Seed-specific expression of truncated OsGAD2 produces GABA-enriched rice grains that influence a decrease in blood pressure in spontaneously hypertensive rats
- First Online:
- 830 Downloads
Gamma-aminobutyric acid (GABA) is a four-carbon amino acid that is commonly present in living organisms and functions as a major inhibitory neurotransmitter in mammals. It is understood to have a potentially anti-hypertensive effect in mammals. GABA is synthesized from glutamate by glutamate decarboxylase (GAD). In plants, GAD is regulated via its calmodulin-binding domain (CaMBD) by Ca2+/CaM. We have previously reported that a C-terminal truncated version of one of the five rice GAD isoforms, GAD2ΔC, revealed higher enzymatic activity in vitro and that its over-expression resulted in exceptionally high GABA accumulation (Akama and Takaiwa, J Exp Bot 58:2699–2607, 2007). In this study, GAD2ΔC, under the control of the rice glutelin promoter (GluB-1), was introduced into rice cells via Agrobacterium-mediated transformation to produce transgenic rice lines. Analysis of the free amino acid content of rice grains revealed up to about a 30-fold higher level of GABA than in non-transformed rice grains. There were also very high levels of various free protein amino acids in the seeds. GABA-enriched rice grains were milled to a fine powder for oral administration to spontaneously hypertensive rats (SHRs) and normotensive Wistar-Kyoto rats (WKYs). Six weeks of administration showed that transgenic rice brings about a 20 mmHg decrease in blood pressure in two different kinds of SHRs, while there was no significant hypotensive effect in WKYs. These results suggest an alternative way to control and/or cure hypertension in humans with GABA-enriched rice as part of a common daily diet.
KeywordsGABA γ-Aminobutyric acid Glutelin promoter Glutamate decarboxylase Hypertension Transgenic rice
Hypertension is respected as one of the most critical risk factors accompanying severe diseases: when it is left untreated for a long period of time, the risk of other cardiovascular diseases will undoubtedly increase. About one billion people are now suffering from hypertension and it is expected to increase 60% by 2,025 worldwide (Kearney et al. 2005). Patients with severe hypertension have to depend upon various kinds of blood pressure-lowering drugs such as angiotensin converting enzyme (ACE) inhibitors, calcium channel blockers, diuretics and others, whose costs are increasing in ratio to total expenses of drugs each year.
Foods or diets that contain bioactive substances with anti-hypertensive effects have received more attention to date. For example, some natural peptides detected in many different kinds of foods such as egg, fish protein and milk (review, Hong et al. 2008) have ACE inhibitory potential.
Recently, γ-aminobutyric acid (GABA) has received attention as a healthy functional compound: GABA itself, a non-protein amino acid, is ubiquitously present in organisms ranging from bacteria to higher organisms like rice and humans. In mammals, GABA is distributed mainly in the brain and is known to function as a major inhibitory neurotransmitter in the central nervous system (Curtis and Johston 1974). Administration of GABA reportedly lowers blood pressure in experimental animals (Takahashi et al. 1955), and GABA-enriched products have been found to have hypotensive effects in Dahl salt-sensitive rats and spontaneously hypertensive rats (SHRs) (Abe et al. 1995; Hayakawa et al. 2004), in addition to improving brain function in senescent monkeys (Leventhal et al. 2003).
GABA is produced from glutamate via decarboxylation by glutamate decarboxylase (GAD) in the cells. Interestingly, plant GAD generally contains an extended region (about 30 amino acids long) at the C-terminus, called the calmodulin-binding domain (CaMBD), which plays a vital role in the regulation of GAD activity in vitro and in vivo (Bouché and Fromm 2004). Baum et al. (1996) analyzed the function of the CaMBD in vivo and found that over-expression of petunia GAD lacking a CaMBD in transgenic tobacco resulted in increased amounts of GABA in plant cells, but lower levels of glutamate, implying that the CaMBD plays a role in autoinhibition.
