Interactions between partner switcher orthologs BtrW and BtrV regulate type III secretion in Bordetella

Abstract

We have recently described a multicomponent cascade that regulates type III secretion in Bordetella. This cascade includes a group of proteins, BtrU, BtrW, and BtrV, that contain an array of domains that define partner-switching complexes previously characterized in gram-positive bacteria. BtrU contains a PP2C-like serine phosphatase domain, BtrW contains a serine kinase/anti-sigma factor motif, and BtrV includes an anti-sigma factor antagonist domain. On the basis of genetic studies and sequence similarity with the RsbU-RsbW-RsbV and SpoIIE-SpoIIAB-SpoIIAA partner switchers of Bacillus subtilis, a series of interactions between Bordetella orthologs have been proposed. Bacterial two-hybrid analysis, tagged protein pull-downs, and in vitro phosphorylation assays were used to characterize interactions between BtrW and BtrV. In addition, BtrV mutants predicted to mimic a constitutively phosphorylated form of BtrV or to be nonphosphorylatable and BtrW mutants defective in serine kinase activity or the ability to bind BtrV were constructed and analyzed. Our results demonstrate that (i) BtrW and BtrV interact with each other, (ii) BtrW phosphorylates BtrV at serine S55, (iii) the conserved serine residue S55 of BtrV plays a key role in BtrV-BtrW interactions, and (iv) the ability of BtrW to phosphorylate BtrV and disrupt BtrV-BtrW binding is essential for the type III secretion process.

FIG. 1.

Partner switchers. (A) The B. subtilis partner-switching cycle regulates σ^B activity in the general stress response pathway (see the text). (B) Predicted protein structures of Bordetella BtrV, BtrW, and BtrU compared with B. subtilis orthologs. The conserved residues of BtrV and BtrW that were mutated and analyzed in this paper are shown. BtrV contains an anti-σ factor antagonist domain (STAS). BtrW contains an HPK-like serine kinase domain that forms a Bergerat fold with four motifs, called N, G1, G2, and G3 boxes, participating in binding to nucleotide ligands. BtrU contains two putative transmembrane domains separated by a stretch of mainly hydrophilic residues, a HAMP domain, and a PP2C-like serine phosphatase domain.

FIG. 2.

BtrW interacts with BtrV and can form homodimers. (A and B) Characterization of interactions between wild-type BtrW and BtrV using a bacterial two-hybrid system. The BacterioMatch two-hybrid reporter strains containing bait and target vectors that were empty or carried the indicated fusions of λcI to Gal4, BtrV, or BtrW (bait fusions) or fusions of the RNA polymerase α subunit to Gal11^P, BtrV, or BtrW (target fusions) were assayed for β-galactosidase activity. The values are the averages of three independent samples. The unpaired t test was used to compare differences in activities of the negative controls versus the samples carrying both bait and target fusions for each fusion arrangement; *, P < 0.005; **, P < 0.05. In some cases, the error bars (standard errors) are too small to be seen. (A) BtrW interacts with BtrV. The black bars represent positive and negative controls for proper functioning of the bacterial two-hybrid system. The white bars represent samples in which BtrW was the bait and BtrV was the target, while the gray bars represent the reciprocal arrangement. (B) BtrW can form homodimers (gray bars), but BtrV does not (white bars). (C and D) GST and MBP pull-down assays confirm the BtrW-BtrV interaction. (C) Glutathione-Sepharose was incubated with affinity-purified GST-BtrV and MBP-BtrW. For controls, glutathione-Sepharose was mixed with GST-BtrV and MBP-β-Gal α, GST tag and MBP-BtrW, or MBP-BtrW alone. Separated by 10% SDS-PAGE, protein complexes were probed with anti-MBP or anti-GST antibody. A band corresponding to MBP-BtrW is seen only in a sample containing GST-BtrV. (D) Amylose slurry was mixed with affinity-purified MBP-BtrW and GST-BtrV. To test the specificity of binding, amylose slurry was also mixed with purified MBP-BtrW and GST tag, MBP-β-Gal α and GST-BtrV, or GST-BtrV alone. The samples were separated by 10% SDS-PAGE and probed with anti-MBP or anti-GST antibody. A band corresponding to GST-BtrV was detected only in the sample containing MBP-BtrW.

