Cell-cell signal-dependent dynamic interactions between HD-GYP and GGDEF domain proteins mediate virulence in Xanthomonas campestris

Abstract

RpfG is a paradigm for a class of widespread bacterial two-component regulators with a CheY-like receiver domain attached to a histidine-aspartic acid-glycine-tyrosine-proline (HD-GYP) cyclic di-GMP phosphodiesterase domain. In the plant pathogen Xanthomonas campestris pv. campestris (Xcc), a two-component system comprising RpfG and the complex sensor kinase RpfC is implicated in sensing and responding to the diffusible signaling factor (DSF), which is essential for cell-cell signaling. RpfF is involved in synthesizing DSF, and mutations of rpfF, rpfG, or rpfC lead to a coordinate reduction in the synthesis of virulence factors such as extracellular enzymes, biofilm structure, and motility. Using yeast two-hybrid analysis and fluorescence resonance energy transfer experiments in Xcc, we show that the physical interaction of RpfG with two proteins with diguanylate cyclase (GGDEF) domains controls a subset of RpfG-regulated virulence functions. RpfG interactions were abolished by alanine substitutions of the three residues of the conserved GYP motif in the HD-GYP domain. Changing the GYP motif or deletion of the two GGDEF-domain proteins reduced Xcc motility but not the synthesis of extracellular enzymes or biofilm formation. RpfG-GGDEF interactions are dynamic and depend on DSF signaling, being reduced in the rpfF mutant but restored by DSF addition. The results are consistent with a model in which DSF signal transduction controlling motility depends on a highly regulated, dynamic interaction of proteins that influence the localized expression of cyclic di-GMP.

Fig. 1.

The interaction of the HD-GYP domain of RpfG with a specific subset of GGDEF domains requires the GYP motif but not the HD motif. (A) Schematic representation of RpfG indicating the receiver (REC) and HD-GYP domains with the positions of the HD dyad and the GYP motif. (B and C) Y2H analysis of the role of the signature residues in the interactions of the HD-GYP domain with GGDEF domains from specific proteins. Baits were the HD-GYP domain and variants with alanine substitutions. Preys were (1) wild type; (2) the HD-GYP domain; and the GGDEF domains of (3) XC_0249; (4) XC_0420; (5) XC_0613; (6) XC_1803; (7) XC_2866; (8) XC_2793; (9) XC_2949; and (10) XC_2324. Yeast cultures were grown to midlogarithmic phase in selective media, and expression of the GAL4-LacZ reporter gene was assayed by measuring beta-galactosidase activity as Miller units. Values are the means ± SD of triplicate measurements. In addition, the degree of interaction was assessed by the number of amino acid auxotrophs recovered (indicated by values above the bars).

Fig. 2.

Interactions within Xcc between the wild-type and variant HD-GYP domain from RpfG and different full-length GGDEF-domain proteins as measured by FRET. The HD-GYP domain and variants were fused at the C terminus to CFP, whereas the GGDEF-domain proteins were fused at the C terminus to YFP. (A) The HD-GYP domain shows significant interaction with XC_0249 and XC_0420 but not with other GGDEF-domain proteins. (B and C) Interaction of the HD-GYP domain with the full-length XC_0249 protein (B) or XC_0420 protein (C) is affected by alanine substitution in the GYP motif but not in the HD motif. Values given are means ± SD of triplicate measurements of the apparent FRET efficiency (Eapp).

Fig. 3.

Interaction between RpfG and XC_0249 or XC_0420 occurs in vivo and is dependent on the presence of the DSF signal. For these FRET experiments, RpfG was fused at the C terminus to CFP, whereas the GGDEF-domain proteins XC_0249 and XC_0420 were fused at the C terminus to YFP. FRET signals are lower when tested in the rpfF mutant of Xcc, which cannot make the DSF signal, than in the wild type. Addition of exogenous DSF to the rpfF mutant restores the level of interaction to that seen in the wild-type background. Values given are means ± SD of triplicate measurements of the apparent FRET efficiency (Eapp)

Fig. 4.

Subcellular localization of RpfG in Xcc depends upon the REC domain and DSF signaling. Xcc strains expressing GFP fusions of RpfG, the HD-GYP domain, or the sensor RpfC were spotted onto slides coated with a thin cushion of 1% agarose gel before visualization of the fluorescent proteins. The HD-GYP domain shows no specific localization in the wild type, rpfF mutant, or rpfF mutant grown with exogenously added DSF, whereas RpfC shows a polar location under all these conditions. RpfG shows a polar location in the wild type and in the rpfF mutant after growth with exogenously added DSF but not in the rpfF mutant without DSF addition.

Fig. 5.

Effects of mutation of the signature residues of the HD-GYP domain on the regulation of motility and extracellular enzyme synthesis in Xcc. (A) Expression of the HD-GYP domain with a His6 tag in the rpfG mutant restores motility to that seen in the wild-type 8004 with the empty pLAFR3 vector. Alanine substitution within the HD motif abolishes this regulatory function. Results for the AD-GYP variant are shown. (B) Alanine substitution within the GYP motif also abolishes the regulatory effect on motility. Results for the HD-AYP variant are shown. (C) Expression of the HD-GYP domain with a His6 tag in the rpfG mutant restores extracellular endoglucanase production to that seen in the wild-type 8004 with the empty pLAFR3 vector. Alanine substitutions within the GYP motif have little or no effect on this regulatory activity, although substitutions within the HD motif abolish it (6). (D) Western blot analysis with an anti-His6 antiserum shows that all variant proteins are expressed in Xcc. Lanes 1–7 in D correspond to the strains in C.

Fig. 6.

The GGDEF-domain proteins XC_0249 and XC_0420 that interact with RpfG collectively regulate motility and virulence in Xcc. (A) Although mutation of the genes encoding XC_0249 and XC_0420 alone has no effect on motility, the double mutant has the highly reduced motility seen with the rpfG mutant. (B) Symptom production on leaves clipped 14 days after inoculation with (from Left to Right) wild type, rpfG mutant, XC_0249 mutant, XC_0420 mutant and XC_0249, and XC_0420 double mutant. (C) The average lesion lengths caused by the XC_0249, XC_0420 double mutant are significantly shorter than the wild type and single mutants but are similar to those caused by the rpfG mutant. Values are the means ± SD of 140 measurements for each strain.