Activation of MKK9-MPK3/MPK6 enhances phosphate acquisition in Arabidopsis thaliana

Transcription

1 Research Activation of MKK9-MPK3/MPK6 enhances phosphate acquisition in Arabidopsis thaliana Lei Lei, Yuan Li, Qian Wang, Juan Xu, Yifang Chen, Hailian Yang and Dongtao Ren State Key Laboratory of Plant Physiology and Biochemistry, College of Biological Sciences, China Agricultural University, Beijing , China Author for correspondence: Dongtao Ren Tel: Received: 27 March 2014 Accepted: 29 April 2014 doi: /nph Key words: Arabidopsis thaliana, MKK9, MPK3, MPK6, phosphate, Pi responses, signal transduction pathway. Summary Despite the abundance of phosphorus in soil, very little is available as phosphate (Pi) for plants. Plants often experience low Pi (LP) stress. Intensive studies have been conducted to reveal the mechanism used by plants to deal with LP; however, Pi sensing and signal transduction pathways are not fully understood. Using in-gel kinase assays, we determined the activities of MPK3 and MPK6 in Arabidopsis thaliana seedlings under both LP and Pi-sufficient (Murashige and Skoog, MS) conditions. Using MKK9 mutant transgenic and crossed mutants, we analyzed the functions of MPK3 and MPK6 in regulating Pi responses of seedlings. The regulation of Pi responses by downstream components of MKK9-MPK3/MPK6 was also screened. LP treatment activated MPK3 and MPK6. Under both LP and MS conditions, mpk3 and mpk6 seedlings took up and accumulated less Pi than the wild-type; activation of MKK9- MPK3/MPK6 in transgenic seedlings induced the transcription of Pi acquisition-related genes and enhanced Pi uptake and accumulation, whereas its activation suppressed the transcription of anthocyanin biosynthetic genes and anthocyanin accumulation; WRKY75 was downstream of MKK9-MPK3/MPK6 when regulating the accumulation of Pi and anthocyanin, and the transcription of Pi acquisition-related and anthocyanin biosynthetic genes. These results suggest that the MKK9-MPK3/MPK6 cascade is part of the Pi signaling pathway in plants. Introduction Phosphorus is an essential macronutrient for plant growth, development and reproduction. It is a structural constituent of many macromolecules (e.g. nucleic acids, adenosine-5 0 -triphosphate and phospholipids) and is indispensable during energy transfer, metabolic regulation and signal transduction (Marschner, 1995; Chiou & Lin, 2011). Phosphate (H 2 PO 4 and HPO 4 2, hereafter Pi) is the major form of phosphorus taken up by plants from the soil (Bieleski, 1973; Holford, 1997; Vance et al., 2003). Despite the abundance of phosphorus in the soil, very little is present as Pi; therefore, plants often experience low Pi (LP) conditions (Marschner, 1995; Shen et al., 2011). To deal with the limited availability of Pi, plants have evolved a variety of adaptive strategies to facilitate Pi acquisition, usage and remobilization, such as by increasing root absorption surfaces, inducing Pi transporter gene expression, elevating phosphatase and ribonuclease activities, increasing the secretion of organic acids and reducing Pi requirements via alerting the metabolism (Misson et al., 2005; Yuan & Liu, 2008; Lin et al., 2009; Yang & Finnegan, 2010; Chiou & Lin, 2011). Experiments with either Phi or methylphosphate, non-metabolized analogs of Pi, reveal that Pi itself is functional as an initial signal (Carswell et al., 1996; Ticconi et al., 2001; Varadarajan et al., 2002; Pratt et al., 2004). External Pi can either be transported across the plasma membrane and sensed by an intracellular sensor, or sensed directly by an unknown sensor(s) that is localized in the plasma membranes of plant cells (Svistoonoff et al., 2007; Chiou & Lin, 2011). The signal is transduced via signaling transduction pathway(s) and subsequently regulates an extensive array of responses (Yuan & Liu, 2008; Chiou & Lin, 2011). Phytohormones, such as ethylene, gibberellic acid, cytokinins and auxin, have been determined to be important mediators in the integration of Pi signaling with Pi responses (Martın et al., 2000; Lopez-Bucio et al., 2002; Franco-Zorrilla et al., 2005; Devaiah et al., 2009; Nagarajan & Smith, 2012). Sugar signaling has been reported to regulate root system architecture (RSA) and the expression of Pi starvation-induced genes (Hammond & White, 2008, 2011). MicroRNAs can either be a systemic signal in the transduction of Pi deficiency signaling or can target and direct the cleavage of some Pi-responsive genes (Lin et al., 2008; Pant et al., 2008; Kuo & Chiou, 2011). Transcription factors, such as MYB family members (PHR1, PHL1, MYB62) and WRKY family members (WRKY75), have also been shown to be important signaling components in the regulation of Pi responses (Devaiah et al., 2007, 2009; Bustos et al., 2010). Some members of the SPX domain-containing protein families have recently been shown to be involved in Pi signaling (Kant et al., 2011; 1146

2 New Phytologist Research 1147 Secco et al., 2012; Wang et al., 2012). Our knowledge of the Pi signaling components in plants has been greatly improved during the past few years; however, to fully understand the mechanism of plant responses to Pi, research on the interaction between these components and the identification of the novel signaling machinery for the regulation of Pi responses is still needed (Chiou & Lin, 2011). Mitogen-activated protein kinase (MAPK) cascades, consisting of MAPKKK, MAPKK and MAPK, are highly conserved signaling modules in eukaryotes. These cascades have been demonstrated to be modules that link upstream receptors or sensors to downstream processes in various ways. MAPK cascades have important functions in regulating plant stress responses, growth and developmental processes (Colcombet & Hirt, 2008; Pitzschke et al., 2009; Andreasson & Ellis, 2010; Rodriguez et al., 2010; Komis et al., 2011; Tena et al., 2011). The Arabidopsis genome contains 60 MAPKKKs, 10 MAPKKs and 20 MAPKs (namely MKKK, MKK and MPK) (MAPK-Group, 2002). Previous reports have shown that MPK3 and/or MPK6 can be activated by different MKKs and participate in specific signaling pathways: MKK3 activates MPK6 to regulate jasmonic acid signaling (Takahashi et al., 2007); MKK2 activates MPK4/MPK6 to mediate cold and salt stress tolerance (Teige et al., 2004); MKK4/MKK5 activate MPK3/MPK6 to regulate H 2 O 2 production (Ren et al., 2002), H 2 O 2 -induced NO production (Wang P et al., 2010), stomata and ovule development (Wang et al., 2007, 2008) and defense responses (Asai et al., 2002; Wang Y et al., 2010); MKK4/MKK5/MKK9 activate MPK3/MPK6 to induce ethylene and camalexin biosynthesis (Liu et al., 2008; Xu et al., 