PhcQ mainly contributes to the regulation of quorum sensing‐dependent genes, in which PhcR is partially involved, 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 via the methyltransferase PhcB and senses the chemical through the sensor histidine kinase PhcS. This leads to functionalization of the LysR family transcriptional regulator PhcA, regulating QS‐dependent genes responsible for the QS‐dependent phenotypes including virulence. The phc operon consists of phcB, phcS, phcR, and phcQ, with the latter two encoding regulator proteins with a receiver domain and a histidine kinase domain and with a receiver domain, respectively. To elucidate the function of PhcR and PhcQ in the regulation of QS‐dependent genes, we generated phcR‐deletion and phcQ‐deletion mutants. Though the QS‐dependent phenotypes of the phcR‐deletion mutant were largely unchanged, deletion of phcQ led to a significant change in the QS‐dependent phenotypes. Transcriptome analysis coupled with quantitative reverse transcription‐PCR and RNA‐sequencing revealed that phcB, phcK, and phcA in the phcR‐deletion and phcQ‐deletion mutants were expressed at similar levels as in strain OE1‐1. Compared with strain OE1‐1, expression of 22.9% and 26.4% of positively and negatively QS‐dependent genes, respectively, was significantly altered in the phcR‐deletion mutant. However, expression of 96.8% and 66.9% of positively and negatively QS‐dependent genes, respectively, was significantly altered in the phcQ‐deletion mutant. Furthermore, a strong positive correlation of expression of these QS‐dependent genes was observed between the phcQ‐deletion and phcA‐deletion mutants. Our results indicate that PhcQ mainly contributes to the regulation of QS‐dependent genes, in which PhcR is partially involved.


| INTRODUC TI ON
Quorum sensing (QS) allows bacterial cells to communicate for the cooperative regulation of physiological processes coordinating various bacterial community activities (Ham, 2013;Waters & Bassler, 2005).
To recognize their own populations, bacterial cells produce and secrete QS signals, which are small, diffusible molecules. Bacteria monitor QS signals to track changes in their cell numbers and to activate QS for the synchronous control of the expression of genes beneficial for vigorous replication and adaptation to environmental conditions, including virulence, such as the formation of biofilms and the production of virulence factors (Galloway et al., 2011;Rutherford & Bassler, 2012).
Ralstonia has developed a genus-specific QS system consisting of Phc QS cascade regulatory elements that respond to a unique fatty acid derivative signal (Flavier et al., 1997). Each strain of the RSSC produces either methyl 3-hydroxypalmitate (3-OH PAME) or methyl 3-hydroxymyristate (3-OH MAME) as a QS signal (Flavier et al., 1997;Kai et al., 2015;Schell, 2000;Ujita et al., 2019). RSSC strains synthesize the QS signal by the methyltransferase PhcB and sense the chemical through the sensor histidine kinase PhcS, activating QS (Genin & Denny, 2012;Kai et al., 2015;Schell, 2000;Ujita et al., 2019). The sensor histidine kinase PhcK is required for full expression of phcA, which encodes the LysR family transcriptional regulator PhcA, independently of QS signal production of PhcB (Senuma et al., 2020). In the active state of QS, PhcA activated through QS signal sensing of PhcS regulates QS-dependent genes responsible for QS-dependent phenotypes including virulence (Genin & Denny, 2012). This process leads to the induction of the production of ralfuranones, which are aryl-furanone secondary metabolites, and the major exopolysaccharide EPS I, which is involved in the virulence of RSSC strains (Genin & Denny, 2012;Kai et al., 2014Kai et al., , 2015Pauly et al., 2013;Schell, 2000;Wackler et al., 2011). Furthermore, these secondary metabolites are also associated with the feedback loop of QS-dependent gene regulation (Hayashi et al., 2019b;Mori et al., 2018).
The phc operon consists of phcB and phcS, which encode Phc QS cascade regulatory elements, along with phcR and phcQ Salanoubat et al., 2002;Tang et al., 2020). PhcR is an intracellular soluble two-component protein composed of a sensor histidine kinase domain and a receiver domain without a DNAbinding domain. PhcQ is an intracellular soluble regulator containing a receiver domain without a DNA-binding site. Schell (2000) proposed a model for the QS signalling pathway of the phylotype IIA strain AW1 of R. solanacearum, which produces 3-OH PAME as the QS signal (Flavier et al., 1997). At low levels of 3-OH PAME, the regulator protein PhcR may interact with PhcA, leading to PhcA dysfunction.
At the threshold concentration of 3-OH PAME, the sensor histidine kinase PhcS senses 3-OH PAME and carries out the phosphorylation by itself. This event may induce the ability of PhcS to phosphorylate the cognate PhcR, resulting in functional PhcA. Furthermore, Tang et al. (2020) demonstrated that PhcQ is involved in the dynamics of activation of PhcA in response to the bacterial density of phylotype I strain GMI1000 of R. pseudosolanacearum, which produces 3-OH MAME as the QS signal . However, the effects of these regulator proteins on the transcription of QS-dependent genes have not been validated experimentally.
In this study, we aimed to elucidate the function of PhcR and PhcQ in the regulation of QS-dependent genes. To achieve this goal, we first created phcR-deletion (ΔphcR) and phcQ-deletion (ΔphcQ) mutants from the phylotype I strain OE1-1 of R. pseudosolanacearum, which produces 3-OH MAME as the QS signal , and analysed their QS-dependent phenotypes as well as their virulence on tomato plants. We then analysed the transcriptomes of R. pseudosolanacearum strains using quantitative reverse transcription-PCR (RT-qPCR) and RNA-sequencing (RNA-seq).  Figure S1a) as well as PhcQ ( Figure S1b) showed that the 34 strains were divided into four clades, consistent with their phylotypes.

