Comparative Genomic Analysis Reveals Preserved Features in Organohalide-Respiring Sulfurospirillum Strains.

Sulfurospirillum species strains are frequently detected in various pristine and contaminated environments and participate in carbon, sulfur, nitrogen, and halogen elements cycling. Recently we obtained the complete genome sequences of two newly isolated Sulfurospirillum strains, ACSDCE and ACSTCE, capable of dechlorinating tetrachloroethene to cis-1,2-dichloroethene and trichloroethene under low-pH conditions, but a detailed analysis of these two genomes in reference to other Sulfurospirillum genomes for an improved understanding of Sulfurospirillum evolution and ecophysiology has not been accomplished. Here, we performed phylogenetic and pangenome analyses with 12 completed Sulfurospirillum genomes, including those of strain ACSTCE and strain ACSDCE, to unravel the evolutionary and metabolic potentials in the genus Sulfurospirillum. Based on 16S rRNA gene and whole-genome phylogenies, strains ACSTCE, ACSDCE, and JPD-1 could be clustered into a single species, proposed as "Candidatus Sulfurospirillum acididehalogenans." TimeTree analysis suggested that the organohalide-respiring (OHR) Sulfurospirillum might acquire the ability to use chlorinated electron acceptors later than other energy conservation processes. Nevertheless, the ambiguity of the phylogenetic relations among Sulfurospirillum strains complicated the interpretation of acquisition and loss of metabolic traits. Interestingly, all OHR Sulfurospirillum genomes except the ones of Sulfurospirillum multivorans strains harbor a well-aligned and conserved region comprising the genetic components required for the organohalide respiration chain. Pangenome results further revealed that a total of 34,620 gene products, annotated from the 12 Sulfurospirillum genomes, can be classified into 4,118 homolog families and 2,075 singleton families. Various Sulfurospirillum species strains have conserved metabolisms as well as individual enzymes and biosynthesis capabilities. For instance, only the OHR Sulfurospirillum species strains possess the quinone-dependent pyruvate dehydrogenase (PoxB) gene, and only "Ca. Sulfurospirillum acididehalogenans" strains harbor urea transporter and urease genes. The plasmids found in strain ACSTCE and strain ACSDCE feature genes coding for type II toxin-antitoxin systems and transposases and are promising tools for the development of robust gene editing tools for Sulfurospirillum. IMPORTANCE Organohalide-respiring bacteria (OHRB) play critical roles in the detoxification of chlorinated pollutants and bioremediation of subsurface environments (e.g., groundwater and sediment) impacted by anthropogenic chlorinated solvents. The majority of known OHRB cannot perform reductive dechlorination below neutral pH, hampering the applications of OHRB for remediating acidified groundwater due to fermentation and reductive dechlorination. Previously we isolated two Sulfurospirillum strains, ACSTCE and ACSDCE, capable of dechlorinating tetrachloroethene under acidic conditions (e.g., pH 5.5), and obtained the complete genomes of both strains. Notably, two plasmid sequences were identified in the genomes of strain ACSTCE and strain ACSDCE that may be conducive to unraveling the genetic modification mechanisms in the genus Sulfurospirillum. Our findings improve the current understanding of Sulfurospirillum species strains regarding their biogeographic evolution, genome dynamics, and functional diversity. This study has applied values for the bioremediation of toxic and persistent organohalide pollutants in low-pH environments.

INTRODUCTION

Members within the genus Sulfurospirillum are capable of versatile energy metabolism (e.g., nitrate reduction and organohalide respiration), enabling them to thrive in various pristine and contaminated environments (1–3). Consequently, Sulfurospirillum species strains are considered excellent candidates for biotechnological applications such as oil field souring control and bioremediation of sites impacted by chlorinated contaminants (4). For example, some Sulfurospirillum strains had demonstrated nitrate reduction to nitrite, which inhibits the growth of sulfate-reducing bacteria and consequently prevents the souring of oil reservoirs (5). A subset of Sulfurospirillum strains (e.g., S. multivorans strain DSM 12446, Sulfurospirillum sp. strain ACSTCE, and Sulfurospirillum sp. strain ACSDCE) can perform organohalide respiration (OHR) with tetrachloroethene (PCE) and trichloroethene (TCE) as electron acceptors, particularly at a pH as low as 5.5, suggesting that they play key roles for the natural attenuation of anthropogenic chlorinated solvents in acidic environments (2, 4, 6–8). In addition, some Sulfurospirillum strains are capable of synthesizing the unconventional norpseudo-type cobamide (i.e., vitamin B12 derivative), which can be used to explore the impacts of cobamides on the activities of corrinoid-dependent enzymes and microbial community structures (9). In the contaminated environments, Sulfurospirillum strains were frequently found to coexist with obligate organohalide respiring bacteria (OHRB) phylotypes such as Dehalococcoides mccartyi and Dehalobacter, which might be explained by their ability to provide the required nutrients, such as cobamides and hydrogen, for these dechlorinators (4). The hypothesis that hydrogen produced by Sulfurospirillum via anaerobic oxidation of carboxylic acids enabling syntrophic growth with a hydrogenotrophic partner was later demonstrated in the coculture of S. multivorans strain DSM 12446 and Methanococcus voltae (10). Recently, the cross-feeding of hydrogen, acetate, and cobamides was reported during the cocultivation of S. multivorans strain DSM 12446 and Dehalococcoides mccartyi strain 195 (11). Based on these observations, Sulfurospirillum strains are considered suitable partner populations to explore OHR-related microbial ecology (9–11). Great efforts have been made to investigate the environmental distribution of the genus Sulfurospirillum; however, the physiological traits, ecological relevance, and evolutionary history of Sulfurospirillum strains are not well understood.

