Comparative Study of Early Cold-Regulated Proteins by Two-Dimensional Difference Gel Electrophoresis Reveals a Key Role for Phospholipase Dα1 in Mediating Cold Acclimation Signaling Pathway in Rice*

To understand the early signaling steps that regulate cold responses in rice, two-dimensional difference gel electrophoresis (2-D DIGE)1 was used to study early cold-regulated proteins in rice seedlings. Using mass spectrometry, 32 spots, which represent 26 unique proteins that showed an altered expression level within 5 min of cold treatment were identified. Among these proteins, Western blot analyses confirmed that the cellular phospholipase D α1 (OsPLDα1) protein level was increased as early as 1 min after cold treatment. Genetic studies showed that reducing the expression of OsPLDα1 makes rice plants more sensitive to chilling stress as well as cold acclimation increased freezing tolerance. Correspondingly, cold-regulated proteomic changes and the expression of the cold-responsive C repeat/dehydration-responsive element binding 1 (OsDREB1) family of transcription factors were inhibited in the pldα1 mutant. We also found that the expression of OsPLDα1 is directly regulated by OsDREB1A. This transcriptional regulation of OsPLDα1 could provide positive feedback regulation of the cold signal transduction pathway in rice. OsPLDα1 hydrolyzes phosphatidylcholine to produce the signal molecule phosphatidic acid (PA). By lipid-overlay assay, we demonstrated that the rice cold signaling proteins, MAP kinase 6 (OsMPK6) and OsSIZ1, bind directly to PA. Taken together, our results suggest that OsPLDα1 plays a key role in transducing cold signaling in rice by producing PA and regulating OsDREB1s' expression by OsMPK6, OsSIZ1, and possibly other PA-binding proteins.

In comparison with what is known in Arabidopsis, the mechanism for cold acclimation is not well studied in rice. Sequence homology searches have identified homologous OsHOS1, OsICE1, and OsDREB1s genes in the rice genome (17)(18)(19). Similar to what has been observed in Arabidopsis, OsHOS1 is able to regulate the protein abundance of OsICE1 (19). Also, the expression of OsDREB1A and OsDREB1B is induced by cold stress, and overexpression of these two genes increases the cold stress tolerance of transgenic rice (20), which suggests that the CBF/DREB1 cold signaling pathway is conserved in rice. In addition to OsDREB1s and OsICE1, a number of other proteins such as OsCOIN (21), OsMYBS3 (22), OsPRP3 (23), and OsMAPK6 (24) have been shown to regulate cold stress responses in rice. Whether these proteins function as COR genes or as primary regulators of early cold responses in rice needs to be further explored.
Despite considerable progress in studying plant cold signal transduction mechanisms, our understanding of the cold stress signal perception and early signaling transduction steps leading to CBF/DREB1s activation is still very limited. With the development of mass spectrometry technologies, quantitative proteomic approaches have proven to be very powerful tools, complementing traditional genetic approaches, in finding new signaling components as well as revealing new signaling mechanisms (25,26). There has been a number of reports studying cold-regulated proteins in rice using proteomic approaches (27). However, the rice plants in these studies were cold-stress-treated for hours or days. It was reported that cold induces an increase in cytoplasmic calcium levels as early as 40 s after cold treatment (1) and the expression of CBFs in Arabidopsis is dramatically increased within 15 min after cold treatment (10). These results suggest that early cold stress signal transduction occurs within 15 min or less after a plant is exposed to cold temperature. Therefore, it is likely that cold-regulated proteins previously identified in rice using proteomic approaches were mostly late cold-regulated proteins whose expression was regulated by DREB1s and other cold-regulated transcription factor-mediated transcriptional networks.
In order to help better understand early cold stress signaling, two-dimensional difference gel electrophoresis (2D DIGE) technology was applied to characterize rapid cold-regulated proteomic changes and identify novel early cold-regulated proteins in 1-week-old rice seedlings. A total of 32 spots, which represent 26 unique proteins whose abundance changed in response to a five-minute cold stress treatment, were identified by mass spectrometry. Additional genetic, biochemical, and physiological evidence suggest one of these early cold-regulated proteins, OsPLD␣1, plays a key role in the early cold stress signal transduction pathway in rice.

EXPERIMENTAL PROCEDURES
Protein Extraction and Two-Dimensional Difference Gel Electrophoresis-Because we were interested in finding proteins whose abundance changed rapidly in response to cold stress, a water bath was used for the cold treatment to make sure all the rice seedlings were instantly exposed to a similar temperature at the same time. Wild-type (Oryza sativa japonica cv. Nipponbare) rice seedlings were first germinated in double distilled water in darkness for 3 days. Seedlings were then transferred to Hoagland's liquid growth medium to grow 1 more week at 28°C, 50% humidity and long-day conditions. After being submerged in double distilled water at 28°C or 0°C for 5 min, the seedlings were quickly tap-dried with tissue paper and snap frozen using liquid nitrogen. Cold-treated and untreated samples were handled in parallel for protein extraction and separation to reduce variations.
