Rice WRKY13 Regulates Crosstalk between Abiotic and Biotic Stress Signaling Pathways by Selective Binding to Different cis -Elements 1

Plants use a complex signal transduction network to regulate their adaptation to the ever-changing environment. Rice WRKY13 plays a vital role in the crosstalk between abiotic and biotic stress signaling pathways by suppressing abiotic stress resistance and activating disease resistance. However, it is not clear how WRKY13 directly regulates this crosstalk. Here, we show that WRKY13 is a transcriptional repressor. During rice response to drought stress and bacterial infection, WRKY13 selectively bound to certain site- and sequence-specific cis -elements on the promoters of SNAC1 , the overexpression of which increases drought resistance, and WRKY45-1 , the knockout of which increases both bacterial disease and drought resistance . WRKY13 also bound to two cis -elements of its native promoter to autoregulate the balance of its gene expression in different physiological activities. WRKY13 was induced in leaf vascular tissue, where bacteria proliferate, during infection, and in guard cells, where transcriptional factor SNAC1 enhances drought resistance, during both bacterial infection and drought stress. These results suggest that WRKY13 regulates the antagonistic crosstalk between drought and disease resistance pathways by directly suppressing SNAC1 and WRKY45-1 and autoregulating its own expression via site- and sequence-specific cis -elements on the promoters of these genes in vascular tissue and guard cells. ( P SNAC1 ) or WRKY13 ( P WRKY13 ) promoter. Mock, mock-transformation of rice calli carrying P SNAC1 :GUS or P WRKY13 :GUS . GUS activity was determined by measuring the amount of 4-methylumbelliferone (Mu) produced under the catalysis of GUS in 1 mg total protein per minute. Each data point represents mean (three biological replicates with each replicate containing approximately 20 g calli) ± standard deviation. C, (B), and WRKY45-1 (C) promoters under different physiological conditions analyzed by electrophoretic mobility shift assay. The sequences of DNA probes harboring different cis element (underlined) and their corresponding point-mutated probes are presented. WBOX1 and WBOX2 are cis -elements on the SNAC1 promoter; WBOX-C, PRE4, and PRE2 are cis elements on the WRKY13 promoter; WBOXa and WBOXc-C are cis -elements on the WRKY45-1 promoter. Samples were from rice variety Mudanjiang 8 that was untreated (control), 2 d after inoculation with Xoo (strain PXO61), and 2 d after drought stress (withholding water). The cis -elements WBOX2 and PRE2 are negative controls. +, plus 100 ng anti-WRKY13 antibody; –, without using anti-WRKY13 antibody.

7 own gene expression in rice. It is also unclear which cis-elements on the promoter of WRKY45 are essential for WRKY13 binding during rice responses to abiotic and biotic stresses.
To address these questions, we analyzed the interactions of WRKY13 with candidate promoters in rice and the tissue-and cell-specific expression patterns of WRKY13 in abiotic and biotic stress responses. Our results suggest that WRKY13 is a negative regulator of transcription. It transcriptionally regulates rice responses to both drought stress and bacterial infection by specifically binding to certain cis-acting elements on the promoters of the drought resistance-related gene SNAC1, defenseresponsive gene WRKY45-1, and its own gene.

Abiotic Stresses Influenced WRKY13 Expression
WRKY13 (also named EI12I1 for its cDNA) is transcriptionally induced in ricepathogen interactions. The increased transcript level is closely associated with the presence of a major disease resistance gene Xa3/Xa26 against Xoo and positively correlated with the transcript level of Xa3/Xa26 (Wen et al., 2003;Cao et al., 2007;Cai et al., 2008). To determine whether WRKY13 was also transcriptionally responsive to abiotic stress, its expression in rice varieties Mudanjiang 8 and Minghui 63 was examined after drought and high salinity (150 mM NaCl) stresses. WRKY13 expression was significantly suppressed (P < 0.05) at 12 h to 5 d after drought stress in both rice varieties ( Fig. 1, Supplemental Fig. S1). Salt stress also significantly suppressed (P < 0.05) WRKY13 at 3 to 36 h after treatment in the two rice varieties (Fig. 1, Supplemental   Fig. S1). These results suggest that WRKY13 is involved in rice responses to abiotic stresses. The suppressed expression of WRKY13 in response to salt stress in Mudanjiang 8 is consistent with the phenotype of WRKY13-transgenic plants after stress.
Overexpression of WRKY13 in the genetic background of Mudanjiang 8 resulted in growth retardation of these transgenic plants after salt stress (Qiu et al., 2008).

