https://orcid.org/0000-0002-8905-2055
Figures
Abstract
The prevailing view in bacterial pathogenesis is that antimicrobial resistance and virulence are constrained by evolutionary trade-offs, with resistance mechanisms imposing fitness costs that attenuate pathogenic potential. Herein we document that contemporary Vibrio cholerae clinical isolates from the ongoing seventh pandemic have circumvented this paradigm by coupling multidrug resistance with hypervirulence. We examined five geographically diverse Wave 3 isolates collected between 2017 and 2019 and compared them to early pandemic strains. These contemporary isolates exhibited both broad-spectrum antimicrobial resistance and markedly enhanced colonization capacity in the infant mouse model. Phylogenetic analysis of 67 O1 El Tor genomes spanning 1960–2019 confirmed that the isolates cluster within a representative Wave 3 sublineage. We identified the VexB RND efflux pump as a mediator of this coupled phenotype. Elevated vexB expression in the contemporary isolates conferred resistance to multiple antibiotic classes, while vexB inactivation simultaneously impaired resistance and colonization. This dual function was not observed in early pandemic strains, consistent with a recent evolutionary adaptation. VexB-mediated hypervirulence occurred through multiple pathways independent of cholera toxin and toxin-coregulated pilus production levels. VexB deletion impaired bacterial adherence to intestinal epithelial cells, impaired motility, and increased susceptibility to membrane-active antimicrobials. In contrast, laboratory evolution under antibiotic pressure alone generated resistant but avirulent strains, demonstrating that complex selective forces in nature enabled the co-optimization of resistance and virulence. These findings establish VexB as a molecular link between antimicrobial resistance and hypervirulence in pandemic V. cholerae, highlighting efflux pumps as dual-function therapeutic targets whose inhibition could both restore antibiotic activity and attenuate disease.
Author summary
Antimicrobial resistance is often assumed to come at the cost of virulence, with resistant bacteria paying a fitness price that limits their ability to cause disease. Our work shows that this assumption does not hold for contemporary Vibrio cholerae, the pathogen responsible for cholera. We studied recent clinical isolates from the ongoing pandemic and found that they are both highly multidrug-resistant and unusually virulent, causing enhanced intestinal colonization compared to earlier pandemic strains. We identified the VexB efflux pump as the key factor linking these traits. In contemporary strains, mutating vexB simultaneously reduced antibiotic resistance and impaired colonization by decreasing bacterial adherence to intestinal cells, reducing motility, and increasing susceptibility to gut antimicrobials. This indicates that VexB has acquired new functions that allow modern isolates to optimize both traits at once. Laboratory evolution under antibiotic pressure produced resistant but weakened strains, showing that additional natural pressures are needed to drive this dual adaptation. These findings highlight efflux pumps as promising therapeutic targets that could both restore antibiotic activity and lessen disease severity.
Citation: Weng Y, Bina XR, Van Allen ME, Bina JE (2026) RND-mediated efflux couples antimicrobial resistance and hypervirulence in contemporary Vibrio cholerae. PLoS Pathog 22(3): e1014031. https://doi.org/10.1371/journal.ppat.1014031
Editor: Jun Zhu, University of Pennsylvania Perelman School of Medicine, UNITED STATES OF AMERICA
Received: January 26, 2026; Accepted: February 23, 2026; Published: March 9, 2026
Copyright: © 2026 Weng et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: Next generation sequencing file are available from the NCBI Sequence Read Archive under BioProject PRJNA1284390. All other relevant data are within the manuscript and its Supporting Information files.
Funding: This work was supported by the National Institute of Allergy and Infectious Diseases of the National Institutes of Health (NIH) under awards R01AI132460 and R21AI166889 (to JEB). MEV was supported by NIH training grant T32AI060525. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exists.
Introduction
Antimicrobial resistance (AMR) continues to pose a significant challenge to global health, with the expansion of multidrug-resistant (MDR) bacterial infections raising concerns about the future efficacy of existing therapeutics [1,2]. The emergence and spread of MDR pathogens have led to increased morbidity, mortality, and healthcare burdens worldwide. However, a significant knowledge gap remains in understanding the molecular mechanisms by which high-level antimicrobial resistance evolves and how it influences bacterial virulence [3,4].
A long-standing principle in bacterial pathogenesis has been that antimicrobial resistance and virulence involve evolutionary trade-offs, where enhanced resistance usually comes at the cost of pathogenic potential due to fitness burdens [5–8]. Resistance determinants are thought to impose a metabolic load and disrupt homeostasis, particularly in the absence of antimicrobial pressure, thereby attenuating a pathogen’s ability to cause disease. However, multiple recent studies suggest that this principle does not always hold true, demonstrating that resistance and virulence are not always antagonistic and that context-dependent exceptions exist [9–12]. Clinical isolates that maintain, or amplify virulence while acquiring multidrug resistance underscore the complexity of the relationship between resistance and virulence. This emerging phenomenon challenges the widely accepted trade-off paradigm by exposing its limitations as an overly simplified predictive model.
Understanding the molecular mechanisms by which high-level antimicrobial resistance evolves requires examining both acquired and intrinsic resistance systems. The evolution of MDR in bacteria is driven by multiple mechanisms including antibiotic target site mutations and the acquisition of resistance genes by horizontal gene transfer (HGT). While HGT accelerates the dissemination of resistance determinants across bacterial populations [13,14], the evolution of MDR pathogens often involves the activation of intrinsic (i.e., innate) resistance mechanisms [15]. These mechanisms include reduced membrane permeability, enzymatic inactivation of antibiotics, and active efflux systems. Among Gram-negative bacteria, the Resistance-Nodulation-Division (RND) family of efflux pumps represents a particularly important intrinsic resistance mechanism [16]. These tripartite, proton-motive force-driven transporters often exhibit broad substrate specificity allowing them to export a remarkably diverse array of antimicrobial compounds, providing protection against multiple classes of antibiotics simultaneously.
Numerous studies across diverse bacterial pathogens have demonstrated that RND efflux systems function beyond antimicrobial resistance, modulating adaptive responses that affect a broad range of virulence-associated phenotypes [17]. Consistent with this, genetic or chemical impairment of RND-mediated efflux frequently results in attenuated virulence [18–25]. This dual role of the RND systems in antimicrobial resistance and pathogenesis raises the possibility that genetic adaptations enhancing RND-mediated efflux could simultaneously influence pathogenic potential. This dual role of RND systems suggests that antimicrobial pressure may inadvertently select for increased virulence and that RND pumps may serve as molecular links coupling multidrug resistance with pathogenic traits, potentially circumventing classical resistance-virulence trade-offs. However, the relationship between RND-mediated efflux, multidrug resistance, and virulence remains poorly understood in most pathogens, including the Gram-negative enteric pathogen V. cholerae.
V. cholerae is the causative agent of cholera, an acute diarrheal disease that has been around since antiquity. Cholera continues to be a global health burden causing an estimated 2.8 million cholera cases and 91,000 deaths annually [26,27]. There have been seven recorded global cholera pandemics since the early 1800’s [28]. The first six pandemics were caused by V. cholerae strains belonging to the O1 Classical biotype. The ongoing seventh cholera pandemic, which began in the early 1960s with the emergence of the O1 El Tor biotype, has been driven by a clonal lineage that originated in the Bay of Bengal region and subsequently spread worldwide through three successive waves [29]. Early pandemic isolates (Wave 1) were largely antibiotic-susceptible, but contemporary Wave 3 strains are extensively MDR which has been linked to the acquisition of mobile genetic elements carrying resistance determinants and antibiotic target site mutations [30]. Recent genomic studies have demonstrated that Wave 3 strains evolved hypervirulence within this clonal framework, contributing to enhanced colonization, elevated virulence factor production and increasingly severe clinical outcomes in affected populations [31–33], contradicting traditional predictions that highly resistant strains should exhibit attenuated virulence.
The V. cholerae core genome encodes six independent RND efflux systems: VexAB, VexCD, VexEF, VexGH, VexIJK, and VexLM, each encoding their cognate periplasmic adapter protein (vexIJK encodes two adapter proteins) and integral inner membrane transporter (i.e., VexBDFHKM) [34]. The six RND efflux systems are distributed across both chromosomes: VexAB, VexCD, VexEF, VexGH, and VexIJK are encoded on the large chromosome (Chr I), while VexLM is encoded on the small chromosome (Chr II). All of the transporters share TolC as their outer membrane pore protein [35], which is encoded independently on the large chromosome. Studies with the archetypical Wave 1 strain N16961 established that the RND transporters contribute to both intrinsic antimicrobial resistance and virulence [34]. When RND-mediated efflux was impaired through chemical inhibition or mutation [36,37], N16961 showed hypersensitivity to multiple antimicrobials and reduced production of the critical virulence factors cholera toxin (CT) and the toxin co-regulated pilus (TCP). These observations raised a critical question: how might RND efflux systems mechanistically link antimicrobial resistance to virulence factor regulation?
Subsequent studies have proposed a novel mechanism to explain this link: RND transporters function as critical mediators in a metabolite feedback circuit that coordinates V. cholerae adaptive responses to environmental challenges (reviewed in [22]). In this model, RND transporters continuously remove endogenous metabolites and exogenous compounds from the periplasmic space, maintaining low periplasmic concentrations of these molecules under normal conditions. When RND efflux activity is impaired, through mutation, chemical inhibition, energy depletion, or competitive substrate inhibition, these metabolites accumulate in the periplasm. There, they are sensed by periplasmic sensor proteins, initiating signal transduction cascades that modulate expression of genes involved in virulence, stress response, and cellular homeostasis. Thus, RND systems serve a dual role, providing direct protection through antimicrobial efflux while simultaneously functioning as regulators of adaptive gene expression by modulating the availability of signaling metabolites to periplasmic sensors. This feedback mechanism may allow V. cholerae to coordinately sense and respond to complex environmental conditions encountered during host colonization and environmental survival. Despite these findings, the role of RND efflux systems in the evolution of contemporary Wave 3 MDR strains remains unknown.
In this study, we examined the contributions of RND efflux systems to antimicrobial resistance and virulence in contemporary V. cholerae clinical isolates, as compared to early pandemic strains. Using genetic, phenotypic, and comparative analyses, we investigated whether RND-mediated efflux contributed to both multidrug resistance and increased pathogenic potential exhibited by contemporary pandemic strains. Our results reveal that VexB functions as a molecular link coupling resistance and virulence in Wave 3 strains through mechanisms that include both direct antimicrobial efflux and modulation of adaptive responses via metabolite-dependent feedback signaling. Our findings demonstrate that antimicrobial resistance and hypervirulence can be mechanistically coupled through adaptations effecting RND-mediated efflux, revealing an underappreciated pathway by which antibiotic pressure may drive the evolution of more dangerous pathogens.
Results
Contemporary V. cholerae isolates exhibit coupled resistance and hypervirulence
To investigate whether contemporary Wave 3 isolates differ from early pandemic strains in their pathogenic properties, we obtained five recent V. cholerae O1 El Tor clinical isolates collected between 2017 and 2019 from travelers returning to the United States from cholera-endemic regions (S1 Table). These isolates represent geographically diverse Wave 3 strains and provided an opportunity to assess both antimicrobial resistance profiles and virulence capacity in contemporary pandemic isolates.
Antimicrobial susceptibility testing revealed that all five Wave 3 isolates exhibited increased resistance, as measured by an increase in minimum inhibitory concentration (MIC), to multiple antibiotics compared to the early pandemic strain C6706 (Table 1). This included elevated resistance to ciprofloxacin (50–100-fold increase) and tetracycline (4–20-fold increase), two antibiotics commonly used for treating enteric infections. Additionally, four isolates showed increased resistance to erythromycin (2–4-fold) and chloramphenicol (4–10-fold), while all five exhibited increased ampicillin resistance (3–4-fold) and modest increases in kanamycin resistance (2-fold).
All of the clinical isolates displayed reduced polymyxin B resistance (approximately 4-fold decrease in MIC) compared to C6706, consistent with reports of diminished polymyxin B resistance in recent Wave 3 strains [38–41]. This phenotype represents a significant phenotypic shift, as it emerged around 2012 in Kolkata, India, and marks a departure from the polymyxin resistance that defined the El Tor biotype since the start of the seventh pandemic [38,41]. This phenotype correlated with significantly reduced expression of almE and carR genes that are involved in lipopolysaccharide modification and high level polymyxin B resistance (S1 Fig), suggesting that alterations in this membrane remodeling system are responsible for the attenuated antimicrobial peptide resistance [42–44].