Cy3 and Cy5 labeling of the proteins and two-dimensional electrophoresis were performed as described previously (29,30). 2D DIGE images were acquired using a Typhoon Trio scanner (GE Healthcare, Piscataway, NJ). The images were analyzed using DeCyder 6.5 software (GE Healthcare). Spots that showed consistent cold stress regulated changes in at least four biological repeat samples were picked using a robotic spot picker (GE healthcare).
Mass Spectrometry and Protein Identification-Selected protein spots were in-gel digested by trypsin as previously described (28,29). The extracted peptides were dissolved in 10 l 0.1% formic acid. The peptides were loaded onto a C18 reverse phase column (100 m ϫ 150 mm, Thermo Fisher Scientific, Waltham, MA) coupled online to a LTQ-XL linear ion trap mass spectrometer (Thermo Fisher Scientific) at a flow rate of 350 nl/min. Peptides were separated using a gradient from 100% of A (0.1% formic acid) to 45% of B (0.1% formic acid, 95% acetonitrile) for 60 min. MS1 spectra were acquired in a positive mode using the data dependent automatic survey MS scan in the m/z range between 400 and 2000. The 10 most intense ions that over 500 counts were selected for MS2 acquisition using normalized collision induced dissociation (CID) with 35% collision energy, and activation Q value was set to 0.25. CID product ions were analyzed on the linear ion trap in centroid mode. The dynamic exclusion activation time was set to 30 s to prevent same m/z ions from being selected after its acquisition.
LC-MS/MS peak lists were searched against database generated from Oryza sativa subset of the NCBInr database (date 20/8/2010, 84,086 entries searched), using the SEQUEST search algorithm. Peptides were selected in the mass range between 400 -5000 amu. The following search parameters were applied: mass tolerance for precursor ions and fragment ions were set to 2.00 amu and 1.00 amu, respectively; two incomplete cleavages were allowed; alkylation of cysteine by carbamidomethylation and oxidation of methionine were considered as possible modification. Search result option filter was set as the following: Delta CN Ն 0.1; Xcorr (Ϯ1, 2, 3) ϭ 1. 5, 2.0, 2.5; peptide probability Յ0.01.
Cold Tolerance Assays-Cold tolerance assays of rice seedlings were performed as described with modifications (21,32,33). For germination assays, a minimum of 40 wild-type or pld␣1 mutant seeds were submerged in 25 ml double distilled water in a glass Petri dish in darkness at 28°C or 16°C for up to 14 days. The seeds were checked daily for their germination rate. Seed germination was determined by emergence of coleoptile tips. For cold sensitivity assays, seedlings were first germinated in water in darkness for 3 days at 28°C. A minimum of 30 seedlings with similar coleoptile length were then transferred to Hoagland's liquid growth medium at 28°C or 12°C, with a daily photoperiodic cycle of 16 h light and 8 h dark (long day) with 50% humidity for 30 days before coleoptile length and seedling weight measurements were made. Alternatively, the seedlings were allowed to grow in Hoagland's liquid growth medium at 28°C, long day with 50% humidity for 7 days before being subjected to 4°C, long day with 50% humidity and allowed to continue growing for up to 7 days. The cold-treated seedlings were allowed to recover at 28°C, long day with 50% humidity conditions for 7 more days before taking pictures and calculating the survival rates. All the experiments shown in this study have been performed at least three times with similar results. Representative data from one repetition are shown.
Freezing Tolerance Assays-Freezing tolerance assays of rice seedlings were performed according to Li et al. (34). Seedlings were grown in soil at 28°C, long-day conditions with 50% humidity for 6 weeks. Freezing stress was performed using a programmable freezing chamber (model AR33L, Percival Scientific, Perry, IA). The plants were first kept at Ϫ1°C for 3 h. Temperature was decreased at Ϫ1°C per hour to Ϫ2°C, Ϫ4°C, or Ϫ6°C. After holding at the designated freezing temperature for 3 h, the temperature was increased at 1°C per hour to Ϫ1°C then incubated at 12°C for 12 h before transferring to a greenhouse to recover for 7 days at 28°C, long-day conditions with 50% humidity. Ion leakage was determined immediately after overnight incubation at 12°C. Approximately 1 g of leaves was harvested and shaken in a falcon tube with 20 ml double distilled water at room temperature overnight. Leaves were then transferred to a new falcon tube with 20 ml double distilled water and boiled for 15 min. Conductivity of the solution was determined using a Leici conductivity meter (DDS-IIA, Shanghai, China). Relative ion leakage was calculated by dividing the conductivity collected from room temperature samples by the sum of conductivities collected from room temperature samples and boiled samples.
