Introduction

Starch, proteins and lipids which are stored in plant seeds are used as nutrient and energy source for human beings. Phytic acid, which is a storage compound of phosphorus, is known as a source of phosphorus pollution in water environment. To control their amount, it is informative to understand regulatory mechanisms of their synthesis.

Starch accumulates in the starchy endosperm from approximately 5 days after flowering (DAF) in rice (Jeon et al. 2010; Zhu et al. 2011). The weight of starch increases in parallel with increasing grain weight and eventually accounts for approximately 75 % of grain weight (Zhu et al. 2011). Seed storage proteins accumulate from 4 DAF and are classified into insoluble glutelin, alcohol-soluble prolamin, salt-soluble globulin and water-soluble albumin (Yamagata et al. 1982; Luthe 1983). The protein content of milled rice is approximately 8 %, and most of it is glutelin in the starchy endosperm (Villareal and Juliano 1978; Shewry and Halford 2002; Agboola et al. 2005). Seed storage proteins also accumulate in the embryo and aleurone layer, but these proteins are different from endosperm proteins (Cagampang et al. 1966). Lipids are stored from approximately 5 DAF as oil bodies consisting of triacylglycerols (TAGs) and surface proteins, oleosin and caleosin, in the embryo and aleurone layer (Choudhury and Juliano 1980; Chuang et al. 1996; Frandsen et al. 2001; Ichihara et al. 2003; Chen et al. 2012). Phytic acid, myo-inositol-1,2,3,4,5,6-hexakisphosphate, is a storage form of phosphorus and accumulates in the embryo and aleurone layer (O’Dell et al. 1972; Ogawa et al. 1979; Iwai et al. 2012).

A previous study using an antibody to abscisic acid (ABA) for inactivating it showed that the synthesis of seed storage proteins and lipids during seed development is affected by ABA activity (Phillips et al. 1997). Moreover, ABA is involved in conferring desiccation tolerance, dormancy and suppression of viviparous phenotype in developing seeds and stress response in vegetative tissues (Finkelstein et al. 2002; Gutierrez et al. 2007; Nakashima and Yamaguchi-Shinozaki 2013). Interactions between ABA and VP1/ABI3 or FUS3 induce the expression of seed storage proteins in Arabidopsis (Suzuki et al. 2003; Kagaya et al. 2005). Accumulation of starch during seed development positively correlates with accumulation of ABA (Yang et al. 2006; Tang et al. 2009). The expression of sucrose synthase, which is considered as a key enzyme of starch synthesis in developing seeds, is regulated by ABA and sucrose (Tang et al. 2009).

Starch synthesis was activated in suspension cells of rice cultured with ABA and sucrose (Akihiro et al. 2005, 2006). Rice cells cultured with ABA also activate phytic acid synthesis similar to that in developing seeds as the expression of genes involved in phytic acid synthesis in developing seeds is also up-regulated in cells cultured with ABA (Matsuno and Fujimura 2014). These results suggest that the suspension cells react with ABA in the same manner as developing seeds.

In the present study, starch synthesis in the suspension cells was investigated in detail to test this hypothesis and showed that starch content in suspension-cultured cells increased with increased concentrations of ABA. The expression analysis of genes involved in starch synthesis in suspension cells cultured with ABA indicated that the regulation of starch synthesis in the suspension cells was different from that in the endosperm. The expression analysis of the gene involved in synthesis of lipids and storage proteins showed that the former was activated in the suspension cells cultured with ABA without activation of the latter. Higher expression levels of master regulators of embryogenesis was observed, suggesting interactions between the master regulators and ABA-activated phytic acid and storage oil synthesis.

Materials and methods

Cultured cells

Rice (Oryza sativa L. ssp. japonica cv. Nipponbare) suspension cells were induced and cultured as described previously (Matsuno and Fujimura 2014). Rice seeds were sterilised by immersion in 70 % ethanol for 1 min and then in sodium hydrochloride (5 % available chloride) for 20 min. After five rinses with sterile deionised water, the seeds were placed on a solid N6 medium (Chu et al. 1975) supplemented with sucrose (30 g/L), proline (10 mM), casein hydrolysate (300 mg/L) and 2,4-dichlorophenoxyacetic acid (2 mg/L) and were incubated at 28 °C for 3 weeks. Friable calli formed were transferred to 15 mL of a liquid N6 medium in a 100-mL flask. The calli were cultured in the dark at 28 °C on a reciprocal shaker (115 strokes per min), and the resulting suspension-cultured cells were subcultured every 5 days.

