A HU‐like protein is required for full virulence in Xanthomonas campestris pv. campestris

Abstract Bacteria harbour several abundant small DNA‐binding proteins known as nucleoid‐associated proteins (NAPs) that contribute to the structure of the bacterial nucleoid as well as to gene regulation. Although the function of NAPs as global transcriptional regulators has been comprehensively studied in the model organism Escherichia coli, their regulatory functions in other bacteria remain relatively poorly understood. Xanthomonas campestris pv. campestris (Xcc) is a gram‐negative bacterium that causes black rot disease in almost all members of the crucifer family. In previous work, we demonstrated that a Fis homologue protein, which we named Fis‐like protein (Flp), contributes to the regulation of virulence, type III secretion, and a series of other phenotypes in Xcc. Here we have examined the role of XC_1355, which is predicted to encode a DNA‐binding protein belonging to the HU family herein named HU‐like protein (Hlp). We show that mutation of XC_1355 in Xcc reduces the virulence, extracellular polysaccharide production, and cell motility, but has no effect on the production of extracellular enzymes and induction of the hypersensitive response. These data together with transcriptome analysis indicate that hlp is a previously uncharacterized gene involved in virulence that has partially overlapping and complementary functions with flp in Xcc, although the two regulators have opposite effects on the expression of genes involved in type III secretion. The findings add to our understanding of the complex regulatory pathways that act to regulate virulence in Xcc.

structuring (H-NS), heat unstable (HU), and integration host factor (IHF). NAPs of the HU, IHF, and H-NS families are often referred to as histone-like proteins (HLPs), because in common with eukaryotic histones they exhibit low molecular mass, high copy number, high electrostatic charge, and strong DNA-binding ability (Grove, 2011;Kamashev et al., 2017).
It is known that Fis influences DNA topology by directly binding and bending DNA (for review, see Dillon & Dorman, 2010). HU binds to DNA nonspecifically but with a higher affinity for nicked, gapped, and cruciform DNA structures. H-NS is also described as binding nonspecifically to DNA but prefers intrinsically curved DNA.
IHF usually binds and bends DNA strongly at specific sequences.
In addition to their roles in maintaining bacterial chromosome architecture and regulating DNA transactions such as recombination and DNA replication, there is an increasing body of work describing how NAPs or their homologues also play important roles in gene regulation. For example, Fis is involved in controlling the expression of virulence factors in a number of pathogens (Duprey et al., 2014;Leng et al., 2019;Lv et al., 2018), where it regulates gene expression by modulating the level of DNA supercoiling in the cell and interacting with RNA polymerase at the position of its binding site (Dillon & Dorman, 2010). HU regulates the expression of hundreds of genes across the whole genome in E. coli by remodelling bacterial nucleoids, modulating the 3D arrangement of DNA, facilitating DNA looping in a promoter region, trapping free supercoils, indirectly altering supercoiling through DNA topoisomerases, or cooperating with transcription regulators (Berger et al., 2016;Lioy et al., 2018;Remesh et al., 2020).
NAPs in a single organism often have overlapping and complementary functions, although the underlying mechanisms are still unknown.
In E. coli, null mutations in genes encoding each of the NAPs generally have minor consequences for the cell, showing that each individual protein is not essential for this bacterium. However, disruption of multiple NAPs causes more severe effects (Yasuzawa et al., 1992).
Although the function of NAPs as global transcriptional regulators has been comprehensively studied in the model organism E. coli, their regulatory function and associated mechanisms in other bacteria remain poorly understood.
Xanthomonas campestris pv. campestris (Xcc) is a gram-negative bacterium that causes black rot disease in almost all members of the crucifer family (Brassicaceae), including important vegetables such as broccoli, Brussels sprouts, cabbage, cauliflower, kale, mustard, radish, and rape, and the model plant Arabidopsis thaliana (Vicente & Holub, 2013). This phytopathogen infects host plants via wounds or hydathodes. After infection, the bacterial cells multiply in the intercellular spaces, spread via the vascular system, and lead to the development of typical disease symptoms: vein blackening and Vshaped chlorotic and necrotic lesions extending from leaf margins along veins (Chan & Goodwin, 1999).
Xcc encodes a number of virulence factors, such as type III secretion system (T3SS)-dependent effectors, cyclic glucans, lipopolysaccharides, extracellular polysaccharide (EPS, also called xanthan gum), and a series of extracellular enzymes including amylase, endoglucanase, polygalacturonate lyase, and protease (Büttner & Bonas, 2010;Ryan et al., 2011). These virulence factors act in diverse ways to promote Xcc virulence. Complex regulatory networks, including quorum-sensing pathways, multiple two-component systems, and transcriptional regulators, act to ensure appropriate expression of these virulence factors to promote successful invasion and proliferation during pathogenesis.
Although the regulation of virulence in Xcc has been studied over the past several decades, our understanding of this complex of regulatory pathways is not complete. In particular, the role of NAPs has received comparatively little attention. In a previous work, we demonstrated that a Fis homologue protein, which we named Fislike protein (Flp), contributes to the regulation of virulence, expression of T3SS genes, and a series of other phenotypes in Xcc (Leng et al., 2019). Here we have examined the role of XC_1355, which is predicted to encode a DNA-binding protein belonging to the HU family herein named HU-like protein (Hlp). We show that mutation of XC_1355 in Xcc reduces the virulence, EPS production, and cell motility, but has no effect on the production of extracellular enzymes and induction of the hypersensitive response (HR). Moreover, overexpression of XC_1355 in the flp deletion mutant Δflp partially restored the virulence towards wild type. These data together with transcriptome analysis indicate that hlp is a previously uncharacterized gene involved in virulence that has partially overlapping and complementary functions with flp in Xcc.