Protein Purification and In Vitro Lipid Overlay Assay-Full-length coding sequence of OsDREB1A, OsMPK6, AtSIZ1, OsSIZ1, OsOST1, and Arabidopsis RCN1 without stop codon were amplified by PCR, cloned into pENTRY/S.D./D-TOPO vector (Thermo Fisher Scientific), and subcloned into Gateway-compatible vectors pGEX-4T-1 by LR clonase (Thermo Fisher Scientific). The GST-tagged proteins were expressed and purified from Escherichia coli using glutathione Sepharose 4B beads by standard protocols. For in vitro lipid overlay assay, phospholipids (Avanti polar lipids, Alabaster, AL) dissolved in chloroform were spotted on a nitrocellulose membrane and dried at room temperature. The membrane was incubated with blocking solution containing 3% (w/v) fatty-acid-free BSA (Sigma-Aldrich, St. Louis, MO), 10 mM Tris-HCl, pH 8.0, 140 mM NaCl, and 0.1% Tween-20 for 1 h. The membranes were incubated with 2 g/ml GST-tagged proteins in blocking buffer for 12 h at 4°C with gentle shaking. After washing three times with blocking buffer, the membrane was incubated with a rabbit HRP conjugated anti-GST antibody (Sigma-Aldrich 1:5000 in blocking buffer) for 1 h. Binding of proteins to phospholipids was visualized with FluorChemQ MultiImage III (Alpha Innotech, San Leandro, CA) using the SuperSignal West Dura chemiluminescent reagent (Thermo Fisher Scientific).
Electrophoretic Mobility Shift Assay-Double-stranded oligonucleotide sequence for electrophoretic mobility shift assay (EMSA) was: 5Ј-cgccacttggaccgacctcgggacgacgac-3Ј. EMSA assay was performed using a modified 10 l mixture, which contained 1.5 l (0.5 g/l) purified protein, 1 l biotin-labeled oligonucleotides, 1 l 10 ϫ binding buffer (100 mM Tris, 500 mM KCl, and 10 mM DTT, pH 7.5), 0.5 l 50% glycerol, 0.5 l 1% Nonidet P-40, 0.5 l 1 M KCl, and 5 l ultrapure water. The reactions were incubated at 4°C for 60 min and loaded onto a 6% native polyacrylamide gel in TBE buffer (45 mM Tris, 45 mM boric acid, and 1 mM EDTA, pH 8.3). The gel was transferred to a nylon membrane (Millipore, Billerica, MA) using 0.5 ϫ TBE buffer at 100 V for 60 min in cold room. Nylon membrane was cross-linked at 120 mJ/cm 2 for 45-60 s using a UV cross-linker. Biotin-labeled DNA was detected by using the Light Shift Chemiluminescent EMSA kit (Thermo Fisher Scientific) according to manufacturer's instruction.
Transient Luciferase Expression Assay-To generate pOsPLD␣1: firefly luciferase, a 2.5 kb promoter upstream of translation start codon ATG of OsPLD␣1 was PCR amplified, cloned into pENTRY/ S.D./D-TOPO vector (Thermo Fisher Scientific), and subcloned into pHGWL7.0 vector by LR clonase (Thermo Fisher Scientific). A fulllength coding sequence of OsDREB1A without stop codon was PCR amplified and inserted into a p35S:EGFP-HA expression vector by EcoRI ϩ SalI to replace the enhanced green fluorescent protein (EGFP) sequence to generate a p35S:OsDREB1A-HA expression vector. Transient luciferase expression assay was performed according to Yoo et al. (35). In brief, rice protoplasts were prepared from leaves of 7-day-old wild-type rice seedlings (Dongjin) plants grown at 28°C under long-day conditions with 50% humidity. The expression vectors were transformed into rice protoplasts by polyethylene glycol (PEG) 4000 (Sigma-Aldrich) mediated transfection. An p35S-driven renilla luciferase expression vector was cotransformed to calibrate the transfection efficiency, and 16 h after transformation, luciferase luminescence was recorded for 10 s using a microplate luminometer (Centro LB 960, Berthold technologies, Bad Wildbad, Germany). Quantitation of luminescent signal from each of the luciferase reporter enzymes was performed using a dual-luciferase® reporter assay kit (Promega, Fitchburg, WI) according to manufacturer's instruction.

OsPLD␣1
Is an Early Cold-Regulated Protein-To investigate the early proteomic response of rice to cold stress, 1-week-old rice seedlings were treated with 0°C water bath for 5 min. Cellular proteins were then separated into soluble, microsomal, and pellet (which include most of the organelles) fractions to reduce the complexity of the protein samples, and analyzed independently by two-dimensional difference in gel electrophoresis (2D DIGE). Samples from six independent preparations were analyzed. On average, 2683 Ϯ 120 spots were detected in our 2D DIGE gels, proteins that showed consistent cold-regulated changes (over 1.5 fold) in abundance in at least four independent experiments were chosen for protein identification.
