A novel role for oleosins in freezing tolerance of oilseeds in Arabidopsis thaliana

Authors

  • Takashi L. Shimada,

    1. Department of Botany, Graduate School of Science, Kyoto University, Kyoto 606-8502, Japan, and
    Search for more papers by this author
  • Tomoo Shimada,

    1. Department of Botany, Graduate School of Science, Kyoto University, Kyoto 606-8502, Japan, and
    Search for more papers by this author
  • Hideyuki Takahashi,

    1. Department of Botany, Graduate School of Science, Kyoto University, Kyoto 606-8502, Japan, and
    Search for more papers by this author
  • Yoichiro Fukao,

    1. Department of Botany, Graduate School of Science, Kyoto University, Kyoto 606-8502, Japan, and
    2. Graduate School of Biological Sciences, Nara Institute of Science and Technology, Ikoma 630-0101, Japan
    Search for more papers by this author
  • Ikuko Hara-Nishimura

    Corresponding author
    1. Department of Botany, Graduate School of Science, Kyoto University, Kyoto 606-8502, Japan, and
      *(fax +81 75 753 4142; e-mail ihnishi@gr.bot.kyoto-u.ac.jp).
    Search for more papers by this author

*(fax +81 75 753 4142; e-mail ihnishi@gr.bot.kyoto-u.ac.jp).

Summary

Oil bodies in seeds of higher plants are surrounded with oleosins. Here we demonstrate a novel role for oleosins in protecting oilseeds against freeze/thaw-induced damage of their cells. We detected four oleosins in oil bodies isolated from seeds of Arabidopsis thaliana, and designated them OLE1, OLE2, OLE3 and OLE4 in decreasing order of abundance in the seeds. For reverse genetics, we isolated oleosin-deficient mutants (ole1, ole2, ole3 and ole4) and generated three double mutants (ole1 ole2, ole1 ole3 and ole2 ole3). Electron microscopy showed an inverse relationship between oil body sizes and total oleosin levels. The double mutant ole1 ole2, which had the lowest levels of oleosins, had irregular enlarged oil-containing structures throughout the seed cells. Germination rates were positively associated with oleosin levels, suggesting that defects in germination are related to the expansion of oil bodies due to oleosin deficiency. We found that freezing followed by imbibition at 4°C abolished seed germination of single mutants (ole1, ole2 and ole3), which germinated normally without freezing treatment. The treatment accelerated the fusion of oil bodies and the abnormal-positioning and deformation of nuclei in ole1 seeds, which caused seed mortality. In contrast, ole1 seeds that had undergone freezing treatment germinated normally when incubated at 22°C instead of 4°C, because degradation of oils abolished the acceleration of fusion of oil bodies during imbibition. Taken together, our findings suggest that oleosins increase the viability of over-wintering oilseeds by preventing abnormal fusion of oil bodies during imbibition in the spring.

Introduction

Oilseed crops have received increasing attention as a bio-energy and biomass resource in recent years. Much effort has focused on cryopreservation and desiccation tolerance of oilseed crops (Ellis et al., 1991; Leprince et al., 1998; Roberts and Ellis, 1989). On the other hand, plants that live in temperate and cool-temperate zones need to survive the harsh winter. They produce over-wintering seeds. The seeds have acquired resistant systems against freezing, which is a severe multiple stress involving mechanical stress, low-temperature stress and drought stress. However, the mechanism underlying the freezing tolerance of oilseeds is not known. Clarification of the mechanism is important from agricultural and ecological points of view.

Arabidopsis thaliana, an oilseed plant, accumulates a large amount of storage lipids, mainly triacylglycerols, in the oil bodies of seeds. The lipids are used as sources of carbon and energy for seed germination and subsequent seedling growth. Oil bodies in seeds are usually 0.5–2.0 μm in diameter, and are surrounded by a phospholipid monolayer overlaid with oil-body membrane proteins (Huang, 1996; Jolivet et al., 2004; Lin et al., 2002; Naested et al., 2000; Tzen et al., 1997). The most abundant proteins in oil bodies of seeds are oleosins (Huang, 1992). Oleosins are plant-specific proteins. Arabidopsis has 16 oleosins (five seed-type oleosins, eight tapetum-type oleosins and three seed/microspore-type oleosins) (Kim et al., 2002) (Figure S1). All oleosins have three domains: a hydrophilic N-terminal domain, a hydrophobic central domain and a hydrophilic C-terminal domain (Abell et al., 2004). The central domain, which is highly conserved, is essential for targeting to oil bodies (Abell et al., 1997; van Rooijen and Moloney, 1995). It is composed of two anti-parallel β-strands connected by a proline knot motif, and is inserted into the triacylglycerol matrix of oil bodies, forming a hairpin-like structure (Frandsen et al., 2001; Tzen et al., 1992).