Traditional antimicrobial resistance-virulence trade-off models would predict that highly resistant strains might exhibit attenuated virulence due to fitness costs associated with enhanced antimicrobial resistance mechanisms [5–8]. However, competitive colonization assays in the infant mouse model revealed that all five Wave 3 isolates demonstrated significantly enhanced virulence compared to the early pandemic strain C6706 (Fig 1A). The contemporary isolates outcompeted C6706 by approximately 30-fold (strain 2017) to more than 100-fold (strains 2018–1, 2018–2, 2018–3, and 2019), indicating a hypervirulent phenotype. Growth curves in LB medium showed similar kinetics for all strains, indicating that colonization differences reflected host-specific adaptation rather than general fitness defects.
(A) Competitive indices (CI) of contemporary clinical isolates (purple circles) and their isogenic ΔvexB mutants (green open circles) against the Wave 1 reference strain C6706 in the infant mouse model. The CI for the C6706 ΔvexB mutant against its wild-type parent is shown (blue circles). Each circle represents one mouse; horizontal bars indicate geometric means. A CI value below 1.0 (red line) indicates a colonization defect. (B) TcpA immunoblot and (C) cholera toxin (CT) quantification of indicated strains after growth under virulence-inducing conditions. The immunoblot is representative of three independent experiments. CT data are presented as the mean ± SD from three independent experiments. Statistical significance was determined by the Mann-Whitney U test (A) and an unpaired, two-tailed Student’s t-test (C), comparing each ΔvexB mutant to its respective parental strain. *, P < 0.05; **, P < 0.01; ***, P < 0.001.
Surprisingly, enhanced colonization capacity was independent of classical virulence factor production levels. Quantification of CT and TcpA production (TcpA is the pilin subunit of the TCP) under virulence-inducing conditions revealed no correlation between production of these classical virulence effectors and colonization efficiency (Fig 1B and 1C). Strains 2017 and 2019 produced significantly higher levels of CT and TcpA compared to strains 2018–1, 2018–2, and 2018–3, yet exhibited lower competitive indices than these strains (Fig 1A). Furthermore, the 2018 strains produced virulence factors at levels similar to the reference strain C6706 yet still demonstrated a hypervirulence phenotype.
The ToxR regulon is a hierarchical regulatory system that regulates CT and TCP production in response to environmental stimuli (reviewed in [45]). ToxR is a membrane-associated transcription factor with a periplasmic sensing domain that works in concert with TcpP to activate toxT expression [46,47]. ToxT then activates the expression of ctxAB (encoding CT) and tcpA-F (encoding TCP) [48]. Upstream regulators AphA and AphB control tcpP expression and are themselves responsive to environmental cues [49,50]: AphB in response to oxygen [51] while aphA is regulated by leuO and quorum sensing [52]. Quantifying the expression of aphA, aphB, ctxA, tcpA, tcpP, toxR, and toxT provides comprehensive insight into each regulatory node in the ToxR regulon. qRT-PCR analysis of the ToxR regulon in the contemporary strains confirmed the CT and TCP bioassays, showing elevated ToxR regulon expression including ctxA and tcpA expression in strains 2017 and 2019 whereas strains 2018–1, 2018–2, and 2018–3 exhibited ToxR regulon expression levels comparable to the reference strain C6706, with the exception of a minor decrease in aphA expression in strains 2018–2 and 2018–3 (Table 2).
Overall, these findings reveal an unexpected link between increased antimicrobial resistance and enhanced colonization in contemporary V. cholerae clinical isolates, challenging the traditional view that resistance comes at the cost of virulence. Significantly, the enhanced colonization occurred independent of high-level virulence factor production observed in two of the five Wave 3 strains (e.g., 2017 and 2019), suggesting that complementary virulence mechanisms may drive this phenotype.
V. cholerae’s ability to survive and persist in aquatic ecosystems contributes to both endemicity and sustained transmission during outbreaks [28]. To determine whether enhanced colonization capacity was accompanied by improved environmental survival, we performed competitive fitness assays in sterile pond water using a limited number of representative isolates (S2 Fig). In contrast to their increased fitness during host colonization, all Wave 3 isolates tested exhibited equal or reduced fitness in pond water relative to C6706, with strain 2017 showing marked attenuation (CI = 0.02). These findings suggest that the epidemiological success of Wave 3 strains was unlikely to be driven by enhanced environmental fitness; however, analysis of additional isolates will be required to determine whether this perceived lack of environmental advantage is broadly conserved across Wave 3 strains.
Phylogenetic context for Wave 3 clinical isolates
To establish a phylogenetic framework for interpreting genomic variation in our clinical isolates, we performed a maximum-likelihood phylogenetic analysis using a core genome alignment of 3,863,204 bp from 67 V. cholerae O1 El Tor genomes spanning the seventh pandemic (1960–2019) (S3 Fig). This analysis resolved three major, well-supported pandemic wave clades, with strong bootstrap support (>95%) at the primary branching nodes. All five clinical isolates (2017, 2018–1, 2018–2, 2018–3, and 2019) clustered as a coherent sublineage within the Wave 3 clade (2004–2019; red), forming a monophyletic group distinct from the Wave 1 (1960–1992; blue) and Wave 2 (1992–2004; green) pandemic waves. These isolates were separated by only 4–11 SNPs from one another and from other contemporary Wave 3 strains, indicating close phylogenetic relatedness consistent with recent pandemic transmission. This phylogenetic placement argues against the interpretation that the genomic variants identified in these isolates reflect strain-specific anomalies or assembly artifacts. Instead, their localization within a well-defined, strongly supported Wave 3 subclade suggests that these variants are representative of broader Wave 3 evolutionary features. This phylogenetic framework provides the context for subsequent comparative genomic analyses aimed at distinguishing wave-associated adaptations from strain-specific variation.
The VexB efflux pump couples antimicrobial resistance with hypervirulence in Wave 3 strains
Given the dual roles of RND efflux systems in both antimicrobial resistance and virulence [34], we investigated whether altered RND efflux activity might contribute to the coupled resistance-hypervirulence phenotype observed in Wave 3 strains. Examination of RND expression patterns in the contemporary clinical strains revealed distinct profiles compared to the early pandemic strain C6706 (Fig 2). Most notably, four of the five isolates (2017, 2018–1, 2018–3, and 2019) exhibited significantly elevated basal vexB expression compared to C6706, with the remaining isolate (2018–2) showing expression levels comparable to the reference strain. By contrast, the remaining RND pumps showed either decreased expression or no significant changes relative to C6706, with the exception of vexK, which was marginally increased in strain 2017.
Expression of RND efflux pump genes during growth in LB medium. Strains were cultured to mid-logarithmic phase when RNA was isolated for qRT-PCR analysis as described in the Methods. Data are presented as fold-change in expression for each RND gene in each clinical strain relative to the corresponding gene in C6706 and represent the mean ± standard deviation from three independent experiments. Statistical significance was determined using a one-sample t-test to assess whether the fold-change for each gene in each strain differed significantly from a theoretical ratio of 1.0. *, P < 0.05; **, P < 0.01; ***, P < 0.001.
To assess whether elevated vexB expression contributes to the enhanced antimicrobial resistance observed in Wave 3 isolates, we generated vexB deletion mutants in each clinical strain and evaluated their antimicrobial susceptibility profiles. The results showed that vexB deletion resulted in substantial reductions in resistance across multiple antibiotic classes (Table 3), with the most dramatic effects observed in isolates 2018–1, 2018–2, and 2018–3, where vexB deletion decreased ampicillin resistance by 9–15-fold and erythromycin resistance by 10-17.5-fold across isolates. Complementation studies confirmed that VexB directly mediates the observed resistance phenotypes (S2 Table).
To determine whether VexB-mediated resistance extended to the bactericidal activity of membrane-active antimicrobials that are representative of antimicrobial compounds present in the GI tract, we performed killing assays with sodium dodecyl sulfate (SDS) and polymyxin B, both of which disrupt bacterial membranes. Deletion of vexB significantly increased bacterial killing by SDS and polymyxin B across the contemporary strains (Fig 3). These data demonstrate that VexB provides protection against membrane-targeting antimicrobial compounds, which V. cholerae encounters during colonization of the intestine.
Survival of wild type (WT, black bars) and vexB deletion mutant (ΔvexB, red bars) V. cholerae strains following treatment with (A) SDS or (B) polymyxin B. Data represent mean ± SD from three independent experiments performed in triplicate. Statistical significance determined by unpaired t-test (*p < 0.05, **p < 0.01, ***p < 0.001).
We next investigated whether VexB contributes to the enhanced colonization observed in Wave 3 strains. If VexB mediates the coupling between resistance and colonization capacity, deletion should impair both phenotypes simultaneously. Infant mouse colonization assays revealed that vexB deletion impaired the colonization capacity of four of the five contemporary clinical isolates compared to wildtype C6706 retaining vexB (Fig 1A, open green circles). The most dramatic effects were observed in strains 2018–2 (>10-fold reduction) and 2019 (~10-fold reduction), while strains 2018–1 and 2018–3 showed substantial but more moderate reductions (6-fold and 4-fold, respectively). Significantly, vexB deletion in the Wave 1 pandemic strains C6706 and N16961 had no significant effect on colonization (Fig 1A, cyan circles and [34]), indicating that VexB’s contribution to colonization represents a functional adaptation specific to contemporary strains.
Assessment of CT and TcpA levels revealed strain-specific effects of VexB on virulence factor production. Across most strains, vexB deletion had minimal impact on TcpA production and showed only modest effects on CT levels (Fig 1B and 1C). qRT-PCR analysis of ToxR regulon genes corroborated these findings, showing no significant changes in ctxA or tcpA expression in most vexB deletion mutants (Table 2). However, strain 2019 exhibited a distinct phenotype: vexB deletion resulted in decreased ctxA expression (2-fold) and reduced CT production, while TcpA levels remained unchanged. Interestingly, the 2019ΔvexB mutant also showed significant upregulation of aphA (4.3-fold) and its downstream target tcpP (1.9-fold) (Table 2). The increased aphA expression in this mutant likely explains the elevated tcpP levels, as AphA positively regulates tcpP expression. However, this creates an apparent paradox as despite elevated aphA and tcpP expression, ctxA expression was reduced rather than increased. This unexpected pattern suggests that VexB may influence virulence gene expression through multiple regulatory pathways. The reduced toxR expression observed in the 2018–3ΔvexB and 2019ΔvexB mutants may contribute to altered virulence factor production. However, ToxR is also subject to post-transcriptional and translational regulation [53], including proteolytic degradation [54] and redox-dependent stability [55,56]. Whether VexB influences ToxR stability through effects on periplasmic redox state or other post-transcriptional mechanisms is unknown and warrants further investigation.
The mechanism linking VexB-mediated efflux to aphA regulation, as well as the disconnect between elevated tcpP expression and reduced ctxA, remains unclear. We also observed approximately 2-fold decreases in toxR expression in both the 2018–3ΔvexB and 2019ΔvexB mutants. Because toxR is generally constitutively expressed in rich medium, this reduction suggests that VexB may influence basal toxR transcript levels through an unknown regulatory mechanism. Given that ToxR is subject to post-transcriptional and post-translational control, including proteolytic degradation and redox-dependent stability [53–56], it is possible that VexB-mediated efflux affects virulence gene expression by altering periplasmic conditions that impact ToxR abundance or activity. However, whether VexB directly influences ToxR stability or acts through parallel regulatory pathways remains to be determined.
Taken together, these findings established VexB as the first identified molecular mechanism mediating both enhanced antimicrobial resistance and colonization capacity in contemporary V. cholerae pandemic isolates. While VexB contributes primarily to antimicrobial resistance in earlier pandemic strains without significantly influencing colonization, in current isolates it functions as a key determinant of both phenotypes.
VexB contributes to adherence to intestinal epithelial cells
Since reduced intestinal colonization can result from impaired interactions with the mucosal epithelium, we examined whether deletion of vexB affected bacterial adherence to intestinal epithelial cells. We used differentiated HT29-MTX cells, a human enterocyte-like cell line that produces a mucin-rich surface layer, providing a model for V. cholerae interactions with the intestinal epithelium during infection. Our results showed that deletion of vexB significantly reduced adherence efficiency in all five contemporary Wave 3 strains, with decreases ranging from ~40–80% relative to their respective wild-type parents (Fig 4). By contrast, no adherence defect was observed in the C6706 vexB mutant, consistent with the evolved, strain-specific role of VexB in Wave 3 isolates. Given that bacterial motility is required to access and engage epithelial surfaces, the observed adherence defect in vexB mutants likely reflects a combination of impaired motility (see below) and potential undefined alterations in bacterial surface properties that influence host cell interactions.
Adherence efficiency of V. cholerae wild type (WT, black bars) and vexB deletion mutant (ΔvexB, red bars) to HT29-MTX cells. Bacteria were incubated with confluent cell monolayers for 30 minutes at 37°C, and adherence efficiency was calculated as the percentage of input bacteria that remained associated with cells after washing. Data represent the mean ± standard deviation of three independent experiments. Statistical significance between wild type and the vexB mutant was determined by paired t-test (*p < 0.05, **p < 0.01, ***p < 0.001).
