At harvest, sunflower (Helianthus annuus L.) seeds are dormant and unable to germinate at temperatures below 15°C. Seed storage in the dry state, known as after-ripening, is associated with an alleviation of embryonic dormancy allowing subsequent germination at suboptimal temperatures. To identify the process by which dormancy is broken during after-ripening, we focused on the role of reactive oxygen species (ROS) in this phenomenon. After-ripening entailed a progressive accumulation of ROS, namely superoxide anions and hydrogen peroxide, in cells of embryonic axes. This accumulation, which was investigated at the cellular level by electron microscopy, occurred concomitantly with lipid peroxidation and oxidation (carbonylation) of specific embryo proteins. Incubation of dormant seeds for 3 h in the presence of hydrogen cyanide (a compound that breaks dormancy) or methylviologen (a ROS-generating compound) also released dormancy and caused the oxidation of a specific set of embryo proteins. From these observations, we propose a novel mechanism for seed dormancy alleviation. This mechanism involves ROS production and targeted changes in protein carbonylation patterns.
After-ripening is an intriguing phenomenon as it occurs at low seed moisture contents (MC), generally <0.10 g H2O/g dry weight (DW). Under these extreme conditions, water is generally not available for biochemical reactions, and very little is known about the cellular and molecular mechanisms involved in this process. However, changes in gene expression and/or protein synthesis during after-ripening have been shown to occur in seeds of Nicotiana tabacum (Leubner-Metzger, 2005), Nicotiana plumbaginifolia (Bove et al., 2005) and Arabidopsis thaliana (Cadman et al., 2006; Chibani et al., 2006), which presumably reflects the existence of hydrated pockets within cells or tissues of the mature seeds (Leubner-Metzger, 2005). Non-enzymatic reactions are also likely to occur during dry storage of seeds, such as lipid peroxidation (McDonald, 1999; Priestley, 1986; Wilson and McDonald, 1986) or the Amadori and Maillard reactions associated with free radical production and oxidation processes (Esashi et al., 1993; Murthy and Sun, 2000; Murthy et al., 2003; Sun and Leopold, 1995). Furthermore, several studies have documented the production of reactive oxygen species (ROS) during seed storage in the dry state (Bucharov and Gantcheff, 1984; Hendry, 1993; McDonald, 1999; Pukacka and Ratajczak, 2005). ROS can react with virtually all biological molecules including lipids, DNA and proteins. Because proteins have numerous biological functions, their oxidation may result in modification of their enzymatic and binding properties and lead to diverse functional changes. Oxidation of proteins can occur through a number of different mechanisms, such as the formation of disulfide cross-links and glycoxidation adducts, nitration of tyrosine residues, and carbonylation of specific amino acid residues (Davies, 2005). Recent studies have indicated that protein oxidation is not necessarily a deleterious phenomenon in plants (Job et al., 2005; Johansson et al., 2004). Moreover, cellular ROS may show some selectivity with respect to their targets. For example, H2O2, which is an oxidant, can react with specific molecules at specific sites (Davies, 2005; Halliwell and Gutteridge, 1999). Importantly, ROS have been invoked to play a role in cellular signaling (for review, see Bailly, 2004), raising the hypothesis that these compounds can facilitate the shift from a dormant to a non-dormant status in seeds.
Sunflower (Helianthus annuus) seeds provide an excellent system for studying dormancy because they are deeply dormant at harvest, but this dormancy is progressively lost during dry storage (Corbineau et al., 1990). Their dormancy results from both seed coat- and embryo-imposed dormancy, the latter being involved in the failure to germinate at 10–15°C. However, the molecular mechanisms of embryo dormancy and of its release during after-ripening are still largely unknown.
The objective of this study was to investigate whether the production of ROS and lipid autoxidation occur after harvest, and whether this can, in turn, bring about modifications in seed protein oxidation patterns that facilitate alleviation of seed dormancy. Moreover, in order to assess a putative causal association between ROS production, protein oxidation and dormancy alleviation, dormant sunflower seeds were imbibed in the presence of hydrogen cyanide, a compound that breaks dormancy (Bethke et al., 2006; Bogatek et al., 1991; Esashi et al., 1991; Taylorson and Hendricks, 1973), or in the presence of methylviologen (MV), a ROS-generating compound (Slooten et al., 1995). Although MV has not yet been reported as a dormancy-breaking chemical, we have used it here as a tool to assess the proposition that ROS generation is involved in dormancy alleviation in this species. The general finding is that dormancy release, both in dry and imbibed states, is associated with ROS production and the carbonylation of specific embryo proteins.
Figure 1 shows the germination behavior of dormant and non-dormant de-coated (naked) sunflower seeds at various temperatures. At 15, 20 and 25°C, dormant seeds germinated slowly and reached 80–90% germination within 6–8 days (Figure 1a). Only about 20% of dormant naked seeds were able to germinate at 10°C within 10 days (Figure 1a), but dry storage markedly enhanced their germination at all temperatures tested, and, in particular, they became able to fully germinate at 10°C within 4 days (Figure 1b).
