Gene Expression of ADP-glucose Pyrophosphorylase and Starch Contents in Rice Cultured Cells are Cooperatively Regulated by Sucrose and ABA

Abstract

Six cDNA clones encoding two small subunits and four large subunits of ADP-glucose pyrophosphorylase (AGPase) were mined from the database of rice full-length cDNAs, cloned and subsequently named: OsAPS1, OsAPS2, OsAPL1, OsAPL2, OsAPL3 and OsAPL4. Expression patterns of the six genes were examined by Northern blot analysis with gene-specific probes. OsAPL3 was predominantly expressed in the middle phases of seed development, and OsAPS1, OsAPL1 and OsAPL2 were expressed later in seed development. OsAPS2 and OsAPL4 were constitutively expressed and these isoforms were coordinated with starch accumulation in the developing rice seed. In order to clarify the effect of sugars and plant hormones on AGPase gene expression more precisely, a rice cell culture system was used. OsAPL3 transcript significantly accumulated in response to increased levels of sucrose and abscisic acid (ABA) concentration in the medium; however, the transcripts of other AGPase genes did not show significant accumulation. Under identical conditions, starch contents in the cultured cells also increased. Interestingly, ABA alone did not affect the gene expression of OsAPL3 and starch content. Collectively, these results indicated that the expression level of OsAPL3 and starch content in the cultured cells were cooperatively controlled by alterations in the concentration of both sucrose and ABA.

Introduction

ADP-glucose pyrophosphorylase (AGPase; EC 2.7.7.27) catalyzes one of the main regulatory steps of starch biosynthesis in plants (Ghosh and Preiss 1966, Preiss 1984a). Specifically, AGPase catalyzes the conversion of glucose (Glc)-1-phosphate and ATP to ADP-glucose and pyrophosphate. Bacterial AGPase is formed as a homotetramer and is composed of one subunit that is encoded by a single gene (Preiss 1984b). In contrast, plant AGPase is a heterotetrameric enzyme that is composed of two small subunits and two large subunits. Unlike the bacterial AGPase, the plant ortholog is encoded by different genes. Three mechanisms are known to regulate AGPase activity: (i) transcriptional regulation (Müller-Röber et al. 1990, Scheible et al. 1997, Nielsen et al. 1998, Sokolov et al. 1998, Miyazawa et al. 1999, Harn et al. 2000, Fritzius et al. 2001, Rook et al. 2001, Li et al. 2002); (ii) allosteric regulation, with glycerate-3-phosphate (activator) and inorganic phosphate (inhibitor) (Sowokinos and Preiss 1982); and (iii) post-translational redox modification in response to sugars (Tiessen et al. 2002, Hendriks et al. 2003). Despite the fact that the relationship of post-transcriptional regulation of AGPase enzyme activities and its role in the control of starch biosynthesis is well established, little is known about transcriptional regulation of AGPase. Progress in the studies of transcriptional regulation of AGPase has been hampered by the lack of knowledge about the exact number of highly similar AGPase isoforms which exist in plants and by difficulties in the design of isoform-specific probes for Northern blot analysis.

AGPase genes are known to accumulate transcript in response to sugars (Müller-Röber et al. 1990, Sokolov et al. 1998, Harn et al. 2000, Fritzius et al. 2001, Li et al. 2002) and plant hormones (Miyazawa et al. 1999, Rook et al. 2001) and to decrease transcript in response to nitrate (Harn et al. 2000) and phosphate exposure (Nielsen et al. 1998). Although the aforementioned AGPase gene response patterns have been well characterized, the precise regulation of each individual gene is not known.

Rice is perhaps one of the most important food crops in the world. For this reason, a focused research effort was initiated through multiple sequencing projects to provide genomic and full-length cDNA sequences. Due to the wealth of genomic resources, rice is a suitable plant model system that can be used for understanding the transcriptional regulation of individual AGPase isoforms in various plant tissues.

In this study, we data-mined and isolated six AGPase cDNAs from rice cultured cells and characterized the expression patterns for all isoforms. All genes were expressed in developing seeds, but each gene was differentially regulated in various tissues. Based on the expression patterns in the developing seeds and putative subcellular localizations as predicted by alignment analysis and by the programs, it is possible that OsAPL2 and OsAPL3 are major forms in the middle and late phases of seed development, respectively. Transcripts of OsAPL3 in rice cultured cells were synergistically regulated by increased concentrations of sucrose (Suc) and abscisic acid (ABA) in the medium and they were directly correlated with starch quantity.

