OsWRKY74 , a WRKY transcription factor, modulates tolerance to phosphate starvation in rice

1Phosphorus (P) is one of the essential macronutrients for plant growth and development. It is a constituent of key molecules such as ATP, nucleic acids and phospholipids. Plants take up P exclusively in the form of inorganic phosphate (Pi). Although the overall P content in soil is generally high, in many natural and agricultural ecosystems, plants often have to face conditions in which availability of Pi is at lower extremities. To cope with Pi deficiency, plants have evolved numerous strategies to optimize Pi acquisition from soil and its distribution to different organs and sub-cellular compartments. For example, stimulation of lateral roots and root hairs leads to profound changes in root system architecture to maximize root surface area for Pi uptake under Pi-deficient conditions. Exudation of organic anions and phosphatases, as well as acidification of the rhizosphere have been used to liberate plant available P by solubilizing Pi bounded to soil particles. In addition, Pi-starved plants can regulate multiple metabolic processes to reprioritize utilization of internal Pi and maximize acquisition of external Pi to adapt to low Pi environments. The complex network of regulatory genes necessary to sense and respond to Pi deficiency is being dissected. For example, in Arabidopsis thaliana, a major transcriptional regulatory system that involves PHR1, SIZ1, miR399 and PHO2 in response to Pi deficiency has been identified. In contrast to Arabidopsis, recent studies have shown that the PHR1-miR399-PHO2 signalling pathway also operates in rice plants in response to Pi deficiency. For instance, demonstrated that OsPHR2, the homologue of AtPHR1, is a key regulator involved in Pi starvation signalling in rice. OsSPX1 is associated with Pi homeostasis and suppresses the function of OsPHR2 by regulating OsPT2 expression. The PHR1-miR399-PHO2 pathway is a central component of the Pi starvation response, but several lines of evidence demonstrate that some Pi-responsive transcriptional factors (TFs) are not involved in the PHR1-miR399-PHO2 pathway. These include OsPTF1 in rice, and MYB62, WRKY75, ZAT6 and BHLH32 in Arabidopsis. These TFs function in crosstalk between Pi-starvation signalling and signalling cascades associated with phytohormones and photosynthates, to govern physiological responses to Pi limitation.

2WRKY TFs are a large family of regulatory proteins in plants. The most prominent feature of these proteins is the presence of the WRKY domain, which is a 60 amino acid region with a strongly conserved amino acid sequence WRKYGQK in its N-terminal, and a novel potential C-C-H-H/C zinc-finger motif in its C-terminal. Based on the number of WRKY domains and the type of their zinc-finger motif, WRKY proteins are classified into three distinct groups. Compared with other multigene families of plant TFs, the percentage of WRKY gene family members that are responsive to biotic stress is relatively high, implying that WRKY proteins may play key roles in the regulation of biotic stresses. For example, 49 out of 72 examined WRKY genes in Arabidopsis are responsive to bacterial infection or salicylic acid (SA) treatment, and the majority of WRKY genes within group III of the WRKY family are responsive to both SA and pathogen infection. In contrast to biotic stresses, little information is known about the roles of WRKY proteins in plant responses to abiotic stress in general and Pi starvation in particular. In Arabidopsis, only four WRKY proteins, i.e. AtWRKY75, AtWRKY6, AtWRKY45 and AtWRKY42, have been reported to be involved in Pi starvation. For example, WRKY75 acts as a positive regulator of Pi stress responses and RNAi suppression of WRKY75 results in impaired Pi starvation responses in Arabidopsis. AtWRKY6 negatively regulates Pi starvation response by modulating Arabidopsis PHOSPHATE1 (PHO1) expression. Moreover, AtWRKY6 has been demonstrated to be positively involved in the regulation of response to boron deficiency. WRKY45 regulates Pi uptake by modulating expression of PHT1;1 in Arabidopsis. WRKY42, a homologue of WRKY6 in Arabidopsis, modulates Pi homeostasis by regulating the expression of PHO1 and PHT1;1.