Cells cultured were passed through a 1.5-mm mesh to remove larger cell masses and then cultured in a fresh medium for 2 days. They were then transferred to a fresh medium after another sieving through a 1.75-mm mesh and cultured for 24 h to minimise the transfer effect. ABA was then added to the medium, and the culture was harvested after various intervals by filtration under reduced pressure. The cells were desiccated at 80 °C to measure starch contents or frozen in liquid nitrogen and stored at −80 °C for gene expression analyses.

Plant materials

Rice plants were grown in the field at the Agricultural and Forestry Research Center (University of Tsukuba, Ibaraki, Japan) under natural environmental conditions from May to September, 2012. Seeds were harvested 7 DAF, immersed in RNAlater (Ambion®) and kept at 4 °C. Dehulled seeds were separated into three parts: embryo, starchy endosperm and aleurone layer. The embryo was first isolated with a pair of tweezers. The seed coat, its inner part rinsed with RNAlater, was harvested as the aleurone layer. The rinse containing the inner part of the seed coat was harvested as the starchy endosperm.

Another rice plant was cultivated hydroponically with Yoshida nutrient solution (Yoshida et al. 1976) in a growth chamber (16 h light at 27 °C). At the six-leaf stage, the fifth leaf was harvested and stored at −80 °C for gene expression analyses.

Quantitative and semi-quantitative RT-PCR

Total RNA was extracted from the suspension cells with the Isogen® reagent (Nippon Gene), from the embryo, the aleurone layer or fifth leaf with the RNeasy Plant Mini Kit (Qiagen) or from the endosperm with the Isogen® reagent and High salt precipitation solution (Nippon Gene). The RNAs were then reverse transcribed using the PrimeScript® RT Master Mix (Takara Bio) following the manufacturer’s instructions after treatment with DNase I.

Expression levels of genes (Tables 1, 2) were quantified by real-time PCR (RT-PCR) using the SYBR® Premix Ex Taq™ II (Takara Bio) and the Thermal Cycler Dice® Real-Time System (Takara Bio). The standard curve method or the 2−ΔΔCT method (Livak and Schmittgen 2001) was applied, and the rice actin 1 gene (OsACT1, AK100267) or ubiquitin 5 gene (OsUBQ5, AK061988) (Jain et al. 2006) was amplified with 5′-CCCAAGGCCAATCGTGAGAAG-3′ and 5′-ACCATCACCAGAGTCCAACACAA-3′ or 5′-AAGGAAGGAGGAGGAAATCG-3′ and 5′-GGGCATCACAATCTTCACAG-3′, respectively, and used as an internal control. Semi-quantitative RT-PCR was performed with Tks Gflex® DNA Polymerase (Takara Bio) for 35 cycles. OsACT1 and OsUBQ5 were used as internal controls.

Table 1 Primers for qPCR of genes involved in starch synthesis
Full size table
Table 2 Primers for PCR and qPCR
Full size table

Measurement of starch content

An aliquot of desiccated and ground cells was immersed in 80 % (v/v) ethanol at 82 °C for 5 min, and then the supernatant was discarded to remove soluble sugars. The treatment was repeated twice. The total starch content of the cells was determined with a total starch assay kit (Megazyme International Ireland Ltd.) by the KOH method following the manufacturer’s instructions.

Results

The accumulation of starch in the suspension cells cultured with ABA

The suspension cells cultured without ABA contained smaller amounts of starch. The starch content increased with increasing concentrations of ABA 24 h after ABA treatment (Fig. 1) by 1.4-, 2.7- and 2.9-fold at 0.17, 17 and 50 µM ABA, respectively, compared with that without ABA. Given that starch accumulation differed little between 17 and 50 µM ABA, 17 µM ABA was used for subsequent experiments.

Fig. 1
figure 1

Starch contents in rice cells cultured in a medium containing ABA. Cells were cultured with various concentrations of ABA (0–50 μM) for 24 h, and their starch content (mg/g dried cells) was measured. Values are means, and error bars represent standard deviations (n = 4)

Full size image

Starch contents in the suspension cells cultured with 17 µM ABA increased 3.1-fold from their initial value 24 h after adding ABA (Fig. 2a) and increased to 7.1-fold by 96 h (Fig. 2a) with a temporary decrease at 90 min (Fig. 2b).