The size of oil bodies is correlated with oleosin content in seeds (Ross et al., 1993; Ting et al., 1996; Tzen et al., 1993), and therefore might be regulated by oleosins. Maize has two types of oil kernels, high-oil kernels and low-oil kernels. The former have a high ratio of oil to oleosin and are large and spherical, while the latter have a low ratio of oil to oleosin and are small with irregular surfaces (Ting et al., 1996). Oil bodies of avocado mesocarp, which have no oleosins, are very large (20 μm in diameter) (Platt-Aloia and Thompson, 1981). It has been suggested that steric hindrance and electrical repulsion by oleosins are involved in preventing fusion of oil bodies (Tzen and Huang, 1992; Tzen et al., 1992). Siloto et al. (2006) confirmed that oleosins function as a size regulator of oil bodies in Arabidopsis seeds. However, the physiological role of oleosins of oilseeds is unclear. Oilseed crops such as coffee, cocoa and neem, which are designated as ‘recalcitrant’, are sensitive to freezing and desiccation (Ellis et al., 1991; Roberts and Ellis, 1989). These oilseed crops that originate in the tropics have low oleosin contents (Leprince et al., 1998). This suggests a relationship between oleosins and freezing tolerance.

In this study, we generated an oleosin-deficient mutant series of Arabidopsis. Analyses of the mutant series revealed an inverse relationship between oil-body sizes and total oleosin levels, and positive association of germination rates with oleosin levels. Interestingly, we found that freezing followed by imbibition at 4°C caused seed mortality of single mutants, which germinated normally without freezing treatment. Our findings provide not only a valuable insight into the freezing tolerance of oilseeds, which is required for over-wintering, but also suggest possibilities for producing genetically modified oilseed crops for developing bio-energy sources.

Results

Seed-type oleosin-deficient mutants of Arabidopsis

Arabidopsis thaliana has five seed-type oleosins (Figure S1) (Kim et al., 2002). Because of the extremely low level of the expression of At5G51210, we focused on the remaining four oleosins (At4G25140, At5G40420, At3G27660 and At3G01570). In a reverse-genetic approach, we generated four T-DNA insertion mutants for these oleosin genes (Figure 1a). To determine the abundance of each oleosin in seeds, we isolated oil bodies from the seeds of wild-type (Col-0) and oleosin-deficient mutants, and subjected them to SDS–PAGE followed by Coomassie blue staining. The wild-type oil bodies gave four bands with molecular masses between 15 and 20 kDa on the SDS gel (Col-0, Figure 1b). One of the bands disappeared in oil bodies from each oleosin-deficient mutant (Figure 1b). The results indicated that each of the four bands correspond to the oleosins. We designated the oleosins OLE1, OLE2, OLE3 and OLE4 (corresponding to At4G25140, At5G40420, At3G27660 and At3G01570, respectively) in decreasing order of their abundance in dry seeds (Figure 1b).

Figure 1.

 Arabidopsis seed-type oleosins and their knockout mutants.
(a) Seed-type oleosin mutants with a T-DNA insertion (ole1, ole2, ole3 and ole4). Shaded boxes, exons; open boxes, introns.
(b) Oil bodies containing 350 μg of oils from seeds of wild-type (Col-0) and single and double mutants were subjected to SDS–PAGE with Coomassie blue staining. *1, steroleosins (At5G50600 and At5G50700); *2, caleosin (At4G26740); *3, AWPM-19-like (At1G04560).
(c) Immunoblot of homogenates from two seed grains of wild-type (Col-0), single mutants and double mutants with either anti-OLE1 antibody or anti-OLE2 antibody. The ole1 and ole2 mutants were transformed with pOLE1::OLE1 and pOLE2::OLE2, respectively, for complementation.

We further generated three double mutants (ole1 ole2, ole1 ole3 and ole2 ole3) by crossing each pair of single mutants except ole4, because the OLE4 content in seeds was too low for any phenotype to be apparent (Figures 3 and 5). We also isolated oil bodies from seeds of these double mutants. Each double mutant lacked two oleosin bands on the SDS gel with Coomassie blue staining (Figure 1b). These results indicated that the single and double mutants that we generated were oleosin-deficient mutants. Of all the single and double mutants, the double mutant ole1 ole2 had the lowest total oleosin content.

Figure 3.

 Double mutants ole1 ole2 and ole1 ole3 have a defect in seed germination.
Seeds of wild-type (Col-0), single mutants, double mutants and complemented plants were stratified and then germinated in the absence (a) or presence (b) of sucrose at 22°C under continuous light. Germination rates were determined at 3, 5, 7, 9 and 11 days. ole1 ole2 seeds, which had the lowest oleosin content among the oleosin mutants, hardly germinated at all, and ole1 ole3 seeds, which had the second lowest content of oleosins, exhibited germination rates of approximately 50%. Supplying sucrose did not significantly affect the germination rates.

Figure 5.

 The levels of freezing-induced delay of germination increased in inverse proportion to oleosin levels.
(a, b) Seeds of wild-type (Col-0), single mutants, double mutants and complemented plants were stored at −30°C for 1 day (freezing treatment) and then were stratified before germination in the absence (a) or presence (b) of sucrose. The symbols for each oleosin mutant are the same as in Figure 3.

We raised antibodies against each of OLE1 and OLE2. Immunoblots using these antibodies showed that seeds of ole1, ole1 ole2 and ole1 ole3 lacked the OLE1 band, while seeds of ole2, ole1 ole2 and ole2 ole3 lacked the OLE2 band (Figure 1c). Seeds of ole2, ole3, ole4 and ole2 ole3 accumulated about the same amount of OLE1 as did wild-type seeds, while seeds of ole1, ole3, ole4 and ole1 ole3 accumulated about the same amount of OLE2 as did wild-type seeds (Figure 1c).

