The WRKY6 transcription factor modulates PHOSPHATE1 expression in response to low Pi stress in Arabidopsis.

Arabidopsis thaliana WRKY family comprises 74 members and some of them are involved in plant responses to biotic and abiotic stresses. This study demonstrated that WRKY6 is involved in Arabidopsis responses to low-Pi stress through regulating PHOSPHATE1 (PHO1) expression. WRKY6 overexpression lines, similar to the pho1 mutant, were more sensitive to low Pi stress and had lower Pi contents in shoots compared with wild-type seedlings and the wrky6-1 mutant. Immunoprecipitation assays demonstrated that WRKY6 can bind to two W-boxes of the PHO1 promoter. RNA gel blot and beta-glucuronidase activity assays showed that PHO1 expression was repressed in WRKY6-overexpressing lines and enhanced in the wrky6-1 mutant. Low Pi treatment reduced WRKY6 binding to the PHO1 promoter, which indicates that PHO1 regulation by WRKY6 is Pi dependent and that low Pi treatment may release inhibition of PHO1 expression. Protein gel blot analysis showed that the decrease in WRKY6 protein induced by low Pi treatment was inhibited by a 26S proteosome inhibitor, MG132, suggesting that low Pi-induced release of PHO1 repression may result from 26S proteosome-mediated proteolysis. In addition, WRKY42 also showed binding to W-boxes of the PHO1 promoter and repressed PHO1 expression. Our results demonstrate that WRKY6 and WRKY42 are involved in Arabidopsis responses to low Pi stress by regulation of PHO1 expression.


INTRODUCTION
Phosphorus (P), as a major essential nutrient for plant growth and development, serves various basic biological functions in the plant life cycle (Raghothama, 1999). Phosphate (H 2 PO 4 2 , or in short, Pi) is the major form that is most readily taken up and transported in the plant cell (Ullrich-Eberius et al., 1981;Tu et al., 1990). The Pi concentration in the soil, typically 10 mM or less, results in Pi starvation for plant growth and survival, which is one of major limiting factors for crop production in the cultivated soils. A number of studies have shown that plants have evolved different strategies to overcome limited Pi availability. In response to low Pi stress or Pi starvation, plants may increase the Pi uptake from the soil by alteration of root architecture and function (Ló pez-Bucio et al., 2003;Ticconi and Abel, 2004;Osmont et al., 2007). Under Pi-limiting conditions, plants may also increase their Pi acquisition by changing their metabolic and developmental processes , such as increasing phosphatase activity (Lipton et al., 1987) and secretion of organic acids (Marschner, 1995). PHOSPHATE1 (PHO1) has been shown to play roles in Pi translocation from root to shoot (Hamburger et al., 2002), which is also important for plant adaptation to a low Pi environment. A single nuclear recessive mutation in PHO1 led to its inability to load Pi into xylem (Poirier et al., 1991;Hamburger et al., 2002).
PHO1 is predominantly expressed in the stellar cells of the root and the lower part of the hypocotyls and is believed have a role in Pi efflux out of root stellar cells for xylem loading (Hamburger et al., 2002). However, PHO1 shares no homology with any previously described Pi transporter proteins in plants and fungi (Hamburger et al., 2002). It is interesting that PHO1 contains a SPX domain, which can be found in several proteins that are involved in phosphate transport and/or Pi signaling pathways in plants and yeast. For example, an SPX protein in yeast named PHO81 is a key regulator in transporting and sensing phosphate, as well as in sorting proteins to endomembranes (Lenburg and O'Shea, 1996;Wykoff and O'Shea, 2001). In Arabidopsis, the SPX proteins SPX1-SPX3 are involved in Pi signaling pathways and regulate the expression of the Pi transporter genes Pht1;4 and Pht1;5 (Duan et al., 2008). Thus, the possibility cannot be excluded that PHO1 may not be a direct Pi transporter but rather may regulate Pi loading of the xylem either by directly influencing the activity of transporter proteins or via signal transduction.
WRKY proteins are plant-specific transcription factors encoded by a multigene family comprising 74 members in Arabidopsis (http://www.Arabidopsis.org/browse/genefamily/WRKY-Som. jsp), and many of them have been found to play important roles in plant responses to biotic and abiotic stresses. In addition to a number of WRKY genes that have been demonstrated to be involved in plant responses to pathogen infection and other defense-related stimuli (Dong et al., 2003;Kalde et al., 2003;Li et al., 2004;Eulgem and Somssich, 2007), some WRKY genes have also been shown to function in plant responses to various abiotic stress, such as drought (Pnueli et al., 2002;Rizhsky et al., 2002;Seki et al., 2002;Mare, et al., 2004), cold (Huang and Duman, 2002;Seki et al., 2002;Mare, et al., 2004), heat (Rizhsky et al., 2002), salinity (Seki et al., 2002), wounding (Hara et al., 2000), and Pi starvation (Devaiah et al., 2007a). However, little is known about the specific interaction of a given WRKY protein with a defined target gene. Recent studies using an oligodeoxynucleotide decoy strategy have revealed that SUSIBA, a WRKY protein, can bind to SURE (sugar responsive) and W-box elements in the iso1 promoter (Sun et al., 2003). Petroselinum crispum WRKY1 has been shown to bind to the W-box of its native promoter as well as to that of Pc WRKY3 and Pc PR1-1 based on chromatin immunoprecipitation (ChIP) analysis (Turck et al., 2004). Candidate target genes for At WRKY53 were identified with a pull-down assay (Miao et al., 2004), and electrophoretic mobility shift assays identified candidate targets for Hordeum vulgare WRKY transcription factor WRKY38 (Zou et al., 2008), Arabidopsis WRKY26 and WRKY11 (Ciolkowski et al., 2008), Nicotiana tabacum WRKY1, WRKY2, and WRKY4 (Yamamoto et al., 2004), and O. sativa WRKY71 (Zhang et al., 2004).
At WRKY6 was first reported to be associated with senescence-and defense-related processes, and it could activate the expression of its target gene SIRK, a receptor-like protein kinase in the process of senescence (Robatzek and Somssich, 2002). Here, we report a previously unknown function of WRKY6 in plant responses to low Pi stress. We demonstrate that plants overexpressing WRKY6 become more sensitive to low Pi stress and display a similar phenotype as the pho1 mutant. WRKY6 negatively regulates PHO1 expression by binding to two W-box consensus motifs within the PHO1 promoter, and the repression of PHO1 expression by WRKY6 is released under low Pi conditions.