| The HU family protein Hlp influences the virulence of Xcc
In our previous work we identified and characterized a Fis-like protein (named Flp) that plays an important role in virulence and T3SS gene expression in Xcc (Leng et al., 2019). Analysis of Xcc genome sequences (da Silva et al., 2002;Qian et al., 2005;Vorhölter et al., 2008) revealed a further 12 genes that probably encode NAPs (Table 1). Homologues of some of these proteins have been implicated in the virulence of other Xanthomonas species. In particular, HupB is implicated in flagellar motility and virulence in Xanthomonas citri (Conforte et al., 2019), whereas XvrA, XvrB, and XvrC are implicated in virulence in Xanthomonas oryzae (Feng et al., 2009;Kametani-Ikawa et al., 2011;Liu et al., 2016).
As a first step to explore the roles of these other NAPs in Xcc, we screened a series of insertional mutants that were constructed using a suicide vector pK18mob (Windgassen et al., 2000) or transposon Tn5gusA5 strategy (Table S1). Virulence assays in the host plant Chinese radish using a leaf-clipping method showed that besides flp and hupB (Leng et al., 2019;Zhang et al., 2020), mutation of XC_1355, which is predicted to encode a DNA-binding protein belonging to the HU family, also caused a reduction of disease symptoms when compared with the wild-type Xcc strain (Figure 1). XC_1355 was named HU-like protein (Hlp). In order to explore the detailed function of hlp, an in-frame deletion mutant of hlp was constructed by using the suicide vector pK18mobsacB; the derived mutant strain was designated Δhlp. Simultaneously, a complemented strain was constructed by introducing a recombinant pLAFR3 plasmid carrying the hlp coding sequence into Δhlp. The resulting complemented strain was named CΔhlp (Table S1). Virulence assays revealed that the Δhlp mutant caused markedly reduced disease symptoms ( Figure S1) compared with the wild type: the mean lesion length caused by Δhlp was significantly shorter than that caused by the wild-type strain 8004 ( Figure 2a). Furthermore, CΔhlp produced similar virulence symptoms to the wild type. These data confirmed the findings reported above that hlp is important for the virulence of Xcc.
To get more insight into the regulatory effects of Hlp on virulence in Xcc, the effects of Hlp on the proliferation of the pathogen in host tissues were assessed. To do this, the bacterial cell numbers of the wild-type strain, the Δhlp mutant, and the CΔhlp complemented strain in infected Chinese radish leaves were determined on each day for 10 days after inoculation. For the Δhlp mutant, the number of bacterial cells recovered from the infected leaves was 10-to 100fold less compared to the wild-type strain up to 5 days postinoculation ( Figure 2b). After 5 days the differences in bacterial numbers were smaller. The complemented strain CΔhlp showed similar cell numbers to the wild type ( Figure 2b). These data indicated that loss of Hlp has the strongest influence in the early invasive stage of the infection, where it reduces bacterial fitness ( Figure 2b).
The growth characteristics of the Δhlp mutant and the wild-type strain in the minimal medium XVM2 were also compared. Results revealed that the Δhlp strain displayed similar growth properties to the wild-type strain in all growth phases ( Figure S2). These findings indicated that the impact of mutation of hlp on growth in planta was not due to a general growth defect.

