The XA21 binding protein XB25 is required for maintaining XA21-mediated disease resistance


For correspondence (e-mail


Plant genomes encode a large number of proteins that potentially function as immune receptors in the defense against pathogen invasion. As a well-characterized receptor kinase consisting of 23 tandem leucine-rich repeats, a transmembrane domain and a serine/threonine kinase, the rice (Oryza sativa) protein XA21 confers resistance to a broad spectrum of Xanthomonas oryzae pv. oryzae (Xoo) races that cause bacterial blight disease. We report here that XA21 binding protein 25 (XB25) belongs to the plant-specific ankyrin-repeat (PANK) family. XB25 physically interacts, in vitro, with the transmembrane domain of XA21 through its N–terminal binding to transmembrane and positively charged domain (BTMP) repeats. In addition, XB25 associates with XA21 in planta. The downregulation of Xb25 results in reduced levels of XA21 and compromised XA21-mediated disease resistance at the adult stage. Moreover, the accumulation of XB25 is induced by Xoo infection. Taken together, these results indicate that XB25 is required for maintaining XA21-mediated disease resistance.


Plant innate immunity is activated by the recognition of microbe-associated molecule patterns (MAMPs) by the host pattern recognition receptors (PRRs). PRRs often appear to be receptor-like kinases (RLKs), consisting of an extracellular ligand-binding domain, a transmembrane (TM) domain and an intracellular serine/theronine-kinase domain (Bittel and Robatzek, 2007; Boller and Felix, 2009). Well-known pairs of MAMP–PRR interactions include the Arabidopsis flagellin sensitive 2 (FLS2) and elongation factor receptor (EFR) that recognize 22- and 18-amino-acid (aa) peptides (flg22 and elf18) derived from the bacterial flagellin and elongation factor Tu (EF-Tu), respectively (Gómez-Gómez et al., 2001; Zipfel et al., 2006). The perception of MAMPs by host PRRs results in multiple downstream responses. For example, a number of Arabidopsis proteins, including AtPHOS32, AtPHOS34 and AtPHOS43, are rapidly phosphorylated in response to flg22 treatment (Peck et al., 2001; Merkouropoulos et al., 2008). AtPHOS32 is highly similar to AtPHOS34. Biochemical data revealed that AtPHOS32 serves as a substrate of AtMPK3 and AtMPK6, the two mitogen-activated protein kinases (MPKs) that are activated by flg22. AtPHOS43 belongs to a plant-specific ankyrin-repeat (PANK) family, the members of which are characterized by a conserved ankyrin repeat domain at the C terminus (Wirdnam et al., 2004). Evidence indicates that the flg22-induced phosphorylation of AtPHOS43 is dependent on FLS2, the direct connection with FLS2 and the precise role of AtPHOS43 in flg22/FLS2-mediated innate immunity remains to be demonstrated.

As a paralog of AtPHOS43, the PANK protein AKR2A has also been implicated in the regulation of disease resistance, but in a negative manner (Yan et al., 2002). AKR2A interacts with a 14–3–3 binding protein and with ASCORBATE PEROXIDASE 3 (APX3), a peroxisomal membrane-bound enzyme that plays an important role in hydrogen peroxide (H2O2) scavenging. Reduction of AKR2A by antisense leads to increased H2O2 production, necrotic leaves, and enhanced disease resistance. In addition to the 14–3–3 protein and APX3, AKR2A interacts with a number of other proteins localized on the membranes of diverse organelles, including microsomes, mitochondria, peroxisomes and chloroplasts (Bae et al., 2008; Shen et al., 2010). All of these AKR2A binding proteins (except 14–3–3) contain a single TM domain followed by a short sequence tail rich in positively charged amino acid (aa) residues. This structure, reminiscent of the membrane peroxisomal targeting signal (mPTS) of peroxisomal integral membrane proteins (Dyer et al., 1996), has been shown to be responsible for binding to the N-terminal half of AKR2A. Evidence from genetic and cell biology studies supports the hypothesis that AKR2A acts as a chaperone for the biogenesis of intrinsic organelle membrane proteins. The questions of whether AKR2A is involved in PRR-mediated immunity and what sequence characteristics of the N–terminal region of AKR2A are responsible for binding to its target proteins remain to be addressed.