Results

Data-mining, isolation and characterization of rice AGPase genes

Since the genome sequence and full-length cDNA project of Oryza sativa ssp. japonica cv. Nipponbare are almost finished, we selected this cultivar for data-mining. Sequence information of two small subunits (AK073146 and AK103906) and four large subunits (AK069296, AK071497, AK100910 and AK121036) of AGPase cDNAs was obtained from the database of rice full-length cDNA clones (http://cdna01.dna.affrc.go.jp/cDNA/) (Kikuchi et al. 2003) (Table 1). All clones were successfully amplified by reverse transcription–polymerase chain reaction (RT–PCR) using specific primer pair sets and cDNA from RNA of cultured cells. AK073146 and AK103906 were named OsAPS1 and OsAPS2, respectively (Table 1), and AK069296, AK071497, AK100910 and AK121036 were named OsAPL1, OsAPL2, OsAPL3 and OsAPL4, respectively (Table 1).

The chromosome localization of the six rice AGPases was investigated by using rice genome sequences (http://www.ncbi.nlm.nih.gov). OsAPS1, OsAPS2, OsAPL1, OsAPL2, OsAPL3 and OsAPL4 are encoded by independent genes and are localized on chromosome 9, 8, 3, 1, 5 and 7, respectively (Table 1). Because there is only one P1-derived artificial chromosome (PAC) contig which corresponds to each gene in the rice genome database (http://cdna01.dna.affrc.go.jp/PIPE/) (Table 1), it was concluded that each gene exists as a single copy. The six clones contained putative domains that are known to exist in AGPases. Specifically, ATP-binding site (Frueauf et al. 2001, Frueauf et al. 2003), catalytic, Glc-1-phosphate and activator site (Sivak and Preiss 1998) are conserved in all six rice AGPases (Fig. 1).

OsAPS1 and OsAPS2 differ only in one and three amino acid residues from pVK1 that was previously isolated from ssp. japonica cv. Kinmaze (AY028315; Sikka et al. 2001) and from RSc6 which was isolated from ssp. japonica cv. Biggs M201 (J04960; Anderson et al. 1989), respectively. Since OsAPS1 and OsAPS2 were isolated from ssp. japonica cv. Nipponbare, it was estimated that OsAPS1 and OsAPS2 are alleles of pVK1 and RSc6, respectively.

Comparison of six rice AGPases with known extra-plastidial form AGPases

A comparison of the deduced amino acid sequences of OsAPS1 and OsAPS2 proteins showed a high homology, but their N-terminal regions showed no significant homology (Fig. 1A). The N-terminal amino acid sequence of OsAPS1 shows significant homology with the N-terminal amino acid sequence (VSDSQNNSDQ) of the spinach leaf small subunit as determined by Edman degradation (Morell et al. 1987). In contrast, no obvious sequence homology was observed between the N-terminal region of OsAPS2 and the spinach leaf small subunit (Fig. 1A). Maize BT2 and wheat AGPase small subunit AGP.S.1aS (wheat 1aS) are extra-plastidial AGPase small subunits (Choi et al. 2001, Burton et al. 2002). The N-terminal amino acid sequence of OsAPS2 shows significant homology with wheat 1aS (Fig. 1A).

A comparison of the deduced amino acid sequences of OsAPL1, OsAPL2, OsAPL3 and OsAPL4 proteins showed a high homology except their N-terminal regions (Fig. 1B). The N-terminal amino acid sequence of OsAPL1 shows significant homology with the barley leaf type AGPase large subunit bpl14 (Eimert et al. 1997) (data not shown). Both N-terminal amino acid sequences contain VAAA which is a consensus motif for proteolytic processing of proteins that are post-translationally targeted to the plastid compartment (Bairoch 1992, Eimert et al. 1997). OsAPL2 and OsAPL3 show significant homology with the cytosolic wheat AGPase large subunit APL.1 (CAA79980; Ainsworth et al. 1993, Burton et al. 2002). OsAPL2 and OsAPL3 may have no plastid-targeting signal sequence which is predicted by PSORT (http://psort.nibb.ac.jp/), ChloroP (http://www.cbs.dtu.dk/services/ChloroP/) and iPSORT (http://hc.ims.u-tokyo.ac.jp/iPSORT/) programs. OsAPL3 was predicted to have a plastid target signal sequence by TargetP (http://www.cbs.dtu.dk/services/TargetP/); however, reliability is very low (<0.2). OsAPL4 has a plastid target signal sequence as predicted by the PSORT, ChloP, iPSORT and TargetP programs. Collectively, it is estimated that OsAPS1, OsAPL1 and OsAPL4 are plastidial forms and OsAPS2, OsAPL2 and OsAPL3 are extra-plastidial forms of AGPase.

