Expression analysis and functional characterization of the monosaccharide transporters, OsTMTs, involving vacuolar sugar transport in rice (Oryza sativa)


Author for correspondence:
Jong-Seong Jeon
Tel: +82 31 201 2025


  • In Arabidopsis, the compartmentation of sugars into vacuoles is known to be facilitated by sugar transporters. However, vacuolar sugar transporters have not been studied in detail in other plant species.
  • To characterize the rice (Oryza sativa) tonoplast monosaccharide transporters, OsTMT1 and OsTMT2, we analysed their subcellular localization using green fluorescent protein (GFP) and expression patterns using reverse-transcription polymerase chain reaction (RT-PCR), performed histochemical β-glucuronidase (GUS) assay and in situ hybridization analysis, and assessed sugar transport ability using isolated vacuoles.
  • Expression of OsTMT–GFP fusion protein in rice and Arabidopsis revealed that the OsTMTs localize at the tonoplast. Analyses of OsTMT promoter-GUS transgenic rice indicated that OsTMT1 and OsTMT2 are highly expressed in bundle sheath cells, and in vascular parenchyma and companion cells in leaves, respectively. Both genes were found to be preferentially expressed in the vascular tissues of roots, the palea/lemma of spikelets, and in the main vascular tissues and nucellar projections on the dorsal side of the seed coats. Glucose uptake studies using vacuoles isolated from transgenic mutant Arabidopsis (tmt1-2-3) expressing OsTMT1 demonstrated that OsTMTs are capable of transporting glucose into vacuoles.
  • Based on expression analysis and functional characterization, our present findings suggest that the OsTMTs play a role in vacuolar glucose storage in rice.


In higher plants, sugars are produced in the mesophyll cells of mature leaves which are known as source organs. Heterotrophic cells, such as roots and seeds, are sink organs and rely on the supply of sugars for their nutrition. Thus, the adequate production, storage and transport of sugars are essential to sustain growth and development in plants.

Various sugar transport systems exist in plants. Sugars are transported into surrounding cells apoplastically by sugar transporters across membranes or symplastically through the plasmodesmata. Sugars in phloem sieve cells are loaded apoplastically by sugar transporters and/or symplastically, depending on the plant species, and are transported over long distances from source to sink organs (ap Rees, 1994; Williams et al., 2000; Lalonde et al., 2004; Lim et al., 2006; Büttner, 2007). In addition to these short- and long-distance transport events, sugar molecules are distributed into different subcellular organelles in cells depending on the nutrient requirements. For example, transitory starch is degraded into maltose in chloroplasts, which is then exported into the cytosol at night via the maltose transporter (Niittyläet al., 2004; Weise et al., 2006).

In many plants, the disaccharide sucrose is the principal form of sugar for long-distance transport in phloem sieve cells, whereas the major monosaccharide sugar forms are glucose and fructose. Sugar transporters constitute a large gene family. In Arabidopsis for example, 53 putative monosaccharide transporter genes belonging to seven different subfamilies (Büttner, 2007) and more than 10 disaccharide transporters (Lalonde et al., 2004) have been identified. In addition, rice harbors more than 20 monosaccharide transporter genes and five disaccharide transporters (Hirose et al., 1997; Furbank et al., 2001; Aoki et al., 2003; Lim et al., 2006).

In mature plant cells, vacuoles occupy a large part of the cell volume and are separated from the cytosol by the tonoplast, a semipermeable vacuolar membrane. Vacuoles serve as a main long-term and also temporary storage pool in the cell (i.e. metabolites produced in excess are transported into vacuoles and released according to metabolic requirements; Cairns et al., 2000; Pollock et al., 2000; Wormit et al., 2006; Martinoia et al., 2007; Neuhaus, 2007). In this process, specific tonoplast sugar transporters are required for the vacuolar transport of these nutrients. However, whereas a large number of plasma membrane-localized sugar transporters have been well characterized in various plants (Lalonde et al., 2004; Lim et al., 2006), the existence of tonoplast sugar transporters has only recently been identified in the tonoplast proteomes of Arabidopsis and barley (Carter et al., 2004; Endler et al., 2006). Moreover, to date, only two types of vacuolar sugar transporters, the Arabidopsis thaliana tonoplast monosaccharide transporter (AtTMT) and vacuolar glucose transporter (AtVGT), have been functionally characterized in any detail (Wormit et al., 2006; Aluri & Büttner, 2007).

The tonoplast localization of AtTMT and AtVGT was previously examined by transient expression analysis of green fluorescent protein (GFP) fusion constructs using protoplasts and an active glucose transport ability of AtVGT1 has been demonstrated in isolated yeast vacuoles. Gene expression and mutant analyses of AtVGT1 have further suggested a role in seed germination and flowering (Aluri & Büttner, 2007). By contrast, AtTMT1 and AtTMT2, which possess a very long hydrophilic central loop, were found to be upregulated in response to stress treatments such as drought, salt and cold. The sugar import activities of AtTMT1 have been demonstrated using vacuoles isolated from cold-treated tmt1 mutants. These data collectively suggest a major role for the AtTMTs during stress responses in Arabidopsis that is distinct from that of AtVGT1 (Wormit et al., 2006). Notably also, unlike the AtTMTs, SsGTR, a tonoplast monosaccharide transporter (TMT) homolog of the cyanobacterium Synechocystis sp. PCC6803 that acts as glucose transporter, resides on the plasma membrane of cyanobacteria. SsGTR catalyses proton-coupled sugar import into bacterial cells (Schmetterer, 1990). In other plants, the role of the TMTs remains unknown and the characterization of these genes remains an important issue.

