Actuation of CRP activating region 3 by acetylation modulates V. cholerae sugar utilization and virulence

V. cholerae, cAMP receptor protein, lysine acetylation, virulence, acetate switch, 47 small RNA 48 49 50 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2026. ; https://doi.org/10.64898/2026.01.27.701997doi: bioRxiv preprint

51 The cyclic AMP receptor protein or CRP is a global regulator of bacterial metabolism that 52 activates transcription of genes required for utilization of alternative carbon sources in response 53 to the second messenger cAMP, which is synthesized in the setting of glucose scarcity. CRP 54 activates transcription through contact with RNA polymerase at three sites termed activating 55 regions (ARs) 1-3. AR3 was previously reported to be functional only when CRP K52 was 56 mutated to a neutral residue and to be essential for transcription only in the absence of AR1 and 57 AR2. Multiple proteomic studies have reported acetylation of CRP K52. This post-translational 58 modification is predicted to activate AR3. To probe the role of K52 acetylation (K52QAc) and AR3 59 at the genome level, we used ChIP-seq and RNA-seq analysis to compare WT CRP with a CRP 60 K52Q mutant that mimics CRP K52Ac. We report that CRP K52Q binds to hundreds of new sites 61 on the chromosome, resulting in increased abundance of known as well as previously unknown 62 transcripts. These transcripts increase uptake and metabolism of dietary sugars such as maltose 63 and galactose, repress acetate consumption, and augment virulence gene expression. We 64 attribute the repression of acetate consumption to a novel small RNA, CrbZ, which is positively 65 regulated by CRP K52Q in LB broth and by WT CRP specifically in minimal medium containing 66 maltose. This study highlights the role of post-translational modifications in molding the CRP 67 regulon to optimize pathogen metabolism and virulence gene expression in the human intestine in 68 response to nutritional cues. 69 Significance statement: As a model in the field of bacterial transcription, the structure and 70 function of the cAMP receptor protein (CRP), a global transcription regulator, has been 71 exhaustively investigated. These studies have established three activating regions (ARs) where 72 CRP contacts RNA polymerase, of which only two were thought to participate in transcription 73 activation by native CRP. Here we provide evidence that post-translational acetylation of V. 74 cholerae CRP lysine 52 actuates AR3, enabling occupancy of hundreds of novel CRP binding 75 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2026. ; https://doi.org/10.64898/2026.01.27.701997doi: bioRxiv preprint sites and the transcription of genes encoding novel small RNAs. These changes alter virulence 76 gene expression, promote utilization of dietary carbon sources, and delay acetate uptake. We 77 propose that acetylation of CRP K52 engages AR3, thus optimizing V. cholerae fitness in the 78 human intestine. 79 80 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2026. ; https://doi.org/10.64898/2026.01.27.701997doi: bioRxiv preprint

81 The cyclic AMP receptor protein (CRP) is a conserved global regulator of bacterial transcription 82 (1). To regulate transcription, CRP dimers complexed with the second messenger cyclic 83 adenosine monophosphate (cAMP) bind to the chromosome at two inverted repeats with the 84 consensus sequence ATGTGA separated by a 6 bp spacer (Fig 1A) (2). 85 CRP activates uptake and catabolism of many sugars when glucose is unavailable, and glucose 86 transport is intimately tied to generation of cAMP. The phosphoenolpyruvate phosphotransferase 87 system (PTS), a regulatory multi-component phosphotransfer cascade that terminates in 88 phosphorylation of the incoming sugar, mediates transport of a variety of sugars (Fig 1B) (3). 89 PTS transport of some sugars depends on CRP , while transport of others such as glucose does 90 not. The PTS cascade begins with phosphorylation of Enzyme I (EI) by phosphoenolpyruvate. 91 This phosphate is passed to Histidine protein (HPr) and from there to sugar-selective enzymes llA 92 and B (EllA and B). Sugar transport through the Enzyme llC results in passing of phosphate from 93 Enzyme llB directly to the incoming sugar. Enzymes ll A, B, and C may be independent proteins 94 or domains within a single multidomain protein. In V. cholerae, the sugar specificities of many Ell 95 components have been determined (4, 5). 96 When sugars transported by the PTS are abundant, sugar-specific PTS proteins exist principally 97 in their unphosphorylated state, while they are phosphorylated when these sugars are absent. 98 Both the phosphorylated and unphosphorylated states of these proteins regulate cell physiology 99 by direct protein-protein interactions. PTS transport plays a key role in regulation of transcription 100 by CRP because phosphorylated glucose-specific Enzyme llA interacts directly with adenylate 101 cyclase to increase intracellular levels of cAMP (6, 7). Thus, when glucose is scarce, CRP is 102 active. Association of cAMP with the N-terminus of CRP increases RNA polymerase binding and 103 transcription, which activates hundreds of genes including many that transport and metabolize 104 alternative carbon sources (8). In many pathogens, CRP also regulates virulence gene 105 expression (9-13). 106 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2026. ; https://doi.org/10.64898/2026.01.27.701997doi: bioRxiv preprint In initial studies of Escherichia coli CRP, two classes of promoters were identified that make 107 distinct contacts with the α -subunit of RNA polymerase (Fig 1A). At class I promoters, the 108 consensus binding sequence of CRP is centered at -61.5 bp with respect to the transcription start 109 site (TSS). This results in contact between the downstream CRP monomer and the C-terminal 110 domain of the α subunit (α -CTD) of RNA polymerase (RNAP). CRP residues 156-164, which 111 interact with the α -CTD are termed activating region 1 (AR1) and are essential for transcription 112 activation at class l promoters (14). 113 At class ll promoters, the CRP consensus binding sequence is centered at -41.5 bp with respect 114 to the TSS. Here the α -CTD of RNAP wraps around the CRP dimer via a flexible loop, and the 115 AR1 contact forms between the upstream CRP monomer and the α -CTD of RNAP . At class ll 116 promoters, two additional activating regions located in the downstream CRP monomer (AR2 and 117 AR3) have been identified. AR2, which consists of CRP residues 19, 21, 96, and 101, contacts 118 the N-terminal domain of the α subunit of RNAP and is essential for transcription at class ll 119 promoters (15, 16). In contrast, AR1 is only essential at a subset of class ll promoters. CRP AR3 120 is defined by amino acid residues 52-58 which contact the σ -subunit of RNAP (17, 18). AR3 was 121 reported to activate transcription at class ll promoters only when K52 is changed to a neutral 122 amino acid and transcription activation by AR1 or AR2 is diminished by mutation (19, 20). This 123 body of work led to the conclusion that CRP AR3 is dispensable for transcription activation by 124 native CRP at class ll promoters (14). 125 The genome of the diarrheal pathogen Vibrio cholerae encodes a CRP homolog. Of the 210 126 amino acids that comprise V. cholerae CRP, 201 are identical to those of Escherichia coli CRP 127 including the residues defining AR3. CRP is crucial for V. cholerae’s response to the aquatic 128 environment, mammalian hosts, environmental hosts, and Vibrio-specific phages through its 129 regulation of metabolism, quorum sensing, horizontal gene transfer, biofilm formation, and 130 virulence (21-32). Interestingly, although CRP represses many virulence determinants including 131 the toxin co-regulated pilus, cholera toxin, the RTX toxin, and biofilm formation under laboratory 132 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2026. ; https://doi.org/10.64898/2026.01.27.701997doi: bioRxiv preprint conditions, a Δ crp mutant, in which these virulence factors are highly expressed, is massively 133 defective for colonization of both the mammalian and zebra fish intestines (27, 33). Two 134 explanations are that V. cholerae CRP is essential for optimal utilization of nutrients to promote 135 growth in the human intestine and/or that CRP does not actually repress virulence gene 136 expression in vivo. 137 We previously reported that V. cholerae CRP is membrane-associated under specific growth 138 conditions and sequesters the dual function transcription factor peptidase A to the membrane 139 (34). In these studies, we found that K52 was succinylated. A proteomic study detailing the 140 acetylome of the closely related species Vibrio parahemolyticus as well as proteomic studies of 141 purified E. coli CRP have reported acetylation of K52 (35, 36). Because acetylation neutralizes 142 the positive charge of K52, this post-translational modification is predicted to engage AR3. 143 To better understand the impact of K52 acetylation on the function of V. cholerae CRP , we defined 144 the regulon of a CRP mutant with K52 substituted for Q to mimic acetylation. Here we report that 145 a CRP K52Q substitution generates hundreds of novel chromosomal binding sites, some of which 146