By LC-MS/MS, 32 spots, which represent 26 unique proteins, were identified. Interestingly most of these spots were found in the soluble and pellet fractions of our samples (Figs. 1 A-1C and Table I). Only six spots whose abundance was consistently regulated by cold treatment were identified from the microsomal fraction (Fig. 1B), suggesting a relative delay in the expression changes of membrane proteins to cold stress. Functional classification showed these proteins are involved in processes such as cellular metabolism, stress responses, protein folding, cytoskeleton rearrangement, and lipid metabolism. A comparison of our results with previous cold stress proteomic studies in rice showed that 20 (76.9%) of our identified proteins had not been previously reported. For the proteins we identified that had also been identified by previous cold stress proteomic studies, our results indicate that these proteins are actually early cold-regulated proteins and thus likely contribute to early cellular responses to cold stress treatment.
Of all the early cold-regulated proteins identified, OsPLD␣1 was chosen for further studies (Fig. 1D) because it was previously reported that suppressing the expression of PLD␣1 in Arabidopsis made plants more resistant to freezing temperature (5). We first validated the proteomic results by Western blot analysis, using a polyclonal antibody raised specifically against the conserved C-terminal 12-amino acid sequence found in both AtPLD␣1 and OsPLD␣1 (36). Surprisingly, the increase of the OsPLD␣1 protein level inside the cell can be seen as early as 1 min after cold stress treatment (Figs. 1E and 1F). This increase in OsPLD␣1 protein level was sustained for 5 min and then decreased to return to the basal level 10 min after cold treatment. We also tested whether the abundance of OsPLD␣1 transcript is rapidly regulated by cold stress. Semi-quantitative RT-PCR results showed that the expression of OsPLD␣1 is decreased by cold treatment for 1 min, possibly due to sudden increase in protein translation, and then continues to increase from 5 min to 30 min after cold treatment (Figs. 1G and 1H).
Expression Pattern and Subcellular Localization of OsPLD␣1-To examine the tissue-specific expression pattern of OsPLD␣1, genomic sequence, without the stop codon and 3Јuntranslated region (UTR), of OsPLD␣1, which includes 2537-bp promoter and 5Ј UTR sequence before ATG and two introns, was cloned, fused with a ␤-glucuronidase (GUS) reporter enzyme at the C terminus, and introduced into wild-type rice plants. T2 homozygous pOsPLD␣1:OsPLD␣1-GUS transgenic plants were used for histochemical staining for GUS activity. As shown in Fig. 2, GUS expression was mostly detected at the tips of roots and coleoptiles of rice seedlings that had just germinated ( Fig. 2A). When plants were 7 days old, strong GUS activity was detected in whole roots. Except for weak OsPLD␣1-GUS expression at the tips of the leaves, no detectable GUS activity was found in coleoptiles and young leaves of 7-day-old rice seedlings grown under normal conditions. A 5 min treatment with 0°C water bath induced a significant increase in the expression of OsPLD␣1-GUS in the coleoptile and leaf, which further confirmed that the level of OsPLD␣1 protein is increased by cold treatment (Fig. 2B). During tillering stages, OsPLD␣1-GUS is also expressed in stem. A close examination of hand cross-dissected stem showed that the expression of OsPLD␣1 in rice stems is very strong in vascular tissues (Fig. 2C). When plants enter the flowering stage, OsPLD␣1 promoter activities can be detected in hulls but not in stamens and pistils (Fig. 2D).
We also generated pOsPLD␣1:OsPLD␣1-GFP-expressing transgenic rice and observed the subcellular localization pat- tern of OsPLD␣1 in vivo. Using confocal microscopy, we found OsPLD␣1-GFP protein is distributed in the cytoplasm, but excluded from the nucleus of cells in the primary root elongation zone (Fig. 2E).
Suppressing the Expression of OsPLD␣1 Impaired the Chilling Tolerance of Rice Plants-To investigate the possible role of OsPLD␣1 in rice cold stress responses, a transfer DNA (T-DNA) insertion mutant of OsPLD␣1 (PFG-1A-21508) was isolated. Although the T-DNA is inserted in the 5ЈUTR of the OsPLD␣1 genome sequence (Fig. 3A), reverse transcription PCR (RT-PCR) analysis of the mutant showed that the expression of OsPLD␣1 was barely detectable (Fig. 3B). Using PLD␣1-specific antibody, we further confirmed that the OsPLD␣1 protein is not detectable in the pld␣1 mutant (Fig. 3C).
