Reduction in sucrose contents by downregulation of fructose-1,6-bisphosphatase 2 causes tiller outgrowth cessation in rice mutants lacking glutamine synthetase1;2.

BACKGROUND
Our previous transcriptomic analysis revealed that downregulation of nitrogen and carbon metabolism in the basal portions of the shoots inhibited cytosolic glutamine synthetase1;2 (GS1;2), which severely reduced rice tiller number. In the present study, we used rice mutants lacking GS1;2 (gs1;2 mutants) to determine the contribution of carbon metabolism to tiller growth.


RESULTS
Metabolomic analysis indicated the effects of carbon metabolism disorder such as reductions in the levels of sugar metabolites (e.g., sucrose and glucose 6-phosphate) in the shoot basal portions of the gs1;2 mutant seedlings. Decrease in sucrose caused by the lack of GS1;2 was successfully restored to the wild-type levels by introducing OsGS1;2 cDNA into the mutants. In the basal portions of the shoots, the lack of GS1;2 caused low expression of cytosolic fructose 1,6-bisphosphatase2 (OscFBP2), which is a key cytosolic sucrose synthesis enzyme; it is especially important in the phloem companion cells of the nodal vascular anastomoses. NH4+ supply upregulated OscFBP2 expression in the shoot basal portions of the wild type but not in those of the gs1;2 mutants. Rice mutants lacking cFBPase2 presented with ~ 30% reduction in total cFBPase activity in the basal portions of their shoots. These mutants displayed reductions in sucrose levels of the basal portions of their shoots but not in their leaf blades. They also had relatively lower tiller numbers at the early growth stage.


CONCLUSIONS
Metabolomic analysis revealed that the lack of GS1;2 reduced sucrose metabolism in the basal portions of the shoots. Our results indicated that sucrose reduction was caused by the downregulation of OscFBP2 expression in the basal portions of the gs1;2 mutant shoots. The reduction in sucrose content caused by the lack of cFBPase2 resulted in lower tiller number at the early growth stage. Therefore, adequate sucrose supply via cFBPase2 may be necessary for tiller growth in the basal portions of rice shoots.


