Vibrio cholerae’s mysterious Seventh Pandemic island (VSP-II) encodes novel Zur-regulated zinc starvation genes involved in chemotaxis and autoaggregation

Vibrio cholerae is the causative agent of cholera, a notorious diarrheal disease that is typically transmitted via contaminated drinking water. The current pandemic agent, the El Tor biotype, has undergone several genetic changes that include horizontal acquisition of two genomic islands (VSP-I and VSP-II). VSP-I and -2 presence strongly correlates with pandemicity; however, the contribution of these islands to V. cholerae’s life cycle, particularly the 26-kb VSP-II, remains poorly understood. VSP-II-encoded genes are not expressed under standard laboratory conditions, suggesting that their induction requires an unknown signal from the host or environment. One signal that bacteria encounter under both host and environmental conditions is metal limitation. While studying V. cholerae’s zinc-starvation response in vitro, we noticed that a mutant constitutively expressing zinc-starvation genes (Δzur) aggregates in nutrient-poor media. Using transposon mutagenesis, we found that flagellar motility, chemotaxis, and VSP-II encoded genes are required for aggregation. The VSP-II genes encode an AraC-like transcriptional activator (VerA) and a methyl-accepting chemotaxis protein (AerB). Using RNA-seq and lacZ transcriptional reporters, we show that VerA is a novel Zur target and activator of the nearby AerB chemoreceptor. AerB interfaces with the chemotaxis system to drive oxygen-dependent autoaggregation and energy taxis. Importantly, this work suggests a functional link between VSP-II, zinc-starved environments, and aerotaxis, yielding insights into the role of VSP-II in a metal-limited host or aquatic reservoir. Author Summary The Vibrio Seventh Pandemic island was horizontally acquired by El Tor pandemic strain, but its role in pathogenicity or environmental persistence is unknown. A major barrier to VSP-II study was the lack of stimuli favoring its expression. We show that zinc starvation induces expression of this island and describe a transcriptional network that activates a VSP-II encoded aerotaxis receptor. Importantly, aerotaxis may enable V. cholerae to locate more favorable microenvironments, possibly to colonize anoxic portions of the gut or environmental sediments.


Introduction 43
The Gram-negative bacterium Vibrio cholerae, the causative agent of cholera (Harris et al. (Twedt et al. 1981; Hood et al. 1981), and arthropods (Purdy and Watnick 2011; Broza and Halpern 49 therefore sought to identify factors that were required for Δ zur to aggregate. Aggregation (quantified as 107 the ratio of optical densities (OD 600nm ) in the supernatant before and after vortexing) was alleviated by 108 complementing zur in trans, excluding polar effects resulting from zur deletion (Fig. 1B). We next 109 examined the role of zinc availability on aggregate formation. Since metals can absorb to the surface of    2015)) was reversed by zinc supplementation (Fig. 1C). In contrast, the Δ zur mutant, which 116 constitutively expresses zinc starvation genes, aggregated in both the presence and absence of 117 exogenous zinc. These data indicate that aggregation occurs in minimal medium when the Zur regulon 118 is induced (i.e., during zinc deficiency or in a zur deletion strain) and is not a direct consequence of zinc 119 availability per se. Surprisingly, none of the annotated members of the Zur regulon were required for 120 aggregation (Fig. 1D, Fig. S1B), suggesting that there may be other Zur-regulated aggregation genes 121 yet to be identified.   (Fig. S1C). Collectively, these 141 data suggest that motility and chemotaxis are required for Δ zur aggregation in minimal medium.

142
We noted that the Δ zur phenotype resembles aggregation in E. coli rough mutants, which have 143 reduced expression of lipopolysaccharides (Nakao et al. 2012). We observed similar aggregation in V.