Expression properties of AGPase genes

The expression properties of the six rice AGPase genes in developing seeds, leaves, stems and roots were analyzed with Northern blot analysis and isoform-specific probes. In developing seeds, the six AGPase genes were all expressed; however, they exhibited differential patterns of gene regulation (Fig. 2). Specifically, OsAPS1 and OsAPL2 are expressed during 10–20 days after flowering (DAF). OsAPS2, OsAPL1, OsAPL3 and OsAPL4 genes were expressed from the early to late phase of seed development (Fig. 2). Gene expression of OsAPS2 and OsAPL1 reached a maximum at 10–12 DAF and then gradually decreased. The OsAPL3 gene expression reached a maximum level at 5 DAF and decreased gradually thereafter. The expression level of the OsAPL4 gene was lower than that of other AGPase isoforms in the developing seeds. Expression patterns of SSII-3, RBE1 and OsGBSSI were also analyzed as a positive control for seeds. The individual expression patterns of controls in the developing seeds were unique and they differed from the six AGPase genes.

The expression properties of the six rice AGPase genes in various tissues from rice were analyzed with Northern blot analysis. The six AGPase genes were expressed specifically in each tissue (Fig. 2, 3). OsAPL1 and OsAPL4 were expressed in seeds, leaves and stems but not in roots. OsAPL2 was expressed only in developing seeds. OsAPS1 and OsAPS2 were expressed in both seeds and leaves (Fig. 2, 3). OsAPL3 was expressed in both seeds and stems (Fig. 2, 3).

Diurnal effects were also studied in order to determine if they influenced AGPase gene expression. OsAPS1, OsAPS2 and OsAPL3 transcripts were higher in the light periods than in the dark periods (Fig. 3A, B). However, such changes in relation to day length were not observed for OsAPL1 and OsAPL4.

OsAPL3 expression is induced by sugars and ABA

The effects of sugars on the expression levels of AGPase genes were evaluated by using cultured rice cells as a model system (Yu et al. 1991). Rice cultured cells were grown in modified medium containing 0–10% (w/v) Suc for 24 h. Under these conditions, transcripts of OsAPL1 and OsAPL2 were not detected with Northern blot analysis (Fig. 4A). Since OsAPL1 and OsAPL2 transcripts are only detectable with RT–PCR (data not shown), their expression levels may be very low in cultured cells. OsAPS1, OsAPL3 and OsAPL4 transcripts accumulated in response to an increased concentration of Suc in the medium (Fig. 4A). The effect of Glc on OsAPS1, OsAPL3 and OsAPL4 gene expression was also evaluated (Fig. 4B). Rice cells were cultured in the modified medium containing 0–10% (w/v) Glc for 24 h. As shown in Fig. 4B, OsAPS1, OsAPL3 and OsAPL4 transcripts accumulated in response to increased Glc concentrations in the medium. Additions of 10% (w/v) polyethylene glycol MW6000 (PEG) and 10% (w/v) mannitol (Man) in the medium did not influence the OsAPL3 gene expression (Fig. 4C). In contrast, OsAPS1 and OsAPL4 transcripts also increased in response to 10% (w/v) PEG and 10% (w/v) Man in the medium (data not shown).

The starch contents in the cultured cells were measured as the reference to levels of OsAPL3 gene expression. Greater amounts of starch were detected in the cultured cells when the concentrations of sugars were increased in the medium (Fig. 5A, B). It is important to note that the expression pattern of OsAPL3 was directly correlated with starch accumulation.

Rice cells were cultured in the modified medium containing various plant hormones and AGPase gene expression was evaluated with Northern blot analysis. Under these conditions, OsAPL1 and OsAPL2 transcripts were not detected; however, OsAPL2 and OsAPL4 transcripts were detected and were found to be constitutively expressed (Fig. 6). On the contrary, the gene expression patterns of OsAPL3 and OsAPS1 were enhanced by exogenous applications of ABA (Fig. 6). In particular, OsAPL3 transcripts only increased in response to ABA when Suc was also supplemented in the medium (Fig. 7). The starch content in the cell cultures with Suc and ABA was much higher than that of the cells cultured with Suc alone (Fig. 8). The starch contents in the cells that were treated solely with ABA were decreased (Fig. 8).

