RpoN1 and RpoN2 play different regulatory roles in virulence traits, flagellar biosynthesis, and basal metabolism in Xanthomonas campestris

Abstract Homologous regulatory factors are widely present in bacteria, but whether homologous regulators synergistically or differentially regulate different biological functions remains mostly unknown. Here, we report that the homologous regulators RpoN1 and RpoN2 of the plant pathogen Xanthomonas campestris pv. campestris (Xcc) play different regulatory roles with respect to virulence traits, flagellar biosynthesis, and basal metabolism. RpoN2 directly regulated Xcc fliC and fliQ to modulate flagellar synthesis in X. campestris, thus affecting the swimming motility of X. campestris. Mutation of rpoN2 resulted in reduced production of biofilms and extracellular polysaccharides in Xcc. These defects may together cause reduced virulence of the rpoN2 mutant against the host plant. Moreover, we demonstrated that RpoN1 could regulate branched‐chain fatty acid production and modulate the synthesis of diffusible signal factor family quorum sensing signals. Although RpoN1 and RpoN2 are homologues, the regulatory roles and biological functions of these proteins were not interchangeable. Overall, our report provides new insights into the two different molecular roles that form the basis for the transcriptional specialization of RpoN homologues.


| INTRODUC TI ON
Homologous regulatory factors are widely present in bacterial genomes. These homologous regulators may display functional redundancy or different biological functions, and some homologues even have no biological function, for example the two homologous ArsR regulators of Pseudomonas putida KT2440 bind arsenite with similar affinities (Fernandez et al., 2016). AgaR2 and AgaR1 are two novel regulators that are specific for GalN/GalNAc catabolism and have been assigned distinct roles in bacterial infection . NagC and Mlc, homologous members of the ROK (repressors, open reading frames [ORFs] and kinases) family of proteins with almost identical helix-turn-helix DNA-binding motifs, specifically regulate genes for the transport and utilization of N-acetylglucosamine and glucose (Brechemier-Baey et al., 2015). Two homologous members of the cyclic-AMP receptor protein (CRP) family of transcription factors (MSMEG_0539 and MSMEG_6189) show differences in cAMP binding affinity, trypsin sensitivity, and binding to a CRP site in Mycobacterium smegmatis (Sharma et al., 2014). Three homologous LysR-type transcriptional regulators (MetR, CysR, and HomR) control sulphur amino acid supply in Streptococcus mutans (Sperandio et al., 2010), while rpoN1 and rpoN2 of Xanthomonas citri subsp. citri play different roles in virulence, nutrient utilization, and cell motility (Gicharu et al., 2016). However, some homologous transcription factors still have unknown biological and transcriptional functions.
Here, we use Xanthomonas campestris pv. campestris (Xcc) sigma factor 54 (σ 54 ) proteins (XCC RpoN1 and XCC RpoN1) as a model to illustrate the characteristics of homologous transcription factors in a single bacterial genome.
Sigma factor 54, alternatively named RpoN, is also required for the initiation of transcription at specific DNA sequences (Hirschman et al., 1985). The σ factor facilitates transcription at specific DNA sequences by binding to the core RNA polymerase (RNAP) to form the σ-RNAP holoenzyme, recognizing and binding to a specific DNA sequence adjacent to the transcription start site, called the promoter element, and opening the double-stranded DNA to initiate transcription (Barrios et al., 1999;Doucleff et al., 2005;Wiesler et al., 2012;Schaefer et al., 2015).
RpoN can regulate the transcription and expression of many genes, and is historically known for its role in nitrogen assimilation (Kohler et al., 1989;Rajeev et al., 2015). This protein has been shown to be involved in the regulation of other important lifestyle-associated functions in bacteria, such as in the regulation of the type III secretion system (Lee et al., 2016) and type VI secretion system (Dong and Mekalanos, 2012;Sana et al., 2013). RpoN also affects the growth, virulence, motility, and biofilm formation of bacteria (Hao et al., 2013;Hayrapetyan et al., 2015;Ray et al., 2015). In addition, RpoN can affect the assembly of bacterial flagella (Yang et al., 2009;Schulz et al., 2012), quorum sensing (QS) (Cai et al., 2015), amino acid utilization, the regulation of catalase expression at the transcriptional level (Diep et al., 2015), and carbohydrate metabolism (Stevens et al., 2010;Hayrapetyan et al., 2015).
Analysis of the complete genome sequences of several microorganisms has shown that the rpoN gene (encoding the σ 54 factor) is widely distributed among bacteria (Studholme et al., 2000;Buck, 2000a, 2000b;Dombrecht et al., 2002;Studholme and Dixon, 2003). A few organisms have two copies of rpoN, such as Ralstonia solanacearum (Ray et al., 2015), Burkholderia pseudomallei (Diep et al., 2015), Bradyrhizobium japonicum (Kullik et al., 1991), and X. campestris (Yang et al., 2009). In B. japonicum, these copies are highly similar and functionally interchangeable (Kullik et al., 1991). The genome of R. solanacearum contains two copies of rpoN that have different functions and are involved in twitching motility, natural competence, growth on nitrate, and virulence (Ray et al., 2015). In contrast, the four different σ 54 factors of Rhodobacter sphaeroides are not functionally interchangeable (Poggio et al., 2002), and this unusually high number of rpoN gene copies does not seem to be the result of recent duplication events because the products of these genes show degrees of similarity ranging from 48% to 87% (Poggio et al., 2006). However, why is it that among highly similar RpoNs some are functionally interchangeable and some are not? The molecular determinants that confer this specificity remain unknown.
A majority of bacterial species belonging to the genus Xanthomonas are plant pathogens. Among these species, X. campestris pv. campestris causes black rot disease in crucifers (Zhou et al., 2015a;Wang et al., 2018). Two rpoN homologues genes encoding σ 54 were also identified in the genome of X. campestris. To distinguish these two genes, the rpoN gene XCC2802, located in a phosphotransferase system (PTS), was named rpoN1 and the other (XCC1935), located in a large flagellar gene cluster, was named rpoN2 (Figure 1a,b) (da Silva et al., 2002;Yang et al., 2009). The flagellar biogenesis of X. campestris requires the alternative sigma factor RpoN2 (da Silva et al., 2002;Yang et al., 2009). However, the mechanism by which RpoN2 regulates flagellar biogenesis and the role of rpoN1 in X. campestris remain unknown. Whether these two different σ 54 factors of X. campestris are functionally interchangeable and the molecular determinants that may confer specificity on RpoN1 and RpoN2 have not yet been investigated.
In this study, we describe for the first time the distinct roles of RpoN1 and RpoN2 and their different regulatory functions in basal metabolism, flagellar biosynthesis, extracellular polysaccharide (EPS) formation, biofilm formation, and virulence in X. campestris.
Our study therefore reveals that these functions of the two homologous RpoN proteins are not redundant in X. campestris. Overall, the results of the present study improve the current understanding of RpoN-mediated regulation in X. campestris.

