Comparative Proteomics Analysis of Differentially Expressed Proteins in Chickpea Extracellular Matrix during Dehydration Stress*S

Water deficit or dehydration is the most crucial environmental factor that limits crop productivity and influences geographical distribution of many crop plants. It is suggested that dehydration-responsive changes in expression of proteins may lead to cellular adaptation against water deficit conditions. Most of the earlier understanding of dehydration-responsive cellular adaptation has evolved from transcriptome analyses. By contrast, comparative analysis of dehydration-responsive proteins, particularly proteins in the subcellular fraction, is limiting. In plants, cell wall or extracellular matrix (ECM) serves as the repository for most of the components of the cell signaling process and acts as a frontline defense. Thus, we have initiated a proteomics approach to identify dehydration-responsive ECM proteins in a food legume, chickpea. Several commercial chickpea varieties were screened for the status of dehydration tolerance using different physiological and biochemical indexes. Dehydration-responsive temporal changes of ECM proteins in JG-62, a relatively tolerant variety, revealed 186 proteins with variance at a 95% significance level statistically. The comparative proteomics analysis led to the identification of 134 differentially expressed proteins that include predicted and novel dehydration-responsive proteins. This study, for the first time, demonstrates that over a hundred ECM proteins, presumably involved in a variety of cellular functions, viz. cell wall modification, signal transduction, metabolism, and cell defense and rescue, impinge on the molecular mechanism of dehydration tolerance in plants.

Water deficit or dehydration is the most crucial environmental factor that limits crop productivity and influences geographical distribution of many crop plants. It is suggested that dehydration-responsive changes in expression of proteins may lead to cellular adaptation against water deficit conditions. Most of the earlier understanding of dehydration-responsive cellular adaptation has evolved from transcriptome analyses. By contrast, comparative analysis of dehydration-responsive proteins, particularly proteins in the subcellular fraction, is limiting. In plants, cell wall or extracellular matrix (ECM) serves as the repository for most of the components of the cell signaling process and acts as a frontline defense. Thus, we have initiated a proteomics approach to identify dehydrationresponsive ECM proteins in a food legume, chickpea. Several commercial chickpea varieties were screened for the status of dehydration tolerance using different physiological and biochemical indexes. Dehydration-responsive temporal changes of ECM proteins in JG-62, a relatively tolerant variety, revealed 186 proteins with variance at a 95% significance level statistically. The comparative proteomics analysis led to the identification of 134 differentially expressed proteins that include predicted and novel dehydration-responsive proteins. This study, for the first time, demonstrates that over a hundred ECM proteins, presumably involved in a variety of cellular functions, viz. cell wall modification, signal transduction, metabolism, and cell defense and rescue, impinge on the molecular mechanism of dehydration tolerance in plants.

Molecular & Cellular Proteomics 6:1868 -1884, 2007.
Environmental stress is a primary cause of crop loss worldwide, resulting in average yield losses of more than 50% for major crops every year (1,2). Among the unfavorable environmental conditions, water deficit or dehydration is the most significant factor that adversely affects plant growth, devel-opment, and productivity. In recent years, the physiological and molecular basis for plant responses to dehydration tolerance has been the subject of intense research (3,4). It is now well established that plants, being sessile, have evolved many adaptations to counteract dehydration. These adaptations are classified into four categories: dehydration avoidance (developmental and physiological traits), dehydration tolerance (physiological and biochemical adaptations), dehydration escape, and dehydration recovery (5). Dehydration response in plants is a complex phenomenon, and the exact structural and functional modification caused by dehydration is poorly understood. Thus, understanding the responses of plants to dehydration is of importance not only for basic research but also an attractive target for improving such stress tolerance. A few characteristics such as osmotic adjustment (OA) 1 and cell membrane stability are recognized as effective components of dehydration tolerance in many crops. These are expressed in terms of relative water content (RWC) of the plant, accumulation of compatible solutes like proline, and increased permeability of ions and electrolytes. In addition the status of photosynthetic machinery has been considered as an ideal index to monitor the health and vitality of plants during dehydration (6).
