Comparative Transcriptome Profiling Reveals Different Expression Patterns in Xanthomonas oryzae pv. oryzae Strains with Putative Virulence-Relevant Genes

Xanthomonas oryzae pv. oryzae (Xoo) is the causal agent of rice bacterial blight, which is a major rice disease in tropical Asian countries. An attempt has been made to investigate gene expression patterns of three Xoo strains on the minimal medium XOM2, PXO99 (P6) and PXO86 (P2) from the Philippines, and GD1358 (C5) from China, which exhibited different virulence in 30 rice varieties, with putative virulence factors using deep sequencing. In total, 4,781 transcripts were identified in this study, and 1,151 and 3,076 genes were differentially expressed when P6 was compared with P2 and with C5, respectively. Our results indicated that Xoo strains from different regions exhibited distinctly different expression patterns of putative virulence-relevant genes. Interestingly, 40 and 44 genes involved in chemotaxis and motility exhibited higher transcript alterations in C5 compared with P6 and P2, respectively. Most other genes associated with virulence, including exopolysaccharide (EPS) synthesis, Hrp genes and type III effectors, including Xanthomonas outer protein (Xop) effectors and transcription activator-like (TAL) effectors, were down-regulated in C5 compared with P6 and P2. The data were confirmed by real-time quantitative RT-PCR, tests of bacterial motility, and enzyme activity analysis of EPS and xylanase. These results highlight the complexity of Xoo and offer new avenues for improving our understanding of Xoo-rice interactions and the evolution of Xoo virulence.


Introduction
The gram-negative plant pathogenic Xanthomonas oryzae pv. oryzae (Xoo) is the causal agent of bacterial blight disease on rice [1]. Bacterial blight is the most serious bacterial disease of rice in tropical Asian countries where high-yielding rice cultivars are often highly susceptible, and it has the potential to reduce rice yields by as much as 50% [2]. The complete genome sequences have been published for Japanese race 1 [3], Korean race 1 [4], and PXO99 A , a 5-azacytidine-resistant derivative of the Philippines' race 6 [5]. These genomes have helped to elucidate the molecular interactions between a pathogen and a monocotyledonous plant and have greatly advanced the understanding of the molecular interactions between rice and Xoo. Several virulencerelated factors have been identified, such as the hypersensitive response and pathogenicity (hrp) genes [6], type III (T3) effectors [7,8], genes associated with the production of exopolysaccharides (EPS), and genes associated with motility and extracellular enzymes [9,10].
The type II (T2S) and type III (T3S) secretion systems are important for the virulence of Xoo. The T3S system, encoded by hrp genes, plays an important role in interactions between Xoo and rice by injecting T3 effectors into plant cells, whereas the T2S system may play a role in the secretion of other virulence factors, such as extracellular enzymes like xylanase [11]. The T3S system is transcriptionally induced in certain minimal media and in plants [12], and the ompR-type response regulator HrpG, which is activated by unknown plant signals, controls the genome-wide regulon, including hrps, T3 effectors and putative virulence genes [13].
The collection of T3 effectors in Xanthomonas are designed as Xanthomonas outer proteins (Xop). Sixteen and 18 candidate Xop effectors were identified in Xoo strains MAFF311018 and PXO99 A , respectively. Among them, XopZ PXO99 was demonstrated to contribute to the virulence of Xoo strains [14,15]. Besides Xop genes, there is another important type T3 effector, the transcriptional activator-like (TAL) effectors in Xoo, which contain a central repeat domain in which amino acids 12 and 13 [known as the repeat variable diresidue (RVD)] of each repeat, and have been shown to transcriptionally activate the corresponding host genes for host disease susceptibility or resistance by recognizing and binding specific DNA sequences within the promoters of host target genes with RVDs [16][17][18][19][20].
The transcriptional regulation of putative virulence-relevant genes is critical to Xoo for infection and proliferation in rice varieties. Although the complete genome sequences of three Xoo strains from Asia and a draft genome sequence from Africa have been analyzed [3][4][5]21], so far only microarray analysis has revealed that a greater number of Xoo genes are differentially expressed in XOM2 relative to PSB [22], and little is known about the transcriptome patterns of different strains. Analysis and comparisons of gene expression profiles in different strains will provide new insight into the pathogen's virulence strategies. Here, we report the transcriptional expression profiling of genes involved in the virulence of three Xoo strains, one from China and two from the Philippines, induced on XOM2 minimal media.

