• Open Access

Foreign gene products can be enhanced by introduction into low storage protein mutants

Authors

  • Yoshifumi Tada,

    1. Department of Plant Biotechnology, National Institute of Agrobiological Sciences, Kannondai 2-1-2, Tsukuba Ibarki 305-8602, Japan,
    Search for more papers by this author
  • Shigeru Utsumi,

    1. Division of Agronomy and Horticultural Science, Graduate School of Agriculture, Kyoto University, Uji, Kyoto 611-0011, Japan
    Search for more papers by this author
  • Fumio Takaiwa

    Corresponding author
    1. Department of Plant Biotechnology, National Institute of Agrobiological Sciences, Kannondai 2-1-2, Tsukuba Ibarki 305-8602, Japan,
    Search for more papers by this author

Correspondence (fax: +81 298 38 8397; e-mail: takaiwa@nias.affrc.go.jp)

Summary

Transgenic rice expressing soybean glycinin in its endosperm was crossed with two types of low-glutelin mutants to determine how much storage the protein mutants can contribute to increases in glycinin accumulation. The glycinin level (102 µg/100 mg seed) in the parental transgenic line was enhanced to ≈ 224–237 µg/100 mg seed within a genetic background deficient in glutelin (i.e. of low glutelins). The enrichment of this foreign gene product was compensated by a decrease in the expression of other endogenous prolamine and globulin storage proteins, resulting in an almost equivalent total amount of seed storage proteins. These results show that low storage protein mutants can provide potentially useful hosts for the expression of foreign genes, allowing a higher-level accumulation, because they can provide wider space for the accumulation of foreign gene products than in the normal host plant.

Introduction

Plant seeds reserve large quantities of protein to provide a source of nitrogen and carbon for germinating seedlings: as such, they have been exploited as the primary source of nutrition by both humans and livestock. Crop nutritional value has been improved by introducing heterologous storage protein genes or modified storage protein genes which are rich in limited essential amino acids (Tabe and Higgins, 1998). The accumulation level of the introduced gene products is generally not particularly high, usually being less than a few per cent of the total seed protein (Habben and Larkins, 1995). The higher accumulation levels of transgenic products are required not only for crop improvement, but also for molecular farming.

Increased accumulation levels of transgene products could be primarily achieved by enhancement of the transcription level. Screening for strong seed-specific promoters, improvements in promoter strength by modification of the cis-elements, or increasing the copy number of transgenes are required for high level expression. The stability of the transgene product is also important for stable and high level accumulation. There are many factors affecting transgene product stability, such as the design of the molecular structure of the transgene product, co-introduction of related subunits for assembly, and control of the secretory pathway, which allows the transgene products to accumulate at the appropriate subcellular locality (Moloney and Holbrook, 1977).

It is clear that much effort has been carried out towards determining ways of enhancing the accumulation of transgene products. However, the actual capacity of seeds to accumulate transgene products has not thus far been studied in any great depth.

According to their solubility, seed storage proteins are classified into albumins (water soluble), globulins (saline soluble), prolamins (alcohol soluble) and glutelins (residue). Glutelin is the major storage protein of rice, accounting for ≈ 60–80% of the total endosperm protein (Takaiwa et al., 1999). Based on recently established molecular and biochemical properties, the storage proteins can be classified into two major groups, the globulins and the prolamins (Shewry et al., 1995). Seed storage proteins are packaged into organelles called protein bodies (PBs) (Okita and Rogers, 1996; Vitale et al., 1993). In the case of rice, glutelin and globulin are stored in a vacuolar compartment (PBII), whereas prolamins are packaged in ER-derived PBI (Okita and Rogers, 1996; Tanaka et al., 1980).

To date, many storage protein mutants have been isolated from seeds treated with chemical mutagens or irradiated with gamma rays. A low-protein rice (LGC-1: low glutelin content-1) is a unique mutant selected from ethyleneimine-treated rice (Iida et al., 1993). In this mutant, several glutelin subunits, which are controlled by a single dominant gene, are severely suppressed. Another mutant, α123 (α-1, 2, 3 less), also lacks three glutelins. This mutant was developed by successive crossings of three different mutants lacking different acidic subunits of glutelin, named α-1, α-2 and α3 (Iida et al., 1997), which are controlled by three independent recessive genes.