As an approach to elucidating the possible role of TMT proteins during the growth and development of rice, we have isolated and characterized two rice (Oryza sativa) TMTs, OsTMT1 and OsTMT2. The subcellular localization of OsTMT–GFP fusion protein was then examined in both rice and Arabidopsis. Detailed expression patterns were also analysed by reverse-transcription polymerase chain reaction (RT-PCR) and histochemical analysis of transgenic rice plants expressing OsTMT promoter–GUS fusions, and by additional in situ hybridization experiments. The sugar transport ability across tonoplasts was tested in isolated vacuoles from the transgenic Arabidopsis TMT-deficient mutant line (tmt1-2-3) expressing OsTMT1, and complementation tests of sugar levels were performed also in the transgenic Arabidopsis lines. Finally, based on the data we have obtained, the involvement of OsTMTs in sugar retrieval during assimilate transport is discussed.

Materials and Methods

Plant materials

Wild type rice (Oryza sativa L. cv Dongjin) grown in a glasshouse at 30°C during the day and at 20°C at night in a light–dark cycle of 14 h : 10 h was used in all experiments. Immature seeds were harvested at different developing stages and flowers were also collected immediately before heading. The flag leaves of mature plants were collected, while the roots were taken at the four-leaf seedling stage. Arabidopsis wild type (ecotype Columbia) and tmt1-2-3 mutant plants (Wormit et al., 2006) were used for transformation. All Arabidopsis plants were grown on soil at 22°C under a 14 h light : 10 h dark cycle unless otherwise specified.

Cloning of OsTMT full-length cDNA clones

Full-length cDNAs of the OsTMT1 and OsTMT2 genes were isolated by PCR using gene-specific primers encompassing the translation start codon and 3′ untranslated region (UTR): OsTMT1, 5′-AAATCTCCCCTAAAAGCTTCC-3′ and 5′-GAACTAGTACTCGGCTATGCTAA-3′; and OsTMT2, 5′-AAGAGGTGGAAGAAGAGGGAT-3′ and 5′-TATTATGAACCCCAACATAGTAGC-3′. All cDNAs synthesized were subcloned into the pGEM-T easy vector (Promega) and sequenced. The cDNA sequences of OsTMT1 and OsTMT2 have been deposited in the NCBI database with the accession numbers GU066765 and GU066766, respectively.

Sequence alignment and phylogenetic tree construction

The deduced amino acid sequences of the OsTMT genes were aligned with previously reported genes from other species using the clustal w program (Thompson et al., 1994). A phylogenetic tree was constructed using mega software version 4.0 (Tamura et al., 2007) via the neighbor-joining method. Bootstrap analysis was performed with 1000 replicates and bootstrap values are shown as percentages. The accession numbers of the sequences used for the construction of phylogenetic tree are: AtTMT1 (At1g20840; Z50752), AtTMT2 (At4g35300; AJ532570) and AtTMT3 (At3g51490; AJ532571) from Arabidopsis; HvSTP1 (AJ534445) and HvSTP2 (AJ534446) from barley; PST type 2a (AY165599) from sugarcane; and VvHT6 (AY861386) from grape.

Subcellular localization of OsTMT–GFP fusion proteins

To examine the subcellular localization of OsTMT1 and OsTMT2 in rice and Arabidopsis, each full-length cDNA fragment excluding the stop codon and 3′ UTR was amplified and then cloned into the region between the CaMV35S promoter and sGFP of the JJ461 binary vector (Cho et al., 2009). The primers used for the construction of OsTMT-GFP fusion vectors were: OsTMT1, 5′-GCTCTAGACTTGGTGGTAAGATTCGCCG-3′ and 5′-CCCTCGAGAGTCCTCCTTGGCCTGCTTTGC-3′; OsTMT2, 5′-GCTCTAGAGGGGTGAAGATGTCGGGTGCT-3′ and 5′-CCCTCGAGAGGCCTTTGTAGCCTGCATTTG-3′. Transgenic calli and plants expressing OsTMT-GFP fusion constructs and protoplasts isolated from transgenic Arabidopsis plants were examined using a confocal microscope (LSM 510 META; Zeiss). Chlorophyll autofluorescence was used as a chloroplast marker. In transgenic rice calli, the plasma membrane was stained with a lipophilic styryl dye, FM4-64 (Molecular Probes, Carlsbad, CA, USA).

Promoter cloning and construction of GUS fusion vectors

Using rice genomic DNA as a template, 5′ upstream regions of the OsTMT1 and OsTMT2 genes, of 2445 bp and 2481 bp in length, respectively, were amplified by PCR. The proof-reading DNA polymerase pfu (SolGent, Daejeon, Korea) was used to reduce PCR errors. The primers used were: OsTMT1, 5′-CCCTCGAGAAACCTAAACTTGAAATGTGCTA-3′ and 5′-CGCCTAGGCGAATCTTACCACTGCACAAGAA-3′; and OsTMT2, 5′-GCGTCGACATCGTGATTCTTTCTTTCAAAC-3′ and 5′-GCTCTAGACTTCACCCCTCTCCGATCCC-3′. The amplified products were cloned into pGEM T-easy and sequenced. The cloned inserts were digested with XhoI/AvrII or SalI/XbaI and then subcloned into the SalI and XbaI sites of the JJ376 vector containing the promoterless ß-glucuronidase (GUS) reporter gene, which is derived from the binary vector pC1300intC (Ouwerkerk et al., 2001). As a result, promoter regions of OsTMT1 and OsTMT2 genes were fused to a GUS gene linked to the terminator of the nopaline synthase (nos) gene. These constructs containing OsTMT1 promoter-GUS and OsTMT2 promoter-GUS fusions were thus designated pOsTMT1-GUS and pOsTMT2-GUS, respectively.

Rice transformation

To produce transgenic rice plants expressing CaMV35SOsTMT–GFP or pOsTMT–GUS fusion constructs, Agrobacterium tumefaciens LBA4404 strains harboring each of these vectors were grown on AB media (Chilton et al., 1974) supplemented with 25 mg l−1 kanamycin for 3 d at 28°C, and transgenic calli were obtained via the Agrobacterium-mediated co-cultivation method as described by Jeon et al. (2000). Transgenic rice plants were regenerated from the transformed calli on selection media containing 50 mg l−1 hygromycin and 250 mg l−1 cefotaxime.