Due to the lack of efficient genetic manipulation systems and difficulties in studying OHR Sulfurospirillum using classical biochemical techniques, bioinformatic analysis based on metagenome-assembled genomes (MAGs) and complete genomes are alternative methods to investigate the physiology and metabolic potentials of geographically distributed Sulfurospirillum (12). For instance, a genome-wide comparison of Sulfurospirillum strains indicated that the non-OHR Sulfurospirillum barnesii strain SES-3, which was isolated from anoxic sediment contaminated with arsenate and selenate, uniquely harbors a rare hybrid gene cluster encoding polyketide synthases (PKS) and nonribosomal peptide synthetases (NRPS). Based on this genome-to-metabolite approach, the first anoxically biosynthesized NRPS-PKS-derived natural product, barnesin A, was identified by omics-guided isolation and total synthesis (13). Additionally, Goris et al. conducted a thorough comparative genomic analysis of S. multivorans strain DSM 12446 with closely related non-OHR S. deleyianum strain DSM 6946 and S. barnesii strain SES-3, which revealed the presence of an ∼50-kbp region containing genes required for OHR and cobamide cofactor biosynthesis and horizontally acquired genes, enabling the catabolic flexibility in S. multivorans strain DSM 12446 (14). The genome of S. cavolei strain MES, assembled from the metagenomic sequences of an electrosynthetic microbiome, was compared with other 10 complete or draft Sulfurospirillum genomes featuring conserved and divergent physiologies and metabolisms. This pangenome analysis revealed a total of 6,264 homolog families, including 1,082 homolog families shared among all 11 Sulfurospirillum genomes (i.e., core clusters), 1,991 homolog families shared among part of 11 genomes (i.e., accessory clusters), and 3,191 singleton families (i.e., unique clusters), indicating the commonalities in general functions as well as ecological pressures induced acquisition of unique gene sets (15). Combined with transcriptomics data, a recent Sulfurospirillum comparative genomics study on the region encoding OHR genetic components identified a two-component regulator that is responsible for PCE-induced gene expression in OHR Sulfurospirillum strains (16).

To date, at least 47 complete or draft Sulfurospirillum genomes are publicly available (www.ncbi.nlm.nih.gov/genome). Two novel Sulfurospirillum organisms, strain SL2-1 and strain SL2-2, which performed PCE-to-TCE and PCE-to-cis-1,2-dichloroethene (cDCE) dechlorination, respectively, were recently enriched from a PCE-dechlorinating consortium maintained for 10 years (8). The assembled genomes of strain SL2-1 and strain SL2-2 are highly identical, representing a new Sulfurospirillum species proposed as “Candidatus Sulfurospirillum diekertiae” (8). The increased numbers of Sulfurospirillum genomes require an updated genomic examination and offer the opportunity for comprehensive comparison of OHR and non-OHR Sulfurospirillum. Such efforts are promising to provide new insights into the physiological traits, metabolic potentials, and evolution characteristics in the genus Sulfurospirillum.

In this study, we performed comparative pangenome analysis on 12 complete Sulfurospirillum genomes, including those of two new Sulfurospirillum isolates, strain ACSDCE and strain ACSTCE, capable of dechlorinating PCE to TCE and cDCE, respectively, under low-pH conditions (e.g., pH 5.5). We found that a well-aligned and conserved region comprising the genetic components required for the organohalide respiration chain is present in all OHR Sulfurospirillum genomes except the ones in S. multivorans strains. Genomic differences between non-OHR and OHR Sulfurospirillum strains as well as variations among OHR Sulfurospirillum strains were observed and discussed. Findings of this study will advance our understanding of members of the genus Sulfurospirillum regarding their evolutionary traits, genome dynamics, and functional diversity.

RESULTS AND DISCUSSION

Proposition of “Candidatus Sulfurospirillum acididehalogenans” as a new Sulfurospirillum species.

Multiple tools (e.g., JSpeciesWS, TYGS, and GTDB) and analyses (e.g., 16S rRNA genes and whole-genome sequences) were performed to reveal the phylogenetic placement of strain ACSDCE and strain ACSTCE in reference to other Sulfurospirillum species strains. Pairwise comparison of 16S rRNA gene sequences demonstrated that strains ACSDCE, ACSTCE, and JPD-1, which share 99.8 to 99.9% identities to each other, are clustered into a distinct subclade with 90.9% to 98.9% identities to other Sulfurospirillum species strains (Fig. 1A). Phylogenetic inference with complete genome sequences further demonstrated that strains ACSDCE, ACSTCE, and JPD-1 can be placed into a single cluster with “Candidatus Sulfurospirillum diekertiae” strain SL2-1 and strain SL2-2 as the closest relatives (Fig. 1B; see also Fig. S2 in the supplemental material). Pairwise comparison of genome sequences performed with TYGS found that strain ACSDCE and strain ACSTCE shared 99.3 to 99.7% dDDH (digital DNA-DNA hybridization) based on three different GBDP formulas (Table S2); by comparison, strain JPD-1 shared 69.3% and 84.7% dDDH with strain ACSDCE and 69.3% and 83.4% dDDH with strain ACSTCE (Table S2). ANIm, ANIb, and orthoANI analyses by JSpeciesWS and orthoANI demonstrated that the calculated ANI values for each pair of strains ACSTCE, ACEDCE, and JPD-1 were above the 95% threshold for species delineation (Tables S3 and S4, Fig. S3). Based on these results, we proposed to unify strains ACSDCE, ACSTCE, and JPD-1 into a new Sulfurospirillum species, designated “Candidatus Sulfurospirillum acididehalogenans.”

FIG 1

(A) 16S rRNA gene phylogenetic tree of Sulfurospirillum species strains. (B) The genome-based phylogenetic tree was inferred with FastME 2.1.6.1 from GBDP distances calculated from genome sequences. The branch lengths are scaled in terms of GBDP distance formula d5. The numbers above branches in panel B are GBDP pseudobootstrap support values of >60% from 100 replications, with an average branch support of 81.7%. Strain ACSTCE and strain ACSDCE are colored blue.

