ORIGINAL RESEARCH published: 10 January 2019 doi: 10.3389/fmicb.2018.03332 Origin of the Mobile Di-Hydro-Pteroate Synthase Gene Determining Sulfonamide Resistance in Clinical Isolates Miquel Sánchez-Osuna 1 , Pilar Cortés 1 , Jordi Barbé 1* and Ivan Erill 2* 1 Departament de Genètica i de Microbiologia, Universitat Autònoma de Barcelona, Bellaterra, Spain, 2 Department of Biological Sciences, University of Maryland, Baltimore County, Baltimore, MD, United States Sulfonamides are synthetic chemotherapeutic agents that work as competitive inhibitors of the di-hydro-pteroate synthase (DHPS) enzyme, encoded by the folP gene. Edited by: Resistance to sulfonamides is widespread in the clinical setting and predominantly Laurent Poirel, mediated by plasmid- and integron-borne sul1-3 genes encoding mutant DHPS Université de Fribourg, Switzerland enzymes that do not bind sulfonamides. In spite of their clinical importance, the Reviewed by: genetic origin of sul1-3 genes remains unknown. Here we analyze sul genes and their Jose L. Martinez, Spanish National Research genetic neighborhoods to uncover sul signature elements that enable the elucidation Council (CSIC), Spain of their genetic origin. We identify a protein sequence Sul motif associated with sul- Rune Henning Sørum, Norwegian University of Life Sciences, encoded proteins, as well as consistent association of a phosphoglucosamine mutase Norway gene (glmM) with the sul2 gene. We identify chromosomal folP genes bearing these *Correspondence: genetic markers in two bacterial families: the Rhodobiaceae and the Leptospiraceae. Jordi Barbé Bayesian phylogenetic inference of FolP/Sul and GlmM protein sequences clearly jordi.barbe@uab.cat Ivan Erill establishes that sul1-2 and sul3 genes originated as a mobilization of folP genes erill@umbc.edu present in, respectively, the Rhodobiaceae and the Leptospiraceae, and indicate that the Rhodobiaceae folP gene was transferred from the Leptospiraceae. Analysis of Specialty section: This article was submitted to %GC content in folP/sul gene sequences supports the phylogenetic inference results Antimicrobials, Resistance and indicates that the emergence of the Sul motif in chromosomally encoded FolP and Chemotherapy, a section of the journal proteins is ancient and considerably predates the clinical introduction of sulfonamides. Frontiers in Microbiology In vitro assays reveal that both the Rhodobiaceae and the Leptospiraceae, but not Received: 14 September 2018 other related chromosomally encoded FolP proteins confer resistance in a sulfonamide- Accepted: 21 December 2018 sensitive Escherichia coli background, indicating that the Sul motif is associated with Published: 10 January 2019 sulfonamide resistance. Given the absence of any known natural sulfonamides targeting Citation: Sánchez-Osuna M, Cortés P, DHPS, these results provide a novel perspective on the emergence of resistance to Barbé J and Erill I (2019) Origin of the synthetic chemotherapeutic agents, whereby preexisting resistant variants in the vast Mobile Di-Hydro-Pteroate Synthase Gene Determining Sulfonamide bacterial pangenome may be rapidly selected for and disseminated upon the clinical Resistance in Clinical Isolates. introduction of novel chemotherapeuticals. Front. Microbiol. 9:3332. doi: 10.3389/fmicb.2018.03332 Keywords: sulfonamide, resistance, antibiotic, phylogeny, mobile element, integron, plasmid Frontiers in Microbiology | www.frontiersin.org 1 January 2019 | Volume 9 | Article 3332 Sánchez-Osuna et al. Origin of Clinical Sulfonamide Resistance INTRODUCTION Resistance to sulfonamide through increased production of PABA was reported in the early 1940’s (Landy et al., 1943), but the Antibiotic resistance is a pressing problem in modern healthcare most commonly reported mechanism of sulfonamide resistance (Carlet et al., 2014; Rossolini et al., 2014). Bacterial cells are mutations to the chromosomal folP gene (Huovinen et al., present several mechanisms to cope with exposure to antibiotics 1995; Sköld, 2000). Mutations to the chromosomal folP gene or chemotherapeutic agents, which may be acquired through have been shown to provide varying degrees of trade-off mutation or, most frequently, via lateral gene transfer on between resistance and efficient folate synthesis, decreasing mobile genetic elements (Davies and Davies, 2010). These DHPS affinity for sulfonamide while maintaining or increasing its mechanisms include modification of the antimicrobial target, affinity for PABA (Sköld, 2000). These mutations have occurred degradation or chemical modification of the antimicrobial independently in multiple bacterial genera and target multiple molecule, targeted reduction of antimicrobial uptake, active conserved areas of the DHPS protein (Sköld, 2000). However, export of the antimicrobial through efflux pumps and use of similar mutational profiles, such as two-amino acid insertions in alternate pathways and enzymes (Davies and Davies, 2010). Neisseria meningitidis and Streptococcus pneumoniae, have been It is widely accepted that many antibiotic resistance genes reported (Rådström et al., 1992; Haasum et al., 2001), and in both present today in pathogenic bacteria originated from homologs these genera there is evidence of extensive recombination within evolved over eons in either the microbes that naturally produce folP genes (Fermer et al., 1995; Swedberg et al., 1998). the antibiotics or their natural competitors (Aminov and In spite of the multiple instances of chromosomal folP Mackie, 2007). When coupled with the high plasticity of resistant variants, clinical resistance to sulfonamides is bacterial genomes and their co-existence with a large variety predominantly plasmid-borne and mediated by sul genes of genetic mobile elements, the availability of a readily evolved encoding alternative sulfonamide-resistant DHPS enzymes pool of antibiotic resistance genes set the stage for the rapid (Sköld, 2000). Four different sul genes have been described to proliferation of multi-resistant strains in the clinical setting date, with sul1 and sul2 being the predominant forms in clinical shortly after the commercial introduction of antibiotics (Aminov isolates (Rådström et al., 1991). The sul1 gene is typically found in and Mackie, 2007). In contrast, the origins of resistance against class 1 integrons and linked to other resistance genes (Rådström chemotherapeutic agents are harder to pinpoint. Since these were et al., 1991), whereas sul2 is usually associated to non-conjugative designed in vitro, it seems unlikely that a large pool of genes plasmids of the IncQ group (Enne et al., 2001) and to large conferring resistance to chemotherapeutic agents existed before transmissible plasmids like pBP1 (Treeck et al., 1981). The their introduction. After their discovery in the 1960’s, resistance sul3 gene was characterized in the Escherichia coli conjugative to quinolones was initially rare and limited to chromosomal plasmid pVP440. It was shown to be flanked by two copies of mutations in DNA gyrase, topoisomerase IV or efflux pumps the insertion element IS151/26 and to be widespread in E. coli (Hooper, 1999). However, in the early 2000’s plasmid-borne qnr isolates from pigs in Switzerland (Perreten and Boerlin, 2003). genes were first detected and spread rapidly to clinical pathogens. Recently, a sul4 gene was identified in a systematic prospection of Qnr is a member of the pentapeptide repeat family and was class 1 integron-borne genes in Indian river sediments, but this shown to confer resistance by binding to DNA gyrase and sul variant has not yet been detected in clinical isolates. Genomic limiting the effect of quinolone drugs. The origin of plasmid- context analyses revealed that the sul4 gene had been recently borne qnr genes has been traced to environmental homologs mobilized and phylogenetic inference pinpointed its putative and these are thought to have derived from genes originally origin as part of the folate synthesis cluster in the Chloroflexi targeting antibiotics, such as microcin B17 (Tran and Jacoby, phylum (Razavi et al., 2017). 