3Rice is a model monocot plant and one of the most important food crops in Asia. Rice growth, development and productivity are seriously affected by Pi availability in many areas worldwide, but the Pi starvation signalling pathway is still largely unknown in rice. In the present study, we identified a WRKY TF belonging to group III of the WRKY family, designated OsWRKY74, in rice. Our results demonstrated that overexpression of OsWRKYP74 in rice conferred the transgenic plants greater tolerance to low Pi stress by activating Pi starvation-induced genes and modulating root system architecture. In addition to the involvement of OsWRKY74 in multiple Pi starvation responses, analysis of transgenic plants with overexpressing and RNAi OsWRKY74 provided direct evidence that OsWRKY74 was also involved in Fe deficiency and cold stress. These results demonstrate that OsWRKY74 participates in the regulation of multiple nutrient starvation responses and cold stress, highlighting the possible roles of OsWRKY74 in crosstalk between P and Fe, and P and cold stress.

4Japonica rice cv. Zhonghua 10 was used in physiological experiments and rice transformation throughout this study. For hydroponic culture of the seedlings, rice seeds were surface sterilized for 5min with ethanol (75%, v/v) and for 10min with commercially diluted (1:3, v/v) NaClO, followed by several rinses with sterile water. Seeds were germinated in the dark at 28°C for 72h. Thereafter seedlings were grown in a greenhouse. Then, the 7-d-old seedlings were transferred to nutrient solution containing 1.425mM NH4NO3, 0.513mM K2SO4, 0.998mM CaCl2, 1.643mM MgSO4, 0.168mM Na2SiO3, 0.125mM Fe-EDTA, 0.019mM H3BO3, 0.009mM MnCl2, 0.155mM CuSO4, 0.152mM ZnSO4 and 0.075mM Na2MoO4, pH 5.5, supplemented with 0.323mM NaH2PO4 (HP) or 0.016mM NaH2PO4 (LP). The hydroponic experiments were carried out in a growth room with a 16-h-light (30°C)/8-h-dark (22°C) photoperiod and the relative humidity was controlled at ~70%. The solution was refreshed every 3 d.

5To minimize recycling of P from seed endosperm, the seed endosperm was removed prior to transfer of rice seedlings to Pi-deficient medium. The optimal time and concentration used for the low-Pi stress were determined following protocols described by. The concentration of Pi deficiency was set at 0.016mM throughout this study. One-week-old wild-type (WT) and transgenic rice plants were exposed to the low-Pi solution (0.016mM Pi) for 30 d or 14 d. For analyses of root system architecture and RT-PCR, rice seedlings grown in the low-Pi (0.016mM Pi) solution for 14 d were used. P concentration, shoot biomass and root biomass were measured after 30 d of Pi starvation.

6To determine the effect of deprivation of other mineral nutrients, including nitrogen (N), iron (Fe), and potassium (K) on OsWRKY74 expression, 1-week-old WT and transgenic rice plants were exposed to solution containing no nitrogen (−N), no potassium (−K), and no iron (–Fe), respectively. Plants were harvested for RNA extraction after the treatments for varying periods (0, 6 and 12h; 1, 3, 5 and 7 d).

7For the treatment of salt and osmotic stress, seedlings were exposed to solution containing 200mM NaCl or 15% PEG 6000 and leaves were sampled after 5h, respectively. For cold stress, seedlings were transferred to a growth chamber at 4°C for 5, 12, 24 and 72h, and sampled for further analysis.

8For pot experiments in soil, the experiments were performed in an experimental field of the Institute of Botany, Chinese Academy of Sciences. At 14 d after germination (DAG), WT and transgenic seedlings were transferred into pots with two Pi levels; 60mg Pi kg-1 soil as KH2PO4 and 15mg Pi kg -1 soil. Each pot received the equivalent of 200 mgN kg-1 soil (as urea) and 130mg K2O kg-1 (as K2SO4).

9Three biological replicates, each comprising five individual plants, were used for quantitative real-time PCR. Total RNA was extracted using Trizol reagent (Invitrogen). 2 μg of total RNA was treated with DNAase I (Promega) and then transcribed in a total volume of 20 μl with 1 μg oligo (Dt)18, 10mM deoxynucleotide triphosphate, and 200 units SuperScriptsTM II reverse transcriptase (Invitrogen). The cDNA samples were diluted to 2 and 8ng μl-1. Triplicate quantitative assays were performed on 1 μl of each cDNA dilution with the SYBR Green Master Mix or TaqMan reagents (TaKaRa) and an ABI 7900 sequence detection system according to the manufacturer’s protocol (Applied Biosystems). The relative quantification method (Delta-Delta cycle threshold) was used to evaluate quantitative variation between the replicates examined. The PCR signals were normalized to those of Actin or rice polyubiquitin1 (RubQ1). All the primers used for the quantitative RT-PCR (RT-qPCR) are listed in Supplementary Tables S1, S2 at JXB online. Detection and quantification of mature miR399 were performed as previously described, Briefly, RNA was reversely transcribed using One Step PrimeScript miRNA cDNA Synthesis Kit (TaKaRa). This kit adds poly (A) to the 3′ end of miRNAs and starts to reverse transcribe. The reverse transcription was led by a kind of special oligo-dT ligated with a known sequence at its 5′ end. RT-qPCR was performed using SYBR Premix Ex Tag II (TaKaRa).