The isolated T-DNA mutant was backcrossed with wildtype (Dongjin cultivar) plants twice to clean up the background. Segregated homozygous pld␣1 mutant plants, which contained only one T-DNA insertion site (based on the segregation ratio of F2 hygromycin resistant seedlings) and showed no obvious growth phenotype, were used for chilling tolerance studies. One-week-old wild-type and pld␣1 mutant rice seedlings were exposed to chilling stress (4°C) for various durations. After recovering at 28°C for 7 days, the seedlings were photographed and their survival rate was calculated. Compared with slightly reduced survival rate of wildtype plants, about half of the pld␣1 mutant seedlings died after growing at 4°C for 5 days (Figs. 3D and 3E). Similar results were also observed for seeds germinated at 16°C or seedlings grown at 12°C for 30 days. On average, the ger- To rule out the possibility that the reduced chilling tolerance of pld␣1 mutant is due to additional mutation site in the mutant background, we generated OsPLD␣1 RNAi transgenic rice plants. Western blot analysis using PLD␣1 antibody showed the expression of OsPLD␣1 protein in our transgenic RNAi rice seedlings is either very weak or not detectable (Fig.  3I). Based on the expression level, R4b and R14b RNAi lines were chosen for chilling tolerance assays. As shown in Fig. 3J, on average, around 37.5% of wild-type seedlings survived after growing in 4°C for 6 days. In comparison, the survival rate of our RNAi transgenic seedlings was 25% for line R4b and 9.7% for line R14b. Taken together, these results demonstrate that suppressing the expression of OsPLD␣1 makes rice plants more sensitive to chilling stress.
Suppressing the Expression of OsPLD␣1 Reduces the Cold Acclimation Increased Freezing Tolerance of Rice Plants-We also tested the freezing tolerance of the pld␣1 mutant under acclimation and nonacclimation conditions. Six-week-old rice seedlings were subjected to various freezing temperatures for 3 h and allowed to recover at 28°C for 7 days. Because of the space limitation in the growth chamber used for freezing treatment, a maximum of eight plants (four mutants and four wild type) could be treated at the same time. Thus, the experiment was independently repeated five times, and the average number of green leaves per plant from all of the experiments after recovery was calculated and used to evaluate the freezing tolerance of the mutant.
Suppressing the expression of OsPLD␣1 in rice seedlings had no significant effect to the development of leaves. Sixweek-old wild-type and pld␣1 mutant rice plants developed 8.25 Ϯ 1.21 and 8.08 Ϯ 1.83 leaves on average, respectively. There was no obvious difference in the green leaves of freezetreated wild-type and pld␣1 mutant rice plants under nonacclimation conditions at all freezing temperatures tested (Figs. 4A and 4B), suggesting the expression level of OsPLD␣1 does not regulate the basal freezing tolerance of rice. Cold acclimation at 12°C under long-day conditions for 7 days significantly increased the freezing tolerance of wild-type rice plants. However, compared with the wild-type control, the increased freezing tolerance in pld␣1 mutant was greatly reduced (Figs. 4A and 4B). Correspondingly, freeze-induced electrolyte leakage, an indicator of plasma membrane damage, in the pld␣1 mutant was dramatically increased under cold acclimation condition compared with wild-type plants. Based on the chilling and freezing tolerance assays of the pld␣1 mutant, we conclude that OsPLD␣1 is a positive regulator of the rice cold acclimation processes.
Seedlings Overexpressing OsPLD␣1-GFP Are Hypersensitive to Cold and Freezing Stresses-To further assess the role of OsPLD␣1 in cold and freezing tolerance of rice plant, full-length coding sequence of OsPLD␣1 with a C-terminal green fluorescent protein (OsPLD␣1-GFP) was overexpressed under the control of p35S cauliflower mosaic virus promoter. The expression of OsPLD␣1-GFP was detected by GFP antibody (Fig. 5A). Rice seedlings were subjected to cold (4°C) treatment or cold acclimation at 12°C for 7 days followed by freezing (Ϫ6°C) treatment. Unexpectedly, after recovery at 28°C for 7 days, the survival rate of OsPLD␣1-GFP overexpressing rice seedlings was lower than wild-type control seedlings but higher than the pld␣1 mutant seedlings (Figs. 5B and 5C). It appears that the protein level of PLD␣1 has to be carefully controlled and maintained at an appropriate level. Overexpressing OsPLD␣1-GFP may cause overaccumulation of PA inside the cell and the corresponding decrease in phosphatidylcholine (PC) concentration in the membrane, which might stimulate the formation of hexagonal phases of membrane and increase membrane damage when plant cells encounter low temperature (37).
Cold-Stress-Regulated Proteomic Changes in pld␣1 Mutant-Using 2D-DIGE, we compared the cold-stress-regulated proteomic changes in wild-type (Dongjin) and pld␣1 mutant seedlings. Because we are interested in finding coldregulated proteins downstream of PLD␣1, 1h cold treatment in 0°C water bath was chosen for the proteomic analyses. We prepared and analyzed six independent biological repeats of soluble protein sample sets. Spots that showed consistent cold-regulated abundance changes in at least four independent experiments were chosen for future protein identification. LC-MS/MS identified 25 spots, which represents 18 proteins, OsPLD␣1 Regulates Cold Acclimation Signaling in Rice whose abundance was regulated by 1 h cold treatment from wild-type seedlings (Fig. S1 and Table II). Compared with Table  I, only eight spots, which represents seven proteins (38.9%) including a villin (Os03g24220), an actin 3 (Os03g61970), an actin 7 (Os05g01600), an aminotransferase (Os08g41990), an ATP-dependent Clp protease (Os04g32560), a putative subtilisin (Os02g53860), and a sucrose synthase (Os03g28330), showed similar up-or down-regulated patterns. The differences in cold-stress-regulated proteomic changes we observed might be due to using a different cultivar of rice or the difference in the treatment duration. Of the 1 h cold-treatment-regulated proteins identified, the abundance changes of nine spots, which represents eight proteins (44.4%) showed significant differences between wild-type control and pld␣1 mutant seedlings (Table II and Fig. S2). These results further support that cold regulates the abundance change of a group of proteins through PLD␣1 in rice.
