Capsicum annuum transcription factor WRKYa positively regulates defense response upon TMV infection and is a substrate of CaMK1 and CaMK2

Plants are constantly exposed to pathogens and environmental stresses. To minimize damage caused by these potentially harmful factors, plants respond by massive transcriptional reprogramming of various stress-related genes via major transcription factor families. One of the transcription factor families, WRKY, plays an important role in diverse stress response of plants and is often useful to generate genetically engineered crop plants. In this study, we carried out functional characterization of CaWRKYa encoding group I WRKY member, which is induced during hypersensitive response (HR) in hot pepper (Capsicum annuum) upon Tobacco mosaic virus (TMV) infection. CaWRKYa was involved in L-mediated resistance via transcriptional reprogramming of pathogenesis-related (PR) gene expression and affected HR upon TMV-P0 infection. CaWRKYa acts as a positive regulator of this defense system and could bind to the W-box of diverse PR genes promoters. Furthermore, we found Capsicum annuum mitogen-activated protein kinase 1 (CaMK1) and 2 (CaMK2) interacted with CaWRKYa and phosphorylated the SP clusters but not the MAPK docking (D)-domain of CaWRKYa. Thus, these results demonstrated that CaWRKYa was regulated by CaMK1 and CaMK2 at the posttranslational level in hot pepper.

T he plant defense response against pathogen infection is triggered by the recognition of pathogen-associated molecular patterns (PAMPs) or effectors 1,2 . When a plant recognizes the PAMP, PAMP-triggered immunity (PTI) activates basal defense such as production of reactive oxygen species (ROS), phytoalexins, and pathogenesis-related (PR) gene expression to prevent pathogen penetration and proliferation 3 . In contrast, when plant resistance (R) proteins recognize effectors, effector-triggered immunity (ETI) activates stronger and faster resistance responses often accompanied by hypersensitive response (HR) cell death 1 . In both basal defense and R gene-mediated resistance, plant mitogen-activated protein kinase (MAPK) cascades play a central role in signal transduction via three-kinase phosphorelay systems composed of the MAPK, MAPK kinase (MKK) and MAPK kinase kinase (MKKK) 4 . Approximately 23 MAPKs, 10 MKKs, and 80 MKKKs exist in Arabidopsis and the cascade combinations of the components make complex signal transduction networking 5 . These cascade events affect various cellular processes such as resistance to biotic or abiotic agents and plant development via phosphorylation or protein-protein interaction with diverse target proteins. However, few target proteins of MAPKs have been identified and regulation of downstream signaling pathway is poorly understood.
WRKY transcription factors have a conserved WRKY domain of about 60 amino acids containing a 'WRKYGQK' heptapeptide and zinc finger-like motif that can bind to W-box (TTGAC[C/T]) in target genes to regulate gene expression upon abiotic/biotic stress conditions 6,7 . In plants, the WRKY family has been identified in almost all the plant species including Arabidopsis (74 members), soybean (197 members), poplar (104 members), sorghum (68 members), barley (more than 45 members), and rice (more than 100 members). WRKY proteins are mainly classified into group I, IIa 1 IIb, IIc, IId 1 IIe, and III based on sequence homology [8][9][10][11] . The WRKY genes from several plant species are highly upregulated by various stimuli such as pathogen infection, wounding stress, and exogenous defense-related plant hormone treatments. WRKYs also participate in defense responses as a positive or negative regulator through regulation of downstream gene expression 7,12 . Furthermore, WRKY proteins are activated or inactivated by MAPK-dependent phosphorylation or the conformational change of binding partner protein. Substrates of MAPKs were enriched in WRKY transcription factors and validated phosphorylation of WRKYs by MAPKs 13,14 . Recently, it was shown that salicylic acid (SA)-induced protein kinase (SIPK), wound-induced protein kinase (WIPK), and tobacco mitogen-activated protein kinase 4 (NTF4) directly phosphorylated Nicotiana benthamiana WRKY8 (NbWRKY8), which is involved in plant innate immunity 14,15 . Interestingly, the interaction of NbWRKY8 with MAPKs is dependent on a MAPK-docking domain (D-domain) and phosphorylation efficiency is affected by serine-proline (SP) clusters located in the NbWRKY8 N-terminal region 15 . In Arabidopsis, MPK3 and MPK6 can phosphorylate AtWRKY33 upon Botrytis cinerea infection and phosphorylated AtWRKY33 activates phytoalexin deficient 3 (PAD3) expression 16 . The NbWRKY8 and AtWRKY33 are classified as group I WRKY family members. Group I members contain conserved D-domain and at least five SP clusters at N-terminal region 17 . Thus, WRKY transcription factors of group I could have a possibility of being substrates of MAPKs.