Phylogenetic tree of Sulfurospirillum species based on GTDB-tk classification and GTDB analysis. Strains JPD-1, ACSTCE, and ACSDCE were classified into a single species, designated “Sulfurospirillum sp002309535” (colored in blue), with strain JPD-1 as the type strain, while strains SL2-1 and SL2-2 were classified into another species, designated “Sulfurospirillum sp002162315” (i.e., “Sulfurospirillum diekertiae”). Download FIG S2, TIF file, 0.3 MB.

Classification and ANI values for selected Sulfurospirillum strains performed by OrthoANI software. Download FIG S3, TIF file, 0.3 MB.

OHR Sulfurospirillum species diverged from non-OHR Sulfurospirillum recently.

Genetic processes (e.g., mutation, horizontal gene transfer) and extreme geological events (e.g., great oxygenation event, neoproterozoic oxygenation event) can affect the long-term evolution of microorganisms; however, our understanding of microbial evolution is limited by the lack of geological and biological evidence (e.g., fossils, genetics) through the geological time scale. We applied the TimeTree analysis with the RelTime method to evaluate the evolution time frame of several newly sequenced Sulfurospirillum species strains by following the well-established protocols (17, 18). The TimeTree analysis suggested that the ancestor of the new species “Candidatus Sulfurospirillum acididehalogenans” emerge approximately between the Eocene and Miocene (e.g., between 49.41 and 6.33 million years ago [MYA]) (Fig. 2), which was much later than the estimated appearance time (e.g., the Neoproterozoic era, 1,000 to 541 MYA) of obligate organohalide respiring Dehalococcodia (18). The TimeTree analysis is based on 16S rRNA gene sequences and intended to infer the divergence times of strains within the genus Sulfurospirillum. Such an approach can only provide a rough estimate; however, this effort reflected our interests in unresolved questions, including (i) how does the organohalide respiration metabolism evolve, (ii) when does this energy conservation emerge in geoscale time, and (iii) when were the organohalide-respiratory genes transferred into microorganisms of different genera (e.g., Sulfurospirillum). To the best of our knowledge, no solid evidence is available to predict or support when an insertion event (e.g., the organohalide respiration region) in the Sulfurospirillum genome occurs, since dating the hypothetical insertion event is still difficult. Nevertheless, the dechlorinating Sulfurospirillum species strains could not obtain organohalide respiration genes before their own existence; therefore, the insertion event or organohalide respiration genes horizontally transferred occurred roughly at the same time or later than the divergence time of Sulfurospirillum species strains.

FIG 2

Phylogeny and molecular clock of selected 16 rRNA genes from Sulfurospirillum species strains and Campylobacter lari from Proteobacteria. Acidobacterium capsulatum ATCC 51196 was chosen as the outgroup. Branch lengths represent the divergence times (MYA) approximated by the RelTime method using Mega X. The OHR and non-OHR clusters of Sulfurospirillum species strains were colored in light orange and gray, respectively.

Furthermore, a previous genomic comparison study based on the genome sequence of Sulfurospirillum multivorans discovered a 50-kbp organohalide-respiring region (14), which was also identified in the genome of another species, Sulfurospirillum halorespirans (6), indicating the ability to respire halogenated organic compounds was horizontally acquired by some ancestors of the genus Sulfurospirillum recently. Ancestors of the OHR Sulfurospirillum strains may diverge from the non-OHR Sulfurospirillum strains via horizontal acquisition of the genetic components required for organohalide respiration chain and de novo cobamide biosynthesis, and OHR Sulfurospirillum strains were subsequently distributed to different niches on the planet Earth, which could be inferred by the observation that different OHR Sulfurospirillum strains with conserved arrangements of the gene cluster responsible for OHR were isolated from geographically distinct origins. Nonetheless, the non-OHR Sulfurospirillum sp. strain UCH001, which was isolated from chlorinated ethene-contaminated groundwater in Japan but could not dechlorinate chlorinated ethenes (19), was closely related to the OHR Sulfurospirillum strains. Such an inconsistency may be due to the loss of gene clusters responsible for organohalide respiration and requires further investigation.

Shared features in the chromosomes of strain ACSDCE and strain ACSTCE.

The chromosome of strain ACSTCE has 2,998 features, including 2,853 coding genes, 3 CRISPR (clustered regularly interspaced short palindromic repeats) arrays, 26 CRISPR repeats, 23 CRISPR spacers, and 50 noncoding RNAs. Only 1,338 coding sequences could be assigned with a SEED (https://pubseed.theseed.org) annotation ontology across 910 distinct SEED functions. By comparison, the chromosome of strain ACSDCE harbors 2,993 features, including 2,852 coding genes, 3 CRISPR arrays, 26 CRISPR repeats, 23 CRISPR spacers, and 55 noncoding RNAs. Only a total of 1,351 coding sequences could be assigned with a SEED annotation ontology across 911 distinct SEED functions. Functioning-based metabolic reconstruction comparison between strain ACSDCE and strain ACSTCE suggested that both strains share a total of 956 functioning roles defined within the SEED subsystems (e.g., biotin biosynthesis, cobamide synthesis, coenzyme A biosynthesis, and heme and siroheme biosynthesis) (Table S5). BLAST of the annotated coding sequences showed that the homologs of 70 genes in the chromosome of strain ACSDCE were not found in the chromosome of strain ACSTCE, including those genes encoding hydroxylamine reductase, arginyl-tRNA protein transferase, DNA helicase restriction/modification system component YeeBC, and 50 hypothetical genes with unknown functions (Table S6). Likewise, 23 coding genes in the chromosome of strain ACSTCE were not matched in the chromosome of strain ACSDCE, such as genes encoding cytochrome c552 precursor, anaerobic C4-dicarboxylate transporter DcuA, heterodisulfide reductase subunit B-like protein/putative succinate dehydrogenase subunit, and other 10 hypothetical proteins (Table S7). Genes encoding tetrathionate reductase, nitrate reductase, formate dehydrogenase, and [NiFe] hydrogenase were present in both chromosomes of strains ACSTCE and strain ACSDCE. However, no genes related to nitrite ammonification were identified, suggesting that they can reduce nitrate as an electron acceptor and produce nitrite only. In pure cultures, strain ACSTCE and strain ACSDCE could transform up to 60% of an initial 5 mM nitrate to nitrite (Fig. S4 and S5). Whether the accumulation of nitrite from nitrate transformation inhibits the activities of Sulfurospirillum and whether strains ACSDCE and ACSTCE can transform nitrite to ammonia require further investigation.

