Background

Bacterial blight of rice, one of the most devastating rice diseases, is caused by the gram-negative bacterium Xanthomonas oryzae pv. oryzae (Xoo). Xoo colonizes and spreads along the vascular tissue of rice leaves, causing severe loss of yield and damage to grain quality (Mew 1987; Nino-Liu et al. 2006). The most economical and environmentally friendly method of controlling bacterial blight of rice is utilizing genetically inheritable resistance (R) genes. More than 40 R genes for bacterial blight of rice were identified, and 15 were cloned and characterized, including Xa1, Xa1–2, Xa4, xa5, Xa7, xa13, Xa14, Xa3/Xa26, Xa10, Xa21, Xa23, xa25, Xa27, Xa31(t), and xa41(t). Xa4, widely utilized in the 1980s, encodes a cell wall-associated kinase (WAK) protein that confers resistance to Xoo strains by enhancing cell wall strength (Hu et al. 2017). Xa7, Xa10, Xa27, and Xa23 are executor R genes that can confer bacterial blight resistance once induced by Xoo strains (Gu et al. 2005; Tian et al. 2014; Wang et al. 2015; Chen et al. 2021). The wide utilization of the R gene in rice breeding leads to the co-evolution of Xoo virulence.

In China, Xoo strains are divided into nine pathotypes according to the disease response of Xoo on five rice cultivars, including Nangeng 15, Java14, JG30, IR26, and Tetep (Fang et al. 1990). In the 1980s and 1990s, the Xoo strain of pathotype IV was the most virulent and epidemic Xoo strain type in South China. The wide utilization of Xa4, which is resistant to Xoo strains of pathotype IV while susceptible to that of pathotype V, has led to an increase in pathotype V Xoo strains. In the last decade, various R genes were utilized for resistance rice breeding, including Xa3, xa5, Xa7, Xa21, and Xa23, etc. In recent years, the percentage of pathotype V in the fields of South China has been declining while that of pathotype IX is increasing (Zeng et al. 2002). It is urgent to characterize the mechanism underlying the interaction between rice-Xoo and develop resistant rice cultivars to secure rice production.

The major virulent effectors in Xoo are secreted by the type III secretion system (T3SS) encoded by the hrp gene cluster. Type III effectors are mainly divided into two types, including transcription activator-like (TAL) effectors and non-TAL effectors (Furutani et al. 2009). TAL effectors are highly conserved at the nucleotide and amino acid levels and consist of three functional domains, including an N-terminal secretion and translocation domain, a central repeat domain-containing varying numbers of central repeats with 34 amino acids, and a C-terminal domain with nuclear localization signal and transcription activation domain. The 12th and 13th amino acids (the repeat variable di-residues, or RVD) of each repeat can recognize one nucleotide specifically. Accordingly, the repeat number and composition of each repeat’s 12th and 13th amino acids determine the TAL effector binding element (EBE) sequence in the promoter of DNA targets (Boch et al. 2009). For example, TAL effectors AvrXa7, PthXo3, TalC, and Tal5 specifically recognize EBE-AvrXa7, EBE-PthXo3, EBE-TalC, and EBE-Tal5b in the promoter region of bacterial blight susceptible gene OsSWEET14, respectively, to induce susceptible response of rice plants (Eom et al. 2019). Meanwhile, several non-TAL effectors have also been identified and characterized to suppress the pattern-triggered immunity (PTI) of rice. For example, XopK of PXO99A, which harbors E3-ubiquitin-ligase activity, can ubiquitinate and degrade embryogenic receptor kinase 2 (OsSERK2), thereby inhibiting PTI upstream of the MAPK cascade, and also, XopZ is reported to be essential for the full virulence of PXO99A (Song and Yang 2010; Qin et al. 2018).

AP2/ERF (APETALA2/ethylene-responsive element binding factors) transcription factor family is a large family of plant-specific transcription factors involved in plant development and stress response (Riechmann and Meyerowitz 1998; Nakano et al. 2006). According to the number of AP2/ERF domains, the superfamily is divided into four families: AP2, ERF, RAV, and soloist (Sharoni et al. 2011). The ERF family members, which have only one AP2/ERF domain, play an important role in abiotic and biotic stress. In biotic stress, OsERF123 was identified as a bacterial blight susceptible gene; it was induced by TAL effector TalB from African Xoo strain MAI1, leading to enhanced virulence on rice plants (Tran et al. 2018). OsERF922, induced by the rice blast fungus, negatively regulates rice blast resistance (Liu et al. 2012). Overexpression of OsERF922 represses the expression of defense-related genes and makes rice plants more susceptible to blast fungus. OsERF3 is rapidly induced by the rice striped stem borer pathogen (Lu et al. 2011). It positively affects two (mitogen-activated protein kinases) MAPKs and two WRKY gene expression levels, as well as jasmonate (JA) and salicylate (SA) concentrations. These findings suggest that ERF family genes play key roles in biotic stress response, including bacterial blight of rice.