2008); and MKK9 activates MPK3/MPK6 to regulate leaf senescence (Zhou et al., 2009) and ethylene signaling (Yoo et al., 2008). Because Pi is involved in many cellular processes, LP conditions cause severe nutrient stress during plant growth and development. Genome-scale transcriptional profiling has revealed that multiple genes in these MAPK modules are induced by LP (Misson et al., 2005; Graham et al., 2006). However, it remains unknown whether MAPK signaling cascades are involved directly in the regulation of plant responses to LP. In this study, we found that MPK3 and MPK6 could be activated in wild-type (WT) seedlings under LP conditions. Constitutive activation of MPK3 and MPK6 via upstream MKK9 in MKK9 transgenic seedlings suppresses LP-induced anthocyanin production. After MKK9-MPK3/MPK6 activation, the transcription of several LP-responsive genes was induced under Pi-sufficient (Murashige and Skoog, MS) conditions and enhanced under LP conditions, and Pi uptake and accumulation were increased under both LP and MS conditions. Further analysis showed that the MAPK module regulated Pi responses through WRKY75. Materials and Methods Plant materials and growth conditions WT and mutant Arabidopsis thaliana (L.) Heynh (ecotype, Columbia-0) seeds were surface sterilized and sown on 1% agar plates containing MS and 1% sucrose, ph 5.7. After cold treatment at 4 C for 2 d, the seeds were germinated and grown on plates at 22 C in a growth room with a 16-h photoperiod at a photon flux density of 100 lmol m 2 s 1 for 7 d. The MS and LP media used for the Pi treatments were prepared as described by Chen et al. (2009). The Pi concentration in MS medium was 1.26 mm and in LP medium was 10 lm. For Pi response experiments, 7-d-old seedlings of similar sizes were transferred onto new plates containing either MS or LP medium, and grown for the additional days as indicated. For MKK9 mutant protein induction, 0.02 lm dexamethasone (DEX) was pre-added to the media. Photographs were taken, and seedlings were harvested and used for anthocyanin and Pi content measurements. For soil growth, seedlings were transferred from MS medium to soil and grown in a growth room with a 12-h photoperiod at a photon flux density of 100 lmol m 2 s 1 for 7 d. Soil-grown plants were used for genetic crossing and seed setting. T-DNA insertion mutants and genetic crosses Vector, MKK9 KR, MKK9 DD, MKK9 DD /mpk3, MKK9 DD /mpk6, MKK9 DD /ein2, mpk3, mpk6 and mkk9 plants were generated as described previously (Xu et al., 2008). The T-DNA insertion mutants, including wrky33 (SALK_006603), wrky75 (SALK_101367) and phr1 (SALK_067629), were obtained from the Arabidopsis Biological Resource Center. The homozygous phr1 and wrky75 mutants were screened using genomic PCR and confirmed by reverse transcription-polymerase chain reaction (RT-PCR) using gene-specific primers. Homozygous wrky33 was identified according to Zheng et al. (2006). All mutants were crossed into the MKK9 DD background and the homozygous double mutants were used for the experiments. The sequences of the primers used are listed in Supporting Information Table S1. RNA isolation, RT-PCR and real-time quantitative RT-PCR (Q-PCR) Total RNA was isolated from samples using Trizol reagent (Invitrogen). First-strand cdna was synthesized by M-MLV reverse transcriptase (Promega) using oligo dt(16) as the primer and total RNA as the template. The RT-PCR experiments were conducted to confirm gene knockouts using WRKY33- and WRKY75-specific primers. Real-time Q-PCR was performed in the presence of SYBR Green Mix (Takara, Dalian, China) to monitor Pi-responsive gene expression. Amplification was conducted in real time with a 7500 real-time PCR system (Applied Biosystems). Expression levels of Ubiquitin5 were used as the internal control. The sequences of the primers are listed in Table S1. Measurement of anthocyanin and total Pi content For the anthocyanin content assay, the seedlings were harvested and quickly frozen in liquid nitrogen. Anthocyanin content in samples was measured as described by Teng et al. (2005). For the

3 1148 Research New Phytologist measurement of Pi content, the shoots and roots of seedlings were separately collected and dried at 80 C for 48 h. After recording their dry weight, these samples were flamed to ash and dissolved in 0.1 M HCl. Pi in the samples was measured as described by Ames (1966). Pi uptake assay For pretreatment, 7-d-old seedlings that had germinated on MS medium were transferred to either MS or LP medium with or without DEX for an additional 1 or 3 d. Groups of 10 seedlings were weighed and used as one biological sample. Pi uptake assay was performed as described by Devaiah et al. (2007) and with modifications. In this study, 0.02 lm DEX was added to both the pretreatment and uptake solutions, and 0.3 lci [ 32 Pi]orthophosphate was used for each sample. Radioactivity was measured using a Hidex 300SL Automatic Liquid Scintillation Counter (Hidex, Turku, Finland). Immunoblot and in-gel kinase assays Protein extraction, separation, immunoblot and in-gel kinase assays were performed as described previously (Xu et al., 2008). An anti-flag M2 monoclonal antibody (1 : ) was used as the primary antibody, and a horseradish peroxidase-conjugated goat anti-mouse antibody was used as the secondary antibody (1 : ). The protein membranes were visualized using an Enhanced Chemiluminescence Kit (Roche) and exposed to X-ray film. For in-gel kinase assays, myelin basic protein was embedded in the separating gel as a substrate for the kinases. Microarray analysis WT and MKK9 DD seedlings were grown on MS medium for 7 d, and WT seedlings were then transferred to either MS or LP medium and grown for an additional 3 d. MKK9 DD seedlings were then transferred to MS medium with or without DEX and grown for an additional day. Samples were harvested and total RNA was isolated. RNA purification and chip hybridization were performed using Affymetrix Microarray Services (CapitalBio Co., Beijing, China). Microarray data were deposited at the ArrayExpress database ( under accession number E-MTAB The identification of differentially expressed genes was performed using Significance Analysis of Microarrays software 2.10 with a fold change of 2.0 or 1.5 and P < 0.01 as cut-off values. Accession numbers Sequence data from this article can be found in the Arabidopsis Genome Initiative or GenBank/EMBL databases under the following accession numbers: MKK9, At1g73500; MPK3, At3g45640; MPK6, At2g43790; WRKY75, At5g13080; WRKY33, At2g38470; Pht1;1, At5g43350; Pht1;4, At2g38940; At4, At5g03545; mir399d, At2g34202; MYB62, At1g68320; IPS1, At3g09922; DFR, At5g42800; LDOX, At4g22880; PHR1, At4g28610; UBQ5, At3g