Transcriptional profiling of Vibrio cholerae O1 following exposure to human anti- lipopolysaccharide monoclonal antibodies

ABSTRACT

Following an episode of cholera, a rapidly dehydrating, watery diarrhea caused by the Gram-negative bacterium, Vibrio cholerae O1, humans mount a robust anti-lipopolysaccharide (LPS) antibody response that is associated with immunity to subsequent re-infection. In neonatal mouse and rabbit models of cholera, passively administered anti-LPS polyclonal and monoclonal (MAb) antibodies reduce V. cholerae colonization of the intestinal epithelia by inhibiting bacterial motility and promoting vibrio agglutination. Here we demonstrate that human anti-LPS IgG MAbs also arrest V. cholerae motility and induce bacterial paralysis. A subset of those MAbs also triggered V. cholerae to secrete an extracellular matrix (ECM). To identify changes in gene expression that accompany antibody exposure and that may account for motility arrest and ECM production, we subjected V. cholerae O1 El Tor to RNA-seq analysis after treatment with ZAC-3 IgG, a high affinity MAb directed against the core/lipid A region of LPS. We identified > 160 genes whose expression was altered following ZAC-3 IgG treatment, although canonical outer membrane stress regulons were not among them. ompS (VCA1028), a porin associated with virulence and indirectly regulated by ToxT, and norR (VCA0182), a σ54-dependent transcription factor involved in late stages of infection, were two upregulated genes worth noting.

enteric, immunity, cholera, antibody, vaccine, mucosal, transcription

INTRODUCTION

The Gram-negative bacterium, Vibrio cholerae serotype O1, is the causative agent of cholera, a profuse watery diarrhea that can be fatal in children and adults in the absence of intravenous fluid rehydration therapy. V. cholerae O1 strains are grouped into two biotypes, classical, the causative agent of the first six pandemics, and El Tor, the primary etiological agent of the seventh (and ongoing) pandemic (Clemens et al. 2017). The bacterium, which is transmitted to humans via ingestion of contaminated food or water, colonizes the small intestine through a series of coordinated events. Mucus penetration is facilitated by a single polar flagellum and expression of mucinases, while the so-called toxin co-regulated pilus (TCP) is involved in inter-bacterial attachment and micro-colony formation; subsequent attachment to the intestinal epithelium is mediated by adhesins (Rhine and Taylor 1994; Krebs and Taylor 2011). Following epithelial attachment and microcolony formation, the bacterium secretes cholera toxin (CT), which preferentially targets enterocytes and is ultimately responsible for the onset of fluid loss that is the hallmark of the disease (Harris et al. 2012). Although the classical and El Tor biotypes each produce CT, they differ in disease severity, virulence gene regulation, sensitivity to cationic antimicrobial peptides (CAMPs) and their tendency to form biofilms (Pradhan et al. 2010; Ayala et al. 2018).

Following a bout of cholera, humans develop serotype-specific immunity consisting of mucosal IgA and serum IgG (and IgM) antibodies against two predominant antigens: CT and lipopolysaccharide (LPS) (Kauffman et al. 2016). Although anti-CT antibodies have toxin-neutralizing activity in vitro, the LPS-specific response is considered the main determinant of protection against reinfection. The O-antigen or O-specific polysaccharide (OSP) of the V. cholerae O1 serogroup consists of (1→2)-linked moieties of 4-amino-4,6-dideoxy-alpha-D-mannopyranose (D-perosamine) in which the amino groups are acylated with 3-deoxy-L-glycero-tetronic acid (Kenne et al. 1982). Within the O1 serogroup, there are two predominant serotypes, Ogawa and Inaba, that differ in the presence (Ogawa) or absence (Inaba) of methylation of the terminal non-reducing D-perosaminyl moiety. The OSP side chain is anchored to the outer leaflet of V. cholerae’s outer membrane (OM) via a conserved core/lipid A moiety (Chatterjee and Chaudhuri 2003).

In the neonatal mouse model of cholera, anti-LPS polyclonal and monoclonal IgA and IgG (MAb) antibodies reduce the ability of V. cholerae O1 to colonize the intestinal epithelium by several orders of magnitude (Apter et al. 1993; Bishop et al. 2010; Weil, Becker and Harris 2019). A total of two mechanisms have been proposed to explain how these antibodies interfere intestinal colonization: motility arrest and agglutination. Upon exposure to anti-LPS polyclonal or monoclonal IgG and IgA antibodies in liquid medium, individual bacteria become completely paralyzed and stop swimming within minutes (Bishop et al. 2010; Levinson, De Jesus and Mantis 2015; Levinson, Baranova and Mantis 2016; Wang, Lazinski and Camilli 2016). Soon thereafter, micro- and macro-agglutination occurs as a result of cell-cell collisions and flagellum entanglements (Levinson, De Jesus and Mantis 2015). It should be noted that V. cholerae is unique in that its polar flagellum is sheathed with LPS and, therefore, a target of anti-LPS antibodies.

While two studies from the laboratory of Andrew Camilli have made a compelling case that motility arrest accounts for the attenuation of V. cholerae by anti-LPS antibodies in the neonatal mouse model (Bishop et al. 2010; Wang, Lazinski and Camilli 2016), we postulate that additional factors may be at play. For example, classical and El Tor strains of V. cholerae O1 secrete an extracellular matrix (ECM) following exposure to ZAC-3 IgG, a gut-derived mouse MAb directed against the core/lipid A region of LPS (Levinson, Baranova and Mantis 2016; Baranova, Levinson and Mantis 2018; Baranova et al. 2020). The ECM is enriched in OSP and appears to form a capsule that renders the bacterium resistant to secondary insults like complement-mediated lysis. ECM production is triggered by the combination of motility arrest and agglutination. Although no specific signal transduction pathway has been associated with the onset of ECM production, the fact that ZAC-3 IgG-treated cells display surface blebs has led us to speculate that one or more canonical extra cytoplasmic and/or OM stress response pathways may be activated.

In this report, we sought to better understand the responses of classical and El Tor strains of V. cholerae O1 to ZAC-3 IgG that may account for their competitive disadvantage within the gut environment. We were particularly intrigued with the hypothesis that rapid motility arrest induced by ZAC-3 IgG results in a concomitant change in gene expression reflecting a transition from motile to non-motile phenotypes.