ROS production during dry storage
To characterize ROS production during after-ripening, hydrogen peroxide and superoxide anion contents were determined in axes (see Figure 2) excised from dormant and non-dormant seeds (Table 1). Dormancy release was associated with a marked enhancement in both compounds, as the level of H2O2 doubled and that of O2−· increased by about 50%. Dry storage was also associated with a slight but significant increase in malondialdehyde (MDA) content, indicating the occurrence of lipid peroxidation (Table 1).
Table 1. Hydrogen peroxide, superoxide and malondialdehyde (MDA) contents in dry dormant and non-dormant sunflower axes
H2O2 (µmol g DW−1)
O2−· (µmol g DW−1)
MDA (µmol g DW−1)
Values are means of five replicates ± SD.
1.37 ± 0.21
5.29 ± 0.25
76.75 ± 3.65
2.10 ± 0.14
8.66 ± 0.51
90.62 ± 0.72
In situ accumulation of O2−· in whole seeds (embryos) and in sections through axes during dry storage is shown in Figure 2. At the whole-tissue level, accumulation of formazans occurred almost homogeneously in axes of non-dormant seeds (Figure 2c). In contrast, staining of dormant axes by nitroblue tetrazolium (NBT) was only very slight (Figure 2b). Transverse sections through the axes confirmed the presence of formazan deposits at the cellular level. In dormant axes, no NBT precipitation was visible within the cells (Figure 2d,f), whereas numerous dark spots indicated the presence of superoxide anion within the cells of non-dormant axes (Figure 2e,g).
The detection of hydrogen peroxide was carried out by CeCl3 staining and revelation by transmission electron microscopy (Bestwick et al., 1997; Pellinen et al., 2002). Cells from axes of both dormant and non-dormant seeds were filled with large lipid bodies and smaller protein bodies, which somewhat occluded observation of the cytoplasm (Figure 3a,b). However, in the presence of CeCl3, H2O2 was clearly evident as electron dense (cerium perhydroxide) spots (arrowed) within the cytoplasm of the cells (Figure 3c–f).While some H2O2 could be detected in dormant seeds (Figure 3c,e), there was much higher staining in the cytoplasm of cells from axes of non-dormant seeds, revealing endogenous H2O2 production mainly in the non-dormant tissues (Figure 3d,f).
To determine whether ROS accumulation during dry storage caused seed dormancy alleviation, and was not simply a side-effect accompanying seed storage, whole dormant sunflower seeds were stored at 75% RH (relative humidity) or 5% RH at 25°C for various durations, and the H2O2 content in the axes was determined (Table 2). Storage of seeds at 75% RH was associated with a progressive reduction of dormancy, whereas dormancy was maintained when seeds were stored at 5% RH. After 6 weeks of storage at 25°C and 75% RH, dormancy was lost, as these seeds germinated at 10°C almost as fast as the non-dormant ones (i.e. the seeds obtained after 2 months of dry storage at 25°C and 60% RH; Table 3 and Figure 1b). Under these conditions, the hydrogen peroxide content of the embryos increased from 1.3 to 2.2 µmol g DW−1, whereas this content remained unchanged when dormant seeds were stored for the same duration but at 5% RH, i.e. conditions that do not permit seed dormancy release (Table 3).
Table 2. Germination after 7 days at 10°C and hydrogen peroxide content of axes of dormant sunflower seeds stored for 3 and 6 weeks under 75% and 5% RH at 25°C
Duration and conditions of storage
Germination (%) at 10°C after 7 days
H2O2 content (µmol g DW−1)
Values are means of four (germination) or five (H2O2) replicates ± SD
None (after harvest)
19 ± 2.2
1.37 ± 0.21
3 weeks, 75% RH
42 ± 1.1
2.01 ± 0.12
3 weeks, 5% RH
22 ± 1.5
1.41 ± 0.04
6 weeks, 75% RH
70 ± 2.0
2.21 ± 0.22
6 weeks, 5% RH
21 ± 0.5
1.40 ± 0.03
Table 3. Identification of carbonylated proteins in embryo axes of dry dormant and non-dormant sunflower seeds
Change in carbonylation during after-ripening
Peptide sequences were identified by MS-MS sequencing, see for the peptide sequences. Experimental molecular weight (MW) and pI are indicated.
To determine whether ROS production during after-ripening could be associated with protein oxidation (carbonylation), one-dimensional (1D) and two-dimensional (2D) PAGE of seed protein extracts were performed, and the presence of carbonyl groups was detected by Western blotting using the 2,4-dinitrophenylhydrazine (DNPH) immunoassay (Korolainen et al., 2002).
The protein patterns obtained by 1D gel electrophoresis of soluble proteins from dormant and non-dormant embryonic axes are shown in Figure 4. Coomassie blue staining showed the soluble proteins from both dormant and non-dormant embryonic axes to be very similar. Soluble proteins from dormant axes exhibited three faint carbonylated bands of about 75, 60 and 55 kDa (Figure 4). In contrast, the extent of protein carbonylation was much higher in soluble proteins from non-dormant axes, and a number of new carbonylated bands were detected (Figure 4).