Arabidopsis transformation

Agrobacterium tumefaciens GV3101 strains harboring one of the CaMV35SOsTMT–GFP, CaMV35S–OsTMT1, or CaMV35S–OsTMT2 constructs were grown to stationary phase in LB (Luria-Bertani; 10 g l−1 tryptone, 5 g l−1 yeast extract, 10 g l−1 NaCl) liquid culture containing 25 mg l−1 kanamycin at 28°C. For CaMV35S-OsTMT-GFP, wild type Arabidopsis plants (ecotype Columbia) were transformed by the floral dip method as described by Clough & Bent (1998). For CaMV35S–OsTMT1 and CaMV35S–OsTMT2 constructs, the Arabidopsis tmt1-2-3 mutant line was used for transformation (Wormit et al., 2006). All transgenic plants were selected on Gamborg B5 medium including vitamins (Duchefa Biochemie, Haarlem, The Netherlands) containing 25 mg l−1 hygromycin and then grown on soil at 22°C under a 14 h light : 10 h dark cycle. To examine the subcellular localization of OsTMTs in Arabidopsis, protoplasts were isolated from well expanded leaves of 3- to 4-wk-old transgenic plants, according to the method of Sheen (2001).

Histochemical GUS assay

Various organs from transgenic plants were collected and stained in GUS reaction solution containing 0.2 m sodium phosphate (pH 7.0), 10 mM EDTA, 20% methanol, 2% dimethyl sulfoxide (DMSO), 0.1% 5-bromo-4-chloro-3-indol ß-glucopyranoside (X-Gluc), as described by Jeon et al. (1999). Samples were incubated at 37°C from 1 h to overnight until sufficient staining was observed and then transferred to 70% ethanol to remove chlorophyll. After staining, the samples were fixed in a solution containing 50% ethanol, 5% acetic acid and 3.7% formaldehyde. The flowers were observed under a dissecting microscope (SZX12; Olympus) and photographed with ImagePro Express (Olympus). For longitudinal and transverse sectioning, the stained samples were embedded in Paraplast plus (Sigma) and 10 μm sections were prepared using a microtome (HM360; Microm, Waldorf, Germany). The sections were subsequently mounted on slides and observed under a compound microscope.

Sugar treatments

For sugar treatment, seedlings at the four-leaf stage were cultured for 48 h under dark conditions to deplete their endogenous sugars. The leaves were then excised in 1 cm pieces and treated in the dark at 28°C with liquid Murashige and Skoog (MS) media containing 175 mM sucrose, glucose, fructose or mannitol, according to the methods of Dian et al. (2003) and Cho et al. (2006). The samples were harvested after 12 h or 24 h of treatment.

RT-PCR analysis

For RT-PCR analysis, total RNA was prepared from havested samples using Trizol reagent and reverse-transcribed using the First-Strand cDNA Synthesis Kit (Roche) and oligo-dT primers. First-strand cDNAs were used in PCR reactions with gene-specific primers and control primers for the housekeeping genes Act1 (McElroy et al., 1990) and OsUBQ5 (Jain et al., 2006), and a sugar-inducible gene OsGBSSII (Dian et al., 2003). Gene-specific primers used were: OsTMT1, 5′-GAAGTCATCACCGAGTTCTTC-3′and 5′-TACAGGGAGAAGAATTCCAAA-3′; OsTMT2, 5′- GTCTTTGGTATATACGCAGTTGTT-3′ and 5′- AGTGGACATGAATAACATGAAAAT-3′; OsGBSSII, 5′-TAGGTGTCACTGAATGGCTAGAAC-3′ and 5′-TGGCCCACATCTCTAAGTAACATC-3′; Act1, 5′-GGAACTGGTATGGTCAAGGC-3′ and 5′-AGTCTCATGGATACCCGCAG-3′; and OsUBQ5, 5′-GACTACAACATCCAGAAGGAGTC-3′ and 5′-TCATCTAATAACCAGTTCGATTTC-3′. For Arabidopsis transgenic plants, UBQ (At5g20620) primers, 5′-GTGGTGCTAAGAAGAGGAAGA -3′ and 5′-TCAAGCTTCAACTTCTTCTTT-3′, were used as a PCR control (Cho et al., 2009). The RT-PCR analyses were performed using previously described methods (Cho et al., 2005) and repeated at least three times, giving similar results.

In situ hybridization

Harvested 3 DAF (days after fertilization) and 5 DAF caryopses were fixed for 1 h in 4% paraformaldehyde and 0.25% glutaraldehyde in 50 mM sodium phosphate buffer, dehydrated in a graded series of ethanol, and finally embedded in Paraplast plus (Sigma). The embedded samples were then carefully cut into transverse or longitudinal sections using a microtome (HM360; Microm). Sliced 7 μm sections were mounted on slides and used for in situ hybridization using the modified methods of Kouchi & Hata (1993) and Hirose et al. (2002). To generate the hybridization probe, a 523-bp fragment of the central loop region of OsTMT2 cDNA was subcloned into the BamHI and EcoRI site of pBluescript II SK (+) (Stratagene, La Jolla, CA, USA). This constructed vector was linearized by digestion with BamHI or EcoRI and labeled by in vitro transcription using digoxigenin (DIG)-labeling mix (Roche) and T3 or T7 polymerases. Sense and antisense probes labeled with DIG were synthesized from the T3 or T7 promoter of the construct, respectively, and then hydrolysed to approx. 100–200 bases. The sections were hybridized with the DIG-labeled sense and antisense probes at 42°C for 16 h. Hybridized probes were immunologically detected with anti-DIG alkaline phosphatase conjugate (Roche). Nitro blue tetrazolium chloride–5-bromo-4-chloro-3-indolyl phosphate, toluidine salt (NBT/BCIP) ready-to-use tablets (Roche) were used as the substrate.

