Involvement of hrpX and hrpG in the Virulence of Acidovorax citrulli Strain Aac5, Causal Agent of Bacterial Fruit Blotch in Cucurbits

Acidovorax citrulli causes bacterial fruit blotch, a disease that poses a global threat to watermelon and melon production. Despite its economic importance, relatively little is known about the molecular mechanisms of pathogenicity and virulence of A. citrulli. Like other plant-pathogenic bacteria, A. citrulli relies on a type III secretion system (T3SS) for pathogenicity. On the basis of sequence and operon arrangement analyses, A. citrulli was found to have a class II hrp gene cluster similar to those of Xanthomonas and Ralstonia spp. In the class II hrp cluster, hrpG and hrpX play key roles in the regulation of T3SS effectors. However, little is known about the regulation of the T3SS in A. citrulli. This study aimed to investigate the roles of hrpG and hrpX in A. citrulli pathogenicity. We found that hrpG or hrpX deletion mutants of the A. citrulli group II strain Aac5 had reduced pathogenicity on watermelon seedlings, failed to induce a hypersensitive response in tobacco, and elicited higher levels of reactive oxygen species in Nicotiana benthamiana than the wild-type strain. Additionally, we demonstrated that HrpG activates HrpX in A. citrulli. Moreover, transcription and translation of the type 3-secreted effector (T3E) gene Aac5_2166 were suppressed in hrpG and hrpX mutants. Notably, hrpG and hrpX appeared to modulate biofilm formation. These results suggest that hrpG and hrpX are essential for pathogenicity, regulation of T3Es, and biofilm formation in A. citrulli.


