Introduction

Plastids play essential roles in plant metabolic processes such as photosynthesis, starch biosynthesis, nitrogen assimilation, sulfate assimilation, fatty acid synthesis, and terpenoid synthesis. Plastids are specialized for each metabolic process in plants. For example, chloroplasts of green tissues are specialized for photosynthesis while amyloplasts of non-photosynthetic organs such as tubers and seeds are specialized for storage of a great amount of starch. Starch is a major end-product of photosynthesis in the chloroplast of source organs. Starch metabolism in the chloroplast and amyloplast is closely related to other metabolic events in the cytosol such as sucrose metabolism, glycolysis, and glyconeogenesis. Therefore, a number of translocators on the plastid envelope membrane play important roles in coordinating starch metabolism in the plastid with carbohydrate metabolism in the cytosol (Fig. 1). However, the precise mechanism of how plastidic translocators coordinate the carbohydrate metabolism between plastids and the cytosol is still unknown.

Fig. 1
figure 1

Schematic representation of starch metabolism-related plastidic translocators in the carbon flow in source (a) and sink (b) tissues of plants (modified from Neuhaus and Wagner 2000; Fischer and Weber 2002; Weber et al. 2004). ADPG ADP-glucose, CB cycle Calvin-Benson cycle, E4P erythrose 4-phosphate, FBP fructose 1,6-bisphosphate, F6P fructose 6-phosphate, Glc glucose, G1P glucose 1-phosphate, G6P glucose 6-phosphate, Mal maltose, MOS malto-oligosaccharides, PPi inorganic pyrophosphate, S6P sucrose 6-phosphate, UDPG UDP-glucose

Full size image

Phosphate translocators such as the triose phosphate/phosphate translocator (TPT), the glucose 6-phosphate/phosphate translocator (GPT), and the phosphoenolpyruvate/phosphate translocator (PPT), which counter-exchange triose phosphate, glucose 6-phosphate, and phosphoenolpyruvate, respectively, with inorganic phosphate, have been characterized in details (for review, see Flügge 1999). These phosphorylated compounds could be intermediates for starch metabolism. Other plastidic translocators including the putative ADP-glucose translocator (or Brittle-1, BT1), the plastidic nucleotide transport protein (NTT), plastidic glucose translocator (pGlcT), and the maltose translocator (MT) have been characterized (Neuhaus and Wagner 2000; Fischer and Weber 2002; Weber 2004), and are also thought to play important roles in starch metabolism.

The BT1 gene, which was identified by an analysis of the maize mutant brittle-1 (Sullivan et al. 1991), was suggested to code for the ADP-glucose translocator on the amyloplast membrane (Sullivan et al. 1995; Shannon et al. 1998), which presumably counter-exchanges ADP-glucose with AMP (Mohlmann et al. 1997; Emes et al. 2003). The shrunken-2 mutant of maize was found to be defective in the large subunit of cytosolic ADP-glucose pyrophosphorylase (AGPase) in endosperm, and produces shriveled kernel with reduced starch content (Bhave et al. 1990). These observations indicate that the supply of ADP-glucose mediated by cytosolic AGPase and BT1 from cytosol into the amyloplast stroma is essential for starch biosynthesis in maize endosperm. Therefore, BT1 and AGPase are closely related to starch productivity in the maize seed.

NTT in the plastidic membrane, formerly designated as ATP/ADP translocator protein, was characterized in Arabidopsis (Kampfenkel et al. 1995; Neuhaus et al. 1997; Reiser et al. 2004). In contrast to the mitochondrial ADP/ATP carrier, the plasitidic NTT imports extra-plastidic ATP from cytosol probably for the biosynthesis of starch or fatty acids in plastids (for review, Winkler and Neuhaus 1999). The importance of plastidic NTT in starch metabolism has been further elucidated by an analysis of transgenic potato plants in which the NTT activity was increased or decreased (Tjaden et al. 1998; Geigenberger et al. 2001).

The gene encoding the pGlcT was cloned from spinach, potato, tobacco, maize, and Arabidopsis (Weber et al. 2000), and from olive (Butowt et al. 2003). pGlcT is considered to be crucial for the efflux of glucose produced from the degradation of transitory starch in the chloroplast into cytoplasm in the dark (Weber et al. 2000; Butowt et al. 2003) and is possibly involved in the influx of extra-plastidial glucose into plastids in heterotrophic tissues (Fischer and Weber 2002).

