A Conserved Threonine Residue in the Juxtamembrane Domain of the XA21 Pattern Recognition Receptor Is Critical for Kinase Autophosphorylation and XA21-mediated Immunity*

Despite the key role that pattern recognition receptors (PRRs) play in regulating immunity in plants and animals, the mechanism of activation of the associated non-arginine-aspartate (non-RD) kinases is unknown. The rice PRR XA21 recognizes the pathogen-associated molecular pattern, Ax21 (activator of XA21-mediated immunity). Here we show that the XA21 juxtamembrane (JM) domain is required for kinase autophosphorylation. Threonine 705 in the XA21 JM domain is essential for XA21 autophosphorylation in vitro and XA21-mediated innate immunity in vivo. The replacement of Thr705 by an alanine or glutamic acid abolishes XA21 autophosphorylation and eliminates interactions between XA21 and four XA21-binding proteins in yeast and rice. Although threonine residues analogous to Thr705 of XA21 are present in the JM domains of most RD and non-RD plant receptor-like kinases, this residue is not required for autophosphorylation of the Arabidopsis RD RLK BRI1 (brassinosteroid insensitive 1). The threonine 705 of XA21 is conserved only in the JM domains of plant RLKs but not in those of fly, human, or mouse suggesting distinct regulatory mechanisms. These results contribute to growing knowledge regarding the mechanism by which non-RD RLKs function in plant.

kinases are regulated by autophosphorylation of the activation segment, a centrally located loop positioned close to the catalytic center.
In contrast to RD kinases, non-RD kinases typically carry a cysteine or glycine in place of the arginine. We have previously reported that non-RD kinases are associated with the control of early signaling events in both plant and animal innate immunity (7). For example, in humans, non-RD kinases associate with proteins belonging to the Toll-like receptor (TLR) family, which contain leucine-rich repeats in the extracellular domain and Toll-interleukin receptor intracellular domains (8). TLRs recognize pathogen-associated molecular patterns at the cell surface and then activate a common signaling pathway through an association with non-RD kinases to induce a core set of defense responses (9). TLR1, TLR3, TLR5, TLR6, TLR7, TLR8, and TLR9 function through the non-RD interleukin-1 receptor-associated kinase (IRAK1), whereas TLR3 and TLR4 function through the non-RD receptor interacting-protein 1 (10). In plants, RLKs demonstrated to function in mediating innate immunity fall into the non-RD class (10) or are associated with non-RD RLKs (11)(12)(13). These include the Arabidopsis pattern recognition receptors (PRRs), flagellin sensitive 2 (FLS2) (14) and elongation factor-Tu receptor (15), the rice PRRs, such as XA21 (16), XA26 (17), and Pid2 (or called Pi-d2) (18), the barley PRG1 (resistance to Puccinia graminis f. sp. tritici) (19), and the Arabidposis RD RLK BAK1 that associates with the non-RD RLK FLS2 (11). Unlike RD kinases, non-RD kinases do not autophosphorylate their activation segments (6,10,20). Given the demonstrated importance of the non-RD class of kinases in innate immunity, there is great interest in understanding their mode of action.
RLKs contain a juxtamembrane (JM) domain located between the transmembrane and kinase domains. It is now clear that the JM domain can play an important role in regulating the function of kinase. For example, deletion of the JM domain of the epidermal growth factor (ErbB-1) kinase (an RD RLK) results in a severe loss of tyrosine phosphorylation (1). Two conserved tyrosine phosphorylation sites Tyr 605 and Tyr 611 of EphB2 are essential for EphB2 kinase autophosphorylation and biological responses (21,22). Phosphorylation of the JM domain of the type I transforming growth factor-␤ (T␤R-I, an RD RLK) eliminates the binding site for the FKBP12 (12-kDa FK506-binding protein) inhibitor protein, leading to activation of the T␤R-I kinase (23,24). To date, the role of the JM domain in non-RD RLK phosphorylation has not yet been elucidated. The rice PRR XA21 is a non-RD RLK that binds the sulfated peptide, called Ax21 (activator of XA21-mediated immunity) (25,26). XA21 is not autophosphorylated in the activation segment (20). Here we investigate the role of the XA21 JM domain in regulating XA21 autophosphorylation and XA21-mediated immunity. We found that the JM domain is essential for XA21 autophosphorylation in vitro. We identified Thr 705 in the JM domain as an essential site for XA21 autophosphorylation and XA21-mediated innate immunity. We further identified that Thr 705 is an autophosphorylation site and proposed a model for Thr 705 to activate XA21 function. Although the RLKs of most plant species surveyed contain a residue analogous to Thr 705 , this threonine is not critical for autophosphorylation of the Arabidopsis RD RLK BRI1.