Increasing evidence suggests that the recognition of the pathogen Xanthomonas oryzae pv. oryzae (Xoo) by the Oryza sativa (rice) receptor kinase XA21 also belongs to the category of MAMP–PRR interactions. In this case, XA21 recognizes a sulfated peptide (axYS22) embedded in the N–terminal region of the bacterial polypeptide Ax21 and initiates defense responses, leading to resistance to this pathogen (Lee et al., 2009). In contrast to the genes encoding FLS2 and EFR, Xa21 was originally identified as a resistance gene in nature, which confers broad resistance to bacterial blight disease of rice (Song et al., 1995; Wang et al., 2006). Cell biology studies indicate that XA21 is localized on the plasma membrane and the endoplasmic reticulum (ER; Park et al., 2010; Chen et al., 2010a). Increasing evidence shows that XA21 functions in a protein complex for its stability as well as for signaling. For instance, XA21 interacts, through its intracellular kinase domain (XA21K), with the E3 ubiquitin ligase XB3 in planta (Wang et al., 2006). XB3 can be phosphorylated by the kinase domain of XA21 in vitro. Downregulation of the Xb3 gene significantly destabilizes XA21 and compromises Xa21-mediated disease resistance. Peng et al. (2008) identified the rice transcription factor OsWRKY62 as another XA21 binding protein, namely XB10. Overexpression of OsWRKY62 compromises both basal and Xa21 resistance. A third XA21 binding protein, XB15, is a protein phosphatase 2C that can dephosphorylate the autophosphorylated XA21 in vitro (Park et al., 2008). The interaction between XB15 and XA21 has been confirmed in planta. Overexpression of Xb15 reduces Xa21-mediated disease resistance. Two additional XA21 binding proteins that have been characterized are XB24 and BiP3 (Chen et al., 2010b; Park et al., 2010). XB24 appears to be an ATPase in which the association with XA21 is disrupted in response to pathogen inoculation, whereas BiP3 is an ER chaperone. Recently, a genome-scale gene network study identified three regulators of XA21 (ROX), of which ROX1 encodes a thiamine pyrophosphokinase, ROX2 belongs to a member of the NOL1/NOL2/sun gene family and ROX3 encodes a nuclear migration protein, nudC (Lee et al., 2011). Genetics studies suggest that ROX1 and ROX2 positively regulate XA21-mediated disease resistance, whereas ROX3 plays a negative role in the XA21 signaling pathway. Thus far, most of the identified XBs differ from the components described in FLS2 signaling. Here we report an XA21-binding protein, named XB25 for XA21 binding protein 25, and its role in XA21-mediated signaling.


XB25 is closely related to AtPHOS43 and AKR2A

Two different regions of XA21K have been previously used as bait to screen yeast two-hybrid libraries, and yielded distinct XBs (Wang et al., 2006; Peng et al., 2008; Chen et al., 2010b; Park et al., 2010). To identify additional XBs, a truncated version of XA21 (XA21KTM), spanning the TM domain and XA21K [aa 651–1025], was fused to the DNA binding region (DB) of the GAL4 transcription factor. Yeast two-hybrid screenings using this bait identified a cDNA, 25–1, which contains a complete open reading frame (ORF) and 46 bp of a 5′ untranslated region (5′–UTR). This ORF (Os09g33810) was previously named OsBIANK1, the expression of which is induced by benzothiadiazole treatment or Magnaporthe grisea infection (Zhang et al., 2010). Because this gene has not been functionally characterized, we use the name Xb25 in this paper.

Xb25 encodes a product (XB25) of 329 aa in length with a theoretical molecular weight of 35 kDa (Figure 1a). It belongs to the PANK family and shares high sequence similarities with other members (Wirdnam et al., 2004; Table S1). Overall, XB25 shares 73 and 65% sequence similarity and identity, respectively, with AtPHOS43. Four ankyrin repeats (aa 205–329), well-known for protein–protein interactions (Sedgwick and Smerdon, 1999), constitute the C–terminal half of XB25 (Figure 1a). A motif database search did not find any known domain in the N–terminal portion of XB25. However, detailed sequence analysis revealed interesting characteristics. First, this region is rich in glycine and proline (Figure 1a). More importantly, there are 11 repeats of 12–22 residues, in which one or two hydrophobic residues followed by a negatively charged amino acid are located in the center of the repeats (Figure 1a). The potential function of these repeats will be discussed below.

Figure 1.

XB25 specifically interacts with XA21KTM in yeast. (a) Predicted amino acid (aa) sequence of XB25. The deduced domains are indicated as follows: I, unknown; II, binding to transmembrane and positively charged domain (BTMP) repeats; III, negatively charged domain; IV, ankyrin repeats. The conserved hydrophobic and negatively charged residues in the BTMP domain are highlighted in bold red. Glycine and proline in BTMP repeats are underlined. The conserved aa of the ankyrin repeats are indicated in bold blue. (b) Schematic representations of XA21, Pi-d2, XB25 and their derivatives. LRR, leucine-rich repeats; TM, transmembrane domain; JM, juxtamembrane domain. The TM sequence of XA21 is underlined and shown above. The 12 aa following the TM domain is also shown with the positive residues highlighted in bold orange. Autophosphorylated residues of XA21: S686, Ser686; T688, Thr688; and S689, Ser689. Pi-d2, a rice receptor-like kinase conferring resistance to blast disease, was used as a control. Boxes in pink denote the BTMP repeats of XB25, whereas boxes in green represent the ankyrin repeats of this protein. (c) Yeast two-hybrid analysis of the interaction of XA21KTM and XB25. The constructs indicated were co-transformed into the yeast strain CG1945. Cells growing up on the double-deficient medium (SD, –Trp, –Leu were replicated onto the triple-deficient medium (SD, –Trp, –Leu, –His. Cells capable of growing on this medium indicate positive interactions. AD, activation domain of the yeast GAL4 transcription factor; BD, DNA binding domain of the yeast GAL4 transcription factor.

