Gene dosage compensation calibrates four regulatory RNAs to control Vibrio cholerae quorum sensing

Abstract

Quorum sensing is a mechanism of cell-to-cell communication that allows bacteria to coordinately regulate gene expression in response to changes in cell-population density. At the core of the Vibrio cholerae quorum-sensing signal transduction pathway reside four homologous small RNAs (sRNAs), named the quorum regulatory RNAs 1-4 (Qrr1-4). The four Qrr sRNAs are functionally redundant. That is, expression of any one of them is sufficient for wild-type quorum-sensing behaviour. Here, we show that the combined action of two feedback loops, one involving the sRNA-activator LuxO and one involving the sRNA-target HapR, promotes gene dosage compensation between the four qrr genes. Gene dosage compensation adjusts the total Qrr1-4 sRNA pool and provides the molecular mechanism underlying sRNA redundancy. The dosage compensation mechanism is exquisitely sensitive to small perturbations in Qrr levels. Precisely maintained Qrr levels are required to direct the proper timing and correct patterns of expression of quorum-sensing-regulated target genes.

Figure 1

Model of the core of the V. cholerae quorum-sensing circuit. The backbone of the quorum-sensing signalling pathway is depicted in black. Auto-inducer inputs are ultimately transmitted to LuxO. At LCD, LuxO-P functions together with σ⁵⁴ to activate transcription of the genes encoding the four Qrr sRNAs. The Qrr sRNAs, in conjunction with Hfq, repress translation of hapR mRNA. When hapR translation is derepressed, HapR controls downstream target genes. The previously defined feedback loops are shown in blue. HapR and LuxO auto-repress the hapR and luxO promoters, respectively (see discussion for details on the LuxO auto-repression loop). HapR also enhances qrr transcription through an unknown factor, denoted by ‘X'. The feedback loop between the Qrr sRNAs and LuxO identified in this work is shown in red. Arrows indicate positive interactions, T-bars indicate negative interactions.

Figure 2

Qrr sRNA levels in wild-type and triple qrr deletion strains. (A) Northern blots showing Qrr levels in wild-type V. cholerae (lane 1), V. cholerae qrr triple deletion strains, possessing only the qrr gene encoding the sRNA indicated on the right (lane 2), and V. cholerae qrr single deletion strains, lacking only the qrr gene encoding the Qrr sRNA indicated on the right (lane 3). Total RNA was visualized with ethidium bromide as the loading control (not shown). (B) Northern blot showing Qrr4 levels in a V. cholerae qrr4 single deletion strain expressing qrr4 from the Ptac promoter (lane 1) and a V. cholerae Δqrr1–4 quadruple deletion expressing qrr4 from the Ptac promoter (lane 2). 5S RNA is shown as a loading control. Total RNA was collected from the indicated strains at OD₆₀₀=0.1.

Figure 3

Dosage compensation is insensitive to the origin of the Qrr sRNAs. Light production from the indicated qrr–lux constructs was measured at OD₆₀₀=0.1 in a V. cholerae Δqrr1–4 mutant carrying the vector (white bars), V. cholerae wild type carrying the vector (black bars), a V. cholerae Δqrr1–4 mutant expressing qrr4 under control of the endogenous qrr4 promoter on the vector (striped bars), and a Δqrr1–4 mutant expressing qrr4 under control of the Ptac promoter on the same vector (dotted bars). Each bar shows the average light production from three independent cultures. Error bars indicate one standard deviation from the mean. RLU: relative light units.

Figure 4

luxOU mRNA is a target of Qrr sRNA translational repression. (A) Alignment of the reverse complement of the conserved pairing region of the Qrr sRNAs with the 5′-UTR of two known target mRNAs, hapR and vca0939, and the 5′-UTR of luxOU mRNA. The region of the Qrr sRNAs that pair with the hapR and vca0939 5′-UTRs is completely conserved among the four Qrrs. Nucleotides in the target mRNAs that are complementary to the Qrr sRNAs are highlighted in white on black background. The underlined sequence (UAGG) of luxOU mRNA is mutated to AUCC in the luxO^AUCC mutant. (B) Degradation of the luxOU mRNA was measured by northern blot in V. cholerae wild-type, Δqrr1–4 and Δqrr1–4, Ptac-qrr4 following transcription termination. The indicated times are seconds after addition of rifampicin. 5S RNA is shown as a loading control. (C) E. coli SLS1277 carrying plasmids harbouring either LuxO–GFP (pSLS146) or LuxO^AUCC–GFP (pSLS152) protein fusions were grown overnight in either LB (black bars) or LB supplemented with 0.4% arabinose (white bars). The experiment was performed in duplicate on three separate occasions. Error bars indicate one standard deviation from the mean of all six measurements.

Figure 5

The LuxO-Qrr feedback loop contributes to Qrr dosage compensation. Light production from the indicated qrr–lux fusions was measured at OD₆₀₀=0.1 in V. cholerae wild-type (black bars), ΔhapR (white bars), luxO^AUCC (grey bars) and ΔhapR, luxO^AUCC (striped bars) strains containing or lacking the four chromosomal qrr genes. In each case, Qrr dosage compensation is calculated as the light produced from the Δqrr1–4 mutant divided by the light produced from the isogenic qrr1–4⁺ strain. Regarding the qrr3–lux data, the two slashes indicate that the bars extend beyond the scale of the y axis. The fold-dosage-compensation value is indicated above the corresponding bar. Light production from each strain was measured in triplicate. Error bars indicate one standard deviation from the mean. The standard deviation is ±0.39 for qrr3–lux in wild type, and ±0.09 for qrr3–lux in the luxO^AUCC mutant.

Figure 6

Qrr dosage compensation is not accurate in the qrr triple deletion mutants. (A) The total level of Qrr sRNAs in wild-type (denoted WT), Δqrr2,3,4 (denoted qrr1⁺), Δqrr1,3,4 (denoted qrr2⁺), Δqrr1,2,4 (denoted qrr3⁺), and Δqrr1,2,3 (denoted qrr4⁺) V. cholerae strains grown to OD₆₀₀=0.1 was measured by northern blot with a probe for the 32-bp region that is completely conserved among Qrr1–4 sRNAs. 5S RNA is shown as a loading control. (B) The level of hapR mRNA in the same strains shown in (A) and in a Δqrr1–4 V. cholerae strain grown to OD₆₀₀=0.1 was measured by RT–PCR. Error bars indicate one standard deviation from the mean of triplicate measurements. The experiment was repeated three times with similar results.

Figure 7

Dosage compensation is sensitive to the loss of any individual qrr gene. (A) Northern blot showing Qrr4 levels in V. cholerae wild-type, single (Δqrr3), double (Δqrr2,3), and triple (Δqrr1,2,3) qrr mutants grown to OD₆₀₀=0.1. 5S RNA is shown as a loading control. (B) Relative levels of Qrr1–4 and hapR mRNA were measured by RT–PCR in the same strains shown in (A) as well as in a Δqrr1,2,3,4 mutant strain carrying a plasmid-borne Ptac-qrr4 overexpression construct (Δqrr1–4 Ptac-qrr4). Error bars indicate one standard deviation from the mean of triplicate measurements. The standard deviation is ±25.1 for Qrr4 levels in the Δqrr1–4 Ptac-qrr4 V. cholerae strain. (C) Western blot showing HapR protein levels in the same strains as (B) at LCD (top panel, OD₆₀₀=0.1) and at intermediate cell density (bottom panel, OD₆₀₀=0.4).