The MT gene was cloned only recently in Arabidopsis following the analysis of a mutant with an elevated level of maltose (Niittyla et al. 2004), although its protein has been suggested to exist in the plastidic membrane two decades ago (Herold et al. 1981). MT is also considered to be important basically in the same way as pGlcT. MT mediates the export of maltose, the product of starch hydrolysis by α-amylase and/or β-amylase, from source to sink organs (Weise et al. 2004).

The starch metabolic system in plastids consists of a network of reactions catalyzed by numerous enzymes with distinct functions (Nakamura 2002; Ball and Morell 2003). Previous studies have demonstrated that several plastidic translocators are involved in starch metabolism in plants (Neuhaus and Wagner 2000; Fischer and Weber 2002; Weber et al. 2004; Fig. 1). To date, no comprehensive analysis of the expression patterns of genes encoding plastidic translocators involved in starch metabolism has been reported. In the present study, to gain insights into the mechanism of how plastidic translocators regulate starch biosynthesis, the expression patterns of almost all of the plastidic translocator candidate genes identified in the database were analyzed by quantitative real-time PCR in both photosynthetic and non-photosynthetic organs of rice.

Materials and methods

Plant material

Japonica-type rice plants (Oryza sativa L. cultivar Kinmaze; MAFF Genebank-Plant, National Agrobiological Sciences, Tsukuba, Japan; http://www.gene.affrc.go.jp/plant/) were grown in a paddy field at Akita Prefectural University, Akita, Japan. Leaf blades, leaf sheaths, and seeds at mid-milking stage [7–9 days after flowering (DAF)] were sampled at noon for total RNA extraction. For preparation of total RNA from root, 2-week-old seedlings were cultured aseptically under continuous illumination (ca. 60 μmol photons m−2 s−1) at 28°C in MS media (Murashige and Skoog 1962) with agar.

Gene search

Based on previous reports on starch metabolism-related plastidic translocators in maize and Arabidopsis (i.e., Arabidopsis TPT, Knappe et al. 2003a; maize TPT, Fischer et al. 1994; Arabidopsis GPT, Knappe et al. 2003b; maize GPT, Kammerer et al. 1998; Arabidopsis PPT1, Fischer et al. 1997; Arabidopsis PPT2, Knappe et al. 2003a; maize PPT, Fischer et al. 1997; maize BT1, Sullivan et al. 1991; Arabidopsis NTT1, Neuhaus et al. 1997; Arabidopsis NTT2, Mohlmann et al. 1997; Arabidopsis and maize pGlcT, Weber et al. 2000; Arabidopsis MT1, Niittyla et al. 2004) candidates for plastidic translocator genes in rice were first searched in the rice full-length cDNA (KOME) database (Kikuchi et al. 2003; http://cdna01.dna.affrc.go.jp/cDNA/), and the Gramene database (Ware et al. 2002; http://www.gramene.org/). Subsequently, the protein databases Rice Membrane Protein Library (RMPL; http://www.cbs.umn.edu/rice/), and the ARAMEMNON (http://aramemnon.botanik.uni-koeln.de/) were accessed for confirmation. The data obtained from these databases were scrutinized, compared, and filtered. It was also ensured that none of the genes code for the rice chloroplast genome.

Construction of phylogenetic trees

Multiple alignment of deduced amino acid sequences was performed with the Clustal W program (http://www.ddbj.nig.ac.jp; Thompson et al. 1994) to construct phylogenetic trees which were displayed using the TREEVIEW program (Page 1996).

Total RNA extraction and cDNA synthesis

Approximately 100 mg of leaf, leaf sheath, seed, or root were sampled for total RNA extraction using the RNeasy Plant Mini Kit (QIAGEN) following the attached protocol. Two micrograms of the total RNA were subsequently used to synthesize cDNAs, using the iScript cDNA synthesis kit (Bio-Rad) in total reaction volumes of 40 μl. Sterilized distilled water was added to dilute the cDNA tenfold to obtain 400 μl of cDNA solution.