Pathogenicity Assays and Growth Curve of Xoo
Thirty rice varieties ( Figure 1A) were used to evaluate the pathogenicity of Xoo strains P6, P2 and C5. The plants were grown in the screenhouse of the Institute of Crop Sciences, Chinese Academy of Agricultural Sciences, Beijing, China in the summer of 2011. For evaluating bacterial blight resistance, seeds of the rice varieties were sown in the seedling nursery and 30-day-old seedlings were transplanted in the screenhouse with 9 plants in each row at spacing of 20617 cm. At the tillering stage (plant age was 65 days), four to five of the uppermost leaves of each plant were inoculated with Xoo strains by the leaf-clipping method [23]. Inoculum of each race was prepared by suspending the bacterial mass in sterile water at a concentration of 10 8 cells ml 21 . Five central plants of each line were inoculated with each race for three replications. The lesion lengths (LL) were measured on all inoculated leaves 2 weeks after inoculation when lesions became obvious and stable in the susceptible variety IR24, and the average LL of each plant was calculated according to its three longest lesions. Evaluation of the resistance level of each variety was based on the average LL of 15 plants. LL!1 cm, 1 cm !LL,5 cm, 5 cm!LL,10 cm, 10 cm!LL,15 cm, 15 cm!LL,20 cm, and LL$20 cm represent highly resistant, resistant, moderately resistant, moderately susceptible, susceptible and highly susceptible, respectively.
The growth curve of Xoo was produced according to the method described by Song et al. [24]. Five inoculated leaves of Hanyou715 and Hanyou53 from three plants were collected at 1, 2, 3, 4 and 5 days post-inoculation (dpi). The collected leaves were immediately frozen in liquid nitrogen, and then kept at 270uC. For each time point, the bacterial populations were determined by grinding three leaves separately, plating the resulting extract on potato sucrose agar media containing 200 mM azacytidine, and counting colony-forming units (CFUs) after 48 h at 28uC.

Library Preparation and Illumina Sequencing
The ribosomal RNA (rRNA) was removed from 3 mg of total RNA with Ribo-Zero TM Magnetic kit for Gram-Negative Bacteria (EpicentreBio) by the manufacturer's instructions. The RNA library was constructed according to the TruSeq TM RNA Sample Preparation kit (Illumina) with minor modification. Briefly, RNA was fragmented and the first cDNA strand was synthesized by using the random hexamers and SuperScript II Reverse Transcriptase (Invitrogen), then RNA template was removed and a replacement strand was synthesized to generate double-stranded (ds) cDNA. After end repair and 39 end adenylation, the indexed adapter was ligated with the dsDNA. Fragments of 300,350 bp were excised and enriched by PCR for 12 cycles. The yield and size distribution of PCR products were checked by QUBIT and Agilent 2100 Bioanalyzer respectively. The produced libraries were performed cluster generation on cBot and sequenced on HiSeq 2000 platform (illumina) with 100 bp paired-end reads by CapitalBio Corporation, Beijing, China. Illumina Casava(version 1.7) was used for basecalling, then all the sequencing data was processed by removing sequencing adapters for further analysis.

Analysis of Illumina Sequencing
The analysis of RNA-seq sequencing data was performed as described by Zhao et al. [25] ( Figure S1). All the tags mapped to reference sequences by Burrows-Wheeler Aligner [26,27] with a maximum of five nucleotide mismatch. For gene expression analysis, the value of reads per kilo bases per million reads (RPKM) [28] was calculated. DEGseq [29] was applied to identify differentially regulated genes between two samples using the two classes unpaired MA-plot-based method to detect and visualize gene expression difference with significant P values less than 0.001. The whole genome sequence of Xoo was downloaded from the National Center for Biotechnology Information (www.ncbi.nlm. nih.gov), and coding regions were annotated according to the annotated protein data sets of Xoo strain PXO99 A .