The protein targets of carbonylation during after-ripening were characterized by 2D gel electrophoresis and LC/MS–MS analyses. The 2D protein patterns of soluble proteins extracted from dormant and non-dormant axes were very similar (Figure 5a,c). However, in agreement with the data obtained by 1D gel electrophoresis (Figure 4), after-ripening was associated with increased carbonylation of several proteins (red arrows and red circles in Figure 5). There was also a decrease in the carbonylation level of some proteins during this process (yellow arrows in Figure 5). Finally, some carbonylated proteins remained at a constant level during after-ripening (green arrows in Figure 5). It is clear that not all seed proteins detected by silver staining were carbonylated, testifying to the specificity of protein oxidation during after-ripening.
Twenty spots representing carbonylated proteins were excised from 2D gels and analyzed by mass spectrometry. Identification failed for 12 proteins. Table 2 and Table S1 list the oxidized proteins that were identified by this approach, mainly by sequence homology with proteins from species other than sunflower. However, several proteins could be identified using the sunflower EST collections (http://cgpdb.ucdavis.edu/; http://genoplante-info.infobiogen.fr/), highlighting the usefulness of such collections as a tool for proteomics in the absence of extensive genomic sequence information. The data showed that the carbonylation level of elongation factor 2 (EF2, spot number 3 in Figure 5), pyruvate orthophosphate dikinase (PPDK, spot number 4 in Figure 5) and a protein corresponding to a sunflower EST, which has a strong homology with a 7S globulin of Sesamum indicum and with various vicilin-like proteins (spot number 15 in Figure 5) increased during dry after-ripening. In contrast, seed dormancy alleviation during dry storage was associated with decreased carbonylation of several proteins including a 20S proteasome α-subunit (spot number 46 in Figure 5). Proteins whose carbonylation level remained constant during dry storage included 81 and 70 kDa heat shock proteins (HSPs, spot numbers 7 and 9 in Figure 5), enolase (spot number 12 in Figure 5) and a protein with strong homology to a globulin-like protein from carrot (Daucus carota) (spot number 18 in Figure 5).
Effect of cyanide and methylviologen on the protein oxidation pattern
To further document a link between protein oxidation and release of seed dormancy, dormant sunflower seeds were imbibed in the presence of hydrogen cyanide, a compound known to break dormancy in many species (Bethke et al., 2006; Bogatek et al., 1991; Côme et al., 1988; Esashi et al., 1991; Taylorson and Hendricks, 1973). A 3 h imbibition of the dormant sunflower seeds in the presence of 1 mm hydrogen cyanide efficiently released dormancy (Figure 6). Furthermore, this incubation triggered the carbonylation of specific proteins in dormant axes only (e.g. the blue circles in Figure 7). This carbonylation pattern was specific as it was not seen with non-dormant seeds incubated either in water or in the presence of hydrogen cyanide (Figure 7d). To further assess the correlation between ROS accumulation, protein carbonylation and break of dormancy, dormant and non-dormant seeds were imbibed for 3 h in the presence of methylviologen, a compound known to induce ROS production (Slooten et al., 1995). After this treatment, the H2O2 content of dormant axes was 2.7 ± 0.11 µmol g DW−1, whereas it was 2.0 ± 0.13 µmol g DW−1 when dormant seeds were imbibed on water (data not shown). Figure 6 shows that this compound was very effective in promoting dormancy release in sunflower seeds, lending further support to the hypothesis that dormancy release is associated with the generation of ROS. Remarkably, not only did this compound break dormancy, but this also entailed the same specific pattern of protein oxidation as seen with hydrogen cyanide (blue circle in Figure 7). Oxyblots also revealed that imbibition of dormant and non-dormant embryos on water was associated with an increase in protein carbonylation (compare Figures 5 and 7). However, the level of protein oxidation was higher in non-dormant axes, which displayed many new oxidized proteins ranging from approximately 15–40 kDa (Figure 7b). Conversely, around 20 proteins that were found to be carbonylated during imbibition in dormant axes were not detected in non-dormant axes (Figure 7a, yellow circles). Carbonylated proteins identified by mass spectroscopy are listed in Table 4. EF2 (spot number 3), PPDK (spot number 4), 81 and 70 kDa HSP (spot numbers 7 and 9), enolase (spot number 12) and 7S globulin (spot number 15), previously identified in dry-after ripened seeds, were also carbonylated in the axes of all imbibed seeds (Figure 7, Tables 3 and 4). The globulin-like protein (spot number 18) was also always carbonylated in imbibed axes, but the level of oxidation appeared to be stronger in non-dormant axes isolated from seeds imbibed on water (Figure 7b). Treatment of imbibed dormant seeds with cyanide and MV also resulted in the carbonylation of proteins not detected in dry after-ripened seeds. For example, proteins that show homology with 11S storage protein (spot number 36 in Figure 7) and epoxide hydrolase (spot number 38 in Figure 7) were detected in dormant seeds but did not appear in the axes of non-dormant seeds treated with cyanide and MV. Other proteins sharing homology with an 11S globulin precursor, an alcohol dehydrogenase and a basic 2S albumin (spot numbers 19, 16 and 23, respectively, in Figure 7) were oxidized in axes of all seeds able to germinate. Spot number 21 was detected only in imbibed non-dormant seeds (Figure 7b,d,f), and this protein was found to be similar to dehydrin (Table 4).