MATERIALS AND METHODS

Bacterial strains and culture conditions

Bacterial strains used in this study are shown in Table S1 (Supporting Information). Bacteria were cultured in non-toxin inducing conditions (non-TIC): LB medium (pH 7) at 37°C with aeration. For the classical strain O395, toxin-inducing conditions (TIC) were defined as incubation at 30°C in LB medium (pH 6.5) in an orbital shaking incubator (200 rpm; Miller and Mekalanos 1988). For El Tor strains, TIC were prepared as follows, for each technical replicate a single colony was inoculated into 10 mL AKI medium, (0.5% NaCl, 0.3% NaHCO3, 0.4% yeast extract and 1.5% Bacto-Peptone) and incubated at 37°C in static conditions for 3.5 h. After 3.5 h, 7 mL were removed from the tube and the remaining 3 mL were placed in a 37°C incubator at 200 rpm for 4 h, at which time experiments were performed (Iwanaga et al. 1986). As necessary, media were supplemented with streptomycin (100 μg/mL), gentamicin (10 μg/mL) or chloramphenicol (2 μg/mL).

Antibodies

ZAC-3 IgG1 and the isotype control, SyH7, against ricin toxin were produced in the Nicotiana benthamiana-based rapid antibody-manufacturing platform (RAMP), as described (Sully et al. 2014; Levinson et al. 2015). Rabbit anti-TcpA antiserum was kindly provided by Dr. Ron Taylor (Taylor et al. 2004).

RNA-seq library preparation

Overnight cultures of V. cholerae C6706 were grown in LB broth at 37°C on a rotatory shaker (200 rpm). The following day, subcultures were grown to early log phase (OD₆₀₀ of 0.3–0.4) before being treated with ZAC-3, or SyH7 IgG (9 μg/mL) for 60 min. RNA-seq libraries were prepared as previously described using cells collected from two independent induction experiments (Fitzgerald, Bonocora and Wade 2014). Total RNA was extracted and purified using a modified hot phenol method (Stringer et al. 2014). Resulting RNA was treated with DNase (TURBO DNA-free kit, Life Technologies, Carlsbad, CA) for 45 min at 37°C and then phenol extracted and precipitated with ethanol. The RiboZero kit (Epicentre, Madison, WI) was used to remove ribosomal RNA from the preps and cDNA libraries were constructed using the ScriptSeq kit (Epicentre, Madison, WI). Single read sequencing was done on the Illumina Hi-Seq 2500 platform at the Genomics and Bioinformatics Core at the University at Buffalo. Two independent libraries were constructed for each antibody treatment with RNA collected from two independent induction experiments.

RNA-seq analysis

Sequence reads were mapped and visually checked for library quality using Signal Map (NimbleGen Systems, Madison WI). Differential expression analysis was performed using Rockhopper utilizing the V. cholerae N16961 reference genome (McClure et al. 2013; Tjaden 2015) set to the default parameters and genes were considered significantly differentially expressed if the q-value was < 0.01 and there was ≥ 2log2(fold change) compared to both the SyH7 IgG treated control groups. q-values are p-values that have been adjusted for false discovery rate using the Benjamini-Hochberg Procedure. Predicted RNAs, rRNAs and tRNAs were excluded from analysis.

Generation of V. cholerae ompS and norR deletion strains

Unmarked and in-frame ompS (VCA1028) and norR (VCA0182) gene deletion strains were engineered into the V. cholerae classical (O395) and El Tor (C6706) backgrounds through allelic exchange using the suicide vector, pCVD442 (Donnenberg and Kaper 1991). Allelic exchange plasmids were generated essentially as described for pGW65 using primers listed in Table S2 (Supporting Information; Willsey et al. 2018). Briefly, ∼1 bp regions immediately upstream and downstream of the ompS and norR open reading frames (ORFs) were amplified from V. cholerae C6706 and O395 genomic DNA using Q5 polymerase (NEB) and primers listed in Table S2 (Supporting Information), resulting in the V. cholerae O395 and C6706 ompS allelic exchange plasmids, pGW93 and pGW94, and V. cholerae O395 norR allelic exchange plasmid, pGW117. Each allelic exchange plasmid was transformed into E. coli S17-1 by electroporation and mobilized into V. cholerae C6706 or O395 through conjugative mating (Simon, Priefer and Pühler 1983). Following overnight incubation at 37°C, V. cholerae merodiploids were selected on TCBS agar (Millepore-Sigma, Darmstadt, Germany) supplemented with carbenicillin (100 µg/mL). Sucrose-resistant, carbenicillin-sensitive colonies were then screened for the absence of ompS or norR via colony PCR using Taq Quickload (NEB) using the screening primer sets shown in Table S2 (Supporting Information).

Construction of V. cholerae P_ompStranscriptional reporter strains

A P_ompS-lacZ transcriptional reporter plasmid was created to examine ompS promoter activity in V. cholerae O395 and C6706 resulting from treatment with ZAC-3. The intergenic region immediately upstream of the ompS ORF is identical in sequence in V. cholerae O395 and C6706, therefore only one P_ompS transcriptional reporter plasmid was engineered. Briefly, the ∼400 base non-coding region immediately upstream of the ompS Shine-Dalgarno sequence was amplified from V. cholerae O395 genomic DNA using Q5 polymerase (NEB) and the ompS primer sets shown in Table S2 (Supporting Information). The resulting DNA amplicon was gel purified via GeneJet Gel Extraction kit (ThermoFisher, Waltham, MA), digested with BamHI and KpnI (NEB) and ligated into the promoter-less -lacZYA transcriptional reporter vector, pMW5 (Wargo et al. 2009). This plasmid was then electrotransformed into V. cholerae O395 and C6706.

β-galactosidase assays

β-galactosidase assays were done as previously described (Stringer et al. 2014). Overnight V. cholerae cultures were diluted 1:100 into fresh LB broth with appropriate antibiotics. For the ompS, norR, hmpA and nnrS studies bacteria were subcultured to mid-log phase (OD₆₀₀ ∼0.5) and then treated for 1 or 2 h as indicated with ZAC-3 IgG, 10 µg/mL. For V. cholerae O395 toxT:lacZ and tcpA studies, as well as C6706 tcpA studies in non-TIC, bacteria were subcultured to early mid-log phase and then treated with in the presence or absence of ZAC-3 IgG until cultures reached an OD600 ∼0.6. For C6706 tcpA studies done in TIC, bacteria were grown as described above in the presence or absence of antibody. We used 10 μg/mL of ZAC-3 for toxT and tcpA studies in the O395 strain and 20 μg/mL for C6706. This was done to account for the fact that TIC for C6706 necessitated growing the bacteria for 8 h in the presence of antibody, thus bacteria grew to a higher OD and concentration of antibody was increased.