| Transcriptomic profiling of the rpoN deletion strains revealed the regulatory role of this gene in basal metabolism and flagellar synthesis
To investigate whether there are new genes/functions that are controlled by rpoN1 and rpoN2 in X. campestris, we performed transcriptomic analyses (RNA-Seq) of the wild-type and rpoN mutant strains. The examination of each sample was repeated twice. The results show that compared to the wild type, a total of 385 genes were significantly differentially altered at the transcriptional level in the ΔrpoN1 mutant strain, with 274 genes up-regulated and 111 genes down-regulated ( Figure 3a and Table S3). We also analysed and compared the transcriptomes of the ΔrpoN2 mutant strain and wild-type strain using RNA-Seq. Differential gene expression analysis showed that 283 genes were up-regulated and 179 genes were down-regulated in the ΔrpoN2 mutant compared with the wild-type strain, whereas 468 genes were up-regulated and 346 genes were down-regulated in the ΔrpoN1N2 double mutant compared with the wild-type strain (Figure 3a, Tables S4 and S5). We also compared the transcriptome profiles using Venn diagrams, which showed the overlap of 67 up-regulated and three down-regulated genes in different mutant backgrounds (Figure 3a). These results were further partly confirmed by quantitative reverse transcription PCR (RT-qPCR) analysis of eight randomly selected candidate genes ( Figure S1). Further bioinformatics analyses showed that the products of these differentially expressed genes (DEGs) belong to three major functional categories. In addition, each major category contains rather diverse subfunctional groups ( Figure S2a The function of rpoN1 and rpoN2 in flagellum-dependent motility was investigated. We then tested the swimming motility of the ΔrpoN1, ΔrpoN2, and ΔrpoN1N2 mutants on 0.3% semisolid agar plates. Compared with the wild-type strain, deletion of rpoN2 abolished swimming motility in X. campestris, suggesting that the motility was significantly impaired (Figure 4a,b). The ΔrpoN1 mutant remained as motile as the wild type. The motility defect phenotypes F I G U R E 4 Deletion of rpoN2 affected Xanthomonas campestris swimming ability and flagellar synthesis. (a) and (b) Assay of swimming motility of the wild-type Xc1; mutants ΔrpoN1, ΔrpoN2, and ΔrpoN1N2; and complemented strains ΔrpoN1/N1, ΔrpoN1/N2, ΔrpoN2/N1, ΔrpoN2/N2, ΔrpoN1N2/N1, and ΔrpoN1N2/N2. The swimming zones were recorded after bacterial growth for 4 days on semisolid plates at 28 °C. Error bars, means ± SD (n = 3). Different letters indicate significant differences between treatments with the least significant difference at p = .05. All experiments were repeated three times with similar results. (c) Observation of flagella using transmission electron microscopy of the ΔrpoN2 and ΔrpoN1N2 mutants could be restored to wildtype levels by introducing rpoN2 into ΔrpoN2 and ΔrpoN1N2, respectively. Moreover, overexpression of rpoN1 in the ΔrpoN2 and ΔrpoN1N2 mutants resulted in no substantial difference between the mutant and complemented strains of X. campestris in terms of swimming motility (Figure 4a,b). These results suggest that RpoN2, but not RpoN1, is required for the swimming motility of X. campestris.
To further demonstrate the biological function of RpoN, we performed electrophoretic mobility gel shift assays (EMSA) without a core RNAP (Cannon et al., 1993). The RpoN1 and RpoN2 proteins were then expressed in E. coli BL21 (DE3), and the N-terminal-