10The whole coding sequence of OsWRKY74 was amplified with two primers, 5′-GCTCTAGAATGGAGAGCATGGAGGGC-3′ (XbaI site underlined) and 5′-CGG GTACCTGCGAAGAAG CTGGTGATATC-3′ (KpnI site underlined). The PCR product was subcloned into the pBI221 vector to generate pBI221- OsWRKY74-GFP, containing an OsWRKY74-GFP fusion construct under the control of the CaMV 35S promoter. The construct was confirmed by sequencing and used for transient transformation of onion (Allium cepa) epidermis via a gene gun (Bio-Rad). Transformed onion cells were observed with a confocal microscope (Nikon).

11For OsWRKY74 RNAi, a fragment of 337bp was amplified from OsWRKY74 with two primers, 5′-GGGGTACCACTA GTATGGAGAGCATGGAGGG-3′ (Kpn I and Spe I sites underlined) and 5′-CGGGATCCGAGCTCTAATCTGATGCCTCTTC-3′ (BamH I and Sac I sites underlined), containing two restriction enzymes at their 5′ ends, respectively. The hairpin structure consisting of an antisense OsWRKY74 fragment, a rice intron and an OsWRKY74 sense fragment were inserted between the maize ubiquitin promoter and the nopaline synthase terminator of the vector pTCK303 (Supplementary Fig. S1A). For OsWRKY74 overexpression, the full-length cDNA of OsWRKY74 was amplified using two primers, 5′-CGCGGATCCATGGAGAGCATGGAGGGCAATGG-3′, (BamH I site underlined) and 5′-CGGGGTACCTCATGC GAAGAAGCTGGTGAT-3′, (Kpn I site underlined), by RT-PCR with Pyrobest DNA Polymerase (TaKaRa), directionally cloned into the KpnI-BamHI sites of a pUN1301 under the control of the ubiquitin promoter (Supplementary Fig. S1B). These constructs were electroporated into Agrobacterium tumefaciens EHA105 and transformed into rice (Zhonghua 10). Generation of transgenic rice plants was performed as described by. T2 and T3 seeds were used for subsequent experiments.

12Genomic DNA isolated from 2-week-old rice seedlings was digested with Hind Ш, fractioned electrophoretically on 0.8% (w/v) agarose gel, and blotted onto a nylon membrane (Amersham Pharmacia Biotech). α-32P-ATP- and CTP-labelled GUS amplified from pUN1301 was used as a probe for hybridization. The membrane was exposed to x-ray film (Eastman-Kodak) at −70°C for 3–7 d.

13The dry root and shoot samples were separated and digested with concentrated nitric acid and hydrogen peroxide, and total P and Fe were determined by using inductively coupled plasma mass spectrometry following the protocols described by.

14Root APase staining was analysed according to. The roots were excised from 14-d-old Pi-supplied and Pi-deprived seedlings and incubated with a 5-bromo-4-chloro-3-indolyl-phosphate (BCIP) agar overlay solution containing 50mM sodium acetate (pH 5.5) with 10mM MgCl2, 0.6% agar and 0.08% BCIP at room temperature for 20min. The blue colour on the root surface, formed by hydrolysis of BCIP, was photographed using a PENTAX k-7 camera (Pentax Corporation, Tokyo, Japan).

