A LuxR‐type regulator, AcrR, regulates flagellar assembly and contributes to virulence, motility, biofilm formation, and growth ability of Acidovorax citrulli

Abstract LuxR‐type regulators regulate many bacterial processes and play important roles in bacterial motility and virulence. Acidovorax citrulli is a seedborne bacterial pathogen responsible for bacterial fruit blotch, which causes great losses in melon and watermelon worldwide. We identified a LuxR‐type, nonquorum sensing‐related regulator, AcrR, in the group II strain Aac‐5 of A. citrulli. We found that the acrR mutant lost twitching and swimming motilities, and flagellar formation. It also showed reduced virulence, but increased biofilm formation and growth ability. Transcriptomic analysis revealed that 394 genes were differentially expressed in the acrR mutant of A. citrulli, including 33 genes involved in flagellar assembly. Our results suggest that AcrR may act as a global regulator affecting multiple important biological functions of A. citrulli.

luxR genes not associated with a luxI gene are responsible for modulating the expression of virulence factors, biofilm formation, and host immune responses (Fuqua et al., 2001;Santos et al., 2012).
In Salmonella enterica, a LuxR-type transcriptional regulator, RflM, harbours a conserved HTH domain at the C-terminal, whereas it does not contain any domain at the N-terminal. RflM negatively regulates the expression of flhDC, which encodes FlhD 4 C 2 , the flagellar master regulatory complex, which, in turn, affects the flagellar assembly (Singer et al., 2013).
Bacterial fruit blotch is a seedborne disease that causes significant yield losses in melon and watermelon worldwide. Its causal agent is Acidovorax citrulli, a gram-negative bacterium with a single polar flagellum (Bahar and Burdman, 2010). Based on DNA-fingerprinting profiles, whole-cell fatty-acid composition, carbon-source utilization, pathogenicity assays, pulsed-field gel electrophoresis, and multilocus sequence typing, strains of A. citrulli are divided into two groups: group I includes strains isolated mainly from nonwatermelon hosts and are pathogenic to most cucurbit hosts, and group II is composed of strains isolated mainly from watermelon and more aggressive to watermelon than the group I strains (Walcott et al., 2004;Burdman et al., 2005;Feng et al., 2009;Yan et al., 2013). The role of QS in A. citrulli has been studied in both the group I strain XJL12 and the group II strain Aac-5 (Chen et al., 2009;Wang et al., 2016). Mutants of both the luxI/luxR and aacI/aacR showed reduced virulence and motility, indicating the important role QS plays in A. citrulli (Chen et al., 2009;Wang et al., 2016).
In addition to luxR and aacR, many other LuxR-type regulators in the genome of A. citrulli have been identified in our preliminary studies (data not shown). In this study, we explored the regulatory role of one of the LuxR-type regulators, AcrR, in the group II strain Aac-5 of A. citrulli by mutational analysis and RNA sequencing. Our results showed that deletion of acrR affected flagellar biosynthesis, cell motility, biofilm formation, and virulence of Aac-5 of A. citrulli.
In addition, our RNA sequencing (RNA-Seq) results revealed that the expression of many genes involved in various functions of strain Aac-5 of A. citrulli, including flagellar assembly, were changed in acrR deletion mutant strain, further suggesting that acrR functions as a global transcriptional regulator, especially for flagella-related functions in A. citrulli.