These two mutants have a unique composition regarding the relative amounts of their individual seed storage proteins, in which 13 kDa prolamine and 26 kDa globulin are enriched following the suppression of glutelin. It is interesting to note that the total amounts of seed proteins do not differ from those of normal rice (approximately 6% of the seed's weight) (Iida et al., 1997). The increase in the quantities of prolamine and globulin appear to compensate for the decrease in glutelin content. A similar observation has also been made by suppressing seed storage protein genes (Kohno-Murase et al., 1994) or other storage protein mutants in other crops (Hartweck and Osborn, 1997; Ogawa et al., 1989). Reduction of rice glutelin synthesis by an antisense construct results in an increase in other endogenous storage proteins to maintain the total amount of seed storage proteins (Maruta et al., 2001). Furthermore, it has recently been demonstrated that a greater accumulation of the Phaseolus vulgaris arcelin (arc)5-I gene product was obtained by the co-introduction of an antisense gene for an endogenous seed storage protein (2S albumin) in Arabidopsis thaliana seeds (Goossens et al., 1999).

We previously demonstrated that glycinin can be accumulated at significant levels in the endosperm tissue of tobacco and rice seeds under the control of the glutelin GluB-1 promoter (Katsube et al., 1999; Takaiwa et al., 1995). However, significant differences in many of the physico-chemical properties of the seed proteins, such as lipid, fibre, ash and carbohydrate could not be detected between transgenic rice accumulating soybean glycinin and non-transgenic rice (Momma et al., 1999). Glycinin was specifically synthesized in the endosperm tissue and co-localized with glutelins in PBII, and it has been shown that both the soybean glycinin and rice glutelin proteins can be assembled with each other to form soluble hexameric oligomers or insoluble aggregates (Katsube et al., 1999). Therefore, it was suggested that the glutelin in PBII might compete with the glycinin over space for accumulation or sources of amino acids for synthesis; low-glutelin mutants might thus have a greater capacity for glycinin accumulation because sufficient space remains for the accumulation of foreign proteins. There are also sufficient amounts of amino acids in the low glutelin mutant when it is used as a host for recombinant protein production.

In this paper, the suitability of low-glutelin mutants as hosts was examined using the soybean glycinin as the model of a foreign protein. For this purpose, a genetically homozygous line expressing the glycinin gene, 11-5, was crossed with two types of low-glutelin mutant, LGC-1 and α-123. We demonstrate that glycinin accumulation levels were enhanced approximately twofold within a low-glutelin genetic background, proving its feasibility as a host for a high production system. These results suggest that the deposition space of foreign gene products is the crucial factor determining the accumulation levels of foreign gene products.

Results

Expression of soybean glycinin gene in low-glutelin mutants

To determine whether the expression level of the soybean glycinin transgene is enhanced within a low storage protein genetic background, a transgenic line (11-5) expressing glycinin at a level of 102 ± 22 µg/100 mg seed (Katsube et al., 1999) was crossed with two low-glutelin mutants (LGC-1 and α123). Two homozygous lines expressing glycinin within a low-glutelin genetic background were screened for the presence of glycinin and the deletion of some glutelins characteristic of individual mutants by SDS-PAGE electrophoresis and Western blotting analysis. The endosperm half of the seed was used as the material for this analysis. It is notable that the suppression of some glutelin genes in the LGC-1 is controlled by a dominant gene, whereas the low level of glutelins in the α123 mutant is caused by the structural mutations of three glutelin genes. The desired genotypes expressing glycinin within a low glutelin background were obtained from T4 or T5 generation lines by continuous self-crossing (Figure 1).

Figure 1.

SDS-PAGE analysis of total proteins extracted from the parental lines (11-5, LGC-1, α123) and their hybrid lines (LGC-1 × 11-5, α123 × 11-5). Total seed proteins from matured seeds were analysed on 15% SDS–PAGE electrophoretic gel. The expression levels of soybean glycinin were analysed by Western blotting using antiglycinin antibody.

The transfer of the glycinin gene to these two low-glutelin mutants was confirmed by Southern analysis. As shown in Figure 2, at least three copes of the glycinin genes were integrated into the genome of the 11-5 transgenic rice line. The DNA restriction pattern of the glycinin gene was identical between the parental transgenic line (11-5) and its hybrids (LGC × 11-5 and α123 × 11-5), indicating that the glycinin gene was accurately transferred to the low-glutelin mutants (Figure 2). The low-glutelin genetic background was also confirmed by the pattern of total seed proteins on a SDS-PAGE electrophoresis gel, which showed that the acidic and basic subunits of glutelins characteristic of the LGC-1 and α123 mutants were severely suppressed (i.e. decreased to approximately one-third of their normal levels).

Figure 2.

