JCM18981 had a large number of gene rearrangements, insertions and/or deletions. This result indicates multiple recombination events and a rich evolutionary history of A. ferridurans species. strains JAGS, JCM18981 and IO-2C are all A. ferridurans species as they shared ANI values >98% with each other (Table 1 and Table S1), which is larger than the reported threshold of ≥96% for classification [14]. Genome alignment among the complete-genome sequences of A. ferridurans JAGS, A. ferridurans JCM18981, A. ferrooxidans ATCC53993 and A. ferrooxidans ATCC23270 was performed using progressiveMauve [24] and is shown in Figure 2. Surprisingly, the genomic arrangement of A. ferridurans JAGS had better co-linearity with A. ferrooxidans ATCC53993 and A. ferrooxidans ATCC23270 than with A. ferridurans JCM18981. The same-color blocks suggest high conservation of gene orders among multiple genomes that are likely inherited through vertical transfer, while A. ferridurans JCM18981 had a large number of gene rearrangements, insertions and/or deletions. This result indicates multiple recombination events and a rich evolutionary history of A. ferridurans species.

Pan-Genome and Functional Gene Analysis
Pan-genome analysis was carried out using the Bacterial Pan-genome Analysis (BPGA) tool to provide insights regarding genomic features, diversity and evolution [32]. It is well accepted that more than five genomes in a pan-genome analysis could provide suf-ficient data for extrapolation of the information for species [40]. In this study, genomes of A. ferridurans JAGS and eight other strains of A. ferridurans and A. ferrooxidans were used for the pan-genome analysis since we wanted to investigate the genetic diversity and ecological adaption of these two species. As shown in Figure 3A,B, according to the Heaps' Law function (f(n) = 2819.13n 0.22 ), the pan genome is open as the γ was calculated as 0.22, which means that the addition of new genomes will provide novel genes and it indicates evolutionary changes in these genomes [41]. The pan genome contains a total of 4601 genes, of which 1982 genes are in the core genome and 1064 genes are in the unique genome. The richness of unique genes in A. ferridurans JAGS, JCM18981, IO-2C and A. ferrooxidans ATCC23270 suggests that they may actively exchange genes with other genera to adapt to different environmental conditions.

Genetic Mechanisms of Acid Stress and Metal Resistance.
In response to acidic heavy metal stress, acidophiles have developed different genetic mechanisms to survive and they are described, which can be very complex [43,44]. The metabolic diversity and adaptive mechanisms of Acidithiobacillus spp. responding to extremely acidic environments have been reviewed [45] and are beyond the scope of this study. Here, we focused on five major mechanisms ( Figure 4A) for acid and heavy metal resistance in representative Acidithiobacillus strains by analyzing related gene clusters: (1) a membrane barrier created by outer membrane proteins (Omp40) and hopanoids; (2) maintenance of a membrane potential by influx of potassium and sequestration of metal ions intra-/extracellularly; (3) active removal by antiporters or exporters; (4) decarboxylation and detoxification; and (5) DNA and protein repair systems.  Figure 3C, the numbers represent the percentage of each category. A, RNA processing and modification; B, Chromatin structure and dynamics; C, Energy production and conversion; D, Cell cycle control and mitosis; E, Amino acid metabolism and transport; F, Nucleotide metabolism and transport; G, Carbohydrate metabolism and transport; H, Coenzyme metabolism; I, Lipid metabolism; J, Translation; K, Transcription; L, Replication and repair; M, Cell wall/membrane/envelope biogenesis; N, Cell motility; O, Post-translational modification, protein turnover, chaperone functions; P, Inorganic ion transport and metabolism; Q, Secondary structure; T, Signal transduction; U, Intracellular trafficking and secretion; V, Defense mechanisms. The categories of "R, General functional prediction only" and "S, Function unknown" were omitted.
