Biological and transcriptomic studies reveal hfq is required for swimming, biofilm formation and stress response in Xanthomonas axonpodis pv. citri

Background Hfq is a widely conserved bacterial RNA-binding protein which generally mediates the global regulatory activities involv ed in physiological process and virulence. The goal of this study was to characterize the biological function of hfq gene in Xanthomonas axonpodis pv. citri (Xac), the causal agent of citrus canker disease. Results An hfq mutant in Xac was generated by plasmid integration. The loss of hfq resulted in attenuation of bacterial growth, motility and biofilm formation. In addition, the hfq mutation impaired Xac resistance to H2O2 and both high and low pH environments, but did not affect the virulence to citrus. RNA-Seq analyses indicated that Hfq played roles in regulating the expression of 746 genes. In hfq mutant, gene expression related to chemotaxis, secretion system, two-component system, quorum sensing and flagellar assembly were repressed, whereas expression of ribosomal genes were significantly up-regulated. The down-regulated expression of three bacterial chemotaxis related genes and seven flagella genes, which involved in cell growth and biofilm formation, were further validated by RT-qPCR. Conclusions The study demonstrated that hfq was involved in multiple biological processes in Xac. The results could serve as initiate points for identifying regulatory sRNAs and genes controlled by Hfq-sRNA interactions in Xac. Electronic supplementary material The online version of this article (10.1186/s12866-019-1476-9) contains supplementary material, which is available to authorized users.


Background
The RNA chaperone Hfq was originally discovered in Escherichia coli as a host factor essential for replication of the bacteriophage Qβ [1,2] and afterwards identified in a large number of Gram-positive and Gram-negative bacterial species [3]. Hfq forms a hexameric ringshaped doughnut structure that mediates the global post-transcriptional regulation involved in numerous physiological and biochemical functions in bacteria [2]. Inactivation of hfq gene exhibits broadly pleiotropic phenotypes in Escherichia coli, e.g. alteration in growth rate and tolerance to UV or high osmolarity stresses [4]. Many hfq mutants from a broad spectrum of bacterial pathogens show a general role in bacterial physiology and virulence [3,[5][6][7]. Notably, hfq deficiency impairs the stability and functional activation of many Hfq-dependent small non-coding RNAs (sRNAs) which are usually encoded in the intergenic regions of bacterial genomes [8]. The majority of these Hfq-dependent sRNAs can modulate the expression of target mRNAs by base pairing mechanisms, and then affect the downwards cellular processes [9,10]. Thus, Hfq protein in facilitating the interaction between small non-coding RNAs (sRNAs) and target mRNAs is currently considered as its most prominent function.
Xanthomonas spp. are economically important phytopathogens and are grouped into pathovars (pv.) based on their specific host ranges [11,12]. X. oryzae pv. oryzae (Xoo), X. campestris pv. vesicatoria (Xcv), X. campestris pv. campestris (Xcc) and X. axonpodis pv. citri (Xac) (synonym X. citri subsp. citri) are economical important pathogens among the genus Xanthomonas. hfq is conserved in Xanthomonas spp. and has been investigated in Xoo and Xcv. Inactivation of hfq in Xoo affects it growth in complex medium, but does not disrupt its virulence [13]. Similarly, hfq gene in Xcv does not involve in virulence [14]. Although many Hfq-dependent sRNAs were identified in Xcv, Xcc, and Xoo [15][16][17], their involvements in regulation and virulence function have been poorly characterized. Xanthomonas axonpodis pv. citri (Xac), the causal agent of citrus canker, is an important bacterial pathogen that severely affects citrus production worldwide. The two key regulators, HrpG and HrpX, regulating type III secretion system (T3SS), are known to play a critical role in Xac infection [16]. However, the roles of global regulator hfq and its related sRNAs remain to be determined.
In this study, the function of hfq in Xac biology and gene expression was characterized by using an hfq mutant constructed in strain Xac29-1. The inactivation of hfq resulted in the phenotypic alterations in bacterial growth, swimming motility, biofilm formation and stress response. Results of RNA-Seq analyses indicated that Hfq plays an important role in multiple biological processes including chemotaxis, flagellar assembly and secretion systems.