Data from the rice genome project show that there exist five GAD genes in the rice (Project IRGS 2005). OsGAD2 that is encoded in one of these genes is not able to bind to Ca2+/CaM (Akama et al. 2001). But that truncation of the C-terminal peptide confers higher enzymatic activity both in vitro and in vivo on OsGAD2, and over-expression of this enzyme resulted in an extremely large amount of GABA to be accumulated in cells (Akama and Takaiwa 2007). However, the plants that regenerated from callus cells that expressed this mutant gene (i.e., a truncated form of OsGAD2, OsGAD2ΔC) had aberrant phenotypes (dwarfism and pale green leaves) and were infertile, so there was no way to evaluate the GABA content of the next generation of the seeds. In this study, in order to avoid this situation, we used a rice storage protein promoter, the glutelin promoter (GluB-1), to obtain endosperm-specific expression of OsGAD2ΔC in order to increase the amount of GABA in the rice grain. As a result, high levels of GABA in rice grains were detected. The GABA-enriched milled rice was used for evaluating changes in blood pressure of SHRs.
Materials and methods
Vector construction and rice transformation
Plant materials and growth conditions
Oryza sativa L. cv. Nipponbare was used in this study. Rice seeds were sterilized, germinated and grown in a growth-chamber as described by Akama et al. (2001). Shoots that regenerated after plant transformation (described below) were transferred to artificial soil (Bonsol; Sumitomo Chemical Co. Ltd, Osaka, Japan) and were grown for about 2 months, at which time seeds were harvested. Rice plants of T0, T1 and T2 generations after self-pollination were cultivated in plant growth chambers (26°C with 12 h [light] and 12 h [dark]). In order to harvest a sufficient quantity of rice grains in the T3 plants for milling and administration to the rats, seedlings were planted in large pots (25 cm in diameter) containing soil in a non-contained greenhouse (40 m2 in area) in the campus of Shimane University (Matsue, Japan) under natural condition from April to September. In the cultivation of T3 generation, several parameters for rice yield were investigated.
Analysis of the transgene in rice T1 plants
Total DNA was extracted from rice seedlings grown on Murashige and Skoog solid medium (Murashige and Skoog 1962) using the CTAB method (Murray and Thompson 1980). Isolated DNA was used for Southern blot analysis, essentially according to Akama et al. (2001). DNA probe (800 bp in length) was PCR-labeled with a digoxigenin-UTP as a template for hygromycin phosphotransferase (HPT) gene using the primers 5′-CGTTATGTTTATCGGCACTTT-3′ and 5′-GGGCGTCGGTTTCCACTATCG-3′ with cycling parameters as follows: 30 cycles at 95°C (30 s), 60°C (30 s) and 72°C (60 s). In parallel, PCR analysis was performed to confirm the presence of the OsGAD2ΔC, using the plant DNAs and the Ti plasmid as templates with combination of primer set of P1 (5′-GACCAGCCAACCTTCACGCTC-3′) and P2 (5′-AAGTAACCCGGACGACGACG-3′) or P1 and P3 (5′-TTCCCGATCTAGTAACATAG-3′) under the PCR conditions of 35 cycles at 95°C (30 s), 60°C (30 s) and 72°C (45 s).
Protein analysis and in situ hybridization in the T2 seeds
Total protein was extracted from T2 generation of two different fractions (milled rice fraction and bran fraction) from rice grain 4 weeks after fertilization using extraction buffer (50 mM Tris–HCl [pH 7.5], 5 mM EDTA, 1 mM DTT, 1 mM PMSF). In parallel, seeds were sampled every week between 2 and 7 weeks after flowering in order to explore expression of transgene during seed maturation at the protein level. Protein samples were separated by 12% SDS/PAGE and transferred to a membrane and western blotted using an anti-GAD2 polyclonal antibody, as described by Akama and Takaiwa (2007). Localization of protein over-expressed in the seeds was determined utilizing in situ western hybridization as described by Qu et al. (2003).
GABA and amino acid analysis
Rice grains (T2 to T4) were homogenized using 8% (v/v) trichloroacetic acid (TCA). After centrifugation, the supernatant was extracted twice using diethylether to remove the remaining TCA. Quantitative determination of GABA content was performed essentially as described by Graham and Aprison (1966). Aliquots of a sample were mixed with reaction buffer (100 mM NaPO4 [pH 9.0], 0.5 mM NADP+, 0.6 mM α-ketoglutarate, 1.2 mM 2-mercaptoethanol, 0.005 units of GABase [SIGMA]) then incubated at 37°C for 1 h. Reduction of NADP+ to NAPDH during the two-step enzymatic reaction of GABA to succinate via succinic semialdehyde was detected fluorometrically using a fluorescence reader (Genios FL; TECAN, Salzburg, Austria). In parallel, the levels of some amino acids in polished rice (T4) were determined using an automated amino acid analyzer (JLC-300; JOEL, Tokyo, Japan).