2002). Despite the importance of sulfonamides in human and Aryl sulfonamides are synthetic antibacterial compounds animal therapy, the putative origin of the three sul genes that presenting a structure similar to para-amino benzoic acid account for the vast majority of reported clinical resistance to (PABA), and containing a sulfonamide group linked to an sulfonamide remains to be elucidated. In this work we leverage aromatic group. Commonly referred to as sulfonamides or sulfa comparative genomics, phylogenetic analysis and in vitro drugs due to their clinical relevance, synthetic aryl sulfonamides determination of minimal inhibitory concentrations (MIC) of function as competitive inhibitors of the di-hydro-pteroate sulfamethoxazole to unravel the origin of the sul1, sul2, and synthase (DHPS) enzyme, encoded in bacteria by the folP gene sul3 genes. Our analysis indicates that chromosomally encoded (Sköld, 2000). DHPS participates in folate synthesis using PABA folP genes conferring resistance to sulfonamide originated in as a substrate, and the competitive inhibition of DHPS by members of the Leptospiraceae family and were transferred to the sulfonamides results in growth arrest (Mitsuhashi, 1993; Sköld, Alphaproteobacteria Rhodobiaceae family more than 500 million 2000). Experiments in mice in the 1930’s demonstrated the years ago. These isolated sources of chromosomally encoded effectiveness of sulfonamide against bacteria, and sulfonamide sulfonamide-resistant DHPS were mobilized independently, became the first antibacterial chemotherapeutic to be used leading to the broadly disseminated sul1, sul2, and sul3 resistance systemically (Domagk, 1935; Aminov, 2010). It remained in genes. Our results hence indicate that resistance to synthetic use throughout World War II, but by the end of the 1940’s chemotherapeutic agents may be available in the form of resistant strains started to emerge and sulfonamides were rapidly chromosomally encoded variants among the extremely diverse displaced in favor of the newly discovered antibiotics (Sköld, bacterial domain, and can be rapidly disseminated upon the 2000; Davenport, 2012). release of novel synthetic drugs. Frontiers in Microbiology | www.frontiersin.org 2 January 2019 | Volume 9 | Article 3332 Sánchez-Osuna et al. Origin of Clinical Sulfonamide Resistance MATERIALS AND METHODS service (Castresana, 2000; Dereeper et al., 2008). Bayesian phylogenetic inference on trimmed alignments was performed Data Collection with MrBayes (MrBayes, RRID:SCR_012067) (Ronquist and FolP, GlmM, and Sul1-3 homologs were identified in complete Huelsenbeck, 2003). Four Metropolis-Coupled Markov Chain GenBank sequences (GenBank, RRID:SCR_002760) through Monte Carlo runs with four independent chains were carried out BLASTP (BLASTP, RRID:SCR_001010) (Altschul et al., for 30,000,000 generations, and the resulting consensus tree was 1997) using the E. coli FolP (WP_000764731) and GlmM plotted with FigTree (FigTree, RRID:SCR_008515). (WP_000071134) proteins as the query. Putative homologs were detected as BLASTP hits passing stringent e-value DNA Sequence Analyses (<1e−20) and query coverage (75%) thresholds. FolP and Analysis of %GC in synonymous and non-synonymous GlmM chromosomally encoded proteins were identified on a patterns and K a /K s divergence were performed according to representative genome of all bacterial orders with complete the Nei–Gojobori computation method (Nei and Gojobori, genome assemblies on RefSeq (RefSeq, RRID:SCR_003496), of 1986) and the standalone PAL2NAL program for codon-based each bacterial family for the Proteobacteria, of any bacterial alignments (Suyama et al., 2006), using custom Python scripts species where chromosomally encoded sulfonamide resistance for pipelining. Analyses of %GC content were performed on mutants had been reported, and on all available complete all sampled bacterial genomes, computing genome-wide %GC genomes for clades of interest (Rhodobiaceae, Spirochaetes, and statistics and comparing them to folP estimates. Analyses of Chlamydiae) (Supplementary Table S1). All protein coding gene K a /K s divergence were performed on pair-wise alignments sequences for these genomes were downloaded for %GC analysis. of the N- and C-terminal ends of the glmM gene sequence Sul proteins encoded by mobile sul genes were identified on of all sampled bacterial groups. One-sided Mann–Whitney complete plasmid, transposon, and integron GenBank sequences U-tests were performed using GraphPad Prism (GraphPad, (GenBank, RRID:SCR_002760). RRID:SCR_002798) to determine whether differences between folP and chromosomal %GC content were significantly different Identification and Visualization of in the presence and absence of Sul-like signature motifs, and Sul-Like Signatures in FolP Sequences whether the N- and C-terminal regions presented different To identify sequence motifs associated with Sul proteins, mutational profiles. The scripts used for the analysis are we performed a CLUSTALW alignment (Clustal Omega, available at the GitHub ErillLab repository. Nucleotide sequence RRID:SCR_001591) using a non-redundant (<99% identity) identities percentages were computed on gapless positions of subset of the Sul1-3 homologous sequences detected previously PAL2NAL codon-based alignments using a custom Python and FolP sequence sampled from each bacterial clade. Following script. Amelioration times were estimated using the Ameliorator visual inspection of the resulting alignment, a Sul-like motif program (Lawrence and Ochman, 1997) under different conserved in several chromosomally encoded FolP proteins was selection modes. K a and K s values were estimated from visualized using iceLogo (iceLogo, RRID:SCR_012137) (Colaert pairwise alignments of orthologs between the Parvibaculum et al., 2009) and a consensus motif was derived and encoded lavamentivorans and Leptospira interrogans genomes as into a PROSITE-format pattern (PROSITE, RRID:SCR_003457). determined by the OMA Orthology database (OMA Orthology The inferred PROSITE pattern was used to seed a Pattern database, RRID:SCR_016425) (Altenhoff et al., 2018) and species Hit Initiated BLAST (NCBI PHI-BLAST; RRID:SCR_004870) divergence times were inferred from published molecular clock search against the NCBI non-redundant Protein database phylogenies (Battistuzzi et al., 2004). (Protein, RRID:SCR_003257) using as a query the protein sequences of Sul1-3 reported in the literature (WP_001336346, Cloning, Transformation and WP_010890159, and WP_000034420) and conservative e-value Complementation of the folP Gene for (<1e−20) and query coverage (75%) limits. Only chromosomal Broth Microdilution Assays hits with the identified signature characteristic of sul gene The L. interrogans serovar Lai str. 56601 folP and Chlamydia products were retained for further analysis. trachomatis D/UW-3/CX folKP gene were synthesized and adapted to E. coli codon usage at ATG:biosynthetics GmbH, Multiple Sequence Alignment and Germany; whereas P. lavamentivorans DS-1 (DSMZ 13023) and Phylogenetic Inference Rhodobacter sphaeroides 2.4.1 (gently provided by Professor For phylogenetic inference, multiple sequence alignments S. Kaplan; Health Science Center. University of Texas) folP of identified FolP/Sul1-3 and GlmM homologous sequences genes were amplified from genomic DNA. The sul2 gene was were performed with CLUSTALW (Clustal Omega, amplified from the RSF1010 plasmid (Kushner, 1978; Honda RRID:SCR_001591) (Thompson et al., 1994) using variable et al., 1991) and used as a positive control. The folP/sul genes (5, 10, and 25) gap opening penalties. These alignments were amplified using suitable primers (Supplementary Table were then integrated with local LALIGN alignments with S2), purified PCR products were then digested with NdeI T-COFFEE (T-Coffee, RRID:SCR_011818) (Notredame et al., and BamHI (New England Biolabs, RRID:SCR_013517) and 2000), and the resulting alignment was trimmed using the cloned into a dephosphorylated pUA1108 vector (Mayola et al., “less stringent selection” parameters of the Gblocks online 2014), previously cut with the same restriction enzymes. The Frontiers in Microbiology | www.frontiersin.org 3 January 2019 | Volume 9 | Article 3332 Sánchez-Osuna et al. Origin of Clinical Sulfonamide Resistance ligation was introduced by transformation into competent signature Sul motif as FolP∗ , and to their encoding gene E. coli DH5α cells and recombinant plasmids were extracted as folP∗ . with the NZYMiniprep kit (NZYTech, RRID:SCR_016772), In order to gain further insight into the possible chromosomal sequenced (Macrogen, RRID:SCR_014454) and introduced by origins of sul genes, we performed tBLASTX searches against the transformation in competent E. coli K-12 (CGSC 5073) (CGSC, NCBI RefSeq Genome Database (RefSeq, RRID:SCR_003496) RRID:SCR_002303). Susceptibility to sulfamethoxazole (Sigma- using the genetic surroundings (5,000 bp) of sul1, sul2, and Aldrich, RRID:SCR_008988) for the strains containing the sul3 genes with at most 90% similarity to those reported in the folP/sul genes was determined using broth microdilution tests literature (Supplementary Table S3). This search did not return in Mueller-Hinton broth (MH) with half serial dilutions consistent results for the sul1 and sul3 genetic surroundings, of sulfamethoxazole ranging from 512 to 0.125 mg/L, but it identified a conserved gene fragment encoding the following the CLSI guidelines (Clinical and Laboratory N-terminal region of the phosphoglucosamine mutase GlmM Standards Institute, 2003). Colonies were grown on Luria- protein downstream of sul2 in multiple plasmids harboring Bertani (LB) agar for 18 h and then suspended in sterile this resistance gene. These sul2-associated GlmM sequences 0.9% NaCl solution to a McFarland 0.5 turbidity level. lack the entire GlmM C-terminal region, including three of These suspensions were diluted at 10−2 in