Introduction

The gram-negative facultative anaerobic bacterium Vibrio cholerae is the causative agent of cholera, a severe waterborne diarrheal disease that has caused seven pandemics since 1817. The current, seventh pandemic started in 1961 and continues to be a significant cause of death globally, with 3–5 million new infections and more than 100,000 deaths every year1. Among pandemic strains, only those from the sixth and seventh cholera pandemics are available for modern scientific study, and phylogenetic analyses suggested that pandemic strains originated from nonpandemic strains2,3. V. cholerae has a complex life cycle involving transitions between the aquatic environment, such as surface seawater, and the human small intestine4. V. cholerae exhibits the ability to survive in the aquatic environment year-around5,6. To survive in the environment, V. cholerae has to respond to changing conditions, including temperature shifts, osmotic stress, and nutrient limitation. The bioavailability of carbon sources is limited in seawater7. Several strategies enabling V. cholerae to acquire or store carbon sources in this environment have been reported, such as utilizing chitin in the exoskeleton of marine crustaceans and using accumulated intracellular glycogen stores6. In addition, extracellular DNA and dissolved organic matter of seawater may also support the growth of V. cholerae8,9. However, the mechanisms that V. cholerae utilizes to efficiently uptake and metabolize carbon sources of seawater are still not fully understood.

Once ingested by humans, V. cholerae passes through the stomach, grows in the small intestine, transits through the mucus, colonizes the surface of the intestinal epithelial cells, and causes severe diarrhea10. V. cholerae undergoes a significant shift in environmental conditions as it transitions from the aquatic environment into and within the human host. Adaptation to these changes involves changes in gene expression including the precise induction of virulence genes in the small intestine11. Like other enterobacteria12,13, V. cholerae mainly utilizes sugars present in the human intestine as the major carbon and energy sources14,15.

In bacteria, there are three types of active sugar transport systems: phosphoenolpyruvate: carbohydrate phosphotrasferase system (PTS), major facilitator superfamily (MFS) transporters, and ATP-binding cassette transporters. There are 13 distinct PTS transporters in V. cholerae, including a PTS specific for fructose16,17. Sugar transport via PTS lowers the activity of adenylate cyclase that is responsible for the synthesis of cyclic adenosine monophosphate (cAMP)18. cAMP receptor protein (CRP) binding with cAMP negatively regulates the expression of several virulence genes in V. cholerae19,20,21. MFS transporters catalyze sugar translocations across the membrane against its concentration gradient by using the energy released from the energetically downhill movement of a cation22. LacI-type regulators play important roles in bacterial carbon utilization23. They can coordinate available sugars with the expression of corresponding transport and catabolic genes, benefiting bacteria by conserving energy for mRNA and protein production24,25. There is a huge difference in the availability and concentration of sugars between seawater and human intestine12,26,27, and it remains to be understood whether LacI-type regulators play a role in regulating sugar uptake and utilization of pandemic V. cholerae for adapting to changes between these two environments.

Previous studies on pandemic V. cholerae strains have mostly focused on the identification of virulence factors expressed by the bacteria in the human intestine; these studies have shown that the generation and evolution of pandemic strains is associated with the acquisition of virulence elements, such as the ctx phage encoding cholera toxin, Vibrio pathogenicity island (VPI), and Vibrio 7th Pandemic islands I and II (VSP-I and II)2,28. Pandemic V. cholerae strains also acquired several abilities that benefit their survival and persistence in the environment, such as formation of biofilms6. It is highly likely that acquisition of factors to efficiently uptake and utilize sugars is essential for the survival of pandemic V. cholerae in the aquatic environment where carbon source is limited; however, this was often overlooked in previous research studies.

Here we aimed to identify the carbon uptake transporter that might confer an advantage to pandemic V. cholerae strains as well as the mechanisms by which carbon transport is regulated during the bacteria life cycle. We first performed a large-scale comparative genomics analysis on all 762 available pandemic V. cholerae genomes and 433 available nonpandemic V. cholerae genomes. We found that fruT encoding a fructose/H+ symporter and fruI encoding a LacI-type regulator that negatively regulates fruT expression are present in all pandemic strains, while are only present in 27.3% of nonpandemic strains. Fructose, whose concentration is in the range of 0.12–0.55 μM, is a major sugar present in seawater26. Fructose uptake assays revealed that FruT is responsible for uptaking low-concentration fructose. We showed that in seawater, the expression of fruT is turned on as the repression of fruT by FruI is released in response to high O2 concentration, and deletion of fruT made pandemic V. cholerae exhibit a substantial survival disadvantage in seawater. Animal experiments showed that deletion of fruI significantly inhibited colonizing ability of V. cholerae in the small intestine of mice. Through a series of molecular biology experiments, we showed that in the small intestine, FruI directly downregulates fruT expression by sensing the low O2 signal. This promotes the uptake of high-concentration fructose via PTS in the small intestine, and enhanced PTS activity leads to a decrease in the intracellular cAMP level, and upregulation of virulence gene expression. In summary, our work revealed that acquisition of this fructose uptake system not only promotes bacterial survival in seawater, but also maintains bacterial pathogenicity in animal intestine.

Results

FruI is a LacI-type regulator involved in fructose utilization

There are 11 genes encoding LacI-type regulators (VCA0673, VC0289, VC0909, VC1286, VCA0132, VCA0519, VC2677, VC2337, VC1721, VC1557, and VCA0654) in the genome of V. cholerae strain N16961, a representative strain of the seventh cholera pandemic29. To reveal whether some of these 11 LacI-type regulator genes are more associated with pandemic V. cholerae than nonpandemic V. cholerae, a comparative genomics analysis was performed using all 762 publicly available genomes of pandemic strains and 433 publicly available genomes of nonpandemic strains. The comparative genomics analysis showed that these 11 genes are present in all pandemic strains (Supplementary Data 1). They are also present in all or most nonpandemic strains, except for VCA0673 that is present in 27.3% of nonpandemic strains (Supplementary Data 1). It suggests that VCA0673 is considerably associated with pandemic strains and may play a role in the pathogenicity and/or environmental survival of pandemic strains. However, the function of VCA0673 remains unclear.

BlastP analysis showed that VCA0673 shares sequence similarity to LacI-type regulators FruR, TreR, MalR, and RbsR, which regulate the transport and/or utilization of fructose, trehalose, maltose, and ribose, respectively. To assess whether VCA0673 is indeed involved in the transport and/or utilization of these sugars, we constructed a deletion mutant strain Δvca0673 in strain El2382, a seventh-pandemic strain stored in our lab whose genome has been sequenced30. Then we assessed how deletion of this gene affected growth of V. cholerae in medium with each of those four sugars as the only carbon source, by growing the wild-type (WT) strain and Δvca0673 in M9 medium supplemented with 10 or 30 mM fructose, trehalose, maltose, or ribose, respectively. There is no significant difference in bacterial growth curves between Δvca0673 and WT in M9 medium supplemented with trehalose, maltose, or ribose, respectively, indicating VCA0673 is not involved in the regulation for the transport and utilization of these sugars (Supplementary Fig. 1a–h). However, in M9 medium supplemented with fructose, the bacterial growth curves showed that Δvca0673 underwent a longer exponential phase than WT, and reached a higher final cell density than WT (Fig. 1a and Supplementary Fig. 1i). This effect was reversed by complementation of the Δvca0673 (Fig. 1a and Supplementary Fig. 1i). This result suggests that Δvca0673 exhibits better ability to utilize fructose than WT. Thus, the LacI-type regulator encoded by vca0673 may be involved in the regulation for fructose transport and/or utilization in V. cholerae and, consequently, the gene was named fruI and the Δvca0673 was renamed ΔfruI.