INTRODUCTION
Bacterial fruit blotch (BFB), caused by the gram-negative bacterium Acidovorax citrulli (Schaad et al., 1978(Schaad et al., , 2008Willems et al., 1992), is a globally occurring destructive disease of cucurbit crop species (Burdman and Walcott, 2012). Multiple studies have evaluated methods to mitigate the impact of BFB, including pathogen exclusion by quarantine, seed health testing, seed treatment, and field applications of biological control agents (Dutta et al., 2014;Jiang et al., 2015;Tian et al., 2016). Unfortunately, the success of these approaches has been limited, and BFB still poses a serious threat to commercial watermelon and melon production worldwide (Burdman and Walcott, 2012).
Despite the economic importance of BFB, little is known about the molecular mechanisms of A. citrulli pathogenesis (Burdman and Walcott, 2012). Recent studies have shown that important pathogenicity and virulence determinants of A. citrulli include the type II (T2SS) (Johnson, 2010), type III (T3SS) (Johnson et al., 2011), and type VI (T6SS) secretion systems (Tian et al., 2015); type IV pili (T4P) (Bahar et al., 2009a), polar flagella (Bahar et al., 2011), and quorum sensing (QS) (Wang T. et al., 2016). Most gram-negative, biotrophic, phytopathogenic bacteria rely on a functional T3SS to promote disease or trigger a hypersensitive response (HR) in susceptible and resistant plants, respectively (Büttner and Bonas, 2002). Importantly, gram-negative phytopathogenic bacteria secrete type 3-secreted effectors (T3Es) directly into plant host cells via the T3SS (Block et al., 2008;Van Engelenburg and Palmer, 2010;Nomura et al., 2011;Jiang et al., 2013;Xin et al., 2016). These T3Es may function to overcome pathogen-associated molecular patterntriggered immunity (PTI) and effector-triggered immunity (ETI) (Cui et al., 2013) or promote effector-triggered susceptibility (Jehle et al., 2013). The genes encoding the components of the T3SS are named hypersensitive response and pathogenicity (hrp) genes (Kim et al., 2003;Tampakaki et al., 2010). The hrp genes are located in large clusters, generally 20-25 kb in size (Büttner and Bonas, 2002). On the basis of gene organization, sequence, and regulation, hrp clusters can be divided into two classes: class I contains clusters of Pseudomonas syringae and enteric plant-pathogenic bacteria, whereas class II contains the hrp genes of Xanthomonas spp. and Ralstonia solanacearum (Bogdanove et al., 1996;Büttner and Bonas, 2002). Analysis of the genome sequence of the A. citrulli group II strain AAC00-1 revealed the presence of a hrp-T3SS (Burdman and Walcott, 2012). Moreover, sequence and cluster organization analyses showed that the A. citrulli hrp cluster belongs to class II (Burdman and Walcott, 2012). The hrp genes are induced in plant leaves and in T3SS-inducing, nutrient-poor medium. Generally, T3SS is suppressed in rich medium (Schulte and Bonas, 1992;Clarke et al., 2010). Gram-negative bacterial plant pathogens of the genera Erwinia, Pseudomonas, Ralstonia, and Xanthomonas require a fully active T3SS for pathogenicity, and the recent establishment of an inducing medium has facilitated the study of hrp gene expression and regulation in in-vitro experiments (Yuan and He, 1996;Rossier et al., 1999;Murata et al., 2006;Ancona et al., 2015).
In Xanthomonas spp., the expression of hrp genes is controlled by two regulators: HrpG and HrpX. HrpG is an OmpR family regulator that activates the expression of hrpX in X. campestris pv. vesicatoria (Wengelnik et al., 1996b) and in X. campestris pv. campestris (Huang et al., 2009). HrpX, an AraC-type transcriptional activator, controls the expression of other hrp genes and some effector genes (Wengelnik et al., 1996a). The HrpG regulon regulates the expression of hrp genes, as shown by cDNA-amplified fragment length polymorphism analysis (Noël et al., 2001). Regulated genes include the hrp gene cluster, effector genes, and genes encoding proteases. Thirty HrpG-induced and five HrpG-repressed genes have been identified in X. campestris pv. vesicatoria. Phosphorylated HrpG is predicted to activate the expression of hrpX (Büttner and Bonas, 2010;Tampakaki et al., 2010). HrpX binds to a conserved cis-regulatory element called the plant-inducible promoter (PIP) (TTCGC-N15-TTCGC) or a less conserved PIPlike motif (TTCGC-N8-TTCGT) in promoter regions associated with Xanthomonas virulence (Fenselau and Bonas, 1995). However, the mechanism differs from that of the HrpL/HrpSdependent systems in Erwinia carotovora and Pseudomonas spp., which belong to class I hrp clusters (Tampakaki et al., 2010). Recently, Guo et al. (2011) showed that hrpG and hrpX in Xanthomonas axonopodis pv. citri are involved in the regulation of multiple physiological functions, such as biofilm formation, by genome-wide microarray analyses. Thus, hrp gene regulation may be involved in a wide array of functions.
A. citrulli relies on a functional T3SS for pathogenicity (Ren et al., 2009;Burdman and Walcott, 2012;Eckshtain-Levi et al., 2014). Unfortunately, not much is known about the T3SS and T3Es in A. citrulli as compared to other plantpathogenic bacteria (Burdman and Walcott, 2012) because of the lack of a T3SS-inducing medium for A. citrulli until recently (Chen, 2016). Recent research has focused mainly on the characterization of genetically distinct groups of A. citrulli strains and on host preference (Silva et al., 2016;Zivanovic and Walcott, 2017). In addition, the arsenal of T3S effectors in A. citrulli strains were reported to be related to host preference on melon fruits (Yan et al., 2017). More specifically, group I A. citrulli strains can cause water-soaked lesion on detached immature watermelon rinds, while group II strains cannot. Previously, we reported that the A. citrulli group II strain Aac5, which was isolated from watermelon in Taiwan also relies on the T3SS for pathogenicity. As expected, Aac5 lost the ability to infect watermelon when the homologous hrcN gene was deleted (Yan et al., 2015). In addition, biofilm formation was crucial for Aac5 virulence and may be related to the T3SS (Wang T. et al., 2016). Based on sequence analysis of A. citrulli strain AAC00-1, hrpX and hrpG homologs were identified. We hypothesized that the hrpG and hrpX homologs in A. citrulli strain Aac5 may function as T3SS regulators. Hence, the objective of this study was to characterize the roles of these homologs in hrp gene regulation and pathogenicity.

Molecular Manipulations
Molecular manipulations were carried out using standard procedures (Sambrook and Russel, 2001