| Hlp is not required for HR in nonhost plants
Our finding that Hlp is required for full virulence and in planta growth of Xcc in host plants prompted us to examine the influence of Hlp on the ability to induce HR in nonhost plants. For these experiments, the Δhlp mutant was infiltrated at a cell concentration of 10 7 cfu/ml into the leaves of the pepper cultivar ECW-10R. The results showed that this mutant elicited similar HR symptoms to those seen with the wild-type strain (Figure 2c)

| Hlp influences extracellular polysaccharide production, cell motility, and stress tolerance, but not extracellular enzyme production
To explore the role of Hlp in the pathogenesis of Xcc, we conducted a series of basic phenotypic tests to examine the influence that mutation of hlp might have on the production of EPS, the production of extracellular enzymes, cell motility, and the tolerance to stress and antimicrobials.
The results showed that the mutant strain Δhlp produced slightly less EPS ( Figure 3a) and had decreased motility (swimming and swarming, tested on 0.28% (wt/vol) agar plates and 0.6% (wt/vol) agar plates, respectively; Figure 3b). EPS production and motility of the CΔhlp strain were similar to wild-type values under the conditions tested (Figure 3a,b).
The ability of the wild-type strain, the Δhlp mutant, and the CΔhlp complemented strain to tolerate environmental stresses was assessed by using growth assays.
The growth of bacterial strains on agar plates supplemented with different concentrations of environmental stressors was examined.  Figure 3c).
Comparison of the production of extracellular enzymes (endoglucanase and amylase) showed no differences between the wildtype strain, the Δhlp mutant, and the CΔhlp complemented strain ( Figure S5).

| Functional redundancy between Hlp and Flp
To examine any functional redundancies between Flp and Hlp (which both regulate virulence), the phenotypic effects of expression of hlp in the deletion mutant Δflp was tested. To do this, a construct comprising the vector pLAFR3 carrying the hlp gene was transferred into deletion mutant Δflp (Table S1). As a first step, this strain was tested for virulence in the host Chinese radish using a leaf-clipping assay. Results revealed that the expression of F I G U R E 1 Mean lesion lengths (disease symptoms) caused by different Xanthomonas campestris pv. campestris (Xcc) mutant strains. The Xcc wild-type strain 8004 and a series of mutant strains (Δflp, 1046pk, 195B09, 1355pk, 1656pk, 1806pk, 1860nk, 247H03, 3262nk, 100B06, 147G06, 4203pk; for the corresponding genes, see Table S1) were cultured in NYG medium overnight and then adjusted to 10 7 cfu/ ml in sterile distilled water. Chinese radish (Raphanus sativus) was inoculated with bacterial suspensions of different Xcc strains by the leaf-clipping method. The lesion lengths were measured at 10 days postinoculation. Values given are the mean and SD from 15 inoculated leaves in one experiment. Significance was determined by analysis of variance and Dunnett's post hoc test for comparison to the wild type. *p < .05; **p < .01. The experiment was repeated three times with similar results hlp partially restored the virulence of the Δflp mutant: the Δflp/ pLChlp strain produced more severe disease symptoms with a mean lesion length of 7.2 mm than the Δflp mutant with a mean lesion length of 2.8 mm, but less severe disease symptoms compared to the wild type with a mean lesion length of 12.3 mm ( Figures 4a and S6).