Southern blot analysis of glycinin A1aB1b gene in transgenic rice. The genomic DNA isolated from the parental transgenic line (11-5), its hybrid lines with low glutelin mutants (LGC-1 × 11-5 and α123 × 11-5) and wild-type line (non-transgenic rice) were digested with EcoRI and HindIII and fractionated by electrophoresis on a 0.8% agarose gel. The coding region of pre-proglycinin A1aB1b cDNA (A1aB1b) was used as a probe.

Quantification of glycinin and endogenous storage protein levels accumulated in mature seed against different genetic backgrounds

To estimate the accumulation levels of glycinin within the low-glutelin background, total proteins were extracted from the matured seeds of the parental 11-5 transgenic plant and the two hybrid lines (LGC × 11-5, α123 × 11-5). Crude seed extracts were successively diluted, spotted on to nitrocellulose membranes with a dilution series of glycinin standard protein, and then subjected to immunoblotting using antiglycinin (A1aB1b) antibody (Figure 3). Glycinin quality was determined from the standard curve. The levels of glycinin in the hybrid lines were 237 ± 40 µg and 224 ± 29 µg per 100 mg seed, respectively, representing a greater than twofold enhancement compared to the parental line, 11-5 (102 ± 22 µg) (Table 1). We then examined whether the total protein in each seed and seed weight were affected by these low glutelin backgrounds. The level of total protein was 6.5–7.4% of seed weight and did not significantly differ in respect to genetic background or glycinin expression level (Table 1), suggesting that the reduction of some glutelins might be compensated by an increase in other seed proteins, thus maintaining similar overall levels of seed proteins. Furthermore, there was no significant difference (19–24 mg/seed) in mature seed weights among the two hybrid mutant lines, the parental transgenic line or the wild-type rice. These results suggest that the expression of the foreign genes in the low glutelin mutants used here have little or no effect on protein yield per seed, thus indicating that the use of these mutants could be advantageous as a production platform.

Figure 3.

Dot blot analysis of glycinin in transgenic rice seed. Total seed protein was extracted and spotted on to a nitrocellulose membrane using a Multi-pin-blotter. The glycinin level in each dot was determined by comparing with sequentially diluted purified glycinin used as standard.

Table 1.  Comparison of accumulation levels of glycinin in matured seeds, total proteins in matured seed and weight of 100 hulled brown rice seeds from wild-type, the parental transgenic line (11-5) and its hybrids with low glutelin mutants (LGC1 × 11-5, α123 × 11-5). Data are provided as the mean value ± SD of glycinin in the total seed protein. Four independent experiments were performed
 Accumulation levels of glycinin (µg/100 mg seed)Total protein (g/100 g seed)Weight of seed grain (g/100 seeds)
Wild-type07.127 ± 0.0612.3
11-5102 ± 226.804 ± 0.0472.4
LGC-1 × 11-5237 ± 406.523 ± 0.2582.3
α123 × 11-5224 ± 297.362 ± 0.4061.9

To examine whether such high glycinin expression had any effects on the expression pattern of other seed proteins, total seed proteins were analysed by electrophoresis on SDS-PAGE gel. As shown in Figure 1, the pattern of seed proteins on the gel was fundamentally similar to that of low-glutelin mutants except for the accumulation of glycinin (LGC-1 vs. LGC-1 × 11-5; α123 vs. α123 × 11-5). It is clear that the acidic and basic subunits of glycinin accumulated in greater quantities in the hybrid lines than in the parent line. Quantities of both acidic and basic glutelin subunits were significantly decreased, whereas the levels of prolamine and globulin were enhanced. It should be noted that the prolamine level was slightly decreased by the accumulation of glycinin in the hybrid lines compared to those of the parental low-glutelin mutants (Figures 1 and 4).

Figure 4.

Relative accumulation levels of the albumin-globulin, prolamine and glutelin fractions in matured seeds from the wild-type line, parental lines (11-5, LGC-1 and α123) and their hybrid lines (LGC-1 × 11-5 and α123 × 11-5). Step-wise extraction of these storage proteins were performed by extraction with 1 m NaCl solution, 60% n-propanol and 1% lactic acid. Data are means ± SD of four independent replicates.

To determine which seed proteins were changed by crossing, we separated them into three fractions (albumin-globulin, prolamine and glutelin) by a step-wise extraction. Quantitative determination of the proteins in each fraction demonstrated that glutelin content was specifically decreased for the low-glutelin lines, LGC-1, α123 and the hybrid lines, LGC × 11-5, α123 × 11-5. The changes in protein composition were greater between LGC-1 and its hybrid line, LGC × 11-5, than between α123 and its hybrid line, α123 × 11-5 (Figure 4).