Analysis of general composition of rice T4 grains
Basic composition such as water, protein, fat, carbohydrate, and mineral and vitamin content of the grains from non-transformed Nipponbare (NT) and GABA-enriched rice was analyzed by the Food Analysis Technology Center (Osaka, Japan).
Oral administration of GABA-enriched rice (T4) to rats
The experiment plan of the present study was approved by the Animal Experiment Committee of the Department of Experimental Animals, Center for Integrated Research in Science, Shimane University, Izumo, Japan. The entire experiment followed the guidelines issued by the committee. Three kinds of rats (all male and 8 weeks old) were used in this study. Two types were spontaneously hypertensive rats, SHR/NcrlCrlj (SHR) and SHR/NDmcr-cp (SHR/cp) while the remaining were normotensive Wistar-Kyoto rats (WKY). SHRs and WKYs were purchased from Charles River Japan, Inc., Yokohama, Japan. SHR/cps were kindly provided by Japan SLC, Inc., Hamamatsu, Japan. Order of animal experiments was first with SHR, second with SHR/cp and last with WKY, using transgenic rice from line 47–52 for SHR and 78–87 for SHR/cp and WKY. Rats were housed one per cage (temperature at 23 ± 2°C, humidity at 55 ± 10%, 7–19: light, 19–7: dark) and fed a standard diet (MF; Oriental Yeast Industry, Tokyo, Japan) and received tap water ad libitum. After 2 weeks of acclimation, rats were divided into four experimental groups (n = 6 rats) except for WKY. For oral administration, rice grain samples were prepared as follows: rice grains were subjected to a rice milling machine (KG-16, National Ltd, Kobe, Japan) and then processed to a fine powder with a blender, which was then suspended in distilled water. A 10-ml suspension of rice grain powder was equally administered and the amount of GABA in each rice sample was calculated per 1 kg rat fresh weight. Low dose rice containing 0.1 mg GABA, high-dose rice containing 0.5 mg GABA, 0.5 mg GABA reagent and NT rice containing 0.016 mg GABA, respectively, was administered to each corresponding group with a sonde. In the case of GABA reagent, NT milled rice powder was supplemented as an equal weight to GABA-enriched rice. An oral administration was conducted once daily continuing for 6 weeks. Once a week, the body weight and systolic blood pressure (SBP) of the rats were measured. To measure SBP, each rat was first kept at 37°C for 10 min, and then SBP of each rat was determined by the tail-cuff method using a blood pressure analyzer (BP-98A, Softron, Co. Ltd, Tokyo, Japan).
Statistical analyses of data
The data were analyzed by using Student’s unpaired t-test in the Microsoft Excel and differences were considered to be significant at P < 0.05 or P < 0.01.
Construction of GAD2ΔC controlled by a seed-specific promoter and rice transformation
Harvest of rice T4 grains in greenhouse
The number of pots for lines 47–52 and 87–78 to be cultivated were 35 and 87 in the greenhouse, respectively. After 1 month of self-pollination, seeds were harvested to pool each line. Final yields (brown grain) were 1.13 kg (47–52) and 3.59 kg (78–87). Table S1 shows yield components of the non-transformed line (NTL) and two transgenic lines. There was no significant difference in agricultural traits examined between NTL and transgenic plants.