In barley, MLA10 encodes a nucleotide binding (NB) and leucinerich repeat (LRR) type R protein that recognizes a fungal avirulence A10 (AVR A10 ) effector protein and then localizes to the nucleus. The Coiled-coil (CC) domain of MLA10 interacts with the transcriptional repressors HvWRKY1 or HvWRKY2, and then another unknown WRKY protein might activate defense gene expression 18 . The Arabidopsis genome encodes the Resistant to Ralstonia solanacearum 1-R (RRS1-R) protein which has an N-terminal Toll and interleukin-1 receptor, resistance protein (TIR) NB-LRR domain and C-terminal WRKY DNA binding domain 19 . The sensitive to low humidity 1 (slh1) mutant, which possesses an additional amino acid in the WRKY domain of RRS1-R, exhibits impaired DNA binding affinity and causes activation of defense response and hypersensitive cell death 20 . Thus, WRKY domains might be involved in both basal defense and effector-triggered immunity by R protein.
Hot pepper (Capsicum annuum L.) is important crop. The hot pepper plant contains L gene alleles which encode CC-NB-LRR type resistance proteins and the L gene confers resistance to Tobacco mosaic virus (TMV) by restricting virus spread at the primary infection site 21 . Previously, we verified WRKY and MAPK genes were responding to TMV-P 0 infection via DNA microarray analysis of a hot pepper expressed sequence tag (EST) database 22 . Specifically, some WRKY genes were involved in HR upon TMV-P 0 infection and positively regulated expression of PR genes [22][23][24] . In this study, we performed functional study of CaWRKYa via virus-induced gene silencing (VIGS) and demonstrated CaWRKYa positively regulated PR gene expressions upon TMV-P 0 infection. Furthermore, CaMK1 and CaMK2 directly interacted with and phosphorylated CaWRKYa which contains D-domain and SP clusters at N-terminus. These results indicated that phosphorylation of CaWRKYa by two TMV-P 0 -responsive MAPKs (CaMK1 and CaMK2) could play a role in TMV defense response in hot pepper plant.

Results
Sequence analysis of CaWRKYa. We previously isolated a CaWRKYa cDNA clone from library screening and deduced the amino acid sequence of CaWRKYa, which contains two 'WRKY' domains and thus CaWRKYa was assigned to group I compared with other groups (Fig. 1a). Recently some group I member WRKYs were found to contain MAPK-docking sites (D-domain) and Ser or Thr followed by Pro (SP) clusters at the N-terminus 15,17 . By amino acid analysis, we found CaWRKYa contained a D-domain and five SP clusters at N-terminal region (Fig. 1a). CaWRKYa exhibited 91% and 90% amino acid sequence homology with Solanum chacoense (Sc) WRKY1 and Lycopersicon peruvianum (Lp) WRKY1, respectively (Fig. 1b). LpWRKY1 was phosphorylated by unknown 44 kD and 67 kD protein kinases which are shown to be transiently activated in response to an elicitor-preparation derived from the wilt inducing fungus Fusarium oxysporum lycopersici (E-FOL) 25 . These results suggest that CaWRKYa is a member of group I WRKY and has putative phosphorylation sites.
