Reciprocal control of motility and biofilm formation by the PdhS2 two-component sensor kinase of Agrobacterium tumefaciens

Abstract

A core regulatory pathway that directs developmental transitions and cellular asymmetries in Agrobacterium tumefaciens involves two overlapping, integrated phosphorelays. One of these phosphorelays putatively includes four histidine sensor kinase homologues, DivJ, PleC, PdhS1 and PdhS2, and two response regulators, DivK and PleD. In several different alphaproteobacteria, this pathway influences a conserved downstream phosphorelay that ultimately controls the phosphorylation state of the CtrA master response regulator. The PdhS2 sensor kinase reciprocally regulates biofilm formation and swimming motility. In the current study, the mechanisms by which the A. tumefaciens sensor kinase PdhS2 directs this regulation are delineated. PdhS2 lacking a key residue implicated in phosphatase activity is markedly deficient in proper control of attachment and motility phenotypes, whereas a kinase-deficient PdhS2 mutant is only modestly affected. A genetic interaction between DivK and PdhS2 is revealed, unmasking one of several connections between PdhS2-dependent phenotypes and transcriptional control by CtrA. Epistasis experiments suggest that PdhS2 may function independently of the CckA sensor kinase, the cognate sensor kinase for CtrA, which is inhibited by DivK. Global expression analysis of the pdhS2 mutant reveals a restricted regulon, most likely functioning through CtrA to separately control motility and regulate the levels of the intracellular signal cyclic diguanylate monophosphate (cdGMP), thereby affecting the production of adhesive polysaccharides and attachment. We hypothesize that in A. tumefaciens the CtrA regulatory circuit has expanded to include additional inputs through the addition of PdhS-type sensor kinases, likely fine-tuning the response of this organism to the soil microenvironment.

Keywords: Agrobacterium tumefaciens; biofilm; development; motility; phosphorelay; sensor kinase.

Fig. 1.

The PdhS kinases of C. crescentus and A. tumefaciens localize differentially and affect phenotypic outputs through the response regulators DivK, PleD and CtrA. (a) Cartoon model of known localization of the namesake PdhS kinases from C. crescentus , PleC and DivJ, and three PdhS kinases from A. tumefaciens , PdhS1, PdhS2 and DivJ. Kinases represented as coloured ovals with a black border experimentally localize to the indicated poles. The PleC oval without a border has not been experimentally demonstrated to localize in A. tumefaciens . As a result of this localization, the phosphorylation status of the direct PdhS kinase targets, DivK and PleD, and the indirect target, CtrA, may be differentially affected. (b) Multiple sequence alignment of the HisKA domain from the PdhS kinases of C. crescentus and A. tumefaciens . Sequences were aligned using the Clustal Omega web service hosted by the European Molecular Biology Laboratory (EMBL) – European Bioinformatics Institute. The four PdhS kinases from A. tumefaciens plus PleC and DivJ from C. crescentus were included. Also included were two additional predicted PdhS kinases, CC_0652 and CC_1062, from C. crescentus . The EnvZ sensor kinase is included for comparison. Yellow highlighting indicates residues that define the PdhS kinases. The conserved histidine and threonine residues mutated in this work are in bold.

Fig. 2.

Evaluation of roles for PdhS2 kinase and phosphatase activities and genetic interactions with divK. (a) The ability of plasmid-borne wild-type PdhS2 (p-pdhS2), the kinase-null allele (p-pdhS2, K⁻P⁺), or the phosphatase-null allele (p-pdhS2, K⁺P⁻) to complement the ΔpdhS2 biofilm formation (black bars) and swimming motility (white bars) phenotypes was evaluated using P_lac-driven expression of each allele. Static biofilm formation was measured after 48 h (black bars) and swim ring diameter was measured after 7 days (white bars). Adherent biomass on PVC coverslips was determined by adsorption of crystal violet. Crystal violet was then solubilized and A_{600 nm} values were normalized to culture density (OD₆₀₀). The data are the mean of three independent experiments, each of which contained three technical replicates (n=3). Swim ring diameters were measured after single-colony inoculation into low density swim agar and incubation at room temperature. The data are the mean of nine independent experiments (n=9). (b) Biofilm formation (black bars) and swimming motility (white bars) were evaluated in the indicated strains. Experiments were performed and data were analysed as described for (a) above. (c) The effect of plasmid-borne wild-type PdhS2 (p-pdhS2), the kinase-null allele [p-pdhS2 (K⁻P⁺)], or the phosphatase-null allele [p-pdhS2 (K⁺P⁻)] on biofilm formation (black bars) and swimming motility (white bars) when expressed from the P _lac promoter in the ΔdivK mutant background was evaluated as in (a) and (b) above. For presentation, all data have been normalized to the wild-type (WT) and they are expressed as %WT ± standard error of the mean (se). 1=P<0.05 compared to the wild-type strain or the wild-type strain carrying empty vector. 2=P<0.05 compared to the ΔpdhS2 strain carrying empty vector (a), or compared to the ΔpdhS2 mutant strain (b), or compared to the ΔdivK strain carrying empty vector (c). Statistical significance was determined using Student’s t-test.