Fig. 1: FruI represses the expression of the putative fructose transporter FruT.
figure 1

a Growth curves of WT, ΔfruI, and ΔfruI+ in M9 medium containing 30 mM fructose as the only carbon source. Data represent the mean ± SD (n = 3). b qRT-PCR expression level of fruT in WT, ΔfruI, and ΔfruI+ in M9 medium containing 30 mM fructose as the only carbon source. Data represent the mean ± SD (n = 3). c Growth curves of WT, ΔfruT, and ΔfruT+ in M9 medium containing 30 mM fructose as the only carbon source. Data represent the mean ± SD (n = 3). d Growth curves of WT, Δfpr, and ΔfruTΔfpr in M9 medium containing 30 mM fructose as the only carbon source. Data represent the mean ± SD (n = 3). e EMSA of the specific binding of purified FruI protein to the promoter region of fruT and kana (negative control). f FruI binds to a motif in the fruT promoter region. The protected region shows a significantly reduced peak intensities (blue) pattern than seen in compared with those of the control (red). The identified FruI-binding motif is shown in a box at the bottom of the figure. Images are representative of three independent experiments (e). Significance is determined by two-tailed unpaired Student’s t test (b) or two-way ANOVA (a, c, d) and indicated as the P value; n.s. means no significant difference. Source data are included in Source Data file.

FruI represses the expression of the putative fructose transporter FruT by directly binding to its promoter

To investigate the molecular mechanisms through which FruI regulates fructose transport and/or utilization, we compared the global gene expression of ΔfruI and WT at the stationary phase in M9 medium supplemented with 30 mM fructose as the only carbon source. We identified 456 genes that were significantly differentially expressed between ΔfruI and WT (Supplementary Data 2). Among these 456 genes, vca0669 is the most upregulated gene (394-fold) in ΔfruI compared to that of WT. vca0669 is located 1.3 kb upstream of fruI (Supplementary Fig. 1k), and both of them are within a genomic island (GI-7) revealed by comparative genomics analysis on V. cholerae31. qRT-PCR assays confirmed that the expression of vca0669 exhibited a 2.8- and 17.4-fold increase in ΔfruI compared to WT at the stationary phase in M9 medium supplemented with 10 or 30 mM fructose (Fig. 1b and Supplementary Fig. 1l). Complementation of ΔfruI with fruI restored vca0669 expression to the WT level (Fig. 1b). VCA0669 encodes a potential MFS sugar transporter29. Bacterial growth curves showed that deletion of vca0669 did not influence bacterial growth in M9 medium supplemented with 10 or 30 mM glucose, trehalose, maltose, and ribose, respectively (Supplementary Fig. 1a–h). However, in M9 medium supplemented with 10 or 30 mM fructose, Δvca0669 underwent a shorter exponential phase than WT, and the final cell density of Δvca0669 was also lower than that of WT (Fig. 1c and Supplementary Fig. 1j), indicating the deletion of vca0669 impairs bacterial ability to utilize fructose partially. This effect was reversed by complementation of Δvca0669 (Fig. 1c and Supplementary Fig. 1j). These data suggest that FruI represses the expression of vca0669 and that vca0669 may encode a fructose transporter. Consequently, vca0669 was named fruT and the Δvca0669 and complemented strains were named ΔfruT and ΔfruT+, respectively.

Since the ΔfruT could still grow in M9 medium with fructose, even though it entered into stationary phase earlier than the WT, we evaluated if fructose PTS plays a role during the exponential phase growth. fpr encodes a multidomain protein Fpr included in fructose PTS of V. cholerae, which consists of an enzyme IIA (EIIA)-like domain and a HPr domain17. Enzyme I (EI) autophosphorylates in the presence of phosphoenolpyruvate and then subsequently transfers this phosphate to FPr17. Bacterial growth curves in M9 medium supplemented with 10 or 30 mM fructose showed that both Δfpr and ΔEI exhibited a lower growth rate than WT in the exponential phase, and reached a lower final cell density than WT (Supplementary Fig. 1m, n). In addition, ΔfruTΔfpr barely grew in M9 medium supplemented with 10 or 30 mM fructose, which reached a 1.55- to 1.83-fold lower final cell density than that of WT, Δfpr, and ΔfruT (Fig. 1d and Supplementary Fig. 1o), indicating that both FruT and PTS are required for pandemic V. cholerae growth with fructose as the only carbon source.

To evaluate whether FruI directly regulates fruT expression, we performed electrophoretic mobility shift assays (EMSAs) using purified 6×His-tagged FruI and the potential promoter region of fruT corresponding to 200 bp upstream of fruT. The result showed that FruI specifically binds to the fruT promoter (Fig. 1e). A dye-based DNase I footprinting assay revealed a specific FruI-bound sequence containing an 31–base pair motif (5-AAATCTATTCTTCTTCGCGAAGCAAGGTTGA-3; −106 to −74 from the fruT translational start site) in the fruT promoter region (Fig. 1f). These data indicate that FruI in V. cholerae directly downregulates the expression of fruT.

FruT is a H+ symporter

To investigate the specific function of FruT, we performed protein domain structure analysis using SMART, which indicates that FruT is a hypothetical H+ symporter. We also found that FruT shares 32.6% similarity with MstA, a H+ symporter identified in Aspergillus niger, which is able to transporter the sugar simultaneously with the movement of protons32. To corroborate this finding, symport assays were performed to investigate whether fructose transport mediated by FruT is associated with H+ movement. Changes in the extracellular pH of unbuffered cell suspensions of WT, ΔfruT, and ΔfruT+ upon the addition of fructose were monitored. Significant extracellular alkalinization reflecting a reduction in the extracellular H+ concentration was observed in the media of WT and ΔfruT+ but not in the medium of ΔfruT (Fig. 2). These data suggest that FruT transports fructose by a H+ symport mechanism.

Fig. 2: Fructose symport activity of FruT.
figure 2

Effect on the extracellular pH elicited measured before and after the addition of fructose to a final concentration of 30 mM to unbuffered aqueous cell suspensions of WT, ΔfruT, and ΔfruT+. The extracellular pH was measured by Seahorse XF analyzer. The dotted lines indicate the times of fructose addition. Data are shown as the mean ± SD (n = 3). Significance is determined by two-way ANOVA and indicated as the P value; n.s. means no significant difference. Source data are included in Source Data file.

FruI contributes to the virulence of V. cholerae by repressing the expression of fruT and decreasing the intracellular cAMP level

The above findings indicated that FruT likely encodes a fructose/H+ symporter that is required for V. cholerae to grow in fructose in vitro; we also found that fruT expression is negatively regulated by FruI, whose deletion enhances V. cholerae growth in fructose in vitro. Next, we investigated whether FruI and FruT are required for V. cholerae colonization in small intestine by performing competitive infection assays in mice. The competitive index (CI) value of ΔfruI versus WT was 0.45 in the small intestine of mice (Fig. 3a), while the CI value of complemented strain (ΔfruI+) versus WT was 1.19 (Fig. 3d). This indicates that compared with the WT, ΔfruI exhibits a significantly defective colonizing ability in the small intestine and that FruI contributes to intestinal colonization of V. cholerae. The CI value of ΔfruT versus WT was 3.18 in the small intestine of mice (Fig. 3b), while the CI value of ΔfruT+ versus WT was 0.86 (Fig. 3d), indicating that ΔfruT exhibited a significant colonization advantage over WT. Furthermore, the competitive infection assays showed that the CI value of the double mutant ΔfruIΔfruT versus WT was 3.26 in the small intestine of mice, which is similar to the CI value of ΔfruI versus WT (Fig. 3c), and ΔfruIΔfruT competed evenly with ΔfruT in the small intestine of mice (CI = 1.12) (Fig. 3e). These data indicate that in the absence of the fruT, deletion of fruI does not have an effect on V. cholerae colonizing ability. In addition, qRT-PCR assays showed that fruT expression exhibited a significant increase in ΔfruI compared to WT in the small intestine of mice, indicating FruI represses the expression of fruT in the small intestine (Fig. 3f). These data suggest that FruI contributes to intestinal colonization of V. cholerae by downregulating fruT expression.