Quantitative real-time PCR

Expressions of genes were analyzed by using the SYBR Green kit (QIAGEN) and iCycler (Bio-Rad) following the manufacturers’ instructions. One microliter of the synthesized cDNA and 1.25 μl of 6.25 μM primer solutions (forward and reverse; summarized in Table 1) were used for each real-time PCR (total volume of 50 μl) with an annealing temperature of around 58–62°C. The melting curve was generated to check the specificity of the amplified fragments. As an internal control gene, rice actin (Accession; X16280) or the α-tubulin (Accession; AK067721) gene was used. Absolute quantification of the transcript number was performed with external calibration standards generated by quantitative real-time PCR from the amplified fragment cloned in pGEM-T vector (Promega). The identity of the cloned transcript was calculated as the ratio to actin (Fig. 2) or α-tubulin (Fig. 3) control. All the primers were designed using Primer Express 2.0 (Applied Biosystems).

Table 1 Oligonucleotide primers for real-time PCR used in this study
Full size table
Fig. 2
figure 2

Expressions of genes possibly encoding starch metabolism-related plastidic translocators in various organs of rice at mid-milking stage of the seed (7–8 DAF). The relative amount of each mRNA (vertical bars) was defined as the ratio to that of rice actin gene (Accession no. X16280). Values are means ± SE of at least four replicates. Le leaf, LS leaf sheath, Se seed, Ro root

Full size image
Fig. 3
figure 3

a, b Expressions of OsBT1-1, OsGPT1, OsGPT2-1 & 2-2, and OspGlcT during seed maturation of rice. Total RNAs from rice seeds at 5, 10, 15 DAF were prepared for real-time PCR analysis. a Relative amount of each mRNA was defined as the ratio to that of rice α-tubulin gene (Accession no. AK067721), which was used as the internal control. b Expression profile of each gene during seed maturation

Full size image

Results

Starch metabolism-related plastidic translocator genes in rice

Available information on amino acid and genomic sequences of TPT, GPT, PPT, BT1, NTT, pGlcT, and MT of Arabidopsis and maize were used as in silico probes to find their counterparts in rice. The putative rice translocator genes were named according to their Arabidopsis and/or maize homologs and annotated using the TIGR loci, thus: OsTPT1 (LOC_Os01g13770) and OsTPT2 (LOC_Os05g15160) for the TPT genes, OsGPT1 (LOC_Os08g08840) and OsGPT2-1, 2-2, 2-3 (LOC_Os07g34010, LOC_Os07g33960, LOC_Os07g33910, respectively) for the GPT genes, OsPPT1,2, 3, 4 (LOC_09g12600, LOC_Os08g25630, LOS_Os01g07730, LOC_Os05g07870, respectively) for the PPT genes, OsBT1-1, 1-2, 1-3 (LOS_Os02g10800, LOC_Os05g07900, LOC_Os06g40050, respectively) for the BT1 genes, OsNTT 1, 2 (LOC_Os01g45910, LOC_Os02g11740, respectively) for the NTT genes, OspGlcT (LOC_Os01g04190) for the pGlcT gene, and OsMT (LOC_Os04g51330) for the MT gene. These genes were considered as starch metabolism-related plastidic translocator genes in rice, based on the high similarities of their amino acid sequences with those of maize and Arabidopsis, and of their exon/intron splice sites with their respective counterpart genes in Arabidopsis.

One advantage of using the genome databases is that the sequences besides the coding region, such as 5′-upstream and 3′-downstream sequences, are easy to discern. Although the TATA-less promoters were reported to be restricted to photosynthetic and plastid ribosomal genes (Nakamura et al. 2002; Achard et al. 2003), only OsGPT and OsBT1-2 genes had an apparent TATA-box motif (data not shown).

Phylogenetic trees of the rice gene products and their homologs in other plants were constructed (data not shown). Overall, the rice translocator proteins identified in this study, except for OsBT1-2 and OsBT1-3, were most similar with their homologs in monocots such as maize and wheat.

Expression profiles of starch metabolism-related plastidic translocator genes in rice

Expression profiles of the rice genes were analyzed by quantitative real-time PCR using the oligonucleotide primers listed in Table 1. The actual amounts of each transcript in different organs of rice were calculated. Data were normalized using the expression level of actin (Fig. 2) or α-tubulin (Fig. 3) as internal control because both genes were expressed constitutively in various tissues. Although the amount of transcript is not always correlated with the protein amount or activity, the data in this study could be strong indicators of their physiological roles.

Levels of the OsTPT1 transcript in all organs examined were by far greater than those of OsTPT2 (Fig. 2). The expression levels of both genes in source organs such as the leaf blade were more than tenfold greater in comparison to the levels in sink organs such as seeds and roots. The OsTPT1 transcript levels in the source organs were one order higher than that of the actin gene. These results suggest that OsTPT1 is the major functional form playing an important role in the source organs in carbohydrate metabolism in rice plants.