15Protein was isolated with ice-cold extraction buffer (100mM potassium acetate, pH 5.5, 20mM CaCl2, 2mM EDTA, 1mM dithiothreitol, 0.1mM phenylmethylsulfonyl fluoride and 1.5% (w/v) polyvinylpolypyrrolidone) from 0.5mg of roots of 14-d-old seedlings. The protein content was determined using the method of, with BSA as an internal standard. APase activity was analysed as described by. Acid phosphatase activity was assayed by adding 1 μg of protein to 620 μl of reaction buffer (50mM NaAc pH 5.5 and 10mM MgCl2), and 10 ul of p-nitrophenol phosphate (10mg ml-1 pNPP; Sigma). After incubation at 37°C for 10min, the reaction was stopped by 1.2ml of 1M NaOH, and then absorbance was measured at 412nm wavelength. Phosphatase activity was expressed as ng of pNPP accumulated μg–1 soluble protein min–1. These experiments were replicated three times.

16For statistical analyses, the SPSS Statistics Base software package (version 16) was used. Significant differences were evaluated using one-way ANOVA and Duncan’s test at P≤0.05.

17OsWRKY74 (LOC_Os09g16510) was identified from a low Pi-responsive rice microarray. A Blastp search revealed that OsWRKY74 protein had a highly conserved WRKYGQK motif and a characteristic C2-HC zinc finger motif (Fig. 1A). Therefore, this gene belongs to group III of the WRKY TFs. A search of PROSITE (http://www.expasy.org/prosite) revealed that the OsWRKY74 protein contained 13 potential protein kinase phosphorylation sites and an Ala-rich domain in its C-terminal (Supplementary Fig. S2).

18Most WRKY proteins studied so far have been implicated in regulating biotic stress responses. However, recent studies revealed that WRKY TFs are closely associated with abiotic stresses such as nutrient deficiency, cold, salt and drought stresses. To better understand the role of OsWRKY74 under conditions of various abiotic stresses (cold, drought, salinity or Pi starvation), we analysed WRKY proteins involved in abiotic stress from rice based on the report by, and WRKY proteins have been characterized for their functions in tolerance to cold stress and Pi starvation in rice and Arabidopsis. Subsequently, a phylogenetic tree was constructed using the MEGA 6.0 program. Phylogenetic analysis revealed that OsWRKY74 was not grouped with AtWRKY42, AtWRKY45, AtWRKY6 and AtWRKY75 WRKY proteins involved in Pi homeostasis; rather, it formed a separate branch with OsWRKY69 belonging to group III of the WRKY proteins with unknown function (Fig. 1B), although OsWRKY69 was induced by drought and salt stress.

19To determine its subcellular localization, OsWRKY74 was fused in frame to a 5′ terminus of the GFP reporter gene under the control of the cauliflower mosaic virus 35S (CaMV 35S) promoter. The recombinant constructs of the OsWRKY74-GFP fusion gene and GFP alone were introduced into onion (Allium cepa) epidermal cells by the particle bombardment. The OsWRKY74-GFP fusion protein accumulated mainly in the nucleus, whereas GFP alone was present throughout the whole cell (Fig. 1C), suggesting that OsWRKY74 is a nucleus-localized protein. This result is also consistent with the predicted function of OsWRKY74 as a TF.

20Quantitative real-time RT-PCR analysis showed that OsWRKY74 was expressed in all organs examined, with the highest expression in roots and lowest in flowers (Fig. 2A). The expression patterns of OsWRKY74 under Pi-sufficient and Pi-deficient conditions were evaluated by real-time RT-PCR using RNA samples extracted from roots and leaves. As shown in Fig. 2B, Pi starvation-induced expression of OsWRKY74 was observed in roots and leaves up to 7 d. The increases in OsWRKY74 transcripts peaked at 6h of Pi starvation and displayed a gradually decline thereafter (Fig. 2B). To determine whether the up-regulation of OsWRKY74 gene was specific to Pi-starvation, the responsiveness of OsWRKY74 expression to deprivation of other mineral nutrients, including nitrogen (N), potassium (K), and iron (Fe), was also investigated. Similar to Pi deprivation, the expression of OsWRKY74 was markedly enhanced by deprivation of Fe (Fig. 2C). In contrast, the expression of OsWRKY74 was suppressed by deprivation of N (Fig. 2D). No significant difference in OsWRKY74 mRNA by deprivation of K up to 7 d was observed (Supplementary Fig. S3).

