JXB Advance Access originally published online on April 18, 2007
Journal of Experimental Botany 2007 58(7):1873-1892; doi:10.1093/jxb/erm012
RESEARCH PAPER |
Proteomic analysis reveals differences between Vitis vinifera L. cv. Chardonnay and cv. Cabernet Sauvignon and their responses to water deficit and salinity
1Department of Biochemistry and Molecular Biology, MS 200, University of Nevada, Reno, NV 89557, USA
2Institute of Biotechnology, University of Ankara, 06500 Besevler-Ankara, Turkey
3UMR de Genetique Vegetale, INRA/CNRS, F-91190 Gif-sur-Yvette, France
4Department of Genetics and Genomics, Boston University School of Medicine, Boston, MA 02118, USA
* To whom correspondence should be addressed. E-mail: cramer{at}unr.edu
Received 29 August 2006; Revised 12 December 2006 Accepted 9 January 2007
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussionConclusionsSupplementary dataReferences |
|---|
The impact of water deficit and salt stress on two important wine grape cultivars, Chardonnay and Cabernet Sauvignon, was investigated. Plants were exposed to increasing salinity and water deficit stress over a 16 d time period. Measurements of stem water potentials, and shoot and leaf lengths indicated that Chardonnay was more tolerant to these stresses than Cabernet Sauvignon. Shoot tips were harvested every 8 d for proteomic analysis using a trichloroacetic acid/acetone extraction protocol and two-dimensional gel electrophoresis. Proteins were stained with Coomassie Brilliant Blue, quantified, and then 191 unique proteins were identified using matrix-assisted laser desorption ionization time of flight/time of flight mass spectrometry. Peptide sequences were matched against both the NCBI nr and TIGR Vitis expressed sequence tag (EST) databases that had been implemented with all public Vitis sequences. Approximately 44% of the protein isoforms could be identified. Analysis of variance indicated that varietal difference was the main source of protein expression variation (40%). In stressed plants, reduction of the amount of proteins involved with photosynthesis, protein synthesis, and protein destination was correlated with the inhibition of shoot elongation. Many of the proteins up-regulated in Chardonnay were of unclassified or of unknown function, whereas proteins specifically up-regulated in Cabernet Sauvignon were involved in protein metabolism.
Key words: 2-DE, grapevine, MALDI-TOF/TOF, salinity, shoot, water deficit
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionConclusionsSupplementary dataReferences |
|---|
Grapevine (Vitis vinifera L.) is a perennial woody vine that produces the most economically important fruit crop in the world. Relative to other crop plants, grapevines are fairly tolerant to water deficit, but are susceptible to significant damage from long-term salinity. With the shortage of freshwater resources and the increase of saline soils in irrigated regions around the planet, it is important to develop crops that are able to sustain productivity under such challenging conditions.
Many recent proteomic studies have been performed on various species and under different abiotic stresses, such as water deficit (Salekdeh et al., 2002; Hajheidari et al., 2005; Vincent et al., 2005; Ali and Komatsu, 2006; Jorge et al., 2006), high salinity (Abbasi and Komatsu, 2004; Lee et al., 2004; Kim et al., 2005; Yan et al., 2005; Askari et al., 2006; Parker et al., 2006), low temperature (Tafforeau et al., 2002; Bae et al., 2003; Imin et al., 2004; Renaut et al., 2004; Cui et al., 2005; Amme et al., 2006), heat (Majoul et al., 2003,2004), high light (Nam et al., 2003; Phee et al., 2004), ozone (Agrawal et al., 2002), and metal toxicity (Requejo and Tena, 2005; Sarry et al., 2006). To our knowledge, there are only five proteomic studies using grapevine (Tesnières and Robin, 1992; Sarry et al., 2004; Carvalho et al., 2005; Castro et al., 2005; Vincent et al., 2006). Although powerful mass spectrometry-based techniques have emerged recently, such as isotope-coded affinity tagging (ICAT; Gygi et al., 1999), mass-coded abundance tagging strategies (MCAT; Cagney and Emili, 2002), stable isotope labelling by amino acids in cell culture (SILAC; Ong et al., 2002), or isobaric tags for relative and absolute quantitation (iTRAQ; Ross et al., 2004), as well as direct isotopic labelling methods using 18O (Rao et al., 2005) or D2O (Che and Fricker, 2005), two-dimensional electrophoresis (2-DE) remains one of the most efficient strategies to isolate proteins showing quantitative variation in response to a treatment (Rabilloud, 2002; Görg et al., 2004). The reproducibility of 2-DE and the accuracy of sample comparison were recently improved with the introduction of the difference gel electrophoresis (DIGE; Unlu et al., 1997) technique.
Varietal differences have been exploited for centuries in agriculture. Today, many breeding strategies use genetic variations between cultivars to improve the performance of an elite variety. In grapevine in particular, and in fruit crops in general, the choice of a variety is of critical importance because such differences often drive consumer preference for a particular table grape or wine. An intense artificial selection and a vegetative mode of reproduction appear to account for the genetic structure of cultivated V. vinifera varieties (Aradhya et al., 2003; This et al., 2006). Improved disease resistance, abiotic stress tolerance, or quality factors are important targets for genetic improvement of grapevine cultivars (Vivier and Pretorius, 2002; Sarry et al., 2004). Several groups have investigated abiotic stress effects in grapevine (Benson and Roubelakis-Angelakis, 1994; Flexas et al., 1999; Papadakis and Roubelakis-Angelakis, 1999; Gaudillère et al., 2002; Lesniewska et al., 2004) or have studied grapevine shoot development (Castelan-Estrada et al., 2002; Lovisolo et al., 2002; Lebon et al., 2004); yet, none of them used any of the ‘omics’ technologies available today. Here, a proteomic study of the response to abiotic stresses of two of the most prominent V. vinifera cultivars, Chardonnay and Cabernet Sauvignon, is presented. Both varieties display physiological differences when exposed to water deficit and salinity, with Chardonnay appearing to be more tolerant to water deficit and salinity. Quantitative analysis of two-dimensional gels from shoots showed that varietal differences mainly account for changes in protein expression patterns. Some of the stress- and cultivar-responsive proteins are discussed in relation to the physiological adaptations of Chardonnay and Cabernet Sauvignon.