WRKY13-Transgenic Plants Showed Altered Response to Drought Stress
To test the putative role of WRKY13 in drought stress, we examined the 1 with the genetic background of japonica rice variety Mudanjiang 8); the survive rates of the two WRKY13-RNAi lines were 19% and 20% compared to 14% for the corresponding control (wild-type 2 with the genetic background of indica rice variety Minghui 63; Fig. 2A). We then measured water loss rates of detached leaves from the transgenic plants. Consistent with the phenotypes of these transgenic plants, WRKY13oe plants had a significantly higher (P < 0.05) water loss rate than the corresponding control at 1 to 1.5 h after detaching the leaves, whereas WRKY13-RNAi plants had a significantly (P < 0.05) lower water loss rate than the corresponding control at 0.5 to 2 h after detaching the leaves (Fig. 2B). These results suggest that WRKY13 negatively regulates rice response to drought stress.

WRKY13 Functioned as a Transcriptional Repressor
Rice SNAC1 transcription factor enhances drought resistance by promoting stomatal closure (Hu et al., 2006). WRKY13-oe plants showed markedly suppressed expression of SNAC1 without pathogen infection and in rice−Xoo interaction (Qiu et al., 2008). The present study showed that WRKY13-RNAi plants have significantly enhanced (P < 0.05) expression of SNAC1 compared to control plants without infection ( Fig. 3A). The opposite expression patterns of SNAC1 in WRKY13-oe and WRKY13-RNAi plants as compared to its expression in control plants suggest that WRKY13 negatively and transcriptionally regulates SNAC1.
To further examine whether WRKY13 was a negative transcriptional regulator, we analyzed the interaction of WRKY13 protein and the SNAC1 promoter by transient expression of WRKY13 driven by the constitutive 35S promoter in rice calli generated from transgenic lines carrying a SNAC1 promoter-reporter (β-glucuronidase, GUS) gene (P SNAC1 :GUS) construct via Agrobacterium-mediated transformation. The expression of GUS driven by the SNAC1 promoter in the calli was rapidly suppressed (2 h after transformation of WRKY13) compared to that in the calli immediately before the transformation (0 h, Fig. 3B). The GUS expression in the transformed calli reached the lowest level at 12 h after transformation of WRKY13. However, the expression of GUS 9 in the mock-transformed (control) calli was maintained at a relatively high and stable level throughout the time courses examined. The GUS level in the control calli was more than 6-fold higher than that in WRKY13-transformed calli at 12 h after transformation.
WRKY13 protein can also bind to its native promoter in vitro (Qiu et al., 2007;Cai et al., 2008). To ascertain whether WRKY13 could also suppress its own expression, we transiently expressed WRKY13 in rice calli carrying a P WRKY13 :GUS construct via Agrobacterium-mediated transformation. The expression of GUS driven by the WRKY13 promoter in the calli was markedly suppressed compared to that in untransformed calli (ck, Fig. 3B). Mock-transformation (control) did not markedly influence GUS expression driven by the WRKY13 promoter. The GUS level in the control calli was approximately 5-fold higher than that in WRKY13-transformed calli at 12 h after transformation.
WRKY13 can also bind to WRKY45-1 promoter in rice (Tao et al., 2009). To examine whether WRKY13 could also suppress WRKY45-1 expression, we transiently expressed WRKY45-1 promoter-reporter (green fluorescence protein, GFP) gene (P WRKY45-1 :GFP) in rice calli overexpressing WRKY13 via Agrobacterium-mediated transformation. The transcript level of GFP driven by the WRKY45-1 promoter in the calli was increased at 12 h after transformation (Fig. 3C). However, the GFP transcript level in the wild-type calli (control) was significantly higher (P < 0.05) than that in WRKY13-oe calli. These results suggest that WRKY13 is a transcriptional repressor that not only suppresses SNAC1 and WRKY45-1 but perhaps it own gene as well.