Localization of truncated GAD2 protein in the seeds
Determination of GABA concentration
Determination of GABA content in rice seeds
NTL (μmol/g FW grain)
47–52 (μmol/g FW grain)
78–87 (μmol/g FW grain)
0.05 ± 0.04
0.50 ± 0.15
0.56 ± 0.14
0.07 ± 0.08
0.85 ± 0.35
0.42 ± 0.22
0.16 ± 0.04
2.16 ± 0.71
1.12 ± 0.77
Free amino acid and other components in the T4 generation
Quantity of free amino acids in non-transformant (NT) and transformed seeds
NT (nmol/g FW)
47–52 (nmol/g FW)
78–87 (nmol/g FW)
1,148.8 ± 484.5
952.7 ± 251.8
840.3 ± 324.3
26.2 ± 10.7
114.0 ± 22.3**
82.7 ± 25.3*
484.4 ± 17.5
992.6 ± 146.4**
800.2 ± 102.3**
974.5 ± 30.8
1,040.8 ± 151.0
1,133.4 ± 149.9
217.3 ± 7.7
275.9 ± 42.2*
166.0 ± 24.6*
37.1 ± 12.4
353.6 ± 124.8**
266.2 ± 26.2**
148.0 ± 50.2
942.5 ± 246.1**
577.7 ± 233.6*
48.4 ± 2.6
249.0 ± 43.9**
178.1 ± 25.1**
2.0 ± 1.8
4.1 ± 3.0
2.5 ± 3.1
2.0 ± 1.8
21.0 ± 16.6
19.2 ± 16.1
19.8 ± 1.8
112.2 ± 16.3**
79.1 ± 9.4**
18.0 ± 1.8
156.6 ± 23.6**
116.4 ± 11.8**
19.3 ± 0.6
87.4 ± 11.9**
65.0 ± 7.3**
12.9 ± 2.5
87.9 ± 9.6**
58.9 ± 4.0**
52.3 ± 2.9
1,445.2 ± 196.9**
687.2 ± 63.5**
30.8 ± 1.5
63.3 ± 10.5**
47.1 ± 5.8**
15.0 ± 5.8
125.8 ± 14.7**
89.2 ± 3.9**
69.2 ± 3.0
209.7 ± 31.5**
171.6 ± 26.7**
72.9 ± 51.5
393.8 ± 138.3**
259.4 ± 81.3**
3,398.9 ± 691.4
7,628.2 ± 1,501.1**
5,640.3 ± 1,144.2**
3,346.6 ± 688.5
6,183.0 ± 1,304.2**
4,953.1 ± 1,080.7**
In parallel, other components including basic compounds, minerals and vitamins were examined as summarized in supplemental Table S2. There was no significant difference in the amount of water, protein, lipid, total minerals and carbohydrate among the NT and GABA-enriched grains. Evaluating mineral content, sodium was about two times higher in NTL than both of the transgenic lines, while phosphorus, calcium, potassium and magnesium levels were relatively higher in transgenic rice.
Oral administration with milled T4 rice
In this study we established transgenic rice plants that displayed higher concentrations of GABA in the endosperm of the seeds in order to evaluate the effect of orally administered rice grains in hypertensive rats. In our previous reports (Akama and Takaiwa 2007), GAD2ΔC driven by 35S promoter showed an extremely high level of GABA in the vegetative tissues. However, due to its high accumulation, plants showed aberrant phenotype and did not reach the reproductive stage. In the present study, the mutant gene construct was controlled by rice GluB-1 promoter, which is well characterized in seed-specific expression (Takaiwa et al. 1991; Washida et al. 1999). During the vegetative stage, there was no difference in plant growth and phenotypes between transgenics and NT. Even during the seed maturation stage, analyzing several parameters for rice yield (Table S1), there was no significant difference between the number and weight of seeds per plants, and sterility. This indicates that tightly regulated GluB-1 promoter and GABA accumulation did not affect the seed-maturation process in rice plants. Over-expression of the introduced gene’s protein products in edible parts of the seeds was confirmed both with western blot and in situ histochemical analyses (Fig. 3). Likewise there have been many prior reports of the successful over-expression of GluB-1-controlled foreign genes in the rice endosperm (Takagi et al. 2005; Wakasa et al. 2006b).
GABase analysis of the transgenic seeds indicated about 0.5–2.2 μmol of GABA per gram grain, equivalent to 5–22 mg of GABA per 100 g grain in three successive generations (Table 1). Interestingly, the increasing GABA level in these successive generations from both non-transformed and transformed lines was observed. This increase in GABA level might be influenced by different growth conditions (growth-chamber and greenhouse in T1 or T2 and T3, respectively), although further analysis in the same natural conditions must be required. After milling the brown rice, the GABA content of polished rice was measured at ~0.7 and 1.45 μmol/g (from 687.2 to 1,445.2 nmol/g in Table 2), with most of the GABA content remaining. Although several groups have achieved increased GABA levels in the vegetative tissues of tobacco using a transgenic approach (Baum et al. 1996; MacLean et al. 2003), as far as we know this is the first report of extremely high levels of GABA in grains, especially in milled rice that is an ordinary form of the diet for human.