Background
Rice plant tiller number is a critical agronomic trait. It defines grain yield and is influenced by nitrogen and carbon availability (Mae 1997;Sakamoto and Matsuoka 2008;Liu et al. 2011). In rice, the axillary buds of the tillers emerge from the basal portions of the shoots. Axillary bud outgrowth may be correlated with the efficiency of metabolite use and with hormone signaling networks (Domagalska and Leyser 2011;Evers et al. 2011;Ohashi et al. 2017).
Inorganic NH 4 + is the major nitrogen source in paddy fields. It is assimilated by the coupled reactions of glutamine synthetase (GS) (which catalyzes an ATP-dependent conversion of glutamate to glutamine) and glutamate synthase (Yamaya and Kusano 2014). A small gene family encodes cytosolic GS (GS1) in rice (Tabuchi et al. 2007;Yamaya and Kusano 2014) and the major species of chloroplastic GS (GS2) in green tissue of rice. The evidence from previous study showed that the three GS1 isoenzymes (GS1;1, GS1;2 and GS1;3) in rice have distinct functions (Yamaya and Kusano 2014). Of these, GS1;2 is most important in primary NH 4 + assimilation in rice roots (Ishiyama et al. 2004;Funayama et al. 2013).
The lack of GS1;2 suppressed tiller axillary bud outgrowth, which, in turn, substantially decreased tiller number and rice yield (Funayama et al. 2013;Ohashi et al. 2015b). In the basal portions of rice shoots where the axillary buds emerge, GS1;2 protein localized in the phloem companion cells of the nodal vascular anastomoses joining the axillary bud vessels (Ohashi et al. 2015b). Transcriptomic analysis showed downregulation of genes regulating nitrogen and carbon metabolism in the basal portions of the gs1;2 mutant shoots (Ohashi et al. 2015b). Therefore, the metabolic disorder caused by the lack of GS1;2 severely reduced tiller axillary bud outgrowth triggered by imbalanced primary metabolism (Ohashi et al. 2015b). The lack of GS1;2 significantly lowered glutamine and asparagine content in the basal portions of the shoots (Ohashi et al. 2017). The low glutamine levels repressed the active cytokinin required for tiller bud outgrowth (Ohashi et al. 2017). In contrast, the low asparagine content caused by asparagine synthetase1 (OsAS1) downregulation did not significantly inhibit tiller axillary bud outgrowth (Ohashi et al. 2018). In conclusion, it is highly probable that glutamine is responsible for tiller axillary bud outgrowth. Glutamine is a nitrogen source and signals for axillary bud development by modulating cytokinin synthesis (Ohashi et al. 2015b(Ohashi et al. , 2017(Ohashi et al. , 2018. Sufficient sucrose and glutamine may be needed for tiller axillary bud outgrowth. However, this process also depends on phytohormonal, environmental, and developmental signals (Kebrom et al. 2012;Mason et al. 2014). Nevertheless, the mechanisms regulating tiller development involved in carbon metabolism remain unidentified. Sucrose is a major exporting and storage form of soluble carbohydrate in higher plants. It is imported into sink tissues and then used to maintain cellular metabolism, cell wall biosynthesis, and respiration. It is also converted to starch for storage (Ruan 2014;MacNeill et al. 2017). Sucrose biosynthesis is balanced with the photosynthetic rate because it must be produced from the triose phosphate exported from chloroplasts (Ruan 2014;MacNeill et al. 2017). Cytosolic triose phosphates are substrates for sucrose synthesis. The rate-limiting enzymes in this process are cytosolic fructose-1,6-bisphosphatase (cFBPase) and the sucrose phosphate synthase-sucrose phosphate phosphatase (SPP) enzyme complex (Stitt and Quick 1989;Daie 1993;Strand et al. 2000;Nägele and Weckwerth 2014;Maloney et al. 2015).
Our previous study revealed that the lack of GS1;2 in the basal portions of rice shoots downregulated the genes regulating sucrose synthesis, namely OscFBP2, which encodes cFBPase (Ohashi et al. 2015b). Rice has two cFBPase isozymes (OscFBP1 and OscFBP2) (Lee et al. 2008). In rice, the cFBPase1 may play a major role in the cytosolic conversion of trioseP to sucrose in the leaves during the daytime (Lee et al. 2008). The rice mutants lacking cFBPase1 exhibited severe growth retardation and had significantly reduced levels of sucrose, glucose, fructose, and starch in their leaves (Lee et al. 2008;Koumoto et al. 2013). In contrast, the role of cFBPase2 is unclear (Lee et al. 2008).
In this study, we used metabolomic analysis to elucidate the carbon metabolism disorder in basal portions of rice shoots caused by the lack of GS1;2. To understand the contribution of carbon metabolism to tiller growth in gs1;2 mutants, we focused on the reduction in sucrose levels in the shoot basal portions of rice mutants lacking GS1;2 (gs1;2 mutants). We also used rice mutants lacking either GS1;2 or cFBPase2 (oscfbp2 mutants) to determine (1) the mechanisms of sucrose reduction in the shoot basal portions of gs1;2 mutants, and (2) the contribution of cFBPase2 to tiller number.