VexB is the predominant RND efflux pump in V. cholerae
To assess the relative contributions and potential redundancy among the six RND efflux systems, we generated a comprehensive panel of mutants including six strains each retaining only a single RND pump (referred to as Δ5 mutants) and a strain lacking all six RND pumps (ΔRND). The substrate specificity of each RND pump was determined by comparing the MICs of the Δ5 mutants with that of the ΔRND mutant, where increased MIC values relative to ΔRND indicated that the tested antimicrobial was a substrate for the retained pump.
The ΔRND mutant exhibited hypersensitivity to multiple chemically diverse compounds, with markedly increased susceptibility to erythromycin (32-fold), ampicillin (125-fold), ciprofloxacin (5-fold), polymyxin B (4-fold), Triton X-100 (>125-fold), cholate (250-fold), deoxycholate (500-fold), and SDS (4-fold) compared to wildtype (S3 Table). The ΔRND mutant and all Δ5 mutants retained wildtype resistance to kanamycin, tetracycline, and chloramphenicol, confirming that the observed susceptibility changes were specific to RND substrates and not due to unlinked mutations or generalized cell envelope or fitness defects.
Among the Δ5 mutants, the Δ5vexB+ strain fully phenocopied wildtype resistance levels across all tested compounds, demonstrating that VexB alone can compensate for the loss of the other five RND pumps. Notably, this mutant displayed a reproducible 2-fold increase in resistance to ampicillin and other β-lactam antibiotics compared to wildtype. While the mechanism underlying this enhanced resistance remains to be determined, we hypothesize it may result from reduced competition for TolC in the absence of other RND transporters, potentially facilitating increased formation of functional VexAB-TolC tripartite complexes. Analysis of the remaining Δ5 + mutants revealed functional redundancy among three additional RND pumps with more limited substrate specificities. The Δ5vexD+ mutant fully restored resistance to deoxycholate and partially restored cholate resistance, consistent with previous identification of VexD as a bile acid-specific efflux system [57,58]. The Δ5vexH+ mutant partially restored resistance to ampicillin and ciprofloxacin, while the Δ5vexK+ mutant partially restored resistance to Triton X-100 and cholate. In contrast, neither VexF nor VexM contributed to resistance against any tested antimicrobials, suggesting they efflux substrates not included in our panel or have dedicated non-resistance roles in V. cholerae physiology.
Investigation of RND gene expression following antimicrobial exposure revealed a coordinated regulatory strategy where relevant contributing efflux pumps are induced while non-contributing pumps are simultaneously repressed (S4 Fig), a finding that is consistent with the RND systems being regulated in response to their efflux substrates. Together, these results establish VexB as the predominant broad-spectrum RND efflux pump mediating intrinsic antimicrobial resistance in V. cholerae, while VexD, VexH, and VexK exhibit more limited substrate specificities that functionally overlap with VexB.
RND-mediated efflux potentiates horizontally acquired resistance mechanisms
To investigate potential synergistic interactions between intrinsic and acquired resistance mechanisms, we assessed how RND-mediated efflux affects resistance levels conferred by a naturally occurring resistance plasmid (p2017) isolated from V. cholerae strain 2017 (S1 Table). Preliminary sequencing of p2017 identified three β-lactamase genes and an erythromycin methylase gene, conferring resistance to β-lactams and macrolides; both antibiotic classes are known RND substrates. Introduction of p2017 into C6706 and its isogenic ΔRND mutant revealed synergistic effects between RND-mediated efflux and plasmid-encoded resistance determinants (Table 4). While p2017 increased resistance to ampicillin and erythromycin in both strains, the fold-increase was substantially greater in the wildtype strain than in the ΔRND mutant, demonstrating that intrinsic resistance provided by RND efflux systems enhances the efficacy of acquired resistance determinants.
Contemporary strains exhibit VexB-dependent enhanced motility
Because enhanced intestinal colonization occurred independently of CT and TCP production, we next examined motility, another virulence-associated trait critical for host colonization. Contemporary altered El Tor (Wave 3) strains have been reported to be hypermotile relative to earlier pandemic isolates [32]. Consistent with this, soft agar motility assays showed that all five Wave 3 clinical isolates exhibited significantly greater motility than the early pandemic Wave 1 strain C6706, with motility zone diameters ranging from 132% to 251% of C6706 levels (Table 5).
To determine whether VexB contributes to this hypermotile phenotype, we assessed motility in vexB deletion mutants. Deletion of vexB significantly impaired motility in all contemporary strains, with reductions ranging from 31% to 81% relative to their respective parental strains (Table 5). In contrast, deletion of vexB in C6706 resulted in only a modest 17% motility reduction, indicating that VexB plays a disproportionately larger role in motility among contemporary Wave 3 isolates. The most pronounced effect was observed in strain 2019, which exhibited an 81% decrease in motility upon vexB deletion.
Collectively, these data demonstrate that VexB efflux contributes to the hypermotility phenotype of contemporary V. cholerae strains, providing a potential mechanistic basis for their enhanced colonization capacity that is independent of classical CT and TCP virulence factors.
In vitro evolution for antibiotic resistance produces classical resistance-virulence trade-offs
The unexpected coupling of resistance and hypervirulence in contemporary clinical isolates raised the question of whether enhanced antimicrobial resistance favors the development of increased pathogenic capacity. To investigate this relationship under controlled laboratory conditions, we conducted experimental evolution studies by subjecting the Wave 1 strain C6706 to increasing concentrations of ampicillin, ciprofloxacin, or erythromycin during serial passage in LB medium.
Serial passage under increasing antibiotic pressure successfully generated strains with enhanced resistance to the selection antimicrobials (Table 6). The evolved strains also acquired cross-resistance to multiple unrelated antibiotics, indicating activation of non-specific intrinsic resistance mechanisms. qRT-PCR analysis revealed that enhanced resistance in all strains correlated with vexB upregulation (Fig 5), with vexB expression increases of 9-fold in C6706Amp and C6706Cip, and 14.6-fold in C6706Ery. To confirm the direct contribution of VexB to MDR in the evolved strains, we deleted vexB from the ciprofloxacin-evolved strain and assessed changes in MICs. The ciprofloxacin-evolved strain was selected for this analysis because vexB was singularly upregulated in this strain, eliminating potential confounding effects from multiple upregulated pumps observed in the other evolved strains.
The Wave 1 strain C6706 and its laboratory-evolved derivatives (C6706Amp, C6706Cip, and C6706Ery) were grown to mid-logarithmic phase in LB medium. Total RNA was isolated and qRT-PCR was used to quantify the expression of all six RND efflux pump genes. Data are presented as the fold-change in expression relative to the wild-type C6706 parent and represent the mean ± standard deviation from three independent experiments. Statistical significance was determined using a one-way ANOVA with Dunnett’s multiple comparisons post-hoc test, comparing each evolved strain to the wildtype. *, P < 0.05; **, P < 0.01; ***, P < 0.001.
Deletion of vexB decreased MICs of ampicillin, ciprofloxacin and erythromycin in both wildtype and C6706Cip, confirming the critical role of VexB in intrinsic and evolved resistance (Table 7). In C6706Cip, vexB deletion reduced ampicillin and erythromycin MICs to levels comparable to those of the C6706 ΔvexB mutant (S3 Table), indicating that the elevated resistance of the evolved strain was largely due to increased vexB expression. Despite deletion of vexB, the C6706Cip ΔvexB mutant exhibited a 40-fold higher ciprofloxacin MIC than the parental ΔvexB strain and a 10-fold higher MIC than wild-type C6706 (Table 6), suggesting that additional mechanisms contribute to ciprofloxacin resistance. We hypothesized that mutations in canonical fluoroquinolone resistance genes might arise under ciprofloxacin selection. DNA sequence analysis of gyrase (gyrA, gyrB; VC1258, VC0015) and topoisomerase (parC, parE; VC2430, VC2431) genes in C6706Cip identified four mutations: GyrA S83I, a 3 bp deletion in gyrB (nucleotides 1396–1398), ParC S85L, and ParE L445P. Among these mutations, GyrA S83I and ParC S85L are established determinants of fluoroquinolone resistance [59] and were conserved among five contemporary strains (see below). Thus, while vexB upregulation promoted cross-resistance to multiple antimicrobials, additional target-site mutations amplified ciprofloxacin resistance under sustained pressure, underscoring how efflux and target site mutations together contribute to high-level resistance.
Assessment of the evolved strains in the infant mouse model showed that contrary to the hypervirulent phenotype observed in contemporary clinical isolates, all three laboratory-evolved strains exhibited severe colonization defects in the infant mouse competition model (Fig 6A). The evolved strains showed competitive indices more than 100-fold lower than the parental C6706 strain, indicating severe virulence attenuation despite enhanced antimicrobial resistance. Analysis of virulence factor production in the evolved strains revealed impaired CT and TcpA production and reduced expression of ToxR regulon genes (Fig 6B and Table 8). Similarly, the evolved strains showed approximately 3-log reduced survival in pond water compared to the parental C6706 strain (S2 Fig).
(A) Competitive indices (CI) of laboratory-evolved strains (C6706Amp, C6706Cip, and C6706Ery) against the parental strain C6706 in the infant mouse model. Each circle represents one mouse; horizontal bars indicate geometric means. A CI value below 1.0 indicates a colonization defect. (B) Cholera toxin (CT) production and TcpA immunoblot (inset) of indicated strains after growth under virulence-inducing conditions. The immunoblot is representative of three independent experiments. CT data are presented as the mean ± SD from three independent experiments. Statistical significance was determined by a one-way ANOVA with Dunnett’s multiple comparisons test (A) and an unpaired, two-tailed Student’s t-test (B), comparing each evolved strain to the parental C6706 strain. *, P < 0.05; ***, P < 0.001.
Together these results demonstrate that laboratory selection for antibiotic resistance produces the classic resistance-fitness trade-off, where enhanced antimicrobial resistance is achieved at the expense of both pathogenic capacity and environmental fitness. The sharp divergence between the laboratory-evolved strains (resistant but attenuated) and the contemporary clinical isolates (resistant and hypervirulent) indicates that antibiotic pressure alone cannot account for the coupled resistance and hypervirulence phenotype observed in Wave 3 strains.
Genomic analysis reveals molecular adaptations in contemporary strains
Our identification of VexB as mediating both antimicrobial resistance and hypervirulence in contemporary isolates raised the question of what molecular adaptations enable VexB’s dual function in Wave 3 strains. To identify genetic changes contributing to the coupled phenotype, we performed whole-genome sequencing on all five clinical isolates and compared sequences to the Wave 1 reference strain C6706 using Breseq [60].
Breseq analysis identified 279 conserved genetic variants among all five isolates (S4 Table), including 121 nonsynonymous substitutions, 103 synonymous changes, 7 nonsense mutations, 20 intergenic variants, and 30 indels distributed across both chromosomes (230 on chromosome 1, 49 on chromosome 2). All strains possessed canonical altered El Tor (AET) Wave 3 markers: the ctxB7 allele (H20N, Y39H, I68T), the N89S TcpA substitution, and rtxA frameshift mutation (W4534stop) eliminating functional MARTX production [29,61–63].
All strains shared a conserved 49.3 kb deletion on chromosome 1 (positions 1,494,671–1,543,947) encompassing 56 genes corresponding to a cryptic prophage element (S5 Table). The deleted region encodes core phage structural components, including terminase subunits, portal protein, major capsid protein, head completion protein, prohead protease, and head-tail connector, as well as proteins involved in DNA metabolism (DNA polymerase, recombinase, exonuclease, toprim domain protein) and an antirestriction protein (ArdA). The prophage lacks identifiable lysis genes (holin or lysin) and integration or excision functions (integrase or excisionase), indicating that it represents a degenerate element unlikely to undergo excision or lytic replication. This element is present in several commonly used Wave 1 strains (e.g., C6709 and A1552) but absent from the reference strain N16961. The deleted region does not encode known virulence factors, motility regulators, or c-di-GMP metabolism proteins, suggesting that its loss is unlikely to directly account for the enhanced motility or colonization phenotypes observed in Wave 3 strains.
We found no mutations within any of the six RND pump genes or their immediate promoter regions, suggesting elevated vexB expression in Wave 3 strains results from trans-acting regulatory changes rather than cis-regulatory mutations. Beyond vexB regulation, conserved mutations were identified in known MDR determinants including quinolone resistance mutations (GyrA S83I and ParC S85L), consistent with elevated ciprofloxacin MICs [64,65]. A conserved D89N substitution in CarR, a transcriptional activator of the almEFG lipid A modification operon, impairs function and reduces polymyxin B resistance [40]. Additionally, a nonsynonymous mutation in ompV, an outer membrane protein linked to polymyxin B resistance, was identified [66].