Unexpectedly, an increase in not only GABA content, but also other free amino acid content was observed in the transgenic rice (Table 2). The increase in other amino acids analyzed seems to be closely related to that in GABA: the increase in free amino acids was roughly proportional to the GABA content in two different transgenic lines tested. With the exception of aspartate (Asp), glutamate (Glu) and glutamine (Gln), other protein amino acids also accumulated in high levels. In our previous study in which GAD2ΔC was over-expressed in callus cells, we found that GABA content increased, but that the remaining protein amino acids were present at almost the same levels or lower levels than in NT calli (Akama and Takaiwa 2007). These findings show that the accumulation of GABA in rice grains, acting as sink organs, may differentially influence the composition of the amino acid pool, because both grains and calli also accumulated a large amount of GABA as well. Very recently, Akihiro et al. (2008) reported that during fruit development in tomatoes not only GAD but also α-ketoglutarate-dependent GABA transaminase (GABA-TK) activity plays a critical role in the accumulation of GABA in tomato fruit. Kisaka et al. (2006) have developed transgenic tomato fruits where GAD expression was antisense suppressed. Amino acid analysis of the fruit revealed an increase not only in Glu, but also in other free amino acids, in three out of four transgenic lines. Taking these reports into account, although it has remained unclear regarding GABA-TK activity in rice seeds, a combination of the high activity of truncated GAD2 and putative GABA-TK activity might result in the rapid turnover of Glu as an amino residue donor to a wide range of amino acids, i.e., sufficient supply of 2-oxoglutarate and other 2-oxo acids (Forde and Lea 2007). On the other hand, it was shown that over-expression of the transgene encoding α subunit of rice anthranilate synthase (AS) results not only in a marked increase in tryptophan (Trp) than is expected, but also in most of the other protein amino acid content as well (Wakasa et al. 2006a). It was speculated that Trp accumulation in rice seeds increases the transcription of genes that are involved in amino acid synthesis. It has also been indicated that Asp kinase 2 and Dof 1 transcription factor are critical for amino acid synthesis or nitrogen assimilation (Wang et al. 2001; Yanagisawa et al. 2004). On the basis of these findings, it would be tempting to speculate that GABA plays a role in regulating homeostasis of amino acid metabolism via control of key factor(s), especially in seeds. To clarify these points, further analyses of transgenic lines, i.e., assessing carbon and nitrogen assimilation and amino acid metabolism in concert with the expression of genes involved in these metabolic pathways, is required.
We demonstrated that SHRs administered a 0.5-mg dose of GABA-enriched rice grain (per kg BW) resulted in about a 20-mmHg antihypertensive effect of systolic blood pressure (SBP) both in SHR and SHR/cp. However, neither the lower dose (0.1 mg GABA) nor 0.5 mg GABA reagent showed a significant SBP-lowering effect. On the other hand, Hayakawa et al. (2004) reported that 0.5 mg GABA reagent/kg BW in SHR has a lowering effect on BP at 4–8 h after oral administration and also at 4 weeks of administration, although we cannot clearly explain the contradiction among the two experiments. Rice grains established in this study contain not only higher amount of GABA but also increase in most of protein amino acids, which may be related to induce blood pressure lowering effect (discussed in detail in below). Nevertheless, it should be noted that a decrease in SBP did not occur in WKY. It has already been shown that GABA has no lowering effect on BP in the case of humans with normal BP (Tsuchida et al. 2003). Until now, there have been many reports on the blood pressure lowering effect of GABA-containing foods (rice germ soaked in water, anaerobically treated tea, and fermented milk product) in which GABA accumulation showed about a 10–15% reduction in SBP compared with the control group of SHRs (Saikusa et al. 1994; Abe et al. 1995; Hayakawa et al. 2004). Because GABA poorly crosses the blood-brain barrier (Kuriyama and Sze 1971), it has been speculated that exogenous GABA acts within peripheral tissues such as blood vessel (Manzini et al. 1985). In order to explore the mechanism of GABA-mediated suppression of BP, we evaluated the level of ACE, nitronic oxide (NOx) and aldestrone in plasma, which are all critical parameters in blood pressure regulation. However, these were not changed significantly in the rats fed with GABA-enriched rice grains, suggesting an antihypertensive effect observed in this study might not be explained based on these parameters. Although we can only speculate to what extent GABA brings about a BP lowering effect, it is highly possible that a combination of GABA with unknown factors could induce anti-hypertensive effects in this study. It has been reported that amino acids such as arginine (Arg) and nicotianamine have an antihypertensive effect via vasodilation and ACE inhibition, respectively (Costa et al. 1998; Hayashi and Kimoto 2007). A recent study has also disclosed that amino acids have an unexpected new role in mammals, i.e., as a putative trigger for insulin secretion (e.g., Newsholme et al. 2005). Because basic components, vitamins, and minerals (except for sodium) are almost the same between Nipponbare and GABA-enriched rice, the transgenic rice established in this study possess not only GABA but also many kinds of amino acids, implying one or several amino acids could cooperatively interact with GABA, inducing an antihypertensive effect in SHR with long-range oral administration. It would be interesting to evaluate which amino acids actually relate to the reduction of SBP in order to evaluate its independent role in decreasing SBP in SHRs as well as mechanisms that are involved.