In the rice genome databases, there are two Xb25 paralogs, named Xbos251 (Os03g63480) and Xbos252 (Os08g42690). They share similar structures and high levels of identity (68 and 67%) with XB25 at the amino acid level (FigureS1; Table S1).

XB25 specifically interacts with XA21KTM in yeast

We confirmed the XA21KTM-XB25 interaction in yeast by re-transformation experiments. Different from XB3, whose ankyrin domain is bound to XA21K (Wang et al., 2006), the N–terminal region of XB25 (XB25N, aa 1–214), rather than the C–terminal ankyrin repeats (XB25C, aa 195–329), is responsible for interacting with XA21KTM (Figures 1b, c). Neither XB25 nor XB25N interacted with XA21K. Furthermore, neither XB25 nor XB25N interacted with the kinase-inactive mutant XA21KTMK736E or Pi-d2, an RLK involved in resistance to the fungal pathogen M. grisea, which causes blast disease of rice (Liu et al., 2002; Chen et al., 2006). However, they all interacted with the autophosphorylation mutant XA21KTMS686A/T688A/S689A (Liu et al., 2002; Xu et al., 2006). Taken together, these results demonstrate that the interaction between XA21KTM and XB25 is specific, and that the TM domain of XA21 is required for the binding.

XB25N can interact with the TM domain of XA21

In vitro pull-down assays were used to further characterize the XA21KTM–XB25 interaction. XA21KTM and its truncated derivatives were expressed as maltose binding protein (MBP) fusions. As indicated in Figure 1b, XA21TMCF (aa 651–688) spans the TM domain and 12 C–terminal flanking amino acids rich in positively charged residues, whereas XA21TM (aa 651–676) encompasses the TM domain only. Amylose resin with purified fusion proteins was incubated with FLAG-tagged XB25N. Following repeated washings, MBP fusions on the resin were detected using anti-MBP antibodies, and their binding proteins were analyzed using anti-FLAG antibody. As shown in Figure 2, FLAG-XB25N can be individually pulled down by MBP-XA21KTM, MBP-XA21TMCF or even MBP-XA21TM, but not by a similar level of MBP-XA21K, MBP-Pi-d2KTM and MBP alone. Notably, the quantity of FLAG-XB25N purified by MBP-XA21TMCF is higher than that purified by MBP-XA21TM. These results indicate that XB25N interacts with the TM region of XA21 and that the positive residues downstream of the TM domain also contribute to the binding. Consistent with the yeast two-hybrid data described above, XB25N does not interact with the corresponding region of Pi-d2, indicating that the interaction between XA21TMCF and XB25N is specific.

Figure 2.

XB25N interacts with the transmembrane domain (TM) of XA21 in vitro. The MBP-fusion proteins indicated were individually expressed in Escherichia coli and purified by affinity chromatography. The purified protein was mixed with equal quantities of bacterial extracts containing FLAG-XB25N and pulled down by amylase resin. After extensive washing, the pull-down products were subjected to protein blot analysis using the indicated antibodies. One-tenth of the input of bacterial extracts expressing XB25N was loaded as a control. The predicted molecular weight for each fusion protein is as follows: MBP-XA21KTM, 85.3 kDa; MBP-XA21K, 83.7 kDa; MBP-Pi-d2KTM, 87.2 kDa; MBP-XA21TMCF, 47.4 kDa; MBP-XA21TM, 46.1 kDa; MBP, 49.1 kDa. The molecular weight of MBP is larger than that of MBP-XA21TMCF or MBP-XA21TM because an extra 61-aa-long tail, encoded by the sequence after the multiple cloning site of the pMAL-C2x vector, was co-translated with MBP.

XB25 interacts with XA21 in planta

Polyclonal antibodies (anti-XB25M) against a middle region of XB25 (aa 82–214) were raised. These antibodies specifically recognized bacterially expressed XB25 (Figure S2) and two bands of 42 and 48 kDa in total protein extracts prepared from the rice variety Taipei309 (TP309) (Figure 3a). To determine which of these two proteins is XB25, the expression of Xb25 was downregulated by using an established RNA interference (RNAi) system (Wang et al., 2006). A 336–bp sequence derived from a 3′–UTR of Xb25 was used as the gene-specific probe. This probe shares less than 50% sequence identity with the corresponding regions of Xbos251 and Xbos252. No sequence stretch of more than 18 bp is identical among the Xb25, Xbos251 and Xbos252 regions. A 979–bp fragment from the bacterial uidA gene was inserted between two inverted Xb25 probes, and the resultant construct, RNAiXB25, was placed in a rice expression cassette under the control of a Zea mays (maize) ubiquitin promoter.

Figure 3.