Chilling-Induced Expression of OsDREB1s Is Impaired in Ospld␣1 Mutant-Cold acclimation has been shown to increase plant chilling/freezing tolerance by up-regulating the expression of CBF/DREB family transcription factors. There are 10 CBF/DREB homologs (OsDREB1A to OsDREB1J) in the rice genome. It has been shown that the expression of OsDREB1A, OsDREB1B, OsDREB1E, OsDREB1F, and OsDREB1G is regulated by cold acclimation (17,38). To investigate whether the hypersensitive phenotype of the pld␣1 One-week-old OsPLD␣1 RNAi seedlings were grown in 4°C, long-day condition for 6 days. Survival rate was calculated after recovery at 28°C for 7 days. In E and J, one-way ANOVA test was performed. Statistically significant differences are indicated by different lowercase letters (p Ͻ 0.05). Error bars represent ϮS.D. mutant in response to chilling/freezing stress is related to the endogenous level of OsDREBs, the chilling-induced expression of OsDREBs was examined in the pld␣1 mutant. Quantitative real-time PCR analysis showed 12°C chilling stress induced a dramatic increase of OsDREBs's expression in wild-type rice seedlings. In comparison, the increase in expression of OsDREBs was greatly reduced in pld␣1 mutant (Fig. 6), which suggests OsPLD␣1 is required for the chillinginduced increase in OsDREBs expression during the cold acclimation process. We also found that the expression of OsPLD␣1 in wild-type plants increased by chilling treatment but remained at a basal level in the pld␣1 mutant (Fig. 6). As the cold acclimation process is sustained for 6 h, the expression of OsPLD␣1 decreases, which could help prevent the overhydrolysis of PC in the membrane.
PA Interacts with OsMPK6 and OsSIZ1 In Vitro-PLD␣1 is a phospholipase that hydrolyzes phospholipids, such as phos-phatidylcholine to produce the signaling molecule phosphatidic acid (PA). PA is able to bind to a variety of signaling proteins, regulating their subcellular localization or their activity to transduce signals to downstream components (39). To investigate whether a similar mechanism exists in the rice cold-stress-signaling pathway, recombinant OsMPK6, At-SIZ1, a rice SIZ1 homologous protein (Os05g03430), and a rice OST1 homologous protein (Os03g41460) were purified using a GST tag and tested for their ability to interact directly with PA. A previously demonstrated PA-binding protein, Arabidopsis RCN1, was used as a positive control. As shown, RCN1, OsMPK6, AtSIZ1, and OsSIZ1 all bound strongly to PA in a dose-dependent manner but not to other lipids such as phosphotidylcholine (PC), phosphatidylethanolamine (PE), and phosphatidylglycerol (PG). In contrast, OsOST1 does not bind to any of the lipids we tested, similar to the GST protein control (Fig. 7). Expression of OsPLD␣1 Is Directly Regulated by OsDREB1A-CBF/DREB family transcription factors are able to bind to a CCGAC core sequence of C-repeat (CRT)/dehydration respon-sive elements (DRE) and regulates COR gene expression (40). The finding that the expression of OsPLD␣1 is up-regulated by chilling treatment motivated us to search the promoter of One-week-old transgenic rice seedlings overexpressing OsPLD␣1-GFP were transferred to 4°C, long-day condition and allowed to continue to grow for 5 days. Survival of the seedlings was examined after recovery at 28°C for 7 more days. (C) Six-week-old transgenic rice seedlings overexpressing OsPLD␣1-GFP were cold acclimated and treated with freezing temperature (Ϫ6°C) for 3 h. Seedlings were allowed to recover at 28°C for 7 days, and the green leaves per plants were calculated. Shown are average results from three independent experiment with at least 24 plants per treatment. One-way ANOVA test was performed. Statistically significant differences are indicated by different lowercase letters (p Ͻ 0.05). Error bars represent ϮS.D.

OsPLD␣1 Regulates Cold Acclimation Signaling in Rice
OsPLD␣1 for potential CRT/DRE consensus sequences. Using an online analysis tool (http://bioinformatics.psb.ugent.be/ webtools/plantcare/html/), we found three copies of a typical CRT/DRE consensus sequence in the 1500-bp upstream of the ATG start codon of OsPLD␣1 (Fig. 8A). In comparison, the CRT/DRE consensus sequence was not found in promoter sequences 1500-bp upstream translation start codon of PLD␣1 in Arabidopsis (AtPLD␣1), which explains the relatively stable expression level of AtPLD␣1 in Arabidopsis in response to cold treatment (41). This observation suggests that the role of PLD␣1 in regulating cold acclimation is different in rice and Arabidopsis.