Nitrate reduction by Sulfurospirillum strain ACSTCE with different starting nitrate concentrations (i.e., 1, 5, 10, and 20 mM). Download FIG S4, TIF file, 0.2 MB.

Nitrate reduction by Sulfurospirillum strain ACSDCE with different starting nitrate concentrations (i.e., 1, 5, 10, and 20 mM). Download FIG S5, TIF file, 0.2 MB.

Potential roles of plasmids in strain ACSDCE and strain ACSTCE.

Except for strain ACSTCE and strain ACSDCE, plasmids are not found in Sulfurospirillum species strains, presumably due to the absence of plasmid or difficulty in faithful assembly and identification of plasmid solely using Illumina short reads (20). In contrast, circular plasmids with sizes of 38,046 bp and 39,868 bp were assembled from the genome sequencing data of strain ACSTCE and strain ACSDCE, respectively, using PacBio long reads and Illumina short reads (21, 22). Plasmid sequence alignment by MAUVE demonstrated that the synteny of the two plasmid sequences was not conserved, and only two regions with a total size of 9.5 kbp could be mapped between the two plasmid sequences. Plasmids of strain ACSTCE and strain ACSDCE carried 94 and 57 coding sequences, of which only 19 and 20 encoded hypothetic proteins, respectively. By comparison, the eggNOG-mapper pipeline found that 37 coding sequences of strain ACSTCE plasmid and 32 of strain ACSDCE plasmid (Table S8 and S9) matched with orthologous groups in the eggNOG database. While most of the orthologous genes related to the coding sequences in the plasmids of strain ACSTCE and strain ACSDCE were found in Proteobacteria, a few coding sequences (e.g., parA and mcrB) present in these two plasmids were closely related to those found in Clostridia of Firmicutes and Bacteroidetes (Tables S8 and S9). Therefore, we hypothesized that these two plasmids originate from Proteobacteria with additional coding sequences horizontally acquired from other phyla (e.g., Firmicutes). Two prophage-related regions with lengths of 20,367 bp and 26,549 bp were found in the plasmids of strain ACSTCE and strain ACSDCE, respectively. In addition, the plasmid and chromosomal sequences of strain ACSTCE shared an identical repeat region with an approximate size of 1 kbp and was in strain ACSDCE, indicating the exchange of genetic materials between the plasmid and the chromosome in both strains. Sequences encoding components (e.g., a toxic protein and its cognate antitoxin protein) of various type II toxin-antitoxin (TA) systems (e.g., HicA-HicB, YafQ-DinJ, and RelE-RelB) are both present in the plasmids of the strain ACSTCE and strain ACSDCE. Type II TA systems have been proposed to play roles in genome stabilization, abortive phage infection, stress modulation, and antibiotic persistence (23); however, how such a system is related to the survival of strain ACSTCE and strain ACSDCE is unknown. Four and three genes coding for transposases were annotated from the plasmids of strain ACSTCE and strain ACSDCE, respectively. A phylogenetic analysis on transposases annotated from Sulfurospirillum genomes indicated that five transposases from the plasmid sequences were clustered within the IS30/IS982 family. One IS21 family transposase on the plasmid of strain ACSDCE was clustered with the transposases annotated on the chromosomes of strain ACSDCE and strain ACSTCE (Fig. S6). Restriction sites for restriction endonucleases such as BamHI, BglII, EcoRI, PvuI, and SalI were identified in both plasmids, providing retrospective in silico evidence of the movement of genomic sequences between the plasmid and chromosome in strain ACSTCE and strain ACSDCE. Overall, these results improve our understanding of the genomic characteristics of Sulfurospirillum species strains and are promising for the development of molecular tools for editing Sulfurospirillum genomes.

Phylogenetic analysis of transposase sequences encoded on the plasmids of strain ACSTCE and strain ACSDCE (colored in blue) and transposases from the chromosomes of select Sulfurospirillum species strains. Transposases were verified using ISFinder server and database (https://www-is.biotoul.fr/about.php) with two representative families labeled with red and green colors. The scale bar represents substitution per site. Download FIG S6, TIF file, 0.5 MB.

Pangenome analysis revealed conserved and differed features in OHR Sulfurospirillum.