WRKY6 Overexpression Plants Showed Similar Phenotypes as the pho1 Mutant under Low Pi Conditions
The growth of the aerial portion of the Arabidopsis WRKY6overexpressing line (35S:WRKY6-9) was impaired when the plants were grown in a potting soil mixture ( Figure 1A). However, the overall growth of the WRKY6 null mutant wrky6-1, an En-1 insertion mutant, was obviously better than wild-type plants ( Figure 1A). The measured free Pi concentration in the potting soils was ;10 mM in this study. Thus, the plants grown under the described conditions as shown in Figure 1A were actually experienced low Pi stress. Under this low Pi stress condition, the 35S:WRKY6-9 line displayed thinner stalks and smaller leaves compared with wild-type plants ( Figure 1A). Figure 1B shows that the En-1 insertion in wrky6-1 disrupted the transcription of the WRKY6 gene.
Besides the 35S:WRKY6-9 line, two more WRKY6-overexpressing lines, Super:WRKY6-13 and Super:WRKY6-18, were included in our further experiments. The elevated expression of WRKY6 mRNA in these transgenic lines is shown in Figure 1C. Transcription of WRKY6 in either wild-type or the wrky6-1 plants was not detectable in our RNA gel blot experiments. Among the three WRKY6-overexpressing lines, 35S:  displayed the highest, medium, and the lowest WRKY6 expression, respectively.
Plants usually accumulate more anthocyanin in their aerial portions in response to low Pi stress (Marschner, 1995), and this can result in brown-colored leaves. When grown on Murashige and Skoog (MS) medium with sufficient Pi supply, all tested plants showed no difference in their phenotypes (top panel in Figure 1D). After the low Pi treatment, the WRKY6-overexpressing plants, particularly Super:WRKY6-18 and 35S:WRKY6-9 lines (both have much higher WRKY6 expression than does the Super:WRKY6-13 line), displayed dark-brown leaves similar to the phenotype of the pho1 mutant (bottom panel in Figure 1D). To confirm further the effects of WRKY6 overexpression on the low Pi response phenotype, another group of WRKY6-overexpressing lines (35S:WRKY6-3, 35S:WRKY6-5, and 35S:WRKY6-9; Robatzek and Somssich, 2002) were tested. As shown in Supplemental Figure 1 online, overexpression of WRKY6 indeed increased plant sensitivity to low Pi stress. Furthermore, increase of transgenic plant sensitivity to low Pi stress was closely related to WRKY6 expression level (see Supplemental Figure 1 online).
A defect in Pi transfer from root to shoot has been reported in the pho1 mutant (Poirier et al., 1991;Hamburger et al., 2002), resulting in reduced Pi content in the shoot and smaller plant size. Under either Pi-sufficient or Pi-deficient conditions, the WRKY6overexpressing lines showed similar reduced Pi contents in shoots as the pho1 mutant (Figures 2A and 2B). As a result, the ratios of Pi content in shoot to that in root (Pi shoot /Pi root ) for both WRKY6-overexpressing lines and the pho1 mutant were significantly lower than the ratio determined in wild-type plants, particularly under low Pi condition ( Figures 2C and 2D). In addition, four Super:PHO1 lines (Super:PHO1-1, -7, -9, and -13) with differential PHO1 expression were selected for the Pi content assay. As shown in Supplemental Figure 2 online, all Super:PHO1 lines and the wrky6-1 mutant displayed higher Pi contents in shoots, whereas the pho1 mutant showed the lowest Pi content in shoots under both MS and low Pi (LP) conditions. The results demonstrate that Pi content in shoots indeed correlates with PHO1 expression. These data suggest that WRKY6 may play a role in plant responses to Pi starvation at least partially through regulating PHO1-dependent Pi transfer.
It should be noted that the wrky6-1 mutants were obviously growing better than wild-type plants under the low Pi condition (Figures 1A and 1D,bottom panel) in our experiments, although they showed no difference under Pi-sufficient conditions ( Figure  1D, top panel).
However, Somssich (2001, 2002) had not observed phenotype difference between the wrky6-1 mutants and wild-type plants. After we have grown the plants under different environmental conditions, we believe that this difference mainly resulted from growth conditions, particularly light period. In the studies by Somssich (2001, 2002), plants were grown first under short-day conditions followed by long-day periods, while plants were grown under a constant long-day (18 h light) condition in our experiments. When Arabidopsis plants were grown under a short-day condition (10-h light), almost no phenotype difference between wild-type and wrky6-1 plants was observed (as shown in Supplemental Figure 3 online).