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionConclusionsSupplementary dataReferences |
|---|
Plant material and experimental conditions
Two-year-old rooted cuttings of dormant V. vinifera L. cv. Chardonnay (clone #6) and cv. Cabernet Sauvignon (clone #8) were obtained from Inland Desert Nursery (Benton City, WA, USA) and kept at 4 °C under low light in a cold room until potted in soil (Scotts MetroMix 200 medium). Cuttings were grown in a greenhouse in individual pots under controlled temperature and supplemental sodium halide light (day/night=25/17 °C; 16 h/8 h at a minimum PAR of 200 µmol m–2 s–1). Plants were irrigated every third day with tap water. Every 30 d, the plants were irrigated with 0.5 l of a nutrient solution as described in Gibeaut et al. (1997). Upon producing an unbranched solitary shoot of 80–100 cm, the shoot was pruned to 60 cm as described in Cramer et al. (2005). Lateral shoots, buds, and young flowers were removed, leaving only winter buds and leaves. Plants were fertilized with 1.0 l of Gibeaut's solution every 2 d, a total of five times, to promote growth of a new shoot. The experiment began when new shoots grew and gave rise to a minimum of two new nodes (young leaves). Fertilization was not maintained over the course of the experiment.
The experiment consisted of a completely randomized factorial design consisting of two cultivars (Chardonnay and Cabernet Sauvignon)x3 treatments (control, water deficit, and salinity)x3 times (day 0, day 8, and day 16), with 3–6 biological replicates. Three biological replicates were used for the proteomic analyses. On day 0, only three control plants were measured (no stress treatments as they were not initiated yet), bringing the total number of plants to analyse to 42 (for each cultivar: three controls on day 0, three controls+three water-stressed+three salt-stressed on day 8, and three controls+three water-stressed+three salt-stressed on day 16). The subsequent effects of water deficit and salinity on the growth, water potentials, and proteomes of Chardonnay and Cabernet Sauvignon over the course of 16 d were studied (October 2003). For water deficit treatments, watering was completely stopped at the onset of the experiment (day 0). Control plants and salt-stressed plants were watered daily with 4.0 l of tap water or salt solution throughout the experiment. Stem water potentials of the plants were measured, and a calculated concentration of salt solution was added to the salinity-treated plants so that the water potentials of these plants matched those of the water deficit-treated plants. Consequently, salt concentration was increased over time as follows. On day 1, salt-treated vines were given 10 mM NaCl and 1 mM CaCl2. From day 2 to day 6, they received 20 mM NaCl and 2 mM CaCl2 daily. During days 7 and 8, 55 mM NaCl and 5.5 mM CaCl2 were supplied to the plants. From day 9 to day 12, 75 mM NaCl and 7.5 mM CaCl2 were added daily to the watering solution. During days 13 and 14, the vines received 175 mM NaCl and 17.5 mM CaCl2, and 250 mM NaCl and 25 mM CaCl2 on days 15 and 16.
Shoot tips (cut below the fourth leaf, thus including the apex, stem, four leaves, and tendrils) were harvested on days 0, 8, and 16. Two shoots were harvested per plant, pooled in a 50 ml Falcon tube, instantly frozen in liquid nitrogen, and stored at –80 °C until further use.
Physiological measurements
Stem water potential
Fully mature leaves from between the second and sixth oldest nodes of the main shoot were selected for stem water potential measurement as described in McCutchan and Shackel (1992). Each measurement was performed using a pressure chamber (3005 Plant Water Status Console, Soilmoisture Equipment Corp., Santa Barbara, CA, USA). Xylem sap was also collected to measure the solute potential of the sap with a vapour pressure osmometer (Model 5500, Wescor Inc., Logan, UT, USA). Solute potential was calculated as in Cramer and Bowman (1991), and the water potential of the xylem was corrected for variation in solute accumulation in the salt-treated plants (data not shown). The water potential was recorded every 4 d, using different plants in each measure.
Growth measures
Shoot length (mm) and leaf length (mm) were measured every 4 d. A leaf was not considered apparent until it had reached at least 10 mm in length; otherwise, it was assumed to be indistinguishable from the shoot apex. Reference starting points along the shoots for growth measurements were marked with tape, resulting in all shoot tips bearing exactly two leaves from the reference point on day 0.
Protein analysis
Protein extraction
Grapevine shoots were finely ground in liquid nitrogen, then the proteins were extracted following the procedure described by Damerval et al. (1986). Proteins were resuspended in R2D2 buffer developed by Méchin et al. (2003). For each sample, 90 µl of resolubilization solution was used to resuspend 1 mg of pellet. Once resolubilized, the protein extract was stored at –80 °C until use. Protein concentration was determined using an EZQ Protein Quantitation Kit (Invitrogen, Carlsbad, CA, USA) with ovalbumin as standard, according to the manufacturer's instructions.
Two-dimensional PAGE
Before use, the protein extracts were allowed to thaw at room temperature; they were then centrifuged at 14 000 g for 15 min at 24 °C. A volume of 185 µl (
0.2 mg of protein) of supernatant was loaded onto immobilized pH gradient (IPG) strips (ReadyStrip 110 mm, pH 4–7, Bio-Rad, Hercules, CA, USA). Isoelectric focusing (IEF) was performed using a Protean IEF Cell (Bio-Rad, Hercules, CA, USA) system. Strips were passively rehydrated with the protein extracts for 1 h. Wicks were then added to the electrodes and strips were covered with mineral oil and further rehydrated for 12 h at 50 V. IEF conditions were: 50–250 V linear gradient for 20 min, 250–4000 V linear gradient for 2 h, and 4000–8000 V rapid gradient until a total of 18 000 Vh was reached. The strips were then equilibrated twice for 15 min each as described in Görg et al. (1987). The second dimension was run on 12.5% Criterion TRIS-HCl resolving gels (Bio-Rad, Hercules, CA, USA) for 1 h at 200 V in a Criterion Dodeca cell. Two-dimensional gels were stained using a colloidal Coomassie brilliant blue (CBB) G-250 procedure, according to Neuhoff et al. (1988).
Two-dimensional image processing and statistical analyses
Two-dimensional gels were digitized with a Versadoc Imaging system and processed using PDQuest software V.7.3 (Bio-Rad, Hercules, CA, USA). Spot quantity normalization occurred throughout the whole matchset, which included all 42 two-dimensional gels. The gel used as reference in the PDQuest matchset corresponded to a co-migration of protein extracts from both Chardonnay and Cabernet Sauvignon control plants, and was subsequently used to excise spots for protein identification (see below).
