OsBLS6.2:A rice bacterial leaf streak resistance gene identified by GWAS and RNA-seq

Huin Xie, Chunzi Lin, Wenyu Lu, Zhiki Hn, Dnhong Wei, Xing Huo, Tinjio Li, Jin Zhng,Yongqing He, Chun Chen, Hui Wng, To Guo,*, Jifeng Wng,*

a National Engineering Research Center of Plant Space Breeding, Guangdong Key Laboratory of Plant Molecular Breeding, South China Agricultural University, Guangzhou 510642, Guangdong, China

b Guangdong Key Laboratory of New Technology in Rice Breeding, Guangdong Rice Engineering Laboratory, Rice Research Institute, Guangdong Academy of Agricultural Sciences,Guangzhou 510640, Guangdong, China

c Guangxi Key Laboratory of Agro-environment and Agro-products Safety, College of Agriculture, Guangxi University, Nanning 530004, Guangxi, China

Keywords:Rice Bacterial leaf streak Xanthomonas oryzae pv.oryzicola Genome-wide association study RNA sequencing

ABSTRACT Bacterial leaf streak(BLS),caused by Xanthomonas oryzae pv.oryzicola(Xoc),is a bacterial disease affecting rice production in Asia and Africa, whose severity is expected to increase with climate change.Identification of new quantitative-trait loci (QTL) or resistance genes for BLS resistance is essential for developing resistant rice.A genome-wide association study to identify QTL associated with BLS resistance was conducted using phenotypic and genotypic data from 429 rice accessions.Of 47 QTL identified,45 were novel and two co-localized with previously reported QTL or genes conferring BLS resistance.qBLS6.2 on chromosome 6 explained the greatest phenotypic variation.Combined analysis of differential expression and annotations of predicted genes near qBLS6.2 based on haplotype and disease phenotype identified OsBLS6.2(LOC_Os06g02960)as a candidate gene for qBLS6.2.OsBLS6.2 knockout plants showed higher resistance to Xoc than wild-type plants.Many other candidate genes for resistance to Xoc were identified.

1.Introduction

Rice,a crop that sustains more than half of humanity[1],is subject to bacterial leaf streak(BLS),caused by Xanthomonas oryzae pv.oryzicola (Xoc).BLS was first observed in the Philippines in 1918,and subsequently reported in tropical and subtropical regions of Asia, northern Australia, and west Africa [2].In China, it was first discovered in Guangdong province and has recently become one of the major diseases in southern China [3,4], causing yield losses of 10%–32% [5].Although the simplest method to prevent BLS is chemical pesticide application, their excessive use pollutes the environment and poses food safety concerns [6].Because Xoc infects rice leaves mainly through stomata or wounds, large-scale infections caused by adverse weather conditions such as typhoons are difficult to control effectively with pesticides.Breeding and improving disease-resistant varieties is the most effective way to prevent and control the disease.

At least 15 BLS resistance QTL have been mapped using classical QTL mapping methods [3,4,7–9], and key genes, qBlsr5a and bls1,have been identified.The functional gene of qBlsr5a is believed[4] to be xa5, which is also a bacterial blight resistance gene.bls1 was reported [10] to be controlled by OsMAPK6 for resistance to Xoc in 2021.Because qBlsr5a and bls1 are both recessive genes,their application in hybrid rice BLS resistance improvement is limited.A dominant resistance gene Xo1 showed complete resistance to African but not Asian strains [9].Some 29 QTL conferring BLS resistance have been identified in genome-wide association studies(GWAS),but there is insufficient understanding of the potential functional genes among them [6,11,12].At least 21 genes have been reported[13–27]to be associated with BLS resistance in rice,and a non-host R gene called Rxo1 was isolated in maize [28].Because BLS resistance genes or QTL may be strain-specific, BLS control may depend on identifying resistance QTL or genes that confer resistance to region-specific strains.

Most studies of resistance to Xoc employ parental-inheritance mapping populations.However, large-scale germplasm resources can help us identify more resistance genes and QTL for improving broad-spectrum, durable resistance in rice by marker-assisted selection and aggregation of multiple resistance QTL and genes.With the sharp decline in sequencing costs, high-density DNA markers have been produced on a large scale, a development that favors the detection of minor-effect QTL [29].The use of GWAS based on whole-genome resequencing has become widespread in plant disease resistance research, leading to the identification of resistance to stalk rot in sorghum[30],ear rot in maize[31],white mold in soybean [32], and blast [33–35] and bacterial blight [36–38] in rice.However, due to the large interval of gene mapping,pinpointing candidate genes can be challenging.GWAS cannot identify a unique target gene.Transcriptome analysis can overcome this limitation by detecting and distinguishing the expression of candidate genes in multiple genotypes [39].