The relative accumulation levels of glutelin contents in the hybrid lines were almost the same as those of their respective parental lines. These results suggest that glycinin expression had little effect on the accumulation levels of glutelin within these genetic backgrounds.

It should be noted that the prolamine content in the low-glutelin and hybrid lines was greatly enhanced compared to those of the wild-type or the original transformant, 11-5. The level of change between the LGC-1 and its hybrid line, LGC × 11-5, was greater than that between α123 and its hybrid line, α123 × 11-5 (Figure 4). The prolamine content increased in inverse proportion to the level of glutelin decrease.

Furthermore, it is interesting to note that the increase in the albumin-globulin fraction containing glycinin resulted in a slight reduction of either prolamine (LGC-1 background) or glutelin (α123 background) contents (Figure 4). Taking these results together, we conclude that transgene expression results in the reduction of endogenous seed proteins so as to maintain the total amount of nitrogen accumulated in the seed storage proteins.

We also determined which seed proteins in the albumin-globulin fraction were changed by crossing. This fraction contains glycinin, 26 kDa globulin and 14–16 kDa albumins. Most of the glycinin in this fraction can be selectively recovered by precipitation with 50% acetone. This precipitated fraction contains not only glycinin but also several high molecular weight enzymes, whereas the 26 kDa globulin and 14–16 kDa albumin are recovered as the residual albumin-globulin fraction (data not shown). It is notable that the acetone-precipitable fraction was enhanced in the hybrid lines compared to that of their parental line, 11-5 (Figure 5A).

Figure 5.

Accumulation levels of (A) glycinin fraction and (B) residual albumin-globulin fraction in matured seeds from the wild-type line, parental lines (11-5, LGC-1 and α123) and their hybrid lines (LGC-1 × 11-5 and α123 × 11-5). Data are means ± SD of four independent replicates.

The amounts of residual albumin-globulin fraction were also increased within the low-glutelin background (LGC-1, LGC × 11-5, α123 and α123 × 11-5) (Figure 5B). Furthermore, it is interesting to note that the expression of the transgene led to a slight reduction of the residual albumin-globulin fraction, because the amount of this fraction in the hybrid line (LGC × 11-5) was lower than that of its parental line (LGC-1).

Analysis of mRNA and proteins during seed development

To determine whether the expression of the soybean glycinin gene is affected by different genetic backgrounds or whether it has an effect on the specific expression patterns of endogenous seed storage protein genes, mRNA levels of two types of glutelins (GluA-2 and GluB-1), 13 kDa prolamin, 26 kDa globulin and soybean glycinin were examined by Northern blot analysis during seed development from 5 to 25 DAF (days after flowering). As Figure 6 shows, glutelin mRNA levels in the low-glutelin mutants were highly suppressed. It was notable that expression of the GluB-1 gene was restricted to the early maturation stage from 5 to 10 DAF for the LGC-1 line, and that mRNA was not detected thereafter. This suppression was also observed for the GluA-2 mRNA. This is in remarkable contrast with the mRNA expression patterns in the non-transgenic and the 11-5 transgenic rice lines, in which glutelin mRNA levels gradually increased from 5 DAF to a peak at 15 DAF, then decreased towards seed maturation. This evidence suggests that the suppression of glutelin genes found in the LGC-1 might be regulated in a post-transcriptional manner. As Figure 6 shows, the transcript level of the GluA-2 gene was severely suppressed in both the α123 mutant and its hybrid with the 11-5 transgenic line (α123 × 11-5). This low level of GluA mRNA was due to the loss of two genes belonging to the GluA family in the α123 mutant.

Figure 6.

Northern blot analysis of seed RNAs during seed development. Total RNAs were isolated from maturing seeds of the wild-type line, low glutelin mutants (LGC-1 and α123), transgenic rice (11-5) and their hybrid lines (LGC-1 × 11-5 and α123 × 11-5). They were fractionated by electrophoresis on a 1.2% agarose gel containing formaldehyde and then blotted on to nylon membranes. The RNAs were hybridized with the coding regions of glycinin A1aB1b cDNA (A1aB1b), glutelin GluA-2 cDNA (pREE99), glutelin GluB-1 cDNA (pREEK1). 26 kDa globulin cDNA (26 k globulin) and 13 kDa prolamine cDNA.

When the expression level of the soybean gene was examined for the two hybrid lines, the glycinin mRNA levels were not found to be influenced by the low glutelin background because they were almost identical to those of parental mutants and wild-type lines (Figures 6 and 7). These results indicate that the glycinin gene directed by the GluB-1 promoter was normally expressed, even although the endogenous glutelin genes belonging to the GluB family were highly suppressed. Taken together, this evidence demonstrates that the activity of the GluB-1 promoter expressing the glycinin gene was not affected by the low-glutelin genetic background, thus indicating that foreign genes can be expressed at almost identical levels to that of wild-type rice.