Fig. 3: FruI contributes to the virulence of V. cholerae by repressing the expression of fruT and decreasing the intracellular cAMP level.
figure 3

ad Competition assay comparing the colonizing ability of WT, ΔfruI (a), ΔfruT (b), ΔfruIΔfruT (c), ΔfruI+ (d), and ΔfruT+ (d) in infant mouse intestine (n = 6 mice per group). CI is defined as the output ratio of mutant strains to WT lacZ- divided by the input ratio of mutant strains to WT lacZ-, or output ratio of complemented strains to WT lacZ- with pBAD33 divided by the input ratio of complemented strains to WT lacZ- with pBAD33. Each symbol represents the CI in an individual mouse; horizontal bars indicate the median. e Competition assay comparing the colonizing ability of ΔfruIΔfruT and ΔfruT in infant mouse intestine (n = 6 mice per group). CI is defined as the output ratio of ΔfruIΔfruT or ΔfruT to ΔfruT lacZ- divided by the input ratio of ΔfruIΔfruT or ΔfruT to ΔfruT lacZ-. Each symbol represents the CI in an individual mouse; horizontal bars indicate the median. f qRT-PCR expression level of fruT in WT and ΔfruI in mouse small intestine. Data represent the mean ± SD (n = 3). g, h qRT-PCR expression level of virulence genes in WT, ΔfruT, and ΔfruI in M9 medium containing 20 mM fructose as the only carbon source (g) and in mouse small intestine (h). Data represent the mean ± SD (n = 3). i Representative western blotting image and quantitative analysis of cholera toxin in WT, ΔfruT, and ΔfruI in M9 medium containing 20 mM fructose as the only carbon source. RNA polymerase (RNAP) was used as a loading control. Data represent the mean ± SD (n = 3). j Competition assay comparing the colonizing ability of WT, Δfpr, and ΔfruK in infant mouse intestine (n = 7 mice per group). CI is defined as the output ratio of these strains to WT lacZ- divided by the input ratio of these strains to WT lacZ-. Each symbol represents the CI in an individual mouse; horizontal bars indicate the median. k qRT-PCR expression level of fpr in WT, ΔfruI, and ΔfruT in M9 medium containing 20 mM fructose as the only carbon source. Data represent the mean ± SD (n = 3). l Representative western blotting image and quantitative analysis of Fpr in WT, ΔfruT, and ΔfruI in M9 medium containing 20 mM fructose as the only carbon source. RNA polymerase (RNAP) was used as a loading control. Data represent the mean ± SD (n = 3). m Phosphorylation status analysis of Fpr in WT, ΔfruI, and ΔfruT in M9 medium containing 20 mM fructose as the only carbon source. P phosphorylated protein, N non-phosphorylated protein. Data represent the mean ± SD (n = 3). n qRT-PCR expression level of fpr in WT, ΔfruI, and ΔfruT in mouse small intestine. Data represent the mean ± SD (n = 3). o cAMP concentrations in WT, ΔfruI, and ΔfruT in M9 medium containing 20 mM fructose as the only carbon source. Data represent the mean ± SD (n = 3). p qRT-PCR expression levels of virulence genes in ΔfruT in M9 medium containing 20 mM fructose with or without 1 mM cAMP. Data represent the mean ± SD (n = 3). Significance is determined by two-sided Mann–Whitney U test (ae, j) or two-tailed unpaired Student’s t test (fI, kp) and indicated as the P value; n.s. means no significant difference. Source data are included in Source Data file.

To understand the mechanism underlying the colonization advantage of ΔfruT, we analyzed the expression of the virulence genes toxR, tcpP, and toxT, which is required for intestinal colonization of V. cholerae, by qRT-PCR in ΔfruT compared to WT in M9 medium supplemented with 20 mM fructose, which mimics the concentration of fructose in the small intestine environment33, and in the small intestine of mice. We found that the expression of the virulence genes was significantly higher in ΔfruT compared to the WT strain in vitro and in vivo (Fig. 3g, h). In consistent, western blotting assays showed that the production of cholera toxin, a major virulence factor of V. cholerae, was significantly higher in ΔfruT compared to WT in M9 medium supplemented with 20 mM fructose (Fig. 3i). These results suggest that in the absence of fruT, the increased expression of virulence factors confers a colonization advantage to ΔfruT. Conversely, the increased expression of fruT in ΔfruI could explain the defective colonizing ability of this strain (Fig. 3a), as the expression of the virulence genes and the production of cholera toxin analyzed above was significantly lower in ΔfruI compared to WT in M9 medium supplemented with 20 mM fructose and in the small intestine of mice (Fig. 3g, h). qRT-PCR assays and western blotting assays also showed that the expression of virulence genes or production of cholera toxin in ΔfruT and ΔfruI exhibited no significant difference compared to that of WT in M9 medium supplemented with 20 mM glucose (Supplementary Fig. 2a, b), indicating that FruT and FruI influence expression of virulence factors via modulating bacterial fructose uptake.

The findings above suggested that the fructose transporter FruT was not required for V. cholerae growth in the intestine; in fact, ΔfruT showed a colonization advantage compared to the WT strain, raising the question of whether V. cholerae actually uses fructose via PTS when growing in the small intestine. To this end, we performed competitive infection assays in mice between WT and Δfpr and between WT and ΔfruK; fruK encodes fructose-1-phosphate kinase, which converts fructose-1-phosphate to fructose-1,6-bisphosphate in one of the key steps of the fructose catabolism34. The results showed that the CI values of Δfpr and ΔfruK versus WT were 0.39 and 0.28, respectively (Fig. 3j), in the small intestine of mice. These results indicate that Δfpr and ΔfruK are significantly defective in the ability to colonize the small intestine of mice and that fructose catabolism and fructose transport via the PTS are essential for intestinal colonization by V. cholerae.

The above findings suggested that V. cholerae utilizes PTS but not FruT to uptake fructose in the small intestine of mice, and the expression of fruT is downregulated by FruI in the small intestine of mice. Then, we investigated whether the repression of fruT leads to the enhanced fructose PTS expression and activity in vitro and in vivo. qRT-PCR and western blotting assays showed that fpr expression and Fpr protein level exhibited a significant increase in ΔfruT but exhibited a significant decrease in ΔfruI compared to WT in M9 medium supplemented with 20 mM fructose (Fig. 3k, l), while fpr expression and Fpr protein level in ΔfruT or ΔfruI exhibited no difference compared to that of WT in M9 medium supplemented with 20 mM glucose (Supplementary Fig. 2c, d). The decrease of phosphorylation level of EIIA is directly correlated with the increase of the sugar uptake mediated by PTS35. Western blotting showed that the phosphorylation level of Fpr, which consists of a EIIA-like domain, in ΔfruT and ΔfruI growing in M9 medium supplemented with 20 mM fructose was significantly decreased or increased compared to that of WT (Fig. 3m), while Fpr in ΔfruT, ΔfruI, and WT exhibited no phosphorylation in M9 medium supplemented with 20 mM glucose (Supplementary Fig. 2e). In addition, qRT-PCR assays showed that the expression of fpr exhibited a significant increase and a significant decrease in ΔfruT and ΔfruI, respectively, compared to WT in the small intestine of mice (Fig. 3n). These data indicate that deletion of fruI, resulting in fruT overexpression, inhibits the expression and utilization of the fructose PTS, while deletion of fruT promotes the expression and utilization of the fructose PTS.

