Index References

The expression of seed storage protein genes is regulated spatially and temporally during embryogenesis (Perez-Grau and Goldberg, 1989). Both transcriptional and posttranscriptional processes are involved in the regulation (Walling et al., 1986). In addition, multiple cellular events including translocation, processing, folding, assembly, and packaging of proteins occur subsequent to the protein synthesis (reviewed by Shewry et al., 1995); thus, regulation of these events may also affect protein accumulation. In embryos, mRNA and/or protein are sometimes present at different levels in different cell types such as the cotyledon and embryonic axis. Although both cis-acting (Bustos et al., 1991; Bäumlein et al., 1992; Kawagoe et al., 1994; Chandrasekharan et al., 2003) and trans-acting (Nambara et al., 1995; Vicente-Carbajosa et al., 1997; Suzuki et al., 1997; Parcy et al., 1997; Ezcurra et al., 2000; Reidt et al., 2000; Kroj et al., 2003) factors involved in the control of embryo-specific transcription have been identified, little is known about mechanisms that govern the relationship between differentiation into cell types in the embryo and accumulation of proteins. Analyses of Arabidopsis thaliana mutants that exhibit an altered organization of the embryo body plan have led to the understanding that plant embryos form from disparate regions (modules) that develop independently of each other (reviewed by Goldberg et al., 1994). According to this concept, a possible mechanism underlying cell-type specific protein accumulation in the embryo involves activation of gene expression controlled by transcriptional regulatory networks unique to each module.

β-Conglycinin, a major component of seed storage proteins in soybean (Glycine max), is a trimeric protein composed of various combinations of three subunits: α, α′, and β (Higgins, 1984; Thanh and Shibasaki, 1976a, 1976b). Differences in the spatial and temporal pattern of protein accumulation in developing soybean embryos have been detected among these three subunits (Ladin et al., 1987; Meinke et al., 1981; Harada et al., 1989). While the α′ subunit protein is present at similar levels in the cotyledons and the embryonic axis, the β subunit protein is present in the cotyledons but not in the embryonic axis. The α subunit protein is present in both the cotyledons and the embryonic axis but the level is lower in the latter (Meinke et al., 1981). Thus, the α subunit is synthesized and/or accumulated in a more complex manner than the other two subunits. In vitro translation of mRNA isolated from the cotyledons produced a higher level of α subunit protein compared with that of mRNA isolated from the embryonic axes (Ladin et al., 1987). This experiment indirectly indicated that the accumulated mRNA level of the α subunit gene is higher in the cotyledons than in the embryonic axis, and this may account for the higher level of α subunit protein in the cotyledons. The expression pattern of β-glucuronidase (GUS) in the embryos of transgenic petunia or A. thaliana plants carrying the α′ subunit gene or β subunit gene promoters fused to the GUS gene in fact did reflect the spatial distribution of the proteins in soybean embryos (Naito et al., 1988, 1994). However, mechanisms responsible for the spatial and temporal control of the transcription of the β-conglycinin subunit genes within the embryo have not yet been elucidated.

Regulatory elements involved in the transcription of the α′ subunit and β subunit genes have been studied by means of a reporter gene assay in transgenic plants and a binding assay using nuclear extracts (Allen et al. 1989; Chamberland et al. 1992; Chen et al., 1989; Fujiwara and Beachy, 1994; Lessard et al., 1991). In contrast, despite its importance in terms of nutritional value and allergenic activity of the α subunit in humans, until recently, little was known about the regulatory elements involved in the transcriptional control of the α subunit gene. We previously identified the β-conglycinin α subunit gene (Yoshino et al., 2001, 2002, 2006; DDBJ/EMBL/GenBank accession number AB237643). To understand the regulatory mechanisms that control expression of the α subunit gene during seed development, we made transgenic A. thaliana plants containing reporter gene constructs comprising the upstream sequence of the α subunit gene and the GUS gene. We found that embryo-specific transcriptional activation of the α subunit gene promoter is closely associated with the presence of the RY sequence, a cis-acting element involved in the regulation of various seed storage protein genes (Yoshino et al., 2006), as previously reported for the α′ subunit gene promoter (Chamberland et al., 1992; Chen et al., 1989; Fujiwara and Beachy, 1994). Because of the quantitatively different distribution of α subunit protein between the cotyledons and embryonic axis in soybean embryos, analysis of the α subunit gene promoter in the transgenic plants provides a good tool for understanding mechanisms of unequal distribution of the protein between different embryonic tissues. Here, we analyzed spatial and temporal control of α subunit gene expression during embryogenesis using transgenic A. thaliana plants carrying the promoter–reporter construct. We report that unequal distribution of α subunit protein within the embryos is established as a consequence of differential transcriptional activation controlled by a short proximal promoter region of the α subunit gene in different embryonic tissues.