We created the stable plant transformation vector pCami-bia1300:pXA21-Myc-Xa21 according to the method described previously (27). Briefly, the DNA fragment encoding the 13-amino acid Myc peptide was synthesized and cloned into the unique DraIII site in pC822 to create plasmid pC822-Myc. The KpnI fragment containing Myc-Xa21 was cloned into pCAM-BIA1300 to create pCambia1300:pXA21:Myc-XA21 (designated pC1300:Myc-XA21). The 1641-base pair DNA fragment covering the region encoding the JM domain was recovered from plasmid pC1300:Myc-XA21 by NcoI digestion and cloned, after sequence verification, into pGEM TM -T Easy vector pre-digested with NcoI, yielding construct pEGM-T:XA21-1641bp. After the 1641-bp sequence was removed from pCAMBIA1300:pXA21:Myc-XA21 by NcoI, the product was designated pC1300:Myc-XA21-8.3kb. The single amino acid mutants, pEGM-T:Xa21-1641bp T705A and pEGM-T:Xa21-1641bp T705E , were obtained by site-directed mutagenesis using pEGM-T:XA21-1641bp as the template DNA and primer pairs for T705A and T705E, respectively, as described above. After the sequences were verified, the DNA fragments of XA21-1641bp containing the single amino acid mutations were recovered by NcoI digestion and cloned into vector pC1300:Myc-XA21-8.3kb separately. After the sequence and orientations were verified by digestion and sequencing, the resulting constructs were designated pC1300:Myc-XA21 T705A and pC1300: Myc-XA21 T705E , respectively.
Yeast Two-hybrid Assay-The Matchmaker LexA two-hybrid system (Clontech) was used for our yeast two-hybrid assay. Briefly, pEGY48/p8op-lacZ (Clontech) yeast cells were co-transformed with the BD and AD vectors using the yeast transformation kit, Frozen-EZ yeast transformation II (Zymol Research), and spread onto plates containing SD/-His/-Leu/-Ura medium. After the plates were incubated for 2-3 days, the clones were picked and streaked onto new SD/-His/-Leu/-Ura plates. The yeast clones were then streaked onto test plates containing SD/Gal/Raf/-His/-Leu/-Ura/ϩX-GalϩBU medium, which allowed LexA reporter selection.
Expression and Purification of GST-fused Proteins-BL21 CodonPlus Escherichia coli cells (Invitrogen) were transformed with the recombinant GST plasmids and grown to an A 600 of 0.6. Transformed bacterial cells in 200 ml of medium were induced with 200 M isopropyl ␤-D-thiogalactopyranoside for 7 h at 30°C. Cells were harvested by centrifugation and resuspended into 1ϫ PBS buffer (125 mM NaCl, 2.5 mM KCl, 10 mM Na 2 HPO 4 , and 2 mM KH 2 PO 4, pH 7.4) supplemented with 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, and 1 mM dithiothreitol. The resuspended cells were lysed by sonication, followed by addition of Triton X-100 to a final concentration of 0.1% and gentle agitation at 4°C for 15 min. The lysate was centrifuged at 13,000 ϫ g for 15 min at 4°C. The supernatant was mixed with glutathione-agarose beads (Sigma) followed by incubation for 1 h at 4°C. After washing twice with 1ϫ PBS buffer containing 0.05% Triton X-100 and once with 0.5ϫ PBS buffer, some of the glutathione-agarose beads bound with proteins were subjected to an immunodetection assay using anti-GST, whereas others were stored on ice or at 4°C for later use.
Stable Transformation of Rice-Rice transformation was performed as described previously (28). The Xa21 gene was isolated from the wild rice Oryza longistaminata (16), most rice cultivars including Kitaake, do not contain Xa21. Thus, we used Kitaake as the transformation recipient rice in our study. Agrobacterium EHA105 was used to infect calli from the rice Kitaake for transformation. Transformants of the rice carrying Myc-XA21, Myc-XA21 T705A , and Myc-XA21 T705E were selected using hygromycin.
PCR-based Genotyping and Development of Homozygous Transgenic Lines-The presence of the transgene in genomic DNA was examined by PCR using primer pairs 5Ј-ATAGCA-ACTGATTGCTTGG-3Ј/5Ј-CGATCGGTATAACAGCAAA-AC-3Ј. There were two major PCR products after 0.8% agarose gel separation. One product was 1.2 kb and was specific to transgenic Myc-XA21 or its site mutants, Myc-Xa21 T705A and Myc-Xa21 T705E . The other product was about 1.1 kb, which was a nonspecific band amplified from rice genomic DNA. The transgenic lines with a segregation ratio of ϳ3:1 (transgene: non-transgene) in the T1 generation were selected to develop the homozygous lines. The homozygous lines were obtained from the T3 generation, of which all the progeny plants contained the target transgene.
