Overexpression of the chloroplastic 2-oxoglutarate/malate transporter in rice disturbs carbon and nitrogen homeostasis

The chloroplastic oxaloacetate/malate transporter (OMT1 or DiT1) takes part in the malate valve that protects chloroplasts from excessive redox poise through export of malate and import of oxaloacetate (OAA). Together with the glutamate/malate transporter (DCT1 or DiT2), it connects carbon with nitrogen assimilation, by providing α-ketoglutarate for the GS/GOGAT reaction and exporting glutamate to the cytoplasm. OMT1 further plays a prominent role in C4 photosynthesis. OAA resulting from PEP-carboxylation is imported into the chloroplast, reduced to malate by plastidic NADP-MDH, and then exported for transport to bundle sheath cells. Both transport steps are catalyzed by OMT1, at the rate of net carbon assimilation. Therefore, to engineer C4 photosynthesis into C3 crops, OMT1 must be expressed in high amounts on top of core C4 metabolic enzymes. We report here high-level expression of ZmOMT1 from maize in rice (Oryza sativa ssp. indica IR64). Increased activity of the transporter in transgenic rice was confirmed by reconstitution of transporter activity into proteoliposomes. Unexpectedly, over-expression of ZmOMT1 in rice negatively affected growth, CO2 assimilation rate, total free amino acid contents, TCA cycle metabolites, as well as sucrose and starch contents. Accumulation of high amounts of aspartate and the impaired growth phenotype of OMT1 rice lines could be suppressed by simultaneous over-expression of ZmDiT2. Implications for engineering C4-rice are discussed.


Introduction
6 Rootrainers (http://rootrainers.co.uk/). After 2 weeks, plants were transplanted to 7 L 167 soil pots. Plants were grown at Heinrich-Heine University (HHU) Düsseldorf,168 Germany under semi-controlled greenhouse conditions (16h day/8h night and 25°C). 169 Assessment of leaf gas exchange, as well as metabolite, C:N ratio, total free amino 170 acids, and transporter activity measurements were performed at HHU. cycles of 95°C for 15 sec and 60°C for 30 sec followed by the measurement of the 197 melting curve after 40 cycles for primer specificity. The primer efficiency was 198 calculated as described by Udvardi et al., (2008) using different dilutions of cDNA 199 together with the a highly stable housekeeping gene from rice, OseEF-1a 7 (Os03g0177500) that was identified in a previous experiment by Jain et al., (2006). 201 The mean normalized expression (MNE) for calculation of average CT was used as 202 described by Simon (2003). 203 204 205 To detect the mRNA expression of ZmOMT1 and ZmDiT2 genes in OMT1/DiT2 206 double cross lines, RT-PCR analysis was performed in 8 week-old plants. RNA was 207 extracted from leaf materials using TRIzol reagent (Invitrogen, USA) and treated with 208

Real-time PCR (RT-PCR)
DNase (Promega, USA). A 1 µg RNA was used to synthesize cDNA using a first-209 stand cDNA synthesis kit (Roche Diagnostics, Switzerland). The cDNA was 210 normalized to 100 ng µl -1 and used for PCR analysis in a 10 µl reaction with gene- The presence of the AcV5-tagged ZmOMT1 protein in leaf membrane extracts of rice 230 lines overexpressing ZmOMT1 was checked by fractionating the isolated protein on 231 12% SDS-PAGE gel, followed by Western-blot analysis. Primary mouse anti-AcV5 232 tag 1:2,000 (Abcam plc, UK) and peroxidase-conjugated secondary (Goat anti-8 Mouse IgG (H+L) HRP, 1:2,500, ThermoFisher Scientific, Germany) antibodies were 234 used for the detection of the AcV5 tag. Visualization of the stained protein on 235 nitrocellulose membranes was carried out by a LAS-4000 Mini luminescence image 236 analyzer (GE Healthcare, Germany) using the ECL Western Blotting Detection 237 Reagents (GE Healthcare, Germany). 238 239 DNA blot analysis 240 Genomic DNA was extracted from the leaves of mature rice plants using the 241 potassium acetate method as described by Guillemaut and Maréchal-Drouard, 242 (1992

