Induction of Xa10-like Genes in Rice Cultivar Nipponbare Confers Disease Resistance to Rice Bacterial Blight

Bacterial blight of rice is one of the most destructive bacterial diseases throughout the major rice-growing regions in the world. The causal agent of bacterial blight of rice is Xanthomonas oryzae pv. oryzae. The X. oryzae pv. oryzae strains enter rice leaves, typically through hydathodes at leaf tips and margins or wounds, and multiply in leaf veins and xylem, causing blockage and plant wilting (Niño-Liu et al. 2006). During infection, the X. oryzae pv. oryzae strains deliver AvrBs3/PthA-like effectors into host cells via a bacterial type III secretion system (Yang and White 2004). The AvrBs3/PthA-like effectors, also referred to as transcription activator-like (TAL) effectors (Yang et al. 2006), function as transcription factors that enhance the virulence of plant-pathogenic Xanthomonas spp. through the induction of host susceptibility (S) genes (Antony et al. 2010; Cernadas et al. 2014; Hu et al. 2014; Kay et al. 2007; Streubel et al. 2013; Sugio et al. 2007; Yang et al. 2006; Yu et al. 2011; Zhou et al. 2015). The specificity of TAL effector-mediated host gene induction is determined by the repetitive central region of each TAL effector, which consists of near-perfect direct repeats of 33- to 35-amino-acid (aa) residues. The two hypervariable amino acids at positions 12 and 13 of each repeat, termed repeat-variable diresidue (RVD), preferentially associate with a different nucleotide to define the length and sequence of effector binding elements (EBE) in the promoters of the targeted host genes (Boch et al. 2009; Moscou and Bogdanove 2009). A few TAL effector-dependent S genes have been identified in rice for susceptibility to X. oryzae pv. oryzae and, out of these, the most important and well-studied S genes belong to the MtN3 or SWEET family (Antony et al. 2010; Yang et al. 2006; Yu et al. 2011; Zhou et al. 2015). The SWEET genes encode sugar transporters with seven transmembrane helices and can divert nutrition from the host to facilitate bacterial growth in apoplast of plant cells (Chen et al. 2010).

Plants have coevolved recessive disease resistance (R) genes to overcome this TAL effector-dependent susceptibility. For instance, the recessive mutations in the EBE of SWEET genes in rice abolish the specific recognition between TAL effectors and the cognate SWEET genes and, therefore, cause the failure of X. oryzae pv. oryzae to induce SWEET gene expression for pathogen propagation (Chu et al. 2006; Hutin et al. 2015; Zhou et al. 2015). In another example, the recessive allele of the general transcription factor OsTFIIAγ5 gene in rice, xa5, provides disease resistance to incompatible X. oryzae pv. oryzae strains through the attenuation of TAL effector-dependent SWEET gene induction (Huang et al. 2016; Iyer and McCouch 2004). Plants have also coevolved dominant R genes to adapt TAL effector activity to counter bacterial infection (Zhang et al. 2015). The gene products of TAL effector-dependent dominant R genes, referred to as “executor” R genes (Bogdanove et al. 2010), trigger cell-death-associated hypersensitive response (HR) and activate a defense response in plants (Gu et al. 2005; Römer et al. 2007; Strauss et al. 2012; Tian et al. 2014; Wang et al. 2015). Thus far, three TAL effector-dependent dominant R genes have been cloned in rice (Gu et al. 2005; Tian et al. 2014; Wang et al. 2015). Xa27 is the first cloned executor R gene in plants that provides broad-spectrum disease resistance to X. oryzae pv. oryzae strains (Gu et al. 2004, 2005). The XA27 protein is an apoplastic protein that depends on its N-terminal signal-anchor-like sequence to localize to the apoplast (Wu et al. 2008). The susceptible allele of Xa27 (xa27) in ‘IR24’ rice encodes a protein identical to XA27 (Gu et al. 2005). The induction of xa27 by X. oryzae pv. oryzae strains that deliver the artificially designed TAL effector (dTALE) dTALE-xa27, which targets a DNA element in the xa27 promoter, triggered resistance response in IR24 (Li et al. 2013). Following a similar approach, four dTALE genes were designed to induce the four annotated Xa27-like genes in ‘Nipponbare’ rice but none of the four induced Xa27-like genes conferred resistance to the dTALE-containing X. oryzae pv. oryzae strains (Li et al. 2013). The Xa10 gene in rice confers race-specific but narrow-spectrum resistance to X. oryzae pv. oryzae strains containing TAL effector gene avrXa10 (Hopkins et al. 1992; Yoshimura et al. 1983). The XA10 protein locates to the endoplasmic reticulum (ER) membrane, induces ER Ca²⁺ depletion, and triggers cell death in rice, Nicotiana benthamiana, and human HeLa cells (Tian et al. 2014). The Xa23 gene is the third executor R gene in rice, which was recently cloned from ‘CBB23’ (Wang et al. 2015). Xa23 was derived from wild rice (Oryza rufipogon) and confers broad-spectrum resistance to X. oryzae pv. oryzae strains (Wang et al. 2014b; Zhang et al. 2001). Xa23 encodes a 113-aa protein, which shares 50% identity with XA10 (Wang et al. 2015). By comparing the genetic loci of Xa10 and Xa23 in the Nipponbare genome, the two R genes were physically mapped to the same locus or adjacent loci on rice chromosome 11 (Gu et al. 2008; Wang et al. 2014a), indicating that the two TAL-dependent R genes may be allelic to each other.