A total of 12 complete Sulfurospirillum genomes were selected for genomic comparison analysis to unravel the core functions and core protein families using the OrthoMCL tool, including seven experimentally verified OHR Sulfurospirillum genomes (i.e., strains ACSDCE, ACSTCE, JPD-1, SL2-1, SL2-2, S. multivorans DSM 12446, and S. halorespirans DSM 13726) and four non-OHR Sulfurospirillum genomes (i.e., strains UCH001, UCH003, S. deleyianum DSM 6946, and S. barnesii SES-3). S. multivorans strain N was demonstrated to be incapable of dechlorinating chlorinated ethenes despite its genome harbors two homologous reductive dehalogenase genes. For pangenome analysis, we grouped strain N with other OHR Sulfurospirillum species strains, since strain N may have lost its OHR capability recently by a transposition event (16). Generally, all examined Sulfurospirillum species strains shared a variety of conserved sequence regions (Fig. 3). Furthermore, whole-sequence alignment of all OHR Sulfurospirillum species strains by Mauve demonstrated that all of them except S. multivorans strain DSM 12446 and S. multivorans strain N have a larger well-aligned and conserved regions (i.e., 330 kbp) containing a previously proposed 50-kb OHR region (Fig. S7). A 148-kbp block that was not found in other OHR Sulfurospirillum genomes was inserted into the end of the OHR region only in the genomes of S. multivorans strain DSM 12446 and S. multivorans strain N. The OHR regions of eight OHR strains were estimated to be between 63 kbp (e.g., S. halorespirans) and 74 kbp (e.g., S. multivorans). A total of 34,620 coding sequences (i.e., 32,545 homolog gene sequences and 2,075 singleton gene sequences), annotated from 12 Sulfurospirillum genomes, was classified into 6,193 families (i.e., 4,118 homolog families and 2,075 singleton families). The average numbers of gene sequences, genes in homologs, genes in singletons, and homolog families for the total of 12 Sulfurospirillum genomes were 2,885, 2,712, 173, and 2,625, respectively. By comparison, the average numbers of gene sequences, genes in homologs, and homolog families, but not genes in singletons, found in the eight OHR Sulfurospirillum genomes were larger than those of the four non-OHR Sulfurospirillum genomes, probably because all OHR Sulfurospirillum species strains have larger genome sizes than those of non-OHR Sulfurospirillum species strains by horizontally acquiring the genomic regions responsible for OHR (Table S1). The shared homolog families for each pair of Sulfurospirillum genomes were summarized in Table 1.

FIG 3

Circle view of pangenome with S. multivorans strain DSM 12446 as the base genome. The order of the genes in the rest of the pangenome is aligned to the position of its ortholog in the base genome. Genomes 1 to 11 are the following: 1, strain N; 2, S. halorespirans; 3, strain SL2-1; 4, strain SL2-2; 5, strain JPD-1; 6, strain ACSDCE; 7, strain ACSTCE; 8, strain UCH003; 9, strain UCH001; 10, strain SES-3; 11, S. deleyianum. The OHR colored in orange stands for the organohalide respiration region.

TABLE 1

Pairwise comparisons of 12 Sulfurospirillum genomes for the shared homolog families

Homolog family	G1	G2	G3	G4	G5	G6	G7	G8	G9	G10	G11	G12
G1, S. deleyianum	2,085	1,847	1,849	1,747	1,776	1,781	1,794	1,794	1,773	1,746	1,854	1,845
G2, strain N	1,847	3,146	2,050	2,089	2,133	2,248	2,309	2,308	2,227	2,187	2,534	3,140
G3, strain SES-3	1,849	2,050	2,255	1,812	1,847	1,862	1,906	1,903	1,827	1,813	2,001	2,050
G4, strain UCH001	1,747	2,089	1,812	2,307	2,030	2,062	2,104	2,102	2,070	2,052	2,113	2,088
G5, strain UCH003	1,776	2,133	1,847	2,030	2,389	2,096	2,109	2,107	2,074	2,057	2,154	2,133
G6, strain JPD-1	1,781	2,248	1,862	2,062	2,096	2,632	2,485	2,484	2,427	2,376	2,303	2,243
G7, strain SL2-1	1,794	2,309	1,906	2,104	2,109	2,485	2,797	2,779	2,394	2,351	2,347	2,306
G8, strain SL2-2	1,794	2,308	1,903	2,102	2,107	2,484	2,779	2,791	2,391	2,351	2,343	2,305
G9, strain ACSDCE	1,773	2,227	1,827	2,070	2,074	2,427	2,394	2,391	2,628	2,565	2,253	2,224
G10, strain ACSTCE	1,746	2,187	1,813	2,052	2,057	2,376	2,351	2,351	2,565	2,578	2,214	2,187
G11, S. halorespirans	1,854	2,534	2,001	2,113	2,154	2,303	2,347	2,343	2,253	2,214	2,750	2,531
G12, S. multivorans	1,845	3,140	2,050	2,088	2,133	2,243	2,306	2,305	2,224	2,187	2,531	3,141

Complete alignment of genome sequences of OHR Sulfurospirillum species strains as computed by Mauve. Download FIG S7, TIF file, 0.9 MB.

A total of 63 homolog families was only found in the genomes of OHR Sulfurospirillum species strains, including homologous genes encoding reductive dehalogenase, quinone-dependent pyruvate dehydrogenase (i.e., PoxB), NADPH-dependent FMN reductase, short-chain dehydrogenases/reductase, transcriptional regulator AraC family, Ser-tRNA(Ala) deacylase/Gly-tRNA(Ala) deacylase, phosphinothricin N-acetyltransferase, proteins associated with cobamide transport and biosynthesis (e.g., vitamin B12-ABC transporter permease component BtuC, B12-binding component BtuF, uroporphyrinogen-III methyltransferase/uroporphyrinogen-III synthase, and cobalt-precorrin-6A reductase), and a cluster related to propanoate metabolism (e.g., acetyl-coenzyme A synthetase, methylisocitrate lyase, 2-methylcitrate synthase, and 2-methylcitrate dehydratase). The previously reported 50-kbp gene region (e.g., 54 to 61 coding sequences starting from the gene representing carboxymuconolactone decarboxylase family protein) containing reductive dehalogenase genes and the gene cluster coding for (nor)cobamide biosynthesis from uroporphyrinogen III (14, 24) were conserved in all eight OHR Sulfurospirillum genomes (Fig. 4) but not present in the non-OHR Sulfurospirillum genomes. The only exception is that S. barnesii strain SES-3 of the non-OHR Sulfurospirillum group has several genes encoding de novo cobamide biosynthesis (Fig. 3) (14). All eight OHR Sulfurospirillum species strains and the non-OHR S. cavolei strain UCH003 possess the complete gene set for nitrogen fixation. All examined Sulfurospirillum species strains except S. barnesii strain SES-3 and S. deleyianum strain DSM 6946 possess the complete gene set for tetrathionate reductase.