CaWRKYa is involved in L-mediated resistance upon TMV-P 0 infection. The roles of CaWRKYa in the L gene-mediated resistance against TMV-P 0 infection in hot pepper plants were studied by using VIGS system based on Tobacco rattle virus (TRV) vector using a partial fragment of CaWRKYa containing 3' untranslated region (UTR) (Fig. 2a). The expression level of CaWRKYa was reduced about 67% in TMV-P 0 -infected TRV2-CaWRKYa plants compared with the TMV-P 0 -infected TRV2 control plants (Fig. 2b). To investigate whether silencing of CaWRKYa affects HR cell death upon TMV-P 0 infection trypan blue staining and conductivity measurement were performed. At 3 days post-inoculation (dpi), the HR numbers of CaWRKYa-silenced plants were reduced by about 40% and ion conductivities were decreased by about 30% compared with TMV-P 0 -inoculated TRV2 plants ( Fig. 2c and d). To test whether silencing of CaWRKYa affects L-mediated resistance upon TMV-P 0 infection, we performed RT-PCR analysis of TMV-P 0 -coat protein (CP) gene expression in the local inoculated leaves and systemic non-inoculated upper leaves. In the local inoculated leaves of TRV2-CaWRKYa plants, TMV-P 0 -CP mRNA was still detected compared with TRV2 control plants (Fig. 2d). In the systemic non-inoculated upper leaves, TMV-P 0 -CP was detected only in the CaWRKYa-silenced plants but not in the TRV2 control plants (Fig. 2e). Thus, L-mediated resistance was affected by silencing of CaWRKYa. These results imply that suppression of CaWRKYa expression leads to reduced resistance to TMV-P 0 and CaWRKYa is a component of the L-mediated TMV resistance response in hot pepper.
CaWRKYa positively regulated expression of PR genes and some plant hormone related genes upon TMV-P 0 infection. WRKY transcription factors play a crucial role in regulating multiple defense response genes such as PR genes 9 . We have shown the L-mediated resistance was diminished in hot pepper plants upon TMV-P 0 infection when the expression of CaWRKYa was suppressed in CaWRKYa-silenced plants by VIGS ( Fig. 2d and e). Furthermore, we showed that CaWRKYa bound to the W-box (TTGAC[C/T]) but not to a mutated version in vitro (Supplementary Fig. S1a and b). Following infection with TMV-P 0 , we tested whether expression of PR genes was affected by suppression of CaWRKYa expression. Expressions of CaPR1, 2, 4, 5 and 10 were strongly reduced in the CaWRKYasilenced plants compared to the TRV2 control plants (Fig. 3). To determine whether CaWRKYa activates these PR genes, we checked activity of the CaPR10promoter-GUS reporter when co-expressing CaWRKYa-GFP in tobacco leaves by Agrobacterium-mediated infiltration 24 . The activity of CaPR10pro-GUS was increased about 37% when it was co-transformed with CaWRKYa-GFP compared to GFP ( Supplementary Fig. S1c). These results suggest that CaPR10 might be one of the target genes of CaWRKYa and CaWRKYa might be a positive regulator of other PR genes by binding to the W-box-containing promoter.