In conclusion, we established GABA-enriched rice in this study. Compared with SHRs that were administered common rice, SHRs that were fed GABA-enriched rice showed a significant effect in the lowering of BP. Therefore, transgenic rice lines established in this study are likewise expected to protect humans from diseases that occur due to high BP. Besides, the latest report indicates that metabolism of GABA and γ-hydorxybutyric acid (GHB) which is known as a drug of abuse, are tightly linked even in plants (Ludewig et al. 2008). Although there has been no clear report on health problem of GABA-enriched foods from plants by a human consumption so far, it should be considered to determine the level of GHB in the GABA enriched rice developed in this study for future practical use.
This research was supported by grants for Functional Analysis of Genes Relevant to Agriculturally Important Traits in the Rice Genome from the Ministry of Agriculture, Forestry and Fisheries of Japan (Green Technology Project IP-2004).
This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the original author(s) and source are credited.
- Akama K, Takaiwa F (2007) C-Terminal extension of rice glutamate decarboxylase (OsGAD2) functions as an autoinhibitory domain and overexpression of a truncated mutant results in the accumulation of extremely high levels of GABA in plant cells. J Exp Bot 58:2699–2707. doi:10.1093/jxb/erm120 CrossRefPubMedGoogle Scholar
- Akihiro T, Koike S, Tani R, Tominaga T, Watanabe S, Iijima Y, Aoki K, Shibata D, Ashihara H, Matsukura C, Akama K, Fujimura T, Ezura H (2008) Biochemical mechanism on GABA accumulation during fruit development in tomato. Plant Cell Physiol 49:1378–1389. doi:10.1093/pcp/pcn113 CrossRefPubMedGoogle Scholar
- Kisaka H, Kida T, Miwa T (2006) Antisense suppression of glutamate decarboxylase in tomato (Lycopersicon esculentum L.) results in accumulation of glutamate in transgenic tomato fruits. Plant Biotechnol 23:267–274Google Scholar
- Takagi H, Hiroi T, Yang L, Tada Y, Yuki Y, Takamura K, Ishimitsu R, Kawauchi H, Kiyono H, Takaiwa F (2005) A rice-based edible vaccine expressing multiple T cell epitopes induces oral tolerance for inhibition of Th2-mediated IgE responses. Proc Natl Acad Sci USA 102:17525–17530. doi:10.1073/pnas.0503428102 CrossRefPubMedGoogle Scholar
- Tsuchida T, Mashiko K, Yamada K, Hiratsuka H, Shimada T, Itagaki Y, Fujinuma H, Samejima K, Nakamura T, Hasegawa T, Matsubayashi T (2003) Clinical study of γ-aminobutyric acid-rich Chlorella for subjects with high-normal blood and mild hypertension. J Jpn Soc Nutr Food Sci 56:97–102 (text in Japanese with English abstract)Google Scholar
- Wakasa K, Hasegawa H, Nemoto H, Matsuda F, Miyazawa H, Tozawa Y, Morino K, Komatsu A, Yamada T, Terakawa T, Miyagawa H (2006a) High-level tryptophan accumulation in seeds of transgenic rice and its limited effects on agronomic traits and seed metabolite profile. J Exp Bot 57:3069–3078. doi:10.1093/jxb/erl068 CrossRefPubMedGoogle Scholar