FIG 4

Illustration of the preserved OHR-related gene regions in the OHR Sulfurospirillum species strains. These regions encode carboxymuconolactone decarboxylase family protein (light blue), followed by the reductive dehalogenase with associated putative membrane anchor protein (green), iron-sulfur cluster assembly scaffold protein (light gray), sensor histidine kinase and response regulator transcription factor (purple), DUF4405 domain-containing protein (dark blue), components of a putative quinol dehydrogenase (orange), norcobamde biosynthesis clusters (light green), and transposase (red). The OHR region of strain ACSTCE possesses two additional reductive dehalogenase gene clusters presented in the dashed box.

The eight OHR Sulfurospirillum genomes contain a total of 4,113 gene clusters (Fig. 5). Genomes of “Candidatus Sulfurospirillum acididehalogenans” strains ACSTCE, ACSDCE, and JPD-1 shared 59 clusters consisting of 179 coding sequences (Fig. 5). Notably, “Candidatus Sulfurospirillum acididehalogenans” strains ACSDCE, ACSTCE, and JPD-1 have a 14-genes cluster encoding urea transporter UrtABCDEFG, urease, two-component sensor histidine kinase, and a hybrid sensor histidine kinase/response regulator. The presence of urease is hypothesized to allow strains ACSDCE, ACSTCE, and JPD-1 to tolerate low-pH conditions by yielding ammonia to neutralize protons (25); however, S. multivorans also can grow and dechlorinate PCE under acidic conditions (i.e., pH 5.5) and does not possess the urease gene cluster. “Candidatus Sulfurospirillum acididehalogenans” and S. multivorans probably could possess additional acid tolerance mechanisms, such as involvement of F0F1-ATPase in pH homeostasis, amino acid-dependent decarboxylase/antiporter systems, and deiminase and deaminase systems. Genes encoding F0F1-ATP synthase and various decarboxylase (e.g., arginine decarboxylase, aspartate decarboxylase) and agmatine deiminase family proteins were present in the genomes of “Candidatus Sulfurospirillum acididehalogenans” and S. multivorans. Comparatively, a total of 246 coding sequences from all OHR Sulfurospirillum species strains except “Candidatus Sulfurospirillum acididehalogenans” strains ACSDCE, ACSTCE, and JPD-1 were classified into 49 clusters, including genes for the l-proline glycine betaine ABC transport system (i.e., ProVWX), arsenite oxidase, 2-oxoglutarate/malate translocator, ferric iron ABC transporter, respiratory arsenate reductase Mo binding, and FeS subunits ArrAB. Strain ACSTCE and strain ACSDCE shared 152 orthologous gene clusters containing 310 coding sequences, most of which code for CRISPR-associated proteins, transposases, and mobile element proteins. Two adjacent genes annotated as carbon monoxide dehydrogenases, CooS, and carbon monoxide dehydrogenase accessory protein, CooC, were present in the genomes of strain ACSDCE and strain ACSTCE but not in other Sulfurospirillum genomes. By comparison, all the OHR Sulfurospirillum species strains, except strain ACSDCE and strain ACSTCE, harbor genes encoding polysulfide reductase NrfD, potassium-transporting ATPase, nitrous oxide reductase NosZ, and nitrous oxide reductase maturation protein NosL.

FIG 5

Orthologous gene clusters across eight OHR Sulfurospirillum strains. The filled and gray blocks indicate the presence and absence of orthologous gene clusters in each genome. SH, SM, and N stand for S. halorespirans, S. multivorans, and strain N, respectively.

Differences between PCE-to-cDCE and PCE-to-TCE dechlorinating reductive dehalogenases.

All examined OHR Sulfurospirillum genomes possess two copies of reductive dehalogenase homologous (RdhA) genes, except that four RdhA copies were found in the genome of strain ACSTCE (Fig. 4). Each of these RdhA genes is adjacent to a gene with a size of 150 to 225 bp encoding reductive dehalogenase membrane anchor protein RdhB. Phylogenetic analysis indicated that all 18 RdhAs encoded by the genomes of known Sulfurospirillum species strains can be grouped into two distinct clusters, with several RdhAs belonging to the genera Desulfovibrio, Desulfomonile, and Desulforhopalus as the closet relatives (Fig. 6). Seven RdhAs in cluster I (colored in blue in Fig. 6) were 100% identical to each other, but the RdhA from strain ACSTCE shares 99.8% amino acid similarity with the other seven RdhAs in cluster I. The substrate spectrum of the eight putative RdhAs in cluster I remains elusive. By comparison, the three identical RdhAs (ACSTCE_3, ACSTCE_17, and ACSTCE_2843) of strain ACSTCE and the other five putative RdhAs (Shal_1516, JPD-1_1501, SL2-2_1608, ACSDCE_8 and Smul_N_1565) in cluster II (colored red in Fig. 6) were grouped with the characterized PceA (Smul_1557) of S. multivorans strain DSM 12446, indicating that they can dechlorinate PCE. The putative RdhAs of Shal_1516, JPD-1_1501, SL2-2_1608, ACSDCE_8, and Smul_N_1565 shared 92.1%, 94.0%, 96.2%, 97.2%, and 99.8% similarities, respectively, with PceA of S. multivorans. The three identical RdhAs of strain ACSTCE (colored red in Fig. 6) share 97.8% similarity with the RdhA of strain SL2-1 (SL2-1_1591), and these four RdhAs were predicted to be responsible for dechlorinating PCE to TCE based on the fact that strain ACETCE and strain SL2-1 could only dechlorinate PCE to TCE. Only four critical residue differences (Ser279 versus Ala279, Gly286 versus Cys286, Ser312 versus Cys312, and Pro320 versus Ala320) were identified among the four putative PCE-to-TCE dechlorinating RdhAs (SL2-1_1591, ACSTCE_3, ACSTCE_17, and ACSTCE_2843) and the other six putative PCE-to-cDCE dechlorinating RdhAs (Shal_1516, JPD-1_1501, SL2-2_1608, ACSDCE_8, Smul_1557, and N_1565). The architecture of the active site of PCE-to-TCE and PCE-to-cDCE dechlorinating RdhAs of Sulfurospirillum appears to be similar (26). Based on the structure of PceA in S. multivorans (27), one has to assume that the residues identified here are not directly involved in the formation of the enzyme’s active-site cavity. However, these residues are located in close vicinity to amino acids lining the substrate binding site and might influence the substrate specificity of the enzyme indirectly. Whether the residue substitution changed the substrate specificity could not be confirmed yet. Therefore, the development of efficient tools to edit the genetic contents of Sulfurospirillum will assist us in understanding these versatile microorganisms regarding organohalide respiration.