F I G U R E 2
Hlp regulates virulence and in planta growth in host plants, but not the hypersensitive response (HR) in nonhost plants. (a) Lesion lengths caused by tested Xanthomonas campestris pv. campestris (Xcc) strains. The Xcc wild-type strain 8004, the hlp deletion mutant Δhlp, the complemented strain CΔhlp, and control strain Δhlp/pLAFR3 were cultured in NYG medium overnight and then adjusted to 10 7 cfu/ml in sterile distilled water. Xcc strains were inoculated into Chinese radish by the leaf-clipping method. Lesion lengths were scored 10 days postinoculation. Data are shown as the mean ± SD from 15 inoculated leaves in one experiment. Significance was determined by analysis of variance and Dunnett's post hoc test for comparison to the wild type. **p < .01; n.s., not significant. The experiment was repeated three times with similar results. (b) Growth of Xcc strains within host plants. The Xcc wild-type strain 8004, the hlp deletion mutant Δhlp, and complemented strains CΔhlp were cultured in NYG medium overnight and then adjusted to 10 7 cfu/ml in sterile distilled water. Xcc strains were inoculated onto radish leaves using the leaf-clipping method. Five leaves were collected from every group of clipped leaves daily and homogenized in sterile water. The homogenates were diluted and plated on NYG plates. Bacterial cfu were counted after incubation for 3 days. Data are shown as the mean and SD from three replicates. (c) HR induction by Xcc strains in nonhost plants. Bacterial cells from overnight culture were resuspended in 10 mM sodium phosphate buffer to a density of 10 7 cfu/ml. Approximately 5 μl bacterial resuspension was infiltrated into pepper (Capsicum annuum 'ECW-10R') leaf mesophyll tissue with a blunt-end plastic syringe. HR symptoms were recorded at 8 and 24 hr postinoculation (hpi) (top). There was electrolyte leakage from pepper leaves inoculated with Xcc strains (bottom). The conductivity of the infiltrated spots was measured by a DDS-307A conductometer at 8, 16, and 24 hpi, with four 0.4-cm 2 leaf discs collected from the infiltrated area for each sample. Three samples were taken for each measurement in each experiment. Data are shown as the mean ± SD of three replicates from a representative experiment, and similar results were obtained in two other independent experiments F I G U R E 3 Hlp regulates a series of phenotypes in Xanthomonas campestris pv. campestris (Xcc). (a) Extracellular polysaccharide (EPS) yield of tested Xcc strains. Xcc strains were cultured in NY medium containing 2% glucose for 3 days before EPS was extracted and quantified. Data are shown as the mean ± SD of three replicates from a representative experiment. Significance was determined by analysis of variance (ANOVA) and Dunnett's post hoc test for comparison to the wild type. **p < .01; n.s., not significant. The experiment was repeated three times with similar results. (b) Motility of tested Xcc strains. Two microlitres of culture suspension (10 9 cfu/ml) of Xcc strains was stabbed into "swim" (0.28% agar) medium and incubated for 4 days at 28 °C or inoculated onto "swarm" (0.6% agar) plates and incubated for 3 days at 28 °C. Colony diameters were measured. Data are shown as the mean ± SD of 10 measurements from a representative experiment. Significance was determined by ANOVA and Dunnett's post hoc test for comparison to the wild type. **p < .01; n.s., not significant. The experiment was repeated three times with similar results. (c) Sodium dodecyl sulphate (SDS) tolerance of tested Xcc strains. Survival experiments were performed by subculturing strains overnight on fresh NYG agar plates supplemented with different SDS concentrations. The surviving bacterial colonies on the plates were counted after incubation for 3 days The virulence and in planta growth of the Δflp mutant, the cross-complemented strain Δflp/pLChlp, and the control strain Δflp/ pLAFR3 in infected leaves were also estimated. Consistent with our previous work (Leng et al., 2019), the Δflp mutant and the control strain Δflp/pLAFR3 had lower population numbers in planta compared to the wild type in the first 5 days after inoculation ( Figure 4b).
Interestingly, the population size of the cross-complemented strain Δflp/pLChlp was near to that of the wild type during the same period, indicating that in trans expression of the hlp gene in the Δflp mutant restored in planta fitness in the early, invasive stage.
These experiments were extended to examine the effects of expression of other NAPs on the virulence of the Δflp mutant (Table S1). However, none of these strains showed a change in virulence compared to the Δflp mutant ( Figure 4a).
In parallel, the expression of hlp in the cross-complemented strain

| Hlp controls the expression of genes involved in virulence and various adaptation processes in Xcc
To gain a better understanding of the regulatory role of Hlp in Xcc, a set of global gene expression profiles was generated using collected, and total RNA was extracted from two independent biological replicates. Following library construction and sequencing, the generated data were analysed to assess differential gene expression. were cultured in NYG medium overnight and then adjusted to 10 7 cfu/ml in sterile distilled water. Xcc strains were inoculated onto radish leaves using the leaf-clipping method. Five leaves were collected from every group of clipped leaves daily and homogenized in sterile water. The homogenates were diluted and plated on NYG plates. Bacterial cfu were counted after incubation for 3 days. Data are shown as the mean and SD from three replicates