In this study, the genome of three Xoo strains, GD0201, GD1358, and GD0202, belonging to Chinese pathotypes IV, V, and IX, were sequenced and assembled. Based on genomic information and effector analysis, we found that GD0202 contains only 10 complete TAL effector coding genes, far fewer than the number of TAL effector coding genes in GD0201 and GD1358. The transcriptomic analysis of ZH11 leaves inoculated with GD0201, GD1358, and GD0202 revealed that 40 ERF genes were repressed by all three Xoo strains, which strongly suggests that Xoo strains facilitate infection by repressing ERF gene expression. Through the weighted gene co-expression network analysis (WGCNA) and protein-protein interaction (PPI) analysis, 14 hub genes were identified. The expression of several hub genes was validated to be induced by all three Xoo strains with RT-qPCR. These results provide insights into the molecular mechanism underlying the interaction between rice and Xoo and the gene expression pattern of Xoo infection.

Results

Isolation and classification of GD0201 and GD0202 strain

Xoo strains were isolated from fields in Guangzhou, Guangdong Province, China, in 2017 and 2020. The strain isolated in 2017 was designated as GD0201, while that isolated in 2020 was designated as GD0202. The pathotype of GD0201 and GD0202 was determined by inoculating rice cultivars, including Nangeng 15, Java14, JG30, IR26, and Tetep (Fang et al. 1990). According to the classification criteria, GD0201, incompatible with IR26 and compatible with the other four rice cultivars, belongs to pathotype IV (Table 1). GD0202, compatible with all five rice cultivars, was classified as pathotype IX (Table 1). GD1358 is a strain of Chinese pathotype V isolated decades ago (Zeng et al. 2002).

Table 1 Pathotype classification of Xoo strains according to the disease response of rice
Full size table

The virulence of GD0201, GD1358, and GD0202 was tested on rice cultivars containing various R genes (Additional file 1: Table S1). Rice cultivars IRBB5, IRBB8, CBB23, and IRBB27, which harbour R gene xa5, Xa8, Xa23, and Xa27, respectively, conferred strong resistance to GD0201. IRBB5, CBB23, and IRBB27 also confer strong resistance to both GD1358 and GD0202. These results imply that currently, IRBB5 (xa5), CBB23 (Xa23), and IRBB27 (Xa27), are able to confer strong resistance to Xoo in fields of China and are good choices for resistance rice breeding.

To further verify the virulence of GD1358 and GD0202, 206 rice cultivars collected worldwide, which belong to the 3000 Rice Genomes Project, were inoculated with GD1358 and GD0202 in the field in 2018 and 2020, respectively. All rice varieties tested were susceptible to GD0202, while 89.8% of the cultivars were susceptible to GD1358 (Additional file 1: Table S2). This indicates that both GD1358 and GD0202 are highly virulent Xoo strains, while the virulence of GD0202 is more potent than that of GD1358.

Sequencing, function annotation, and phylogenetic analysis of GD0201, GD1358, and GD0202 genome

Genomes of the three Xoo strains were sequenced with third- and second-generation sequencing to figure out the difference between the three genomes. The sequencing and assembly information are listed in Additional file 1: Table S3, while the graphical representation of genomes is shown in Fig. 1. The genome size of GD0201 is ~ 5.05 Mb and contains 4866 predicted genes and 19 TAL effector coding genes (tals). GD1358 has a genome size of ~ 5.15 Mb and is predicted to contain 4902 genes and 18 tals. In contrast, GD0202 possesses a smaller genome size of ~ 4.97 Mb with an estimated 5017 genes, surpassing the gene number of both GD0201 and GD1358 (Table 2).

Fig. 1
figure 1

Graphical representation of GD0201, GD1358, and GD0202 genomes. The outmost circle is the size of the genome; the second and third circles represent coding sequence (CDS) on the sense and antisense strands, respectively; different colors represent CDS with different COG (Clusters of Orthologous Genes) functional classifications; the fourth circle are rRNA and tRNA; the fifth circle shows the GC content. The outward red bars indicate GC content is higher than the whole genome, and the higher the bar, the larger the difference. The inward blue bars indicate GC content is lower than that of the whole genome; the higher the bars, the larger the difference. The innermost circle is the GC skew, calculated by G-C/G + C; when the value is positive, the sense strand tends to transcribe CDS, while when the value is negative, the antisense strand tends to transcribe CDS