Table 4. Identification of carbonylated proteins in embryo axes of dormant and non-dormant sunflower seeds imbibed for 3 h on water (W) or in the presence of hydrogen cyanide (CN) or methylviologen (MV) at 10°C
D, dormant seeds; ND, non-dormant seeds. Peptide sequences were identified by MS-MS sequencing; see Table 3 for characteristics of spots 3, 4, 7, 9, 12, 15 and 18, and for the peptide sequences. Experimental molecular weight (MW) and pI are indicated.
Dormancy is a characteristic feature of sunflower seeds at their harvest. This dormancy exhibits both seed coat and embryo components, and progressively appears during seed development on the mother plant (Corbineau et al., 1990). The seed-coat-imposed dormancy acts through the effects of phenolics on oxygen availability for the embryo, and is therefore expressed to a greater extent at high temperatures, while the embryo-imposed dormancy is likely to be involved in the restriction of germination at temperatures below 15°C (Corbineau et al., 1990; Gay et al., 1991). In agreement, dormant naked seeds assayed immediately after harvest only germinated very poorly at 10°C (Figure 1a), and after-ripening in dry conditions (20°C, 60% RH, 2 months) resulted in a widening of the temperatures permitting germination (Figure 1b).
Our results show that after-ripening is associated with an accumulation of superoxide anions and hydrogen peroxide in the embryonic axes (Table 1, Figures 2 and 3). O2−· appeared as formazan precipitates within the cells (Figure 2), and ultrastructural studies showed that hydrogen peroxide accumulated in the cytoplasm of embryonic axis cells (Figure 3). By storing the seeds at various RH, it was possible to modulate the extent of dormancy release (Table 2). This suggests that the water status within the embryo cell is likely to play a critical role in the process of after-ripening, and shows that hydrogen peroxide accumulation is tightly associated with the breaking of embryo dormancy as this compound only accumulated under conditions associated with dormancy release. Thus a causal link between ROS production, or at least H2O2 accumulation, and after-ripening is likely to exist, and this process is not just related to seed storage. This demonstrates that, in sunflower seeds, ROS production in the dry state is initiated after harvest, and that, in agreement with previous proposals that ROS can act as cell messengers (reviewed in Bailly, 2004), these molecules could therefore act as a signal to allow dormancy release and favor subsequent seed germination. However, it is also likely that prolonged storage of seeds in the dry state would be associated with a sustained production of ROS, which would lead to oxidative stress and to the related deteriorative events known to occur during seed aging (Bailly, 2004).
In plants, and more especially in seeds, ROS may originate from the mitochondrial respiratory chain or be produced through the action of enzymes such as NADPH oxidase (Bailly, 2004). However, in dry tissues, such as in mature orthodox seeds, respiration is prevented and enzymes are presumably not active, which suggests the occurrence of other mechanisms for ROS production. At low moisture content, non-enzymatic reactions are known to occur, such as the Amadori and Maillard reactions (Priestley, 1986; Sun and Leopold, 1995) and lipid peroxidation (McDonald, 1999). Enzymatic oxidation of lipids through lipoxygenase is also possible when water activity is as low as 0.4 (Drapron, 1985). Using model systems, such as those based on oil encapsulation in a glassy matrix, it has been demonstrated that glasses do not prevent oxygen diffusion, thus allowing autoxidation reactions of lipids (Andersen and Skibsted, 2002; Andersen et al., 2000; Nelson and Labuza, 1992). As the MDA content increased during dormancy alleviation (Table 1), we assume that after-ripening is associated with lipid peroxidation, which could be a consequence of ROS accumulation and may, in turn, be involved in free radical formation.
To test the hypothesis that ROS accumulation during after-ripening is involved in dormancy alleviation, and to assign a possible mechanism whereby this might occur, we characterized the oxidized proteome of embryonic axes from dormant and non-dormant seeds. Protein carbonylation results from oxidative attack on Arg, Lys, Pro or Thr residues of proteins, which can affect enzyme activities or alter susceptibility of the modified proteins to proteolysis (Berlett and Stadtman, 1997; Davies, 2005; Dunlop et al., 2002). Our present data clearly indicate the occurrence of carbonylation of specific embryonic proteins during after-ripening of sunflower seeds (Figures 4 and 5). An alternative possibility to account for our results could be a redistribution of carbonylated proteins in the various protein fractions (i.e. soluble and non-soluble proteins) because of cell structural changes occurring during after-ripening that would render the carbonylated proteins more soluble from the non-dormant axes than from the corresponding dormant ones. However, the fact that no marked ultrastructural changes occurred during after-ripening (Figure 3a,b) and that the soluble proteomes revealed by silver nitrate staining were very similar for the dormant and non-dormant axes does not favor this idea (Figure 5). We suggest that, in the sunflower seed system, protein carbonylation may result from an accumulation of ROS themselves, and from accumulated lipid peroxidation products such as MDA (Table 1), which is known to react with lysine residues to form carbonyl derivatives (Burcham and Kuhan, 1996; Liu and Wang, 2005). In animals, protein carbonylation has been widely used as a measurement of oxidative damage, and it has been shown to increase in aging tissues (Dalle-Donne et al., 2003; Ding et al., 2006; Nyström, 2005). In marked contrast, protein carbonylation may not be an inevitable consequence of tissue aging in plants. Thus, in Arabidopsis, patterns of protein carbonylation vary widely during progression of the life cycle, and the extent of protein carbonylation drops abruptly prior to the vegetative to reproductive transition (Johansson et al., 2004). Carbonylation of numerous proteins also occurs during Arabidopsis seed germination, although the germinated seeds gave rise to vigorous plantlets (Job et al., 2005).