RNA Extraction and Reverse Transcription-PCR-Total RNA was extracted from leaf tissue using TRIzol reagent (Invitrogen) and reverse transcribed to cDNA using Moloney murine leukemia virus reverse transcriptase (Invitrogen). PCR analyses were performed using Xa21-specific primer pairs 5Ј-TTGTACACTGTGATATTAAATC-3Ј/5Ј-TCAGAATTCA-AGGCTCCCACCTTC-3Ј, which could amplify the 666-bp cDNA transcript from a 1408-bp genomic DNA sequence with a 742-bp intron of Xa21.
Plant Cultivation, Xoo Inoculation, and Disease Resistance Determination-Rice plants were grown in the greenhouse until 6 weeks of age and transferred to the growth chamber before Xoo inoculation. Bacterial suspensions (A 600 of 0.5) of Xoo strains, PXO99AZ carrying Ax21 (25,26), were used to inoculate rice by the scissors-dip method (16) separately. The disease lesion length and bacterial population accumulated in rice leaves were evaluated according to previously reported methods (16).
For immunodetection of AD-or BD-protein expression in yeast cells, the yeast cells co-transformed with AD and BD vectors were grown in liquid medium (20 g/liter of galactose, 10 g/liter of raffinose, 6.7 g/liter of Difco TM yeast nitrogen base without amino acids (BD Bioscience), and 0.7 g/liter of drop out supplement -Ura/-His/-Leu (BD Bioscience)) to an A 600 of 1.0, and then harvested. The yeast cells were washed twice with cold 1ϫ PBS, and resuspended in 1ϫ Laemmli loading buffer and boiled for 5 min. SDS-PAGE (10% gel) was used to separate the proteins. Western blotting was performed as described previously (29) using anti-LexA (Clontech), anti-XB24 (Pacific Immunology), or anti-XB15 (Pacific Immunology) antibodies.
Immunoprecipitation of Rice Proteins-To immunoprecipitate MycXA21 and its mutant proteins, 10 g of rice leaf tissue was harvested from 5-week-old transgenic plants carrying pC1300:Myc-XA21 or its site mutation proteins pC1300:Myc-XA21 T705A or pC1300:Myc-XA21 T705E . Protein extracts were isolated by grinding leaf tissue in liquid nitrogen and thawing in 20 ml of extraction buffer (20 mM Tris-HCl, pH 7.6, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% Triton X-100, 10 mM ␤-mercaptoethanol, 20 mM NaCl, 1 mM phenylmethylsulfonyl fluoride, 1% protease mixture (Sigma), 10 M leupeptin, and 5% glycerol). After filtering through double layers of Miracloth (Calbiochem) followed by centrifugation twice at 13,000 ϫ g for 15 min at 4°C, the supernatant was mixed with 50 l of agaroseconjugated anti-Myc antibodies (Santa Cruz) and incubated at 4°C for 1 h. The agarose beads were washed four times each in 1 ml of extraction buffer lacking protease inhibitors. Some of the washed beads were boiled, and the proteins were separated by SDS-PAGE followed by immunodetection with anti-Myc (Santa Cruz), anti-XB15, and anti-XB24 antibodies, respectively. The other beads were stored on ice or at 4°C for later use.
In Vitro Kinase Autophosphorylation Assay-The purified GST-or Myc-fused proteins bound to agarose beads were washed twice with kinase buffer (50 mM HEPES, pH 7.4, 10 mM MgCl 2 , 10 mM MnCl 2 , 1 mM dithiothreitol). The autophosphorylation assay was carried out in a 30-l reaction mixture containing 20 l of agarose bead-bound proteins and 15 Ci of [␥-32 P]ATP (PerkinElmer Life Sciences). The reaction was stopped after 1 h by addition of cold (4°C) washing buffer (0.1ϫ PBS containing 2 mM EDTA) and washed five times. The pro-teins bound to agarose beads were separated by SDS-PAGE (10 or 7% gels) after addition of 10 l of Laemmli loading buffer (4 times) and boiling for 5 min. After staining with Coomassie Brilliant Blue G-250 (for the GST-fused proteins), the gel was dried and exposed to x-ray film.
