Increasing resistance to bacterial leaf streak in rice by editing the promoter of susceptibility gene OsSULRT3;6.

Rice (Oryza sativa L.) is an important cereal crop consumed by almost half the world population and is vital for global food security. Bacterial leaf streak (BLS), which is caused by Xanthomonas oryzae pv. oryzicola (Xoc), is a devastating rice disease in Asia, Africa and Australia (Ji et al., 2014; Nino‐Liu et al., 2006). Like other Xanthomonads, Xoc utilizes the type III secretion system (T3SS) to translocate effector proteins directly into host cells to suppress plant immunity (Nino‐Liu et al., 2006; Yuan et al., 2021). Transcription activator‐like effectors (TALEs) act as virulence or avirulence factors and function as eukaryotic transcription factors in plant cell nuclei, where they bind to effector‐binding elements (EBEs) of targeted plant gene promoters via a TALE‐encoded central repeat region (CRR). The CRR consists of highly conserved tandem repeats of 34‐amino acids; the hypervariable 12th and 13th residues in each repeat determine nucleotide binding specificity and are referred to as repeat‐variable diresidues (RVDs) (Boch et al., 2009; Moscou and Bogdanove, 2009). Identification of TALE‐targeted genes in plants is used to facilitate plant disease resistance breeding programmes for X. oryzae pv. oryzae (Xoo), which is closely related to Xoc (Eom et al., 2019; Ji et al., 2016; Oliva et al., 2019; Xu et al., 2019).

Relatively few TALE‐targeted genes have been identified in the interaction of rice with Xoc. One example is OsSULTR3;6, which is up‐regulated by Tal2g in Xoc strain BLS256. OsSULTR3;6 encodes a predicted sulphate transporter in rice and is considered a susceptibility (S) gene for BLS (Cernadas et al., 2014). In the Xoo‐rice pathosystem, the disruption of TALE‐binding elements in three S genes confers broad‐spectrum resistance to Xoo (Oliva et al., 2019; Xu et al., 2019); therefore, we reasoned that a possible strategy for increasing rice resistance to Xoc might be the loss of susceptibility (RLS) genes. We speculated that rice would gain RLS if the EBE sequence of OsSULTR3;6, recognized by Tal2g, is edited by CRISPR/Cas9 technology (Zhou et al., 2014). In the present study, CRISPR/Cas9 was to disrupt the Tal2g‐recognized EBE of OsSULTR3;6 in rice cultivar IRBB10, which is susceptible to Xoc.

The TALE gene tal5d of Xoc strain RS105 was previously characterized (Ji et al., 2014; Wilkins et al., 2015). Tal5d contains 17.5 central repeat units that are nearly identical to Tal2g in Xoc strain BLS256; the obvious difference is that the 10th RVD of Tal5d is ND rather than HD in Tal2g (Figure 1a). We speculated that Tal5d might bind to the same EBE as Tal2g (EBETal2g) based on prediction of the two TALE‐binding sites (Figure 1a). Furthermore, no other variants of Tal2g and Tal5d were identified in the available sequences of Xoc (Ji et al., 2014; Wilkins et al., 2015).

Figure 1

Generation of resistance to bacterial leaf streak in rice using CRISPR/Cas9. (a) RVD sequence alignment of Tal5d and Tal2g in RS105 and BLS256, respectively; the EBETal2g/Tal5d consensus sequence is shown, and differences in the 10th RVD are indicated in red font. (b) Xoc Tal5d binds the EBE recognized by Tal2g (EBETal2g) in the OsSULTR3;6 promoter. The uppermost sequence shows probe derived from OsSULTR3;6 promoter, and the EBETal2g is shown by blue letters. Unlabelled pEBETal2g was used in competition assays. The presence and absence of DNA and proteins are indicated by (+) and (−), respectively. (c) Schematic diagram of pCas9‐gRNA4‐SU. RB, right border; LB, left border; and HPTII, hygromycin phosphotransferase II. The sequence shown below the map is for sgRNA. (d) Functional map of the OsSULTR3;6 promoter in IRBB10 and analysis of edited rice lines. The uppermost diagram shows the OsSULTR3;6 promoter with the EBETal2g/Tal5d locus in blue font. Gene‐specific target sequences of sgRNA and PAM are underlined and boxed. The lower panel shows sequence analyses of homozygous lines SU‐1 to SU‐5. The red hyphen indicates a single nucleotide deletion. Pink letters indicate inserted nucleotides. (e) Disease phenotypes in rice IRBB10 and gene‐edited lines SU‐1 to SU‐5 (T2 generation) inoculated with Xoc strains BLS256 and RS105. (f) Disease lesion lengths on IRBB10, SU‐1, SU‐2, SU‐3, SU‐4 and SU‐5 at 14 dpi. Column height shows mean lesion length (cm), and error bars indicate the standard deviation of six replicates. Data are means ± SD of six biological replicates (Student's t‐test at P < 0.05). (g) Os01g52130 expression in rice lines IRBB10, SU‐1 to SU‐5 infiltrated with Xoc strain BLS256. The expression of Os01g52130 was evaluated by qRT‐PCR at 24 hpi. Data are means ± SD of three biological replicates (Student's t‐test at P < 0.05). (h) Bacterial growth of Xoc strains BLS256 and RS105 in leaves of rice line IRBB10, SU‐1 and SU‐4. Data points represent the mean ± SD from three replicates. All the experiments were repeated three times, and similar results were obtained.