Double mutants ole1 ole2 and ole1 ole3 have greatly enlarged and irregularly shaped oil-containing structures throughout the seed cells

To determine how oleosin deficiency affects formation of oil bodies in seeds, we performed an ultrastructural analysis of seeds of single and double mutants. The cross-sections of oil bodies of single mutants increased in inverse proportion to the oleosin levels in the order ole4, ole3, ole2 and ole1 (Figure 2b–e,k). ole1 seeds had large spherical oil bodies, some of which had cross sections of 4–10 μm2 (Figure 2e,k), while wild-type seeds had uniformly sized oil bodies with cross-sections of approximately 1 μm2 (Figure 2a,k). This phenotype of ole1 is similar to that of OLE1/OLEO1 knockdown seeds (Siloto et al., 2006).

Figure 2.

 Oleosin deficiency causes enlargement of oil bodies in seeds.
(a–j) Electron micrographs of seeds of Col-0 (a), single mutants (b–e) and double mutants (f–h). The sizes of oil bodies of were in an inverse relationship with the oleosin contents in the mutant seeds. The large oil bodies of ole1 and ole2 were reduced by expressing OLE1 and OLE2, respectively, resulting in normal-sized oil bodies (i, j). OB, oil body; PSV, protein storage vacuole; CW, cell wall. Scale bars = 2 μm.
(k) A histogram showing a quantitative analysis of oil-body sizes in cross-sections of electron micrographs.

The oil bodies of the double mutants ole1 ole3 and ole1 ole2 (Figure 2g,h,k) were much larger than those of ole1. The seeds of ole1 ole2, which had the lowest content of oleosins among the mutants, exhibited the severest phenotype. ole1 ole2 seeds had oil-containing structures that were greatly enlarged (Figure 2h). The cross-sections of the structures were >10 μm2 (Figure 2k). These irregularly shaped structures may have formed by fusion of oil bodies during seed maturation. Our results clearly show an inverse relationship between oil-body sizes and total oleosin levels. Oleosin deficiency may promote fusion of oil bodies in the cells during seed maturation.

For complementation, we introduced the OLE1 gene into the ole1 mutant, and designated the transgenic plant ole1(OLE1), and introduced the OLE2 gene into the ole2 mutant and designated the transgenic plant ole2(OLE2). An OLE1 band was detected in ole1(OLE1) seeds and an OLE2 band was detected in ole2(OLE2) seeds (Figure 1c). The sizes of oil bodies in ole1(OLE1) and ole2(OLE2) were the same as those in the wild-type (Figure 2i–k). This indicated that oleosin genes were responsible for the abnormal sizes of oil bodies in these mutants.

Electron micrographs of seeds of these mutants also showed changes in the protein storage vacuoles (PSVs) that accumulate seed-storage proteins. Wild-type seeds have spherical PSVs located near the center of the cells (Figure 2a). However, the mutants had irregularly shaped PSVs dispersed throughout the cells (Figure 2d–h). However, the protein compositions in the seeds of each mutant were not different from wild-type (Figure S2). Conversely, Arabidopsis mutants that have a defect in accumulation of storage proteins in PSVs have smaller oil bodies in seeds than wild-type (Fuji et al., 2007; Li et al., 2006; Shimada et al., 2003, 2006; Tamura et al., 2007). These results suggest that oil bodies and PSVs form in a coordinated manner on the endoplasmic reticulum during seed maturation by competing for the spaces in seed cells.

Double mutants ole1 ole2 and ole1 ole3 have a defect in seed germination

OLEO1/OLE1 knockdown seeds germinated as wild-type seeds when grown in the presence of sucrose, and germinated just 1 day late when grown in the absence of sucrose (Siloto et al., 2006). Unexpectedly, however, we found ole1 ole2 seeds hardly germinated at all and ole1 ole3 seeds germinated poorly in the absence of sucrose (Figure 3a). The final germination rates of ole1 ole2 seeds and ole1 ole3 seeds were only approximately 10 and 50% at 11 days, respectively (Figure 3a). In contrast, the final germination rates of the other mutants and wild-type seeds were more than 90%. Addition of sucrose did not notably affect the germination rates of these mutant seeds (Figure 3b; discussed below). However, seeds of ole1(OLE1) and ole2(OLE2), in which oleosin levels were restored, exhibited final germination rates of more than 90% with or without sucrose.

Given that ole1 ole2 had the lowest oleosin content among the mutants and ole1 ole3 had the second lowest content (Figure 1b), the defects in seed germination would appear to be related to the lower abundance of oleosins, and might be related to the extreme expansion of oil bodies (Figure 2g,h).

Freezing treatment abolishes seed germination in single mutants ole1, ole2 and ole3

Figure 4 shows seeds and/or seedlings of wild-type, ole1 and ole1 ole2 after 5 days of germination in the absence of sucrose. In agreement with the results in Figure 3(a), ole1 exhibited a nearly normal level of germination (Figure 4c), which contrasts with an exceptionally low level of germination in the double mutant ole1 ole2 (Figure 4e). Interestingly, we found that storing ole1 seeds at −30°C for 1 day (freezing treatment) reduced their germination rate by half (Figure 4d). The freezing treatment did not noticeably affect the germination of wild-type seeds, which germinated normally with or without freezing treatment (Figure 4a,b). None of the ole1 ole2 seeds that were exposed to freezing had germinated at 5 days (Figure 4f).