F I G U R E 4 Overexpression of the hlp gene in the
Of the 4,273 annotated genes from the genome of Xcc strain 8004, 535 genes were differentially expressed (|log 2 (fold change) ≥ 1), with 481 genes up-regulated and 54 genes down-regulated according to the transcriptome data ( Figure 6a, Table S2), implying that Hlp mainly acts as a repressor of gene transcription in Xcc. To verify the transcriptome data, 25 differentially expressed genes (DEGs) were randomly selected and semiquantitative RT-PCR was performed to examine the relative expression levels of these genes. All selected genes showed expression changes that were comparable with the transcriptome data (Table S3). Hlp were assigned to functional categories that are based on clusters of orthologous genes. A total of 215 genes were predicted to encode hypothetical proteins or have not been given a functional category to date (Figure 6a, Table S2). The most dominant functional category was "pathogenicity and adaption" (48 genes).

F I G U R E 5
Overexpression of hlp in the Δflp mutant restores its extracellular polysaccharide (EPS) production, cell motility, and stress tolerances. (a) Quantitative reverse transcription-PCR assay to examine the transcription levels of hlp in Xanthomonas campestris pv. campestris (Xcc) strains. RNA was isolated from cultures of Xcc strains grown in XVM2 medium for 24 hr. The relative mRNA levels were calculated with respect to the level of the corresponding transcript in the wild-type strain 8004. Values given are the means ± SD of triplicate measurements from a representative experiment; genes with a twofold change in expression in the mutants compared to the wild type were considered differentially expressed (*, significant). Similar results were obtained in two other independent experiments. (b) EPS yield of Xcc strains cultured in NY medium containing 2% glucose for 3 days. Data are shown as the mean ± SD of three replicates from a representative experiment. Significance was determined by analysis of variance (ANOVA) and Dunnett's post hoc test for comparison to the wild type. **p < .01; n.s., not significant. Similar results were obtained in two other independent experiments. (c) Measurements of colony diameters of Xcc strains in "swim" (0.28% agar) medium and on "swarm" (0.6% agar) medium after 4 and 3 days of incubation at 28 °C, respectively. Data are shown as the mean ± SD of 10 measurements in a representative experiment. Significance was determined by ANOVA and Dunnett's post hoc test for comparison to the wild type. **p < .01; n.s., not significant. Similar results were obtained in two other independent experiments. (d) Stress tolerance test of Xcc strains on fresh NYG agar plates supplemented with different concentrations of sodium dodecyl sulphate (SDS) (i), phenol (ii), and CuSO 4 (iii). Values given are the means ± SD of triplicate measurements from a representative experiment. Similar results were obtained in two other independent experiments

| Hlp and Flp convergently regulate genes involved in different phenotypes
The influence of Flp on the Xcc transcriptome has been characterized in our previous work (Leng et al., 2019). Here we compared the general overlap of genes controlled by both Flp and Hlp. In total, 55 genes were found to be regulated by both Flp and Hlp (Figure 6b, genes were found to be differentially expressed (|log 2 (fold change)| ≥ 1) in the Δhlp mutant (Table S2). These genes were broadly categorized according to their biological function (He et al., 2007;Qian et al., 2005) Figure 6c). Results demonstrated that the expression of these selected genes was consistent with the data from the transcriptome analyses (Figure 6c).

| Hlp and Flp divergently regulate the type III secretion system in Xcc
The gene transcription profile data revealed that the expression of T3SS-related genes was increased in the Hlp mutant but downregulated in the Δflp mutant compared to the wild type (Tables S2   and S4 F I G U R E 7 Mutation of hlp in the Δflp mutant background restores the expression of type III secretion system (T3SS) genes. (a) The expression levels of hrp and type III effector (T3E) genes in Xanthomonas campestris pv. campestris (Xcc) wild-type strain 8004, the Δflp mutant, and the ΔflpΔhlp double mutant as measured by quantitative reverse transcription-PCR. RNA was isolated from cultures of Xcc strains grown in XVM2 medium for 24 hr. Relative gene expression with respect to the corresponding wild-type levels were calculated. Values given are the means ± SD of triplicate measurements from a representative experiment. Genes were considered to be differentially expressed if |log 2 (fold change)| ≥ 1 compared to the wild type (*, significant). Similar results were obtained in two other independent experiments. (b) Hypersensitive response (HR) symptoms observed after infiltration. The Xcc wild-type strain 8004, the ΔflpΔhlp double mutant, the Δflp mutant, and the Δhlp mutant were infiltrated into pepper leaves. Three replications were performed in each experiment, and each experiment was repeated three times. The results presented are from a representative experiment, and similar results were obtained in all other independent experiments. (c) Electrolyte leakage from pepper leaves inoculated with Xcc strains. For each sample, four 0.4-cm 2 leaf discs were collected from the bacteria-infiltrated area and incubated in 5 ml distilled water. Conductivity was measured with a DDS-307A conductivity meter. Three samples were taken for each measurement in each experiment. Results presented are the mean ± SD of three replicates from a representative experiment, and similar results were obtained in two other independent experiments