21The transcript of OsWRKY74 was down-regulated after 5h of cold treatment and the down-regulation lasted up to 72h of cold treatment (Fig. 2E). To further validate this experiment, we used MYBS3 (Fig. 2F), a gene encoding MYB protein from rice, as a positive control. MYBS3 is induced by cold stress. In contrast, no response of OsWRKY74 was detected when treated with salt and dehydration stress for 5h (Supplementary Fig. S4). Taken together, these results suggest that OsWRKY74 is induced by deficiency of Pi and Fe, while it is suppressed by N deficiency and cold stress.

22To investigate the function of OsWRKY74 in planta, we overexpressed and suppressed OsWRKY74 in rice under the control of a ubiquitin promoter of maize. Transgenic rice lines of OsWRKY74 were confirmed by hygromycin selection and Southern blotting. Southern blotting was performed using the DNA digested with Hind III and the GUS gene as a probe. Two overexpressed lines and two RNA interference (RNAi) lines were randomly selected, and different hybridized patterns to the GUS probe were observed. In contrast, no signals were detected in WT rice plants under the same conditions (Fig. 3A). Therefore, the two overexpressed transgenic lines and two RNAi transgenic lines are likely to be independent. Furthermore, real-time PCR analysis showed that expression of OsWRKY74 was markedly increased in the two independent overexpressing lines, while its expression was suppressed in the two RNAi transgenic lines (Fig. 3B).

23To examine the phenotypes of transgenic lines, T3 transgenic lines and the WT were grown in a greenhouse under identical conditions. Seedlings of both overexpressing and knockdown lines showed comparable phenotypes to WT plants under non-stressed, control conditions (Fig. 3C), suggesting that alteration of OsWRKY74 expression has no impacts on their phenotypes under normal, non-stressed conditions.

24To functionally characterize the role of OsWRKY74 in response and adaptation to Pi starvation, 1-week-old plants of T3 transgenic lines and the WT were exposed to a hydroponic solution containing a high level of Pi (HP; 0.323mM Pi) and a low level of Pi (LP; 0.016mM Pi) for 30 d. In the hydroponic culture solution with a high-Pi level, no significant differences in plant height, root and shoot biomass were observed between WT and transgenic plants (Fig. 4A and Table 1). With a low-Pi level, however, the OsWRKY74-overexpressing lines (OE-1 and OE-2) exhibited better growth than WT plants, as evidenced by the greater plant height, P concentration, root and shoot biomass of the transgenic plants compared to the WT plants (Fig. 4 and Table 1). In contrast to the overexpressing lines, growth of the RNAi transgenic lines (Ri-4 and Ri-6) was less than that of the WT plants, as evidenced by the shorter plant height, and lower P concentration, shoot and root biomass of the knockdown transgenic plants compared to the WT plants when grown in LP medium (Fig. 4 and Table 1). These results suggest that interference of OsWRKY74 renders rice seedlings more sensitive to Pi deficiency.

25To further confirm the tolerance of OsWRKY74 in response to Pi deficiency, the transgenic lines and WT plants were grown in soil with two levels of Pi supply (high and low Pi soils contained 60 and 15mg Pi kg-1 soil, respectively) for the entire growth period (~150 d). In the low-Pi soils, plant height, tiller number, shoot and root biomass, grain weight, and P concentration of the OsWRKY74-overexpressing plants were about 13%, 45%, 37%, 35%, 33% and 24% higher than those in WT plants, respectively (Fig. 5 and Table 2). In contrast, the RNAi knockdown lines exhibited greater growth inhibition than WT plants. For instance, the OsWRKY74 RNAi transgenic lines had fewer tillers (1.7 per plant on average) than WT plants (1.2 per plant on average) in the whole growth period. In addition, the grain weight, P concentration and shoot biomass were also lower in the RNAi lines than in the WT (Fig. 5 and Table 2). In the high-Pi soils, however, no significant differences in tiller number, shoot biomass, grain weight and plant height were observed between WT and transgenic plants (Fig. 5 and Table 2). Taken together, these results suggest that the expression level of OsWRKY74 in rice plants is positively correlated with tolerance to low Pi stress.