Discussion

Two small and four large subunits of AGPase genes were identified in the rice genome by using the rice full-length cDNA database (Kikuchi et al. 2003) (Table 1). They are encoded by independent genes. Using them, detailed expression analyses were performed with Northern blot analysis with isoform-specific probes. Previously, six different AGPase subunits had been detected with an AGPase–antibody reaction in developing seeds (Nakamura and Kawaguchi 1992). Our studies confirm that six different AGPase genes are expressed in rice seeds. Based on the alignment analysis (Fig. 1) and subcellular localizations as predicted by programs (PSORT, ChloP, TargetP and iPSORT), OsAPS2, OsAPL2 and OsAPL3 are extra-plastidial forms and OsAPS1, OsAPL1 and OsAPL2 are plastidial forms of AGPase. Since the majority (90%) of the AGPase activity in the developing rice seeds was extra-plastidial (Sikka et al. 2001), OsAPS2, OsAPL2 and/or OsAPL3 may be major forms in the rice endosperm. Approximately 30% of the total starch in the seeds accumulated within 10 DAF (Nakamura and Yuki 1992, Hirose and Terao 2004). AGPase activities in the developing rice seeds increased at 7–8 DAF, reached a maximum at 13 DAF and then decreased (Nakamura et al. 1989). Time course expression analysis of OsAPL2 and OsAPL3 indicated their importance in starch synthesis during middle and late phase of developing seeds, respectively (Fig. 2). OsAPL1 and OsAPL4, both putative plastidial AGPases, were expressed from the early to late phase in developing seeds (Fig. 2); however, OsAPS1 was expressed from the middle to late phase (Fig. 2). These results indicated that plastidial AGPase is only present at the middle to late phase of seed development. Since wheat and barley extra-plastidial AGPase small subunit proteins increased in abundance during the middle to late phase (Burton et al. 2002, Johnson et al. 2003), cereal plastidial AGPase may play an important role in middle to late phases of seed development.

It is generally accepted that AGPase, soluble starch synthase (SS), rice starch branching enzyme (RBE) and granule-bound starch synthase (GBSS) may play a key role in starch biosynthesis. The expression properties of eight SSs (Hirose and Terao 2004), three RBEs (Mizuno et al. 2001) and two GBSSs in developing seeds (Hirose and Terao 2004) have already been characterized. Since the expression properties of six rice AGPase gene were examined, we were able to describe comprehensive expression profiles of genes related to starch biosynthesis. Taken together with these expression properties and their respective enzyme activities (Nakamura and Yuki 1992), we classified seed development into three phases: I (early), II (middle) and III (late) (Fig. 3). We categorized gene expression patterns with the following designations: (i) SSII-2, SSIII-1 and GBSSII are expressed in phase I; (ii) OsAPL3, RBE1, RBE3, RBE4, SSII-3 and SSIII-2 are expressed in phase II; (iii) OsAPS1, OsAPL1, OsAPL2 and OsGBSSI are expressed in phase III; and (iv) OsAPS2, OsAPL4SSI, SSII-1, SSIV-1 and SSIV-2 are expressed constantly. All of them were expressed in concert and alternated step-wise during seed maturation.

All AGPase isoforms were expressed in the developing seeds; however, the expression of each isoform was tissue specific for leaves, stems and roots. Large subunit sequences are divided into four groups; stem/tuber, leaf, fruit/root and monocot endosperm isoforms (Park and Chung 1998, Singh et al. 2002, Cross et al. 2004). Although OsAPL1 closely aligned with leaf isoforms that have been identified from other plants (data not shown), it was expressed in not only seeds, but also leaves and stems. OsAPL2 and OsAPL3 closely aligned with the monocot endosperm isoforms (data not shown). OsAPL2 is seed specific, while OsAPL3 was expressed not only in the developing seeds, but also in stems. Although OsAPL4 was closely aligned with stem/fruit isoforms, it was expressed in stems, but it was also detected in leaves and developing seeds. These results indicated that classification of rice AGPase isoforms based on sequence information alone is not a practical method to predict tissue specificity. Furthermore, these data determined that rice AGPase isoforms are coordinately and/or pleiotropically regulated.

The influence of day and night cycles on the expression of AGPase genes was evaluated (Fig. 3A, B). Transcripts for iAGPSI and iAGPSII from sweet potato in leaves (Bae and Liu 1997), ApS, Apl2 and Apl3 from Arabidopsis in leaves (Sokolov et al. 1998), sAGP from potato in leaves (Nakata and Okita 1995) and pvagpS1 from kidney bean in leaves (Omoto et al. 2003), transcripts for pvagpL1 in stems and roots (Omoto et al. 2003) and transcripts for iAGPLI-1 from sweet potato in stems (Harn et al. 2000) were much higher in light than in the dark periods. However, relatively minor changes were observed for transcripts of OsAPS1, OsAPS2 and OsAPL3 during day and night cycles, and such changes were not observed for OsAPL1 and OsAPL4 (Fig. 3). These results suggest that the activity of AGPase in the rice leaves and stems is not significantly regulated at a transcriptional level in response to diurnal changes of daylength.