Both OsGPT1 and OsGPT2-1∼2-3 were expressed in all organs examined, although the amount of OsGPT1 transcript was higher in the seed than in the leaf (Fig. 2). A similar pattern was also observed for OsGPT2-1 & 2-2 although it was impossible to independently measure the levels of OsGPT2-1 and OsGPT2-2 because of the difficulties in designing the specific primers for each transcript. In contrast, OsGPT2-3 was expressed the most in the leaf although the overall levels were lower than the other OsGPT genes.

The transcript levels of the four OsPPTs shared similar patterns in that their transcripts were predominantly expressed in source organs such as the leaf and leaf sheath rather than in sink organs such as the seed and root (Fig. 2). These results suggest that OsPPTs play important parts in carbohydrate metabolism during photosynthesis.

OsBT1-1 gene was expressed almost exclusively in seed while the OsBT1-2 and OsBT1-3 were expressed mainly in the leaf and leaf sheath, although the expression level of OsBT1-1 in seed was markedly higher than those of OsBT1-2 and OsBT1-3 in the leaf (Fig. 2). These observations suggest that OsBT1-1 is essential for starch biosynthesis in the rice endosperm by translocating ADP-glucose from the cytosol into the amyloplast.

The OsNTT 2 gene was expressed by far greater than that of the OsNTT 1 gene (Fig. 2). The transcript levels of all the OsNTT genes were higher to some extent in the leaf and leaf sheath than in the seed and root.

OspGlcT was mostly expressed in the seed, but lesser in the leaf and leaf sheath (Fig. 2), suggesting that OspGlcT plays some roles in both the photosynthetic and non-photosynthetic tissues of rice. OsMT was expressed mainly in the leaf and leaf sheath, but the overall expression of OsMT was seemingly quite low (Fig. 2). These results suggest the distinct physiological roles of OspGlcT and OsMT in carbohydrate metabolism in rice plants.

Changes in expression levels of BT1, GPT, and pGlcT during seed development

In an attempt to clarify the possible roles of OsBT1-1, OsGPTs, and OspGlcT, we followed changes in their transcript levels during three different developmental stages of the rice endosperm, as shown in Fig. 3. It was noted that the expression of OsBT1-1 was greatly higher at 10 and 15 DAF when starch most vigorously accumulates in the endosperm as compared with the very early endosperm development prior to significant starch production in the endosperm (at 5 DAF). The expression of OsGPT1 gradually increased during endosperm development whereas the transcripts of OsGPT2-1 and 2-2 were constant throughout the three stages. The expression of OspGlcT increased with the progress of endosperm development, but the increase was not dramatic as compared with OsBT1-1. All these data strongly suggest a specific role of the OsBT1 protein in starch biosynthesis in the rice endosperm.

Discussion

Arabidopsis has approximately one hundred plastidic translocators (Ferro et al. 2002; Koo et al. 2002; Schwacke et al. 2003). In this study, homologs of these plastidic translocator proteins in rice were searched for in the database. Rice has approximately 32,000–50,000 genes (Goff et al. 2002), and is predicted to have at least a hundred plastidic translocator genes like Arabidopsis.

The plastidic translocators TPT, GPT, PPT, BT1, NTT, pGlcT, and MT were examined in this study because of their important roles in coordinating starch metabolism in plastids with carbohydrate metabolism in the cytosol (Flügge 1999; Neuhaus and Wagner 2000; Fischer and Weber 2002; Emes et al. 2003; Weber 2004; Weber et al. 2004). The existence of glucose 1-phosphate/phosphate translocator, which was postulated to be a plastidic translocator, has been reported in wheat (Tetlow et al. 1994, 1996; Tyson and ap Rees 1988) and in potato (Naeem et al. 1997), but its sequence has not been determined. Although the presence of the deduced translocator has been doubted (Huber et al. 1992; Kofler et al. 2000), its existence cannot be ruled out until annotation of the Arabidopsis and rice genome is completed. Xylulose 5-phosphate/phosphate translocator (XPT), while present in Arabidopsis (Eicks et al. 2002), has not been reported in the TIGR rice database.

