The sorting of rice storage protein RNAs to distinct ER subdomains, and the packaging of these proteins into separate endomembrane compartments, suggests that RNA localization and protein localization are interrelated. Such a relationship is obvious for prolamine RNAs and the immediate assembly of the coded protein product upon translocation into the lumen. Additionally, proteins synthesized on the cisternal ER would presumably be in close proximity to the transitional ER, a requirement for their export to the Golgi complex or storage vacuole. Because storage protein RNA localization requires specific cis elements, this hypothesis relating RNA and protein localization can be directly tested. A hybrid 10-kDa 8-zein RNA containing a 3'UTR from the nopaline synthase (NOS) gene is normally localized to the PB-ER in developing rice endosperm as mentioned before. Immunocytochemical studies of ultrathin sections and transmission electron microscopic analysis showed that 10-kDa 8-zein polypeptides are localized to PB-I, the site of their synthesis (Fig. 3). Replacement of the 3'NOS sequence with the glutelin 3'UTR results in the displacement of the RNA from the PB-ER to cisternal ER (Hamada et al. 2003b). This change in RNA localization is mediated by the cis element(s) contained within the glutelin 3'UTR, and takes place even though the 10-kDa 8-zein coding sequence contains three cis elements that direct PB-ER localization of the RNA. Further analysis has shown that this altered RNA
localization pattern also changes the localization pattern of the 10-kDa 8-zein protein. The 10-kDa 8-zein was not found in PB-I but in the storage vacuole PB-II (Fig. 3). The final destination of the 10-kDa 8-zein protein therefore depends on which specific ER subdomain is used as the initial site of protein synthesis.
Several questions that arise from this study center on the mechanism by which the 10-kDa 8-zein protein, which is normally deposited as part of an intracisternal inclusion granule, is transported from the cisternal ER to PB-II. One property of the 10-kDa 8-zein is that it appears to be incapable of forming a protein body-like structure by itself in transgenic plants. Lending and Larkins (1989) have demonstrated that the synthesis of the various zein classes shows a strict temporal and spatial pattern during endosperm development, with the P- and y-zeins laid down first followed by the a- and 8-zeins, which displace the former zein classes to the peripheral regions of the protein bodies. Coexpression of y- or P-zein is needed for stable a- and 8-zein accumulation and aggregation to form protein bodies in transgenic tobacco (Bagga et al. 1995; Coleman et al. 1996; Bagga et al. 1997; Kim et al. 2002; Coleman et al. 2004), suggesting that the 10-kDa 8-zein (as well as a-zeins) is incapable of self-assembly in a larger macromolecular structure and requires
Fig. 3 Protein-A gold immunocytochemistry showing the localization of 10-kDa 8-zein in distinct protein bodies. The 10-kDa 8-zein and hybrid zein-glutelin constructs were introduced into rice by Agrobacterium-mediated transformation. In situ RT-PCR analysis revealed that 10-kDa 8-zein RNAs were localized only to the PB-ER, whereas the hybrid zein-glutelin RNAs were targeted to cisternal ER (Hamada et al. 2003b). 8-Zein polypeptides coded by RNAs containing the NOS 3'UTR are observed in PB-I (a) , while those coded by RNAs containing the glutelin 3'UTR are localized to PB-II (b). Scale bar = 1 |im
Fig. 3 Protein-A gold immunocytochemistry showing the localization of 10-kDa 8-zein in distinct protein bodies. The 10-kDa 8-zein and hybrid zein-glutelin constructs were introduced into rice by Agrobacterium-mediated transformation. In situ RT-PCR analysis revealed that 10-kDa 8-zein RNAs were localized only to the PB-ER, whereas the hybrid zein-glutelin RNAs were targeted to cisternal ER (Hamada et al. 2003b). 8-Zein polypeptides coded by RNAs containing the NOS 3'UTR are observed in PB-I (a) , while those coded by RNAs containing the glutelin 3'UTR are localized to PB-II (b). Scale bar = 1 |im heterotypic interactions with other proteins. In the absence of interacting partners, the 10-kDa 8-zein is likely to be competent for ER export and transport to the storage vacuole.
A second question relates to how 10-kDa 8-zein is targeted to the storage vacuole PB-II. In general, the mechanism by which storage proteins are sorted to the storage vacuole is poorly understood. Both 7S and 11S globulins are exported from the ER where they are concentrated into dense vesicles at the ci's-Golgi (Hillmer et al. 2001). Only phaseolin, a 7S globulin, has been shown to possess a vacuolar targeting signal at its C-terminus (Frigerio et al. 1998), although such targeting signals have also been suggested for the 11S globulins (Saalbach et al. 1991). The presence of a vacuolar sorting signal indicates the existence of a receptor that is responsible for ER export and dense vesicle formation. The 10-kDa 8-zein polypeptides may have a cryptic peptide signal that targets them to the storage vacuoles or they may be escorted to this organelle by their heterotypic interaction with glutelins. Immunocytochem-ical studies showed that the 10-kDa 8-zeins are embedded in the crystalline (glutelin-containing) parts of PB-II, suggesting the possible interaction of these proteins. Conversely, transport to the storage vacuole in rice may be simply a default process. Further studies on protein transport to the storage vacuole in rice should provide information on the existence of any of these pathways.
The preceding description clearly demonstrates that at least for one protein, the 10-kDa 8-zein, the localization of its RNA has a profound impact on where the protein is deposited in the cell. The converse experiment to support the relationship between RNA localization and protein localization is to mistarget an RNA from the cisternal ER to the PB-ER. The sunflower seed 2S albumin (SSA) is stored in the protein storage vacuole in both sunflower and transgenic rice. Closer examination showed that SSA protein is localized on the periphery of PB-II (Washida et al. unpublished) much like the 26-kDa globulin (Krishnan et al. 1986). Consistent with this location in the storage vacuole, its RNA is observed on the cisternal ER much like that seen for glutelin RNAs. Targeting of SSA RNA to the cisternal ER suggests the presence of cis elements functionally equivalent to those present in glutelin RNA (Hamada et al. 2003b)
The SSA RNA was modified to contain the 5' coding sequences of the 10-kDa 8-zein RNA, which contains one or two PB-ER cis elements. In both instances, a portion of the hybrid 10-kDa 8-zein-SSA RNA was found to be partially mislocalized to the PB-ER in addition to the remainder which was targeted to its normal location on the cisternal ER. Cytologic examination using SSA antibodies showed that the SSA protein was found in both PB-I and PB-II (Washida et al. unpublished). Although SSA RNA could not be totally displaced to the PB-ER, the results do support the hypothesis that the PB-ER delimits a unique ER subdomain for the confined localization of proteins.
In PB-I, SSA was distributed to the periphery of the inclusion granule. Such a spatial arrangement is similar to that seen for the zein protein bodies, where the P- and y-zeins are displaced to the periphery around the a- and 8-zeins. The peripheral distribution pattern also indicates that SSA does not interact with prolamines and self-assembles around the more hydrophobic prolamine inclusion granule. The physical processes that account for the self-assembly of SSA within PB-I are not known, but it is likely that they are identical to those used for the concentration of SSA in Golgi-derived dense vesicles.
Was this article helpful?