Mueller-Hinton its functional domains (Mehra-Chaudhary et al., 2011), and it (MH) broth, and 50 µl (5.104 cells) were inoculated onto can therefore be safely assumed that they are not functional as microtiter plates containing 50 µl of MH broth supplemented phosphoglucosamine mutases. This genetic arrangement has with 1024–0.250 mg/L of sulfamethoxazole. To determine been reported previously as a feature of sul2 isolates (Kehrenberg growth, absorbance at 550 nm was measured after 24 h and Schwarz, 2005; Hu et al., 2016), and it is strongly conserved incubation at 37◦ C (Sunrise plate reader; Tecan Life Sciences, in the genomic surroundings of chromosomal folP genes RRID:SCR_016771). in the Gammaproteobacteria, the Betaproteobacteria and several Alphaproteobacteria lineages (Figure 1C). Analysis of the folP genetic surroundings in complete genomes of RESULTS the Spirochaetes and the Alphaproteobacteria shows clear differences between the genes coding for the identified Identification of Putative Chromosomal Rhodobiaceae and Leptospiraceae FolP∗ proteins harboring Origins for sul1-3 Genes the two-amino acid insertion pattern and those without it To identify putative chromosomal homologs of sul1-3 genes, (Figure 1C). The Leptospiraceae show a conserved arrangement we performed a multiple sequence alignment including any with folP∗ flanked by a peptidoglycan-associated lipoprotein protein sequences with at most 90% similarity to those and a tetratricopeptide repeat-containing domain protein, encoded by sul1-3 genes reported in the literature and by whereas in most other Spirochaetes folP is flanked by a 1- chromosomal folP genes from a representative of each bacterial deoxy-D-xylulose-5-phosphate synthase and a diadenylate order. Inspection of the resulting alignment (Figure 1A and cyclase. In contrast, the Alphaproteobacteria yield several Supplementary Data 1) revealed the presence of a two- distinct syntenic regions for folP. In the Rhodobiaceae, amino acid insertion in proteins encoded by sul1-3 genes folP∗ is flanked by genes coding for either a FtsH-family that is not present in those encoded by sul4 or the analyzed metallopeptidase or a TetR-family transcriptional repressor and chromosomal folP genes. This two-amino acid insertion is the phosphoglucosamine mutase glmM. In the Rhodobacterales, located in a conserved region of the FolP protein (residues folP is flanked by a dihydroneopterin aldolase and glmM, but in R171-N211 of the E. coli FolP protein [WP_000764731]) that the Rhizobiales it is flanked by a Zn-dependent protease and the presents other signature changes in sul-encoded proteins with dihydroneopterin aldolase. This last arrangement, in which the respect to chromosomally encoded FolP proteins (Figures 1A,B dihydroneopterin aldolase is followed by a 2-amino-4-hydroxy- and Supplementary Data 1) (Rådström and Swedberg, 1988; 6-hydroxymethyldihydropteridine diphosphokinase is also part Morgan et al., 2011). We derived a PROSITE-format pattern of the genetic surroundings of folP in most Actinobacteria (PROSITE, RRID:SCR_003457) (Supplementary Data 2) of (Figure 1C). the identified Sul motif to seed a Pattern Hit Initiated BLAST (NCBI PHI-BLAST; RRID:SCR_004870) search against the NCBI non-redundant (NR) protein database (NCBI, Phylogenetic Analysis of sul/folP and RRID:SCR_006472). This search identified several proteins glmM Genes encoded by Rhodobiaceae family members that presented a The presence of a signature two-amino acid insertion similar insertion pattern. BLASTP (BLASTP, RRID:AB_141607) characteristic of sul gene products in chromosomally encoded searches with these Rhodobiaceae FolP sequences matched FolP∗ proteins and the identification of a genetic environment proteins in several members of the Leptospiraceae and the for sul2 genes that is conserved in multiple bacterial genomes Chlamydiae. However, analysis of the resulting multiple sequence suggested that it might be possible to pinpoint the evolutionary alignment showed that only the Leptospiraceae FolP protein origin of sul genes. To further investigate this possibility, sequences displayed the identified two-amino acid insertion we performed a rigorous phylogenetic analysis of FolP/Sul pattern (Supplementary Data 3). Heretofore, we refer to protein sequences. We sampled a representative genome of these chromosomally encoded FolP proteins containing the all bacterial orders with complete genome assemblies, of each Frontiers in Microbiology | www.frontiersin.org 4 January 2019 | Volume 9 | Article 3332 Sánchez-Osuna et al. Origin of Clinical Sulfonamide Resistance FIGURE 1 | (A) Segment of the multiple sequence alignment including any sul genes with at most 90% similarity to reported sul genes and a representative chromosomal folP gene for all bacterial phyla with complete genomes available in NCBI RefSeq. (B) IceLOGO highlighting the difference in amino acid frequency at each position of the region of the folP protein sequence containing the identified insertion between the multiple sequence alignment of sul gene products and the chromosomally encoded FolP proteins. The upper part of the iceLOGO plot shows residues overrepresented in the sul-encoded FolP proteins; the bottom part shows residues overrepresented in chromosomally encoded FolP proteins for all bacterial phyla with complete genomes available. Only differences with significant z-score under a confidence interval of 0.01 are shown. (C) Schematic representation of the genetic environment of sul2 genes, similar arrangements in chromosomally encoded folP genes of the Gammaproteobacteria, Betaproteobacteria, and Alphaproteobacteria, and arrangements in other major phyla. Arrow boxes indicate coding regions. When available, gene names or NOG identifiers are provided. Boxes for folP genes containing the two-amino acid insertion are designated as folP∗ . bacterial family for the Proteobacteria and all available complete RRID:SCR_011818) and MrBayes (MrBayes, RRID:SCR_012067) genomes for clades of interest (Rhodobiaceae, Spirochaetes, and (Supplementary Data 5). Chlamydiae), and we identified chromosomally encoded FolP The resulting tree (Figure 2) provides strong support for homologs in each of these genomes using BLASTP (BLASTP, the hypothesis that sul1-3 genes originated in the Rhodobiaceae RRID:SCR_001010) with the E. coli FolP protein as a query. We and Leptospiraceae families. In particular, the topology inferred used a distance tree generated with CLUSTALW (Clustal Omega, by MrBayes suggests that the Leptospiraceae folP∗ gene gave RRID:SCR_001591) to identify and discard a set of protein rise to both sul3 and the folP∗ gene encountered in the sequences from duplicated folP genes in the Actinobacteria Rhodobiaceae, most likely through a lateral gene transfer event (Supplementary Data 4), and we performed multiple sequence in an ancestor of this Alphaproteobacteria family. According to alignment and Bayesian phylogenetic reconstruction of the the reconstructed FolP phylogeny, the Rhodobiaceae folP∗ gene remaining FolP/Sul sequences with T-COFFEE (T-Coffee, was subsequently mobilized as sul2, and later evolved into the Frontiers in Microbiology | www.frontiersin.org 5 January 2019 | Volume 9 | Article 3332 Sánchez-Osuna et al. Origin of Clinical Sulfonamide Resistance FIGURE 2 | Consensus tree of Sul/FolP protein sequences. Branch support values are provided as Bayesian posterior probabilities. For clarity, only posterior probability values higher than 0.8 are displayed. Proposed lateral gene transfer and mobilization events are shown by means of superimposed continuous and dotted, respectively, arrows. integron-borne sul1 gene (Orman et al., 2002). The fact that the phylogenetic evidence thus indicates that the sul2 gene was Leptospiraceae FolP∗ sequences branch independently of other excised with the N-terminal fragment of the glmM gene during Spirochaetes sequences and immediately after the Chlamydiae the mobilization event that led to their incorporation into suggests that the Leptospiraceae folP∗ gene might have originated plasmid vectors. Given that the folP-glmM arrangement is only as a result of lateral gene transfer event from the Chlamydiae, seen in the Proteobacteria, this also excludes the possibility and that it subsequently incorporated the signature two-amino that the sul2 gene was mobilized directly from a Leptospiraceae acid insert present in sul-encoded DHPS proteins. Importantly, background, where the folP gene presents an unrelated, yet the trimmed multiple sequence alignment used for FolP/Sul conserved, genomic environment (Figure 1C). Bayesian phylogenetic inference (Supplementary Data 5) does not incorporate the two-amino acid insertion of the Sul motif, indicating that the joint branching of Sul1-3 sequences with Analysis of sul/folP and glmM Gene chromosomally encoded Rhodobiaceae and Leptospiraceae FolP Sequences proteins is based on sequence similarity beyond this insertion and The phylogenetic analysis of FolP and GlmM sequences