Most studies on dehydration to date have mainly focused on the changes in gene expression, whereas there is far less information available on their functional products. The changes in gene expression are regulated by a number of different, and potentially overlapping, signal transduction pathways (7,8). Nevertheless the significance of these genes in dehydration tolerance is incomplete without the knowledge of their function. Hence proteomics analysis is necessary to reveal the plasticity of gene expression because this allows the global analyses of gene products and physiological state of cells. However, the application of a proteomics approach at the whole cell level is limited by several factors such as protein abundance, size, hydrophobicity, and other electrophoretic properties (9). The compartment-specific proteome is thus recommended because fractionated subsets of proteins provide the suitable information in which they exert their particular function (10 -12). Plant cell wall or extracellular matrix (ECM) is the first compartment that senses the stress signals, transmits them to the cell interior, and eventually influences the cell fate decision (11,13). Increasing evidence suggests that there is continuous cross-talk between ECM and the cytoskeleton network (14). Within the past few years, there have been rapid advances in ECM research (15). However, most research interest on ECM has been in the carbohydrate components due to their structural role and commercial interests, whereas the study of the complexity of ECM proteins remained secondary (16).
Legumes are valuable agricultural and commercial crops mostly cultivated in the arid and semiarid part of the world and are particularly important as nutrient sources for the human diet and animal feed (17). They are among the few selfautonomous organisms capable of fixing carbon and nitrogen, and the estimates of nitrogen fixation by food legumes is about 300 kg/hectare/year. Chickpea is the most important food legume and ranks third in terms of total global production (18,19). The susceptibility of chickpea to dehydration severely reduces the yield, and its productivity has remained historically low. Although the simultaneous changes in gene expression and physiological responses strongly suggest that induced proteins play a role in these responses, the correlation between their expression and the level of stress tolerance in different genotypes has rarely been studied. In this study, dehydration-induced physiological responses were monitored in eight commercial chickpea varieties using various parameters, viz. RWC, proline accumulation, lipid peroxidation, electrolyte leakage, and the status of photosynthetic pigments. The altered physiological status of the cell in response to dehydration was correlated with temporal changes in the ECM/cell wall proteome in a tolerant variety, JG-62. In a previous study, we reported the ECM proteome map of chickpea (11). Here we used the ECM-specific comparative proteome of chickpea to identify novel components involved in dehydration tolerance with a wider aim to use them in a future crop improvement program. The quantitative image analysis revealed 163 protein spots that changed their intensities significantly (p Ͻ 0.05) by more than 2.5-fold at least at one time point during dehydration. A total of 134 differentially expressed ECM proteins were identified during the course of dehydration using classical two-dimensional electrophoresis (2-DE) coupled with LC-MS/MS. The comparison of the dehydration-responsive ECM proteome in chickpea revealed predicted and unexpected components indicating their possible role in dehydration tolerance.

EXPERIMENTAL PROCEDURES
Plant Growth, Maintenance, and Dehydration Treatment-Seeds of eight commercial chickpea (Cicer arietinum L.) cultivars (Vijay, WR-315, Annigiri, CPS-1, K-850, JG-62, C-235, and ICCV2) were obtained from the International Crops Research Institute for the Semi-Arid Tropics, Hyderabad, India and grown in a mixture of soil and soilrite (2:1, w/w; 10 plants/1.5-liter-capacity pot with 18-cm diameter) in an environmentally controlled growth room. The seedlings were maintained at 25 Ϯ 2°C, 50 Ϯ 5% relative humidity under a 16-h photoperiod (270 mol m Ϫ2 s Ϫ1 light intensity). The pots were provided with 100 ml of water every day that maintained the soil moisture content to approximately 30%. A gradual dehydration condition was applied on the 3-week-old seedlings for a period of 7 days by withdrawing water, and tissues were harvested at every 24 h. The pots were then rewatered allowing complete recovery of the stressed seedlings for 24 h. The unstressed and the stressed plants were kept in parallel in the same growth room. The samples from the unstressed (control) plants were collected every day during the course of the dehydration experiment and were finally pooled. The tissues were harvested, instantly frozen in liquid nitrogen, and stored at Ϫ80°C unless described otherwise.
Determination of Relative Water Content-Chickpea tissues were collected and immediately weighed (fresh weight (FW)). The tissues were rehydrated in water for 24 h until fully turgid, surface-dried, reweighed (turgid weight (TW)) followed by oven drying at 80°C for 48 h, and reweighed (dry weight (DW)). The RWC was calculated by the following formula: RWC (%) ϭ (FW Ϫ DW/TW Ϫ DW) ϫ 100 (20). The experiment was carried out in triplicates.
Proline Estimation-Free proline content was measured as described earlier (21). The tissues were homogenized in 3% aqueous sulfosalicylic acid, and the homogenate was centrifuged at 9000 ϫ g. The reaction mixture consisted of 2 ml of supernatant, 2 ml of acid ninhydrin, and 2 ml of glacial acetic acid, which was boiled at 100°C for 1 h. After termination of the reaction on ice, the reaction mixture was extracted with 4 ml of toluene, and the absorbance was read at 520 nm. The assays were done in triplicates using corrected weight calculated for the actual moisture content of tissue at each time point. Proline was calculated according to the following formula: proline (mg/g) ϭ 36.6 ϫ A 520 ϫ volume/2 ϫ fresh weight.