SNAC1, WRKY13, and WRKY45-1 Promoters during Abiotic and Biotic Stresses in Vivo
Sequence analysis revealed that the promoter region of SNAC1 harbored five cis-acting elements putative for WRKY protein binding, the W-like boxes. The five boxes include two "TTGACT" type W-like boxes (named WBOX1 and WBOX3 according to their locations in the promoter region) and three "TGACC" type W-like boxes (named WBOX2, WBOX4, and WBOX5; Fig. 4A chromatin immunoprecipitation (ChIP) assays using anti-WRKY13 antibody. Four segments of the SNAC1 promoter, which harbor W-like boxes, were analyzed by quantitative PCR before immunoprecipitation (input) or after immunoprecipitation (IP) with or without anti-WRKY13 antibody (Fig. 4A, Supplemental Fig. S2). The IP percentage (IP%), which was the percentage of PCR product from IP relative to that from input, was used to evaluate the binding intensity of WRKY13 to the target DNA segment. After immunoprecipitation with anti-WRKY13 antibody, we detected 4.6-, 3.4-, and 7.7-fold increased IP% of the DNA segment harboring WBOX1 from untreated, Xoo-inoculated, and drought-stressed samples, respectively, as compared to their corresponding controls (samples after precipitation without anti-WRKY13 antibody; Fig. 4A). Compared to untreated sample, there was less of an increase in IP% of the DNA segment harboring WBOX1 in the Xoo-inoculated sample and more of an increase in IP% of this segment in the drought-stressed sample. No significant or obvious increased IP% of the DNA segments harboring WBOX2, WBOX3, or WBOX4 and WBOX5 was detected; thus, WBOX2, WBOX3, WBOX4, and WBOX5 could serve as controls in this assay. WBOX1 and WBOX3 are the same type of cis-elements, but with different flanking sequences (Supplemental Fig. S2), indicating the binding specificity of WRKY13 to the DNA segment harboring WBOX1. These results suggest that WRKY13 binds to the promoter of SNAC1 in vivo; it may preferentially bind to the WBOX1 cis-element of the SNAC1 promoter both with and without abiotic and biotic stresses.
WRKY13 can bind to its native promoter at two sites, a novel cis-element PRE4 (TACTGCGCTTAGT) and a DNA probe harboring the complementary strand (TCTAGAACGTCAAATAAAA; named WBOX-C in this paper) of a W-like box WBOX (TTTTATTTGACGTTCTAGA; W-like box is underlined) as analyzed using electrophoretic mobility shift assay (EMSA;Qiu et al., 2007;Cai et al., 2008). To determine whether WRKY13 protein physically bound to its native promoter (Supplemental Fig. S3) in vivo, we performed ChIP assays using the same samples used for studying the interaction of WRKY13 and the SNAC1 promoter. After immunoprecipitation with anti-WRKY13 antibody, we detected 1.8-and 4.2-fold increased IP% of the DNA segments harboring PRE4 and WBOX-C from untreated samples, respectively, as compared to their corresponding controls (Fig. 4B). Activation of WRKY13 suppressed the expression of the defense-responsive gene WRKY45-1, and this suppression was reduced after Xoo infection (Qiu et al., 2009).
In addition, WRKY13 protein can bind to the WRKY45-1 promoter in vivo without stress (Tao et al., 2009). To ascertain whether WRKY13 also bound to site-specific W or W-like boxes on the WRKY45-1 promoter during rice responses to abiotic and biotic stresses, four segments of the WRKY45-1 promoter, which harbored W-like boxes Increased IP% of the DNA segments harboring WBOXa and WBOXc-C was also detected in Xoo-inoculated and drought-stressed samples. However, compared to the untreated control, there was less of an increase in IP% of the DNA segments harboring WBOXa and WBOXc-C in treated samples. These results suggest that WRKY13 preferentially binds to W-like boxes other than the W box on the WRKY45-1 promoter under both stressed and unstressed conditions.