puts its immediate vicinity (Figures 1A,B). forward an evolutionary scenario wherein the Leptospiraceae The existence of a genetic environment for sul2 genes folP∗ was transferred to the members of the Rhodobiaceae conserved in bacterial chromosomes provides the means to family before being mobilized independently into the sul3- independently assess the likelihood of the evolutionary scenario and sul1/2-harboring mobile genetic elements reported in inferred from the FolP phylogeny. Using the same sampling sulfonamide-resistant clinical isolates. To further investigate this methods utilized for sul/folP protein products, we collected hypothesis, we undertook a systematic analysis of folP and protein sequences for phosphoglucosamine mutase (GlmM) glmM coding sequences. We compiled folP gene sequences homologs and performed Bayesian phylogenetic inference on the for all the FolP proteins included in the phylogenetic analysis aligned N-terminal regions. The resulting GlmM tree (Figure 3) (Figure 2), as well as any sul gene sequences with less provides further support for a Rhodobiaceae origin of the than 90% identity to those reported in the literature and sul2 gene, with the sul2-associated GlmM sequences branching any chromosomal folP∗ genes encoding a DHPS with the with the Rhodobiaceae GlmM protein sequences deep within signature Sul motif (Figure 1A) for which there were at least an otherwise monophyletic Alphaproteobacteria clade. Taken 1 Mbp of whole genome shotgun sequence data (Supplementary together, the consistent branching with the Rhodobiaceae of the Data 6). We computed the overall and codon-position %GC protein sequences encoded by both sul2 and its accompanying content on both the folP/sul coding sequences and all the glmM gene fragment firmly establish this Alphaproteobacteria available coding sequences in their respective genome assembly family as the chromosomal origin for the sul2 gene. The (Supplementary Table S4). The %GC content data (Figure 4A) Frontiers in Microbiology | www.frontiersin.org 6 January 2019 | Volume 9 | Article 3332 Sánchez-Osuna et al. Origin of Clinical Sulfonamide Resistance FIGURE 3 | Consensus tree of N-terminal GlmM protein sequences. Branch support values are provided as Bayesian posterior probabilities. For clarity, only posterior probability values higher than 0.8 are displayed. The placement of sul2-encoded proteins is indicated by a shaded ellipse. reveals that sul1/2 sequences have a high %GC content from the Leptospiraceae into a Rhodobiaceae ancestor. In this (60.76 SD ± 1.42) that is consistent with their origin as context, the substantial divergence in %GC content between the mobilized Rhodobiaceae folP∗ sequences (%GC content: 62.02 chromosomal folP∗ genes of both clades indicates a long process SD ± 2.22). Similarly, sul3 sequences display a %GC content of amelioration. In fact, statistical analysis of the differences (38.14 SD ± 0.55) consistent with their mobilization from in codon position %GC content between folP genes and all a Leptospiraceae folP∗ background (39.39 SD ± 4.17). This available coding sequences in their respective genomes shows similarity in %GC content precludes the dating of these that Leptospiraceae and Rhodobiaceae folP∗ genes encoding mobilization events through analysis of DNA amelioration proteins with the Sul motif cannot be distinguished from other rates. However, examination of DNA sequence similarity folP genes (one-sided Mann–Whitney U-test p > 0.05 for GC1, (Table 1) reveals identity values of 50–60% between the GC2, and GC3) (Figure 4B) (Supplementary Table S4). We used posited chromosomal donors and their mobile counterparts. Ameliorator (Lawrence and Ochman, 1997) to estimate the time These values are in the lower range of sequence identities required for the observed amelioration via forward simulation for previously described mobilization events (Table 1), and from Leptospiraceae codon position %GC values. Even under hence point to an early mobilization of folP∗ genes that is assumptions of fast evolutionary change, the software provides a consistent with the almost universal association of sul genes lower bound of 476 million years for the observed amelioration with class 1 mobile integrons (Lévesque et al., 1995). Together of the Leptospiraceae folP∗ gene into the Rhodobiaceae one. with the phylogenetic inference results, these data provide strong Statistical analysis of synonymous and non-synonymous support for an independent mobilization of sul1/2 and sul3 mutation patterns in the N- and C-terminal regions of the genes from, respectively, Rhodobiaceae and Leptospiraceae family glmM gene also shows that mutation patterns in each region chromosomal backgrounds. of the Rhodobiaceae glmM gene are indistinguishable from The independent mobilization of sul1/2 and sul3 is those observed in other glmM genes (one-sided Mann–Whitney underpinned by a preceding lateral gene transfer of folP∗ U-test p > 0.05). Since the glmM gene fragment associated to Frontiers in Microbiology | www.frontiersin.org 7 January 2019 | Volume 9 | Article 3332 Sánchez-Osuna et al. Origin of Clinical Sulfonamide Resistance FIGURE 4 | (A) %GC content of folP and all other chromosomal coding sequences in different clades. The %GC content of sul1/2 and sul3 genes is shown adjacent to that of the Rhodobiaceae and the Leptospiraceae. (B) Correlation between the %GC content of folP genes and that of all other coding sequences in their respective genomes. The data points corresponding to folP∗ genes from the Rhodobiaceae and the Leptospiraceae are shown as squares. sul2 genes is likely to be non-functional and subject to genetic fragments in sul1 isolates supports in turn the notion that sul1 drift, the absence of diverging substitution patterns between the derived from sul2. This is consistent with the branching pattern N- and C-terminal regions of Rhodobiaceae glmM sequences observed in the FolP/Sul tree (Figure 2) and the observed DNA indicates that the glmM and sul2 genes were transferred from identity values (Table 1), which define a scenario of independent the Rhodobiaceae to sul2-harboring vectors, and not vice versa mobilization of sul3 from the Leptospiraceae and sul2 from the (Supplementary Table S5). Lastly, given that gene loss is much Rhodobiaceae, with the subsequent uptake of sul1 by class 1 more likely than gain (Kannan et al., 2013), the absence of glmM integrons. Frontiers in Microbiology | www.frontiersin.org 8 January 2019 | Volume 9 | Article 3332 Sánchez-Osuna et al. Origin of Clinical Sulfonamide Resistance TABLE 1 | Percent GC content and nucleotide identity between posited chromosomal donors and the corresponding mobilized resistance genes for a number of antibiotics. Putative donor %GC Mobile determinant %GC % Identity PMID Sulfonamide Leptospira interrogans 32.62 Salmonella enterica sul3 (WP_117344859.1) 37.75 51.91 (This work) Leptospiraceae 39.39 ± 4.17 sul3 group 38.14 ± 0.55 49.73 ± 1.55 (This work) Spirochaetales 52.22 ± 9.24 sul3 group 38.14 ± 0.55 47.28 ± 2.63 (This work) Parvibaculum lavamentivorans 64.48 Pasteurella multocida sul2 (WP_010890206.1) 60.61 63.33 (This work) Rhodobiaceae 62.02 ± 2.22 sul2 group 59.6 ± 2.70 60.46 ± 2.37 (This work) Alphaproteobacteria 62.34 ± 13.20 sul2 group 59.6 ± 2.70 48.14 ± 3.28 (This work) Parvibaculum lavamentivorans 64.48 Acinetobacter baumannii sul1 (AFB76381.1) 61.78 54.65 (This work) Rhodobiaceae 62.02 ± 2.22 sul1 group 61.72 ± 0.09 54.06 ± 0.51 (This work) Alphaproteobacteria 62.34 ± 13.20 sul1 group 61.72 ± 0.10 46.50 ± 3.48 (This work) Vancomycin Amycolatopsis orientalis 64.86 vanX 64.59 76.57 22303296 Quinolones Shewanella algae 53.12 qnrA 53.27 95.6 16048974 Vibrio splendidus 44.44 qnrS 43.99 72.78 17452482 β-lactams Kluyvera cryocrescens 54.02 CTX-M-1 54.11 99.3 11709346 Kluyvera ascorbata 56.62 CTX-M-2 56.73 99.2 12183268 Kluyvera georgiana 57.07 CTX-M-8 57.08 98.5 12435721 Kluyvera georgiana 57.64 CTX-M-9 60.5 99 15855541 Kluyvera georgiana 57.53 CTX-M-25 57.99 95.9 20421403 Groups of homologs for sul genes correspond to sequences with at most 90% similarity to reported sul genes. For each donor clade and target group, the most likely mapping is shown. Where appropriate, the PudMed identifier (PMID) for the reference establishing the correspondence between putative chromosomal donor and mobile resistance determinant is provided. Sulfonamide Resistance of of conferring resistance to sulfonamide in E. coli. In contrast Chromosomal folP Genes with the Leptospiraceae and the Rhodobiaceae folP∗ genes, the chromosomal folKP gene of the Chlamydiae, which encodes Phylogenetic and sequence analysis results indicate that a DHPS lacking the Sul motif, does not confer resistance to chromosomal folP∗ genes encoding proteins with the signature sulfamethoxazole (Table 2). This is in agreement with abundant Sul motif were independently mobilized into the sul1-3- reports of sulfonamide susceptibility in several Chlamydia harboring mobile elements found in