Determination of Lipid Peroxidation-Lipid peroxidation was estimated in terms of malondialdehyde (MDA) production. Fresh chickpea tissues were ground in 5% TCA. One part of the supernatant was mixed with four parts of 0.5% thiobarbituric acid prepared in 20% TCA and heated at 95°C for 30 min. Absorbance at 532 nm was measured and corrected for nonspecific absorbance at 600 nm. The pellet was washed in 100% chilled acetone and suspended in 0.1 N NaOH to estimate total protein using Bradford reagent. Oxidized MDA was calculated according to the following formula: MDA (g/mg) ϭ ((A 532 Ϫ A 600 ) ϫ volume ϫ 100)/(155 ϫ total protein (mg)). The experiments were carried out in triplicates using corrected weights calculated for actual moisture content of the seedlings at each time point (22).
Electrolyte Leakage Assay-Electrolyte leakage was assayed by estimating the ions leaching from the leaflets into Milli-Q water. Plant materials were placed in 20 ml of Milli-Q in two sets. The first set was kept at room temperature for 4 h, and its conductivity (C1) was recorded using a conductivity meter. The second set was autoclaved, its conductivity was also recorded (C2), and electrolyte leakage (1 Ϫ (C1/C2)) ϫ 100 was calculated. The experiments were carried out in triplicates.
Isolation of ECM Fraction and 2-DE-The ECM protein fraction was essentially isolated as described earlier (11). In brief, the tissues were ground to powder in liquid nitrogen with 0.3% (w/w) polyvinylpolypyrrolidone and transferred to an open mouthed 50-ml centrifuge tube. Immediately tissue powder was homogenized in homogenizing buffer (5 mM K 3 PO 4 , pH 6.0, 5 mM DTT, 1 mM PMSF) for 1-2 min. The cell wall fraction was recovered by differential centrifugation at 1000 ϫ g for 5 min at 4°C. The pellet thus obtained was washed 10 times with excess deionized water. The purified ECM fraction was suspended in 3 volumes (w/v) of extraction buffer (200 mM CaCl 2 , 5 mM DTT, 1 mM PMSF, 0.3% (w/w) polyvinylpolypyrrolidone) and extracted on a shaking platform for 45 min at 4°C. Proteins were separated from the insoluble ECM fraction by centrifugation (10,000 ϫ g) for 10 min at 4°C and filtered through a 0.45-m filter. The filtrate was concentrated using Centricon YM3 and then dialyzed overnight against 1000 volumes of deionized water with one change. The concentration of the protein extract was determined by Bradford assay (Bio-Rad).
Isoelectric focusing was carried out with 125 g of protein sample in 250 l of two-dimensional rehydration buffer for 13-cm gel strips. Protein was loaded by the in-gel rehydration method onto IEF strips (pH range, 4 -7). Electrofocusing was performed using an IPGphor system (Amersham Biosciences) at 20°C for 30,000 V-h. The focused strips were subjected to reduction with 1% (w/v) DTT in 10 ml of equilibration buffer (6 M urea, 50 mM Tris-HCl (pH 8.8), 30% (v/v) glycerol and 2% (w/v) SDS) followed by alkylation with 2.5% (w/v) iodoacetamide in the same buffer. The strips were then loaded on top of 12.5% polyacrylamide gels for SDS-PAGE. The electrophoresed proteins were stained with a Silver Stain Plus kit (Bio-Rad).
Image Acquisition and Data Analysis-Gel images were digitized with a Bio-Rad FluorS system equipped with a 12-bit camera. PDQuest version 7.2.0 (Bio-Rad) was used to assemble a first level match set (master image) from three replicate 2-DE gels. Experimental molecular mass and pI were calculated from digitized 2-DE images using standard molecular mass marker proteins. Each spot included on the standard gel met several criteria: it was present in at least two of the three gels and was qualitatively consistent in size and shape in the replicate gels. We defined "low quality" spots as those with a quality score less than 30; these spots were eliminated from further analysis. The remaining high quality spot quantities were used to calculate the mean value for a given spot, and this value was used as the spot quantity on the standard gel. The first level match set spot densities were normalized against the total density in the gel image. The replicate gels used for making the first level match set had at least a correlation coefficient value of 0.8. After obtaining the first level match sets, a second level match set that allowed a comparison of the standard gels from each of the time points was obtained. A second normalization was done with a set of three unaltered spots identified from across the time points. From this match set, the filtered spot quantities from the standard gels were assembled into a data matrix of high quality spots from the eight time points for further analysis.