| RpoN2 is involved in biofilm formation and EPS production
Bacterial biofilms are involved in adaptation to complex environments and are used by pathogenic bacteria to colonize host cells (Tao et al., 2010). To test whether RpoN plays a role in biofilm formation, three independent assays were carried out. In the first assay, the frequently used crystal violet (CV) staining approach was applied.
The results showed that the ΔrpoN2 and ΔrpoN1N2 mutant strains exhibited 5.8-fold and 14.9-fold reductions, respectively, in biofilm formation on the polystyrene surface after staining with CV, compared with the wild type ( Figure 6a); however, inactivation of rpoN1 had no effect on biofilm formation (Figure 6a). These data reveal that RpoN2 is required for full biofilm formation.
EPS is a critical virulence factor for pathogenicity (Rai et al., 2012;Cai et al., 2017). Yu et al. found that RpoN (σ 54 ) is required for floc formation but not for EPS biosynthesis in a floc-forming Aquincola tertiaricarbonis strain . To further investigate the function of Xcc RpoN in EPS production, we examined EPS synthesis in the rpoN mutants. As shown in Figure 6b, the EPS production ability of the ΔrpoN2 and ΔrpoN1N2 mutants was significantly decreased, but EPS production by the ΔrpoN1 strain was not significantly different compared with that by the wild-type strain. These findings suggest that RpoN2 is required for full EPS production. showing that RpoN2 and RpoN1 directly regulate fliC. RpoN2 and RpoN1 (0, 1, 2, 4, or 8 μM) were added to reaction mixtures containing 50 ng of probe DNA, and the reaction mixtures were separated on polyacrylamide gels. (e) The effect of RpoN on fliC gene expression was measured by assessing the β-galactosidase activity of the fliC-lacZ transcriptional fusions in the Xc1 wild-type, ΔrpoN1, ΔrpoN2, and ΔrpoN1N2 strains. Error bars, means ± SD (n = 3). *p < .05, **p < .01, ****p < .0001, assessed by one-way analysis of variance. All experiments were repeated three times with similar results produced by the mutant strain ΔrpfCΔrpoN1 and double-mutant strain ΔrpfCΔrpoN1N2 were significantly higher than those produced by the Xc1 ΔrpfC strain (p < .001). These findings indicated that rpoN1 deletion increased the synthesis of DSF family signals in X. campestris, especially the amount of DSF.
Our data show that in-frame deletions of the rpoN2 coding sequence caused a significant decrease in bacterial virulence and in planta growth in a susceptible cabbage variety (B. oleraceae 'Jingfeng No. 1') (Figure 2), whereas the ΔrpoN1 mutant did not cause a substantial decrease in virulence or in planta growth. However, why does the rpoN2 mutation cause a decrease in X. campestris virulence? To understand the underlying mechanisms, the role of rpoN2 in X. campestris virulence was examined. We evaluated several pathogenicity-related virulence factors produced by the rpoN mutant strains. The activities of extracellular cellulase, amylase, and protease enzymes produced by the rpoN mutant strains were not significantly different from those of the enzymes produced by the wild-type strain Xc1 ( Figure S5). However, we evaluated EPS production in the rpoN mutant strains, and found that the ability of the ΔrpoN2 and ΔrpoN1N2 mutants to produce EPS was significantly decreased (Figure 6b). EPS act as a matrix material and is essential for biofilm formation in most gram-negative bacteria (Pratt and Kolter, 1999;Donlan, 2002;Singh et al., 2006). Does this finding indicate that the rpoN2 mutant also affects biofilm formation in X. campestris? We tested the biofilm formation ability of ΔrpoN1, ΔrpoN2, and ΔrpoN1N2, and found that ΔrpoN2 and ΔrpoN1N2 exhibited significantly reduced biofilm formation compared with the wild type ( Figure 6a). Bacterial biofilms are involved in adaptation to complex environments and are used by pathogenic bacteria to colonize host cells (Tao et al., 2010). Thus, we hypothesized that reduction in EPS and biofilm formation was one of the reasons why X. campestris exhibited impaired virulence against host plants. However, further investigation is required to understand these mechanisms in greater detail.
RpoN has been implicated in the regulation of flagellar and motility genes in many species. We also showed that the absence of native RpoN2 significantly reduced motility-associated phenotypes (Figure 4a-c). For many bacteria, motility is essential for survival, growth, virulence, biofilm formation, and intra/interspecies interactions and allows the bacteria to position themselves in appropriate locations at appropriate times (Nan and Zusman, 2016;Kilmury and Burrows, 2018). In prokaryotes, flagellation is essential for swimming motility. Flagellar biogenesis is a complex process that involves over 40 genes. The phytopathogen X. campestris possesses a single polar flagellum (Yang et al., 2009). RNA-Seq analysis revealed that the ΔrpoN2 and ΔrpoN1N2 mutations affected a group of genes associated with flagellar assembly, and 55.6% and 100% of these genes, respectively, were down-regulated (Tables S4 and S5). This result suggests that the ΔrpoN2 mutant could be disruptive to the flagellar system with subsequent disorganization of flagellar assembly. To confirm the effect of the ΔrpoN2 mutation on X. campestris flagellation, we used TEM to observe the filaments of the Xc1, ΔrpoN1, ΔrpoN2, and (a) Deletion of rpoN2 in X. campestris Xc1 affects biofilm formation in nutrient yeast glycerol (NYG) broth, as determined by crystal violet (CV) staining. (b) Deletion of rpoN2 in X. campestris affects EPS