Glucose uptake into isolated leaf mesophyll vacuoles and soluble sugar quantification

To test the glucose transport ability of OsTMT1, intact mesophyll vacuoles were isolated from wild type, tmt1-2-3, and OsTMT1 transgenic plants. Plants were grown for 4–5 wk on soil at 22°C under an 8 h light : 16 h dark cycle (90 μmol m−2 s−1) and transferred for 5 d to 50 μmol m−2 s−1 light before vacuole isolation. Isolation of vacuoles and their use in transport experiments using 14C-glucose were performed as described by Wormit et al. (2006). Transport was conducted for 15 min with 125 μm14C-glucose (specific activity 300 mCi mmol−1) and 1 mM Mg2+-ATP.

For the measurement of soluble sugar contents, Arabidopsis leaves harvested from wild type, tmt1-2-3, and OsTMT1 transgenic plants were ground to a fine powder and frozen using a TissueLyser (Qiagen-Retsch GmbH, Haan, Germany) with precooled adapters. A total of 100 mg of leaf powder was then extracted with 80% ethanol at 80°C for 1 h and extracts were evaporated and dissolved in sterile water. Spectroscopic quantification of the sugar contents was performed as described elsewhere (Lee et al., 2008).


Cloning of the rice TMT genes, OsTMT1 and OsTMT2

Systematic blast searches of the rice annotated sequence database using AtTMT isoforms as the reference sequence (Wormit et al., 2006) identified four Oryza sativa TMT genes: OsTMT1 (LOC_Os10g39440), OsTMT2 (LOC_Os02g13560), OsTMT3 (LOC_Os03g03680) and OsTMT4 (LOC_Os11g40540). To compare the similarity between these OsTMTs and the AtTMTs, and to identify TMT sequences from other plant species, we constructed a rooted phylogenetic tree using the neighbor-joining method, based on amino acid sequence comparisons. This analysis revealed that of the four OsTMTs we identified, OsTMT1 and OsTMT2 showed a closer evolutionary relationship with the AtTMTs than OsTMT3 or OsTMT4 (Fig. 1). OsTMT1 and OsTMT2 in fact formed a monocot subgroup in our tree with the two putative sugar transporters of barley, HvSTP1 and HvSTP2, which have been proposed to be involved in controlling sugar ratios between maternal and filial tissues of caryopses during early seed development (Weschke et al., 2003). This monocot subgroup also contains a putative sugar transporter from sugarcane, PST type 2a, which is homologous to the AtTMTs (Casu et al., 2003). The tonoplast localization of HvSTP1, HvSTP2, and PST type 2a, and the sugar transporter function of each remains to be determined. The AtTMTs and a putative sugar transporter, VvHT6, from grape (Vitis vinifera) formed a dicot subgroup in our analysis (Fig. 1). This finding prompted us to further characterize the localization and function of OsTMT1 and OsTMT2. In addition, OsTMT1 and OsTMT2 did not show any significant homology with the Arabidopsis VGT group, a known vacuolar glucose transporter subfamily, or the Arabidopsis STP group, a plasma membrane sugar transporter subfamily (data not shown).

Figure 1.

 Phylogenetic tree of tonoplast monosaccharide transporter (TMT) genes constructed using the neighbor-joining method. Scale bar corresponds to a distance of 10 changes per 100 amino acid positions. Numbers in the branches indicate bootstrap values as percentages. Accession numbers are given in the Materials and Methods.

The full length cDNAs of OsTMT1 and OsTMT2 were isolated by RT-PCR using gene specific primers encompassing the upstream region of the translation start codon and the downstream region of the stop codon. Sequence analysis and comparison with their genomic clones showed that both genes have genomic structures composed of six exons and contain an intron in the 5′ UTR (see the Supporting Information, Fig. S1). The deduced amino acid sequences revealed that OsTMT1 (GenBank accession no. GU066765) and OsTMT2 (no. GU066766) contain 740 and 746 amino acids, respectively, with 12 predicted transmembrane domains and an extraordinarily large, hydrophilic central loop of c. 340 amino acid residues between transmembrane domains 6 and 7, which is characteristic of the TMT family (Fig. S2). Their substantial structural similarity to the AtTMTs further suggests that OsTMT1 and OsTMT2 are members of monosaccharide transporters localizing to the tonoplast.

Subcellular localization of OsTMT–GFP fusion proteins

To determine the subcellular localization of OsTMT1 and OsTMT2, we generated transgenic rice calli expressing OsTMT1–GFP or OsTMT2–GFP constructs under the control of the CaMV35S promoter. Our analysis of these transgenic calli showed that OsTMT1-GFP localizes at the tonoplasts of the small and large vacuoles in the calli (Fig. 2a). To distinguish the tonoplast-localized GFP signals from the plasma membrane, the transgenic rice calli were next stained with a lipophilic styryl dye, FM4-64, which primarily stains the plasma membrane and is known to follow an endocytic pathway in various eukaryotes including plants (Ueda et al., 2001; Uemura et al., 2004). However, the GFP signals were not well merged with those of FM4-64 in many regions (Fig. 2b), indicating that OsTMT1 does not localize at the plasma membranes. The tonoplast localization of OsTMT1 was also demonstrated in protoplasts isolated from the mesophyll cells of transgenic Arabidopsis plants overexpressing OsTMT1-GFP. This was clearly indicated by the presence of vacuolar tonoplasts indented by chloroplasts with chlorophyll autofluorescence (Fig. 2c). Similar observations were found in the OsTMT2-GFP-expressing rice and Arabidopsis cells (Fig. S3). These results clearly demonstrate that OsTMT1 and OsTMT2 reside on tonoplasts and are not targeted to the plasma membrane.

Figure 2.

