Generation of rice plants expressing OASA1D
Transgenic rice plants that express OASA1D were generated and subjected to the analyses of Southern blot and northern blot as described previously (Tozawa et al., 2001; Yamada et al., 2004). Given that expression of OASA1D confers resistance to 0.3 mM 5-methyltryptophan (5MT), some transgenic plants were generated from calli selected with 5MT instead of with hygromycin (Yamada et al., 2004). Transgenic lines selected by growth in the presence of 5MT or hygromycin are denoted by M or H, respectively.
Spikelet fertility of greenhouse-grown plants
Almost all regenerated plants of >120 transgenic lines grown in pots exhibited normal growth, with exceptional instances of dwarfism or slow growth presumably being attributable to somaclonal mutations (Phillips et al., 1994). However, for plants grown in a greenhouse, the spikelet fertility of transgenic plants tended to be lower than that of the wild type (Table 1). The mean spikelet fertility of the transgenic plants was thus only 31%, compared with a value of 76% for seed-grown Nipponbare. Given that a decrease in spikelet fertility has previously been observed in transgenic rice plants generated by tissue culture (Hiei et al., 1994; Urushibara et al., 2001), part of the reduction in spikelet fertility apparent in OASA1D-expressing plants might be due to the regeneration process. However, transgenic rice plants that overexpress the wild-type gene (OASA2) for another subunit of rice AS but do not accumulate Trp in calli or leaves (Tozawa et al., 2001) exhibited a spikelet fertility (61%) higher than that of plants that express OASA1D, but lower than that of Nipponbare, suggesting that accumulation of Trp might also contribute to the reduced spikelet fertility of the OASA1D transgenic plants. Analysis by genomic Southern hybridization of R0 (regenerated) plants of OASA1D transgenic lines with a spikelet fertility of >50% revealed that they had a relatively high copy number (three to seven) of the transgene (data not shown), suggesting that the copy number was not a principal cause of low spikelet fertility.
Trp content of seeds produced by greenhouse-grown plants
Seeds of 12 OASA1D transgenic lines with a spikelet fertility of >50% were analysed for free and total Trp contents in the second or third generation (R2 or R3 seeds) of plants grown in a greenhouse. All lines showed a marked increase in the amount of free Trp, with the mean free Trp content ranging from 3037 to 23 705 nmol g–1 of dry seed weight (Table 2); these values correspond to increases of 55- to 431-fold compared with the free Trp content of non-transgenic Nipponbare seeds (55 nmol g–1). The amount of free Trp as a percentage of total Trp in the transgenic seeds (34–87%) was also greatly increased compared with that in wild-type seeds (1.3%). Given that seeds of a transgenic line expressing GUS, which encodes ß-glucuronidase, exhibited a Trp content similar to that of Nipponbare seeds, it was unlikely that the Trp accumulation apparent in seeds of OASA1D transgenic plants resulted from an abnormality caused by the transformation process.
The transgenic seeds analysed for Trp content were a mixture of those homozygous or heterozygous for OASA1D. To examine the possible influence of genotype on Trp content, seeds from a single transgenic plant of the H17 line in the R1 generation were divided into two groups. One group of 26 seeds showed segregation of OASA1D by PCR analysis (14 positive, 12 negative). The other group of 32 seeds, which also should have been a mixture of transgene genotypes, all contained an increased level of free Trp (9134±3740 nmol g–1). These results thus indicated that Trp accumulation in seeds is determined primarily by the genotype of the mother plant. However, given that the free Trp content of seeds of the H17 or M34 lines showed some variability (Table 2), the level of free Trp in seeds might also be influenced by seed genotype. Although the physiological conditions of plants and seeds also affect amino acid content, the increase in the amount of Trp in the seeds of transgenic rice expressing OASA1D was sufficiently high to be attributed to the activity of the transgene.
It is the total Trp content, including both free Trp and Trp in proteins, that is important for the nutritional value of seeds. The increase in the amount of free Trp in seeds of OASA1D transgenic plants was accompanied by an increase in the total Trp content (Table 2). The total Trp content of seeds of the various transgenic lines thus ranged from 8100 to 48 519 nmol g–1, values that correspond to increases of 1.9- to 11.6-fold compared with that for Nipponbare (4188 nmol g–1).