production in NYG broth. Error bars, means ± SD (n = 3). ****Significant differences (p < .0001, assessed by one-way analysis of variance). All experiments were repeated three times with similar results F I G U R E 7 Deletion of rpoN1 affected the synthesis of branched-chain fatty acids and increased the production of diffusible signalling factor (DSF) family signals in Xanthomonas campestris. (a) Fatty acid composition of the total lipid extracts from the wild-type Xc1 and ΔrpoN1 mutant strains grown in nutrient yeast glycerol (NYG) medium. (b) Fatty acid composition of the total lipid extracts from the Xc1 and ΔrpoN2 mutant strains grown in NYG medium. (c) Fatty acid composition of the total lipid extracts from the Xc1 and ΔrpoN1N2 doublemutant strains grown in NYG medium. Total lipids were extracted and transesterified to fatty acid methyl esters, and the products were identified by gas chromatography-mass spectrometry (GC-MS). Values are percentages of total fatty acid levels and are means ± SD of three independent experiments. n-C 14:0 3-OH, 3-hydroxyltetradecanoic acid; n-C 14:0 , tetradecanoic acid; iso-C 15:0 , 13-methyl-tetradecanoic acid; anteiso-C 15:0 , 12-methyl-tetradecanoic acid; n-C 15:0 , pentadecanoic acid; n-C 16:1 , cis-9-hexadecenoic acid; n-C 16:0 , hexadecanoic acid; iso-C 17:1 cis-9, cis-9-15-methyl-hexadecenoic acid; n-C 18:1 cis-11, cis-11-octadecenoic acid; n-C 18:0 , octadecanoic acid. (d) DSF signals produced by the rpoN1 deletion mutant ΔrpoN1 grown in nutrient broth (NB) medium. Supernatants (50 ml) of strain ΔrpfCΔrpoN1 grown in NB medium for 36 hr were collected and DSF signals were detected. (e) DSF signals produced by the rpoN2 deletion mutant ΔrpoN2 grown in NB medium. Supernatants (50 ml) of the strain ΔrpfCΔrpoN2 grown in NB medium for 36 hr were collected and DSF signals were detected. (f) DSF signals produced by the rpoN1 and rpoN2 double-mutant ΔrpoN1N2 grown in NB medium. Supernatants (50 ml) of strain ΔrpfCΔrpoN1N2 grown in NB medium for 36 hr were collected and DSF signals were detected. DSF, cis-11-methyl-2-dodecenoic acid; BDSF, cis-2-dodecenoic acid; IDSF, cis-10-methyl-2-dodecenoic acid. Error bars, mean ± SD (n = 3). *p < .05, **p < .01, ***p < .001, assessed by one-way analysis of variance. All experiments were repeated three times with similar results ΔrpoN1N2 strains and the rpoN1-and rpoN2-complemented strains.
TEM observation showed that deletion of rpoN2 led to defects in flagellar morphology compared with the wild-type strain (Figure 4a). We found that RpoN1 and RpoN2 are not functionally interchangeable for X. campestris flagellation (Figure 4a). This discovery can provide experimental reference information for analysis of the biological functions of RpoN2 for X. campestris flagellar assembly in the future.
RpoN binds to cognate promoters containing a TGGCACN 5 TTGCW motif (Barrios et al., 1999;Doucleff et al., 2005;Wiesler et al., 2012;Schaefer et al., 2015). To determine the genome-wide variation caused by X. campestris RpoN, the conserved RpoN consensus binding motif was also used to search the complete genomic sequences of X. campestris to identify the conserved GC sequences in the promoter of Xcc fliC. We hypothesized that RpoN2 binds to the promoter of Xcc fliC to regulate X. campestris flagellation. We demonstrated by EMSA that RpoN2 could positively regulate the expression of Xcc fliC by directly binding to the promoter of this gene, but RpoN1 exhibited weak binding to the Xcc fliC promoter ( Figure 5). We found that RpoN2 can also bind to the promoter of Xcc fliQ ( Figure S4). Taken together, these results suggest that RpoN2 might be a transcriptional activator that modulates bacterial flagellation by directly binding to the promoters of Xcc fliC and Xcc fliQ in X. campestris. However, further details need to be studied in the future.
Fatty acids and DSF family QS signals are critical secondary metabolites for bacterial physiological metabolism (Zhou et al., 2015a;Li et al., 2017;Hu et al., 2018). In this study, in-frame deletion of rpoN1 led to significantly reduced production of branched-chain amino acids (BCAAs) and increased production of DSF family QS signalling molecules, especially cis-11-methyl-2-dodecenoic acid. The precursors of the DSF family signals originate from the fatty acid synthesis pathway (Yu et al., 2016;Li et al., 2017). However, BCAAs are the primary precursors for the synthesis of BCFAs (Zhu et al., 2005). RpoN also affects amino acid utilization and carbohydrate metabolism (Stevens et al., 2010;Diep et al., 2015;Hayrapetyan et al., 2015). Therefore, we speculate that rpoN1 directly regulates the biosynthesis of amino acids and basal metabolism to affect BCFA and DSF synthesis. RNA-Seq showed that the rpoN1 mutation significantly affected the expression of a valine, leucine, and isoleucine biosynthesis-related gene (leuA) and a long-chain fatty acid-CoA ligase gene (rpfB) (Figure S6 and Tables S3 and S5). These results suggest that RpoN1 and RpoN2 might play different regulatory roles in the basal metabolism of X. campestris. However, elucidation of the detailed mechanism requires further investigation.
Based on our results, we propose a schematic model, as shown in Figure 8. In summary, X. campestris RpoN2 modulates biofilm formation and EPS production, and regulates bacterial flagellation via direct binding to the promoters of Xcc fliC and Xcc fliQ. These factors together cause weakening of X. campestris virulence. Moreover, RpoN1 could regulate fatty acid synthesis and DSF production in X. campestris.