To test this hypothesis, plasmid pET30a‐tal5d was constructed and used for purifying His‐Tal5d. Electromobility shift assays (EMSA) showed that the His‐Tal5d bound to the Cy5‐labelled pEBEtal2g fragment, and binding was reduced by adding unlabelled pEBEtal2g (Figure 1b). These results demonstrated that Tal5d binds the OsSULTR3;6 promoter at the EBEtal2g locus, which was then designated EBETal2g/Tal5d (Figure 1a).

In order to disrupt Tal2g and Tal5d binding, we designed a sgRNA targeting the OsSULTR3;6 promoter near EBETal2g/Tal5d and constructed binary vector pCas9‐gRNA4‐SU (Zhou et al., 2014) to edit the EBETal2g/Tal5d sequence (Figure 1c,d). Five homozygous rice lines of IRBB10 (T1 generation) were obtained and named SU‐1 to SU‐5 (Figure 1d). PCR amplification and sequencing of the EBETal2g/Tal5d region showed the following changes relative to the wild‐type EBE: SU‐1 contained a 14‐bp deletion and 12‐bp insertion; SU‐2 contained a 1‐bp insertion (cytosine); SU‐3 was missing a single nucleotide (thymine deletion); and SU‐4 and SU‐5 had 17‐ and 3‐bp deletions, respectively (Figure 1d).

The wild‐type IRBB10 and five edited lines (SU‐1 to SU‐5) were inoculated by pin‐pricking method (Pan et al., 2018) with Xoc strains BLS256 and RS105, respectively, to investigate potential resistance. At 14 days post‐inoculation (dpi), the lesions induced by BLS256 and RS105 on the five edited rice lines (SU‐1 to SU‐5) were significantly smaller than those on IRBB10 (Figure 1e, 1f). It should be noted that SU‐2 line showed less resistance significantly than IRBB10, but the lesion length formed was obviously longer than those in SU‐1, SU‐3, SU‐4 and SU‐5 lines (Figure 1e,f). Sequence analysis showed that the single nucleotide insertion in SU‐2 occurred at the terminal nucleotide of the EBE (Figure 1d), suggesting that the modification may have less effect on TALE binding than the changes in the other four edited lines. The expression levels of Os01g52130 in SU‐1 to SU‐5, infiltrated with BLS256, were significantly lower than those in IRBB10 (Figure 1g), suggesting that the edited EBE loci disrupt the activation of Os01g52130 by Tal2g of BLS256 strain. For this, Xoc strains BLS256 and RS105, containing tal2g and tal5d, respectively, were then inoculated to IRBB10, SU‐1 and SU‐4 and bacterial growth was measured. Growth of Xoc BLS256 and RS105 was remarkably reduced in rice lines SU‐1 and SU‐4, respectively, as compared to the wild‐type IRBB10 (Figure 1h). These results indicated that the homozygous mutations in the EBETal2g/Tal5d locus disarmed the recognition of Tal2g and Tal5d in rice nuclei, resulting in resistance to Xoc infection.

In summary, we first generated mutations in the EBE of the OsSULTR3;6 promoter in IRBB10 and created new germplasm that exhibits resistance to Xoc strains containing virulence factors either Tal2g or Tal5d. Our findings show that genetic modification of the EBETal2g/Tal5d sequence via CRISPR/Cas9 technology may be used to develop rice lines with broad‐spectrum resistance to bacterial leaf streak in rice by disrupting the EBEs of TALE‐matched S genes in rice.

Conflict of interest

The authors declare no conflicts of interest.

Author contributions

X.X. and G.C. designed the experiments. X.X., Z.X., Z.L. and M.Z. performed the experiments. X.X. and Z.X. wrote the manuscript. L.Z. and G.C. revised the manuscript.