Validation of Expression Patterns of DEGs Using Quantitative Real-time RT-PCR
To validate the results of the Illumina sequencing experiment, a subset of differentially expressed genes (DEGs) were verified by quantitative real-time RT-PCR (qRT-PCR). An independent set of cell cultures of the three Xoo strains were cultured following the same protocol as for the Illumina analysis. QRT-PCR followed the methods described by Swarbrick et al. [30]. The sequence of each gene was obtained from the X. oryzae pv. oryzae PXO99 A database (http://www.ncbi.nlm.nih.gov), and the sequences from each gene were used for designing primers by Primer 5 software (http:// frodo.wi.mit.edu/) (Table S1). RNA samples from three independent replicates for each treatment were pooled before cDNA synthesis. Thirty-three Xoo genes were tested in 50 ml reactions using the SYBRH Green PCR Master Mix kit (Applied Biosystems, CA, USA) following the manufacturer's protocol. The correlation coefficient between the qRT-PCR and RNA-Seq results was calculated.

Motility Analysis
Fresh colonies from PS agar plates were stabbed into swarm plates composed of 0.03% (wt/vol) Bactopeptone, 0.03% yeast extract, and 0.3% agar. The inoculated cells were cultured at 28uC and examined for bacteria swarming away from the inoculation site at 12, 24, 48 and 72 h after inoculation [31]. This study was repeated three times for reproducibility.

Quantitative Determination of EPS and Xylanase Activity
The fresh colonies of Xoo strains were grown at 28uC with shaking at 200 rpm in nutrient-rich PSB until the OD 600 equaled 2.0. They were then washed twice and immediately transferred into XOM2. After growth for 16 h, the bacterial cultures were collected and supernatants were prepared by centrifugation at 5000 rpm for 10 min. The extracellular xylanase was measured by using 4-O-methyl-D-glucurono-D-xylan-Ramezol Brilliant Blue R (RBB-xylan; Sigma Co.) according to the methods described by Biely et al. [32]. The production of EPS was determined according to the methods described by He et al. [33].

Pathogenicity Testing of Xoo Strains
To evaluate the pathogenicity of P6, P2 and C5, 30 rice varieties, including 29 recently developed varieties in China, and a cultivar susceptible to all Philippine Xoo races from the International Rice Research Institute (IRRI), IR24, were inoculated at the tillering stage. Also included were CBB23, Xinhuangzhan, and Hua201S-1, which carry the bacterial blight resistance gene Xa23, a single completely dominant resistance gene identified from wild rice species of Oryza rufipogon [34,35]; Xa21 from the wild rice strain XF10450 [36] and O. longistaminata [37], and Xa7 from IRBB7 [38]. Only four varieties were resistant to the three Xoo strains. Among them, three varieties including CBB23, Xinhuanzhan and XF10450, exhibited a typical hypersensitive reaction (HR) with less than 0.5 cm LL. The LL of Hua201S-1 inoculated with P6, P2 and C5 was 3.061.1 cm, 3.160.9 cm and 0.460.4 cm, respectively. Except for Teqing, which was moderately resistant to P6 and P2 and susceptible to C5, the other 25 varieties displayed a moderate or high susceptibility to the three strains with LLs ranging from 12.1 to 43.5 cm.
The 30 rice varieties were placed in three groups depending on the LL phenotypes exhibited after inoculation (Figure 1 A)  42.063.6 cm and 0.460.4 to 32.162.4 cm against P6, P2 and C5, respectively, also consisted of 12 varieties: Hua2018, Yunguang20, Tianyou20, Tianyou145, Yunjing30, Hanyou73, Liangfengyou339, Bohuyou813, CDR22, Hanyou715, Ha-nyou113, and Hanyou53. The LLs of the same variety were significantly longer when inoculated with two of the Xoo strains than when infected by the third strain. Eleven of the varieties had significantly longer LLs when infected by P6 or P2 than when infected by C5. The exception was the variety Liangyou1813 ( Figure 1A). Bacterial growth curve analysis indicated that the growth of P2 and P6 increased by more than three-fold compared with C5 in Hangyou715 36 hr post-inoculation, and the growth of P2 and P6 increased by more than 20-fold compared with C5 in Hangyou53 36 h post-inoculation ( Figure 1B, C).
Group III had six varieties. The LLs caused by one strain were significantly longer than when infected by the other two strains. The LLs of Wuhuashandao and Yunzijing20 when infected by P6 were significantly longer than when infected by P2 and C5, and the LLs of Yunguang101 and Hanyou79 when infected by P2 were significantly longer than when infected by P6 and C5. However, the LLs of Teqing and Hanyou713 when infected by C5 were significantly longer than when infected by P6 and P2.
Our results indicated that the virulence of P2 was similar with that of P6 when infecting 21 rice varieties, and the LLs of five and four varieties were significantly longer and shorter against P6 than against P2, respectively. However, there was no significant difference between the virulence of C5 and P6 when infecting 14 rice varieties, C5 was significantly weaker than P6 when infecting 13 rice varieties, and C5 was stronger than P6 when infecting two varieties.