Since CRP that responds to intracellular cAMP level is a key regulator of metabolic and virulence gene expression in bacteria in response to environmental signals, and that sugar transport via PTS is known to lower the activity of the adenylate cyclase responsible for the synthesis of cAMP18,36, we investigated whether fruI and fruT influence the intracellular cAMP level. We found that when growing in M9 medium supplemented with 20 mM fructose, which mimics the concentration of fructose in the small intestine environment, ΔfruI exhibited significantly higher intracellular cAMP level than WT (Fig. 3o), while ΔfruT exhibited significantly lower intracellular cAMP level than WT (Fig. 3o). In contrast, the intracellular cAMP level of ΔfruI and ΔfruT exhibited no difference compared to that of WT in M9 medium supplemented with 20 mM glucose (Supplementary Fig. 2f). Consistent with cAMP-CRP negatively regulating the expression of virulence genes in V. cholerae36, the above finding showed that the expressions of toxR, tcpP, and toxT were upregulated or downregulated in ΔfruT and ΔfruI, respectively, compared to WT, both in the small intestine of mice and in M9 medium supplemented with 20 mM fructose (Fig. 3g, h). We also showed that the upregulation of virulence gene expression in ∆fruT was inhibited by addition of 1 mM cAMP to M9 medium supplemented with 20 mM fructose (Fig. 3p). These results suggest that deletion of fruI and fruT, which inhibits or promotes the expression and utilization of the fructose PTS, increases or decreases the intracellular cAMP level, which in turn downregulates or upregulates the expression of virulence genes.

To confirm that FruI and FruT influence the virulence of V. cholerae by regulating the intracellular cAMP level in the small intestine of mice, we performed competitive infection assays in mice between a mutant in the adenylate cyclase gene cyaAcyaA) and the double mutants ΔfruIΔcyaA and ΔfruTΔcyaA, respectively. We found that ΔfruIΔcyaA and ΔfruTΔcyaA competed evenly with ΔcyaA in the small intestine of mice (CI = 1.25 and 1.09, respectively) (Supplementary Fig. 2g, h), suggesting that in the absence of the adenylate cyclase, deletion of fruI or fruT does not have an effect on V. cholerae colonizing ability. Consistent with this result, no differences in the expression of virulence genes were observed among ΔcyaA, ΔfruIΔcyaA, and ΔfruTΔcyaA in either the small intestine of mice or M9 medium supplemented with 20 mM fructose (Supplementary Fig. 2i–l). Taken together, these data suggest that FruI regulates the virulence of V. cholerae in the small intestine by downregulating fruT expression, which in turn affects the intracellular cAMP level.

fruI expression is induced by a low O2 signal in the small intestine via ArcA

The majority of LacI-type regulators bind to the promoters of carbohydrate metabolic/transport genes in the absence of the specific carbohydrate substrate and repress their expression; this repression is released when the corresponding carbohydrate is available and binds to the carbohydrate-binding domain in the LacI-type regulator23. As FruI is a LacI-type regulator, it is expected that the FruI-mediated repression of fruT expression would be released in the small intestine, where fructose is present. However, the above finding showed that fruT expression was significantly upregualted in ΔfruI compared to that of WT in the small intestine of mice (Fig. 3f). To confirm that FruI-mediated repression of fruT expression cannot be released by fructose, we also performed EMSAs, and the result showed that the binding of FruI to the fruT promoter was not inhibited by fructose and its analogs (Fig. 4a). These data indicate that the FruI-mediated repression of fruT expression is not released by fructose and fructose analogs, which would result in FruI-mediated repression of fruT expression in the small intestine, where fructose is present.

Fig. 4: fruI expression is induced by a low O2 signal in the small intestine via ArcA.
figure 4

a The influence of fructose, fructose-1-phosphate (F-1-P) and fructose-6-phosphate (F-6-P) on the binding of purified FruI protein to the promoter region of fruT. EMSAs of 0.2 μM purified FruI protein complexed with 30 ng of the DNA fragment in the presence of different concentrations of putative effectors (0–10 mM) were conducted. b EMSA of the binding of purified ArcA protein to the promoter region of fruI and kana (negative control). c qRT-PCR expression level of fruI in WT in the small intestine of mice and M9 medium containing 20 mM fructose as the only carbon source. Data represent the mean ± SD (n = 3). d ArcA binds to a motif in the fruI promoter region. The protected region shows a significantly reduced peak intensities (blue) pattern than seen in compared with those of the control (red). The identified FruI-binding motif is shown in a box at the bottom of the figure. e qRT-PCR expression level of fruI in WT, ΔarcA and ΔarcA+ in M9 medium containing 20 mM fructose as the only carbon source. Data represent the mean ± SD (n = 3). f qRT-PCR expression level of fruI in WT and ΔarcA in M9 medium containing 20 mM fructose as the only carbon source under anaerobic conditions or aerobic conditions. Data represent the mean ± SD (n = 3). g qRT-PCR expression level of fruI and fruT in WT, ΔarcA, and ΔarcA+ in mouse small intestine. Data represent the mean ± SD (n = 3). h Competition assay comparing the colonizing ability of WT, ΔarcA, and ΔarcAΔfruT in infant mouse intestine (n = 7 mice per group). CI is defined as the output ratio of mutant strains to WT lacZ- divided by the input ratio of mutant strains to WT lacZ-. Each symbol represents the CI in an individual mouse; horizontal bars indicate the median. Images are representative of three independent experiments (a, b). Significance is determined by two-tailed unpaired Student’s t test (c, eg) or two-sided Mann–Whitney U test (h) and indicated as the P value; n.s. means no significant difference. Source data are included in Source Data file.

Since fruT regulation by FruI in the small intestine did not appear to depend on the presence or absence of fructose, we investigated whether it depended on regulation of the expression of fruI. To this end, we investigated whether fruI expression is induced in V. cholerae in the small intestine compared to bacteria incubated in M9 medium supplemented with 20 mM fructose, which mimics the concentration of fructose in the small intestine. qRT-PCR assays showed that fruI expression in the WT strain was significantly upregulated in the small intestine of mice compared to that in M9 medium supplemented with 20 mM fructose (Fig. 4c).

To investigate the mechanism that regulates fruI expression in the small intestine, we performed DNA pulldown assays to identify potential regulators that bind to the promoter region of fruI. For two regulator proteins (ArcA and Fur) identified, only ArcA specifically binds to the fruI promoter, as revealed by EMSAs (Fig. 4b and Supplementary Fig. 4). A ArcA-bound sequence containing an 26–base pair motif (5-TATGTGACGTTGTGTTAATTGAATTG-3; −67 to −42 from the fruI translational start site) in the fruI promoter region was revealed by a dye-based DNase I footprinting assay (Fig. 4d), which is similar to the identified ArcA binding consensus sequence37. To study whether ArcA actually regulates fruI expression, we measured fruI expression by qRT-PCR in a ΔarcA strain and the WT strain grown in M9 supplemented with 20 mM fructose. We found that fruI expression was significantly downregulated in the ΔarcA strain and this effect was restored after complementation of ΔarcA with arcA (Fig. 4e). This indicates that the expression of fruI is under direct positive regulation by ArcA.