in increased abundance of putative small RNAs (sRNAs). This significantly reshapes the 147 regulon of CRP leading to increased transcription of genes that encode virulence factors and 148 sugar metabolism genes and repression of those involved in acetate utilization. We show that the 149 latter are regulated by the novel sRNA CrbZ, which is in turn positively regulated by CRP K52Q in 150 cells grown in LB broth and by wild-type (WT) CRP in cell grown in minimal medium containing 151 maltose. Although CRP K52Q reverses inhibition of the V. cholerae virulence factors by WT CRP , 152 this mutant has no competition advantage in the infant mouse intestine. We hypothesize that CRP 153 K52 is acetylated in vivo resulting in engagement of AR3 to optimize V. cholerae fitness in the 154 mammalian intestine. 155 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2026. ; https://doi.org/10.64898/2026.01.27.701997doi: bioRxiv preprint

156 An amino acid substitution that mimics CRP K52 acetylation modulates DNA occupancy, 157 gene expression, and virulence without altering CRP membrane association. We previously 158 reported that, when V. cholerae was cultured in defined medium supplemented with sucrose, CRP 159 was found principally in the membrane fraction (34). When lysines 22, 26, 35, and 52 were 160 mutated to Q (4XQ) to mimic acetylation, CRP relocated to the cytoplasm, leading to the 161 hypothesis that post-translational modifications dictate the subcellular localization of CRP. In V. 162 cholerae LB broth cultures, approximately 26% of CRP was found in the membrane (Fig S1). To 163 determine if similar residues might determine the subcellular localization of CRP in LB broth, we 164 tested chromosomal 4XQ and 4XR mutants designed to mimic acetylated and deacetylated forms 165 of CRP, respectively. A V5 affinity tag was added to the C-terminus of WT CRP and all mutants to 166 facilitate subsequent analysis. The 4XQ mutant was found principally in the cytoplasmic fraction, 167 while the 4XR mutant was more abundant in the membrane fraction (Fig S1). 168 Because a sizeable amount of CRP was cytoplasmic for all V. cholerae strains cultured in LB 169 broth, we hypothesized that ChIP-seq under these conditions might elucidate the impact of these 170 substitutions independent of differences in membrane association. Importantly, the V5 tag used 171 for immunoprecipitation had only a small effect on gene expression (Fig S2). ChIP-seq analysis 172 was carried out on biological triplicates of WT, 4XR, and 4XQ mutants cultured in LB broth to mid-173 log phase. Enrichment in DNA occupancy was measured relative to an untagged control and a 174 minimum of 4-fold enrichment (EF≥ 4) in each of three biological replicates was considered 175 significant (Table S1). As shown in Fig 1C, all three alleles were significantly enriched at the 176 majority of chromosomal binding sites, proving that all alleles are functional. WT CRP and the 177 4XR mutant were most similar with both showing enrichment at an additional 48 sites. In 178 contrast, while WT CRP and the 4XR mutant bound very few sites uniquely, the 4XQ substitution 179 enhanced binding at multiple novel chromosomal locations. Twenty-three sites were common 180 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2026. ; https://doi.org/10.64898/2026.01.27.701997doi: bioRxiv preprint between the 4XR and 4XQ CRP mutant alleles, indicating a unique role for lysine in inhibiting 181 binding at these sites. 182 To investigate the functional significance of the additional 4XQ binding sites, we undertook RNA-183 seq analysis of exponential phase cultures of WT CRP as well as 4XR, 4XQ, and Δ crp mutants 184 (Table S2). Because the 4XQ mutant includes K52Q, which is predicted to engage AR3, we 185 included K52Q and 3XQ mutants (K22Q, K26Q, and K35Q) in our analysis (15, 16, 18, 37). WT 186 CRP and all mutants were expressed at comparable levels, grew at comparable rates, and did 187 not alter acetylation at other sites (Fig S3). Furthermore, the K52Q substitution did not alter CRP 188 membrane association (Fig S1). We defined significantly differentially expressed genes as those 189 having at least a 2-fold change in expression and a false discovery rate of 0.05 or less. Table 1 190 shows the numbers of total and gene-associated transcripts that increased and decreased in 191 abundance relative to a Δ crp mutant. WT CRP and the 4XR mutant had similar numbers of 192 differentially regulated genes, while the 3XQ mutant differentially regulated far fewer genes than 193 WT CRP, and the K52Q mutant differentially regulated more genes than any of the other alleles. 194 The 4XQ mutant regulated fewer genes than WT and K52Q and more than 3XQ, indicating a 195 phenotype intermediate between the K52Q and 3XQ mutants. 196 To further explore the RNA-seq data, we carried out principal component analysis. The WT, 4XR 197 and 3XQ CRP transcriptomes clustered together, while the 4XQ and K52Q transcriptomes were 198 separate and closer to each other (Fig. 1D). Taken together, the ChIP-seq and RNA-seq data 199 suggest that the K52Q substitution, and by extension post-translational acetylation of K52, has a 200 large impact on transcription regulation by CRP. We, therefore, focused on the role of the CRP 201 K52Q substitution in chromosomal occupancy and gene regulation. 202 We first performed a ChIP-seq experiment comparing chromosomal binding of WT CRP and the 203 K52Q protein (Table S3). As shown in Fig 2A, using an enrichment factor of 4 (EF>4) as a 204 threshold, WT CRP associated with a mean of 66 chromosomal sites, while CRP K52Q bound to 205 263 sites, an increase of almost 4-fold. Considering sites that reached the threshold of EF>4 in 2 206 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2026. ; https://doi.org/10.64898/2026.01.27.701997doi: bioRxiv preprint of 3 replicates, only 2/3 of WT binding sites were shared by K52Q (Fig 2B). We then categorized 207 the positions of the chromosomal binding sites of WT CRP and the K52Q mutant with respect to 208 putative TSSs using a previous comprehensive report of V. cholerae TSSs as a reference (38). 209 Putative promoters were defined as 200 bp upstream of a TSS (Fig 2C). The 5/i5 untranslated 210 region (5/i5 -UTR) was considered to extend from the TSS to the start of the coding sequence 211 (CDS) of a gene. The majority of WT CRP binding sites were within promoters (Fig 2C). While 212 very few WT CRP binding sites were located outside promoters, almost half of the K52Q mutant 213 binding sites were either in the 5/i5 -UTR or CDS of genes. 214 In a CRP K52Q mutant, AR3 is predicted to form an additional contact with the σ subunit of RNAP 215 at Class ll promoters. In the few promoters previously studied, K52 and AR3 did not participate in 216 transcription activation by native CRP at Class ll promoters (18, 37). Nevertheless, we were 217 curious whether Class ll promoters might be over-represented in K52Q binding sites. We defined 218 Class l promoters by the location of a CRP binding site within -69 to -52 bp of a TSS and class ll 219 promoters as a binding site within -50 bp to -32 bp of the TSS. Based on these definitions, we 220 identified binding at more class ll than class I promoters for both WT CRP and CRP K52Q (Fig 221 2D). While the majority of WT CRP binding sites at Class l and Class ll promoters were shared by 222 CRP K52Q, CRP K52Q had approximately twice as many Class l and Class ll binding sites as WT 223 CRP, suggesting that CRP K52Q does not show a preference for binding at Class ll promoters. 224 CRP K52Q tolerates more degeneracy in operator sequences than WT CRP. We questioned 225 whether the operator sequences bound by WT CRP were distinct from those bound by CRP 226 K52Q. Based on 53 binding sites, Meme analysis of WT CRP yielded an operator consensus 227 sequence that was similar to that previously defined for E. coli and V. cholerae CRP (Table S4, 228 Fig 2E) (2, 27, 39). Meme analysis of the CRP K52Q binding sites yielded two consensus 229 sequences. One of these (Meme l), based on 90 binding sites, was closely related to that of WT 230 CRP. The second consensus sequence (Meme ll), based on 60 binding sites, displayed 231 degeneracy in one half of the inverted repeat, suggesting that, at some sites, binding of one half 232 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2026. ; https://doi.org/10.64898/2026.01.27.701997doi: bioRxiv preprint of the CRP K52Q dimer to the chromosome may be adequate to stabilize the interaction with 233 DNA. 234 We reasoned that WT CRP and the K52Q mutant might share some of the binding sites on which 235 their respective consensus sequences were based. In fact, out of the 53 sites used to identify WT 236 Meme l, 41 were shared with CRP K52Q (Fig 2F). In contrast, only 29 of the 90 sites used to 237 identify CRP K52Q Meme l and only 4 of the 60 sites used to identify CRP K52Q Meme ll were 238 shared. 239 The σ 70 subunit of RNAP , which interacts with CRP AR3, binds to the -35 and -10 elements within 240 promoters where it stabilizes the open complex of DNA known as the transcription bubble, which 241 is essential for transcript elongation (40). Promoter escape after synthesis of the first 11 242 nucleotides of the mRNA was reported to coincide with release of σ 70 from RNAP, at least partly 243 due to displacement by the growing mRNA transcript (41). However, recent in vitro studies 244 suggest that the σ 70 subunit of RNAP can be retained throughout transcript elongation and that 245 the interaction of σ 70 with sequences similar to σ 70 binding sites, which are found throughout the 246 genome, may induce pausing of the elongation complex (42-44). We, therefore, hypothesized 247 that, if the Meme ll operator was oriented appropriately, CRP K52Q binding to the DNA, even at 248 intragenic sites, could be stabilized by the interaction of CRP K52Q AR3 with σ 70 (Fig 2G). 