The weight of individual dehulled seeds tended to be smaller for OASA1D transgenic lines than for Nipponbare (Table 2). However, a reduced seed weight was also apparent for plants harbouring a GUS transgene, suggesting that this effect might be attributable to the transformation process. OASA1D transgenic line M34, whose seeds showed the highest Trp content, also manifested the lowest seed weight. However, the seeds of line M21, which exhibited a medium level of Trp accumulation, were also of low weight. No clear correlation between Trp content and seed weight was thus apparent.
Generation of homozygous OASA1D transgenic rice lines for evaluation of agronomic traits
Two OASA1D transgenic lines, M121 and H41, were advanced to obtain homozygotes for evaluation of agronomic traits. The resulting homozygous lines were designated HW1 for M121 and HW5 for H41. The early generations of these homozygous lines exhibited a spikelet fertility of >50% and normal morphological features. Genomic Southern blot analysis of selfed progenies and of F2 plants of F1 hybrids between either HW1 or HW5 and Nipponbare revealed that HW1 contained three copies of the transgene and HW5 harboured four copies. No segregation of hybridized bands was observed in F2 plants, indicating that the transgenes were integrated at one locus (data not shown).
Trp content of seeds produced by field-grown OASA1D transgenic plants
The free and total Trp contents as well as the nitrogen content of R5 and R6 seeds of the HW lines grown in a greenhouse were higher than those of Nipponbare seeds (Table 3), and all of these values were slightly higher for HW1 than for HW5. The levels of free and total Trp in seeds were stable during growth of the HW lines in a greenhouse for 2 years (data not shown). In the field condition, the free Trp content of seeds of the HW lines was increased about 2-fold compared with the corresponding values for seeds of greenhouse-grown plants (Table 3). This increased accumulation of free Trp in the seeds of field-grown plants was not accompanied by a similar increase in the total Trp content. The level of total Trp in seeds of the HW lines was thus stable under different growth conditions. The nitrogen content of seeds of field-grown HW lines and Nipponbare was lower than the corresponding values for seeds of greenhouse-grown plants. The amounts of free and total Trp in Nipponbare seeds were also lower in the field condition than in the greenhouse, with the result that the relative values to Nipponbare for seeds of the field-grown HW lines were increased accordingly (Table 3). The levels of Trp in the seeds of Nipponbare grown in another field were similarly low (data not shown).
The opposite effects of field growth on the free Trp content of seeds of the HW lines and of Nipponbare seeds might be attributable, in part, to a response of the transgene promoter to the cooler temperatures or to the difference in temperature between night and day in the field. The OASA1D gene was driven by the ubiquitin gene promoter, which is responsive to stress (Takimoto et al., 1994) and might therefore be activated by low temperatures, resulting in increased expression of OASA1D and a greater accumulation of free Trp.
Amino acid composition of seeds produced by field-grown OASA1D transgenic plants
The increased Trp content of rice calli expressing OASA1D does not result in substantial changes in the amounts of other amino acids (Tozawa et al., 2001). The marked accumulation of free Trp in the seeds of the HW1 and HW5 lines grown under field conditions was accompanied by an increase in the amounts of other amino acids to some extent (Table 4). Their increases were relatively small compared with that of Trp. The total amount of free amino acids was increased 4.2- and 2.9-fold in HW1 and HW5, respectively, compared with the value for Nipponbare, with the maximal change in the content of any one amino acid (other than Trp) being limited to a 5.2-fold increase. However, with the exception of Ala (and Trp), the ratio of the amount of each amino acid to the total amount of free amino acids was reduced or remained virtually the same in the transgenic lines compared with Nipponbare.
The levels of Lys and Phe were low in Nipponbare seeds but their absolute amounts were increased in the HW seeds, with the result that their percentage contributions to the total amount of free amino acids were the same in the HW lines and in the wild type. Given that Gln is the amino donor for the synthesis of anthranilate, its abundance might have been expected to be changed in the seeds of the transgenic lines. Its absolute amount in seeds of the HW lines was similar to that in Nipponbare seeds, however. Serine is also a precursor for Trp synthesis, but its amount in HW seeds was not significantly increased compared with that in Nipponbare seeds.
