Comparative Proteomic Analysis Provides New Insights into Chilling Stress Responses in Rice*

Low temperature is one of the major abiotic stresses limiting the productivity and the geographical distribution of many important crops. To gain a better understanding of chilling stress responses in rice (Oryza sativa L. cv. Nipponbare), we carried out a comparative proteomic analysis. Three-week-old rice seedlings were treated at 6 °C for 6 or 24 h and then recovered for 24 h. Chilling treatment resulted in stress phenotypes of rolling leaves, increased relative electrolyte leakage, and decreased net photosynthetic rate. The temporal changes of total proteins in rice leaves were examined using two-dimensional electrophoresis. Among ∼1,000 protein spots reproducibly detected on each gel, 31 protein spots were down-regulated, and 65 were up-regulated at least at one time point. Mass spectrometry analysis allowed the identification of 85 differentially expressed proteins, including well known and novel cold-responsive proteins. Several proteins showed enhanced degradation during chilling stress, especially the photosynthetic proteins such as Rubisco large subunit of which 19 fragments were detected. The identified proteins are involved in several processes, i.e. signal transduction, RNA processing, translation, protein processing, redox homeostasis, photosynthesis, photorespiration, and metabolisms of carbon, nitrogen, sulfur, and energy. Gene expression analysis of 44 different proteins by quantitative real time PCR showed that the mRNA level was not correlated well with the protein level. In conclusion, our study provides new insights into chilling stress responses in rice and demonstrates the advantages of proteomic analysis.

Among various abiotic stresses, low temperature (chilling and freezing temperature) is a major stress that limits the productivity and the geographical distribution of many important crops such as rice and maize. Chilling temperatures that range from 0 to 12°C are common during the growing season in temperate regions and can substantially decrease plant productivity (1). To defend against the stress, plants use several strategies, one of which is regulation of gene expression. The hypothesis that cold-responsive proteins are likely to be involved in cold tolerance has led to great efforts to study the gene expression profile during cold stress (2). Identification of novel responsive genes, determination of their expression patterns, and understanding their functions in stress responses will provide the molecular basis of effective engineering strategies leading to greater stress tolerance (3).
The transcriptome analyses of gene expression at the mRNA level have contributed greatly to our understanding of the cold responses in Arabidopsis and rice (4 -8). However, the level of mRNA does not always correlate well with the level of protein, the key player in the cell (9 -12). Therefore, it is insufficient to predict protein expression level from quantitative mRNA data. This is mainly due to post-transcriptional regulation mechanisms such as nuclear export and mRNA localization, transcript stability, translational regulation, and protein degradation (13). Proteome studies aim at the complete set of proteins encoded by the genome and thus complement the transcriptome studies. In the studies on the Arabidopsis proteome, the abundance of 38 plasma membrane proteins and 54 nuclear proteins was found to be altered after cold stress (14,15). In addition to nuclear and membrane proteins, many proteins are located in the cytosol and other organelles. Therefore, proteomic analysis of total proteins may provide new insights into the chilling stress responses.
Rice is not only an important crop worldwide but also a model plant for monocots because of its relatively small genome size. During the past several years, considerable research efforts have been made to analyze the rice proteome, and remarkable progresses have been achieved (16). Recently the effects of early cold stress on the maturation of rice anthers were investigated using proteomic approaches, and 70 differentially expressed protein spots were found (17).
In the present study, we analyzed the temporal changes of total proteins in rice leaves after chilling treatment and recovery. Eighty-five differentially expressed proteins were identified, including many novel cold-responsive proteins.

EXPERIMENTAL PROCEDURES
Plant Material and Stress Treatment-Rice seeds (Oryza sativa L. cv. Nipponbare, which was used in the International Rice Genome Sequencing Project) were allowed to germinate in the dark for 48 h at 28°C before being transplanted into nutrient solution. The seedlings were grown in a growth chamber with 28/25°C (day/night), photon flux density of 350 -400 mol m Ϫ2 s Ϫ1 , 16-h photoperiod, and relative humidity of 60ϳ80%. Three-week-old seedlings were used in the experiment. Low temperature treatment was started after 1 h of illumination by setting the temperature to 6°C, which was reached about 30 min later. The temperature was returned to 28°C after 24 h of treatment.