This study was based on an unbalanced experimental design in which the first time point (day 0) was represented only by control plants. All three treatments (control, water deficit, and salinity) were measured during the successive temporal stages (days 8 and 16). A balanced experiment was obtained by converting protein levels of days 8 and 16 into percentages with respect to the protein levels of the controls measured on day 0 for both varieties (Chardonnay and Cabernet Sauvignon). Controls on day 0 were then removed from the data set. Therefore, spot quantities for each stress treatment on days 8 and 16 were expressed relative to their controls on day 0. It was verified that the variation due to the technique (using three technical replicates, SE=0.05) was insignificant compared with the variation resulting from the experimental conditions (three biological replicates; SE=0.18). Spots were considered reproducible when present in at least two of the three biological replicates. Statistical analyses were performed on those 758 spots deemed reproducible.
A full factorial analysis of variance (ANOVA) was performed on each protein to determine whether varietal effects across time were of significance, using the statistical programming language R. The following linear model was used for the ANOVA:
 | (1) |
i2, 1 j2, and 1 k3. The terms Vi, Dj, and Tk measure the effect of the variety, day, and treatment, respectively. The interaction terms (VD)ij, (VT)ik, and (DT)kj account for the interaction between variety and day, variety and treatment, and day and treatment, respectively. Similarly, the term (VDT)ijk represents the interaction of variety, day, and treatment. An ANOVA was performed on each protein using the linear model above. An adjustment for the false discovery rate (Benjamini and Hochberg, 1995) was made on the P-values of the individual effects. Any protein with an adjusted P-value of <0.05 associated with a specific effect was deemed to have a significant change attributed to the specified effect. Additionally, a forward stepwise regression was performed to determine which proteins could be used as predictors of shoot length and water potential, using JMP software (SAS Institute, Cary, NC, USA).
Protein identification
To the best of our knowledge, no proteomic map of grapevine shoot exists. It was decided to establish a working map by identifying the proteins which were expressed abundantly at the onset of treatment and then to follow their expression over time in relation to water or salt stresses in both cultivars. A total of 202 spots were excised from a gel in which two samples had co-migrated (equal proportions of protein samples obtained from Chardonnay and Cabernet Sauvignon both grown under control conditions and collected on day 0) under the electrophoretic conditions described above. To increase the chance of protein identification, the 186 most abundant spots were excised from the reference gel. An additional 16 spots proven significant in the two stepwise regression analyses were subsequently excised. Excised spots were trypsin digested and analysed using matrix-assisted laser desorption ionization time of flight/time of flight (MALDI-TOF/TOF) mass spectrometry (MS) as described in Vincent et al. (2006). The peptide lists generated by the MASCOTTM software were searched against both the non-redundant (nr) protein database of NCBI and the TIGR Vitis expressed sequence tag (EST) database.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionConclusionsSupplementary dataReferences |
|---|
Physiological responses of grapevines subjected to water and salt stresses
Stem water potential
The stem water potential of control plants ranged between –0.16 MPa and –0.37 MPa over the course of the experiment (Fig. 1). No significant differences were detected between cultivars at any time or between control and stressed grapevines until after day 4. On day 8, both water deficit-treated and salt-stressed plants had lower stem water potentials than the controls. On days 12 and 16, the stem water potential of the water deficit-treated grapevines decreased more rapidly than that of NaCl-stressed vines and reached a maximum of –1.0 MPa to –1.2 MPa depending on the cultivar. The stem water potential of the salt-treated plants did not decrease to the same extent and reached a maximum of –0.85 MPa in both cultivars. The general trend after day 8 was an intermediate water status of the salt-stressed plants between that of the controls and water deficit-treated plants.
|
Shoot and leaf length
The shoot length response to water deficit and salinity treatment differed between the two cultivars (Fig. 2A). Under optimal conditions, Cabernet Sauvignon shoots grew slightly faster than Chardonnay, but this difference was not statistically significant. With the salt treatment, the shoot growth of Chardonnay was not affected relative to controls, whereas Cabernet Sauvignon shoot length was significantly reduced by day 12. With water deficit, Chardonnay shoot length was affected by day 8 relative to its control, whereas Cabernet Sauvignon shoot elongation decreased significantly as early as day 4.
|
When individual leaf lengths were measured, the responses to the treatments differed from one leaf to another, as well as between cultivars. Leaves were observed to grow faster in well-watered Cabernet Sauvignon grapevines than in Chardonnay, particularly at later developmental stages. Leaf 1 length appeared to be reduced only by water deficit on days 12 and 16 (Fig. 2B). The same observation held true for leaves 2 and 3, which showed a more acute reduction under water stress (Fig. 2C, D). Then the pattern changed for leaf 4 and the younger leaves. Salt stress reduced the length of leaf 4 in both cultivars (Fig. 2E), although to a lesser extent than water deficit. As previously observed in shoots, Cabernet Sauvignon seemed to be more affected by salt and water stresses than Chardonnay when the growth of leaves 4–8 was considered (Fig. 2F–I). In Cabernet Sauvignon plants, elongation of leaves 6 and 7 was stopped by water deficit and no new leaves were produced after leaf 7. The inhibition in Chardonnay occurred at later developmental stages than that in Cabernet Sauvignon; water deficit stopped the elongation of leaves 7 and 8 and no new leaves were produced after leaf 8.
Protein identified in grapevine shoots
Database search and cross-contamination
A consensus map of both varieties was generated from a reference gel (Fig. 3A). Out of 202 excised spots, no MS data were obtained for three spots (spot #47, spot #63, and spot #91). The remaining 199 spots were examined for spot cross-contamination (several proteins identified in a single spot) by observing the MS spectra. If cross-contamination was found, the quantitative data associated with the spot was discarded. A total of 191 proteins led to a unique hit (including proteins of unknown function). Protein identities are indicated in Table 1.
|
|
A number of protein identities had low protein scores, as expected, due to the incomplete sequencing effort of the grapevine genome (http://www.evry.inra.fr/public/projects/genome/grape_fr.html), and nucleotide sequence data are available only from a limited number of EST sequences in Vitis. The EST libraries do not represent all genes of a given organism, and the gene sequences, whose quality varies, are only partial.