Figure 7.

Comparison of mRNA levels of soybean glycinin (A) and 13 kDa prolamine (B) at 15 DAF and 20 DAF among the non-transgenic line, the low glutelin mutants, the transgenic line and its hybrid lines.

It has been reported that levels of prolamine transcripts are enhanced in the low-glutelin genetic background (Maruta et al. (2001). However, when the mRNA level of 13 kDa prolamine in the parental 11-5 line and wild-type rice were compared with those of the two hybrid lines (α123 × 11-5 and LGC × 11-5) during seed maturation from 5 to 25 DAF, no marked differences between the levels of these transcripts could be detected (Figure 7). The minor difference observed can not explain the remarkable increase in 13 kDa prolamins in the low-glutelin background. This suggests that the high accumulation of 13 kDa prolamin in the low-glutelin mutant might be mainly regulated at the post-transcriptional level.

Relative increases in the accumulation levels of glutelin, globulin, prolamine and glycinin were calculated from the corresponding band density to investigate the levels of protein synthesis (data not shown). In the 11-5 line, glutelin, globulin and glycinin are synthesized at similar times, whereas the synthesis of 13 kDa prolamine begins slightly later than the other storage proteins and continues to the matured stage (Figure 8). In the case of the LGC-1 × 11-5 hybrid line, glutelin synthesis was only restricted at the early stages of seed development and reached a plateau at 15 DAF, which might be reflected by the observed mRNA level shown in Figure 6. Globulin and glycinin were actively synthesized in the same way as in the 11-5 line. It is notable that the prolamine level was greatly enhanced at the early maturation stage (10–15 DAF) in the LGC-1 × 11-5 hybrid line compared to that of the 11-5 line (Figure 8). Furthermore, the synthesis of globulin, prolamine and glycinin were prolonged in the hybrid. In the α123 × 11-5 hybrid, it is notable that the accumulation levels of glutelin and glycinin reached a peak at 20 DAF, whereas the synthesis of globulin and prolamine were prolonged even beyond 25 DAF (Figure 8). Taken together, these results suggest that the high accumulation of prolamine and globulin in the low-glutelin genetic background could be accounted for by their prolonged synthesis to the matured stage.

Figure 8.

SDS-PAGE analysis of total proteins extracted from maturing seeds (5–25 DAF) of the transgenic line (11-5) and its hybrid lines (LGC-1 × 11-5, α123 × 11-5). Total seed proteins were fractionated by electrophoresis on 15% polyacrylamide gel and stained by Coomassie Brilliant Blue.

Discussion

Low-glutelin trait can enhance the accumulation of glycinin

It has been reported that the suppression of certain storage proteins by mutation or antisense technology can give rise to a compensatory increase in the contents of the other endogenous seed proteins to ensure that a sufficient amount of nitrogen is present in the matured seed (Hartweck and Osborn, 1997; Iida et al., 1993; Maruta et al., 2001; Ogawa et al., 1989). Goossens et al. (1999) introduced the bean seed protein arcelin-5 (arc5-I) gene into A. thaliana and generated lines that accumulated Arc5 protein in large quantities. They showed that expression of the arc5-I gene was substantially increased by the simultaneous reduction of endogenous 2S albumin genes by antisense technology, compared to plants that only expressed the arcelin gene, indicating that the suppression of endogenous storage protein genes facilitated a high expression of the foreign gene. These results suggest that the introduction of a foreign gene into a low storage protein mutant or transformant is expected to enhance its expression levels, compared to its introduction into the wild-type plant, because greater biosynthetic capacity would be allocated to the transgene. However, to date there have been no reports regarding the contribution of low storage protein mutants to the accumulation levels of a transgene. If a foreign gene were directly introduced into a low storage protein mutant, expression levels would be influenced by many factors, including the site of integration (position effect) and copy number: thus we cannot easily estimate the effect of mutants on accumulation levels.

To make a quantitative assessment of how reduced endogenous storage protein production might affect accumulation in transgenic mutants, we carried out an experiment using low-glutelin hybrid lines obtained by crossing the homozygous transformant, 11-5, which expresses soybean glycinin, with two different low-glutelin mutants, LGC-1 and α123 (Figure 1). As described in Figure 3 and Table 1, these results clearly indicated that transgene products were more than twofold enhanced within low storage protein mutants compared with the wild-type rice, irrespective of different genetic backgrounds. It is important to note that the total seed protein content and seed weight in these low storage protein mutants did not significantly differ from the wild-type (Table 1).