FIG 6

Phylogenetic tree of RdhAs from OHR Sulfurospirillum and select OHRB. The putative RdhAs were retrieved from the OGs database (29). Clusters I and II of Sulfurospirillum RdhAs were colored in blue and red, respectively.

Despite physiological and evolutionary differences among the eight OHR Sulfurospirillum species strains, the second copies of RdhAs (e.g., cluster I) are highly conserved (i.e., >99% amino acid sequence identity) in all eight OHR Sulfurospirillum strains (14, 28). Generally, reductive dehalogenases are known to be phylogenetically diverse (29), and we did not expect to observe that the second copies of RdhAs with unknown functions were highly conserved in OHR Sulfurospirillum species strains distributed in geographically distinct locations. Despite the detection of the second copy of RdhA in the transcriptomic and proteomic studies on S. multivorans grown on PCE (14, 24), the function of the second copy of RdhA remains unclear. Hypothetically, the two genes encoding RdhAs (i.e., pceA and the second copy of the rdhA gene) were acquired horizontally before the speciation of the genus Sulfurospirillum, and more evolutionary pressures could be forced on the functional pceA gene than the second copy of the rdhA gene.

Genome-inferred metabolic capacities and electron transport mechanism in Sulfurospirillum.

In addition to OHR, Sulfurospirillum species strains can conserve energy using a variety of electron acceptors (e.g., fumarate, dimethyl sulfoxide [DMSO], thiosulfate, arsenate, selenate, and nitrate) and electron donors (e.g., hydrogen, formate, hydrogen sulfide, lactate, and pyruvate) (4). For example, S. halorespirans strain PCE-M2 possesses a gene cluster encoding the SoxCDXYZAB proteins for thiosulfate oxidation. By comparison, other Sulfurospirillum species strains, including strain ACSDCE and strain ACSTCE, have only two copies of the phsA gene encoding thiosulfate reductase. Sulfurospirillum species strains have been implicated in arsenate reduction with the potential of mobilizing arsenic in underground aquifers. While a single copy of the arsenate reductase gene arsC is present in the genomes of strains ACSDCE, ACSTCE, and JPD-1, other OHR Sulfurospirillum species strains harbor multiple copies of the arsC gene as well as arsenite oxidase genes aioAB (14). The ability to oxidize arsenite to arsenate at the presence of azurin protein as an electron acceptor was not observed in pure culture study, demonstrating that aioAB genes may not be functioning.

The electron transport chain of OHR can be generally classified into two categories: quinone dependent, represented by Sulfurospirillum, and quinone independent, represented by Dehalococcoides (12, 26, 30). OHR Sulfurospirillum species strains are quinone dependent and likely express the NapGH-like quinol dehydrogenases to transfer electrons from the menaquinone pool to PceA, similar to the function of NapGH quinol dehydrogenase in nitrate reduction (12, 24, 26). One of the possible electron sources for replenishing the electron pool is through oxidizing pyruvate via two different enzymes: pyruvate:ferredoxin/flavodoxin oxidoreductase (PFOR) and ubiquinone-dependent pyruvate dehydrogenase (PoxB) (14). The PFOR protein catalyzes the conversion of pyruvate to acetyl-CoA and carbon dioxide with simultaneous transfer of two electrons to ferredoxin or flavodoxin, while the PoxB enzyme could convert pyruvate into acetate and carbon dioxide accompanied by transferring the generated electrons directly to the menaquinone pool (31). The gene encoding PFOR has been found in all 12 Sulfurospirillum genomes; by comparison, the poxB genes are only present in the eight OHR Sulfurospirillum genomes, suggesting that poxB is probably related to the electron transfer chain of OHR. The inferred poxB products of OHR Sulfurospirillum species strains share 86.9% to 100% amino acid similarities to each other and are closely related to the ones identified in Malaciobacter marinus and gammaproteobacterial microorganisms (Fig. S8). The coexistence of both genes in a bacterial genome is uncommon, and the advantages for OHR Sulfurospirillum species strains possessing the pfoR and poxB genes simultaneously are not clear.

Phylogenetic analysis of PoxB from selected microorganisms. The PoxB homologs from Sulfurospirillum species strains were colored in blue, and the PoxB homologs from Gammaproteobacteria were colored in olive. The scale bar represents substitution per site. Download FIG S8, TIF file, 0.9 MB.

Summary.