Many studies demonstrated that pooling multiple extreme materials in equal proportions to create two extreme pools for transcriptome sequencing can yield candidate differentially expressed gene (DEG) associated with the target trait.To increase the correlation between identified DEGs and the target trait, the majority of RNA-seq studies utilize transcriptome analysis involving two or more extreme materials(or those subjected to relevant treatments) [39–41].However, acquiring individual transcriptomes for multiple materials incurs high sequencing costs, produces large amounts of data, and presents complex analytical challenges.In certain cases,due to variations in research materials and specific research goals, some researchers choose to conduct transcriptome analysis by pooling samples from multiple extreme materials.In the transcriptome of pooled extreme materials, the distinct allelic differences in these extremes lead to an enrichment of target genes associated with the trait within these extreme materials.As the number of materials increases, the background noise tends to become diluted.Consequently, the transcriptomes of multiple extreme materials might clearly highlight genes with significant effects, while being less suitable for genes with minor effects.Li et al.[42] conducted RNA-seq with a LN (Low Nitrogen)-responsive pool (five rice accessions) and LNunresponsive pool (five rice accessions).By integrating GWAS and RNA-seq, they identified 22 candidate genes associated with N utilization.Combining linkage disequilibrium and transcriptome analyses in four extremely high and low HI(harvest index)Brassica napus lines, Lu et al.[43] identified 33 functional candidate genes controlling HI-associated traits.Chen et al.[44] performed RNAseq in two pools: an L pool (consisting of 13 Recombinant Inbred Lines (RILs) with low levels of percentage of grains with chalk and degree of endosperm chalkiness (PGWC and DEC) and an H pool (comprising 13 RILs with high levels of PGWC and DEC).The identified 33 DEGs that were co-localized within the regions of three specific QTL (qDEC5, qDEC7, and qDEC8).Two genes linked to the chalky trait, SWEET3a and OsUAM3, were identified.

In this study, 429 rice accessions were inoculated with the strain GDXC-1 isolated from the Southern China rice BLS epidemic area.Their resistance to Xoc was assessed based on lesion length and lesion area.Subsequently, a GWAS was conducted, resulting in the identification of 47 QTLs associated with BLS resistance.Upon performing a transcriptome analysis of BLS-resistant and susceptible groups inoculated with Xoc, OsBLS6.2 was confirmed as the candidate gene for qBLS6.2.Further BLS resistance analysis showed that knocking out OsBLS6.2 can enhance the resistance of rice to GDXC-1, indicating that OsBLS6.2 is a promising BLS resistant gene with significant potential applications.

2.Materials and methods

2.1.Plant materials and pathogen

A panel of 429 rice accessions from 15 countries was used,consisting of 216 landraces and 213 cultivars(Table S1).The bacterial strain GDXC-1 was isolated and identified from diseased samples collected from the epidemic area in southern China.

2.2.Evaluation of rice resistance to Xoc

All accessions were sown in a nursery and then transplanted at 30 days.Each accession was planted in a row with a row spacing of 20 cm, and each row contained 3 holes, with 2 plants per hole.Standard local agricultural practices were used throughout the growing season, but no fungicides and bactericides were applied.Field management including irrigation, fertilization, and disease and pest control followed conventional production practice.The Xoc strain was cultured on nutrient agar medium (containing glucose) at 28 ℃for two days and diluted with phosphate-buffered saline (PBS) solution to OD600= 1 as inoculum.Inoculation was by pressure infiltration [45], On day 45 day after transplantation,a disposable syringe was employed to draw up inoculum.The syringe nozzle was placed against the leaves and the plunger was depressed to propel the inoculum into the intercellular spaces.Four to five fully expanded leaves at the top of each plant were inoculated.Disease symptoms were evaluated 21 days after inoculation.At least nine inoculated leaves from each accession were cut to appropriate lengths, affixed to white paper, and scanned with a scanner.Image-Pro Plus was employed to measure the length and area.Mean lesion length was used for evaluating drug resistance levels [12].