The increases in the absolute amounts of free amino acids in the seeds of the transgenic lines suggest the existence of regulatory mechanisms that increase amino acid synthesis in response to Trp accumulation. The opaque-2 mutation in maize (Oh545o2) is associated with an increased level of free amino acids in mature endosperm (Wang and Larkins, 2001). Genetic analysis suggests that the gene for aspartate kinase 2 is the gene responsible for the effect of this mutation on free amino acid content (Wang et al., 2001). A mutation in a transcriptional regulator of AS genes also renders Trp biosynthesis insensitive to Trp concentration (Bender and Fink, 1998). Our results therefore suggest that Trp accumulation in rice seeds might increase the transcription of genes that encode enzymes responsible for amino acid synthesis.
Agronomic traits of field-grown OASA1D transgenic plants
HW lines grown in a greenhouse appeared similar to Nipponbare with regard to most agronomic traits analysed (Table 5). For plants grown under field conditions, however, differences in traits related to seed productivity were apparent between the transgenic lines (especially HW1) and the wild type. Both HW lines grown in the field thus exhibited a spikelet fertility lower than that of Nipponbare, although pollen fertility (Table 5) and anther size (reflecting the number of pollen grains) (data not shown) for the transgenic lines were similar to those for Nipponbare. Moreover, the average spikelet number per panicle was significantly smaller for the HW lines than for Nipponbare. The reduction in the number of spikelets per panicle and the low spikelet fertility likely contributed to an observed decrease in harvested seed weight for HW1 and HW5 to 52.5% and 69.6% of the value for Nipponbare, respectively. The harvested plant weight was similar for the HW lines and Nipponbare (data not shown). Whereas the individual brown seed weight varied among transgenic lines in early generations (Table 2), it did not differ markedly among HW lines and Nipponbare under field conditions (Table 5).
The high concentration of Trp in the HW lines is likely to be the primary cause of the differences in agronomic traits between these lines and Nipponbare grown under field conditions. Spikelet number per panicle is determined at an early stage of development of the inflorescence apex and is influenced by several conditions, such as nitrogen and carbon availability as well as temperature (Takeoka et al., 1993). Accumulation of Trp might increase the sensitivity of plants of the HW lines to environmental stress and thereby reduce spikelet number and fertility in the field condition.
The culm length of HW lines grown in the field was smaller than that of Nipponbare, although this difference was statistically significant only for HW1 (Table 5). No difference in plant height was observed between HW lines and Nipponbare (data not shown). A short culm length has often been observed in plants regenerated from tissue culture (Phillips et al., 1994), suggesting that this characteristic of the HW lines might be attributable to somaclonal mutation.
Germination of seeds from field-grown OASA1D transgenic plants
The growth condition markedly influenced seed germination in HW lines. Whereas the germination percentage for seeds of greenhouse-grown plants was similar for HW lines and Nipponbare, it was greatly reduced for seeds of field-grown HW1 plants compared with that for field-grown HW5 or Nipponbare plants (Table 6). The time to germination was also increased for seeds from both HW lines grown under field conditions, as well as for seeds from HW1 plants grown in the greenhouse.
Poor seed germination has been observed in other plants with increased levels of an essential amino acid. Transgenic soybean with a large increase in free Lys content thus manifested reduced seed viability (Falco et al., 1995). Seeds of transgenic Arabidopsis thaliana with a high Lys content also exhibited retarded germination and seed establishment (Zhu and Galili, 2003). The seeds of HW1 plants grown in the field showed the lowest frequency of and greatest delay in germination as well as the highest content of free Trp compared with seeds of HW1 or HW5 plants grown in the greenhouse and HW5 plants grown in the field. These observations suggest that an increase in Trp content over a certain threshold level might substantially influence germination. The importance of amino acid metabolism such as Gln, Lys, and Met in maize germination efficiency has been shown (Limami et al., 2002; Anzala et al., 2006). Trp has been shown to be an endogenous inhibitor of embryo germination in white wheat (Morris et al., 1988).