ArcA and ArcB constitute a two-component regulation system that can sense changes in O2 availability and regulate the expression of downstream genes38. Therefore, we studied whether this two-component regulation system regulates fruI expression in response to the low O2 concentration in the small intestine38. qRT-PCR assay showed that fruI expression in the WT strain significantly increased under low O2 conditions compared to high O2 conditions in vitro (Fig. 4f). This effect was not observed in the ΔarcA strain (Fig. 4f), indicating that low O2 activates fruI expression via ArcA. Furthermore, qRT-PCR assays showed that the expression of fruI and fruT was significantly decreased and increased, respectively, in ΔarcA compared to WT in the small intestine of mice (Fig. 4g). Complementation of ΔarcA with arcA restored the expression level of fruI and fruT to the WT level in the small intestine of mice (Fig. 4g). Competitive infection assays in mice showed that compared to the WT, ΔarcA exhibited a significant decreased ability to colonize the small intestine of mice (Fig. 4h). These results indicate that V. cholerae induces fruI expression in the small intestine in response to the low-O2 conditions via ArcB/ArcA, which promotes intestinal colonization of V. cholerae.

ArcA is a global transcription regulator in the Enterobacteriaceae, which controls the expression of genes involved in several different pathways39,40. For instance, ArcA directly regulates the expression of vpsT involved in biofilm formation of V. cholerae41, which contributes to bacterial intestinal colonization42,43. It is highly likely that the influence of ArcA on the virulence of V. cholerae might only be partially dependent of its regulation on fruI and fruT expression. Competitive infection assays showed that the CI value of the double mutant ΔarcAΔfruT versus WT was 0.41 in the small intestine of mice (Fig. 4i), which was significantly higher than the CI value (0.16) of ΔarcA versus WT (Fig. 4i), and significantly lower than the CI value (3.18) of ΔfruT versus WT (Fig. 3b). These data confirm that in addition to fruI and fruT, ArcA also influences the virulence of V. cholerae via other regulatory pathways.

V. cholerae utilizes FruT to take up fructose at low concentrations in seawater

The above findings suggested that FruT was not used for fructose transport during V. cholerae colonization of the small intestine. H+ symporters always operate at low sugar concentrations suggesting that fructose transport by FruT might be important in the seawater where nutrients44, including carbon sources, are limited. To assess this, we first compared the growth of Δfpr and ΔfruT in M9 medium supplemented with 30 mM, 10 mM, 3 μM, or 0.1 μM fructose under aerobic conditions. The results showed that ΔfruT cannot grow in medium supplemented with 0.1 μM fructose, which is different from Δfpr and WT (Fig. 5a). It exhibited a growth disadvantage compared to Δfpr in medium supplemented with 3 μM fructose, and reached a lower final cell counts than Δfpr and WT (Fig. 5b). In contrast, Δfpr exhibited a growth disadvantage in exponential phase compared to ΔfruT in medium supplemented with 30 or 10 mM fructose, and it reached almost the same final cell density as that of ΔfruT in these two conditions when fructose is almost exhausted (Fig. 5c, d and Supplementary Fig. 5a, b). We also constructed a fruT-overexpressing strain (fruT++) by introducing a plasmid containing fruT into WT, and found that it exhibited a growth advantage compared to WT in medium supplemented with 0.1 or 3 μM fructose, and reached a higher final cell counts than WT (Fig. 5a, b). In contrast, fruT++ exhibited almost same growth ability as WT in M9 medium supplemented with 30 or 10 mM fructose (Fig. 5c, d). Furthermore, using fructose uptake assays we found that WT, fruT++, and Δfpr exhibited higher fructose uptake efficiency than ΔfruT in M9 medium with 0.1 μM fructose (Fig. 5e). In addition, we showed that there was no significant difference in the growth among WT, ΔfruT, and fruT++ in M9 medium supplemented with 30 mM, 10 mM, 3 μM or 0.1 μM glucose, trehalose, maltose, or ribose (Supplementary Fig. 3a–p). These data indicate that FruT contributes to the uptake of fructose, but not other sugars, at low concentrations.

Fig. 5: FruT is responsible for the bacterial uptake of fructose at low concentrations in seawater.
figure 5

ad Growth curves of WT, ΔfruT, Δfpr, and fruT++ in M9 medium containing 0.1 μM (a), 3 μM (b), 10 mM (c), 30 mM (d) fructose as the only carbon source under aerobic conditions. Cell growth was followed by plating and determining the cell counts on LB agar plates (a, b) or measuring the absorbance at 600 nm (c, d). Data represent the mean ± SD (n = 3). e Fructose consumption by WT, ΔfruT, Δfpr, and fruT++ grown in M9 medium containing 0.1 μM fructose as the only carbon source. Data represent the mean ± SD (n = 3). Significance is determined by two-way ANOVA and indicated as the P value; n.s. means no significant difference. Source data are included in Source Data file.

We then investigated whether FruT contributes to the fructose uptake of V. cholerae in seawater. To this end we used samples of seawater from Bohai Sea, which presented fructose concentrations of 0.085 ± 0.017 μM (Supplementary Fig. 5c), which is within the ranges previously reported26. As natural surface seawater is an aerobic environment45, fructose uptake assays were performed under aerobic conditions. A total of 54.3%, 54.1%, and 70.5% of the fructose in 500 ml of seawater was utilized by WT, Δfpr, and fruT++ in 12 h, respectively, while only 21.8% of the fructose was utilized by ΔfruT in 12 h (Supplementary Fig. 5d). These data indicate that V. cholerae utilizes mainly FruT to take up fructose at low concentrations in seawater.

FruT provides a survival advantage to pandemic strains growing in seawater under aerobic conditions

Since expression of fruI, which represses fruT expression, is induced by low O2, we hypothesized that in an aerobic surface seawater environment, the repression of fruT by FruI is released and that fructose uptake via FruT may benefit the survival of V. cholerae in seawater. Indeed, qRT-PCR assays showed that the expression of fruT and fruI in the WT strain growing aerobically was significantly upregulated and downregulated, respectively, compared to that in the small intestine of mice (Fig. 6a). To test whether fruT induction in seawater conferred a survival advantage to V. cholerae, we compared the survival ability of WT and ΔfruT in seawater under aerobic conditions using competition assays. To mimic the conditions in wide extension of seawater, where the fructose concentration remains constant despite fructose utilization by low abundant microorganisms, WT and ΔfruT were located within a membrane-enclosed environment in 500 ml of seawater that was replaced by fresh seawater every 6 h. The CI values of ΔfruT versus WT (from 0.211 to 0.009) continued decreasing during a span of 10 days (Fig. 6b). The results indicated that relative to WT, ΔfruT exhibited a substantial survival disadvantage when growing aerobically in seawater. The similar result was obtained when competition assays was performed in artificial seawater medium that was supplemented with 0.1 μM fructose as the only carbon source every 6 h under aerobic conditions (Fig. 6c). However, WT almost competed evenly with ΔfruT in artificial seawater medium supplemented with 0.1 μM glucose as the only carbon source every 6 h under aerobic conditions (Fig. 6d), confirming that FruT confers pandemic V. cholerae a survival advantage in seawater through enabling bacteria to efficiently uptake low-concentration fructose under aerobic conditions. In contrast, ΔfruT exhibited the same survival ability as WT when growing anaerobically in artificial seawater medium supplemented with 0.1 μM fructose as the only carbon source every 6 h (Fig. 6e).