Full size image
Table 2 Summary of three Xoo genome genomes
Full size table

Besides GD0201, GD1358, and GD0202, the genomes of 12 Xoo strains originated from different regions were subjected to Average Nucleotide Identity (ANI) analysis. ANIb analysis showed that GD0201, GD1358, and GD0202 had average identities of 99.47, 99.53, and 99.39% to PXO99A (Additional file 1: Table S4). ANIm analysis revealed that GD0201, GD1358, and GD0202 exhibited average identities of 99.49, 99.57, and 99.50%, respectively, to PXO99A (Additional file 1: Table S4). The similarity between the three strains and Asian Xoo strains exceeded 99%, while the similarity between the three strains and African Xoo strains fell below 98%. This suggests that GD0201, GD1358, and GD0202 are evolutionarily close to Asian Xoo strains. Therefore, GD0201, GD1358, and GD0202 are identified as Xoo strains based on ANI analysis.

To investigate the population structure of Xoo strains further, we performed phylogenetic analyses of GD0201, GD1358, GD0202, and 83 other Xoo strains with their protein sequences (Fig. 2). According to the phylogenetic tree, all Xoo strains were divided into two lineages. One lineage is African Xoo strains. In the phylogenetic tree, one clade contains only African strains, such as AXO1947, MAI1, and BAI3, indicating that African Xoo strains were phylogenetically close to each other. The other lineages predominantly comprise Asian strains. GD1358, a type V strain isolated more than 30 years ago, clusters closely with Xoo strains from the Philippines, including PXO86, PXO145, PXO236, PXO83, and PXO211, which were also isolated decades ago (Oliva et al. 2019). While GD0201 and GD0202, isolated within the last 10 years, exhibit close phylogenetic proximity to Xoo strains GXO2001 and GXO2006, isolated in Guangxi Province, China, the neighboring province of Guangdong. This suggests that the evolution of Xoo strains was influenced by the local environment over time.

Fig. 2
figure 2

Phylogenetic analysis of 86 Xoo strains from different regions worldwide. African strains, pink; Chinese strains, orange; the Philippines strains, blue; Indian strains, purple; Korean strains, green; Japanese strains, yellow. The brown square represents the strains containing complete TalAG effector. The green circle represents the strains that contain truncated TalAG effector. The blue star represents the strains that do not contain TalAG effector

Full size image

Non-TAL and TAL effectors encoded by the GD0201, GD1358, and GD0202

Non-TAL effectors are secreted by the T3SS of Xoo and are involved in host immunity suppression (Furutani et al. 2009; Song and Yang 2010). To identify non-TAL effectors encoded by the three Xoo strains, HMM files corresponding to various non-TAL effectors in the National Center for Biotechnology Information (NCBI) protein family models were utilized to search three Xoo strain genomes. The HMM accession number, sequence cutoff, and domain cutoff for each non-TAL effector are listed in Additional file 1: Table S5. According to the results, the number of non-TAL effectors in GD0202 is fewer than that of GD0201 and GD1358. Besides the three Xoo strains, we have also analyzed the non-TAL effectors in other Chinese Xoo strains. We found AvrBs2, XopF, XopK, XopL, XopN, XopV, XopX, XopZ, XopAA, XopAD, XopAE, and XopAF were presented in all the analyzed Xoo strains, indicating the importance of these non-TAL effectors for the virulence of Xoo strains. While XopW presents in 8 out of 18 Chinese Xoo strains, and XopI only presents in African strain AXO1947. Notably, GD0202 harbors the fewest non-TAL effector coding genes among all the analyzed Xoo strains, totaling 12. Further investigations were needed to study whether the number of non-TAL effectors is related to Xoo strains’ virulence.

According to the genome sequence of three Xoo strains, TAL effectors form clusters in the genome (Fig. 3). RVDs of TAL effectors from the three Xoo strains are listed in Additional file 1: Table S6–S8. Interestingly, GD0202 harbors 14 tals, but only 10 of these genes encode TAL effectors with RVDs; the remaining four genes are pseudogenes encoding truncated TAL effectors lacking the central repeat domain. Furthermore, tal1c-GD0201 and tal1c-GD0202 are the same as AvrXa23D, and tal8b-GD1358 is AvrXa23A. Both AvrXa23D and AvrXa23A are capable of inducing the expression of the R gene Xa23 (Xu et al. 2022). As a result, all three Xoo strains are incompatible with CBB23. Also, all three Xoo strains harbor an AvrXa27-like effector coding gene in the genome, leading to an incompatible response with IRBB27.