2.3.DNA sequencing, quality control and SNP calling

Fresh leaves of 21-day-old rice seedlings were collected for genomic DNA extraction by the CTAB method [46].DNA samples were fragmented into 350-bp pieces by ultrasonication and then used to construct libraries with the NEB Next Ultra DNA Library Prep Kit (New England Biolabs, Ipswich, MA, USA).Wholegenome sequencing was performed on an Illumina NovaSeq PE150 sequencer (BerryGenomics, Beijing, China) by paired-end sequencing to a sequencing depth of 10×.Each sample yielded about 3 Gb of reads.The reads were aligned to the reference genome (version IRGSP-1.0; https://rapdb.dna.affrc.go.jp/download/irgsp1.html) using Clean Reads.SNP detection was performed using the GATK software package [47].Vcftools [48] was used to select 3,576,716 SNPs with a minimum allele frequency (MAF)greater than 5% and a missing rate less than 20% from the preliminary population SNPs.

2.4.Population structure analysis and linkage disequilibrium decay analysis

Using the SNPs, a phylogenetic tree was constructed by the maximum likelihood method using the pipeline Snphylo [49].Structure analysis of the panel was performed with Admixture[50], using the Bayesian algorithm to estimate the number of subpopulations as K = 2–10.The optimal number of subpopulations was determined based on cross validation error value.Principal component analysis(PCA)analysis was performed with GCTA software [51].PopLDdecay [52] was used to calculate r2within the panel, and the degree of LD within the population was estimated following Hill [53].The physical distance of LD block was determined as the distance when its r2value drops to 45% of the maximum r2value, which was approximately 48 kb.

2.5.Genome-wide association study and QTL localization

463,995 high-quality SNPs were selected after filtering with PLINK[54]based on LD block distance to reduce the amount of calculation.Using these SNPs as genotypes and disease lesion length and lesion area as phenotype data, GWAS was performed with EMMAX software [55].The individual kinship coefficient matrix(kinship matrix) was used as a covariate to adjust the association results,and the association threshold was set at 5.Manhattan and quantile–quantile plots were created with the R package CMplot(https://github.com/YinLiLin/R-CMplot).LD blocks containing significantly associated SNPs were defined as QTL regions.Candidate genes were identified by searching within a 48 kb region upstream of the significant SNP with the smallest physical position to a 48 kb region downstream of the significant SNP with the largest physical position in each QTL.

2.6.SNP classification and haplotype analysis

SNPs were classified into five categories by importance [56].Based on the SNP classifications, haplotypes of candidate genes were classified into major groups, each containing at least 10 accessions.ANOVA was fitted to compare lesion length among haplotypes and identify genes showing lesion length differences.

2.7.RNA extraction and transcriptome analysis

Fifteen resistant and 15 susceptible rice varieties were selected by resistance phenotype and grown in a greenhouse for 45 days for inoculation (Table S1).PBS inoculation was used for the control group and bacterial inoculation for the experimental group.One cm of leaf tissue from the inoculation site was collected 12 h after inoculation.Total RNA was extracted with TRIzol (Invitrogen,Carlsbad, CA, USA) and accession samples were mixed equally,resulting in a total of 12 samples (two resistance classes of rice×two types of inoculation×three biological replicates).Sample detection,cDNA library construction,and sequencing were performed by Biomarker Technology Company (Beijing, China) using standard protocols.Clean reads were aligned to the Nipponbare reference genome (version IRGSP-1.0; https://rapdb.dna.affrc.go.jp/download/irgsp1.html)using HISAT[57].The abundance of gene expression levels for each gene in each sample was evaluated as FPKM (Fragments Per Kilobase of transcript per Million fragments mapped).Differential expression analysis was performed with edgeR software [58], with transcripts showing a fold change ≥ 1.5 and P value < 0.05 assigned as differentially expressed.Functional annotation and enrichment analysis of differentially expressed genes were performed using the Kyoto Encyclopedia of Genes and Genomes (KEGG) database (https://www.genome.jp/kegg/genes.html).Transcription factor prediction was performed with the plantTFDB database (https://planttfdb.gaolab.org/index.php).