Fig. 6: V. cholerae turns on FruT in response to O2 for growth in seawater.
figure 6

a qRT-PCR expression level of fruI and fruT in the mouse small intestine and seawater under aerobic conditions. Data represent the mean ± SD (n = 3). b Competition assay comparing the survival ability of WT and ΔfruT within a membrane-enclosed environment in seawater that was replaced by fresh seawater every 6 h under aerobic conditions. c, d Competition assay comparing the survival ability of WT and ΔfruT located within a membrane-enclosed environment in artificial seawater medium that was supplemented with 0.1 μM fructose (c) or 0.1 μM glucose (d) as the only carbon source every 6 h under aerobic conditions. e Competition assay comparing the survival ability of WT and ΔfruT located within a membrane-enclosed environment in artificial seawater medium that was supplemented with 0.1 μM fructose under anaerobic conditions. CI was determined as the output ratio of ΔfruT to WT lacZ- cells divided by the input ratio of ΔfruT to WT lacZ- cells. Data represent the mean ± SD (n = 3). Significance is determined by two-tailed unpaired Student’s t test (a) and indicated as the P value; n.s. means no significant difference. Source data are included in Source Data file.

Overall, these results indicate that in aerobic surface seawater environment, the repression of fruT by FruI is released and that FruT-mediated fructose uptake plays an important role in the survival of V. cholerae in aerobic seawater environments.

Discussion

Pandemic V. cholerae lives in aquatic environment and causes diseases in human small intestine. Here, we elucidated the molecular mechanisms that enable pandemic V. cholerae strains to use a fructose/H+ symporter FruT to uptake low-concentration fructose in seawater, conferring them a survival advantage in that nutrient-limited environment. Pandemic V. cholerae turns off fruT via FruI and uptake high-concentration fructose via PTS in the mouse small intestine, resulting in the decrease of intracellular cAMP level and upregulation of virulence gene expression, and maintaining high pathogenicity in vivo.

V. cholerae spends much of its life cycle outside of the human host in aquatic environment, such as seawater, and must acquire nutrients from its environment6. The bioavailability of carbon sources is limited in seawater7. In this study, we showed that acquiring the fructose/H+ symporter FruT enables pandemic V. cholerae strains to efficiently take up fructose at low concentrations in seawater. There are other sugars present in seawater; however, their concentrations (0.01–0.24 μM) are much lower than that of fructose26, except for glucose, whose concentration (0.03–0.67 μM) is similar to that of fructose26. Future research is needed to determine, in addition to glucose PTS, whether pandemic V. cholerae strains have a transporter specific for uptaking low-concentration glucose.

LacI-type regulators are widely used in bacteria for control of carbon source utilization in response to different environments. Through comparative genomics analysis, we showed that, among 11 genes encoding LacI-type regulators in V. cholerae, fruI is significantly associated with pandemic strains. fruT, whose expression is under negative regulation by fruI, is located 1.3 kb upstream of fruI. Comparative genomics analysis showed that the distribution of fruT in pandemic and nonpandemic V. cholerae is identical to that of fruI, indicating fruI and fruT were introduced into V. cholerae genome via a single lateral gene transfer event. As pandemic V. cholerae needs to repress fruT expression using FruI in the small intestine of mice, which benefits the virulence gene expression and bacterial colonizing ability in vivo, simultaneous acquisition of fruI and fruT enabled pandemic V. cholerae strains to obtain a survival advantage in seawater while maintaining high pathogenicity in animal intestine.

Both fruT and fruI are located within a predicted genomic island (GI-7). GIs are clusters of genes that are mobile via mechanisms of lateral gene transfer46. Thus, it is highly likely that fruI and fruT were introduced into V. cholerae genome via a single lateral gene transfer event. Over 400 potential GIs have been predicted in different pandemic or nonpandemic V. cholerae strains47. However, the role of most of these GIs in the evolution of V. cholerae remains unclear, with slight exceptions that are related to virulence or antibiotic resistance, such as VPI that encodes toxin-coregulated pilus, VSP-I that includes a gene encoding a dinucleotide cyclase required for bacterial intestinal colonization, and SXT that carries antibiotic resistance genes48,49. In this study, we provided evidence that fruI and fruT in GI-7 give V. cholerae advantages for survival in seawater and maintain bacterial pathogenicity in animal intestine. In addition to fruI and fruT, there are other four genes in GI-7, whose function are unknown, and whether they are related to the virulence or environmental survival of V. cholerae needs to be investigated in future studies.

cAMP is a universal second messenger that is used by diverse forms of life, and it has central roles in regulating gene expression related with many biological process, such as carbon metabolism, virulence, and biofilm formation, in bacteria36. In V. cholerae, cAMP-CRP represses the expression of several virulence genes, such as ctxAB encoding the cholera toxin and tcpA encoding the repeating subunit of toxin-coregulated pilus, via influencing the expression of genes encoding regulator TcpP and ToxR19,20,21. cAMP-CRP also represses the biofilm formation of V. cholerae by regulating the expression of a set of genes, such as vpsR encoding a regulator and cdgA encoding a cyclic diguanylate cyclase that promotes the production of cyclic di-GMP50. cAMP-Crp was also reported to contribute to the resistance of V. cholerae to environmental bacteriophages51. PTS is a major factor involved in the regulation of intracellular cAMP level in bacteria52, and uptake of sugar via PTS results in the decrease of bacterial intracellular cAMP level18,36. PTS components play multiple regulatory roles in enteric bacteria53. In V. cholerae, EIIA of glucose PTS influences biofilm formation via modulating the activity of cyclic di-GMP phosphodiesterase PdeS and regulating the decay of CsrB/C sRNA54,55. In this study, we showed that in response to the low O2 signal in animal small intestine, V. cholerae turns off fruT and uptakes fructose via PTS, leading to the decrease of intracellular cAMP level and the enhanced expression of virulence genes. It is likely that in addition to inducing virulence gene expression found in this study, the elevated fructose PTS activity of V. cholerae in animal small intestine also influences bacterial pathogenicity via other mechanisms, such as regulating the formation of biofilm.

Compensatory relationships have been observed between glucose transporters in mammalian cells; however, the underlying mechanisms are unclear56. There is a previously known fructose PTS in V. cholerae17. In this work, we identified a fructose/H+ symporter FruT in V. cholerae, and showed that the repression of fruT expression by FruI enhances the expression of fructose PTS. Furthermore, qRT-PCR assays showed that the deletion of fpr encoding a component of fructose PTS led to the elevated expression of fruT, but not fruI, in M9 medium supplemented with 20 mM fructose, while the expression of both fruT and fruI in Δfpr exhibited no significant change compared to that of WT in M9 medium supplemented with 20 mM glucose (Supplementary Fig. 6a, b), indicating fructose PTS also influences the expression of fruT via modulating fructose uptake. It is highly likely that there is a compensatory relationship between FruT and fructose PTS. The repression of either of these two fructose transports leads to the increased expression of the other transporter. The exact compensatory mechanism between FruT and fructose PTS is currently unclear; however, our data showed that FruI does not play a role in that mechanism. We hypothesize that the reciprocal regulation between fructose PTS and FruT maybe mediated via the change of bacterial intracellular fructose concentration, which will be the subject of future studies.