Plant AGPase genes are induced by sugars (Li et al. 2002, Müller-Röber et al. 1990, Chen et al. 1998, Park and Chung 1998, Sokolov et al. 1998, Harn et al. 2000). We investigated the effect of sugars on AGPase gene expression. Transcripts for OsAPL3 in the cultured cells clearly responded to increased sugar concentrations in the medium. Similarly, cellular starch contents also increased under identical conditions (Fig. 4). Since the transcripts of other AGPase isoforms exhibited little to no induction by sugar treatment, OsAPL3 may play an important role for starch biosynthesis in rice cells. Since the rice endosperm enzyme is allosterically regulated (Sikka et al. 2001), differences in OsAPL3 transcript levels under various sugar conditions do not directly explain the alteration in starch content. However, our results indicate that OsAPL3 transcript levels were directly correlated with starch quantity, transcriptional regulation, either directly or indirectly, affecting the rate of starch accumulation in the rice cells.

The transcripts of α-amylase genes in rice cells are known to increase under sugar starvation, and in contrast they decrease under sugar-rich conditions (Yu et al. 1991). From these results, the machinery related to starch accumulation and degradation in the rice cells can be easily manipulated by altering the sugar level in the medium. With such a system, we analyzed the regulatory mechanisms of gene expression related to starch biosynthesis.

The gene expression of OsAPL3 was synergistically regulated by exogenous applications of Suc and ABA (Fig. 6, 7). It was reported that the Apl3 gene, which encodes one of the large subunits of AGPase in Arabidopsis, is also regulated by increased Suc and ABA concentration in leaves (Rook et al. 2001). From these results, it was concluded that higher plants generally contain such a controlling system of starch synthesis; however, the molecular mechanism of this system and its physiological role is not well known. The rice 1L-myo-inositol-1-phosphate synthase gene (Rino1) is also regulated by concentrations of Suc and ABA in suspension cell cultures (Yoshida et al. 2002). 1L-myo-inositol-1-phosphate synthase (EC 5.5.1.4) catalyzes the formation of 1L-myo-inositol-1-phosphate from Glc-6-phosphate and is a key enzyme in phytin biosynthesis in rice seeds. Phytin serves as an important form of accumulated phosphate in the rice seed. The expression pattern of Rino1 in developing seeds is also similar to that of OsAPL3 (Yoshida et al. 1999). At 4–12 DAF, high levels of Suc [100–400 µmol (g DW)^–1; Hirose et al. 2002] and ABA (0–5 ng grains^–1; Yang et al. 2001) are accumulated in developing seeds and the expression levels of OsAPL3 and Rino1 are at their maximum at that time. Although it is not clear, it is likely that OsAPL3 and Rino1 are cooperatively regulated by the concentration of Suc and ABA in the developing seeds.

Drastic increases in cell division and elongation of caryopsis occur in the early phase of grain filling (1–5 DAF) (Hirose and Terao 2004). The rapid accumulation of starch in the developing seeds is followed by the cessaton of cell division around 5 DAF (Hirose and Terao 2004). This transition is correlated with increased ABA content in the seeds (Yang et al. 2001, Finkelstein et al. 2002). The peak values of ABA significantly correlate with the maximum grain filling rates (Yang et al. 2001). Although it is well known that ABA accumulation in seeds during grain development could promote grain filling, the molecular control of this mechanism is still unknown. Since genes encoding key enzymes of starch biosynthesis and phytin biosynthesis are synergistically regulated by Suc and ABA in the cultured cells, it is possible that this synergistic regulation system may be involved in the promotion of grain filling by ABA.

Materials and Methods

Plant materials and growth condition

Germinated rice seedlings (O. sativa L. ssp. japonica cv. Nipponbare) were grown either in the field at the Agricultural and Forestry Research Center of the University of Tsukuba, Japan, in September 2002 or in a controlled growth chamber with a day/night photoperiod of 16/8 h at 300 µmol photons m^–2 s^–1 and temperature regimes of 28/24°C, respectively. Developing seeds were collected from field-grown rice plants 1–20 DAF. For growth chamber plants, leaves, stems and roots were collected at the 3rd to 4th leaf stage.