Gene Expression during Abiotic and Biotic Stresses
To determine whether the positive segments for WRKY13 binding in the promoter regions of target genes identified by ChIP assays were due to known ciselements, we performed EMSA. Total nuclear protein samples from untreated control, proteins from both untreated and treated plants bound to the probe harboring WBOX1 but not to the probe harboring WBOX2 from the SNAC1 promoter (Fig. 5A). The protein binding to WBOX1 probe was reduced or abolished by adding anti-WRKY13 antibody. The binding of protein from drought-stressed and Xoo-inoculated plants was abolished by mutating the "TTGACT" sequence of WBOX1 element in the probe to "TTGACC" (M7 probe) that was a typical W box. However, protein from untreated plants showed a weak binding to M7 probe, and this binding was abolished by adding anti-WRKY13 antibody (Figs. 5A, C). Consistent with the results from ChIP assays ( Fig. 4A), there was more protein from untreated and drought-stressed plants bound to the WBOX1 probe compared to the binding intensity of the protein from Xoo-inoculated plants (Fig. 5A). These results suggest that the binding protein is primarily WRKY13, and WRKY13 specifically and preferentially binds to the W-like box WBOX1 on the SNAC1 promoter under different physiological conditions. The nuclear proteins also bound to DNA probes harboring cis-element WBOX-C and PRE4, respectively, but not PRE2 on the WRKY13 promoter (Fig. 5B). Anti-WRKY13 antibody reduced or abolished this binding. Mutation of WBOX-C (M8 probe) or PRE4 (Mc probe) abolished this binding, which is consistent with previous reports (Qiu et al., 2007;Cai et al., 2008). Consistent with the results from ChIP assays (

13
To determine whether the differential binding of WRKY13 protein to different cis-elements affected transcriptional regulation of these promoters, we constructed five site-mutated promoters, in which the core sequences of cis-elements WBOX1 on the SNAC1 promoter, PRE4 and WBOX-C on the WRKY13 promoter, and WBOXa and WBOXc-C on the WRKY45-1 promoter were mutated in the same way as DNA probes used for EMSA assays (Fig. 6, Supplemental Fig. S5). Four site-mutated promoters, in which the core sequences of WBOX2 and WBOX3 on the SNAC1 promoter, PRE2 on the WRKY13 promoter, and WBOXb on the WRKY45-1 promoter were mutated, were used as negative controls (Fig. 6, Supplemental Fig. S5). Rice callus has been proved to be a reliable material for studying rice−Xoo interaction (Yuan et al., 2011), and it is also frequently used for studying abiotic stresses, such as salt and water stresses (Ahmad et al., 2007;Wani et al., 2010;Maragathamani and Khurana, 2012). These mutated promoters were fused with GFP and transiently expressed in the calli generated from the WRKY13-oe line and wild-type Mudanjiang 8. Its expression was also slightly induced in rice calli after Xoo treatment (Fig. 6). However, mutation of WBOX2 (P SNAC1△298G ) or WBOX3 (P SNAC1△645G ) on the SNAC1 promoter did not change the transcript level of GFP. The GFP transcripts was also significantly increased (P < 0.05) after mutation of WBOXa (P WRKY45△30G ) and WBOXc-C (P WRKY45△408C ), but not WBOXb (P WRKY45△93G ), on the WRKY45-1 promoter ( suppress SNAC1 through the WBOX1 element and suppress WRKY45-1 through the WBOXa and WBOXc-C elements. Mutation of WBOX-C (P WRKY13△541T ), but not PRE2 (P WRKY13△371G ), on the WRKY13 promoter also resulted in a significantly increased (P < 0.01) GFP transcript level in untreated and Xoo-or drought-treated calli compared to the GFP transcript level regulated by the native promoter (P WRKY13 ; Fig Cai et al., 2008). In addition to WRKY13, several other nuclear proteins can also bind to PRE4, but with different binding core sequences (Cai et al., 2008). These results suggest that WRKY13 may suppress its own gene expression via the WBOX-C element and it may cooperate with other proteins to maintain or increase its gene expression via PRE4 in rice−Xoo interactions.

WRKY13 Was Transcriptionally Induced in Vascular Tissue and Guard Cells during Xoo Infection and Drought Stress
To ascertain in which tissue WRKY13 regulated the crosstalk between abiotic and biotic stress signaling pathways, we expressed GFP driven by the promoter of WRKY13 (P WRKY13 ) in rice variety Mudanjiang 8. Without biotic and abiotic stress (control), a strong GFP signal was observed in various tissues (i.e., callus, node, stem, collar, ligule, root, pistil, lemma, and palea) in the transgenic plants carrying P WRKY13 :GFP, but no obvious GFP signal was observed in the leaf tissue (Fig. 7A). Two hours after inoculation of Xoo, a detectable GFP signal was observed in the vascular tissue of leaf immediately next to the infection site, whereas no obvious GFP signal difference was observed in other tissues compared to the tissues from uninfected plants.
However, 2 h after drought stress, the GFP expression pattern in various tissues showed no obvious difference from that in control plants (Fig. 7A through hydathodes or wounds and spreads via the vascular system. Thus, the tissuespecific induction of WRKY13 by Xoo infection is associated with its function in bacterial resistance. Enhanced SNAC1 expression specifically in guard cells is associated with increased stomatal closure in SNAC1-mediated drought resistance (Hu et al., 2006). To ascertain whether WRKY13 had a similar cell-specific expression pattern as SNAC1 in rice response to drought stress, we analyzed GFP expression in guard cells of transgenic plants carrying P WRKY13 :GFP. No GFP expression was detected in untreated (control) plants, whereas a strong GFP signal was observed in guard cells 2 h after drought stress ( Fig. 7B). Expression of P WRKY13 :GFP in guard cells was also observed at 30 min after Xoo infection. These results suggest that expression of WRKY13 in guard cells is specifically induced by both biotic and abiotic stresses.