Western blot analysis of TCP

Vibrio cholerae strains were grown to mid-log phase (OD₆₀₀ ∼0.6) in TIC or non-TIC in the absence or presence ZAC-3 IgG. Cells were collected by centrifugation, and the resulting bacterial pellet was resuspended in PBS and 2X Laemmli sample buffer (Bio-Rad, Hercules, CA). The samples were boiled for 10 min before being subjected to SDS-PAGE and transferred to nitrocellulose membrane, as described dx.doi.org/10.17504/protocols.io.bcgaitse. The membranes were probed with rabbit anti-TcpA antiserum (1:10 000 dilution), kindly provided by Dr. Ron Taylor (Taylor et al. 2004). Goat anti-rabbit IgG conjugated to HRP was utilized as a secondary (RRID: AB_2 819 160). The membranes were developed using the Pierce ECL Plus Western Blotting Substrate (Thermo Fisher Scientific, Cat. No. 32 132) and imaged using the iBright FL 1500 (Thermo Fisher Scientific, Waltham, MA). Image analysis was done using Fiji (ImageJ 1.52p) (Schindelin et al. 2012; Rueden et al. 2017).

GM₁ ELISA

Immunolon 4HBX 96-well microtiter ELISA plates were coated overnight with monosialoganglioside GM₁ (10 µg/mL) from bovine brain (Sigma-Aldrich Cat. No. 37 758-47-7). Plates were probed for 1 h with serial dilutions of cholera toxin (50 µg/mL; Millipore Sigma, Darmstadt, Germany, Cat. No. 227 036) or bacterial supernatants (3 mL) that had been concentrated (final volume, 150–300 µL) using an Amicon Ultra-4 Centrifugal Filter with Ultracel-10 Membrane (Millipore Sigma, Darmstadt, Germany, Cat. No. UFC801096). Captured CT was detected using a CT-specific human IgG mAb (AT13.1.B12; 1 µg/mL), kindly provided by Jens Wrammert (Emory University; Kauffman et al. 2016), followed by a goat anti-human IgG (H + L) cross-adsorbed to HRP (AB_228 265). Plates were developed using SureBlue Microwell Peroxidase Substrate and analyzed using a Spectramax 250 spectrophotometer with Softmax Pro 5.0 software (Molecular Devices, San Jose, CA).

Crystal violet (CV) assays

CV assays were done as described (Baranova, Levinson and Mantis 2018). Statistical significance across stains was determined by either one-way or two-way ANOVA, as stated in the figure legends, followed by a Tukey multiple comparison test, with GraphPad Prism version 8.3.0 for Windows (GraphPad Software, La Jolla CA).

Motility assays

Semi-solid agar (ssAgar) and liquid motility assays were done as described previously (Levinson, De Jesus and Mantis 2015; Bishop et al. 2010).

Neonatal mouse model of intestinal colonization

Neonatal mouse colonization studies were done essentially as described previously (Levinson, Baranova and Mantis 2016). Briefly, 3–5 day old BALB/c pups were gavaged with a 1:1 mixture of V. cholerae O395 derivative KKV598 (ΔlacZ) and either GGW350 (ΔompS) or GGW446 (ΔnorR). The pups were incubated for 24 h in a 30°C chamber and then euthanized. Mouse intestines were immediately removed and homogenized. There lysates were plated on LB agar containing streptomycin (100 μg/mL) and X-gal (40 μg/mL) for enumerating CFUs, as a surrogate for intestinal colonization (Klose 2000).

Mouse studies were conducted under strict compliance with the Wadsworth Center's Institutional Animal Care and Use Committee (IACUC). The Wadsworth Center complies with the Public Health Service Policy on Humane Care and Use of Laboratory Animals (Assurance #A3183-01). The Wadsworth Center is fully accredited by the Association for Assessment and Accreditation of Laboratory Animal Care (AAALAC). Obtaining this voluntary accreditation status reflects that Wadsworth Center's Animal Care and Use Program meets all standards required by law and goes beyond the standards as it strives to achieve excellence in animal care and use. Mice were euthanized by carbon dioxide asphyxiation followed by cervical dislocation, as recommended by the Office of Laboratory Animal Welfare (OLAW), National Institutes of Health.

RESULTS

Treatment of V. cholerae El Tor or classical strains with ZAC-3 IgG MAb arrests bacterial motility (< 5 min), promotes agglutination (30–60 min) and stimulates ECM production (30–60 min) (Baranova, Levinson and Mantis 2018; Levinson, Baranova and Mantis 2016). However, ZAC-3 IgG is unusual in that it targets the core/lipid A region of V. cholerae LPS. The vast majority anti-V. cholerae LPS antibodies are directed against OSP (Kauffman et al. 2016; Bishop et al. 2010; Aktar et al. 2016; Johnson et al. 2012). For example, all 24 human LPS-specific MAbs identified by Kauffman and colleagues from plasmablasts of cholera patients were specific for OSP (Ogawa and/or Inaba). While those MAbs displayed a range of V. cholerae agglutinating and vibriocidal activities (Table 1), it was not determined whether they also arrest bacterial motility and/or trigger ECM production.

Figure 2.

RNA-seq analysis and functional grouping of genes differentially regulated with ZAC-3 treatment in V. cholerae C6706. Early log-phase V. cholerae C6706 were grown in the presence of ZAC-3 IgG (9 µg/mL) or an isotype control for 1 h at 37°C with aeration. RNA-seq analysis was done as described in the Materials and Methods. (A) Expression values for untreated samples were plotted against ZAC-3 IgG-treated samples. Genes whose expression was affected by > 4-fold up (green) or down (red) are indicated (B) Categories and numbers of genes per category in V. cholerae C6706 that were affected by ZAC-3 IgG treatment.

Table 1.

In vitro properties of anti-OSP human MAbs on V. cholerae.