Relative Electrolyte Leakage Assay and Photosynthesis Measurement-The leaves were cut into 1-cm segments and washed three times with ultrapure water. The segments were placed in tubes containing 5 ml of ultrapure water and incubated at 28°C. Two hours later, the electrical conductivity of the bathing solution (L t ) was measured. Then the tubes were incubated at 100°C for 20 min and subsequently at 28°C for 1 h, and the electrical conductivity (L 0 ) was measured again. The relative electrolyte leakage was calculated by the formula L t /L 0 ϫ 100%. Five replicates were performed for each sample.
Leaf net photosynthetic rates, stomatal conductance, and intercellular CO 2 concentration were measured by a portable gas analysis system, LI-COR 6400 with a light-emitting diode light source (LI-COR Inc., Lincoln, NE). Eight leaves for each sample were measured.
Protein Extraction and Two-dimensional Electrophoresis-The leaf proteins were extracted using a modified trichloroacetic acid/acetone procedure as described previously (18). For 2-DE, 1 80 and 800 g of proteins were loaded onto analytical and preparative gels, respectively. The Ettan IPGphor system and pH 4 -7 IPG strips (13 cm, linear, Amersham Biosciences) were used for IEF. SDS-PAGE was performed with 12% gels using the PROTEAN II xi Cell system (Bio-Rad). The electrophoresis was performed as described previously (18). Protein spots in analytical gels were visualized by silver staining (19). The preparative gels were stained by a modified colloidal Coomassie Blue G-250 staining called Blue Silver (20). At least triplicate gels were performed for each sample.
Gel Scanning and Image Analysis-The gels were scanned by ScanMaker 4 (Microtek). Image analysis was accomplished using PDQuest 7.3 software (Bio-Rad). After automated detection and matching, manual editing was carried out. Three well separated gels of each sample were used to create "replicate groups." Statistic, quantitative, and qualitative "analysis sets" were created between the control group and each treated group. In the statistic sets, the Student's t test and significance level of 95% were chosen. In the quantitative sets, the upper limit and the lower limit were set to 1.5 and 0.66, respectively. Then the Boolean analysis sets were created between the statistic sets and the quantitative or qualitative sets. The spots from the Boolean sets were compared among three biological replicates. Only spots displaying reproducible change patterns were considered to be differentially expressed proteins.
MALDI-TOF/TOF MS Analysis and Database Searching-Mass spectra were acquired on a MALDI-TOF/TOF tandem mass spectrometer, ABI 4700 Proteomics Analyzer, and analyzed by GPS Explorer 2.0 software (Applied Biosystems, Foster City, CA,). The instrument was operated in reflectron mode and calibrated using the 4700 calibration mixture (Applied Biosystems). Both MS and MS/MS data were acquired with a neodymium:yttrium-aluminum-garnet laser at 200-Hz sampling rate. MS/MS mode was operated with 1-kV collision energy. The collision-induced dissociation was performed using air as the collision gas. For MS spectra, the peaks were calibrated by trypsin autodigestion peaks and smoothed. The signal-to-noise criterion was set to 25 or greater. The monoisotopic masses were processed for identification. For MS/MS spectra, the peaks were calibrated by default and smoothed. All peaks were deisotoped.
Database searching and both peptide mass fingerprinting (PMF) and MS/MS were performed using the MASCOT program included in GPS Explorer 2.0 software. The database was set to National Center for Biotechnology non-redundant (NCBInr) (updated on May 5, 2005), which contained 2,464,940 sequences including 33,175 sequences of rice and 154,691 sequences of green plant. The searching was performed first taking rice as taxonomy. If no positive results were obtained, the searching was performed again taking green plant as taxonomy. The other parameters for searching were enzyme of trypsin; one missed cleavage; variable modifications of carbamidomethyl (Cys), oxidation (Met), and pyro-Glu (N-terminal Glu); peptide tolerance of 0.15 Da; MS/MS tolerance of 0.2 Da; peptide charge of 1ϩ; and monoisotopic. Only significant hits, as defined by the MASCOT probability analysis (p Ͻ 0.05), were accepted.
Western Blotting-After 2-DE, the proteins were transferred to nitrocellulose membranes using a TE 70 semidry transfer unit (Amersham Biosciences). The membranes were subsequently blocked with 5% (w/v) nonfat milk for 1 h, incubated with anti-Rubisco large subunit antibody (a kind gift from Dr. Klimentina Demirevska-Kepova or Dr. Gen-Yun Chen) at 1:500 dilution for 2 h, and incubated with alkaline phosphatase-conjugated goat anti-rabbit IgG at 1:1,000 dilution for 1 h, and the signal was detected using the nitroblue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate method.