Breseq analysis revealed multiple alterations in cyclic-di-GMP (c-di-GMP) metabolism genes, with mutations affecting at least eight genes encoding GGDEF (diguanylate cyclase) or EAL/HD-GYP (phosphodiesterase) domain proteins. A conserved null mutation in the diguanylate cyclase gene cdgL (A77stop) would attenuate c-di-GMP synthesis. A frameshift mutation restored production of the phosphodiesterase pseudogene ABDM36_RS08015, predicted to further reduce c-di-GMP. Additional nonsynonymous mutations were found in two phosphodiesterases (cpdB and ABDM36_RS14480) and two diguanylate cyclases (ABDM36_RS02970 and ABDM36_RS17815) with unknown functional effects. All isolates carried the L79F substitution in phosphodiesterase VieA [32], which dysregulates c-di-GMP degradation [67]. Collectively, these mutations suggested reduced intracellular c-di-GMP in Wave 3 strains.
Since low c-di-GMP promotes virulence [68], we assessed c-di-GMP levels using a fluorescent reporter as previously described [69,70]. All Wave 3 strains exhibited significantly lower c-di-GMP compared to C6706 (Fig 7), correlating with their enhanced motility and hypervirulent phenotype. This finding aligns with reported c-di-GMP repression in the 2010 Haiti strain [32], suggesting reduced c-di-GMP is a conserved adaptation enhancing transmission and acute infection. Additionally, the conserved S105G substitution in global repressor H-NS, which was also present in the Haiti strain, may derepress virulence gene expression and affect c-di-GMP signaling, further contributing to hypervirulence [71].
The indicated strains bearing a dual-fluorescent c-di-GMP biosensor (pFY_4535) were grown to mid-exponential phase in LB medium, and fluorescence was quantified as described in Methods. Data are presented as the mean Relative Fluorescence Intensity (RFI), calculated as the ratio of TurboRFP to AmCyan fluorescence. Error bars represent the standard deviation from at least three independent biological replicates. Statistical significance was determined using two separate analyses. An unpaired, two-tailed Student’s t-test was used to compare each mutant strain to its respective parental wild-type strain. A one-way ANOVA with Dunnett’s multiple comparisons test was used to compare all strains to the C6706 control. *, P<0.05, ***, P<0.001.
We hypothesized that the coupled MDR and hypervirulence phenotypes might be mechanistically linked through VexB-dependent effects on c-di-GMP, particularly given the significant motility impairment upon vexB deletion. We therefore assessed whether VexB-dependent modulation of c-di-GMP and its downstream phenotypic consequences were conserved or strain-specific. However, vexB deletion affected c-di-GMP levels in only two isolates (2017 and 2019; Fig 7), while colonization defects occurred in all five strains (Fig 1A). This strain-specific variability in c-di-GMP response, despite consistent virulence and motility phenotypes across all vexB mutants, indicates that VexB’s contribution to pathogenesis operates through multiple mechanistic pathways whose relative importance varies by genetic background. This suggests that c-di-GMP regulation represents one potential regulatory pathway by which VexB can influence motility and colonization, but that additional VexB-dependent pathways contribute in parallel. Furthermore, the absence of mutations in RND genes themselves suggests regulatory adaptations affecting RND expression are mediated by trans-acting factors.
Discussion
The emergence of pandemic pathogens that combine MDR with heightened virulence represents a critical global health threat while challenging fundamental assumptions about pathogen evolution under antibiotic pressure. Yet the molecular mechanisms enabling this coupling have remained elusive. Here, we identify the VexB RND efflux pump as a molecular determinant that simultaneously mediates multidrug resistance and hypervirulence in contemporary V. cholerae, revealing a previously unrecognized evolutionary strategy that circumvents classical resistance-virulence trade-offs. In early pandemic Wave 1 strains, VexB functions primarily in antimicrobial resistance, whereas in contemporary Wave 3 isolates it has acquired dual functionality, contributing to both resistance and colonization independent of CT/TCP production. This functional adaptation demonstrates that the complex selective pressures encountered during the pathogenic life cycle (e.g., host immunity, microbial competition, intermittent antimicrobial exposure, and environmental antibiotic contamination) can drive adaptations that optimize both antibiotic resistance and pathogenic potential. By establishing VexB as a molecular link between resistance and virulence, our work highlights RND efflux systems as dual function targets whose inhibition could both restore antibiotic susceptibility and reduce pathogenic potential.
The marked contrast between our laboratory-evolved strains and clinical isolates underscores the critical role of natural selection in shaping pathogen evolution. Under controlled, single-stressor antibiotic pressure, V. cholerae developed classic resistance-fitness trade-offs, acquiring enhanced resistance but with attenuated virulence. By contrast, the complex selective pressures encountered during natural transmission, including host immunity, microbial competition, and intermittent antimicrobial exposure, likely drove the accumulation of genetic adaptations that balanced both resistance and virulence. Adding to these selective pressures, recent studies have reported high concentrations of antibiotics in surface waters in Bangladesh and other cholera-endemic regions [72,73], suggesting that V. cholerae populations experience strong selection in their native aquatic ecosystems as well as within the human GI tract. RND efflux systems are well positioned to mediate these adaptive responses, as they can export both antibiotics and cellular metabolites. Just as importantly, when efflux is impaired, the intracellular accumulation of these efflux substrates can engage environmental sensors that reshape global regulatory networks (reviewed in [22]). In this way, RND pumps function not only as a defensive barrier against antimicrobial pressure but also as a conduit for integrating environmental signals into transcriptional responses that influence adaptation and virulence [28]. This dual role of RND systems positions them to serve in integrating resistance with virulence. The metabolite feedback model [22] predicts that RND pumps functioning under these complex selective pressures would evolve not merely as active efflux conduits, but as regulators whose expression levels and substrate specificities directly shape both antimicrobial resistance and pathogenic outputs. This complexity is exemplified by the distinct functional evolution of VexB itself, which appears to have acquired new roles in contemporary strains beyond its ancestral functions.
Our mechanistic investigations reveal that VexB contributes to colonization through multiple pathways. First, VexB-mediated efflux protects against membrane-active antimicrobials including detergents and antimicrobial peptides, both encountered in the intestinal environment. This protection would enhance survival during transit through the stomach and small intestine. Second, vexB mutation impairs adherence to intestinal epithelial cells, likely through both reduced motility and potential alterations in surface properties. Third, the enhanced motility conferred by VexB facilitates intestinal colonization, as motility is critical for penetrating the mucus layer and reaching epithelial surfaces. Together, these mechanisms explain in part how VexB simultaneously enhances both antimicrobial resistance and colonization capacity.
Our findings reveal that VexB’s contribution to colonization in Wave 3 strains operates through mechanisms fundamentally distinct from the previously characterized RND-mediated feedback regulation of the ToxR virulence regulon. Prior studies established that RND-exported metabolites, including cyclo(Phe-Pro) and indole, accumulate in the periplasm when efflux is impaired and function to repress ToxR regulon expression through periplasmic sensor-mediated signaling cascades [37,52,74,75]. In contrast, vexB deletion in contemporary Wave 3 strains resulted in significant colonization defects (4–10 fold reductions) while leaving ToxR regulon expression largely unchanged (Table 2). This decoupling of colonization capacity from ToxR-dependent virulence factor production indicates that VexB has acquired novel regulatory functions in Wave 3 strains that extend beyond the canonical metabolite feedback circuits characterized in Wave 1 isolates.
While we have not yet identified the specific VexB substrates or periplasmic sensors mediating these phenotypes, the functional evidence clearly establishes VexB’s central role in coordinating multiple colonization-relevant traits. The strain-specific variations in phenotypic magnitude (e.g., c-di-GMP changes in only 2/5 isolates despite universal colonization defects) suggest VexB may function through partially overlapping pathways whose relative contributions vary by genetic background. This complexity is consistent with RND pumps functioning as regulatory hubs integrating multiple signals rather than simple efflux transporters.
Our identification of VexB as having a dual role in antimicrobial resistance and hypervirulence in contemporary isolates provides a novel molecular entry point for understanding how these traits are coupled. It is noteworthy that this dual function appears strain-specific: in early pandemic Wave 1 strains (e.g., C6706 and N16961), VexB primarily contributes to resistance [34,57], whereas in contemporary Wave 3 isolates it influences both resistance and colonization. This shift suggests that Wave 3 strains have acquired novel regulatory or functional adaptations that co-opt RND-mediated efflux for enhanced pathogenesis. Several mechanisms could explain VexB’s evolved dual functionality. First, the substantially elevated basal vexB expression in Wave 3 strains (Fig 2A) may simply reflect quantitative differences; higher efflux activity could more effectively modulate the periplasmic metabolite pool, effecting feedback signaling to regulatory networks that control motility and colonization. Second, the coordinated mutations in c-di-GMP metabolism may have created a regulatory environment more responsive to VexB-mediated metabolite flux, effectively amplifying signals that were previously too weak to impact virulence in Wave 1 strains. Third, the loss or gain of specific transcriptional regulators in Wave 3 strains could have established new regulatory connections between vexB expression and virulence gene networks. Distinguishing among these possibilities will require comparative genomic and functional analyses of Wave 1 versus Wave 3 strains, identification of key periplasmic substrates, and systematic mapping of the regulatory networks connecting efflux activity to motility and colonization.
The independence of VexB-mediated hypervirulence from virulence factor production (i.e., CT and TCP) points towards RND-mediated efflux affecting additional facets of pathogenesis that complement the canonical virulence factors (e.g., TCP and CT). Our discovery that VexB deletion significantly impairs motility in contemporary strains offers one potential explanation for the reduced colonization capacity of the vexB mutants in the contemporary strains. Motility is critical for V. cholerae pathogenesis, facilitating dissemination and adherence within the intestinal tract [76–80]. The direct correlation between basal vexB expression levels and the degree of motility impairment upon its deletion supports a functional relationship. Although the precise molecular mechanisms linking efflux and flagellar function remain unclear, changes in c-di-GMP do not appear to mediate this phenotype, as we did not detect significant differences in c-di-GMP levels upon vexB mutation except in strain 2019. There have been mixed results for the effects of RND efflux on motility in other organisms. For example, reduced motility was observed in an RND mutant in Acinetobacter baumannaii [81]. By contrast, mutation of acrB in E. coli in increased motility compared to the wild-type strain via a regulatory mechanism linked to AcrR [82,83] and a similar increase motility was found in a Pseudomonas aeruginosa mexEF mutant [84]. We speculate that metabolite feedback may contribute to these phenotypes, but the specific metabolites involved, and whether additional mechanisms play a role, remain important areas for future investigation.
The coordinated mutations identified in genes involved in c-di-GMP metabolism across contemporary V. cholerae isolates suggest a regulatory shift that may influence the balance between virulence, motility, and biofilm formation. Our genomic analysis revealed alterations in at least nine genes encoding predicted diguanylate cyclases (GGDEF domain proteins) and phosphodiesterases (EAL or HD-GYP domain proteins). Among these, the L79F substitution in the EAL domain-containing protein VieA is particularly notable [31,32,67]. While the biochemical impact of this mutation has not been fully characterized, previous studies have implicated VieA in the regulation of intracellular c-di-GMP levels, with downstream effects on motility and virulence gene expression [67,85,86]. The consistent reduction in c-di-GMP levels observed across all Wave 3 isolates (Fig 7) supports the hypothesis that these mutations contribute to a physiological state favoring acute infection. This low c-di-GMP state is associated with increased motility and virulence factor expression, including CT, while suppressing biofilm formation, a phenotype that may enhance transmission and host colonization during epidemic spread [68].
The relationship between c-di-GMP signaling and VexB function illustrates the multi-pathway nature of VexB’s contribution to pathogenesis. While vexB deletion increased c-di-GMP levels in strains 2017 and 2019 but not in the other isolates, motility defects and colonization impairment occurred across all strains regardless of c-di-GMP changes. This demonstrates that VexB influences pathogenesis through multiple overlapping mechanisms, with c-di-GMP regulation serving as one regulatory mechanism alongside other VexB-dependent pathways. These additional pathways may include direct effects on antimicrobial resistance, modulation of adhesins, or influence on other metabolite-mediated signaling cascades. Whether c-di-GMP or its metabolic precursors serve as substrates for RND-mediated efflux, or whether VexB activity influences the periplasmic signaling environment that regulates c-di-GMP metabolizing enzymes represents a critical question for future studies. The ability to influence pathogenesis through multiple pathways would provide robustness to VexB’s function, ensuring that Wave 3 strains maintain enhanced pathogenic capacity during exposure to the diverse selective pressures encountered during infection and environmental survival.
