The transcription regulator ChpA affects the global transcriptome including quorum sensing-dependent genes in Ralstonia pseudosolanacearum strain OE1-1

Abstract

The gram-negative plant-pathogenic β-proteobacterium Ralstonia pseudosolanacearum strain OE1-1 produces methyl 3-hydroxymyristate as a quorum sensing (QS) signal through methyltransferase PhcB and senses the chemical via the sensor histidine kinase PhcS. This leads to activation of the LysR family transcription regulator PhcA, which regulates the genes (QS-dependent genes) responsible for QS-dependent phenotypes, including virulence. The transcription regulator ChpA, which possesses a response regulator receiver domain and also a hybrid sensor histidine kinase/response regulator phosphore-acceptor domain but lacks a DNA-binding domain, is reportedly involved in QS-dependent biofilm formation and virulence of R. pseudosolanacearum strain GMI1000. To explore the function of ChpA in QS of OE1-1, we generated a chpA-deletion mutant (ΔchpA) and revealed that the chpA deletion leads to significantly altered QS-dependent phenotypes. Furthermore, ΔchpA exhibited a loss in its infectivity in xylem vessels of tomato plant roots, losing virulence on tomato plants, similar to the phcA-deletion mutant (ΔphcA). Transcriptome analysis showed that the transcript levels of phcB, phcQ, phcR, and phcA in ΔchpA were comparable to those in OE1-1. However, the transcript levels of 89.9% and 88.9% of positively and negatively QS-dependent genes, respectively, were significantly altered in ΔchpA compared with OE1-1. Furthermore, the transcript levels of these genes in ΔchpA were positively correlated with those in ΔphcA. Together, our results suggest that ChpA is involved in the regulation of these QS-dependent genes, thereby contributing to the behaviour in host plant roots and virulence of OE1-1.

Keywords: Ralstonia pseudosolanacearum; ChpA; quorum sensing; virulence.

FIGURE 1

Biofilm formation (a), production of the major exopolysaccharide EPS I (b), swimming motility (c), and the in vitro growth rate (d) of Ralstonia pseudosolanacearum strain OE1‐1, the phcA‐deletion (ΔphcA) and chpA‐deletion (ΔchpA) mutants, and ΔchpA transformed with native chpA (chpA‐comp). (a) Biofilm formation of R. pseudosolanacearum incubated in quarter‐strength M63 medium was quantified by measuring absorbance at 550 nm (A₅₅₀) and normalized to the number of cells (optical density at 600 nm, OD₆₀₀). Three replicate experiments were conducted using independent samples, with seven technical replicates per experiment. (b) R. pseudosolanacearum strains were incubated on quarter‐strength M63 medium solidified with 0.25% wt/vol agar. EPS I content of the supernatant was quantified by ELISA. Three replicate experiments were conducted using independent samples, with seven technical replicates per experiment. (c) R. pseudosolanacearum strains were grown on quarter‐strength M63 medium solidified with 0.25% wt/vol agar. Three replicate experiments were conducted using independent samples, with five technical replicates per experiment. (d) Overnight cultures of R. pseudosolanacearum strains were diluted to OD₆₀₀ = 0.01 with quarter‐strength M63 and grown with shaking. The in vitro growth rate of bacterial strains was quantified based on OD₆₀₀ of the bacterial culture. Three replicate experiments were conducted using independent samples, with three technical replicates per experiment. Bars indicate standard error. Means were analysed for significant differences between R. pseudosolanacearum strains by analysis of variance followed by Tukey–Kramer's honestly significant difference test. Statistically significant differences are indicated by different lowercase letters (p < 0.05).

FIGURE 2

Gene transcript levels in wild‐type and mutant strains. (a) Transcript levels of chpA in Ralstonia pseudosolanacearum strain OE1‐1 and the phcB‐deletion (ΔphcB), phcQ‐deletion (ΔphcQ), phcR‐deletion (ΔphcR), phcK‐deletion (ΔphcK), and phcA‐deletion (ΔphcA) mutants. (b) Transcript levels of quorum sensing (QS)‐related genes phcB, phcQ, phcR, phcK, and phcA, and QS‐dependent genes epsB, ralA, lecM, and fliC, in OE1‐1 and the chpA‐deletion mutant (ΔchpA). Gene transcript levels were normalized to that of rpoD. The experiments were performed with three biological replicates and two technical replicates. Bars indicate standard errors. Asterisks indicate a significant difference from OE1‐1 (p < 0.05, t test).