WRKY6 Interacts with Two W-Box Motifs of the PHO1 Promoter
To test the hypothesis that WRKY6 regulates PHO1 expression, we first tested whether WRKY6 could bind the PHO1 promoter. It is known that WRKY proteins usually bind to the W-box motifs of their target gene promoters (Eulgem et al., 2000). Analysis of the primary sequence of the PHO1 promoter revealed six W-box consensus motifs within the PHO1 promoter and four of them (named W Q , W X , W Y , and W Z , respectively) are located at the very end of promoter nearing the coding region ( Figure 3A). The in vivo interaction between WRKY6 and the W-box motifs of the PHO1 promoter was investigated using the ChIP-qPCR (chromatin immunoprecipitation quantitative PCR) method. As shown in Figure 3B, WRKY6 strongly interacted with the PHO1 promoter when the primer combinations encompassing either W Y or W Z were applied, while no interaction was observed between WRKY6 and PHO1 promoter containing only W Q or W X box. These results demonstrated that WRKY6, as a transcription factor, can bind to two (W Y and W Z ) W-box motifs within the PHO1 promoter nearing the coding region, suggesting regulation of PHO1 transcription by WRKY6.
(B) RT-PCR test of WRKY6 expression in the wrky6-1 mutant and wild-type seedlings. Seven-day-old seedlings were used for RNA extraction. EF1a was amplified for the control. transcription. To test this hypothesis, we first compared the transcription of PHO1 in the roots of WRKY6-overexpressing lines, wrky6-1, and wild-type plants, since both WRKY6 and PHO1 are highly expressed in roots (Robatzek and Somssich, 2001;Hamburger et al., 2002). As shown in Figure 4A, the transcription of PHO1 in roots was repressed in the WRKY6overexpressing lines. Repression of PHO1 expression was also closely related to the WRKY6 expression levels in WRKY6overexpressing lines, with the strongest repression in 35S: WRKY6-9 plants and the weakest repression in Super: WRKY6-13 plants.
We further tested if WRKY6 binding to PHO1 W-box motifs was required for its function in regulation of PHO1 transcription. Different truncated PHO1 promoter fragments (indicated above each panel of Figures 4B to 4D) driving the b-glucuronidase (GUS) reporter gene were transformed into 35S:WRKY6-9, wrky6-1, and wild-type plants. In wild-type plants, the GUS reporter gene was expressed when driven by PHO1 promoter fragment containing all four W-box motifs (W Q , W X , W Y , and W Z ), two W-box motifs (W Y and W Z ), or no W-box motifs, respectively (Figures 4B to 4D). When all four W-box motifs were deleted from the PHO1 promoter, the expression of the reporter gene in 35S: WRKY6-9 and wild-type roots was dramatically increased, indi-cating the removal of negative regulation ( Figures 4B to 4E). On the other hand, strong GUS expression was detected in wrky6-1 roots regardless of which promoter fragment was used (Figures 4B to 4E). The GUS expression level was much higher in wrky6-1 roots than in wild-type roots when the reporter gene was driven by promoters containing W Y and W Z ( Figures 4B, 4C, and 4E). However, there was almost no difference in the GUS staining among the roots of all three different types of plants when no W-box motif existed ( Figures 4D and 4E). More importantly, in the roots of 35:WRKY6-9 plants expressing the GUS reporter gene driven by W Y -and W Z -containing promoter fragments, only weak GUS staining can be detected ( Figures 4B, 4C, and 4E). The results demonstrate that binding to PHO1 W-box motifs was required for WRKY6 regulation of PHO1 transcription. All these data support the notion that WRKY6 is the negative regulator of PHO1 transcription.

Repression of PHO1 Transcription by WRKY6 Is Removed under Low Pi Stress
Consistent with previous reports (Stefanovic et al., 2007;Ribot et al., 2008), we observed that PHO1 transcription was induced in response to low Pi stress. The PHO1 transcription level in wild-type roots was increased after the plants had been transferred to the low Pi medium for 3 d ( Figure 5B; Ribot et al., 2008). It was further proposed that the low Pi stress might trigger the plant responses through suppression of WRKY6 expression. However, as shown in Figure 5A, after wild-type plants were challenged with low Pi stress, WRKY6 expression level was increased during the first 3 h and then decreased, but stayed above its zero time expression level for ;48 h. Another hypothesis we proposed was that the Pi starvation inhibits WRKY6 functioning in suppression of PHO1 expression, such as through a possible blockage of WRKY6 binding to W-box motifs of the PHO1 promoter or a low Pi-induced WRKY6 protein degradation. ChIP-qPCR experiments were conducted to test whether WRKY6 protein still can bind to W Y and W Z boxes of the PHO1 promoter under the low Pi condition. As shown in Figure 5C, the interaction between the WRKY6 protein and W Y or W Z box of the PHO1 promoter was severely impaired under the low Pi condition. To confirm further the interaction of WRKY6 with the W Y and W Z boxes of the PHO1 promoter, ChIP-qPCR experiments were performed using the wild type, wrky6-1 mutant, and three WRKY6-overexpressing lines. As shown in Figure 6, the strong interaction of WRKY6 with W Y or W Z boxes of PHO1 promoter was displayed again in all three WRKY6-overexpressing lines under the normal conditions (on MS medium), while this interaction was reduced under the low Pi condition (on LP medium).
It was further proposed that low Pi stress may induce degradation of WRKY6 protein so that repression of PHO1 transcrip-tion could be weakened. Protein gel blot analysis was performed using anti-WRKY6 serum in the total proteins extracted from the roots of seedlings grown on the low Pi medium. As shown in Figure 7A, the low Pi treatment induced a time-dependent decrease of WRKY6 protein content. Yeast two-hybrid assays (see Supplemental Table 1 online) showed that WRKY6 interacted with a RING-type finger E3 ligase (At1g74410), indicating that WRKY6 protein degradation may be mediated by the 26S proteosome. Addition of 10 mM MG132, a 26S proteosome inhibitor (Lee et al., 2009), blocked the low Pi-induced decrease of WRKY6 protein ( Figures 7B and 7C), suggesting that a 26S proteasome-mediated WRKY6 proteolysis is involved in WRKY6-regulated PHO1 expression in response to low Pi stress.