Our study utilized five geographically diverse clinical isolates collected between 2017 and 2019, which limits the generalizability of our findings to all Wave 3 strains globally. However, several factors suggest these results reflect genuine adaptations rather than strain-specific anomalies. First, the phenotypes were remarkably consistent across the isolates despite their geographic diversity (i.e., travelers returning from different cholera-endemic regions). Second, the conserved genetic markers, including the VieA L79F mutation, ctxB7 allele, GyrA S83I and ParC S85L substitutions, and CarR D89N mutation, align precisely with previously documented Wave 3 characteristics and mirror findings from the Haiti and Yemen epidemics [87,88]. Third, the mechanistic link between elevated vexB expression and both resistance and colonization were reproducible across multiple strains, with the magnitude of phenotypic effects correlating with vexB expression levels. These lines of evidence suggest our findings represent a broader evolutionary adaptation rather than isolated occurrences.
Nevertheless, validation with larger, systematically collected strain panels from multiple endemic regions over extended timeframes is essential to determine the prevalence and stability of VexB-mediated coupling. Mechanistic investigations should prioritize defining VexB’s specific substrates in Wave 3 versus Wave 1 strains, elucidating the regulatory connections between efflux activity and motility that operate independently of c-di-GMP, and identifying the periplasmic sensors that transduce metabolite signals into transcriptional responses. Additionally, longitudinal studies tracking vexB expression and associated phenotypes in environmental and non-toxigenic V. cholerae strains would clarify how environmental pressures shape this adaptation in natural settings.
In conclusion, our findings reveal a novel adaptation strategy in V. cholerae that bypasses classical evolutionary trade-offs through the co-evolution of resistance and virulence, mediated by the RND efflux pump VexB. The dual function of VexB identifies it as a promising therapeutic target; its inhibition could potentially restore antibiotic efficacy while simultaneously reducing disease severity. This work underscores the sophisticated regulatory integration possible when efflux systems evolve beyond simple resistance mechanisms to coordinate complex physiological responses, creating new challenges for infectious disease control.
Materials and methods
Ethics statement
All animal procedures were conducted in accordance with protocols approved by the University of Pittsburgh Institutional Animal Care and Use Committee (Protocol number 15015310). The studies compiled with standards for humane animal care and use established by the Animal Welfare Act and followed guidelines set forth in the NIH Guide for the Care and Use of Laboratory Animals.
Bacterial stains and culture conditions
The bacterial strains used in this study are listed in S1 Table. V. cholerae O1 El Tor strain C6706 was used as the WT control in all experiments. E. coli strain EC100 was used as the host strain for DNA cloning experiments and E. coli strain SM10 λpir was used to conjugate plasmids into V. cholerae. Bacteria were grown on LB agar or in LB broth at 37°C.
For virulence-related assays requiring induction of the ToxR virulence regulon, bacteria were grown in AKI medium (15 g Bacto peptone, 4 g Difco yeast extract and 5 g of NaCl per liter, pH 7.4). The culture conditions were as follows: Test strains were grown in LB broth at 37°C overnight, then individually diluted (1:10,000) into 10 mL AKI broth in 150 x 15 mm glass tubes. Cultures were incubated statically for 4 hours until the OD600 readings exceeded 0.08, then transferred to 125 mL Erlenmeyer flasks and incubated with shaking at 37°C for additional time as indicated or overnight for CT and TCP quantification.
Antibiotics were used at the following concentrations: streptomycin, 100 μg/mL; carbenicillin, 100 μg/mL; chloramphenicol (Cml), 1 µg/mL for V. cholerae or 20 µg/mL for E. coli. Kanamycin (Kan), ampicillin (Amp), erythromycin (Ery), ciprofloxacin (Cip), tetracycline (Tet), polymyxin B (PXB), deoxycholate (DOC), and sodium dodecyl sulfate (SDS) were used at concentrations indicated in the respective experiments.
Chemicals and reagents
Chemicals and antibiotics were purchased from Sigma-Aldrich (St Louis, MO, USA), and enzymes were purchased from New England Biolabs (Beverly, MA, USA). Oligonucleotides were synthesized by Integrated DNA Technologies (Coralville, IA, USA) and were designed based upon the V. cholerae N16961 genome sequence [89].
Plasmid and mutant construction
The plasmids and primers used in this study are listed in S1 Table. In-frame deletions of the genes that encoding each RND efflux pump was accomplished as previously described [34]. Briefly, SM10 λpir bearing the individual deletion constructs in pWM91 (i.e., vexB (VC0164), vexD (VC1757), vexF (VC0629), vexH (VC0914), vexK (VC1673), vexM (VC0638)) were conjugated into V. cholerae C6706 and cointegrants were selected on LB agar containing Cb and Sm. The Cb/Sm resistant colonies were then streaked onto LB agar (without NaCl) containing 5% sucrose to select for the resolution of the integrated plasmid. Several sucrose-resistant colonies were screened for Cb sensitivity to verify plasmid loss before the target gene deletion was confirmed by PCR using flanking primers. The pBAD33-vexAB expression vector was constructed by a two-step process. First, the 164F-XbaI and 164R-SalI PCR primer pair and the 165F-SacI PCR primer pair were used with C6706 genomic DNA as a template to amplify a 2.1 kb and 2.4 kb amplicons representing overlapping regions of the N- and C-terminal regions of the vexAB locus. The 2.1 amplicon was restricted with XbaI and SalI restriction enzymes and cloned into similarly digested pBAD33. The resulting plasmid was then restricted with SacI and XbaI and ligated with the 2.4 kb PCR product that was similarly digested to generate pBAD33-vexAB (S1 Table). The resulting plasmid was verified by DNA sequencing.
Plasmid p2017 was isolated from V. cholerae strain 2017V-1003 using standard miniprep procedures. Whole-genome sequencing was performed on an Illumina platform as described below, and reads were assembled using SPAdes v3.15.5 in plasmid mode [90], yielding an approximately 134 kb plasmid carrying an IncC-type replicon. Antimicrobial resistance genes were identified using AMRFinderPlus [91], revealing three β-lactamase genes (blaCTX-M-15, blaOXA-10, and one additional β-lactamase), a macrolide phosphotransferase (mph(A)), trimethoprim resistance genes (dfrA14, dfrA15), sulfonamide resistance genes (sul1, sul2), and chloramphenicol resistance genes (cmlA5, catA1). Plasmid p2017 was introduced into recipient strains C6706 and ΔRND by electroporation, and transconjugants were selected on LB agar containing ampicillin.
CT and TcpA quantification
Virulence factor production (CT and TcpA) was assessed in all strains following overnight growth under AKI conditions. For Western blot analysis, cultures were normalized to OD600 = 1.0 prior to sample preparation. Normalized cultures (1 mL) were pelleted by centrifugation at 16,000 × g for 5 minutes, and cell pellets were resuspended in a defined volume of 1x SDS-PAGE sample buffer (50 mM Tris-HCl pH 6.8, 2% SDS, 10% glycerol, 1% β-mercaptoethanol, 0.02% bromophenol blue) to ensure equal loading based on cell density and heated at 95°C for 5 minutes. Equal volumes (10 µL) of the normalized lysates were loaded in each lane of 12% SDS-PAGE gels to ensure equivalent sample loading. Following electrophoresis, proteins were transferred to PVDF membranes and TcpA production was quantified by Western immunoblotting using polyclonal anti-TcpA antibody (1:5,000 dilution). Immunoreactive proteins on the Western blots were visualized using the LI-COR detection kit (LI-COR Biosciences) with images captured using a LI-COR Odyssey CLx-2770 imaging system (LI-COR Biosciences). All Western blots shown are representative of at least three independent biological replicates.
CT production was quantified using a GM1 enzyme-linked immunosorbent assay as previously described, with purified CT used as a standard for quantitation [92]. CT ELISA results were measured using a Synergy II plate reader (BioTek), and data were analyzed using GraphPad Prism 10 software. Statistical significance was determined by one-way analysis of variance (ANOVA) comparing CT production levels between wild-type and ΔvexB mutant strains for each isolate.
Antimicrobial susceptibility tests
Antimicrobial susceptibility tests were performed using the micro broth dilution method as previously described or using gradient agar plates [34]. Briefly, gradient agar plates were generated by adding 35 ml of molten LB agar to each 9 cm x 9 cm square petri plate and the plate was then incubated with one end elevated on a 5 ml disposable pipet until solidified. The petri plates were then leveled and 35 ml molten LB agar containing the specific test antibiotic was added to each plate and the plates were allowed to solidify. The plates were then allowed to equilibrate at room temperature for 2–4 hours before inoculation. The gradient agar plates were inoculated by dipping a cotton-tipped swab into fresh broth cultures of the test strains and streaking the cotton swap across the gradient. A maximum of seven test strains plus the WT control strain were inoculated onto the surface of each plate. After inoculation, the plates were incubated overnight at 37°C before growth was measured using a ruler. The MIC was calculated by the percentage growth across the plate multiplied by the antimicrobial concentration used in the plate. The MIC determination was performed a minimum of three independent times and representative results from one experiment are presented.
Growth analysis
Strains were grown overnight in LB medium and normalized to OD600 of 1.0. The cultures were then diluted 1:10,000 into LB broth with or without arabinose as indicated. 100 µL of each strain were then distributed in triplicate wells of a flat bottom 96-well microtiter plate. The 96-well plate was then incubated in a BioTek ELx808 plate reader at 37°C with intermittent shaking and the growth was monitored by measuring the absorbance of OD600 every 30 min for 16h. The presented data is the average of triplicate samples for each time point. Results were analyzed using GraphPad Prism 10.
Antimicrobial killing assays
V. cholerae strains were cultured overnight at 37°C in LB medium, before being subcultured in fresh LB and grown to mid-log phase (OD600 of 1.0). For the killing assays, 100 μl of log-phase culture was added to 900 μl of LB medium containing SDS (8x MIC) or polymyxin B (PxB, 4x MIC) at the indicated concentrations and mixed thoroughly. The cultures were then incubated at 37°C for 15 min (PxB) or 30 min (SDS) when culture aliquots were serially diluted in LB and plated onto LB agar plates to enumerate surviving bacteria. Untreated control cultures were processed identically without antimicrobial agents. Bacterial survival was then calculated as the percentage of viable bacteria remaining after treatment: (treated CFU/untreated CFU) × 100%. All experiments were performed as three independent experiments, each conducted in triplicate. Statistical significance was determined by unpaired t-test comparing survival percentages between wild type and ΔvexB mutant strains for each antimicrobial treatment.
Pond water survival assay
The pond water survival assay was used to access environmental fitness [93,94] and performed as follows. Pond water was collected from North Park Lake (Allegheny County, Pennsylvania), filtered, and autoclaved prior to use. Test strains and the control strain C6706 were grown to mid-log phase in LB medium, pelleted by centrifugation, and independently resuspended in sterile pond water to equivalent optical densities. Test strains and C6706 were mixed at a 1:1 ratio and used to inoculate pond water at a final concentration of 1 × 105 CFU/mL per strain. Input ratios were verified by plating serial dilutions on LB agar supplemented with X-gal (40 μg/mL), incubating overnight at 37°C, and enumerating blue (test strains, lacZ⁺) and white (C6706, lacZ⁻) colonies. Cultures were incubated statically at 30°C for 7 days. Following incubation, aliquots were serially diluted and plated as described above to enumerate surviving bacteria. The competitive index (CI) was calculated as (test strain CFUoutput/C6706 CFUoutput)/ (test strain CFUinput/C6706 CFUinput). Experiments were performed in triplicate. Statistical significance was determined using a one-sample t test comparing CI values to 1.0 (p < 0.05).
Gene expression analysis
V. cholerae strains were grown in LB medium under the indicated conditions. Total RNA was isolated from cultures at each time point using TRIzol reagent according to the manufacturer’s instructions. RNA quality and quantity were assessed using a NanoDrop 2000 spectrophotometer, ensuring an A260/A280 ratio of ~2.0 and an A260/A230 ratio of 1.8-2.0. Prior to cDNA synthesis, RNA samples were treated with DNase for 5 min at 37°C to remove genomic DNA contamination. cDNA was generated using the SuperScript III RT Kit (Invitrogen) according to the manufacturer’s instructions. Briefly, 2X Maxima Enzyme Premix was added to the DNase-treated RNA, gently mixed, briefly centrifuged, and then amplified in a thermal cycler using the recommended protocol.
The expression level of each specific gene was quantified by amplifying 25 ng of cDNA with 0.5 μM gene-specific primers (S1 Table) using SYBR Green PCR mix (Fisher Scientific) on a StepOnePlus Real-Time PCR System (Applied Biosystems). The rpoB gene was used as an internal reference for normalization and gene expression changes were calculated using the 2−ΔΔCT method. Results are presented as means ± standard error of three biological replicates, with each biological replicate consisting of three technical replicates. Statistical significance was determined using Student’s t-test.