Methods

Bacterial strains, plasmids, and growth conditions

V. cholerae O1 El Tor strain El2382 isolated in 1994 was provided by the Shanghai Municipal Centers for Disease Control & Prevention. Escherichia coli DH5a/λpir strain was used for DNA manipulation, and E. coli S17/λpir strain was used for conjugation with V. cholerae. Bacterial strains were grown in Luria-Bertani (LB) broth, M9 medium, or artificial seawater medium (Sigma; S9883) with different concentrations of glucose, fructose, trehalose, maltose, or ribose and desterilized seawater from the Bohai Sea, respectively. Bacteria were inoculated into fresh medium and grown at 37 °C with shaking (for aerobic conditions) or without shaking (for anaerobic conditions). Antibiotics (Sigma) were used at the following concentrations, unless otherwise indicated: polymyxin B, 40 μg/ml; ampicillin, 50 μg/ml; chloramphenicol, 30 μg/ml (for E. coli) or 5 μg/ml (for V. cholerae).

Mutant construction, complementation, and overexpression

All primers used in this study can be found in Supplementary Table 2. For mutant construction, the DNA fragments were cloned into the suicide vector pRE11257. Plasmids were maintained in E. coli S17-1/λpir for conjugation to V. cholerae. Integration was selected on LB plates containing polymyxin B (40 μg/ml) and 20% sucrose for sacB counterselection. PCR and DNA sequencing were used to confirm the presence of the induced mutations. For complementation, genes with native promoter were amplified by PCR and cloned into vector pBAD3358. The complementary plasmids were introduced into V. cholerae via electroporation. For overexpression of fruT, fruT was amplified by PCR and cloned into vector pTrc99a59. The plasmid was introduced into V. cholerae via electroporation.

RNA isolation and RNA sequencing

V. cholerae were harvested from small intestinal tissue or after culturing at 37 °C for LB broth or M9 medium or at 22 °C for seawater samples. Total RNA was isolated using TRIzol Reagent (Invitrogen; 15596026) according to the manufacturer’s protocol. RNA was quantified and qualified using an Agilent 2100 Bioanalyzer (Agilent Technologies), a NanoDrop (Thermo Fisher Scientific Inc), and 1% agarose gel electrophoresis. One microgram of total RNA with an RIN value >6.5 was used for library preparation. rRNA (including 16S and 23S rRNA) was depleted from total RNA using Ribo-off rRNA Depletion Kit (Bacteria) (Vazyme; N407-01). Libraries were constructed and analyzed by NOVOGENE, Inc (TianJin, China). Differentially expressed genes in ΔfruI compared to WT were identified using edgeR (|fold change| > 2 and P < 0.05). P values were adjusted using the Benjamini and Hochberg approach to control the false discovery rate.

qRT-PCR

qRT-PCR analysis was conducted on Applied Biosystems ABI 7500 (Applied Biosystems) with SYBR green fluorescence dye. cDNA templates were denatured at 95 °C for 10 min, followed by 45 cycles of 95 °C (30 s) and 60 °C (60 s). rrsA gene was used as a reference control for sample normalization60, and the relative expression level was calculated as fold change values using the 2−ΔΔCT method61. Each experiment was carried out in triplicate.

Growth curve

To determine the growth curve of each strain in 10 or 30 mM sugars, overnight cultures were washed with PBS three times and diluted to 106/ml in a flask containing 20 ml of M9 medium with 10 or 30 mM fructose, glucose, trehalose, maltose, or ribose. When necessary, IPTG was added to a final concentration of 0.1 mM. A 200 μl aliquot was added to a 96-well microplate and incubated at 37 °C with shaking at 180 rpm for 24 h. The absorbance at 600 nm was recorded. Experiments were independently performed three times.

To determine the growth curve of each strain in 0.1 or 3 μM sugars, overnight cultures were washed by PBS for three times and diluted to 106/ml in a flask containing 200 ml of M9 medium containing 0.1 or 3 μM fructose, glucose, trehalose, maltose, or ribose and incubated at 37 °C with shaking at 180 rpm. A 100 μl aliquot was removed from the flask and suitable dilutions were plated on LB agar plates. The growth rate was determined by cell counts every 3 h. Experiments were independently performed three times.

EMSA

The 6×His-tagged FruI, ArcA, and Fur proteins were expressed and purified in E. coli BL21 (DE3). DNA target fragments were amplified by PCR and purified using a SPARKeasy Gel DNA Extraction Kit (Sparkjade; AE0101). Purified PCR fragments (40 ng) were incubated at 25 °C for 30 min with 6×His-tagged FruI or ArcA protein at concentrations ranging from 0 to 2 µM in 20 μl reactions containing Binding buffer (10 mM Tris-HCl [pH 7.5], 0.2 mM dithiothreitol, 5 mM MgCl2, 10 mM KCl, and 10% glycerol). When necessary, 30 mM acetyl phosphate was added to the binding buffer before incubation. The protein-DNA fragments were electrophoretically separated on a native polyacrylamide gel at 4 °C and 80 V/cm. The gel was stained for 10 min in a solution of 0.1% GelRed (Biotium; 41003), and protein bands were visualized by ultraviolet transillumination. During effector screening studies, the purified 6×His-tagged FruI protein was incubated at 25 °C for 10 min with 0–10 mM fructose, fructose-1-phosphate, or fructose-6-phosphate, before being added to binding buffer.

Dye primer-based DNase I footprinting assay

About 200-bp fragment of the fruI or fruT promoter regions was generated by PCR with 6-FAM primers. Forty nanograms of 6-FAM-labeled fruI or fruT promoter was incubated with different amounts of ArcA or FruI protein ranging from 0 to 1 μM in a binding buffer (10 mM Tris-HCl [pH 7.5], 0.2 mM dithiothreitol, 5 mM MgCl2, 10 mM KCl, and 10% glycerol). 0.05U DNase I (Thermo; EN0521) was added to a 20-μl reaction for 5 min at room temperature. The reaction was stopped by heating at 65 °C for 10 min in the presence of 250 mM ethylenediaminetetraacetic acid (EDTA). DNA fragments were purified with the QIAquick PCR Purification kit (Qiagen; 28104) and eluted in 15 μl distilled water. Samples were analyzed by MAP Biotech CO., Ltd (Shanghai China). The results were analyzed with Peak Scanner (v1.0).

Intestinal colonization assay

Both sexes of CD-1 infant mice (5 days old) were purchased from the Beijing Vital River Laboratory Animal Technology (Beijing, China). Mice were housed under specific pathogen-free conditions with a 12 h light/dark cycle, at a temperature of 30 °C incubator, and a relative humidity of 50 ± 5%. All animal studies were conducted according to protocols approved by the Institutional Animal Care Committee of Nankai University (Tianjin, China) and performed under protocol no. IACUC 2016030502. Briefly, the V. cholerae lacZ+ strains (WT and the mutants) and lacZ− strains (ΔlacZ) were grown overnight at 37 °C in LB broth. Approximately 105 lacZ+ strains were mixed with an equal amount of WT lacZ− strain and the mixtures were intragastrically administered to groups of six anesthetized mice. After 24 h of incubation, the small intestine was removed, homogenized, and plated onto LB agar plates containing 5-bromo-4-chloro-3-indoyl-β-d-galactopyranoside (X-gal) or chloramphenicol (5 μg/ml) to enumerate the recovered bacteria and to obtain the output ratios. The CI was determined as the output ratio of lacZ+ to lacZ− cells divided by the input ratio of lacZ+ to lacZ− cells. The CI of complemented strains was determined as the output ratio of lacZ+(pBAD33) to lacZ−(pBAD33) cells divided by the input ratio of lacZ+(pBAD33) to lacZ−(pBAD33) cells.