T-DNA insertion lines used here are: mpk3 (SALK_100651), mpk6 (SALK_127507), wrky75 (SALK_101367), phr1 (SALK_067629) and wrky33 (SALK_006603). Results Pi deficiency activates MPK6 and MPK3 To determine whether MAPKs or other kinases are involved in the response of Arabidopsis plants to different Pi conditions, we grew Col-0 WT seedlings on MS medium for 7 d, transferred them to new plates with either MS or LP medium, and took samples at various times. Kinase activities in these seedlings were analyzed using in-gel kinase activity assays. The results showed that WT seedlings grown on MS medium before transfer to new medium contained a lower basal level activity of a 46-kDa kinase (Fig. 1a, 0 d); the 46-kDa kinase activity was increased slightly in WT seedlings after transfer to MS medium and was increased greatly in WT seedlings after transfer to LP medium. A 40-kDa kinase was newly induced, and its activity reached similar levels in WT seedlings transferred to either MS or LP medium; however, a 43-kDa kinase was newly activated in WT seedlings only after transfer to LP medium. The activation of these 43- and 46-kDa kinases specifically on LP treatment suggests that their activities might be involved in seedling responses to LP conditions. The sizes of the activated 43- and 46-kDa kinases are similar to MPK3 and MPK6, two MAPKs that have been reported to be activated in response to multiple stresses in Arabidopsis. The mpk3 and mpk6 mutants, MPK3 and MPK6 T-DNA insert null mutants which have been shown to lack MPK3 and MPK6 activities (Xu et al., 2008), were used in our Pi treatment experiments. As shown in Fig. 1(a), LP treatment failed to activate either the 43-kDa kinase in the mpk3 seedlings or the 46-kDa kinase in the mpk6 seedlings. These results confirm that the 43- and 46-kDa kinases are indeed MPK3 and MPK6, respectively. In addition to the MPK3 and MPK6 activities, the transcription levels of the two kinase genes in WT seedlings treated with LP for 24 h were also assessed. The results showed that transcription of both genes was induced by LP treatment (Fig. 1b). The activation of MPK3 and MPK6 kinase activities and the induction of their gene transcription by LP treatment indicate that MPK3- and MPK6- mediated MAPK cascades may be needed for Arabidopsis seedling responses to Pi conditions. Activities of both MPK3 and MPK6 are needed for Pi acquisition by seedlings To explore the roles of MPK3 and MPK6 in the responses of seedlings to Pi, Pi concentrations in shoots and roots of WT, mpk3 and mpk6 seedlings were further measured. The results showed that Pi concentrations (lmol Pi g 1 DW) in both roots and shoots of mpk6 seedlings and in roots of mpk3 seedlings under MS conditions were reduced significantly relative to those in WT seedlings; Pi concentrations in roots of mpk3 and mpk6

4 New Phytologist Research 1149 (a) (b) (c) Fig. 1 Low phosphate (Pi; H 2 PO 4 and HPO 4 2 ) stress activates MPK3 and MPK6, and loss of their activities reduces seedling Pi accumulation. Seven-day-old Arabidopsis Columbia-0 wild-type (WT), mpk3 and mpk6 seedlings germinated on Murashige and Skoog (MS) medium were transferred to Pi-sufficient (MS) or low Pi (LP) medium, and the samples were taken at the indicated times. (a) Kinase activity in WT, mpk3 and mpk6 seedlings was detected by an in-gel kinase assay with myelin basic protein (MBP) as the substrate. Each lane was loaded with 10 lg total protein. Coomassie-stained Rubisc L (large subunit of ribulose-1,5- biphosphate carboxylase) was used as the loading control. CCB, colloidal Coomassie blue. (b) Transcription levels of MPK3 and MPK6 in WT seedlings were monitored by real-time quantitative reverse transcriptionpolymerase chain reaction (Q-PCR). (c) Pi content in shoots and roots of seedlings after transfer into MS or LP medium for an additional 7 d. Data represent the means SD of three biological replicates of each treatment. Statistically significant differences between WT and either mpk3 or mpk6 seedlings (paired t-test): **, P < seedlings under LP conditions were reduced significantly relative to those in WT seedlings (Fig. 1c). Because of the embryonic lethal phenotype of the mpk3/mpk6 double-knockout mutant, we could not include it in our experiments. The reduction in Pi levels in mpk3 and mpk6 seedlings suggests that the activities of MPK3 and MPK6 are needed for Pi accumulation in seedlings. Total Pi content in plants is primarily dependent on the uptake of Pi. Because the Pi content in WT seedlings was higher than that in mpk3 and mpk6 seedlings, Pi uptake by seedlings of WT and the two mutants were further measured. Seven-day-old seedlings, which were initially grown on MS medium, were transferred to either MS or LP medium and pretreated for an additional 3 d. Seedlings were then transferred into a Pi uptake solution with 32 Pi, and Pi uptake by seedlings over a 2-h period was measured. Figure 2 shows the Pi uptake results. After incubation of the seedlings in the Pi uptake solution for 1 h, Pi uptake amounts in seedlings of WT, mpk3 and mpk6 pretreated with MS conditions did not show significant differences; however, Pi uptake in seedlings of WT pretreated with LP conditions was significantly higher than that in mpk3 and mpk6 seedlings. After incubation of the seedlings in the Pi uptake solution for 2 h, Pi uptake amounts in seedlings of WT pretreated with either MS or LP conditions were significantly higher than those in mpk3 and mpk6 seedlings. These data suggest that the loss of MPK3 and MPK6 activities in mpk3 and mpk6 seedlings reduces Pi uptake and thereby may cause the lower Pi accumulation. Plants grown under lower Pi availability have evolved a series of responses to Pi deficiency. The accumulation of anthocyanin in the aerial portion of the plants (visualized as a purple-colored aerial portion) is a common response to lower Pi and other stresses (Dixon & Paiva, 1995; Marschner, 1995). As shown in Fig. S1(a), the aerial portions of WT, mpk3 and mpk6 seedlings turned purple under LP conditions, creating unrecognizable differences between WT and the two mutants. Alterations of the RSA, such as the inhibition of primary root elongation and increase (under moderate Pi deficiency, 50 lm Pi) or decrease (under severe Pi deficiency, 6 lm Pi) of lateral root formation, in plants grown under lower Pi conditions, are adaptive responses to Pi deficiency (Peret et al., 2011; Gruber et al., 2013). The LP conditions used in this study represent severe Pi deficiency conditions (LP medium contains 10 lm Pi). Figure S1 shows that WT, mpk3 and mpk6 seedlings grown under LP conditions have a shorter primary root length relative to seedlings under MS conditions; LP treatment decreases the lateral root density in WT and mpk3 seedlings, and increases the lateral root density in mpk6 seedlings. However, when comparing mpk6 and mpk3 with WT, the two mutant seedlings show shorter primary root length and higher lateral root density under either MS or LP conditions. Previously, M uller et al. (2010) have shown that mpk6 seedlings grown on MS-phytogel medium have a shorter primary root length and slightly decreased lateral root number, whereas Lopez-Bucio et al. (2014) have reported that mpk6 seedlings grown on MS-agar medium have longer primary root length and higher lateral root density. The different results obtained by different groups may be a result of the use of different experimental conditions. The similar phenotypes in mpk3

