The RavA/VemR two-component system plays vital regulatory roles in the motility and virulence of Xanthomonas campestris

Abstract

Xanthomonas campestris pv. campestris (Xcc) can cause black rot in cruciferous plants worldwide. Two-component systems (TCSs) are key for bacterial adaptation to various environments, including hosts. VemR is a TCS response regulator and crucial for Xcc motility and virulence. Here, we report that RavA is the cognate histidine kinase (HK) of VemR and elucidate the signalling pathway by which VemR regulates Xcc motility and virulence. Genetic analysis showed that VemR is epistatic to RavA. Using bacterial two-hybrid experiments and pull-down and phosphorylation assays, we found that RavA can interact with and phosphorylate VemR, suggesting that RavA is the cognate HK of VemR. In addition, we found that RpoN2 and FleQ are epistatic to VemR in regulating bacterial motility and virulence. In vivo and in vitro experiments demonstrated that VemR interacts with FleQ but not with RpoN2. RavA/VemR regulates the expression of the flagellin-encoding gene fliC by activating the transcription of the rpoN2-vemR-fleQ and flhF-fleN-fliA operons. In summary, our data show that the RavA/VemR TCS regulates FleQ activity and thus influences the expression of motility-related genes, thereby affecting Xcc motility and virulence. The identification of this novel signalling pathway will deepen our understanding of Xcc-plant interactions.

Keywords: RavA/VemR; motility; signalling pathway; two-component system; virulence.

FIGURE 1

RpoN2/FleQ and VemR antagonistically regulate virulence and motility in Xanthomonas campestris pv. campestris (Xcc). (a) Virulence phenotypes of the wild type (WT), rpoN2, vemR and fleQ single, double, and triple deletion mutants, and the corresponding complementation strains 14 days postinoculation (dpi) on broccoli cv. Wenxin leaves. (b) Lengths of the lesions on the 14 dpi‐infected leaves caused by the above Xcc strains, as shown in (a). (c) Swarming zones of the above Xcc strains inoculated on swarming plates (NY medium containing 2% glucose and 0.6% agar) at 28°C for 3 days. (d) Average diameters of the swarming swimming zones. (e) Swarming zones of the abovementioned strains on swimming plates (0.03% Bacto peptone, 0.03% yeast extract, and 0.28% agar incubated at 28°C for 5 days). (f) Average diameters of the swimming zones. The values given are the means ± SD from triplicate experiments. C indicates the indicated gene(s) complementation strain of a given mutant (including the indicated gene(s) deletion). As our focus was the relationship between vemR and fleQ/rpoN2, we only analysed the significance of the difference between ΔvemR and the double mutants ΔvemR/ΔfleQ and ΔvemR/ΔrpoN2. **p < 0.01, ***p < 0.001

FIGURE 2

VemR directly interacts with FleQ. (a) The rpoN2 promoter activities in the wild type (WT), ΔravA, ΔrpoN2, ΔvemR, and ΔfleQ strains were determined by measuring the β‐glucuronidase (GUS) activity of rpoN2p‐gusA. The values given are the means ± SD from triplicate experiments. Significant differences between a given mutant and the WT strain are shown. **p < 0.01, ***p < 0.001. (b) Bacterial two‐hybrid (B2H) assay for analysis of the interactions of FleQ and VemR. All strains were inoculated on LB and M63 supplemented with X‐gal (LX and M63X, respectively) for 24 h at 30°C. Each spot was inoculated with 2 µl of each of a 10‐fold dilution series (i.e., 10⁻¹, 10⁻², 10⁻³, 10⁻⁴, 10⁻⁵, and 10⁻⁶‐fold, from left to right) of cells in logarithmic growth phase (OD at 600 nm = 0.8). D11K, D56A, and D11K/D56A are VemR phosphorylation‐related site (D11 and D56) mutants. The plasmid pairs T18C/T25 and T18CZIP/T25ZIP were used as negative and positive controls, respectively. T18C‐X and T25‐Y indicate the vectors T18C and T25 harbouring the X and Y genes, respectively. (c) β‐Galactosidase activity of each reporter B2H strain. The values given are the means ± SD from triplicate experiments. Significant differences between the given bacteria and the negative control strain (T18C/T25) are shown. **p < 0.01. (d) Pull‐down assays for analysis of the interactions of FleQ and VemR. Cell lysates containing His‐FleQ (55 kDa) and MBP (maltose‐binding protein)‐VemR (58 kDa) or MBP (44 kDa) were incubated overnight at 4°C with amylose resin. After elution, the protein samples were boiled and separated by SDS‐PAGE and immunoblotted with anti‐MBP and anti‐His antibodies