To investigate whether the expression of OsPLD␣1 can be directly regulated by OsDREB1s, RNAi technology was used to knock down the expression of OsDREB1A in rice plants. Quantitative real time PCR analysis showed that when endogenous OsDREB1A transcript was knocked down by RNAi, chilling (12°C)-induced expression of OsPLD␣1 is inhibited (Fig. 8B). We also tested whether OsDREB1A activates the expression of OsPLD␣1 using a transient expression assay. A 2537-bp promoter sequence upstream ATG of OsPLD␣1 was fused upstream of the firefly luciferase gene. The generated pOsPLD␣1:firefly luciferase reporter construct was cotransformed into protoplast generated from 1-week-old rice seedlings with a p35S:OsDREB1A-HA or p35S:EGFP-HA expression vector. As shown in Fig. 8C, expression of OsDREB1A-HA led to a fivefold increase of firefly luciferase activity compared with the EGFP-HA control, indicating that OsDREB1A protein activates OsPLD␣1 promoter in vivo. We also showed using gel mobility shift assay (EMSA) that recombinant OsDREB1A protein purified from E. coli was able to bind specifically with a CRT/DRE containing 30-bp promoter sequence (-618 to -589 bp) of OsPLD␣1 (Fig. 8D). Together, these results suggest that OsDREB1A activates the expression of OsPLD␣1 by binding directly to OsPLD␣1 promoter. DISCUSSION The CBF/DREB1-dependent cold signaling pathway has been shown to play an important role in regulating cold acclimation in both Arabidopsis and rice. Many studies have investigated the regulation of CBF/DREB1 by ICE1 and their upstream regulators HOS1, SIZ1, and OST1, but the understanding about the early cold signal perception and signaling mechanisms leading to activation of HOS1, SIZ1, and OST1 is FIG. 6. The chilling-induced expression of OsDREB1s is inhibited in the pld␣1 mutant. Seven-day-old rice seedlings grown at 28°C were exposed to 12°C chilling stress for the indicated time. The expression of OsDREB1s and OsPLD␣1 were quantified using quantitative real-time RT-PCR. Data represent the mean value from three independent experiments. One-way ANOVA test was performed. Statistically significant differences are indicated by different lowercase letters (p Ͻ 0.05). Error bars represent ϮS.D. 7. OsMPK6, OsSIZ1, and AtSIZ1 proteins bind directly to PA in vitro. The phospholipids were dot blotted onto nitrocellulose membrane. The membrane was blocked with fat-free BSA, probed with recombinant GST-RCN1, GST-OsMPK6, GST-OsOST1, GST-AtSIZ1, GST-OsSIZ1, or GST proteins, followed by horseradish-peroxidase labeled anti-GST antibody. Phospholipids used were: PA, phosphatidic acid; PC, phosphatidylcholine; PE, phosphatidylethanolamine; PG, phosphatidylglycerol.

OsPLD␣1
Regulates Cold Acclimation Signaling in Rice still very limited. Using 2D-DIGE, we identified 26 unique proteins from rice seedlings, whose abundance is regulated rapidly by a 5 min chilling treatment. Of these proteins, genetic studies showed reducing the expression of OsPLD␣1 by RNAi or T-DNA insertion impaired chilling tolerance and cold acclimation increased freezing tolerance of rice plants. In addition, chilling-induced expression of OsDREB1s is greatly reduced in pld␣1 mutant, suggesting OsPLD␣1 plays an important role in regulating the cold acclimation signaling pathway leading to activation of OsDREB1s in rice.
The activity of PLD inside a cell is regulated by intrinsic and extrinsic signals (42). Once activated, PLD␣1 hydrolyzes phospholipids such as phosphatidylcholine to produce a lipid messenger phosphatidic acid (PA), which has been shown to regulate cellular processes by promoting membrane fusion, altering protein-membrane interactions and modulating protein activities (39). By lipid overlay assay, we demonstrated that OsMPK6 and OsSIZ1, which are cold-stresssignaling components upstream of OsDREB1s, are able to bind specifically to PA in a dose-dependent manner, thus providing insight into the mechanism by which OsPLD␣1 regulates cold signaling pathway in rice. It will be interesting to investigate whether the cellular localization or protein activity of additional components in the rice cold signaling pathway are also regulated by PA.