26An increase in activity of root acid phosphatases (APase) is a common phenomenon in the response of plants to Pi starvation. To test whether APase activity is involved in the OsWRKY74-mediated responses to Pi deficiency, APase activities in roots of WT and transgenic plants were visualized in plants grown in both HP and LP media. The APase activities in OsWRKY74-overexpressing roots were much higher than in the WT and RNAi plants under both HP and LP conditions, as indicated by the stronger blue staining of OsWRKY74-overexpressing roots than that of WT and RNAi plants (Fig. 6A). Higher APase activities in the OsWRKY74-overexpressing roots than in WT and RNAi roots in HP and LP media were also obtained by determining hydrolytic activity in roots on substrate pNPP (Fig. 6B). Consistently, the gene coding for purple acid phosphatase10a (OsPAP10a), was also up-regulated in OsWRKY74-overexpressing roots in HP and LP solutions (Fig. 6C). Together, these results indicate that OsWRKY74 can stimulate acid phosphatases to facilitate Pi acquisition.

27Root system architecture (RSA) is an important root trait that is sensitive to Pi status in growth medium. To test whether the greater growth of transgenic rice plants grown in LP medium is related to changes in the root system architecture, 1-week-old WT and transgenic plants with overexpressing and RNAi lines of OsWRKY74 grown in hydroponic solution containing high and low Pi for 14 d were used to compare the number of adventitious roots, primary root length and total length of the three longest adventitious roots. No significant differences in root system architecture were observed between WT and the transgenic lines when grown in HP medium (Fig. 7). However, the elongation of primary and adventitious roots was significantly enhanced in OsWRKY74-overexpressing plants compared with WT plants under LP conditions (Fig. 7B, C). Moreover, the number of adventitious roots in the OsWRKY74-overexpressing plants was much greater than that in WT plants in LP medium (Fig. 7D). These results suggest that the improved growth of the OsWRKY74-overexpressing lines in Pi-deficient conditions may be attributable, at least partially, to the larger root system.

28To further elucidate the mechanisms underlying the regulation of Pi starvation response by OsWRKY74, the expression of several Pi-starvation inducible (PSI) genes was monitored by real-time PCR. The Pi-starvation inducible genes, including Oryza sativa UDP-sulfoquinovose synthase (OsSQD), OsIPS1, OsmiR399a, OsmiR399d, OsmiR399f and OsmiR399j, were markedly induced in both WT and OsWRKY74-overexpressing transgenic plants when grown in LP medium (Fig. 8). This observation is in line with results reported in previous studies. However, the expression of OsSQD, OsIPS1, OsmiR399a, OsmiR399d, OsmiR399f and OsmiR399j in both roots and shoots of OsWRKY74-overexpressing plants was significantly higher than in WT plants in LP medium (Fig. 8; Supplementary Fig. S5). In contrast to OsWRKY74-overexpressing transgenic plants, the expression of OsSQD, OsIPS1, OsmiR399a, OsmiR399d, OsmiR399f and OsmiR399j in both roots and shoots of RNAi OsWRKY74 was significantly lower than that in WT plants in both HP and LP media (Fig. 8; Supplementary Fig. S5). These results suggest that the OsWRKY74 protein may modulate Pi starvation by partially altering expression of downstream Pi-starvation inducible genes.

29Thirteen putative genes encoding high-affinity Pi transporters have been identified in rice. We evaluated the effect of OsWRKY74 on expression of these genes. We failed to detect the expression of OsPT11 and OsPT13 in both WT and transgenic plants grown in both HP and LP media. These results are consistent with reports demonstrating that OsPT11 and OsPT13 are exclusively induced in roots by inoculation with arbuscular mycorrhiza fungi. Disruption of OsWRKY74 expression in transgenic rice plants altered the expression of OsPT3, OsPT4 and OsPT10 under both HP and LP conditions (Fig. 8; Supplementary Fig. S5), while expression of the remaining PHT genes in WT plants was comparable to that in the transgenic rice plants under both HP and LP conditions (data not shown). Thus, OsWRKY74 may regulate Pi acquisition by targeting OsPT3, OsPT4 and OsPT10 transporters under LP conditions.

30The WRKY protein family is a plant-specific transcription factor, and the emerging evidence indicates the important roles played by WRKY proteins in response to nutrient deficiency. In the rice genome, the WRKY TF family comprises 109 members, but the roles of WRKY TFs involved in the maintenance of Pi homeostasis in rice are poorly understood. Our results showed that the expression of OsWRKY74 was rapidly up-regulated by Pi and Fe deprivation, and repressed by N deprivation (Fig. 2), implying its involvement in regulating multiple nutrient starvation responses in rice. This expression pattern differs from that of other Pi-responsive WRKY TFs previously reported. For instance, AtWRKY6 expression is activated in response to Pi deficiency, boron and arsenate starvation. AtWRKY75 is strongly induced upon Pi starvation and pathogen infection. AtWRKY45 is induced during Pi starvation, mainly in the roots. The transcript of AtWRKY42 was suppressed under Pi-deficient conditions. Thus, this is the first report, to our knowledge, showing that OsWRKY74 is involved in the regulation of multiple nutrient starvations in rice.