Furthermore, gene expression level of CaICS1 was also diminished in CaWRKYa-silenced plants upon TMV-P 0 infection (Fig. 3). In general, PR1, PR2, and PR5 are involved in salicylic acid (SA)dependent signaling pathway and isochorismate mutase 1 (ICS1) encodes one of the major enzymes of SA biogenesis 26 . Non-racespecific disease resistance 1 (NDR1), a major regulator of R genemediated resistance, was also downregulated in the CaWRKYasilenced plant 27 (Fig. 3). Lipoxygenase (LOX) genes are involved in jasmonic acid (JA) biogenesis pathway and associated with plant defense 28,29 . In CaWRKYa-silenced plant, CaLOX2 induction was reduced (Fig. 3). These data demonstrate that CaWRKYa regulates the expression of defense-related genes and signaling pathway genes. These regulations could contribute to the decreased resistance of CaWRKYa-silenced plant to TMV-P 0 infection.   31 . The amino acid sequence analysis between CaMK1 and LeMPK3 or WIPK showed 95% and 94% identity, respectively ( Supplementary Fig. S2). Thus, we speculated that CaMK1 and To elucidate CaMK1 and CaMK2 function in L-mediated resistance, we checked whether CaMK1 and CaMK2 could be induced specifically during HR upon TMV-P 0 infection using qRT-PCR. We first confirmed that CaWRKYa expression level was induced by an avirulent strain, TMV-P 0 , but not by a virulent strain, PMMoV-P 1,2,3 ( Fig. 4a) 32 . The mRNA levels of CaMK1 and CaMK2 were increased within 24 h after TMV-P 0 inoculation ( Fig. 4b and c). Thus, CaMK1 and CaMK2 are responsive genes to TMV-P 0 infection. In particular, CaMK1 gene expression level was highly increased by TMV-P 0 inoculation at 24 h and slightly induced upon PMMoV-P 1,2,3 infection (Fig. 4b). Furthermore, mRNA levels of CaWRKYa and CaMK1 were increased at early time points by exogenous application of salicylic acid (SA), methyl jasmonic acid (MeJA), and ethephon (ET) treatments but CaMK2 mRNA level was not increased ( Supplementary Fig. S3a, c, and d). CaPR1 was used as a positive control for treatments ( Supplementary Fig. S3b) 23 . As CaMK1 expression is induced upon treatment with plant hormones and TMV infection, and CaMK2 expression is only induced upon TMV infection, this suggests that they play different signaling roles in the plant. Furthermore, we showed that CaMK1 or CaMK2 silen-  Fig. S4). We thus postulated that function of CaWRKYa might be linked to the CaMK1 and CaMK2 and regulated by CaMK1/CaMK2 via post-translational modification.
CaMK1 and CaMK2 can interact with CaWRKYa in vivo. CaWRKYa can be classified as a WRKY group I member, having two WRKY domains, and D-domain and SP clusters at the Nterminal region (Fig. 1a). Many substrates of MAPK were identified as transcription factors, such as WRKYs, in the nucleus to regulate gene expression 14,33,34 . To determine whether CaWRKYa is localized in the nucleus, as other WRKY proteins are, a construct expressing GFP-CaWRKYa fusion was transformed into hot pepper protoplasts along with Tsi1-RFP construct as a nucleus-target control 35,36 . Signal of GFP-CaWRKYa predominantly was merged with Tsi1-RFP signal in the nucleus although the CaWRKYa did not contain any obvious nuclear localization signal (NLS) (Fig. 5a, panel i). Thus, CaWRKYa localizes to the nucleus in hot pepper protoplasts. Most of interacting proteins generally co-localize to the same or interactive subcellular space. Consequently, we tested the subcellular localization of CaMK1 and CaMK2. The construct of GFP fused to N-terminus of CaMK1 or CaMK2 was transformed into hot pepper protoplasts by polyethylene glycol (PEG)-mediated transformation with Tsi1-RFP construct as nucleus-localization marker. GFP signals of GFP-CaMK1 and GFP-CaMK2 protein were merged with red signals of Tsi1-RFP but some GFP signals were also detected in the cytosolic region (Fig. 5a, panels ii  and iii). These results demonstrate that CaMK1, CaMK2, and CaWRKYa show nucleus localization, suggesting that they could interact in planta.
To confirm the association between CaWRKYa and hot pepper MAPKs, CaMK1 and CaMK2 in vivo, we utilized the bimolecular fluorescence complementation (BiFC) assay in hot pepper protoplasts with modified BiFC constructs 37 . N-and C-terminal half fragments of yellow fluorescent protein (YFP) were fused to CaMK1, CaMK2, and CaWRKYa and YFP N -fused CaMK1 and CaMK2 were separately expressed in the hot pepper protoplasts along with CaWRKYa-YFP C . When the combination of constructs YFP N -CaMK1 and CaWRKYa-YFP C were transiently coexpressed in hot pepper protoplasts by PEG-mediated transformation, reconstituted YFP fluorescence signal was detected in the nucleus by fluorescence microscopy analysis (Fig. 5b, panel i). The reconstituted YFP fluorescence signal was also detected in the nucleus when the combination of constructs CaMK2-YFP N and CaWRKYa-YFP C were coexpressed in hot pepper protoplasts (Fig. 5b, panel ii). The BiFC constructs of the basic leucine zipper 63 (bZIP63) transcription factor form homodimer in the nucleus. Constructed bZIP63-YFP N and bZIP63-YFP C were used as a positive control for BiFC assay (Fig. 5b, panel iii) 37 . To confirm reconstituted YFP fluorescence signals in the nucleus, coimmunoprecipitation (co-IP) assays were performed in hot pepper protoplasts with CaWRKYa and CaMK1/CaMK2. As in Fig. 5c, CaWRKYa co-IPs with both CaMK1 and CaMK2 in vivo. These results suggest that CaWRKYa could associate with CaMK1 and CaMK2 in the nucleus.