The materials and methods used in this study are as follows. Transgenic A. thaliana plants containing reporter gene constructs comprising the upstream sequence of the β-conglycinin α subunit gene up to –1357 relative to the transcription start site (Yoshino et al., 2001) or a series of its 5′-deleted derivatives and the GUS gene (Yoshino et al., 2006) were used. The 5′-deleted derivatives included constructs containing the upstream regions up to –867, –545, –402, –245, –161, or –73. The T₂ seeds were sown on soil, and the plants were grown under 16 h light and 8 h dark at 24°C. T₃ seeds were used for the GUS activity assay. For each promoter–reporter construct, two or three lines were randomly chosen from six independently transformed lines and were used for histochemical staining. Histochemical staining for GUS activity was carried out as described by Jefferson et al. (1987). Embryos were dissected from T₃ seeds under a microscope and then stained with GUS staining solution (Jefferson et al., 1987) at 37°C overnight. After staining, embryos were immersed twice in 95% ethanol at 37°C for 30 min to remove chlorophyll. Thirty embryos were dissected from T₃ seeds at the fully mature stage for each transgenic line and were used for the assay. To examine changes in GUS expression during seed development, five or six embryos at each time point after flowering were analyzed for each line. To analyze GUS activity independently in the cotyledons and embryonic axis, embryos were dissected from T₃ seeds, and the isolated embryos were then divided into cotyledons and embryonic axes under a microscope. The isolated cotyledons and embryonic axes were soaked in extraction buffer (Jefferson et al., 1987) at room temperature, and ground with a mortar and pestle in the presence of extraction buffer. The following enzymatic assay was done as described previously (Yoshino et al., 2006). For each material, two embryos were used for the assay, and the assay was repeated three times.

To test the hypothesis that the distribution of α subunit protein within embryos is primarily controlled by transcription of the gene, spatial distribution of GUS expression was analyzed by histochemical staining using transgenic A. thaliana plants carrying reporter gene constructs comprising the upstream sequence of the α subunit gene and the GUS gene. We have reported that the seed-specific expression of the α subunit gene is maintained in transgenic A. thaliana plants and that transgenic A. thaliana plants are useful for analyzing the upstream regulatory elements of the gene (Yoshino et al., 2006) as previously shown for the promoters of seed-storage protein genes, such as the β-conglysinin α′ subunit and β subunit genes of soybean (Naito et al., 1994) and the β-phaseolin gene of bean (Chandrasekharan et al., 2003). Prominent GUS activity was detected in the entire embryo in the fully mature stage of development (14 days after flowering; DAF) in plants containing the upstream region up to –1357 (Fig. 1a), an upstream region that sufficiently confers embryo-specific expression of the gene (Yoshino et al., 2006). In contrast, only portions of embryos were stained when the embryos were at earlier stages of development. This observation suggests that distribution of GUS expression changes during embryogenesis, which prompted us to analyze time-course-dependency of the pattern of GUS expression.

View Details

Fig. 1 Distribution of GUS expression in fully mature embryos in transgenic A. thaliana plants and putative regulatory elements in the proximal promoter region. (a) Diagram of the upstream regions of the α subunit gene that were fused to the GUS gene and histochemical staining of embryos indicating the GUS activities conferred by respective reporter constructs at 14 DAF. Locations of the RY elements and the TATA box are indicated by vertical lines. The embryo of a plant containing no reporter construct was shown as a control. A scale bar denotes 100 μm. Note that GUS expression was detected in the entire embryo when 245-bp or longer sequences of the upstream region were fused to the GUS gene. (b) Distribution of putative regulatory elements in the upstream region up to –245. The locations of the RY elements, the G-box-like sequence, the CCAAAT box, the CACA element, the vicilin box, the ACGT motif, and the TATA box are shown. Nucleotide positions are numbered relative to the major transcription start site (Yoshino et al., 2001).