Fig. 3
figure 3

TAL effectors in Xoo strains. Arrangement of tals in genomes of three Xoo strains. Each box represents one cluster of tals in the genome of Xoo. One arrow indicates one tal gene, and the arrow’s orientation represents the tal gene’s orientation. Tals with the same color are identical. White tals are strain-specific effectors based on the genome sequence. Pseudogenes have a premature stop codon and do not encode any central repeats, which are indicated in grey boxes

Full size image

Besides the three Xoo strains, we analyzed more than 15 Chinese Xoo strains. All TAL effectors were categorized into different classes utilizing annoTALE (Grau et al. 2016) and listed in Additional file 1: Table S9. Consistent with previous studies, TAL effectors in the African Xoo strain AXO1947 were classified into distinct classes compared to those of Asian Xoo strains. The Philippines strains PXO99A and PXO71 harbor TAL effectors classified into classes that are absent from Chinese Xoo strains. This indicates that location is an important factor that influences the evolution of tals. Furthermore, the absence of AvrXa23-like effectors in strains C9-3, HuN37, and JL33, as well as the lack of AvrXa27-like effectors in strains JL25, JL28, and ScYc-b, suggests these strains may defeat the resistance of rice cultivars harboring Xa23 or Xa27. Thus, novel R genes should be utilized to prevent the outbreak of strains lacking AvrXa23 and AvrXa27. According to the analyzed results, effectors assigned to TalAP present in all the Asian Xoo strains analyzed, and those assigned to TalAE and TalAG classes also present extensively in Asian Xoo strains. We analyzed 59 Asian strains and found that 51 harbor effectors assigned to TalAG, and 2 strains (PXO99A and PXO71) contain truncated TalAG, while six Xoo strains lack this TAL effector from this class (Fig. 2). This suggests effectors belonging to TalAP, TalAE, and TalAG classes may be important for the virulence of Asian Xoo strains.

ERF genes may be repressed by Xoo strains to facilitate infection

To identify key genes and pathways regulated by Xoo stains to facilitate infection, leaves of Zhonghua11 (ZH11) were inoculated with GD0201, GD1358, and GD0202, respectively, with syringes. Inoculated leaves were harvested 1 hour and 48 hours post-inoculation (hpi) for cDNA library preparation and sequencing. After a quality check, the clean reads were mapped to the Nipponbare reference genome. Over 90% of clean reads were successfully mapped to the Nipponbare genome.

To understand the pattern of gene expression induced by GD0201, GD1358, and GD0202, a pairwise comparison was made between ZH11 leaves inoculated with each strain at 1 hpi vs 48 hpi. A total of 10,239 significantly differential expressed genes (DEGs) (5083 down-regulated and 5156 up-regulated) were identified between ZH11 leaves inoculated with GD0201 1 hpi and 48 hpi. For ZH11 leaves infected with GD1358 1 hpi and 48 hpi, 8662 genes (4930 down-regulated and 3732 up-regulated) were identified. Meanwhile, 9936 genes (5126 down-regulated and 4810 up-regulated) were differently regulated by GD0202. In total, 3422 and 2808 genes were down-regulated and up-regulated by all three Xoo strains, respectively (Additional file 2: Figure S1).

The top 10 most highly up-regulated and down-regulated genes by the three Xoo strains were selected, and replicates were deleted. Of the 20 downregulated gene sets, 6 genes were found to encode proteins belonging to the ERF protein family (Fig. 4a). Interestingly, 40 out of 138 ERF genes in the rice genome were commonly downregulated by all three Xoo strains (Fig. 4b). Only two ERF genes were commonly up-regulated. Eight ERF genes, including LOC_Os08g36920, LOC_Os08g43200, LOC_Os06g03670, LOC_Os02g52670, LOC_Os02g54050, LOC_Os01g66270, LOC_Os01g12440, and LOC_Os11g06770 were strongly repressed by all three Xoo strains (Fig. 4b). Several ERF genes were reportedly involved in rice response to pathogens (Lu et al. 2011; Liu et al. 2012). This strongly suggests that the three Xoo strains may facilitate the infection process by repressing ERF gene expression.

Fig. 4
figure 4

Heatmap of the top 10 regulated genes and members of the ERF family regulated by three Xoo strains after inoculation 48 hours at the tillering stage. a The top 10 most highly up-regulated and down-regulated genes are found in three Xoo strains. b The heatmap of 42 ERF genes regulated by three Xoo strains

Full size image

Modules responsible for biotic and abiotic stress were correlated with Xoo infection

WGCNA was performed to construct co-expressed networks and identify co-expression modules. The sample hierarchical clustering analysis suggested by the resulting trees did not identify strong outliers except for one sample treated with water for 2 days, but it had little effect on subsequent results (Additional file 2: Figure S2). The gene expression matrix contains 17,764 genes. After being filtered using the MAD (median absolute deviation) method, the top 15,000 genes were retained.