MAb	Target	Origina	EC50 (μg/mL)b	Agglut. (μg/mL)c	Arrestd	ECMe
CF29.1.C04	OSP	huIgG1	0.13	5	High	−
CF29.1.B02	OSP	huIgA2	0.16	2.5	High	+
CF29.1.G02	OSP	huIgA1	0.15	2.5	High	+
CF21.2.F01	OSP	huIgA2	0.11	5	High	+++

a

As reported in Kauffman et al (2016); ^b, binding activity to whole cells via ELISA, as reported by Kauffman et al.; ^c, minimum agglutinating concertation; ^d, motility arrest as measured in liquid medium with MAb @10 μg/mL and reported in Fig. 1; ^e, ECM production in response to MAbs@10 μg/mL.

To address these questions, four OSP-specific human IgG1 MAbs, kindly provided to us by Dr. Jens Wrammert (Emory University School of Medicine), were tested for their ability to bind to V. cholerae O395 and inhibit motility in liquid medium (Table 1; Fig. 1A, B; Figure S1, Supporting Information). Briefly, V. cholerae O395 expressing mCherry were spotted onto microscope slides after treatment with each individual mAb (10 μg/mL). Three second exposures were taken at 5 min intervals, and fluorescent ‘tails’ were scored as motile cells. We found that treatment of V. cholerae with MAbs CF29.1.C04, CF29.1.B02, CF29.1.G02 and CF21.2.F02, resulted complete arrest of bacterial motility at 5 min post-treatment.

Figure 1.

Effects of anti-OSP IgG human MAbs on V. cholerae O395 motility and ECM production. (A) Reactivity of anti-OSP IgG human MAbs with V. cholerae O395 by whole cell ELISA. Data are representative of two biological replicates, each with two technical replicates. (B) Motility of V. cholerae DB303 (O395 expressing mCherry from pGW104) in liquid medium after 5 min treatment with indicated MAbs (10 μg/mL). Each column represents the average of three biological replicates, each with two technical replicates. All anti-LPS MAbs significantly inhibited bacterial motility, as compared to untreated cells (Unt.) (C) Microtiter wells with V. cholerae O395 were treated with indicated MAbs at indicated concentrations (40–0.6 μg/mL) for 1 h with aeration then stained with CV. For each antibody the bottom bar represents untreated cells, the next bar represents 40 μg/mL and then each above bar represents a 2-fold dilution of that antibody as denoted by color of outlines. Each column represents the average of two biological replicates, each with three technical replicates. Significance was determined for B and C by two-way ANOVA followed by Tukey's multiple comparison test. *; P < 0.05; ns, not significant.

To examine ECM production, strains of V. cholerae O395 were incubated with ZAC-3 IgG or the four anti-OSP MAbs (10 μg/mL) for 1 h at 37⁰C with aeration. The microtiter plates were then treated with CV, which serves as an indirect indicator of ECM deposition. Three of the four MAbs stimulated V. cholerae ECM production to varying degrees (low to medium), although none was as potent as ZAC-3 IgG (Fig. 1C). CF21.2.F01 IgG induced mild, dose-dependent ECM production. CF21.2.F01 also had the lowest EC₅₀ among the four human MAbs. MAbs CF29.1.B02 and CF29.1.G02 induced low but significant ECM at the highest concentrations tested, but did display a demonstrable dose response, possibly because they were under a critical binding (avidity) threshold. Finally, CF29.1.C04 IgG did not trigger V. cholerae O395 to secrete any detectable levels of ECM, even though the same MAb caused motility arrest and macroscopic agglutination (Table 1). In summary, our results demonstrate that human-derived anti-OSP MAbs expressed as IgG1 constructs are potent inhibitors of V. cholerae motility and a subset are capable of stimulating ECM production.

Impact of ZAC-3 IgG on V. cholerae O1 transcription

We have postulated that V. cholerae reprograms gene expression in response to anti-LPS antibody exposure (especially ZAC-3 IgG) to adapt to abrupt changes in motility, cell–cell contact (i.e. agglutination), and possibly OM stress. Indeed, elucidating the transcriptional response of V. cholerae to ZAC-3 IgG may reveal insights into the mechanism(s) underlying antibody-mediated protection. To investigate this hypothesis, mid-log phase cultures of V. cholerae El Tor strain C6706 grown in LB medium were treated with ZAC-3 IgG (10 μg/mL) for 60 min, and then subjected to RNA-seq analysis, as described in the Materials and Methods. As a control, cells were incubated with an isotype matched Ab directed against an irrelevant antigen (SyH7 IgG1). We used the annotated genome of NC16961 as a reference (Heidelberg et al. 2000). The dataset was sorted for genes whose expression changed > 4-fold and had associated q values < 0.01. In total there were 164 genes that met these criteria: 71 whose expression was upregulated and 93 whose expression was downregulated (Table 2; Fig. 2; Appendix 1–3).

Table 2.

Partial list of V. cholerae C6706 genes identified by RNA-seq analysis whose expression changed following ZAC-3 IgG treatment.

Gene	Annotation	Δ log2a
VCA1028	Maltoporin (OmpS)	5.20
VC0911	Trehalose-6-phosphate hydrolase	4.22
VCA0247	Transcriptional repressor UlaR	4.01
VC1649	Serine protease (vesC)	3.83
VC0910	PTS system trehalose(maltose) transporter subunits IIBC	3.65
VCA0860	Alpha-amylase (MalS)	3.44
VCA0945	Maltose ABC transporter substrate-binding protein (MalE)	3.43
VCA0908	Hypothetical protein	3.4
VCA0944	Maltose transporter membrane protein (MalF)	3.26
VC0606	Nitrogen regulatory protein P-II	3.23
VCA0136	Glycerophosphodiester phosphodiesterase (GlpQ)	−3.54
VCA0137	sn-glycerol-3-phosphate transporter (GlpT)	−3.56
VC0089	Cytochrome c551 peroxidase	−3.58
VC1778	Hypothetical protein	−3.65
VC0795	Citrate/sodium symporter	−3.70
VC0797	Citrate lyase subunit gamma	−3.77
VC0796	Citrate (pro-3S)-lyase ligase	−3.79
VCA0684	Regulatory protein UhpC	−4.15
VCA0685	Hexose ABC transporter substrate-binding protein	−4.94
VCA0687	Hexose ATP-binding subunit	−5.11
VC0713	Hypothetical protein	−5.26
VCA0686	Hexose ABC transporter permease	−5.36

a

Genes with highest fold change in response to ZAC-3 IgG treatment, as compared to cells treated with an isotype control. Genes were whose expression changed > 2-fold log2 and with a q value < 0.01 were designated as significant.