DISCUSSION
WRKY13 plays multiple roles in rice biological activities. It is involved in the regulation of antagonistic crosstalk between the pathways of abiotic stress response and biotic resistance, and it reduces rice plant height and delays flowering time (Qiu et al., 2007(Qiu et al., , 2008. Although WRKY13 possesses the structural characteristics of WRKY-type transcription factors and has DNA-binding ability, its transcriptional regulation function has not been examined previously (Qiu et al., 2007(Qiu et al., , 2009Cai et al., 2008;Tao et al., 2009). The present results suggest that WRKY13 is a transcriptional repressor. This conclusion is also supported by a previous report that WRKY13 preferentially bound to the promoters of genes whose expression was downregulated in WRKY13-oe lines analyzed by yeast one-hybrid assays (Qiu et al., 2009). The transcriptional repressor function of WRKY13 appears to be associated with its direct binding to site-and sequence-specific W-like box-type and PRE4 cis-elements on the promoters of genes functioning in abiotic and biotic stress responses and on its native promoter (Fig. 8).

WRKY13 Specifically Binds to Multiple Types of cis-Elements in Rice
WRKY-type transcription factors are well known to bind to W and W-like boxes (Eulgem et al., 2000;Eulgem and Somssich, 2007;Yuan and Wang, 2012 several studies have revealed that some WRKY proteins can bind to non-W or non-Wlike box cis-elements (Sun et al., 2003;Cai et al., 2008;van Verk et al., 2008). For example, a sucrose-regulated barley WRKY transcription factor, SUSIBA2, can bind to both W box and a sugar-responsive element (Sun et al., 2003). Rice WRKY13 was shown to bind to cis-element PRE4 that is not the sequence homolog of W and W-like boxes in yeast cells and in vitro (Cai et al., 2008). In addition, WRKY13 appears to bind to the complementary strand (WBOX-C: GTCAA) of a W-like box (WBOX: TTGAC) on its native promoter in vitro (Qiu et al., 2007(Qiu et al., , 2008. WRKY proteins in plants can bind to W box core sequence TGAC or its complementary sequence GTCA (Rushton et al., 1996). The present results further suggest that WRKY13 can bind to PRE4 and Although WBOX1 and WBOX3 are the same type of W-like boxes on the SNAC1 promoter (Supplemental Fig. S2), WRKY13 only bound to WBOX1. This selective binding of WRKY13 on the SNAC1 promoter may be due to the fact that WBOX1 and WBOX3 have different flanking sequences. Ciolkowski et al. (2008) reported that the W-box consensus alone is insufficient for the binding of WRKY proteins, and additional neighboring nucleotides or space between adjacent W-box elements also contribute to high-affinity binding. This explanation is also supported by the evidence that mutation of WBOX1 to TTGACC (probe M7), the typical W box, abolished WRKY13 binding to this probe under Xoo-inoculated or drought-stress conditions (Fig. 5A) different flanking sequences. These results suggest that WRKY13 can specifically bind to at least five types of cis-elements in vivo (Fig. 8). Furthermore, WRKY13 binding has DNA site specificity, which may be at least partly due to flanking sequence sensitivity.