Mapping of mRNA-seq Reads and Statistical Testing to Detect Differentially Expressed Genes
To compare the transcriptome profile of Xoo strains with different virulence levels, RNA sequencing libraries were constructed for P6, P2 and C5. Each library generated about 28.6 to 29.2 million reads, which were mapped to the PXO99 A genome sequence (NCBI Reference Sequence: NC_010717.1) with 78.9%, 80.1% and 81.6% matched reads to NCBI annotated gene regions, respectively (GEO database: accession number GSE44215). There were 5,083 protein-coding genes predicted in the genome of PXO99 A by genome analysis [39]. In total, 4,781 transcripts were identified in this study. With a threshold of more than five reads mapped to the CDS regions of a given gene in each sample, a total of 4,605, 4,503 and 4,467 protein-coding genes were detected in P6, P2 and C5, respectively (Figure 2A).
The R-package DEGseq [29] was used to identify DEGs. The list of genes with significantly different expression levels between two strains was refined using the criterion of P value ,0.001 in t tests, resulting in 1,151, 3,076 and 3,112 DEGs when P6 was compared with P2, P6 was compared with C5, and P2 was compared with C5, respectively (Table S2, S3, S4). And all the DEGs between each of the comparisons shared 782 genes in common ( Figure S2).
Gene ontology (GO) assignments were used to classify the function of the DEGs. Based on sequence homology, the DEGs can be categorized into 39 functional groups ( Figure 2B). In the three main categories, cellular component, molecular function and biological process, of GO classifications, ''cell part'', ''catalytic'' and ''metabolic process'' terms are dominant, respectively. We also noticed a high-percentage of genes from categories of ''cell'', ''binding'' and ''cellular process'' as well as a few genes from ''reproduction'', ''reproductive process'' and ''viral reproduction''.

Validation of Expression Patterns by qRT-PCR
To validate the Illumina sequencing results, qRT-PCR was used to independently assess expression levels for 33 genes involved in motility, the T3S system and T3 effectors (genes and primer sets used are shown in Table S1; the results of qRT-PCR are shown in Figure S3). RNA samples that were used in the Illumina sequencing experiment as well as RNA samples extracted from three additional replicate sets of cultures were used as templates. There was a good correlation between the qRT-PCR and the mRNA-seq results (correlation coefficient was 0.8297) ( Figure 2C). Although the amplitude of gene expression fold change between the two techniques is different, as might be expected since qRT-PCR is not a reliable measure of quantitative differences, the general trend of gene expression is consistent.