249 Sixteen Meme ll sites were in promoters, while 44 were in coding sequences (Table S4). We 250 examined whether each Meme ll consensus sequence was oriented appropriately to allow an 251 interaction between AR3 of the unbound CRP K52Q and σ 70. We found that 38/60 or 252 approximately 2/3 of the Meme ll sites were appropriately oriented, while the other 1/3 were not. 253 This suggests that binding of CRP K52Q at Meme ll sites is independent of an interaction with 254 σ 70. Last, we examined whether CRP K52Q is functional at Meme ll operator sites. CRP K52Q 255 enrichment at Meme ll sites positioned within genes was not correlated with an effect on transcript 256 abundance. In contrast, 18 Meme ll binding sites were within promoters. Of these, 9 were 257 correlated with either positive or negative changes in transcript abundance as compared with WT 258 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2026. ; https://doi.org/10.64898/2026.01.27.701997doi: bioRxiv preprint CRP. Taken together, these observations suggest that the CRP K52Q substitution stabilizes 259 binding to Meme ll operator sites independent of an interaction with σ 70 and further that, when 260 positioned within promoters, binding of CRP K52Q to Meme ll sites can alter transcription. 261 CRP K52Q activates gene transcription at Class ll promoters. To determine whether CRP 262 K52Q was more likely than WT CRP to increase transcript abundance, we plotted transcript 263 abundance against the presence of WT CRP binding only, CRP K52Q and WT binding, or CRP 264 K52Q binding only (Fig 3A). In each case, both increases and decreases in transcript abundance 265 were observed, suggesting that CRP K52Q does not serve exclusively as an activator of 266 transcription. We then questioned whether CRP K52Q was more likely to activate transcription at 267 Class ll promoters than Class l promoters. At the few binding sites shared by WT CRP and CRP 268 K52Q within promoters meeting the definitions of Class l or Class ll, gene expression was equally 269 likely to be positively or negatively regulated at both types of promoters (Fig 3B). In contrast, at 270 sites bound only by CRP K52Q, transcription activation was more likely than repression at Class ll 271 but not Class l promoters (Fig 3C). This is consistent with engagement of AR3 by CRP K52Q 272 and, possibly, CRP K52Ac, at Class ll promoters. Interestingly, in the majority of cases, CRP 273 K52Q enrichment at both Class l and Class ll promoters had no effect on transcription. This 274 suggests either that these promoters were incorrectly classified or that they are not active during 275 growth in LB broth. Last, we explored the multitude of CRP K52Q binding sites that are within 276 CDSs to determine whether these affect expression of the corresponding gene. As shown in Fig 277 3D, only 20 out of the 93 CRP K52Q binding sites in CDSs altered transcription. Twelve of these 278 positively regulated transcription while 8 negatively regulated transcription. 279 The CRP K52Q substitution alters V. cholerae sugar utilization. CRP is best known for 280 activating utilization of carbon sources other than glucose. The CRP K52Q substitution 281 dramatically reshaped the transcription profile of sugar transport and utilization genes, suggesting 282 a shift in carbon preferences. To illustrate this, we constructed a heat map comparing regulation 283 of such genes by CRP (CRP/Δ crp) and the additional regulation imposed by a CRP K52Q 284 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2026. ; https://doi.org/10.64898/2026.01.27.701997doi: bioRxiv preprint substitution (CRP K52Q/WT CRP). As compared with the Δ crp mutant, WT CRP activated genes 285 involved in transport and catabolism of citrate, maltose and galactose, which are transported 286 independently of the PTS, as well as the PTS-dependent sugars mannitol, mannose, and 287 cellobiose (Fig S4). In addition, CRP activated transcription of an alternative EllC glucose 288 transporter, VC1821, that is reportedly independent of PTS Enzyme l, and nearby EllA and B 289 homologs (VC1822 and VC1823) with unknown sugar specificity (5). In contrast, CRP K52Q 290 principally activated transcription of genes involved in transport and utilization of PTS-291 independent carbon sources such as citrate, maltose, and galactose and repressed those 292 involved in utilization of PTS-dependent carbon sources such as mannitol, mannose, glucose, 293 and cellobiose. In addition, transcription of the fructose repressor, FruR, was increased, which 294 would be predicted to decrease utilization of the PTS-dependent sugar fructose. Transcription of 295 gene clusters involved in uptake and catabolism of only two PTS-dependent sugars, trehalose 296 and the cell wall building block N-acetylmuramic acid, was activated by CRP K52Q. Both are 297 implicated in the cellular response to osmotic stress (45, 46). 298 To test the functional implications of our observations, we measured V. cholerae growth on 299 various sugars. We found that a CRP K52Q mutant grew more poorly than WT on minimal 300 medium supplemented with fructose but better on that supplemented with trehalose (Fig S5). 301 CRP K52Q grew better on several concentrations of maltose but only when inoculated from a pre-302 culture containing maltose. We conclude that the CRP K52Q substitution and, by extension, CRP 303 K52Ac remodels sugar utilization by V. cholerae. 304 The CRP K52Q substitution reverses repression of virulence gene expression by CRP but 305 does not increase virulence in an infant mouse model. V. cholerae depends on the toxin co-306 regulated pilus (TCP) for colonization of the infant mouse and rabbit intestines (47). The VPS-307 dependent biofilm, whose role in attachment to abiotic surfaces has been intensively studied, is 308 also conditionally required for colonization of the mammalian intestine (48-50). The toxins RTX 309 and hemolysin are minor virulence factors (51). CRP represses expression of the TCP, biofilm, 310 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2026. ; https://doi.org/10.64898/2026.01.27.701997doi: bioRxiv preprint and RTX toxin genes in laboratory cultures and activates expression of a hemolysin gene. 311 Nevertheless, a Δ crp mutant is defective in colonization of the infant rabbit intestine (23, 33, 52). 312 This suggests that CRP may not repress virulence gene expression in vivo. To determine how 313 the CRP K52Q point mutation impacted virulence gene expression, we compared transcript 314 abundance in WT CRP and the K52Q mutant (Fig S6A). In spite of repression of the transcription 315 regulators ToxR (VC0984) and TcpP (VC0826), which are activators of virulence gene 316 transcription, the CRP K52Q substitution activated virulence gene expression, bringing 317 transcription closer to that of a Δ crp mutant (53, 54). To determine how this might translate into 318 virulence in the infant mouse model, we carried out competition experiments co-inoculating the 319 4XQ or K52Q mutant and WT V. cholerae. As shown in Fig S6B, the 4XQ mutant had a slight 320 colonization defect, while colonization by the K52Q mutant was similar to that by WT V. cholerae. 321 While colonization is a complex phenotype, these results are consistent with acetylation of CRP 322 K52Q in vivo. This would allow V. cholerae to retain maximal virulence gene expression while 323 optimizing carbon utilization of carbon sources available in the intestine. 324 The CRP K52Q mutant blocks the acetate switch. Like E. coli, in the presence of an excess of 325 carbon sources, V. cholerae excretes acetate as a byproduct of fermentation rather than 326 channeling it into the tricarboxylic acid (TCA) cycle (55, 56). When these carbon sources are 327 expended, V. cholerae initiates acetate consumption by upregulating the acetyl-CoA synthesis 328 gene acs1 in a process known as the acetate switch (55). In V. cholerae, the acetate switch is 329 activated by the two-component system CrbRS and CRP (31, 56). We reasoned that CRP K52Ac 330 might be more abundant during sugar fermentation and, therefore, that one role of both CRP 331 K52Ac and CRP K52Q might be to modulate the acetate switch. To examine this hypothesis, we 332 explored the transcription of switch-specific genes in our RNA-seq data set. In fact, the 333 abundance of transcripts involved in acetate utilization such as crbR and the acetyl-CoA synthase 334 gene acs1 was much higher in WT V. cholerae as compared with the CRP K52Q mutant (Fig 4A). 335 We confirmed these results by qRT-PCR (Fig 4B). To demonstrate that this pattern of gene 336 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2026. ; https://doi.org/10.64898/2026.01.27.701997doi: bioRxiv preprint regulation did, in fact, delay the acetate switch, we measured acetate in the supernatants of LB 337 cultures (Fig 4C). Acetate consumption was delayed in the CRP K52Q mutant as compared with 338 WT V. cholerae. Growth on acetate involves synthesis of acetyl-CoA, funneling of acetyl-CoA into 339 the TCA cycle, and upregulation of the glyoxylate shunt to avoid the carbon wasting steps in the 340 TCA cycle (57). The CRP K52Q mutant was unable to use acetate as a sole carbon source for 341 growth (Fig 4D). Taken together, these results are consistent with a model in which acetylation of 342 CRP K52 functions to block the acetate switch. 