VpsR and cyclic di-GMP together drive transcription initiation to activate biofilm formation in Vibrio cholerae

Abstract

The small molecule cyclic di-GMP (c-di-GMP) is known to affect bacterial gene expression in myriad ways. In Vibrio cholerae in vivo, the presence of c-di-GMP together with the response regulator VpsR results in transcription from PvpsL, a promoter of biofilm biosynthesis genes. VpsR shares homology with enhancer binding proteins that activate σ54-RNA polymerase (RNAP), but it lacks conserved residues needed to bind to σ54-RNAP and to hydrolyze adenosine triphosphate, and PvpsL transcription does not require σ54 in vivo. Consequently, the mechanism of this activation has not been clear. Using an in vitro transcription system, we demonstrate activation of PvspL in the presence of VpsR, c-di-GMP and σ70-RNAP. c-di-GMP does not significantly change the affinity of VpsR for PvpsL DNA or the DNase I footprint of VpsR on the DNA, and it is not required for VpsR to dimerize. However, DNase I and KMnO4 footprints reveal that the σ70-RNAP/VpsR/c-di-GMP complex on PvpsL adopts a different conformation from that formed by σ70-RNAP alone, with c-di-GMP or with VpsR. Our results suggest that c-di-GMP is required for VpsR to generate the specific protein-DNA architecture needed for activated transcription, a previously unrecognized role for c-di-GMP in gene expression.

Figure 1.

Summary of in vitro primer extension and DNase I and KMnO₄ footprinting. (A) Sequence of P_vpsL from −60 to +30. Bold and underlined A with black arrow at +1 and bold G (+3) represent the TSS determined by primer extensions; the −10 element and the −35 region are labeled and boxed in green; sequences in bold and red denote the VpsR binding site. Protection sites from DNase I footprinting and hypersensitive sites are depicted as rectangular boxes and triangles, respectively, either above (non-template) or below (template) the sequences: gray, RNAP with or without c-di-GMP or VpsR; black, RNAP with VpsR and c-di-GMP; red, VpsR with or without c-di-GMP. The open transcription bubble detected using KMnO₄ footprinting is shown as separated ssDNA from position −11 to +2 with sites of KMnO₄ cleavage indicated as purple asterisks. (B) Summary of positions of protection and hypersensitive sites on non-template and template strand DNA.

Figure 2.

VpsR and c-di-GMP activate transcription at P_vpsL by ∼7-fold in vitro. (A) Representative gel showing P_vpsL RNA obtained after multiple round in vitro transcription reactions using plasmid template P_vpsL with RNAP alone (lane 1), RNAP and c-di-GMP (lane 2), RNAP and VpsR (lane 3), and RNAP, VpsR, and c-di-GMP (lane 4). (B) Graph showing the level of P_vpsL transcription relative to that with RNAP alone (basal) obtained from at least three independent experiments (one-way ANOVA with Tukey’s HSD (honest significant difference) posthoc analysis, *P < 0.05).

Figure 3.

Identification of +1 TSS using in vitro and in vivo primer extensions. RNA was isolated from in vitro transcriptions (A) or V. cholerae (B). Two major primer extension products, which are observed only in the presence of both VpsR and c-di-GMP are indicated with arrows.

Figure 4.

KMnO₄ footprinting assigns the +1 TSS at the A located 59 bp upstream of the vpsL translation start site. (A) Reactive thymines within the transcription bubble are observed at positions −11, −4, −3, -2, −1, +1 and +2 on template DNA. (B) Reactive thymines within the transcription bubble are also observed at positions −6 and −7 on non-template DNA. GA corresponds to G+A ladder.

Figure 5.

VpsR forms dimers in vitro with or without c-di-GMP. (A) Samples containing 1.5 μM VpsR with and without 50 μM c-di-GMP were treated with the chemical crosslinker BS³ as indicated and separated on a 10–20% (wt/vol) Tricine gel that was stained with Colloidal Blue. (B) Samples containing 0.6 μM VpsR with and without 12.5 μM c-di-GMP in transcription buffer were treated with the chemical crosslinker BS³ as indicated and separated on a 10–20% (wt/vol) Tricine gel that was silver stained. Far left lane of each panel contains marker proteins, whose molecular weights are indicated. Black arrows indicate position of VpsR monomer (∼50 kDa) and gray arrows indicate position of VpsR dimer (∼100 kDa). Each sample was repeated independently three times, and a representative gel image is shown.

Figure 6.

VpsR binds P_vpsL DNA with similar affinity with or without c-di-GMP. Representative gels showing the retardation of ³²P-labeled DNA harboring -97 to +103 of P_vpsL with increasing VpsR concentrations from 0 to 2 μM either in the absence (lanes 1–5) or presence (lanes 6–10) of 50 μM c-di-GMP. Black arrows indicate retarded complexes while gray arrow indicates free DNA. (B) Quantitation of EMSAs. Apparent DNA-binding dissociation constants (K_d(app)) were calculated as the concentration of VpsR needed to retard 50% of the free DNA. Values from at least three EMSAs were analyzed using one-way ANOVA with Tukey’s HSD posthoc analysis (ns, not significant).

Figure 7.

DNase I footprinting of P_vpsL complexes on (A) nontemplate DNA and (B) template DNA. GA corresponds to G+A ladder. VpsR, c-di-GMP and/or RNAP are present as indicated. To the right of each gel image, a schematic indicates the −10 and −35 regions and the +1. The VpsR binding site is indicated as a dashed black line. DNase I protection regions and hypersensitive sites seen with the activated complex of RNAP, VpsR, c-di-GMP and DNA are depicted as black rectangles and horizontal arrows, respectively. The dashed red boxes indicate the regions of DNA where the protection/enhancement within and immediately adjacent to the VpsR binding site changes when comparing complexes containing RNAP with or without VpsR or c-di-GMP to the activated complex. (See text for details.)