Rice suspension culture

Suspension cultures were initiated from the callus tissue derived from whole rice seeds. The seeds were incubated on solid N6D medium (Toki 1997) containing 3% (w/v) Suc and 2 mg l^–1 auxin (2,4-D) and were maintained at 28°C. After 14–21 d, friable calli formed and they were transferred to 100 ml volume flasks containing 20 ml of fresh liquid medium. Cultures were subsequently maintained in the dark at 28°C on a rotary platform shaker which operated at 120 rpm. In order to maintain the uniformity of small cell cluster size, suspension-cultured cells were passed through mesh and were subcultured every 7 d. In order to lower the internal cell levels of soluble sugars and starch prior to the final treatment, rice cells were cultured for 3 d in the medium containing a reduced amount of Suc from 3 to 1% (w/v). In order to determine the effect of sugars on AGPase gene expression, the concentration of either Suc or Glc was altered in the medium. To evaluate the osmotic effects on AGPase gene expression, carbon sources of the medium were changed from Suc to 22% (w/v) PEG or 10% (w/v) Man. The osmotic pressure of these two media is equal to that of the medium containing 10% (w/v) Glc (1.56 MPa). The osmotic pressures were measured by using a dew point microvolt meter (Model HR33T, WESCOR, Inc.). As a method to examine the effects of exogenous applications of hormones, 1 mg l^–1 cytokinin, N(6)-benzyladenine (BA) or 0–100 µM ABA, was added to the medium instead of 2,4-D.

Characterization of cDNA clones

cDNA fragments encoding the genes related to starch biosynthesis: SSII-3 (AP003509), RBE1 (D10838), OsGBSSI (X53694), OsGBSSII (AP005325) and six AGPase isoforms were obtained by RT–PCR. Total RNA was isolated using ISOGEN (Nippon Gene) from 24-h-old cultured cells. A 10 µg aliquot of total RNA was reverse transcribed using a reverse transcription kit (Invitrogen) with an oligo(dT) primer according to the manufacturer’s instructions. PCR was performed using the following oligonucleotide primers: 5′-ATATGGTGATCTGATCTCCAGTGC-3′ and 5′-CATCAAAATACCCACATTTCAGACACGG-3′ for OsAPS1; 5′-TCTTTTGTTGCCCATTCATCTGG-3′ and 5′-TGATTCCAAGCACACTCTCATCGAC-3′ for OsAPS2; 5′-GGAAAGGTTCCTATTGGAATCG-3′ and 5′-GGAGGGCTTTATTCCACCTCAG-3′ for OsAPL1; 5′-TAGATAGGCCTTGGAATCGCACC-3′ and 5′-TAGAGTTCCCATTCCAAAACAAACC-3′ for OsAPL2; 5′-GGTGTCCAAGAAAGTGATCG-3′ and 5′-AGTTGCTGCTGCTACTTCACTCG-3′ for OsAPL3; 5′-GGTTGCCAGCTTATGAATGAGGC-3′ and 5′-TTGCGGTACAAGATAAACCC-3′ for OsAPL4; 5′-ATTTTGATCTGAACGAACCG-3′ and 5′-GTCCAGCAGCCTTGTAGTATTTCC-3′ for SSII-3; 5′-GAAGGCAACAACTGGAGCTATG-3′ and 5′-CGGGGAAAGGACTTTGAATGAG-3′ for RBE1; 5′-GATCAAGGTTGCAGACAGGTACG-3′ and 5′-CGGGGAAAGGACTTTGAATGAG-3′ for OsGBSSI; and 5′-CTGAAATTCTCCATGACGGTGCC-3′ and 5′-CAAGAAATGCAGTACGAGGACAGC-3′ for OsGBSSII.

For PCR amplification of the cDNAs, the samples were denatured at 94°C for 3 min initially, and then for 1 min in subsequent cycles. Primer annealing was carried out at 66°C (OsAPL1, OsAPL3 and OsGBSSI), 64°C (OsAPL2), 58°C (RBE1,OsAPS1, OsAPS2 and OsAPL4), 57°C (SSII-3) or 54°C (OsGBSSII) for 1 min each, respectively. Extension of the primers with GeneTaq DNA polymerase (Nippon Gene) was carried out at 72°C for 1 min. After 30–35 cycles, the amplified DNA fragments were subcloned into the pGEM-T Easy vector (Promega) according to the manufacturer’s instructions and were subsequently sequenced for confirmation of sequence integrity.