| Identification of AcrR, a putative LuxR-type regulator of A. citrulli, and generation of acrR mutant and complemented strains
An open reading frame (ORF) containing 209 amino acids was identified in the genome sequence of the A. citrulli group II strain AAC00-1 (GenBank accession number CP000512.1), located from nucleotide 4874913 to 4875542, with the locus tag of Aave_4382.
It has a typical domain structure of the LuxR-family response regulators: a receiver domain at the N-terminus and an HTH DNAbinding domain at the C-terminus. In addition, a BLASTp search showed that this ORF has 100% amino acid sequence identity with the protein ADX48163.1 in Acidovorax avenae subsp. avenae ATCC 19860, annotated as a two-component transcriptional regulator in the LuxR family. The acrR has no similarities with the quorumsensing signal receptor aacR (located from nucleotide 4232080 to 4232808, with the locus tag of Aave_3810 in genome of AAC00-1, Wang et al., 2016) at the nucleotide or protein level. This suggests that the ORF in A. citrulli strain AAC00-1 is also a LuxRtype transcriptional regulator and was designated AcrR because it is an A. citrulli regulator that is different from aacR in the group II strain Aac-5 of A. citrulli.
The successful construction of the acrR mutant strain ∆acrR was confirmed by PCR amplification of strain ∆acrR with the LacrR-F and RacrR-R primers (Table 1) and subsequent sequencing of the PCR product. The product is 1,868 bp in size, containing 472-bp upstream and 541-bp downstream fragments of the acrR gene, separated by an 855-bp gentamicin cassette (data not shown). This was different from the PCR product of 1,643 bp in size that was amplified from the wildtype strain Aac-5, which contained the 630-bp acrR gene as well as the same 472-bp upstream and 541-bp downstream sequences.
The complementation strain ∆acrRcomp showed resistance to kanamycin, suggesting the successful transfer of the plasmid pBBR-acrR into ∆acrR (Table 2). The presence of pBBR-acrR in ∆acrRcomp was further confirmed by PCR amplification of the strain ∆acrRcomp with the primers acrR-F and acrR-R (Table 1)  to the wild-type strain Aac-5 ( Figure 1).
To determine whether the mutant is defective in its ability to grow in planta, we injected watermelon cotyledons with bacterial cell suspensions of the wild-type strain Aac-5, the mutant strain ∆acrR, and the complementation strain ∆acrRcomp, as well as sterile water as a negative control. No symptoms were observed in inoculated cotyledons 1 and 24 hr after inoculation (hai) in all treatments, whereas water-soaking necrosis appeared 48 hai, and lesions started to develop 72 and 96 hai ( Figure 2a) in cotyledons inoculated with the wild-type and complementation strains, but not in those inoculated with water and the mutant strain. The results from our quantitative bacterial in planta assay revealed that the populations of the Aac-5, the ∆acrR, and the ∆acrRcomp strains in cotyledons were not significantly different until 72 hai, when the populations of the ∆acrR and the ∆acrRcomp strain were significantly lower than that of the Aac-5 strain (Figure 2b). At 96 hai, the population of the ∆acrR strain remained similar to its population level at 72 hai but was significantly lower than that of the Aac-5 and the ∆acrRcomp strains ( Figure 2b).

| Biofilm formation and growth rate of the acrR mutant of A. citrulli were increased
The wild-type strain Aac-5 did not form any biofilm when measured both qualitatively and quantitatively in our study ( Figure 3).
When the acrR gene was mutated, however, the mutant strain ∆acrR formed a visible ring of biofilm on the inner wall of a flask, while no such ring was observed for the complementation strain ∆acrRcomp ( Figure 3a). This observation was confirmed by our quantitative biofilm assay, since the mean absorption value of the biofilm by ∆acrR was 2.23, while the absorption was below the detection levels for the wild-type strain Aac-5 and the complementation strain ∆acrRcomp (Figure 3b). Our results showed that the mutation of the acrR gene enhanced the biofilm formation of A. citrulli.
The growth ability of Aac-5, ∆acrR, and its complementation strain ∆acrRcomp was determined by measuring the optical density of cell suspensions incubated in King's B broth at 28 °C. The mutant strain ∆acrR was increased in growth ability, with the OD 600 value reaching 1.14 at 12 hr of incubation, and in the meantime the OD 600 value of the complementation strain ∆acrRcomp reached 0.53, whereas the wild-type strain Aac-5 had an OD 600 value of 0.11 at  Figure 4).

| The acrR mutant of A. citrulli lost the ability to twitch and swim, as a result of the loss of flagella formation
We compared the wild-type strain Aac-5 to its mutant strain ∆acrR and the complementation strain ∆acrRcomp for the formation of corrugated trajectories or halos around their colonies as each bacterium migrated via twitching motility. Strain Aac-5 produced typical corrugated haloes, while smooth haloes were produced by the ∆acrR strain and the ∆acrRcomp strain ( Figure 5a). These results show that the A. citrulli strain lost the twitching ability when its acrR gene was mutated.
Our assay for swimming motility revealed that the mutant strain ∆acrR did not spread 36 hr after inoculation of 10 µl of the bacterial suspension into the centre of a soft agar plate (0.3% agar), whereas the wild-type strain Aac-5 and the complementation strain ∆acrRcomp spread to approximately one quarter of the F I G U R E 1 Effect of acrR on virulence of Acidovorax citrulli on watermelon leaves. (a) Appearance of watermelon seedlings 10 days after inoculation with sterile water, the wild-type strain Aac-5, the acrR mutant strain ∆acrR, and the mutant complementation strain ∆acrRcomp of A. citrulli. (b) Virulence of A. citrulli strains 10 days after inoculation, calculated based on a disease index of 0 to 100 (Wang et al., 2016). The bars represent standard errors of the means from three experiments, each containing six inoculated watermelon seedlings in three pots per tested strain. **p < .01 by Student's t test After 36 hr incubation at 28 °C, the average diameter of the bacterial lawn was 2.97 cm for the wild-type strain Aac-5, significantly larger than 1.16 cm for the mutant strain ∆acrR, but was similar (c) Swimming motility measured by colony diameters of each strain on basal medium plate. The bars represent standard errors of the means from three experiments, and each experiment contained three replicates for each strain. **p < .01 by Student's t test in the wild-type and complemented strains but not in the mutant strain ( Figure 6).