XB25 interacts with XA21 in Oryza sativa (rice) protein extracts. (a) RNA blot analysis shows reduced Xb25 transcripts in three Xb25 RNA interference lines: S34, S41 and S42. Total RNA isolated from the indicated lines was hybridized with an Xb25-specific probe (Top). Xb25 transcripts are indicated. Total RNA is used as a loading control (bottom). (b) Immunodetection of XB25 with anti-XB25M in rice. Total protein extracts were prepared from the indicated lines and probed with anti-XB25M (top). The XB25 band of 42 kDa is indicated by an arrow. Ponseau S staining of the same blot is shown as a loading control (lower). n.s., a non-specific product cross-reacting with anti-XB25M. (c) Co-immunoprecipitation confirms the association of XA21 and XB25. Total protein extracts from the lines indicated were immunoprecipitated with IgG beads. One-fifth of the precipitates were subjected to protein blot analysis using PAP (Pellet, top) or anti-XB25M (Pellet, bottom) antibodies. Total protein extracts from the same lines were used as a control (Supernatant). ProA-XA21, TAP-Os06g48590 and XB25 are indicated by arrows. The asterisk shows a degraded product of ProA-XA21. n.s., a non-specific product. The experiments were repeated three times with similar results.

RNAiXB25 was transformed into TP309, which is highly susceptible to Xoo. More than 60 independent transgenic lines were generated, most of which expressed drastically decreased Xb25 transcripts as compared with TP309. However, no visible phenotypes were observed in these plants (Figure S3). Three lines, S34, S41 and S42, were randomly selected for further studies, and their levels of Xb25 transcripts are shown in Figure 3a. To determine the specificity of RNAiXb25-mediated interference, the transcript levels of Xb25, Xbos251 and Xbos252 were measured by quantitative reverse transcription polymerase chain reaction (RT-PCR). Consistent with the results from RNA blot analysis, the Xb25 level was dramatically reduced in S34, S41 and S42, whereas the expression levels of Xb25 paralogs were comparable between TP309 and the three transgenic lines (Figure S4). Consistent with reduced Xb25 transcripts, the abundance of the 42–kDa band in all three RNAiXB25 lines was significantly decreased (Figure 3b, upper panel). By contrast, the 48–kDa band remained at the same level in TP309 and RNAiXB25 lines. These observations indicate that the 42–kDa polypeptide detected in rice protein extracts is XB25, despite its slower migration in the SDS-PAGE gel than the predicted XB25 (35 kDa). The 48–kDa band observed might be a product that non-specifically cross-reacts with anti-XB25M.

To demonstrate the XA21–XB25 interaction in planta under physiological conditions, co-immunoprecipitation analysis was carried out using the established ProA-tagged XA21 line 716–1 (Wang et al., 2006). The XA21 complex was immunoprecipitated using IgG beads. The presence of XA21 in this complex was confirmed by protein blot analysis using the peroxidase-antiperoxidase (PAP) antibody (Figure 3c). The 110–kDa band detected by the same antibody is likely to come from the degradation of ProA-XA21 (Wang et al., 2006). As a control, no product was recognized by the PAP antibody in the precipitates from the recipient line TP309. To examine the presence of XB25 in the same complex, the immunoprecipitates were also probed with anti-XB25M. As expected, the XB25 protein of 42 kDa in length was detected in the 716–1 line, but not from TP309 (Figure 3c). The nonspecific band of 48 kDa was not found in the precipitates, indicating that this protein is not part of the XA21 complex. To exclude the possibility that XB25 may bind to the 128–aa ProA tag, we conducted a similar co-immunoprecipitation assay using the transgenic line A6 that expresses a TAP-tagged kinase (Os8g37800) unrelated to XA21. XB25 was not found in the Os08g37800 precipitates (Figure 3c). We conclude that XB25 physically interacts with XA21 in rice.

XB25 is weakly transphosphorylated by XA21KTM in vitro

We performed transphorylation assays to determine whether XA21 can phosphorylate XB25. The purified MBP-XA21KTM and FLAG-XB25 fusion proteins were incubated with [γ-32P]ATP. After the reaction, the samples were subjected to SDS-PAGE analysis followed by autoradiography. Consistent with our previous observations, MBP-XA21KTM was strongly 32P-labeled, and also caused labeling of the known substrate XB3, indicating that the purified XA21KTM is an active kinase capable of autophosphorylation and transphosphorylation of substrates (Wang et al., 2006). However, only a weak radio-labeled signal was observed from the product corresponding to FLAG-XB25 in the presence of MBP-XA21KTM. This product was absent when FLAG-XB25 was incubated with either the dead kinase mutant MBP-XA21KTMK736E or with itself (Figure 4). Because there is no serine or threonine present in the FLAG tag, these results indicate that XB25 is weakly phosphorylated by XA21.

Figure 4.

XB25 is phosphorylated by XA21KTM in vitro. FLAG-XB25, MBP-XA21KTM and the other fusion proteins indicated were expressed in Escherichia coli and affinity purified. Phosphorylation assays were performed in the presence of [γ–32P]ATP. Autoradiography (Autorad) and Coommassie Brilliant blue staining (CBB staining) of the same gel are shown. FLAG-XB3 was used as a positive control, with the kinase-deficient mutant XA21KTMK736E as a negative control. FLAG-XB25, FLAG-XB3 and MBP-XA21KTM are indicated by arrows. n.s., a non-specific product. The experiments were repeated three times with similar results.