| Deletion of phcQ, but not phcR, led to significant changes in QS-dependent phenotypes
We created ΔphcR (Table 1) and ΔphcQ (Table 1) mutants of strain OE1-1 and analysed their QS-dependent phenotypes. We first assayed biofilm formation with crystal violet staining of R. pseudosolanacearum strains grown in quarter-strength M63 medium. The ΔphcR mutant exhibited slightly less biofilm formation than wildtype strain OE1-1 ( Figure 1a). The ΔphcQ mutant produced significantly less biofilm than strain OE1-1, similar to the ΔphcB mutant (Table 1; Kai et al., 2015) and the ΔphcA mutant (  Figure 1b) and more ralfuranone A (Figure 1c), which is one of the ralfuranones, than strain OE1-1, whereas the phcQ deletion led to significantly reduced production of EPS I (p < .05, t test; Figure 1b) and ralfuranone A (p < .05, t test; Figure 1c), similar to the effects of phcB and phcA deletions. Compared with strain OE1-1, the ΔphcR mutant, when grown on quarter-strength M63 medium solidified with 0.25% agar, exhibited slightly enhanced swimming motility ( Figure 1d). The swimming motility of the ΔphcQ mutant was significantly greater than that of strain OE1-1, similar to the ΔphcB and ΔphcA mutants (p < .05, t test).
The induced expression of two genes-epsB, which is part of the eps operon and required for EPS I biosynthesis (Huang & Schell, 1995), and ralA, which encodes a ralfuranone synthase (Kai et al., 2014;Wackler et al., 2011)-is dependent on QS. In addition, the expression of the flagellar motility-related gene fliC, which encodes flagellin, is suppressed in the active state of QS (Tans-Kersten et al., 2001). To analyse expression levels of these genes in R. pseudosolanacearum strains grown in quarter-strength M63 medium until OD 600 = 0.3, we conducted RT-qPCR assays. Expression levels of ralA and epsB in the ΔphcQ mutant but not in the ΔphcR mutant were significantly lower than those in OE1-1 (p < .05, t test; Figure 2). In contrast, fliC was more highly expressed in the ΔphcQ mutant, but not in the ΔphcR mutant, compared with the OE1-1 strain (p < .05, t test).

| Transformation of native PhcQ recovered QSdependent phenotypes of the ΔphcQ mutant
We next transformed the ΔphcQ mutant with pUC18-mini-Tn7T-Gm-phcQ (Table 1) harbouring the native phcQ gene fused with the promoter of the phcBSRQ operon  to generate the complemented ΔphcQ mutant strain phcQ-comp (Table 1).
Transformation of the ΔphcQ mutant with the pUC18-mini-Tn7T-Gm-phcQ construct led to enhanced biofilm formation ( Figure 1a) and production of EPS I ( Figure 1b) and ralfuranone A ( Figure 1c) and reduced swimming motility (Figure 1d).