343 A medium rich in fermentable sugars increases lysine acetylation but not of K52. Because 344 a CRP K52Q mutant delays the acetate switch, we hypothesized that we might detect CRP K52 345 acetylation in a medium rich in a fermentable sugar. To test this, we purified CRP-V5 from 346 exponential and stationary phase cultures in minimal medium supplemented with 0.4% sucrose 347 and sent the purified protein for proteomic analysis. As shown in Figure S7, acetylation of K35, 348 K89, and K188 and succinylation of K188 was detected in exponential phase. In stationary phase 349 when the sucrose is depleted, the acetylation sites were no longer observed. However, 350 acetylation of CRP K52 was not observed in any of these cultures. This demonstrates that while 351 fermentable sugars in the environment are correlated with V. cholerae CRP lysine acetylation, 352 acetylation of K52 is not detectable in the setting of sucrose abundance. We hypothesize that 353 acetylation of CRP K52 demonstrates sugar specificity. 354 CRP K52Q activates transcription of the novel sRNA gene crbZ, which is transcribed within 355 and antisense to crbS. We noted that several binding sites unique to CRP K52Q were close to 356 TSSs of novel transcripts. Six of these, shown in Fig S8, were positioned either within an 357 intragenic region or within a CDS but running antisense to the CDS. The distance of all the CRP 358 K52Q binding sites from a TSS was most consistent with that of a Class ll promoter. This is 359 supports our previous observations that the CRP K52Q mutant activates transcription from Class 360 ll promoters, possibly by engaging AR3. 361 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2026. ; https://doi.org/10.64898/2026.01.27.701997doi: bioRxiv preprint We further explored a putative sRNA encoded anti-sense to crbS (Fig 5A). This gene was 362 designated crbZ to reflect its location. We first confirmed its transcript abundance by qRT-PCR 363 using primers designed to amplify the crbS region containing crbZ and primers designed to 364 amplify a region of crbS downstream of crbZ. As shown in Fig 5B, when primers spanning crbZ 365 were used to measure crbZ+crbS transcript abundance, that measured for the CRP K52Q mutant 366 was approximately 13-fold higher than that for WT V. cholerae or a Δ crp mutant. In contrast, 367 when primers in the histidine kinase domain (HK) outside crbZ were used, there was no 368 significant difference in transcript abundance between the CRP K52Q mutant and WT V. cholerae 369 or a Δ crp mutant. This supports the presence of a sRNA in the N-terminus of crbS that is 370 regulated by CRP K52Q. To further establish the presence of this sRNA, we performed a ligation 371 to circularize CrbZ, generated cDNA, and amplified the region spanning the ligated ends using 372 outward facing primers at the 5/i5 and 3/i5 ends of crbZ (Fig 5C). As shown in Fig 5D, the primers 373 amplified a product. To prove that this product was ligated CrbZ, it was sequenced. The 374 sequence of the amplification product matched the ligated ends of CrbZ (Fig S9). This provides 375 strong evidence that crbZ is transcribed. 376 Like CRP K52Q, CrbZ inhibits the acetate switch and improves growth on maltose. Based 377 on its position within the crbS gene, we hypothesized that CrbZ might regulate acetate uptake. To 378 test this, we cloned crbZ into an IPTG-inducible vector. We first confirmed overexpression of crbZ 379 from the plasmid (Fig S10A). We then tested its impact on crbS, crbR, and acs1 expression and 380 acetate uptake. crbZ overexpression decreased both acs1 transcript abundance and delayed 381 acetate uptake but did not affect expression of crbRS (Fig S10A and B). Similar to the CRP K52Q 382 mutant, crbZ overexpression improved growth in maltose (Fig 10C). Taken together, these data 383 suggests that CrbZ plays a role in repression of acetate uptake by the CRP K52Q mutant. 384 crbZ is expressed in minimal medium containing maltose. Our attention was drawn to crbZ 385 precisely because it was transcribed when the CRP K52Q mutant but not WT V. cholerae was 386 cultured in LB broth. We hypothesized that crbZ might be transcribed by WT V. cholerae in 387 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2026. ; https://doi.org/10.64898/2026.01.27.701997doi: bioRxiv preprint minimal medium supplemented with maltose, a carbon source that promoted growth of the CRP 388 K52Q mutant. In fact, we found that in exponential phase cultures of WT V. cholerae in minimal 389 medium supplemented with maltose but not trehalose, the region of crbS that includes crbZ was 390 expressed approximately 3-fold more highly than in the LB overnight culture that was used to 391 inoculate the minimal medium cultures (Fig 5E). In contrast, the distal region of crbS was not 392 differentially transcribed for any of these conditions. To measure the transcript abundance of crbZ 393 and crbS separately, we performed strand-specific qRT-PCR. To validate the technique, we first 394 measured crbZ and crbS transcript abundance in WT V. cholerae and the CRP K52Q mutant. In 395 WT V. cholerae the ratio of crbZ/crbS was approximately 0.2, while in the CRP K52Q mutant it 396 was approximately 14 (Fig S11A). This is consistent with our RNA-seq results. In minimal 397 medium supplemented with 0.8% maltose, the ratio of crbZ/crbS was approximately 14 in 398 exponential phase cultures and approximately 6 in stationary phase cultures (Fig S11B). This 399 confirms upregulation of crbZ when V. cholerae is cultured in the presence of maltose. 400 Furthermore, acs1 was downregulated only in the minimal medium culture supplemented with 401 maltose but not trehalose (Fig 5E). Taken together, our results suggest that crbZ is expressed in 402 a sugar-specific manner and represses the acetate switch. 403

404 CRP is one of the best studied bacterial transcription regulators. It activates transcription at Class 405 l and Class ll promoters through ARs 1 and 2, which have been defined by structure-function 406 studies (16, 37, 58). Approximately 25 years ago, investigators observed that substitution of CRP 407 K52 for a neutral residue resulted in transcription activation specifically at Class ll promoters when 408 CRP contained inactive ARs 1 and 2. This led to the discovery of AR 3, which was dismissed as 409 dispensable for transcription activation by native CRP (37). Here we show that a CRP K52Q 410 substitution, which mimics acetylation of K52, results in hundreds of new chromosomal binding 411 sites. Approximately half of these binding sites are in promoters within intergenic regions, while 412 the other half are within genes. Almost all of the novel binding sites that impact transcription are 413 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2026. ; https://doi.org/10.64898/2026.01.27.701997doi: bioRxiv preprint located proximal to TSSs in positions consistent with class ll promoters. Furthermore, the 414 occupancy of CRP K52Q at many of these sites positively regulates the production of RNAs that 415 have not previously been reported. Our results suggest that acetylation of CRP K52 engages 416 AR3 leading to transcription of novel RNAs that rewire sugar utilization, block the acetate switch, 417 and increase virulence gene expression. We further show that at least one sRNA activated by 418 CRP K52Q, crbZ, is expressed by WT V. cholerae in maltose-containing medium and represses 419 the acetate switch. All of these findings are consistent with a role for CRP K52Ac in remodeling V. 420 cholerae carbon utilization and virulence to optimize survival in the human intestine. 421 In LB broth cultures, CRP K52Q occupancy was noted at hundreds of sites not occupied by WT 422 CRP. Using MEME analysis to examine the consensus binding sequences for WT CRP and CRP 423 K52Q, we discovered that many CRP K52Q binding sites belong to one of two consensus 424 sequences, one that is similar to that of WT CRP and an additional one in which half of the 425 inverted repeat is conserved and the other half is highly degenerate. Initial studies describing 426 activation of transcription by AR3 documented an interaction with region 4 of the σ 70 subunit of 427 RNAP at Class ll promoters (18, 59). However, not all the binding sites unique to the CRP K52Q 428 mutant are located in Class ll promoters near TSSs (38). We examined the possibility that the 429 interaction of the downstream CRP monomer with the σ 70 subunit of RNAP positioned either at a 430 promoter or within a gene in a paused elongation complex might stabilize binding of CRP K52Q to 431 DNA. Our results suggest this is not the case. Instead, CRP K52Q appears to be able to bind 432 DNA when only one half of the consensus inverted repeat is a good match regardless of an 433 interaction with σ 70. Furthermore, it is able to modulate transcript abundance when bound at 434 these sites if they are positioned within promoters demonstrating that CRP K52Q is competent to 435 activate transcription at these sites. Further studies are required to understand the mechanism 436 by which the K52Q substitution promotes binding and activates transcription at degenerate 437 operator sites. 438 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2026. ; https://doi.org/10.64898/2026.01.27.701997doi: bioRxiv preprint The CRP K52Q substitution, which mimics acetylation, modifies the V. cholerae transcriptome by 439 altering expression of many genes known to be controlled by WT CRP and generating several 440 putative novel RNAs. One of these sRNAs, CrbZ, is responsible for a subset of the changes 441 observed in the CRP K52Q transcriptome. Because so many of CRP K52Q’s binding sites are 442 within genes and activate antisense transcription, it is interesting to speculate that sRNA 443 regulation plays a dominant role in modulation of the CRP regulon by K52Ac. By using 444 conditionally expressed sRNAs, CRP can retain WT expression of most of its regulon, while 445 modifying transcription of only those genes that are required to respond to a change in carbon 446 source availability. 