Infant mouse colonization model
Overnight cultures of mutant (lacZ positive) and wildtype (WT) reference (lacZ negative) V. cholerae strains were diluted individually 1:100 in LB broth. The inoculum was prepared by combining 10 μL of each diluted strain with 975 μL LB containing 5 μL of blue food coloring (for visualization of gastric delivery). Six-day-old infant CD1 mice, separated from their mothers approximately 2 hours prior to the procedure, were lightly anesthetized with isoflurane and inoculated orally with 50 μL of the bacterial inoculum. Concurrently, an aliquot of the inoculum was serially diluted and plated on LB agar containing X-Gal (20 μg/mL) to determine the input ratio of bacterial strains.
Infected mice were housed at 30°C for ~14 h before euthanasia. The small intestine of each mouse was excised above the cecum, homogenized in 5 mL PBS, serially diluted, and plated on LB agar containing streptomycin (100 μg/mL) and X-Gal. Following overnight incubation at 37°C, colonies were enumerated and differentiated as mutant (blue) or WT (white). The competitive index (CI) was calculated as the ratio of WT to mutant bacteria in the output divided by the ratio of WT to mutant bacteria in the input.
In vitro growth competition assays
In vitro growth competition assays were performed using a protocol similar to the infant mouse colonization model. Overnight cultures of mutant (lacZ positive) and wild-type (WT) reference (lacZ negative) V. cholerae strains were diluted individually 1:100 in LB broth. For each competition, 10 μL of each diluted strain was mixed in 980 μL LB medium. An aliquot of this mixture was serially diluted and plated on LB agar containing X-Gal (20 μg/mL) to determine the input ratio. The mixed cultures were incubated at 37°C for ~16 h in parallel with the in vivo mouse experiments. Following incubation, cultures were serially diluted and plated on LB agar containing X-Gal. After overnight incubation at 37°C, colonies were counted and classified as mutant (blue) or WT (white). The competitive index was calculated as the ratio of WT to mutant bacteria in the output divided by the ratio of WT to mutant bacteria in the input.
Enterocyte adherence assay
V. cholerae strains were cultured overnight at 37°C in LB medium, then subcultured into fresh LB and grown to mid-log phase (OD600 of 1.0). Bacterial cultures (1 ml) were collected and pelleted by centrifugation, washed once in Hank’s Balanced Salt Solution (HBSS), before being resuspended in 1 ml of Hank’s HBSS. HT29-MTX-E12 cells that were purchased from ATCC were cultured in 6-well plates for 3 weeks to achieve confluent monolayers. Prior to infection, the cell monolayers were washed three times with HBSS, and 3 ml of fresh DMEM was added to each well. Control wells containing DMEM but no cells were included to determine bacterial viability during the incubation period. Bacterial suspensions (10 μl) were added to wells containing confluent cell monolayers and to control wells, and plates were incubated at 37°C for 30 minutes. Following incubation, the HT29-MTX wells were washed three times with HBSS to remove non-adherent bacteria before adherent bacteria were recovered by adding 1 ml of HBSS to each well and scraping cells from the plate surface using a cell scraper. Cell-bacteria suspensions were collected, serially diluted in HBSS, and plated onto LB agar plates for enumeration. Bacteria from control wells were similarly collected and plated to determine the viable bacterial count after 30 minutes of incubation in DMEM. Adherence efficiency was calculated as the percentage of bacteria that remained adherent to cells relative to the control wells: (adherent CFU/control well CFU) x 100%. All experiments were performed in triplicate.
Experimental in vitro evolution of antibiotic-resistant V. cholerae
In vitro evolved antibiotic-resistant V. cholerae variants were generated through serial passage under increasing selective pressure provided by the addition of ampicillin, ciprofloxacin or erythromycin. Overnight cultures of V. cholerae C6706 were diluted 1:500 into three parallel test tubes containing 5 mL LB broth supplemented with 0.5 x MIC of the respective antibiotic. After 24 h of aerobic incubation at 37°C with shaking at 200 rpm, 10 μL of each culture was transferred to fresh LB medium containing an antibiotic concentration 0.5 x MIC higher than the previous passage. This serial passage continued with antibiotic concentrations incrementally increasing until the cultures reached a MIC increase of ≥4-fold higher compared to the parental strain. Following the final passage, single colonies were isolated on LB agar plates. Individual colonies were then verified for stable resistance through three successive passages on antibiotic-free media followed by MIC determination. The verified resistant strains (designated C6706Amp, C6706Cip, and C6706Ery) were stored in 25% glycerol at -80°C and used for all subsequent analyses. In parallel, the parental strain was cultured under identical conditions but without antibiotics to serve as a control for adaptations unrelated to antibiotic selection.
Measurement of intracellular c-di-GMP levels
Intracellular cyclic-di-GMP (c-di-GMP) concentrations were determined using a dual-fluorescent reporter plasmid, pFY_4535, which contains a Bc3–5 c-di-GMP biosensor as previously described [69]. V. cholerae strains C6706, 2017, 2018–1, 2018–2, 2018–3, and 2019, plus their corresponding vexB deletion mutants harboring the pFY_4535 plasmid was grown overnight at 37°C with aeration in LB supplemented with 15 µg/mL gentamicin. The overnight cultures were then diluted 1:200 into fresh LB medium and grown aerobically to the mid-exponential phase (OD600 of 0.6). Fluorescence measurements were then taken from 200 µL of culture in clear-bottom, black, polystyrene 96-well microtiter plates using a Biotek Synergy plate reader. Fluorescence intensity for the normalizer, AmCyan, was measured using an excitation/emission of 460/480 nm, and for the reporter, TurboRFP, using an excitation/emission of 550/580 nm. The intracellular c-di-GMP level is reported as Relative Fluorescence Intensity (RFI), calculated as the ratio of the fluorescence intensity of TurboRFP to AmCyan. The experiment was performed with at least three independent biological replicates, each with three technical replicates.
Next generation sequencing and variant identification
Genomic DNA was extracted from V. cholerae strains and submitted to SeqCenter (Pittsburgh, PA, USA) for Illumina paired-end sequencing. Sequencing was performed on an Illumina NovaSeq 6000 platform generating 2 × 150 bp paired-end reads. Raw sequencing reads were analyzed using the breseq computational pipeline (v0.35.7) [60] with default settings to identify genetic variants among the clinical isolates relative to reference strain C6706 (NCBI Reference Sequences NZ_CP157384.1 and NZ_CP157385.1). Variants conserved in all five isolates are reported in S4 Table. Raw sequencing data are available in the NCBI Sequence Read Archive under BioProject PRJNA1284390.
Phylogenetic analysis
To establish the phylogenetic context of our five contemporary clinical isolates within the seventh pandemic, we performed maximum-likelihood phylogenetic analysis of 67 V. cholerae O1 El Tor genomes spanning 1960–2019. Reference genomes representing Wave 1 (n = 4, 1960–1994), Wave 2 (n = 6, 1990–2004), and Wave 3 (n = 52, 2004–2019) were selected from published datasets [29,39] (S6 Table). Whole-genome SNP variation relative to reference strain C6706 was identified using Snippy v4.6.0 (https://github.com/tseemann/snippy), generating a core genome alignment of 13,870 SNP positions (S7 Table). Maximum-likelihood phylogenetic reconstruction was performed using IQ-TREE v2.3.0 with automatic model selection (GTR + F + ASC + G4) and branch support assessed by 1,000 ultrafast bootstrap replicates [95,96]. The tree was visualized using iTOL (Interactive Tree of Life) [97], with wave assignments based on published classifications [29,39].
Supporting information
S1 Fig. Expression of carR and almE in contemporary O1 El Tor V. cholerae isolates.
The indicated isolates were cultured in LB medium to mid-logarithmic phase when RNA was isolated and used for qRT-PCR to quantify expression of carR (A) and almE (B). Expression was normalized to the Wave 1 control strain C6706. Results represent the mean ± standard deviation of three independent biological replicates. Statistical analysis was performed using one-sample t-tests comparing each strain to a theoretical ratio of 1.0. ***, P < 0.001.
https://doi.org/10.1371/journal.ppat.1014031.s001
(TIF)
S2 Fig. Competitive fitness of V. cholerae strains in pond water survival assay.
Competitive index (CI) values of the indicated V. cholerae test strains compared to reference strain C6706 following 7-day incubation at 30°C in sterile pond water collected from North Park Lake. Each strain was mixed 1:1 with the lacZ⁻ reference strain C6706 and inoculated into pond water at 1x105 CFU/ml of each strain. CI values were calculated as (test strain CFU output/C6706 CFU output)/(test strain CFU input/C6706 CFU input). A CI of 1 indicates equal fitness to the reference strain, while CI values <1 indicate reduced competitive fitness. Data represent the mean and standard deviation from three independent experiments. Statistical significance was determined by one-sample t-test comparing CI values to 1.0. *, P < 0.05.
https://doi.org/10.1371/journal.ppat.1014031.s002
(TIF)
S3 Fig. Phylogenetic placement of contemporary V. cholerae isolates within seventh-pandemic waves.
Maximum-likelihood phylogenetic tree of 67 V. cholerae O1 El Tor genomes spanning the seventh pandemic (1960–2019). The tree was generated from a 3,863,204-bp core genome alignment using IQ-TREE2 with automatic model selection (TVM + F + I + R4). The tree is rooted on Wave 1 strains and shows three major clades corresponding to pandemic waves: Wave 1 (blue, 1960–1994), Wave 2 (green, 1990–2004), and Wave 3 (red, 2004–2019). Bootstrap support values (1,000 ultrafast replicates) are shown at internal nodes. The five study isolates (Vc_2017, Vc_2018–1, Vc_2018–2, Vc_2018–3, Vc_2019) are highlighted in yellow and cluster within the Wave 3 clade, indicating that these contemporary isolates form a coherent sublineage, suggesting that their genomic variants are wave-associated adaptations rather than strain-specific artifacts.
https://doi.org/10.1371/journal.ppat.1014031.s003
(TIF)
S4 Fig. Differential regulation of V. cholerae RND efflux pump genes in response to subinhibitory antimicrobial exposure.
Wild-type V. cholerae C6706 was cultured in LB medium containing 0.5x MIC of each indicated antimicrobial compound (Amp, ampicillin; Cip, ciprofloxacin; Ery, erythromycin; Kan, kanamycin). RNA was isolated at mid-logarithmic phase (OD600 0.6) and expression levels of each RND efflux pump gene were quantified by qRT-PCR. Values are normalized to expression levels in untreated cultures and presented as fold-change. Data represent mean ± standard deviation from three independent experiments. Statistical significance was determined by one-way ANOVA with Dunnett’s multiple comparisons test comparing each treatment to the untreated control. *, P < 0.05; **, P < 0.01; ***, P < 0.001.
https://doi.org/10.1371/journal.ppat.1014031.s004
(TIF)
S1 Table. Strains, plasmids, and oligonucleotides used in this study.
Complete list of all bacterial strains, plasmids, and primers employed for molecular cloning, genetic manipulation, and quantitative reverse transcription-PCR analyses.
https://doi.org/10.1371/journal.ppat.1014031.s005
(XLSX)
S2 Table. Ectopic vexB expression restores antimicrobial resistance in ΔvexB mutants.
Complementation data for the five contemporary V. cholerae ΔvexB mutants expressing ectopic vexAB. Demonstrates restoration of antimicrobial resistance phenotypes in complemented mutant strains.
https://doi.org/10.1371/journal.ppat.1014031.s006
(XLSX)
S3 Table. Minimum inhibitory concentrations (MICs) of antimicrobial compounds against V. cholerae C6706 RND efflux pump mutants.
Antimicrobial resistance profiles of C6706 individual RND pump mutants and the Δ5RND quintuple pump mutants determined by broth microdilution assays against ampicillin, ciprofloxacin, erythromycin, kanamycin, and polymyxin B.
https://doi.org/10.1371/journal.ppat.1014031.s007
(XLSX)
S4 Table. Genetic variants conserved among the five Wave 3 clinical isolates identified by Breseq analysis.
Complete list of conserved genetic variants identified across all five contemporary V. cholerae isolates, including 121 nonsynonymous substitutions, 103 synonymous substitutions, and 55 additional variants.
https://doi.org/10.1371/journal.ppat.1014031.s008
(XLSX)
S5 Table. Genes within the 49.3 kb cryptic prophage deletion conserved across Wave 3 isolates.