On occasion, the MS analysis of one particular spot led to multiple protein identifications (eight spots out of 199, 4.02%), indicating that several proteins had co-migrated (otherwise known as cross-contamination). Spot density was high because of the large amount of proteins loaded on small gels (11 cm), although the 2-D gel resolution was improved by using a narrow pH gradient (4–7). In general, spot resolution was of high quality on the 2-D gels; however, some overlaps were detectable. The phenomenon was more dramatic when different proteins co-migrated at precisely the same location on the 2-D gel. Technical improvements in the sensitivity and accuracy of MS analyses allowed the detection and identification of proteins co-localizing in the same spot on 2-D gels. This phenomenon is not frequently discussed in the literature. Méchin et al. (2004) reported that almost 10% (60 spots out of 632 excised spots) of the spots contained cross-contamination in maize seeds. Parker et al. (1998) proposed a method to detect cross-contaminations. Cross-contamination can be partially overcome by reducing the complexity of the protein extract by performing some pre-fractionation methods, or increasing the 2-D gel resolution by using both a narrower pH gradient in the first dimension and a wider percentage polyacrylamide gradient of the SDS–polyacrylamide gel in the second dimension. In this study, analysis of MS/MS spectra revealed such co-migration, and the protein quantity for these particular spots was considered inaccurate, and therefore disregarded (eight spots). Of the 191 remaining spots, 142 (74%) led to a known protein annotation.
Protein isoforms
Protein identification revealed several protein isoforms (Table 1). Out of 191 identified spots, 83 bore the same annotation as another spot (43.5%, see inset in Fig. 3B). In some cases, such as for ribulose 1,5-bisphosphate carboxylase-oxygenase (RuBisCO) large subunit (hereafter referred to as RBCL; spot #11, spot #65, spot #88, spot #90, and spot #173), up to five isoforms of the same protein were isolated. One of the isoforms (spot #11) exhibited a molecular mass (Mr) of 20 000 and an isoelectric point (pI) of 5.8, whereas the four others presented a lower Mr (15 000) and were distributed along a pI from 4.6 to 5.4. According to the theoretical Mr and pI provided by the ExPASy website (http//:www.expasy.org/tools/pi_tool.html), RBCL should bear an Mr of 51 000 and a pI of 7.5, which is beyond the pI range of the gels used in this study. Therefore, those spots probably corresponded to degraded forms of RBCL. Among the 202 excised spots, none corresponded to undegraded forms of RBCL. Because the protein map did not cover all the spots resolved under the electrophoretic conditions (Fig. 3A), it is possible that they were missed. Alternatively, the undegraded form of RBCL could have been expressed in such low levels in grape shoots that its corresponding spot on 2-D patterns was neither excised nor stained. Perhaps the totality of RBCL molecules present in the shoot samples were completely degraded due to the growing conditions (artificial light of the greenhouse). Because most of the identified proteins resolved at their expected pI and Mr, protein degradation related to technical handling tended to be ruled out. A more likely explanation is to be found in the nature of the plant material being dealt with. Grapevine shoots are mainly composed of stem tissues, whose main function is not photosynthesis, hence a low RuBisCo content. The expression profile varied from one RBCL isoform to another, with differences due to the cultivars, treatments, or developmental stages (see Supplementary Fig. S4A–E at JXB online). Houtza and Portis (2003) have reported that a number of co- and post-translational modifications can affect RBCL. RBCL can also be subject to proteolysis (Alburquerque et al., 2001), which is enhanced under oxidative conditions (Penarrubia and Moreno, 1990). Degradation products of RBCL, though of greater Mr than in the present study, were also found in leaves of micropropagated grapevine plantlets (Hajduch et al., 2001; Carvalho et al., 2005).
Other proteins such as heat shock protein 70 kDa (HSP70, Mr 70 000, spot #13 and spot #114), chaperonin 60 kDa (CPN60a, Mr 60 000, spot #6), translationally controlled tumour protein homologue (TCTP, Mr 19 000, spot #21 and spot #23), actin (ACT, Mr 42 000, spot #10 and spot #186), and ascorbate peroxidase (APX, Mr 28 000, spot #34 and spot #161) were located at the correct Mr. Shifts of protein pI and/or Mr of some isoforms are indicated in Fig. 3B. A tool calculating those shifts on 2-D gels has been devised (Halligan et al., 2004). Examples of protein isoform expression patterns are presented in Supplementary Figs S2–S6 at JXB online.
Functional classification
Figure 4 presents the functional classification of the proteins according to the annotation in the MIPS Arabidopsis thaliana database (MAtDB http://mips.gsf.de/proj/thal/db; Schoof et al., 2002). The highest percentage (27%) corresponded to proteins of unknown function. Up to 10% of the identified proteins were involved in photosynthesis; this was expected, as the shoot tip, a photosynthetically active part of the grapevine, was collected. Unclassified proteins (9%) correspond to gene products of known function, but acting in various pathways, hence belonging to multiple categories. Protein metabolism (2% transcription, 12% protein synthesis, 8% protein destination, and 3% protein degradation) represented 25% of the identified proteins, which is consistent with the fact that the shoot tip is an actively growing organ where both cell production and elongation take place. Interestingly, in the control samples, 7% of the proteins identified have been previously reported to be involved in stress, indicating possible roles for these proteins in general metabolism in addition to their role in plant adaptation to severe conditions.
|
Statistical analyses using reproducible spots
Stepwise regression and correlation analyses
The present study demonstrated that abiotic stresses affected Chardonnay and Cabernet Sauvignon at both physiological and protein levels over time. It was hypothesized that some of the changes in protein expression levels could be due to the physiological adjustments occurring in response to the abiotic stresses. A forward stepwise regression model was used to identify which, if any, protein expression levels (independent variable) were best correlated with each of the physiological parameters (dependent variables) of water potential or shoot length. This analysis was performed with all 758 quantified proteins. Shoot length was completely predicted (R2=0.998) by eight proteins (Table 2A), and water potential was completely predicted (R2=1.0) by eight proteins (Table 2B).
|
Pairwise correlations were computed across the expression profiles of these eight proteins, and none was found to be above 0.5, with P-values below 0.05 (data not shown). The eight proteins retained by each model were excised, and identified using MS analysis. No obvious molecular pattern could be found in the model describing water potential (Table 2A), whereas most proteins were involved in protein metabolism (either synthesis or destination) in the model fitting shoot length measurements (Table 2B).