It might prove possible to ameliorate the unfavourable traits observed in the HW lines grown under field conditions by controlling the extent and tissue distribution of Trp accumulation with the use of a different promoter to drive OASA1D expression. Potential promoters for this purpose include those of embryo-specific genes or of genes that are not responsive to stress.
Metabolite profile of seeds from field-grown OASA1D transgenic plants
Changes in the composition of aromatic components in the seeds of field-grown HW1 and HW5 plants were analysed by reversed-phase HPLC. Elution was monitored with a photodiode array detector over a wavelength range of 190–400 nm. Typical chromatograms obtained at 280 nm, the most effective wavelength for detection of changes in the composition of anthranilate-related metabolites, are shown in Fig. 1. The aromatic metabolite profiles of dehulled seeds revealed no apparent marked accumulation of components other than Trp in the transgenic seeds (Fig. 1, inset). Magnification of the chromatograms revealed small differences between HW lines and Nipponbare (Fig. 1). A peak with a retention time of 10 min, for example, was specifically detected in both transgenic lines. Essentially, identical results were obtained by monitoring elution at wavelengths other than 280 nm (data not shown). These results are surprising given the high levels of Trp in the transgenic seeds and that the Trp biosynthetic pathway gives rise to various secondary metabolites, such as the indole alkaloids and indole glucosinolates, in many plants. To date, no remarkable secondary metabolites of Trp origin have been reported in rice plants and the present study suggested that the absence of such Trp-derived secondary metabolites was probably not due to a shortage of Trp supply, but to a very low capability of relevant Trp utilization in rice. Furthermore, Trp decarboxylase probably plays more important role in the divergence of Trp-related carbon flow into the secondary metabolism, as has been observed in transgenic plants overexpressing Trp decarboxylase (Chavadej et al., 1994; Yao et al., 1995)
It was specifically investigated whether the accumulation of Trp in the seeds of the field-grown HW lines affected the amount of IAA, given the close relations between the Trp biosynthetic pathway and IAA production. Therefore the levels of free IAA, of free IAA plus its ester conjugates, and of total IAA (including amide conjugates) were measured in the seeds of the HW lines and Nipponbare (Fig. 2A). The amounts of free and conjugated forms of IAA were each increased about 2-fold in the seeds of both HW lines compared with those in Nipponbare seeds. The increase in the level of free IAA in seeds was consistent with our previous demonstration of IAA accumulation in rice calli expressing OASA1D (Morino et al., 2005). Increased auxin content has been associated with Trp accumulation in cultured carrot and potato cells resistant to 5MT (Widholm, 1977; Sung, 1979). The level of IAA conjugates was also found to be increased in the Arabidopsis mutant Amt1, which expresses a feedback-insensitive AS and accumulates Trp (Kreps and Town, 1992; Ludwig-Muller et al., 1993). In addition, 5MT-resistant mutants of Lemna gibba showed an approximately 3-fold increase in the amount of free IAA (Tam et al., 1995). Seeds of rice and maize normally contain higher concentrations of IAA than do other organs of these plants (Bandurski and Schulze, 1977). The increase in the amounts of free and total IAA apparent in the seeds of the HW lines thus suggests that rice seeds are able to accumulate IAA to especially high levels.
Given that IAA affects multiple aspects of plant growth, the low spikelet fertility and density as well as the impaired seed germination of the HW lines might reflect the increased abundance of IAA rather than that of Trp. IAA markedly inhibited the germination of wheat embryos excised from caryopses that were highly dormant (Ramaih et al., 2003); Trp, the precursor of IAA, was shown to be equally inhibitory in this instance, however. In the case of these HW lines, the impairment in seed agronomic traits was greater for HW1 than for HW5, whereas the amount of total IAA was slightly higher in the seeds of HW5 than in those of HW1. The impairment thus appeared to be more correlated with Trp content than with total IAA (Fig. 2). Regardless, the growth of HW1 and HW5 seedlings after germination overtook that of Nipponbare and no differences in plant growth at the harvesting stage were detected between Nipponbare and the HW lines (data not shown).