In this study, we performed phylogenetic, pangenomic, and evolutionary analyses on 12 complete Sulfurospirillum genomes, including two newly sequenced ones of strain ACSDCE and strain ACSTCE and proposed a new species, “Candidatus Sulfurospirillum acididehalogenans,” represented by strains ACSTCE, ACSDCE, and JPD-1. The relatively preserved region identified in all examined OHR Sulfurospirillum species strains, but not S. multivorans strains, suggested that they have a common ancestor that acquired the OHR capability recently and an additional insertion event(s) occurred in S. multivorans strains. Comparison between the PCE-to-TCE RdhAs and PCE-to-cDCE RdhAs identified the differences of four amino acid residues, but how these residue differences affect the substrate specificity of reductive dehalogenases remains unclear. The versatile metabolisms in Sulfurospirillum species strains ensure their great potentials in biotechnological applications, including the cleanup of soil and groundwater contaminated with a combination of chemicals (e.g., mixtures of arsenate, chlorinated ethenes, nitrate, and selenate). The ability to grow under unfavorable conditions (e.g., low pH) further emphasizes their roles and functions in specialized environmental settings. Nevertheless, some predictions on the metabolisms and capabilities of Sulfurospirillum were solely based on in silico genomic analysis, and experimental evidence is warranted in future studies. The plasmids harbored by strain ACSTCE and strain ACSDCE represent a promising tool for developing a robust gene editing tool that can advance the understanding of physiology and metabolism in the genus Sulfurospirillum.

MATERIALS AND METHODS

Sequence data set and phylogenetic analysis.

As of January 2022, 47 Sulfurospirillum assemblies are publicly available (https://www.ncbi.nlm.nih.gov/assembly/?term=Sulfurospirillum). CheckM (version 1.0.18) (32) assessment showed that the average completeness and contamination of 35 assemblies were 85.5% and 0.9%, respectively. Since the incompleteness and contamination of metagenome-assembled draft genomes (33) may affect the analysis of core, auxiliary, and singleton families, we only focused on the completed Sulfurospirillum genomes in this study. Among them, 12 completed Sulfurospirillum genomes, including the ones of strain ACSTCE and strain ACSDCE, were selected for the following analyses (see Table S1 in the supplemental material and supplemental tables at https://doi.org/10.6084/m9.figshare.17014352.v1). Genome sequences of strain ACSTCE and strain ACSDCE were recently deposited and are publicly available in the NCBI genome database (21, 22). Phylogenetic analysis of Sulfurospirillum 16S rRNA gene sequences, which were retrieved from the Ribosomal Database Project (RDP) database, release 11 update 5 (34), was performed using PhyML with the general time-reversible (GTR) substitution model (35). Whole-genome-based taxonomy was analyzed using the Type (Strain) Genome Server (TYGS) (36) and Genome Taxonomy Database (GTDB), as described previously (37, 38). Briefly, RNAmmer v1.2 was applied to extract Sulfurospirillum 16S rRNA gene sequences (39), which were compared with those of available 10,997 type strains in the TYGS database for finding additional closely related type strains. Genomes were selected for pairwise comparisons using GBDP (Genome BLAST Distance Phylogeny) and accurate intergenomic distances (40), and the results were used to construct phylogenetic trees with a balanced minimum evolution via FASTME 2.1.4 (41). Species boundary was defined as 70% DNA-DNA hybridization (DDH). Average nucleotide identity (ANI) (e.g., ANIm [42], ANIb [43]) was calculated using JSpeciesWS (44) and orthoANI (45) to evaluate if two or more genomes can be classified into the same species. The ANI threshold for species boundary is defined as 95%.

Molecular clock analysis.

MEGA X (46) for molecular evolutionary genetics analysis was applied to predict the divergence times using the RelTime method (47, 48) and the Tamura-Nei model (49). The 16S rRNA gene sequences were aligned with MUSCLE using the unweighted pair group method using average linkages algorithm and then analyzed for phylogeny reconstruction with the minimum evolution method or neighbor-joining method using Mega X. The nucleotide sequence alignment and phylogenetic tree were used as the input for RelTime-ML. The divergence times shown in Fig. S1 for several Sulfurospirillum species were predicted by TimeTree (50, 51) and were applied to set the divergence time calibration constraints by following the published approach (18, 52).

Phylogenetic tree and estimated divergent time of selected Sulfurospirillum species strains by TimeTree with all default settings (www.timetree.org). The predicted divergent times for labeled nodes were retrieved from the TimeTree database. Download FIG S1, TIF file, 0.7 MB.

Whole-genome comparison and pangenome analysis.

Sulfurospirillum genomes were reannotated using the RAST (Rapid Annotation using Subsystem Technology) tool with default parameters (53, 54) and eggNOG-mapper (55–57) to ensure annotation conformity with formats and consistency across all genomes. Annotated coding sequences were verified by BLAST search (58) against NCBI nonredundant protein sequences and by UniProt database search (e.g., UniProtKB reference proteomes plus Swiss-Prot) (59). All 12 completed Sulfurospirillum genomes were used for pangenome analysis, which was constructed and performed by OrthoMCL (60) with default parameters using the KBase platform (61). The protein sequences of the key functions were retrieved and analyzed using KEGG (https://www.kegg.jp) (62, 63) and BRENDA (www.brenda-enzymes.org) (64) to identify the key pathways and modules among various Sulfurospirillum species strains.

Analysis of functional sequences.

Protein sequences for building the phylogenetic inference tree were retrieved from the UniProt database (www.uniprot.com) (59). All protein sequences for building the phylogenetic trees of reductive dehalogenases are listed in the Text S1 (see also Data Set 1 at https://doi.org/10.6084/m9.figshare.17014145.v1) (65). The phylogenetic trees for reductive dehalogenases and transposases were built with Geneious software version 11.1.5 using the MUSCLE and FastTree or PhyML with default settings (Biomatters Inc., Newark, NJ, USA).

Materials and methods of cultivation, metabolite detection, analytical methods, and data availability. Download Text S1, DOCX file, 0.02 MB.