FIG. 3.

BtrW phosphorylates BtrV in vitro. (A) Affinity-purified MBP-BtrW, GST-BtrV, or a mixture of both was incubated with [γ-³²P]ATP as described in Materials and Methods. After proteins from each sample were separated by SDS-PAGE, a ³²P-labeled band was detected by autoradiography (right). Coomassie blue staining (left) indicated that this band corresponded to GST-BtrV. (B) Multiple sequence alignment between BtrV and its B. subtilis orthologs. The DNA-derived amino acid sequence of BtrV (GenBank database accession number NP_888191) was aligned with the sequences of B. subtilis RsbV, RsbS, and SpoIIAA (GenBank accession numbers P17903, P42410, and P10727, respectively) using CLUSTAL_X (34). Identical residues are shown below the alignment in one-letter amino acid code, and “:” indicates the positions of strongly conserved residues. The conserved serine residues that align with serine 55 of BtrV are shown in boldface. (C) μLC-MSMS of the singly phosphorylated form of the peptide spanning amino acids 50 to 59 of BtrV. The presence of an ion of m/z 583.1 corresponding to y5 indicates that S55 rather than S54 is a major site of phosphorylation.

FIG. 4.

Serine 55 of BtrV plays a key role in BtrV-BtrW interactions. (A) Bacterial two-hybrid analysis of the interactions between BtrV S55 mutants and BtrW. The β-galactosidase activities of the reporter strains containing either empty or recombinant vectors carrying wild-type BtrV, BtrV(S55A), or BtrV(S55D) as bait fusions and BtrW as a target fusion were measured. The unpaired t test was used to compare differences in the activities of the strains carrying BtrV(S55A)-BtrW and BtrV(S55D)-BtrW pairs versus the strain carrying the wild-type BtrV-BtrW pair; *, P < 0.005. In addition, the unpaired t test was used to compare differences between the β-galactosidase activities of the negative controls (gray bars) versus the reporter strains containing fusions of BtrV mutants and BtrW (white bars); √, P < 0.005. In some cases, the error bars (standard errors) are too small to be seen. (B) In vitro serine kinase assay to compare GST-BtrV wild type, GST-BtrV(S55A), and GST-BtrV(S55D) as phosphorylation substrates. The affinity-purified GST fusions of wild-type and mutant forms of BtrV were tested for the ability to be phosphorylated by MBP-BtrW as described in Materials and Methods. Coomassie blue staining (left) indicated that the radiolabeled band on the autoradiogram (right) corresponded to GST-BtrV. Neither GST-BtrV(S55A) nor GST-BtrV(S55D) was phosphorylated by MBP-BtrW.

FIG.5.