Construction of hrpG and hrpX Gene Deletion Mutants and Complementation
Markerless hrpG and hrpX mutants of A. citrulli strain Aac5 were generated through allele exchange (Ren et al., 2014). Primers were designed on the basis of A. citrulli AAC00-1 sequences for Aave_0445 and Aave_0444, which encode hrpG and hrpX homologs, respectively, and their flanking regions. The flanking regions were cloned from Aac5 and were aligned to those of to AAC00-1 using DNAMAN version 5.2.2 (Lynnon Biosoft, Quebec, Canada), and the sequence similarity was 100%. For the hrpG gene (Aave_0445), the upstream and downstream flanking regions were amplified by PCR using the primer pairs hrpG-1F/hrpG-1R and hrpG-2F/hrpG-2R (Supplementary Table 1), respectively. The two fragments were concatenated by overlap PCR (Zhang et al., 2013) and subsequently ligated into the suicide plasmid, pk18mobsacB, to generate pk18hrpG. pk18hrpG was transformed into A. citrulli strain Aac5 by tri-parental mating with an E. coli strain carrying the helper plasmid, pRK600. Single crossover colonies were selected on KB plates containing Amp and Km. Individual transformants were then sub-cultured continuously in KB liquid medium containing Amp alone to promote double crossover homologous recombination. The hrpG deletion mutant, hrpG, which was resistant to Amp but sensitive to Km, was confirmed by PCR using the primer pairs hrpG-TF/hrpG-TR, hrpG-1F/hrp-2R, km-F/km-R, and WFB1/WFB2 (Supplementary Table 1).
To construct a complementation vector, the complete hrpG gene (1,074 bp) including its native promoter, was amplified with primers HBhrpG-F/HBhrpG-R (Supplementary Table 1) and inserted into the shuttle vector, pBBRNolac-4FLAG, to create pBBRHBhrpG. The pBBRNolac-4FLAG was constructed by deleting the lac promoter of pBBR1MCS-2 using a KOD-Plus-Mutagenesis Kit (Toyobo, Shanghai, China), and inserting a 4× FLAG tag into the vector with C-terminal fusion expression, so that only the native promoter drove target gene expression. The complementation vector pBBRHBhrpG was introduced into the hrpG mutant to generate the complemented strain hrpGcomp. HrpG expression was verified in the complemented strain by western blotting (Sambrook and Russel, 2001). Similar methods were used to generate hrpX and hrpX-comp mutant strains from A. citrulli Aac5.

Roles of hrpG and hrpX in A. citrulli Virulence and Hypersensitive Response Induction
To determine the roles of hrpG and hrpX in seed-to-seedling BFB transmission, seed transmission assays were performed as described previously (Bahar et al., 2009b), with slight modifications. The WT, hrpG, and hrpX Aac5 strains, were cultured in liquid KB medium and adjusted to 5 × 10 6 CFU/mL with sterilized distilled water. Watermelon seeds (cv. Ruihong, n = 24) were soaked in a bacterial cell suspension (∼5 × 10 6 colony-forming units [CFU]/mL) at room temperature for 2 h with continuous agitation at 60 rpm using a rotary shaker (Ecotron; Infors HT, Bottmingen, Switzerland). The inoculated seeds were then air-dried for 24 h and sown in plastic pots (Rongxiangruihe, Beijing, China, six seeds per pot) filled with a potting mix (vermiculite, soil, and peat at a 1:1:1 ratio). The pots were placed in a growth chamber for 12 days under the following conditions: 25-30 • C; 65% mean relative humidity, and 16 h light/8 h darkness cycle during the experiments. BFB severity was visually assessed 12 days after seedling emergence, as described previously (Bahar et al., 2009b).
To evaluate the effects of hrpG and hrpX deletion on HR induction, HR assays were performed by injecting a bacterial cell suspensions into the leaves of 3-week-old tobacco (Nicotiana tabacum), as described previously (Ren et al., 2014). Leaves were visually examined for HR (tissue collapse) 24 h after infiltration. Electrolyte leakage was quantified for detecting HR on 3-weekold tobacco leaves injected with wild-type A. citrulli Aac5, hrpG, and hrpX, and the respective complementation strains, as described previously (Stork et al., 2015). In brief, the strains were inoculated into tobacco leaves, and four leaf disks from each treatment were harvested after 24 h, washed, and floated on 5 ml of distilled water for 4 h with gentle shaking prior to measuring the conductivity of the bath water. As a negative control (mock treatment), 10 mM MgCl 2 was used.

Assay for Reactive Oxygen Species (ROS)
ROS were measured by chemiluminescence as described previously , with slight modifications. N. benthamiana leaves were inoculated with cell suspensions of A. citrulli Aac5, hrpG, and hrpX, and respective complementation strains (∼10 8 CFU/mL). P. syringae pv. tomato DC3000 was used as a negative control. At 24 hpi, leaf disks (0.4 cm in diameter) were excised and placed into wells of 96-well plates pre-loaded with 100 µL SDW. Subsequently, 100 µL of 0.5 mM L-012 (Wako, Guangzhou, China) in 10 mM MOPS-KOH buffer (pH 7.4) were added to each well. The intensity of ROS generation was determined as described by Liu et al. (2015). N. benthamiana leaf disks infiltrated with P. syringae pv. tomato DC3000 were used as a negative control.

Evaluation of Biofilm Formation
To determine the roles of hrpG and hrpX in biofilm formation, assays were performed using A. citrulli wild-type Aac5, hrpG, and hrpX, and respective complementation strains, as previously described (Bahar et al., 2009a), with slight modifications. Briefly, 24-well plates (Corning, NY, USA) were pre-loaded with liquid T3SS-inducing medium and inoculated with a 1:1,000 dilution of overnight cultures of A. citrulli Aac5, hrpG, and hrpX, and the respective complementation strains. The plates were incubated in a tilted position at 28 • C for 48 h, without agitation. Then, 0.1% crystal violet was added to each well for 30 min, after which the wells were washed with distilled water. Biofilm formation for each strain was compared quantitatively by solubilizing the stained biofilms with 100% ethanol and measuring the OD 590 of the stained-cell suspensions with a spectrophotometer.