447 Bacterial carbon overflow metabolism is defined by fermentation of sugars during rapid growth in 448 oxygen-rich environments (60). This results in generation of large amounts of short chain fatty 449 acids such as acetate and is predicted to increase protein acetylation in general. We show here 450 that CRP acetylation is increased during exponential growth in defined medium supplemented 451 with sucrose. Because we did not detect acetylation of CRP K52 under these conditions, we 452 hypothesize that the abundance of V. cholerae CRP K52Ac depends on the availability of specific 453 sugars. 454 When carbon sources become scarce, uptake of acetate is initiated in a process known as the 455 acetate switch (55). We predict that acetylation of CRP signals excess carbon availability to 456 enhance carbon overflow metabolism while suppressing the acetate switch. In fact, the response 457 of a K52Q mutant is concordant with these predictions. Interestingly, while the transcript 458 abundance of genes involved in uptake and utilization of many PTS-dependent sugars is 459 decreased, the abundance of those involved in uptake of trehalose is increased. This is intriguing 460 because trehalose has been shown to counteract the detrimental effects of protein acetylation 461 (61). Therefore, a selective increase in trehalose uptake and generation may be beneficial in the 462 setting of excess carbon availability. Taken together, these data suggest that CRP K52Ac 463 prepares V. cholerae to thrive in specific carbon-rich environments. 464 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2026. ; https://doi.org/10.64898/2026.01.27.701997doi: bioRxiv preprint CRP represses virulence gene expression and yet, in spite of increased virulence gene 465 expression, a ∆ crp mutant has a massive colonization defect in vivo (23, 33, 52). Furthermore, 466 transposon insertions in maltose metabolism genes, whose transcript abundance is increased in 467 the CRP K52Q mutant, had a competitive disadvantage in colonization of the infant rabbit 468 intestine (33). This suggests that virulence factors may not be repressed by CRP in vivo and that 469 regulation of sugar utilization by CRP is essential for V. cholerae growth in the mammalian 470 intestine. Carbon sources such as citrate, galactose, and maltose, which may promote CRP K52 471 acetylation, are abundant in the human diet and principally absorbed in the small intestine where 472 V. cholerae TCP-dependent colonization is observed (62). Acetylation of CRP in vivo would allow 473 for maximal virulence gene expression while preserving regulation of sugar utilization by CRP. 474 V. cholerae is found in human and environmental host intestines but is also cultured from the 475 aquatic environment during epidemics. As carbon is limiting in most natural aquatic 476 environments, it seems more likely that activation of carbon overflow metabolism by acetylated 477 transcription factors such as CRP is operative in the host intestine. Therefore, we propose that 478 acetylation of CRP K52 in vivo activates a transcriptional pattern that aids both colonization of 479 and replication in the mammalian intestine. 480 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2026. ; https://doi.org/10.64898/2026.01.27.701997doi: bioRxiv preprint

481 A complete reference list of strains used in this study is provided in Table S5. Vibrio cholerae 482 derived from the O1 biovar El Tor strain C6706str2 was cultured in Lysogeny Broth (LB; Difco 483 244620; 10 g tryptone, 5 g yeast extract, 10 g NaCl per liter) or in M9 minimum medium (1× M9 484 salts; Difco) supplemented with 1 mM MgSO4, 0.1 mM CaC2 and 0.4% [w/v] of the indicated 485 carbon source. Escherichia coli strains used for cloning were grown in LB only. Cultures were 486 incubated at 27 ˚C (V. cholerae) or 37 ˚C (E. coli) with shaking at 200 rpm. V. cholerae strains 487 were maintained in the presence of streptomycin (100 µg/mL), and, when required, carbenicillin 488 (100 µg/mL) was added for plasmid selection. Biological triplicates were performed for all 489 experiments. 490 Plasmid construction and mutant generation. PCR-amplified fragments were cloned into the 491 XhoI site in the pWM91 vector to generate constructs for in-frame deletions or chromosomal 492 complementation using New England Biolabs (NEB) HiFi DNA Assembly (NEB, E2621L). For 493 gene overexpression, PCR products containing the gene and its native promoter were cloned into 494 the XhoI site of the pFLAG-CTC IPTG-inducible vector. V. cholerae strains were generated via 495 homologous recombination using the Sanger-sequenced plasmids introduced by conjugation from 496 E. coli SM10 λ pir donor strains. Selection with 12% sucrose and PCR screening were used to 497 confirm successful in-frame deletion. Chromosomal complementation was achieved by allelic 498 exchange at the native locus in a deletion background strain. 499 Chromatin immunoprecipitation and sequencing (ChIP-seq). ChIP was performed as 500 previously described with minor modifications (34). Briefly, protein-DNA complexes were cross-501 linked and immunoprecipitated with anti-V5 antibodies. The cross-links were subsequently 502 reversed and co-purifying DNA was purified using the Qiaquick PCR Purification kit. DNA 503 libraries were prepared with the NEBNext Ultra II DNA Library Prep Kit (NEB E7645S) according 504 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2026. ; https://doi.org/10.64898/2026.01.27.701997doi: bioRxiv preprint to the manufacturer’s instructions and sequenced on an Illumina NovaSeq 6000 V1.5 SP 100 505 cycles Full Flow Cell (650 million reads) at the Harvard Biopolymers Facility. 506 ChIP-seq computational analysis. Paired-end reads were aligned to the V. cholerae reference 507 genome (NC_002505.1; NC_002506.1) using Bowtie2 (v2.4.1). Following mapping, an in silico 508 size selection was performed to retain fragments shorter than 400 nt in length. The 3 mock 509 biological replicates were merged and subsampled (samtools v1.10, using htslib 1.10.2), so that 510 they had approximately 2 million reads more than any single ChIP replicate file. Peak calling was 511 performed with QuEST (version 2.424) to identify putative DNA-binding sites defined as regions 512 exhibiting more than two-fold enrichment relative to background from an untagged mock sample. 513 Peaks assignments were performed by finding the closest gene to the point of maximal 514 enrichment using custom perl scripts and/or manual curation of the data. Peaks positioned such 515 that no genes were located within 1000 bp of the site of maximal enrichment were excluded. For 516 each strain, peaks with enrichment factor (EF) > 4 and reproducible in at least two of three 517 biological replicates were retained for downstream analyses. All custom Python scripts used for 518 data processing, analysis, and figure generation are available at 519 https://github.com/renatoerss/vch-crp-acetylation, and summarized below. 520 Peaks identified in each strain were paired by genomic overlap based on the reported region start 521 and end coordinates. Shared and strain-specific datasets were summarized with Upset plots or 522 Venn diagrams. Transcription start sites (TSSs) were obtained from the differential RNA-seq 523 dataset of V. cholerae and used to compute TSS distances and define promoter classes (Class I 524 and II) (38). Genomic categories were defined as promoter (-200 to -1 nt relative of TSS), 5’-UTR 525 (+1 from TSS to the base preceding the annotated start codon), coding sequence (CDS), or other, 526 if not within these regions. 527 For motif discovery, 100-bp windows centered on the ChIP-seq peak maximum position were 528 extracted and analyzed with MEME-ChIP (MEME suite, https://meme-suite.org/meme/), using 529 MEME (anr model, 16-22 bp motif width), Centrimo for positional enrichment, and Tomtom for 530 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2026. ; https://doi.org/10.64898/2026.01.27.701997doi: bioRxiv preprint motif comparison against curated bacterial motif databases (63, 64). Integration with RNA-seq 531 used per-gene log2 FC (K52Q/WT). Peak categories (WT-only, WT + K52Q, and K52Q-only) are 532 represented in the figure. 533 RNA isolation and transcriptome analysis. Total RNA from mid-exponential-phase (OD600 0.4-534 0.6) culture was isolated using the Direct-zol RNA Miniprep Plus kit (Zymo Research) following 535 the manufacturer’s instructions. Genomic DNA was removed by in-column DNase treatment, and 536 RNA quality was assessed using a TapeStation system (Agilent Technologies). 537 RNA-seq libraries were prepared using a modified version of the RNAtag-Seq protocol starting 538 from 500/i5 ng of total RNA (65). Briefly, RNAs were fragmented, dephosphorylated, and ligated to 539 DNA adaptors carrying 5’-AN₈ -3’ barcodes with a 5’ phosphate and a 3’ blocking group. Barcoded 540 samples were pooled and subjected to rRNA depletion using the Ribo-Zero rRNA Removal Kit 541 (Epicentre). 542 Sequencing was performed on an Illumina NextSeq 500 platform, generating paired-end reads. 543 The average sequencing depth was ~12 million fragments/sample, with three biological replicates 544 per condition. Reads were aligned to the Vibrio cholerae O1 biovar El Tor str. N16961 (accession 545 numbers NC_002505.1 and NC_002506.1) genome using BWA , and read counts were assigned 546 to annotated genomic features using custom scripts provided by the Microbial ‘Omics Core 547 (MOC) at the Broad Institute (66). Differential expression analysis was performed using DESeq2, 548 and expression values were reported as fragments per kilobase per million mapped reads 549 (FPKM)(67). Visualization of RNA-seq coverage traces and manual gene-level inspection was 550 performed using the Integrative Genomics Viewer (68). 