Results
Reductions in nitrogen and carbon metabolite levels in the basal portions of shoots caused by the lack of GS1;2 A metabolomics analysis was performed on the shoot basal portions of both wild type and gs1;2 mutants at the fourth leaf stage. Relative to the wild type, levels of sugars such as sucrose and glucose-6-phosphate were lower in the basal portions of the gs1;2 mutant shoots (Fig. 1). There were no significant changes in the levels of maltose, glucose, fructose, trehalose, fructose-6-phosphate, or other sugars (Fig. 1).
Reduced OscFBP2 expression in the phloem companion cells of the nodal vascular anastomoses in the basal portions of the shoots lacking GS1;2 The transcriptomic analysis of the previous study (Ohashi et al. 2015b) revealed the downregulation of two genes relating to sucrose metabolism (OscFBP2 and OsSPP2) in the basal portions of the gs1;2 mutant shoots. The expression levels of OscFBP1, OscFBP2, OsSPP1, OsSPP2, and OsSPP3, all of which are encoded in the rice genome, were measured by quantitative real-time PCR (qPCR) in the basal portions of the wild type and the gs1;2 mutant shoots. The expression of OscFBP2 in the basal portions of the gs1;2 mutant shoots at the fourth leaf stage was 50% lower than that of the wild type (Fig. 3a). OscFBP1 expression is dominant in wild-type rice leaves (Lee et al. 2008). However, OscFBP2 expression was > 10× higher than that of OscFBP1 in the basal portions of the wild-type shoots (Fig. 3a). The expression of OscFBP1 in the gs1;2 mutants was comparatively less reduced ( Fig. 3a). OsSPP2 expression in the basal portions of the gs1;2 mutant shoots at the fourth leaf stage was~70% lower than that of the wild-type rice. Nevertheless, both OsSPP2 and OsSPP3 expression levels were substantially lower than that of OsSPP1 (Fig. 3b). OsSPP2 was not dominant relative to OsSPP1; therefore, the effect of OsSPP2 downregulation was limited only in the gs1;2 mutants. There were no significant differences between the wild type and gs1;2 mutants regarding Actin1 expression levels in their basal portions of the shoots (Fig. 3c).
In situ hybridization revealed strong OscFBP2 transcript signals in the phloem companion cells of the nodal vascular anastomoses in the basal portions of the wild-type shoots (Fig. 4a). Weak OscFBP2 signals were also detected in the internodal parenchyma cells (Fig. 4b) and the lower regions of the leaf sheaths in wild-type rice (Fig. 4c). In the gs1;2 mutants, the OscFBP2 signal intensity in the phloem companion cells of the nodal vascular anastomoses was much lower than that of the wild type (Fig. 4a, d). The OscFBP2 signal intensities in the internodal parenchyma cells and the lower regions of the leaf sheaths were only at background level in the gs1;2 mutants ( Fig. 4e, f). OscFBP1 transcript signals were detected in the phloem companion cells of the nodal vascular anastomoses in both the gs1;2 mutants and wild-type rice but not in the internodal parenchyma cells or the lower regions of the leaf sheaths (Additional file 2: Figure S2a-f). The OscFBP1 transcript signal intensities did not significantly differ between the gs1;2 mutants and the wild-type rice (Additional file 2: Figure S2a-f). When the sense probes were used as a control, only background levels of the OsFBP2 and OscFBP1 transcript were detected (Fig. 4g, h, i; Additional file 2: Figure S2g, h, i,).
Disappearance of induction of OscFBP2 expression after NH 4 + supply to rice lacking GS1;2 Adenosine phosphate isopentenyltransferases4 (OsIPT4) and OsAS1 are induced in response to NH 4 + supply. They were downregulated by the lack of GS1;2 (Kamada-Nobusada et al. 2013;Ohashi et al. 2017Ohashi et al. , 2018 response to NH 4 + supply were measured for the basal portions of the wild type and the gs1;2 mutant shoots. NH 4 + supply increased OscFBP2 five-fold in the wild type. In contrast, NH 4 + supply had comparatively less impact on OscFBP1 expression (Fig. 5a, b). NH 4 + supply did not induce OscFBP2 expression in the gs1;2 mutants ( Fig. 5b). There were no significant changes in Actin1 expression in response to NH 4 + supply (Fig. 5c).