Fig. 2
figure 2

Accumulation of starch in suspension-cultured cells after addition of ABA. Cells were cultured for 120 h (a) and 24 h (b) with 17 μM ABA (solid line) or without ABA (dotted line). The values are means, and error bars represent standard deviations; a n = 4, b n = 3

Full size image

Expression analysis of genes involved in starch synthesis after adding ABA

Expression levels of genes involved in starch synthesis (Table 1) reported by Ohdan et al. (2005) were evaluated in cells cultured with 17 µM ABA. OsGBSSI expression in the suspension cells was not detected by qPCR. Eighteen genes, OsAGPL1, OsAGPL3, OsAGPS1, OsAGPS2a, OsAGPS2b, OsISA2, OsISA3, OsPUL, OsGBSSII, OsBEIIa, OsBEIIb, OsDPE1, OsDPE2, OsSSI, OsSSIIa, OsSSIIb, OsSSIIc and OsSSIVa responded to the addition of ABA (Fig. 3). Ten of these, OsAGPL1, OsAGPL3, OsAGPS1, OsAGPS2a, OsISA2, OsISA3, OsBEIIa, OsDPE1, OsSSI and OsSSIIa, maintained higher expression levels than the control for 24 h after adding ABA, seven, OsABS2b, OsPUL, OsGBSSII, OsBEIIb, OsDPE2, OsSSIIc and OsSSIVa, increased temporarily and OsSSIIb decreased temporarily.

Fig. 3
figure 3figure 3

Expression analysis of genes involved in starch synthesis in suspension-cultured cells after addition of ABA. The expression levels of genes were evaluated by qRT-PCR in the suspension cells cultured with 17 μM ABA (represented by triangles and solid lines) or without ABA (represented by diamonds and dotted lines). The expression levels were calculated by the standard curve method. OsACT1 was used as an internal control, and the 0-h value was used as the reference value. Values shown are means and standard deviations of each qPCR (n = 3). Cultivation and sampling of cells were performed in two independent experiments for each treatment

Full size image

Comparison of the expression of genes involved in starch synthesis in rice suspension-cultured cells with ABA in the endosperm and in silico

Expression patterns of genes involved in starch synthesis were clustered based on the data set RXP_0001 (accession number GEO21396, Gene Expression Omnibus) and were analysed with RiceXPro (Sato et al. 2011, 2013) (NIAS, http://ricexpro.dna.affrc.go.jp/). OsSSIIIa, OsBEIIb, OsBEI, OsAGPS2a/b, OsSSIIa, OsAGPL2, OsGBSSI, OsPUL and OsISA1 constituted a cluster, and the expression of these genes increased rapidly at 5 DAF in the ovary and occurred predominantly in the endosperm (Fig. 4, clade I). The expression pattern was spatially and temporally consistent with starch synthesis in the rice endosperm. Expression levels of those genes were evaluated with ad hoc qRT-PCR and proved higher in the endosperm than the embryo, except for AGPS2a (Fig. 5). Thus, it was inferred that they may play roles in starch synthesis in the endosperm. In the rice suspension cells, 10 genes, OsAGPL1, OsAGPL3, OsAGPS1, OsAGPS2a, OsISA2, OsISA3, OsBEIIa, OsDPE1, OsSSI and OsSSIIa, increased for 24 h after adding ABA (Fig. 3), suggesting that these genes play roles in activated starch synthesis in cells. Only two of these genes, OsAGPS2a and OsSSIIa, were included in clade I and the other eight were excluded from it. These results showed that the set of genes involved in starch synthesis activated in the cells cultured with ABA was different from the set of genes involved in starch synthesis in the endosperm.

Fig. 4
figure 4

Expression levels of genes involved in starch synthesis. Twenty-eight genes were subjected to clustering by RiceXPro (http://ricexpro.dna.affrc.go.jp/) based on dataset RXP_0001 to generate a dendrogram (a). Given that OsAGPS2a and OsAGPS2b, transcripts of the locus Os08g0345800, were not distinguishable by probes used in the microarray analysis, they are described as OsAGPS2a/b. Genes in clusters labelled I, II, III and IV were expressed mainly in the ovary and endosperm, ovary and embryo, embryo and leaf in vegetative stage, respectively. Genes with plus sign (+) in column (b) showed abnormal features in the endosperm when they were mutated (Sano et al. 1986; Kubo et al. 1999; Nishi et al. 2001; Umemoto et al. 2002; Satoh et al. 2003; Lee et al. 2007). Genes with plus sign (+) in column (c) were considered to function in starch synthesis in the endosperm (Ohdan et al. 2005), and genes with plus sign (+) in column (d) were up-regulated in the cultured cells by ABA treatment for 24 h