31An increase in metallic element acquisition is another adaptive response to Pi starvation in plants. Pi deprivation leads to an increase in Fe contents and activation of Fe-responsive genes. In our study, we found that overexpression of OsWRKY74 enhanced the accumulation of Fe in both roots and shoots regardless of Pi status in the growth medium (Fig. 9), implying that OsWRKY74 may also play a regulatory role in Fe homeostasis in addition to control of P acquisition. MiR399 has been established as an important component in the Pi starvation signalling network. Recent studies reported that miR399 is significantly induced by Fe starvation in both roots and shoots and the concentration of Fe is increased in the OsmiR399-overexpressing plants. Our results revealed that OsmiR399 genes were up-regulated in OsWRKY74-overexpressing plants (Fig. 8; Supplementary Fig. S5). Therefore, OsWRKY74 may play an important role in crosstalk between the Fe and Pi signalling cascades, which merits further study.

32Interestingly, the expression of OsWRKY74 was repressed upon exposure to cold stress (Fig. 2E). Two rice WRKY genes, OsWRKY45 and OsWRKY76, have been functionally characterized for their roles in cold stress. The OsWRKY74 protein cannot be grouped with the OsWRKY45 and OsWRKY76 proteins; it was located in a separate branch (Fig. 1B). It has been proposed that Pi starvation and cold stress might share some common regulatory cascades, and that P may participate in the acclimatization to cold stress, as some cold-responsive genes are also regulated by Pi deficiency. In Arabidopsis and rice, the DREB1/CBF-dependent pathway is a central component of cold response, and overexpression of some CBFs results in enhancement of tolerance to freezing, salt and drought stress by activating associated target genes. In the present study, the expression levels of OsDREB1A, OsDREB1B and OsDREB1C were significantly up-regulated in the OsWRKY74-overexpressing plants compared to the WT plants under LP conditions (Fig. 10). Therefore, OsWRKY74 may play an important role in linking cold stress to Pi starvation signal transduction pathways.

33Improvement of tolerance to low Pi stress by expressing Pi deficiency-induced TFs and/or protein kinase genes such as PSTOL1, OsPTF1 and PHR has been reported in the literature. In this study, OsWRKY74 overexpressors exhibited greater tolerance to low Pi stress in both the hydroponic and pot experiments, as indicated by an increased shoot, root biomass, grain weight and tiller number (Tables 1, 2; Figs 4, 5). The number of rice tillers is an important indicator for Pi nutrient status in plants, and is positively correlated with the tolerance to low-Pi stress. In addition, rice genotypes with high Pi use efficiency often exhibit a better grain yield under conditions of low P availability in soils. In the OsWRKY74-overexpressing plants, an increase in grain yield was found under low-Pi conditions (Table 2). These results support the notion that OsWRKY74 is a positive regulator of Pi starvation responses. Furthermore, the enhanced tolerance of OsWRKY74-overexpressing plants to low-Pi stress coincides with the up-regulation of low-Pi-responsive genes, including OsPAP10a and OsSQD (Figs 6C, 8). OsPAP10a encodes an acid phosphatase that is activated by Pi starvation in rice. OsSQD is involved in sulfolipid biosynthesis. Enhanced production of acid phosphatases and activation of scavenging systems are adaptive mechanisms to maximize Pi availability for plants under low Pi conditions. Thus, the enhanced tolerance of OsWRKY74 transgenic plants to low Pi stress may depend in part on changes in the expression of these genes.

34In addition, overexpression of OsWRKY74 in rice plants conferred higher Pi content than WT plants under Pi-deficient conditions (Figs 4B, C, 5E). Pi transporters are directly responsible for Pi acquisition and transport in plants. Plants generally up-regulate the expression of Pi transporters to enhance Pi uptake and transport efficiency to cope with low Pi supply in soils. In this study, several Pi transporters, such as OsPT3, OsPT4 and OsPT10, may account for the Pi uptake in roots of OsWRKY74-overexpressing plants under Pi-deficient conditions, as evidenced by the greater up-regulation of these genes in OsWRKY74-overexpressing plants than in WT plants (Fig. 8). Therefore, OsWRKY74 may regulate Pi acquisition by targeting PHT at the transcriptional level.