5 1150 Research New Phytologist Fig. 2 Loss of MPK3 or MPK6 activity in Arabidopsis seedlings reduces phosphate (Pi; H 2 PO 4 and HPO 4 2 ) uptake. Seven-day-old wild-type (WT, closed circles), mpk3 (open squares) and mpk6 (open circles) seedlings germinated on Murashige and Skoog (MS) medium were transferred and pretreated with either Pi-sufficient (MS) or low Pi (LP) medium for an additional 3 d. Seedlings were incubated in Pi uptake solution, and Pi uptake was monitored over a 2-h period. Data represent the means SD of three biological replicates for each treatment. Statistically significant differences between WT seedlings and either mpk3 or mpk6 seedlings (paired t-test): *, P < 0.05; **, P < seedlings may also be explained in the same way. However, the primary root and lateral root phenotypes shown in mpk3 and mpk6 seedlings under both MS and LP conditions suggest that the loss of MPK3 and MPK6 causes RSA changes independent of Pi status in the medium. Activation of MPK3 and MPK6 by MKK9 enhances Pi accumulation The activation of MAPKs requires dual phosphorylation of threonine (Thr) and tyrosine (Tyr) residues in the TXY motif by activated upstream MAPKKs (Cobb & Goldsmith, 1995). Many reports have shown that multiple MKKs, such as MKK4, MKK5 and MKK9, activate both MPK3 and MPK6 in Arabidopsis (Ren et al., 2002; Xu et al., 2008). Previously, Xu et al. (2008) generated transgenic plants carrying mutated MKK9 genes (either MKK9 DD or MKK9 KR ) under the control of a steroid-inducible promoter. MKK9 DD is a constitutively active MKK9 kinase mutant, and MKK9 KR is an inactive MKK9 kinase mutant. In MKK9 DD plants, induction of MKK9 DD, by application of DEX, led to long-lasting activation of MPK3 and MPK6, whereas, in the control (MKK9 KR plants), induction of MKK9 KR could not activate either MPK3 or MPK6. In this study, anthocyanin and Pi concentrations in MKK9 mutant transgenic seedlings were compared. Seven-day-old seedlings, initially grown on MS medium, were transferred to either MS or LP medium in the presence or absence of DEX, and grown for an additional 7 d. Two independent transgenic lines for MKK9 KR and MKK9 DD were used in the experiments, and an empty vector transgenic line (hereby referred to as Vector) was used as the vector control. Color observation showed that the aerial portions of the Vector, MKK9 KR and MKK9 DD seedlings on MS medium, with or without DEX, were all green, whereas Vector, MKK9 KR and MKK9 DD seedlings on LP medium without DEX and Vector and MKK9 KR seedlings on LP medium with DEX were all purple, but MKK9 DD seedlings on LP medium with DEX were green (Fig. 3a). When we further measured the anthocyanin content in seedlings, seedlings grown on MS medium with or without DEX produced lower but comparable levels of anthocyanin; compared with seedlings grown on MS medium with DEX, the anthocyanin content in Vector, MKK9 KR and MKK9 DD seedlings grown on LP medium with DEX increased. However, the anthocyanin content in MKK9 DD seedlings differed significantly from that in Vector and MKK9 KR, with Vector and MKK9 KR seedlings accumulating over 200% anthocyanin relative to MKK9 DD seedlings (Fig. 3b). Pi content results showed that MKK9 DD seedlings grown on either MS or LP medium with DEX accumulated significantly higher levels of Pi than Vector and MKK9 KR seedlings (Fig. 3c,d). RSA changes in different transgenic plants were compared and did not show obvious differences between the genotypes (Fig. S2). DEX treatments did not alter Pi and anthocyanin levels in WT seedlings (Fig. S3). These results suggest that the activation of MKK9 suppresses anthocyanin accumulation under LP conditions and enhances Pi accumulation under both MS and LP conditions. As the activity of MKK9 was shown to be involved in the regulation of anthocyanin and Pi contents, mkk9, a previously identified MKK9 T-DNA insertion mutant (Xu et al., 2008), was further analyzed in this study (Fig. S4). Significantly, induction of MKK9 transcription in WT seedlings by LP treatment suggests the involvement of MKK9 in LP responses; however, anthocyanin and Pi contents in WT and mkk9 seedlings under both MS and LP conditions did not show significant differences. Possibly, other closely related MKKs, such as MKK4 and MKK5, which have been shown previously to regulate ethylene and camalexin production, similar to MKK9, are functionally redundant with MKK9 in the regulation of the seedling response to Pi (Liu et al., 2008; Ren et al., 2008; Xu et al., 2008). Upregulation of MKK4 and MKK5 mrna expression in the mkk9 mutant, as shown by Xu et al. (2008), supports this hypothesis. To explore the roles of MPK3 and MPK6 in the MKK9- induced changes in anthocyanin and Pi concentrations, we further compared MKK9 KR, MKK9 DD, MKK9 DD /mpk3 and MKK9 DD /mpk6 seedlings. On MS medium with DEX, seedlings of MKK9 KR, MKK9 DD, MKK9 DD /mpk3 and MKK9 DD /mpk6 were all green, whereas, on LP medium with DEX, seedlings of MKK9 KR, MKK9 DD /mpk3 and MKK9 DD /mpk6 were purple, but only MKK9 DD seedlings were green (Figs 4a, S5). The four