In this study, the fitting degree of the scale-free topological model was 0.85, and the soft threshold for network construction was selected as 14. The obvious modules were identified, and different colors were used to represent them. A total of 13 modules including MEblack (287 DEGs), MEblue (5387 DEGs), MEbrown (912 DEGs), MEgreen (619 DEGs), MEgreenyellow (90 DEGs), MEgrey (418 DEGs), MEmagenta (123 DEGs), MEpink (273 DEGs), MEpurple (96 DEGs), MEred (448 DEGs), MEtan (36 DEGs), MEturquoise (5639 DEGs), MEyellow (672 DEGs) were obtained (Additional file 2: Figure S3). These modules were positively or negatively correlated with the response to the bacterial blight.

ZH11 inoculated with ddH2O at 1 hpi and 48 hpi and inoculated with three Xoo strains at 1 hpi were set as controls, and ZH11 inoculated with three strains at 48 hpi were set as case for gene modules and phenotype correlation analysis. When the absolute values of cor >0.5 and p <0.01 were set as the strict criterion, several specific modules were identified as significantly correlated with rice-Xoo interaction. The MEblue (cor = 0.81, p = 1.9e-6), MEpurple (cor = 0.76, p = 1.5e-5), MEbrown (cor = 0.58, p = 3.2e-3) modules showed a high correlation with the case. The MEyellow (cor = 0.85, p = 1.1e-7), MEturquoise (cor = 0.76, p = 1.7e-5), MEpink (cor = 0.62, p = 1.1e-3), and MEblack (cor = 0.57, p = 3.5e-3) modules showed high correlation with control (Fig. 5a).

Fig. 5
figure 5

WGCNA and PPI analysis of ZH11 leaf samples inoculated with three Xoo strains. a Heatmap of the correlation between Eigengene and disease phenotype of ZH11 inoculated with ddH2O and three Xoo strains. b The 14 hub genes identified from the interaction network analysis within the purple module. Nodes represent genes. Degree defines the node size. Betweenness centrality defines the color of the node. The combined score defines the thickness of the edge

Full size image

Gene Ontology analysis revealed that genes belonging to the MEblue, MEturquoise, and MEyellow modules are mainly related to metabolic processes. The genes belonging to the MEpruple module were found to be mainly related to stress and defence response. A protein-protein interactions network of expressed genes in the MEpurple module was constructed for hub gene identification. Genes in the MEpurple module were imported into the STRING database for protein interaction analysis and were divided into three clusters using the K-means clustering algorithm. Clustered results were then imported into Cytoscape for analysis. The protein-protein interaction networks of expressed genes in the MEpurple module were present in Fig. 5b. Fourteen genes have been identified as hub genes that are potentially related to rice response to the three Xoo strains. According to the trait ontology, most of these hub genes are related to disease response and abiotic tolerance (Additional file 1: Table S10) (Kurata and Oryzabase 2006). Furthermore, we validated the expression pattern of several hub genes in response to three Xoo strains. Os03g0743500 (OsCML4), Os06g0727400 (OsRLCK220), Os01g0660200 (PR31), Os01g0134700 (CBP-like), and Os04g0611400 (VSR7) exhibited induced expression in response to all three Xoo strains (Fig. 6). Notably, the fold change of OsCML4 is highest among the analyzed hub genes, suggesting a crucial role of OsCML4 played in the biotic stress imposed by GD0201, GD1358, and GD0202. Further investigation is needed to verify the involvement of hub genes in Xoo response.

Fig. 6
figure 6

The expression level of several hub genes in response to three Xoo strains. a–g The relative expression level of Os03g0743500 (OsCML4), Os06g0727400 (OsRLCK220), Os06g0727400 (OsRLCK220), Os01g0134700 (CBP-like), Os04g0611400 (VSR7), Os06g0288100 (SOBIR1), and 08 g0386200 (WRKY69) in ZH11 at 0, 3, and 4 days post-inoculation (dpi) with the GD0201, GD1358, and GD0202. Bars represent average expression obtained from three independent RNA samples, with standard deviation. The letters above the bars represent the result of Tukey’s HSD test. Identical letters correspond to means that are not significantly different from each other (α = 0.05). This experiment was repeated three times with similar results

Full size image

Discussion

GD1358 is a Xoo strain isolated over 30 years ago in Guangdong Province, China (Fang et al. 1990). GD0201 and GD0202 are two Xoo strains isolated in Guangdong Province in recent years. Since the 1980s, rice cultivars moderately resistant or resistant to epidemic Xoo strains were planted in fields in China to ensure the yield of rice country-wide. As a result, epidemic Xoo strains were subjected to selection pressure from resistant rice cultivars. Therefore, the virulence of Xoo strains is evolving. From the 1990s to the present, the major virulent Xoo strain types in South China have been pathotypes IV, V, and IX sequentially. GD0201, GD1358, and GD0202 are classified as pathotypes IV, V, and IX, respectively. The complete genome assembly of three Xoo strains, identification of effectors and transcriptomic analysis of DEGs regulated by three Xoo strains are significant for understanding the mechanism underlying Xoo-rice interaction and providing insights into bacterial blight-resistant rice breeding.