The cluster of orthologous group (COG) distributions of the annotated genes for each studied Acidithiobacillus strain are illustrated in Figure 3C. These COGs fell into 20 COG classes, not including "General functional prediction only" and "Function unknown". Escherichia coli MG1655 was used as a reference, which showed that E. coli had the highest portion of genes corresponding to (K) Transcription, while Acidithiobacillus strains showed higher portions of functional genes related to (L) Replication, recombination and repair (10.7-14%) and (M) Cell wall/membrane/envelope biogenesis (9.4-10.9%). This was not surprising given that these categories of proteins have been reported to be necessary for acid and heavy metal resistance and, likely, long-term adaptation mechanisms to extreme environments [42]. When the two Acidithiobacillus species were compared, A. ferridurans had more genes associated with functions supporting (C) Energy production and conversion (9.7-9.9%), while A. ferrooxidans had more genes related to the function of (P) Inorganic ion transport and metabolism (8.5-9.2%).

Genetic Mechanisms of Acid Stress and Metal Resistance.
In response to acidic heavy metal stress, acidophiles have developed different genetic mechanisms to survive and they are described, which can be very complex [43,44]. The metabolic diversity and adaptive mechanisms of Acidithiobacillus spp. responding to extremely acidic environments have been reviewed [45] and are beyond the scope of this study. Here, we focused on five major mechanisms ( Figure 4A) for acid and heavy metal resistance in representative Acidithiobacillus strains by analyzing related gene clusters: (1) a membrane barrier created by outer membrane proteins (Omp40) and hopanoids; (2) maintenance of a membrane potential by influx of potassium and sequestration of metal ions intra-/extracellularly; (3) active removal by antiporters or exporters; (4) decarboxylation and detoxification; and (5) DNA and protein repair systems.  Annotation of arsenic resistance clusters included arsHBRCDA, arsHRBC, arsCDA, arsRC and arsM in the studied genomes. A. ferridurans JAGS possessed all of these clusters, while the arsCDA and arsRC clusters were absent in A. ferridurans JCM18981 and the studied A. ferrooxidans strains. The largest abundance of ars clusters in isolate JAGS compared with other strains might contribute to its dominance in our metal-rich mine drainage sample. Besides, the analysis of arsC protein sequences further suggested that the ars clusters arsHBRCDA, arsHBRC and arsRC are likely acquired from different donors during evolu- Figure 4. Overview of adaptive strategies for acid and heavy metal resistance. Potential resistance mechanisms (A) and comparisons of operons for kdp (B), mer (C) and ars (D) genes among Acidithiobacillus strains. Kdp, a high-affinity K + transport system; mer, Hg 2+ -resistant genes; ars, As 2+/3+ -resistant genes. The mechanisms responsible for acid stress resistance are complex. We analyzed relevant genes in A. ferridurans JAGS and listed them in Table S2 (Supplementary Materials). OMP40 (Gene ID: F6A13_00370) was discovered and it was previously reported as an anionic porin in the outer membrane to restrict the influx of protons in A. ferrooxidans [46]. A number of genes coding for hopanoid-synthesis proteins were identified in the isolate JAGS genome, including a glycosyltransferase (HpnB, Gene ID: F6A13_01630) and a cluster for hopanoid-associated proteins, a squalene-hopene cyclase and a squalene synthase (HpnMHNKJIAG-Sch-Sqs, Gene IDs: F6A13_09105-09160). Hopanoid is an important type of bacterial lipid that can alter membrane fluidity and permeability to restrict H + influx. It was reported to be synthesized from squalene by SHC, HpnG and other proteins enriched in the outer membrane by a transporter HpnN, although the functions of some factors are still unknown [47].