Mass Spectrometry (MS) and Identification of Phosphorylation Sites-LC-MS/MS-based protein identification was carried out at the Genome Center Proteomics Core at the University of California, Davis. Proteins were prepared for MS analysis using standard reduction, alkylation, and tryptic digest procedures (30). Digested peptides were analyzed by LC-MS/MS on a LTQ-FT with a Waters UPLC. Peptides were identified by searching the MS/MS spectra against a data base containing XA21 protein sequence using TurboSEQUEST as provided in BioWorks 3.3 (Thermo Finnigan). All of the spectra were analyzed using the software Scaffold Viewer 2.0 following the way as previously reported (31).
Determination of the Threonine Sites Analogous to Thr 705 of XA21 in Human, Arabidopsis, and Rice RLKs-The amino acid sequences of fly, human, and mouse were from kinase.com (32,33). The transmembrane helices were predicted using TMHMM Serve version 2. Only those containing transmembrane helices were considered RLKs for JM domain analyses. The sequences of rice and Arabidopsis RLKs were obtained from published reports (10). The amino acid residues that were located one (similar as that of Pid2) or three amino acids before the kinase domains of RLKs were sites equivalent to Thr 705 of XA21. The kinase domains of RLKs were determined through PFAM domain research using the simple modular architecture research tool (SMART). The list of RLKs is provided in supplemental Fig. S1.

RESULTS
Both the JM and Kinase Domains of XA21 Are Required for XA21 Autophosphorylation in Vitro-We have previously shown that XA21K690, a truncated XA21 protein containing a truncated JM and the entire kinase domain, possesses kinase activity (20). To determine whether the JM domain is essential for XA21 autophosphorylation, we expressed and purified several variants of the XA21 proteins: XA21K668 (containing the entire JM and kinase domains), XA21K690, XA21K705 (containing the last three amino acids of the JM domain and the entire kinase domain), XA21K668 K736E containing a catalytically inactive mutant (20,34), and XA21JM (containing only the JM domain). These proteins were expressed as GST fusion proteins and used for autophosphorylation assays (Fig. 1A). We found that GST-XA21K668 and GST-XA21K690 undergo in vitro autophosphorylation, as previously demonstrated (20) (Fig. 1B). In contrast, GST-XA21K705 (labeled K705), XA21JM (labeled JM), and GST-K668 K736E (labeled K668 K736E ) all lost the ability to autophosphorylate in vitro. These results indicate that the presence of both the JM and kinase domains of XA21 as well as residue Lys 736 in the kinase domain are required for XA21 autophosphorylation in vitro.
Thr 705 Is Essential for Kinase Autophosphorylation-It was previously reported that the triple sites mutation of the phosphorylation residues, Ser 686 , Thr 688 , and Ser 689 in the JM domain does not affect XA21 autophosphorylation (35). To determine whether the other four candidate phosphorylation sites, Thr 680 , Ser 697 , Ser 699 , and Thr 705 in the JM domain affect(s) XA21 kinase autophosphorylation, we changed these sites into alanine, or the negatively charged aspartic acid or glutamic acid, one by one in GST-XA21K668 and obtained a set of single amino acid mutations of GST-XA21K668. We performed kinase autophosphorylation assays on the E. coli produced proteins of these mutants and found that only one residue, Thr 705 , abolished XA21 autophosphorylation in vitro (E. coli-produced GST-K668 and GST-K668 S686A/T688A/S689A were included as positive controls and the catalytic dead mutant GST-K668 K736E as negative control). As expected, GST-K668 and GST-K668 S686A/T688A/S689A show autophosphorylation, whereas GST-K668 K736E does not (Fig. 2). The mutation of Ser 697 into alanine or aspartic acid slightly reduces XA21 autophosphorylation, whereas the mutation of Thr 680 or Ser 699 enhances XA21 autophosphorylation. More interestingly, both mutations of Thr 705 to alanine and glutamic acid abolished XA21 autophosphorylation (Fig. 2). The same results were obtained from three repeats of kinase autophosphorylation analyses for the purified protein GST-XA21K668 T705A and GST-XA21K668 T705E (Fig. 3A).