Sucrose content reduction in the basal portions of the shoots and tiller axillary bud outgrowth in the absence of OscFBP2 at the early growth stages The endogenous retrotransposon Tos17-inserted rice mutants for OscFBP2 were isolated, and their sugar content and tiller number were measured at the seedling stage. OscFBP2 mutants were screened by searching the flanking sequence database (Miyao et al. 2003) of the mutant panel (https://tos.nias.affrc.go.jp/~miyao/pub/tos17/) of the Project for Rice Genome Research. Mutant lines were produced by random insertion of Tos17 into the rice (Oryza sativa L. cv Nipponbare) genome (Hirochika et al. 1996) and collected in the mutant panel. Tos17 was inserted into the exon 5 and 5′ untranslated regions of OscFBP2 in lines NF9351 (oscfbp2-m1) and NF2064 (oscfbp2-m2), respectively (Fig. 6a). In line NF9351 (oscfbp2-m1), sequencing of the reverse transcription (RT)-PCR fragment for OscFBP2 showed a 114-bp deletion in the OscFBP2 mRNA, which corresponded to the deletion of 38 amino acids in the cFBPase2 sequence (Additional file 3: Figure S3). In NF2064 (oscfbp2-m2), OscFBP2 expression decreased by~50% in the basal portions of the shoots (Fig. 6a, b). Two oscfbp2 mutant lines showed slight fluctuations in OscFBP1 and a b c Fig. 3 qPCR analysis of OscFBP and OsSPPs in the basal portions of the shoots. The qPCR analyses were performed on the OscFBP1 and OscFBP2 genes (a), the OsSPP1, OsSPP2, and OsSPP3 genes (b) and the control Actin1 (c) in the basal portions of the shoots of wild-type plants (WT: black column) and gs1;2 mutants (gs1;2: gray column). Seedlings were grown hydroponically in the presence of 1 mM NH 4 Cl until the fourth leaf stage. Mean values with the SE for four independent plants are shown. Asterisks denote statistically significant differences between the WT and the gs1;2 mutants (*, P < 0.05 by Student's t-test) Actin1 expression in the basal portions of the shoots (Fig. 6b). To confirm the effect of either OscFBP2 absence or repression, an assay of total cFBPase activity was carried out in two oscfbp2 mutant lines. The cFBPase activity of both oscfbp2 mutant lines decreased by~30% in the basal portions of the shoots (Fig. 6c). These results indicate that cFBPase2 did not function in either mutant line.
The two oscfbp2 mutant lines had reduced sucrose content in the basal portions of their shoots but not in their leaf blades ( Fig. 6d; Additional file 4: Figure S4a). The glucose and fructose levels were not significantly  Figure S4b, c). When the primary tillers began to appear under soil culture conditions iñ 60% of the wild type at the fourth leaf stage, they appeared in~40% or 15% of oscfbp2-m1 and oscfbp2-m2, respectively. Tillers only gradually appeared in oscfbp2 mutants. Tiller number and leaf stage were measured in hydroponically raised oscfbp2 mutants during their early growth stages. Two oscfbp2 mutant lines receiving 1 mM NH 4 + had~30% reduction in tiller number at 50 days after germination (Fig. 6e). The leaf stage was not significantly affected in either oscfbp2 mutant 50 days after germination (Fig. 6f ). When the two oscfbp2 mutant lines were grown for extended periods, the reduction in tiller number gradually declined and finally disappeared.
Therefore, the effect that the lack of cFBPase2 in the basal portions of the shoots has on tiller growth is most pronounced during the early growth stages.