The Na + /H + antiporter (Gene IDs: F6A13_09475, 04755) can export excess protons by coupling the uptake of Na + , while decarboxylases (SpeA, PanD and Psd; Gene IDs: F6A13_06090, 14595, 04545) will consume excess protons in the cytoplasm. For most of the proteins, the interspecies identities ranged from 94% to 98%, while intraspecies identities were 100%. Interestingly, we found the A. ferridurans JAGS genome to process three kdp clusters (kdpEDFABC, Gene IDs: F6A13_09590-09605; kdpFABC, Gene IDs: 11020-11005; kdpFA, GeneIDs: 11030-11025), similar to other A. ferrooxidans strains. However, A. ferridurans JCM18981 and IO-2C only had one kdp cluster ( Figure 4B). The kdp clusters code for a high-affinity K + transport system, which could generate a reversed membrane potential through the active influx of K + to cope with acid resistance [48]. In Figure S2 (Supplementary Material), a neighbor-joining (NJ) phylogenetic tree is constructed based on the kdpA protein sequences. Even though A. ferridurans JAGS had a similar pattern of kdp clusters to A. ferrooxidans species (Figure 4B), its kdpA (I) showed higher sequence identity with the other A. ferridurans strains, suggesting that kdp clusters can be acquired more than once in these genomes and have redundancy.
We also tried to find genes involved in heavy metal resistance pathways in A. ferridurans JAGS (Figure 4 and Table S2). In total, eight genes (Gene IDs: F6A13_04860, 04905, 04925, 04945, 08625, 10865, 10890, 11740) were predicted to code for p-type ATPases to transport Pb 2+ , Cd 2+ , Zn 2+ , Hg 2+ and Cu 2+ [49]. Similar numbers of these genes were detected in ATCC53993 (8) and JCM18981 (9), while only 5-6 genes were found in other strains. Some of these genes are highly similar, such as the genes F6A13_10865 and F6A13_04945, which indicates gene duplication within a genome. There are several other clusters belonging to the Resistance-nodulation-division (RND) transporter system, in the czc or znu families, which corresponds to Ni 2+ , Mn 2+ , Fe 2+/3+ , Mo 2+ , Co 2+ , Cd 2+ and Zn 2+ and CorAC for Mg 2+ /Co 2+ export. We also noticed that these Acidithiobacillus strains possessed two types and several copies of czcABC clusters located at different sites, suggesting that czc clusters may be acquired more than once from different origins. Overall, this pattern of redundancy of resistance clusters in Acidithiobacillus suggests that it is part of the adaptive strategy for survival in acidic heavy metal conditions.
We also found genes coding for proteins related to mercury (Hg 2+ ) and arsenic (As 2+/3+ ) resistance. The operons we found for Hg 2+ resistance in Acidithiobacillus strains are summarized in Figure 4C. There were two subgroups of mer clusters detected, merRTPA and merCAD. A. ferridurans JAGS and A. ferrooxidans CCM4253, Hel18, RVS1 and YQH-1 possessed both mer clusters, while A. ferridurans JCM18981 and IO-2C lacked the merCAD cluster. An intact merRTPA cluster was not detected in A. ferrooxidans ATCC23270 and ATCC53993, although ATCC53993 had two copies of merR. The merA in the merCAD cluster exhibited 100% identity in all selected strains, suggesting that merCAD might come from the same donor. Therefore, we further investigated the genetic context of the cluster merCAD and found it adjacent to kdpCBAFAF clusters. The DNA sequences of the kdpCBAFAF-merCAD clusters (11,906 bp) in A. ferridurans JAGS share 100% identity and 100% coverage with A. ferrooxidans ATCC53993, ATCC23270, RVS1 and CCM4253 and 100% identity with less coverage for A. ferrooxidans Hel18 and YQH-1, possibly due to incomplete sequencing. This indicates that the kdpCBAFAF-merCAD clusters might come from a same donor via a horizontal gene transfer (HGT) event [14]. Two predicted mobile element proteins (Gene IDs: F6A13_10980, 10985) were found upstream of the kdpCBAFAF-merCAD clusters in isolate JAGS, supporting the HGT hypothesis.