CaMK1 and CaMK2 phosphorylated the SP clusters of CaWRKYa.
To test whether CaMK1 and CaMK2 are catalytically active in vitro, the MBP-CaMK1 and MBP-CaMK2 fusion proteins were purified from E. coli (Supplementary Fig. S5a). Next the proteins were subjected to an in vitro auto-phosphorylation activity assay. Both MBP fusion recombinant CaMK1 and CaMK2 proteins exhibited auto-phosphorylation activity (Fig. 6a). To test if CaMK1 and CaMK2 are able to trans-phosphorylate the target protein the myelin basic protein (MyBP), which is a commonly used model protein substrate of MAPKs, was used as a substrate. MyBP was phosphorylated by CaMK1 and CaMK2 in the presence of [c-32 P]-ATP (Fig. 6b). Thus, recombinant CaMK1 and CaMK2 could be verified as catalytically active protein kinases that can phosphorylate MyBP as a substrate. Former results clearly demonstrated that CaMK1 and CaMK2 interacted with CaWRKYa in vivo (Fig. 5b and c). To assess the role of interaction between hot pepper MAPKs (CaMK1 and CaMK2) and CaWRKYa, we focused on the possibility of CaWRKYa as a substrate of CaMK1 and CaMK2. We expressed CaWRKYa N fused with MBP protein to test the possibility that CaWRKYa acts as a substrate of CaMK1 and CaMK2 protein kinases (Fig. 6c). When affinity-purified MBP-CaWRKYa N protein was incubated with each of CaMK1 and CaMK2 in the presence of [c-32 P]-ATP, both protein kinases could phosphorylate CaWRKYa N in vitro (Fig. 6d), indicating that CaWRKYa is indeed a substrate of CaMK1 and CaMK2.
Some group I WRKY proteins contain the conserved D domain and SP clusters at the N-terminal region 15 and CaWRKYa also has these domains (Fig. 1a). The D-domain is known to be important for the interaction of a substrate with MAPKs and phosphorylation of SP clusters [38][39][40] . To investigate the phosphorylation site of CaWRKYa, we made mutant protein of MBP-CaWRKYa N-terminal region, which was mutated either at the D-domain (mD) or SP clusters (mSP) (Fig. 6c). Interaction with MAP kinase could be achieved through the putative docking domain of CaWRKYa located at the N-terminus. Thus, if we mutate the corresponding region, we could expect significant decrease of enzymatic activity of both MAP kinases. However, WT CaWRKYa N or CaWRKYa N-mD protein were phosphorylated by recombinant CaMK1 and CaMK2 (Fig. 6d).
These findings suggested that CaMK1 and CaMK2 might use nonidentified docking motif in CaWRKYa. To further investigate the role of SP clusters for phosphorylation, a set of SP cluster mutants were constructed and tested for phosphorylation. MBP-fused recombinant CaWRKYa N-mSP protein, which has four mutations in the SP cluster region (Fig. 6c), did not show phosphorylation signal by CaMK1 and CaMK2 (Fig. 6d). Recombinant MBP fused D-domain mutant CaWRKYa N-mD was not affected in phosphorylation by CaMK1 and CaMK2 (Fig. 6d). These data indicated that the Ddomain of CaWRKYa is not required for phosphorylation of CaMK1 and CaMK2 although interaction site is not clear in our experiments. However, multiple SP clusters of CaWRKYa have an important role in phosphorylation of CaWRKYa by CaMK1 and CaMK2.