| Materials
The antibiotics were obtained from Sigma-Aldrich. Takara Biotechnology Co. provided the molecular biology reagents. Novagen provided the pET vectors. Ni-agarose columns were obtained from Sigma-Aldrich. Bio-Rad provided Quick Start Bradford dye reagent.
All other reagents were of the highest available quality. Genscript (Nanjing, Jiangsu, China) synthesized the oligonucleotide primers.

| Bacterial strains, plasmids, and growth conditions
The strains and plasmids used in this study are listed in Table S1.
E. coli strains were grown in Luria Bertani medium (10 g/L tryptone, F I G U R E 8 Schematic of the proposed RpoN-mediated regulation of flagellar synthesis, virulence, extracellular polysaccharide (EPS) production, biofilm formation, quorum-sensing (QS), and basal metabolic pathways. The potential regulatory pathways and interactions of RpoN are proposed according to our observations and previous studies. RpoN1 could regulate branched-chain fatty acid production and modulate the synthesis of diffusible signalling factor (DSF) signals in Xanthomonas campestris. In the present study, we showed that RpoN2 directly bound to the fliC and fliQ promoter regions and positively regulated the expression of the flagellar assembly genes fliC and fliQ. In addition, RpoN regulated the production of EPS and biofilms, which are also used by X. campestris and other pathogenic bacteria as pathogenicity-related virulence factors against host plants 5 g/L yeast extract, 10 g/L NaCl, pH 7.0) at 37 °C. X. campestris strains were grown at 28 °C in NYG medium (5 g/L peptone, 3 g/L yeast extract, 20 g/L glycerol, pH 7.0) or nutrient broth agar (NA) (5 g/L peptone, 3 g/L beef extract, 10 g/L sucrose, 1 g/L yeast extract, pH 7.0). For culture medium preparation, tryptone, peptone, beef extract, and yeast extract were purchased from Sangon Biotech. When required, antibiotics were added at the following concentration: 100 μg/ml sodium ampicillin, 30 μg/ml kanamycin sulphate, and 30 μg/ml gentamycin for E. coli or 50 μg/ml rifampicin for X. campestris.