Fig. 3.

A kinase-locked allele of CckA fails to suppress the PdhS2-dependent biofilm and motility phenotypes. (a) Biofilm formation was evaluated in the indicated strains as described in Fig. 2. *, P<0.05 compared to background strain carrying vector alone. Statistical significance was determined using Student’s t-test. (b) Swimming motility was evaluated in the indicated strains as described in Fig. 2. *, P<0.001 compared to background strain carrying vector alone. Statistical significance was determined using Student’s t-test.

Fig. 4.

PdhS2 intersects with the activity of multiple diguanylate cyclases. (a) Biofilm formation was quantified for the wild-type (WT) and indicated mutant strains as described in Fig. 2. PleD, DgcA and DgcB have demonstrated in vivo diguanylate cyclase enzymatic activity. Thus far, the conditions under which DgcC is active have yet to be identified. P<0.05 compared to WT (1), ΔpdhS2 (2), or the corresponding diguanylate cyclase mutant (3). (b) The effect on biofilm formation of plasmid-borne expression of wild-type dgcB (p-dgcB) or a catalytic mutant allele of dgcB (p-dgcB*) was evaluated. Expression of each dgcB allele was driven by the P_lac promoter. Biofilm formation was evaluated as described in Fig. 2. *, P<0.05 compared to vector alone.

Fig. 5.

Loss of pdhS2 enhances biofilm formation in the absence of both dgcB and pleD. Biofilm formation and swimming motility was evaluated in the wild-type (WT) and indicated mutant strains as described in Fig. 2. *, P<0.05 compared to the wild-type background. Statistical significance was determined using Student’s t-test.

Fig. 6.

The unipolar polysaccharide is required for PdhS2-dependent biofilm formation. (a) Biofilm formation was evaluated in the presence (+) or absence (−) of pdhS2 in combination with mutants deficient in production of the indicated polysaccharides. WT=wild type, Cel⁻=cellulose mutant, ChvAB⁻=cyclic-β-glucan mutant, CrdS⁻=curdlan mutant, ExoA⁻=succinoglycan mutant, UPP⁻=unipolar polysaccharide mutant, EPS⁻=mutant lacking all of the above polysaccharides. (b) Swimming motility was evaluated in the same strains as in (a). *, P<0.05 compared to the PdhS2+ background strain. Statistical significance was determined using Student’s t-test.

Fig. 7.

An alternative model for PdhS2 regulation of CtrA activity. Our data are consistent with PdhS2 intersecting the DivK–CtrA regulatory pathway at one of two points. Pathway A: canonical genetic model with PdhS2 interacting with DivK. The phosphorylation status of DivK then modulates CtrA activity through the CckA–ChpT–CtrA axis. Pathway B: DivK-independent model of CtrA regulation by PdhS2 through an unidentified response regulator, RR-X. Both routes to the regulation of CtrA activity ultimately affect the phosphorylation status of CtrA, affecting occupancy at CtrA-regulated promoters, and finally leading to inverse regulation of attachment (primarily through cdGMP pools) and separately motility. Regulatory proteins: blue text; histidine kinases; orange text, histidine phosphotransferase (Hpt), green text, response regulators. RR-X indicates a putative response regulator, yet to be identified.