Quantitative Real Time PCR-Total RNA in leaves was extracted using RNAex reagent and systems (Watson Biotech, Shanghai, China). The potential contaminating genomic DNA was treated with DNase I using a DNA-free kit (Ambion, Austin, TX). The cDNA was synthesized using oligo(dT) 18 primer and ReverTra Ace Moloney murine leukemia virus reverse transcriptase (Toyobo, Osaka, Japan) according to the manufacturer's recommendation. Quantitative real time PCR (qPCR) was performed using gene-specific primers (Table  I) on a Chromo4 four-color real time PCR system (MJ Research, San Francisco, CA) using SYBR Green Realtime Master Mix (Toyobo). Polyubiquitin (UBI, NCBI accession number D12776) was used as the reference gene. The relative gene expression was evaluated using the comparative cycle threshold method (21).

The Physiological Responses Induced by Chilling Stress in
Rice-As a tropical and subtropical plant species, rice is sensitive to low temperature. Initially we treated rice seedlings at 4°C and found that they displayed very severe symptoms and could not fully recover after the temperature was returned to 28°C (data not shown). Therefore, we treated the seedlings at 6°C in the following experiments. The rice leaves rolled after 6 or 24 h of cold treatment and fully expanded after 24 h of recovery (Fig. 1A). To evaluate the adverse effects of chilling stress quantitatively, we performed the relative electrolyte leakage assay. The relative leakage increased from 1.5 to 14% after 6 h of cold treatment (Fig. 1B). However, the leakage at the 24-h time point (9%) was lower than that at 6 h. There might be two reasons for this. First, the seedlings had adapted to the chilling stress to some extent after 24 h of treatment and could protect the membrane from further damage. Second, it might be related to the concurrent effects of illumination because light may strengthen the damage of low temperature. After 12 h of darkness at night, the damage was alleviated. The leakage decreased and was close to the basal level after 24 h of recovery (2%).
changes of stomatal conductance (Gs) correlated with Pn, the intercellular CO 2 concentration showed few changes ( Fig. 1, D and E). This suggested that Gs was not the limiting factor for Pn in this condition. The decreased Pn probably resulted from other factors such as the activity of Rubisco and the availability of ATP.

Two-dimensional Electrophoresis Analysis of Total Proteins in Rice
Leaves-To investigate the temporal changes of protein profiles during chilling stress and recovery, we carried out 2-DE analysis of the total proteins in rice leaves from three biological replicates. For each sample, at least triplicate gels were performed, and they showed a high level of reproducibility. The representative gels are shown in Fig. 2. More than 1,000 protein spots were reproducibly detected by PDQuest 7.3 software on silver-stained gels. Quantitative image analysis revealed a total of 93 protein spots that changed their intensities significantly (p Ͻ 0.05) by more than 1.5-fold at least at one time point (Fig. 2). Four typical regions are en-larged in Fig. 3. Although most spots showed quantitative changes, some spots showed qualitative changes. For example, 10 spots (spots 34, 41, 47, 53, 58, 64, 72, 76, 83, and 84) absent in the control sample were induced after treatment (Figs. 2 and 3). Fig. 4 shows the Venn diagram analysis of the differentially expressed spots at different time points. All together, there were 31 down-regulated spots and 65 up-regulated spots. Three of them (spots 12, 27, and 30) were down-regulated at some time points but up-regulated at other time points. There were large overlaps between these spots. For the down-regulated spots, 10 of them were down-regulated both at 6 and 24 h, and 10 spots were down-regulated at all time points (Fig. 4A). Among the 65 up-regulated spots, 43 spots were up-regulated at all time points (Fig. 4B). Note that, although some spots were also down-or up-regulated after 24 h of recovery (R24 h), the changes were less pronounced than that of the 6-and 24-h time points (Fig. 3).