Quantitative Reverse-Transcription (qRT-)PCR
Total RNA was extracted from wild-type A. citrulli Aac5, hrpG, and hrpX cultured in T3SS-induced liquid medium up to OD 600 = 0.45, using TRIzol reagent (Invitrogen, Waltham, MA, USA). Contaminant DNA was digested and cDNA was synthesized using ReverTra Ace qPCR RT MasterMix with gDNA Remover (Toyobo, Shanghai, China). The mRNA levels were quantified by qPCR using KOD SYBR qPCR Mix (Toyobo, Shanghai, China) on an Applied Biosystems 7500 instrument (ABI, Waltham, MA, USA). The primers used for qRT-PCR are listed in Supplementary Table 1. Each sample was assayed in triplicate. rpoB was used as a reference gene. Relative expression of genes of interest was calculated by the 2 − CT method (Livak and Schmittgen, 2001).

β-Glucuronidase (GUS) Reporter Activity Assay For Detecting Promoter Activity
To construct a GUS reporter vector, the GUS gene was inserted into PBBRNolac (without 4×FLAG-tag) to generate the PBBRNolacGUS vector using restriction enzymes XbaI and HindIII. The native promoter sequence of the putative T3S effector Aac5_2166 (GenBank accession MG879253) was cloned from genomic DNA of A. citrulli Aac5 by PCR, and the product was ligated into PBBRNolacGUS to generate pBBR2166GUS. pBBR2166GUS and PBBRNolacGUS (negative control vector) were transformed into Aac5, hrpG, and hrpX to generate WT-2166GUS, hrpG-2166GUS, hrpX-2166GUS, and WT-GUS strains (negative control strain). Promoter activity was detected as previously described (Zhang et al., 2013).

Assay of HrpG and HrpX Protein Production by A. citrulli Strains in T3SS-Inducing Media
HrpX and HrpG protein expression was quantified by western blotting. A. citrulli strains (OD 600 = 1.0) were pre-incubated for ∼12-18 h in KB broth and washed twice. Twenty microliters of the bacterial suspension were transferred into 10 mL of T3SS-inducing medium and incubated at 28 • C until the OD 600 reached 0.5. To extract intracellular protein, 4 mL of the bacterial suspension as cell lysate was treated with Protease Inhibitor Cocktail (Bimake, Shanghai, China). The lysate was centrifuged at 12,000 × g and 4 • C for 3 min, and the cells were resuspended in 200 µL of 4× Laemmli Sample Buffer (Bio-Rad, Beijing, China) and heated for 10 min.

Statistical Analysis
All experiments were conducted three times. Six biological replicates were used in each biofilm assay, electrolyte leakage quantification assay, and promoter activity. Twelve biological replicates were used for ROS detection. Twenty-four leaves as biological replicates were used for assessing BFB severity. For the other assays, three biological replicates were evaluated. Data were analyzed by one-way analysis of variance (ANOVA) and Tukey's honest significant difference (HSD) tests. For qRT-PCR, data were analyzed by the independent-samples t-test. Statistical analyses were conducted using SPSS version 17.0 (SPSS, Chicago, IL, USA) and GraphPad PRISM 5.0 software (GraphPad Software, La Jolla, CA, USA). Differences with p < 0.05 were considered significant.

Sequence Analysis of hrpG and hrpX Homologs in A. citrulli Strain Aac5
The group II A. citrulli strain Aac5 has been routinely used in studies on pathogenicity (Wang T. et al., 2016). BLAST searches using the AAC00-1 genome (GenBank accession NC_008752) identified Aave_0445 and Aave_0444, which encode response regulator receiver protein (A1TJB4), which showed 100% amino acid sequence identity with HrpG (Q5EF45), and transcriptional regulator, AraC family (A1TJB3), which showed 100% amino acid sequence identity with HrpX (Q5EF44) in UniProt, respectively. The predicted HrpG (Q5EF45) and HrpX (Q5EF44) in UniProt were also from A. citrulli. To verify whether homologous genes coding HrpG and HrpX exist in strain Aac5, we cloned the sequences from Aac5 and found that they were 100% identical to the sequences of Aave_0445 and Aave_0444 from AAC00-1, named hrpG (GenBank accession no.: MG879251) and hrpX (MG879252), respectively. Amino acid sequence alignment to AAC00-1 using DNAMAN version 5.2.2 (Figures 1A,B) revealed 100% identity. PCR with the primer pair hrpG-TF/hrpG-TR yielded a 417-bp amplicon when using DNA template from wild-type Aac5, but not when the DNA was from the hrpG mutant. DNA from wild-type Aac5 and hrpG yielded 2,249-and 3,027-bp PCR amplicons, respectively, with the primer pair hrpG-1F/hrp-2R, and no amplicons were generated with the primer pair km-F/km-R. To confirm the identity of the wild-type Aac5 and hrpG mutant, a 360bp amplicon was produced with the diagnostic primer pair WFB1/WFB2. The hrpX gene deletion mutant was confirmed using the same approach.