XB25 contributes to the accumulation of XA21

XB3 is required for the stability of XA21 at the adult stage of rice plants (Wang et al., 2006). To test whether XB25 has a similar role, the RNAiXB25 lines S34 and S41 were crossed with the homozygous Myc-XA21 line 4021–3 (pollen donor). S34 and S41 are the T0 generation of transgenic lines, and are often heterozygous for RNAiXB25. Indeed, the levels of XB25 were significantly reduced in four out of six of the F1 progeny at both the seedling and the adult stages of rice plants (Figure 5a, b, top panels). All the F1 progeny examined expressed XA21, as evidenced by protein blot analysis at the seedling stage (1 month old; Figure 5a, middle panel). We then monitored the XA21 abundance of the F1 plants at the adult stage (4 months old). As shown in Figure 5b (middle panel), the XA21 levels were dramatically reduced only in the progeny, showing lower levels of XB25. Quantitative RT-PCR assays showed that Xa21 transcripts were comparable in the F1 progeny (Figure 5c). These results indicate that XB25 is required for stabilizing XA21 at the adult stage.

Figure 5.

XB25 contributes to the accumulation of XA21- and XA21-mediated disease resistance. The heterozygous Xb25 RNA interference lines S34 and S41 were individually crossed with the homozygous Myc-Xa21-containing line 4021–3. (a, b) XB25 and XA21 levels in the F1 progeny, TP309 and TP309/4021–3 (the progeny from the TP309 and 4021–3 cross) were immunodetected by anti-XB25M antibodies (top) or anti-c–Myc antibody (middle) at the seedling stage (1 month old; a) and adult stage (4 months old; b). Ponseau S staining of the same blots is shown as a loading control (bottom). (c) Real-time quantitative PCR analysis of Xa21 transcripts. Results were normalized to actin expression, then to the value of TP309/4021–3 crossing line, which was set to ‘1’. Error bars represent standard derivations. (d) Lesion length of corresponding lines 2 weeks post-inoculation with PXO99A. Each data point represents three replications. The standard deviations are indicated. (e) Photograph of rice leaves showing lesion development 2 weeks after inoculation with PXO99A. Leaves 1, TP309/4021–3; leaves 2, TP309; and leaves 3, S34/4021–3–6, with reduced XA21 and XB25. (f) Growth of PXO99A in S34/4021–3–6 and control lines. Each point represents three independent replications and standard deviations are indicated. P < 0.05 (Student's t–test) for TP309/4021–3 versus S34/4021–3–6 after day 2, P < 0.01 (Student's t–test) for TP309 versus S34/4021–3–6 after day 2. (g) Photograph of rice leaves showing lesion development 4 weeks post inoculation with PXO99A. Leaves 1, TP309/4021–3; leaves 2, TP309; and leaves 3, S34/4021–3 (-7 and -8), with reduced XB25. (h) XB25 and XA21 levels in the indicated lines were immunodetected by anti-XB25M antibodies (top) or anti-c–Myc antibody (middle) at the seedling stage (1 month old). The lesion length of the corresponding lines was scored 4 weeks post inoculation with PXO99A. Each data point represents three replications. The standard deviations are indicated. The asterisk indicated that the entire leaves of TP309 were dead 4 weeks post inoculation.

XB25 is required for maintaining XA21-mediated disease resistance

F1 plants and their parents were inoculated with the Xoo strain PXO99A containing Ax21 to determine whether XB25 is involved in XA21-mediated disease resistance. S34 and S41 were equally susceptible as TP309 (Figure S5). We have previously shown that XA21-mediated resistance is dose dependent (Wang et al., 2006). All of the F1 plants with reduced levels of XA21 and XB25 showed longer lesions, compared with the heterozygous Xa21 line (TP309/4021-3-6) generated from a cross of TP309 and 4021–3 (Figure 5d). Bacterial growth curve analysis confirms that Xoo levels were higher in the F1, lines with longer lesions than those of the TP309/4021-3-6 line (Figure 5e, f).

To test whether the subtle difference of disease lesion development observed between the RNAiXB25/4021–3 and TP309/4021–3 lines will increase further when observed more than 2 weeks post inoculation, we repeated the inoculation experiments with the F1 progeny produced from the same cross described above. Only the lines with expressed XA21 and downregulated XB25 were chosen (Figure 5g, h). Consistent with the results from the previous inoculation experiments, the F1 plants with reduced XB25 levels displayed slightly longer lesions as compared with the heterozygous Xa21 line 2 weeks after inoculation. However, such differences were significantly greater when the same plants were scored 4 weeks after inoculation (Figure 5g). These results indicate that XB25 is required for maintaining XA21-mediated disease resistance.

XB25 is induced by infection with PXO99A

Xoo infection can alter the expression of a large number of rice genes (Li et al., 2006; Yang et al., 2006; Kottapalli et al., 2010; Gan et al., 2011). We monitored Xb25 transcript levels and the XB25 abundance in both TP309 and the XA21-expressing line 4021–3 after PXO99A inoculation. As shown in Figure 6, both RNA transcript and protein levels of XB25 are increased 12 h after inoculation in both TP309 and 4021–3 lines, indicating that the expression of Xb25 is induced by pathogen inoculation in an Xa21-independent manner. This elevated expression continues for at least 3 days.