FIGURE 3

The rpoN2‐vemR‐fleQ operon functions downstream of ravA in the regulatory pathway of bacterial virulence and motility. (a) Virulence phenotypes of the wild type (WT), ΔravA, double mutants (ΔrpoN2/ΔravA, ΔvemR/ΔravA, ΔfleQ/ΔravA), triple mutants (ΔrpoN2/ΔvemR/ΔravA, ΔvemR/ΔfleQ/ΔravA, ΔrpoN2/ΔfleQ/ΔravA), quadruple mutant (ΔrpoN2/ΔvemR/ΔfleQ/ΔravA), and corresponding complementation strains 14 days postinoculation (dpi) on broccoli cv. Wenxin leaves. C indicates that the indicated gene (rpoN2, vemR, and fleQ) complements a given mutant. (b) Lengths of the lesions on the 14 dpi‐infected broccoli leaves caused by the above strains, as shown in (a). (c, d) The swarming colony phenotypes (c) and average zone diameters (d) of the above strains, as shown in (a). (e, f) The swimming colony phenotypes (e) and average zone diameters (f) of the above strains, as shown in (a). The values given are the means ± SD from triplicate experiments. As our focus was the relationship of ravA and rpoN2, vemR or fleQ, we only compared the significant difference between ΔravA and the double mutants ΔravA/ΔrpoN2, ΔravA/ΔvemR, and ΔravA/ΔfleQ. **p < 0.01

FIGURE 4

RavA interacts with and phosphorylates VemR. (a) Bacterial two‐hybrid (B2H) assays for analysis of the interaction between RavA and VemR in Escherichia coli. All strains were inoculated on LX and M63X at 30°C for 24 h, as shown in Figure 2. D11K, D56A, and D11K/D56A are VemR phosphorylation‐related site (D11 and D56) mutants, while H164A is a RavA phosphorylation site mutant. T18C‐X and T25‐Y indicate the vectors T18C and T25 harbouring the X and Y genes, respectively. (b) β‐Galactosidase activities of each reporter B2H strain. The plasmid pairs T18C/T25 and T18CZIP/T25ZIP were used as negative and positive controls, respectively. The values given are the means ± SD from triplicate experiments. Significant differences between the given bacteria and the negative control strain (T18C/T25) are shown. **p < 0.01. (c) Analysis of the interactions of RavA and VemR by maltose‐binding protein (MBP) pull‐down assays. The experiments were performed as shown in Figure 2d, but His‐VemR (15 kDa) and MBP‐RavA (89 kDa) were used instead of His‐FleQ and MBP‐VemR, respectively. (d) In vivo and in vitro assays for detecting phosphotransfer from RavA to VemR. In vivo: Xanthomonas campestris pv. campestris (Xcc) strains and E. coli strains expressing VemR‐His were cultured in M4M and LB media, respectively. VemR‐His protein was purified with Ni‐NTA resin, and elution samples were separated on Phos‐tag and standard SDS‐PAGE and then transferred to PVDF membranes for western blotting with anti‐His antibody. In vitro: 50 ng of purified RavA was autophosphorylated for 30 min, then 50 ng of purified His‐VemR protein was added and incubated at 28°C. At the indicated time, SDS (1× final concentration) sample buffer was added to stop the reaction. SDS‐PAGE and western blotting were performed as in vivo assays. −, without the RavA protein