Differential Expression of Two-component Systems between Xoo Strains
Two-component systems (TCSs) are widespread signal transducers in prokaryotes that serve as a basic stimulus-response coupling mechanism to allow organisms to sense and respond to changes in many different environmental conditions [40]. They typically consist of a membrane-bound histidine kinase that senses a specific environmental stimulus and a corresponding response regulator that mediates the cellular response [41]. In this study, 55 transcripts associated with TCSs were identified in three Xoo strains (Table S5), and DGE analysis revealed that the transcriptome profile of C5 was considerably different from P6 and P2, which exhibited similar expression patterns. Nineteen TCSs differentially expressed in P2 compared with P6. Among them there were two and 17 significantly up-and down-regulated genes, respectively. Forty-four genes differentially expressed in C5 compared with P6; 11 and 33 were significantly up-and downregulated genes in C5, respectively. Similarly, of the 44 TCSs differentially expressed in C5 compared with P2, 11 and 33 were significantly up-and down-regulated genes in C5, respectively ( Figure 3A).
The TCS transcriptome profiling of three Xoo strains revealed two interesting aspects. First, the expression pattern of the RpfC/ RpfG two-component regulatory system associated with quorum sensing (QS) was different between P2, P6 and C5 (Table 1). Genetic and genomics evidence suggest that Xoo might use the diffusible signal factor (DSF) QS system to regulate virulence factor production [33]. QS is a complicated bacterial group behavior for producing, sensing and responding to multifarious chemical signals, which increases their chances of survival and propagation, and it provides bacterial pathogens an obvious competitive advantage over their hosts in pathogen-host interactions [42]. The RpfC/RpfG two-component regulatory system is implicated in sensing and responding to DSF perception and signal transduction [43,44]. RpfC negatively controls DSF biosynthesis by binding to rpfF at a low cell density [33]. Several GGDEF, EAL and HD-GYP domain proteins of X. campestris pv. campestris (Xcc) are hypothesized to compose a network of signal transduction systems for response to different environmental cues to modulate the level of the second messenger cyclic di-GMP [45]. Our results indicated that a TCS regulatory protein with a HD-GYP domain (PXO_00476) had significantly down-regulated expression in C5 compared with P6 and P2. However, a TCS regulatory protein with a GGDEF domain (PXO_00466) had significantly up-regulated expression in C5 compared with P6 and P2. Moreover, cyclic di-GMP phosphodiesterase A (PXO_00058) had significantly up-regulated expression in C5 compared with P6 and P2.
Second, significant up-regulation of the phoPQ-regulated protein (PXO_01585) was detected in P6 and P2 when compared with C5 (Table 1). This protein is not only required for AvrXA21 activity, but also controls virulence through the regulation of hrpG gene expression [46]. It also regulates numerous cellular activities in Salmonella and other species as a master regulator of virulence [47,48]. AvrXA21 requires a regulatory TCS called RaxRH to regulate expression of 10 rax (required for AvrXA21 activity) genes. Our results indicated that raxH (PXO_04467), raxH2 (PXO_02837), raxR (PXO_04469) and raxR2 (PXO_02836) were significantly up-regulated in P6 and P2 compared with C5. Additionally, raxA (PXO_04478), raxB (PXO_04477), raxC (PXO_02621), raxP (PXO_02134), raxQ (PXO_02135) and raxST (PXO_04479) were also significantly up-regulated (Table 1). However, there was no significant difference in the expression of the phoPQ-regulated protein and the rax genes, except for raxR and raxST between P6 and P2. This suggests the expression pattern of genes involved in AvrXA21 activity and hrpG expression exhibited by C5 was different from P6 and P2.

A set of Genes Possibly Related to Chemotaxis and Bacterial Motility had Significantly Up-regulated Expression Levels in C5
The number of DEGs involved in chemotaxis and motility were differentially expressed between the three strains and exhibited interesting expression patterns (Table S6; Figure 3B). These genes mainly included chemoreceptors, chemotaxis proteins, twitching motility proteins, flagellar motor proteins, pilus biogenesis proteins and pilus assembly proteins. Of the differentially expressed genes, 38 genes had down-regulated expression levels in P2 when compared with P6. In addition, 40 genes had up-regulated expression levels and four genes had down-regulated expression levels in C5 when compared with P6. Finally, 44 genes had up-regulated expression levels and three genes had down-regulated expression levels in C5 when compared with P2. In general, the structural genes encoding motility systems are clustered within large transcriptional units allowing co-regulation of their expression [49]. We found that the expression of the chemoreceptor glutamine deamidase CheA (PXO_00032) increased greater than 2.89-fold in C5 when compared with P6 and P2 ( Figure 4A). Consistent with this finding, 15 and 17 genes involved in the encoding and synthetic metabolism of chemotaxis proteins, including CheD (PXO_00056) ( Figure 4B), were up-regulated in C5 compared with P6 and P2, respectively. Additionally, many pil genes involved in bacterial movement [50], including pilG (PXO_01602), pilH (PXO_01603), pilL (PXO_01607), pilV (PXO_01321), pilX (PXO_01323), pilY1 (PXO_01324) and pilZ (PXO_00049), were also up-regulated in C5 compared with P6 and P2. In addition, four genes encoding flagellar motor proteins, including MotA (PXO_03068), MotB (PXO_03067), MotC (PXO_00026) and MotD (PXO_00027) ( Figure 4C, D), and two genes encoding twitching motility proteins, including PXO_01994 and PXO_01993, were significantly up-regulated in C5.
Swarm plate analysis displayed that the swimming diameter of C5 was significantly larger than those of P6 and P2, and they tended to form larger swarming colonies at 24, 48, 72 and 96 h post-inoculation ( Figure 4E). This confirms that the motility of C5 was significantly stronger as reflected by the enhanced expression levels of many genes encoding chemotaxis proteins, pil proteins and flagellar motor proteins.