WRKY42 Interacts with the PHO1 Promoter and Negatively Regulates PHO1 Transcription
To identify other proteins that interact with WRKY6, we performed yeast two-hybrid experiments using WRKY6 as bait in fusion with the Gal4 DNA binding domain. As listed in Supplemental Table 1 online, there are at least a dozen proteins that interact with WRKY6. Among these WRKY6 interacting proteins, WRKY42, as the closest homolog of WRKY6 (Eulgem et al., 2000), may have similar function to WRKY6. To test this hypothesis, we tested possible binding of WRKY42 with the PHO1 promoter. The ChIP-qPCR experiments showed that, similar to WRKY6, WRKY42 can bind to both the Y and Z W-box motifs within the PHO1 promoter but not to the Q and X W-box motifs ( Figure 8A). To further test possible function of WRKY42 on regulation of PHO1 expression, transient expression experiments in tobacco leaves were performed. The results showed that, similar to WRKY6, WRKY42 inhibited PHO1 promoter activity ( Figure 8B). The coinjection of Super:WRKY6 and Super:WRKY42 showed much stronger repression on ProPHO1: GUS expression than did injection of either Super:WRKY6 or Super:WRKY42 alone ( Figure 8B). However, WRKY75 had no effect on PHO1 expression and did not influence the inhibition of PHO1 expression by WRKY6 ( Figure 8B).
Taking all these results together, we concluded that WRKY6 functions in plant responses to low Pi stress by negatively regulating PHO1 expression. Under normal conditions with sufficient Pi supply, WRKY6 (and probably also WRKY42) can bind to the W-box motifs W Y and W Z within the PHO1 promoter and represses the transcription of PHO1. Under Pi-deficient conditions, WRKY6 protein content is decreased via a 26S proteosome-mediated proteolysis, and the interaction of WRKY6 and the PHO1 is limited. As a result, repression of PHO1 transcription by WRKY6 is relieved, which might be important for plant adaptation to a Pi-deficient environment.

DISCUSSION
Plant-specific WRKY transcription factor family proteins have been implicated in the regulation of genes involved in plant responses to biotic as well as abiotic stresses, such as pathogen-induced stress (Dong et al., 2003;Eulgem and Somssich, 2007), drought, cold, and salinity stresses (Seki et al., 2002;Dong  (A) Diagram of the PHO1 promoter region showing the relative positions of four of six W-boxes (Q, À1718 to À1625; X, À1269 to À1181; Y, À966 to À936; and Z, À775 to À618). W-boxes are marked by black rectangles, and the untranslated region and exons of PHO1 are marked by gray boxes.
(B) ChIP-qPCR analysis of the PHO1 promoter sequence. ChIP assays were performed with chromatin prepared from wild-type Arabidopsis roots. The gray and black bars represent the ChIP signals with (WRKY6) and without (NoAB) addition of anti-WRKY6 serum, respectively. The experiments were repeated three times, and three replicates were included for each sample in each experiment. The data are presented as means 6 SE (n = 3). et al., 2003). WRKY factors act primarily by binding to conserved W-box elements in the promoters of specific targets to direct temporal and spatial expression of these genes (Ulker and Somssich, 2004). Among 74 members in the Arabidopsis WRKY family, only WRKY75 has been reported to be involved in modulation of Pi acquisition and root development (Devaiah et al., 2007a). This study demonstrated that WRKY6 (and probably also WRKY42) plays an important role in modulation of plant responses to low Pi stress via regulation of PHO1 expression.

WRKY6 Is a Negative Regulator for PHO1 Transcription
Plant responses to Pi starvation involve the transcriptional regulation of numerous genes to establish an adaptive mechanism (Franco-Zorilla et al., 2004). Several Arabidopsis transcription factors were identified functioning in the Pi starvation response, such as PHR1 (Rubio et al., 2001), ZAT6 (Devaiah et al., 2007b), BHLH32 (Chen et al., 2007), MYB62 (Devaiah et al., 2009), and a WRKY family protein WRKY75 (Devaiah et al., 2007a). WRKY75 can be induced by low Pi stress and is believed act as a positive regulator of Pi acquisition under Pi-deficient conditions (Devaiah et al., 2007a).
In this study, we first observed that the WRKY6 overexpression lines displayed similar phenotypes as the pho1 mutant under low Pi stress, including growth inhibition and anthocyanin accumulation. We further demonstrated that WRKY6 protein can bind to two W-boxes of the PHO1 promoter and that PHO1 transcription was repressed by overexpression of WRKY6 under normal Pi supply conditions. This repression of PHO1 transcription by WRKY6 was relieved under low Pi conditions, indicating that the regulation of PHO1 transcription by WRKY6 is Pi dependent. In addition, protein blot analysis showed that the low Pi treatmentinduced WRKY6 decrease was inhibited by a 26S proteosome inhibitor MG132. This suggests that the low Pi-induced release of PHO1 repression may result from 26S proteosome-mediated WRKY6 proteolysis. Such a Pi-dependent mechanism may make WRKY6 a key regulator for plant responses to low Pi stress.