The present data document the effectiveness of both hydrogen cyanide and methylviologen in breaking sunflower seed dormancy (Figure 6). Hydrogen cyanide has already been shown to alleviate seed dormancy in several species, including Arabidopsis (Bethke et al., 2006; Bogatek et al., 1991; Côme et al., 1988; Esashi et al., 1991; Taylorson and Hendricks, 1973). Methylviologen, on the other hand, is a compound known to generate oxidative stress in plants in the light through the photosynthetic electron transport chain (Xiong et al., 2007), as well as in the dark, but by an unknown mechanism (Slooten et al., 1995). However, its utilization as a compound that breaks seed dormancy has not been reported to date, and the advantage of using methylviologen in our study is that it supplies evidence that ROS probably play a fundamental role in seed dormancy release.
The oxyblots presented in Figure 7 clearly show that carbonylation of specific proteins occurred upon treating the dormant sunflower seeds with either hydrogen cyanide or methylviologen (e.g. proteins shown in the blue circles in Figure 7). This specific carbonylation pattern was not seen by incubating dormant seeds in water only (Figure 7a), although this latter incubation also resulted in an increase in protein carbonylation compared with the dry seeds (compare Figures 5b and 7a). It is also interesting that the pattern of protein carbonylation is similar in all of the non-dormant axes (Figure 7), regardless of the treatment. This has significance in the present experiments and in interpreting the data as they provide treatment controls illustrating the reproducibly of the technique and the biological behavior. In summary, seed dormancy release during dry storage and imbibition in the presence of hydrogen cyanide or methylviologen was associated with the appearance of specific carbonylation patterns, although these patterns differed when dormancy was broken in dry and in imbibed state conditions (compare Figures 5 and 7). It would nevertheless be interesting to further characterize the specific features of the carbonylated proteome during breaking of dormancy by carrying out time-course experiments during the after-ripening process. The similarity of the effects of HCN and MV on the oxidized proteome also suggests that these compounds share a common mechanism for triggering dormancy release in imbibed dormant seeds.
The role of protein oxidation in dormancy alleviation could be discussed in relation to the nature of the carbonylated proteins presently identified. EF2, PPDK and 7S globulin (spot numbers 3, 4 and 15, respectively, in Figure 5, Table 2 and Table S1) exhibited an increased carbonylation level during after-ripening. EF2 catalyzes peptidyl-tRNA translocation of the ribosome during the elongation phase of protein translation. Its level is known to decrease dramatically during wheat seed development, particularly during the desiccation phase (Gallie et al., 1998). Carbonylation of EF2 during storage in the dry state after completion of seed development might therefore terminate protein synthesis that is associated strictly with developmental processes. PPDK plays a role in photosynthesis in C4 plants but its function in C3 plants is not fully elucidated. Interestingly, this enzyme appears to be absent from many dicot seeds, including Arabidopsis, with the remarkable exception of sunflower seeds (Chastain et al., 2006). As PPDK is presumed to be involved in seed development rather than in seed germination (Chastain et al., 2006), its carbonylation during after-ripening could provide a means to down-regulate some residual enzyme activity. In this context, it has been shown in rice that, as seed development progresses towards late maturation, the enzyme undergoes post-translational down-regulation in terms of activity and amount via regulatory phosphorylation (PPDK inactivation) and protein degradation (Chastain et al., 2006). Carbonylation of storage proteins has previously been reported in dry mature Arabidopsis seeds, and it was suggested that carbonylation of these proteins facilitates their mobilization during germination (Job et al., 2005). Thus, in sunflower seeds, breaking of dormancy in the dry state may be associated with preparation for storage protein mobilization.
The carbonylation level of some proteins decreased during after-ripening. For example, this was the case for a 20S proteasome α-subunit (spot number 46; Figure 5, Table 2 and Table S1). This observation is consistent with data showing a requirement of for proteasome activity in sunflower embryos for both the breaking of dormancy by ethylene and the progression of germination (Borghetti et al., 2002).