Figure 6.

XB25 is induced by Xoo infection in an Xa21-indpendent manner. Rice samples were collected in the Myc-Xa21-containing line 4021–3 and its recipient line TP309 at different time points after inoculation. (a) The expression of Xb25 was analyzed by real-time quantitative PCR. The expression of Xb25 was normalized by actin gene expression, then to the value of time 0, set as ‘1’. (b) The protein levels of XB25 were immunodetected by using anti-XB25M antibodies (top). The intensity of the XB25 band in each lane was calculated with ImageJ, and then normalized to the value of TP309-0h, which is set as ‘1’ (middle). Ponseau S staining of the same blot was used as a loading control (bottom).

The N–terminal regions of XBOS251 and XBOS252 specifically interact with XA21KTM in yeast

To test whether XBOS251 and XBOS252 are capable of binding with XA21KTM, the N–terminal regions of XBOS251 (XBOS251N, aa 1–235) and XBOS252 (XBOS252N, aa 1–216), corresponding to XB25N, were subjected to yeast two-hybrid assays with XA21KTM. Like XB25N, both XBOS251N and XBOS252N interacted with XA21KTM or XA21KTMS686A/T688A/S689A. Neither of them bound to XA21K, XA21KTMK736E, Pi–d2 or the empty activation domain (AD) (Figure 7). Therefore, the interactions between XB25 paralogs and XA21KTM are specific.

Figure 7.

Both XBOS251N (aa 1–235) and XBOS252N (aa 1–216) specifically interact with XA21KTM in yeast. Yeast cells, transformed with the constructs indicated, were replicated on the double-deficient medium (SD, –Trp, –Leu) and the triple-deficient medium (SD, –Trp, –Leu, –His), respectively. Positive interactions are evidenced by cells capable of growing on the triple-deficient medium. AD, activation domain of the yeast GAL4 transcription factor; BD, DNA binding domain of the yeast GAL4 transcription factor.


Members of the PANK family may serve as common signaling regulators in plant PRR-mediated immunity. AtPHOS43 is phosphorylated in response to flg22, although the kinase(s) responsible for this phosphorylation has not yet been identified (Peck et al., 2001). In Solanum lycopersicum (tomato) and rice, proteins cross-reacting with anti-AtPHOS43 antibodies are also phosphorylated after treatment with flg22 and chitin, respectively, suggesting that AtPHOS43-related proteins may be involved in elicitor-triggered signaling in other plant species (Peck et al., 2001). We show here that the PANK member XB25 physically associates with and functionally stabilizes XA21. Consequently, XA21-mediated disease resistance is compromised when XB25 abundance is reduced. Only weak phosphorylation of XB25 by XA21KTM was detected in vitro, which contrasts with the strong phosphorylation of XB3 in similar assays. Whether XB25 is a true substrate of the XA21 kinase in vivo awaits further study.

We found that an accumulation of XB25 is significantly induced by Xoo inoculation, even in the absence of Xa21. To test the potential role of XB25 in basal defense, we inoculated the RNAiXB25 plants with PXO99A. Under our experimental conditions, we did not observe any significant differences in lesion development between the RNAiXB25 plants and TP309. Therefore, it is unclear whether XB25 has a role in basal defense.

What is the role of XB25 in XA21 signaling? We propose that XB25, as AKA2A, may function as a chaperone protein to stabilize XA21. Our data indicates that XB25 acts similarly to AKR2A. Both of these two proteins use their N–terminal domain to interact with the transmembrane region of membrane-spanning proteins. Although AKR2A is required for APX3 stability (Shen et al., 2010), XB25 has a role in stabilizing XA21. Additional signals are thought to be required for AKR2A-mediated protein trafficking, as AKR2A can target multiple substrates to various destinations, including chloroplasts, peroxisomes and even the nucleus (Bae et al., 2008; Shen et al., 2010). In the case of the chloroplast outer-envelope membrane protein 7 (OEP–7), AKR2A binds to the chloroplasts through its C–terminal ankyrin repeats, although the molecular details of this binding remain obscure (Bae et al., 2008). XB25 also carries a conserved ankyrin-repeat domain at the C terminus. Thus, the induction of XB25 by Xoo might trigger dynamic trafficking of XA21, which allows XA21 to interact and regulate other downstream substrates located in the organelles. In line with this prediction, we previously found that XA21K interacts in yeast with rice catalase B (W.-Y. Song and P.C. Ronald, unpublished data), a major H2O2-scavenging enzyme thought to reside in peroxisomes to regulate the concentration of reactive oxygen species in the cell (Chen et al., 2002; Nyathi and Baker, 2006).