Taking these findings together, it was concluded that storage protein mutants are superior to the normal one in their ability to express large quantities of highly valuable foreign proteins, such as pharmaceuticals and industrial enzymes, thus providing a new production platform for high-value recombinant proteins.

Although the yield of recombinant proteins per seed is expected be enhanced in low glutelin mutants, it is not clear whether there is any penalty in the yield of recombinant proteins obtained from harvested seeds measured per unit area. Because the total recombinant protein per unit area is important from a molecular farming perspective, field tests to examine this measure will be required in the near future.

Transgene accumulation levels are mainly regulated at the post-translational level

To determine what molecular mechanisms are involved in these elevated expression levels, the mRNA levels of the transgene and several storage protein genes were examined during seed maturation. Maruta et al. (2001) demonstrated that the mRNA level of the 13 kDa prolamine was enhanced in a low-glutelin transgenic rice line compared to the wild-type. However, we did not find any noticeable differences in the mRNA levels of 13 kDa prolamine and 26 kDa globulin between the normal and low-glutelin lines (Figure 5). Taking these findings together, we conclude that enhancement of endogenous storage proteins is mainly controlled at the post-transcriptional level, although there is little difference in mRNA levels due to differences in physiological conditions, such as the developmental stage used for analysis.

Our data clearly demonstrate that a high accumulation of glycinin in low storage protein trait seeds does not result from the enrichment of the mRNA level, because the glycinin mRNA levels at 15 DAF and 20 DAF showed little difference between the two hybrid lines and their parental 11-5 line. Post-transcriptional regulation might be mainly involved in this enrichment of glycinin accumulation. We have recently developed transgenic rice lines that accumulate glycinin mRNA at levels approximately fivefold higher than that of the 11-5 line under the control of the modified strong glutelin GluB-1 promoter (manuscript in preparation). It is notable that the accumulation levels of glycinin in this line are only ≈ 20% higher than in the 11-5 line (data not shown). This lends further weight to the hypothesis that the accumulation level of glycinin is controlled at the post-transcriptional level; for example, during translation, packaging or storing of foreign gene products into specific cell compartments.

Analysis of seed proteins on SDS-PAGE electrophoretic gel indicated that the differences in the seed protein expression patterns between the parental transformant 11-5 and its low glutelin mutant hybrids were clearest in the middle and later stages of seed maturation (Figure 8). Elevated levels of prolamine and globulin appeared to be achieved at later stages of seed development. Investigation of the increased accumulation ratio of glutelin, globulin, prolamine and glycinin, calculated from the density of bands on the gel, indicated a prolonged period of protein synthesis in the hybrid lines. Prolonged protein synthesis might arise from the use of amino acids that would otherwise be used for glutelin synthesis in the wild-type. It is notable that protein storage bodies are not fully stocked in the low-glutelin mutants, conversely that it might not be possible to completely package synthesized storage proteins into protein bodies in the wild-type seed, and that they might not be fully usable during maturation. Because the low-glutelin mutants have extra space, it is possible to accumulate foreign proteins at higher levels than in the normal plants.

Accumulation levels of individual seed proteins are influenced by the expression levels of other storage proteins

The LGC-1 and α123 lines both possess low levels of glutelin, but their total protein content is almost the same (6%) as the original cultivar (Iida et al., 1997). When total amino acid levels in mature seeds were examined for hybrid mutant lines, parental transgenic rice 11-5 and wild-type, there was little significant difference among them. Their levels are 6.5–7.4% of dry seed weight, which are similar to those (6–8%) of many japonica rice varieties reported so far. Furthermore, it is interesting to note that the genetic backgrounds of low level glutelin have little effect on the weights of the seeds. In wild-type rice, glutelin accounts for 60–80% of the total protein seed content. Prolamine and globulin levels are about 25% and 15%, respectively (Figure 4). In the seeds of the low-glutelin lines, the accumulation levels of prolamine and globulin are higher than those of the wild-type rice. This suggests that deficiency in one seed protein gives rise to the enhancement of other endogenous seed proteins. The prolamine contents in the low-glutelin mutants were particularly enhanced. However, it is interesting to note that the enhancement levels of the prolamine and globulin contents of low-glutelin mutants were not equivalent.

When the transgene was introduced into the low-glutelin mutants, the accumulation levels of glycinin were more than twofold increased compared to that in the parental transformant, 11-5 (Figure 3, Table 1). It is notable that the accumulation levels of endogenous seed proteins decreased. For example, prolamine or glutelin contents were slightly decreased in both hybrids when compared to those of the parental low-glutelin lines LGC-1 and α123 (Figure 4).