Nipponbare is a typical japonica cultivar that is susceptible to many X. oryzae pv. oryzae strains and no bacterial blight R gene has been identified in this cultivar. Two Xa10-like genes, Os11g37570 and Os11g37620, were annotated at the corresponding Xa10 locus in the Nipponbare genome. The two genes are 22,177 bp from each other and have the same transcription direction as Xa10, which points toward the centromere (Tian et al. 2014). They encode hypothetical proteins ABA94452 and ABA94457, respectively (Tian et al. 2014). Recently, an alternative open reading frame (ORF) within the predicted ORF of Os11g37620 was found to be the ORF of the susceptible allele of the Xa23 gene encoding a protein identical to XA23 (Wang et al. 2015). In this article, we designated Os11g37570 as Xa10-Ni and the susceptible allele of the Xa23 gene in Nipponbare as Xa23-Ni and characterized their function by using approaches of synthetic biology, forward and reverse genetics, plant pathology, as well as cellular biology.

RESULTS

Induction of Xa10-Ni or Xa23-Ni in Nipponbare confers disease resistance to X. oryzae pv. oryzae strains.

dTALE were specifically designed and constructed to induce Xa10-Ni or Xa23-Ni expression through bacterial inoculation. dTALE-Xa10-Ni targets a DNA element (EBE_{dTALE-Xa10-Ni}) in the Xa10-Ni promoter (Fig. 1A; Supplementary Fig. S1A; Supplementary Sequence S1). Disease evaluation showed that Nipponbare was resistant to X. oryzae pv. oryzae strain PXO99^A(pHM1dTALE-Xa10-Ni) but was susceptible to the control X. oryzae pv. oryzae strain PXO99^A(pHM1) (Fig. 1B; Supplementary Table S1). Quantitative reverse-transcription polymerase chain reaction (qRT-PCR) analysis also showed that Xa10-Ni was specifically induced by PXO99^A(pHM1dTALE-Xa10-Ni) but not by PXO99^A(pHM1) (Fig. 1C). The two experiments demonstrated that Xa10-Ni was specifically induced by dTALE-Xa10-Ni and the expression of Xa10-Ni activated disease resistance to PXO99^A(pHM1dTALE-Xa10-Ni). A 517-bp cDNA of Xa10-Ni was isolated using 5′ and 3′ rapid amplification of cDNA ends (RACE) and RT-PCR. The Xa10-Ni cDNA contains a 12-bp 5′ untranslated region (UTR), a 405-bp ORF, and a 111-bp 3′ UTR. The ORF of Xa10-Ni cDNA encodes an XA10-like protein identical to ABA94452 containing 134-aa residues (XA10-Ni).

Download as PowerPoint

The TAL effector-dependent gene products were previously found to induce cell death in N. benthamiana (Römer et al. 2007; Tian et al. 2014). However, our initial experiment on transient expression of ABA94457 in N. benthamiana failed to induce cell death (data not shown). Compared with XA10, ABA94457 has an extended N-terminal region, which is presumably encoded by the predicted exon 1 and part of predicted exon 2 of Os11g37620 (Tian et al. 2014). We suspected that this extended N-terminal region in ABA94457 might not be present in the gene product of the Xa10-like gene. We then designed a dTALE, dTALE-Xa23-Ni-1, by targeting a DNA element (EBE_{dTALE-Xa23-Ni-1}) in the predicted intron 1 of Os11g37620 (Fig. 1A; Supplementary Sequence S2). Disease evaluation demonstrated that Nipponbare was resistant to PXO99^A(pHM1dTALE-Xa23-Ni-1) but not to PXO99^A(pHM1) (Fig. 1B). The results suggested that a functional R gene was induced by dTALE-Xa23-Ni-1, which triggered disease resistance to PXO99^A(pHM1dTALE-Xa23-Ni-1). Three independent cDNA (cDNA1, cDNA2, and cDNA3) were isolated by 5′ and 3′ RACE and RT-PCR. After comparing DNA sequences of the three cDNA clones with the genomic sequence at the downstream of EBE_{dTALE-Xa23-Ni-1}, cDNA1 was found to be derived from the Xa23-Ni gene without any intron splicing, whereas cDNA2 and cDNA3 were generated from alternative splicing in the 3′UTR or coding region of the Xa23-Ni gene. A common ORF was identified in cDNA1 and cDNA2, which encodes a 113-aa protein that is identical to XA23 (Wang et al. 2015). For consistency, we designated this 113-aa protein as XA23-Ni in this study. The 5′UTR of cDNA1 of Xa23-Ni is 237 bp in length and the transcription initiation site is at 85 bp downstream of the EBE_{dTALE-Xa23-Ni-1}. To confirm the results, we designed a second dTALE, dTALE-Xa23-Ni-2, by targeting another DNA element (EBE_{dTALE-Xa23-Ni-2}) in the Xa23-Ni promoter (Fig. 1A). Disease evaluation demonstrated that Nipponbare was resistant to PXO99^A(pHM1-Xa23-Ni-2) (Fig. 1B). qRT-PCR analysis revealed that Xa23-Ni was specifically induced by PXO99^A(pHM1dTALE-Xa23-Ni-1) and PXO99^A(pHM1dTALE-Xa23-Ni-2) but not by PXO99^A(pHM1) (Fig. 1D and E).