Proteins that were similarly oxidized in dry and imbibed dormant and non-dormant sunflower axes were also characterized. Among them, HSP 70 and 81 (spot numbers 7 and 9, Figures 5 and 7, Tables 3 and 4 and Table S1) and enolase (spot number 12, Figures 5 and 7, Tables 3 and 4 and Table S1) were found. Molecular chaperones are known to be targets of carbonylation in yeast and bacteria challenged by oxidative stress (Cabiscol et al., 2000; Tamarit et al., 1998). Thus, it is possible that carbonylation of these HSPs in both dormant and non-dormant sunflower seeds reflects the occurrence of an oxidative stress during the desiccation phase of seed development. It has been suggested that the protective role these proteins play is as shields protecting other proteins against ROS damage (Cabiscol et al., 2000). Enolase catalyzes the dehydration of 2-phosphoglycerate to PEP and is therefore the second enzyme involved in metabolism, after PPDK, that has been found to be oxidized in sunflower axes. This result is consistent with the observation that enolase is one of the most prominent carbonylated proteins in dry mature Arabidopsis seeds (Job et al., 2005).
The use of hydrogen cyanide or methylviologen also allowed identification of carbonylated proteins that were specifically associated with breaking of dormancy in the imbibed state. This was the case for alcohol dehydrogenase (ADH, spot number 16, Figure 7, Table 4 and Table S1), which was carbonylated in all axes of seeds undergoing germination. Consistent with this, the beneficial effect of alcohol on breaking the dormancy of seeds (Cohn et al., 1989; Corbineau et al., 1991) or buds (Claassens et al., 2005) is well documented. Two identified proteins, globulin precursor isoform 4 (spot number 19 Figure 7, Table 4 and Table S1) and a basic 2S albumin (spot number 23, Figure 7, Table 4 and Table S1) exhibited the same pattern of oxidation as ADH, supporting the finding that germination requires the oxidation of specific reserve proteins. A putative epoxide hydrolase was found to be carbonylated in dormant axes treated by hydrogen cyanide and methylviologen (spot number 38, Figure 7, Table 4 and Table S1). Epoxide hydrolases catalyze the conversion of epoxides to diols. The known functions of such enzymes include detoxification of xenobiotics, drug metabolism, synthesis of signaling compounds, intermediary metabolism and responses to oxidative stress (Newman et al., 2005). In plants, epoxide hydrolases are thought to participate in general defense systems (Newman et al., 2005). Therefore, the specific carbonylation of this protein is consistent with the occurrence of oxidative stress during dormancy alleviation.
In conclusion, taken together, our results allow us to propose a novel mechanism for seed dormancy release. This mechanism involves a change in proteome oxidation, resulting from an accumulation of ROS during after-ripening. Hence, ROS accumulation appears to be a key signal governing cell activity during after-ripening. Furthermore, this hypothesis may provide a more general model to account for breaking of seed dormancy in view of the similar results obtained during breaking of dormancy in the dry state and the imbibed state.
Two batches of sunflower (Helianthus annuus L., cv LG5665) seeds, harvested in 2004 and 2005 near Montélimar (Drôme, France) and purchased from Limagrain (http://www.limagrain.com), were used in this study. Seeds were dormant at harvest and their moisture content was approximately 4% on a fresh weight (FW) basis. After-ripening was performed by placing the seeds after harvest at 25°C and 60% relative humidity (RH) for 2 months (Corbineau et al., 1990). Seeds were also placed at 25°C over saturated solutions of NaCl and ZnCl2 in tightly closed jars, giving RH of 75\% and 5%, respectively (Vertucci and Roos, 1993). All the results presented in this study represent means of the data obtained from seeds harvested in both 2004 and 2005.
Germination assays were performed with naked seeds (i.e. without pericarp) in darkness in 9 cm Petri dishes (25 seeds per dish, eight replicates) placed on a layer of cotton wool moistened with deionized water. A seed was considered as germinated when the radicle had elongated to 2–3 mm. Germination counts were made daily for 10 days.
Cyanide and methylviologen treatments
The treatment of sunflower embryos by gaseous 1 mm HCN was carried out as described by Bogatek et al. (1991). Naked dry seeds were placed in a tightly closed glass jar (500 ml volume) on a layer of sterile cotton wool moistened with deionized water (50 seeds per jar). A glass tube containing 5 ml of 0.1 m KCN solution placed in the jar was used as a source of gaseous HCN, which was produced by acidifying the KCN solution with 5 ml of lactic acid (10% v/v). After 3 h of treatment in darkness at 10°C, the jars were opened and gaseous cyanide released, and the seeds were rinsed carefully three times with deionized water before germination tests or biochemical analyses. Treatment by methylviologen was carried out by placing naked seeds in darkness at 10°C on a cotton wool moistened with a solution of 0.1 mm MV for 3 h. Embryos were also rinsed with deionized water after treatment.