Figure 4.

 Freezing treatment reduces the germination rates of ole1 seeds.
(a, c, e) Seeds of wild-type (Col-0) (a), ole1 (c) and ole1 ole2 (e) were imbibed at 4°C for 3 days (stratification) before germination in the absence of sucrose. Photographs were taken at 5 days of germination.
(b, d, f) Seeds of wild-type (Col-0) (b), ole1 (d) and ole1 ole2 (f) were stored at −30°C for 1 day (freezing treatment) and then were treated with stratification before germination in the absence of sucrose. Photographs were taken at 5 days of germination.

To understand the effect of freezing treatment on seed germination of oleosin-deficient mutants in greater detail, we exposed seeds of the ole mutants and wild-type to freezing conditions and then measured germination rates. Seeds of ole4, ole1(OLE1) and ole2(OLE2) germinated normally, as did the wild-type (Figure 5). Freezing treatment did not cause any damage to seeds of ole4, ole1(OLE1), ole2(OLE2) or the wild-type. In contrast, all the seeds of the double mutants ole1 ole2 and ole1 ole3 failed to germinate after freezing treatment (Figure 5), although approximately 10% of ole1 ole2 seeds and 40–50% of ole1 ole3 seeds did finally germinate without freezing treatment (Figure 3). Freezing treatment caused some damage to the mutant seeds in which oleosin contents were very low (Figure 1b).

Freezing treatment delayed germination of seeds of the other mutants ole1, ole2, ole3 and ole2 ole3 (Figure 5a). The final germination rate of ole1 was approximately 70%, while the germination rates of ole2 and ole3 seeds recovered to the wild-type level (Figure 5a). The germination delay of the mutants increased in inverse proportion to the oleosin levels in the order ole3, ole2, ole2 ole3 and ole1 (Figure 1b), showing that freezing treatment arrests germination of seeds that lack oleosins. These results suggest that oleosins provide seeds with freezing tolerance.

The seeds were grown in the absence (Figure 5a) or presence (Figure 5b) of sucrose. However, addition of sucrose did not have a significant effect on germination rates. The freezing-induced defect in seed germination of the mutants was not eliminated by adding sucrose (Figure 5b; discussed below).

Stratification causes a freezing-induced defect in seed germination of ole1

We focused on the single mutant ole1 to study what occurs in oleosin-deficient mutant cells after freezing treatment. We stored ole1 seeds at −30°C for 3 h or 1 day before germination in the absence of sucrose. The 3 h freezing treatment also caused a delay of germination of ole1 seeds (Figure 6a), but less than the delay caused by the 1-day freezing treatment. The freezing time-dependent germination delay of ole1 seeds suggests that freezing treatment causes some damage to ole1 seed cells.

Figure 6.

 Freezing followed by stratification promotes reduction of the germination rate of ole1 seeds.
(a) ole1 seeds that had been stored at −30°C for 0 h, 3 h or 1 day (freezing treatment) were germinated in the absence of sucrose after stratification.
(b) ole1 seeds were treated with freezing and then subjected to one of the following four treatments before germination: no treatment (I), stratification (II), incubation at 4°C for 3 days in the dry state (III), and incubation at 4°C for 3 days in the dry state followed by stratification (IV).

Generally, seeds of Arabidopsis thaliana are imbibed at 4°C for 3 days (stratification) to cause them to germinate synchronously. To investigate the effect of stratification on the freezing-induced germination delay of seeds, we subjected ole1 seeds to freezing and then treated some of them with stratification before germination. We found that ole1 seeds germinated normally after freezing treatment alone (Figure 6b, I), but germinated poorly after freezing followed by stratification (Figure 6b, II). This suggested that stratification is a key step for freezing-induced germination delay in oleosin-deficient seeds.

Stratification causes imbibition of seeds at low temperature. To determine the effect of imbibition, we treated ole1 seeds by freezing and then kept them in the dry state at 4°C for 3 days instead of stratification. Interestingly, these ole1 seeds germinated normally (Figure 6b, III), as for ole1 seeds without stratification (Figure 6b, I). Finally, we kept freeze-treated ole1 seeds in the dry state at 4°C for 3 days followed by stratification. These ole1 seeds germinated poorly (Figure 6b, IV), as for ole1 seeds after freezing followed by stratification (Figure 6b, II). This suggests that imbibition at low temperature causes the freezing-induced germination delay of oleosin-deficient seeds.

Fusion of oil bodies in ole1 seeds is accelerated by imbibition at low temperature after freezing treatment

To determine morphological changes in oil bodies, we performed light microscopic observations of ole1 seeds with toluidine blue staining. No morphological changes were observed in oil bodies in ole1 seeds with no treatment (Figure 7a), or in ole1 seeds after stratification (Figure 7b). Unexpectedly, freezing itself caused no morphological changes in oil bodies of ole1 seeds (Figure 7c).

Figure 7.