Pairwise sequence alignment analyses were conducted between the promoters and ORF of the Xa10, Xa10-Ni, and Xa23-Ni genes. The identity in the 500-bp promoters analyzed was 41.1% between Xa10 and Xa10-Ni promoters, 44.1% between Xa10 and Xa23-Ni, and 46.4% between Xa10-Ni and Xa23-Ni, whereas the identity in the ORF of the three genes was 59.2% between Xa10 and Xa10-Ni ORF, 66.5% between Xa10 and Xa23-Ni ORF, and 53.3% between Xa10-Ni and Xa23-Ni ORF (Supplementary Figs. S2 to S7). The statistics data indicated that the promoters of Xa10, Xa10-Ni, and Xa23-Ni are less conserved than their ORF. A similar phenomenon was observed in the previous study, in which the susceptible Xa23 allele in Nipponbare (Xa23-Ni) and the resistant Xa23 allele in CBB23 shared an identical ORF but showed significant difference in the promoter region (Wang et al. 2015). Functional EBE_AvrXa10 and EBE_AvrXa23 were identified in Xa10 and Xa23 promoters, respectively, which are specifically recognized by AvrXa10 and AvrXa23 (Tian et al. 2014; Wang et al. 2015); however, the two EBE or their consensus sequences were not found in either the Xa10-Ni or Xa23-Ni promoter (Wang et al. 2015).

Transgenic rice plants carrying coding regions of Xa10-Ni or Xa23-Ni under the Xa10 promoter are resistant to X. oryzae pv. oryzae strain delivering TAL effector AvrXa10.

TAL effector AvrXa10 binds to EBE_AvrXa10 in the Xa10 promoter and induces Xa10 expression, which results in race-specific disease resistance to X. oryzae pv. oryzae strains that deliver AvrXa10 (Tian et al. 2014). To further confirm the function of Xa10-Ni or Xa23-Ni in disease resistance, the coding region of Xa10 in its genomic clone SA4671 (Tian et al. 2014) was replaced with the ORF in cDNA of Xa10-Ni or the ORF in cDNA1 of Xa23-Ni to generate transcriptional fusion genes P_Xa10:Xa10-Ni:T_Xa10 or P_Xa10:Xa23-Ni:T_Xa10. The fusion genes were used to produce transgenic plants in Nipponbare genetic background. After bacterial blight inoculation, 18 of 47 transgenic T0 plants of P_Xa10:Xa10-Ni:T_Xa10 and 3 of 19 transgenic T0 plants of P_Xa10:Xa23-Ni:T_Xa10 were resistant to PXO99^A(pHM1avrXa10) (Supplementary Table S2). In addition, the specific resistance of P_Xa10:Xa10-Ni:T_Xa10 or P_Xa10:Xa23-Ni:T_Xa10 to PXO99^A(pHM1avrXa10) but not to PXO99^A(pHM1) was detected in the T1 generation (Fig. 2). These results demonstrated that both P_Xa10:Xa10-Ni:T_Xa10 and P_Xa10:Xa23-Ni:T_Xa10 are functional in conferring disease resistance to PXO99^A(pHM1avrXa10). In another experiment, the Xa10 coding region was replaced with the ORF in cDNA3 of Xa23-Ni (ORF2), which encodes a hypothetical protein with 70-aa residues, to generate P_Xa10:Xa23-Ni-ORF2:T_Xa10 gene. Seventy-five T0 plants of P_Xa10:Xa23-Ni-ORF2:T_Xa10 were produced; however, none of transgenic T0 plants was resistant to PXO99^A(pHM1avrXa10) (Supplementary Fig. S8). The result indicated that the 70-aa protein encoded by the ORF2 in cDNA3 of Xa23-Ni does not provide disease resistance to X. oryzae pv. oryzae.