ANOVA
Table 3 shows the number of spots with statistically significant main effects and/or interaction effects associated with the model in equation (1). Protein expression was mainly affected by the varietal (V) effect (305 spots out of 758, 40%); among those, 164 spots were up-regulated in Chardonnay, 137 spots were up-regulated in Cabernet Sauvignon, and four spots showed no variation between both cultivars. The day (D) and treatment (T) effects affected protein expression levels to a lesser extent (199 and 115 spots, respectively). Supplementaey Tables S1 and S2 (available at JXB online) list the identified proteins showing significant variation for the T and V effects, respectively. Interestingly, 120 spots showed significant variation when the dayxtreatment (DT) interaction was considered. This suggested that the grapevine response to the treatment also depended on the developmental stage of the plant or the duration of the stress. The varietyxday (VD) interaction was shown to be a significant effect for fewer proteins (89 spots), indicating that the cultivars were not subjected to as much variation as the other factors over time at the protein level. Similarly, 79 proteins showed a significant varietyxtreatment (VT) interaction, thus indicating that a number of proteins were cultivar specific in their response to abiotic stresses. Differences in salt and water deficit tolerance might reside within this set of proteins.
|
Most of the proteins showed significant effects for several main factors and/or interactions, potentially highlighting the complex regulatory network of grapevine adaptation to the abiotic stresses at the protein level. For instance, more than a third of the spots with a significant V effect (305 spots) showed evidence of a significant D effect (106 spots). Similarly, a quarter of the proteins (81 spots) had a significant T effect (data not shown). For those spots with significant effect(s), the functional classification of the identified proteins has been represented according to the MAtDB molecular function annotation (see Supplementary Fig. S7 at JXB online).
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionConclusionsSupplementary dataReferences |
|---|
Chardonnay appears to be more tolerant than Cabernet Sauvignon to water deficit and salinity at the physiological level
In this study, grapevine growth was significantly affected by the abiotic stresses. The stem water potential of control vines remained steady over time, whereas water deficit-treated and salt-stressed plants had significantly lower stem water potentials on day 8. The decrease of stem water potential in water deficit-treated plants was approximately twice that of salt-stressed plants on days 12 and 16.
At the molecular level, a pathogenesis-related protein characterized in tobacco, NtPRp27 (Okushima et al., 2000), accounted for 82% of the stepwise regression model for water potential. This protein was up-regulated at the last developmental stage under the stress conditions. In addition to the pathogen and wound responses in tobacco, NtPRp27 transcripts were also observed to accumulate following water deficit stress or abscisic acid (ABA) treatment, but not under salinity conditions (Okushima et al., 2000). The greater effect of water stress on NtPRp27 abundance compared with salt treatment reflects what was observed in grapevine shoots.
While stress levels were not different between cultivars, as indicated by the equal water potentials for the different stress treatments, shoot length measurements showed a clear distinction between cultivars. Chardonnay shoot elongation was slowed down by water deficit, but not by salinity, whereas Cabernet Sauvignon shoot length was significantly reduced by both salinity and water deficit, as compared with control. When the leaves along the shoot were considered, leaves that were developed before the treatment onset (leaves 1–3) seemed to cope better with the stresses than the younger ones, indicating that cell production and elongation may have been affected in the young leaves (leaves 4–8). As observed for shoot growth, salinity had almost no effect on leaf growth of Chardonnay. With water deficit, leaf length was altered to a lesser extent in Chardonnay than in Cabernet Sauvignon, even though water potentials were lower in Chardonnay than in Cabernet Sauvignon. Therefore, Chardonnay appears to be more resistant to both water deficit and salinity than Cabernet Sauvignon.
Stepwise regression indicated that the inhibition of shoot growth was best correlated with several proteins involved either in protein synthesis, 60S acidic ribosomal P0 and P1 proteins, or in protein destination, two 70 kDa heat shock proteins (HSP70) and a 60 kDa LS2 chaperonin (CPN60). Because of their well-known role as molecular chaperones, HSP70 and CPN60 were deliberately classified in the protein destination category. However, many studies reported their involvement in heat stress initially, and various biotic or abiotic stresses later on (Wang et al., 2004). Because the chaperones and ribosomal proteins included in the stepwise regression model accumulated under the present stress conditions, it can be hypothesized that they played an active role in maintaining protein metabolism in grapevine shoots.
It is a common observation that wine quality benefits from mild drought periods, thus improving berry composition without any significant yield loss. Therefore, monitoring the water status of grapevine plants, by following the carbon isotope composition (Gaudillère et al., 2002) or by using thermal and visible imagery (Jones et al., 2002; Grant et al., 2006; Möller et al., 2007), should favour grape cultivation for wine production. Drought effects are not limited to berry; vegetative organs are affected as well. Stoll et al. (2000) showed that drying roots of grapevine plants produced chemical signals, such as ABA concentration in xylem sap (Correia et al., 1995), which were transported to the leaves, thus decreasing stomatal conductance. Drought caused the reduction of CO2 assimilation and the enhancement of O2 uptake, due to stomatal closure, along with a decrease in the activity of various enzymes of the Calvin cycle (Flexas et al., 1999; Maroco et al., 2002). Grapevine shoot organogenesis was shown to be under the influence of soil water deficit, and not related to competition for assimilates between axes of various sink strength (Lebon et al., 2006); however, unlike what is presented in the study reported here, no varietal difference was reported in their experiments.
Protein identification of grapevine shoots reveals many proteins of unknown function, and a significant proportion of isoforms
In the present study, the largest category (27%) of the identified proteins corresponded to genes of unknown function, with a rate comparable with the one reported for genome annotation. In the fully sequenced genome of A. thaliana (Arabidopsis Genome Initiative, 2000) and draft sequence of the rice genome (Goff et al., 2002), 30% of the genes were of unknown function. Koller et al. (2002) found that 30% of proteins had no known function or poor homology in rice seeds. Among grapevine ESTs, 19% of the genes did not match any annotation (Goes da Silva et al., 2005), which was lower than the proportion of unknown proteins. The functional categorization drawn from a grape berry 2-D map (Sarry et al., 2004) differed significantly from the present classification in grape shoots. Among other things, the proportion of proteins participating in protein metabolism was small in grape berries (13%) compared with grape shoot (25%), whereas the proportion of proteins involved in stress was greater in berries (19%) than in shoots (7%). These differences may be accounted for by the different organ sources (berry mesocarp versus shoot) of the tissue analysed as well as by grapevine cultivar difference (Gamay Noir versus Chardonnay/Cabernet Sauvignon).