Protein Identification and Expression Clustering-Protein samples were destained and trypsin-digested, and peptides were extracted according to standard techniques (23). Peptides were analyzed by electrospray ionization time-of-flight mass spectrometry (LC/MS/ TOF) using an Agilent 1100 Series HPLC system (Agilent Technologies) coupled to a Q-STAR Pulsar imass spectrometer (Applied Biosystems). Tryptic peptides were loaded onto a Zorbaxா SB-C 18 column (15-cm length; Agilent Technologies) and separated with a linear gradient of water, acetonitrile, 0.1% formic acid (v/v). The MS/MS data were extracted using Analyst Software version 1.4.1 (Applied Biosystems). Peptides were identified by searching the peak list against the Mass Spectrometry Protein Sequence Database (MSDB) September 29, 2005 (2,344,227 sequences; 779,380,795 residues) using the Mascot version 2.1 (Matrix Science) search engine. Because the chickpea genome sequence is not known, a homology-based search was performed. The database search parameters were taxonomy, Viridiplantae (green plants; 195,697 sequences); peptide tolerance, Ϯ1.2 Da; fragment mass tolerance, Ϯ0.6 Da; maximum allowed missed cleavage, 1; instrument type, ESI-Q-TOF. Protein scores were derived from ion scores as a nonprobabilistic basis for ranking protein hits and protein scores as the sum of a series of peptide scores. The score threshold to achieve p Ͻ 0.05 is set by the Mascot algorithm and is based on the size of the database used in the search. The details regarding the precursor ion mass, expected molecular weight, theoretical molecular weight, delta, score, rank, charge, number of missed cleavages, p value, and the peptide sequence for proteins identified with a single peptide are listed in Supplemental Document 1. In addition the fragment spectra for these proteins are provided in Supplemental Document 2. Wherein there were more than one accession number for the same peptide, the match was considered in terms of putative function. In the case of the same protein being identified in multiple spots where several peptides were found to be shared by the isoforms, a differential expression pattern was observed for each of the candidates, and thus the proteins were listed as independent entities. The function of each of the identified protein was analyzed in view of the metabolic role of the candidate protein in the extracellular matrix. Because the functional annotation is based on Pfam and InterPro, the functional redundancy, if any, is thus greatly minimized.
SOTA (self-organizing tree algorithm) clustering was performed on the log-transformed -fold induction expression values across eight time points by using MultiExperiment Viewer (MEV) software (The Institute for Genomic Research). The clustering was done with the Pearson correlation as distance with 10 cycles and a maximum cell diversity of 0.8 (24).

RESULTS AND DISCUSSION
Screening of Chickpea Varieties for Dehydration Tolerance-Dehydration tolerance, in plants, is attained by two major physiological events: (a) the maintenance of a high water status during stress and (b) the maintenance of function at low water status. RWC is considered to be the best integrated measure of plant water status, which represents variations in water potential (WP), turgor potential, and OA. The advantage of RWC over WP for assessing genetic differences in dehydration resistance is that although having the same WP genotypes may vary in their RWC due to the respective difference in OA. The choice of RWC as the best representation of plant water status in terms of genetic variation is also supported by genetic association between RWC and plant production under dehydration (25). In this study, eight commercial varieties of chickpea were evaluated for their relative tolerance to dehydration. The initial soil moisture content was ϳ30%; this gradually decreased to zero after 72 h of withdrawal of water. Upon applying dehydration stress, the seedlings of all the varieties showed a steady increase in RWC around 72 h after a gradual decline at initial stages. This could be substantiated by the observation that proline content increased during 24 -72 h causing probable OA. When com-pared with the other varieties, maximum proline accumulation was observed in JG-62 during dehydration; this can be attributed to the maximum (85%) increase in RWC (Fig. 1, A and B).
The capacity to avoid or repair membrane damage during dehydration processes is pivotal for the maintenance of membrane integrity, especially for those membranes in which functional proteins are embedded. To determine the level of membrane integrity, the status of electrolyte leakage was monitored in all eight varieties under dehydration. The electrolyte leakage was found to be maintained at an almost constant level in JG-62 until 96 h and only showed a marginal increase in later stages of dehydration (Fig. 1D). By contrast, ICCV2 showed a sharp rise in electrolyte leakage accompanied by maximal lipid peroxidation (Fig. 1, C and D). Considering electrolyte leakage and lipid peroxidation, which are considered to be a direct indicator of dehydration tolerance, JG-62 displayed maximum maintenance of cell membrane integrity under dehydration.