253
The seventh leaf at the mid-tillering stage was fixed and prepared for 254 immunolocalization analysis as described in Lin et al., (2016 Total leaf membrane protein isolation 263 The fully expanded 3 rd leaf of rice at the mid-tillering stage was used for protein 264 extraction. Leaves were homogenized to a fine powder using a nitrogen-cooled 265 mortar and pestle. The powder was used as starting materials and total leaf 266 membrane protein was isolated using an extraction buffer consisting of 250 mM Tris 267 (HCl, pH=8.5), 25 mM EDTA, 30 % (w/v) sucrose, 5 mM DTT, and appropriate 268 protease inhibitors. Two subsequent centrifugation steps at 10,000 g and 100,000 g 269 were then performed, using a bench top centrifuge and ultra-centrifuge, respectively. 270 Ultimately, the isolated membrane was re-suspended in 50 mM HEPES (KOH, pH 271 7.5), 5 mM EDTA, 2 mM DTT together with protease inhibitors (for detailed 272 procedure see Furbank et al., 2001 andRoell et al., 2017)

Reconstitution of total leaf membrane into liposomes 277
In-vitro analysis of transporter activity was carried out using a freeze-thaw-sonication 278 reconstitution procedure in concert with forward exchange of the substrate (Palmieri 279 et al., 1995). Following reconstitution, the proteoliposomes were preloaded with 280 unlabeled malic acid to a final concentration of 30 mM (pH=7.5). Reconstituted 281 proteins were separated from the non-reconstituted ones utilizing the size-based 282 column chromatography technique (Sephadex G-25M columns (PD-10 column, GE 283 Healthcare, USA) (for detailed procedure see Roell et al., 2017). 284 285

Radioactive labeled [ 14 C]-malate uptake measurement
286 Uptake of radiolabeled substrates in counter-exchange with non-labeled substrates 287 was carried out during the course of one hour at six different time points (2,4,8,16,288 32 and 64 minutes). The reaction was started by adding 950 µl of proteoliposomes 289 into 50 µl of [ 14 C]-malate diluted in transport medium (7 mM malic acid, pH=7.5), and 290 stopped at each of the above-mentioned time points by loading an 150 µl aliquot of 291 the reaction mixture to an anion exchange resin column (acetate form, 100-200 292 mesh, Dowex AG1-X8 Resin, Bio-Rad, UAS). The resin column was previously 293 equilibrated five times using 150 mM sodium acetate (pH=7.5). Unincorporated [ 14 C]-294 malate was replaced by acetate in the resin column and the incorporated label was 295 washed through a scintillation vial containing 10 ml Rotiszint® eco plus scintillation 296 cocktail (Carl Roth, Germany). Finally, the uptake of radio-labeled substrate was 297 measured as counts per minute (CPM) by scintillation counting. To correct for 298 background and false positives, the entire experiment was repeated using 299 proteoliposomes without pre-loading of the substrate of interest (for detailed 300 procedure see Roell et al., 2017). The uptake data were further assessed relative to 301 both internal standards and total protein content (mg) in each sample. Related 302 graphs were made using the one-phase association equation in GraphPad Prism 6 303 (http://www.graphpad.com/prism/ prism.htm). 