 Subcellular localization of the rice tonoplast monosaccharide transporter (OsTMT1)–green fluorescent protein (GFP) fusion protein in transgenic rice (Oryza sativa) calli and mesophyll protoplasts isolated from transgenic Arabidopsis plants expressing OsTMT1-GFP. (a,b) Localization of the OsTMT1–GFP fusion protein in transgenic rice calli. Note that the plasma membrane stained using FM4-64 does not overlap with OsTMT1-GFP. (c) Localization of the OsTMT1–GFP fusion protein in Arabidopsis protoplasts. Chlorophyll autofluorescence and FM4-64 were used as chloroplast and plasma membrane markers, respectively. The GFP signals are indicated in green, and plasma membranes stained with FM4-64 are shown in red.

Expression analysis of OsTMT1 and OsTMT2

To examine the spatiotemporal expression profile of the OsTMT1 and OsTMT2 genes in rice plants, we performed semiquantitative RT-PCR. The results revealed significant levels of OsTMT1 and OsTMT2 transcripts in all of the rice organs tested, including leaves, roots, flowers and immature seeds (Fig. 3a). Further analysis of immature seeds showed that OsTMT1 and OsTMT2 are expressed preferentially in the seed coat and not in the endosperm (Fig. 3b). A series of rice seeds collected at different developmental stages up to 15 DAF were subjected to RT-PCR analysis of OsTMT1 and OsTMT2. Rice seeds are mainly elongated longitudinally with the cellularization and proliferation of endosperm cells occurring during the pre-storage phase (1–6 DAF) and then thickened during the starch-filling milky phase (7–15 DAF) (Hirose et al., 2002; Lim et al., 2006). The expression of both OsTMTs was found to be higher during the prestorage phase compared with the starch-filling phase (Fig. 3c), suggesting a possible role of OsTMT1 and OsTMT2 in all organs, particularly in seed coats, during early seed development.

Figure 3.

 Expression analysis of rice tonoplast monosaccharide transporter genes OsTMT1 and OsTMT2. (a) Expression patterns of OsTMT1 and OsTMT2 in rice (Oryza sativa) tissues, including the leaf, root, flower and immature seed mixtures of 1- to 12-d after fertilization (DAF) caryopses. (b) Expression patterns of OsTMT1 and OsTMT2 in the endosperm and seed coat. (c) Changes in the gene expression levels of OsTMT1 and OsTMT2 during rice seed development. Rice seeds were harvested over a time course up until 15 DAF. Act1 is a PCR control. (d) Effects of sugars upon the expression of the OsTMT1 and OsTMT2 genes in rice leaves. Excised leaves were treated with Murashige and Skoog (MS) media containing sucrose, glucose, fructose or mannitol and harvested after 12 h and 24 h. MS, MS media; MS + S, MS media + 175 mM sucrose; MS + G, MS media + 175 mM glucose; MS + F, MS media + 175 mM fructose; MS + M, MS media + 175 mM mannitol. OsUBQ5 is a PCR control.

To next examine the effects of soluble sugars on OsTMT gene expression, excised leaves with depleted endogenous sugars were treated with MS media containing 175 mM sucrose, glucose, fructose or mannitol. The RT-PCR analysis revealed that, similar to the expression of the known sugar-inducible gene OsGBSSII (Dian et al., 2003), the levels of OsTMT1 and OsTMT2 are increased by treatment with sucrose, glucose or fructose for 12 h and 24 h, whereas mannitol had no effect (Fig. 3d).

It has been previously demonstrated that AtTMT genes are significantly upregulated by environmental stresses such as salt, drought and cold (Wormit et al., 2006). We thus investigated whether the expression of OsTMT1 and OsTMT2 is altered by such stimulus. We did not observe any significant or reproducible change in OsTMT1 and OsTMT2 expression in rice samples exposed to cold (4°C), salt (250 mM NaCl) or drought, possibly indicating that the OsTMTs have functions that are distinct from the AtTMTs (Fig. S4). Neither the available public microarray nor Massively Parallel Signature Sequencing (MPSS) data indicate any significant alteration of these genes in response to these environmental stresses (data not shown).

Histochemical analysis of OsTMT promoter-GUS transgenic rice plants

The expression of OsTMT1 and OsTMT2 was further examined by histochemical GUS analysis of transgenic rice plants expressing the OsTMT1 promoter–GUS (pOsTMT1–GUS) and the OsTMT2 promoter–GUS (pOsTMT2–GUS) fusion constructs, respectively. An approx. 2.5 kb 5′ upstream fragment of both OsTMT1 and OsTMT2 was transcriptionally fused to the GUS reporter gene. There was no significant variation found in GUS expression among transgenic rice plants expressing pOsTMT1-GUS or pOsTMT2-GUS (data not shown). A representative line of each with the highest GUS intensity was mainly used in our subsequent analyses. By histochemical analysis, we further revealed that GUS activities are detectable in all rice organs examined, including leaves, roots, flowers and immature seeds (Figs 4,5).

Figure 4.

 Histochemical localization of GUS expression in the leaves, roots, and flowers of transgenic rice (Oryza sativa) plants carrying the pOsTMT1–GUS or pOsTMT2–GUS construct. (a,b) Cross-section of leaves from the pOsTMT1–GUS (a) or pOsTMT2–GUS (b) lines. (c,d) Cross-section of roots from the pOsTMT1–GUS (c) or pOsTMT2–GUS (d) lines. (e,f) Spikelet of pOsTMT1–GUS. (g,h) Spikelet of pOsTMT2–GUS. BS, bundle sheath cell; CC, companion cell; MC, mesophyll cell; X, metaxylem; SE, sieve element; VP, vascular parenchyma.

Figure 5.