To determine whether the XA21 protein purified from rice was affected by T705A and T705E mutations as observed in the E. coli-produced, truncated XA21, we generated transgenic rice lines expressing Myc-XA21 T705A , Myc-XA21 T705E , and the control protein Myc-XA21 in Kitaake background. Because the Kitaake rice cultivar does not contain Xa21 and is susceptible to Xoo pathogens, it was used as a recipient in transformation experiments to study the function of Xa21 (29,36,37). In these lines, the transgenes are under control of the Xa21 native promoter. The (T0) lines were genotyped by PCR, and transgene expression was confirmed by reverse transcription-PCR. We immunoprecipitated Myc-tagged proteins (Myc-XA21, Myc-XA21 T705A , and Myc-XA21 T705E ) from 10 g of rice leaf tissue of the Myc-XA21, Myc-XA21 T705A , and Myc-XA21 T705E transgenic lines, respectively, by using agarose-conjugated anti-Myc antibody. The Myc-tagged proteins were then subjected to kinase autophosphorylation assays. We found that although the Myc-XA21 protein was capable of autophosphorylation, both Myc-XA21 T705A and Myc-XA21 T705E lost this ability (Fig. 3B). These results confirm that Thr 705 is required for XA21 autophosphorylation and the phosphorylation mimic state of Thr 705 cannot retain the autophosphorylation ability of XA21.
Both the XA21 T705A and XA21 T705E Variants Abolish the Interactions of XA21 with Its Binding Proteins in Yeast-We have previously characterized several XA21-binding proteins (XBs), including XB3, an E3 ubiquitin ligase (27); XB10 (OsWRKY62), a WRKY transcription factor (36); XB15, a protein phosphatase 2C, which negatively regulates cell death and XA21-mediated innate immunity (29); and XB24, an ATPase that modulates XA21-mediated innate immune response. 4 To determine whether the XA21 T705A and XA21 T705E variants affect the ability of XA21 to interact with these XBs, we performed yeast two-hybrid analyses. For these experiments, we used XA21K668 T705A and XA21K668 T705E mutants as baits and XBs as preys in a yeast two-hybrid assay. We found that both XA21K668 T705A and XA21K668 T705E lost interaction with XB3, XB10, XB15, and XB24 (Fig. 4). These results demonstrate that Thr 705 is critical for interaction of XA21 with these XBs.   T705A and T705E Abolish the Interaction of XA21 with XB15 and XB24 in Vivo-To determine whether Thr 705 affects the interaction of XA21 with XB15 and XB24 in vivo (XB3 and XB10 were not tested because antibodies are not available), we performed immunoprecipitation experiments using 10 g of rice tissue harvested from each of the Myc-XA21, Myc-XA21 T705A , and Myc-XA21 T705E transgenic lines. These lines are homozygous with a single transgene insertion. Tissue was harvested 18 h after inoculation with Xoo strain PXO99AZ, which carries the sulfated Ax21 pathogen-associated molecular pattern. The anti-Myc antibody was used to immunoprecipitate Myc-tagged XA21 and its variants XA21 T705A and XA21 T705E . The presence of XB15 and XB24 proteins in the precipitates was monitored with anti-XB15 and anti-XB24 antibodies, respectively. The anti-Myc antibody detected a 140-kDa band corresponding to the Myc-XA21 protein in all three tested lines (labeled XA21, T705A, and T705E in Fig. 5, A and B, upper panels). The anti-XB15 antibody detected the XB15 protein (about 68 kDa) in the precipitates prepared from the wild-type XA21 line, but not in those from XA21 T705A or XA21 T705E (Fig. 5A, lower panel). As XB24 can be clearly detected in the immunoprecipitates from XA21 rice tissue without inoculation, 4 we harvested leaf tissue from the three transgenic lines prior to inoculation for co-immunoprecipitation of XB24. Western blotting was performed using anti-Myc and anti-XB24 antibodies. Proteins XA21, XA21 T705A , and XA21 T705E were detected in the precipitates using the anti-Myc antibody, as described previously (27) (Fig.  5B, upper panel). Similar to the results obtained for XB15, the anti-XB24 antibody detected the XB24 protein (about 21 kDa) in precipitates of the wild-type XA21 line, but not in those of XA21 T705A or XA21 T705E (Fig. 5B, lower panel).