304 305 Photosynthetic CO 2 assimilation, light response and dark 306 respiration rates 307 Two individual fully expanded leaves per plant and three plants per line were 308 measured for leaf photosynthetic CO 2 assimilation and dark respiration rates during 309 the tillering stage using a LI-6400XT portable photosynthesis system (LI-COR 310 Biosciences, USA) in which a single leaf was clamped in the standard LI-COR leaf 311 chamber. Measurements were performed on the mid-portion of the leaf blade 312 between 08:00 h and 13:00 h at a constant airflow rate of 400 µmol s -1 , leaf 313 temperature of 30°C, and a leaf-to-air vapor pressure deficit of between 1.0 and 1.5 314 kPa. Leaves were acclimated in the cuvette for 30 min before measurements were 315 started. The response curves of the net rate of CO 2 assimilation (A, µmol CO 2 m -2 s -1 ) 316 to changing intercellular CO 2 concentration (C i , µmol CO 2 mol -1 ) were acquired by 317 decreasing C a (CO 2 concentration in the cuvette) from 2,000 to 20 µmol CO 2 mol -1 at 318 a photosynthetic photon flux density (PPFD) of 2,000 µmol photons m -2 s -1 . The 2 % 319 oxygen entering the cuvette was set by mixing different concentration of nitrogen and 320 oxygen in the CO 2 free airstream through two mass flow controllers (model GFC17, 321 Aalborg Mass Flow Systems, USA) at a flow rate of 1.5 ml m -1 . Maximum Rubisco 322 activity (Vcmax) and maximum electron transport activity (Jmax) were determined 323 using the PsFit Model (Bernacchi et al., 2001(Bernacchi et al., , 2003Farquhar et al., 1980). The light-324 response curves were measured by increasing the PPFD from 20 to 2,000 µmol 325 photons m -2 s -1 at a C a of 400 µmol CO 2 mol -1 . The carboxylation efficiency (CE, µmol 326 CO 2 m -2 s -1 µmol CO 2 mol -1 ), CO 2 compensation point (Γ, µmol CO 2 m -2 s -1 ) and 327 quantum yield ( , mol CO 2 mol -1 photons) were calculated as described by Lin et al., 328 (2016). The dark respiration rate (R d ) measurements were made on leaves in 329 darkness following an acclimation at a photosynthetic photon flux density (PPFD) of 330 1,000 µmol photons m -2 s -1 for 10 min at a C a of 400 µmol CO 2 mol -1 . The dark 331 respiration rate (R d , µmole CO 2 m -2 s -1 ) was calculated over a period of 1100-1200 s 332 in the dark. In order to measure photosynthetic CO 2 assimilation and collect the samples for 339 metabolite analysis under steady-state conditions, a custom gas exchange chamber 340 was interfaced with a LI-COR 6400XT portable photosynthesis system (LI-COR 341 Biosciences, USA) (Fig. S1B). The custom gas exchange chamber encased the leaf 342 to be measured within a low-gas permeable sausage casing (5 cm diameter 343 Nalophan, Kalle GmbH, Germany) to allow for rapid freeze-quenching of the sample. 344 The chamber was constructed using two stainless-steel pipe sections fitted with 345 Swagelok connections to the LI-COR sample line, one of which was capped on the 346 end with a welded end cap. Prior to each measurement, a ~20 cm section of 400 or 1,000 µmol CO 2 mol -1 . Leaf surface area was determined by taking a 359 photograph and analyzing in ImageJ v1.51m9 (Schneider et al., 2012). Leaf 360 temperature was not controlled but ranged between 25-27°C as determined from 361 energy balance calculations. Leaves were sealed within the chamber until steady-362 state conditions were reached (as determined from a constant net CO 2 fixation rate) 363 and gas exchange measurements logged. After logging gas exchange data, the 364 liquid nitrogen-cooled piston was inserted rapidly through the ring light onto the leaf 365 and onto a plastic anvil, and then transferred rapidly to an aluminum-foil pouch and 366 into liquid nitrogen. To avoid potential diurnal artifacts, all measurements (genotypes 12 and CO 2 treatments) were randomized and performed only during the peak 368 photosynthetic activity of the rice plants between 9:00 am to 3:00 pm. 369 370