Annotation of arsenic resistance clusters included arsHBRCDA, arsHRBC, arsCDA, arsRC and arsM in the studied genomes. A. ferridurans JAGS possessed all of these clusters, while the arsCDA and arsRC clusters were absent in A. ferridurans JCM18981 and the studied A. ferrooxidans strains. The largest abundance of ars clusters in isolate JAGS compared with other strains might contribute to its dominance in our metal-rich mine drainage sample. Besides, the analysis of arsC protein sequences further suggested that the ars clusters arsHBRCDA, arsHBRC and arsRC are likely acquired from different donors during evolution. Since it was also reported that gene copy number alterations can benefit microorganisms' survival under selective pressure [50], we hypothesize that A. ferridurans JAGS might have gained ars clusters from other species during adaption to the metal-rich environments.
In summary, compared to the other Acidithiobacillus strains that we studied, A. ferridurans JAGS had several genes and considerable redundancy that likely contributes to its acid and heavy metal resistance, which highlights its strong potential for usage in biomining processes, especially for cinnabar (HgS) or arsenopyrite (FeAsS) tailings. Additionally, we speculate that A. ferridurans JAGS might be an intermediate species between A. ferrooxidans and A. ferridurans based on the evidence of the gene cluster types and genomic structure ( Figure 4) and the genome alignment result (Figure 2).

Mobile Genetic Elements Analysis
Mobile genetic elements (MGEs) play a great role in genome plasticity and evolution, shaping both genes and genomes to respond to drastic changes in environmental conditions [51]. MGEs, including insertion sequences (ISs) and genomic islands (GIs), are listed in Table 2. The number of ISs per strain ranged from 38 (A. ferridurans IO-2C) to 78 (A. ferridurans JAGS and A. ferridurans JCM18981), which might be due to the genome assembly level. High similarity regarding IS type was observed in all studied strains: IS1595, IS21, IS3, ISL3 and Tn families, which were the most common IS families. IS3 was the most abundant family. However, closer inspection demonstrated several differences. For instance, A. ferridurans JAGS has a much higher number of IS1595 when compared with other species. The nine genomes harbored 19-26 GIs ranging from 4 to 158 kb in size, representing many versatile gene pools. Several ISs, such as IS110 and IS66, were presented in the predicted GIs, suggesting that these putative GIs were likely acquired by horizontal gene transfer. In addition, GIs carrying mercury resistance genes (merRTPA) were found in all A. ferridurans species but not in the model strains A. ferrooxidans ATCC23270 and A. ferrooxidans ATCC53993, which might provide A. ferridurans with an adaptive advantage in mercury-rich environments.

Conclusions
In the present study, we provide the growth characteristics and genomic insights of a newly isolated strain, A. ferridurans JAGS. The growth features of isolate JAGS in 9K-Fe 2+ and 9K-S 0 liquid media are similar to the A. ferrooxidans type strain ATCC23270, but it shows a higher oxidation rate under elevated concentrations of Fe 2+ and low pH. Genomic comparison and pan-genome analysis among nine strains of two species of A. ferridurans and A. ferrooxidans revealed obvious genetic differences between the two species. A. ferridurans JAGS showed a closer evolutionary relationship with other A. ferridurans species but a higher similarity of genomic structure with the A. ferrooxidans strains. This suggests that A. ferridurans JAGS might be an intermediary strain. Investigations of gene clusters (kdp, mer and ars) and mobile genetic elements indicated that there have been frequent gene transfers between their genomes during evolution. The high abundance of acid and metal resistance genes found in A. ferridurans JAGS points to its unique abilities to survive in harsh mining environments, which highlights its strong potential for applications in biomining processes. Further transcriptomic and proteomic analyses are required to find the exact genes, proteins and possible mechanisms that lead to the increased resistance of the isolate JAGS strain.  Figure S2: Neighbor-joining (NJ) phylogenetic tree of the kdpA protein sequences derived from nine Acidithiobacillus strains. Bootstrap values indicated at each node are based on a total of 500 bootstrap replicates. Table S1: Average nucleotide identity (ANI) (%) based on whole-genome alignments among Acidithiobacillus strains by JSpeciesWS. Table S2: Genes predicted to be involved in acid and heavy metal tolerance in A. ferridurans JAGS.