Overexpression of the RNA Polymerase Alpha Subunit Reduces Transcription of Bvg-Activated Virulence Genes in Bordetella pertussis

Abstract

Overexpression of the RNA polymerase alpha subunit in Bordetella pertussis reduces expression of the virulence factor pertussis toxin. Here we show that this reduction is at the level of transcription, is reversed by overexpression of the transcriptional activator BvgA, and is dependent on the C-terminal domain of alpha.

In Bordetella pertussis, expression of virulence factors is regulated by the Bvg two-component signal transduction system, comprising the sensor BvgS and the transcriptional activator BvgA (1, 11). The Bvg system is modulated (with loss of virulence factor expression) by reduced temperature (<30°C) or the presence of sulfate ions or nicotinic acid in the growth medium (11). Previously we showed that mutant B. pertussis strains with reduced expression of pertussis toxin (Ptx) and adenylate cyclase/hemolysin toxin, but not of other Bvg-regulated virulence factors such as filamentous hemagglutinin (Fha), had mutations upstream of the rpoA gene, which encodes the alpha subunit of RNA polymerase (RNAP) (3). These mutations caused a two- to threefold overexpression of alpha through an increase in translation of the rpoA gene (3). We also showed that inducible overexpression of alpha from a recombinant plasmid in B. pertussis had the same effect (3). The alpha subunit is a common site of interaction of RNAP with transcription activator proteins (6). We therefore hypothesized that the observed effect on virulence factor expression was due to interaction of the excess alpha with BvgA, effectively reducing the level of BvgA present in cells for functional interactions with RNAP. To obtain further evidence that the excess alpha affects BvgA-dependent transcription activation, we first assessed the effect of overexpressing alpha on transcription of both the ptx and fha genes.

Overexpression of alpha reduces transcription of both ptx and fha.

We introduced a ptx-lac transcriptional fusion (8) into the chromosome of wild-type (Tohama I) and mutant (alpha-overexpressing strains BC75 and RPV3 and the bvg knockout strain Tohama I Δbvg) B. pertussis strains as previously described (8). We also introduced a fha-lac transcriptional fusion into the chromosome of the same set of strains, by allelic exchange from the plasmid pSS1581 (kindly provided by Scott Stibitz). The fusion strains were grown at 37°C in SS medium (9) to mid-log phase (nonmodulating conditions that allow full Bvg activity), and then β-galactosidase assays (8) were performed on the cultures to determine the level of ptx and fha transcription. As seen in Fig. 1, the level of ptx transcription is significantly reduced in both alpha-overexpressing mutants (approximately twofold in RPV3 and threefold in BC75), but the level of fha transcription is not significantly reduced in these strains. Since there is only a modest (two- to threefold) overexpression of alpha in RPV3 and BC75 (3), we introduced the plasmid pNMD120 (encoding IPTG [isopropyl-β-d-thiogalactopyranoside]-inducible expression of B. pertussis rpoA) (3), as well as the vector control plasmid pNMD121, into the Tohama I ptx-lac and fha-lac fusion strains. IPTG induction of Tohama I(pNMD120) results in a greater (approximately fivefold) overexpression of alpha (3). Cultures of these strains were grown in the absence or presence of 50 μM IPTG, and β-galactosidase assays were performed as before. As seen in Fig. 1, the higher level of alpha overexpression from pNMD120 significantly reduces fha transcription, though still not to the same extent as ptx transcription. The control vector pNMD121 had no significant effect on ptx or fha transcription. We conclude from these data that overexpression of alpha can reduce Bvg-activated transcription both of genes that require a high concentration of BvgA (ptx) and of genes that require a lower concentration of BvgA (fha) for transcription activation. Therefore, the effect of alpha overexpression likely acts through Bvg rather than at specific loci.

FIG. 1.

Effect of alpha overexpression on ptx and fha transcription in B. pertussis. Wild-type, mutant, and plasmid-containing strains carrying either a ptx-lac or a fha-lac fusion were assayed for β-galactosidase production (indicated as lac units). Shaded bars, cultures with no IPTG; solid bars, cultures grown in the presence of 50 μM IPTG. Results are means of at least three experiments with standard deviations. Asterisks mark values that are significantly (P < 0.05) reduced from values for wild-type (Tohama I) or vector control strains [Tohama I(pNMD121)].

Overexpression of BvgA in alpha-overexpressing mutants restores wild-type ptx expression.

A prediction of the hypothesis that interaction of excess alpha with BvgA causes the reduction in Bvg-activated transcription is that this effect would be reversed by a compensatory overexpression of BvgA. To test this, we constructed pNMD124, which contains IPTG-inducible bvgAS genes, and introduced either this plasmid or the vector control pNMD121 into the alpha overexpressing mutants BC75 and RPV3 carrying the ptx-lac fusion. Cultures were grown and β-galactosidase assays were performed as before. As seen in Fig. 2, IPTG induction of Bvg expression from pNMD124 significantly increased ptx transcription in both RPV3 and BC75 to near wild type levels [the level in Tohama I(pNMD121)]. These data are consistent with the hypothesis that excess alpha interacts with BvgA to reduce ptx transcription. Interestingly, IPTG induction of BvgA expression alone (without BvgS) from a different construct (pNMD123) did not cause any significant increase in ptx transcription in the same strains (data not shown). An explanation for this observation may be that excess BvgS is required to allow phosphorylation of the excess BvgA and that only phosphorylated BvgA can interact with alpha and RNAP. This would be consistent with the observation that only phosphorylated BvgA mediates transcription of Bvg-activated genes in in vitro transcription experiments (2, 10). However, from these data we cannot rule out the alternative, less likely explanation that BvgS, rather than BvgA, interacts with the excess alpha to reverse the reduction of ptx transcription.