Results
The deletion of hfq attenuates the Xac growth and swimming A mutant named XacΔhfq was generated by plasmid integration and confirmed by PCR and Southern blot (Additional file 1: Figure S1). Wild type Xac29-1, mutant XacΔhfq and its complemented strain XacΔhfq-C were cultured in NB media to examine their growth curve. As shown in Fig. 1, the loss of hfq led to remarkably reduced growth rate, while was restored by the complemented strain. The growth rate of hfq mutant was very close to wild type at stationary growth stage (36 h).
The cell motility ability was evaluated on 0.3% (w/v) NA plates. The diameters of the motility zone derived from hfq mutant reduced by almost 60% when compared with the wild type. The complemented mutant strain XacΔhfq-C restored the motility (Fig. 2).
hfq gene was involved in Xac biofilm formation, but did not affect Xac virulence To assess biofilm formation, the strains were grown statically in borosilicate glass tubes in NB medium for 3 days. Staining of bacterial cells with crystal violet (CV) stain showed that Xac29-1 and XacΔhfq-C produced much more biofilms of cell mass adhered to the glass surface than those produced by XacΔhfq strain. Accordingly, the absorbance value of crystal violet from wild type was over three times greater than that of the hfq mutant (Fig. 3). A cell-counting kit (CCK)-8 assay was conducted to evaulate the cell viability, and the results demonstrated that cell viability was only slightly inhibited in hfq mutant and no significant difference was found between Xac29-1 and hfq mutant (Data not shown).
For pathogenicity test, wound infection assay in sweet orange leaves was used. At 4 days post inoculation, all strains induced spongy-like canker symptoms, indicating that hfq does not contribute to the virulence of Xac (Additional file 2: Figure S2).
The hfq mutation impairs bacterial resistance to hydrogen peroxide (H 2 O 2 ) and pH Compared with the NB medium, the growth of the hfq mutant almost showed no difference at the low concentration (0.001 mM) of H 2 O 2 , whereas the growth of the mutant was inhibited at 0.01 mM H 2 O 2 . The resistance of Xac to high and low pH was significantly affected by the mutation of hfq although the growth of complementation strain was also slightly affected at the high and low pH conditions (Fig. 4, Additional file 3: Figure S3). Taken together, the hfq mutation impairs Xac resistance to H 2 O 2 and pH.
Global RNA expression changes in hfq mutant of Xac RNA-Seq data have been submitted to the NCBI database and the accession numbers of wild type and hfq mutant are PRJNA477585 and PRJNA477663, respectively. Disruption of hfq significantly changed expression of 746 genes. Of them, 662 genes were down-regulated and 84 genes were up-regulated. The differential expressed genes (DEGs) were enriched into different function categories through Gene ontology (GO) enrichment analysis (Additional file 3: Figure S3). The most significant GO terms in cellular component GO terms included cellular component (GO:0005575), membrane (GO:0016020), membrane part (GO:0044425) and intrinsic to membrane (GO:0031224). Besides, localization (GO:0051179) and transport (GO:0006810) in biological process GO term, receptor activity (GO:0004872) in molecular function GO term were also enriched (Additional file 4: Figure S4). Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis showed that most of DEGs in pathways related to bacterial chemotaxis, bacterial secretion system, two-component system, quorum sensing and flagellar assembly were repressed, while the ribosome related genes were significantly up-regulated in hfq mutant (Additional file 5: Table S1).
To validate the RNA-seq data, 26 genes related with chemotaxis, flagella biosynthesis, secretion system and ribosome were chosen for RT-qPCR (Fig. 5). The expression trends of 24 genes were similar with those revealed by RNA-seq results, demonstrating the reliability of RNA-seq analysis. The down-regulation expression of XAC29_11660 (50S ribosomal protein L36) and XAC29_17265 (50S ribosomal protein L31) by RT-qPCR was inconsistent with those obtained from the RNA-seq.