One theme that emerged from the RNA-seq data set was that exposure to ZAC-3 IgG influenced expression of genes related to energy metabolism, sugar transport and even nitric oxide metabolism, all of which are peripherally associated with onset of hypoxia and even early stages of biofilm formation (Fig. 2; Bueno, Pinedo and Cava 2020). For example, the gene encoding maltoporin (VCA1028;OmpS) was upregulated more than 30-fold (see section below) and the VCA0685–0687 gene cluster, which encodes a hexose-6-phosphate ABC transport system, was downregulated ∼5-fold. It has been recently reported that the hexose-6-phosphate ABC transport system increases survival of V. cholerae in phosphate limiting conditions, possibly allowing for increased chances of survival when transmitted from human hosts to pond water (Moisi et al. 2013). Vibrio cholerae genes related to oxidative stress and iron homeostasis and known to be associated with OM damage also exhibited low level changes in response to ZAC-3 binding (Sikora et al. 2009). However, contrary to our predictions, the RNA-seq analysis was not consistent with activation of canonical OM stress response pathways like Cpx (Acosta, Pukatzki and Raivio 2015; Peschek et al. 2019).

We also observed downregulation of several chemotaxis related genes, although most did not reach significance (Appendices S1–S3; Boin, Austin and Häse 2004). Therefore, the RNA-seq dataset was not consistent with a model in which ZAC-3 IgG down regulates flagellar synthesis genes. Nonetheless, repression of chemotaxis or interference with metabolism genes could, in principle, dampen virulence of V. cholerae upon egress from a human host. V. cholerae shed from a mammalian host (human or animal) have been shown to be hyper-infective (Merrell et al. 2002); it is intriguing to speculate that anti-LPS antibodies in the intestinal lumen might render the bacteria ‘hypoinfective’. The dataset also raises the possibility that antibody-mediated motility arrest occurs, at least in part, at the post-transcriptional level (e.g. c-di-GMP) via a signaling pathway that functions to repress of flagellar synthesis (Correa et al. 2000; Prouty, Correa and Klose 2001).

Induction of V. cholerae ompS in response to ZAC-3 IgG

RNA-seq analyses did reveal that expression of the VCA1028 allele in V. cholerae C6706 was increased ∼30-fold following exposure to ZAC-3 IgG, making it the most upregulated gene in the dataset. VCA1028 encodes a maltoporin, a maltose-inducible (glucose-repressible) outer membrane protein, also known as OmpS (Lång and Palva 1987). It has been reported that an intact maltose regulon is required for full virulence of classical V. cholerae O1, as mutations in maltose transport genes, including ompS, affected TCP production and CT secretion (Lång et al. 1994). Furthermore, Davies et al showed ompS expression increased ∼120-fold when ToxT is over expressed ectopically (Davies et al. 2012).

To investigate OmpS further, we generated a V. cholerae O395 strain carrying a P_VCA1028-lacZ reporter plasmid (pGW118). In E. coli, ompS is induced in the presence of maltose. To validate the ompS reporter in V. cholerae, we examined β-galactosidase activity in cells grown in LB versus LB supplemented with 0.4% maltose. We found that expression of ompS in V. cholerae O395 was indeed enhanced upon the addition of maltose (Figure S2A, Supporting Information; Lå ng and Palva 1987). Furthermore, ompS was induced following ZAC-3 IgG treatment in LB medium, as evidenced by increased levels of β-galactosidase activity (Fig. 3A). However, the induction of ompS expression in V. cholerae O395 in response to ZAC-3 IgG was not dependent on ToxT, as β-galactosidase activity was similar between the wild type and a toxT null background (Fig. 3B). Thus, ompS is likely regulated by an alternative signal transduction pathway in response to ZAC-3 IgG.

Figure 3.

Regulation of ompS in V. cholerae O395 following ZAC-3 IgG treatment. β-galactosidase activity of mid-log phase V. cholerae (A) O395 or (B) toxT mutants harboring ompS:lacZ fusions grown in non-TIC, with or without ZAC-3 (10 µg/mL) for 1 h. There was a significant increase in relative β- galactosidase activity in both strains. Data are composed of four biological replicates, with two technical replicates each. (C) CV assay of mid-log phase V. cholerae O395 or isogenic ompS mutant treated with ZAC-3 IgG at concentrations ranging from 2.5 to 40 µg/mL in shaking conditions at 37^○C for 1 h. Both strains exhibited significant increase in CV signal in the presence of ZAC-3, and there was no significant difference in CV signal between strains at any antibody concentration tested. Data are comprised of three biological replicates and three technical replicates each. Significance was determined for A and B using unpaired, two-tailed Students t-test. (D) Semi-solid agar (0.3%) motility assays of V. cholerae in the presence or absence of ZAC-3 (10 µg/mL) over 6 h. Data are comprised of two biological replicates and at least two technical replicates each. No significant difference was observed between WT and the ompS strain under any condition tested. (E) Neonatal mouse colonization assay in which 5-day-old mice were inoculated with a 1:1 mixture of V. cholerae O395 strain KKV598 carrying a mutation in chromosomal lacZ locus (Table S1, Supporting Information) and the V. cholerae O395 ompS mutant, premixed or not with ZAC-3 IgG (10 µg/mL). Colonization of both strains was significantly reduced in the presence of ZAC-3 IgG, as revealed by a decrease in CFUs. However, there were no significant differences observed between the wild type and ompS mutant strains. Data are composed of two biological replicates with at least three mice per group. Significance was determined for C–E using two-way ANOVA followed by Tukey's multiple comparisons test. (F) Competitive index showing the ratio of ompS mutant bacteria recovered from the intestine in comparison to WT cells. No significant difference was seen in either treatment group. Statistical significance was determined using students t-test.