Annotation and functional categorization of the 56 genes encompassed by the 49.3 kb chromosomal deletion (positions 1,494,671-1,543,947 on chromosome 1) that is present in reference strain C6706 but absent in all five contemporary Wave 3 isolates. Includes predicted gene functions and genomic coordinates.
https://doi.org/10.1371/journal.ppat.1014031.s009
(XLSX)
S6 Table. Strain metadata for isolates used to generate the phylogenetic tree in S3 Fig.
Metadata for the 67 V. cholerae O1 El Tor strains analyzed in the phylogenomic analysis, including strain identification, year of isolation, country of origin, and assignment to pandemic wave classification (Wave 1, Wave 2, or Wave 3).
https://doi.org/10.1371/journal.ppat.1014031.s010
(XLSX)
S7 Table. Pairwise SNP distance matrix derived from the core genome alignment of analyzed V. cholerae O1 El Tor strains in S3 Fig.
Complete symmetric matrix of pairwise single nucleotide polymorphism distances calculated from the 3,863,204-bp core genome alignment. Distances represent the number of SNPs separating each strain pair. Associated metadata columns include year of isolation, country of origin, and pandemic wave classification for each strain.
https://doi.org/10.1371/journal.ppat.1014031.s011
(XLSX)
References
- 1. GBD 2021 Antimicrobial Resistance Collaborators. Global burden of bacterial antimicrobial resistance 1990-2021: a systematic analysis with forecasts to 2050. Lancet. 2024;404(10459):1199–226. pmid:39299261
- 2. Kavanagh K. The rise of “nightmare bacteria”: antimicrobial resistance in five charts. Nature. 2025;646(8085):526–7. pmid:41057704
- 3. Ahmad N, Joji RM, Kumar V, Pagani L, Shahid M. Editorial: evolution, molecular mechanisms and the strategies to combat antimicrobial resistance (AMR): a One Health approach. Front Cell Infect Microbiol. 2025;15:1582613. pmid:40529308
- 4. Nazir A, Nazir A, Zuhair V, Aman S, Sadiq SUR, Hasan AH, et al. The global challenge of antimicrobial resistance: mechanisms, case studies, and mitigation approaches. Health Sci Rep. 2025;8(7):e71077. pmid:40704322
- 5. Geisinger E, Isberg RR. Interplay between antibiotic resistance and virulence during disease promoted by multidrug-resistant bacteria. J Infect Dis. 2017;215(suppl_1):S9–17. pmid:28375515
- 6. Basra P, Alsaadi A, Bernal-Astrain G, O’Sullivan ML, Hazlett B, Clarke LM, et al. Fitness tradeoffs of antibiotic resistance in extraintestinal pathogenic Escherichia coli. Genome Biol Evol. 2018;10(2):667–79. pmid:29432584
- 7. Fuzi M. The fitness connection of antibiotic resistance. Front Microbiol. 2025;16:1556656. pmid:40276228
- 8. Vogwill T, MacLean RC. The genetic basis of the fitness costs of antimicrobial resistance: a meta-analysis approach. Evol Appl. 2015;8(3):284–95. pmid:25861386
- 9. Jordana-Lluch E, Barceló IM, Escobar-Salom M, Estévez MA, Zamorano L, Gómez-Zorrilla S, et al. The balance between antibiotic resistance and fitness/virulence in Pseudomonas aeruginosa: an update on basic knowledge and fundamental research. Front Microbiol. 2023;14:1270999. pmid:37840717
- 10. Xia P, Yi M, Yuan Y, Huang J, Yang B, Liao J, et al. Coexistence of multidrug resistance and virulence in a single conjugative plasmid from a hypervirulent Klebsiella pneumoniae isolate of sequence type 25. mSphere. 2022;7(6):e0047722. pmid:36472445
- 11. Beceiro A, Tomás M, Bou G. Antimicrobial resistance and virulence: a successful or deleterious association in the bacterial world?. Clin Microbiol Rev. 2013;26(2):185–230. pmid:23554414
- 12. Kochan TJ, Nozick SH, Valdes A, Mitra SD, Cheung BH, Lebrun-Corbin M, et al. Klebsiella pneumoniae clinical isolates with features of both multidrug-resistance and hypervirulence have unexpectedly low virulence. Nat Commun. 2023;14(1):7962. pmid:38042959
- 13. Dimitriu T. Evolution of horizontal transmission in antimicrobial resistance plasmids. Microbiol (Reading). 2022;168(7). pmid:35849537
- 14. Partridge SR, Kwong SM, Firth N, Jensen SO. Mobile genetic elements associated with antimicrobial resistance. Clin Microbiol Rev. 2018;31(4):e00088-17. pmid:30068738
- 15. Cox G, Wright GD. Intrinsic antibiotic resistance: mechanisms, origins, challenges and solutions. Int J Med Microbiol. 2013;303(6–7):287–92. pmid:23499305
- 16. Nikaido H, Pagès J-M. Broad-specificity efflux pumps and their role in multidrug resistance of Gram-negative bacteria. FEMS Microbiol Rev. 2012;36(2):340–63. pmid:21707670
- 17. Nanjan P, Bose V. Efflux-mediated multidrug resistance in critical gram-negative bacteria and natural efflux pump inhibitors. Curr Drug Res Rev. 2024;16(3):349–68. pmid:38288795
- 18. Piddock LJV. Multidrug-resistance efflux pumps - not just for resistance. Nat Rev Microbiol. 2006;4(8):629–36. pmid:16845433
- 19. Alvarez-Ortega C, Olivares J, Martinez JL. RND multidrug efflux pumps: what are they good for?. Front Microbiol. 2013;4:7. pmid:23386844
- 20. Rampioni G, Pillai CR, Longo F, Bondì R, Baldelli V, Messina M, et al. Effect of efflux pump inhibition on Pseudomonas aeruginosa transcriptome and virulence. Sci Rep. 2017;7(1):11392. pmid:28900249
- 21. Nishino K, Latifi T, Groisman EA. Virulence and drug resistance roles of multidrug efflux systems of Salmonella enterica serovar Typhimurium. Mol Microbiol. 2006;59(1):126–41. pmid:16359323
- 22. Bina XR, Bina JE. Vibrio cholerae RND efflux systems: mediators of stress responses, colonization and pathogenesis. Front Cell Infect Microbiol. 2023;13:1203487. pmid:37256112
- 23. Vieira A, Ramesh A, Seddon AM, Karlyshev AV. CmeABC multidrug efflux pump contributes to antibiotic resistance and promotes Campylobacter jejuni survival and multiplication in Acanthamoeba polyphaga. Appl Environ Microbiol. 2017;83(22):e01600-17. pmid:28916560
- 24. Bina XR, Lavine CL, Miller MA, Bina JE. The AcrAB RND efflux system from the live vaccine strain of Francisella tularensis is a multiple drug efflux system that is required for virulence in mice. FEMS Microbiol Lett. 2008;279(2):226–33. pmid:18179581
- 25. Padilla E, Llobet E, Doménech-Sánchez A, Martínez-Martínez L, Bengoechea JA, Albertí S. Klebsiella pneumoniae AcrAB efflux pump contributes to antimicrobial resistance and virulence. Antimicrob Agents Chemother. 2010;54(1):177–83. pmid:19858254
- 26. Siamalube B, Ehinmitan E. Vibrio cholerae: understanding a persistent pathogen in Sub-Saharan Africa and the East mediterranean region. Pathog Dis. 2025;83:ftaf004. pmid:40145130
- 27. Ali M, Nelson AR, Lopez AL, Sack DA. Updated global burden of cholera in endemic countries. PLoS Negl Trop Dis. 2015;9(6):e0003832. pmid:26043000
- 28. Kaper JB, Morris JGJr, Levine MM. Cholera. Clin Microbiol Rev. 1995;8(1):48–86. pmid:7704895
- 29. Mutreja A, Kim DW, Thomson NR, Connor TR, Lee JH, Kariuki S, et al. Evidence for several waves of global transmission in the seventh cholera pandemic. Nature. 2011;477(7365):462–5. pmid:21866102
- 30. Das B, Verma J, Kumar P, Ghosh A, Ramamurthy T. Antibiotic resistance in Vibrio cholerae: Understanding the ecology of resistance genes and mechanisms. Vaccine. 2020;38 Suppl 1:A83–92. pmid:31272870
- 31. Son MS, Megli CJ, Kovacikova G, Qadri F, Taylor RK. Characterization of Vibrio cholerae O1 El Tor biotype variant clinical isolates from Bangladesh and Haiti, including a molecular genetic analysis of virulence genes. J Clin Microbiol. 2011;49(11):3739–49. pmid:21880975
- 32. Satchell KJF, Jones CJ, Wong J, Queen J, Agarwal S, Yildiz FH. Phenotypic analysis reveals that the 2010 haiti cholera epidemic is linked to a hypervirulent strain. Infect Immun. 2016;84(9):2473–81. pmid:27297393
- 33. Alam M, Islam A, Bhuiyan NA, Rahim N, Hossain A, Khan GY, et al. Clonal transmission, dual peak, and off-season cholera in Bangladesh. Infect Ecol Epidemiol. 2011;1:10.3402/iee.v1i0.7273. pmid:22957115
- 34. Bina XR, Provenzano D, Nguyen N, Bina JE. Vibrio cholerae RND family efflux systems are required for antimicrobial resistance, optimal virulence factor production, and colonization of the infant mouse small intestine. Infect Immun. 2008;76(8):3595–605. pmid:18490456
- 35. Bina JE, Mekalanos JJ. Vibrio cholerae tolC is required for bile resistance and colonization. Infect Immun. 2001;69(7):4681–5. pmid:11402016
- 36. Bina XR, Bina JE. The cyclic dipeptide cyclo(Phe-Pro) inhibits cholera toxin and toxin-coregulated pilus production in O1 El Tor Vibrio cholerae. J Bacteriol. 2010;192(14):3829–32. pmid:20453095
- 37. Bina XR, Philippart JA, Bina JE. Effect of the efflux inhibitors 1-(1-naphthylmethyl)-piperazine and phenyl-arginine-beta-naphthylamide on antimicrobial susceptibility and virulence factor production in Vibrio cholerae. J Antimicrob Chemother. 2009;63(1):103–8. pmid:19010827
- 38. Samanta P, Ghosh P, Chowdhury G, Ramamurthy T, Mukhopadhyay AK. Sensitivity to Polymyxin B in El Tor Vibrio cholerae O1 strain, Kolkata, India. Emerg Infect Dis. 2015;21(11):2100–2. pmid:26488385
- 39. Weill F-X, Domman D, Njamkepo E, Almesbahi AA, Naji M, Nasher SS, et al. Genomic insights into the 2016-2017 cholera epidemic in Yemen. Nature. 2019;565(7738):230–3. pmid:30602788
- 40. Samanta P, Mandal RS, Saha RN, Shaw S, Ghosh P, Dutta S, et al. A point mutation in carR Is involved in the emergence of polymyxin B-sensitive Vibrio cholerae O1 El Tor biotype by influencing gene transcription. Infect Immun. 2020;88(5):e00080-20. pmid:32094260
- 41. Samanta P, Saha RN, Chowdhury G, Naha A, Sarkar S, Dutta S, et al. Dissemination of newly emerged polymyxin B sensitive Vibrio cholerae O1 containing Haitian-like genetic traits in different parts of India. J Med Microbiol. 2018;67(9):1326–33. pmid:29927375
- 42. Hankins JV, Madsen JA, Giles DK, Brodbelt JS, Trent MS. Amino acid addition to Vibrio cholerae LPS establishes a link between surface remodeling in gram-positive and gram-negative bacteria. Proc Natl Acad Sci U S A. 2012;109(22):8722–7. pmid:22589301
- 43. Henderson JC, Fage CD, Cannon JR, Brodbelt JS, Keatinge-Clay AT, Trent MS. Antimicrobial peptide resistance of Vibrio cholerae results from an LPS modification pathway related to nonribosomal peptide synthetases. ACS Chem Biol. 2014;9(10):2382–92. pmid:25068415
- 44. Bilecen K, Fong JCN, Cheng A, Jones CJ, Zamorano-Sánchez D, Yildiz FH. Polymyxin B resistance and biofilm formation in Vibrio cholerae are controlled by the response regulator CarR. Infect Immun. 2015;83(3):1199–209. pmid:25583523
- 45. Hsiao A, Zhu J. Pathogenicity and virulence regulation of Vibrio cholerae at the interface of host-gut microbiome interactions. Virulence. 2020;11(1):1582–99. pmid:33172314
- 46. Miller VL, Mekalanos JJ. Synthesis of cholera toxin is positively regulated at the transcriptional level by toxR. Proc Natl Acad Sci U S A. 1984;81(11):3471–5. pmid:6374658