Mechanism of PHO1 Regulation by WRKY6 and WRKY42
Under normal growth conditions, WRKY6 represses PHO1 expression to balance Pi homeostasis through its binding to two W-boxes at the end of the coding region of the PHO1 promoter. When a low Pi stress signal is sensed by an unknown signaling mechanism and relayed to the E3 ubiquitin ligase, a 26S proteosome-mediated WRKY6 protein degradation is activated and WRKY6 binding to the PHO1 promoter W box motifs is weakened so that PHO1 transcription is induced to cope with the Pi-deficient environment. As a result, PHO1-facilitated Pi loading from root to xylem occurs and translocation of Pi from root to shoot could be promoted so that plants can adapt to a Pideficient environment. The plant may use WRKY6 (and probably also WRKY42) as a key regulator that responds to varied Pi supply conditions and regulates Pi distribution in different organs via regulation of PHO1 as well as other unknown components.
Our results also showed that, similar to WRKY6, WRKY42 can bind to both the Y and Z W-box motifs of the PHO1 promoter but not to the Q and X W-box motifs ( Figure 8A). In addition, WRKY42 alone can also inhibit PHO1 expression ( Figure 8B). Alternatively, considering that WRKY6 can interact with WRKY42 (see Supplemental Table 1 online), one may wonder if these two factors can form heterocomplexes to regulate PHO1 expression. Xu et al. (2006b) reported that three different kinds of WRKY proteins (WRKY18, WRKY40, and WRKY60) can interact with each other and form heterocomplexes, and the interactions between these WRKY factors influence their DNA binding activities. Robatzek and Somssich (2002) showed that WRKY6 can act as a negative regulator of its own and WRKY42 expression even though the mechanism and function remains unknown.
To test possible synergic effects of WRKY6 and WRKY42 on PHO1 expression, we coinjected Super:WRKY6 and Super: WRKY42 in tobacco leaves to test their effects on ProPHO1: GUS expression. The results ( Figure 8B) indicate that WRKY6 and WRKY42 together had stronger repression on PHO1 expression. However, still we cannot conclude that they work together (by forming heterodimers) or work independently at this point. The ChIP-qPCR data showed that WRKY6 and WRKY42 had differential interactions with Y and Z W-boxes within the PHO1 promoter. WRKY6 displayed a stronger interaction with the Y box than with the Z-box ( Figures 3B and 5C), while WRKY42 displayed a stronger interaction with the Z-box than with the Y-box ( Figure 8A). These results indicate that WRKY6 and WRKY42 may regulate PHO1 expression in different ways. To clarify further if WRKY6 and WRKY42 work independently or together as a complex in regulation of PHO1 expression is an important issue for comprehensively understanding complex mechanisms of PHO1 regulation by WRKY factors.

Regulation of PHO1 by Other Possible Regulatory Factors
Although the Q and X W-boxes within the PHO1 promoter do not bind to WRKY6, we have noticed that the deletion of the sequences containing the Q-and X-boxes reduced the inhibitory effect of WRKY6 on PHO1 expression ( Figures 4B, 4C, and 4E), particularly for wild-type plants. One of possible explanation for this phenomenon is that, within the deleted sequences, some other regulatory elements related to plant responses to low Pi stress may exist. The analysis of the deleted sequences using TESS (http://www.cbil.upenn.edu/cgi-bin/tess/tess) shows that, in addition to Q and X W-boxes, there are a number of regulatory elements for the following transcription factors: GATA factors, CCAAT-box transcription factor, multiprotein bridging factor 1 (MBF1), homeodomain-leucine zipper (HD-Zip) transcription factor, MYB transcription factor, Dof (DNA binding with one finger) factor, heat shock transcriptional factor (HSF), etc. Although there is no report so far regarding PHO1 regulation by these TFs, at least two of them have been reported involving plant responses to low Pi stress. The HD-Zip factor has been reported to bind to the phosphate response domain of the soybean (Glycine max) VspB tripartite promoter (Tang et al., (C) ChIP-qPCR assays to detect the association between WRKY6 and W-boxes within the PHO1 promoter in wild-type plants under the normal (MS) and LP conditions. The ChIP signals with (WRKY6) and without (NoAB) addition of anti-WRKY6 serum are indicated. The data are presented as means 6 SE (n = 3). The experiments were repeated three times, and three replicates were included for each sample in each experiment. 2001). Nilsson et al. (2007) reported that increased expression of the MYB-related transcription factor PHR1 resulted in enhancement in phosphate uptake in Arabidopsis. It is plausible to further hypothesize that one or more of these regulatory elements (which were deleted together with Q and X W-boxes in the experiments shown in Figure 4C) may directly or indirectly be involved in PHO1 regulation.