A previous report suggested that phosphorylation of WRKY proteins affected DNA binding affinity of WRKY proteins to the W-box 15 . To test whether DNA binding affinity of CaWRKYa is enhanced by CaMK1 and CaMK2, we performed EMSA using purified GST fused with full length recombinant CaWRKYa protein with or without MBP-CaMK1 and -CaMK2 (Supplementary Fig. S5a). CaWRKYa incubated with CaMK1 or CaMK2 show slightly enhanced binding affinity to the W-box in vitro ( Supplementary Fig. S5b). On the contrary, CaWRKYa with higher amount of CaMK1 was inhibited in DNA binding activity but CaMK2 still positively affects DNA binding affinity ( Supplementary Fig. S5b). This result indicated that CaMK1 and CaMK2 function might be related to the regulation of stability and/or DNA binding affinity of CaWRKYa.

Discussion
In the tobacco and TMV interaction model system, MEK1, NTF6, WRKY, and MYB are involved in N-mediated resistance to TMV 41 . This suggests that TMV resistance response is regulated by diverse regulators and especially that transcription factor-MAPK cascades might be a major regulation pathway of R-mediated resistance via posttranslational step. We have studied the hot pepper-TMV-P 0 interaction model system and two WRKY proteins that were previously reported as being involved in L-mediated resistance to TMV-P 0 via regulating PR genes expression 23,24 . In this study, we characterized CaWRKYa as a positive regulator in L-mediated resistance to TMV-P 0 by regulating PR genes and SA/JA biogenesis genes. Furthermore, CaWRKYa was identified as a substrate of CaMK1 and CaMK2 in hot pepper. However, it is still largely unknown how other WRKY proteins are regulated by MAPKs or other regulators in hot pepper.
WRKY transcription factors are localized to the nucleus and then bind to W-boxes of target gene promoters to regulate gene expression upon stress conditions. However, WRKYs do not work alone in recognizing the signals from the stimulus. R-mediated signal transduction or MAPK signal cascade could be necessary for relaying the defense signal transduction to WRKYs or other transcription factors 42 . However, some WRKYs could directly regulate R-mediated signal transduction. The slh1mutant was isolated as a single amino acid insertion mutant in the WRKY domain of RRS1-R, and slh1 showed enhanced disease resistance to pathogen infection 20 . Thus, this WRKY protein has evolved the capacity to play a dual role in defense signal recognition and regulation for plant immunity.
LeMPK3 can be activated in Cf-4/Avr4 interaction in tomato and is involved in HR upon pathogen infection 43 . WIPK is activated in N gene-mediated cell death in tobacco upon TMV infection 3 . CaMK2 also belongs to group A and shows high identity to LeMPK2 (95%), NTF4 (94%), and NtSIPK (88%) (Supplementary Fig. S2). LeMPK2 is required for prosystemin-mediated resistance to Manduca sexta 44 . NTF4 is known to be involved in pathogen-induced HR cell death in tobacco 45 . Furthermore, WIPK and SIPK also are activated by TMV and positively regulate N gene-mediated TMV resistance in tobacco 46 . Interestingly, NbWRKY8 was identified as a substrate of www.nature.com/scientificreports SCIENTIFIC REPORTS | 5 : 7981 | DOI: 10.1038/srep07981 WIPK, SIPK, and NTF4. The D-domain of NbWRKY8 is known to be important for the interaction of these MAPKs 15 . Some group I WRKYs contain D-domain and SP clusters at the N-terminal region 17 . AtWRKY33 is also classified as a group I WRKY which contains D-domain and SP clusters. Similarly, AtWRKY33 is a substrate of MPK3/MPK6. However, phosphorylation of AtWRKY33 does not affect DNA binding affinity of AtWRKY33 to the Wbox 16,47 . CaWRKYa is slightly affected in DNA binding affinity by CaMK1 and CaMK2 (Supplementary Fig. S5b). Furthermore, regulation of DNA binding affinity might be dependent on MAPK protein level because CaMK1 and CaMK2 differently regulated CaWRKYa DNA binding activity at higher protein level (Supplementary Fig.  S5b). These results imply that MAPKs might regulate stability and/ or transcriptional activation ability of WRKY proteins via phosphorylation and physical association.