Full size image
Fig. 5
figure 5

qRT-PCR analysis of the expression of genes involved in starch synthesis. Immature seeds at 7 DAF were harvested and dissected into three parts: embryo (black bars), aleurone layer (white bars) and endosperm (grey bars). RNAs extracted were analysed by qRT-PCR with primer sets for the same genes as in Fig. 4. The expression levels were calculated by the 2−ΔΔCt method with the references of the embryo tissue and the internal control of OsUBQ5. Values shown are means and standard deviations of each qPCR (n = 3)

Full size image

RPBF (Kawakatsu et al. 2009), OsbZIP58 (Wang et al. 2013) and RSR1 (Fu and Xue 2010) have been reported as regulators of starch synthesis in the endosperm. To investigate the effects of these regulators on the activation of starch synthesis in the suspension cells cultured with ABA, their expression levels were evaluated (Fig. 6a). OsbZIP58 was expressed at a low level in the cells, but no expression of RPBF was detected. OsbZIP58 and RPBF were accordingly considered to be of little relevance to the activation of starch synthesis in the suspension cells. RSR1, which is a negative regulator, was expressed to some degree in the suspension cells, but the level decreased after ABA addition (Fig. 6a). This result raised the possibility that RSR1 functions in the activation of starch synthesis in the suspension cells.

Fig. 6
figure 6

Expression levels of regulatory genes in three parts of immature seeds and the suspension cells. Seeds at 7 DAF were harvested and dissected into three parts: embryo, aleurone layer and endosperm. RNAs were extracted from them and from the suspension cells cultured with (+) or without (−) 17 μM ABA for 12 h. Expression levels of regulators of starch synthesis (a), seed storage proteins (b), seed oil bodies (c), master regulators (d) and internal control (e) were evaluated by RT-PCR. Numbers of PCR cycles were 26 for OsUBQ5, 28 for OsACT1 and 35 for others. Samples from seed parts were triplicated and from cells were duplicated

Full size image

Expression of genes involved in seed storage proteins and lipids

As starch and phytic acid were accumulated in the suspension cells cultured with ABA, similar to developing seeds (Akihiro et al. 2005; Matsuno and Fujimura 2014), synthesis of seed storage proteins and lipids in the suspension cells was also expected to be activated by addition of ABA. However, the expression of OsGlb-1 and OsGluB-1a, encoding globulin and glutelin, respectively, (Nakase et al. 1996; Kawakatsu et al. 2008) was not detected in the suspension cells with or without ABA (Fig. 6b). The expression of OsWRI1a/b, which is regarded as a regulator of fatty acid synthesis in rice, OsOLE16, OsOLE18 and OsCAL32, encoding 16 and 18 kDa oleosins composing oil bodies (Chuang et al. 1996) and encoding 32 kDa caleosin (Chen et al. 2012), respectively, were also analysed (Figs. 6c, 7). OsOLE16, OsOLE18 and OsCAL32 were expressed clearly in the suspension cells but weakly in leaves (Fig. 7). After ABA treatment, their expression levels were strongly up-regulated in the suspension cells, and the expression of OsWRI1a/b was up-regulated threefold compared with that of the control.

Fig. 7
figure 7

Expression levels of genes in suspension-cultured cells. RNAs were extracted from three parts of immature seeds at 7 DAF: embryo, aleurone layer and endosperm, suspension cells cultured with (+) or without (−) 17 μM ABA for 12 h and the fifth leaf of young rice plants. Expression levels were evaluated by qRT-PCRs using the 2−ΔΔCt method with the value of ABA (−) as reference. Values shown are means and standard deviations of each qPCR (n = 3)

Full size image

Master regulators of seed development in the rice suspension cells

In Arabidopsis, four master regulators of embryogenesis, LEC1, LEC2, FUS3 and ABI3, have been well investigated and are known to be involved in the synthesis of seed storage substances (Meinke et al. 1994; Gutierrez et al. 2007). To investigate the ability of master regulators to function in rice cells, expression levels of OsLEC1A (Xie et al. 2008), OsLEC1B (Xie et al. 2008), OsLFL1 (Peng et al. 2007, 2008) and OsVP1 (Hattori et al. 1994), homologues of Arabidopsis master regulators, were evaluated. The expression of OsLEC1A was not detected in the suspension cells (Fig. 6d). OsLFL1 and OsVP1 were expressed at levels comparable with those in the embryo at 7 DAF and were not altered by ABA addition in the suspension cells (Fig. 7). OsLEC1B was expressed at a higher level in the ABA-free medium than in the embryo, and its expression was reduced in the suspension cells by ABA addition (Fig. 7).