Differential Expression of a Gum Gene Cluster Involved in EPS Synthesis
The gum gene cluster involved in EPS synthesis functions as a virulence determinant in Xanthomonas [51]. EPS synthesis in Xcc is directed by genes within the gum cluster, which contains 12 genes and has a major promoter upstream of the first gene, gumB [52]. Similarly, the Xoo gum cluster is composed of 14 ORFs that constitute an operon expressed from a promoter located upstream of gumB, but the cluster also has internal promoters upstream of gumG, gumH and gumM [46,53]. In our study, 12 gum genes were differentially expressed in C5 compared with P6 and P2. Except for the gumB (PXO_01391) up-regulated gene, the other 11 genes from gumC to gumM (PXO_01392-PXO_01403) were all down-regulated (Table S7). By contrast, five genes (gumD, gumH, gum K, gumL, and gumM) were up-regulated in P2 compared with P6. In addition, two genes encoding xylanase (PXO_03864 and PXO_04558) were down-regulated in C5 compared with P6 and P2, and PXO_04558 also had down-regulated expression in P2 compared with P6.

Hrp Genes and T3 Effectors had Significantly Upregulated Expression Levels in P6 and P2
Knowledge of hrp genes in Xanthomonas arises mainly from studies of the X. campestris species [55]. Hrp genes are essential for pathogenicity in both Xoo and Xoc. HrpF has been found to be a putative type III translocon protein required for pathogenicity [56]. HrpG activates the expression of hrpA and hrpX. HrpX encodes a protein belonging to the AraC family of positive transcriptional activators and controls the expression of operons hrpB to hrpF as well as avrXv3 and a number of putative virulence factors. We examined expression of hrp genes and tested the expression of several genes by qRT-PCR ( Figure 5C, D, E, F; Figure S3). Interestingly, the expression of 29 genes associated with hrp genes (Table S8), including genes encoding hrpF (PXO_03417), hrpG (PXO_01951) (Figure 5C), the hrpA type III secretion outer membrane pore (PXO_03393) and hrpX (PXO_01953) (Figure 5D), were significantly down-regulated in C5 compared with P6 and P2. However, 14 of 29 genes were up-regulated in P2 compared with P6.
We also specifically examined the expression levels of T3 effectors. Fifteen Xop genes were significantly down-regulated in C5 when compared with P6 and P2, and 10 Xop genes were downregulated in P6 when compared with P2 (Table 2). These effectors were expressed in a HrpX-dependent manner, suggesting the coregulation of effectors and the T3S system. Consistent with the results of sequencing, the expression difference between XopN (PXO_02760) and XopW (PXO_03356) among the three strains was also confirmed by qPCR ( Figure 5G, H). In addition, 11 and 12 genes encoding TAL effectors were down-regulated in C5 compared with P2 and P6, respectively. Among them, pthxo1 (PXO_03922) increases bacterial populations in plants. However, six genes were down-regulated in P6 compared with P2, including pthxo1 (Table 2).