WRKY13 Differentially Regulates Abiotic and Biotic Stress Responses by Selectively Binding to Different cis-Elements of Target Genes
Although it is a transcriptional repressor, WRKY13 enhances biotic resistance and decreases abiotic stress resistance (Fig. 2;Qiu et al., 2007Qiu et al., , 2008. The opposite functions of WRKY13 appear to be at least partly performed by directly suppressing the expression of two transcription factor genes: WRKY45-1, which decreases both bacterial disease and abiotic stress resistance (Tao et al., 2009(Tao et al., , 2011, and SNAC1, which increases drought resistance (Hu et al., 2006). It does this by binding to site-and sequence-specific cis-elements on their promoters in vascular tissue or guard cells (Fig.   8). This inference is supported by the following evidence. WRKY13 also bound to WBOX1 without abiotic and biotic stresses, suggesting that WRKY13 may help to maintain the balance of SNAC1 expression. However, this balance control may not occur in guard cells, because no P WRKY13 :GFP expression was detected without stress (Fig. 7B). Increased expression of SNAC1 in guard cells is associated with SNAC1-mediated drought resistance in rice (Hu et al., 2006). Xoo infection of rice and rice resistance to Xoo are not associated with stomata. Thus, Xoo- 18 induced rapid expression of P WRKY13 :GFP in guard cells suggests that WRKY13 may suppress SNAC1 function in guard cells to temporarily ensure metabolic resources for a host defense response to pathogen infection. However, the total binding of WRKY13 to the SNAC1 promoter in leaf tissue was reduced in rice−Xoo interactions compared to that under the untreated condition (Figs. 4A, 5A). Further research is required to determine whether the reduced total binding of WRKY13 to the SNAC1 promoter is associated with reallocating its function to guard cells. Drought stress also induced the expression of P WRKY13 :GFP in guard cells. A large amount of WRKY13 protein bound to WBOX1 during drought stress (Figs. 4A, 5A). These results suggest that WRKY13 is the suppressor of SNAC1 not only in rice-bacteria interactions but also in drought stress.
WRKY13 had different binding intensities to the same cis-element in different physiological activities (Figs. 4,5), which may be due to different states of this protein (e.g., phosphorylated, dephosphorylated, or phosphorylation of different amino acid residues) or the coexistence of other proteins or cofactors under different physiological conditions (Yuan and Wang, 2012). This inference is supported by the evidence that mutation of WBOX1 to a typical W box (M7 probe) abolished WRKY13 binding in Xoo and drought stresses, but WRKY13 bound to M7 probe in untreated samples (Fig. 5A, C).

WRKY13 Transcriptionally Autoregulates Its Own Expression
WRKY proteins transcriptionally regulate not only other genes but also their own genes, as seen in parsley WRKY1 (Turck et al., 2004). The in vivo binding of WRKY13 to its native promoter (Fig. 4B), repression of its native promoter by transcriptional activation of WRKY13 (Fig. 1B), and altered expression of WRKY13 by mutation of target cis-elements on its promoter (Fig. 6C, Supplemental Fig. S5C) suggest that WRKY13 also has this autoregulation function. There are two noteworthy aspects of this autoregulation. First, the strong binding to and strong effect on WBOX-C element of WRKY13 without stress (Figs. 4B,5B,6,8) suggest that this protein may help to maintain its own gene expression at a low level in some tissues, such as leaves. After biotic stress, the level of WRKY13 autoregulation of its own gene through WBOX-C was reduced. However, the large amount of WRKY13 binding to WBOX-C after drought stress suggests that this protein may also autoregulate its own gene

CONCLUSIONS
WRKY13 is a transcriptional repressor that functions on the node of the disease and abiotic stress resistance pathways. It directly suppresses the expression of two important genes, SANC1 and WRKY45-1, which are involved in abiotic stress resistance and rice−bacterium interaction, by binding to site-and sequence-specific W-like-type ciselements on the promoters of these genes. WRKY13 also autoregulates its own gene expression, which may be associated with balancing its function as rice plants face varying environments. This study provides further insight into the complicated regulation of the crosstalk between signaling pathways leading to abiotic and biotic stress responses in rice. Furthermore, the present results also suggest that rice callus is a suitable tissue to study environmental defense responses.

Abiotic Stress Treatment
The stress treatments were performed as reported previously (Tao et al., 2011).
In brief, rice plants grown in sandy soil were kept in a greenhouse with light intensity maintained at 12000-14000 lux and with a 14 h light/10 h dark cycle at 25°C until the three-to four-leaf stage. The humidity in the greenhouse was maintained at 50-60%.
For studying WRKY13 expression after abiotic stress, water was withheld from Mudanjiang 8 and Minghui 63 for 4 h to 6 d for drought stress or plants were irrigated with a solution containing 150 mM NaCl for 3 to 12 h for salt stress.
For studying the role of WRKY13 in drought stress, water was withheld from WRKY13-transgenic and control plants growing in the same pot for 6 d (until almost all the leaves in the pot became completely rolled). Plants were then recovered by providing water for 5 to 9 d. The plants with more than 20% green leaves were considered to have survived and others were considered not to have survived. The

Water Loss Rate
Plants were grown in sandy soil in a greenhouse until the three-to four-leaf stage. Only fully expanded leaves were cut for measuring the loss of water. At the indicated time points, water loss rates of detached leaves from the plants were measured by monitoring the fresh weight loss: ([initial fresh leaf weight -leaf weight after water loss]/initial fresh leaf weight] × 100) (Xiang et al., 2008).