 Histochemical localization and in situ hybridization analysis of GUS expression in immature seeds from transgenic rice (Oryza sativa) plants carrying the pOsTMT1–GUS or pOsTMT2–GUS construct. (a,b) GUS expression in 7–8 d after fertilization (DAF) immature seeds of pOsTMT1–GUS (a) or pOsTMT2–GUS (b). (c,d) Longitudinal section of a 5 DAF immature seed of pOsTMT1–GUS (c) or pOsTMT2–GUS (d). (e,f) Transverse section of a 5 DAF immature seed of pOsTMT1–GUS (e) or pOsTMT2–GUS (f); dark field image. (g,h) Longitudinal sections of 12–15 DAF immature seeds from pOsTMT1–GUS (g) or pOsTMT2–GUS (h). Arrowheads indicate GUS expression in the dorsal sides. (i) Detection of OsTMT2 transcripts by in situ hybridization of a transverse section of a 3 DAF seed. (j) A similar section hybridized to the sense probe shows no positive labeling. (k) Detection of OsTMT2 transcripts by in situ hybridization of a longitudinal section of a 5 DAF seed. (l) A similar section hybridized to the sense probe shows no labeling. C, cross cell layer; E, epidermis; Em, embryo; En, endosperm; I, integument; M, mesocarp; N, nucellar epidermis; NP, nucellar projection; T, tube cell layer; V, main vascular bundle.

In mature leaves of the pOsTMT1–GUS and pOsTMT2–GUS transgenic rice lines, GUS activities were detected in most cells including the mesophyll cells, bundle sheath cells, and vascular bundles (the vascular parenchyma cells and phloem companion cells but not the mature metaxylem cells) (Fig. 4). The highest OsTMT1 expression was observed in the bundle sheath cells. In epidermal cells, GUS expression was very marginal in the pOsTMT1–GUS line (Fig. 4a). In the pOsTMT2–GUS line however, higher levels of reporter expression were observed in the vascular parenchyma cells and phloem companion cells compared with other regions including mesophyll cells and epidermal cells (Fig. 4b).

In the roots of the pOsTMT1–GUS and pOsTMT2–GUS lines, strong GUS activities were observed in the companion cells of the vascular bundle tissues (Fig. 4c,d) whereas in the flowers, reporter expression was detectable only weakly in the palea and lemma, but showed relatively high expression in the veins. No GUS activity was observed in other inner organs of the pOsTMT1–GUS and pOsTMT2–GUS lines, including the stamens (Fig. 4e–h).

In the developing immature seeds of both transgenic pOsTMT1–GUS and pOsTMT2–GUS rice lines, GUS expression was found to be abundant in vascular tissues of the 7–8 DAF seed coats (Fig. 5a,b). To further examine the expression of OsTMT1 and OsTMT2, immature seeds of the transgenic pOsTMT1–GUS and pOsTMT2–GUS rice lines were longitudinally or transversely sectioned. In 5 DAF immature seeds, OsTMT1 and OsTMT2 were highly expressed in the main vascular cells of the dorsal side (Fig. 5c,d) and in the nucellar projection, cross cell layers and tube cell layers of the maternal portion, but not in the aleurone or subaleurone cells of the filial part (Fig. 5e,f). The GUS activities in both transgenic lines were highest in the vascular tissues of the dorsal sides (black arrowheads) of the 12–15 DAF immature seeds (Fig. 5g,h). Notably, GUS expression in the embryos of these immature seeds was observed only in the pOsTMT2-GUS line (Fig. 5h).

The endogenous expression of OsTMT2 was further examined by in situ hybridization studies of immature rice seeds. These experiments showed that OsTMT2 is mainly expressed in the main vascular bundles and nucellar projection of the dorsal side of 3 DAF immature seeds, and in the vascular parenchyma tissues of the lateral side surrounding the integument (Fig. 5i,j). In the 5 DAF immature seeds, OsTMT2 expression was also observed in the cross cell layers and tube cell layers of the lateral side of the maternal part (Fig. 5k,l). This was very consistent with the results of our earlier RT-PCR (Fig. 3b) and histochemical analysis (Fig. 5a–h) of GUS expression.

Analysis of monosaccharide transport using isolated vacuoles from the transgenic Arabidopsis mutant tmt1-2-3 expressing OsTMT1

Several unsuccessful attempts to demonstrate the activities of the sugar transporters belonging to the TMT family via functional complementation of yeast hexose transporter-deficient mutants have been made previously (Casu et al., 2003; Wormit et al., 2006). Similarly, we found in our current experiments that OsTMT1 and OsTMT2 do not complement the glucose transport activity in the yeast mutant RE700A lacking endogenous glucose transporter. It has been shown in a previous report that the glucose uptake into vacuoles isolated from a cold-treated Arabidopsis tmt1-2-3 triple knockout mutant is significantly lower than those from wild-type controls (Wormit et al., 2006). We thus designed a complementation experiment for OsTMT1 using transgenic Arabidopsis tmt1-2-3 lines expressing CaMV35S promoter–OsTMT1, and selected two independent homozygous transgenic lines with moderate (OsTMT1-OX1) and strong (OsTMT1-OX2) transgene expression for analysis by RT-PCR (Fig. 6a).

Figure 6.

 Complementation of glucose transport into isolated vacuoles and sugar levels of Arabidopsis tmt1-2-3 triple mutant plants overexpressing OsTMT1. (a) OsTMT1 expression in transgenic tmt1-2-3 lines expressing OsTMT1, OsTMT1-OX1 and OsTMT1-OX2. UBQ is a PCR control. (b) Time course analysis of 14C-glucose uptake into vacuoles isolated from Arabidopsis wild type, tmt1-2-3 mutant, OsTMT1-OX1 and OsTMT1-OX2 lines. The dashed gray line with open squares, dashed black line with open diamonds, solid gray line with closed triangles and solid black line with closed circles represent vacuoles isolated from tmt1-2-3 triple mutant, wild type, OsTMT1-OX1 and OsTMT1-OX2, respectively. (c) Recovery of sugar levels in transgenic Arabidopsis tmt1-2-3 lines expressing OsTMT1. Dark gray bars represent wild type, white bars represent the tmt1-2-3 triple mutant, black bars represent OsTMT1-OX1 and light gray bars represent OsTMT1-OX2. Each data point represents the mean ± SD from three separate experiments.