2.8.qRT-PCR validation

Total RNA was extracted with TRIzol (Invitrogen, Carlsbad, CA,USA) according to the manufacturer’s instructions, and reverse transcription was performed using the Evo M-MLV RT kit with gDNA Clean for qPCR II(Agbio,Changsha,Hunan,China).The cDNA templates were quantified by qRT-PCR using SYBR Green Premix(Agbio) and the ABI StepOne Plus system (Applied Biosystems,Waltham, MA, USA).The primers used are listed in Table S9.The 2-ΔΔCTmethod [59] was used to identify differential mRNA expression, with β-actin (LOC_Os03g50885) [60] as the internal control.

Fig.1.Population structure of 429 diverse rice accessions.(A) Phylogenetic tree of 429 rice accessions.(B)Principal component analysis for the first two components of all accessions.(C)Overall chromosome-wide LD decay estimated from the SNP genotypes of all associations.(D)Subgroups(K=2)inferred using admixture software.The blue,red and black colors represent respectively the indica, japonica and all accessions.

Fig.2.BLS disease scores of 429 rice accessions inoculated with Xoc strains.(A) Frequency distribution of 429 rice accessions based on lesion length.(B) Frequency distribution of 429 rice accessions based on lesion area.(C) Comparison of lesion length between all landraces and all modern cultivars.(D) Comparison of lesion length between landraces and modern cultivars from Guangdong province.Box edges represent the 0.25 and 0.75 quantiles with median values indicated by bold lines.*,P<0.05;**,P < 0.01; ***, P < 0.001 (Student’s t-test).

Fig.3.GWAS analysis for BLS resistance in 429 rice accessions.(A)Manhattan plots of GWAS association statistics based on lesion length.(B)Quantile–quantile(Q–Q)plots of GWAS results based on lesion length.(C) Manhattan plots of GWAS based on lesion area.(D) Quantile–quantile (Q–Q) plots of GWAS based on lesion area.Black arrows represent important QTL.Red arrows represent cloned genes.The black dotted line indicates the significance threshold set at P = 1.0 × 10-5.

2.9.Vector construction and rice transformation

A CRISPR/Cas9 vector targeting OsBLS6.2 was constructed as previously described [61,62].A 20-bp DNA sequence harboring a protospacer adjacent motif within the exon of OsBLS6.2, was combined with a U3-gRNA box.The resulting constructs of U3-OsBLS6.2-gRNA were cleaved with BsaI and inserted into the pRGEB32 vector.The pRGEB32-OsBLS6.2 plasmid was then delivered into the Agrobacterium tumefaciens EHA105 strain for infecting Zhonghua 11 (ZH11) callus.Regeneration of transgenic plants from the callus was achieved on selection medium containing 50 mg L-1hygromycin and 250 mg L-1cefotaxime [63].The primers used are listed in Table S9.The CRISPR-GE database(https://skl.scau.edu.cn/home/) and the RIGW database (https://rice.hzau.edu.cn/rice_rs1/) were used to predict potential offtarget sites by BLAST analysis of the rice genome using our designed single guide RNA (gRNA) (20 bp).This search identified 15 genes as potential off-target sites.Examination revealed that Cas9-gRNAOsBLS6.2mismatch splice sites were not present in the predicted sequences of off-target genes in the OsBLS6.2-ko lines(Table S10).

3.Results

3.1.Population structure analysis of rice accessions and their responses to BLS

Fig.4.DEGs in resistant and susceptible rice accessions 12 h after inoculation with Xoc and PBS.(A)The distribution of 3250 differentially expressed genes among the four combinations.BLS_R_up: Gene expression levels after inoculation with Xoc in resistant accessions were higher than those after inoculation PBS.BLS_R_down: Gene expression levels after inoculation with Xoc in resistant accessions were lower than those after inoculation with PBS.BLS_S_up:Gene expression levels after inoculation with Xoc in susceptible accessions were higher than those after inoculation with PBS.BLS_S_down:Gene expression levels after inoculation with Xoc in susceptible accessions were lower than those after inoculation with PBS.(B) Statistics for transcription factor prediction of the DEGs.(C) KEGG analysis of DEGs.

After DNA sequencing, the quality statistics Q20 > 96% and Q30 > 91% showed satisfactory read quality.The genotyping rate(The mean proportion of available SNPs for individuals.) of the SNPs was 99.1%,and they were distributed throughout the genome(Fig.5A) at a mean density of 0.84 kb/SNP (Table S11).The phylogenetic tree assigned the 429 rice accessions mainly to the japonica and indica subgroups (Fig.1A).Among them,there were 40 japonica accessions mainly from Japan,and 389 indica accessions mainly from China and the Philippines (Table S1).Based on PCA and population structure analysis(Fig.1 B,D),the 429 rice accessions were divided into the indica and japonica subgroups.As shown in Fig.1C,when r2decreases to 45%of its maximum value,the genome-wide LD decay distance of all accessions is about 48 kb,the LD decay distance of the japonica subpopulation is about 281.9 kb, and the LD decay distance of the indica subpopulation is about 32.6 kb.The genetic diversity of indica subgroups was high.