Northern blot analysis

Total RNA from developing seeds was extracted by the cetyltrimethylammonium bromide method as previously described (Chang et al. 1993). Leaves, stems and roots were harvested at midday or at the middle of the night cycle, were frozen in liquid nitrogen and were stored at –80°C until further use. Total RNAs from leaves, stems, roots and cultured cells were isolated using ISOGEN (Nippon Gene). A 10 µg aliquot of total RNA in a volume of 3.3 µl was denatured by incubation with 1.5 µl of 6 M glyoxal, 1.2 µl of sodium phosphate buffer (0.1 M, pH 7.0) and 6 µl of dimethylsulfoxide (DMSO) at 55°C for 1 h. The RNA solution was chilled on ice and was separated by electrophoresis through a 1.2% agarose gel with 10 mM phosphate buffer. Afterwards, the RNA was transferred onto a Hybond N⁺ membrane (Amersham Biosciences) and was probed with [α-³²P]dCTP-labeled DNA using the Bca Best labeling kit (Takara) according to the manufacturer’s instructions in hybridization buffer [5× SSPE (SSPE is 0.15 M NaCl, 10 mM sodium phosphate, 1 mM EDTA), 1× Denhalt’s solution, 0.1% (w/v) SDS and 2 ng ml^–1 DNA solutions from salmon sperm (Nippon gene)] for 24 h at 60°C. The blots were washed once in 2× SSC (20× SSC is 3 M NaCl and 300 mM trisodium citrate) for 5 min at room temperature, once with 2× SSC, 0.1% (w/v) SDS for 15 min at 60°C, once with 1× SSC, 0.1% (w/v) SDS for 15 min at 60°C and lastly with 0.1× SSC, 0.1% SDS (w/v) for 15 min at 60°C. Autoradiography was performed at –80°C using BioMax film (Kodak, Rochester, NY, USA) with an intensifying filter. The band intensities were quantified by using ‘ImageJ’ software (http://rsb.info.nih.gov/ij/).

Starch extraction and measurement of starch content

For quantitative analysis of starch contents, soluble sugars were extracted in 80% (v/v) ethanol at 80°C for 15 min and the supernatant was discarded. This step was repeated three times. A 100 mg (DW) aliquot of precipitate was boiled in 5 ml of DMSO for 10 min. After boiling, the mixture was homogenized with a Polytron (Kinematica, Switzerland) at 30,000 rpm for 3 min at room temperature and was then homogenized in a glass homogenizer (Iwaki Glass Inc., Tokyo Japan). The supernatant was transferred to a new 15 ml centrifuge tube and was centrifuged at 1,500 rpm for 10 min at room temperature. The supernatant from this centrifugation step was used for the measurement of starch contents with a total starch assay kit (Megazyme, International Ireland, Ltd) according to the manufacturer’s instructions. In order to observe the starch granules in cells directly, we performed KI–I₂ staining. Cultured cells were immersed in KI–I₂ solution [0.4% (w/v) KI/0.02% (w/v) I₂] for 10 min at 4°C and the supernatant subsequently was discarded. The stained cultured cells were photographed with a Nikon Coolpix 950 digital camera (Nikon, Tokyo Japan).

Acknowledgments

The authors would like to thank Mrs.Tomomi Satozawa for much helpful discussion about AGPase clones (accession No. D50317).

Fig. 1 Comparison of the predicted amino acid sequences of six rice AGPase proteins with cereal extra-plastidial form AGPases. Multiple sequence alignment of AGPase (A) small subunits and (B) large subunits were data-mined using the ClustalW program. The sequences were obtained from the NCBI database; maize_BT2 (AAK69627), wheat_1aS (CAA46879), maize_SH2 (P55241) and wheat_AGP.L.1 (CAA79980). The subcellular localization of these genes is extra-plastidial (Choi et al. 2001, Burton et al. 2002). The conserved ATP-binding (I) (Frueauf et al. 2001, Frueauf et al. 2003), catalytic (II), Glc-1-phosphate (III) and activator site (IV) (Sivak and Preiss 1998) are represented by boxes, respectively. The N-terminal amino acid sequence of OsAPS1 shows significant homology with the N-terminal amino acid sequence (VSDSQNSDQ) of the spinach leaf small subunit as determined by Edman degradation (Morell et al. 1987). Consensus motifs for proteolytic processing of protein targeting plastidial stroma (Ainsworth 1993) are underlined with a thick line. An arrow indicates a putative transit peptide cleavage site.

Fig. 2 Northern blot analysis of the expression of six rice AGPase genes in developing seeds of rice plants. Total RNAs (10 µg each) of developing seeds (1–20 DAF) were electrophoresed on an agarose gel, transferred to a nylon membrane and were probed with ³²P-labeled isoform-specific cDNA fragments. Expression patterns of SSII-3, RBE1 and OsGBSSI were also analyzed as a control for seeds. The bottom panel represents an ethidium bromide-stained gel demonstrating that equal amounts of RNA are present in each lane. Phases I, II and III of seed development (as described in the Discussion) are indicated by arrows.

Fig. 3 Expression of rice AGPases in leaves, stems and roots during light/dark periods. (A) Leaves, stems and roots were harvested at midday (L) or in the middle of the night (D). Total RNAs (10 µg each) from leaves, stems and roots were electrophoresed on an agarose gel, transferred to a nylon membrane and were probed with ³²P-labeled isoform-specific cDNA fragments. Expression of OsGBSSII was also analyzed as a control for gene expression in leaves under light conditions (Dian et al. 2003). The bottom panel represents an ethidium bromide-stained gel demonstrating that equal amounts of RNA are present in each lane. (B) The intensity of expression levels of OsAPS1 and OsAPS2 in the leaves and of OsAPL3 in the stems was quantified from Northern blot analysis by using the ‘ImageJ’ software (http://rsb.info.nih.gov/ij/) with reference to the intensity of the band under light conditions.