Plasmid Construction and Rice Transformation
Two GUS reporter constructs, P WRKY13 :GUS and P SNAC1 :GUS, were generated by ligating the promoter regions of WRKY13 (a 1700-bp fragment located at -1497 to +203; the nucleotide immediately upstream of the translation start codon is numbered as "-1") and SNAC1 (a 1635-bp fragment located at -1371 to +264) amplified using promoterspecific primers (Supplemental Table S1)

Pathogen Inoculation
Rice plants were inoculated with Xoo strain PXO61 at the seedling or booting Rice shoots from seedlings at the three-leaf stage were used for ChIP assay after inoculation with Xoo.

Gene Expression Analysis
Total RNA was isolated from rice leaves using TRIzol reagent (Life technical replicates) for control were determined. The relative transcript level for treatment is presented. Standard deviation represents the variation of three technical replicates.

Agrobacterium-Mediated Transient Expression
Rice calli generated from the seeds of a transgenic line carrying P SNAC1 :GUS and a transgenic line carrying P WRKY13 :GUS were used to study the regulation of WRKY13

Analysis of GUS Activity
Leaf fragments (about 1 cm long) adjacent to the inoculation sites or rice calli were used for analysis of GUS expression. Quantitative analyses of GUS activity were conducted as described previously (Cai et al., 2007). Total protein concentration in the supernatant was quantified with the Bradford assay (Bradford, 1976).  Table S2).