| D ISCUSS I ON
The miscellaneous interactions between NAPs and DNA and the dynamic nucleoid architecture lead to global regulation of bacterial gene expression affecting a variety of cellular processes (Kisner & Kuwada, 2020 (Dillon & Dorman, 2010;Meyer et al., 2018;Remesh et al., 2020), and HU proteins have been implicated in virulence and adaptation to environmental conditions in several mammalian pathogens (Hołówka et al., 2018;Sakatos et al., 2018). Nevertheless, there are relatively few reports on the role of HU proteins in plant pathogens.
Previous work has shown that a 90-amino-acid HU-like protein (named HupB in Xanthomonas citri subsp. citri and HU Xcc in Xcc) has been reported to regulate virulence in Xanthomonas spp. (Conforte et al., 2019;Zhang et al., 2020). Here deletion of HU Xcc (XC_3262) in the Xcc wild-type strain 8004 reduced disease symptoms on the host plant Chinese radish using the leaf-clipping method (Figure 1). This is similar to the previous findings that the insertion inactivation mutant of HU Xcc exhibited partially weakened virulence on the host cabbage by the infiltration method (Zhang et al., 2020 Our work has also addressed the regulatory interplay between Hlp and Flp, a Fis homologue in Xcc that we have previously shown to be a global regulator implicated in virulence and HR induction (Leng et al., 2019). In this work, we focused on the effects of hlp mutation on the virulence and T3SS of Xcc. We demonstrated that it plays key roles in these processes. Furthermore, we focused on the interplay between Hlp and Flp and their regulatory effects on the T3SS.
However, more work is needed to understand more precisely how Hlp specifically regulates virulence and how the specific interplay between Hlp and Flp affects the expression of T3SS genes in plantpathogenic bacteria.

| Bacterial strains, plasmids, and growth conditions
The bacterial strains and plasmids used in this work are listed in Table S1. E. coli strains were cultured in Luria Bertani medium (Miller, 1972) at 37 °C. Xcc strains were cultured in NYG medium (Daniels et al., 1984), NY medium (NYG medium without glycerol), and the mimic medium XVM2 (Wengelnik & Bonas, 1996) at 28 °C and 200 rpm. Antibiotics were used at the following concentrations as required: kanamycin at 25 μg/ml, rifampicin at 50 μg/ml, ampicillin at 100 μg/ml, spectinomycin at 50 μg/ml, and tetracycline at 5 μg/ml for Xcc strains or 15 μg/ml for E. coli strains.

| Nucleic acid manipulations
The nucleic acid manipulations followed the procedures described by Sambrook et al. (1989). Conjugation between the Xcc and E. coli strains was performed as described by Turner et al. (1985). The restriction endonucleases, T4 DNA ligase, and Pfu DNA polymerase were provided by Promega. Total RNA was extracted from cultures of Xcc strains with a total-RNA extraction kit (Invitrogen) and cDNA was generated using a cDNA synthesis kit (Invitrogen). These kits were used with reference to the manufacturer's instructions.
Semiquantitative RT-PCR and RT-qPCR were carried out as previously described (Cui et al., 2018;Lu et al., 2009). For semiquantitative RT-PCR, the obtained cDNA was diluted and used as a template with selected primers for target genes (Table S5). For RT-qPCR, the SYBR Green-labelled PCR fragments were amplified using primer sets (

| Deletion mutant construction and crosscomplementation
The construction of an in-frame deletion mutant of hlp (XC_1355) was carried out using a previously described method (Leng et al., 2019).
The 439-bp (BamHI and XbaI) upstream sequence and the 340-bp (XbaI and HindIII) downstream sequence of the hlp gene were PCRamplified and cloned together into the suicide plasmid pK18mobsacB (Schäfer et al., 1994) (Table S1).