FIGURE 5

RavA regulates fliC expression via RpoN2‐VemR‐FleQ. (a) Plate‐based detection of the effects of rpoN2, vemR, fleQ, and ravA deletion mutations on fliC promoter activity, as determined by measuring β‐galactosidase activity. The XC1214 (Xanthomonas campestris pv. campestris lacZ) gene was integrated downstream of the fliC promoter, and lacZ‐reporter strains were constructed. (b, c) Quantitative analysis of the effects of rpoN2, vemR, fleQ, and ravA single, double, triple, and quadruple mutations on fliC promoter activity. The values given are the means ± SD from triplicate experiments. In rpoN2‐vemR‐fleQ‐related mutants (b), only the vemR single mutation had a significant effect on fliC promoter activity. In ravA‐related mutants (c), ravA mutation significantly increased fliC expression, but fleQ or rpoN2 mutation in the ravA mutant abolished this induction. **p < 0.01, ***p < 0.001

FIGURE 6

Effects of fliA mutation on bacterial virulence and motility. (a) The flhF‐fleN‐fliA operon promoter activity in the wild type (WT), ΔravA, ΔrpoN2, ΔvemR, and ΔfleQ strains. The flhFp‐gusA construct (flhF promoter‐gusA fusion) was transformed into the mentioned strains, and β‐glucuronidase (GUS) activity was detected. (b) Lesion lengths (14 days postinoculation) on broccoli leaves infected by the indicated strains (WT, ΔfliA, ΔravR, ΔravA, ΔvemR, ΔfleQ, ΔfliA/ΔravR, ΔfliA/ΔravA, ΔfliA/ΔvemR, and ΔfliA/ΔfleQ). (c) Swarming zone diameters of the above strains. (d) Swimming zone diameters of the above strains. (e) The fliC promoter activity of the above strains in NYG medium. The values given are the means ± SD from triplicate experiments. Significant differences between a single‐gene mutant and the corresponding double‐gene mutant (ΔravR vs. ΔfliA/ΔravR, ΔravA vs. ΔfliA/ΔravA, ΔvemR vs. ΔfliA/ΔvemR, and ΔfleQ vs. ΔfliA/ΔfleQ) are shown. A significant difference between the fliA mutant and the WT strain (ΔfliA vs. WT) is also shown. **p < 0.01, ***p < 0.001

FIGURE 7

clp is epistatic to ravA/vemR in the regulation of Xanthomonas campestris pv. campestris (Xcc) virulence but not motility. (a) Lesion lengths (14 days postinoculation) of broccoli leaves infected with wild type (WT), Δclp, ΔravR, ΔravA, ΔvemR, ΔfleQ, Δclp/ΔravR, Δclp/ΔravA, Δclp/ΔvemR, and Δclp/ΔfleQ. (b, c) Swarming and swimming zone diameters of the above strains, respectively. (d) The fliC promoter activity was analysed by measuring the β‐galactosidase activity of the above strains. The values given are the means ± SD from triplicate experiments. Significant differences between a single‐gene mutant and the corresponding double‐gene mutant (ΔravR vs. Δclp/ΔravR, ΔravA vs. Δclp/ΔravA, ΔvemR vs. Δclp/ΔvemR, and ΔfleQ vs. Δclp/ΔfleQ) are shown. A significant difference between the clp mutant and the WT strain (Δclp vs. WT) is also shown. *p < 0.05, **p < 0.01, ***p < 0.001

FIGURE 8

Model for the cross‐talk among HpaS‐, RavA‐, and RavS‐mediated signalling pathways in Xanthomonas campestris pv. campestris (Xcc). RavA/VemR regulates flagellin fliC expression via FleQ/RpoN2 and FliA to influence Xcc swimming and regulates FleQ/RpoN2 and Clp activities to affect bacterial swarming. RavA/RavR and RavS/RavR control cellular c‐di‐GMP levels and thus regulate bacterial motility via the c‐di‐GMP receptor Clp. HpaS/VemR regulates bacterial swimming via direct interaction with the flagellum protein FliM. HpaS phosphorylates HrpG and regulates the expression of the type III secretion system (T3SS) and Xcc virulence. positive regulation; negative regulation; solid lines, having experimental evidence; dotted lines, supposed relationship. Components of a given pathway are indicated in the same colour