Download as PowerPoint

Xa10-Ni or Xa23-Ni knock-out mutants abolishes disease resistance to X. oryzae pv. oryzae

The knock-out mutants of Xa10-Ni or Xa23-Ni were generated using transcription activator-like effector nuclease (TALEN) or clustered regularly interspaced short palindromic repeats (CRISPR)-associated protein-9 nuclease (CRISPR/Cas9) technology (Li et al. 2012; Ma et al. 2015). Four independent knock-out mutants of Xa10-Ni were generated using TALEN technology (Fig. 3A). They had a 4- to 14-bp deletion in the interspace region between two TALEN target sites in the coding region of Xa10-Ni, which resulted in frame-shift mutation (Fig. 3A). Three independent knock-out mutants of Xa23-Ni were produced using CRISPR/Cas9 technology (Fig. 3A). They had either 1-bp insertion or 17-bp deletion in the gRNA target site (Fig. 3A). Disease evaluation demonstrated that the Xa10-Ni knock-out mutants were susceptible to PXO99^A(pHM1dTALE-Xa10-Ni) and the Xa23-Ni knock-out mutants were susceptible to PXO99^A(pHM1dTALE-Xa23-Ni-2) (Fig. 3B; Supplementary Table S3). qRT-PCR analysis indicated that the mutated Xa10-Ni or Xa23-Ni gene was still specifically induced by PXO99^A that delivered the matching dTALE (Fig. 3C and D). Nine Xa10-Ni and Xa23-Ni double knock-out mutants were generated by genetic crossing (Table 1). They were susceptible to both PXO99^A(pHM1dTALE-Xa10-Ni) and PXO99^A(pHM1dTALE-Xa23-Ni-2) (Supplementary Fig. S9). It was noted that all of the double knock-out mutants had normal morphological phenotypes in growth and development (data not shown). The results indicated that, except for functioning in disease resistance, Xa10-Ni or Xa23-Ni might not be required for normal plant growth and development.

Download as PowerPoint

Table 1. Xa10-Ni and Xa23-Ni double knock-out mutants

Mutant line	Mutation in Xa10-Ni	Mutation in Xa23-Ni
L1	ATTGCGATTGTCAT, deletion	T, insertion
L2	ATTGCGATTGTCAT, deletion	A, insertion
L3	ATTGCGATTGTCAT, deletion	C, insertion
L4	ATTGCGATTGTCAT, deletion	G, insertion
L5	ATAT, deletion	C, insertion
L6	GCGAT, deletion	G, insertion
L7	GCGAT, deletion	T, insertion
L8	GCGAT, deletion	A, insertion
L9	GCGAT, deletion	CCGG, deletion

View as image HTML

XA10-Ni and XA23-Ni induce cell death in N. benthamiana.

XA10-Ni and XA23-Ni were transiently expressed in leaf cells of N. benthamiana via Agrobacterium tumefaciens-mediated infiltration. Both XA10-Ni and XA23-Ni were able to induce cell death at the infiltrated area on leaves of N. benthamiana (Fig. 4). The early cell death with the damage of cell membrane in leaves of N. benthamiana could be detected by trypan blue staining at 24 h after infiltration (HAI) (Fig. 4). N. benthamiana infiltrated with A. tumefaciens harboring an empty vector only left infiltration marks due to mechanical wounding (Fig. 4). Compared with XA23-Ni, both XA10 and XA10-Ni have an extended 19-aa N-terminal region (Fig. 5A). Based on the results of transmembrane helix prediction using the SOSUI program (http://harrier.nagahama-i-bio.ac.jp/sosui/sosui_submit.html), this 19-aa N-terminal region belongs to the first transmembrane helix in XA10 or XA10-Ni (Supplementary Fig. S10). The XA10 mutant with this N-terminal region removed (D19XA10) still caused cell death in N. benthamiana (Fig. 5B). The XA10-Ni mutant with the 19-aa N-terminal region deleted (D19XA10-Ni) slightly delayed its induction of cell death in N. benthamiana (Fig. 5B). XA10, XA10-Ni, and XA23-Ni have a short hydrophilic C-terminal regions (Fig. 5A). Mutation in the aspartic acid/glutamic acid (DE)-rich motif of XA10, such as XA10C113T and XA10NQ, abolished cell death induction in N. benthamiana and disease resistance in rice (Tian et al. 2014). Domain swapping for the C-terminal regions was conducted between members of the three XA10-like proteins. The resulted hybrid proteins (nXA10-cXA10-Ni, nXA10-cXA23-Ni, nXA10-Ni-cXA10, nXA10-Ni-cXA23-Ni, nXA23-Ni-cXA10, and nXA23-Ni-cXA10-Ni) were still able to induce cell death in N. benthamiana (Fig. 5B). The results demonstrated that the 19-aa N-terminal regions in XA10 and XA10-Ni are dispensable and the short hydrophilic C-terminal regions of XA10-like proteins are functionally conserved and interchangeable.