sulfonamide-resistant species (Hammerschlag, 1982; Fan et al., 1992; Sandoz and clinical isolates, but they do not address whether the presence Rockey, 2010; Marti et al., 2018). Since the Chlamydiae folKP of this motif is associated with sulfonamide resistance. To gene is the most closely related chromosomal folP gene to the investigate this possibility, we cloned the folP gene coding cluster encompassing the sul genes and the Leptospiraceae and for DHPS in the Rhodobiaceae P. lavamentivorans DS-1 the Rhodobiaceae folP∗ (Figure 2), the lack of resistance in (WP_012111048), the Leptospiraceae L. interrogans serovar Chlamydiae folKP genes strongly suggests that changes in the Lai str. 56601 (WP_000444207), the Rhodobacteraceae region encompassing the Sul motif may be responsible for the R. sphaeroides 2.4.1 (WP_011337038) and the Chlamydiae observed resistance. This region is located in a connector loop C. trachomatis D/UW-3/CX (WP_009871981). Following within the N-terminal ‘pole’ of the eight-stranded α/β barrel of Clinical and Laboratory Standards Institute (CLSI) guidelines DHPS, which is involved in sulfonamide recognition (Rådström (Clinical and Laboratory Standards Institute, 2003), we then and Swedberg, 1988; Morgan et al., 2011). The two-amino acid performed broth microdilution assays to determine the minimal insertion might hence result in decreased affinity for sulfonamide inhibitory concentration (MIC) of sulfamethoxazole. The results by locally disrupting folding, as has been proposed previously shown in Table 2 reveal that both P. lavamentivorans and for similar insertions in chromosomal folP genes (Achari et al., L. interrogans chromosomal folP∗ genes confer resistance to 1997). sulfamethoxazole in an E. coli strain sensitive to sulfonamides. These data are in agreement with previous reporting of sulfonamide resistance in multiple L. interrogans strains (Chakraborty et al., 2010, 2011; Wuthiekanun et al., 2015), DISCUSSION and suggest that the observed resistance was likely due to mutations in the Leptospiraceae chromosomal folP∗ Prevalence of Sulfonamide Resistance in gene rather than to the presence of plasmid-borne sul Ancestral Bacteria genes. Moreover, our results show that complementation The evidence presented here converges toward an evolutionary with folP genes from another Alphaproteobacteria family scenario in which sul1-3 genes from clinical isolates derive lacking the Sul motif, the Rhodobacteraceae, does not confer from ancestral chromosomal mutations in the folP∗ gene of the resistance to sulfamethoxazole. These results reveal that the Leptospiraceae and the Rhodobiaceae (Figure 5). The emergence chromosomal folP∗ genes that gave rise to sul genes are capable and maintenance of a sulfonamide-resistant folP∗ gene in the Frontiers in Microbiology | www.frontiersin.org 9 January 2019 | Volume 9 | Article 3332 Sánchez-Osuna et al. Origin of Clinical Sulfonamide Resistance TABLE 2 | Broth microdilution assays. The emergence and persistence of mutations conferring resistance against chemotherapeutic agents prior to their Sulfamethoxazole (mg/L) discovery may also have taken place in the absence of selection Escherichia coli CGSC5073 8 for the resistance-conferring mutations. The appearance of Escherichia coli CGSC5073 pUA1108 8 sulfonamide-resistance mutations in chromosomal folP genes Escherichia coli CGSC5073 pUA1108::folPPl >512 has been amply documented (Huovinen et al., 1995; Sköld, Escherichia coli CGSC5073 pUA1108::folPLi 512 2000), and these were in fact the primary drivers of sulfonamide Escherichia coli CGSC5073 pUA1108::folKPCt 4 resistance following the introduction of sulfa drugs (Sköld, Escherichia coli CGSC5073 pUA1108::folPRs 8 2000). Furthermore, it has been documented that the presence Escherichia coli CGSC5073 pUA1108::sul2 >512 of sulfonamide resistant DHPS does not necessarily impose a Minimum inhibitory concentrations (MICs) of sulfamethoxazole in wild-type fitness cost on bacteria (Enne et al., 2004). Structural studies Escherichia coli CGSC5073 carrying different versions of pUA1108::folP; have suggested that most sulfonamide resistance mutations act Pl, Parvibaculum lavamentivorans; Li, Leptospira interrogans; Ct, Chlamydia by modulating accessibility of sulfonamides to the PABA-binding trachomatis; Rs, Rhodobacter sphaeroides. pocket without hindering PABA binding (Baca et al., 2000; Morgan et al., 2011). It is hence conceivable that naturally Leptospiraceae and its subsequent transfer to the Rhodobiaceae occurring mutations conferring resistance to sulfonamide might suggests that it might have conveyed some selective advantage not be selected against in the absence of this chemotherapeutic prior to the introduction of sulfonamides, but the advent of agent. Subsequent complementary changes to adjust the affinity mutations providing significant resistance to sulfonamide and for PABA of the altered DHPS molecule may have resulted their subsequent spread could also have been fortuitous. In both in fixation of the original mutations conferring resistance to scenarios, formally variants of an evolutionary exaptation process sulfonamide (Andersson, 2006). (Gould and Vrba, 1982), a resistance-causing mutation may arise Alternatively, sulfonamide resistance mutations in folP may and be maintained in the population in the absence of direct have arisen and persisted in response to naturally occurring antibiotic selection. Upon the clinical introduction of the relevant sulfonamides produced by competing organisms. Sulfonamides antibiotic, selection favors the rapid spread of the resistance are rare in nature, with only eight known natural sulfonamides determinant. reported to date (Petkowski et al., 2018). Of these, only The emergence and maintenance of resistance against two naturally occurring sulfonamides are aryl sulfonamides, synthetic chemotherapeutic agents prior to their clinical produced in very small amounts by recombinant Streptomyces deployment may hence reside in the evolution of resistance species harboring the complete xiamycin biosynthesis gene as a side-effect of mutations providing some other fitness cluster (Baunach et al., 2015). Although these sulfonamides show benefit. A case in point is the evolution and rapid mobilization potent antimicrobial activity, their bulky substitution pattern via plasmid-borne qnr genes of resistance to quinolones suggests that their mode of action and molecular target are (Nordmann and Poirel, 2005). Qnr proteins bind to DNA likely different from synthetic aryl sulfonamides (Baunach et al., gyrase, the target of quinolones, early in the gyrase catalytic 2015). cycle and decrease binding of quinolones to the enzyme- DNA complex (Tran et al., 2005). Other members of the pentapeptide repeat family are known to bind DNA gyrase and Mobilization of Ancestral Resistance provide some level of quinolone resistance (Hegde et al., 2005). Reservoirs Conversely, qnr-encoded proteins can provide cross-protection The phylogenetic inference and genomic analysis results against naturally occurring antibiotics and synthetic agents reported in this work uphold an evolutionary scenario wherein targeting DNA gyrase (Hooper and Jacoby, 2016). The likely chromosomally encoded sulfonamide resistant folP variants were chromosomal origins of qnrA and qnrS genes have been independently mobilized from Leptospiraceae and Rhodobiaceae traced to water-borne isolates of, respectively, Shewanella backgrounds, and that mobile folP∗ genes subsequently spread algae and Vibrio splendidus (Cantón, 2009) (Table 1). It has very rapidly following the clinical introduction of synthetic aryl been hypothesized that these ancestral qnr genes and other sulfonamides, giving rise to the sul1/2 and sul3 genes routinely pentapeptide repeat family members may have evolved to provide detected in clinical isolates (Figure 5). The rapid mobilization resistance against other compounds targeting DNA gyrase, such and dissemination of genes conferring resistance to antibiotic as the plasmid-borne microcin B17 (Heddle et al., 2001), as and chemotherapeutic agents upon the clinical or agricultural elements facilitating or modulating the normal function of use of these compounds has been amply documented (Aminov DNA gyrase (Aminov and Mackie, 2007) or as functional and Mackie, 2007; Stokes and Gillings, 2011). Mobilization and components of the adaptation of Shewanella species to cold spread may be mediated by plasmids encoding transposons and environments (Kim et al., 2011). In this context, the sequence integrons, as well as integrative and conjugative elements, mobile determinants associated with chromosomal folP∗ genes may have pathogenicity islands and bacteriophages, but the common tenet originally enhanced PABA binding or another aspect of folate is that sustained exposure of bacterial populations to antibiotics biosynthesis (Sköld, 2000), provided protection against hitherto or chemotherapeutic agents induces a strong selective pressure unknown antibiotics