Based on the Arabidopsis database, Knappe et al. (2003a) showed the gene structures and related data on the plastidic phosphate translocators TPT, GPT, PPT, and XPT. The same strategy was used in this study, which focused on starch metabolism-related plastidic translocators BT1, NTT, pGlcT, and MT, in addition to the phosphate translocators they examined. Although the regulation of starch metabolism, including the structures and functions of numerous enzymes involved, has been extensively studied in various plants, there are no reports on comprehensive gene expressions of plastidic translocators. Having considered that the regulatory mechanism and composition of isoforms in each class of enzymes for starch synthesis are known to be different between dicots and monocots (James et al. 2003), a comprehensive analysis of the expression of plastidic translocator genes in rice as a model monocot plant would generate information for comparison with the available data on plastidic translocators in the model dicot plant Arabidopsis.

Figure 2 shows that OsTPT1 was expressed predominantly in the leaf and leaf sheath, but only slightly in the seed and root, consistent with the results obtained in Arabidopsis (Knappe et al. 2003b), cauliflower (Fischer et al. 1997), maize (Fischer et al. 1997; Kammerer et al. 1998), pea (Knight and Gray 1994), potato (Schulz et al. 1993; Schunemann et al. 1996), tobacco (Knight and Gray 1994), and tomato (Schunemann et al. 1996). OsTPT2 expression pattern was similar to that for OsTPT1, although the expression level of the former was much lower (Fig. 2). Thus, the functional TPT in rice might be OsTPT1 only, and TPT2 is probably a non-functional pseudogene. In Arabidopsis only a single TPT gene was identified (Knappe et al. 2003a).

The expression of the four OsPPT genes was much higher in source organs than in sink organs (Fig. 2), the expression patterns sharply contrasting those in maize and cauliflower (Fischer et al. 1997; Kammerer et al. 1998). RT-PCR analysis demonstrated that the two Arabidopsis PPTs exbibit different patterns of expression; PPT1 is expressed ubiquitously whereas PPT2 is expressed mainly in the leaf (Knappe et al. 2003b). No marked expression of OsPPT genes was detected in the rice seed while the Arabidopsis PPT1 is significantly expressed in the seed, which may reflect the importance of PPT1 for the shikimate pathway and fatty acid synthesis in Arabidopsis seed.

The present observations strongly suggest that both OsTPT and OsPPT fulfill crucial roles in photosynthetic carbon metabolism in source cells. The contribution of OsTPT1 to the export of triose phosphate synthesized from CO2 in the chloroplast into cytoplasm might be greater than that of OsTPT2, because the transcript level of OsTPT1 was about 77-fold and 17-fold higher than that of OsTPT2 and actin transcripts, respectively (Fig. 2).

OsGPT1 and OsGPT2-1 & 2-2 were most abundantly expressed in the seed, but their transcripts in the leaf and leaf sheath were also substantial (Fig. 2), which is in contrast to the finding of Kammerer et al. (1998) that GPT is expressed almost exclusively in non-photosynthetic tissues of maize. One possible explanation for this discrepancy is that the OsGPT expression in rice might arise from the probable substantial expression of GPTs in leaf guard cells, as observed in pea (Overlach et al. 1993). Alternatively, other translocator(s) might perform the function of GPT in maize, considering that the phosphate translocators, TPT, GPT, PPT, and XPT, often share substrate specificity to some extent (Fischer et al. 1997; Kammerer et al. 1998). In Arabidopsis, microarray data demonstrated that GPT1 (At5g54800) is expressed in both source and sink organs while GPT2 (At1g61800) is expressed almost exclusively in the seed (Zimmermann et al. 2004). Taken together, the expression pattern of Arabidopsis GPT1 is similar to that of OsGPTs, while that of Arabidopsis GPT2 resembles that of the maize GPT gene.

The expression of OsBT1-1 was entirely seed-specific, and its expression pattern was very similar to that of maize BT1 (Fig. 2), which was hypothesized to transport the ADP-glucose produced by cytosolic AGPase into amyloplast for starch biosynthesis in maize and probably all cereal endosperms (Sullivan et al. 1991; Cao and Shannon 1997). However, this postulated function of maize BT1 remains to be proven in direct transport experiments. Results of searches for cis-elements in the three OsBT1 gene promoters (ca. 1.5 kb of the 5′-upstream region from the transcription start site) on the PLACE database (Higo et al. 1999; http://www.dna.affrc.go.jp/PLACE/) revealed that the OsBT1-1 promoter has one cis-element for endosperm-specific expression, thus explaining the seed-specific expression of BT1-1. Interestingly, microarray data indicate that the putative Arabidopsis BT1 gene (At4g32400), which seems to be a single copy gene, is expressed in both source and sink organs (Zimmermann et al. 2004). However, the BT1 homolog from potato recently characterized by Leroch et al. (2005) was proposed to be a plastidic ATP uniporter. While the maize plastidic BT1 remains to be elucidated, it appears that the metabolic role and the transport mechanism of BT1 might differ between the two plant species.