Discussion
Although the inactivation of hfq in different bacterial species has exhibited a pleiotropic phenotype [4,[17][18][19], its deletion in the genus Xanthomonas displayed similar alterations in growth and motility [13,14]. In this study, the mutation in hfq gene resulted in remarkably reduced bacterial growth rate. Additionally, the deletion of hfq led to a reduction of cell swimming ability by 60% and biofilm formation was reduced by almost 70%. Transcriptome and RT-qPCR analysis showed that three bacterial chemotaxis related genes, cheR, cheA, cheW, and seven flagella genes, fliD, fliR, flhA, flhB, chpA, motA, motB, were significantly repressed in the hfq mutant (Additional file 5: Table S1). CheA, CheW and CheR are core proteins of chemosensory pathways which are essential for motility and pathogenicity in many bacteria [20]. Transcriptome data related to chemotaxis and flagellar assembly strongly supported the biological results of the attenuation of cell motility and biofilm formation (Fig. 5, Additional file 5: Table S1).
Hfq is a common regulator of virulence in bacteria [3]. However, the virulence-related defects are not common in plant-pathogenic bacteria and only reported in a few bacteria, i.e., Agrobacterium tumefaciens and Erwinia spp. [5,21]. Hfq significantly regulated type III secretion system and the other related pathogenicity determinants in E.amylovora [5]. hfq mutant attenuates the tumor formation and influences the virulence of A. tumefaciens, but the DNA-transferring type IV secretion system is not affected [21]. Transcriptome analysis showed that hfq significantly regulated chemotaxis, bacterial secretion systems, two-component system, quorum sensing and flagellar assembly in Xac. Regarding the expression profile of the genes associated with secretion systems (type I to type VI), those from type IV and type V secretion systems were unaltered, whereas two genes in type I were slighted repressed and 18 genes from type II, III, VI secretion systems were significantly decreased (over 2 folds) in hfq mutant (Additional file 5: Table S1). Similar with previous studies on Xoo and Xcv (13,14), the virulence of the hfq mutant of Xac also did not affected through wound infection assay. It should be noted that virulence phenotype between Xac wild type and waxcO mutant had no difference by wound infiltration, whereas significant differences in lesion numbers were observed by spray assay [22]. Non-wound inoculation, spraying or swabbing, remains to be used to check the pathogenicity phenotype of hfq mutant of Xac, Xoo and Xcv.
It is interesting that nine ribosomal protein genes were significantly up-regulated through transcriptome data. Of the nine proteins, RpsE codes for 30S subunit ribosomal protein S5 and the other eight proteins are components of the ribosomal large subunit. It should be note that there was about four-fold expression increase for rplU-rpmA, encodes for 50S subunit r-proteins L21 and L27, respectively (Additional file 5: Table S1). RT-qPCR also proved the high expression of the operon genes in hfq mutant (Fig. 5). L27, consisting of a C-terminal β-sandwich domain and a long N-terminal arm, plays a key role in tRNA substrate stabilization during the peptidyl transfer reaction [23,24]. RpsE has been implicated in tRNA selection and translation fidelity in E. coli [25]. Although it remains unclear how Hfq affects these ribosomal proteins of Xac, the expression differences of ribosomal proteins, probably resulted in reduced translation accuracy, might also play certain roles in the pleiotropic defects of hfq mutants.
Many sRNA candidates in genus Xanthomonas have been generated by high-throughput transcriptome sequencing approaches [13][14][15]. Of these sRNAs, 44 sRNAs were so far experimentally verified in Xoo, Xcv and Xcc [26]. It should be noted that the accumulation and activity of only six sRNAs were closely related with hfq whereas some sRNAs involved in virulence, i.e. sX12 and sX13, were assumed to act Hfq-independent (14,26). It might partially explain why the hfq mutant strain of Xanthomonas spp. was not altered in the induction of virulence phenotype. So far, the function of Hfq is still obscure in Xanthomonas spp. and its physiological roles and RNA-binding capability is needed to be further addressed. This study is a good starting point for identifying regulatory sRNAs and genes controlled by Hfq-sRNA interactions in Xac.

Conclusions
In this study, biological analyses of the hfq mutant clearly point toward the requirement for Hfq function in multiple biological processes in Xac, i.e. motility, biofilm formation and stress response. RNA-seq data showed that 746 genes were regulated by hfq gene, reflecting its global regulation role in Xac. In particular, the expression of genes associated with bacterial chemotaxis and flagellar assembly were significantly down-regulated in the hfq mutant, consistent with the reduction of swimming and biofilm formation.

Bacterial strains, plasmids and growth conditions
The bacterial strains and plasmids used in this work are listed in Table 1. Xanthomonas strains were grown at 28°C in nutrient broth (NB) medium or on NA (NB with 1.5% Agar). Escherichia coli strains were cultured at 37°C in Luria-Bertani (LB) medium or on LA (LB with 1.5% Agar). When required, the antibiotics kanamycin (50 μg/ml) and gentamycin (10 μg/mL) were added to the growth media.

Construction of the hfq deletion mutant and its complemented strain
The hfq mutant was generated from Xac29-1 wild type strain by allelic homologous recombination. Briefly, two hfq flanking regions were amplified by PCR using the primer pairs up F/R and down F/R ( Table 1). The PCR products of upstream and downstream were digested with BamHI/KpnI and KpnI/XbaI, respectively. The digested fragments were ligated into the suicide vector pK18mobSacB to obtain the recombinant plasmid pK18mobSacB-Δhfq. The plasmid was transformed into wild type strain Xac29-1 by electroporation. The hfq mutant, named Xac29-1Δhfq, was obtained after two recombination events and confirmed by PCR and Southern blotting.
To complement the hfq mutant, DNA fragment containing the entire hfq gene and its upstream promoter was amplified. The amplified fragment was digested with HindIII/EcoRI enzyme and cloned into HindIII/EcoRI-digested pBBR1MCS-5 [27], resulting in pBBR1MCS-5 + hfq plasmid. The complementary plasmid was transformed and one complemented mutant strain, named Xac29-1Δhfq-C, was selected on NA plate with Gentamycin resistance.