| Protein expression and purification
Protein expression and purification were performed as described previously (Li et al., 2017). To clone the Xcc rpoN1 and Xcc rpoN2 genes, genomic DNA extracted from X. campestris was used for PCR amplification with Pfu DNA polymerase using the primers listed in Table S2. PCR products were inserted into pET-28b(+) to produce the plasmids pET-rpoN1 and pET-rpoN2. The Xcc rpoN1 and Xcc rpoN2 genes were verified via nucleotide sequencing by Genscript. Xcc rpoN1 and Xcc rpoN2 with vector-encoded His 6 -tagged N-termini were expressed in E. coli BL21 (DE3) and purified with Ni-NTA agarose (Sigma-Aldrich) using a nickel-ion affinity column. Protein purity was monitored by sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE).

| Deletion of Xcc rpoN genes and complementation
To disrupt the Xcc rpoN1 and Xcc rpoN2 genes, the pK18mobsacBborne in-frame deletion suicide plasmids pK18-ΔrpoN1 and pK18-ΔrpoN2 were constructed. The 500-bp DNA fragments flanking the Xcc rpoN1 and Xcc rpoN2 genes were amplified with Pfu DNA polymerase using X. campestris genomic DNA as a template, and either rpoN1 BamHI and rpoN1 up1 (for Up rpoN1), rpoN1 down1 and rpoN1 HindIII (for Down rpoN1), rpoN2 BamHI and rpoN2 up1 (for Up rpoN2), or rpoN2 down1 and rpoN2 HindIII (for Down rpoN2), as primers (Table S2). Fragments were purified and joined by overlap PCR. The fused fragment was digested with BamHI and HindIII, and inserted into pK18mobscaB (Schafer et al., 1994) to obtain the plasmids pK18-ΔrpoN1 and pK18-ΔrpoN2. The resulting constructs were transferred into X. campestris by electroporation and kanamycin was used to select for integration of the nonreplicating plasmid into the recipient chromosome. A single-crossover integrant colony was spread on NYG medium without kanamycin at 28 °C for 36 hr, and after appropriate dilution the culture was spread on NYG plates containing 15% sucrose. Colonies sensitive to kanamycin were screened by PCR using the primers listed in Table S2, and the Xcc rpoN1 and Xcc rpoN2 deletion strains (ΔrpoN1 and ΔrpoN2) were obtained. For complementation of the Xcc rpoN1 and Xcc rpoN2 mutants, the coding regions of Xcc rpoN1 and Xcc rpoN2 were amplified by PCR and cloned into the versatile pBBR1MCS5 plasmid (Kovach et al., 1995). The resulting plasmid was transferred into the X. campestris strain by electroporation. The ΔrpoN1/N1 andΔrpoN2/N2 strains were obtained. The ΔrpoN1N2 double-mutant and complemented strains ΔrpoN1/N2, ΔrpoN2/N1, ΔrpoN1N2/N1, and ΔrpoN1N2/N2 were obtained by the same method.