Because OsBT1-1 was expressed specifically in maturing seeds while OsBT1-2 and OsBT1-3 were expressed in every tissue in very low levels (Fig. 2), the functions of the three OsBT1 are most likely different.

Based on these observations, we presume that OsBT1-1 and OsGPTs could possibly mediate the transport of ADP-glucose and glucose 6-phosphate, respectively, from the cytosol into the amyloplast in the rice endosperm, where ADP-glucose and glucose 6-phosphate serve as substrates for starch biosynthesis and the pentose phosphate pathway, respectively, or both compounds become the precursor of starch in the amyloplast. However, BT1-deficient mutants of maize (Sullivan et al. 1991; Shannon et al. 1998) and barley (Patron et al. 2004), and cytosolic AGPase-deficient mutants from maize (Bhave et al. 1990) and rice (Yano et al. 1984; Satoh et al. 2003; Kawagoe et al. 2005) have shriveled seeds with a markedly reduced starch content, suggesting that the predominant pathway for the supply of ADP-glucose for starch biosynthesis in the amyloplast of cereal endosperm is via cytosolic AGPase and BT1, whereas the contributions of GPT and/or plastidic AGPase are limited, if any. In this connection, it is particularly interesting to note that the expression of OsBT1-1 sharply increased at 10 DAF and this high expression continued until 15 DAF whereas the increases in transcripts in OsGPT1, OsGPT2-1 & 2-2, and OsGlcT were less marked (Fig. 3). The results indicate that the timing for the increase in starch production in the endosperm is closely related to the level of the BT1 transcript, suggesting a specific role of OsBT1-1 in starch biosynthesis of the rice endosperm.

The expression levels of the two OsNTT s were higher in source organs than in sink organs (Fig. 2), which is basically consistent with the data for Arabidopsis (Kampfenkel et al. 1995; Reiser et al. 2004) and potato (Tjaden et al. 1998). The result tempts us to speculate that in rice NTT facilitates the transfer of cytosolic ATP derived from mitochondria into chloroplasts at night.

OspGlcT was expressed mainly in the leaf, leaf sheath, and seed (Fig. 2), in agreement with the previous reports in tobacco (Weber et al. 2000). pGlcT may play an important role in both the export and import of glucose through the plastid envelope of rice (Fig. 2). OsMT was expressed mainly in the leaf and leaf sheath, but the overall level was relatively lower than those of the other genes (Fig. 2). Despite their possible important roles, pGlcT and MT were revealed to have only one copy each in the rice genome. In Arabidopsis, since the MT-deficient mutant (MEX1) grows more slowly than wild-type plants and has a reduced amount of chlorophyll, it might have quite an important role(s) in starch metabolism, especially in starch degradation at night (Niittyla et al. 2004). The transcript level of OspGlcT was more than threefold greater than that of OsMT in every organ examined (Fig. 2). However, microarray data demonstrated that in Arabidopsis MT is expressed much more than pGlcT (Zimmermann et al. 2004), suggesting that the mechanism of starch degradation may be different between Arabidopsis and rice.

All the results in the present study allowed us to identify the genes involved in, and thus provide insights into, the regulation of the carbohydrate metabolism network encompassing multiple cellular compartments such as the cytosol and chloroplast or amyloplast. The present results strongly suggest the distinct roles of different plastidic translocators in coordinating carbohydrate metabolism in plastids and the cytosol, although these transporters possess varying degrees of importance to the metabolism, and their expression patterns in various tissues seem to be plant species-specific. It is also true that gene transcript levels are merely suggestive of the involvement of the individual gene product in tissue-specific metabolism, and the identification of the gene is undoubtedly an essential initial step prior to its functional characterization. To understand the mechanism of how plastidic translocators coordinate carbohydrate metabolism occurring in plastids and in the cytosol in plants, additional studies such as gene silencing using RNAi methodology and the molecular characterization of these transporters are necessary.