Metabolite analysis (GC/MS) 371
The GC/MS-based metabolite measurements were performed as described by Fiehn, 372 (2007), using ribitol as an internal standard. Leaf samples were collected by rapid 373 freeze-quenching from the custom gas exchange chamber describe above. Freeze-374 quenched tissue was ground into a fine powder in liquid nitrogen using a mortar and 375 pestle. Extracted metabolites were injected into a gas chromatograph (Agilent 7890B 376 GC System, Agilent Technologies, USA) that was in line with a mass spectrometer 377 (Agilent 7200 Accurate-Mass Q-TOF GC/MS, Agilent Technologies, USA). Metabolite 378 peaks were evaluated using Mass Hunter Software (Agilent Technologies, UAS). The 379 relative amount of each metabolite was calculated from the peak area, taking into 380 account both the initial fresh weight used for extraction and the internal standard. 381 382

Total free amino acid (FAA) content
383 FAA contents were measured using the Ninhydrin colorimetric method as described 384 by Smith and Agiza, (1951), with minor changes. Briefly, 10 µl of metabolite extract 385 together with 40 µl of methanol: water mixture (2.5:1 ratio) was added to 50 µl of 1 M 386 citrate (NaOH, pH=5.2) and 100 µl of 1% (w/v) Ninhydrin (prepared in methanol: 387 H 2 O, 2.5:1 ratio), and then heated to 95°C for 20 min. The solution was then 388 transferred to a micro-well plate after a short centrifugation of 10 sec at 10,000 rpm. 389 The total amino acid content was then measured in a Synergy HT plate reader 390 After centrifugation, the supernatant was discarded, and the pellet was air dried and 13 resuspended in 500 µl of water. The starch sample was gelatinized by boiling for 4 402 hours and hydrolyzed overnight at 37°C with 0.5 U of amyloglucosidase and 5 U of α -403 amylase. The starch content was measured as described in Smith and Zeeman, 404 (2006). The soluble fraction containing sucrose was neutralized to pH=6 with 405 neutralization buffer (2 M KOH, 0.4 M MES,0.4 M KCl). After centrifugation at 21,100 406 g for 10 min at 4°C, the supernatant was transferred into a new tube and the 407 remaining insoluble potassium perchlorate was discarded. The supernatant was 408 assayed for sucrose content by enzymatic determination as described by Smith and 409 Zeeman, (2006). Rice seeds were germinated in petri dishes in distilled water for 4 days and then 421 placed on a floating net in distilled water in a 19 L bin in greenhouses at the 422 University of Toronto. Seedlings were fertilized with 1/3 strength hydroponic media at 423 day three after transfer and then with full strength media every 4 days (Makino and 424 Osmond, 1991). Plants were sampled from 09:30 am to 11:00 am when day length 425 was over 11.5 h and light intensity in the unshaded greenhouse regularly exceeded 426 1,400 μ mol photons m -2 s -1 . The middle section of the most recently fully expanded 427 leaf was dissected into 2 mm pieces and prepared for transmission electron 428 microscopy as previously described Khoshravesh et al., (2017). Leaf sections were 429 fixed in 1 % glutaraldehyde, 1 % paraformaldehyde in cacodylate buffer (pH=6.9) and transcript levels observed as in wild-type rice (Fig. 1A). ZmOMT1 transcripts 464 accumulated in all three lines with the highest levels in OMT1-79 and the lowest in 465 OMT1-80 (Fig. 1A). To test whether the high amounts of ZmOMT1 mRNA in the 466 transgenic lines was accompanied by increased transporter protein abundance, the 467 amounts of ZmOMT1 protein in extracted total membrane leaf protein were examined 468 via Western-blot, taking advantage of the C-terminal AcV5-tag. The ZmOMT1 protein was clearly detectable in all three lines (OMT1-79, OMT1-80 and OMT1-87) by 470 immunoblotting (Fig. 1B). As with the transcript levels, OMT1-79 and OMT1-87 lines 471 accumulated more ZmOMT1 protein than the OMT1-80 line. We further examined 472 the spatial localization of ZmOMT1 in the transgenic lines by immunolocalization. Fig.  473 1C shows that the ZmOMT1 protein accumulated primarily in chloroplasts of M cells. 474 Collectively, these data show that the ZmPEPC promoter drives expression of 475  To test whether expression of the ZmOMT1 transgene led to increased OMT1 495 transporter activity in transgenic lines, we measured malate counter-exchange 496 activity in liposomes reconstituted with membrane proteins isolated from wild-type 497 and overexpressing lines ( Fig. 2A). We detected significantly higher malate-malate 498 counter-exchange activity in liposomes reconstituted with membrane proteins from 499 overexpression lines as compared to liposomes reconstituted with membrane 500 proteins isolated from the wild types. These data clearly indicate that the 501 recombinantly introduced ZmOMT1 transporter protein is active in rice (Fig. 2B). 502 503 504 505

Slower growth and leaf lesion phenotypes of OMT1 lines 512
The transgenic plants with the highest ZmOMT1 protein levels (OMT1-79 and OMT1-513 87) displayed perturbed phenotypes at the whole plant level. The OMT1-79 and 514 OMT-87 lines were shorter ( Fig. 3A and Table 1) than wild-type and displayed 515 lesions in mature leaves in IRRI (Fig. 3B). An ELISA test for detection of infection 516 caused by tungro virus was negative (data not shown), indicating that the lesions 517 were not caused by tungro virus infection. The OMT1-80 line that accumulates lower 518 levels of ZmOMT1 (Fig. 1) had more and longer tillers compared to wild-type (Fig. 3A  519 and Table 1) and did not have a lesion mimic phenotype (Fig. 3B). Despite the 520 different lesion mimic phenotypes, chlorophyll content was similar in the youngest 521 fully expanded leaves of all three transgenic lines and wild-type (Table 1)

Chloroplast ultrastructure is perturbed in OMT1 transgenic lines
The macroscopic and physiological phenotypes of OMT1 lines were accompanied by ultrastructural changes in M cell chloroplasts. In contrast to wild-type plants, M cell chloroplasts of the OMT1 lines developed a peripheral reticulum (PR; Fig. 5) which is an internal network of tubules and vesicles continuous with the chloroplast inner membrane of chloroplasts (Rosado-Alberio et al., 1968, Laetsch, 1974. Plastoglobules (PG), not observed in wild-type plants, were also present in chloroplasts of the over-expressing lines. PGs are lipid microcompartments posited to function in lipid metabolism, redox and photosynthetic regulation and thylakoid repair and disposal during chloroplast biogenesis and stress (Rottet et al., 2015;van Wijk and Kessler, 2017).