H+ symport assay

The detection of H+ movements associated with initial sugar uptake was assessed by adding sugar pulses to unbuffered cell suspensions62. Bacteria (OD600 of 1.0) were harvested by centrifugation (5000 g, 5 min, 4 °C), washed with ice-cold PBS, resuspended in 0.9% NaCl to a final concentration of 106/ml and kept in 37 °C for 3 h for starvation. After starvation, 500 µl of WT, mutants, complement strains, and 0.9% NaCl (negative control) were seeded into a 24-well plate (Seahorse XF24 Cell Culture Microplates, Agilent). After acquiring baseline measurements, the following drug was used for an extracellular pH test: 30 mM fructose (port A). Cells were assessed for extracellular alkalinization rate using the Seahorse XF analyzer (Agilent)63.

Determination of the Fpr phosphorylation state

A total of 0.2 ml of bacterial culture (OD600 of 1.0) was treated with the addition of 20 μl of 10 M NaOH followed by 1 ml of ethanol and 180 μl of 3 M sodium acetate, pH 5.2. After chilling at –80 °C for 2 h, precipitates were collected by centrifugation, rinsed with 70% ethanol, and suspended in 100 μl of sample buffer (160 mM Tris-HCl, pH 7.5, 4% SDS, 20% glycerol, and 10% 2-mercaptoethanol). To achieve a good separation of the two forms of Fpr, samples were fractionated on SDS-PAGE gels containing 50 μM of Phos-tag (WAKO; 304-93521) and 40 µM MnCl2. Subsequently, gels were washed with western transfer buffer (25 mM Tris, 192 mM glycine, 20% methanol, and 0.1% SDS) containing 5 mM EDTA for 10 min, followed by a second wash with transfer buffer for 20 min.

Western blotting

Bacterial cells were harvested and washed three times in ice-cold PBS, re-suspend by ice-cold PBS, and sonicated with 15 cycles of 30 s on/off at 95% power. Then the cell debris was removed by centrifuging at 12,000 g for 10 min at 4 °C. The supernatants were collected and kept on ice before use. Total cellular proteins (20 ug) were separated by SDS-PAGE and transferred to 0.2 μm polyvinylidene difluoride membranes (Bio-Rad) by electroblotting. Blots for Fpr, RNAP (RNA polymerase), and cholera toxin protein used the anti-Fpr monoclonal antibody (Willget Biotech Co., Ltd.), anti-RNA polymerase beta antibody(ab191598), and anti-cholera toxin antibody (ab123129) at 1:2000 dilution, respectively. Western blotting was detected using horseradish peroxidase-linked goat anti-rabbit IgG secondary antibodies (1:5000 dilution, Sparkjade; EF0002) and Sparkjade ECL plus (Sparkjade; ED0016) as recommended by the manufacturer. Images were acquired using an Amersham™ Imager 600 System (General Electric Company). Western blotting bands were quantified using Image J (v1.8.0) software.

DNA pulldown assay

The fruI promoter regions were amplified by 5’biotin tagged primers. The negative control DNA fragments were amplified in the fruI coding region. The biotin-labeled DNA fragments were purified using a SPARKeasy Gel DNA Extraction Kit and diluted to a final concentration of 30 ng/μl.

Two hundred microliter of Streptavidin-agarose beads (Thermo Fisher) was washed three times in 2×B/W buffer (10 mM Tris-HCl pH 7.5, 1 mM EDTA, 2 M NaCl) and incubated for 1 h with an equal volume of biotin-labeled DNA fragments at room temperature. DNA-treated beads were washed three times in BS-THES buffer (5×BS buffer: 50 mM HEPES, 25 mM CaCl2, 250 mM KCl, 60% glycerol, 1.5% protease inhibitor, 0.2% phosphatase inhibitor; 2.25×THES: 50 mM Tris-HCl pH 7.5, 10 mM EDTA, 20% sucrose, 140 mM NaCl).

The overnight cultures were diluted 1:100 in a flask containing 100 ml fresh LB broth and grown to log phase (OD600 of 0.6–0.8). Bacteria were harvested and washed three times in ice-cold PBS and re-suspend by BS-THES buffer. Cells were sonicated with 15 cycles of 30 s on/off at 95% power. Then the cell debris was removed by centrifuging at 12,000 g for 10 min at 4 °C. The supernatants were collected and kept on ice before use. DNA-treated Beads were incubated in 1 ml supernatants twice for 30 min at room temperature. The beads were then washed three times with BS/THES. Protein-DNA complexes were eluted by Elution buffer (25 mM Tris-HCl pH 7.5, 300 mM NaCl) and further treated with 20% (v/v) trichloroacetic acid in ice for 3 h to precipitate protein. The precipitates were collected by centrifugation at 12,000 g for 10 min at 4 °C. The pellets protein were segregated by SDS-PAGE and the specific bands were collected and analyzed by liquid chromatography-tandem mass spectrometry in Beijing Genomics Institution (Beijing, China).

Intracellular cAMP assay

To determine the intracellular cAMP level of each strain, WT and mutants were harvested by centrifugation (5000 g, 5 min, 4 °C), weighed, washed with ice-cold PBS, resuspended in 100 mM HCl, and ultrasonicated with 15 cycles of 30 s on/off at 95%. Cell debris was removed, and the resulting supernatant was used to determine the cAMP level. Intracellular cAMP level was determined using a competitive cAMP ELISA kit (abcam133051) following the manufacturer’s instructions.

Fructose uptake assay

Overnight cultures were diluted to 106/ml of M9 broth with 0.1 μM fructose or seawater without antibiotics and incubated at 37 or 22 °C with shaking at 180 rpm. When necessary, IPTG was added to a final concentration of 0.1 mM. A 100 μl aliquot was removed from the flask every 3 h, and the concentration of fructose was measured by a PicoProbe Fructose Fluorometric Assay Kit (BioVision; K611) following the manufacturer’s instructions.

Fructose concentration detection in seawater

In total, eight randomly selected water samples from the Bohai Sea were tested. Seawater samples were sterilized by immediately passing through a 0.22-micron bacteria-retentive filter. Then, the samples were frozen before detection. The concentration of fructose was measured by a PicoProbe Fructose Fluorometric Assay Kit.

Survival assay

Collected seawater was prefiltered through a 50-μm quartz filtration system and then filtered through 0.22-μm pore-size membranes (Millipore). The overnight cultures were washed and diluted to 106/ml by sterilized seawater, and approximately 106 ΔfruT was mixed with an equal amount of WT lacZ− strain. The mixtures were located within a membrane-enclosed environment in 500 ml seawater or in artificial seawater medium. The seawater was replaced with fresh seawater and the artificial seawater medium was supplemented with 0.1 μM fructose or 0.1 μM glucose as the only carbon source every 6 h. The sample were incubated at 22 °C with shaking (for aerobic conditions) or without shaking (for anaerobic conditions). A 100 μl aliquot was removed from the membrane-enclosed environment every 24 h, and suitable dilutions were plated onto LB agar plates containing X-gal to enumerate the recovered bacteria and to obtain the output ratios. The CI was determined as the output ratio of ΔfruT to WT lacZ− cells.

Statistical analysis

All data are expressed as the means ± standard deviation (SD). Differences between two groups were evaluated using two-tailed unpaired Student’s t test, Mann–Whitney U test, or two-away ANOVA according to the test requirements (as stated in the figure legends). Significance is indicated as the P value and n.s. means no significant difference. Statistical analysis of Mann–Whitney U test was conducted using the software MedCalc (v12.3.0.0). Statistical analysis of two-tailed unpaired Student’s t test was conducted using the Microsoft Excel 2019 (16.0.14026.20294). Statistical analysis of two-way ANOVA was conducted using the software using GraphPad.Prism.7.0.4 (GraphPad Inc., San Diego, CA).

Reporting summary

Further information on research design is available in the Nature Research Reporting Summary linked to this article.