CaWRKYa also contains D-domain and five SP clusters at the Nterminal region (Fig. 1a). We provided evidence that CaWRKYa is substrate of CaMK1 and CaMK2 (Fig. 6d). The SP cluster(s) of CaWRKYa is required for phosphorylation by CaMK1 and CaMK2 although the D-domain of CaWRKYa did not play an important role for phosphorylation by both CaMK1 and CaMK2 (Fig. 6d). Thus, WRKY-MAPK interactions seem to have been conserved in posttranslational regulation system but each specific interaction might need a platform of specific interaction motif of WRKY protein and it is still not clear. For instance, SIPK also phosphorylated NtWRKY1, which is classified as a group I WRKY, and transactivation activity of NtWRKY1 was enhanced although NtWRKY1 does not have the D domain 48 . Arabidopsis mitogen activated protein kinase kinase kinase 1 (MEKK1) can also phosphorylate AtWRKY53 which is classified as group IIIb in vitro 49,50 . Phosphorylated AtWRKY53 showed increased DNA-binding activity in vitro and in vivo 50 . AtWRKY53 contained four SP clusters at the N-terminal region but did not have a D-domain. Recently, CaWRKY58 was characterized as a negative regulator and classified as a group I WRKY. CaWRKY58 contained five conserved SP clusters but did not have D-domain 51 . CaWRKY2, another group I member, was isolated as a pathogen-inducible transcription factor and contained D-domain and five SP clusters although functional study of CaWRKY2 was not carried out 52 . These group I WRKYs should be further investigated in respect to WRKY-MAPK regulation system in hot pepper. So far, these results provide evidence that MAPKs could interact with D-domain of WRKYs but other unknown motifs of WRKYs might be also important for interaction of MAPKs. The mechanism of the regulation of WRKY proteins, especially group I members, via MAPKs cascade still leaves many questions and will be a subject of further investigations.

Experimental procedures
Plant materials and growth conditions. Pepper (Capsicum annuum L.) cultivar Bugang plants were grown on soil in a growth room at 25uC with 16 h light and 8 h dark photoperiod 23 .
Gene expression analysis upon the plant hormones treatments and TMV inoculation. TMV-P 0 (avirulent) and PMMoV-P 1,2,3 (virulent) strains were prepared with virus-containing sap in virus-inoculation buffer (25 mM phosphate buffer, pH 7.0) and inoculated on the 6-week-old hot pepper plant leaves by rubbing with carborundum (Hayashi Chemical). Samples were prepared at the indicated time points. Control plants were also treated with virus-inoculation buffer containing the carborundum only. For SA, MeJA, and ET treatments, 3-week-old plants were sprayed with solutions of 1 mM SA, 100 mM MeJA, and 1 mM ET, respectively. Control plants were sprayed with water containing 0.01% ethanol. Three independent samples were harvested and total RNA was prepared using Qiagen RNA kit. Quantitative RT-PCR was performed with SYBR Green (Kapa Biosystems) in the LightCycler 480 Real-Time PCR System (Roche Applied Science) according to the manufacturer's instructions. The primers used for real-time PCR reactions are listed in Table S1.
Trypan blue staining. HR cell death was detected with trypan blue staining 23 . TMV-P 0 -inoculated leaves of TRV2 and TRV2-CaWRKYa plant were stained with lactophenol-trypan blue solution (10 ml lactic acid, 10 ml glycerol, 10 ml acidic-phenol, 0.02 g trypan blue, and 10 ml water) and chloral hydrate solution (Fluka chemical) was used as a destaining solution.