Discussion

In the study, starch synthesis in the suspension cells was activated by the addition of ABA as previously reported (Akihiro et al. 2005). A positive correlation between added ABA concentration and starch content was observed in the suspension cells (Fig. 1) as rice kernels matured (Yang et al. 2006; Tang et al. 2009). These results suggest that ABA participates in starch accumulation in both the endosperm and suspension cells.

Developing seeds begin starch accumulation rapidly from approximately 5 DAF in the endosperm and reach the maximum at approximately 21 DAF (Jeon et al. 2010; Zhu et al. 2011). OsAGPL2 (Lee et al. 2007), OsAGPS2b (Lee et al. 2007), OsSSIIa (Umemoto et al. 2002), OsBEI (Satoh et al. 2003), OsBEIIb (Nishi et al. 2001), OsISA1 (Kubo et al. 1999) and OsGBSSI (Sano et al. 1986), which are known to be involved in starch synthesis in the endosperm, constituted a gene cluster in the dendrogram (Fig. 4, cluster I). Genes in the cluster were expressed at higher levels in the endosperm than in the embryo, except for AGPS2a (Fig. 5), and these findings support the assumption that these genes function in starch synthesis in the endosperm (Ohdan et al. 2005). Thus, the genes in the cluster are considered to be also involved in starch synthesis in the endosperm. In the suspension cells cultured with ABA, OsAGPL1, OsAGPL3, OsAGPS1, OsAGPS2a, OsISA2, OsISA3, OsBEIIa, OsDPE1, OsSSI and OsSSIIa were expressed at higher levels than in the control during the 24 h after the addition of ABA (Fig. 3), suggesting that these genes contributed to activated starch synthesis in the suspension cells cultured with ABA. OsbZIP58, which is known as a regulator of starch synthesis in the endosperm, controls gene expression involved in starch synthesis in the endosperm, and starch in the endosperm of an Osbzip58 mutant is altered quantitatively and qualitatively (Wang et al. 2013). OsbZIP58 is expressed specifically in the endosperm at maximum levels at 5–10 DAF (Sato et al. 2011; Wang et al. 2013). In the suspension cells, the expression of OsbZIP58 was low and not altered by ABA treatment (Fig. 6a), suggesting that increased starch synthesis in the suspension cells was not regulated by OsbZIP58. Taking these results together, it is indicated that the regulatory mechanism of starch synthesis in the suspension cells cultured with ABA is different from that in the endosperm.

Most seed storage proteins are classified into glutelin and stored in the endosperm (Villareal and Juliano 1978; Yamagata et al. 1982). OsGlb-1, encoding 26 kDa α-globlin, and OsGluB-1a, encoding glutelin, are expressed specifically in the endosperm and are up-regulated by RPBF and OsbZIP58 synergistically (Wu et al. 1998; Yamamoto et al. 2006; Kawakatsu et al. 2008). The expression analysis showed that the expression levels of OsGluB-1a, OsGlb-1 and RPBF were undetected and that OsbZIP58 expression was detected at a low level in the suspension cells (Fig. 6a, b), suggesting that the seed storage protein was not synthesised in the suspension cells.

Seed storage lipids are stored in the embryo and aleurone layer as oil bodies composed of TAGs surrounded by a single layer of phospholipids and proteins, oleosins and caleosin (Frandsen et al. 2001; Chen et al. 2012). A transcription factor, AtWri1, activates fatty acid synthesis and thereby activates TAG synthesis (Focks and Benning 1998; Cernac and Benning 2004; To et al. 2012). Two homologues of this factor are found in both rice and maize (Pouvreau et al. 2011). As ZmWRI1 activates fatty acid synthesis in maize (Shen et al. 2010), OsWRI1a/b, orthologue of ZmWRI, is presumed to activate fatty acid synthesis in rice. Two isoforms of oleosin, 16 and 18 kDa oleosins, have been detected in the embryo and aleurone layer, and 32 kDa caleosin has been detected in the embryo in rice (Chuang et al. 1996; Qu and Takaiwa 2004; Chen et al. 2012). Expression levels of OsOLE18, OsOLE16 and OsCAL32 were up-regulated by ABA treatment at levels comparable to those in the embryo (Figs. 6c, 7). As the expression of OsWRI1a/b was also up-regulated by ABA treatment (Figs. 6c, 7), oil body synthesis must have been activated in the suspension cells cultured with ABA.