6 New Phytologist Research 1151 (a) (b) Fig. 3 Induction of active MKK9 mutant (MKK9 DD ) expression in transgenic Arabidopsis plants enhances phosphate (Pi; H 2 PO 4 and HPO 4 2 ) accumulation and suppresses anthocyanin accumulation. Transgenic plants carrying the active MKK9 mutant (MKK9 DD ) transgene, the inactive mutant (MKK9 KR ) transgene or the empty vector (Vector) were analyzed. Two independent homozygous transgenic lines for MKK9 DD and MKK9 KR were included for Pi and anthocynin measurement. Seven-dayold seedlings germinated on Murashige and Skoog (MS) medium were transferred to either Pi-sufficient (MS) or low Pi (LP) medium with or without dexamethasone (DEX) for an additional 7 d. (a) Phenotype comparison of Vector, MKK9 DD and MKK9 KR seedlings. (b) Anthocyanin content in Vector, MKK9 DD and MKK9 KR seedlings transferred to either MS or LP medium with or without DEX. (c) Pi content in shoots and roots of Vector, MKK9 DD and MKK9 KR seedlings transferred to MS medium with or without DEX. (d) Pi content in shoots and roots of Vector, MKK9 DD and MKK9 KR seedlings transferred to LP medium with or without DEX. Data represent the means SD of three biological replicates for each treatment. Statistically significant differences between MKK9 DD seedlings and either Vector or MKK9 KR seedlings (paired t-test): **, P < (c) (d) genotype seedlings grown on MS medium with DEX produced lower levels of anthocyanin than seedlings grown on LP medium with DEX. However, MKK9 KR, MKK9 DD /mpk3 and MKK9 DD / mpk6 seedlings grown on LP medium with DEX accumulated over 190%, 150% and 160%, respectively, of the anthocyanin levels of MKK9 DD seedlings (Fig. 4b). Pi content measurements showed that Pi accumulation in MKK9 DD /mpk3 and MKK9 DD / mpk6 seedlings was significantly reduced compared with that in MKK9 DD seedlings and reached the levels seen in MKK9 KR seedlings (Fig. 4c). These results suggest that full activation of MPK3 and MPK6 by MKK9 is required for both the suppression of anthocyanin production under LP conditions and the elevation of Pi accumulation under either MS or LP conditions. Activation of MKK9-MPK3/MPK6 enhances Pi uptake To further understand how the MKK9-MPK3/MPK6 cascade induces Pi accumulation, we analyzed Pi uptake by MKK9 KR, MKK9 DD, MKK9 DD /mpk3 and MKK9 DD /mpk6 seedlings. Sevenday-old seedlings, which were grown initially on MS medium, were transferred and pretreated with either MS or LP medium plus DEX for an additional day. Seedlings were then transferred into the Pi uptake solution containing 32 Pi plus DEX. Pi uptake by seedlings over a 2-h period was measured. Pi uptake in both MKK9 DD and MKK9 KR seedlings pretreated with MS plus DEX was similar in the first hour; however, Pi uptake by MKK9 DD seedlings was 165% of the uptake by MKK9 KR seedlings by the second hour (Fig. 5a). Pi uptake by MKK9 DD seedlings pretreated with LP plus DEX was significantly higher than that by MKK9 KR seedlings at both tested time points, with Pi uptake by MKK9 DD seedlings being 159% and 143% of the uptake by MKK9 KR seedlings (Fig. 5b). Pi uptake by MKK9 DD / mpk3 and MKK9 DD /mpk6 seedlings was significantly reduced compared with that by MKK9 DD seedlings (Fig. 5c,d). At the first and second time points, MKK9 DD /mpk3 seedlings retained 81% and 67%, respectively, of the Pi taken up by MKK9 DD seedlings pretreated with MS plus DEX, and 60% and 66%, respectively, of the Pi taken up by MKK9 DD seedlings pretreated with LP plus DEX; MKK9 DD /mpk6 seedlings retained 68% and 72%, respectively, of the Pi taken up by MKK9 DD seedlings pretreated with MS plus DEX, and 69% and 86%, respectively, of the Pi taken up by MKK9 DD seedlings pretreated with LP plus DEX. These

7 1152 Research New Phytologist (a) (b) (c) Fig. 4 Activation of endogenous MPK3 and MPK6 is required for MKK9-enhanced phosphate (Pi; H 2 PO 4 and HPO 4 2 ) accumulation and MKK9-suppressed anthocyanin accumulation in Arabidopsis seedlings. Seven-day-old MKK9 KR, MKK9 DD, MKK9 DD /mpk3 and MKK9 DD /mpk6 seedlings, germinated on Murashige and Skoog (MS) medium, were transferred to either Pi-sufficient (MS) or low Pi (LP) medium with dexamethasone (DEX) for an additional 7 d. (a) Phenotype comparison of MKK9 mutant transgenic seedlings and crossed seedlings. (b) Anthocyanin content in MKK9 mutant transgenic seedlings and crossed seedlings transferred to either MS or LP medium with DEX. (c) Pi content in shoots and roots of MKK9 mutant transgenic seedlings and crossed seedlings transferred to either MS or LP medium with DEX. Data represent the means SD of three biological replicates for each treatment. Statistically significant differences between MKK9 DD seedlings and MKK9 KR, MKK9 DD /mpk3 or MKK9 DD /mpk6 seedlings (paired t-test): *, P < 0.05; **, P < data demonstrate that activation of MPK3/MPK6 by MKK9 sufficiently enhances seedling Pi uptake and suggest that increased Pi uptake may be a major contributor to MKK9-MPK3/MPK6 activation-enhanced Pi accumulation in seedlings. Activation of MKK9-MPK3/MPK6 regulates some Piresponsive gene transcription Genome-wide transcriptional analyses have revealed that hundreds of genes in plants are either induced or suppressed by Pi deficiency (Hammond et al., 2003; Wu et al., 2003; Misson et al., 2005). Such genes encode functional proteins to act as enzymes in metabolic pathways, transporters for Pi and other ions, transcription factors, components of signaling pathways and regulators of growth and developmental processes. To fully understand how the activation of MKK9-MPK3/MPK6 regulates Pi responses, we performed a transcriptomics analysis of MKK9 DD seedlings using Affymetrix Arabidopsis Gene Chips, and compared it with that of LP-treated WT seedlings. As shown in Fig. 6, when using a two-fold cut-off value (P < 0.01), 398 genes were upregulated and 69 genes were downregulated in MKK9 DD seedlings after MKK9 DD induction, whereas 434 genes were upregulated and 80 genes were downregulated in WT seedlings treated with LP. Among the regulated genes, 125 of the upregulated genes and 12 of the downregulated genes overlapped in both MKK9 DD and WT seedlings. When the cut-off value was relaxed to 1.5-fold (P < 0.01), the up- and downregulated genes reached 699 and 155, respectively, in MKK9 DD seedlings, and 855 and 157, respectively, in WT seedlings; 301 of the upregulated genes and 25 of the downregulated genes overlapped in both MKK9 DD and WT seedlings. Changes in some previously reported Pi-responsive genes are shown in Table 1. These data suggest that the activation of MKK9-MPK3/MPK6 regulates part of the Pi-responsive gene transcription. The transcription levels of several previously reported Pi-responsive genes in MKK9 KR, MKK9 DD, MKK9 DD /mpk3 and MKK9 DD /mpk6 seedlings were also tested using Q-PCR (Fig. 7). Pht1;1 and Pht1;4, two high-affinity Pi transporters, are involved in Pi acquisition (Mitsukawa et al., 1997; Misson et al., 2004; Shin et al., 2004). As shown in Fig. 7, the transcription levels of Pht1;1 and Pht1;4 were significantly induced in MKK9 DD seedlings under LP conditions and slightly induced in MKK9 DD seedlings under MS conditions by MKK9 DD expression (+DEX) compared with MKK9 KR seedlings. WRKY75, encoding a WRKY family transcription factor, is induced by LP and positively regulates plant Pi uptake (Devaiah et al., 2007). Under both MS and LP conditions, WRKY75 transcription was significantly induced in MKK9 DD seedlings by MKK9 DD expression (+DEX) compared with MKK9 KR seedlings. MYB62, which encodes an R2R3 MYB family transcription factor, has been shown to be induced by LP and positively regulates Pi uptake (Devaiah et al., 2009). Here, transcription of MYB62 was strongly induced in MKK9 DD seedlings under both MS and LP conditions by MKK9 DD expression (+DEX). mir399d, which encodes a microrna that targets and directs the cleavage of PHO2 mrna (PHO2 encodes a ubiquitin-conjugating E2 enzyme), is highly induced by LP