 Stratification after freezing treatment promotes fusion of oil bodies in ole1 seeds.
(a) Light micrograph of ole1 dry seeds.
(b) Light micrograph of ole1 seeds after imbibition at 4°C for 3 days (stratification).
(c) Light micrograph of ole1 seeds after freezing at −30°C for 1 day (freezing).
(d, e) Light micrographs of ole1 seeds after freezing followed by stratification. Half of the ole1 seeds had many fused oil bodies (d, asterisks indicate fused oil bodies), but the remaining seeds had mo morphological changes (e). OB, oil bodies. Scale bars = 10 μm.

On the other hand, freezing followed by stratification resulted in significant changes of oil bodies in ole1 seeds (Figure 7d). The changes included many fused oil bodies as indicated by asterisks in Figure 7(d). These fused oil bodies were irregularly shaped, in contrast to spherical oil bodies in ole1 seeds with no treatment (Figure 7a).

For statistic analysis, we observed 400 cells of five seed grains of ole1 for each treatment to determine the proportion of cells with fused oil bodies, and found that 15.8 ± 4.83% of ole1 seeds that had been treated with freezing followed by stratification contained fused oil bodies. ole1 seeds subjected to other treatments exhibited lower rates: 0.75 ± 0.50% for seeds with no treatment, 4.50 ± 2.19% for seeds with stratification, and 0.50 ± 0.31% for seeds with freezing only.

This indicates that oil bodies fuse to each other during imbibition at 4°C for 3 days. As incubating freezing-treatedole1 seeds at 4°C in the dry state had no effect on germination rates (Figure 6b, III), imbibition at low temperature after freezing might cause the fusion of oil bodies. The results of Figures 6 and 7 suggest that freezing treatment caused undetectable damage to the oleosin-deficient membrane of oil bodies and the subsequent imbibition at low temperature accelerated the fusion of oil bodies.

Freezing followed by imbibition at low temperature damages nuclei in ole1 seeds, which causes seed mortality

Why do seeds of oleosin-deficient mutants lose their ability to germinate? We investigated the morphology and subcellular location of nuclei in DAPI-stained seed cells (Figure 8). All nuclei of seeds of wild-type were spherical and located in the center of the cells (Figure 8a). The shape and location did not change either as a result of stratification (imbibition at 4°C for 3 days; Figure 8b) or freezing followed by stratification (Figure 8c). All seeds, with or without freezing treatment, germinated and grew normally (Figures 3 and 5).

Figure 8.

 Oleosin deficiency causes the abnormal-positioning and deformation of nuclei and degeneration of seed cells.
Seeds of wild-type (Col-0, a–c), ole1 ole2 (d–h) and ole1 (i–l) were untreated (a, d, i), or treated with stratification (b, e, f, j) or freezing followed by stratification (c, g, h, k, l). The seeds were stained with DAPI before and after incubation for 5 or 11 days at 22°C on medium without sucrose. The fluorescent images and DIC images were merged. Scale bars = 5 μm.

In contrast all of the ole1 ole2 seeds, which hardly germinated at all with or without freezing treatment (Figures 3 and 5), had irregularly shaped nuclei on the periphery of the seed cells (Figure 8d). The abnormalities in the nuclei were not changed either by stratification (Figure 8e) or by freezing followed by stratification (Figure 8g). We observed more than 1000 seed cells of each of wild-type, ole1 and ole1 ole2 for each treatment. The nuclei disappeared in 50% of seeds after 5 days at 22°C (Figure 8f,h). This result suggests that nuclei of ole1 ole2 seed cells degenerated during incubation at 22°C.

All of the nuclei of ole1 seeds were located in the center of the cells, as were those of wild-type seeds (Figure 8i). Stratification did not change this feature of the nuclei (Figure 8j). However, 40% of ole1 seeds had irregularly shaped nuclei on the periphery of the cells after treatment with freezing followed by stratification (Figure 8k), as had ole1 ole2 seeds (Figure 8g). After 11 days, approximately 30% of ole1 seeds, none of which germinated (Figure 5), had damaged nuclei on the periphery and/or no nuclei in the cells (Figure 8l). These results suggest that stratification after freezing causes eccentric damaged nuclei in oleosin-deficient mutants, which causes seed mortality.

Discussion

Oleosins might confer freezing tolerance to oilseeds, which is required for over-wintering

Freezing injury results from multiple stresses, including cold stress, mechanical stress and severe dehydration stress. Most of the studies on freezing tolerance have examined vegetative tissues and focused on cold-responsive transcriptional cascades for environmental stress responses (for review, see Zhu et al., 2007). In contrast, the mechanisms of freezing tolerance are poorly understood. There is evidence that cold acclimatization is associated with changes in membrane composition, synthesis of proteins with cryoprotective properties, and increases in osmoprotectants such as proline and sugars, all of which are likely to contribute to an increase in freezing tolerance (Thomashow, 1999). However, an understanding of how these changes are integrated to confer freezing tolerance in vegetative tissues and their relative importance is not clear. Moreover, the mechanisms of freezing tolerance of non-vegetative tissue are even less well understood, including the freezing tolerance of seeds, the subject of this study.

It has been shown that oilseed crops with low water content are more resistant to freezing than those with high water content (Ellis et al., 1991; Roberts and Ellis, 1989). Arabidopsis seeds with low water content, which are resistant to freezing under normal conditions, lose the resistance under high humidity conditions (Vernon et al., 1999). Our findings indicate that loss of oleosins decreases resistance to freezing, especially during imbibition at low temperature. These data suggest that freezing tolerance of oilseeds is closely related to both water content and oleosin content in the seeds.