Our study found bacterial blight R genes, including xa5, Xa23, and Xa27, confer strong resistance to pathotypes IV, V, and IX in China. Currently, these R genes are widely utilized in rice breeding in China. However, virulent Xoo strains that are able to overcome the resistance conferred by these R genes are emerging. For example, Xoo strain JNXO and LYG50 overcome the resistance of xa5; strain C9-3 and AH28 can overcome the resistance of Xa23 (Chen et al. 2019; Chen et al. 2022). This suggests the resistance conferred by bacterial blight R genes is not permanent; they can be overcome by the Xoo strains that lose certain avirulent effectors or gain novel effectors. This suggests that more efforts should be made to understand the molecular mechanism underlying the virulence of Xoo strains.

Based on the complete genome sequence of the three Xoo strains we studied, we found there are fewer complete tals in GD0202 compared to GD0201, GD1358, and most Asian Xoo strains. Many of the tals in GD0202 encode pseudogenes, meaning they encode truncated TAL effectors. Loss of avirulent tals may lead to strong virulence of GD0202. Meanwhile, TAL effectors belong to TalAP, TalAE, and TalAG classes and are present extensively in most Asian Xoo strains. This also suggests that TalAP, TalAE, and TalAG class effectors are important for the virulence of Asian Xoo strains, and the effector-binding elements of TalAP, TalAE, and TalAG are good choices for generating broad-spectrum resistant rice plants.

In the transcriptomic analysis of ZH11 leaves inoculated with three Xoo strains, forty members of the ERF protein family genes were repressed by all three Xoo strains. This strongly suggests that Xoo strains may facilitate the infection process by repressing the expression of ERF protein family genes. Several members of the ERF protein family gene have been reported to be involved in plant immunity. For example, OsERF123 is a bacterial blight S gene induced by TAL effector TalB from African Xoo strain MAI1 (Tran et al. 2018). In addition, rice blast pathogen Magnaportha oryzae can induce the expression of OsERF83, which positively regulates rice resistance to rice blast (Tezuka et al. 2019). Ectopic expression of OsAP2/ERF152 is able to induce the expression of SA, JA, and ET-responsive genes and enhance rice resistance to bacterial and fungal infection (Pillai et al. 2020). These results indicate the ERF protein family positively regulates rice resistance to bacteria and fungi. This strongly suggests that GD0201, GD1358, and GD0202 may facilitate its infection process by repressing the expression of ERF protein family genes. We will further explore the function of the ERF protein family in bacterial blight response.

According to WGCNA analysis, fourteen hub genes were identified to be potentially related to response to Xoo strains. According to the gene Trait Ontology (Kurata and Oryzabase 2006), most of the hub genes were related to plant disease response or abiotic stresses. Os03g0743500, highly induced by three Xoo strains, is a calmodulin-like gene (OsCML4) that functions in regulating different abiotic stresses, such as drought and salt stress (Yin et al. 2015; Asif et al. 2022). Overexpression of OsCML4 significantly improves rice growth performance and enhances survival rates under drought conditions. This is achieved by efficiently clearing reactive oxygen species (ROS) and inducing other stress-related genes in an abscisic acid (ABA)-independent manner, imparting drought resistance (Yin et al. 2015). This mechanism may also enhance the resistance of rice to bacterial blight. Further investigation into the specific functions and regulatory mechanisms of OsCML4 in response to Xoo infection could provide valuable insights into the molecular basis of rice resistance to this pathogen. Other hub genes, such as PR31 and OsRLCK220, which are involved in disease response, were also induced by three Xoo strains, suggesting their potential role in bacterial blight. Interestingly, hub genes Os06g0288100 and 08 g0386200 were not induced at 3 and 4 dpi. Os06g0288100 encodes an LRR receptor kinase (OsSOBIR1) and participated in rice PTI (pattern-triggered immunity) response and antiviral defense (Zhang et al. 2021). 08 g0386200 encodes a WRKY transcription factor 69 that has been reported to respond to M. oryzae infection (Li et al. 2021). It is possible that Os06g0288100 and 08 g0386200 were induced at the early stage, and the expression of Os06g0288100 and 08 g0386200 were declined at 3 and 4 dpi. Alternatively, it is possible that these two genes may not be directly responsible for the response to Xoo strains. The function of all hub genes in rice response to Xoo strains needs to be further verified.

Conclusions

This study identified three highly pathogenic Chinese bacterial blight strains, GD0201, GD1358, and GD0202. By genome sequencing and assembly, their phylogenetic relationship and effectors were analyzed. xa5, Xa27, and Xa23 confer strong resistance to all three Xoo strains. The ERF and 14 hub genes were potentially involved in rice response to three Xoo strains with transcriptomic analysis.