Identification of the Differentially Expressed Proteins-The Three-week-old seedlings were treated at 6°C for 0, 6, and 24 h and then were allowed to recover for 24 h (R24 h). The photographs of the same sample were taken at each treatment time point, and the framed regions were enlarged (A). The relative electrolyte leakage, the Pn, the Gs, and the intercellular CO 2 concentration (Ci) are shown in B, C, D, and E, respectively. differentially expressed proteins were excised from the preparative gels, in-gel digested by trypsin, and analyzed by a MALDI-TOF/TOF mass spectrometer. In total, 85 proteins were successfully identified either by PMF or MS/MS analysis (Tables II and III). The results of spot 71 are shown in Fig. 5 as an example. Some of the identified proteins were annotated either as unknown and hypothetical proteins or as proteins without specific function in the database. To gain the functional information about these proteins, we searched their homologues with BLASTP (www.ncbi.nlm.nih.gov/BLAST/) using their protein sequences as queries. Eleven corresponding homologues with the highest homology are shown in Table IV. Most spots except spot 10 shared more than 50% positives with homologues at the amino acid level, indicating that they might have similar function. Besides the unknown proteins, the rest of the identified proteins were classified into several functional categories including signal transduction, RNA processing, translation, protein processing, redox homeostasis, photosynthesis, photorespiration, and metabolisms of carbon, nitrogen, sulfur, and energy (Fig. 6). The largest functional category was proteins involved in photosynthesis (35.3%), which was greatly affected by chilling stress (Fig. 1C).
Interestingly several differentially expressed proteins ap-  peared to be the products of degradation because their observed M r values were much smaller than the theoretical ones. A particular case was the photosynthetic proteins, including Rubisco large subunit (RcbL) and Rubisco activase (RcbA), photosystem II oxygen-evolving complex protein 2, sedoheptulose-1,7-bisphosphatase, ATP synthase ␣ chain, and ATP   5. Identification of spot 71 by MS. The protein excised from gels was digested with trypsin, and the resulting peptides were analyzed using the 4700 Proteomics Analyzer. A, the MS spectra. The ion 2047.08 marked with an asterisk was analyzed by MS/MS. B, MS/MS spectra of ion 2047.08. The y ions (y3-y14) and the corresponding peptide sequence are shown. The protein was identified as ascorbate peroxidase (NCBI accession number BAB17666) after database searching. synthase CF1 ␤ chain. All together, 20 spots were identified as the same protein, RcbL, which is the most abundant protein in leaves. Although the observed molecular mass of spot 9 (53.8 kDa) was similar to the theoretical one (53.7 kDa), the observed molecular masses of other spots were 10.4 -39.4 kDa smaller than the theoretical ones. Notably although spot 9 was down-regulated at 6-and 24-h time points, the other RcbL spots were up-regulated. It seemed that spot 9 represented the intact RcbL and the others were the fragments. The postulation was further confirmed by Western blotting analysis (Fig. 7). Two anti-RcbL antibodies, kind gifts from Dr. Klimentina Demirevska-Kepova and Dr. Gen-Yun Chen, were used, and similar results were obtained. Both the intact RcbL and the fragments (mainly located at 20 -43 kDa) were detected. The intact RcbL decreased after treatment, whereas the fragments increased. The similar situation may happen to RcbA, photosystem II oxygen-evolving complex protein 2, and glycine dehydrogenase P protein of which both the intact proteins and the fragments were identified and showed reverse change patterns. The identified proteins ␤-galactosidase, myosin-like protein, pentatricopeptide repeat (PPR)-containing protein, and putative oxalyl-CoA decarboxylase may also be products of partial degradation as revealed by the discrepancy between their observed M r values and theoretical ones.
Gene Expression Analysis by qPCR-To investigate the changes of gene expression at the mRNA level, we performed qPCR analysis (Fig. 8). The genes encoding 44 different proteins from O. sativa in Tables II and III were analyzed. For the down-regulated proteins, the mRNA of most genes except spots 8 and 10 were also down-regulated under chilling stress (Fig. 8A). However, for the up-regulated proteins, only five genes (spots 33, 41, 52, 54, and 62) were also up-regulated by chilling treatment at the mRNA level. The other 22 spots showed different expression patterns between mRNA and protein at least at one time point (Fig. 8, B and C). Among the 22 spots, five spots (spots 47, 55, 61, 62, and 87) were protein fragments partially degraded by chilling stress (Tables II and  III). Notably after 24 h of recovery (R24 h), although most genes were remarkably up-regulated, RcbL (spot 9) and Rubisco small subunit (spot 31) were down-regulated. Our results confirmed the concept that the mRNA level is not correlated well with protein level (9 -12). DISCUSSION The transcriptome analyses have identified many cold-responsive genes in plants (4 -8) and revealed a complex network involved in cold responses. The present proteomic study identified not only some well known cold-responsive proteins such as enolase, RcbL, RcbA, ascorbate peroxidase, and heat shock proteins but also several novel proteins such as 2-Cys peroxiredoxin, armadillo repeat-containing protein, and putative nascent polypeptide-associated complex ␣ chain. These identified proteins can be classified into two groups. The first group consists of regulatory proteins, such as proteins involved in signal transduction, RNA processing, translation, and protein processing. The second group consists of functional proteins or proteins that probably function in stress tolerance, such as proteins involved in redox homeostasis, photosynthesis, photorespiration, and metabolisms of carbon, nitrogen, sulfur, and energy. In addition, our results demonstrated that many proteins might undergo enhanced degradation under chilling stress.