CO 2 assimilation rate and leaf metabolite profiles of transgenic lines
The photosynthetic rate of the older ZmOMT1 transgenic plants (50-55 days old) measured in our custom-build gas exchange cuvette (Fig. S1B) was affected more than that of younger ones (30-35 days old) ( Fig. 6A and B), in which the photosynthesis rate became significantly lower in ZmOMT1 transgenic lines under ambient CO 2 concentration (400 ppm) (Fig. 6 B). The photosynthetic rate was partially restored under high CO 2 concentration (1000 ppm) for older plants (Fig. 6B).

Metabolite profiles of ZmOMT1 lines and wild-type rice reveal altered steady state pools of TCA intermediates and aspartate
The metabolic state of 30-35-day-old ZmOMT1 transgenic rice lines and wild-type under different CO 2 conditions was examined using GC/MS analysis. Large differences were observed among the measured metabolites of the mitochondrial tricarboxylic acid cycle between the transgenic lines and wild-type. Malic acid, fumaric acid, iso-citric acid, succinic acid, and α -ketoglutarate were significantly lower in all ZmOMT1 transgenic rice lines than wild-type under different CO 2 concentrations (Fig. 7A). Among photorespiratory intermediates, only glyceric acid displayed a lower amount in OMT1 lines. Others, such as glycolic acid, glycine, and serine were similar to the wild type or tended to be higher, in some cases significantly (Fig. 7B). Of the substrates transported by OMT1 and DiT2, apparently, aspartic acid was significantly increased in the overexpression lines (Fig. 7C). Malic acid and α ketoglutarate, as previously mentioned, were significantly lower and glutamic acid remained unchanged for all three OMT1 transgenic rice lines in comparison with wild-type under different CO 2 concentrations ( Fig. 7A and 7C). We further calculated the aspartate/malate ratio for all transgenic rice lines and compared to wild-type. As shown in Figure 7D, the aspartate to malate ratio was significantly higher in transgenic ZmOMT1 lines relative to wild-type under different CO 2 concentrations.

Total free amino acids, carbon:nitrogen ratios, and carbohydrate contents are decreased in leaves of ZmOMT1 lines
The absolute FAA contents of ZmOMT1 lines and wild-type rice were determined to assess the effect of altered plastidial dicarboxylate transport capacity on amino acid metabolism. Levels were lower in older plants of ZmOMT1 lines (50-55 days old) under all CO 2 concentrations but were significantly decreased under ambient CO 2 (400 ppm) compared to wild-type rice (Fig. 8A). As plants aged, the C/N ratio also decreased significantly in ZmOMT1 transgenic lines but the δ 13 C value did not differ between wild-type and transgenic lines (Fig. 8B). Sucrose and starch amounts were significantly reduced in the OMT1 lines compared to wild-type plants (Fig. 8c).