 The alcohol soluble fraction contains 16 kDa, 13 kDa and 10 kDa prolamines. The targeting of these prolamines into protein bodies (PBs) is different from that of globulin and glutelin. In rice, glutelin and 26 kDa globulin are stored in a vacuolar compartment (PBII), whereas prolamines are packaged in ER-derived PBI (Takaiwa et al., 1999; Tanaka et al., 1980) (Figure 9). We expected that low-glutelin lines might have higher capacities for the accumulation of transgene product, because low-glutelin mutants have a larger space for protein accumulation in the PBII and high levels of the amino acids required for protein synthesis, compared to those of the wild-type rice (Figure 9). If accumulation levels were mainly determined by space for protein accumulation, the glycinin accumulation level as a foreign product would be influenced by glutelin and 26 kDa globulin, because they share space with each other in the PBII (Katsube et al., 1999). However, it is noteworthy that prolamines stored in the PBI are dominantly affected, accompanying any decrease in glutelins. Inflow of nitrogen or carbon source in rice seed might be consistent with protein synthesis, because the total amount of seed proteins were not changed and the decrease of glutelin in the low-glutelin lines was compensated for by the synthesis of other proteins, such as prolamine, globulin and transgene product glycinin, irrespective of the differences in the PBI and PBII. Once the source for protein synthesis, such as amino acids, becomes available to the seed, it might be necessary to be fixed as protein and stored in an appropriate cellular compartment, such as the PBI or PBII, to enable subsequent germination of the seedling. The changes in accumulation levels of the prolamines are remarkably dependent upon the contents of the other endogenous storage proteins. Prolamine contents might therefore act primarily as a regulator maintaining the physiological condition of the seed, such as its nitrogen balance.

Figure 9.

Working model for high accumulation of foreign gene product in low storage mutant.

Experimental procedures

Plant materials

Homozygous transgenic rice (Oriza sativa L., cv Matsuyamamii) expressing soybean glycinin was used as the parental host for crossing (Katsube et al., 1999). This transgenic rice (11-5 line) was crossed with two lines of low-glutelin mutants, LGC-1 and α123 (Iida et al., 1993; Iida et al., 1997). The half seed technique was used for the selection of homozygous lines.

Protein analysis

Protein extraction and immunological detection by Dot blotting analysis

Mature grains from each line were ground into a fine powder using a Multi-beads shocker (Yasui Kikai, Osaka). To 25 mg of the milled grain, 1 mL of protein extraction buffer (62.5 mm Tris-HCl buffer (pH 6.8) containing 2.5% (w/v) SDS, 10% (v/v) glycerol and 5% 2-mercaptoethanol) was added. The total seed proteins were then extracted from the milled grain by vigorously mixing for 1 h at room temperature. After centrifugation, the supernatant for the total seed protein (×1) was successively diluted two times (×2) and four times (×4) with protein extraction buffer. Equal amounts of the mixtures (≈ 0.1 µL) were spotted on to a nitrocellulose membrane using a dot-blotter (ATTO Multi-pin-blotter AB-6690). Purified soybean glycinin A1aB1b diluted with ×4 protein extract of wild-type rice (7–28 µg of glycinin/mL) was also spotted on to the same membrane as standard. The glycinin level in each dot was detected immunologically with antiglycinin serum followed by goat anti-rabbit IgG-Horseradish peroxidase conjugate (Promega). The glycinin accumulation levels were determined by comparing the densitometry signals obtained from the ×4 extracts prepared from each line with those of given amounts of the purified glycinin standards. Concentrations of crude total seed proteins and glycinin standards used for this assay were within a linear range. The unknown samples were quantified against the standard curve using a linear curve fitting procedure.

The total amount of seed protein was determined by the Kjerdahl method, using a conversion factor of 5.95 for rice proteins.