We speculated that OmpS may be involved in V. cholerae adaptation to ZAC-3 IgG treatment, since there are reports connecting maltoporins to outer membrane stress (Reimann and Wolfe 2011). To test this hypothesis, we constructed an isogenic ompS mutant in V. cholerae C6706 and evaluated the strain in the absence and presence of ZAC-3 IgG. Overall, the ΔompS mutant, GGW341, was indistinguishable from the wild type strain in terms of motility and ECM production, in the absence or presence of ZAC-3 IgG (Fig. 4). An isogenic ΔompS mutant in V. cholerae strain O395, GGW340, was also generated and shown to behave like the wild type strain (Fig.   3C, D). Finally, growth of the ompS mutants was not affected by the presence of ZAC-3 IgG over the course of 6 h (Figure S3, Supporting Information).

Figure 4.

Role of OmpS in V. cholerae C6706 response to ZAC-3 IgG-induced motility arrest and ECM induction. (A) Crystal violet assay of mid-log phase V. cholerae C6706 or the isogenic ompS mutant treated with ZAC-3 (0–2.5 µg/mL) at 37^○C for 1 h under static conditions. Data are comprised of at least three biological replicates and 3 technical replicates each. (B) Motility of V. cholerae C6706 in semi-solid agar (0.3%) in the presence and absence of ZAC-3 (10 µg/mL) over a 6 h period. Each bar represents the averaged (+/-SEM) of two biological replicates and at least two technical replicates each. No significant difference was seen between WT and ompS strains in untreated or ZAC-3 treated conditions. Significance was determined for A and B using two-way ANOVA followed by Tukey's multiple comparisons test.

While OmpS has been implicated in V. cholerae virulence (Lång et al. 1994), the ability of an ompS null mutant to colonize the intestinal mucosa has never been examined. Therefore, we compared the V. cholerae O395 ΔompS mutant to the wild type parent in the neonatal mouse model. Briefly, 3–5-day old BALB/c pups were gavaged with a 1:1 mixture of V. cholerae O395 derivative KKV598 (ΔlacZ) and GGW350 (ΔompS). The pups were incubated for 24 h in a 30°C chamber and then euthanized. Mouse intestines were immediately removed and homogenized. There lysates were plated on LB agar containing streptomycin (100 μg/mL) and X-gal (40 μg/mL) for enumerating CFUs, as a surrogate for intestinal colonization (Klose 2000). We recovered ∼1.2 × 10⁷ CFU of KKV598 per mouse. The ompS mutant also yielded ∼1.58 × 10⁷ CFU and was statically indistinguishable from wild type (Fig. 3E, F). To determine whether the absence of OmpS enabled V. cholerae to ‘escape’ the effects of ZAC-3 IgG in vivo, we repeated the studies in the presence of ZAC-3 IgG. ZAC-3 IgG significantly reduced ability of V. cholerae O395 ompS mutant to colonize the intestine to a degree that was identical to the wild type strain. We conclude from these studies that OmpS is not required for V. cholerae colonization of the intestinal mucosa in the neonatal mouse model, nor does the maltoporin alone play a significant role in bacterial adaptation to or escape from ZAC-3 IgG.

Role of norR in response of V. cholerae O1 to ZAC-3 IgG

VCA0182 was another V. cholerae gene whose expression was significantly increased following 60 min treatment with ZAC-3 IgG in non-TIC (Table 2). VCA0182 caught our interest because it encodes NorR, a σ54-dependent transcriptional regulator involved V. cholerae nitric oxide stress (Stern et al. 2012). A norR mutant was previously shown to be induced in the late stages of infection in the infant mouse model of V. cholerae infection (Stern et al. 2012; Schild et al. 2007). Moreover, in a separate study, VCA0182 was identified by transposon mutagenesis as having altered responses to ZAC-3 IgG (Baranova and Mantis, unpublished results).

Using a plasmid-encoded norR- lacZ fusion in V. cholerae C6706, we confirmed that norR was upregulated following ZAC-3 IgG, although the NorR regulated genes, hmpA and nnrS, were not (Fig. 5). To test if VCA0182 influences ECM induction or motility arrest in response to ZAC-3, we grew V. cholerae C6706 or the VCA0182 mutant in microtiter plates for 60 min at 37⁰C in static conditions and then performed CV staining. There was no significant difference in either the CV assay or motility in semi-solid agar between wild type and norR isogenic mutants of V. cholerae C6706 in absence or presence of ZAC-3 IgG (Fig. 5B, C). Additionally, we show that VCA0182 mutants in O395 strain shows no difference in ECM induction or neonatal mouse model (Figure S3, Supporting Information). We conclude that VCA0182 does not play a direct role in V. cholerae swimming arrest or ECM production in response to ZAC-3 IgG.

Figure 5.

VCA0182 is upregulated in response to ZAC-3, though ECM and motility arrest are unperturbed in ∆VCA0182 mutant strains of V. cholerae C6706. (A) β-galactosidase activity of mid-log phase V. cholerae C6706 harboring norR:lacZ, hmpA:lacZ and nnrS:lacZ transcriptional reporters treated with or without ZAC-3 (10 µg/mL) for 2 h under non-TIC. We observed significant upregulation of norR, but not nnrS or hmpA, in response to ZAC-3 IgG treatment. Data are comprised of three biological replicates composed of two technical replicates each. (B) CV assay of mid-log phase V. cholerae C6706 or isogenic norR mutant treated with ZAC-3 IgG (2.5–40 µg/mL) in static conditions at 37^○C for 1 h. Both strains exhibited a significant increase in CV staining in the presence of ZAC-3 IgG, although the two strains did not differ from each other under any condition tested. Data are comprised of at least three biological replicates and three technical replicates each. (C) Motility of V. cholerae C6706 in semi-solid agar (0.3%) in the presence and absence of ZAC-3 (10 µg/mL) over 6 h. Data are comprised of three biological replicates, with two technical replicates each. No significant differences were observed between WT and norR strains under conditions tested. Significance was determined using two-way ANOVA followed by Tukey's multiple comparisons test.

Impact of ZAC-3 IgG on virulence factor expression

In addition to antibody-mediated motility arrest and ECM induction, we postulated that changes in virulence gene expression might explain more fully how ZAC-3 IgG treatment attenuates V. cholerae colonization in the neonatal mouse intestine (Levinson, Baranova and Mantis 2016). Although the hypothesis that ZAC-3 IgG down-regulates V. cholerae virulence gene expression was not borne out by the RNA-seq analysis, those studies were performed under non-toxin inducing conditions (non-TIC). Therefore, we performed follow up experiments (in both non-TIC and TIC) to examine the impact of ZAC-3 IgG treatment on key virulence factors.