The frequency distributions of trimmed mean values of lesion length and area are shown in Fig.2.Fourteen accessions were highly resistant, 149 were resistant, 116 were moderately resistant, 96 were susceptible, and 54 were highly susceptible.Among the 150 susceptible and highly susceptible accessions, 60% were from Guangdong, indicating that GDXC-1 showed stronger virulence on accessions from Guangdong than on those from other localities.The mean lesion length of landraces was significantly greater than that of cultivars(Fig.2C),and the difference was more marked in cultivars originating in Guangdong (Fig.2D).

Fig.5.Candidate genes screening by integration of GWAS and RNA-seq.TMLBLS:Genes identified by GWAS based on lesion length.TMABLS: Genes identified by GWAS based on lesion area.DEGs: Differentially expressed genes identified by transcriptome.

3.2.Mapping of QTL for bacterial leaf streak resistance by GWAS

As shown in Fig.3,252 SNPs(Table S2)were identified as associated with resistance.A total of 47 QTL for BLS resistance were detected between the significant SNPs(Table S3),on chromosomes 1,5,6,7,8,9,and 10.Among them,7 QTL were repeatedly mapped in the identification of the two phenotypic data.The intervals containing qBLS6.2 and qBLS6.13 showed the strongest association with the two traits.Based on the number of SNPs and the colocalization results within the intervals, we identified qBLS6.1,qBLS6.2, qBLS6.12, qBLS6.13, qBLS7.3, and qBLS10.2 as important QTL (ImQTL) for BLS resistance.Comparison with the 23 reported BLS resistance genes collected by China Rice Data Center (https://www.ricedata.cn/) (Table S4) showed that OsBGLU19 and OsRPIL1 were located within our QTL region.The Q–Q plot showed that the model effectively controlled the effects of population structure and kinship on association.GWAS was performed using three further models: General Linear Model (GLM), Mixed Linear Model(MLM), and Enriched Composite Maximum Likelihood Model(ECMLM).All three GWAS models identified qBLS6.2 (Fig.S5A).GWAS utsing the maximum and minimum values of lesion length also identified qBLS6.2 (Fig.S5B, C).

3.3.Comparison of transcriptomes of resistant and susceptible lines of rice

Transcriptome sequencing generated 269,012,852 clean reads,with mapped read proportions ranging from 93.3% to 95.2%.DEGs were identified using the standard of P-value ≤ 0.05 and log2FC ≥1.5, with PBS-treated samples as the control group and GDXC-1-inoculated samples as the experimental group.Of 3250 DEGs(Table S5),2098(1625 upregulated and 473 downregulated)were found in resistant accessions and 2289 (1861 upregulated and 428 downregulated) in susceptible accessions.As shown in Fig.4A,1137 DEGs were shared by resistant and susceptible accessions, of which 21 DEGs showed opposite expression trends and 1116 DEGs similar expression trends.The enriched pathways of the DEGs were associated with plant–pathogen interactions,including plant hormone signal transduction, amino acid biosynthesis, carbon metabolism, and phenylpropanoid biosynthesis(Fig.4C).Transcription factor prediction of the DEGs (Fig.4B)revealed the rice transcription factor families NAC, MYB, and WRKY.

Table 1 Six genes that can be identified by ImQTL and transcriptomes.

3.4.Identification of BLS resistance genes

GWAS was performed using lesion length and lesion area separately as phenotype data.Two gene sets, TMLBLS and TMABLS,were obtained using the gene search method described above.Another gene set (Table S6) was obtained by transcriptome analysis.Genes that were identified through both GWAS and transcriptome analysis were considered candidate genes.As shown in Fig.5,of 35 candidate genes, 12 were present in both the TMLBLS and DGE sets, 17 were found in both the TMABLS and DGE sets, and six genes were shared among all three sets, including two cloned genes for resistance to bacterial disease, OsMYB30 and OsRP1L1.OsMYB30 was identified by TMLBLS,TMABLS,and RNA sequencing(RNA-seq), and OsRP1L1 were identified by TMLBLS and RNA-Seq.Two other disease resistance genes(OsACO7 and OsMESL)were also identified.To verify the accuracy of the RNA-seq results,nine genes were selected for qRT-PCR analysis to compare their relative expression levels with RNA-seq results.The expression patterns of these genes after inoculation with GDXC-1 were consistent with the RNA-seq results (Fig.S1).Among these 35 DEGs, six were located in the ImQTL of GWAS,suggesting that they are associated with BLS resistance (Table 1).