Methods

Isolation of Xoo strains

Xoo-infected rice leaves were isolated from infected rice plants in the field. Briefly, infected rice leaves were sterilized by immersing them in 75% ethanol for 1 minute and washed three times using sterile distilled water. Subsequently, infected leaves were homogenized with mortar and pestle. The homogenate and its diluent were spread on peptone sucrose agar (PSA) plates (10 g/L peptone, 10 g/L sucrose, 1 g/L glutamic acid, 16 g/L agar, and pH 7.0) and incubated at 28°C for 2 d until single colonies were formed. Single Xoo colonies were isolated from solid plates and stored under − 80°C for further virulence verification. GD0201 and GD0202 were isolated in 2017 and 2020, respectively, from Guangdong, China. GD1358 is kindly provided by Dr. Yin Zhongchao from Temasek Life Sciences Laboratory.

Xoo strain inoculation and disease assays

Xoo strain inoculation was carried out using the leaf-clip** method. Briefly, Xoo strains were cultivated on PSA plates for 2 d at 28°C. Xoo strains were collected and resuspended in sterile distilled water at an optical density of 0.5 at 600 nm. Xoo strains were applied to rice leaves by cutting about 5 cm from the leaf tips using scissors dipped in the inoculum. At least 3 individual rice plants of one cultivar were inoculated with one Xoo strain. Lesion length was measured 14 days after inoculation with rulers. For RNA-seq analysis, Xoo strains were infiltrated into rice leaves with syringes. Leaves samples were harvested 0, 1 or 48 h after infiltration.

Plants and growth conditions

In 2020 and 2021, rice plants were grown in paddy fields at the South China Botanical Garden in Guangzhou, China. Six-week-old rice plants were utilized for Xoo inoculation and disease scoring.

DNA extraction and genome sequencing

Genomic DNA of Xoo strains was extracted with the Bacteria DNA kit (OMEGA; Cat: #D3494–01) according to the manufacturer’s instructions. DNA quality and quantity were determined using the TBS-380 fluorometer (Turner BioSystems Inc., Sunnyvale, CA). High qualified DNA sample (OD260/280 = 1.8 ~ 2.0, > 6 μg) is utilized to construct fragment library.

For Illumina pair-end sequencing of each strain, at least 3 μg high-qualified DNA was utilized for sequencing library construction. Paired-end libraries with insert sizes of ~ 400 bp were prepared following Illumina’s standard genomic DNA library preparation procedure. Purified genomic DNA is sheared into smaller fragments with a desired size by Covaris, and blunt ends are generated by using T4 DNA polymerase, adapters are ligated to the ends of the DNA fragments. The desired fragments can be purified through gel-electrophoresis, then selectively enriched and amplified by PCR. Finally, the qualified Illumina pair-end library will be used for Illumina NovaSeq 6000 sequencing.

For Pacific Biosciences sequencing, an aliquot of 8 μg DNA was spun in a Covairs g-TUBE (Covaris, MA) at 6000 rpm for 60 s using an Eppendorf 5424 centrifuge. DNA fragments were then purified, end-repaired and ligated with SMR Tbell sequencing adapters. Sequencing libraries were purified three times using 0.45× volumes of Agencourt AMPure XPbeads (Beckman Coulter Genomics, MA). 20 kb insert whole genome shotgun libraries were generated and sequenced on a Pacific Biosciences RS instrument using standard methods at Biozeron (Shanghai, China).

Genome assembly and annotation

The raw reads were trimmed and quality controlled by Trimmomatic (v0.36) (Bolger et al. 2014). ABySS was first used for GD0201 and GD1358 genome assembly with multi-Kmer parameters (Jackman et al. 2017). Canu was subsequently utilized to assemble PacBio-corrected long reads (Koren et al. 2017). Finally, GapCloser software was applied to fill the remaining local internal gaps and correct single-base polymorphism for final assembly results (Xu et al. 2020). Unicycler (0.4.8) (Wick et al. 2017) was used for GD0202 genome assembly.

For the prokaryotic organism, we used the ab initio prediction method to obtain gene models for Xoo strain. Gene models were identified using GeneMark (Besemer et al. 2001). Then all gene models were blastp against the non-redundant (NR in NCBI) database, SwissPort (http://uniprot.org), Kyoto Encyclopedia of Genes and Genomes (KEGG) (http://www.genome.jp/kegg/), and Database of Clusters of Orthologous Genes (COG) (http://www.ncbi.nlm.nih.gov/COG) for functional annotation. In addition, tRNAs were identified using the tRNAscan-SE (v1.23) (Chan et al. 2021), and rRNAs were determined using RNAmmer (v1.2) (Lagesen et al. 2007). The genome diagram was developed by the Circos (V0.69) software (Krzywinski et al. 2009). TAL effector coding sequences were analyzed using ANNOTALE 1.4 (Grau et al. 2016). Non-TAL effectors were searched with HMM files corresponding to different non-TAL effectors from NCBI Protein Family Models. The HMM accession number and corresponding sequence and domain cutoff of each non-TAL effector were listed in Additional file 1: Table S5.