BtrW N-box mutant binds but cannot phosphorylate BtrV, whereas BtrW G1-box mutant preserves its serine kinase activity despite its low affinity for BtrV. (A) Multiple sequence alignment between B. bronchiseptica and B. subtilis anti-sigma factors/serine kinase orthologs. CLUSTAL_X was used to align the DNA-derived amino acid sequence of BtrW (GenBank accession number NP_888190) with the sequences of B. subtilis RsbW, RsbT, and SpoIIAB (GenBank accession numbers P17904, P42411, and P10728, respectively). The conserved N, G1, G2, and G3 boxes of the Bergerat fold are outlined. The sequences of the conserved motifs were identified based on the alignment of the B. bronchiseptica BtrW and B. subtilis RsbW, RsbT, and SpoIIAB sequences with B. stearothermophilus SpoIIAB (GenBank accession number 1L0OB) (5). Identical residues are shown below the alignment in one-letter amino acid code, and “:” indicates the positions of strongly conserved residues. The conserved arginine residue corresponding to N51 of BtrW within the N box and the conserved aspartate and glycine residues corresponding to D82 and D84, respectively, of BtrW within the G1 box are shown in boldface. (B and C) Bacterial two-hybrid analysis of interactions between wild-type BtrV and BtrW N- and G1-box mutants. The BacterioMatch reporter strain was cotransformed with bait and target vectors carrying fusions of wild-type BtrV and wild-type BtrW, the BtrW N-box mutant, or the BtrW G1-box mutant. For the negative controls, the reporter strain was cotransformed with one vector carrying fusions to wild-type or mutant BtrW and another empty vector. The unpaired t test was used to compare differences between the activities of strains carrying BtrV-BtrW(N51A) and BtrV-BtrW(D82G84) pairs versus that of the strain with the wild-type BtrV-BtrW pair. *, P < 0.005; **, P < 0.05. The unpaired t test was also used to compare differences between the activity of the negative controls carrying a single fusion of wild-type or mutant BtrW (gray bars) versus that of the reporter strains carrying that BtrW fusion together with the BtrV fusion (white bars). √, P < 0.005; √√, P < 0.05. Note the differences in scales of β-galactosidase activity for panels B and C. In some cases, the error bars (standard errors) are too small to be seen. (B) β-Galactosidase activities of the reporter strains cotransformed with bait vectors carrying the BtrV fusions and target vectors carrying BtrW fusions. (C) β-Galactosidase activities of the reporter strains containing the reciprocal arrangement of BtrV and BtrW fusions compared to panel B. (D) Test for the ability of BtrW N-box and G1-box mutants to form homodimers. The reporter strain was cotransformed with bait and target plasmids carrying fusions of wild-type BtrW or BtrW N-box or BtrW G1-box mutants, and the β-galactosidase activities of the transformed strains were measured. The unpaired t tests to compare differences between the activity of the reporter strains carrying BtrW mutant-BtrW wild-type or double-BtrW mutant pairs versus the activity of the strain carrying a double-wild-type BtrW pair indicate that the activity levels are comparable. The unpaired t test was used to compare differences between the activities of the negative controls (gray bars) versus the strains carrying fusions on both bait and target vectors (white bars); √, P < 0.005. In some cases, the error bars are too small to be seen. (E) In vitro phosphorylation assay to assess serine kinase activities of MBP-BtrW N-box and MBP-BtrW G1-box mutants. Affinity-purified MBP-BtrW(N51A) and MBP-BtrW(D82AG84A) proteins were tested for the ability to phosphorylate GST-BtrV upon incubation with [γ-³²P]ATP. The ³²P-labeled bands were detected by autoradiography (right). The same amounts of the purified fusion proteins as were used in the kinase reactions were mixed and separated by SDS-PAGE, followed by Coomassie blue staining (left). The Coomassie blue staining indicated that ³²P-labeled bands corresponded to GST-BtrV. The MBP-BtrW G1-box mutant can phosphorylate GST-BtrV, albeit more weakly than the MBP-BtrW wild type. (F) Assay for the ability of BtrV(S55A) to interact stably with the BtrW G1-box mutant in the bacterial two-hybrid system. β-Galactosidase activity was measured in the reporter strain cotransformed with bait vectors carrying the wild type or the S55A mutant of BtrV and the target vector carrying the wild type or the G1-box mutant of BtrW. Negative controls are in gray. The unpaired t test was used to compare the differences between the activity of the strain containing the BtrV(S55A)-BtrW G1-box mutant pair versus the strain containing the BtrV wild type-BtrW G1-box mutant pair; *, P < 0.005. In addition, an unpaired t test was conducted to evaluate the differences between the activities of the strains used in the test described above versus the reporter strain with wild-type BtrV and BtrW fusions; √, P < 0.005. In some cases, the error bars are too small to be seen.

FIG. 6.

B. bronchiseptica btrV(S55A), btrV(S55D), btrW(N51A), and btrW(D82AG84A) mutants are defective in type III secretion. Pellet and supernatant fractions of B. bronchiseptica wild-type and various mutant strains defective in type III secretion were separated by SDS-PAGE, transferred to PVDF membranes, and probed with polyclonal antibodies generated against Bsp22 and BopD.