Plants that are osmotically adjusted and maintain a high photosynthetic rate have a higher growth rate and higher productivity under dehydration (26). To understand the photosynthetic capabilities of the genotypes studied, the status of photosynthetic pigments, viz. chlorophyll a, chlorophyll b, and carotenoid, was determined in all the varieties under dehydration. Better maintenance of the photosynthetic apparatus was observed in C-235 and JG-62, which also showed the highest protein content under dehydration, whereas maximum damage was observed in Annigiri and ICCV2 (Fig. 2,  A-D). All the varieties showed an initial decline in photosynthetic pigments, whereas a general increase in pigments was observed around 72 h. It is interesting to note that during this period there was an increase in RWC presumably as a result of increased proline accumulation. It is likely that maintenance of higher RWC in JG-62 because of OA at lowered WP could maintain growth and metabolic activities including photosynthesis and other physiological processes. After 96 h, a constant decline in RWC and proline accumulation indicated the severity of dehydration where OA might fail to maintain the turgor in affected tissues. All together these results suggest JG-62 as a potential dehydration-tolerant variety as compared with the remaining varieties studied.
Dehydration-induced Changes and 2-DE Analysis-Dehydration-induced responses in plants involve nearly every aspect of plant physiology and metabolism. We subjected 3-week-old JG-62 seedlings to gradual dehydration. There were no visible changes in the seedlings during 24 h of dehydration. The leaflets rolled after 36 h of dehydration, and the damage increased further during 48 -192 h (Fig. 3). When the dehydrated seedlings were rewatered for 24 h, they fully recovered. The temporal changes in the ECM proteome for the dehydrated as well as stress-recovered seedlings were monitored using 2-DE on proteins isolated from the cell wallenriched (purified cell wall) fraction during dehydration (0 -192 h) followed by a recovery stage as detailed under "Experimental Procedures." The purity of the cell wall fraction was estab-lished by transmission electron microscopy and phosphotungstic acid staining (11). The control samples refer to the protein prepared from the pooled unstressed seedlings harvested from different time periods during the course of the experiment to normalize the effect, if any, of growth and development of plants. In addition we compared the ECM Three-week-old seedlings were subjected to dehydration and then were allowed to recover for 24 h. The ECM proteins were isolated from the seedlings challenged with dehydration at every 24 h for 192 h duration followed by a recovery stage (R24). The photographs of the same sample were taken at each time point, and framed regions depict the leaflets from plants under dehydration. An equal amount (125 g) of protein from each time point was separated by 2-DE. Three replicate gels for each time point were computationally combined using PDQuest software to generate the standard gel. proteome of unstressed seedlings of day 0 with that of day 8, which showed no significant difference in protein expression (Supplemental Fig. 1). For each time point, three replicate 2-DE gels were run that were then computationally combined into a representative standard gel (first level match set; Supplemental Fig. 2 and Fig. 3). Only those spots that met several stringent criteria (classified as "high quality" spots) were used to estimate spot quantities; otherwise a large number of protein spots were included in the match set. For example, 271 spots were detected in the control gel, but 257 were classified as high quality (Table I). The spot densities at the lower level were normalized against the total density present in the respective gel to overcome the experimental errors introduced due to differential staining. To make comparisons between the time points, a second level match set was created (Fig. 4). A second normalization was done against the densities of unaltered protein spots (Fig. 6, i-j). From the higher level match set, the filtered spot quantities were assembled into a data matrix that consisted of 420 unique spots indicating change in intensity for each spot during dehydration. The data reveal that nearly 94% of the spots on the standard gels were of high quality reflecting the reproducibility of the experimental replicates (Table I).
As the set of quantitative proteome data in this study is too large to discuss the relative expression of each individual protein by itself, we analyzed the data by creating statistic and quantitative analysis sets. In the statistic sets, 420 spots were subjected to t test at a significance level of 95% resulting in 186 substantially differential spots between the time points. The results revealed 163 DRPs that showed more than 2.5fold differences in expression values at least at one time point. Of the 163 DRPs, 25 proteins were clearly up-regulated, and 38 proteins were down-regulated, whereas 100 proteins showed a mixed pattern of time-dependent expression. MS/MS analysis was carried out for 134 DRPs resulting in 98 proteins with a significant match, whereas the rest showed no significant match in the database.