| Pathogenicity assays
Plant inoculation and virulence assays were conducted as previously described (Cai et al., 2017). In brief, 6-week-old plants of the cabbage cultivar B. oleracea 'Jingfeng No. 1' were used as host plants. The wild-type strain Xc1 and sterile 10 mM MgCl 2 were used as positive and negative controls, respectively. All bacterial strains were cultured overnight in NYG medium containing appropriate antibiotics.
Cells were collected and washed with 10 mM MgCl 2 , and the cell densities were adjusted to OD 600 = 0.1 before inoculation into plant leaves using sterile scissors. Lesion lengths were measured 10 days after inoculation on 20 leaves for each strain tested.

| Quantitative reverse transcription PCR
RT-qPCR was carried out as described in previous studies (Cui et al., 2018). The bacterial cells were collected when the cell optical density (OD 600 ) reached 1.0 in NYG medium. Total RNA was extracted using the TRIzol-based method (Life Technologies). RNA quality control was performed by several steps: (a) RNA degradation degree and potential contamination were monitored on 1% agarose gels; (b) RNA purity (A 260 /A 280 , A 260 /A 230 ) was checked using a NanoPhotometer spectrophotometer (IMPLEN); and (c) RNA integrity was measured using a Bioanalyzer 2,100 (Agilent). The primers used in this assay are listed in Table S2. Quantification of gene expression and melting curve analysis were completed using QuantStudio 6 Flex real-time PCR system (Applied Biosystems), and TransStart Top Green qPCR SuperMix (TransGen Biotech) was used according to the manufacturer's instructions. As a control, RT-qPCR was similarly applied to analyse 16S rDNA gene expression. The relative expression levels of target genes were calculated using the 2 −ΔΔCt method for comparative quantitation.

| RNA-Seq
The RNA-Seq assay was performed as described previously (Yang et al., 2017;Zhou et al., 2017). Briefly, the wild-type Xc1, ΔrpoN1, ΔrpoN2, and ΔrpoN1N2 mutant strains were grown in NYG medium, and their cells were collected when the OD 600 reached 1.0 based on the growth curve. The collected cells were used for RNA extraction by the TRIzol-based method (Life Technologies), and RNA degradation and contamination were monitored on 1% agarose gels. Clustering and sequencing were performed by Genedenovo Biotechnology Co., Ltd (Guangzhou, Guangdong, China). To analyse the DEGs between the wild-type Xc1, ΔrpoN1, ΔrpoN2, and ΔrpoN1N2 mutant strains, the gene expression levels were further normalized using the fragments per kilobase of transcript per million (FPKM) mapped reads method to eliminate the influence of different gene lengths and amount of sequencing data on the calculation of gene expression.
The edgeR package (http://www.r-proje ct.org/) was used to determine DEGs across samples with fold changes ≥2 and a false discovery rate-adjusted p (q value) <.05. DEGs were then subjected to enrichment analysis of gene ontology (GO) functions and KEGG pathways, and q values were corrected using <.05 as the threshold.

| Measurement of extracellular enzymatic activity and swimming motility
The relative activities of extracellular enzymes were assayed as described previously (Wei et al., 2007;Yu et al., 2016). Two microlitres of each X. campestris strain culture (OD 600 ≈ 1.0) was spotted onto NYG agar plates containing 1% (wt/vol) skim milk (for protease), 0.5% (wt/vol) carboxymethylcellulose (for cellulase), or 0.1% (wt/vol) starch (for amylase) and incubated at 28 °C for 24-48 hr. The plates were stained where necessary as previously described (Wei et al., 2007). The zones of clearance around the spots due to degradation of the substrate were photographed. Three plates were inoculated in each experiment, and each experiment was repeated three times.
The relative enzymatic activity was indicated by the diameter of the clearance zone.
Bacteria were inoculated into the centres of NYG plates containing 0.3% agarose. The plates were incubated at 28 °C for 48 hr before the colony diameters were measured.