Download as PowerPoint

Download as PowerPoint

XA10-Ni and XA23-Ni locate to the ER membrane and trigger ER calcium depletion.

The subcellular localization of XA10-Ni and XA23-Ni were detected by coexpressing enhanced yellow fluorescent protein (eYFP)-RcDGAT2, an ER membrane marker (Tian et al. 2014), with functional XA10-Ni-enhanced cyan fluorescent protein (eCFP) or XA23-Ni-eCFP in leaf cells of N. benthamiana. Both XA10-Ni-eCFP and XA23-Ni-eCFP were colocalized with eYFP-RcDGAT2 to the ER membrane in leaf epidermal cells of N. benthamiana (Fig. 6). In addition, XA10-Ni and XA23-Ni showed self-interaction in a bimolecular fluorescence complementation (BiFC) assay (Fig. 7). The fluorescent signal from the N-terminal region of YFP (nYFP)-XA10-Ni/C-terminal region of YFP (cYFP)-XA10-Ni or nYFP-XA23-Ni/cYFP-XA23-Ni interactions had localization patterns on the nuclear envelope and the ER membrane similar to those of XA10-Ni-eCFP or XA23-Ni-eCFP alone (Fig. 7). No signal was detected in the control experiments for the interaction between nYFP and cYFP-XA10-Ni or between nYFP and cYFP-XA23-Ni (Fig. 7). The results indicated that both XA10-Ni and XA23-Ni locate to the ER membrane and form oligomers through self-interaction. ER is one of the important intracellular Ca²⁺ stores in eukaryotic cells. To check whether XA10-Ni and XA23-Ni had a function similar to that of XA10 in inducing Ca²⁺ depletion from ER (Tian et al. 2014), XA10-Ni-mCherry or XA23-Ni-mCherry was transiently coexpressed with the ER luminal Ca²⁺ indicator YC4.60ER or cytosol Ca²⁺ indicator YC3.60 in N. benthamiana. The ER Ca²⁺ concentrations in the XA10-Ni-mCherry- and XA23-Ni-mCherry-expressing cells, measured by the emission ratio of fluorescence resonance energy transfer (FRET)/CFP of YC4.60ER, were 41.3 and 48.6% lower, respectively, than that in the control mCherry-expressing cells (Fig. 8). Meanwhile, the cytosolic Ca²⁺ concentration in the XA10-Ni-mCherry- and XA23-Ni-mCherry-expressing cells, measured by the emission ratio of FRET/CFP of YC3.60, were 123.1 and 77.6% higher, respectively, than that in the control mCherry-expressing cells (Fig. 8). The results indicated that the ER Ca²⁺ depletion in the N. benthamiana cells, which is accompanied by cytosolic Ca²⁺ increase, is associated with the expression of XA10-Ni-mCherry or XA23-Ni-mCherry.

Download as PowerPoint

Download as PowerPoint

Download as PowerPoint

DISCUSSION

The TAL effector-dependent executor R genes rely on their promoters to perceive the presence of corresponding TAL effectors from bacteria, whereas their gene products perform defense function when they are expressed (Gu et al. 2005; Römer et al. 2007; Strauss et al. 2012; Tian et al. 2014; Wang et al. 2015). In this study, we have characterized two Xa10-like executor R genes, Xa10-Ni and Xa23-Ni, in Nipponbare rice using forward and reverse genetics together with synthetic biology. The induction of Xa10-Ni and Xa23-Ni in Nipponbare provided disease resistance to X. oryzae pv. oryzae strains that deliver the matching dTALE. The ORF of Xa10-Ni or Xa23-Ni could be used to replace the coding sequence in the Xa10 gene and the resulting hybrid genes showed AvrXa10-dependent disease resistance specificity, indicating that both Xa10-Ni and Xa23-Ni can function as executor R genes in a TAL effector-dependent manner. Knock-out of Xa10-Ni or Xa23-Ni in Nipponbare abolished dTALE-dependent and XA10-Ni- or XA23-Ni-mediated disease resistance to X. oryzae pv. oryzae. Similar to the XA10 protein, both XA10-Ni and XA23-Ni are transmembrane proteins that locate to the ER membrane, trigger ER Ca²⁺ depletion and cell death, and function broadly in both monocotyledonous and dicotyledonous plants. Considering that Xa10, Xa10-Ni, Xa23-Ni, and Xa23 locate to the same genetic locus or adjacent genetic loci, they form a family of Xa10-like executor R genes that encode ER membrane-localized executor R proteins for triggering cell death or HR and activating disease resistance.