Roles of hrpG and hrpX in A. citrulli Biofilm Formation
Since biofilm formation is involved in A. citrulli virulence and pathogenicity (Bahar et al., 2009a;Tian et al., 2015; Wang T. et al., 2016), we investigated the roles of hrpG and hrpX in biofilm formation. To rule out interference by plasmid introduction, we additionally transformed the empty vector, pBBRNolac-4FLAG, into wild-type A. citrulli Aac5 (WT-EV), hrpG ( hrpG-EV), and hrpX ( hrpX-EV). Biofilm formation was quantified by crystal violet staining, which showed that hrpG-EV and hrpX-EV exhibited significantly enhanced biofilm formation as compared with Aac5. The hrpG-comp strain produced significantly less biofilm than hrpX-EV, but significantly more than the wild-type strain (Figure 2). hrpX-comp produced significantly more biofilm than the wild-type strain.
hrpG and hrpX Mutants of A. citrulli Fail to Induce HR in N. tabacum To determine whether hrpG and hrpX are involved in HR induction by A. citrulli, we examined HR induction by the mutant and wild-type strains in N. tabacum, a non-host plant typically used to test HR induction (Ren et al., 2009). The strains hrpGcomp and hrpX-comp induced an HR in N. tabacum similar to that induced by the WT-EV strain after 24 h. In contrast, hrpG-EV and hrpX-EV did not elicit an HR (Figure 3). These observations suggested that Aac5 requires hrpG and hrpX to induce an HR in N. tabacum.
To substantiate these results, we used electrolyte leakage analysis. As expected, hrpG-comp and hrpX-comp induced electrolyte leakage to levels similar to those induced by WT-EV. In contrast, hrpG-EV and hrpX-EV induced significantly lower levels of electrolyte leakage, similar to that induced by mock treatment and 3-fold lower than the level induced by WT-EV (Figure 4). Overall, these findings revealed that Aac5 requires hrpG and hrpX to induce an HR in N. tabacum.

hrpG and hrpX Negatively Regulate PTI-Associated ROS Generation
As hrpG and hrpX were found to be essential for HR induction by A. citrulli, we investigated whether PTI-associated ROS was altered during the A. citrulli-host interaction. Thus, we examined the ROS levels induced by the A. citrulli strains in N. benthamiana leaves using chemiluminescence . hrpG and hrpX elicited significantly higher levels of ROS (∼12-fold and ∼8-fold, respectively) than the WT strain, which induced ROS to a level similar to that of DC3000 ( Figure 5). These results indicated that hrpG and hrpX act as negative regulators of PTI-associated ROS generation in the interplay between A. citrulli and N. benthamiana.

hrpG and hrpX Contribute to A. citrulli Virulence on Watermelon Seedlings
Based on the above observations, we hypothesized that hrpG and hrpX are critical for A. citrulli virulence. We tested this hypothesis by conducting seed transmission and seedling infiltration assays in watermelon. The mean BFB severity index upon inoculation of hrpG and hrpX was similar to that of mock treatment, but ∼7-fold lower than that upon inoculation of the WT strain ( Figure 6A). Figure 6B shows that the WT strain caused more severe BFB symptoms than the hrpG and hrpX mutants. In the tissue infiltration assays, we observed that hrpG and hrpX did not cause BFB symptoms on watermelon cotyledons between 24 and 96 hpi. In contrast, the WT and complemented strains elicited BFB symptoms by 36 hpi (Supplementary Images 1A,B). Overall, these results showed that hrpG and hrpX contribute to A. citrulli virulence on watermelon seedlings. FIGURE 2 | Effects of hrpG and hrpX genes on Acidovorax citrulli biofilm formation. The biofilm formation of A. citrulli in 24-well multi-dishes as quantified using crystal violet staining. WT-EV, wild-type strain Aac5 transformed with empty vector, pBBRNolac-4FLAG; hrpG-EV, hrpG mutant transformed with empty vector, pBBRNolac-4FLAG; hrpX-EV, hrpX mutant transformed with the empty vector pBBRNolac-4FLAG; hrpG-comp, hrpG complemented with vector, pBBRHBhrpG (native promoter); hrpX-comp, hrpX complemented with the vector, pBBRHBhrpX (native promoter). Each column shows the mean and standard deviation. The experiment was conducted three times. Averages and standard deviations from one of three experiments with similar results are shown. Different letters above bars indicate statistically significant differences as determined by one-way analysis of variance (ANOVA) and Tukey's honest significant difference (HSD), p < 0.05.