Changes in the distribution of GUS expression in the embryos were analyzed at different time points after flowering. GUS expression was not detected in the embryos at 4 DAF (data not shown), but was detected at 6 or 8 DAF as well as later stages of development (Fig. 2, top row). At 6 DAF, GUS expression was detected in the cotyledons, but was not or barely detected in the embryonic axis. Thus, the pattern of GUS expression differed between these tissues. GUS expression was first detected in the outer edge of each cotyledon, and then spread over the cotyledon (Fig. 2), which probably reflects the progressive transcriptional activation of the gene as was suggested previously (Perez-Grau and Goldberg, 1989). At later stages (e.g., 12 DAF), GUS expression was detected in the entire embryo (Fig. 2).

View Details

Fig. 2 Time-course dependent changes in the distribution of GUS expression in developing embryos. Representative results obtained for each promoter-reporter construct at 6 to 12 DAF are shown. A scale bar denotes 100 μm. Transformed lines are named based on the length of the upstream region of the α subunit gene fused to the GUS gene. Note that embryos exhibiting no visible GUS staining and those exhibiting weak staining were both detected in the same plants at 6 DAF, whereas all embryos exhibited GUS staining at 8 DAF, indicating that GUS expression commenced at 6–8 DAF. GUS expression was first detected in the cotyledons, and was also detected in the embryonic axis at the subsequent stages. This pattern was commonly detected when 245-bp or longer sequences of the upstream region were fused to the GUS gene.

Although GUS activity could be assessed using histochemical staining, the intensity of staining appeared to be saturated because of a very high level of GUS expression, particularly in the later stages of embryogenesis. To quantify the level of GUS expression, cotyledons and embryonic axes were isolated from the embryos by dissection under a microscope, and extracts from these tissues were used for the fluorometric assay of GUS activity (Fig. 3). GUS activity was measured at 7, 8, and 14 DAF using line 1357-1, which exhibited the highest GUS activity of 42 independently transformed lines (data not shown). At 7 DAF, GUS activity was obvious in the cotyledons, but not in the embryonic axes (Fig. 3). From 7 to 14 DAF, GUS activity dramatically increased in the cotyledons: there was a 19-fold increase in activity. GUS activity was also detected in the embryonic axes at 8 DAF and increased up to 14 DAF. However, the GUS activity at 14 DAF was still three-fold higher in cotyledons than in the embryonic axes (Fig. 3). Thus, GUS activity was higher in the cotyledons than in the embryonic axis throughout the period of its expression during embryogenesis, which is coincident with the distribution of the α subunit protein in developing soybean embryos. These results suggest that accumulation of α subunit protein at a higher level in the cotyledons than in the embryonic axis in developing soybean embryos can be ascribed primarily to a difference in the extent of transcriptional activation of the gene between these two tissues, although involvement of posttranscriptional expression control has also been assumed for the expression of seed storage protein genes previously (Walling et al., 1986).

View Details

Fig. 3 GUS activities in the cotyledons and the embryonic axis during embryogenesis in transgenic A. thaliana line 1357-1. Cell extracts of the cotyledons and the embryonic axes dissected from developing embryos were used for GUS enzymatic assay. Gray and open bars indicate GUS activities in the cotyledons and the embryonic axes, respectively. Average activity with the standard error obtained from three replicates of the analysis is shown.

To identify the promoter region responsible for the control of transcriptional pattern of the α subunit gene, we tested a series of 5′-deleted derivatives of the upstream region for their ability to transcribe the GUS gene in A. thaliana embryos. Prominent GUS activity was detected in the entire embryos at 14 DAF from plants containing the upstream regions up to –867, –545, –402, and –245, whereas no GUS activity was detected in embryos from plants containing the upstream regions up to –161 and –73, as well as plants containing no reporter construct (Fig. 1a). For each construct, no difference in staining pattern was detected between independently transformed lines (data not shown). The presence or absence of GUS activity depending on the extent of the deletion was consistent with results of previous quantitative analyses of GUS activity using extracts of siliques containing embryos at the same stage (Yoshino et al., 2006). These results indicate that the proximal promoter region up to –245 was sufficient to confer GUS expression in the entire embryos at 14 DAF, and the elements responsible for transcriptional activation in the cotyledons and in the embryonic axis were both localized in the region spanning –245 to –161.

GUS expression in the earlier stages of development (6–12 DAF) was also analyzed for plants containing the upstream regions up to –867, –545, –402, and –245, which exhibited GUS expression at the fully mature stage. A substantially identical pattern of GUS expression, characterized by preferential expression in the cotyledons, was detected for all of these plants, and the pattern was the same as that in plants containing the upstream region up to –1357 (Fig. 2). Therefore, elements located in the proximal promoter region up to –245 were sufficient to control the pattern of transcription, which most likely reflects a difference in the extent of transcriptional activation between the cotyledons and the embryonic axis of the α subunit gene during embryogenesis.