FIG. 2.

Effect of BvgAS overexpression (from pNMD124) on ptx transcription in the alpha-overexpressing mutants RPV3 and BC75. Wild-type and mutant plasmid-containing strains carrying a ptx-lac fusion were assayed for β-galactosidase production (indicated as lac units). Shaded bars, cultures with no IPTG; solid bars, cultures grown in the presence of 50 μM IPTG. Results are means of at least three experiments with standard deviations. Asterisks mark values that are significantly (P < 0.05) reduced from those for the vector control strain [Tohama I(pNMD121)].

Interaction with the alpha CTD causes reduction of ptx expression.

In a previous study of BvgA-mediated transcription from the fha promoter in vitro (2), it was observed that transcription was significantly reduced when the RNAP contained the alpha subunit with either a deletion of the C-terminal domain (CTD) or a substitution of alanine for arginine at position 265 within the CTD. The conclusion was that BvgA interacts with the alpha CTD to activate transcription from this promoter. To test whether the same interaction with the alpha CTD might mediate our observed reduction of ptx transcription in alpha-overexpressing strains, we constructed a series of plasmids encoding IPTG-inducible expression of truncated and mutant alpha subunits (Fig. 3). The alpha-encoding fragments were amplified by PCR and cloned into the vector pNMD121 to derive these plasmids, which were then introduced into B. pertussis Tohama I by conjugation. The IPTG-inducible overexpression of alpha was confirmed for each construct by Western blotting of whole-cell lysates of cultures grown in the absence and presence of 50 μM IPTG, using either monoclonal antibody 4RA2 (Fig. 4a) or polyclonal antiserum (Fig. 4b) against Escherichia coli alpha (cross-reacts with B. pertussis alpha; kindly provided by Nancy Thompson and Richard Burgess). The effect of overexpression of the various alpha derivatives on expression of Ptx was then assessed by Western blotting of trichloroacetic acid-precipitated supernatant proteins (4) from cultures grown in the absence and presence of 50 μM IPTG (Fig. 4c), using monoclonal antibody X2X5 to Ptx S1 subunit (kindly provided by Drusilla Burns). As seen in Fig. 4c, overexpression of B. pertussis alpha dramatically reduces production of Ptx, as we previously observed (3). Overexpression of E. coli alpha (from pNMD126) also reduced Ptx expression, though not to the same extent. This may be due to small differences in expression levels of the different alphas, or possibly to a weaker interaction of BvgA with E. coli alpha than with B. pertussis alpha. Overexpression of either the E. coli alpha N-terminal domain (NTD) (residues 1 to 240) from pNMD128 or E. coli alpha with the R265A mutation in the CTD (pNMD135) did not reduce Ptx expression, whereas overexpression of the B. pertussis alpha CTD (residues 246 to 328 fused with glutathione S-transferase) from pNMD138 did reduce Ptx expression (overexpression of glutathione S-transferase without the alpha CTD had no effect on Ptx production [data not shown]). Collectively these data strongly suggest that the CTD of alpha mediates the reduction of Ptx expression in alpha-overexpressing strains, consistent with the idea that the effect is due to interaction of the alpha CTD with BvgA. To confirm that the effect of overexpression of the alpha derivatives on Ptx production was due to reduction of transcription, we analyzed ptx transcription from the plasmid-containing strains grown in the absence and presence of 50 μM IPTG by reverse transcription (RT)-PCR. RNA was prepared from mid-log-phase cultures and RT-PCR was performed with ptx-specific primers as previously described (7). Primers specific for B. pertussis sodB (5, 7) were used as a Bvg-independent internal control, and quantitation of RT-PCR data for ptx transcription is shown in Fig. 3. The results are consistent with the Ptx production data (Fig. 4c), showing reduction of ptx transcription by IPTG induction of alpha overexpression from pNMD120, pNMD126, and pNMD138 but not from pNMD128 and pNMD135. We conclude that overexpression of the RNAP alpha subunit reduces expression of Bvg-activated virulence genes in B. pertussis and that this effect is mediated by the CTD of alpha (involving residue R265), probably by interaction with BvgA, reducing its effective concentration for productive interaction with RNAP.

FIG. 3.

Effect of overexpression of alpha or alpha fragments on ptx transcription in B. pertussis. Tohama I strains carrying the indicated plasmids and a ptx-lac fusion were assayed for ptx transcription by RT-PCR [indicated as percent transcription, where the value for Tohama I(pNMD121) with no IPTG is 100%]. Shaded bars, cultures with no IPTG; solid bars, cultures grown in the presence of 50 μM IPTG. Results are means of at least three experiments (standard deviations were less than 20%). Asterisks mark values that are significantly (P < 0.05) reduced from those for the vector control strain [Tohama I(pNMD121)]. Bp, B. pertussis; Ec, E. coli; GST, glutathione S-transferase.

FIG. 4.

Western immunoblots to detect overexpressed alpha or secreted Ptx S1 subunit from B. pertussis strains carrying the indicated plasmids. +, cultures grown in the presence of 50 μM IPTG; −, cultures grown without IPTG. (a) Whole-cell lysates to detect overexpressed alpha using monoclonal antibody 4RA2 (specific to the alpha CTD). Bands corresponding to alpha and the glutathione S-transferase-alpha CTD fusion are indicated. (b) Whole-cell lysates to detect overexpressed alpha by using polyclonal antibody to alpha (the monoclonal antibody 4RA2 does not recognize the alpha NTD or the R265A mutant alpha). Bands corresponding to alpha and the alpha NTD are indicated. (c) Trichloroacetic acid precipitates of supernatant proteins to detect Ptx S1 subunit by using monoclonal antibody X2X5.