Dehydration-responsive Differential ECM Proteome-Dehydration resistance mechanisms can be classified into two general categories, avoidance and tolerance mechanisms; and accordingly, the dehydration-responsive ECM proteins could be distinguished. The ECM-resident DRPs are expected to play a variety of functions during cellular adaptation against dehydration. A total of 80 DRPs, identified with a significant match, were grouped into six functional classes based on their putative roles in the dehydration response ( Fig. 5 and Table II). The protein functions were assigned using a protein function database Pfam (pfam.sanger.ac.uk/) or InterPro (www.ebi.ac.uk/interpro/). The criteria used to address the subcellular localization of the cell wall proteins, in this study, were based on their presence in other reported cell wall proteomes in addition to the predicted biochemical and biological functions.
The communication between the cytoskeleton and the ECM   Fig. 3. The boxed areas marked with dotted lines represent the magnified gel sections, and the encircled areas are the representative unaltered spots used for the second normalization. The numbers correspond with the spot numbers mentioned in Table II. is one of the most characteristic features of cellular mechanics that allows cells to respond effectively to various extracellular signals (27). Several candidate components involved in signal transduction were identified in this study, for example,   CaE-20, -63, -78, -92, -117, -189, -214, and -354) and to a 2C-type protein phosphatase in the cytoplasm and form a signalosome complex (28). These results suggest that WAKs might play a critical role in dehydration tolerance. CHRK1 belongs to the family of receptor kinases where it has a chitinase (CaE-206)-specific interacting domain. Both CHRK1 and chitinase were found to be up-regulated between 48 and 120 h of dehydration. These proteins are known to be induced under different environmental stresses (29), and elucidating their possible role in dehydra-tion tolerance is of importance. NDK1 is known to interact with cytosolic catalases and play a role in relieving oxidative stress (30). It is also an important component for maintaining stable GTP levels through nucleotide homeostasis in various GTP-mediated signal transduction pathways (31). The tubbylike protein (CaE-61) was found to be up-regulated at 72 h. It belongs to the F-box family proteins that encode transcription factors and is known in animals to shuttle between the plasma membrane and the nucleus on environmental cue (32). Unusual proteins like translation initiation factor were also observed that may play a more diverse role in ECM than in its classical com-TABLE II-continued a Spot number as given on the 2-D gel images. The first letters (Ca) signify the source plant, C, arietinum, followed by the subcellular fraction, ECM (E). The numerals indicate the spot numbers corresponding to Fig. 4.
b Time kinetics represents the average change in spot density at various time points C (control) and 24,48,72,96,120,192, and R24 (recovery stage) h. The data were taken in terms of -fold expression with respect to the control value and were log-transformed to the base 2 to level the scale of expression and to reduce th noise. c NP represents the number of peptides. d Gene identification number as in GenBank TM . partment. The differential regulation of different components of the signaling network in ECM suggests that there is a complicated mechanism controlling the perception and transduction of various extracellular signals during dehydration. As mentioned above, the JG-62 seedlings showed a drop in cellular RWC after 24 h followed by an increased accumulation of proline in the next 48 h (Fig. 1, A and B). It is suggested that the mechanism of dehydration avoidance involves accumulation of compatible solutes and cell wall modification. A cellulose synthase (CaE-204 and -467) was detected at 24 and 48 h but could not be detected during 72-120 h of dehydration. This could be due to an increase in the accumulation of certain soluble sugars probably through inhibiting enzymes involved in cellulose biosynthesis (33). For example, glycosyl hydrolases such as ␤-galactosidase (CaE-60a; Fig.  6h), xyloglucan hydrolase (CaE-305), and hexosaminidase (CaE-360) utilize cell wall polysaccharides as an alternate carbon source under sugar depletion in the cytoplasm. Furthermore sedoheptulose (CaE-95) earlier identified in the secondary cell walls of pea was implicated to serve as an alternative sugar in the developing xylem (34). It is assumed that under stress ECM serves as an alternative source for sugar moieties that are redirected in the cytosol for turgor maintenance. An alternate mechanism to avoid water loss under dehydration could be the lignification of the cell wall, and a step toward this is the methylation of the lignin monomers. The methylation of the lignin precursors is carried out by S-adenosylmethionine synthetase from S-adenosylmethionine, a common methyl group donor. The expression of genes encoding methionine synthase (CaE-26, -93, -140, -267, and -491) contributes to the production of methionine from which S-adenosylmethionine is generated (35,36). This might be explained by a demand for more methyl groups for lignin methylation.