Fig. 4 Influence of exogenous sugars on six rice AGPase genes in rice cells. Total RNAs were isolated from the rice cells cultured with: (A) 0–10% (w/v) Suc, (B) 0–10% (w/v) Glc or (C) 22% (w/v) PEG or 10% (w/v) Man for 24 h. Total RNAs (10 µg each) were separated by agarose gel electrophoresis, transferred onto a nylon membrane and were then probed with ³²P-labeled isoform-specific cDNA fragments: (A) probed with six rice AGPase genes, (B) probed with OsAPS1, OsAPL3 and OsAPL4; and (C) probed with OsAPL3. Exposure time was 24 h for OsAPS1, OsAPS2 and OsAPL3, and 7 d for OsAPL1, OsAPL2 and OsAPL4.

Fig. 5 Influence of exogenous sugars on starch accumulation in rice cells. Rice cells were cultured in the sugar-free medium or in the medium containing 3–10% (w/v) Suc or 3–10% (w/v) Glc for 24 h (A) Starch granules in the cultured cells were stained with KI–I₂ solution. (B) Starch contents were determined as described in Materials and Methods. Presented values for starch contents are the means of three replications. Vertical bars represent the SD.

Fig. 6 Effect of plant hormones on six rice AGPase genes in rice cultured cells. Total RNAs were isolated from rice cells cultured in the medium containing 3% sucrose with 2,4-D, BA or ABA for 24 h. Total RNAs (10 µg each) were electrophoresed on an agarose gel, transferred onto a membrane and were probed with ³²P-labeled isoform-specific cDNA fragments. The lower panel shows the quantification of ethidium bromide-stained total RNA profiles.

Fig. 7 Synergistic effect of Suc and ABA on OsAPL3 gene expression in rice cells. OsAPL3 mRNAs in the cells at 24 h after 0–100 µM ABA treatment (+ABA) in either the presence (+Suc) or absence (–Suc) of Suc were quantified with Northern blot analysis. Pre-cultured cells were used as a control. The lower panel shows the quantification of ethidium bromide-stained total RNA profiles as a reference.

Fig. 8 Synergistic enhancement of starch accumulation in rice cultured cells with applications of ABA and Suc. The rice cells were cultured in the sugar-free medium containing 50 µM ABA (+ABA), 3% (w/v) Suc (+Suc) and 3% (w/v) Suc plus 50 µM ABA (+Suc+ABA) for 24 h. Samples were collected and starch contents in the rice cells were extracted and measured. Starch content in the pre-cultured cells was used as a control. Values of the amount of starch are the means of three replications. Vertical bars represent the SD.

Table 1

AGPase gene multifamily in rice data-mined from the rice genome and the rice full-length cDNA database of Oryza sativa ssp. japonica cv. Nipponbare

	Gene name		Accession no.			Length		Mol. wt. (kDa)	Chromosome a	Exon no. b
		Full-length cDNAs	BAC	PAC		bp	Amino acids
Small subunit	OsAPS1	AK073146	AP004011	AP004756		1,952	500	54.8	9	10
	OsAPS2	AK103906	AC091687	AP004459		1,749	479	52.9	8	10
Large subunit	OsAPL1	AK069296	AC096689	AP004382		2,067	511	55.4	3	15
	OsAPL2	AK071497	–	AP004317		2,382	518	57.5	1	17
	OsAPL3	AK100910	AC007858	AP120988		2,313	519	57.6	5	16
	OsAPL4	AK121036	AP003574	AP004382		2,132	509	55.8	7	16

^a Chromosome number on which each AGPases gene is located.

^b The number of exons.

Abbreviations

 
• AGPase

  ADP-glucose pyrophosphorylase

 
• BA

  N(6)-benzyladenine

 
• DAF

  days after flowering

 
• DMSO

  dimethylsulfoxide

 
• GBSS

  granule-bound starch synthase

 
• Glc

  glucose

 
• Man mannitol
 
• PAC

  P1-derived artificial chromosome

 
• PEG

  polyethylene glycol

 
• RBE

  rice starch branching enzyme

 
• Rino1

  rice 1L-myo-inositol-1-phosphate synthase gene

 
• RT–PCR

  reverse transcriptase–polymerase chain reaction

 
• SS

  soluble starch synthase

 
• Suc

  sucrose