Up to 43.5% of the excised proteins corresponded to isoforms in this study. This is a common observation reported in proteomic studies on different species such as alfalfa (Valot et al., 2004; Watson et al., 2004), ginseng (SI Kim et al., 2003; Nam et al., 2003), barley (Finnie et al., 2004; Ostergaard et al., 2004), and rice (ST Kim et al., 2003). In maize seeds, isoforms represented 71% of the identified proteins, with one protein being present as 26 isoforms (Méchin et al., 2004). In the present study, it was possible to identify several isoforms without trying to characterize them further. Expression profiles differed from one isoform to another, potentially giving some hints about their function (see Supplementary Figs S1–S6 at JXB online). The presence of isoforms raises the important question of protein redundancy and activity. Those diverse forms of protein could have different origins. They could be various gene products encoded by different genes but having the same function. They could be expressed in various organs or tissues, or act in different cell compartments. For instance, three ascorbate peroxidases were identified; two of them were cytosolic (spot #34 and spot #161) whereas the other one was chloroplastic (spot #197). Isoforms could also arise from alternative splicing, encoded by the same gene but yielding different mRNAs, hence different proteins. Alternatively, they could be explained by post-translational modification (PTM) wherein the same protein adopts several forms by adding non-protein components. PTMs affect many proteins and can result from phosphorylation, glycosylation, sulphation, etc., and are preferentially associated with certain amino acids (Huber and Hardin, 2004). Though 2-DE cannot indicate whether those isoforms are encoded by different copies of the same or related genes or whether they correspond to different forms of the same gene product, such a technique gives the opportunity to isolate post-translationally modified proteins. Other tools, such as affinity chromatography, are needed to characterize them further.
Few proteins respond to abiotic stresses
ANOVAs indicated that only 15% (115 out of 758) of the spots showed significant variation in response to the water deficit and salinity stress treatments; among them, 36 spots were identified and their annotations suggest they belong to various functional categories (see Supplementary Table S1 at JXB online). A few identified proteins, displaying apparent overexpression under the present stress conditions, were of particular interest. A nuclear matrix constituent protein 1-related (NMCP1, spot #82) was significantly up-regulated by both water deficit and salinity stresses at the oldest developmental stage (day 16), regardless of the variety; however, protein levels reached in salt-stressed Cabernet Sauvignon vines on day 16 were equivalent to those expressed on day 0 (see Supplementary Fig. S8A at JXB online). This protein was found to be exclusively localized at the nucleus periphery of carrot cells (Masuda et al., 1997) and seemed to be involved in restructuring of the nuclear envelope during mitosis (Masuda et al., 1999). To our knowledge, no study has mentioned a potential role of NMCP1 in abiotic stress. Another novel result was the overexpression of a ribosomal protein L39 (RPL39; spot #175), under both stresses, but particularly under water deficit, on the last developmental stage analysed (see Supplementary Fig. S8B at JXB online). As far as is known, this is the first time that an increase in the amount of RPL39 under abiotic stress has been reported in plants. In yeast, this protein is a 60S ribosomal subunit implicated in translational accuracy (Dresios et al., 2000). Both NMCP1 and RPL39 bore fairly low MASCOT protein scores (39 and 41, respectively; Table 1), therefore one must remain cautious about the tentative annotations, and the involvement of such proteins in these processes remains to be evidenced.
Three spots identified as photosynthetic enzymes, two RBCL isoforms and a chloroplastic sedoheptulose-1,5-bisphosphatase (SBPase), were up-regulated under both abiotic stress conditions, water deficit in particular (spot #88 Supplementary Fig. S1A, spot #65 Supplementary Fig. S1D, and spot #72 Supplementary Fig. S8C at JXB online). Interestingly, a spot loosely annotated photosystem II (PSII, spot#145, Supplementary Fig. S8D at JXB online) was down-regulated by both water and salt stresses. However, it should be noted that the protein score attached to the MS identification of this spot was poor (protein score=45, Table 1); thus, its implication in photosynthesis needs to be confirmed. Other RBCL isoforms (Supplementary Fig. S1B, C, E at JXB online) or proteins involved in photosynthesis, such as RuBisCO activase, Mg-chelatase, chloroplastic phosphoribulokinase, glyceraldehyde-3-phosphate dehydrogenase, chloroplastic triose phosphate isomerase, and oxygen-evolving protein of photosystem II, were not affected by abiotic stresses, highlighting the diversity of response to stress within a gene family or a particular biochemical pathway. An increased expression of four isoforms of RBCL was observed under drought stress in sugar beet leaves (Hajheidari et al., 2005). Both small and large subunits of RuBisCO were down-regulated under ozone stress in rice leaf (Agrawal et al., 2002). Maroco et al. (2002) reported a reduced activity of several enzymes of the Calvin cycle in drought-stressed grapevine. Photosynthesis dysfunction caused by stress can be associated with oxidative stress (Mittler et al., 2004). While it appears that this is not the case in the present study because none of the antioxidant enzymes identified (ascorbate peroxidase, superoxide dismutase, or peroxiredoxin) appeared to be up-regulated by stress, it is possible that other isozymes, which were not detected here, were affected or that substrate levels rather than protein amounts were responsible for oxidative stress. Interestingly, protein response to abiotic stress is generally greater on the last developmental stage of grapevine shoots, reflecting what was observed at the physiological level.
Protein expression is mainly affected by cultivar differences
ANOVA indicated that protein quantity was most strongly affected by cultivar, then time, and finally treatment. Potentially, proteins being expressed at higher levels in one cultivar compared with the other could account for the differences observed at the morphological and physiological levels. Some functional categories were specific to a particular variety. Proteins involved in protein destination and degradation are specifically expressed in greater amounts in Cabernet Sauvignon (see Supplementary Table S2, as well as inset in Supplementary Fig. S7 at JXB online). A chloroplastic peptidyl-prolyl cis-trans isomerase (PPIase), involved in protein folding, was clearly more abundant in Cabernet Sauvignon, relative to Chardonnay (spot #101, Supplementary Fig. S9A at JXB online). Two other proteins also participating in protein degradation, a proteasome subunit alpha type 4 (PSA4, spot #51, Supplementary Fig. S9B at JXB online) and a ubiquitin-conjugating enzyme family member (CE, spot #92, Supplementary Fig. S9C at JXB online), further exemplified this phenomenon. The apparent activation of the proteolysis machinery of Cabernet Sauvignon is consistent with the observation that this grapevine cultivar was less able to cope physiologically with water and salt stresses. A xyloglucan endotransglycosylase (XET), part of the XTH family of enzymes (Rose et al., 2002), a cell wall-modifying enzyme associated with cell elongation (Wu et al., 2005), was expressed at higher abundance during the initial developmental stage day 0 in Chardonnay and decreased over time regardless of the treatment, compared with Cabernet Sauvignon wherein XET amounts remained low except on day 8 (spot #101; see Supplementary Fig. S9D at JXB online). At the physiological level, shoot growth was better sustained under stress conditions in Chardonnay, compared with Cabernet Sauvignon. It can be hypothesized that the elevated amounts of XET observed in Chardonnay may participate in maintenance of shoot elongation in this species, relative to Cabernet Sauvignon.