8 New Phytologist Research 1153 (a) (b) (c) (d) Fig. 5 Activation of MKK9 enhances phosphate (Pi; H 2 PO 4 and HPO 4 2 ) uptake through endogenous MPK3 and MPK6 in Arabidopsis seedlings. Sevenday-old MKK9 KR, MKK9 DD, MKK9 DD /mpk3 and MKK9 DD /mpk6 seedlings germinated on Murashige and Skoog (MS) medium were transferred and pretreated with either Pi-sufficient (MS) or low Pi (LP) medium with dexamethasone (DEX) for an additional day. Seedlings were incubated in Pi uptake solution with DEX, and Pi uptake was monitored over a 2-h period. (a) Comparison of Pi uptake between MKK9 DD and MKK9 KR seedlings from MS medium with DEX. (b) Comparison of Pi uptake between MKK9 DD and MKK9 KR seedlings from LP medium with DEX. (c) Comparison of Pi uptake among MKK9 DD, MKK9 DD /mpk3 and MKK9 DD /mpk6 seedlings from MS medium with DEX. (d) Comparison of Pi uptake among MKK9 DD, MKK9 DD /mpk3 and MKK9 DD /mpk6 seedlings from LP medium with DEX. Data represent the means SD of three biological replicates for each treatment. Statistically significant differences between MKK9 DD seedlings and MKK9 KR, MKK9 DD /mpk3 or MKK9 DD /mpk6 seedlings (paired t-test): *, P < 0.05; **, P < (Aung et al., 2006; Bari et al., 2006; Chiou et al., 2006). IPS1 (Induced by Phosphate Starvation 1), a member of the Mt4/TPSI1 gene family that encodes a non-coding RNA that sequesters mir399 and inhibits mir399 cleavage activity, has been reported to be strongly induced by LP (Devaiah et al., 2007; Franco-Zorrilla et al., 2007). At4, another member of the Mt4/ TPSI1 gene family, has been shown to regulate Pi distribution between roots and shoots (Shin et al., 2006). In contrast with the genes induced by MKK9 DD expression, transcription levels of IPS1, mir399 and At4 in this study were strongly suppressed in MKK9 DD seedlings under LP conditions by MKK9 DD expression (+DEX) compared with MKK9 KR seedlings. The induction of Pht1;1, Pht1;4, WRKY75 and MYB62 by MKK9 DD expression was compromised in MKK9 DD /mpk3 and MKK9 DD /mpk6 seedlings, whereas the suppression of IPS1, mir399 and At4 transcription was partially or fully rescued in MKK9 DD /mpk3 and MKK9 DD /mpk6 seedlings. The data suggest that the regulation of these genes by MKK9 activation required both MPK3 and MPK6 activities, and the MKK9-MPK3/MPK6 cascade plays an important role in regulating Pi-responsive gene transcription. Because the accumulation of anthocyanin was suppressed after activation of MPK3/MPK6 by MKK9 (Figs 3, 4), the transcription of DFR and LDOX (also known as ANS), which encode dihydroflavonol 4-reductase and leucoanthocyanidin dioxygenase, respectively, major enzymes in the anthocyanin biosynthetic pathway (Holton & Cornish, 1995; Martens et al., 2010), was analyzed by Q-PCR (Fig. 7). Our results show that the transcription levels of DFR and LDOX were suppressed in MKK9 DD seedlings under LP conditions by MKK9 DD expression (+DEX) compared with their induction in MKK9 KR, MKK9 DD /mpk3 and MKK9 DD /mpk6 seedlings. The gene transcription results agreed well with our color observations and anthocyanin measurement results, suggesting that the suppression of LP-induced anthocyanin biosynthetic gene transcription and anthocyanin accumulation by MKK9 activation requires both MPK3 and MPK6 activities. WRKY75 is required for MKK9-MPK3/MPK6-induced alteration of Pi responses Q-PCR and Genechip results revealed that the transcription of a set of LP-responsive genes is regulated by the activation of MKK9-MPK3/MPK6. Among these genes, MYB62 and WRKY75 have been shown previously to positively regulate Pi uptake in plants (Devaiah et al., 2007, 2009). WRKY33 encodes a WRKY transcription factor that can be phosphorylated by

9 1154 Research New Phytologist Table 1 Comparison of the selected phosphate (Pi; H 2 PO 4 and HPO 4 2 )- responsive gene expression in Arabidopsis wild-type (WT) seedlings after Pi starvation and MKK9 DD seedlings after MKK9 DD induction Relative expression (fold) P < 0.01 Functional categories Gene ID Gene symbol WT LP/MS MKK9 DD +DEX/ DEX Fig. 6 Number of genes upregulated or downregulated in Arabidopsis wild-type (WT) seedlings treated with low phosphate (Pi; H 2 PO 4 and HPO 4 2 ) (LP) and in MKK9 DD seedlings treated with dexamethasone (DEX). MPK3 and MPK6 (Mao et al., 2011). To determine whether MYB62, WRKY33 and WRKY75 are involved in MKK9-MPK3/ MPK6-regulated Pi responses, we initially attempted to generate and analyze MKK9 DD /myb62, MKK9 DD /wrky75 and MKK9 DD / wrky33 mutants. Because we failed to obtain T-DNA insertion homozygous mutants of MYB62, we could not generate the MKK9 DD /myb62 mutant; therefore, MKK9 DD /wrky33 and MKK9 DD /wrky75 seedlings were used in the Pi treatment experiments (Figs S5, S6). As shown in Fig. 8(a), the aerial portions of MKK9 DD, wrky75 and MKK9 DD /wrky75 seedlings were all purple after transfer to LP medium without DEX for an additional 7 d, whereas the aerial portions of MKK9 DD seedlings were green after transfer to LP medium with DEX, but the aerial portions of MKK9 DD /wrky75 and wrky75 seedlings were still purple. The anthocyanin content in seedlings on LP medium with DEX showed that MKK9 DD / wrky75 seedlings produced over 169% of the anthocyanin level that accumulated in MKK9 DD seedlings (Fig. 8b). The anthocyanin content results agreed well with the color observations. Measurements of Pi content showed that the Pi accumulation of MKK9 DD and MKK9 DD /wrky75 seedlings was significantly different; shoots and roots of MKK9 DD /wrky75 seedlings accumulated c. 24% and 22% less Pi, respectively, under MS plus DEX conditions, and 34% and 24% less Pi, respectively, under LP plus DEX conditions than the levels seen in MKK9 DD seedlings (Fig. 8c). The accumulation of Pi and anthocyanin in MKK9 DD / wrky33 seedlings did not show a significant difference compared with that in MKK9 DD seedlings under both MS and LP conditions plus DEX (Fig. S7). Loss of WRKY75 function in MKK9 DD /wrky75 seedlings reduced MKK9 