WRKY6 and WRKY75 May Respond to Low Pi Stress via Different Pathways
WRKY75 has been reported to play an important role in the phosphate starvation response, particularly by modulating Pi uptake and root development (Devaiah et al., 2007a). The results presented here demonstrate that WRKY6 responds to low Pi stress by regulation of PHO1 transcription. Repression of WRKY75 expression (by RNA interference methods) resulted in a decrease of Pi uptake (Devaiah et al., 2007a), while overexpression of WRKY6 repressed PHO1 expression and consequently reduced Pi accumulation in shoots. It is known that PHO1 functions in Pi translocation from root to shoot (Poirier et al., 1991;Hamburger et al., 2002). Considering these results together with the fact that WRKY75 did not have an effect on PHO1 promoter activity ( Figure 8B; in addition, WRKY75 did not interact with WRKY6 in the yeast two-hybrid assay), we may further hypothesize that WRKY75 and WRKY6, in response to low Pi stress, function in different regulatory pathways. Identifying the gene(s) whose expression is specifically regulated by WRKY75 as well as other possible transcription factors will help us to clarify the complex mechanisms of plant responses to low Pi stress.

Other Possible Roles of WRKY in Regulation of Pi Starvation Responsive Genes
To test if WRKY6 would play roles in regulation of other Pi starvation responsive genes, we performed comparative transcriptome analyses with various plant materials (35S:WRKY6-9, wrky6-1, and the wild type) using the Affymetrix GeneChip. As shown in Supplemental Table 2 online, among 30 low Pi response genes (Devaiah et al., 2007a;Lin et al., 2009), 11 of them showed expression changes between either wild-type and 35S: WRKY6-9 plants or the wild type and the wrky6-1 mutant. Among the members of the PHT1 family, Pht1;5 and Pht1;8 displayed transcriptional changes in 35S:WRKY6-9 plants and the wrky6-1 mutant compared with wild-type plants (see Supplemental Table  2 online). Both Pht1;5 and Pht1;8 contain W-boxes in their promoters (Devaiah et al., 2007), suggesting that WRKY6 may regulate their transcription. It is known that expression of Pht1;8 was significantly increased in the pho2 mutant Bari et al., 2006), a mutant overaccumulating Pi in leaves, suggesting a possible role of Pht1;8 in WRKY6-and PHO1related Pi mobilization. PS2 and PS3, two members of a phosphatase family, were significantly upregulated in 35S:WRKY6-9 plants and downregulated in the wrky6-1 mutant (see Supplemental Table 2 online), suggesting that WRKY6 also might be involved in plant early responses to low Pi stress (Devaiah et al., 2007). Several low Pi responsive transcription factors listed in Supplemental Table 2 online, including PHR1, ZAT6, WRKY75, and BHLH32, did not show significant changes in their transcription either in 35S:WRKY6-9 plants or in the wrky6-1 mutant.
In addition, among the genes whose expression was upregulated or downregulated by more than two times and showed relevant changes (changes in opposite direction in 35S:WRKY6-9 plants compared with the wrky6-1 mutant), there are a total of 25 genes (listed in Supplemental Table 3 online) whose promoters contain W-box(es). Among these genes, there are 15 genes whose transcriptions were repressed, and transcription of 10 other genes was enhanced in 35S:WRKY6-9 plants. These data indicate that, in addition to its function in plant responses to low Pi stress, WRKY6 may be involved in a broad range of transcriptional regulations related to different processes, such as senescence, pathogen defense, and wounding responses Somssich, 2001, 2002). Seven-day-old wild-type seedlings were transferred to LP medium (A), LP medium with 10 mM MG132 (LP+MG132) (B), or LP medium with DMSO (LP+DMSO) (C). The roots of seedlings were harvested for protein extraction at the indicated time. Protein extracts were analyzed by immunoblots using rabbit anti-WRKY6 serum. Tubulin levels were detected in parallel as a loading control with antitubulin antibody. (A) ChIP-qPCR assays to detect the association between WRKY42 and W-boxes within the PHO1 promoter in wild-type plants under normal conditions. The experiments were repeated three times, and three replicates were included for each sample in each experiment. The data are presented as means 6 SE (n = 3). (B) Transient overexpression of the ProPHO1:GUS fusion together with Super:WRKY6, Super:WRKY42, or Super:WRKY75 in Nicotiana benthamiana leaves. ProPHO1:GUS fusion together with Super1300 vector was taken as the control. The data are presented as means 6 SE (n = 4).
For phenotype tests and seed harvest, Arabidopsis thaliana plants were grown in a potting soil mixture (rich soils:vermiculite = 2:1, v/v) and kept in growth chambers at 228C with illumination at 120 mmol·m 22 ·s 21 for an 18-h daily light period. The relative humidity was ;70% (65%).
Low Pi stress treatment of plants was conducted by growing seedlings on Petri dishes containing Pi-sufficient (MS) or Pi-deficient (low Pi or LP) medium. The seeds were surface sterilized with the mixed solutions of NaClO (0.5%) and Triton X-100 (0.01%) for 10 min followed by washing with sterilized distilled water four times. The sterilized seeds were first incubated on Petri dishes containing MS agar (0.8%) medium (containing 1.25 mM KH 2 PO 4 and 3% sucrose) at 48C for 2 d before germination. The seeds were germinated at 228C under constant illumination at 40 mmol·m 22 ·s 21 and the 7-d-old seedlings were transferred to LP medium. The LP medium was made by modification of MS medium such that the Pi (supplied with KH 2 PO 4 ) concentration in LP medium was 10 mM, and agar was replaced by agarose (Promega) to avoid the contamination of phosphorous.

Quantification of Total Pi
Seven-day-old Arabidopsis seedlings germinated on MS medium were transferred to MS or LP medium for another 7 d, and then the roots and shoots were harvested for Pi content measurements. The samples were oven dried at 808C for 48 h before determination of dry weight, and the samples were ashed in a muffle furnace at 3008C for 1 h and 5758C for an additional 5 h and then dissolved in 0.1 N HCl. The total Pi content in the samples was quantified as described previously (Ames, 1966).