Discussion
Bacterial blight occurs in most rice-growing areas of the world, and Xoo isolates from within and across Asia, Africa, and Australia show a great diversity of genotypes based on the polymorphisms of transposable elements, predominantly insertion sequences (IS), avirulence genes, rep/box elements, and other markers [57]. The great diversity of strains within Xoo undoubtedly reflects adaptation of the pathogen to the diversity of host genotypes as well as the diverse environmental conditions in which rice is grown. Our transcriptome profiling analysis revealed some interesting aspects First, a large set of genes associated with the expression of Hrp genes and T3 effectors were significantly up-regulated in P6 and P2. This finding suggests that hrp genes and genes encoding T3 effector expression may differ for these strains in rice varieties. In phytopathogenic bacteria, the T3S system is encoded by hrp genes for eliciting HR on non-host or resistant host plants and for pathogenesis on susceptible hosts [55]. More and more evidence demonstrates that Hrp proteins and TAL effectors play key roles in host immunity responses or facilitate nutritional or virulence processes in the pathogen. These may trigger a resistance response in plants that contains TAL effector recognition features (such as AvrXa27 and AvrXa10), some of which are critical for virulence (such as PthoXo1, PthoXo6 and PthoXo7), and others of which appear to have more moderate or contextual functions in virulence [58]. The genomic sequences of the published Philippine (PXO99 A ), Japanese (MAFF311018) and Korean (KACC10331) strains contain 19, 17 and 15 TAL effector genes, respectively, and the African strain BAI, may contain eight TAL effector genes [21]. However, the X. oryzae strains in the United States lack TAL effectors and exhibit weak pathogenicity and a severely limited range of host cultivars compared with the Asian and African Xoo strains [59].
Although Illumina data alone were not sufficient to decipher the complicated and repetitive nature of the TAL effector coding sequences in P2 and C5, our results indicate that the expression patterns of hrp genes and T3 effectors were significantly different between Xoo strains from China and the Philippines. Together with the phenotypes of rice varieties, growth curves in Hang-you715 and Hanyou53, and expression patterns of the pathogen, we speculated that the increased virulence of P6 and P2 compared with C5 when infecting some rice varieties was due to the differentially up-regulated expression of hrp genes and T3 effector genes, which might promote Xoo multiplication in rice plants. The finding that three TAL effectors targeting the OsSWEET family of sucrose transporters conferred an increased virulence to the weakly pathogenic USA X. oryzae strain supported our speculation [60]. So far, whether the divergence of T3 effectors of Xoo potentially correlate with the geographic origin and diversity of rice cultivars remains a mystery. Answering this question will be facilitated by the determination of the full genome sequence of more Xoo strains.
Second, a large set of genes encoding chemotaxis and proteins involved in bacterial motility were significantly up-regulated in C5. Motility over solid surfaces is an important bacterial mechanism that allows complex social behaviors and pathogenesis. In some plant-pathogen systems, flagella-driven chemotaxis plays a role in the early interactions with host plants, and motility enables foliar pathogens to reach internal sites in the leaves [61]. Moreover, the bacterial protein flagellin has been found to be a plant elicitor, and plants have a sensitive perception system for this protein [62,63]. Even though the genes encoding Hrp proteins and T3 effectors were significantly down-regulated in C5, there was no significant difference in the LLs of 14 rice varieties infected by C5, P6 and P2. We hypothesize that this is due to C5's stronger motility, which might compensate for the weaker expression levels of Hrp proteins and T3 effectors, and allows C5 to exhibit similar virulence levels with P2 and P6 in some rice varieties. Verdier et al. (2012) recently reported that the plant genetic background affected the level of virulence enhancement by Xoo TAL effectors [60]. This also provides a clue to why there was no significant difference in virulence levels among P2, P6 and C5 when they infected some rice varieties.
Our analysis of the Xoo transcriptomes based on deep transcriptome sequencing led to remarkable insights into the transcriptional landscape of this important model plant pathogen from different countries, and it offers new avenues for improving our understanding of the Xoo-rice pathogenic mechanism and the evolution of Xoo virulence. Further understanding of the roles of bacterial motility and TAL effectors in diverse plant genetic backgrounds would shed light or provide further insights into their roles in interaction with the host, especially when we include the differing or diverse genetic background of the rice varieties and the environment or ecosystem where the crop is grown.