Discussion
Our metabolomic analysis showed that the lack of GS1;2 decreases both nitrogen-and carbon metabolites (sucrose and glucose 6-phosphate) in the basal portions of rice shoots at the seedling stage (Figs. 1 and 2; Additional file 1: Figure S1). The decrease in glucose 6-phosphate content may have been caused by the decrease in sucrose. The decrease in sucrose content was successfully restored to wild-type levels by complementation (Fig. 2). The reductions in sucrose and nitrogen metabolism found here corroborated the previous transcriptomic analysis (Ohashi (Fig. 3). The expression level of OscFBP2 was dominant relative to the other isogenes in the basal portions of the shoots (Fig. 3). The lack of GS1;2 also reduced OscFBP2 expression signals in the phloem companion cells of the nodal vascular anastomoses, the internodal parenchyma cells, and the lower leaf sheath regions in the basal portions of the shoots (Fig. 4). The importance of cFBPase in sucrose biosynthesis has been established (Strand et al. 2000). The cFBPase antisense mutants of Arabidopsis presented with reduced shoot growth and sucrose content compared to the wild type (Strand et al. 2000). Rice mutants lacking cFBPase2 were isolated to determine whether OscFBP2 downregulation was responsible for the low sucrose content. Oscfbp2 mutants, which cannot translate functional cFBPase2 protein, displayed low sucrose content in the basal portions of their shoots ( Fig. 6; Additional file 4: Figure S4). This phenotype resembled that of the gs1;2 mutants (Figs. 1 and 2; Additional file 1: Figure S1.). Therefore, low sucrose metabolism in the basal portions of the gs1;2 mutant shoots was largely caused by OscFBP2 downregulation. Ohashi et al. (2015b) showed that GS1;2 localizes in the phloem companion cells of the nodal vascular anastomoses. This structure circulates metabolites to sink tissues, such as axillary buds, in the basal portions of the shoots (Hoshikawa 1989). OscFBP2 and OscFBP1 were expressed in the same vascular tissues as GS1;2 ( Fig. 4; Ohashi et al. 2015b). NH 4 + supply induced OscFBP2 but not OscFBP1 in the basal portions of the shoots. In the absence of GS1;2, OscFBP2 induction was not detected (Figs. 3 and 5). When inorganic nitrogen is assimilated, sucrose furnishes the carbon skeletons (Nunes-Nesi et al. 2010). Based on OscFPB2 expression and localization, cFBPase2 may provide carbon skeletons via the phloem for nitrogen assimilation. Sucrose may be synthesized in the phloem companion cells of the nodal vascular anastomoses in response to nitrogen availability in the basal portions of the shoots. Plants possess an intricate regulatory mechanism that coordinates nitrogen assimilation and carbon metabolism with the demands of growth and development (Nunes-Nesi et al. 2010). The control of carbon and nitrogen interactions is highly complex. It involves signals from nitrate, ammonium, and nitrogen-containing metabolites such as glutamine and signals from carbon metabolism (Coruzzi and Zhou 2001;Stitt and Krapp 1999;Miller et al. 2008). The mechanism by which NH 4 + induces OscFBP2 should be examined to elucidate the modulation of carbon and nitrogen interaction in response to changes in environmental conditions. Rice mutants lacking cFBPase1 presented with severe reductions in growth, tiller number, and sugar (sucrose, glucose, and fructose) levels in the leaf blades (Lee et al. 2008;Koumoto et al. 2013). OscFBP1 was significantly more expressed than OscFBP2 in the leaves, roots, flowers, and seeds (Lee et al. 2008). Compared with cFBPase1 deficiency, the lack of cFBPase2 may have only a limited role in tiller growth at the early growth stage. Kebrom et al. (2012) indicated that in wheat, the suppression of axillary bud outgrowth is associated with a reduction in sucrose content. Low sucrose levels may, in turn, reduce the expression of sucrose-inducible genes involved in the synthesis, gap2-mitosis, and cytokinesis phases of the cell cycle. Axillary buds were rapidly released in intact Arabidopsis plants supplemented with exogenous sucrose (Mason et al. 2014). Therefore, sucrose may serve as an important cue for axillary bud outgrowth. Reduction in tiller number at the early growth stage in the absence of cFBPase2 indicates that OscFBP2 downregulation delayed tiller development at the early growth stages of gs1;2 mutants via low sucrose availability. Early anatomical analyses reported that normal tiller development in rice depends on adequate nitrogen and carbohydrate (sugars and/or starch) levels in the basal portions of the shoots (Takahashi et al. 1956;Sato, 1959). Previous research has shown that the glutamine generated from the GS1;2 reactions contributes to tiller outgrowth via glutamine-dependent cytokinin synthesis (Ohashi et al. 2017). The findings of the present study indicated that adequate supplies of both glutamine and sucrose are required for rice tiller growth because they function as nutrients and signaling molecules in the basal portions of the shoots.