There are 12 and 17 PLDs encoded in Arabidopsis and rice genome, respectively. Based on differences in their biochemical and structural properties, these PLDs can be further grouped into two subfamilies: C2-PLDs (PLD␣, ␤, ␥, ␦ and ) and PX/PH-PLDs (PLD) (39). Recent discoveries indicate that PLDs play an important role in regulating growth, development, and abiotic stress responses in plants (43). There are also studies showing the involvement of PLDs in plant cold stress responses. For example, the expression or protein activities of PLD in Arabidopsis (5), rice (44), cotton (45), Chorispora bungeana (46), and Jatropha curcas (47) have been shown to be regulated by cold stress. Suppressing the expression of AtPLD␦ in Arabidopsis makes the mutant less tolerant to freezing, whereas overexpression of AtPLD␦ increases freezing tolerance of the transgenic plants (34). In contrast, suppressing the expression of AtPLD␣ in Arabidopsis makes the plant more resistant to freezing stress (5). It remains to be determined whether the substrate selection or tissue-specific localization of AtPLD␣ and AtPLD␦ is responsible for the distinct roles of these two PLDs in regulating freezing tolerance in plants. In this study, we found that cellular OsPLD␣1 protein abundance is increased as early as 1 min after chilling treatment, suggesting a role of OsPLD␣1 in early cold-regulated cellular responses. We also found that chilling treatment increases the expression of OsPLD␣1. Suppressing the expression of OsDREB1A in rice plants diminished the chilling-induced increase of OsPLD␣1 expression. Motif search of the 1500-bp promoter region upstream ATG of OsPLD␣1 identified three potential cold-responsive CRT/DRE elements. By EMSA assay, we demonstrated OsDREB1A protein binds directly with the CRT/DRE element (-609 to -601) that is closest to the ATG translation start site. Together these results suggest that OsPLD␣1 is also a COR gene, and the chilling-induced expression of OsPLD␣1 could provide positive feedback regulation during the cold acclimation process.
It is interesting that OsPLD␣1 and AtPLD␣1 play opposite roles in regulating cold stress responses in Arabidopsis and rice plants. An alignment indicates that OsPLD␣1 and AtPLD␣1 are 79% identical at the amino acid level. Amino acid sequences in domains that are responsible for enzymatic activity of OsPLD␣1 and AtPLD␣1 are even more conservative (Fig. S3). Thus, it is unlikely that differences in enzyme activity or substrate selection contribute to this functional discrep-ancy between OsPLD␣1 and AtPLD␣1. The difference in responses to cold and freezing stresses of transgenic plants with knocking down expression level of OsPLDa1 or AtPLDa1 suggested that the dynamics of phospholipids in responding to cold treatment is differently regulated in monocots and dicots. Meanwhile, examination of the 1500-bp promoter and the 5ЈUTR sequence upstream of the translation start codon of the OsPLD␣1 gene revealed the presence of various stress responsive cis-elements, including three copies of cold-responsive CRT/DRE elements. This suggests OsPLD␣1 might play important roles in regulating responses of rice plant to various environmental stresses. In contrast, no CRT/DRE elements and only a few stress-responsive elements were found in AtPLD␣1. The difference in transcription regulation suggests that tissue-specific expression and/or cellular accumulation levels of OsPLD␣1 and AtPLD␣1 protein are regulated differently, which might also contribute to the different roles of OsPLD␣1 and AtPLD␣1 in regulating cold responses in rice and Arabidopsis.
Given the important role of OsPLD␣1 in regulating cold acclimation in rice plants, it is surprising to find that the expression of OsPLD␣1 is not ubiquitously distributed throughout the whole plant. Histochemical staining of GUS activity in pOsPLD␣1:OsPLD␣1-GUS transgenic rice plants showed the expression of OsPLD␣1-GUS protein was mostly in root, tip of young leaf, and stem and hulls. Very weak GUS activity can be detected in young and old leaves grown under normal conditions. This raises the question, where is the cold signal sensed and how does locally expressed OsPLD␣1 regulate cold acclimation for the whole rice plant? Interestingly, a high level of GUS staining is detected in the vascular tissue of the stem. Recent studies have discovered the existence of phospholipids, such as PC and PA, and lipid-binding proteins in phloem exudates of Arabidopsis (48). Although the presence of long distance transport of lipid signal molecules still needs to be tested, the accumulation of OsPLD␣1 in vascular tissue suggests that it is possible that cold stress increases the local PA concentration in vascular tissue, which could then regulate the cold resistance of surrounding cells and tissues in rice.
In conclusion, we found that the abundance of OsPLD␣1 protein is rapidly increased by cold stress treatment and that OsPLD␣1 is important for cold acclimation of rice plants. It has been demonstrated that the phospholipase activity of PLD␣ is regulated by the intracellular calcium level and that cold stress induces a rapid increase in the intracellular calcium level (49). Based on these observations and our current results, we propose a model for an OsPLD␣1-mediated cold stress signal transduction pathway in rice (Fig. 7E). In this pathway, environmental chilling stress is sensed by a still unknown mechanism, which activates the phospholipase activity of OsPLD␣1, possibly by increasing the intracellular calcium level. Activated OsPLD␣1 uses PC as a substrate to produce PA. Then PA binds to cold signaling components, such as OsMPK6 and OsSIZ1, altering their cellular localization and/or protein activity and increasing the expression of OsDREB1s. The increased level of cellular OsDREB1s then can regulate the expression of OsPLD␣1, which could provide a positive feedback enhancement of cold acclimation processes.