The haplotypes based on SNP classification are shown in Table S7.Only those of OsBLS6.2 and LOC_Os07g11110 showed phenotypic variation(Fig.6).Five significant SNPs in OsBLS6.2 resulted in nonsynonymous mutations.The accessions carrying these five SNPs were roughly divided into two haplotypes, and accessions carrying haplotype 1 had significantly shorter lesion lengths than those carrying haplotype 2.LOC_Os07g11110 could be divided into two haplotypes using six SNPs, and accessions carrying haplotype 1 had significantly shorter lesion lengths than those carrying haplotype 2.These results further suggest that OsBLS6.2 and LOC_Os07g11110 are associated with resistance to Xoc.

3.5.Functional validation of OsBLS6.2

Of 47 QTL, qBLS6.2, located on chromosome 6, showed the strongest signal and the most significant SNPs.Using these 121 SNPs, an LD block was constructed (Fig.7A) that divided qBLS6.2 into two regions.The lead SNP of qBLS6.2 (Chr.6_1093215) is only 590 bp from OsBLS6.2,and there were nine significant SNPs within OsBLS6.2,further suggesting that this gene was a promising candidate for BLS resistance.To confirm the function of OsBLS6.2 in BLS resistance, gene-edited mutants of OsBLS6.2 in ZH11 variety were generated and two independent T2generation homozygous lines(ko-1 and ko-2) were selected for further evaluation (Fig.7D, E).They did not display the Cas9-induced off-target effect(Table S10).Both lines(ko-1 and ko-2)showed markedly increased resistance to Xoc (Fig.7B, C), suggesting that OsBLS6.2 is a promising candidate in the regulation of BLS resistance through an unclear mechanism.

Fig.6.Haplotype analysis of candidate genes.(A) Haplotype analysis of OsBLS6.2.(B) Haplotype analysis of LOC_Os07g11110.Rectangles and lines represent exons and introns, respectively and the coding sequence highlighted in blue.White rectangles and white arrows represent 5′ UTR and 3′ UTR.Box edges represent the 0.25 and 0.75 quantiles with median values indicated by bold lines.The square in the box represents the mean.*, P < 0.05; **, P < 0.01; ***, P < 0.001 (Student’s t-test).

Fig.7.Genomic region of qBLS6.2 and functional verification of candidate gene OsBLS6.2.(A) Manhattan plot (top) and linkage disequilibrium heat map (bottom) of the qBLS6.2 region, with the red arrow indicating the lead SNP position.(B) The phenotypes of wild-type ZH11 and OsBLS6.2 knockout plants after Xoc inoculation.Values are means ± SDs of three biological replicates.(C) Lesion length and lesion area in wild-type ZH11 and transgenic plants after Xoc inoculation.The left side is the length of the lesion, the right side is the area of the lesion, and the middle is separated by a black dotted line.(D-E) Homozygous gene editing line of OsBLS6.2.

4.Discussion

We evaluated 429 rice accessions to determine their resistance levels to the Xoc strain GDXC-1.The finding that breeding accessions showed higher resistance than landraces suggests that breeders have incorporated some BLS resistance into cultivars.The finding that GDXC-1 appears to be more detrimental to local species in Guangdong suggests that that over time, the strain has adapted to the selection pressure of the local host and become the dominant physiological race in the region.

4.1.Novel resistance loci and DEGs provide information for studying rice resistance to Xoc

The SNP Chr.05_1981633 was located in qXO-5–2 previously mapped by Bossa-Castro et al.[11].The distance between qBLS7.3 and the candidate interval qBB-7–1 was only 90.6 kb.LOC_Os07g11110 may be a candidate gene for qBLS7.3.Our findings that Chr.02_24909501 was only 29 kb from the cloned bacterial blight resistance gene OsMYB30, and that two reported BLS resistance-associated genes, OsBGLU19 and OsRPIL1, were located near Chr.05_17539510 [15,18], further support the reliability of our GWAS associations.Lists of the QTL,genes,and significant SNPs identified by GWAS presented in Tables S2, S3, and S6 may aid future efforts to identify BLS resistance genes in rice.