| EPS formation assay
EPS production was measured as described previously (Yu et al., 2016). Each X. campestris strain culture (2 ml, OD 600 ≈ 1.0) was used to inoculate 100 ml of NYG medium containing 4% glucose in a 250 ml flask and kept at 28 °C with shaking at 180 rpm for 4 days. The EPS was precipitated from the culture supernatant by the addition of 4 volumes of ethanol. The pelleted EPS was washed with 70% ethanol, air dried, and weighed. Three flasks were inoculated in each experiment, and each experiment was repeated three times.

| Biofilm formation assay
A biofilm formation assay was performed as described previously with some modifications (Wang et al., 2018). Briefly, bacterial cells were cultured in NYG medium to a final OD 600 of 1.0. Then, 3 ml of the cell suspension was added to sterilized polystyrene tubes. These tubes were kept in a humidified chamber at 28 °C for 7 days without shaking. The cultures were then moved and the tubes were washed three times in tap water. Biofilm formation on the tubes was visualized by staining with 0.1% CV, followed by washing three times in tap water. The CV-stained biofilm in polystyrene tubes was dissolved in methanol-acetic acid-water (4:1:5, vol/vol/vol) and quantified by measuring the A 575 using an 8,453 UV-visible spectrophotometer (Agilent). The average of three replicates was used for quantitative measurement. The biofilm assays were repeated three times and showed similar results with three replicates each time.

| Detection of DSF signal components in the X. campestris culture supernatant
The protocol for extraction and purification of DSF family components was described previously (Zhou et al., 2015a). X. campestris strains were cultured in liquid medium for 36 hr and 50 ml of the bacterial supernatant was collected by centrifugation at 4,000 × g for 15 min at 4 °C. The pH of the supernatants was adjusted to 4.0 by adding hydrochloric acid prior to two extractions with an equal volume of ethyl acetate. The ethyl acetate fractions were collected, and the solvent was removed by rotary evaporation to dryness at 42 °C.
The residue was dissolved in 100 µl of methanol. The crude extract was subjected to 0.22-µm Mini-star filtration, and the filtrate was concentrated to 100 µl. The extract (10 µl) was injected into a C18 reversed-phase HPLC column (4.6 × 250 mm, Agilent Technologies, Inc.) and eluted with water in methanol (23:77 vol/vol, 0.1% formic acid) at a flow rate of 1 ml/min in an HPLC E2695 system (Waters) with a UV220 detector.

| Analysis of fatty acid composition
Bacterial cultures were grown aerobically for 2-4 days. Cells were harvested and washed three times with sterile water. Fatty acid methyl esters were synthesized and extracted as described previously (Li et al., 2017). Cellular lipids were saponified by addition of 2 ml of sodium hydroxide/methanol solution at 100 °C for 40 min with shaking (800 rpm). Fatty acids were methylated by the addition of 4 ml of hydrochloric acid/methanol solution at 80 °C for 20 min and cooled to less than 20 °C. Fatty acid methyl esters were obtained by three extractions with 1 ml of petroleum ether. The solvent was removed under a stream of nitrogen, and the residue was dissolved in 100 µl of hexane. The crude extract was filtered with 0.22-µm Mini-star units, and 2 µl of extract was analysed by GC-MS. Samples were analysed with a GC-MS system (Agilent 5975c) with a DB 5MS chromatographic column. The oven temperature was held at 100 °C for 5 min, changed at 10 °C/min to 200 °C and held for 5 min, then changed at 10 °C/min to 250 °C and held for 5 min. Electron impact ionization (EI+, 70 eV) was used for all samples. Mass spectrometry was carried out at 1 s/ scan, m/z 35-500, and 1 kV, and the data were analysed by the NIST 08 database.

| Statistical analyses
The experimental datasets were subjected to analyses of variance using GraphPad Prism 7.0. The significance of the treatment effects was determined by the F value (p = .05). If a significant F value was obtained, separation of means was accomplished by Fisher's protected least significant difference at p ≤ .05.

ACK N OWLED G EM ENTS
This work was supported by the National Natural Science Foundation of China (31571974).

CO N FLI C T O F I NTE R E S T
The authors declare no competing financial interests.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.