targeting DHPS, or fulfilled an unrelated to elicit the mobilization of resistance determinants (Stokes and function. Gillings, 2011). Frontiers in Microbiology | www.frontiersin.org 10 January 2019 | Volume 9 | Article 3332 Sánchez-Osuna et al. Origin of Clinical Sulfonamide Resistance FIGURE 5 | Schematic diagram of the evolutionary process leading to the emergence of sul1/2- and sul3-harboring mobile genetic elements. (1) A set of mutagenic events in the Leptospiraceae folP gene generates the signature motif observed in folP∗ and sul genes. (2) Following the diversification of the Alphaproteobacteria, the Leptospiraceae folP∗ gene is transferred to the Rhodobiaceae. (3) Upon the clinical and agricultural introduction of sulfonamides, folP∗ genes from the Leptospiraceae and the Rhodobiaceae are independently mobilized, giving rise to the sul–containing mobile elements reported in clinical isolates. This figure was constructed using some Servier Medical Art templates, which are licensed under a Creative Commons - Attribution Unported License. Together with penicillin and tetracycline, sulfonamides (Kumar et al., 2005). The widespread and intensive use of have been the antibacterial agents most frequently used sulfonamides in agriculture, aquaculture and animal husbandry at sub-therapeutic levels in livestock production (Franco since the mid 1960’s, and their persistence in soil, sediments et al., 1990), and it has been reported that sulfonamides and subterranean aquatic communities where Leptospiraceae have higher mobility, lower removal efficiency and deeper and Rhodobiaceae abound, provides an ample window of environmental penetration than most other antibacterial agents opportunity for the spread of chromosomally encoded or already Frontiers in Microbiology | www.frontiersin.org 11 January 2019 | Volume 9 | Article 3332 Sánchez-Osuna et al. Origin of Clinical Sulfonamide Resistance mobilized folP∗ genes within these bacterial communities and the before the clinical introduction of sulfonamides. Nonetheless, subsequent transfer of these mobile resistance determinants to extensive recombination and unusual selection pressures on other bacterial clades. mobile elements could in principle also account for the observed Recent mobilization from a Chloroflexi chromosomal folP sequence divergence within a shorter timescale. What seems background has been postulated as the likely origin of the sul4 abundantly clear is that shortly after the clinical introduction of gene (Razavi et al., 2017), and this result is in agreement with the sulfonamides, sul genes spread rapidly on a variety of mobile phylogenetic analysis reported here (Figure 2). In the case of the elements, as attested by the well-established association between chromosomal folP∗ genes identified here and their mobilization sul genes and integrons (Lévesque et al., 1995; Davies, 2007). into sul-harboring resistance vectors, several sources of evidence Metagenomics analysis and prospective studies of preserved provide additional support for the frequent mobilization of ancient environments, such as permafrost and remote cave chromosomal folP genes. For instance, phylogenetic evidence habitats, have largely displaced the notion that antibiotic (Figure 2) indicates that the Rhodobiaceae folP∗ was incorporated resistance emerges in response to anthropogenic antibiotic use at some point by the Actinobacterium Amycolatopsis, which (D’Costa et al., 2011; Bhullar et al., 2012; Perron et al., 2015; harbors three folP orthologs (Supplementary Data 7). Similarly, Crofts et al., 2017). These studies have conclusively shown that a plasmid broadly distributed among Azospirillum isolates antibiotic resistance predates the use of antibiotics by humans, (e.g., AP010951, FQ311873), a member of the Rhodospirillaceae and that it is widely distributed across the bacterial pangenome. Alphaproteobacteria family, contains a folP gene flanked by genes In a few isolated cases, resistance determinants for synthetic coding for a flagellar export pore protein (FlhB) and the full chemotherapeutic agents that predate or have rapidly arisen upon length phosphoglucosamine mutase (GlmM) (Supplementary human use has been documented, but their existence can be Data 7). This folP does not contain the signature two-amino acid attributed to cross-resistance to naturally occurring antibiotics insertion, indicating that its mobilization occurred independently [e.g., microcin B17 for quinolones (Tran and Jacoby, 2002), of those leading to sul1/2 genes. sisomicin for amikacin (Perron et al., 2015)]. The identification More significantly, a partial genomic sequence from a in this work of ancient chromosomal mutations in folP conferring Pseudomonas aeruginosa isolate (LLMY01000073.1) harbors a resistance to sulfonamide as the likely origins of the sul1-3 genes folP∗ gene with high sequence and genetic neighborhood present in sulfonamide-resistant clinical isolates puts forward similarity to the Rhodobiaceae P. lavamentivorans DS-1 (van an alternative scenario. Given the absence of known naturally Belkum et al., 2015). The genes immediately upstream and occurring aryl sulfonamides targeting DHPS, our results suggest downstream of this P. aeruginosa folP∗ , which contains the that resistance to novel synthetic chemotherapeutic agents may Sul motif, encode a TetR family regulator and a partial be already available in the vast microbial pangenome, and that phosphoglucosamine mutase (GlmM) protein (Supplementary its global dissemination can take place in a very short amount Data 7). These three genes are flanked by IS91 and ISL3 family of time upon the clinical introduction of novel chemotherapeutic transposases. Importantly, the IS91 transposase contains similar compounds. sequence motifs and shares termini identity with ISCR elements, which are present in both sul1 and sul2-harboring plasmids (Toleman et al., 2006; Toleman and Walsh, 2011). Moreover, this P. aeruginosa folP∗ presents 72.76% nucleotide identity with DATA AVAILABILITY STATEMENT the P. lavamentivorans folP∗ , which is significantly higher than The datasets used in this study can all be freely accessed at the one observed between P. lavamentivorans folP∗ and sul2 the NCBI GenBank/RefSeq databases (https://www.ncbi.nlm.nih. (Table 1). It is hence highly likely that the P. aeruginosa folP∗ gov/). All scripts used for analysis can be obtained at the GitHub represents an intermediate step or an independent mobilization ErillLab repository (https://github.com/ErillLab/). of the Rhodobiaceae folP∗ . Taken together, the protein sequence phylogeny and genomic context evidence (Figures 1–3), the absence of Sul-containing motifs in any other chromosomal folP genes (Figure 1), the tight AUTHOR CONTRIBUTIONS alignment in %GC content between the different chromosomal folP∗ and mobile sul genes (Figure 4), and the identification of MS-O and IE performed the in silico analyses. MS-O and IE multiple folP mobilization events consistently point toward an developed scripts for genomic analyses and ran phylogenetic independent mobilization from Leptospiraceae and Rhodobiaceae inference methods. MS-O and PC performed the in vitro analyses. chromosomal backgrounds that gave rise to, respectively, sul3 IE and JB conceived the experiments and coordinated the and sul2/1 genes. The close similarity in %GC content between research. IE and MS-O drafted the manuscript. All authors the putative donor and mobilized sequences (Figure 4), and discussed the findings and interpreted the results. the lack of suitable models for sequence evolution in mobile elements make it difficult to estimate precisely the timing of this mobilization using sequence analysis methods. The relatively FUNDING low nucleotide sequence identity between the Leptospiraceae and Rhodobiaceae chromosomal folP∗ genes and sul genes (Table 1) This work was supported by grant BIO2016-77011-R from suggests that folP∗ genes were mobilized and diversified long the Spanish Ministerio de Economia y Competitividad to JB. Frontiers in Microbiology | www.frontiersin.org 12 January 2019 | Volume 9 | Article 3332 Sánchez-Osuna et al. Origin of Clinical Sulfonamide Resistance MS-O was the recipient of a predoctoral fellowship from the A preprint version of this manuscript prior to peer review is Ministerio de Educación, Cultura y Deporte de España. available on bioRxiv (https://doi.org/10.1101/472472). ACKNOWLEDGMENTS SUPPLEMENTARY MATERIAL The authors wish to thank Joan Ruiz and Susana Escribano The Supplementary Material for this article can be found for their technical support during some of the experimental online at: https://www.frontiersin.org/articles/10.3389/fmicb. procedures, as well as Júlia López for her continued support. 