| Deletion of phcQ, but not phcR, led to a loss of bacterial virulence
To investigate the effects of phcR and phcQ on the virulence of strain OE1-1, we inoculated 5-week-old tomato plants with R. pseudosolanacearum strains by the root-dip method and then assayed the population dynamics and behaviour of these strains in the tomato plants as well as disease development. The population of the ΔphcQ mutant at 3 days after inoculation (DAI) was significantly smaller than that of OE1-1 and the ΔphcR mutant, similar to ΔphcB and ΔphcA mutants (p < .05, t test; Figure 3a). In a plate-printing

| Deletion of phcR or phcQ did not influence the regulation of QS-related genes
The

| RNA-seq transcriptome analysis of R. pseudosolanacearum strains
The deletion of phcR or phcQ did not influence the regulation of QS-related genes ( Figure 2). To analyse the effects of phcR and phcQ deletion on the regulation of QS-dependent genes, we performed an RNA-seq transcriptome analysis of R. pseudosolanacearum strains grown in quarter-strength M63 medium until OD 600 = 0.3.
Mapping of RNA-seq reads of the OE1-1 strain to the GMI1000 genome (Salanoubat et al., 2002) resulted in the identification of 4,437 protein-coding transcripts (Table S2). To extract genes with significant expression changes, the following thresholds were applied: q < .05  Table S3). Among the positively PhcR-regulated genes, 79 genes, including norB, lecM, and xpsR, were positively QS-dependent genes ( Figure 4a; Table S3). Among the F I G U R E 2 Expression levels of the quorum sensing (QS)-related genes phcB, phcK, phcA, phcR, and phcQ, the positively QS-dependent genes ralA, epsB, and lecM, and the negatively QS-dependent gene fliC in Ralstonia pseudosolanacearum strain OE1-1 and phcA-deletion (ΔA), phcR-deletion (ΔR), phcQ-deletion (ΔQ), and phcB-deletion (ΔB) mutants grown in quarter-strength M63 medium until OD 600 = 0.3, as determined by quantitative reverse transcription-PCR. Two replicate experiments conducted using independent samples with eight technical replicates per experiment produced similar results. Results of a single representative sample are provided. Bars indicate standard errors.
Asterisks indicate values significantly different from those of OE1-1 (p < .05, t test) negatively PhcR-regulated genes, 43 genes, including some flagellin biosynthesis-related genes and chemotaxis-related genes, were negatively QS-dependent genes ( Figure 4b; Table S3). The log ( Table S4) and 197 negatively PhcQregulated genes were significantly up-regulated relative to strain OE1-1 (Figure 4b; Table S4). Among the positively PhcQ-regulated genes, 334 genes were positively QS-dependent genes ( Figure 4a; Table S4). Among the negatively PhcQ-regulated genes, 109 genes were negatively QS-dependent genes ( Figure 4b;   Tang et al. (2020) demonstrated that PhcQ contributes to the synthesis of 3-OH MAME by strain GMI1000. In the active state of QS, expression of lecM, which encodes the lectin LecM, is induced and  LecM affects the activation of QS through regulating the stability of secreted 3-OH MAME (Hayashi et al., 2019a). Deletion of phcR or phcQ did not influence the regulation of QS-related genes including phcB (Figure 2). We assayed 3-OH MAME content purified from R. pseudosolanacearum strains. The lecM-M mutant exhibited a significantly lower 3-OH MAME content than strain OE1-1, similar to the ΔphcQ mutant (p < .05, Figure 5). Though the ΔphcR mutant exhibited significantly lower 3-OH MAME content compared to strain OE1-1, 3-OH MAME content in the ΔphcR mutant was higher than that in other mutants, including the ΔphcA mutant.

| Deletion of phcQ led to a significant reduction in 3-OH MAME content
To analyse the expression levels of lecM in R. pseudosolanacearum strains grown in quarter-strength M63 medium until OD 600 = 0.3, we conducted RT-qPCR assays. The expression levels of lecM in the ΔphcQ and ΔphcA mutants was significantly lower than in strain OE1-1 (p < .05, t test; Figure 2). Though the expression level of lecM in the ΔphcR mutant was significantly lower than in strain OE1-1, the lecM expression level in the ΔphcR mutant was higher than in the ΔphcA mutant.