RNA Gel Blot, RT-PCR, and Real-Time PCR Analysis
For RNA gel blot analysis, 7-d-old Arabidopsis seedlings germinated on MS medium were transferred to MS or LP medium for another 7 d and then the seedlings or roots were harvested for extraction of total RNA using the Trizol reagent (Invitrogen). Thirty micrograms of RNA were loaded per lane and transferred to a nylon membrane for hybridization. Gene-specific probes were amplified by PCR using WRKY6-specific primers (the forward primer 59-CTTTGGCGATGTCTAGAATTGA-39 and the reverse primer 59-CCTCACCTACTGCTCTCGTAGG-39) and PHO1specific primers (the forward primer 59-TACTTGATTCTTTCTTACCC-TACTTCTGG-39 and the reverse primer 59-TCCAAGGAACGGTAACGG-TACGGTCTTCACT-39) as the templates, respectively. The probes were labeled with [a-32 P]dCTP using random primer labeling reagents (Pharmacia) and hybridized to RNA gel blotted onto nylon membrane. The rRNA was taken as the control.
For RT-PCR analysis, total RNA was extracted with Trizol reagent (Invitrogen) and then treated with DNase I RNase Free (Takara) to eliminate genomic DNA contamination. The cDNA was synthesized from treated RNA by SuperScript II RNase H 2 reverse transcriptase (Invitrogen) using oligo(dT) 15 primer (Promega). The PCR experiments were conducted with WRKY6-specific primers (the forward primer 59-ATGTTTCGTTTTCCGGTAAGTCTTGGAGGA-39 and the reverse primer 59-TATTGCCTATTGTCAACGTTGCTCGTTGTAACATTA-39) and EF1a -specific primers (the forward primer 59-ATGCCCCAGGACATCGT-GATTTCAT-39 and the reverse primer 59-TTGGCGGCACCCTTACGTG-GATCA-39). EF1a was used as a quantitative control.
For real-time PCR analysis, total RNA extraction was performed as described above, and the RNA was treated with DNase I RNase Free (Takara) to eliminate genomic DNA contamination. The cDNA was synthesized from total RNA by SuperScript II RNase H 2 reverse transcriptase (Invitrogen) using Radom Hexamer Primer (Promega). Quantitative realtime PCR was performed using the Power SYBR Green PCR Master Mix (Applied Biosystems; P/N 4368577) on a 7500 Real Time PCR System machine (Applied Biosystems) following the manufacturer's protocols. The PCR amplification was performed at 958C for 15 s and 608C for 1 min. Relative quantitative results were calculated by normalization to 18S rRNA. qPCR was conducted with WRKY6-specific primers (the forward primer 59-TAGTCACGACGGGATGATGA-39 and the reverse primer 59-ATTAGGAGGCGGAGGTGAGT-39) and PHO1-specific primers (the forward primer 59-TGGTTCTCCGGAACAAGAAC-39 and the reverse primer 59-TGACTTCAAGTGACGCCAAG-39).

Antibody Generation and ChIP-qPCR Assay
The whole coding sequences were amplified by PCR using WRKY6specific primers (the forward primer 59-cccgggCCCGGGATGGACAGAG-GAGGTCT-39 and the reverse primer 59-ctcgagCTCGAGCTATTGATTT-TTGTTGTTTC-39) and WRKY42-specific primers (the forward primer 59-ggatccGGATCCATGTTTCGTTTTCCGGTAAG-39 and the reverse primer 59-gagctcGTCGACTCTTATTGCCTATTGTCAAC-39). The lowercase letters represent the restriction sites. The whole coding sequences of At WRKY6 and WRKY42 were cloned into pGEX-4T-2 (Pharmacia). The reconstructed pGEX-4T-2 plasmid containing WRKY6 or WRKY42 was then transformed into Escherichia coli strain BL21 (DE3) to express proteins by induction with 0.1 mM isopropyl-b-D-thiogalactopyranoside at 188C for 8 h. The resulting glutathione S-transferase (GST) fusion proteins were purified using Glutathione Sepharose 4B (Pharmacia). Protein concentrations were determined using the Bio-Rad protein assay kit. The polyclonal antibodies against WRKY6 or WRKY42, generated by inoculation of rabbits, were purified using the Amino Link Plus Immobilization Kit (Pierce).
The 7-d-old seedlings germinated on MS medium were transferred to MS or LP medium for another 7 d and then the roots were harvested for ChIP experiments. The chromatin samples for ChIP experiments were obtained following the methods by Saleh et al. (2008). The roots of plants seedlings were first cross-linked by formaldehyde, and the purified crosslinked nuclei were then sonicated to shear the chromatin into suitably sized fragments. The antibody that specifically recognizes the recombinant WRKY6-GST or WRKY42-GST was used to immunoprecipitate DNA/protein complexes from the chromatin preparation. The DNA in the precipitated complexes was recovered and analyzed by qPCR methods. Following the methods described by Haring et al. (2007), qPCR analysis was performed using the Power SYBR Green PCR Master Mix (Applied Biosystems; P/N 4368577) in a 20-mL qPCR reaction on the 7500 Real Time PCR System machine (Applied Biosystems) following the manufacturer's protocols. The chosen primer combinations (see Supplemental  Table 4 online) can amplify fragments of 150 to 200 bp within the promoter of PHO1. To ensure the reliability of ChIP data, the input sample and noantibody (NoAB) control sample were analyzed with each primer set. The results were quantified with a calibration line made with DNA isolated from cross-linked and sonicated chromatin.