In this study, 45% (15 out of 33 spots) of the excised proteins observed to be more abundant in Chardonnay shoots did not yield any annotation. One explanation for this low rate of identification is that these gene products were specific to Chardonnay and account, in part, for its enhanced physiological tolerance to abiotic stress. Additionally, some functional categories were common to both cultivars, but were preferentially expressed in one variety and not the other. Although we demonstrated that a significant proportion of protein isoforms were identified, no isoforms were shared between the two cultivars (Supplementary Table S2 at JXB online). For instance, two ATP synthase beta chain isoforms (ATPsb) were more abundant in Chardonnay whereas three RBCL isoforms were up-regulated in Cabernet Sauvignon.
Among the proteins displaying a significant effect for variety factor, 10% (79 out of 758) presented significant effects for varietyxtreatment (VT) interaction as well, of which nine were identified by MS (see Supplementary Table S3 and VT pie chart in Supplementary Fig. S7 at JXB online). These proteins responded to abiotic stresses, albeit displaying varietal differences (see Supplementary Fig. S10 at JXB online). Most of these proteins showed significant effects for treatment (T) factor as well (see Supplementary Table S3 at JXB online, spots #113, 78, 7, 149, and 38). They might explain the cultivar differences in stress adaptation. For instance, an NAC-related protein (NAC, spot #38) was significantly down-regulated under water stress and, to a greater extent, under salinity conditions in Cabernet Sauvignon, but showed no distinct responses to treatments in Chardonnay (see Supplementary Fig. S10H at JXB online). No function was reported for this NAC protein. Likewise, a mitochondrial peroxiredoxin (PRX, spot #119) was strongly induced by salt stress on day 8 in Cabernet Sauvignon, whereas it was almost not affected by any treatment in Chardonnay over time (see Supplementary Fig S10C at JXB online). PRXs belong to the antioxidant system, thus scavenging reactive oxygen species (Dietz et al., 2006). It can be hypothesized that the salt-treated shoots from Cabernet Sauvignon suffered from oxidative stress on day 8, hence presenting more damage at the physiological level; however, it must be emphasized that none of the other antioxidant enzymes were significantly up-regulated under the present stress conditions. Among these proteins showing a significant effect for VT interaction, proteins beside PRX have been reported to be involved in response to abiotic or biotic stresses. Pathogenesis-related proteins 10 (PR-10) have been associated with a variety of stress responses in plants; in ginseng, they displayed an RNase activity (Hoffmann-Sommergruber, 2002). In grapevine, Castro et al. (2005) observed the induction of several PR-10 isoforms upon herbicide applications. Under the present conditions, PR-10 (spot #113) was markedly up-regulated by salt treatment over time and water deficit stress on day 16 in Cabernet Sauvignon shoots, but not in Chardonnay (see Supplementary Fig. S10B at JXB online). Thus, this protein did not seem to participate in the relative stress tolerance displayed by Chardonnay. bHLH proteins correspond to a large family of basic helix–loop–helix transcription factor members which have been involved in plant adaptation to abiotic stresses, such as drought and wounding (Kim and Kim, 2006), salinity (Kiribuchi et al., 2005), and iron deficiency (Yuan et al., 2005; Ogo et al., 2006). In the present experiment, the up-regulation of bHLH protein (spot #7) under stress conditions on day 8 was followed by almost a repression on day 16 in Cabernet Sauvignon, while its abundance was increased under salt conditions in Chardonnay (see Supplementary Fig. S10E at JXB online), potentially explaining the sustained shoot elongation observed in this variety. Interestingly, another protein containing a bHLH domain, GHDEL61 protein (DEL, spot #78), also presented an increased abundance in salt-stressed shoots from Cabernet Sauvignon on day 8 (see Supplementary Fig. S10D at JXB online). The relevance of such an observation in grapevine adaptation to abiotic stress is yet to be elucidated. Another potential candidate in cultivar-dependent responses to abiotic stress might be found in a pentatricopeptide repeat-containing protein (PPR, spot #93), which displayed an up-regulation under salinity on days 8 and 16, as well as under water deficit on day 16 in Chardonnay relative to the controls; almost the opposite pattern could be noted in Cabernet Sauvignon (see Supplementary Fig. S10F at JXB online). PPR proteins contain a degenerate 35 amino acid repeat motif; they bind to RNA and affect the post-transcriptional regulation of chloroplastic or mitochondrial genes (Okuda et al., 2006; Schimtz-Linneweber et al., 2006). A total of 466 PPR proteins were described in Arabidopsis thaliana (Rivals et al., 2006). To our knowledge, this protein has never been reported to participate in the stress response in plants. It can be hypothesized that, when subjected to environmental constraints such as drought or salinity, V. vinifera cv. Chardonnay copes by enhancing RNA editing of the major gene products involved in growth processes. Again, care should be taken when drawing such conclusions, as the protein score at which PPR was identified was low.
Another proteomic study also highlighted the variation in protein amount, albeit to a lesser extent, between various grapevine cultivars (Sarry et al., 2004). None of the aforementioned proteins, however, was identified in berry mesocarp. Major changes in protein pattern from one cultivar to another were also reported in rice during drought (Salekdeh et al., 2002) or salt stress (Abbasi and Komatsu, 2004), as well as in wheat species differing in salt sensitivity (Majoul et al., 2000; Ouerghi et al., 2000). There are many phenotypical differences between Chardonnay and Cabernet Sauvignon, the most obvious being the colour of the berries. Shoot tips and young leaves also differ (Bettiga, 2003; Wolpert, 2003). Genetic variations have also been reported between grapevine cultivars (Aradhya et al., 2003; Adam-Blondon et al., 2004; Riaz et al., 2004; This et al., 2006); Cabernet Sauvignon resulted from a cross between Cabernet Franc and Sauvignon Blanc, while Chardonnay was a cross between Pinot Noir and Gouais Blanc (Bowers et al., 1999). In this study, cultivar differences have been observed at both physiological and protein levels in relation to abiotic stress adaptation over time. These observed variations probably arose from the genetic differences between these two cultivars.