551 To assess sample reproducibility and global transcriptional differences within the CRP regulon, a 552 principal component analysis (PCA) was performed using genes with greater than two fold 553 change and adjusted P-value < 0.05. The analysis was conducted on normalized, log-transformed 554 read counts. The first two principal components accounted for the majority of the variance and 555 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2026. ; https://doi.org/10.64898/2026.01.27.701997doi: bioRxiv preprint separated biological replicates by strain. PCA was computed in R (version 4.4.3) using 556 stats::prcomp. 557 Quantitative reverse transcription PCR (qRT-PCR). Total RNA was extracted as previously 558 described. 500 ng of total RNA was reverse-transcribed using the SuperScript III First-Strand 559 Synthesis System (Invitrogen). qPCR reactions were performed in 10 µL volumes using PowerUp 560 SYBR Green Master Mix for qPCR (Applied Biosystems) on a QuantStudio 5 Real-Time PCR 561 System (Applied Biosystems). Gene-specific and housekeeping primers are listed in Table S5. 562 Relative transcript abundance was calculated by the ∆∆ Ct method, normalized to clpX (VC1921). 563 All reactions were performed in technical duplicates using three independent biological samples. 564 Extracellular acetate quantification. Extracellular acetate levels were determined using 565 EnzyChrom Acetate Assay Kit (BioAssay Systems, EOAC-100) following manufacturer’s protocol. 566 Culture supernatants were collected from mid-exponential (OD600 0.4-0.6) or stationary phase 567 cultures (18 h incubation under standard conditions). Samples were clarified by centrifugation and 568 filtered through 0.22 µm PVDF membranes. Acetate concentration was determined from a sodium 569 acetate standard curve and absorbance was measured at 570 nm using a SpectraMax iD5 570 microplate reader (Molecular Devices). 571 Circularization RT-PCR (cRT-PCR). The 5’ and 3’ ends of the crbZ transcript were mapped 572 according to a previously described RNA circularization protocol (69). Briefly, DNAse I treated 573 total RNA (2 µg) was decapped with 2 U of RNA 5’-pyrophosphohydrolase (RppH; New England 574 Biolabs) in 1× NEBuffer 2 for 1 h at 37 ˚C to generate 5’-monophosphorylated ends. The RNA 575 was purified using the RNA Clean & Concentrator kit (Zymo Research) and circularized overnight 576 at 16 ˚C using 1 U of T4 RNA Ligase 1 (NEB) in 1× RNA Ligase Buffer supplemented with 1 mM 577 ATP, 8 µL PEG 8000 (50%), 2 µL DMSO, and 0.75 U RNase inhibitor in a total volume of 35 µL. 578 Circularized RNA was cleaned again and eluted in nuclease-free water. cDNA synthesis was 579 carried out with 500 ng of circular RNA using the SuperScript III First-Strand Synthesis System 580 (Invitrogen) and random hexamer primers, following manufacturer’s protocol. The resulting cDNA 581 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2026. ; https://doi.org/10.64898/2026.01.27.701997doi: bioRxiv preprint was amplified by PCR using outward-facing primers specific for crbZ and designed to flank the 582 ligated junction (Supplementary Table S4). PCR reactions were carried out with Q5 High Fidelity 583 DNA Polymerase (New England Biolabs) for 30 cycles, and products were resolved on 2% 584 agarose gels. Control PCR reactions with genomic DNA were included to confirm primer 585 specificity. Specific amplicons (150-400 bp) detected in the CRP K52Q sample were purified and 586 sent for DNA sequencing. 587 Statistical analysis: All experiments included a minimum of biological triplicates. Graphs were 588 constructed and statistical analysis was carried out using GraphPad Prism (version 10.6.1). The 589 appropriate statistical test was applied to evaluate significance as noted in the figure legends. 590 Data availability statement. ChIP-seq and RNA-seq data have been deposited in NCBI's Gene 591 Expression Omnibus and are accessible through GEO Series accession number GSE310654. 592 Acknowledgements. This work was supported by NIH R01 AI112652 to P .I.W. and NIH 593 R01AI018045 to J.J.M. 594 595 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2026. ; https://doi.org/10.64898/2026.01.27.701997doi: bioRxiv preprint

596 1. C. L. Lawson et al., Catabolite activator protein: DNA binding and transcription activation. 597 Curr Opin Struct Biol 14, 10–20 (2004). 598 2. H. Youn, M. Carranza, cAMP Activation of the cAMP Receptor Protein, a Model Bacterial 599 Transcription Factor. J Microbiol 61, 277–287 (2023). 600 3. E. Sonnleitner, A Comparative Analysis: Molecular Mechanisms of Carbon Catabolite 601 Repression in Bacteria. Annual review of microbiology 79, 241–262 (2025). 602 4. L. Houot, S. Chang, C. Absalon, P. I. Watnick, Vibrio cholerae PTS control of carbohydrate 603 transport, biofilm formation, and colonization of the germ-free mouse intestine. Infection 604 and immunity (2010). 605 5. C. A. Hayes, T. N. Dalia, A. B. Dalia, Systematic genetic dissection of PTS in Vibrio 606 cholerae uncovers a novel glucose transporter and a limited role for PTS during infection of 607 a mammalian host. Mol Microbiol 104, 568–579 (2017). 608 6. B. S. Pickering, D. R. Smith, P . I. Watnick, Glucose-specific enzyme IIA has unique binding 609 partners in the vibrio cholerae biofilm. mBio 3, e00228–00212 (2012). 610 7. J. Daniel, Enzyme III stimulation of cyclic AMP synthesis in an Escherichia coli crp mutant. 611 Journal of bacteriology 157, 940–941 (1984). 612 8. S. Poncet et al. , Correlations between Carbon Metabolism and Virulence in Bacteria. 613 Contrib Microbiol 16, 88–102 (2009). 614 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2026. ; https://doi.org/10.64898/2026.01.27.701997doi: bioRxiv preprint 9. S. M. Espert, E. A. Elsinghorst, G. P. Munson, The tib adherence locus of enterotoxigenic 615 Escherichia coli is regulated by cyclic AMP receptor protein. Journal of bacteriology 193, 616 1369–1376 (2011). 617 10. Y. Akhter et al., Genome scale portrait of cAMP-receptor protein (CRP) regulons in 618 mycobacteria points to their role in pathogenesis. Gene 407, 148–158 (2008). 619 11. T. J. Kim et al., Direct transcriptional control of the plasminogen activator gene of Yersinia 620 pestis by the cyclic AMP receptor protein. Journal of bacteriology 189, 8890–8900 (2007). 621 12. C. C. Li, D. S. Merrell, A. Camilli, J. B. Kaper, ToxR interferes with CRP-dependent 622 transcriptional activation of ompT in Vibrio cholerae. Mol Microbiol 43, 1577–1589 (2002). 623 13. J. T. Ritzert, W. W. Lathem, Depletion of Glucose Activates Catabolite Repression during 624 Pneumonic Plague. Journal of bacteriology 200 (2018). 625 14. S. Busby, R. H. Ebright, Transcription activation at class II CAP-dependent promoters. Mol 626 Microbiol 23, 853–859 (1997). 627 15. R. M. Williams, V. A. Rhodius, A. I. Bell, A. Kolb, S. J. Busby, Orientation of functional 628 activating regions in the Escherichia coli CRP protein during transcription activation at 629 class II promoters. Nucleic acids research 24, 1112–1118 (1996). 630 16. V. A. Rhodius, D. M. West, C. L. Webster, S. J. Busby, N. J. Savery, Transcription 631 activation at class II CRP-dependent promoters: the role of different activating regions. 632 Nucleic acids research 25, 326–332 (1997). 633 17. R. Jin, K. A. Sharif, J. S. Krakow, Evidence for contact between the cyclic AMP receptor 634 protein and the delta 70 subunit of Escherichia coli RNA polymerase. The Journal of 635 biological chemistry 270, 19213–19216 (1995). 636 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2026. ; https://doi.org/10.64898/2026.01.27.701997doi: bioRxiv preprint 18. V. A. Rhodius, S. J. Busby, Interactions between activating region 3 of the Escherichia coli 637 cyclic AMP receptor protein and region 4 of the RNA polymerase sigma(70) subunit: 638 application of suppression genetics. Journal of molecular biology 299, 311–324 (2000). 639 19. D. West et al., Interactions between the Escherichia coli cyclic AMP receptor protein and 640 RNA polymerase at class II promoters. Mol Microbiol 10, 789–797 (1993). 641 20. R. Williams, A. Bell, G. Sims, S. Busby, The role of two surface exposed loops in 642 transcription activation by the Escherichia coli CRP and FNR proteins. Nucleic acids 643 research 19, 6705–6712 (1991). 644 21. W. Liang, A. Pascual-Montano, A. J. Silva, J. A. Benitez, The cyclic AMP receptor protein 645 modulates quorum sensing, motility and multiple genes that affect intestinal colonization in 646 Vibrio cholerae. Microbiology (Reading) 153, 2964–2975 (2007). 647 22. J. C. Fong, F. H. Yildiz, Interplay between cyclic AMP-cyclic AMP receptor protein and 648 cyclic di-GMP signaling in Vibrio cholerae biofilm formation. Journal of bacteriology 190, 649 6646–6659 (2008). 650 23. K. Skorupski, R. K. Taylor, Cyclic AMP and its receptor protein negatively regulate the 651 coordinate expression of cholera toxin and toxin-coregulated pilus in Vibrio cholerae. 652 Proceedings of the National Academy of Sciences of the United States of America 94, 653 265–270. (1997). 654 24. G. Kovacikova, K. Skorupski, Overlapping binding sites for the virulence gene regulators 655 AphA, AphB and cAMP-CRP at the Vibrio cholerae tcpPH promoter. Mol Microbiol 41, 393–656 407. (2001). 