| RNA-Seq revealed that the acrR gene plays a role in the flagellar formation of A. citrulli
Because AcrR is identical to an annotated LuxR-type regulator in A. avenae subsp. avenae, we explored the possible regulatory role played by AcrR by comparing the acrR mutant strain ∆acrR of A. citrulli with its wild-type strain Aac-5 through transcriptome and gene ontology analyses. A total of 394 genes were differentially expressed in the ∆acrR mutant compared to its wild-type strain, including 219 highly and 175 lowly expressed genes (Table S1) The RNA-Seq data also showed that deletion of the acrR gene affected many genes involved in flagellar biosynthesis. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis revealed that 33 DEGs were involved in flagellar biosynthesis, including 3 highly and 30 lowly expressed genes (Table 3). The lowly expressed genes encode critical components for flagellar assembly, including flagellin, flagella biosynthesis proteins, flagella L-g and P-ring proteins TA B L E 3 Differentially expressed genes involved in flagellar biosynthesis in Acidovorax citrulli ΔacrR strain compared to its wild-type strain Aac-5 Note: Gene ID refers to the locus_tag of a differentially expressed gene in A. citrulli ΔacrR strain compared to strain Aac-5, identified by hits in a BLASTn search against the strain AAC00-1 genome (GenBank accession number CP000512.1). FC, fold-change located in outer membrane and peptidoglycan layer, flagella cap, hook and basal-body rod related proteins, as well as transcriptional regulators FlhDC (Figure 8). The RNA-Seq results were consistent with the loss of flagella formation in the mutant strain when acrR was deleted ( Figure 6). In addition, 46 DEGs were found involved in ribosomal proteins, all of which were significantly highly expressed (Table S1).