| Deletion of phcR led to a slight change in expression levels of QS-dependent genes at lower bacterial density
In the inactive state of QS, PhcR reportedly inhibits PhcA function in strain AW1 (Schell, 2000). To analyse the influence of phcR deletion on expression levels of QS-dependent genes in the inactive state of QS, using RNA isolated from the ΔphcR mutant and strain OE1-1 grown in quarter-strength M63 medium until OD 600 = 0.01, we conducted RT-qPCR assays to assess relative expression levels of lecM, ralA, epsB, and fliC in the ΔphcR mutant compared to strain OE1-1.
The expression levels of the positively QS-dependent genes lecM, ralA, and epsB were slightly but significantly lower in the ΔphcR mutant than in strain OE1-1 (p < .05, t test; Figure 6). In addition, the expression level of the negatively QS-dependent fliC was significantly higher in the ΔphcR mutant than in strain OE1-1 (p < .05, t test).
2.9 | Exogenous 3-OH MAME application did not lead to a change in the QS-dependent phenotypes of ΔphcR and ΔphcQ mutants QS activity in strain OE1-1 is dependent on the exogenous levels of 3-OH MAME (Hayashi et al., 2019b;Kai et al., 2015). We examined the influence of exogenous 3-OH MAME application on QSdependent phenotypes, biofilm formation, EPS I production, and swimming motility of these mutants. Exogenous application of 3-OH MAME at a concentration of 0.1 μM enhanced biofilm formation and EPS I production and reduced swimming motility of strain OE1-1 and the ΔphcB mutant (p < .05, t test; Figure 7a). However, exogenous 3-OH MAME application had no effect on these QS-dependent phenotypes of ΔphcR and ΔphcQ mutants.
To analyse the influence of exogenous 3-OH MAME application on the expression levels of the QS-dependent genes lecM, ralA, epsB, and fliC, we conducted RT-qPCR assays using R. pseudosolanacearum strains grown in quarter-strength M63 medium containing 3-OH MAME at a concentration of 0.1 μM until OD 600 = 0.3. Exogenous 3-OH MAME application significantly enhanced the expression levels of lecM, ralA, and epsB and significantly reduced the expression levels of fliC in strain OE1-1 and the ΔphcB mutant but not in the ΔphcQ mutant (p < .05, t test; Figure 7b). In contrast, in the ΔphcR mutant, exogenous 3-OH MAME application slightly but significantly enhanced the expression levels of lecM and epsB, but not ralA, and slightly but significantly reduced the expression levels of fliC (p < .05, t test). which encodes diaminopimelate decarboxylase, required for staphyloferrin B production, is negatively regulated by PhcA (Bhatt & Denny, 2004). To analyse the influence of PhcR and PhcQ on siderophore-mediated iron acquisition activity, we measured the siderophore-mediated iron acquisition activity of R. pseudosolanacearum strains. The ΔphcB and ΔphcQ mutants had significantly higher siderophore-mediated iron acquisition activity than wildtype strain OE1-1 (p < .05, t test; Figure 8a). The deletion of phcK or phcA significantly enhanced siderophore-mediated iron acquisition activity to a greater extent than did the deletion of phcB or phcQ (p < .05, t test). On the other hand, the ΔphcR mutant had significantly lower siderophore-mediated iron acquisition activity than wild-type strain OE1-1.

| D ISCUSS I ON
Balancing selection is a type of positive selection that favours the maintenance of a high genetic diversity within a given population and functions through spatial/temporal heterogeneity as well as overdominant selection and frequency-dependent selection (Charlesworth, 2006;Hedrick, 2012;Stoeckel et al., 2012). Under balancing selection, the phcBSRQ operon plays a significant role in virulence of RSSC strains. (Castillo & Agathos, 2019). PhcS is subject to strong selection from the plant host (Guidot et al., 2014). Based on the deduced PhcB and PhcS amino acid sequences, RSSC strains are divided into two groups according to their QS signal types . In contrast, the phylogenetic analysis based on the deduced amino acid sequences of PhcR and PhcQ indicates that RSSC strains can be divided according to their phylotypes, similar to that of PhcA (Senuma et al., 2020). Overall, it is thought that extracellular communications through the production and sensing of QS signals reveal the unique evolution among RSSC strains independently of their phylotypes. The following intracellular signalling is under balancing ΔphcB or ΔphcA. We previously demonstrated that the putative sensor histidine kinase PhcK is required for full phcA expression independently of 3-OH MAME content (Senuma et al., 2020). However, the deletion of phcQ had no effect on the expression level of phcA. In Strain AW1 produces staphyloferrin B as a siderophore, and ssd, which is negatively regulated by PhcA, is required for staphyloferrin B production and siderophore-mediated iron acquisition activity (Bhatt & Denny, 2004). The strains used in the present study showed siderophore-mediated iron acquisition activity in the order ΔphcA ≈ ΔphcK < ΔphcB ≈ ΔphcQ < OE1-1 < ΔphcR. PhcA regulated not only ssd but also RSc1271, RSc2918, RSc2919, RSc2920, RSp0419, and Ujita et al., 2019). PhcR  and PhcQ (Tang et al., 2020) are regulator proteins with a receiver domain without a DNA-binding site. Single-receiver-domain proteins play roles as phospho-relays or phospho-sinks (Feldheim et al., 2018).
Considering the results of the present study, we propose the fol-