To determine whether OsTMT1 expression in the tmt1-2-3 mutant background can complement the reduced glucose uptake activity into the tmt1-2-3 mutant vacuoles, we isolated leaf mesophyll vacuoles from wild type, tmt1-2-3 and OsTMT1 transgenic plants. Glucose uptake rates were found to be very low for tmt1-2-3 (0.5 pmol μl−1 vacuoles min−1, R2 = 0.364) and considerably higher for the Arabidopsis wild type plants (2.4 pmol μl−1 vacuoles min−1, R2 = 0.75) (Fig. 6b). This is consistent with previously reported data showing that the tmt1-2-3 mutant line has a significantly reduced capacity to import glucose into vacuoles (Wormit et al., 2006). Importantly, the vacuoles from OsTMT1-OX1 and OsTMT1-OX2 plants exhibited a transport activity that was even higher than that observed for wild-type Arabidopsis plants (3.0 pmol μl−1 vacuoles min−1, R2 = 0.766 and 4.0 pmol μl−1 vacuoles min−1, R2 = 0.979 for OsTMT1-OX1 and OsTMT1-OX2, respectively). It is also noteworthy that glucose uptake into vacuoles from OsTMT1-OX2 lines with the highest expression of OsTMT1 was significantly higher than that in the OsTMT1-OX1 plants with a moderate expression of OsTMT1 (Fig. 6b). This result strongly suggests that OsTMTs transport monosaccharide sugars into vacuoles in rice in a similar manner to the AtTMTs.

Vacuolar glucose transport was reported previously to be strongly upregulated during cold incubation in Arabidopsis (Wormit et al., 2006). To test whether the sugar content in the transgenic tmt1-2-3 mutant lines expressing OsTMT1 is restored to wild-type levels under cold stress conditions, wild-type, tmt1-2-3 and OsTMT1 transgenic Arabidopsis plants were transferred into a cold chamber (9°C) and incubated for 24 h under continuous light. In cold-treated wild-type plants, glucose, fructose and sucrose were found to have accumulated at high concentrations (7.18, 4.41 and 9.50 μmol g−1 fresh weight, respectively; Fig. 6c). However, tmt1-2-3 mutant Arabidopsis plants contained significantly lower levels of hexoses compared with wild-type plants. By contrast, transgenic tmt1-2-3 lines expressing OsTMT1 showed increased hexose levels upon 24 h of cold treatment compared with tmt1-2-3 mutant plants. Notably also, OsTMT1-OX2 lines contained glucose and fructose at levels similar to wild-type plants. This significant increase of hexose levels in transgenic tmt1-2-3 mutant plants expressing OsTMT1 under cold stress further supports the evidence that OsTMT1 can restore the role of AtTMT genes in the tmt1-2-3 mutant.


Assimilated sugar in source leaves is either transiently stored or distributed to heterotrophic sink tissues where it can serve either as an energy source for growth and development or as a supply of nutrient storage reserves. Hence, efficient sugar retention during translocation to release sites is necessary for optimal plant growth and productivity. It is postulated that sugar transporters can function to retrieve sugars that have leaked into the apoplasm (Ritchie et al., 2003; Scofield et al., 2007). Analyses of the biochemical properties of sugar transporters using protoplasts from developing seed coats have indicated that in Phaseolus and Vicia seed coats, they enable the retrieval of sugars that have leaked to the coat apoplasm during the symplasmic passage of these molecules along the post-sieve element pathway. This suggests that sugar recovery contributes to the movement of photoassimilates from the importing sieve elements to the cells responsible for sucrose release that are located at the interface between maternal and filial tissues (Ritchie et al., 2003). Similarly, it has been proposed in rice that the primary role of OsSUT1, a plasma membrane sucrose transporter, is the retrieval of sucrose from the apoplasm and its transport to the phloem sieve cells to maintain supply to the filling grain on the panicle (Scofield et al., 2007). These data raise questions about the nature of sugar retention in vacuoles within these transport paths and the involvement of vacuolar sugar transporters (Thom et al., 1982; Chiou & Bush, 1996).

The TMT family is the first vacuolar sugar transporter group to be characterized, and the TMT proteins themselves possess a large central loop between transmembranes 6 and 7. The TMT proteins have been shown to transport monosaccharides, glucose and fructose in a proton/sugar antiport manner, demonstrated by the finding that the protonophore ammonia (NH4Cl) inhibits glucose transport in isolated vacuoles (Wormit et al., 2006). In our present study, we have characterized two rice TMT genes, OsTMT1 and OsTMT2. Localization experiments using GFP fusion proteins clearly show that both of these rice TMTs reside on tonoplasts (Figs 2, S3). Glucose transport experiments using isolated vacuoles from the Arabidopsis tmt1-2-3 lines expressing OsTMT1 further demonstrated that OsTMT1 exhibits glucose transport ability across the tonoplasts (Fig. 6). Taken together with the earlier findings in Arabidopsis (Wormit et al., 2006), our data strongly suggest that the TMT proteins function in vacuolar monosaccharide transport in both monocot and dicot plants. In addition, our phylogenetic analysis further indicates that the previously reported hexose transporters, HvSTP1 and HvSTP2 from barley, PST type 2a from sugarcane and VvHT6 from grape, are likely also to be tonoplast monosaccharide transporters belonging to the TMT subfamily (Casu et al., 2003; Weschke et al., 2003; Agasse et al., 2009) (Fig. 1). It will be of great interest and value in the future to determine if these proteins all localize at the tonoplasts and retain monosaccharide transport abilities.