In the rice seed, the accumulation of glutelin is mostly observed in the endosperm along with starch accumulation, and the accumulation of seed storage lipids is observed in the embryo and aleurone layer along with phytic acid accumulation. The regulation of phytic acid synthesis in the suspension cells appeared similar to its regulation in developing seeds (Matsuno and Fujimura 2014). In contrast, with respect to starch synthesis, regulatory mechanisms seemed to be different in the suspension cells and in developing seeds as mentioned above. The combined results suggest that the suspension cells mimic embryo-specific expression and ABA response (Fig. 8). The suggestion would imply the possibility that the epigenetic regulation in the suspension cells was influenced by its origin since the suspension cells was derived from embryo.

Fig. 8
figure 8

Predicted regulation mechanisms for synthesis of storage substance in the immature rice seed and suspension cells. a Phytic acid and oil body synthesis in the rice embryo and suspension cells are presumed to be under the control of OsVP1 and OsLFL1. OsVP1 and OsLFL1 may interact with ABA signals and then regulate embryo-specific ABA responses including fatty acid synthesis and expression levels of oleosin and caleosin genes and phytic acid synthesis. b Starch and seed storage protein synthesis in the endosperm are under the control of RSR1, OsbZIP58 and RPBF. Solid lines represent demonstrated relationships (Suzuki et al. 2003; Yamamoto et al. 2006; Kawakatsu et al. 2008, 2009; Wang et al. 2013), and dotted lines have been assumed here

Full size image

In the study, expression of OsVP1, a homologue of AtABI3, was observed at a level comparable to that in the embryo at 7 DAF (Fig. 7), a finding also reported by Nakagawa et al. (1996). VP1/ABI3 regulates gene expression and induces seed-specific ABA response (Suzuki et al. 2003). The expression of oleosin gene in the suspension cells of Brassica napus was activated by ABA treatment and overexpression of ABI3 (Crowe et al. 2000). In the rice suspension cells, it is likely that highly expressed OsVP1 (Fig. 7) enables the expression of oleosin genes as a seed-specific ABA response (Fig. 8). OsLEC1A and OsLEC1B are homologues of AtLEC1 (Xie et al. 2008). Microarray data [RiceXPro (Sato et al. 2011)] have shown that OsLEC1A and OsLEC1B were expressed specifically in the endosperm and embryo, respectively. Our results also showed that OsLEC1B was expressed specifically in the embryo (Figs. 6d, 7). Given that OsLEC1B was expressed at high levels in the suspension cells but OsLEC1A was not, the suspension cells resemble the rice embryo in expression specificity (Figs. 6d, 7). OsLFL1 has high similarity to AtFUS3 and AtLEC2 and is expressed in the embryo, callus, anther and pollen (Peng et al. 2007, 2008). Given that AtLEC2 and AtFUS3 regulates fatty acid metabolism via AtWri1 (Baud et al. 2007; Wang and Perry 2013), OsLFL1 also appears to regulate fatty acid metabolism through OsWri1a/b. The interaction of AtFUS3 with ABA regulates the expression of seed-specific or seed-preferential genes (Yamamoto et al. 2010). In addition to the interaction of OsVP1 and ABA, the interaction of OsLFL1 with ABA may enable the suspension cells to display a seed-specific ABA response (Fig. 8).

In the rice suspension cells, starch synthesis was activated by ABA treatment via regulatory mechanisms different from those in the endosperm. The synthesis of oil bodies was suggested to be activated by ABA treatment, but that of seed storage protein was not. The synthesis of phytic acid in the suspension cells has been reported to be activated by ABA (Matsuno and Fujimura 2014). OsVP1 and OsLFL1 were expressed at levels comparable to those in embryo at 7 DAF, and it was accordingly inferred that the suspension cells mimic embryos rather than the endosperm or aleurone layer in ABA response, including the synthesis of phytic acid and oil bodies. In developing seeds, a study focusing on the relationship between master regulators and synthesis of phytic acid or oil bodies is difficult because master regulators affect many aspects of early seed development including the process of embryogenesis, which is accompanied by morphological changes. Further studies using the suspension cells will clarify the regulatory mechanism of OsVP1 or OsLFL1 in the synthesis of phytic acid or oil bodies.