Western blotting analysis

Each mature seed was ground to a fine powder using a Multi-beads shocker (Yasui Kikai, Osaka). SDS-Urea solution (4% SDS, 8 m Urea, 5% 2-mercaptoethanol, 50 mm Tris-HCl (pH 6.8), 20% (v/v) glycerol) was added to the seed powder and the solution vortexed and centrifuged to collect the supernatant from the total protein extract (Iida et al., 1993), which was then applied to 15% (w/v) SDS–polyacrylamide gels. After electrophoresis, the gels were either stained with Coomassie Brilliant Blue R250 or blotted on to PVDV membranes (Immobilon P; Millipore). After pre-treatment of the blots with 5% (w/v) skimmed milk in TBST (20 mm Tris-HCl (pH 7.5), 0.9% NaCl, 0.01% Tween-20) for 1 h at room temperature, the protein gel blots were probed with polyclonal antibody raised in rabbit against glycinin diluted to 1 : 5000 with TBST containing 5% (w/v) skimmed milk. After 2 h at room temperature, the blots were washed four times with TBST for 10 min at room temperature followed by treatment for 1 h at room temperature with Horseradish peroxidase conjugated goat anti-rabbit IgG antibodies (Promega, USA), diluted to 1 : 2500 with TBST. After incubation, the blots were washed at least four times with TBST for 10 min at room temperature. Finally, the blots were developed using an enhanced chemiluminescence (ECL) system (Amersham Pharamacia Biotech, USA).

Separative protein extraction

Milled rice was ground into fine powder with the Multi-beads shocker. A 25 mg sample of the rice powder was used for protein extraction. 200 µL of albumin-globulin extraction buffer (0.5 m NaCl, 10 mm Tris-HCl pH 6.8) was added to the rice powder, sonicated on ice and centrifuged. This process was repeated and the supernatant was combined. Four times the volume of acetone was added to the supernatant to obtain the albumin-globulin fraction. To separate the glycinin fraction from the total albumin-globulin fraction, an equal volume of acetone was added to the supernatant extracted with albumin-globulin extraction buffer. The residual globulin-albumin fraction was precipitated by the addition of a threefold volume of acetone to the supernatant after extraction of the glycinin fraction.

The prolamine fraction was extracted from the precipitate after extraction of the total albumin-globulin fraction with 60% n-propanol with 5% (w/v) 2-mercaptoethanol. This process was repeated and the supernatants were combined. An equal volume of chilled water was added to the combined supernatant, placed at 4 °C for 1 h, and then centrifuged to obtain the prolamine precipitate.

The glutelin fraction was extracted from the residual precipitate after prolamine extraction with 1% lactic acid. The solubilized fraction was neutralized with 1 n NaOH to precipitate again to obtain the glutelin fraction.

RNA analysis

Isolation of RNA

Total RNA was extracted from maturing seeds from 5 to 25 DAF (days after flowering) as described previously (Takaiwa et al., 1987). The maturing seeds were rapidly frozen in liquid nitrogen and ground in a mortar, then suspended with phenol/chloroform/isoamylalchol (25 : 24 : 1) and extraction buffer (0.1 m Tris-HCl, pH 9.0, 1% SDS, 5 mm EDTA, 0.1 m NaCl). The aqueous phase was corrected by centrifugation and nucleic acids were precipitated with ethanol. Total RNA was specifically precipitated in 2 m LiCl.

Northern analysis

Ten micrograms of total RNA was electrophoresed in a 1.2% (w/v) formaldehyde-containing agarose gel and then blotted on to a nylon membrane (Hybond N+, Amersham Biosciences). The resulting nylon membrane was then hybridized in 6 × SSC, 0.5% SDS, 5 × Denhardt's solution at 65 °C for 8 h with 32P-labelled DNA fragments. Membranes were washed twice in 2 × SSC and 0.1% SDS at room temperature and then twice in 0.1 × SSC and 0.1% SDS at 55 °C.

DNA analysis

Total DNA was prepared from rice leaves according to the CTAB (Hexadecyltrimethylammonium bromide) method (Milligan, 1989) with minor modifications. Rice leaves (≈ 0.5–1 g) were frozen, ground into fine powder, then suspended in 2 × CTAB buffer (0.1 m Tris-HCl (pH 8.0), 20 mm EDTA, 1.4 m NaCl, 2% CTAB, 0.2% 2-mercaptoethanol) and incubated at 60 °C for 20 min. After centrifugation, the supernatant was extracted with chloroform and isoamyl alcohol (24 : 1, v/v). Nucleic acids were precipitated with ethanol, and then treated with RNase. 1 µg of DNA was digested with restriction enzymes and then loaded on to a 0.8% agarose gel. Following electrophoresis, Southern blotting analysis was carried out.

Acknowledgements

We wish to thank Drs T. Nishio and S. Iida for providing the rice mutants LGC-1 and α123, and Ms M. Utsuno, F. Ito, Y. Suzuki for technical assistance. The 13 kDa prolamin and 26 kDa globulin cDNA clones were provided from the NIAS DNA bank. This research was supported by research grants from the Ministry of Agriculture, Forestry and Fishery of Japan (Integrated studies of new rice production technology by breeding superior varieties for the next millennium) to F.T.

Ancillary