No difference in the pattern of GUS expression was detected irrespective of the length of the upstream region of the α subunit gene as long as the regions contained 245 bp or longer sequences upstream of the transcription start site (Fig. 2). On the other hand, our previous analysis indicated that the average level of GUS activity of extracts of T₃ seeds at the fully mature stage, which was obtained by analyzing five independently transformed lines for each construct, varied depending on the length of the upstream region (Yoshino et al., 2006). These results suggest that the transcriptional pattern is controlled by elements located in the upstream region up to –245, and differences in the level of GUS gene expression, rather than differences in the number of cells that express the GUS gene, account for the differences in gross GUS activity in the embryos. Accordingly, elements located in the region upstream of –245 should influence transcription rate without affecting spatial and temporal pattern formation of transcription.

Multiple elements that function as cis-acting elements responsible for activation of transcription in embryos are present in the region spanning –245 to –161 (Fig. 1b). It is very likely that the RY elements, the CCAAAT box, and the CACA element present in this region, which play a positive regulatory role (Chandrasekharan et al., 2003), coordinately participate in the control. In addition, a strong candidate element responsible for the difference in the extent of transcriptional activation between the cotyledons and the embryonic axis may be the G-box-like sequence present at –229 to –224. Site-directed mutagenesis of the G-box in the bean β-phaseolin gene promoter resulted in a dramatic decrease in the level of gene expression, especially in the embryonic axis (Chandrasekharan et al., 2003). The G-box-like sequence of the α subunit gene promoter and the G-box of the β-phaseolin gene promoter are present at similar positions in a sequence context adjacent to two RY elements, a structure also conserved in the Brassica napus napin napA promoter (Ezcurra et al., 1999). However, the G-box-like sequence of the α subunit gene promoter comprises CACGTA, in which the sixth nucleotide is different from the original G-box, CACGTG (Kawagoe et al., 1994). It is possible that this difference may cause a change in affinity between the DNA motif and protein factor(s) and consequently accounts for the reduction in gene expression in the embryonic axis. However, the α′ subunit gene also has the CACGTA sequence at the identical position (for the sequence, see Doyle et al., 1986), and the encoded protein does not accumulate at a lower level in the embryonic axis (Ladin et al., 1987; Meinke et al., 1981). Therefore, assuming that this sequence is involved in quantitative variation in the mRNA level of the α subunit gene in the embryos, it is likely that some other elements in the α subunit gene promoter also play a role in controlling transcription, either positively in the cotyledons or negatively in the embryonic axis. Alternatively, it is also possible that the α′ subunit gene promoter contains elements that compensate for the effect of the base change. Further experiments are required to ascertain which sequences can act as cis-acting elements to give rise to the observed results.

It has been suggested that discrete cis-acting domains are required for transcriptional activation of a gene within each embryonic tissue (Goldberg et al., 1994; Conceição and Krebbers, 1994; Chandrasekharan et al., 2003). A typical example that supports this notion is the soybean major Kunitz trypsin inhibitor gene (Kti3) promoter, in which a sequence covering up to a 2-kb upstream region conferred GUS expression in all regions of a mature transgenic tobacco embryo, whereas deletion of a 0.2-kb region from the 5′ end eliminated GUS expression in the radicle and deletion of another 1-kb region eliminated GUS expression in the cotyledons and shoot meristem (Goldberg et al., 1994). In contrast, the upstream region up to –245 of the β-conglycinin α subunit gene was sufficient to confer transcription to the entire embryo, which is characteristic of this promoter and may be useful for expressing a foreign gene to engineer novel metabolic composition of seeds. The order of transition of mRNA accumulation, from the cotyledons to the entire embryo, of the α subunit gene was also different from that observed for some other seed storage protein genes, e.g., the B. napus napin genes (Fernandez et al., 1991) and the soybean Kti3 gene (Goldberg et al., 1994), indicating diverse patterns of gene expression. A comparative analysis of seed storage protein genes having different patterns of expression, together with experiments involving site-directed mutagenesis, would help understand how a spatial and temporal transcriptional pattern in developing embryos is controlled by multiple cis-acting elements.

We thank Atsushi Nagamatsu for technical assistance, and Michiko Yoshino, Ken-ichi Tsutsumi, and Jun Abe for helpful suggestions. This work was supported in part by Grants-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan, and a grant from Fuji Foundation for Protein Research.