A sharp drop in both RWC and proline accumulation was observed in JG-62 seedlings (Fig. 1, A and B) after 96 h of dehydration, indicating their inability to maintain the turgor in affected tissues. It is conceivable that with a further drop of RWC it becomes increasingly difficult to avoid dehydration, and thus mechanisms to tolerate water deficit become important. Peroxides and other redox compounds play an important role in the stress perception of the apoplast, which acts as a bridge between the environment and the symplast (37). The direct measure of peroxides in the cell comes from estimating lipid peroxidation under dehydration. Under progressive dehydration, lipid peroxidation increased drastically after 96 h with a concurrent drop in RWC (Fig. 1), indicating the severity of the stress condition.
Reactive oxygen species (ROS) produced under dehydration can act as signaling molecules for the stress response, but above a certain threshold they can cause damage to many cellular components. Most of the dehydration tolerance mechanisms primarily involve protection of the cellular structure wherein an important method is the control of the level of ROS or the limitation of damage caused by ROS. It is reported that NADH is generated by the malate dehydrogenase (CaE-21 and -444; Fig. 6f) and used to form H 2 O 2 (38) possibly by NADH oxidase on the plasmalemma. Alternately quinone oxidoreductase (CaE-303) produces peroxides by oxidizing quinones, which are generated by diamine oxidases (39). Plants control ROS levels through sophisticated mechanisms such as scavenging them by peroxidases like ascorbate peroxidase (CaE-54b and -73; Fig. 6b). In addition thioredoxin (CaE-76, -82, and -97; Fig. 6, e and h) is also involved in redox regulation by reducing disulfides on the target protein and relaying the signal to MAPKK (mitogen-activated protein kinase kinase) pathway of stress signaling (40). Earlier we identified a chickpea ECM ferritin (11), which is known to sequester the highly reactive Fe 3ϩ and prevent the formation of toxic OH ⅐ species (41). We showed, for the first time, that the ECM ferritin (CaE-87b, -68, -40, and -143) is induced by dehydration (Fig. 6, b and d) and might play a crucial role in dehydration tolerance. Other redox proteins such as ferredoxin-NADP reductase and NADH dehydrogenase are also involved in the ROS pathway (42,43). Another interesting protein that was expressed only under stress conditions was mannose lectin (CaE-202). There are reports suggesting that cell wall lectins may be involved in compensatory mechanisms, which stabi-

Dehydration-responsive Chickpea ECM Proteome
lize the cytoskeleton structure under stress conditions (44). Furthermore aldolases (CaE-35 and -59) are known to act as adhesion molecules between ECM and the cytoskeleton and may be involved in cross-talk between the extracellular space and the cell interior (45).
The involvement of multiple dehydration-responsive components in a variety of cellular functions is important in understanding and solving the problems of such stress tolerance. There are several reports that suggest extensive crosstalk among environmental stress-responsive pathways (7,46,47). We observed the expression of chitinase (CaE-206), between 48 and 120 h of dehydration, that is known to be induced under osmotic stress (1). The Rps 4 fragment (CaE-328), an R gene, was detected exclusively at 192 h of dehydration. The GRPs form a novel class of proteins that play pleiotropic roles both in osmotic stress and biotic stress (48). Induction of GRPs at different time intervals of dehydration could be attributed to their multiple roles (Fig. 6a). The role of glyoxalase-1 (CaE-192) under salt stress has been well studied (49). Although glyoxalase-1 has been detected in cell wall proteomes of maize and Medicago (50,51), its role in the ECM and its involvement in dehydration response are not understood. Proteins like leucine aminopeptidase (CaE-49) and glutamate synthase (CaE-414) are known to be induced under various abiotic and biotic stresses and modulate the turnover of selected peptide or protein populations under stress (52,53). Glutamate synthase particularly regulates the accumulation of proline in osmotically stressed plants (54). Candidates like epoxide hydrolase (CaE-18) are induced under fungal stress, although its role under dehydration is not known. Roles of miscellaneous dehydration-responsive candidates like aspartate carbamoyltransferase (CaE-294) and ATP synthase (CaE-123, 151) need to be evaluated.
In this study, several differentially expressed proteins appeared to be the products of degradation because their observed M r values were much smaller than the theoretical values. We observed this phenomenon in a few class of proteins including cell wall-modifying enzymes (cellulose synthase and methionine synthase), signaling proteins (wall-associated kinase 4 and receptor kinase CHRK1), and cell defense and rescue (glyoxalase-1) among others. A particular case was that of methionine synthase represented by three spots (Table II). The observed M r values of two of the spots (CaE-93 and -267) were smaller than the theoretical M r values, which might be due to the stress-responsive degradation of proteins (55)(56)(57). As previously noted, this stress-responsive degradation of proteins can be attributed to the ROS-mediated pathway (58). It is very likely that ROS may also account for the degradation of other proteins.