2018.03332/full#supplementary-material REFERENCES Clinical and Laboratory Standards Institute (2003). Methods for Dilution Antimicrobial Susceptibility Tests for Bacteria that Grow Aerobically –Approved Achari, A., Somers, D. O., Champness, J. N., Bryant, P. K., Rosemond, J., and Standard, 6 Edn. Wayne, PA: Clinical and Laboratory Standards Institute. Stammers, D. K. (1997). Crystal structure of the anti-bacterial sulfonamide Colaert, N., Helsens, K., Martens, L., Vandekerckhove, J., and Gevaert, K. (2009). drug target dihydropteroate synthase. Nat. Struct. Biol. 4, 490–497. doi: 10.1038/ Improved visualization of protein consensus sequences by iceLogo. Nat. nsb0697-490 Methods 6, 786–787. doi: 10.1038/nmeth1109-786 Altenhoff, A. M., Glover, N. M., Train, C.-M., Kaleb, K., Warwick Vesztrocy, A., Crofts, T. S., Gasparrini, A. J., and Dantas, G. (2017). Next-generation approaches Dylus, D., et al. (2018). The OMA orthology database in 2018: retrieving to understand and combat the antibiotic resistome. Nat. Rev. Microbiol. 15, evolutionary relationships among all domains of life through richer web and 422–434. doi: 10.1038/nrmicro.2017.28 programmatic interfaces. Nucleic Acids Res. 46, D477–D485. doi: 10.1093/nar/ Davenport, D. (2012). The war against bacteria: how were sulphonamide drugs gkx1019 used by Britain during World War II? Med. Human. 38, 55–58. doi: 10.1136/ Altschul, S. F., Madden, T. L., Schaffer, A. A., Zhang, J., Zhang, Z., Miller, W., et al. medhum-2011-010024 (1997). Gapped BLAST and PSI-BLAST: a new generation of protein database Davies, J. (2007). Microbes have the last word. A drastic re-evaluation of search programs. Nucleic Acids Res. 25, 3389–3402. doi: 10.1093/nar/25.17.3389 antimicrobial treatment is needed to overcome the threat of antibiotic-resistant Aminov, R. I. (2010). A brief history of the antibiotic era: lessons learned and bacteria. EMBO Rep. 8, 616–621. doi: 10.1038/sj.embor.7401022 challenges for the future. Front. Microbiol. 1:134. doi: 10.3389/fmicb.2010.00134 Davies, J., and Davies, D. (2010). Origins and evolution of antibiotic resistance. Aminov, R. I., and Mackie, R. I. (2007). Evolution and ecology of antibiotic Microbiol. Mol. Biol. Rev. 74, 417–433. doi: 10.1128/MMBR.00016-10 resistance genes. FEMS Microbiol. Lett. 271, 147–161. doi: 10.1111/j.1574-6968. D’Costa, V. M., King, C. E., Kalan, L., Morar, M., Sung, W. W. L., Schwarz, C., 2007.00757.x et al. (2011). Antibiotic resistance is ancient. Nature 477:457. doi: 10.1038/ Andersson, D. I. (2006). The biological cost of mutational antibiotic resistance: any nature10388 practical conclusions? Curr. Opin. Microbiol. 9, 461–465. doi: 10.1016/j.mib. Dereeper, A., Guignon, V., Blanc, G., Audic, S., Buffet, S., Chevenet, F., et al. (2008). 2006.07.002 Phylogeny.fr: robust phylogenetic analysis for the non-specialist. Nucleic Acids Baca, A. M., Sirawaraporn, R., Turley, S., Sirawaraporn, W., and Hol, W. G. (2000). Res. 36, W465–W469. doi: 10.1093/nar/gkn180 Crystal structure of Mycobacterium tuberculosis 7,8-dihydropteroate synthase Domagk, G. (1935). Ein beitrag zur chemotherapie der bakteriellen infektionen. in complex with pterin monophosphate: new insight into the enzymatic Dtsch. Med. Wochenschr. 61, 250–253. doi: 10.1055/s-0028-1129486 mechanism and sulfa-drug action. J. Mol. Biol. 302, 1193–1212. doi: 10.1006/ Enne, V. I., Bennett, P. M., Livermore, D. M., and Hall, L. M. C. (2004). jmbi.2000.4094 Enhancement of host fitness by the sul2-coding plasmid p9123 in the absence Battistuzzi, F. U., Feijao, A., and Hedges, S. B. (2004). A genomic timescale of of selective pressure. J. Antimicrob. Chemother. 53, 958–963. doi: 10.1093/jac/ prokaryote evolution: insights into the origin of methanogenesis, phototrophy, dkh217 and the colonization of land. BMC Evol. Biol. 4:44. doi: 10.1186/1471-2148-4-44 Enne, V. I., Livermore, D. M., Stephens, P., and Hall, L. M. (2001). Persistence Baunach, M., Ding, L., Willing, K., and Hertweck, C. (2015). Bacterial synthesis of of sulphonamide resistance in Escherichia coli in the UK despite national unusual sulfonamide and sulfone antibiotics by flavoenzyme-mediated sulfur prescribing restriction. Lancet 357, 1325–1328. doi: 10.1016/S0140-6736(00) dioxide capture. Angew. Chem. Int. Ed Engl. 54, 13279–13283. doi: 10.1002/anie. 04519-0 201506541 Fan, H., Brunham, R. C., and McClarty, G. (1992). Acquisition and synthesis of Bhullar, K., Waglechner, N., Pawlowski, A., Koteva, K., Banks, E. D., Johnston, folates by obligate intracellular bacteria of the genus Chlamydia. J. Clin. Invest. M. D., et al. (2012). Antibiotic resistance is prevalent in an isolated cave 90, 1803–1811. doi: 10.1172/JCI116055 microbiome. PLoS One 7:e34953. doi: 10.1371/journal.pone.0034953 Fermer, C., Kristiansen, B. E., Sköld, O., and Swedberg, G. (1995). Sulfonamide Cantón, R. (2009). Antibiotic resistance genes from the environment: a perspective resistance in Neisseria meningitidis as defined by site-directed mutagenesis through newly identified antibiotic resistance mechanisms in the clinical could have its origin in other species. J. Bacteriol. 177, 4669–4675. doi: 10.1128/ setting. Clin. Microbiol. Infect. Off. Publ. Eur. Soc. Clin. Microbiol. Infect. Dis. jb.177.16.4669-4675.1995 15(Suppl. 1), 20–25. doi: 10.1111/j.1469-0691.2008.02679.x Franco, D. A., Webb, J., and Taylor, C. E. (1990). Antibiotic and sulfonamide Carlet, J., Rambaud, C., and Pulcini, C. (2014). Save antibiotics: a call for action residues in meat: implications for human health. J. Food Prot. 53, 178–185. of the world alliance against antibiotic resistance (WAAAR). BMC Infect. Dis. doi: 10.4315/0362-028X-53.2.178 14:436. doi: 10.1186/1471-2334-14-436 Gould, S. J., and Vrba, E. S. (1982). Exaptation—a missing term in the science of Castresana, J. (2000). Selection of conserved blocks from multiple alignments for form. Paleobiology 8, 4–15. doi: 10.1017/S0094837300004310 their use in phylogenetic analysis. Mol. Biol. Evol. 17, 540–552. doi: 10.1093/ Haasum, Y., Ström, K., Wehelie, R., Luna, V., Roberts, M. C., Maskell, J. P., et al. oxfordjournals.molbev.a026334 (2001). Amino acid repetitions in the dihydropteroate synthase of Streptococcus Chakraborty, A., Miyahara, S., Villanueva, S. Y. A. M., Gloriani, N. G., and pneumoniae lead to sulfonamide resistance with limited effects on substrateKm. Yoshida, S. (2010). In vitro sensitivity and resistance of 46 leptospira strains Antimicrob. Agents Chemother. 45, 805–809. doi: 10.1128/AAC.45.3.805-809. isolated from rats in the philippines to 14 antimicrobial agents. Antimicrob. 2001 Agents Chemother. 54, 5403–5405. doi: 10.1128/AAC.00973-10 Hammerschlag, M. R. (1982). Activity of trimethoprim-sulfamethoxazole against Chakraborty, A., Miyahara, S., Villanueva, S. Y. A. M., Saito, M., Gloriani, N. G., Chlamydia trachomatis in vitro. Rev. Infect. Dis. 4, 500–505. doi: 10.1093/ and Yoshida, S. (2011). A novel combination of selective agents for isolation of clinids/4.2.500 Leptospira species. Microbiol. Immunol. 55, 494–501. doi: 10.1111/j.1348-0421. Heddle, J. G., Blance, S. J., Zamble, D. B., Hollfelder, F., Miller, D. A., Wentzell, 2011.00347.x L. M., et al. (2001). The antibiotic microcin B17 is a DNA gyrase poison: Frontiers in Microbiology | www.frontiersin.org 13 January 2019 | Volume 9 | Article 3332 Sánchez-Osuna et al. Origin of Clinical Sulfonamide Resistance characterisation of the mode of inhibition. J. Mol. Biol. 307, 1223–1234. doi: Nordmann, P., and Poirel, L. (2005). Emergence of plasmid-mediated resistance 10.1006/jmbi.2001.4562 to quinolones in Enterobacteriaceae. J. Antimicrob. Chemother. 56, 463–469. Hegde, S. S., Vetting, M. W., Roderick, S. L., Mitchenall, L. A., Maxwell, A., Takiff, doi: 10.1093/jac/dki245 H. E., et al. (2005). A fluoroquinolone resistance protein from Mycobacterium Notredame, C., Higgins, D. G., and Heringa, J. (2000). T-Coffee: a novel method tuberculosis that mimics DNA. Science 308, 1480–1483. doi: 10.1126/science. for fast and accurate multiple sequence alignment. J. Mol. Biol. 302, 205–217. 1110699 doi: 10.1006/jmbi.2000.4042 Honda, Y., Sakai, H., Hiasa, H., Tanaka, K., Komano, T., and Bagdasarian, M. Orman, B. E., Piñeiro, S. A., Arduino, S., Galas, M., Melano, R., Caffer, M. I., (1991). Functional division and reconstruction of a plasmid replication origin: et al. (2002). Evolution of multiresistance in nontyphoid salmonella serovars molecular dissection of the oriV of the broad-host-range plasmid RSF1010. from 1984 to 1998 in Argentina. Antimicrob. Agents Chemother. 46, 3963–3970. Proc. Natl. Acad. Sci. U.S.A. 88, 179–183. doi: 10.1073/pnas.88.1.179 doi: 10.1128/AAC.46.12.3963-3970.2002 Hooper, D. C. (1999). Mechanisms of fluoroquinolone resistance. Drug Resist. Perreten, V., and Boerlin, P. (2003). A new sulfonamide resistance gene (sul3) in Update 2, 38–55. doi: 10.1054/drup.1998.0068 Escherichia coli is widespread in the pig population of Switzerland. Antimicrob. Hooper, D. C., and Jacoby, G. A. (2016). Topoisomerase inhibitors: Agents Chemother. 