Subcellular localization determination of CaWRKYa, CaMK1, and CaMK2. The CaWRKYa, CaMK1, and CaMK2 coding regions were fused to the green fluorescence protein (GFP)-coding region at the N-terminal region in 326 GFP vector. The DNA of GFP-CaWRKYa, GFP-CaMK1, and GFP-CaMK2 constructs was introduced into the hot pepper protoplasts by modified polyethylene glycol (PEG)mediated transformation 54 . GFP signal was excited at 488 nm laser and was collected using 495-510 nm bandwidths using an Axioplan 2 imaging fluorescence microscope (Carl Zeiss).
To confirm these BiFC interactions, co-IP was performed with the hot pepper protoplasts co-expressing CaWRKYa-YFP C and YFP N -CaMK1 or CaMK2-YFP N constructs. Briefly, after PEGmediated transformations into protoplasts from the hot pepper plant, the cells were harvested by quick spin. The cells were suspended in extraction buffer (150 mM NaCl, 50 mM Tris-Cl pH 7.5, 5 mM EDTA, 1% Triton X-100, 1 mM DTT, and proteinase inhibitor cocktail) and then total proteins were extracted by sonication. After centrifugation, the supernatants were mixed with HA-agarose beads and incubated at 4uC for 2 h. The samples were run in SDS-PAGE gel and then western blot was performed with ECL kit (Amersham).
In vitro phosphorylation assay. The N-terminal CaWRKYa fragment (CaWRKYa N ) protein and its mutant constructs were fused with maltose-binding protein (MBP) using pMAL vector.
For in vitro phosphorylation assay, MBP fused CaWRKYaN, CaMK1, and CaMK2 were expressed in E. coli and recombinant proteins were purified. MBP-CaMK1 (1 mg) or -CaMK2 (1 mg) were incubated with 1 mg of its substrates myelin basic protein (MyBP) or MBP-CaWRKYaN, -CaWRKYaN-mD, and -CaWRKYa-mSP proteins in reaction buffer (50 mM Tris-HCl pH7.5, 10 mM MgCl 2 , 2 mM MnCl 2 , 1 mM DTT, 50 mM ATP, [c-32 P]-ATP) for 1 h at RT. The reaction was stopped by adding the sample buffer. After heating for 5 min, the samples were run in SDS-PAGE gel and then stained and destained shortly. The gels were dried in gel dryer for 1 h and the signal was detected by BAS reader.
Electrophoretic mobility shift assay (EMSA). GST-fused full length CaWRKYa recombinant protein was purified according to the GST gene fusion system protocol provided by Amersham. EMSA was performed as described 55 . Briefly, double-stranded synthetic Wbox and mutant W-box oligonucleotides were labeled with [c-32 P]-ATP using T4 polynucleotide kinase (TaKaRa). The labeled probes were incubated with purified GST-CaWRKYa protein (5 mg) or GST protein (5 mg) at 4uC for 30 min. DNA and protein complexes were resolved on a 5% polyacrylamide gel in 0.5 X TBE buffer. The gels were dried and the signals were detected by BAS reader.
GUS promoter activity assay. CaWRKYa-GFP construct is used as an effector protein and GFP construct is used as a negative control. CaPR10 promoter-GUS is used as a reporter 12 . Each effector and reporter was co-transformed into tobacco leaves by Agrobacteriummediated infiltration. The leaves were harvested at three days after infiltration and extracted with GUS extraction buffer (50 mM NaHPO 4 pH7.0, 10 mM 2-mercaptoethanol, 10 mM Na 2 EDTA, 0.1% sodium lauryl sarcosine, 0.1% Trition X-100) for GUS activity analysis. The 4-methyl-umbelliferyl-glucuronide (4-MU) as a substrate was used and protein concentrations of the extracts were determined by BSA assay. GUS activity was measured using a Mithras LB940luminolmeter (Berthold Technologies) 23 . The excitation wavelength was 365 nm and the emission wavelength 455 nm.
Electrolyte leakage assay. The electrolyte leakage assay was performed as described previously 23 . Briefly, three leaf discs were taken from one leaf of each TMV-P 0 inoculated plant. The conductivity is measured by portable conductivity meter (Thermo Orion). The error bars show the mean value of the standard deviation (SD) of the replicate samples (n 5 3).