| D ISCUSS I ON
In this study, we identified a LuxR-type transcriptional regulator AcrR in A. citrulli. We demonstrated that deletion of the acrR gene resulted in reduced virulence on watermelon seedlings, lost twitching and swimming motilities, and failure to form flagella, but increased biofilm formation and growth ability of the mutant strain compared to its wild-type strain of A. citrulli. Additionally, comparative RNA profiling analysis revealed that 394 genes were differentially expressed, including 33 involved in flagellar assembly, suggesting the regulatory role AcrR plays, especially in flagella-related functions of A. citrulli.
LuxR-type regulators are known to play a role not only in the QS system but also in other biological functions such as flagellar synthesis. For example, VisN and VisR are two regulators in Sinorhizobium meliloti that belong to the LuxR family and act as global regulators of chemotaxis, flagellar, and motility genes (Sourjik et al., 2000). As demonstrated in our study, the AcrR of A. citrulli either positively or negatively regulated multiple biological functions of A. citrulli, because the acrR mutant strain lost swimming motility and failed to twitch and form flagella but increased in biofilm production and growth ability.
The bacterial flagellum is an organelle for cell propulsion (Macnab, 2004). In addition, the flagellum is involved in several F I G U R E 8 Schematic diagram of the differentially expressed genes involved in flagellar assembly. Layers of inner and outer membrane, peptidoglycan layer, C-, MS-, P-, and L-rings, motor proteins (MotAB), rotor (FlgBCF), hollow rod, hook, and distal part of flagellum are shown. Components of flagellum transport apparatus, ATPase complex, chaperones of flagellar transport, and their substrates are presented schematically. Highly expressed genes are shaded grey and lowly expressed genes are highlighted in green functions associated with bacterial pathogenicity, including biofilm formation, protein export, and adhesion (Haiko and Westerlund-Wikström, 2013). The flagellum also serves as a virulence factor in many bacteria, including Salmonella typhimurium, Escherichia coli, Vibrio cholera, and Pseudomonas aeruginosa (Duan et al., 2013). Bahar et al. (2011) found that an intact flagellum was required to achieve full virulence of A. citrulli. They observed that the A. citrulli strain M6 was unable to form an intact flagellum when its fliC gene was mutated, and the mutant was reduced in virulence and twitching motility although not in biofilm formation (Bahar et al., 2011). In our study, the deletion of the acrR gene abolished the ability of A. citrulli strain Aac-5 to form a polar flagellum, which, in turn, may have resulted in reduced swimming motility, contributing to its reduced virulence on watermelon seedlings. Our RNA-Seq data support the role that AcrR plays in flagellar biosynthesis because 33 genes involved in flagellar assembly were differentially expressed in the acrR mutant strain in comparison with the wild-type strain ( Table 3).
The lack of flagella could lead to enhanced growth ability in prokaryotes. Pyrococcus furiosus DSM363 exhibited an enhanced growth ability when its flagella were absent (Lewis et al., 2015).
The transcription regulator SwrA stimulates the transcription of the genes for σ D , which controls when Bacillus subtilis cells enter into the motile state with expression of flagellum biosynthesis genes from a state of growth as long and nonmotile chains (Kearns and Losick, 2005). For growth and maintenance on the host, bacteria may reduce or eliminate flagellar expression (Chaban et al., 2015). In our study, the deletion of the acrR gene not only eliminated the biosynthesis of flagellum of A. citrulli Aac-5 strain but also enhanced its growth ability, whereas the complementation strain showed similar growth ability to the Aac-5 strain, indicating that AcrR may regulate the Aac-5 switch between the growth and motility stages. The fact that 46 ribosomal genes were highly expressed in the mutant strain compared to the wild-type one (Table S1) suggests that these genes may contribute to the enhanced growth ability of the mutant strain.
Flagellar synthesis is a strictly hierarchical process in which more than 50 genes are involved (Wang et al., 2015). These genes are orga-  (Fitzgerald et al., 2014). In A. citrulli, we found that among the 33 differentially expressed genes associated with flagellar assembly, flhC and flhD genes were lowly expressed with 2.77-and 2.83-log fold change, respectively, suggesting that the early stage of flagellar synthesis is regulated by AcrR. The class II sigma-factor fliA gene regulates the transition from early to late-stage flagellar gene expression (Fitzgerald et al., 2014;Osterman et al., 2015). Along with its cognate flgM gene, an anti-sigma factor, fliA was significantly lowly expressed in the acrR mutant, which in turn down-regulated its downstream class III flagellar synthesis genes flgKL, fliDST, motB, and cheW (Table 3).
Why the three genes, fliP, fliQ, and fliR, which code for FliP, FliQ, and FliR, respectively, in flagellar biosynthesis were highly expressed in the acrR mutant in comparison to the wild-type strain is unclear.
FliP, FliQ, and FliR are components of the flagellar secretion apparatus that anchors at the cell membrane (Ward et al., 2018) and belongs to the flagellar type III secretion system (fT3SS). They reside within the MS-ring, a subdomain of the hook-basal-body, which is located within the cytoplasmic membrane (Figure 8) (Zhuang and Shapiro, 1995;Fan et al., 1997). Assembly of the flagellum begins with the MS-ring, comprising several transmembrane proteins, belonging to the fT3SS and inserted into the cellular membrane (Macnab, 2004). Biofilm formation is crucial for the virulence of some plant pathogenic bacteria (Dow et al., 2003;Fujishige et al., 2006). The phytopathogenic bacterium Xanthomonas citri showed reduced disease symptoms when it was unable to form a biofilm (Rigano et al., 2007). In contrast, our study revealed that biofilm formation of the acrR mutant of A. citrulli was not decreased but increased significantly compared to the wild-type strain Aac-5, while the virulence of the mutant was reduced. This result is in agreement with previous findings that some of the group II strains of A. citrulli are not able to form a biofilm, while the group I strain M6 is able to form a biofilm (Bahar et al., 2009;Chen et al., 2009), suggesting potential differences in the trajectory of biofilm formation processes between some of the strains that belong to the two groups of A. citrulli strains.
Twitching is a process that contributes to the adherence of bacterial cells to surfaces and colonization on bacterial hosts (Mattick, 2002;Craig et al., 2004). It is also required for the virulence and biofilm formation of A. citrulli (Bahar et al., 2009). When the acrR gene was mutated in Aac-5 of A. citrulli, no twitching motility was observed and the virulence was reduced, suggesting that acrR is important for twitching motility in A. citrulli, which, in turn, contributes to the virulence of the bacterium. An interesting finding of this study is that the acrR mutant lost the ability to twitch but increased its ability to form a biofilm compared to the wild-type strain Aac-5. This is different from previous findings that A. citrulli group I strain M6 and its mutants lacked the ability to twitch and were also decreased in biofilm formation (Bahar et al., 2009;Rosenberg et al., 2018). It is possible that the twitching motility and biofilm formation are regulated differently in group I and II strains, and the global transcriptional regulator AcrR may regulate many genes, including the ones involved in biofilm formation.
In summary, the acrR gene contributes to the virulence of A. citrulli strain Aac-5, either directly and/or indirectly though positive regulation of the twitching and swimming motilities and flagellar formation and negative regulation of the biofilm formation.
Additionally, our transcriptomic analysis revealed that the acrR gene also positively regulates flagellar assembly in A. citrulli, supporting the role that AcrR plays in flagellar biosynthesis. Because AcrR contains a receptor domain at the N-terminus, possibly the AcrR senses or interacts with certain signals that affect multiple biological functions, including virulence and flagellar formation of A. citrulli. Future research is needed to identify such signals and elucidate the molecular mechanisms behind the regulation of A. citrulli by AcrR for the development of effective control strategies to combat this important bacterial pathogen.