These results indicate that Thr 705 is essential for the interaction of XA21 with XB15 and XB24 in vivo. Rice Lines Carrying XA21 T705A or XA21 T705E Are Altered in XA21-mediated Resistance-The discovery that residue Thr 705 is required for binding of the negative regulators, XB15 and XB24 suggests that Thr 705 is critical for XA21-mediated innate immunity. We hypothesized that if Thr 705 was mutated and XA21 could no longer bind the negative regulators in vivo, rice lines expressing these mutants would display enhanced resistance. To test this hypothesis, we assayed transgenic plants carrying Myc-XA21 T705A and Myc-XA21 T705E for resistance to Xoo. Six-week-old plants were challenged with Xoo strain PXO99AZ. We found that all 10 Myc-XA21 T705A and all 12 Myc-XA21 T705E primary transgenic (T0) lines were susceptible to Xoo (Fig.  6). We then developed homozygous transgenic progeny from these lines carrying single copies of Myc-  XA21 T705A or Myc-XA21 T705E and confirmed that Myc-XA21 T705A or Myc-XA21 T705E are expressed well. To validate the T0 results, we assayed these homozygous transgenic progeny of Myc-XA21 T705A or Myc-XA21 T705E by inoculation with PXO99AZ. Myc-XA21 and Kitaake lines were included as controls. We found that both Myc-XA21 T705A (labeled T705A in Fig. 7) and Myc-XA21 T705E (labeled T705E in Fig. 7) were susceptible to PXO99AZ, similar to the Kitaake control. Fig. 6A shows two typical inoculated leaves (2 weeks post-inoculation) from each of the 6-week-old progeny. As shown in Fig. 7A, both Myc-XA21 T705A and Myc-XA21 T705E display disease lesion lengths (about 20 cm) similar to that of Kitaake, which is much longer than that (about 2.5 cm) of Myc-XA21 (labeled XA21 in Fig. 7). To more accurately quantify these results, we measured leaf lesion lengths (Fig. 7B) and bacterial populations (Fig. 7C) at different time points (0, 4, 8, and 12 days) post-inoculation. The leaf lesion lengths from both Myc-XA21 T705A and Myc-XA21 T705E at the 4-, 8-, and 12-day time points are quite similar to that of Kitaake, although much longer than that of Myc-XA21 (Fig. 7B). Consistently, the accumulation of bacterial populations in Myc-XA21 T705A and Myc-XA21 T705E are similar to that of Kitaake but much higher (Ϸ12 times at 12 days) than that of Myc-XA21. These results demonstrate that Thr 705 in the JM domain is critical for XA21 function, in contrast to our hypothesis.
Thr 705 Is an Autophosphorylation Site-Because the Thr 705 residue is essential for XA21 kinase autophosphorylation and required for XA21 function, we were prompted to test whether this residue can be autophosphorylated. To address this question, we performed in vitro phosphorylation assays on the GST-K668 protein as described above using excess unlabeled ATP. The in vitro autophosphorylated GST-XA21K668 protein was separated by SDS-PAGE after the phosphorylation reaction. In parallel, GST-XA21K668 that was not subjected to the in vitro phosphorylation was used as a control and also separated by SDS-PAGE. The GST-XA21K668 bands were excised, digested with trypsin, and subjected to LC/MS/MS analysis. We detected several phosphorylation sites in the XA21 kinase domain (data not shown). But, we could not detect the phosphorylated Thr 705 after the in vitro autophosphorylation treatment. However, when the same in vitro phosphorylation experiment was performed in the presence of XB24 (an ATPase), which binds to XA21 and promotes XA21 autophosphorylation, 4 we detected the phosphorylation of residue Thr 705 (Fig. 8). Because XB24 is not a kinase but an ATPase that promotes XA21 autophosphorylation, these results indicate that residue Thr 705 is an autophosphorylation site.
Threonine Sites Analogous to Thr 705 in XA21 Are Present in Other Plant Receptor-like Kinases-Given the critical importance of Thr 705 in XA21 autophosphorylation and XA21-mediated immunity, we aligned the XA21 JM sequence with JM sequences of five other known plant RLKs (XA26, Pid2, BRI1, elongation factor-Tu receptor, and FLS2) and found that the other five RLKs contain a threonine residue analogous to Thr 705 in XA21 at the corresponding location in the JM domain. These include residue Thr 880 in Arabidopsis BRI1 (brassinosteroid insensitive 1), an RD RLK that mediates the brassinosteroid response (38,39); Thr 867 in Arabidopsis FLS2 (14,40); Thr 709 in Arabidopsis elongation factor-Tu receptor (15); Thr 804 in rice XA26 (17); and Thr 520 in rice Pid2 (18) (Fig. 9A). We next surveyed the genome sequences of rice, Arabidopsis, fly, human, and mouse to determine whether residues analogous to Thr 705 in XA21 were present and correlated with function of these proteins. We found that 92.7% (507 in 549) of rice RD RLKs and 90.1% (318 in 357) of rice non-RD RLKs contain a threonine residue analogous to Thr 705 in XA21 at the corresponding location in the JM domain (Fig. 9B). In Arabidopsis, 94.2% (425 in 451) of RD RLKs and 84.4% (38 in 45) of non-RD RLKs contained an analogous threonine residue (Fig. 9B). In contrast, only one (the HSER protein) out of 66 mouse RLKs and none of the 22 fly or 71 human RLKs surveyed carried this residue (the JM sequences of these RLKs are described in supplemental Fig. S1). Considering the capability of tyrosine autophosphorylation of the RLKs from fly, human, and mouse, we expected that the JM domains of these RLKs may carry a conserved tyrosine residue, instead of a threonine, which is equivalent to Thr 705 of XA21. However, we found no RLKs from fly, human, or mouse that contain a tyrosine residue equivalent to Thr 705 of XA21. These analyses indicate that analogous Thr 705 residues are highly conserved in plant species but not in fly, human, or mouse RLKs.