|
Conclusions |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionConclusionsSupplementary dataReferences |
|---|
Here, the first proteomic study on grapevine shoots of two of the most important cultivars worldwide, Chardonnay and Cabernet Sauvignon, exposed to water or salt stress for either 8 d or 16 d, is presented. The adverse effects of both treatments were assessed at a physiological level by measuring the stem water potential, and shoot and leaf elongation. Significant differences in leaf elongation were found between cultivars in response to salinity and water deficit, with Chardonnay apparently more stress tolerant than Cabernet Sauvignon. Statistical analyses showed that protein quantities varied mainly in response to the cultivars, then with time, and finally with the stress. Many of the proteins up-regulated in Chardonnay were of unclassified or of unknown function, whereas proteins specifically up-regulated in Cabernet Sauvignon were involved in protein metabolism. Among the proteins responding to stress treatments, some have previously been involved in stress adaptation, such as PR-10, PRX, and bHLH, while other proteins, such as NMCP1, RPL39, DEL, and PPR, were described as being affected by abiotic stress in plants for the first time. Further analyses are needed to determine the functional role of many of these proteins in V. vinifera. Proteins specific to each cultivar are being investigated. By combining physiology and protein analyses, better characterization of the grape shoot response to two major abiotic stresses could be obtained.
|
Supplementary data |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionConclusionsSupplementary dataReferences |
|---|
The Supplementary data mentioned herein can be found at JXB online.
Supplementary Tables S1, S2, and S3 list the identified proteins with significant effect (P <0.10) for, respectively, the treatment (T) factor, the variety (V) factor, and the varietyxtreatment (VT) interaction in the ANOVA model.
Supplementary Figs S1–S6 illustrate the expression patterns of the isoforms identified for the following proteins: RuBisCO large subunit, actin, cytosolic ascorbate peroxidase, RNA-binding protein, heat shock protein 70 kDa, and ATP synthase beta subunit.
Supplementary Fig. S7 presents the functional classification according to MAtDB of the identified proteins significant for each individual effect of the ANOVA model. Supplementary Figs S8, S9, and S10 show examples of proteins responding to treatment, variety, and their interaction, respectively.
|
Acknowledgements |
|---|
The authors are grateful to the reviewers for their helpful comments. This work was supported by NSF Plant Genome Program Grant #0217653, the American Vineyard Foundation, and the Nevada Agricultural Experiment Station. The proteomic equipment used in this study was acquired with NIH/NCRR grants 5P20RR16464 and 1S10RR15919, and NSF Grant EPS-0132556. AE was supported by a TUBITAK-NATO B1 fellowship.
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionConclusionsSupplementary dataReferences |
|---|
Abbasi FM, Komatsu S. A proteomic approach to analyze salt-responsive proteins in rice leaf sheath. Proteomics (2004) 4:2072–2081.[CrossRef][ISI][Medline]
Adam-Blondon A-F, Roux C, Claux D, Butterlin G, Merdinoglu D, This P. Mapping 245 SSR markers on the Vitis vinifera genome: a tool for grape genetics. Theoretical and Applied Genetics (2004) 109:1017–1027.[CrossRef][ISI][Medline]
Agrawal GK, Rakwal R, Yonekura M, Kubo A, Saji H. Proteome analysis of differentially displayed proteins as a tool for investigating ozone stress in rice (Oryza sativa L.) seedlings. Proteomics (2002) 2:947–959.[CrossRef][ISI][Medline]
Alburquerque JA, Esquivel MG, Teixeira AR, Ferreira RB. The catabolism of ribulose bisphosphate carboxylase from higher plants. A hypothesis. Plant Science (2001) 161:55–65.
Ali GM, Komatsu S. Proteomic analysis of rice leaf sheath during drought stress. Journal of Proteome Research (2006) 5:396–403.[CrossRef][ISI][Medline]
Amme S, Maltros A, Schlesier B, Mock H-P. Proteome analysis of cold stress response in Arabidopsis thaliana using DIGE-technology. Journal of Experimental Botany (2006) 57:1537–1546.
Arabidopsis Genome Initiative. Analysis of the genome sequence of the flowering plant Arabidopsis thaliana. Nature (2000) 408:796–815.[CrossRef][Medline]
Aradhya MK, Dangl GS, Prins BH, Boursiquot JM, Walker MA, Meredith CP, Simon CJ. Genetic structure and differentiation in cultivated grape, Vitis vinifera L. Genetic Research (2003) 81:179–192.[CrossRef]
Askari H, Edqvist J, Hajheidari M, Kafi M, Salekdeh GH. Effects of salinity levels on proteome of Suaeda aegyptiaca leaves. Proteomics (2006) 6:2542–2554.[CrossRef][ISI][Medline]
Bae MS, Cho EJ, Choi EY, Park OK. Analysis of the Arabidopsis nuclear proteome and its response to cold stress. The Plant Journal (2003) 36:652–663.[CrossRef][ISI][Medline]
Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. Journal of the Royal Statistical Society B (1995) 57:289–300.
Benson EE, Roubelakis-Angelakis KA. Oxidative stress in recalcitrant tissue cultures of grapevine. Free Radical Biology and Medicine (1994) 16:355–362.[CrossRef][ISI][Medline]
Bettiga LJ. Chardonnay. In: Wine grape varieties in California (2003) University of California.
Bowers J, Boursiquot JM, This P, Chu K, Johansson H, Meredith C. Historical genetics: the parentage of chardonnay, gamay, and other wine grapes of northeastern France. Science (1999) 285:1562–1565.
Cagney G, Emili A. De novo peptide sequencing and quantitative profiling of complex protein mixtures using mass-coded abundance tagging. Nature Biotechnology (2002) 20:163–170.[CrossRef][ISI][Medline]
Carvalho LC, Esquivel MG, Martins I, Ricardo CP, Amancio S. Monitoring the stability of Rubisco in micropropagated grapevine (Vitis vinifera L.) by two-dimensional electrophoresis. Journal of Plant Physiology (2005) 162:365–374.[CrossRef][ISI][Medline]
Castelan-Estrada M, Vivin P, Gaudillère JP. Allometric relationships to estimate seasonal above-ground vegetative and reproductive biomass of Vitis vinifera L. Annals of Botany (2002) 89:401–408.
Castro AJ, Carapito C, Zorn N, Magné C, Leize E, Van Dorsselaer A, Clément C. Proteomic analysis of grapevine (Vitis vinifera L.) tissues subjected to herbicide stress. Journal of Experimental Botany (2005) 56:2783–2795.
Che FY, Fricker LD. Quantitative peptidomics of mouse pituitary: comparison of different stable isotopic tags. Journal of Mass Spectrometry (2005) 40:238–249.[CrossRef][ISI][Medline]<