XA10-like executor R proteins are conserved in structure and function. XA23 and XA23-Ni have three predicted transmembrane helices, whereas XA10 and XA10-Ni have an extended N-terminal region forming a putative additional transmembrane helix. Deletion of this N-terminal transmembrane helix in XA10 or XA10-Ni did not affect their function in induction of cell death. In addition, the short hydrophilic C-terminal regions of XA10-like proteins are interchangeable. Like XA10, both XA10-Ni and XA23-Ni/XA23 locate to the ER membrane and show self-interaction. They also induce the depletion of free Ca²⁺ from the ER to cytoplasm in leaf epidermal cells of N. benthamiana. This ER Ca²⁺ depletion followed by the elevation of cytosolic Ca²⁺ may trigger early signaling that leads to cell death. First, Ca²⁺ is required for protein processing and glycosylation, protein folding, and subunit assembly. The reduction of ER Ca²⁺ levels results in unfolded protein response, an evolutionarily conserved stress response that can trigger apoptosis if ER dysfunction is severe or prolonged (Görlach et al. 2006; Shore et al. 2011). Second, the ER is a key organelle that initiates and modulates apoptosis. ER Ca²⁺ release is an early signaling event for the initiation of apoptosis induced by many apoptotic signals (Pinton et al. 2008). The elevation of cytosolic Ca²⁺ may activate signaling cascades involved in cell-death-associated plant immune responses. For instance, an ER-localized type IIB Ca²⁺-ATPase in N. benthamiana (NbCA1) functions in innate immunity-mediated programmed cell death, and downregulation of NbCA1 results in the modulation of intracellular calcium signaling in response to a cryptogein elicitor (Zhu et al. 2010). In our previous report, XA10 was found to form hexamers on the ER membrane, which may form a pore or affect existing ion pumps or channels on the ER to either suppress the ER Ca²⁺ intake or promote ER Ca²⁺ release (Tian et al. 2014). In this study, both XA10-Ni and XA23-Ni showed self-interaction on the ER membrane. However, the details of how the XA10-like executor R proteins cause the disruption of the ER and the change of Ca²⁺ homeostasis in the ER remain to be investigated.

Members of Xa10-like gene family were physically mapped to the same locus or adjacent loci in the rice genome; among these members, Xa23-Ni and Xa23 encode an identical protein (Gu et al. 2008; Tian et al. 2014; Wang et al. 2014a, 2015). Xa10 was originally identified in Cas209, an indica subspecies of cultivated rice (genome type: AA) (Ogawa et al. 2003; Yoshimura et al. 1983). Xa23 was derived from wild rice O. rufipogon (genome type: AA) (Zhang et al. 2001). Xa10-Ni and Xa23-Ni are present in Nipponbare, a cultivar of the japonica subspecies (genome type: AA). The Xa10-like genes were also found to be present in other cultivated and wild rice species: B8BL96 from O. sativa subsp. indica (genome type: AA); BAT14641, Q2R1Y8, and Q2R1Z3 from O. sativa subsp. japonica (genome type: AA); J3KY37, J3N992, and J3N993 from O. brachyantha (genome type: FF); A0A0E0BKP3 from O. glumipatula (genome type: AA); A0A0E0F739 from O. meridionalis (genome type: AA); A0A0E0J3N1 from O. nivara (genome type: AA); A0A0E0MGN7, A0A0E0MGN8, and A0A0E0MGN9 from O. punctata (genome type: BB); and A0A0E0RA89 and A0A0E0RA92 from O. rufipogon (genome type: AA) (Supplementary Figs. S11 and S12). The presence of Xa10-like genes in different genome types of rice and wild rice species suggests that Xa10-like genes might have originated from a single gene in an ancient rice genome before the divergence of wild and cultivated rice species. It was noted that Xa10-Ni and Xa23-Ni double knock-out mutants did not show any abnormal phenotype, indicating that, except for disease resistance to bacterial blight pathogens, the two Xa10-like genes in Nipponbare are dispensable for normal plant growth and development.