The Regulatory Network Involved in Chilling Stress Responses-Rice can perceive chilling stress signals by putative sensors and transmit them to the cellular machinery by signal transduction to regulate gene expression. Some candidate components involved in signal transduction were identified here, i.e. putative protein phosphatase 2C-like protein (spot 54), protein similar to WAK1 (spot 34), putative armadillo repeat-containing protein (spot 35), and RAB2A (spot 88). It is suggested that protein phosphatase 2C is a negative regulators of abscisic acid signaling, which plays a critical role in cold stress responses (22). RAB2A is a member of the Rab GTPase gene family, which has been implicated in intracellu-lar vesicle trafficking and in the organization of membranes. Increasing evidence indicates that it is also involved in the responses to stresses (23). Although WAK1 and armadillo repeat-containing protein were shown to be involved in other abiotic stress responses (24,25), our data showed, for the first time, that they were up-regulated by chilling stress. The signal transduction pathway is complex. The identified novel components will broaden our knowledge about this process.
Gene expression can be regulated at transcriptional, posttranscriptional, translational, and post-translational levels. Proteins involved in these processes were identified. PPRcontaining protein (spot 25) is an RNA-binding protein targeted to the chloroplast or mitochondrion and may be involved in RNA processing. Putative nascent polypeptide-  Tables II and III were analyzed. Quantitative real time PCR was performed using gene-specific primers (Table I)  associated complex (NAC) ␣ chain (spots 17 and 20), elongation factor 1-␤Ј (spots 18 and 19), putative acidic ribosomal protein P3a (spot 82), and heat shock proteins (spots 7, 38, and 40) were proteins involved in protein translation and processing. It is suggested that NAC is involved in protein sorting and translocation by preventing mistargeting of nascent polypeptide chains to endoplasmic reticulum. The NAC ␣ chain can also function as a transcriptional co-activator (26). Although the NAC ␣ chain and elongation factor 1-␤Ј were down-regulated, the putative acidic ribosomal protein P3a was up-regulated. The differential regulation of different components of the translation machinery suggests that there is a complicated mechanism controlling protein synthesis in response to chilling stress. Heat shock proteins act as molecular chaperones, and their up-regulation may play a pivotal role in preventing aggregation of the denatured proteins and facilitating the refolding under chilling stress.
The Functional Network Involved in Chilling Stress Responses-Through signal transduction and gene expression regulation, the abundance and activities of functional proteins change and might work cooperatively to establish a new cellular homeostasis under the stress condition. Growing evidence suggests that redox homeostasis is a metabolic interface between stress perception and physiological responses (27). Reactive oxygen species (ROS) readily produced in stress conditions can act as signaling molecules for stress responses. However, they can also cause damage to cellular components. Plants can control the ROS level through sophisticated mechanisms such as scavenging them by peroxidases of which three members (spots 23, 29, and 71) were identified in this study. Thioredoxin H-type (spot 92) is also involved in the redox regulation by reducing disulfide bridges on target proteins (28). Furthermore NADP-malate dehydrogenase (spot 45) serves as a redox valve by using excess NADPH to convert oxaloacetic acid to malate (29).
Photosynthesis is greatly inhibited by low temperature in various crops (1). However, the mechanism is still not fully understood. Our proteomic analysis showed that many photosynthetic proteins were partially degraded by chilling stress (see "Results"). Because the photosynthetic components are functionally linked, damage of any components may lead to the overall reduction of photosynthetic activity. Glycine dehydrogenase and RcbA are involved in photorespiration, which has been suggested to be important for maintaining electron flow to prevent photoinhibition under stress conditions (30). The degradation of glycine dehydrogenase P protein and RcbA will lead to the inhibition of photorespiration. This is probably one of the reasons why rice is sensitive to chilling stress.