DD -enhanced Pi accumulation and rescued anthocyanin production, suggesting that WRKY75 is a downstream component of MKK9-MPK3/ MPK6 in the regulation of anthocyanin and Pi accumulation. To determine the role of WRKY75 in MKK9-MPK3/MPK6- regulated Pi-responsive gene transcription, gene transcription (see genes in Fig. 7 with the exception of WRKY75) inmkk9 DD / wrky75 seedlings was analyzed by Q-PCR and further compared Sensing and At5g20150 AtSPX signaling At2g45130 AtSPX Transport At5g43350 Pht1: At2g38940 Pht1: At2g32830 Pht1: At1g76430 Pht1: ** At3g52190 PHF At1g73220 AtOCT Transcription factor At1g71130 AtERF At1g68320 MYB At5g04340 ZAT At5g13080 WRKY At1g13300 HRS At1g62300 WRKY Others At1g68740 PHO1;H At5g09470 DIC At1g08650 AtPPCK Scavenging/ mobilization of Pi At1g56650 AtPAP At1g52940 AtPAP ** At2g01880 AtPAP At2g16430 AtPAP At1g08310 AtPAP At2g27190 AtPAP At2g46880 AtPAP At3g17790 AtPAP At4g13700 AtPAP ** At4g24890 AtPAP At4g36350 AtPAP ** At1g73010 AtPS At1g17710 AtPECP At2g02990 RNS Metabolism At4g15210 BAM At5g42800 DFR At1g56650 CHS At5g54060 UF3GT At3g51240 F3H Seven-day-old WT seedlings were transferred to Murashige and Skoog (MS) or low Pi (LP) medium for 3 d, and MKK9 DD seedlings were transferred to MS medium with or without dexamethasone (DEX) for 1 d. **, P > with MKK9 DD seedlings (Fig. 9). The results revealed that the transcription levels of Pht1;1, Pht1;4 and MYB62, which were induced in MKK9 DD seedlings under both LP and MS conditions by MKK9 DD expression (+DEX), were reduced to varying degrees in MKK9 DD /wrky75 seedlings; IPS1, DFR and LDOX transcription, which was suppressed in MKK9 DD seedlings under LP conditions by MKK9 DD expression (+DEX), was either partially or fully rescued in MKK9 DD /wrky75 seedlings, whereas IPS1 and LDOX transcription levels in MKK9 DD /wrky75 seedlings under MS with DEX conditions did not show any obvious change, and DFR transcription was reduced in MKK9 DD /wrky75 seedlings compared with MKK9 DD seedlings. mir399d

10 New Phytologist Research 1155 Fig. 7 Transcription detection of multiple phosphate (Pi; H 2 PO 4 and HPO 4 2 )- responsive genes in MKK9 mutant transgenic and crossed Arabidopsis seedlings. Sevenday-old MKK9 mutant transgenic seedlings and crossed seedlings, germinated on Murashige and Skoog (MS) medium, were transferred to either Pi-sufficient (MS; open bars) or low Pi (LP; closed bars) medium with dexamethasone (DEX) for an additional 3 d. Transcription levels of Pht1;1, Pht1;4, WRKY75, MYB62, IPS1, At4, mir399d, DFR and LDOX in seedlings were monitored by real-time quantitative reverse transcription-polymerase chain reaction (Q-PCR). Data are the means SD of three biological replicates for each treatment, and represent fold changes normalized to the transcription levels of the MKK9 mutant transgenic seedlings or crossed seedlings in MS medium. transcription, which was suppressed in MKK9 DD seedlings under both LP and MS conditions by MKK9 DD expression (+DEX), could not be rescued in MKK9 DD /wrky75 seedlings; At4 transcription, which was suppressed in MKK9 DD seedlings under both LP and MS conditions by MKK9 DD expression (+DEX), was decreased in MKK9 DD /wrky75 seedlings under LP with DEX conditions and increased in MKK9 DD /wrky75 seedlings under MS with DEX conditions compared with MKK9 DD seedlings. These data suggest that the regulation of the Pi-responsive transcription of certain genes by MKK9-MPK3/MPK6 requires WRKY75. Discussion On limited Pi availability in soil, plants have evolved adaptive mechanisms to sense Pi availability, transduce the signal and subsequently promote a series of responses that maximize Pi acquisition, alter Pi utilization and facilitate survival under Pi-limited conditions (Marschner, 1995; Raghothama, 1999; Poirier & Bucher, 2002; Yuan & Liu, 2008; Lin et al., 2009; Rouached et al., 2010; Yang & Finnegan, 2010; Chiou & Lin, 2011; Kuo & Chiou, 2011; Peret et al., 2011). The transcriptional and/or post-transcriptional regulation of many Pi starvation-responsive (PSR) genes and the post-translational regulation of proteins have been well documented in many plant species in response to Pi availability; however, the Pi sensor(s)/receptor(s) and the signaling pathway(s) in plants are not fully understood. In this study, our results reveal that the MAPK signaling cascade, MKK9- MPK3/MPK6, plays an important role in the regulation of Pi responses in Arabidopsis. MAPK cascades have been reported to be important modules for the perception and transduction of signals from receptors or sensors for the induction of intracellular responses (Colcombet & Hirt, 2008; Pitzschke et al., 2009; Yoo et al., 2009; Andreasson & Ellis, 2010; Rodriguez et al., 2010). In Arabidopsis, activation of MPK3 and/or MPK6 is involved in the regulation of plant responses to multiple stresses, such as pathogen infection (Asai et al., 2002; Ren et al., 2008; Beckers et al., 2009; Galletti et al., 2011), dehydration (Xu & Chua, 2012), salt (Yu et al., 2010), cold (Teige et al., 2004), heat shock (Li et al., 2012), UV-B (Gonzalez Besteiro et al., 2011) and oxidative stress (Lee & Ellis, 2007; Wang P et al., 2010). The increase in MPK3 and MPK6 activities and their gene transcription in seedlings induced by LP (Fig. 1) suggest that these MAPKs may also mediate the regulation of seedling responses to LP conditions. This finding was further supported by the following evidence: compared with WT seedlings, mpk3 and mpk6 mutants accumulated less Pi under LP conditions and had lower levels of Pi uptake under both MS and LP conditions (Figs 1, 2). After transgene induction, MKK9 DD seedlings with overly activated MPK3 and MPK6 accumulated less anthocyanin under LP conditions, and accumulated more Pi under both LP and MS conditions, relative to MKK9 KR seedlings (Fig. 3). Activation of MPK3 and MPK6 in MKK9 DD seedlings enhanced Pi uptake, whereas loss of MPK activity in either MKK9 DD /mpk3 or MKK9 DD /mpk6 seedlings reduced Pi uptake (Fig. 5). Activation of MPK3 and MPK6 by MKK9 DD regulated the transcription of a set of LP-responsive genes in seedlings under either LP or MS conditions (Figs 6, 7). Furthermore, previous genome-wide transcriptional analyses, which have reported the induction of MKKK14, MKKK15, Raf43 and Raf35 in Arabidopsis (Misson et al., 2005), and MAPKK2 and WIPK in common bean (Phaseolus vulgaris L.) (Graham et al., 2006), also support the involvement of MAPK cascades in the Pi stress response. Previous studies have demonstrated that activation of MKK9- MPK3/MPK6 induces camalexin production (Xu et al., 2008) and WRKY33 is required for camalexin induction (Mao et al., 2011). Currently, there is no evidence for a correlation between camalexin biosynthesis and Pi responses; thus, we detected anthocyanin and Pi contents in MKK9 DD /wrky33 seedlings, which