| Bacterial strains, plasmids, growth conditions, and primer design
The bacterial strains and plasmids used in this study are listed in Table 1 (Table 1). Primers GmF and GmR were designed based on gentamicin cassette (Table 1). All primers used in this study were designed using the free online program Primer 3.0 (http://www. simge ne.com/Primer3).

| Construction of the acrR mutant and its complemented strain
The acrR gene was deleted by homologous double recombination as described previously (Wang et al., 2016). Briefly, the 472-bp upstream and 541-bp downstream sequences of the acrR gene were amplified from the wild-type strain Aac-5 using the LacrR-F/ LacrR-R and RacrR-F/RacrR-R primers (Table 3). After confirmation by sequencing, the PCR fragments were digested by appropriate restriction enzymes and ligated into pK18 mobsacB to create the plasmid pK18-acrR-Up&Down (Table 2). The plasmid was digested by BamHI and SalI, and a Gm gene cassette (855 bp) was inserted between the BamHI and SalI sites to create plasmid pK18-acrRGm (Table 3). The plasmid was then introduced from E. coli DH5α into the A. citrulli strain Aac-5 by triparental conjugation using pRK600 as a helper plasmid to create the acrR mutant strain ΔacrR (Table 2).
Transconjugants were screened on KBA supplemented with 10% sucrose and antibiotics (Ap and Gm). The presence of the Gm cassette in the transconjugants was confirmed by PCR and sequencing of the amplified PCR product using the primer pair GmF/GmR (Table 1).
The absence of the acrR gene in the transconjugants was confirmed by the lack of PCR product using primer pair acrR-F/acrR-R.
To generate a complementation strain of ΔacrR, the acrR gene in Aac-5 was amplified using primers acrR-F and acrR-R (Table 1).
The PCR product was digested with HindIII and BamHI and cloned into pBBR1MCS-2 to generate pBBR-acrR, which was transferred into the mutant strain ∆acrR by triparental conjugation ( Table 2).
The successful transconjugant, named ∆acrRcomp, was identified through screening on KBA (amended with Ap, Km, and Gm; Table 2).
All obtained plasmids and A. citrulli strains were confirmed by PCR and DNA sequencing.

| Virulence assays
The virulence of the A. citrulli strains was tested on 3-week-old watermelon seedlings (Citrullus lanatus 'Jingxin#6', provided by the Beijing Academy of Agriculture and Forestry Sciences, Beijing, China). The virulence assay was performed, and the DI was calculated as previously described (Wang et al., 2016). Briefly, A. citrulli strains were grown in KB broth and their OD 600 was adjusted to 0.6 (approximately 10 8 cfu/ml). Two hundred millilitres of each bacterial suspension was sprayed onto watermelon seedlings grown at 28 °C in a growth chamber with 90% relative humidity. Disease symptoms were evaluated at the eighth day after inoculation using a disease severity scale: 0 for no symptoms; 1, 3, 5, and 7 for necrotic lesions on approximately 25%, 50%, 75%, and 100% of the leaves, respectively; and 9 for complete death of the seedling. The disease index was calculated based on the formula DI where A is the disease scale, B is the number of seedlings in each disease scale, and C is the total number of seedlings in each treatment. For each A. citrulli strain, six watermelon seedlings in three pots were inoculated in each experiment, and the experiment was repeated three times.