DISCUSSION
Phosphorylation and dephosphorylation of residues in the JM domain of animal and plant RD RLKs plays an important regulatory role in the signal transduction of these proteins (1)(2)(3)(4)(5). It has been previously reported that XA21 JM amino acids Ser 686 , Thr 688 , and Ser 689 are required to maintain XA21 protein stability (35). Transgenic rice carrying XA21 mutants with either single or triple alanine replacement mutants of these three sites display slightly compromised resistance compared with the wild-type XA21 (35). However, none of these three sites affect XA21 kinase activity to autophosphorylate (35). In this study, we showed that Thr 705 in the JM domain of XA21 is essential for XA21 autophosphorylation (Figs. 3 and 4) and plays an important biological function. This conclusion is further confirmed by in vivo results demonstrating that replacement of Thr 705 either by alanine or glutamic acid disrupts XA21-mediated immunity.
Our results demonstrate that T705E, a mutation that carries a negatively charged, non-transferable carboxyl group (mimicking a constitutively phosphorylated Thr 705 ), has the same effects as T705A, a non-phosphorylatable residue. Both are incapable of transducing XA21-mediated immunity. We also show that phosphorylation of XA21 Thr 705 can be detected in the XB24 ATPase-promoted autophosphorylated XA21K668 (Fig. 8). These results suggest that the Thr 705 is an autophosphorylation site and that the phosphosphoryl group of Thr 705 may be transferable as it is not detectable in the XA21K668 protein subjected to kinase autophosphorylation in the absence of XB24 ATPase. All together, these results suggest that Thr 705 may serve dual roles, to receive and then to donate a phosphoryl group. In other words, after Thr 705 is autophosphorylated, it may then transfer its phosphoryl group to other XA21 residues, which would activate XA21. In this possible model, a glutamic acid cannot replace Thr 705 because it cannot serve as a recipient and donor of the phosphoryl group. It is consistent  with our result that the T705E mutation cannot retain the capability of the XA21 kinase to autophosphorylate. This proposed mechanism is different from most RD kinases (or receptor kinases), whose kinase activation is regulated only by phosphorylation inside the activation segment (6). It is also possible that the T705E mutation may not exactly mimic the stable phosphorylated Thr 705 . If this is the case then the T705E and T705A would be functionally equivalent. In this model Thr 705 would activate XA21 kinase autophosphorylation through another unknown mechanism.
The mechanism of Thr 705 function in activation of the XA21 kinase autophosphorylation is different from that of the conserved tyrosine phosphorylation residues of the JM domains of EphB family RD RLKs. First of all, we have shown that the conserved tyrosine phosphorylation residues (for example, Tyr 605 and Tyr 611 in the JM domain of EphB2) (supplemental Fig. S1) (22) are not equivalent to Thr 705 of XA21. Second, mutation of these tyrosines to phenylalanine impaires kinase activity but mutation to glutamic acid does not disrupt kinase autophosphorylation (21). In contrast, the replacement of Thr 705 with T705E disrupts XA21 kinase autophosphorylation.
In the case of T␤R-I kinase activation, at least three, and perhaps four or five of the serines and threonines in the conserved glycine-serine (GS) region must be phosphorylated by T␤R-II upon the binding of transforming growth factor to the T␤R-I/T␤R-II complex to fully activate T␤R-I (23). Phosphorylation of the GS region activates T␤R-I not by increasing the activity of the kinase domain but rather by converting the GS region into an efficient recruitment motif for the proper Smad substrate. The mode of XA21 kinase activation via Thr 705 is most likely distinct from that of the T␤R-I RD receptor kinase. Thr 705 is also required for the binding of the XA21 negative regulators, XB10, XB15, and XB24, representing another difference between XA21 and T␤R-I regulation.