The specific recognition between EBE in the promoters of host target genes and the RVD in the central repeats of TAL effectors determines disease susceptibility or resistance. The coevolution between plants and pathogens leads to the generation of recessive insensitive alleles of S genes in plants to avoid being targeted or the evolution of new TAL effectors in pathogens to overcome this avoidance (Antony et al. 2010; Iyer and McCouch 2004; Zhou et al. 2015). Similarly, the evolution of TAL effector-dependent R genes is a kind of plant adaptation to the activity of TAL effectors as virulence factors. The mutation in the TAL effectors would enable pathogens to avoid being recognized by the EBE in the R gene promoters and, therefore, overcome the TAL effector-dependent R-gene-mediated disease resistance. Among the four Xa10-like genes characterized, Xa10 and Xa23 confer race-specific disease resistance to field X. oryzae pv. oryzae strains that deliver AvrXa10 and AvrXa23, respectively, and the corresponding TAL effector binding sites, EBE_AvrXa10 and EBE_AvrXa23, were identified in Xa10 and Xa23 promoters (Hopkins et al. 1992; Wang et al. 2014b). However, neither Xa10-Ni nor Xa23-Ni in Nipponbare has been found to confer disease resistance to any field X. oryzae pv. oryzae strains or to be induced by any known natural TAL effector. It is likely that the rapid evolution of TAL effector genes in X. oryzae pv. oryzae has overcome TAL effector-dependent Xa10-Ni- or Xa23-Ni-mediated disease resistance to incompatible bacterial strains (Vera Cruz et al. 2000). Nevertheless, the identification of Xa10-Ni and Xa23-Ni encoding functional executor R proteins provides opportunity to genetically generate novel disease resistance to Xanthomonas pathogens. One approach is to add multiple EBE, which match to the RVD of diffident TAL effectors from X. oryzae pv. oryzae or X. oryzae pv. oryzicola, to the promoters of Xa10-Ni or Xa23-Ni through transgenic or gene editing technology (Li et al. 2016). The engineered Xa10-Ni or Xa23-Ni genes can provide broad-spectrum and perhaps durable resistance to these pathogens. A similar approach has been successfully achieved in genetic engineering of Xa10 and Xa27 promoters for broad-spectrum disease resistance to X. oryzae pv. oryzae or X. oryzae pv. oryzicola (Hummel et al. 2012; Zeng et al. 2015). Because the XA10-like proteins function in N. benthamiana, they may be used to engineer novel disease resistance in dicotyledonous plants to bacterial diseases caused by Xanthomonas strains that rely on TAL effectors for virulence.

MATERIALS AND METHODS

Genes and constructs.

The central repeats in dTALE-Xa10-Ni, dTALE-Xa23-Ni-1, dTALE-Xa23-Ni-2, TALEN 3F, and TALEN 4R were designed according to the codes between RVD and DNA nucleotides (NI = A, NG = T, HD = C, NN/NH = G) (Boch et al. 2009). The central repeats in dTALE and TALEN genes were assembled and cloned into vector pTAL1 or pTAL3 according to the method described previously (Cermak et al. 2011). The SphI fragments of central DNA repeats in pTAL1-dTALE-Xa10-Ni, pTAL1-dTALE-Xa23-Ni-1, and pTAL1-dTALE-Xa23-Ni-2 were isolated and used to replace the central DNA repeats of avrXa10 in pZWavrXa10 to generate pZWdTALE-Xa10-Ni, pZWdTALE-Xa23-Ni-1, and pZWdTALE-Xa23-Ni-2, respectively. These constructs were then digested with HindIII and inserted into the HindIII site of cosmid vector pHM1 (GenBank accession number EF059993) to generate pHM1dTALE-Xa10-Ni, pHM1dTALE-Xa23-Ni-1, and pHM1dTALE-Xa23-Ni-2 (Supplementary Table S4). The cosmid constructs were transferred into X. oryzae pv. oryzae strain PXO99^A by electroporation. The BglII-SacI fragment in pTAL3-3F, which contains the TALEN gene 3F, was isolated and cloned into pZH2Bik to generate pZH2Bik-3F. The HindIII-EcoRI fragment in pZH2Bik was isolated and cloned into pUC198AM to generate pUC198AM-HE. The BglII-SacI fragment in pTAL3-4R, which contains the TALEN gene 4R, was isolated and cloned into pUC198AM-HE to generate pUC198AM-HE-4R. The AscI-MluI fragment in pUC198AM-HE-4R was isolated and cloned into pZH2Bik-3F to generate final binary construct pZH2Bik-3F-4R. The expression cassette that contains target sequence in the coding region of Xa23-Ni with sgRNA under the rice snoRNA U3 promoter in pYLsgRNA-OsU3 was amplified and cloned into vector pYLCRISPR/Cas9P_Ubi-H to generate binary construct pYLCas9-Xa23-Ni, according to the method described previously (Ma et al. 2015). The ER membrane marker for plant cells used in this study was eYFP-RcDGAT2, which was generated by fusion of eYFP with type 2 diacylglycerol acyltransferase from caster bean (Ricinus communis). The nYFP and cYFP genes that were used to fuse with Xa10-Ni or Xa23-Ni genes for BiFC assay were reported previously (Citovsky et al. 2006). Other binary constructs were made based on CAMBIA vector pC1305.1. Binary constructs were introduced into A. tumefaciens stain AGL1 for rice transformation or GV3101 for infiltration of N. benthamiana.

Rice cultivars and growth conditions.

Rice (O. sativa L.) japonica Nipponbare was used in this study. All rice plants, including transgenic plants and plants inoculated with X. oryzae pv. oryzae strains, were grown in greenhouse at 32°C for 12.5 h (light) and 25°C for 11.5 h (dark), with average humidity at 84%.

Rice transformation.

Agrobacterium-mediated transformation of Nipponbare was performed according to the method described previously (Hiei et al. 1994).