47, 1169–1172. doi: 10.1128/AAC.47.3.1169-1172. fluoroquinolone mechanisms of action and resistance. Cold Spring Harb. 2003 Perspect. Med. 6:a025320. doi: 10.1101/cshperspect.a025320 Perron, G. G., Whyte, L., Turnbaugh, P. J., Goordial, J., Hanage, W. P., Hu, L. F., Xu, X. H., Yang, H. F., Ye, Y., and Li, J. B. (2016). Role of sul2 gene Dantas, G., et al. (2015). Functional characterization of bacteria isolated linked to transposase in resistance to trimethoprim/sulfamethoxazole among from ancient arctic soil exposes diverse resistance mechanisms to Stenotrophomonas maltophilia isolates. Ann. Lab. Med. 36, 73–75. doi: 10.3343/ modern antibiotics. PLoS One 10:e0069533. doi: 10.1371/journal.pone.006 alm.2016.36.1.73 9533 Huovinen, P., Sundström, L., Swedberg, G., and Sköld, O. (1995). Trimethoprim Petkowski, J. J., Bains, W., and Seager, S. (2018). Natural products containing a and sulfonamide resistance. Antimicrob. Agents Chemother. 39, 279–289. doi: Nitrogen–Sulfur bond. J. Nat. Prod. 81, 423–446. doi: 10.1021/acs.jnatprod. 10.1128/AAC.39.2.279 7b00921 Kannan, L., Li, H., Rubinstein, B., and Mushegian, A. (2013). Models of gene Rådström, P., Fermér, C., Kristiansen, B. E., Jenkins, A., Sköld, O., and Swedberg, G. gain and gene loss for probabilistic reconstruction of gene content in the last (1992). Transformational exchanges in the dihydropteroate synthase gene of universal common ancestor of life. Biol. Direct 8:32. doi: 10.1186/1745-615 Neisseria meningitidis: a novel mechanism for acquisition of sulfonamide 0-8-32 resistance. J. Bacteriol. 174, 6386–6393. doi: 10.1128/jb.174.20.6386-6393. Kehrenberg, C., and Schwarz, S. (2005). dfrA20, A novel trimethoprim resistance 1992 gene from Pasteurella multocida. Antimicrob. Agents Chemother. 49, 414–417. Rådström, P., and Swedberg, G. (1988). RSF1010 and a conjugative plasmid contain doi: 10.1128/AAC.49.1.414-417.2005 sulII, one of two known genes for plasmid-borne sulfonamide resistance Kim, H. B., Park, C. H., Gavin, M., Jacoby, G. A., and Hooper, D. C. (2011). dihydropteroate synthase. Antimicrob. Agents Chemother. 32, 1684–1692. doi: Cold shock induces qnrA expression in shewanella algae. Antimicrob. Agents 10.1128/AAC.32.11.1684 Chemother. 55, 414–416. doi: 10.1128/AAC.00991-10 Rådström, P., Swedberg, G., and Sköld, O. (1991). Genetic analyses of sulfonamide Kumar, K., Gupta, S. C., Chander, Y., and Singh, A. K. (2005). Antibiotic use in resistance and its dissemination in gram-negative bacteria illustrate new aspects agriculture and its impact on the terrestrial environment. Adv. Agron. 87, 1–54. of R plasmid evolution. Antimicrob. Agents Chemother. 35, 1840–1848. doi: doi: 10.1016/S0065-2113(05)87001-4 10.1128/AAC.35.9.1840 Kushner, S. R. (1978). “An improved method for transformation of. Escherichia coli Razavi, M., Marathe, N. P., Gillings, M. R., Flach, C.-F., Kristiansson, E., with ColE 1 derived plasmids,” in Genetic Engineering, eds H. W. Boyer and S. and Joakim Larsson, D. G. (2017). Discovery of the fourth mobile Nicosia (Amsterdam: Elsevier), 17–23. sulfonamide resistance gene. Microbiome 5:160. doi: 10.1186/s40168-017- Landy, M., Larkum, N. W., Oswald, E. J., and Streightoff, F. (1943). Increased 0379-y synthesis of p-aminobenzoic acid associated with the development of Ronquist, F., and Huelsenbeck, J. P. (2003). MrBayes 3: bayesian phylogenetic sulfonamide resistance in Staphylococcus aureus. Science 97, 265–267. doi: 10. inference under mixed models. Bioinformatics 19, 1572–1574. doi: 10.1093/ 1126/science.97.2516.265 bioinformatics/btg180 Lawrence, J. G., and Ochman, H. (1997). Amelioration of bacterial genomes: Rossolini, G. M., Arena, F., Pecile, P., and Pollini, S. (2014). Update on the rates of change and exchange. J. Mol. Evol. 44, 383–397. doi: 10.1007/PL0000 antibiotic resistance crisis. Curr. Opin. Pharmacol. 18, 56–60. doi: 10.1016/j. 6158 coph.2014.09.006 Lévesque, C., Piché, L., Larose, C., and Roy, P. H. (1995). PCR mapping of integrons Sandoz, K. M., and Rockey, D. D. (2010). Antibiotic resistance in reveals several novel combinations of resistance genes. Antimicrob. Agents chlamydiae. Future Microbiol. 5, 1427–1442. doi: 10.2217/fmb. Chemother. 39, 185–191. doi: 10.1128/AAC.39.1.185 10.96 Marti, H., Borel, N., Dean, D., and Leonard, C. A. (2018). Evaluating the antibiotic Sköld, O. (2000). Sulfonamide resistance: mechanisms and trends. Drug Resist. susceptibility of chlamydia – New approaches for in vitro assays. Front. Updat. 3, 155–160. doi: 10.1054/drup.2000.0146 Microbiol. 9:1414. doi: 10.3389/fmicb.2018.01414 Stokes, H. W., and Gillings, M. R. (2011). Gene flow, mobile genetic elements Mayola, A., Irazoki, O., Martínez, I. A., Petrov, D., Menolascina, F., Stocker, R., and the recruitment of antibiotic resistance genes into Gram-negative et al. (2014). RecA protein plays a role in the chemotactic response and pathogens. FEMS Microbiol. Rev. 35, 790–819. doi: 10.1111/j.1574-6976.2011.00 chemoreceptor clustering of Salmonella enterica. PLoS One 9:e105578. doi: 273.x 10.1371/journal.pone.0105578 Suyama, M., Torrents, D., and Bork, P. (2006). PAL2NAL: robust Mehra-Chaudhary, R., Mick, J., and Beamer, L. J. (2011). Crystal structure conversion of protein sequence alignments into the corresponding of Bacillus anthracis phosphoglucosamine mutase, an enzyme in the codon alignments. Nucleic Acids Res. 34, W609–W612. doi: 10.1093/nar/ peptidoglycan biosynthetic pathway? J. Bacteriol. 193, 4081–4087. doi: 10.1128/ gkl315 JB.00418-11 Swedberg, G., Ringertz, S., and Sköld, O. (1998). Sulfonamide resistance Mitsuhashi, S. (1993). Drug resistance in bacteria: history, genetics and in streptococcus pyogenes is associated with differences in the biochemistry. J. Int. Med. Res. 21, 1–14. doi: 10.1177/030006059302100101 amino acid sequence of its chromosomal dihydropteroate synthase. Morgan, R. E., Batot, G. O., Dement, J. M., Rao, V. A., Eadsforth, T. C., and Hunter, Antimicrob. Agents Chemother. 42, 1062–1067. doi: 10.1128/AAC.42.5. W. N. (2011). Crystal structures of Burkholderia cenocepacia dihydropteroate 1062 synthase in the apo-form and complexed with the product 7,8-dihydropteroate. Thompson, J. D., Higgins, D. G., and Gibson, T. J. (1994). CLUSTAL BMC Struct. Biol. 11:21. doi: 10.1186/1472-6807-11-21 W: improving the sensitivity of progressive multiple sequence alignment Nei, M., and Gojobori, T. (1986). Simple methods for estimating the numbers of through sequence weighting, position-specific gap penalties and weight synonymous and nonsynonymous nucleotide substitutions. Mol. Biol. Evol. 3, matrix choice. Nucleic Acids Res. 22, 4673–4680. doi: 10.1093/nar/22.22. 418–426. doi: 10.1093/oxfordjournals.molbev.a040410 4673 Frontiers in Microbiology | www.frontiersin.org 14 January 2019 | Volume 9 | Article 3332 Sánchez-Osuna et al. Origin of Clinical Sulfonamide Resistance Toleman, M. A., Bennett, P. M., and Walsh, T. R. (2006). ISCR elements: novel antibiotic-resistant Pseudomonas aeruginosa. mBio 6:e01796-15. doi: 10.1128/ gene-capturing systems of the 21st century? Microbiol. Mol. Biol. Rev. 70, mBio.01796-15 296–316. doi: 10.1128/MMBR.00048-05 Wuthiekanun, V., Amornchai, P., Langla, S., White, N. J., Day, N. P. J., Toleman, M. A., and Walsh, T. R. (2011). Combinatorial events of insertion Limmathurotsakul, D., et al. (2015). Antimicrobial disk susceptibility sequences and ICE in Gram-negative bacteria. FEMS Microbiol. Rev. 35, 912– testing of Leptospira spp. using leptospira vanaporn wuthiekanun 935. doi: 10.1111/j.1574-6976.2011.00294.x (LVW). Agar Am. J. Trop. Med. Hyg. 93, 241–243. doi: 10.4269/ajtmh.15- Tran, J. H., and Jacoby, G. A. (2002). Mechanism of plasmid-mediated quinolone 0180 resistance. Proc. Natl. Acad. Sci. U.S.A. 99, 5638–5642. doi: 10.1073/pnas. 082092899 Conflict of Interest Statement: The authors declare that the research was Tran, J. H., Jacoby, G. A., and Hooper, D. C. (2005). Interaction of the plasmid- conducted in the absence of any commercial or financial relationships that could encoded quinolone resistance protein Qnr with Escherichia coli DNA gyrase. be construed as a potential conflict of interest. Antimicrob. Agents Chemother. 49, 118–125. doi: 10.1128/AAC.49.1.118-125. 2005 Copyright © 2019 Sánchez-Osuna, Cortés, Barbé and Erill. This is an open-access Treeck, U., van Schmidt, F., and Wiedemann, B. (1981). Molecular nature of a article distributed under the terms of the Creative Commons Attribution License streptomycin and sulfonamide resistance plasmid (pBP1) prevalent in clinical (CC BY). The use, distribution or reproduction in other forums is permitted, provided Escherichia coli strains and integration of an ampicillin resistance transposon the original author(s) and the copyright owner(s) are credited and that the original (TnA). Antimicrob. Agents Chemother. 19, 371–380. doi: 10.1128/AAC.19.3.371 publication in this journal is cited, in accordance with accepted academic practice. van Belkum, A., Soriaga, L. B., LaFave, M. C., Akella, S., Veyrieras, J.-B., Barbu, No use, distribution or reproduction is permitted which does not comply with these E. M., et al. (2015). Phylogenetic distribution of CRISPR-cas systems in terms. Frontiers in Microbiology | www.frontiersin.org 15 January 2019 | Volume 9 | Article 3332