Conclusions
This study demonstrated that the lack of GS1;2 caused a disorder of carbon metabolism in the basal portions of rice shoots at the seedling stage. The reduction in the sucrose content was caused by the downregulation of OscFBP2 in the basal portions of the gs1;2 mutant shoots. Downregulation of OscFBP2 caused low rates of sucrose synthesis, which, in turn, reduced tiller numbers at the early growth stages. Therefore, sufficient sucrose and glutamine is required for tiller growth in the basal portions of rice shoots.

Plant materials
The wild-type rice (Oryza. sativa L.) cultivar was 'Nipponbare'. A retrotransposon Tos17-inserted line of the gs1;2 mutant and its complementation line produced by reintroducing OsGS1;2 cDNA under the control of its own promoter were used (Funayama et al. 2013;Ohashi et al. 2015b). Seeds were germinated, and seedlings were grown hydroponically in an outdoor greenhouse until either the fourth-or seventh leaf stage. The temperature was set to 26°C during the day and the plants received supplemental light for 13 h per day. The hydroponic culture solution was replenished with 1 mM NH 4 Cl once per week as described in Ohashi et al. (2017). Longitudinal sections (5 mm length) of basal portions of the shoots including axillary buds, internodes, and shoot apical meristems were prepared after removing the primary and secondary leaves, seeds, and roots as described in Ohashi et al. (2017). The basal portions of the shoots at the fourth-or seventh leaf stage were used for sugar content determination, in situ hybridization, and qPCR analysis.
The Project for Rice Genome Research, the Ministry of Agriculture, Forestry and Fisheries of Japan (mutant panel: https://tos.nias.affrc.go.jp/~miyao/pub/-tos17/) provided 20 seeds from each of two rice (Oryza sativa L. cv. Nipponbare) lines (NF2064 and NF9351). In these, the retrotransposon Tos17 may have been inserted into the OscFBP2 gene. To isolate OscFBP2 mutants, wild type and Tos17-inserted rice line seedlings were germinated and raised on hydroponic culture solution (Kamachi et al. 1992) in a greenhouse for 2 weeks. Thereafter, genomic DNA was extracted. The Tos17 insertion points in the OscFBP2 genes of NF2064 and NF9351 were localized by PCR using specific primer pairs against OscFBP2 genomic DNA (Additional file 5: Table S1). They were fully sequenced using the procedure of previous experiments (Funayama et al. 2013). Lines NF2064 and NF9351 were identified as Tos17-insertion mutants for OscFBP2 and named oscfbp2-m1 and oscfbp2-m2, respectively. The oscfbp2-m1 and oscfbp2-m2 lines were segregated homozygotes for Tos17 and used as OscFBP2 mutants. The oscfbp2 mutants were grown hydroponically in the presence of 1 mM NH 4 Cl until the fourth leaf stage. They were used to determine sugar content and for qPCR analyses. The tiller number and leaf stage were measured for the oscfbp2 mutants until 50 days after germination.

Cloning of OscFBP and OsSPP cDNAs
Molecular biological experiments were carried out according to previous protocols (Ishiyama et al. 2004;Ohashi et al. 2015a, b). The cDNAs encoding cFBPase (OscFBP1 and OscFBP2) and SPP (OsSPP1, OsSPP2 and OsSPP3) were isolated by RT-PCR. They were used in the first-strand cDNA synthesis along with specific primer pairs designed according to the rice nucleotide sequence observed in the Rice Annotation Project (RAP-DB; http://rapdb.dna.affrc. go.jp/index.html) and the Rice Genome Annotation Project (RGAP; http://rice.plantbiology.msu.edu/index.shtml). Codes of each gene are listed in Additional file 5: Table S1. Total RNA extraction, RT, and PCR were performed according to previous experiments (Ohashi et al. 2015b). The amplified PCR products were cloned into pCR-Blunt II-TOPO (Thermo Fisher Scientific K.K., Tokyo, Japan) and fully sequenced.

qPCR analysis
The qPCR analysis was performed according to Ohashi et al. (2017). Gene-specific primers for OscFBPs and OsSPPs used in the qPCR analysis are listed in Additional file 1: Table S1. Those used for Actin1 are described in Ohashi et al. (2017).