We were particularly interested in ToxT expression, as it is a master regulator for in V. cholerae O1 and suspected to regulate (indirectly) OmpS. The RNA-seq analysis done in non-TIC revealed that transcription of toxT in the El Tor C6706 strain was unaltered in response to ZAC-3 IgG treatment. We confirmed this observation in a classical strain V. cholerae O395 strain carrying a chromosomally encoded toxT:lacZ transcriptional fusion (Häse and Mekalanos 1998). The strain was grown under TIC and non-TIC inducing conditions in the presence and absence of ZAC-3 IgG, as described in the Materials and Methods. While toxT:lacZ expression was elevated under TIC, in accordance with what is reported in the literature (DiRita et al. 1991), under no conditions did we detect changes in β-galactosidase activity (up or down) following treatment with ZAC-3 IgG (Figure S4, Supporting Information).

Similarly, RNA-seq analysis indicated that ZAC-3 IgG treatment did not affect expression of the tcp locus. To examine this further, we employed classical (O395) and El Tor (C6706) V. cholerae tcpA-lacZ reporter strains grown under non-TIC and TIC. While β-galactosidase activity was significantly increased when V. cholerae tcpA-lacZ reporter strains were grown in TIC, β-galactosidase activity was unchanged upon exposure to ZAC-3 IgG (Fig. 6A, B). Western blot analysis with anti-TcpA polyclonal antibody confirmed that TcpA expression was unaltered in strains treated with ZAC-3 IgG (Fig. 6C–F).

Figure 6.

Analysis of tcpA expression in V. cholerae O395 and C6706 following ZAC-3 IgG treatment. V. cholerae (A) O395 or (B) C6706 strains carrying tcpA:lacZ promotor fusions were grown under non-TIC or TIC in the presence of 10 or 20 µg/mL of ZAC-3 IgG, respectively. Cultures were assessed for β-galactosidase activity, as described in the Materials and Methods. Data are composed of at least three biological replicates and two technical replicates each. Western blot and densitometry of whole cell lysates of (C, E) O395 and (D, F) C6706 strains grown as described above. Western blots are representative of at least three biological replicates. Presumed TcpA band at ∼16 kDa is indicated by black arrows. Graphs of average densitometry from Western blots described in C and D. The band density was not different between non-TIC or TIC. Significance was determined using two-way ANOVA followed by Sidak's multiple comparison test.

As a next step, we investigated whether ZAC-3 IgG treatment impacted actual secretion of CT by V. cholerae (Fig. 7A). In V. cholerae classical O395 and El Tor C6706 strains, there was a significant increase in total CT secretion, as measured by GM₁ ELISA, when cells were cultured under TIC, as compared to non-TIC. However, the addition of ZAC-3 IgG did not influence CT secretion by either V. cholerae strain O395 or C6706 (Fig. 7B–D). Collectively, these results demonstrate that treatment of V. cholerae O1 with ZAC-3 IgG does not significantly impact virulence gene expression and, therefore, cannot explain the effect of antibody treatment on bacterial colonization of the neonatal mouse intestine (Levinson, Baranova and Mantis 2016).

Figure 7.

CT secretion by V. cholerae O395 and C6706 following ZAC-3 IgG treatment. (A) GM₁ ELISA detecting either purified CT as a control, or concentrated supernatants from V. cholerae O395 or C6706 grown in non-TIC or TIC. (B) Concentrated supernatants of V. cholerae C6706 grown in TIC in the presence (open circles) or absence (closed circles) of ZAC-3 (20 µg/mL). Concentrated supernatants of V. cholerae O395 grown in (C) non-TIC or (D) TIC in the presence or absence of ZAC-3 (10 µg/mL). The x-axis indicates concentration of the CT standard. The concentrated supernatants were serially diluted (range 10⁻²–10⁻⁷). CT levels were not different between ZAC-3 IgG-treated and untreated samples. Data are representative of at least three biological replicates and two technical replicates. Significance was determined by two-way ANOVA followed by Tukey's multiple comparison test at each dilution.

DISCUSSION

The importance of anti-LPS antibodies in contributing to intestinal immunity to V. cholerae is well accepted, although the mechanisms involved in protection, especially early in the infection process, remain poorly understood. Serotype-specific IgG and IgM antibodies are vibriocidal in the presence of complement, for example, but this activity is unlikely to be a major factor in preventing bacterial colonization because IgG/IgM and complement levels are normally low in the gut lumen (Clements et al. 1982; Yang et al. 2019). Rather, intestinal immunity has been attributed to the ability of anti-LPS IgA and IgG antibodies to arrest V. cholerae motility and promote bacterial agglutination, resulting in bacterial paralysis and clearance through mucus entrapment (Bishop et al. 2010; Roche et al. 2015). Indeed, in this report, we demonstrated that human-derived, OSP-specific MAbs with known V. cholerae agglutinating activity are also potent inhibitors, in liquid medium, of V. cholerae motility. It should be noted that the four human MAbs studied were originally of various isotypes: one IgG1, one IgA1 and two IgA2 (Table 1). It is possible that if we were to study these MAbs as dimeric IgA or secretory IgA as would be found in the gut, we would see an increase in the magnitude of the antibody's effects, as their polyvalency would be expected to have increased capacity to crosslink LPS. In an invasive model of Salmonella, for instance, we recently reported that an anti-LPS dimeric IgA MAb can block invasion in vivo while its IgG1 isotype variant cannot (Richards et al. 2020).

The central point of this study was, however, to test the hypothesis anti-LPS antibodies like ZAC-3 IgG trigger reprogramming of V. cholerae gene expression in response to antibody-mediated motility arrest (‘paralysis’) and agglutination, which involves cell–cell contact and distortion of the OM. While we identified > 160 V. cholerae genes whose expression was affected by ZAC-3 IgG treatment, no clear evidence of OM stress was evident that would explain the ability of ZAC-3 IgG to reduce colonization of the neonatal mouse intestinal epithelium (Levinson, Baranova and Mantis 2016).