Our findings show that oleosin deficiency leads to loss of freezing tolerance of Arabidopsis seeds, indicating that oleosins are necessary not only for seed germination but also for freezing tolerance of seeds (Figures 3–5). The results suggest that abnormal fusion of oil bodies to form expanded oil structures is dangerous for oilseeds. Oil bodies must be densely surrounded with oleosins to prevent fusion of oil bodies during imbibition at low temperatures after freezing. Consequently, seeds of autumn-flowering oilseed plants might require large amounts of oleosins to over-winter and then to safely imbibe in the spring. Here, we propose oleosin as a key protein conferring freezing tolerance to Arabidopsis seeds.

Oleosins prevent abnormal fusion of oil bodies during imbibition at low temperature, increasing the viability of seed cells

Even ole1 ole2 seeds which have irregularly expanded oil-containing structures and deformed nuclei, synthesize storage oils and storage proteins during seed maturation. This means that oleosin deficiency does not affect seed viability during seed maturation. Stratification (imbibition at low temperature) after the freezing treatment caused the oil bodies of ole1 seeds to fuse, resulting in formation of expanded oil-containing structures (Figure 7d) and deformation and disappearance of nuclei in seed cells (Figure 8k).

The results with single and double oleosin-deficient mutants show that irregularly expanded oil-containing structures push the nuclei to the periphery of the cells, which might prevent nuclei from playing their roles. It is possible that the suppression of nuclear functions leads to degeneration and death of seed cells. These results suggest that oleosins are essential for maintaining seed viability during imbibition after freezing.

The ratio of oil to oleosin is a key factor for oil-body fusions (Ting et al., 1996). Siloto et al. (2006) reported that oil contents in seeds of both ole1 and ole2 were lower than those of wild-type (reduced by 16 and 6%, respectively). We also found that oil contents of ole1 ole2 seeds were much lower than those of wild-type (reduced by 33%). Oleosin deficiencies reduce oil contents of seeds in parallel with the low levels of oleosins, and this may be a means by which oleosin-deficient seed cells might retain their viability. These results suggest that the physiological significance of oleosins in oilseeds is to keep the oil-body size small and prevent severe intracellular changes from the desiccation stage to the imbibition stage.

Freezing treatment causes mechanical stress on oil-body membranes of oleosin-deficient mutants

How do oil bodies that lack oleosins fuse during imbibition after freezing treatment? One possibility is that fusion is due to mechanical stress on the oil bodies resulting from rapid influx of water into the cells. To test this hypothesis, we mechanically stressed the ole1 seeds by shaking them at 25 Hz for 30 min and then determined their germination rate after stratification. The treated ole1 seeds showed delayed germination, although both untreated ole1 seeds and the shaken seeds of wild-type germinated normally (Figure S3). Mechanical stress by shaking treatment followed by stratification leads to a defect in seed germination that mimics the defect caused by freezing treatment followed by stratification.

A combination of freezing and stratification is necessary to delay seed germination of ole1 seeds (Figure 6). Light microscopy showed that just freezing treatment did not alter any intracellular structures in ole1 seeds (Figure 7). However, freezing treatment might damage the oil bodies of ole1 seeds. The damage might accelerate membrane fusion of oil bodies during imbibition at 4°C for 3 days. Imbibition at 22°C instead of 4°C promotes degradation of lipids in oil bodies, resulting in a reduction of the volume of oil bodies and abolishment of their fusion, allowing normal germination of the ole1 seeds (Figure 6b, I). Our results suggest that oleosin is required to prevent abnormal fusion of oil bodies during imbibition at low temperature, which might cause mechanical stress on oil bodies.

Sucrose rescues oleosin-deficient seeds from death at the seedling stage, but does not prevent delay in seed germination

Siloto et al. (2006) reported that a 1-day delay of germination of OLEO1/OLE1 knockdown seeds in the absence of sucrose was rescued by addition of sucrose as an energy source. However, Figures 3 and 5 show that the defect in seed germination of oleosin-deficient mutants is not caused by depletion of sucrose. However, addition of sucrose enhanced the seedling growth of ole1 seeds that had been treated by freezing followed by stratification (Figure S4). After 11 days of germination in the absence of sucrose, approximately 40% of the seedlings survived, approximately 35% died at the seedling stage, and the rest failed to germinate (Figure S4). Interestingly, approximately 70% of the seedlings survived after addition of sucrose to the medium (Figure S4). This result indicates that seed germination is independent of an energy source, while seedling growth depends on it.

It has been shown that hormone-signaling pathways comprising phospholipase D (PLD) and hormones (Li et al., 2007) or protein phosphatase 2A catalytic subunit (PP2Ac-2) and ABA (Pernas et al., 2007) regulate gene expression for seed germination. It is possible that gene expression is inhibited by damage to the nuclei after freezing and stratification. On the other hand, depletion of energy sources might cause a defect in the seedling growth of ole1, in which the abnormally enlarged oil bodies have difficulty in degrading lipids by hydrolytic enzymes. This idea is supported by the finding that sdp1, which lacks a lipase for degradation of storage lipids, exhibits a similar phenotype: seeds of sdp1 die at the seedling stage in the absence of sucrose and grow in the presence of sucrose (Eastmond, 2006).