657 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2026. ; https://doi.org/10.64898/2026.01.27.701997doi: bioRxiv preprint 25. W. Liang, S. Z. Sultan, A. J. Silva, J. A. Benitez, Cyclic AMP post-transcriptionally regulates 658 the biosynthesis of a major bacterial autoinducer to modulate the cell density required to 659 activate quorum sensing. FEBS letters 582, 3744–3750 (2008). 660 26. M. S. Zahid et al., The cyclic AMP (cAMP)-cAMP receptor protein signaling system 661 mediates resistance of Vibrio cholerae O1 strains to multiple environmental 662 bacteriophages. Applied and environmental microbiology 76, 4233–4240 (2010). 663 27. J. Manneh-Roussel et al., cAMP Receptor Protein Controls Vibrio cholerae Gene 664 Expression in Response to Host Colonization. mBio 9 (2018). 665 28. L. M. Walker et al., A simple mechanism for integration of quorum sensing and cAMP 666 signalling in Vibrio cholerae. eLife 12 (2023). 667 29. M. Blokesch, Chitin colonization, chitin degradation and chitin-induced natural competence 668 of Vibrio cholerae are subject to catabolite repression. Environ Microbiol 14, 1898–1912 669 (2012). 670 30. K. Liimatta et al., A Putative Acetylation System in Vibrio cholerae Modulates Virulence in 671 Arthropod Hosts. Applied and environmental microbiology 84 (2018). 672 31. I. Muzhingi et al., Modulation of CrbS-Dependent Activation of the Acetate Switch in Vibrio 673 cholerae. Journal of bacteriology 200 (2018). 674 32. M. Lo Scrudato, S. Borgeaud, M. Blokesch, Regulatory elements involved in the 675 expression of competence genes in naturally transformable Vibrio cholerae. BMC 676 microbiology 14, 327 (2014). 677 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2026. ; https://doi.org/10.64898/2026.01.27.701997doi: bioRxiv preprint 33. Y. Fu, M. K. Waldor, J. J. Mekalanos, Tn-Seq analysis of Vibrio cholerae intestinal 678 colonization reveals a role for T6SS-mediated antibacterial activity in the host. Cell host & 679 microbe 14, 652–663 (2013). 680 34. J. A. Gibson, M. J. Gebhardt, R. Santos, S. L. Dove, P. I. Watnick, Sequestration of a dual 681 function DNA-binding protein by Vibrio cholerae CRP. Proceedings of the National 682 Academy of Sciences of the United States of America 119, e2210115119 (2022). 683 35. J. Pan, R. Chen, C. Li, W. Li, Z. Ye, Global Analysis of Protein Lysine Succinylation Profiles 684 and Their Overlap with Lysine Acetylation in the Marine Bacterium Vibrio parahemolyticus. 685 Journal of proteome research 14, 4309–4318 (2015). 686 36. R. Davis et al., An acetylatable lysine controls CRP function in E. coli. Mol Microbiol 107, 687 116–131 (2018). 688 37. V. A. Rhodius, S. J. Busby, Transcription activation by the Escherichia coli cyclic AMP 689 receptor protein: determinants within activating region 3. Journal of molecular biology 299, 690 295–310 (2000). 691 38. K. Papenfort, K. U. Forstner, J. P. Cong, C. M. Sharma, B. L. Bassler, Differential RNA-seq 692 of Vibrio cholerae identifies the VqmR small RNA as a regulator of biofilm formation. 693 Proceedings of the National Academy of Sciences of the United States of America 112, 694 E766–775 (2015). 695 39. T. L. Bailey et al., MEME SUITE: tools for motif discovery and searching. Nucleic acids 696 research 37, W202–208 (2009). 697 40. S. J. W. Busby, D. F. Browning, Transcription activation in Escherichia coli and Salmonella. 698 EcoSal Plus 12, eesp00392020 (2024). 699 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2026. ; https://doi.org/10.64898/2026.01.27.701997doi: bioRxiv preprint 41. A. Mazumder, A. N. Kapanidis, Recent Advances in Understanding sigma70-Dependent 700 Transcription Initiation Mechanisms. Journal of molecular biology 431, 3947–3959 (2019). 701 42. J. Qian, B. Wang, I. Artsimovitch, D. Dunlap, L. Finzi, Force and the alpha-C-terminal 702 domains bias RNA polymerase recycling. Nature communications 15, 7520 (2024). 703 43. T. T. Harden et al. , Bacterial RNA polymerase can retain sigma70 throughout transcription. 704 Proceedings of the National Academy of Sciences of the United States of America 113, 705 602–607 (2016). 706 44. A. N. Kapanidis et al., Retention of transcription initiation factor sigma70 in transcription 707 elongation: single-molecule analysis. Molecular cell 20, 347–356 (2005). 708 45. A. R. Strom, I. Kaasen, Trehalose metabolism in Escherichia coli: stress protection and 709 stress regulation of gene expression. Mol Microbiol 8, 205–210 (1993). 710 46. N. More et al., Peptidoglycan Remodeling Enables Escherichia coli To Survive Severe 711 Outer Membrane Assembly Defect. mBio 10 (2019). 712 47. B. M. Childers, K. E. Klose, Regulation of virulence in Vibrio cholerae: the ToxR regulon. 713 Future microbiology 2, 335–344 (2007). 714 48. J. C. Fong, K. A. Syed, K. E. Klose, F. H. Yildiz, Role of Vibrio polysaccharide (vps) genes 715 in VPS production, biofilm formation and Vibrio cholerae pathogenesis. Microbiology 156, 716 2757–2769 (2010). 717 49. K. Barrasso et al., Impact of a human gut microbe on Vibrio cholerae host colonization 718 through biofilm enhancement. eLife 11 (2022). 719 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2026. ; https://doi.org/10.64898/2026.01.27.701997doi: bioRxiv preprint 50. J. K. Teschler, C. D. Nadell, K. Drescher, F. H. Yildiz, Mechanisms Underlying Vibrio 720 cholerae Biofilm Formation and Dispersion. Annual review of microbiology 76, 503–532 721 (2022). 722 51. V. Olivier, G. K. Haines, 3rd, Y . Tan, K. J. Satchell, Hemolysin and the multifunctional 723 autoprocessing RTX toxin are virulence factors during intestinal infection of mice with 724 Vibrio cholerae El Tor O1 strains. Infection and immunity 75, 5035–5042 (2007). 725 52. K. Skorupski, R. K. Taylor, Control of the ToxR virulence regulon in Vibrio cholerae by 726 environmental stimuli. Mol Microbiol 25, 1003–1009. (1997). 727 53. V. J. DiRita, Co-ordinate expression of virulence genes by ToxR in Vibrio cholerae. Mol 728 Microbiol 6, 451–458 (1992). 729 54. S. M. Wong, P . A. Carroll, L. G. Rahme, F. M. Ausubel, S. B. Calderwood, Modulation of 730 expression of the ToxR regulon in Vibrio cholerae by a member of the two-component 731 family of response regulators. Infection and immunity 66, 5854–5861 (1998). 732 55. A. J. Wolfe, The acetate switch. Microbiol Mol Biol Rev 69, 12–50 (2005). 733 56. S. Hang et al. , The acetate switch of an intestinal pathogen disrupts host insulin signaling 734 and lipid metabolism. Cell host & microbe 16, 592–604 (2014). 735 57. S. K. Dolan, M. Welch, The Glyoxylate Shunt, 60 Years On. Annual review of microbiology 736 72, 309–330 (2018). 737 58. P. O. Frendorf, I. Lauritsen, A. Sekowska, A. Danchin, M. H. H. Norholm, Mutations in the 738 Global Transcription Factor CRP/CAP: Insights from Experimental Evolution and Deep 739 Sequencing. Comput Struct Biotechnol J 17, 730–736 (2019). 740 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2026. ; https://doi.org/10.64898/2026.01.27.701997doi: bioRxiv preprint 59. M. A. Lonetto et al., Identification of a contact site for different transcription activators in 741 region 4 of the Escherichia coli RNA polymerase sigma70 subunit. Journal of molecular 742 biology 284, 1353–1365 (1998). 743 60. N. B. Chowdhury, W. L. Schroeder, L. Monteiro, K. E. Burnum-Johnson, mGem: Revisiting 744 bacterial overflow metabolism. mBio 16, e0119325 (2025). 745 61. M. Moruno Algara et al., Trehalose protects Escherichia coli against carbon stress 746 manifested by protein acetylation and aggregation. Mol Microbiol 112, 866–880 (2019). 747 62. M. J. Angelichio, J. Spector, M. K. Waldor , A. Camilli, Vibrio cholerae intestinal population 748 dynamics in the suckling mouse model of infection. Infection and immunity 67, 3733–3739 749 (1999). 750 63. T. L. Bailey, J. Johnson, C. E. Grant, W. S. Noble, The MEME Suite. Nucleic acids 751 research 43, W39–49 (2015). 752 64. S. L. Nystrom, D. J. McKay, Memes: A motif analysis environment in R using tools from the 753 MEME Suite. PLoS computational biology 17, e1008991 (2021). 754 65. A. A. Shishkin et al., Simultaneous generation of many RNA-seq libraries in a single 755 reaction. Nat Methods 12, 323–325 (2015). 756 66. H. Li, R. Durbin, Fast and accurate short read alignment with Burrows-Wheeler transform. 757 Bioinformatics 25, 1754–1760 (2009). 758 67. M. I. Love, W. Huber, S. Anders, Moderated estimation of fold change and dispersion for 759 RNA-seq data with DESeq2. Genome biology 15, 550 (2014). 760 68. J. T. Robinson et al. , Integrative genomics viewer. Nature biotechnology 29, 24–26 (2011). 761 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2026. ; https://doi.org/10.64898/2026.01.27.701997doi: bioRxiv preprint 69. J. H. Urban, J. Vogel, Two seemingly homologous noncoding RNAs act hierarchically to 762 activate glmS mRNA translation. PLoS biology 6, e64 (2008). 763 764 765 766 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2026. ; https://doi.org/10.64898/2026.01.27.701997doi: bioRxiv preprint Figure Legends 767 Figure 1: Point mutants designed to mimic unmodified and acetylated CRP modulate DNA 768 occupancy, gene transcription, and virulence. (A) Illustration showing the interaction of CRP 769 activating regions (AR) 1-3 with RNA polymerase (RNAP) subunits α and σ at Class l and Class ll 770 promoters. Created in BioRender. Watnick, P . (2026) https://BioRender.com/2vpegk4. (B) 771 Schematic showing the phosphoenolpyruvate phosphotransfer cascade. Enzyme l (El) is 772 phosphorylated by phosphoenolpyruvate (PEP). This phosphate is passed to Histidine protein 773 (HPr) and from there to enzymes llA and B (EllA, EllB). Enzyme llC (EllC) is an integral 774 membrane transporter that mediates tandem sugar transport and phosphorylation by EllB. 775 Created in BioRender. Watnick, P. (2025) https://BioRender.com/Sm3al62. (C) UpSet plot 776 showing common and unique CRP DNA occupancy sites identified by ChIP-seq for WT, 4×Q, and 777 4×R CRP. Occupancy sites were defined by an enrichment factor (EF) > 4 for at least 2 of 3 778 biological replicates. Shared occupancy events are indicated by connecting lines. (D) Principal 779 Component Analysis of differentially expressed genes (-1 ≤ log2(FC)≥ 1 and FDR < 0.05) based on 780 RNA-seq analysis of the indicated strains cultured in LB medium and harvested at mid-781 exponential growth phase. Biological triplicates are shown. 