Determination of superoxide anion content
Superoxide anion content was determined according to the method developed by Elstner and Heupel (1976). Axes (0.2 g FW) were ground in 4 ml of sodium phosphate buffer (pH 7.8, 50 mm) at 4°C. The extracts were centrifuged at 16 000 g for 15 min, and the resulting supernatants were used for O2−· determination. The supernatant (1 ml) was first incubated at 25°C for 30 min in the presence of 1 mm hydroxylamine hydrochloride in 50 mm sodium phosphate buffer (pH 7.8). Then, 0.5 ml of this reaction mixture was incubated with 0.5 ml of 17 mm sulfanilamide and 0.5 ml of 7 mm 2-naphtylamine at 25°C for 30 min. The absorbance was measured at 540 nm after centrifugation at 13 000 g for 10 min. A calibration curve was established using sodium nitrite. The results are expressed as µmol g−1 DW of seeds and correspond to the means of measurements carried out on five extracts ± SD.
In situ localization of superoxide anion
Whole axes and hand-cut sections of axes were incubated in 6 mm nitroblue tetrazolium (NBT) in 10 mm Tris-HCl buffer (pH 7.4) at room temperature for 30 min. Superoxide anion was visualized as deposits of dark-blue insoluble formazan compounds (Beyer and Fridovich, 1987).
Determination of hydrogen peroxide content
The H2O2 content of excised axes was determined according to the method described by O’Kane et al. (1996). Axes (0.5 g FW) were ground in a mortar and homogenized with 5 ml of 0.2 m perchloric acid. After 15 min of centrifugation at 13 000 g at 4°C, the resulting supernatant was neutralized to pH 7.5 with 4 m KOH and then centrifuged at 1000 g for 3 min at the same temperature. The supernatant was immediately used for spectrophotometric determination of H2O2 at 590 nm using a peroxidase-based assay. The reaction mixture contained 12 mm 3-dimethylaminobenzoic acid in 0.375 m phosphate buffer (pH 6.5), 1.3 mm 3-methyl-2-benzothiazolidone hydrazone, 20 µl (0.25 U) horseradish peroxidase (Sigma; http://www.sigmaaldrich.com/) and 50 µl of the collected supernatant to a total volume of 1.5 ml. The reaction was started by the addition of peroxidase. Increase in absorbance at 590 nm was monitored after 5 min at 25°C and compared with the absorbance obtained with known amounts of H2O2. The results are expressed as nmol H2O2 g−1 DW and correspond to means of the values obtained with five different extracts ± SD.
Subcellular localization of hydrogen peroxide
The localization of H2O2 was determined by CeCl3 staining as described by Bestwick et al. (1997). Sections of axes (approximately 5 mm3) were imbibed in 50 mm MOPS buffer (pH 7.2) containing 5 mm CeCl3 under vacuum until they were fully infiltrated. CeCl3-treated sections and control sections (without staining) were fixed in 6% glutaraldehyde in 25 mm sodium phosphate buffer (pH 7.2) for 24 h at room temperature, washed in the same buffer and post-fixed in 1% osmium tetroxide in 25 mm sodium phosphate buffer (pH 7.2) for 18 h at room temperature. After several washes in double-distilled water, the tissues were progressively dehydrated in ethanol, then soaked in propylene oxide and embedded in Araldite (Sigma). The blocks were sectioned with glass knives at 120 nm using a Reichert Ultractu S (Leica, http://www.leica-microsystems.com), stained with lead citrate and 2% uranyl acetate (Reynolds, 1963), and viewed with a LEO912 transmission electron microscope (Leo Electron Microscopy, http://www.stm.zeiss.com).
Lipid peroxidation was evaluated by measuring malondialdehyde (MDA) content from 0.5 g FW of embryonic axes, according to the method of Heath and Parker (1968). The results are expressed as µmol g−1 DW of seeds and correspond to means of measurements carried out with five extracts ± SD.
Preparation of protein extracts
Axes from dry dormant and non-dormant seeds (140 mg FW, corresponding approximately to 30 axes) were ground in liquid nitrogen using a mortar and pestle. Soluble proteins were extracted from the resulting powder at 4°C in 1.0 ml of a buffer containing 10 mm HEPES, 1 mm EDTA, the protease inhibitor cocktail ‘complete Mini’ from Roche Molecular Biochemicals, 60 U DNAse I (Roche Diagnostics, http://www.roche-applied-science.com), and 6 U RNAse A (Sigma, http://www.sigmaaldrich.com). After 10 min at 4°C, 20 mm dithiothreitol (DTT) was added, and the protein extracts were stirred for 20 min at 4°C and then centrifuged (20 000 g, 15 min) at 4°C. The final supernatant corresponded to the soluble protein extract. Protein concentrations in the various extracts were measured according to the method described by Bradford (1976) using a Bio-Rad assay kit (http://www.bio-rad.com/). Bovine serum albumin was used as a standard.
One- and two-dimensional electrophoresis
One-dimensional SDS–PAGE of seed protein extracts (5 µg protein) was performed using 12% w/v polyacrylamide resolving gels, as described by Laemmli (1970).