| Generation of the phcR-deletion and phcQdeletion mutants
The oligonucleotide primers used for construction of recombinant plasmids are listed in Table S6. The fragments delta-R-1 and delta-R-2 for generation of the ΔphcR mutant and the fragments delta-Q-1 and delta-Q-2 for generation of the ΔphcQ mutant were amplified by PCR from the genomic DNA of strain OE1-1. The fragments delta-R and delta-Q amplified using delta-R-1 and delta-R-2 were ligated into a pK18mobsacB vector (Kvitko & Collmer, 2011) to produce recombinant plasmids pdelta-phcR and pdelta-phcQ, respectively. Each plasmid was electroporated into OE1-1 competent cells, which were prepared as previously described by Mori et al. (2016). Kanamycinsensitive, sucrose-resistant recombinants ΔphcR and ΔphcQ were then selected.

| Differential gene expression analysis
Statistical analysis of the RNA-seq data was performed in the R environment. Genes with zero counts in at least one OE1-1 sample were excluded. RNA-seq read counts of the remaining genes were normalized using the function calcNormFactors (trimmed mean of M value normalization) in the package edgeR (Robinson et al., 2010).
To extract genes with significant expression changes, the following thresholds were applied: q < .05 and |log(FC)| ≥ 2. The false discovery rate (q value) was calculated in edgeR from Benjamini-Hochbergcorrected p values. Hierarchical clustering of all normalized mean expression values based on their relative expression (counts per million) was performed using Cluster v. 3.0 software (de Hoon et al., 2004). The average value of three replicates per strain was used.
Heatmaps were created in TreeView (Eisen et al., 1998). kit. An RT-qPCR assay with gene-specific primers (Table S7) was carried out using the SYBR GreenER qPCR Reagent system (Invitrogen) on a 7300 Real-Time PCR system (Applied Biosystems) as previously described (Hayashi et al., 2019a). All values were nor-

| QS-dependent phenotypes
We examined the in vitro biofilm formation of R. pseudosolanacearum strains grown without shaking in quarter-strength M63 as previously described (Mori et al., 2016). To determine the influence of exogenous 3-OH MAME application on biofilm formation, R. pseudosolanacearum strains were grown without shaking in quarter-strength M63 containing 0.1 μM 3-OH MAME as previously described (Senuma et al., 2020

| 3-OH MAME contents produced by R. pseudosolanacearum strains
Strains of R. pseudosolanacearum grown in B medium at 30 °C for 4-6 hr were diluted to an OD 600 of 1.0 with new medium. The cell suspension at a volume of 50 μl was pipetted onto a BG agar plate (90 mm, 25 ml; Kai et al., 2015), and the plate was incubated for 24 hr at 30 °C. The BG agar was cut into small pieces and soaked in ethyl acetate at a volume of 50 ml for 2 hr twice. The combined extracts were dried over Na 2 SO 4 and concentrated. The residue was dissolved in acetonitrile (100 ml) and subjected to LC-MS analysis.

| Siderophore-mediated iron acquisition activity
The siderophore-mediated iron acquisition activity of R. pseudosolanacearum strains was analysed using a method modified from Wali et al. (2015). R. pseudosolanacearum strains were incubated in PY medium (5 g/L polypeptone and 2 g/L yeast extract) for 18 hr at 30 °C and adjusted to a concentration of 2 × 10 9 cfu/ml with 0.1 M PIPES buffer (pH 6.5). After 6 hr of incubation, each culture was filtered through a 0.2μm pore filter. Next, 100 µl of a culture, or PIPES buffer alone as a reference, was added to 100 μl
We assessed populations of R. pseudosolanacearum strains in inoculated tomato roots according to their observed growth on Hara-Ono medium as described by Hayashi et al. (2019a). For each bacterial strain, three replicate experiments conducted using independent samples with five technical replicates per experiment produced similar results. Results of a single representative sample are provided.
The behaviour of R. pseudosolanacearum strains in tomato plants inoculated by the root-dip method was assessed as described by Hayashi et al. (2019a). A sample from each cut site (Figure 3b) was pressed onto Hara-Ono medium. Twelve plants were analysed per trial, and each assay comprised five successive trials.

| Statistical analysis
The means of all assays were analysed for significant differences between R. pseudosolanacearum strains by Student's t test in Microsoft Excel.

ACK N OWLED G EM ENTS
We gratefully acknowledge the experimental assistance of Nobuko Sato. This work was supported by JSPS KAKENHI (grant no.