Despite the above similarity, our data also indicates that XB25 functions differently from AKR2A. AKR2A negatively regulates basal defense (Yan et al., 2002), whereas XB25 is positively involved in XA21 resistance. The downregulation of AKR2 results in a series of changes, ranging from reduced plant size, curled leaves, an altered response to the catalase inhibitor aminotriazole during germination and enhanced disease resistance against bacterial pathogens (Yan et al., 2002; Shen et al., 2010). All of these alternations may result from peroxisome deficiency caused by a reduced abundance of AKR2A. In contrast, transgenic plants with drastically decreased XB25 did not display obvious morphological phenotypes. These observations suggest that the likelihood for XB25 as a major mediator in biogenesis of intrinsic peroxisomal proteins is low. Another striking difference between these two proteins is subcellular localization. AKR2A is located in both the cytoplasm and the nucleus (Shen et al., 2010). XB25 was associated with the plasma membrane when expressed as a GFP fusion in onion epidermal cells (Zhang et al., 2010), which is consistent with its association with XA21. Thus, XB25 probably functions in XA21-mediated defense signaling, whereas AKR2A is a housekeeping chaperone.

In addition to XB25, both XBOS251 and XBOS252 interact with XA21KTM in yeast, suggesting a possibility that they might also be involved in XA21 signaling. This could provide an explanation for the partial resistance observed in the RNAiXB25/4021–3 lines. However, it is unlikely that these three proteins act in a completely redundant manner. The downregulation of XB25 can cause a significant reduction in XA21-mediated resistance. In Arabidopsis, AKR2A does not function equally with the closely related homologue AKR2B (Shen et al., 2010). Therefore, a more likely possibility is that the PANK family members function in an overlapping but distinct manner in XA21 immunity. It is worthy of note that, in the later stage of this study, translocalization of the intracellular kinase domain to the nucleus was reported (Park and Ronald, 2012).

The N–terminal portions of AKR2A and XB25 (and possibly XBOS251 and XBOS252) clearly represent a domain that uniquely interacts with the hydrophobic TM region and the adjacent positively charged residues of membrane-spanning proteins. However, structural characteristics of this domain have not yet been elucidated. We found that a type of amino acid repeats was embedded in this domain in XB25, which potentially allows the formation of a hydrophobic-negative ridge on one side of the protein. This structure can explain the observed interaction. We therefore propose to name the repeats identified in the N–terminal portion of XB25 ‘binding to transmembrane and positively charged residues’ (BTMP) repeats.

It is noteworthy that the hydrophobic and positively charged stretch in membrane proteins may not be sufficient for binding to a BTMP domain, because we were unable to detect the interaction between XB25 and Pi-d2KTM (Figures 1b,and 2). Additional sequence specificity is probably required for the binding. Therefore, the TM domain of XA21 serves not only as a membrane anchor, but also as a sequence-sensitive protein–protein interaction motif. Consistent with this notion, a single amino acid mutation in the TM domain of Pi-d2 abolishes its resistance against blast disease in rice, but has no apparent effects on the plasma membrane localization of this protein (Chen et al., 2006). It would be intriguing to test whether the TM sequence of XA21 is also crucial for its resistance function.

Experimental procedures

Yeast two-hybrid assays

To make the bait construct BD-XA21KTM, the sequence of XA21KTM was amplified using primers XA21KTM–1 and XA21KTM–2 (Table S2). The PCR product was cloned into the yeast two-hybrid vector pPC97 containing the GAL4 BD domain. To confirm the accuracy, all PCR products described in this study were sequenced. Yeast two-hybrid screening using BD-XA21KTM as bait was performed using the yeast strain CG1945, as previously described (Wang et al., 2006).

A similar strategy was used to construct BD-XA21KTMK60E (primers Xa21KTMK60E–1 and Xa21KTMK60E–2), BD-XA21KTMS686A/T688A/S689A (primers Xa21KTMS686A/T688A/S689A–1 and Xa21KTMS686A/T688A/S689A–2), BD-Pi-d2KTM (primers Pi-d2KTM–1 and Pi-d2KTM–2), AD-XB25N (primers Xb25N–1 and Xb25N–2), AD-XB25C (primers Xb25C–1 and Xb25C–2), AD-XBOS251 (primers Xbos–1F and Xbos251R) and AD-XBOS252 (primers Xbos252F and Xbos252R) for yeast two-hybrid analyses. The primers used to make these constructs are listed in Table S2.

In vitro pull-down assay

To make constructs for the in vitro pull-down assay, Xa21KTM, Xa21K, Pi-d2KTM, Xa21TM and Xa21TMCF fragments were cloned in frame into pMAL86, a derivative of the commercial expression vector pMAL-C2x (New England Biolabs, Xb25N was cloned in frame into pFLAG-MAC vector (Sigma-Aldrich,, with a FLAG tag at the N terminus. All fusion proteins were expressed in the Eschericha coli strain ER2566. MBP-XA21KTM, MBP-XA21K, MBP-Pi-d2KTM, MBP-XA21TM, MBP-XA21TMCF and MBP were purified by amylase resin (New England Biolabs) and incubated with equal quantities of bacterial protein extracts containing FLAG-XB25N recombinant protein. The resin was extensively washed with MBP column binding buffer (20 mm Tris-HCl, pH 7.4, 0.2 m NaCl, 1 mm EDTA and 1 mm DTT) five times, and the MBP fusion protein was eluted with 10 mm of maltose. The eluted fraction was subject to protein blot analysis using either the Anti-FLAG M2 (Sigma-Aldrich) or the anti-MBP (New England Biolabs) antibodies.