FIGURE 3

Virulence of Ralstonia pseudosolanacearum strains on tomato plants (a), microscopic observation of roots of 4‐day‐old tomato seedlings co‐incubated with R. pseudosolanacearum strains (b), and percentages of cell wall‐degraded cortex cells infected with R. pseudosolanacearum strains in the meristematic zone in roots of 4‐day‐old tomato seedlings after 48 h of co‐incubation (HOI) with R. pseudosolanacearum strains (c). (a) Eight‐week‐old tomato plants were inoculated with R. pseudosolanacearum strain OE1‐1, the phcA‐deletion (ΔphcA) and chpA‐deletion (ΔchpA) mutants, and ΔchpA transformed with native chpA (chpA‐comp). Plants were rated according to the following disease index scale: 0, no wilting; 1, 1%–25% wilting; 2, 26%–50% wilting; 3, 51%–75% wilting; 4, 76%–99% wilting; 5, dead. For each bacterial strain, three independent groups were tested, with 12 technical replicates per group. Bars indicate standard error. Means were analysed for significant differences between R. pseudosolanacearum strains by analysis of variance followed by Tukey–Kramer's honestly significant difference test. Statistically significant differences are indicated by different lowercase letters (p < 0.05). (b) Mock and 24, 48, and 120 HOI with R. pseudosolanacearum strains OE1‐1, ΔphcA, and ΔchpA. Longitudinal semithin resin sections were stained with toluidine blue. Red arrows indicate bacterial cells. Bars indicate 20 μm. e, epidermal cell; c, cortical cell; en, endodermis cell; pe, pericycle cell; v, xylem vessel; mBF, mushroom‐shaped biofilm. (c) Micrographs of toluidine blue‐dyed serial semi‐thin sections (800 nm thick) from the meristematic zone in roots were observed. The experiment was repeated five times, each with 100 technical replicates. Bars indicate standard error. Means were analysed for significant differences between R. pseudosolanacearum strains by analysis of variance followed by Tukey–Kramer's honestly significant difference test. Statistically significant differences are indicated by different lowercase letters (p < 0.05).

FIGURE 4

RNA‐sequencing transcriptome analysis of Ralstonia pseudosolanacearum strains grown in quarter‐strength M63 medium until OD₆₀₀ = 0.3. (a) Numbers of genes with transcript‐level log₂(fold change) ≤ −2 in the chpA‐deletion (ΔchpA) and phcA‐deletion (ΔphcA) mutants relative to their transcript levels in OE1‐1 (q < 0.05). (b) Number of genes with transcript‐level log₂(fold change) ≥ 2 in ΔchpA and ΔphcA relative to their transcript levels in OE1‐1 (q < 0.05). (c) Correlation of transcript levels of ChpA‐dependent genes among quorum sensing (QS)‐dependent genes between R. pseudosolanacearum mutants: ΔchpA versus ΔphcA.

FIGURE 5

Global effect of ChpA on gene expression in Ralstonia pseudosolanacearum. A principal component analysis plot for the transcriptome data of positively quorum sensing (QS)‐dependent genes (a) and all genes (b) in OE1‐1 and phcA‐deletion (ΔphcA), phcR‐deletion (ΔphcR), phcQ‐deletion (ΔphcQ), and chpA‐deletion (ΔchpA) mutants. (c) Hierarchical clustering and heatmap analysis of relative transcript levels of genes that were differentially expressed between at least two genotypes (OE1‐1, ΔphcA, and ΔchpA) based on RNA‐sequencing data with the samples grown in quarter‐strength M63 medium until OD₆₀₀ = 0.3. Values are averages of three replicates per strain. Hierarchical clustering by the complete linkage method was applied to z‐scores calculated with normalized log₂(fold change) values of all normalized mean expression values (counts per million) using the hclust R package. The heatmap was created with the R package pheatmap. (d) GO terms enriched in each cluster shown in (c).

FIGURE 6

Siderophore‐mediated iron acquisition activity of Ralstonia pseudosolanacearum strain OE1‐1 and the phcA‐deletion (ΔphcA) and chpA‐deletion (ΔchpA) mutants grown in PY medium. Three replicate experiments were conducted using independent samples, with eight technical replicates per assay. Means were analysed for significant differences between R. pseudosolanacearum strains by analysis of variance followed by Tukey–Kramer's honestly significant difference test. Statistically significant differences are indicated by different lowercase letters (p < 0.05).

FIGURE 7

Hierarchical clustering of relative transcript levels of genes involved in twitching motility governed by type IV pilus genes in Ralstonia pseudosolanacearum strain OE1‐1 and the phcA‐deletion (ΔphcA) and chpA‐deletion (ΔchpA) mutants grown in quarter‐strength M63 medium until OD₆₀₀ = 0.3. Based on RNA‐sequencing transcriptome data for R. pseudosolanacearum strains, fragments per kilobase of open reading frame per million fragments mapped values from R. pseudosolanacearum strains OE1‐1, ΔphcA, and ΔchpA were normalized prior to analysis of differentially expressed genes. Hierarchical clustering of all normalized mean expression values (counts per million) was performed using Cluster v. 3.0 software. Heatmaps were created with TreeView.

FIGURE 8

(a) Swimming motility of Ralstonia pseudosolanacearum strain OE1‐1 and the phcA‐deletion (ΔphcA), chpA‐deletion (ΔchpA), and fliC‐deletion (ΔfliC) mutants grown on modified CPG medium containing 0.3% wt/vol agar and (b) hierarchical clustering of relative transcript levels of flagellar motility‐related genes in OE1‐1, ΔphcA, and ΔchpA grown in quarter‐strength M63 medium until OD₆₀₀ = 0.3. (a) The experiment was repeated three times, each with five technical replicates. Bars indicate standard errors. Means were analysed for significant differences between R. pseudosolanacearum strains by analysis of variance followed by Tukey–Kramer's honestly significant difference test. Statistically significant differences are indicated by different lowercase letters (p < 0.05). (b) Based on RNA‐sequencing transcriptome data for R. pseudosolanacearum strains, fragments per kilobase of open reading frame per million fragments mapped values from OE1‐1, ΔphcA, and ΔchpA were normalized prior to analysis of differentially expressed genes. The average value of three replicates per strain was used. Hierarchical clustering of all normalized mean expression values (counts per million) was performed using Cluster v. 3.0 software. Heatmaps were created with TreeView.