Simultaneous expression of ZmOMT1 and ZmDiT2 in transgenic rice lines restored the wild-type growth phenotype
We hypothesized that the phenotypes observed in rice lines overexpressing and ZmDiT2 were expressed ( Fig. S5 and S6). Notably, double transgenic lines displayed similar physiological phenotypes as wild-type plants when grown under ambient conditions (Fig. 9). Leaf chlorophyll content, number of tillers, and plant height were comparable to wild-type in two of three independent ZmOMT1/ZmDiT2 double over-expressing plants.   (Pottosin and Shabala, 2016). Although PR has been reported to be present in M and bundle sheath cells of other C 3 grasses such as wheat (Szczepanik and Sowinski, 2014), this cellular feature has not been observed in other Oryza species or cultivars (Sage and Sage, 2009;Giuliani et al., 2013). The PR is also present in M and sheath cells of C 4 species of grasses and eudicots, although in comparison to C 3 grasses, the PR in C 4 species is much more abundant (Rosado-Alberio et al., 1968;Laetsch, 1968;Laetsch, 1969;Szczepanik and Sowinski, 2014).
Chloroplast envelope proliferation in association with over-expression of envelope proteins has been previously reported (Breuers et al., 2012), supporting the idea that the ZmOMT1 transporter is accumulating to high amounts in the inner envelope of M chloroplasts. Given that the presence of PR is posited to be correlated with high rates of metabolite exchange (Gracen et al., 1972a, Hilliard andWest, 1971;Laetsch, 1974;Gracen et al., 1972b), the PR phenotype in ZmOMT1 transgenic lines is consistent with the altered metabolic profiles observed. plants Arabidopsis (Kinoshita et al., 2011) and tobacco (Schneidereit et al., 2006), caused an increase in levels of 2-OG and malate and a decrease in levels of aspartate, the opposite trend to that seen in ZmOMT1 overexpressing rice plants.
Surprisingly, any disruption to OMT1 activity (either an increase or decrease) leads to lower photosynthetic rates than wild-type, suggesting that OMT1 transporter activity must be precisely regulated to maintain optimal photosynthetic performance. The reduced photosynthetic rates in ZmOMT1 transgenic rice plants reveal possible relationships between photosynthesis, photorespiration, and cellular redox status.
Differences in photosynthesis were significant in the plants measured in the Philippines and in older plants grown in Düsseldorf, Germany ( Fig. 4A and 6B). This decrease in photosynthesis is only partially explained by increases in R d (Table 2).
Interestingly, this decrease in photosynthesis could be rescued by minimizing photorespiration under some measurement and growth conditions, but not others.
Specifically, the photosynthetic rates of ZmOMT1 transgenic lines were not rescued by elevated CO 2 or reduced O 2 when measured under growth conditions in the Philippines ( Fig. 4B and 4C), but were rescued in the plants grown in Düsseldorf, Germany when measured under elevated CO 2 (Fig. 6). One major difference in these measurements was the light intensity used (2000 μ mol m −2 s −1 for the A-C i curves vs 500 μ mol m −2 s -1 for the metabolite assays), meaning that phenotypic rescue may only occur under sub-saturating light intensities. As photorespiratory rates increase, the increased demand for ATP relative to NAD(P)H pushes the redox status of the NADP + /NADPH pools to be more reduced unless processes either decrease plastidic NADPH (malate valve) or increase ATP production (cyclic electron flux around photosystem I, CEF). The oxidation of NADPH, which could be increased with increased export of malate, must be finely balanced with metabolic demand so as not to directly compete with NADPH pools needed to supply the Calvin-Benson cycle or photorespiration. Under sub-saturating light, there are numerous lines of evidence suggesting that the malate valve regulates this balance, particularly under photorespiratory conditions (Kramer and Evans, 2011;Walker et al., 2014;Shameer et al., 2019). Specially, this event leads to the reduced provision of carbon skeletons for nitrogen assimilation and to a significant reduction of the leaf C/N ratio (Fig. 8B) together with the reduction of FAA in the older OMT1 transgenic lines under 400 ppm CO 2 concentration (Fig. 8A). Principally, both carbohydrate and amino acid biosynthesis are relying on each other (Nunes-Nesi et al., 2010). Correspondingly, in all three OMT1 transgenic lines, both sucrose and starch contents were decreased significantly compared to wild-type rice (Fig. 8C). It is known that a part of the photoassimilated carbon during the day will be partitioned and stored as starch to be used later during the night as a source of energy supply for sink tissues as well as fatty acid and amino acid biogenesis (Stitt and Zeeman, 2012). On the other hand, sucrose biosynthesis is occurring during the day (from the triose-phosphate pathway) and the night (from various enzymatic reactions involved in starch degradation) (Kunz et al., 2014). Therefore, starch and sucrose metabolisms tightly depend on each other and both are orchestrated by the amount of the fixed carbon during photosynthesis. Taken together, apparently too high or too low amounts of OMT1 protein affect the coordination of the C and N assimilation pathways.

Concluding model
Our results present evidences on the crucial roles of OMT1 transporter in rice plants.
We suggest a hypothetical model (Fig. 10) in which aspartate accumulates in chloroplast of single OMT1 transgenic lines in comparison with wild-type rice (Fig.   7D). We propose that the accumulated aspartate impairs the flux between the inside and outside of the chloroplast causing the growth and photosynthetic deficiency phenotypes in single OMT1 transgenic lines. Our assumption is supported by the finding that providing an exit pathway for aspartate by introducing an additional plastidial transporter (ZmDiT2) suppresses the phenotype of OMT1 overexpression (Fig. 10). These double over-expressor OMT1/DiT2 lines grew similar as the wildtype and plant height along with numbers of tiller were recovered (Table 3). Our results indicate that coordinated expression of OMT1 and DiT2 is needed for engineering C 4 -rice plants. were originated from C 4 -maize plant.