Proteins were also analyzed by two-dimensional gel electrophoresis as described previously (Görg et al., 1987; Job et al., 2005). Isoelectrofocusing (100 μg protein) was carried out using gel strips forming an immobilized non-linear pH gradient from 3 to 10 (Immobiline DryStrip, pH 3–10 NL, 18 cm; Amersham Pharmacia Biotech; http://www5.amershambiosciences.com/). Strips were rehydrated for 14 h at 22°C with the thiourea/urea lysis buffer as described previously (Harder et al., 1999), containing 2% v/v Triton X-100, 20 mm dithiothreitol and the protein extracts. Isoelectrofocusing was performed at 22°C in the Multiphor II system (Amersham Pharmacia Biotech) for 1 h at 300 V and 7 h at 3500 V. Then, the gel strips were equilibrated for 2 × 20 min in 2x 100 ml of equilibration solution containing 6 m urea, 30% v/v glycerol, 2.5% w/v SDS, 0.15 m bis-Tris and 0.1 m HCl (Görg et al., 1987; Harder et al., 1999). DTT (50 mm) was added to the first equilibration solution, and iodoacetamide (4% w/v) was added to the second (Harder et al., 1999). Separation in the second dimension was carried out in polyacrylamide gels (10% w/v acrylamide, 0.33% w/v piperazidine diacrylamide, 0.18 m Trizma base, 0.166 m HCl, 0.07% w/v ammonium persulfate and 0.035% v/v Temed). Electrophoresis was performed at 10°C in a buffer (pH 8.3) containing 25 mm Trizma base, 200 mm taurine and 0.1% w/v SDS for 1 h at 35 V and for 14 h at 110 V. Ten gels (200 × 250 × 1.0 mm) were run in parallel (Isodalt system; Amersham Pharmacia Biotech). For each treatment analyzed, 2D gels were run in triplicate.
Detection of carbonylated proteins and Western blotting
The appearance of carbonyl groups in proteins was analyzed by immunodetection of 2,4-dinitrophenylhydrazone (DNP)-derivatized protein as described previously (Job et al., 2005; Korolainen et al., 2002). SDS was added to the protein extract (100 µl, 10 µg µl−1) to a final concentration of 0.8%. Following dialysis, four volumes of 10 mm DNPH (Sigma)/2 m HCl were added. Samples were agitated for 30 min at room temperature, and five volumes of 20/80 ice-cold TCA/acetone containing 1 mm DTT were added to each sample. The samples were centrifuged for 15 min at 15 000 g at 4°C. The precipitated protein was then washed three times with ice-cold acetone containing 1 mm DTT, then with 1 ml of 1:1 v/v ethanol:ethyl acetate, and resolubilized in the thiourea/urea lysis buffer containing 2% v/v Triton X-100 and 20 mm DTT. Proteins were separated by 1D or 2D SDS–PAGE as described above, and transferred to nitrocellulose sheets (Bio-Rad) using standard procedures. Carbonylated proteins were revealed by incubation with rabbit anti-DNP antibodies (Chemicon, http://www.chemicon.com) followed by incubation with anti-rabbit secondary antibodies conjugated to horseradish peroxidase (Sigma) and detection with the ECL kit (Roche Diagnostics) (Job et al., 2005). Relative protein carbonyl levels were quantitated by densitometric analyses of the blots as described above.
Protein identification by mass spectrometry
Bands and spots of interest were excised from 1D and 2D SDS–PAGE gels using sterile tips and placed in 1.5 ml sterile tubes. Each polyacrylamide piece was rinsed with water, then reduced with 10 mm DTT, alkylated with 55 mm iodoacetamide, and incubated overnight at 37°C with 12.5 ng µl−1 trypsin (sequencing grade; Roche Diagnostics) in 25 mm NH4HCO3. The tryptic fragments were extracted, dried, reconstituted with 2% v/v acetonitrile, 0.1% formic acid and sonicated for 10min. Analysis of tryptic peptides by tandem mass spectrometry was performed on a nanoelectrospray ionization quadrupole time-of-flight hybrid mass spectrometer (Q-TOF Ultima Global; Waters Micromass, http://www.waters.com) coupled with a nano-HPLC (Cap-LC; Waters Micromass). The samples were loaded and desalted on a C18 pre-column (LC-Packings PepMap C18, 5 µm, 100 Å, 300 µm × 5 mm; Dionex Corp., http://www.dionex.com) at a flow rate of 20 µl min−1 isocratically with 0.1% formic acid. The peptides were separated on a C18 column (Atlantis dC18, 3 µm, 75 µm × 150 mm Nano Ease; Waters). After washing with solvent A (water/acetonitrile 98/2 v/v, 0.1% formic acid), a linear gradient from 5% to 60% of solvent B (water/acetonitrile 20/80 v/v, 0.1% formic acid) was developed over 80 min at a flow rate of 180 nl min−1. The Q-TOF spectrometer was operated in data-dependent analysis mode using a 1 sec mass spectrometry (MS) survey scan on three different precursor ions. The peptide masses and sequences obtained were either matched automatically to proteins in a non-redundant database National Center for Biotechnology Information, NCBI, http://www.ncbi.nlm.nih.gov using the Mascot MS/MS ions search algorithm (http://www.matrixscience.com) or manual blast searches were performed against the current databases. (NCBI, Swiss-Prot, http://expasy.org/sprot/)