Interaction Between HrpG and HrpX
To evaluate the interaction between HrpG and HrpX, we assessed their translation levels by western blot analysis using anti-HrpX and anti-HrpG antibodies. Total protein was extracted from Aac5, hrpG, and hrpX cultured in T3SS-inducing medium. As shown in Figure 7A, HrpX expression was not detected in hrpG when compared to the wild-type strain (as expected, HrpX was not detected for hrpX), while HrpG expression in hrpX was similar to the WT strain (as expected, HrpG was not detected for hrpG). These findings indicated that HrpG might be the upstream regulator of HrpX in A. citrulli.
qRT-PCR revealed that the HrpX transcript levels in the hrpG mutant were ∼7-fold lower than those in the WT strain (Figure 7B), whereas hrpG transcript levels were not affected in the hrpX mutant as compared to the WT strain ( Figure 7C). These results indicated that hrpG regulates the expression of hrpX, but hrpX does not regulate the expression of hrpG.

hrpG and hrpX Regulate the Expression of T3Es
Based on the above results, we hypothesized that the expression of T3Es in A. citrulli is regulated by HrpG and HrpX. To test this hypothesis, we used qRT-PCR to compare the transcript levels of T3Es in hrpG, hrpX, and the WT strain in T3SS-inducing medium. We transiently expressed Aave_2166 homolog, encoding the YopJ homolog (Eckshtain-Levi et al., hrpG-comp, hrpX-comp were inoculated into the N. tabacum leaves at a concentration of OD 600 = 0.3 and incubated for 24 h. Ten millimolars of MgCl 2 was used as a negative control. The photographs were obtained using EOS 70D (Canon, Japan). The experiment was conducted three times.
2014; Traore, 2014), as a putative effector inducing programmed cell death in N. benthamiana (Supplementary Image 2) in WT, hrpG, and hrpX. Since the gene cloned from Aac5 has 100% sequence identity with the sequence of Aave_2166, it was named Aac5_2166 (GenBank accession MG879253). We then measured the transcript levels of Aac5_2166 in hrpG, hrpX, and the WT strain. The mRNA level of Aac5_2166 was significantly lower in hrpG and hrpX (∼3-fold and ∼2fold, respectively) than in the WT strain. These results indicated that hrpG and hrpX regulate the expression of Aac5_2166 ( Figure 8A).
To verify these results, we analyzed the promoter activity of Aac5_2166. We constructed a reporter vector, pBBRNolacGUS to measure promoter activity. In pBBRNolacGUS, the constitutive lac promoter was deleted and the GUS reporter gene was inserted, which is commonly used to analyze promoter activity in bacteria (Zhang et al., 2013;Wang L. et al., 2016). The results showed that the promoter activities of Aac5_2166 in the hrpG-2166GUS and hrpX-2166GUS strains were significantly lower (∼4-fold and ∼6-fold lower, respectively) than in WT-2166GUS, but similar to that in mock-treated WT-GUS. These data indicated that hrpG and hrpX regulate Aac5_2166 promoter activity ( Figure 8B).
We speculated that the secretion of T3Es might be impaired in the hrpG and hrpX mutants. To test this hypothesis, we cloned Aac5_2166 and its native promoter and inserted them into pBBRNolac-4FLAG to concatenate a C-terminal fusion 4 × FLAG-tag. This construct was transformed into A. citrulli Aac5 and hrpG to generate WT-2166 and hrpG-2166. No signal was detected in the cell lysates and supernatants of hrpG-2166, while a strong signal was detected for the WT-2166 FIGURE 4 | Electrolyte leakage quantification assay for verification of hypersensitive response (HR) in Nicotiana tabacum induced by Acidovorax citrulli strains. Electrolyte leakage quantification for the detection of HR. Values are mean conductivity of bathing water (µS mm −2 leaf ml −1 bath water) ± standard deviation (SD) (n = 3 plants). Different letters above bars indicate statistically significant differences as determined by one-way analysis of variance (ANOVA) and Tukey's honest significant difference (HSD), p < 0.05.
FIGURE 5 | Effects of hrpG and hrpX gene on Acidovorax citrulli inducing ROS on Nicotiana benthamiana. Reactive oxygen species (ROS) generation by the wild-type A. citrulli strain Aac5 and the hrpG and hrpX deletion mutants. ROS assay results are presented as mean and standard deviations based on three replicates from three different plants. Different letters above bars indicate statistically significant differences as determined by one-way analysis of variance (ANOVA) and Tukey's honest significant difference (HSD), p < 0.05. cell lysates and supernatants (Supplementary Image 3). These results indicated that HrpG regulates the expression of Aac5_2166 and that Aac5_2166 could not be synthesized in the hrpG mutant.