782 783 Figure 2. Many chromosomal binding sites unique to K52Q CRP are distinguished by their 784 operator sequence and position within coding sequences. (A) Total number of sites where 785 WT CRP and CRP K52Q binding reach EF >4. Bars indicate the average of three biological 786 replicates. Error bars represent the standard deviation. A student’s t test was used to calculate 787 significance. (B) Venn diagram showing total numbers of binding sites that are unique and 788 shared between WT CRP and a CRP K52Q mutant. These numbers reflect peaks that were 789 present in 2 out of the 3 biological replicates. EF>4 was used as a threshold. (C) Map showing 790 how positions of WT and K52Q binding sites were categorized, and enumeration of WT and K52Q 791 CRP chromosomal binding sites that are in the regions indicated. (D) Numbers of class I and 792 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2026. ; https://doi.org/10.64898/2026.01.27.701997doi: bioRxiv preprint class ll promoters identified for WT CRP and CRP K52Q. Class l promoters were defined as 793 within -52 to -70 of a TSS, while class ll promoters were defined as within -51 to -32 of TSS. (E) 794 Consensus motifs discovered by MEME (Multiple Em for Motif Elicitation, https://meme-795 suite.org/meme/). The number of sites and the corresponding E-values are indicated. E-values 796 are derived from the log-likelihood ratio (LLR) statistic used by MEME’s expectation-maximization 797 (EM) algorithm and represent the expected number of motifs with an equal or higher LLR 798 occurring by chance under the background model. (F) Graph showing numbers of chromosomal 799 binding sites used to build MEME 1 and 2 that are shared by WT CRP and the CRP K52Q mutant 800 or unique to the CRP K52Q mutant. (G) Schematic showing orientation of Meme ll operators in 801 which association of the unbound CRP K52Q monomer is stabilized by an interaction of AR3 with 802 σ 70. The green circle represents AR3. The arrow indicates the direction of transcript elongation. 803 Created in BioRender. Watnick, P . (2025) https://BioRender.com/2vpegk4. 804 Figure 3. CRP K52Q regulates transcription at many new class ll promoters and intragenic 805 binding sites. (A) Scatter plot of differential gene expression near sites of WT CRP occupancy 806 only, CRP K52Q occupancy only, or occupancy of both transcription factors. Numbers of 807 enrichment sites correlated with transcriptional activation or repression at Class I and Class II 808 promoters (B) shared by WT CRP and CRP K52Q and (C) unique to CRP K52Q. (D) Impact of 809 CRP K52Q binding sites within coding sequences (CDS) on transcription. A two-fold change in 810 transcription with FDR < 0.05 was considered to be significant. 811 Figure 4: CRP K52Q delays the acetate switch. (A) Heat map showing differential regulation of 812 genes essential for the acetate switch. Data is the mean of RNA-seq biological triplicates. (B) 813 qRT-PCR analysis of crbR and acs1 transcription in the indicated V. cholerae strains. The mean 814 of biological triplicates is shown. An ordinary one-way ANOVA was used to determine 815 significance. (C) Acetate concentrations in the supernatants of exponential and stationary phase 816 cultures of WT V. cholerae and CRP K52Q, and Δ crp mutants. The mean of biological triplicates 817 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2026. ; https://doi.org/10.64898/2026.01.27.701997doi: bioRxiv preprint is shown. A student’s t test was used to determine significance. (D) Growth over time of the 818 indicated strains in M9 medium supplemented with 0.4 % acetate. The average of biological 819 triplicates is shown. **** p<0.0001, *** p<0.001, ** p<0.01, * p<0.05, ns not significant. 820 Figure 5: The small RNA CrbZ is encoded antisense and within CrbS and activated by CRP 821 acetylation. (A) ChIP-seq and RNA-seq traces showing enrichment of CRP K52Q at a class ll 822 promoter and transcription of the crbZ sRNA gene, respectively. Traces are representative of 823 measurements in biological triplicate. These data are also shown in Fig S 8A. (B) qRT-PCR 824 quantification of transcription of crbZ+crbS and crbS in the indicated strains. The mean of 825 biological triplicates is shown. Error bars represent the mean. A one-way ANOVA or student’s t 826 test was used to assess significance. (C) Schematic showing the experimental design and (D) an 827 agarose gel to document the presence of the crbZ sRNA by ligation and PCR amplification. 828 Arrows indicate the position of the PCR primers. (E) qRT-PCR quantification of crbZ + crbS, crbS, 829 or acs1 transcription in WT V. cholerae cultured for 20h in LB broth or in M9 minimal medium 830 supplemented with 0.4% maltose or trehalose for 2 h. The mean of biological triplicates is shown. 831 Error bars represent the mean. A one-way ANOVA was used to assess significance. **** 832 p<0.0001, *** p <0.001, ** p<0.01, * p< 0.05, ns not significant. 833 Table 1: Genes differentially regulated relative to the Δ crp mutant in strains expressing the 834 indicated V5-tagged CRP isoforms from the chromosomal locus. 835 Figure S1: A CRP K52Q point mutant remains membrane-associated. 836 Figure S2. A CRP C-terminal V5 tag has only a small impact on the V. cholerae 837 transcriptome. 838 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2026. ; https://doi.org/10.64898/2026.01.27.701997doi: bioRxiv preprint Figure S3: Chromosomally-encoded CRP acetylation mutants are expressed at 839 comparable levels, grow at comparable rates in LB broth, and do not affect CRP 840 acetylation at other sites. 841 Figure S4: The CRP K52Q mutant modulates the transcript abundance of genes required 842 for utilization of fructose, trehalose, citrate, maltose mannitol, mannose, and cellobiose. 843 Figure S5: The CRP K52Q substitution alters growth of V. cholerae in a variety of sugars. 844 Figure S6: The CRP K52Q mutant increases the transcript abundance of genes required for 845 synthesis of the toxin co-regulated pilus, the VPS-dependent biofilm, and the RTX toxin. 846 Figure S7: In minimal medium supplemented with sucrose, V. cholerae CRP acetylation 847 decreases in stationary phase. 848 Figure S8: Six putative RNAs are uniquely activated by CRP K52Q. 849 Figure S9: Sequence alignment of the circularized RNA junction. 850 Figure S10: The sRNA CrbZ regulates the acetate switch and maltose metabolism. 851 Figure S11: Strand-specific RT-qPCR demonstrates activation of crbZ transcript 852 abundance in maltose-containing medium. 853 Table S1: WT CRP, CRP 4XR, and CRP 4XQ ChIP-seq analysis. 854 Table S2: Deseq comparisons of WT (C6706) , WT CRP-V5, CRP K52Q-V5, CRP 3XQ-V5, 855 CRP 4XQ-V5, CRP 4XR-V5 and Δ crp transcriptomes in exponential LB broth cultures. 856 Table S3: WT and K52Q ChIP-seq data 857 Table S4: List of meme sites for WT CRP and CRP K52Q 858 Table S5: Strains, plasmids, and primers used in this study 859 860 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2026. ; https://doi.org/10.64898/2026.01.27.701997doi: bioRxiv preprint A B PC2 (12.73%) CD (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2026. ; https://doi.org/10.64898/2026.01.27.701997doi: bioRxiv preprint A # binding EF>4 ✱✱✱ EF>4 WT (n=65) K52Q (n=267) B C # binding sites D # binding sites WT MEM E-1 K52Q M EME -1 K52Q M EME -2 0 25 50 4 29 41 F E G (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2026. ; https://doi.org/10.64898/2026.01.27.701997doi: bioRxiv preprint A Activated Repressed Not sig. # transcripts D Class I Class II 0 10 20 30 3 5 1 22 3 # binding sites Shared Activated Repressed Not sig. B Class I Class II 0 10 20 30 11 22 3 22 10 # binding sites K52Q only Activated Repressed Not sig. C (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2026. ; https://doi.org/10.64898/2026.01.27.701997doi: bioRxiv preprint B OD600 (M9/Acetate) ∆crp Exponential Stationary 0 2 4 6 8 10 12 ✱✱✱✱ ns ✱ WT K52Q Δcrp A C D Relative acs-1 abundance ✱✱ ✱✱ WTK52Q∆crp 0.0 0.5 1.0 1.5 ✱ ✱✱ (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2026. ; https://doi.org/10.64898/2026.01.27.701997doi: bioRxiv preprint A B C RNAcirc MW WT K52Q WT K52Q DNA 0.5 kb – 0.2 kb – D ✱✱✱✱ ✱✱✱✱ E 20h LB 2h maltose 2h trehalos e 0 1 2 3 4 crbZ+crbS Transcript abundance ✱✱ ns 20h LB 2h maltose 2h treh alose 0 1 2 3 ns 20h LB 2h ma ltose 2h tre ha lose 0.0 0.5 1.0 1.5 2.0 acs1 transcript abundance ✱✱ ns (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2026. ; https://doi.org/10.64898/2026.01.27.701997doi: bioRxiv preprint Table 1: Transcripts differentially regulated relative to the Δ crp mutant in strains expressing the indicated chromosomal V5-tagged CRP isoforms. vs Δ crp All transcripts Coding transcripts only Up Down Up Down WT 1405 1035 639 508 4XR 1455 981 658 485 4XQ 1137 754 525 341 3XQ 800 660 395 309 K52Q 1606 1320 729 566 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 27, 2026. ; https://doi.org/10.64898/2026.01.27.701997doi: bioRxiv preprint