| Pathogenicity tests, in planta growth curve, HR assays, and ion leakage assays
The virulence of Xcc to Chinese radish (Raphanus sativus) was tested by the leaf-clipping method (Dow et al., 2003). Xcc strains, collected from overnight culture, were washed and adjusted to the same final density (OD 600 = 0.6, approximately 10 9 cfu/ml). The bacterial resuspension was then diluted to 10 7 cfu/ml. Leaves were cut with scissors dipped in the bacterial suspensions. The lesions and symptoms were measured 10 days postinoculation.
The growth of bacteria in radish leaf tissue was measured by homogenizing a group of leaves (five leaves for each sample) in 9 ml sterile water. Diluted homogenates were plated on NYG agar plates supplemented with corresponding antibiotics, and bacterial colonies were counted after incubation for 3 days.
HR was tested on pepper leaves (Capsicum annuum 'ECW-10R') as previously described (Castañeda et al., 2005;Li et al., 2014). Briefly, bacterial suspensions (10 7 cfu/ml) were infiltrated into the abaxial side of the pepper leaves. These inoculated plants were kept in the greenhouse to observe the HR symptoms and gauge conductivity at 8, 16, and 24 hr after inoculation. For conductivity measurements, samples (leaf discs of 0.4 cm 2 ) were collected and soaked in 10 ml ultrapure water with shaking at 200 rpm. The leaf discs were then removed and the conductivity of water was measured.

| Stress tolerance assay
The minimal inhibitory concentration (MIC) method (Li et al., 2020) was employed to test the sensitivity of the Xcc strains to several environmental stresses. Xcc strains were cultured overnight and diluted to an OD 600 of 0.1. Then 100 μl of the diluted culture was plated on NYG plates supplemented with different concentrations of each reagent. The surviving colonies on the plates were counted after 3 days of incubation at 28 °C.

| Exopolysaccharide and extracellular enzyme assays
EPS and extracellular enzyme assays were performed as previously described (Li et al., 2020;Tang et al., 1991). To estimate EPS production, Xcc strains were inoculated into 100 ml NY liquid medium containing glucose (2% wt/vol) at 28 °C, 200 rpm for 3 days. EPS was precipitated from the culture supernatant, dried, and weighed. For estimation of the activity of the extracellular enzymes endoglucanase (cellulase) and amylase, a radial diffusion assay was used as described by Tang et al. (1991). For quantification of endoglucanase and amylase, Xcc strains were cultured in NYG medium for 12 hr. Ten microlitres of enzyme-containing extracts (supernatant) was added to 200 μl of indicator buffer (citric acid-Na 2 HPO 4 , pH 5.5) containing 1% (wt/ vol) carboxymethylcellulose (for endoglucanase) or 1% (wt/vol) starch solution (for amylase) as the substrate. The reactions were carried out for 30 min at 28 °C. The released reducing sugars were measured as d-glucose equivalents as described by Miller (1959). One unit (U) of endoglucanase or amylase activity was defined as the amount of enzyme releasing 1 μmol of reducing sugar per minute.

| Cell motility assays
Cell motility was detected as previously described (Li et al., 2020).
To test swimming motility, two microlitres of bacterial suspension (10 9 cfu/ml) was stabbed into 0.28% agar plates composed of 0.03% Bacto peptone and 0.03% yeast extract followed by incubation at 28 °C for 4 days. To detect swarming motility, bacterial suspensions were spotted on NY plates containing 2% glucose and 0.6% agar, which were then incubated at 28 °C for 3 days. The diameter of the area occupied by strains was measured and the values were used to indicate the motility of Xcc strains. The experiment was repeated three times.

| Transcriptome analysis of the Hlp mutant
Transcriptome analysis was performed as previously described (Cui et al., 2018). Briefly, Xcc strains were cultured in XVM2 medium to a concentration of OD 600 of 0.6. RNA was prepared and the contaminating genomic DNA was removed. After the quantity was determined and quality was assessed, total RNA was sent to Novogene (Beijing, China) for library construction and strand-specific RNA sequencing. Clean reads were mapped to the genome of Xcc strain 8004 and the reads per kilobase per million mapped reads method was used to calculate the gene expression levels. Genes with false discovery rate ≤ .05 and |log 2 (fold change)| ≥ 1 were considered for differentially expressed.

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.