Bacterial blight inoculation.

Bacterial blight inoculation was performed using the leaf-clipping method (Yang and White 2004), and disease scoring was measured as described previously (Gu et al. 2004). Experiments were repeated three times.

5′ and 3′ RACE and RT-PCR.

The cDNA of Xa10-Ni and Xa23-Ni were isolated by 5′ and 3′ RACE using a SMART RACE cDNA amplification kit (Clontech) as well as by RT-PCR. Both 5′ RACE and 3′ RACE were conducted in accordance with the manufacturer’s instructions. The PCR products were cloned into pGEMT-easy vector (Promega Corp.) and sequenced. The specific primers for 5′ RACE and 3′ RACE of Xa10-Ni were Xa10-Ni-5′-R (5′-ACAATCGCAATATGGACGGC-3′) and Xa10-Ni-3′-F (5′-CGACCCTCTTCTTCCTCCGCA-3′), respectively. The specific primers for 5′ RACE and 3′ RACE of Xa23-Ni were Xa23-Ni-5′-R (5′-GATCATGTGTATACCGGCTACG-3′) and Xa23-Ni-3′-F (5′-TCACCGTTTCCAACAGCCTCCG-3′), respectively. The specific primers for RT-PCR of Xa10-Ni were Xa10-Ni-RT-F (5′-GCTTCTCCTCATCGCCGTCC-3′) and Xa10-Ni-RT-R (5′-GCGTGAGAGGAGCAATGCGG-3′). The specific primers for RT-PCR of Xa23-Ni were Xa23-Ni-RT-F (5′-CGTAGCCGGTATACACATGATC-3′) and Xa23-Ni-RT-R (5′-CGACGAAGCAAGCAAAGCTGGC-3′).

qRT-PCR.

qRT-PCR was performed in accordance with the procedures described previously (Gu et al. 2011). The expression of the rice ubiquitin gene 5 (UBQ5) was used as the internal control. The specific primer pairs for Xa10-Ni, Xa23-Ni, and UBQ5 were Xa10-Ni-Q-F (5′-CGACCCTCTTCTTCCTCCGCA-3′)/Xa10-Ni-Q-R (5′-TGTACCATCTGATGAGCGCGAG-3′), Xa23-Ni-Q-F (5′-GTAGCCGGTATACACATGATCCTC-3′)/Xa23-Ni-Q-R (5′-GTCTCTCGGTCATGCTGTTCTAC-3′), and UBQ5-F (5′-AACCACTTCGACCGCCACT-3′)/UBQ5-R (5′-GTTCGATTTCCTCCTCCTTCC-3′), respectively. The qRT-PCR experiments were performed in triplicate and the data are presented as means ± standard deviation.

Infiltration of N. benthamiana with Agrobacterium and confocal microscopy.

Suspensions of Agrobacterium strain GV3101 harboring binary constructs were infiltrated into leaves of N. benthamiana as described previously (Kay et al. 2007). The infiltrated plants were grown under 16 h of light and 8 h of darkness at 25°C. Phenotype was checked at 24 to 48 HAI. For confocal microscopy, water-mounted leaf sections were examined using an LSM 510 Exciter Upright confocal microscope system (Carl Zeiss). The excitation/emission combinations for eCFP, eYFP, mCherry, and autofluorescence of chlorophyll were band passes of 405/475 to 525, 514/560 to 615, 543/600 to 650, and 514/610 nm, respectively.

Trypan blue staining.

Cell death in plant tissues was detected by trypan blue staining in accordance with the procedures described previously (Tian et al. 2014).

Measurement of calcium in leaf cells of N. benthamiana.

The free Ca²⁺concentration in the ER and cytoplasm of leaf cells of N. benthamiana was measured with Ca²⁺ indicators YC4.60ER and YC3.60 using the method described previously (Iwano et al. 2009). For confocal microscopy, N. benthamiana cells at 20 to 30 HAI were imaged for mCherry, eCFP, eYFP, and FRET using an LSM 510 Exciter Upright confocal microscope system (Carl Zeiss). The excitation/emission combinations for mCherry, eCFP, eYFP, and autofluorescence of chlorophyll were band passes of 543/600 to 650, 458/475 to 525, 488/505 to 530, and 458/505 to 530 nm, respectively. Images analysis and FRET ratio calculation were conducted using the methods described previously (Tian et al. 2014).

ACKNOWLEDGMENTS

We thank Y. Liu for providing plasmids pYLsgRNA-OsU3 and pYLCRISPR/Cas9P_Ubi-H, D. Voytas for the Golden Gate TAL effector kit (through Addgene), and K. H. Ong for critical reading of the manuscript. This research is supported by the National Research Foundation, Prime Minister’s Office, Singapore under its Competitive Research Programme (NRF-CRP7-2010-02) and administered by Temasek Life Sciences and National University of Singapore.