DISCUSSION
The mechanisms by which pathogenic bacteria regulate T3Es are critical for successful infection (Tampakaki et al., 2010) and have been well characterized in some plant-pathogenic bacteria.
FIGURE 6 | Contributions of hrpG and hrpX gene to Acidovorax citrulli virulence. Bacterial fruit blotch severity on watermelon induced by wildtype (Aac5), hrpG and hrpX deletion mutants of A. citrulli in seedling transmission assays. Disease severity was assessed according to (Bahar et al., 2009b). The experiment was conducted three times. (A) Disease severity of plants inoculated with different strains. The 10 mM MgCl 2 was used as a mock treatment. Each column shows the mean and standard deviation. Different letters above bars indicate statistically significant differences as determined by one-way analysis of variance (ANOVA) and Tukey's honest significant difference (HSD), p < 0.05; (B) representative picture was taken after 12 days by EOS 70D (Canon, Japan).
For example, hrpG and hrpX regulate T3Es in Xanthomonas spp. (Xue et al., 2014). Moreover, hrpG, as an OmpR family regulator, is typically regulated as a core gene of the T3SS by phosphorylation. OmpR family regulators, as response regulator (RR) proteins, belong to the two-component systems that consist of a histidine kinase and an RR protein, which serve as a basic stimulus-response coupling mechanism to allow organisms to sense and respond to changes in environmental conditions (Tampakaki et al., 2010;Li et al., 2014). The histidine kinase transfers a phosphoryl group to the RR in a reaction catalyzed by the RR. This phosphotransfer to the RR results in the activation of a downstream effector domain that elicits a specific response (Stock et al., 2000;Laub and Goulian, 2007). In the HrpG-HrpX pathway, HrpG activates HrpX, an AraC-type transcriptional activator, to regulate T3E expression. Unfortunately, the mechanisms that regulate the T3SS in A. citrulli are poorly understood. However, based on the similarity of the hrp clusters of A. citrulli and Xanthomonas spp. (Tampakaki et al., 2010;Burdman and Walcott, 2012), we hypothesized that a similar mechanism exists in A. citrulli. This mechanism likely involves genes homologous to those encoding the core proteins that regulate the T3SS in Xanthomonas spp. Genome sequence analyses identified two genes from the AAC00-1 genome (GenBank accession NC_008752) that FIGURE 7 | The interaction of hrpG and hrpX in Acidovorax citrulli. (A) Western blotting analysis showed that HrpG activated HrpX. The WT, hrpG, and hrpX strains were cultured in liquid T3SS-inducing medium up to OD 600 = 0.5. Total cell extracts were analyzed by SDS-PAGE and immunoblotting, using specific antibodies. The experiment was performed three times. (B) Represents real-time quantitative RT-PCR to determine hrpX mRNA levels. hrpX expression was analyzed through RT-qPCR in wild-type strains Aac5 and the hrpG mutant strain, using specific primers. Bacteria were grown in liquid T3SS-inducing medium to OD 600 = 0.45 and harvested to extract total RNA. (C) Represents real-time quantitative RT-PCR to determine hrpG mRNA levels. hrpX expression was assayed through RT-qPCR in wild-type strain Aac5 and the hrpX mutant, using specific primers. All the data shown are means ± standard deviation of duplicate samples from one representative experiment and are reported as fold induction relative to expression of WT. The rpoB gene was used as a reference gene. Each column shows the mean and standard deviation. The experiment was performed three times, with similar results obtained each time. Asterisks above bars indicate significant differences as determined by t-test, p < 0.05.
were homologs of hrpG and hrpX, designated Aave_0445 and Aave_0444. We cloned the homologs of these genes from A. citrulli strain Aac5 (Wang T. et al., 2016).
In the past two decades, researchers have shown that hrpG and hrpX regulate the expression of T3SS (Wengelnik et al., 1996b;Noël et al., 2001;Huang et al., 2009). However, these molecules also have other functions. Guo et al. (2011) showed that hrpG and hrpX play global roles in coordinating different virulence traits of X. axonopodis pv. citri, including traits involved in biofilm formation. Wang T. et al. (2016) showed that in Aac5, the transcription of hrpE, a core gene of T3SS, consistently is altered when biofilm formation changes, indicating that biofilm formation may be regulated by T3SS. The results of the current study confirm that hrpG and hrpX regulate biofilm formation in A. citrulli strain Aac5.