Expression Profile of Dehydration-responsive Proteins-To achieve a comprehensive overview of expression profile in terms of protein function, SOTA clustering was done for the 80 DRPs. The data were taken in terms of -fold expression with respect to the control expression value. In addition the data sets were log-transformed to the base 2 to level the scale of expression and to reduce the noise. The analysis yielded 11 expression clusters, and only the clusters with n Ͼ 6 were used to study the co-expression patterns for functionally sim-ilar proteins (Fig. 7). Detailed information on proteins within each cluster can be found in Supplemental Fig. 3.
The most abundant group, Clusters 3 and 4, were found to be early dehydration-responsive and showed induction at all time points reaching maxima at 96 -120 h. These groups were found to be enriched in proteins involved in cell defense and rescue. The time span of expression of cell defense proteins entails the kinetics of dehydration tolerance in chickpea, JG-62 in particular. Proteins involved in signaling, cell wall modification, and metabolism displayed a diverse and complex pattern of regulation (Clusters 1, 2, 4, and 11) wherein no clear clustering pattern was observed. The cellular defense strategy seems to be dependent on the advancement of dehydration, and thus both early and late responses could be seen. Also the miscellaneous class of proteins showed no clear clustering patterns, which may be due to heterogeneous composition of this class. However, co-expression patterns were observed for proteins with unidentified function (Clusters 2 and 11), and identification of these proteins can provide valuable insight into kinetics of dehydration tolerance mechanisms.
Conclusions-In plants, the mechanisms of dehydration avoidance versus tolerance form the basis for understanding and interpreting the dehydration-responsive events, which may help in formulating strategies for defense. Nevertheless there are many overlapping mechanisms that fragment the line between these two events of dehydration adaptation. For example, accumulation of compatible solutes such as proline plays an important role in dehydration avoidance by maintenance of RWC. Alternately accumulation of proline has been proposed to play a role in dehydration tolerance by protecting protein and membrane structure, regulating redox status, or acting as a scavenger of ROS (59 -61). It should also not be assumed that dehydration avoidance and/or tolerance occurs in a linear progression in time after dehydration sets in. In the study of plant responses to dehydration, the important consideration would be the method used to impose the dehydration, the severity and duration of the dehydration, the physiological status to be assessed, and how the observed responses of the plant fit into an overall strategy for adaptation to dehydration. Although major attention has been focused on plant responses to dehydration at the cellular level, an integrated approach comprising physiological and molecular biology techniques that may improve the basic understanding of complex genotypes versus dehydration is rare. This study was aimed at understanding cell physiology under dehydration and its effects in terms of changes in organelle-specific protein expressions in chickpea. The results obtained have been illustrated as a representative model to summarize the events modulating dehydration tolerance (Fig. 8). The perception of dehydration condition is presumably regulated by water status of the plant. The fall in RWC in the first 24 h of dehydration seemingly resulted in induction of signaling proteins. Probable osmotic adjustment was seen as a result of compatible solute accumulation in the next 48 h. Plants undergo cell wall modifications to minimize water loss and avoid dehydration condition. Cell defense pathway-related proteins are induced in later stages that perhaps aid the survival of plants under severe dehydration. Also probable linkers between the cytoskeleton and ECM may relay biochemical/mechanical signals induced by dehydration to the interior of the cell. These findings would bridge the gap between the events of ecophysiological experiments and comparative proteomics analysis in elucidating the molecular mechanisms by which plants sense and respond to dehydration stress. Comparative proteomics analysis is a promising tool for determination of components involved in dehydration tolerance, although the number of proteins that can be analyzed by 2-DE is still limited with respect to the predicted number of proteins. A considerable number of candidates (21%) were found to be unknown or hypothetical proteins. As many as 36 DRPs showed no significant match in the existing database. The results are particularly important, at least in part, due to the fact that the genome sequence of chickpea has not been determined yet, and the number of fully annotated expressed sequence tags is very low. Our future efforts will focus on identifying the dynamics associated with the ECM-resident proteins involved in the dehydration response and determining their functions, which would provide the basis for effective engineering strategies for a crop improvement program. * This work was supported by grants from the Department of Biotechnology, Ministry of Science and Technology (to N. C.) and by predoctoral fellowships (to D. B., A. P., and M. K. C.) from the Council of Scientific and Industrial Research, Government of India. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. □ S The on-line version of this article (available at http://www. mcponline.org) contains supplemental material.