| In planta growth assays
The growth ability of A. citrulli strains in 2-week-old watermelon cotyledons (the same cultivar as the one used in virulence assay) was determined using the method of Johnson et al. (2011), with modifications. Briefly, A. citrulli was grown overnight in KB broth at 28 °C to an OD 600 of 0.8. The bacterial cells were then washed three times with sterile water and adjusted to 10 4 cfu/ml with sterile water to make the bacterial inoculum. One millilitre of the inoculum was injected into 30 watermelon cotyledons using sterile syringes.
One 3-mm disc from each cotyledon and six discs in total were collected and homogenized in Lysing Matrix A tubes (MP Biomedicals Co., Ltd) containing 1 ml of sterile water using MP FastPrep-24 T 5 G (MP Biomedicals Co., Ltd). The lysates were serially diluted with sterile water and plated on KBA plates. The plates were incubated at 28 °C for 48 hr. The growth of colonies on the KBA plates was counted as a measurement of the population of A. citrulli strains growing in watermelon cotyledons. The experiment was repeated three times.

| Assay for swimming and twitching motilities and observation of flagella by transmission electron microscopy
Swimming and twitching motilities of the A. citrulli strains were determined using the methods of Wang et al. (2016)

| Biofilm and growth ability assays
The effect of acrR deletion on biofilm formation was determined both qualitatively and quantitatively, and the growth rate of the A. citrulli strains was measured as described by Wang et al. (2016). (2016) using a FastQuant RT kit (TianGen). RT-qPCR analysis was carried out using primers designed for 10 DEGs (Table S2). cDNA was used as a template with SYBR Green added in the PCR, and relative levels of gene expression were determined as previously described (Wang et al., 2016). Three biological replicates were established in each experiment, and the experiment was repeated three times.

| RNA-Seq library construction and sequencing
A total of 3 μg of RNA per A. citrulli strain was used as input material for subsequent RNA sample preparations. The RNA sequencing libraries were constructed and sequenced commercially by Novogene Co., Ltd using NEBNext Ultra RNA Library Prep Kit for Illumina (NEB), and index codes were added to attribute sequences to each sample. The clustering of the index-coded samples was performed on a cBot Cluster Generation System using a TruSeq PE Cluster Kit v. 3-cBot-HS (Illumina). After cluster generation, the library preparations were sequenced on an Illumina HiSeq 2500 platform, and 100-bp paired-end reads were generated.

| RNA-Seq data analysis
The sequencing data were analysed commercially by Novogene Co., Ltd. Briefly, analysis for differential gene expression between the wild-type strain Aac-5 and the acrR mutant strain ΔacrR (three biological replicates per strain) was performed using the DESeq R package (v. 1.10.1). DESeq provides statistical routines for determining differential expression in digital gene expression data using a model based on the negative binomial distribution. The resulting p values were adjusted using the Benjamini and Hochberg's approach for controlling the false discovery rate. Genes with an adjusted p value <.05 found by DESeq were assigned as differentially expressed (Pan et al., 2018).
Gene ontology enrichment analysis of the DEGs was implemented by the goseq R package, in which the gene length bias was corrected. GO terms with a corrected p value less than .05 were considered significantly enriched by the DEGs (Pan et al., 2018).

| Statistical analysis
Statistical analysis was performed using the Student's t test in Excel 2010 software (Microsoft Inc.). Differences were considered statistically significant if p < .01.

ACK N OWLED G EM ENTS
This work was supported financially by National Natural Science

CO N FLI C T O F I NTE R E S T
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

AUTH O R CO NTR I B UTI O N S
W.G., T.W., and T.Z. designed the research. W.G. wrote the paper.

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.

S U PP O RTI N G I N FO R M ATI O N
Additional supporting information may be found online in the Supporting Information section.

FIGURE S1
Relative expression of selected genes by RT-qPCR. The bacterial strains were incubated under the same condition as those for the RNA-Seq experiment (three biological replicates per strain).
The x axis represents log 2 (fold-change) of each gene in acrR mutant compared to its wild-type strain Aac-5

TABLE S1
Differentially expressed genes between the Acidovorax citrulli acrR mutant strain and its wild-type strain Aac-5