Experimental procedures

Plant materials and growth conditions

Seeds of Arabidopsis thaliana, ecotype Columbia (Col-0) and oleosin-deficient mutants were surface-sterilized with 70% ethanol and then sown onto 0.8% w/v agar in Murashige–Skoog (MS) medium (Wako, http://www.wako-chem.co.jp) with or without 1% w/v sucrose. The seeds were incubated at 4°C for 3 days to break seed dormancy (stratification) and then were germinated at 22°C under continuous light (100 mE m−2 sec−1). The seedlings were transferred onto vermiculite for subsequent growth.

Oleosin mutants

Information about the T-DNA insertion mutants of Arabidopsis thaliana (Col-0) was obtained from the Salk Institute Genomic Analysis Laboratory website (http://signal.salk.edu). The seeds of ole1 (SM_3_29875) and ole4 (SM_3_20767) were provided by the Nottingham Arabidopsis Stock Centre (NASC). The seeds of ole2 (SALK_072403) and ole3 (SALK_011533) were provided by Arabidopsis Biological Resource Center (ABRC) at Ohio State University. We generated three double mutants (ole1 ole2, ole1 ole3 and ole2 ole3) by crossing between single mutants.

Plasmid construction and transformation

We used Gateway Technology (Invitrogen; http://www.invitrogen.com/) to generate plasmids. The OLE1 and OLE2 genes were amplified by PCR from Arabidopsis genome DNA using the primer pairs 5′-CACCGCGCTCTTGCTCGTCTGTGTAGTAGT-3′ and 5′-GTGGGTAGAGAAAGGGATAGCACCTGCT-3′ for OLE1 and 5′-CACCTTTCCAGTGAAGGAAATTTCT-3′ and 5′-GAAATGCCAATGGAGCCTAGTTCT-3′ for OLE2. The amplified fragments were inserted into pENTER/D-TOPO (Invitrogen, http://www.invitrogen.com) to produce two entry clones. We performed an LR recombination reaction to transfer the OLE1 and OLE2 fragments of entry clones into the destination vector pHGW (Plant System Biology, http://www.psb.ugent.be, to create expression clones pHGW-OLE1 (OLE1 promoter::OLE1:OLE1 terminator) and pHGW-OLE2 (OLE2 promoter::OLE2:OLE2 terminator). The expression vectors pHGW-OLE1 and pHGW-OLE2 were introduced into Agrobacterium tumefaciens (strain GV3101) by electroporation. Arabidopsis ole1 and ole2 plants were transformed with Agrobacterium tumefaciens containing pHGW-OLE1 and pHGW-OLE2, respectively. T1 seeds were surface-sterilized and sown onto 0.8% agar in MS medium with 1% sucrose and hygromycin B (30 mg/l; Wako) for selection. We selected T2 transformants expressing high levels of OLE1 or OLE2 to establish T3 homozygous lines.

Specific antibodies

Two peptides derived from OLE1 (CKYATGEHPQGSDKLDS) and OLE2 (CHRVDRTDRHFQFQS) were chemically synthesized using a peptide synthesizer (model 431A; Applied Biosystems; http://www.appliedbiosystems.com/). These peptides were cross-linked to BSA using 3-maleimidobenzoic acid N-hydroxysuccinimide ester (Sigma-Aldrich; http://www.sigmaaldrich.com/). The peptide–BSA conjugates were injected into a rabbit subcutaneously with complete Freund’s adjuvant. After 3 weeks, four booster injections with incomplete adjuvant were given at 1-week intervals. One week after the booster injections, blood was drawn and the antibodies were prepared.

Purification of seed oil bodies

For isolation of oil bodies, we used dry seeds of wild-type (Col-0) and oleosin-deficient mutants (ole1, ole2, ole3, ole4, ole1 ole2, ole1 ole3 and ole2 ole3). We homogenized the seeds using a mixer mill (MM300; Resch GmbH, http://www.goliath.ecnext.com) at 25 Hz for 5 min. The homogenates were centrifuged at 10 000 g at 4°C for 30 min to obtain top fractions containing crude oil bodies. The subsequent purification procedures were essentially the same as described previously (Tzen et al., 1997), except that the hexane treatment of the final preparation was omitted. We measured the oil contents of the isolated oil bodies as described below. The oil bodies with 350 μg of oils were suspended in a 7× volume of SDS sample buffer (75 mm Tris/HCl, pH 6.8, 7.5% w/v SDS, 15% v/v glycerol, 7.5% v/v 2-mercaptoethanol), and subjected to SDS–PAGE on 7.5–15% acrylamide gradient gels (BIO CRAFT, http://www.bio-craft.co.jp), followed by staining with Coomassie blue.