RNA extraction, library construction, and sequencing

For mRNA isolation, leaves were ground to powder with liquid nitrogen and 100 mg of powder was utilized for RNA extraction. The total mRNA was isolated with Eastep Super RNA isolation kit (Promega, Cat.# LS1040). The yield and purity of RNA were assessed by recording absorbance at 260 and 280 nm on Nanodrop 8000 Spectrophotometer (Thermo Scientific). Library construction and sequencing were performed by Novogene (Bei**g, China).

Gene expression analyses

A 4 cm leaf section was entirely infiltrated and collected at 0, 3, and 4 days post-inoculation (dpi) of three strains, respectively. Each individual sample contained three independent infiltration areas, and three independent replicate samples were processed. After the grinding of the sample, total RNA was extracted from plant leaves using the Eastep Super total RNA isolation kit (Promega, Cat. # LS1040). 1μg total RNA was reverse transcribed into cDNA using HiScript II 1st Strand cDNA Synthesis Kit (Vazyme, Cat. R212–01). All gene expression studies were performed by RT-qPCR in a LightCycler (ROCHE), using ChamQ SYBR Color qPCR Master Mix (Vazyme, Cat. Q421–02). The average transcript levels of each gene were detected with gene-specific primers (Additional file 1: Table S11). The expression of the OsActin gene was used to standardize the RNA sample (Hu et al. 2017). High-quality (QV > 25) transcriptome libraries from all 24 samples were mapped individually to the Nipponbare genome using hisat2 (v2.2.1) with default parameters (Kim et al. 2019). Transcript abundance and differential gene expression were estimated using the subread and DESeq2 along with FPKM (Fragments Per Kilobase of transcript per Million mapped reads) from reference-guided map** (Love et al. 2014). Significance tests for differential expression were based on a modified exact test. FDR < 0.05 and |Log2FC| > 1 criteria were set to retrieve the significantly differential expressed genes (DEGs) between different time points. The ggplot2, a package of R, was used to produce the heat maps of differentially expressed genes.

Construction of WGCNA co-expression network

The WGCNA co-expression network was constructed using WGCNA shiny in TBtools (Langfelder and Horvath 2008; Chen et al. 2020). Inputting the gene counts format as the expression matrix. Pearson’s coefficient was used to calculate the correlation between genes to determine the co-expression similarities between the two genes. An appropriate adjacency matrix weight parameter β value was selected to satisfy the scale-free network distribution precondition, and the Topological Overlap Matrix (TOM) was constructed to cluster and segment the modules. The correlation between each network module and the sample phenotype were analyzed, to select modules related to the time points and different samples.

Identification and analysis of vital modules and key genes

The significance of the different time points and different rice sample modules were compared; the correlation of module-trait value cor ≥ 0.5, p < 0.05, indicating a statistically significant difference and vital module. The module eigengene (ME) value can be obtained by analyzing the correlation between gene expression and the corresponding module eigengene. The final value of the correlation coefficient is gene significance (GS). GS reflects the correlation between gene expression and phenotype data. The higher the ME and GS, the more relevant the gene is to the research phenotype.

Phylogenetic tree and ANI analysis

Besides GD0201, GD1358, and GD0202, genome sequences of Xoo stains were retrieved from the NCBI database. PhyloPhlAn3.0 (v3.0.67) was utilized to construct phylogenetic trees with all the protein sequences of each strain based on Multiple Sequence Alignment (MSA) results according to maximum likelihood method (Asnicar et al. 2020). iTOL (https://itol.embl.de/) was used to visualize the phylogenetic tree (Letunic and Bork 2021). The average nucleotide identity (ANI) values among 13 genome sequences, including GD0201, GD1358, GD0202, and other Xoo strains were calculated using the JSpeciesWS online service (https://www.ribocon.com/jspeciesws.html) (Richter et al. 2015).

Function enrichment and visualization of key genes

Key genes were subjected to the Gene Ontology (GO) database and the Kyoto Encyclopedia of Genes and Genomes (KEGG) for function and pathway enrichment analysis (Ashburner et al. 2000; Kanehisa and Goto 2000; Consortium et al. 2023); p < 0.05 was the significant threshold. Cluster and visualization of key genes were constructed with the STRING database and analyzed network plugins of Cytoscape (3.9.1) (Shannon et al. 2003).