A Novel Bvg-Repressed Promoter Causes vrg-Like Transcription of fim3 but Does Not Result in the Production of Serotype 3 Fimbriae in Bvg- Mode Bordetella pertussis.

In Bordetella pertussis, two serologically distinct fimbriae, FIM2 and FIM3, undergo on/off phase variation independently of each other via variation in the lengths of C stretches in the promoters for their major subunit genes, fim2 and fim3 These two promoters are also part of the BvgAS virulence regulon and therefore, if in an on configuration, are activated by phosporylated BvgA (BvgA~P) under normal growth conditions (Bvg+ mode) but not in the Bvg- mode, inducible by growth in medium containing MgSO4 or other compounds, termed modulators. In the B. pertussis Tohama I strain (FIM2+ FIM3-), the fim3 promoter is in the off state. However, a high level of transcription of the fim3 gene is observed in the Bvg- mode. In this study, we provide an explanation for this anomalous behavior by defining a Bvg-repressed promoter (BRP), located approximately 400 bp upstream of the Pfim3 transcriptional start. Although transcription of the fim3 gene in the Bvg- mode resulted in Fim3 translation, as measured by LacZ translational fusions, no accumulation of Fim3 protein was detectable. We propose that Fim3 protein resulting from translation of mRNA driven by BRP in the Bvg- mode is unstable due to a lack of the fimbrial assembly apparatus encoded by the fimBC genes, located within the fha operon, and therefore is not expressed in the Bvg- mode.IMPORTANCE In Bordetella pertussis, the promoter Pfim3-15C for the major fimbrial subunit gene fim3 is activated by the two-component system BvgAS in the Bvg+ mode but not in the Bvg- mode. However, many transcriptional profiling studies have shown that fim3 is transcribed in the Bvg- mode even when Pfim3 is in a nonpermissive state (Pfim3-13C), suggesting the presence of a reciprocally regulated element upstream of Pfim3 Here, we provide evidence that BRP is the cause of this anomalous behavior of fim3 Although BRP effects vrg-like transcription of fim3 in the Bvg- mode, it does not lead to stable production of FIM3 fimbriae, because expression of the chaperone and usher proteins FimB and FimC occurs only in the Bvg+ mode.

Activation of Bvg-Repressed Genes in Bordetella pertussis by RisA Requires Cross Talk from Noncooperonic Histidine Kinase RisK.

The two-component response regulator RisA, encoded by open reading frame BP3554 in the Bordetella pertussis Tohama I genomic sequence, is a known activator of vrg genes, a set of genes whose expression is increased under the same environmental conditions (known as modulation) that result in repression of the bvgAS virulence regulon. Here we demonstrate that RisA is phosphorylated in vivo and that RisA phosphorylation is required for activation of vrg genes. An adjacent histidine kinase gene, risS, is truncated by frameshift mutation in B. pertussis but not in Bordetella bronchiseptica or Bordetella parapertussis Neither deletion of risS' or bvgAS nor phenotypic modulation with MgSO4 affected levels of phosphorylated RisA (RisA∼P) in B. pertussis However, RisA phosphorylation did require the histidine kinase encoded by BP3223, here named RisK (cognate histidine kinase of RisA). RisK was also required for expression of the vrg genes. This requirement could be obviated by the introduction of the phosphorylation-mimicking RisAD60E mutant, indicating that an active conformation of RisA, but not phosphorylation per se, is crucial for vrg activation. Interestingly, expression of vrg genes is still modulated by MgSO4 in cells harboring the RisAD60E mutation, suggesting that the activated RisA senses additional signals to control vrg expression in response to environmental stimuli.IMPORTANCE In B. pertussis, the BvgAS two-component system activates the expression of virulence genes by binding of BvgA∼P to their promoters. Expression of the reciprocally regulated vrg genes requires RisA and is also repressed by the Bvg-activated BvgR. RisA is an OmpR-like response regulator, but RisA phosphorylation was not expected because the gene for its presumed, cooperonic, histidine kinase is inactivated by mutation. In this study, we demonstrate phosphorylation of RisA in vivo by a noncooperonic histidine kinase. We also show that RisA phosphorylation is necessary but not sufficient for vrg activation but, importantly, is not affected by BvgAS status. Instead, we propose that vrg expression is controlled by BvgAS through its regulation of BvgR, a cyclic di-GMP (c-di-GMP) phosphodiesterase.