Kinase activation of most mammalian RD RLKs (e.g. vascular endothelial growth factor receptor, platelet-derived growth factor receptor; fibroblast growth factor receptor, and epidermal growth factor receptor) is regulated only by phosphorylation inside the activation segment (6). Tyrosine sites in the JM domain of these receptors are not required for kinase activation nor do they appear to be equivalent to Thr 705 . It is unlikely, therefore, that substitution of these tyrosine phosphorylation sites with glutamic acid would constitutively activate these RD receptors. These data suggest that the mechanism by which the JM domain regulates the autophosphorylation and biological function of plant RLKs, especially non-RD RLKs, is different from that of mammalian RLKs. Thus, our findings represent new insights into the mechanism(s) of kinase activation of non-RD receptor kinases. Because, no substrates of XA21 have yet been identified, it is not yet possible to study how phosphorylation of Thr 705 affects the affinity of XA21 toward its substrates. The failure of XA21 to autophosphorylate results in the inability to interact with XB3, XB10, XB15, and XB24. This result suggests that a conformational change resulting from XA21 autophosphorylation is required for XA21 to interact with its binding proteins. Alternatively, the phosphorylation of XA21 residues may be essential for XBs binding to XA21. For example, the phosphorylation of residue Ser 697 may be required for the binding of some XBs like XB15 (29).
It is well known that the conserved lysine site, which is located inside the kinase domain and is required for ATP binding for both RD and non-RD kinases, is essential for kinase autophosphorylation. When this conserved site is mutated, the kinase fails to autophosphorylate and loses its biologic function. For example, the Lys 911 residue is essential for the RD RLK BRI to autophosphorylate and mediate the signal of brassinosteroid (39). The Lys 736 residue inside the kinase domain is essential for XA21 to autophosphorylate and confer full resistance to Xoo (20). However, there are no previous reports showing that a single amino acid in the JM domain of either RD or non-RD RLKs is essential for kinase autophosphorylation. Although residues analogous to Thr 705 of XA21 are conserved in both RD and non-RD receptor kinases in planta, functional studies of these sites are very limited. For example, in RD receptor kinases, the only example in which the equivalent site has been studied is the BRI1 protein. Threonine 880 in BRI1 is analogous to Thr 705 of XA21 and is an autophosphorylation site. However, the mutation of Thr 880 to alanine does not affect BRI1 kinase autophosphorylation or biological function (39). In contrast, residue Thr 867 of FLS2, a non-RD RLK, is analogous to Thr 705 in XA21 (Fig. 2), and was shown to be essential for the function of FLS2 (40). Transgenic Ws-0/FLS2 T867V Arabidopsis displays an enhanced disease susceptibility phenotype to Pseudomonas syringae (40), suggesting that residue Thr 867 of FLS2 carries a similar function as that of Thr 705 in XA21, although further study is needed to demonstrate that the Thr 867 of FLS2 is essential for FLS2 autophosphorylation. Thus, for the three plant RLKs studied to date, Thr 705 and Thr 867 have been demonstrated to be critical for the function of non-RD RLKs XA21 and FLS2, respectively, but not Thr 880 for the RD RLK BRI1. Whether autophosphorylation of Thr 705 equivalent sites is critical for other non-RD functions of RLKs remains to be further investigated.
Protein XA21K705, which begins at residue Thr 705 and contains the whole kinase domain, loses XA21 autophosphorylation activity (Fig. 1). In contrast XA21K690, containing 15 more amino acids at the N terminus, maintains the autophosphorylation activity. These results suggest that the amino acids immediately flanking Thr 705 are required for the residue to function. This is consistent with previous reports that the flanking sequence is required for kinase autophosphorylation (41,42). XA21 autophosphorylation is essential for its interaction with XB3, a positive regulator of XA21 function (27). The fact that XA21 autophosphorylation is also essential for XA21 to bind to negative regulators XB10, XB15, and XB24 suggests that autophosphorylation of XA21 plays an autoregulatory role in XA21 function. Similar examples have been reported. For example, phosphorylation of insulin receptor substrate 1 (IRS1) on serine residues has a dual role, either enhancing or terminating the effects of insulin. The activation of protein kinase B in response to insulin propagates insulin signaling and promotes the phosphorylation of IRS1 on a specific serine residue. Insulin also activates JNK (c-Jun NH 2 -terminal kinase), ERK (extracellular signal-regulated kinase), protein kinase C, and mTOR (mammalian target of rapamycin). These kinases phosphorylate IRS1 on other specific sites, yet, in contrast to protein kinase B, these proteins inhibit IRS1 function. Thus, a delicate balance exists between positive and negative IRS1 serine phosphorylation (43). Similarly, our results suggest that XA21 autophosphorylation may lead to differential binding of positive and negative regulators.