In situ hybridization of OscFBP genes
RNA probes were prepared as follows. DNA fragments 200 bp long from the 3′-untranslated regions of OscFBP1 and OscFBP2 were amplified by PCR using specific primers (Additional file 5: Table S1). The amplified PCR product for each gene was cloned into pCR®4Blunt-TOPO® (Thermo Fisher Scientific K.K.). Digoxigenin (DIG)-labeled sense-and antisense probes were synthesized with T3 and T7 polymerases (Promega Corporation, Madison, WI, USA) and a DIG RNA labeling kit (Roche Diagnostics K.K., Tokyo, Japan) according to Ohashi et al. (2015a). In situ hybridization analysis of the basal portions of the shoots were performed as described in Ohashi et al. (2015aOhashi et al. ( , 2017.

Metabolite profiling analysis
The shoot basal portions of the wild type and gs1;2 mutants at the fourth leaf stage were used to perform metabolite phenotyping and profiling by gas chromatography-time-of-flight/mass spectrometry (GC-TOF-MS) (Kusano et al. 2007(Kusano et al. , 2011. We used six to eight independent wild type and gs1;2 mutant plants for the analysis. Extracts of 10-30 mg (fresh weight, FW) samples were analyzed by using GC-TOF-MS. Extracts of the equivalent of 60 μg FW of each sample was derivatized and then injected into the GC-TOF-MS. Sample pre-treatment and normalization were performed as described by Kusano et al. (2007). The unprocessed data were imported in succession to MATLAB v. 7.0.4 (Mathworks, http://www.mathworks.com/). All alignments and metabolite identification were performed as described by Kusano et al. (2007Kusano et al. ( , 2011.

Determination of sugar content
Frozen rice tissue samples were powdered in liquid N 2 using a Mixer Mill MM300 (Qiagen K.K., Tokyo, Japan) set to 25 Hz and run for 2 min, then homogenized in 10 volumes of 10 mM HCl. The homogenate was centrifuged and the supernatant fraction was filtered though Amicon Ultra 3 K filter (Merck Millipore Corporation, Darmstadt, Germany) and the resulting filtrate was used for analysis of soluble sugars. Sugars (glucose, fructose, and sucrose) were separated on a COSMOSIL Sugar-D column (4.6 mm diameter × 250 mm length) (Nacalai Tesque, Kyoto, Japan), and isocratically eluted with 100% solvent A (acetonitrile: H 2 O = 75:25) for 17.5 min after injection. After elution, the column was washed with 100% solvent B (100% H 2 O) for 12.5 min and equilibrated with 100% solvent A for 10 min. The flow rate was 0.8 mL min − 1 . The column temperature was maintained at 30°C by incubation in a Hot Pocket (Thermo Scientific Japan, Yokohama, Japan). Column pressure was~6.5 MPa. The sugar peaks were detected by UV/VIS-155 (Gilson, Inc., Middleton, WI, USA; length 192 nm) and analyzed by TRILUTION® LC (Gilson, Inc., Middleton, WI, USA). Linearity of the detector response to the sugar concentration was tested in the range of 1-100 μg per sugar standard.

Measurement of cytosolic fructose-1,6-bisphosphatase activity
Cytosolic fructose-1,6-bisphosphatase activity was determined according to the method of Sharkey et al. (1991) with a slight modification. The basal portions of rice shoots harvested at the four-leaf stage were frozen at − 80°C then milled with a mortar and pestle in 20 mM HEPES (4-(2-hydroxyethyl)-1-piperazine ethanesulfonic acid)-NaOH (pH 7.5), 100 mM KCl, and 0.5 mM EDTA. Each extraction solution was centrifuged at 13,000×g and 4°C for 10 min. The supernatant was added to a reaction mixture consisting of 100 mM HEPES-NaOH (pH 7.5), 100 mM KCl, 5 mM MgCl 2 , 0.5 mM EDTA, 0.5 mM NADP, one unit of phosphoglucoisomerase (Sigma-Aldrich Corp., Tokyo, Japan), and two units of glucose-6-phosphate dehydrogenase. The reaction was initiated by adding 40 μM fructose-1,6-bisphosphate (Sigma-Aldrich Corp., Tokyo, Japan). It was then spectrophotometrically monitored at 340 nm for increases in NADPH concentration for 5 min after the start of the enzymatic reaction. The amount of product formed was calculated from the increase in absorbance using the NADPH extinction coefficient of 6220 L mol − 1 cm − 1 .

