Phenotype of aSTARice
β-carotene is the major compound of Golden Rice. When we endosperm-specifically co-expressed PSY and CrtI genes, the expected phenotype of Golden Rice was obtained (Figure 5). Although re-transformation of Golden Rice using BHY and BKT would get aSTARice (astaxanthin rice), it is time-consuming and labor-intensive. Therefore, the application of transgene stacking the four genes (PSY, CrtI, BHY and BKT) is more effective and easier to directly biosynthesize astaxanthin.


Astaxanthin is an orange-red carotenoid. When we directly stacked the four genes in rice and specially expressed in endosperm, the transgenic polished seeds showed pigmentation of orange-red astaxanthin (Figure 5). Wild-type rice remains colorless without astaxanthin accumulation.
Figure 5   The polished grain phenotype of aSTARice. The yellow arrow indicates production of Golden Rice by stacking PSY and CrtI gene (380MF-PC). The black dot arrow indicates the re-transformation method using BHY and BKT genes. The red arrow indicates direct astaxanthin biosynthesis in rice endosperm by stacking PSY, CrtI, BHY and BKT genes (380MF-BBPC).
Astaxanthin content in aSTARice analyzed by HPLC
To confirm the level of synthetic astaxanthin in aSTARice seeds, we used high performance liquid chromatography (HPLC) to analyze the pigment composition and astaxanthin content from the methanol extracts of aSTARice and wild-type rice HG1 seeds. Astaxanthin is identified on the basis of retention times related to standard sample. Content is determined by integrating peak areas and converted to concentration. According to the retention time of the standard astaxanthin sample, astaxanthin compound of extracts from aSTARice seeds can be confirmed (Figure 6). The HPLC results also showed that astaxanthin was the predominant carotenoid in aSTARice, and did not exist in wild-type rice seeds. The orange–red compound in transgenic seeds mainly is astaxanthin.
Figure 6   HPLC chromatogram of methanol extracts from transgenic aSTARice (red line) and wild-type (blue line) rice seeds. HPLC analysis recorded at 480 nm of extracts.
To calculate astaxanthin content of different transgenic aSTARice lines, we analyzed the relationship between different concentration of the astaxanthin standard sample and its peak areas using HPLC, and we acquired an astaxanthin standard sample curve (Figure 7). By using this curve, we analyzed several transgenic aSTARice seeds. Some lines accumulated astaxanthin up to 5.20 μg/g DW (dry weight). In contrast, no astaxanthin was detected in wild-type rice.
Figure 7   Standard curve of relationship between astaxanthin concentrations and its peak areas.



Investigation of aSTARice in later generation
Figure 8   Relative expression levels of four foreign astaxanthin biosynthetic genes and astaxanthin content in seeds of wild-type and later generation of aSTARice. Their expression levels are normalized to Osactin1 transcript level. All reactions were carried out in triplicate, and each experiment was repeated twice. Error bars indicate ± SEM. DW, dry weight.
The later generation seeds of aSTARice were harvested at the end of September this year and we performed quantitative RT-PCR (qRT-PCR) and HPLC analysis. Several transgenic lines were chosen for investigation. As the results shown in the diagram, astaxanthin content varies from lines, some lines remain stable astaxanthin production while some synthesize in a low level. Supported by the qRT-PCR analysis, the low content of astaxanthin in the endosperm of transgenic lines are associated with poor transcriptional activity of a BHY gene, which no obvious transcriptional activity is detected in four lines. The investigation demonstrates that astaxanthin biosynthesis accomplishes only under the well cooperation of the four genes. When one of the four foreign genes was down regulation of expression, astaxanthin content would be decreased. These four genes are the essential genes for biosynthetic astaxanthin in aSTARice.



Strategy for marker free
The multigene vector 380MF-BBPC has four genes (CrtI, PSY, BKT and BHY) for endosperm-specific synthetic astaxanthin, and a marker-free element comprised of two genes (HPT and Cre) for marker-free deletion. This marker-free element was placed between two loxP sites, and consists of a HPT (hygromycin) resistance gene expression cassette and a Cre gene expression cassette controlled by anther-specific promoter (Figure 9). When Cre gene was expressed in transgenic rice anther, the Cre enzyme deleted the HPT and Cre gene cassettes between two loxP sites. Two primer pairs, F1/R1 and F1/R2, were used to PCR identify the marker-free aSTARice later generations. As is shown in Figure 9a, the expected 400-bp band of F1/R2 PCR indicates marker-free homozygous lines, but the PCR results containing both 540-bp and 400-bp bands indicate marker-free heterozygous lines (Figure 9a). The 400-bp band was further determined by direct sequencing (Figure 9c). And more marker-free homozygous lines would be obtained in progeny.
Figure 9   Marker-free analyses in aSTARice later generations. (A) Gel electrophoresis assays of marker-free specific 400-bp band. M, Marker 2K plus; WT, negative control (wild-type rice HG1); Lane marked as 1-6 are the later generation of aSTARice #8-6. (B) Schematic diagram of marker deletion process. (C). 400-bp PCR sequencing result.
In conclusion, we achieved astaxanthin accumulation in rice endosperm, indicated that rice endosperm bioreactor is an efficient tool for astaxanthin production. Our project demonstrated that rice endosperm is a potential green factory for the economical production of astaxanthin.