In Arabidopsis, AtTMT1 expression is prominent in mesophyll cells, in the cells surrounding the vascular tissue, and in the flower organs including petals, filaments and pollens (Wormit et al., 2006). However, developing Arabidopsis seeds did not exhibit AtTMT1 expression. AtTMT2 has been shown to be expressed in the root stele, in the edges of mature leaves, and in petals and filaments of flowers in Arabidopsis (Wormit et al., 2006). By contrast, AtTMT3 expression was found to be low in all Arabidopsis tissues and not detectable by northern blotting. In particular, the AtTMT1 and AtTMT2 genes have been shown to be upregulated upon cold, drought and salt stresses, and glucose transport into vacuoles was found to be stimulated by cold treatment. Thus, the AtTMT proteins have been suggested to contribute to the molecular response to osmotic stress stimuli (Wormit et al., 2006; Neuhaus, 2007).

In monocots, previous in situ localization analysis has shown that PST type 2a, a putative TMT homolog from sugarcane, is expressed mainly in the phloem companion cells and associated parenchyma cells in the mature stem (Casu et al., 2003). The expression of HvSTP1 and HvSTP2, two putative barley TMT homologs, has also been observed in young developing caryopses by in situ hybridization (Weschke et al., 2003). HvSTP2 transcripts were found to be present mainly in the region of the dorsal vascular bundle of the maternal pericarp tissue during early barley seed development, whereas HvSTP1 was shown to be expressed in the filial syncytial endosperm layer. These data likely suggest the functional involvement of these proteins in sugar translocation in sink organs. In our present study, we analysed transgenic rice, plants expressing pOsTMT1-GUS or pOsTMT2-GUS reporter constructs. Analysis of the GUS expression patterns in these transgenic plants revealed that OsTMT1 and OsTMT2 are highly expressed in vascular tissues and its peripheral regions in both source and sink organs (Fig. 5). These results from RT-PCR, promoter GUS assay and in situ hybridization experiments suggest that OsTMT1 and OsTMT2 are likely also to be centrally involved in sugar translocation in both source and sink organs.

We were unable to observe any significantly altered expression of OsTMT1 and OsTMT2 in response to environmental stress stimuli such as drought, cold or salt treatments, suggesting that they may not function in stress responses in rice. It is noteworthy in this regard that OsTMT1 and OsTMT2 are expressed at considerable levels in the photosynthetic mesophyll cells. It has been shown previously that a number of cereals, including rice, store relatively high ratios of soluble sugars to transitory starch in their leaves, whereas Arabidopsis plants primarily store starch (Nakano et al., 1995; Winder et al., 1998; Murchie et al., 2002; Trevanion, 2002; Lee et al., 2008). This likely suggests that rice requires functioning vacuolar sugar transporters to facilitate sugar turnover in these organelles and thereby sustain plant growth under normal conditions. By contrast, this role of the vacuolar sugar transporters may not be essential for plant growth in starch-storing plants including Arabidopsis. Thus, the TMTs in starch-storing plants may have evolved a stress response role.

The results of studies of plasmodesmatal frequencies and transport pathway experiments using 14C-labeled assimilates have suggested that sucrose moves symplastically from phloem sieve elements of the pericarp to the circumferential maternal nucellus, and is then apoplastically taken up through the aleurone and subaleurone cells into the endosperm cells (Oparka & Gates, 1981a,b; Furbank et al., 2001; Lim et al., 2006). In the apoplastic space in seeds, sucrose is cleaved by a cell-wall invertase such as GRAIN INCOMPLETE FILLING1 (GIF1; also called OsCIN2) in rice, Miniature1 (Mn1) in maize, and HvCWIN1 and HvCWIN2 in barley, to produce hexose compounds (Miller & Chourey, 1992; Hirose et al., 2002; Weschke et al., 2003; Roitsch & Gonzalez, 2004; Cho et al., 2005; Wang et al., 2008). GIF1 expression in rice has also been found to be restricted to the main vascular tissues of seed coats (Wang et al., 2008). Interestingly, the barley HvCWIN1 localization profile was found to be strongly associated with expression of the putative barley TMT homolog, HvSTP1. Similarly, HvCWIN2 was found to be preferentially expressed in seed coats, correlating roughly with the expression of another putative barley TMT homolog, HvSTP2 (Weschke et al., 2003). Hence, our current data and previous findings suggest that sucrose transferred into the apoplastic space in early developing seeds is hydrolysed by cell-wall invertases and that the resulting excess hexose compounds are retrieved and temporally stored in the vacuoles of vascular tissues and peripheral cells by TMTs. The generation of OsTMT1 and OsTMT2 loss-of-function rice plants in which these genes have been suppressed will be very valuable for future studies involving a more detailed characterization of the TMTs in rice, an agronomically vital crop species.

An extremely long hydrophilic loop, which is characteristic of TMT proteins, is also observed in the sucrose transporters localized on plasma membranes such as AtSUT2/AtSUC3 (Barker et al., 2000; Aoki et al., 2003). In addition, an extended large cytoplasmic domain is also found in the carboxy-terminal regions of the yeast glucose sensors SNF3 and RGT2. The existence of these extraordinarily long hydrophilic regions has initiated considerable debate about a possible function as a sugar sensor. However, no robust evidence for this has yet been presented and it will thus be interesting to determine the specific function of this large loop domain in TMT proteins in the future.

In conclusion, we have isolated and characterized the TMT genes in rice, OsTMT1 and OsTMT2, and demonstrated their involvement in vacuolar monosaccharide transport. Significantly, while the Arabidopsis TMTs play a major role in osmotic stress responses, our current data indicate that rice TMTs function in sugar retrieval during sugar translocation. Thus, our data suggest that the TMT transporters are evolutionarily conserved in many plants but are likely to have functionally diverged in monocot and dicot plants.


This work was supported, in part, by grants from the Biogreen 21 Program, Rural Development Administration, from the Crop Functional Genomics Center (CFGC) of the 21st Century Frontier Research Program (CG2111-2), from the Basic Research Program (R01-2007-000-20149-0) of the National Research Foundation, from the World Class University program (R33-2008-000-10168-0) of the Korean Ministry of Education, Science and Technology, from the Global Research Laboratory program of the Korean Ministry of Education, Science and Technology, and from the Swiss National Foundation.