Vector Construction and Arabidopsis Transformation and GUS Assays
PHO1 promoter variants, including the full-length promoter (2282 bp), the promoter containing two W-boxes (Y and Z, 1141 bp), and the promoter without W-box (454 bp), were amplified by PCR from Arabidopsis genomic DNA and cloned into the transformation vector pCAMBIA1381 at the SalI and PstI restriction sites, respectively. All primer sequences used for vector constructions are listed in Supplemental Table 5 online. All tested Arabidopsis plants (Col-0, 35S:WRKY6-9, and the wrky6-1 mutant) were transformed by these three constructed vectors, respectively, using the Agrobacterium tumefaciens-mediated gene-transfer procedure (Clough and Bent, 1998). All transgenic lines used in this study are T3 homozygous plants with single copy insertion. The GUS staining and GUS activity measurements were performed as described previously (Xu et al., 2006a).

SDS-PAGE and Protein Gel Blot Analysis
For MG132 treatment, 10 mM stock solution was made by dissolving MG132 (Calbiochem) in DMSO, and the final MG132 concentration was 10 mM. For the control of MG132 treatment, 1/1000 DMSO was added to the medium. The 7-d-old seedlings germinated on MS medium were transferred to low Pi (LP), low Pi with 10 mM MG132 (Calbiochem) (LP+MG132), or low Pi with DMSO (LP+DMSO) medium, and then the roots were harvested at the indicated time for immunoblot analysis. Total proteins were extracted according to the method of Saleh et al. (2008), and 50 mg proteins of each sample were separated on a 10% SDS-PAGE and analyzed by protein gel blot according to Towbin et al. (1979). Mouse anti-tubulin (Sigma-Aldrich) (with 1:4000 dilution) and rabbit anti-WRKY6 serum (with 1:10,000 dilution) were used as the primary antibodies, and goat anti-mouse (or anti-rabbit) peroxidase-conjugated AffiniPure Goat Anti-Mouse IgG (H+L) (Jackson) (with 1:8000 dilution) was used as the secondary antibody. The membranes were visualized using a Super-Signal West Femto Trial Kit (Themor) following the manufacturer's instructions.

Transient Expression Assays in Nicotiana benthamiana
The constructs of ProPHO1:GUS with Super1300, Super1300:WRKY6, or Super1300:WRKY42 were transformed into A. Agrobacterium (GV3101). Agrobacterium cells were harvested by centrifugation and suspended in the solutions containing 10 mM MES, pH 5.6, 10 mM MgCl 2 , and 200 mM acetosyringone to an optical density (600 nm) of 0.7, incubated at room temperature for 4 h, and then used to infiltrate leaves of N. benthamiana using a needle-free syringe. The GUS activity of the infiltrated leaves was quantitatively determined.

Yeast Two-Hybrid Assay
The WRKY6 full-length cDNA sequence was amplified with WRKY6specific primers (the forward primer 59-GCcatatgATGGACAGAG-GATGGTCTGG-39 and the reverse primer 59-GCgtcgacCTATT-GATTTTTGTTGTTTCCTTCG-39) and cloned into pGBK-T7 vector at NdeI and SalI sites. The lowercase letters in the primers indicate restriction sites. Two-hybrid screening and assays were performed as described (Kohalmi et al., 1997). Strain AH109 (BD Biosciences) of Saccharomyces cerevisiae was used in all yeast two-hybrid experiments.

Microarray Analysis
For microarray analysis, the seeds of WRKY6 overexpression line 35S: WRKY6-9, WRKY6 knockout mutant wrky6-1, and wild-type (Col-0) plants were germinated at 228C under constant illumination at 40 mmol·m 22 ·s 21 and grown for 7 d on MS medium (control conditions), and the seedlings were used for total RNA extraction using the Trizol reagent (Invitrogen). Microarray experiments were conducted once, and three arrays for different plant materials were hybridized according to the Affymetrix GeneChip Expression Analysis Manual (http://www.affymetrix. com). The scanned arrays were analyzed first with Affymetrix GCOS 1.0 (MAS 5.0) software to generate detection calls and normalized using Affymetrix GCOS software, and the TCT value was set to 100. When analyzing the transcriptionally changed genes, the signal ratio between two plant materials was calculated to represent the fold change of this gene for its transcription in the corresponding process, and the fold change P value for each gene was obtained at the same time using GCOS software. A gene was considered to be transcriptionally changed when it was upregulated or downregulated more than onefold. The genes with P values < 0.05 and fold changes > 1 are included in Supplemental Table 3 online.

Accession Numbers
Sequence data from this article can be found in the Arabidopsis Genome Initiative or GeneBank/EMBL database under the following accession numbers: WRKY6 (At1g62300), WRKY42 (At4g04450), WRKY75 (At5g13080), PHO1 (At3g23430), and EF1a (At5g60390). Microarray data from this article have been deposited with the National Center for Biotechnology Information Gene Expression Omnibus data repository (http://www.ncbi.nlm.nih.gov/geo/) under accession number GSE18273.

Supplemental Data
The following materials are available in the online version of this article.   Table 3. Genes with W-Boxes Whose Expression Was Up-or Downregulated More Than Twofold in the 35S:WRKY6-9 Line (OE) and the wrky6-1 Mutant (KO) Compared with Wild-Type Plants.
Supplemental Table 4. Primers Used for ChIP-qPCR Experiments.
Supplemental Table 5. Primers Used for the ProPHO1:GUS Constructs.