In our transcriptome analysis, we identified 3250 DEGs,LOC_Os02g01355 and LOC_Os12g03040 were found to be consistently upregulated in resistant accessions and downregulated in susceptible ones.LOC_Os12g03040 is a NAC transcription factor gene that has been reported to be induced by Magnaporthe oryzae infection[64],while LOC_Os02g01355 is a MADS-box transcription factor.In Arabidopsis, transcription factors of the MADS-box gene family function in regulating development[65].In rice,OsMADS26,another MADS-box transcription factor,serves as an upstream regulator of stress-responsive genes,and impairs the regulation of rice resistance to both M.oryzae and bacterial blight [66].Among all transcription factors in rice,OsWRKY genes show promising effects against diseases from seedling to adult stage [67].Among the 19 WRKY family transcription factors (Table S8), OsWRKY45,

OsWRKY62, OsWRKY72, OsWRKY53, and OsWRKY28 have all been reported [68–72] to function in rice defense responses.With the finding that OsWRKY7,OsWRKY24,OsWRKY67,and OsWRKY69 were all induced to expression after rice was inoculated with bacterial blight(Fig.S2),this finding suggests that these WRKY family transcription factors function in rice defense response to bacterial diseases.

4.2.Potential value of OsBLS6.2 in rice breeding

All five SNPs analyzed for haplotypes were heterozygous.Analysis of the Rice Functional Genomics and Breeding(RFGB)database(https://www.rmbreeding.cn/) revealed that these five sites were divided mainly into two haplotypes [73].Haplotype 2 also had heterozygous sites.Gene-edited plants were shorter than wildtype plants (Fig.S3).OsBLS6.2 is expressed mainly in rice leaves,leaf sheaths, and stems (Fig.S4).Removing part of the top leaves of the rice plant led to downregulation of OsBLS6.2 expression[74].OsBLS6.2 expression was lower in samples inoculated with GDXC-1 than in PBS-treated samples.Gene-edited OsBLS6.2 transgenic plants showed higher resistance to BLS.Under inoculation of ZH11 and gene-edited mutants with three Xoc strains (GX01,GX1901,and GX2001,an epidemic physiological race revealed that mutants were resistant to GX2001 and GX1901, but not to GX01(Fig.S6).The bls1 (OsMAPK6) is a QTL for resistance to BLS [10]and its mutant showe dwarfing compared with the wild type[75].Genotyping OsMAPK6 in ZH11 indicate that it is bls1 rather than BLS1 (Fig.S7A), but there was no significant difference in the relative expression of bls1 between the wild-type and mutant after inoculation with GDXC-1 (Fig.S7B).These findings suggest that OsBLS6.2-mediated resistance is independent of bls1.We hypothesize that OsBLS6.2 functions in responding to external stress and regulating plant height, making it a promising target for molecular breeding to increase rice resistance to BLS.CRediT authorship contribution statement

Huabin Xie:Writing–original draft,Conceptualization,Visualization.Chunzi Lin:Conceptualization,Formal analysis,Data curation.Wenyu Lu:Formal analysis, Data curation.Zhikai Han:Formal analysis, Data curation.Danhong Wei:Formal analysis,Data curation.Xing Huo:Investigation.Tianjiao Li:Investigation.Jian Zhang:Investigation.Yongqiang He:Investigation, Writing– review & editing.Chun Chen:Investigation.Hui Wang:Investigation.Tao Guo:Conceptualization, Investigation, Writing – original draft, Writing – review & editing, Funding acquisition.Jiafeng Wang:Conceptualization, Investigation, Writing – original draft,Writing – review & editing, Funding acquisition.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

This work was supported by the Open Project(2020)of Guangdong Key Laboratory of New Technology in Rice Breeding,the Natural Science Foundation of Guangdong Province, China(2019A1515011825) and the Special Rural Revitalization Funds of Guangdong Province (Seed Industry Revitalization Project) (2022-NPY-00-006).

Appendix A.Supplementary data

Supplementary data for this article can be found online at https://doi.org/10.1016/j.cj.2023.08.007.