During the early stages of infection, plant-pathogenic bacteria colonize the host with the aid of biofilm formation. When hrpG and hrpX were deleted in A. citrulli, we observed significant changes in biofilm formation and virulence. As expected, the hrpG and hrpX mutants lost the ability to induce an HR in N. tabacum, which was confirmed by electrolyte leakage assays. Furthermore, pathogenicity on watermelon seedlings was lost in the mutants. Pathogenic bacteria must overcome the plant's immune systems, including PTI and ETI, to achieve successful infection (Zhang et al., 2007). Physiological markers for PTI typically include ROS generation (Boller and Felix, 2009), and the ETI pathway commonly results in localized, rapid PCD (Stork et al., 2015). The T3SS is essential for inducing a defense response in the host (Boller and Felix, 2009). Bacterial resistance mechanisms involve the secretion of T3Es, which come into contact with host-cell receptors and result in suppression of the host immune system (Dickman and Fluhr, 2013;Jiang et al., 2013). We observed that hrpG and hrpX elicited high-level ROS production, indicating that proteins encoded by these genes affect PTI. Additionally, expression of the putative T3E gene Aac5_2166, which encodes a YopJ homolog, was downregulated at the transcriptional level-via alterations in promoter activityand at the translational level in the hrpG and hrpX mutants. Combined with the observation that the mutants lost the ability to induce an HR in N. tabacum, we concluded that the reduction in virulence in the hrpG and hrpX mutants was because they could not secrete T3Es and therefore, could not suppress PTI and ETI. This suggests that hrpG and hrpX play key roles in the regulation of T3Es.
Our findings demonstrate that A. citrulli strain Aac5 requires hrpG and hrpX to regulate virulence factors. hrpG and hrpX contribute to the expression of the T3SS and are involved in biofilm formation in this strain. Importantly, our findings FIGURE 8 | Effects of hrpG and hrpX genes on the regulation of type 3 effectors in Acidovorax citrulli. (A) Real-time quantitative RT-PCR was used to determine Aac5_2166 mRNA levels. Putative type 3 secreted effector (T3E) Aac5_2166 gene expression was assayed through RT-qPCR in wild-type strain Aac5 and the hrpG and hrpX deletion mutant strains, using specific primers. The data shown are means ±standard deviation of duplicate samples from one representative experiment and are reported as fold induction relative to expression of WT. The rpoB gene was used as a reference gene. The experiment was performed three times, with similar results obtained each time. Asterisks above bars indicate significant differences as determined by t-test, p < 0.05; (B) The detection of promoter activity of the putative T3E Aac5_2166 was determined by assaying GUS activity. The pBBR2166GUS was constructed with Aac5_2166 native promoter and GUS reporter gene. The wild-type strain Aac5 transformed with vector pBBR2166GUS carrying the GUS reporter gene, WT-2166GUS; the hrpG mutant transformed with vector pBBR2166GUS carrying the GUS reporter gene, hrpG-2166GUS; the hrpG mutant transformed with vector pBBR2166GUS carry GUS reporter gene, hrpX-2166GUS; the wild-type strain Aac5 transformed with pBBRNolacGUS without promoter sequence, WT-GUS (negative control). Each column shows the mean and standard deviation. All experiments were performed three times, with similar results obtained each time. Different letters above bars indicate significant differences as determined by one-way analysis of variance (ANOVA) and Tukey's honest significant differences test, p < 0.05. provide, for the first time, strong evidence that hrpG and hrpX play key roles in A. citrulli Aac5 pathogenicity. Though strain Aac5 belongs to the group II strains of A. citrulli (Yan et al., 2013;Wang T. et al., 2016), on the basis of a genome sequence for the group I strain pslb65 reported by Wang et al. (2015) (GenBank accession JYHM00000000), it seems that the T3SS system structures of group I and group II strains are homologous. Moreover, key A. citrulli T3Es may be identified in hrpG and hrpX mutants using transcriptome analysis, which will improve our understanding of the molecular mechanisms of A. citrulli pathogenicity.

AUTHOR CONTRIBUTIONS
XZ and TZ designed the research and wrote the paper. XZ, JY, and LY executed the experiments. XZ, MZ, YY, and WG performed the data analyses. XZ, RW, TZ, and MZ critically reviewed the manuscript. All authors read and approved the final manuscript. bacterial cell supernatants were treated with Protease Inhibitor Cocktail (Bimake), separated by centrifugation (Liu et al., 2017), and the proteins were precipitated with 10% trichloroacetic acid (Sigma, Shanghai, China). The intracellular and secreted protein extracts were analyzed by SDS-PAGE and immunoblotting using specific antibodies. The experiment was conducted three times.