Immunoblotting

Dry seeds of wild-type (Col-0), the oleosin-deficient mutants (ole1, ole2, ole3, ole4, ole1 ole2, ole1 ole3and ole2 ole3), ole1(OLE1) and ole2(OLE2) were homogenized in SDS sample buffer (100 mm Tris/HCl, pH 6.8, 4% w/v SDS, 20% v/v glycerol, 10% v/v 2-mercaptoethnol). The homogenates from two seed grains were subjected to SDS–PAGE on 15% acrylamide gels and then either to Coomassie blue staining (Figure S2) or immunoblotting. For immunoblotting, the separated proteins on the gels were transferred electrophoretically to polyvinylidene difluoride membranes (Immobilon-P; Millipore, http://www.millipore.com/index.do). The membranes were treated in blocking solution (5% skim milk, 100 mm Tris/HCl, pH 7.5, 0.05% v/v Tween-20), and then incubated with antibodies for 1 h. The dilutions of antibodies were 1:2000 for anti-OLE1 antibody and 1:5000 for anti-OLE2 antibody. Horseradish peroxidase-conjugated goat antibodies against rabbit IgG (Pierce, http://www.piercenet.com) that were diluted 1:2000 were used as the secondary antibody. Immunodetection was performed using an enhance chemiluminescence kit (Amersham Biosciences; http://www5.amershambiosciences.com/).

Electron microscopy and light microscopy

Dry seeds of wild-type (Col-0), the oleosin-deficient mutants (ole1, ole2, ole3, ole4, ole1 ole2, ole1 ole3and ole2 ole3), ole1(OLE1) and ole2(OLE2) were cut into slices with a razor blade and the slices were fixed with 4% w/v paraformaldehyde, 10% dimethylsulfoxide and 1% v/v glutaraldehyde in 0.05 m cacodylate buffer, pH 7.4, for 2 h. Similarly, imbibed ole1 seeds were fixed with 4% w/v paraformaldehyde, 10% dimethylsulfoxide and 5% v/v glutaraldehyde in 0.05 m cacodylate buffer, pH 7.4 for 2 h. The fixed slices were subjected to post-fixation with 2% OsO4 for 2 h. The following procedures were essentially the same as those described previously (Hara-Nishimura et al., 1993). The samples were dehydrated in a graded ethanol series at room temperature and embedded in Epon 812 resin embedding kit (TAAB Laboratories Equipment, http://www.taab.co.uk) followed by polymerization at 60°C for 3 days. Thin sections that were stained with toluidine blue were examined with a light microscope. Ultra-thin sections were examined with a transmission electron microscope (model JEM-1015B; JEOL, http://www.jeol.co.jp).

Quantitative analysis of oil-body sizes

We statistically analyzed oil-body sizes using electron micrographs of dry seeds of wild-type (Col-0), the oleosin-deficient mutants (ole1, ole2, ole3, ole4, ole1 ole2, ole1 ole3 and ole2 ole3), ole1(OLE1) and ole2(OLE2). We measured cross-sections of oil bodies in each cell of three seeds of each genotype using Photoshop Elements 5.0 (Adobe Systems Incorporated, http://www.adobe.com). The sizes of the cross-sections of oil bodies were classified into five groups: <1, 1–2, 2–4, 4–10 and >10 μm2.

Oil content

Dry seeds (15 grains) of wild-type (Col-0) and ole1 ole2 were boiled in 60 μl of water at 99°C for 10 min to deactivate hydrolytic enzymes and were then homogenized on ice. We measured the oil content of the homogenates (40 μl) using the triglyceride E-test (Wako), and determined the oil content based on the fresh weight of the seeds.

Freezing treatment

Dry seeds were treated by freezing at −30°C for 1 day. After surface sterilization with 70% ethanol, the seeds were sown onto 0.8% agar in MS medium with or without 1% sucrose, stratified at 4°C for 3 days, and germinated at 22°C under continuous light (100 mE m−2 sec−1). The germination rate was measured over 3, 5, 7, 9 and 11 days.

DAPI staining of seeds

The procedures for DAPI staining of seeds were essentially the same as previously reported (Tamura et al., 2005). After removal of seed coats in 100% glycerol, the seeds were vacuum-infiltrated for 5 min with 2.5 μg ml−1 DAPI (4′,6-diamino-2–phenylindole) in 3.7% v/v formaldehyde and 0.1% Triton X-100, and were then incubated for 30 min. Fluorescent images and differential interference contrast (DIC) images were inspected with a confocal laser scanning microscope (LSM510 META; Carl Zeiss; http://www.zeiss.com/) using the 405-nm-line of a laser diode 405.

Mass spectrometry

Each band on the SDS–PAGE gel of the isolated oil bodies was treated as described previously (Li et al., 2006) and subjected to peptide mass fingerprinting analysis by MALDI-TOF (REFLEX III; Bruker Daltonics, http://www.bdal.de). Proteins were identified by a search of the National Center for Biotechnology Information database using Mascot (http://www.matrixscience.com/cgi/search).

Acknowledgements

We are grateful to Dr Katsuya Okawa (Kyoto University, Japan) for his help with mass spectrography. We are also grateful to the Arabidopsis Biological Recourse Center for providing the seeds of the T-DNA insertion mutants SALK_072403 and SALK_011533 and the Nottingham Arabidopsis Stock Centre for providing the seeds of T-DNA insertion mutants SM_3_29875 and SM_3_20767. This work was supported by Core Research for Evolutional Science and Technology of the Japan Science and Technology Corporation, by Grants-in-Aid for Scientific Research (nos: 16085203 and 17107002) from the Ministry of Education, Culture, Sports, Science and Technology (MEXT) of Japan, and by the Global Center of Excellence Program ‘Formation of a Strategic Base for Biodiversity and Evolutionary Research: from Genome to Ecosystem’ of MEXT.

Ancillary

Advertisement