11 1156 Research New Phytologist (a) (b) (c) Fig. 8 MKK9-MPK3/MPK6 enhanced phosphate (Pi; H 2 PO 4 and HPO 4 2 ) accumulation and suppressed anthocyanin production through WRKY75 in Arabidopsis seedlings. Seven-day-old MKK9 DD transgenic seedlings and crossed seedlings, germinated on Murashige and Skoog (MS) medium, were transferred to low Pi (LP) medium either with or without dexamethasone (DEX) for an additional 7 d. (a) Phenotype comparison of MKK9 DD seedlings and crossed seedlings transferred to LP medium either with or without DEX. (b) Anthocyanin content in MKK9 DD seedlings and crossed seedlings transferred to LP medium with DEX. (c) Pi content in shoots and roots of MKK9 DD seedlings and crossed seedlings transferred to LP medium with DEX. Data represent the means SD of three biological replicates for each treatment. Statistically significant differences between MKK9 DD /wrky75 seedlings and either MKK9 DD or wrky75 seedlings (paired t-test): **, P < produce a rare camalexin after MPK3/MPK6 activation (Ren et al., 2008; Xu et al., 2008; Mao et al., 2011). However, the comparable anthocyanin and Pi contents in MKK9 DD and MKK9 DD /wrky33 seedlings suggest that MKK9-MPK3/MPK6- induced Pi responses are independent of camalexin production (Fig. S7). WRKY75, which encodes a WRKY family transcription factor, can be induced by LP, pathogen infection and senescence (Robatzek & Somssich, 2001, 2002; Guo et al., 2004; Devaiah et al., 2007). Under LP conditions, downregulation of WRKY75 transcription in WRKY75 RNAi plants has been reported to reduce Pi uptake and PSR gene induction, and increase anthocyanin accumulation (Devaiah et al., 2007). These results suggest that WRKY75 is involved in the regulation of LP responses in Arabidopsis (Devaiah et al., 2007). Here, we found that LP-induced WRKY75 transcription was enhanced by MKK9- MPK3/MPK6 activation (Fig. 7; Table 1). To determine whether WRKY75 is a functional downstream component of the MKK9- MPK3/MPK6 cascade in regulating LP responses, the MKK9 DD / wrky75 mutant was analyzed. To avoid the effect of a lower level of WRKY75 mrna retained in the RNAi plants (Devaiah et al., 2007), a T-DNA insertion mutant of WRKY75 was crossed to MKK9 DD. Interestingly, the loss of WRKY75 function in the MKK9 DD background (MKK9 DD /wrky75) reduced the transcription of Pi uptake-related genes (including Pht1;1, Pht1;4 and MYB62 in our experiments) and MKK9-MPK3/MPK6- enhanced Pi accumulation. In addition, the transcription of anthocyanin biosynthetic genes (e.g. LDOX and DFR) and anthocyanin accumulation, which were suppressed by the activation of MKK9-MPK3/MPK6, were rescued in MKK9 DD /wrky75

12 New Phytologist Research 1157 Fig. 9 Transcription detection of multiple phosphate (Pi; H 2 PO 4 and HPO 4 2 )- responsive genes in MKK9 DD and MKK9 DD / wrky75 Arabidopsis seedlings. Seven-dayold MKK9 DD and MKK9 DD /wrky75 seedlings, germinated on 0.59 Murashige and Skoog (MS) medium, were transferred to either Pi-sufficient (MS; open bars) or low Pi (LP; closed bars) medium with dexamethasone (DEX) for an additional 3 d. Transcription levels of Pht1;1, Pht1;4, MYB62, IPS1, At4, mir399d, DFR and LDOX in seedlings were monitored by realtime quantitative reverse transcriptionpolymerase chain reaction (Q-PCR). Data are the means SD of three biological replicates for each treatment, and represent fold changes normalized to the transcription levels of the MKK9 DD or MKK9 DD /wrky75 seedlings in MS medium. seedlings. These results suggest that MKK9-MPK3/MPK6 regulates both anthocynin production and Pi accumulation through WRKY75. MAPK phosphorylation sites generally contain a P-X-S/T-P sequence (Alvarez et al., 1991; Clark-Lewis et al., 1991; Sorensson et al., 2012). The putative WRKY75 protein, however, does not contain the P-X-S/T-P sequence, and MKK9-MPK3/MPK6 cannot phosphorylate recombinant WRKY75 protein (data not shown), thereby excluding WRKY75 as a MAPK substrate, and further suggesting that WRKY75 is genetically downstream of MKK9-MPK3/MPK6 during the regulation of Pi responses. As WRKY75 is not a substrate of MAPK, but transcription of WRKY75 can be induced by MKK9-MPK3/MPK6, we propose that MKK9-MPK3/MPK6 may phosphorylate an unrevealed transcription factor TFa, subsequently promote the transcription of WRKY75 and thereby regulate Pi responses. As PHR1 has been reported to be a key component in the regulation of transcription of a large proportion of Pi-responsive genes and positively regulates Pi accumulation (Rubio et al., 2001; Bustos et al., 2010), we speculate that it may be the TFa candidate. However, the following reasons exclude PHR1 as the candidate (Fig. S8). First, PHR1 did not interact with MPK3 and MPK6 in a yeast two-hybrid system; second, PHR1 could not be successfully phosphorylated by MPK3 and MPK6 (only an extremely weak band was shown compared with the well-used MAPK substrate MBP protein); third, loss of PHR1 function in MKK9 DD /phr1 double mutant did not affect MKK9-MPK3/ MPK6 activation-enhanced Pi accumulation under MS conditions. Based on previous reports and our data, we propose a working model for Pi responses in plants (Fig 10). When the availability of Pi in environment is altered, plants sense the varied Pi state by an unidentified sensor or receptor. The Pi transporter PHO84 in yeast (Giots et al., 2003) and nitrate transporter CHL1 in Arabidopsis (Ho et al., 2009) also function as Pi sensors (called Fig. 10 Proposed model for phosphate (Pi; H 2 PO 4 and HPO 4 2 ) sensing, signaling and responses in Arabidopsis plants. transceptors). Whether the Pi transporters in plants also act as transceptors awaits elucidation. The active sensor/receptor activates either the MAPK cascade or a MAPK-independent signaling module. MKKKx-MKK9-MPK3/MPK6 was suggested as the MAPK cascade in our study. The sensor/receptor can activate the MKKKx-MKK9-MPK3/MPK6 cascade directly or through an additional mediator. MPK3 and MPK6 phosphorylate an unknown transcription factor TFa, thereby promoting the transcription of the WRKY75 gene, and subsequently induce the adaptive responses to facilitate Pi acquisition. Identification of TFa and MKKKx will facilitate the understanding of the mechanism by which MKK9-MPK3/MPK6 regulates Pi responses. Acknowledgements This work was supported by grants from the State Basic Research Program (2012CB114200) and the National Natural Science Foundation of China ( and ) to D.R.; the