Further investigation into the results of the RNA-seq analysis are ongoing. Of particular interest was the downregulation of multiple genes previously implicated in late-stage infection, as high-affinity antibody production occurs during later stages of infection in human hosts, therefore we hypothesized that ZAC-3 treatment may affect similar genes to this stage of pathogenesis (Schild et al. 2007). Here we followed up on VCA0182, which encodes NorR, a transcription factor previously shown to be critical for survival of nitric oxide stress and in long term infection models in the adult mouse (Stern et al. 2012). We show that VCA0182 was upregulated 4.6-fold in RNA-seq experiments, though mutants in VCA0182 were no different than wild type strains in their response to ZAC-3, although an argument could be made that the adult mouse model is more appropriate to examine effects on late stage infection.

OmpS is a maltose-inducible, growth-phase dependent gene that was upregulated in response to ZAC-3 IgG (Lång and Palva 1987; Lång et al. 1988). While our results suggest that OmpS does not play a direct role in mediating the previously characterized responses of V. cholerae to antibody exposure, because an isogenic ompS mutants responded similarly to ZAC-3 IgG in terms of motility arrest, ECM production and colonization in neonatal mice. Nonetheless, we cannot rule out the possibility that a second porin could compensates for the loss of OmpS in the response to ZAC-3 IgG. In support of this idea is the fact that the ompS mutant did not exhibit a growth defect in minimal medium with maltodextrin as the sole carbon source, as would be expected if OmpS were the only porin involved in maltodextrin uptake.

The question still remains how ompS is upregulated in V. cholerae following antibody exposure. In E. coli ompS is under the control of the maltose system regulator MalT, which is positively regulated by cAMP-CRP and negatively regulated by Mlc. Mlc negatively regulates glucose uptake in addition to its role in repressing maltose utilization (Decker, Plumbridge and Boos 1998; El Qaidi and Plumbridge 2008). Collectively these data suggest a possible role for cAMP signaling in the response to ZAC-3 IgG. In addition, the loss of OmpS has been shown to negatively impact CT secretion, MSHA pilin production and hemagglutinin–protease release, while positively regulating TCP production in V. cholerae (Lång et al. 1994). Furthermore, in Salmonella and E. coli OmpS has been shown to play a role in regulation of the OM stress response pathway RpoE (Kenyon et al. 2005; Reimann and Wolfe 2011). This is thought to be due to a general role of OmpS in regulating OM permeability (Boos and Böhm 2000). OmpS is a porin with multiple roles, and it is highly possible that ZAC-3 cross-linking LPS impacts OM permeability, and ompS overexpression could be in direct response to this.

We predicted that the rapid motile to non-motile transition upon ZAC-3 treatment alone would serve as a sensory input (possibly through the flagellum itself) that modulates virulence gene expression, thereby explaining why anti-LPS antibodies are so effective at limiting bacterial colonization of the intestinal epithelium. The only significant changes in virulence-related genes seen in RNA-seq experiments done in non-TIC were hemagglutinin, VCA0865 (∼2.6-fold decrease) and the serine protease VesC, VC1649 (3.8-fold upregulation). Hemagglutinin has been shown to be expressed in late stage infection and is thought to play a role in detachment from cells. While VesC is associated with OM vesicles produced in infection and the ability to induce inflammation and hemorrhage in the rabbit ligated ileal loop model (Syngkon et al. 2010; Mondal et al. 2016). It would be of interest to repeat these RNA-seq experiments in TIC to assess virulence gene regulation more directly in a relevant experimental condition. The follow up studies performed here assessing TCP and CT suggest that at minimum, ZAC-3 does not significantly impact production of these two major virulence factors in vitro.

The mechanism(s) by which ZAC-3 IgG causes motility arrest and upregulation of ECM remain unresolved. From the current study, we postulate that regulation likely occurs at the post-transcriptional level. Certainly antibody coating of the flagellum itself and/or flagellar base could affect flagellum rotation due to steric hindrance (Bishop et al. 2010). In addition, we cannot rule out a possible role for c-di-GMP. It is plausible that perturbation of the OM by ZAC-3 IgG results in an outside-in signaling cascade involving one or more phosphodiesterases (PDE) and/or diguanyl cyclases (DGC). Alterations in the pool of c-di-GMP is known to negatively influence motility and positively influence VPS production (Moisi et al. 2009; Zamorano-Sánchez et al. 2019; Pratt et al. 2007).

ECM production by classical and El Tor V. cholerae strains following ZAC-3 IgG exposure was previously shown to be VPS-independent and not influenced by the master regulator, VieA, or quorum sensing molecules, HapR or While the exact composition of antibody-induced ECM is unknown, it was enriched in LPS molecules and acted much like OSP capsules or VPS-independent biofilms formed in seawater. Our transcriptional profiling studies revealed no significant change in OSP or VPS related genes. However, there are multiple examples of post-transcriptional and translation regulation in response to environmental factors, i.e. bile acids modulating T6SS, or c-di-GMP responsive riboswitches such as Vc1 which controls translation of GbpA, an integral N-acetyl-D-glucosamine (GlcNAc)- binding protein important for binding both mucins and chitin (Zhou et al. 2016).The contribution of c-di-GMP and other types of post-transcriptional regulation of antibody-induced motility arrest and ECM is the subject of ongoing study.

ACKNOWLEDGEMENTS

We thank Dr. Jans Wrammert (Emory University) for sharing anti-OSP antibodies, Dr. Claudia Hase (Oregon State University) for the toxT:lacZ reporter strain, the late Dr. Ronald Taylor (Dartmouth University) for both the tcpA:lacZ reporter strains and anti-TCPa antiserum, Jay Zhu (University of Pennsylvania) for the V. cholerae C6706 norR mutant and norR-related reporter strains, and Karl Klose (University of Texas at San Antonio) for strain KKV598 and the V. cholerae O395 toxT mutant. We gratefully acknowledge the Wadsworth Center's veterinary science staff for accommodating the mouse studies. We thank Richard Cole of the Wadsworth Centers Advanced Light Microscopy and Image Analysis Core Facility for assistance with microscopy. Finally, we extend our thanks to Ryan Schneider for assistance with analysis of ompS gene expression, Dylan Ehrbar for statistical analysis and Angelene Richards for helpful discussions.

FUNDING

This work was supported by the National Institute of Allergy and Infectious Diseases (NIAID) of the National Institutes of Health (NIH) under award number R21-AI109275. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Conflict of interest

None declared.