144
cholerae rough mutants (vc2205::kan), but this aggregation did not require motility to form (Fig. S1D) 145 and is therefore mediated by a distinct mechanism. We anticipated initially that Δ zur pellet formation 146 was a group behavior that may require processes associated with surface attachment (e.g., biofilm 147 formation, attachment pili) or cellular communication (e.g., quorum sensing); however, such mutants  156 S1F). Taken together, these data indicate that Δ zur pellet formation is not a clumping phenomenon Since aggregation appeared to require induction of the Zur regulon, we were surprised that the 160 transposon screen was not strongly answered by genes with an obvious Zur binding site in their 161 promoters. We reasoned, however, that our screen did not reach saturation due to the large number of 162 motility genes encoded in the V. cholerae genome. We therefore refined the screen by pre-selecting for 163 mutants that retained motility on soft agar, followed by a subsequent screen for loss of pellet formation 164 in the motile subset of the mutant pool, as described above. Interestingly, 19 of the 34 transposon 165 insertions answering this screen mapped to the Vibrio Seventh Pandemic island (VSP-II) (Fig. 2B, Fig.   166 S2), a horizontally acquired genomic region that is strongly associated with the El Tor biotype and the 167 current (seventh) cholera pandemic. Transposons concentrated in a section of VSP-II that encodes a 168 putative AraC-like transcriptional activator (VC0513, henceforth "VerA"), two ligand-sensing chemotaxis 169 proteins (VC0512, formerly Aer-1 is henceforth referred to as "AerB", and VC0514), and a cyclic di-

170
GMP phosphodiesterase (VC0515). Notably, the vc0513-vc0515 operon is preceded by a canonical Zur 171 binding site and is thus a novel candidate for Zur-dependent regulation.

172
To validate the VSP-II genes' involvement in  To determine if these VSP-II genes were sufficient to generate aggregation, we overexpressed 177 them in a wild-type and a Δ vsp-II background. Both aerB and verA overexpression caused the wild-type 178 to aggregate, but only the aerB chemoreceptor triggered aggregation in a strain lacking other VSP-II 179 genes (Fig. 2D). These data indicated that AerB drives aggregation and raised the possibility that VerA

231
The RNA-seq analysis also identified 36 differentially expressed genes that lacked canonical 232 Zur binding sites. Nineteen of these genes were significantly up-regulated in Δ zur, including several 233 genes on VSP-II (vc0504-vc0508 and vc0512) ( Fig. 3D-E

257
Interestingly, the aerB promoter lacks a conserved Zur binding site; however, our transcriptomic 258 data suggests that Zur-regulated VerA promotes aerB transcription. To verify this, we constructed a 259 P aerB -lacZ transcriptional reporter. Our initial attempt using a small (400 bp) promoter fragment did not 260 yield detectable signal under inducing conditions ( Fig. S4B-C). 5'-RACE mapping of the transcription 261 start indicated that aerB is part of a much longer transcript (extending >1 kb upstream of the start 262 codon). Thus, we designed a new reporter construct to include this entire region. P aerB activity in 263 standard LB medium fell below our threshold for detection (Fig. 4C, Fig. S4C), but we found that VerA 264 overexpression was sufficient to activate the aerB promoter. P aerB was strongly induced in a Δ zur strain 265 background -consistent with our initial RNA-seq -but only if the strain also carried a native or trans copy of verA. These data indicate that aerB expression is dependent upon VerA-mediated activation. In 267 summary, we found that VerA is a Zur-regulated, transcriptional activator that upregulates four genes 268 (aerB, vc0513-vc0515).

308
6A). Two additional mutants (R61A and H62A), corresponding to E. coli FAD-binding residues, were 309 also unable to aggregate. The requirement for oxygen and these highly conserved FAD-binding 310 residues suggests that AerB may bind FAD or a similar ligand to facilitate aerotaxis.

311
Based on our aggregation phenotype, we hypothesized that in contrast to AerA and Aer EC , AerB  (Fig. 5D, Fig. S6A-B). This is consistent with AerB promoting a negative response to In summary, we propose the following model for ∆ zur aggregation in M9 minimal medium (Fig. 6). In           Table S1. Successful integration of lacZ targeting vectors (pTD101, pJL1) 464 were identified by blue-white screening on plates containing 5-Bromo-4-chloro-3-indolyl-β-d-

557
Plates were incubated overnight at 30˚C and for an additional day at room temperature before being

663
Another set up tubes was prepared anoxically by purging with N 2 (-O 2 ) (see Methods for details).