35The PHR-miR399-PHO2 pathway has been established to play a critical role in regulating Pi starvation response in Arabidopsis and rice. Our results showed that the expression of OsmiR399a, OsmiR399d, OsmiR399f and OsmiR399j was enhanced in the OsWRKY74-overexpressing plants and repressed in the OsWRKY74 RNAi transgenic plants under both HP and LP conditions (Fig. 8). These observations suggest that the positive regulation of miR399 gene expression in response to Pi starvation is mediated at least in part by the TF OsWRKY74. As for OsPHO2, we did not find the OsmiR399-mediated transcript degradation as transcript abundance of OsPHO2 was unchanged despite the change in OsmiR399 levels (Fig. 8). Our data suggest that OsWRKY74 may positively regulate OsmiR399, but OsPHO2 may not be the target of OsmiR399 in controlling Pi homeostasis in rice. It is also conceivable that its activity may be controlled by another level of post-transcriptional regulation that warrants further investigation.

36Changes in the expression of genes involved in Pi signalling, high-affinity Pi transport and mobilization highlight a global role of OsWRKY74 in response to Pi deficiency (Fig. 8). In fact, it has been previously reported that the transcripts of an array of PSI genes involved in Pi sensing, translocation, transport and mobilization were reduced or induced by other TFs such as ZAT6, MYB62, and AtERF070, but the precise mechanisms remain elusive. Thus, although up-regulation of PSI genes in OsWRKY74-overexpressing plants indicates that OsWRKY74 may act as a positive regulator of their expression, more detailed studies are required to identify the targets of OsWRKY74 and the genes controlled by OsWRKY74 during Pi limitation.

37Alterations in both primary and adventitious root elongation are a typical phenomenon in response to Pi deprivation in rice. In this study, the highest expression of OsWRKY74 was observed in roots, but no significant difference in root system architecture was observed between WT and OsWRKY74 transgenic plants in HP conditions. These results differ from those Pi-responsive TFs of WRKY75, ZAT6, MYB62 and OsMYB2P-1 in terms of their regulation of root system architecture (, b,; ). It has been reported that small changes in root system architecture can have marked effects on P uptake. Our results show that transgenic lines with an overexpression of OsWRKY74 displayed significant increases in length of the primary and adventitious roots, as well as adventitious root number in LP medium (Fig. 7A–D). A similar increase in the primary and adventitious root length has been reported for OsPTF1 and OsMYB2P-1 overexpression lines. Thus, the larger root system of OsWRKY74-overexpressing plants grown in Pi-deficient conditions would allow the transgenic plants to exploit more soils and increase root surface area for Pi uptake, thus conferring them more efficient acquisition of Pi under Pi-deficient conditions.

38In summary, this study characterized a rice WRKY protein belonging to group III of the WRKY protein family that is localized at the nucleus in rice. The OsWRKY74 protein acts as a positive regulator of Pi starvation responses, such that overexpression of OsWRKY74 resulted in enhanced tolerance to low Pi stress, larger root system architecture under LP conditions, and improved expression of PSI genes. In addition, our results also showed that OsWRKY74 can function as a regulator of Fe starvation and cold stress. Further works on the role of OsWRKY74 crosstalk between P and Fe, and P and cold stress are under way in our laboratory.

39Supplementary data is available at JXB online.

40Figure S1. Plasmids construction for plant transformation.

41Figure S2. Deduced amino acid sequence of OsWRKY74.

42Figure S3. Response of OsWRKY74 to -K in leaves and roots.

43Figure S4. Expression of OsWRKY74 under cold, NaCl and drought treatments.

44Figure S5. Expression of Pi starvation-induced genes in wild-type (WT) and OsWRKY74 transgenic plants.

45Table S1. Primers used in semi-quantitative RT-PCR and real-time RT-PCR.

46Table S2. Sequences of forward and reverse primers and 6-FAM 5′ end-labelled probes designed for the 3′ UTR of the rice Pi transporter genes and the rice polyubiquitin 1 (RubQ1) gene for quantitative RT-PCR.