The primary metabolisms, such as metabolisms of carbon, nitrogen, sulfur, and energy, need to be modulated to establish a new homeostasis under chilling stress (2). As expected, chilling stress up-regulated expression of enolase (spot 43) and putative 2,3-bisphosphoglycerate-independent phosphoglycerate mutase (spot 41), two enzymes involved in gly-colysis. Their up-regulation might help to produce more energy needed in various defense processes. In contrast, a putative aconitate hydratase (spot 4), which is involved in the tricarboxylic acid cycle, was down-regulated. Aconitate hydratase is exquisitely sensitive to ROS and was shown to be down-regulated by oxidative stress in Arabidopsis (31). Ferredoxin-nitrite reductase (spot 8) and glutamine synthetase shoot isozyme (spot 50) are two enzymes related to nitrogen metabolism. It has been reported that overexpression of a glutamine synthetase gene in rice increased photorespiration and enhanced salt and chilling stress tolerance (32). Two sulfur metabolism-related enzymes, putative plastidic cysteine synthase 1 (spot 57) and S-adenosylmethionine synthetase 2 (spot 48), were up-regulated during chilling stress. Cysteine synthase is responsible for the final step in cysteine biosynthesis, which is a key limiting step in the production of glutathione, a thiol implicated in resistance to biotic and abiotic stresses (33). S-Adenosylmethionine synthetase catalyzes the biosynthesis of S-adenosyl-L-methionine, a precursor for the biosynthesis of ethylene and polyamines. It has been reported that there was a close correlation between the chilling tolerance and polyamine accumulation level in rice under chilling stress (34). Energy metabolism was altered under chilling stress as revealed by the altered expression of adenylate kinase A (spot 77) and ATP synthase ␣ and ␤ chains. Adenylate kinase catalyzes a reversible transphosphorylation reaction interconverting ADP to ATP and AMP, an essential reaction for many processes in living cells, and thus is considered a key enzyme in energy metabolism (35). The ATP synthase ␣ and ␤ chains were degraded under chilling stress, which unavoidably resulted in decreased ATP production through photophosphorylation and thus affected the Calvin cycle in photosynthesis.
Chilling Stress Enhances Protein Degradation-Our proteomic analysis showed that several proteins were partially degraded by chilling stress, especially the components of the photosynthesis apparatus (Tables II and III). Similar to our results, the analysis of the rice anther proteome under cold stress identified seven proteins as breakdown products (17). These results indicated that low temperature enhanced the protein degradation. A typical example is RcbL of which 19 fragments were identified in this study. The degradation of RcbL was also reported in other proteomic studies (36 -38). Recently proteomic analysis of pea mitochondria also showed that the glycine dehydrogenase P protein was degraded by chilling stress (39). Moreover our results also provided evidence for the degradation of other photosynthetic proteins such as RcbA, photosystem II oxygen-evolving complex protein 2, sedoheptulose-1,7-bisphosphatase, and ATP synthase ␣ and ␤ chains for the first time. These results suggest that the photosynthesis apparatus is susceptible to chilling stress, which may be one of the major reasons for decreased net photosynthetic rate under stress condition (Fig. 1C). It has been suggested that ROS may modify Rubisco, facilitating its subsequent degradation by proteases (40). It is highly possible that ROS may also account for the degradation of other proteins under chilling stress. The degradation of proteins cannot be concluded from transcriptome analysis, thus demonstrating the advantages of proteome analysis.
Conclusions-In this study, the molecular responses to chilling stress were investigated at the protein level in rice. Ninety-three differentially expressed proteins were revealed, and 85 of them were further identified by MS analysis. These proteins were involved in several processes that might work cooperatively to establish a new homeostasis under chilling stress. A simple model of chilling stress responses is outlined in Fig. 9. The identification of novel cold-responsive proteins provides not only new insights into chilling stress responses but also a good starting point for further dissection of their functions using genetic and other approaches. Rice can perceive chilling stress signals by putative sensors and transmit them to the cellular machinery by signal transduction to regulate gene expression. Through regulation of transcription, RNA processing, translation, and protein processing, rice change the abundance and activities of functional proteins involved in redox homeostasis, photosynthesis, photorespiration, and metabolisms of carbon, nitrogen, sulfur, and energy. Redox homeostasis can also act as signals. These processes might work cooperatively to establish a new cellular homeostasis under chilling stress.