Additional files
Additional file 1: Figure S1. Sugar levels in the leaf blades of wild-type rice and gs1;2 mutant rice. Seedlings were grown hydroponically in the presence of 1 mM NH 4 Cl until the fourth leaf stage. Sucrose, fructose, and glucose levels were measured in the third expanded leaf blades of wild-type (WT, black column) and gs1;2 mutant (gs1;2, gray column) rice. Mean values plus the SE of five independent plants are indicated. Statistically significant differences were not observed between the WT and the gs1;2 mutants (P < 0.05 by Student's t-test). (PDF 844 kb) Additional file 2: Figure S2. In situ hybridization of OscFBP1 in the basal portions of the shoots. Longitudinal sections of the basal portions of the shoots were prepared from wild-type (WT) (a-c, g-i) and gs1;2 mutant (gs1;2) (d-f) rice grown hydroponically in the presence of 1 mM NH 4 Cl until the fourth leaf stage. The antisense probe for the OscFBP1 transcript was hybridized with the longitudinal sections from the shoot basal portions of wild-type (a-c) and gs1;2 mutant (d-f) rice. The sense probe for the OscFBP1 transcript was hybridized with the sections from the shoot basal portions of the wild type (g-i) as a negative control. The phloem companion cells of the nodal vascular anastomoses (a, d, g), the internodal parenchyma cells (b, e, h), and the leaf sheath (c, f, i) are shown. The red arrowhead in (a, d) indicates the hybridization signal of the OscFBP1 transcript in the phloem companion cells of the nodal vascular anastomoses. Abbreviation: cc, companion cell. Scale bars: 20 μm (a, b, d, e, g, h) and 50 μm (c, f, i). (TIF 5981 kb) Additional file 3: Figure S3. RT-PCR analysis and alignment of the deduced amino acid sequences of the cFBPase2 polypeptides. (a) The RT-PCR analysis of OscFBP2 was performed using the shoot basal portions of wild-type (WT) and oscfbp2 mutant (oscfbp2-m1) rice at the fourth leaf stage. OscFBP2 gene-specific primers for RT-PCR analysis are shown in Additional file 5: Table S1. (b) Deduced amino acid sequences between cFBPase2 polypeptides from wild-type (WT) and oscfbp2 mutant (oscfbp2-m1) rice are denoted as one-letter codes representing each amino acid residue. Asterisks in the cFBPase2 of oscfbp2-m1 indicate amino acid residues identical to those of the wild-type cFBPase2. Deletions of thirty-eight amino acid residues were observed in the deduced amino acid sequences of the cFBPase2 of oscfbp2-m1. (PDF 888 kb) Additional file 4: Figure S4. Analysis of sugar contentsin the OscFBP2 mutants. Seedlings were grown hydroponically in the presence of 1 mM NH 4 Cl until the fourth leaf stage. Sucrose levels (a) were measured in the third expanded leaf blades of wild-type (WT, black column) and oscfbp2 mutant (oscfbp2-m1, gray column; oscfbp2-m2, open column) rice. Glucose (b) and fructose (c) levels were measured in the shoot basal portions (BP) of the WT and the oscfbp2 mutants. Mean values plus the SE of five independent plants are indicated. Statistically significant differences were not observed between the WT and the oscfbp2 mutants (P < 0.05 by Student's t-test). (PDF 868 kb) Additional file 5: Table S1. Primers used in this study. a Rice Annotation Project (http://rapdb.dna.affrc.go.jp/index.html). b Rice Genome Annotation Project (http://rice.plantbiology.msu.edu/index.shtml). (PDF 52 kb) Abbreviations AS: Asparagine synthetase; cFBP: Cytosolic fructose-1,6-bisphosphate; cFBPase: Cytosolic fructose-1,6-bisphosphatase; FBP: Fructose-1,6-bisphosphate; GS: Glutamine synthetase; IPT: Adenosine phosphate isopentenyltransferases; qPCR: Quantitative real-time PCR; RT: Reverse transcription; SPP: Sucrose phosphatase