Statistical Analysis
The significance of differences between control and treated samples or between Mudanjiang 8. The calli were untreated (control) or treated with Xoo (strain PXO61) or drought for 2 h. Asterisks indicate that a significant difference was detected between native and mutated promoters within the same treatment at P < 0.01 (**) and P < 0.05 (*).   indicate that a significant difference was detected between the treated plants and untreated control at P < 0.01 and P < 0.05, respectively.  Overexpressing WRKY13 suppressed the transcript level of GFP driven by the WRKY45-1 promoter (P WRKY45-1 ) analyzed by quantitative reverse transcription-PCR.
Asterisks indicate that a significant difference was detected between GFP expression in WRKY13-oe background and WT background at P < 0.01 (**) and P < 0.05 (*). Each data point was obtained by analyzing approximately 5 g calli. and WRKY45-1 (C) analyzed by ChIP assays. Samples were from rice variety Mudanjiang 8 that was untreated (control), 2 d after inoculation with Xoo (strain PXO61), or 2 d after drought stress (withholding water). The quantitative PCR was conducted before immunoprecipitation (input), after immunoprecipitation (IP) using anti-WRKY13 antibody (white bar), or mock immunoprecipitation (without using anti-WRKY13 antibody, black bar). The presented percentage of PCR product from IP is relative to that from input. Bars represent mean (three replicates) ± standard deviation.
Asterisks indicate that a significant difference was detected between the PCR products from IP with and without using anti-WRKY13 antibody at P < 0.01 (**) and P < 0.05 (*). The numbers in the white bars are fold differences compared to samples without using anti-WRKY13 antibody for IP. 1 to 5, cis-elements WBOX1 to WBOX5 on the SNAC1 promoter. The amplification of DNA fragment harboring a non-WRKY13 binding element PRE2 served as sample quantity control. a to d, cis-elements WBOXa, WBOXb, WBOXc-C, and WBOXd on the WRKY45-1 promoter.  The calli were untreated (control) or treated with Xoo (strain PXO61) or drought for 2 h.
Asterisks indicate that a significant difference was detected between native and mutated promoters within the same treatment at P < 0.01 (**) and P < 0.05 (*).  Bars represent mean (three technical replicates) ± standard deviation. The "a" and "b" indicate that a significant difference was detected between the treated plants and untreated control at P < 0.01 and P < 0.05, respectively.   Figure 3. WRKY13 regulated SNAC1, WRKY13, and WRKY45-1 expression. Each data represent mean (three technical replicates) ± standard deviation for gene expression analyzed by quantitative reverse transcription-PCR (A and C). 0 h, immediately after transformation (B and C). A, Suppressing WRKY13 (WRKY13-RNAi) increased SNAC1 transcript level at the booting stage analyzed by quantitative reverse transcription-PCR. The "a" and "b" indicate that a significant difference was detected between WRKY13-RNAi plants and wild-type control at P < 0.01 and P < 0.05, respectively. B, Transient overexpression of WRKY13 in rice calli suppressed GUS expression driven by the SNAC1 (PSNAC1) or WRKY13 (PWRKY13) promoter. Mock, mock-transformation of rice calli carrying PSNAC1:GUS or PWRKY13:GUS. GUS activity was determined by measuring the amount of 4-methylumbelliferone (Mu) produced under the catalysis of GUS in 1 mg total protein per minute. Each data point represents mean (three biological replicates with each replicate containing approximately 20 g calli) ± standard deviation. C, Overexpressing WRKY13 suppressed the transcript level of GFP driven by the WRKY45-1 promoter (PWRKY45-1) analyzed by quantitative reverse transcription-PCR. WT, wild-type Mudanjiang 8; WRKY13-oe, WRKY13-overexpressing line D11UM1-1. Asterisks indicate that a significant difference was detected between GFP expression in WRKY13-oe background and WT background at P < 0.01 (**) and P < 0.05 (*). Each data point was obtained by analyzing approximately 5 g calli. , and WRKY45-1 (C) analyzed by ChIP assays. Samples were from rice variety Mudanjiang 8 that was untreated (control), 2 d after inoculation with Xoo (strain PXO61), or 2 d after drought stress (withholding water). The quantitative PCR was conducted before immunoprecipitation (input), after immunoprecipitation (IP) using anti-WRKY13 antibody (white bar), or mock immunoprecipitation (without using anti-WRKY13 antibody, black bar). The presented percentage of PCR product from IP is relative to that from input. Bars represent mean (three replicates) ± standard deviation. Asterisks indicate that a significant difference was detected between the PCR products from IP with and without using anti-WRKY13 antibody at P < 0.01 (**) and P < 0.05 (*). The numbers in the white bars are fold differences compared to samples without using anti-WRKY13 antibody for IP. 1 to 5, cis-elements WBOX1 to WBOX5 on the SNAC1 promoter. The amplification of DNA fragment harboring a non-WRKY13 binding element PRE2 served as sample quantity control. a to d, cis-elements WBOXa, WBOXb, WBOXc-C, and WBOXd on the WRKY45-1 promoter.   element (underlined) and their corresponding point-mutated probes are presented. WBOX1 and WBOX2 are cis-elements on the SNAC1 promoter; WBOX-C, PRE4, and PRE2 are ciselements on the WRKY13 promoter; WBOXa and WBOXc-C are cis-elements on the WRKY45-1 promoter. Samples were from rice variety Mudanjiang 8 that was untreated (control), 2 d after inoculation with Xoo (strain PXO61), and 2 d after drought stress (withholding water). The cis-elements WBOX2 and PRE2 are negative controls. +, plus 100 ng anti-WRKY13 antibody; -, without using anti-WRKY13 antibody. The calli were untreated (control) or treated with Xoo (strain PXO61) or drought for 2 h. Asterisks indicate that a significant difference was detected between native and mutated promoters within the same treatment at P < 0.01 (**) and P < 0.05 (*).

Figure 8.
Model of how WRKY13 decreases resistance to drought stress and enhances resistance to bacterial pathogen Xoo by regulating the crosstalk between abiotic and biotic signaling pathways by binding to site-and sequence-specific ciselements on the promoters of SNAC1, WRKY45-1, and its own gene. Arrow, promoting gene expression through specific cis-element; inhibiting marker (⊥), inhibiting gene expression through specific cis-element. The thin and thick lines indicate weak and strong effects of WRKY13 protein on different cis-element based on the results presented in Figure 6 and Supplemental Figure S5. The treatments (no stress, Xoo, drought) shown in different colors indicate that WRKY13 protein has variable effects on a cis-element under different physiological conditions based on the results presented in Figures 4 and 6 and Supplemental Figure S5. Other PRE4 binding protein, based on Cai et al. (2008).