Determination of growth curve
Pellets of Xac29-1, Xac29-1Δhfq and Xac29-1Δhfq-C strains were cultured in NB medium and adjusted to an OD 600 =0.6 and then sub-cultured (1100) in fresh NB for 36 h. The OD 600 values were tested after every 2 h post sub-culturing. All the experiments were repeated at least three times.

Motility assay
To test cell motility, all strains were grown overnight in NB medium and adjusted to an OD 600 =0.6 [10 8 colony-forming units (cfu)/mL]. 2 μL of each cell sample was dropped to 0.3% agar NA plates for the swimming motility tests [18]. The diameters of each colony were measured after 48 h of incubation, and the resulting values were taken to indicate the bacterial motility.

Biofilm formation assay
Biofilms that formed on polystyrene and glass surfaces were examined as previously described [28]. Briefly, 5 μL of each adjusted cell sample was transferred to a glass bottle containing 5 mL fresh NB medium and stationary incubated at 28°C for 3 days. After the medium removal and washing, bacteria were stained with 5 mL crystal violet for 5 min. Excess crystal violet stain was removed to observe a circle of purple material formed on the glass bottle. The crystal violet dye was solubilized by addition of 5 mL organic solvent (anhydrous ethanol: acetone = 70:30, v/v), then crystal violet was quantified by measuring absorbance at 600 nm. All the experiments were repeated three times and the average for each strain was checked by t-test.

Cell viability assay
A CCK-8 assay was used to determine the cell viability of Xac29-1, Xac29-1Δhfq and Xac29-1Δhfq-C strains incubated at 28°C for 3 days. 100 μL of each cell sample and 10 μL CCK-8 (BioDee Biotechnology, Beijing, China) were plated into 96-well plates and incubated at 37°C for 4 h. Then the absorbance was measured at 450 nm wavelength. Each group had three wells and all experiments were repeated three times.

Pathogenicity assay
Strains were cultured and adjusted to an OD 600 =0.6. The full expanded leaves of pineapple sweet orange (Citrus sinensis) were used as host materials. Five wounds were produced on the back of the leaves with a needle. 10 μL of each cell suspension were then placed on the wounds. Disease symptoms were observed and photographed 2 days post inoculation [30]. Each test was repeated at least three times.

RNA extraction, library preparation and RNA sequencing
Bacterial cultures from wild-type and hfq mutant strains were collected in the middle exponential stage (OD 600 = 0.6-0.8). RNAs were extracted using the RNA prep pure Cell/Bacteria Kit (Tiangen Biotech, Beijing, China). Strand-specific RNA-seq libraries were generated using NEBNext® Ultra™ Directional RNA Library Prep Kit for Illumina® (NEB, USA). After cluster generation, the library preparations were sequenced on an Illumina Hiseq platform (Illumina, CA, USA) in Novogene, Tianjing.

Transcriptome data analysis
After cleaning the raw reads, we mapped the clean reads to the complete genome Xac29-1 strain (CP004399.1, CP004402.1, CP004401.1 and CP004400.1) using Bowtie 2 [31] and then calculated gene expression level with RSEM [32]. Differentially expression analysis was performed using the DESeq R package (1.18.0). The p-values were adjusted for the false discovery rate (FDR). Genes with an adjusted p-value < 0.05 and |log 2 (Fold Change)| > 1 were considered to be differentially expressed genes at a statistically significant level. A t-test was performed on log 2 -transformed data to identify the genes with significant changes in expression between wild type and mutant strains. Gene Ontology (GO) terms with corrected P value ≤0.05 were considered significantly enriched by differentially expressed genes.

RT-qPCR
To assess the RNA-seq quality, the same set RNAs for RNA-seq were subjected to a two-step RT-qPCR assay with Bio-Rad iQ5 Real Time PCR System (Bio-Rad, CA, USA) using SYBR green RT-PCR kit (Promega). 26 genes related with bacterial chemotaxis, flagellar assembly, bacterial secretion and ribosome were chosen for RT-qPCR (Additional file 6: Table S2). The 16S rRNA gene was used as an endogenous control. The relative fold change in target gene expression was calculated using the 2 -ΔΔCT method [33].

Additional files
Additional file 1: Figure S1. PCR and southern blotting confirmation of the hfq mutant. (A) The gene deletion scheme. The 480-bp (amplified by hfq-up-F/R) ( Table 1) and 680-bp (amplified by hfq-down-F/R) ( Table 1) DNA fragments were used as the 5′and 3′fragments for homologous recombination, respectively. The 278-bp DNA fragment of hfq gene was deleted in the hfq mutant. The hfq-F/R primer (Table 1)