Protein immunodetection

To generate anti-XB25M antibodies, a region from the middle of Xb25 (aa 82–214) was amplified using primers Xb25M–1 and Xb25M–2 (Table S2). The PCR product was cloned in frame into the expression vectors pGTK (Liu et al., 2002) and pMAL-c2X (New England Biolabs), respectively. The MBP-XB25M and GST-XB25M fusion proteins were individually expressed in the E. coli strain ER2566 and affinity purified. Although MBP-XB25M was used to immunize rabbit (Cocalico Biologicals,, GST-XB25M was used to purify the antisera following the procedure described by Lin et al. (1996).

Co-immunoprecipitation and protein blot analysis were performed as described previously (Wang et al., 2006; Xu et al., 2006 and Park et al., 2010), with modification. Briefly, rice total protein was extracted from 10 g of leaf tissue in 40 mL of ice-cold extraction buffer [20 mm Tris-HCl, pH 8.0, 150 mm NaCl, 0.1% Triton X–100, 2.5 mm EDTA, 2 mm benzamidine (Sigma-Aldrich), 10 mm β–mercaptoethanol, 20 mm NaF, 1 mm phenylmethanesulfonyfluoride (PMSF), 1% Protease Cocktail (Sigma-Aldrich), 10 μm leupeptin, 10% glycerol]. Cell debris was removed by filtering through double layers of Miracloth (Calbiochem, now Millipore,, followed by centrifugation twice at 13 000 g for 10 min at 4°C. The supernatant was applied to a column containing 500 μL of IgG Sepharose beads that were pre-equilibrated with protein extraction buffer without protease inhibitors. The column was then washed three times with protein extraction buffer without protease inhibitors and the IgG-binding protein was eluted with 2 mL of acetic acid (pH 3.4) and concentrated by acetone precipitation. The samples were then subjected to protein blot analysis.

Generation of the RNAiXB25 transgenic lines

A 336–bp fragment from the 3′–UTR of Xb25 was selected to generate the RNAiXB25 construct. This region was PCR amplified in both sense and antisense orientations using primer sets RNAiXb25S–1 and RNAiXb25S–2, and RNAiXb25A–1 and RNAiXb25A–2, respectively (Table S2). The sense and antisense fragments, digested by restriction enzymes XbaI and BamHI, and BglII and EcoRV (New England Biolabs), respectively, were ligated to the bacterial uidA fragment spanning nucleotides 815–1793. The resultant construct was subcloned into the binary vector pCmHU under the control of the maize ubiquitin promoter (Wang et al., 2006). The RNAiXB25 construct was then subjected to rice transformation, as described previously (Xu et al., 2006).

Characterization of transgenic plants

RNA blot analysis was performed according to the standard method. An Xb25-specific probe was PCR amplified from the N terminus of the gene using primers Xb25NT–1 and Xb25NT–2 (Table S2). The probe was then synthesized using the Primer-It II Random Primer Labeling Kit for hybridization (Stratagene, now Agilent

RNAiXB25 lines S34 and S41 were used as the pollen recipient parents to cross with pollen donor 4021–3. Recovered seeds from each cross were germinated for characterization. Protein blot analysis was performed as described above.

Plant inoculation and growth curve analyses were performed as described by Song et al. (1995), except that a bacterial suspension of OD 600 ~ 1.0 was used.

Real-time quantitative PCR analysis

Primers used to amplify Xa21, actin, Xb25, Xb0s251 and Xbos252 were listed in Table S2. Total RNA was isolated as described above. Isolated RNA was treated with 10 units of RNase-free DNase (Qiagen, to eliminate genomic DNA contamination prior to cDNA synthesis. The first-strand cDNA was synthesized using the First-Strand cDNA Synthesis Kit (Applied Biosystems, Real-time quantitative PCR was performed using a MiniOpticon real-time PCR detection system (Bio-Rad, PCR amplification was carried out in a 20–μL reaction system containing 1 μL of diluted cDNA template, 100 nM of each primer and 10 μL of iQ SYBR Green Supermix (Bio-Rad). All PCRs were performed under the following conditions: 3 min at 95°C, 40 cycles of 15 sec at 95°C, 15 sec at 60°C and 15 sec at 72°C. The specificity of each amplicon was confirmed by melting curve analysis after 40 cycles and agarose gel electrophoresis. Three biological replicates for each sample were performed with three technical replicates for each biological replicate. Quantification data were analyzed according to the manufacturer's instructions. Results were normalized to the expression of actin RNA.


We thank Dr. Harry J. Klee and Ms. Terry A. Davoli for their critical reading of the article. We would also like to thank the anonymous reviewers for their valuable comments and suggestions to improve the article. This research was initially supported by the National Science Foundation to W.Y.S.