- 47. Häse CC, Mekalanos JJ. TcpP protein is a positive regulator of virulence gene expression in Vibrio cholerae. Proc Natl Acad Sci U S A. 1998;95(2):730–4. pmid:9435261
- 48. Higgins DE, DiRita VJ. Transcriptional control of toxT, a regulatory gene in the ToxR regulon of Vibrio cholerae. Mol Microbiol. 1994;14(1):17–29. pmid:7830555
- 49. Kovacikova G, Skorupski K. A Vibrio cholerae LysR homolog, AphB, cooperates with AphA at the tcpPH promoter to activate expression of the ToxR virulence cascade. J Bacteriol. 1999;181(14):4250–6. pmid:10400582
- 50. Skorupski K, Taylor RK. A new level in the Vibrio cholerae ToxR virulence cascade: AphA is required for transcriptional activation of the tcpPH operon. Mol Microbiol. 1999;31(3):763–71. pmid:10048021
- 51. Liu Z, Yang M, Peterfreund GL, Tsou AM, Selamoglu N, Daldal F, et al. Vibrio cholerae anaerobic induction of virulence gene expression is controlled by thiol-based switches of virulence regulator AphB. Proc Natl Acad Sci U S A. 2011;108(2):810–5. pmid:21187377
- 52. Bina XR, Taylor DL, Vikram A, Ante VM, Bina JE. Vibrio cholerae ToxR downregulates virulence factor production in response to cyclo(Phe-Pro). mBio. 2013;4(5):e00366-13. pmid:23982069
- 53. Mey AR, Midgett CR, Kull FJ, Payne SM. Vibrio cholerae CsrA controls ToxR levels by increasing the stability and translation of toxR mRNA. mBio. 2024;15(12):e0285324. pmid:39555915
- 54. Almagro-Moreno S, Kim TK, Skorupski K, Taylor RK. Proteolysis of virulence regulator ToxR is associated with entry of Vibrio cholerae into a dormant state. PLoS Genet. 2015;11(4):e1005145. pmid:25849031
- 55. Fengler VHI, Boritsch EC, Tutz S, Seper A, Ebner H, Roier S, et al. Disulfide bond formation and ToxR activity in Vibrio cholerae. PLoS One. 2012;7(10):e47756. pmid:23144706
- 56. Midgett CR, Swindell RA, Pellegrini M, Jon Kull F. A disulfide constrains the ToxR periplasmic domain structure, altering its interactions with ToxS and bile-salts. Sci Rep. 2020;10(1):9002. pmid:32488093
- 57. Bina JE, Provenzano D, Wang C, Bina XR, Mekalanos JJ. Characterization of the Vibrio cholerae vexAB and vexCD efflux systems. Arch Microbiol. 2006;186(3):171–81. pmid:16804679
- 58. Cerda-Maira FA, Ringelberg CS, Taylor RK. The bile response repressor BreR regulates expression of the Vibrio cholerae breAB efflux system operon. J Bacteriol. 2008;190(22):7441–52. pmid:18776020
- 59. Zhou Y, Yu L, Li J, Zhang L, Tong Y, Kan B. Accumulation of mutations in DNA gyrase and topoisomerase IV genes contributes to fluoroquinolone resistance in Vibrio cholerae O139 strains. Int J Antimicrob Agents. 2013;42(1):72–5. pmid:23643392
- 60. Barrick JE, Colburn G, Deatherage DE, Traverse CC, Strand MD, Borges JJ, et al. Identifying structural variation in haploid microbial genomes from short-read resequencing data using breseq. BMC Genomics. 2014;15(1):1039. pmid:25432719
- 61. Ghosh P, Naha A, Pazhani GP, Ramamurthy T, Mukhopadhyay AK. Genetic traits of Vibrio cholerae O1 Haitian isolates that are absent in contemporary strains from Kolkata, India. PLoS One. 2014;9(11):e112973. pmid:25415339
- 62. Naha A, Mandal RS, Samanta P, Saha RN, Shaw S, Ghosh A, et al. Deciphering the possible role of ctxB7 allele on higher production of cholera toxin by Haitian variant Vibrio cholerae O1. PLoS Negl Trop Dis. 2020;14(4):e0008128. pmid:32236098
- 63. Dolores J, Satchell KJF. Analysis of Vibrio cholerae genome sequences reveals unique rtxA variants in environmental strains and an rtxA-null mutation in recent altered El Tor isolates. mBio. 2013;4(2):e00624. pmid:23592265
- 64. Zhou Y-Y, Ma L-Y, Yu L, Lu X, Liang W-L, Kan B, et al. Quinolone resistance genes and their contribution to resistance in vibrio cholerae serogroup O139. Antibiotics. 2023;12(2):416.
- 65. Xu Y, Zheng Z, Sun R, Ye L, Chan EW-C, Chen S. Epidemiological and genetic characterization of multidrug-resistant non-O1 and non-O139 Vibrio cholerae from food in southern China. Int J Food Microbiol. 2024;418:110734. pmid:38759293
- 66. Mathieu-Denoncourt A, Whitfield GB, Vincent AT, Berne C, Pauzé-Foixet J, Mahieddine FC, et al. The carRS-ompV-virK operon of Vibrio cholerae senses antimicrobial peptides and activates the expression of multiple resistance systems. Sci Rep. 2025;15(1):13686. pmid:40258937
- 67. Carignan BM, Brumfield KD, Son MS. Single nucleotide polymorphisms in regulator-encoding genes have an additive effect on virulence gene expression in a Vibrio cholerae clinical isolate. mSphere. 2016;1(5):e00253-16. pmid:27668288
- 68. Conner JG, Zamorano-Sánchez D, Park JH, Sondermann H, Yildiz FH. The ins and outs of cyclic di-GMP signaling in Vibrio cholerae. Curr Opin Microbiol. 2017;36:20–9. pmid:28171809
- 69. Zamorano-Sanchez D, Xian W, Lee CK, Salinas M, Thongsomboon W, Cegelski L. Functional specialization in Vibrio cholerae diguanylate cyclases: distinct modes of motility suppression and c-di-GMP production. mBio. 2019;10(2). pmid:31015332
- 70. Wu DC, Zamorano-Sánchez D, Pagliai FA, Park JH, Floyd KA, Lee CK, et al. Reciprocal c-di-GMP signaling: incomplete flagellum biogenesis triggers c-di-GMP signaling pathways that promote biofilm formation. PLoS Genet. 2020;16(3):e1008703. pmid:32176702
- 71. Ayala JC, Silva AJ, Benitez JA. H-NS: an overarching regulator of the Vibrio cholerae life cycle. Res Microbiol. 2017;168(1):16–25. pmid:27492955
- 72. Salma U, Nishimura Y, Tokumura M, Hossain A, Watanabe K, Noro K, et al. Occurrence, seasonal variation, and environmental risk of multiclass antibiotics in the urban surface water of the Buriganga River, Bangladesh. Chemosphere. 2025;370:143956. pmid:39675576
- 73. Maghsodian Z, Sanati AM, Mashifana T, Sillanpää M, Feng S, Nhat T, et al. Occurrence and distribution of antibiotics in the water, sediment, and biota of freshwater and marine environments: a review. Antibiotics (Basel). 2022;11(11):1461. pmid:36358116
- 74. Bina XR, Bina JE. The cyclic dipeptide cyclo(Phe-Pro) inhibits cholera toxin and toxin-coregulated pilus production in O1 El Tor Vibrio cholerae. J Bacteriol. 2010;192(14):3829–32. pmid:20453095
- 75. Howard MF, Bina XR, Bina JE. Indole inhibits ToxR regulon expression in Vibrio cholerae. Infect Immun. 2019;87(3):e00776-18. pmid:30617203
- 76. Secchi E, Vitale A, Miño GL, Kantsler V, Eberl L, Rusconi R, et al. The effect of flow on swimming bacteria controls the initial colonization of curved surfaces. Nat Commun. 2020;11(1):2851. pmid:32503979
- 77. Millet YA, Alvarez D, Ringgaard S, von Andrian UH, Davis BM, Waldor MK. Insights into Vibrio cholerae intestinal colonization from monitoring fluorescently labeled bacteria. PLoS Pathog. 2014;10(10):e1004405. pmid:25275396
- 78. Butler SM, Camilli A. Both chemotaxis and net motility greatly influence the infectivity of Vibrio cholerae. Proc Natl Acad Sci U S A. 2004;101(14):5018–23. pmid:15037750
- 79. Echazarreta MA, Klose KE. Vibrio flagellar synthesis. Front Cell Infect Microbiol. 2019;9:131.
- 80. Butler SM, Camilli A. Going against the grain: chemotaxis and infection in Vibrio cholerae. Nat Rev Microbiol. 2005;3(8):611–20. pmid:16012515
- 81. Pérez-Varela M, Corral J, Aranda J, Barbé J. Roles of efflux pumps from different superfamilies in the surface-associated motility and virulence of Acinetobacter baumannii ATCC 17978. Antimicrob Agents Chemother. 2019;63(3):e02190-18. pmid:30642939
- 82. Ruiz C, Levy SB. Regulation of acrAB expression by cellular metabolites in Escherichia coli. J Antimicrob Chemother. 2014;69(2):390–9. pmid:24043404
- 83. Maldonado J, Czarnecka B, Harmon DE, Ruiz C. The multidrug efflux pump regulator AcrR directly represses motility in Escherichia coli. mSphere. 2023;8(5):e0043023. pmid:37787551
- 84. Vaillancourt M, Limsuwannarot SP, Bresee C, Poopalarajah R, Jorth P. Pseudomonas aeruginosa mexR and mexEF antibiotic efflux pump variants exhibit increased virulence. Antibiotics (Basel). 2021;10(10). pmid:34680745
- 85. Tischler AD, Camilli A. Cyclic diguanylate (c-di-GMP) regulates Vibrio cholerae biofilm formation. Mol Microbiol. 2004;53(3):857–69. pmid:15255898
- 86. Tischler AD, Camilli A. Cyclic diguanylate regulates Vibrio cholerae virulence gene expression. Infect Immun. 2005;73(9):5873–82. pmid:16113306
- 87. Mukhopadhyay AK, Takeda Y, Balakrish Nair G. Cholera outbreaks in the El Tor biotype era and the impact of the new El Tor variants. Curr Top Microbiol Immunol. 2014;379:17–47. pmid:24710767
- 88. Montero DA, Vidal RM, Velasco J, George S, Lucero Y, Gómez LA, et al. Vibrio cholerae, classification, pathogenesis, immune response, and trends in vaccine development. Front Med (Lausanne). 2023;10:1155751. pmid:37215733
- 89. Heidelberg JF, Eisen JA, Nelson WC, Clayton RA, Gwinn ML, Dodson RJ, et al. DNA sequence of both chromosomes of the cholera pathogen Vibrio cholerae. Nature. 2000;406(6795):477–83. pmid:10952301
- 90. Antipov D, Hartwick N, Shen M, Raiko M, Lapidus A, Pevzner PA. plasmidSPAdes: assembling plasmids from whole genome sequencing data. Bioinformatics. 2016;32(22):3380–7. pmid:27466620
- 91. Feldgarden M, Brover V, Gonzalez-Escalona N, Frye JG, Haendiges J, Haft DH, et al. AMRFinderPlus and the reference gene catalog facilitate examination of the genomic links among antimicrobial resistance, stress response, and virulence. Sci Rep. 2021;11(1):12728. pmid:34135355
- 92. Bina XR, Howard MF, Taylor-Mulneix DL, Ante VM, Kunkle DE, Bina JE. The Vibrio cholerae RND efflux systems impact virulence factor production and adaptive responses via periplasmic sensor proteins. PLoS Pathog. 2018;14(1):e1006804. pmid:29304169
- 93. Schild S, Tamayo R, Nelson EJ, Qadri F, Calderwood SB, Camilli A. Genes induced late in infection increase fitness of Vibrio cholerae after release into the environment. Cell Host Microbe. 2007;2(4):264–77. pmid:18005744
- 94. Nelson EJ, Chowdhury A, Flynn J, Schild S, Bourassa L, Shao Y, et al. Transmission of Vibrio cholerae is antagonized by lytic phage and entry into the aquatic environment. PLoS Pathog. 2008;4(10):e1000187. pmid:18949027
- 95. Nguyen L-T, Schmidt HA, von Haeseler A, Minh BQ. IQ-TREE: a fast and effective stochastic algorithm for estimating maximum-likelihood phylogenies. Mol Biol Evol. 2015;32(1):268–74. pmid:25371430
- 96. Kalyaanamoorthy S, Minh BQ, Wong TKF, von Haeseler A, Jermiin LS. ModelFinder: fast model selection for accurate phylogenetic estimates. Nat Methods. 2017;14(6):587–9. pmid:28481363
- 97. Letunic I, Bork P. Interactive Tree Of Life (iTOL) v5: an online tool for phylogenetic tree display and annotation. Nucleic Acids Res. 2021;49(W1):W293–6. pmid:33885785