mRNA localisation during development

Although there are many differences, mRNA localisations in the Xenopus oocyte show some tantalizing similarities to those occurring in Drosophila development. As in Drosophila, transcripts localise to opposite poles of the oocyte, this localisation is hierarchical and occurs in a multistep process in which localisation is followed by anchoring at the cortex. This distinction between initial transport and long-term maintenance reflects the dynamic nature of the cytoskeleton: the microtubule tracks form and reform according to the needs of the cell so that stable localisation must be mediated by a more constant structure--the cortex. A possible exception is the localisation of gurken mRNA where it is unknown whether there are separate mechanisms for transport to and maintenance at the oocyte nucleus. However, gurken is responsible for the transmission of a transitory signal; once this has been received, and the fate of the recipient follicle cells determined, there is no further need for localisation. It is possible that the time scale over which the localisation machinery is stable is sufficient for transmission of this signal without the need for a separate maintenance phase. The existence of a nanos homologue, Xcat-2 (Mosquera et al., 1993), associated with the Xenopus germ plasm is particularly interesting because of the morphological and functional similarities between Drosophila polar granules, Caenorhabditis P-granules, and Xenopus germ plasm. These electron-dense protein-RNA complexes are maternally supplied and in each case segregate with the germ line. These granules may represent a fundamental conserved pathway to germ-cell specification and it is now at least a possibility that they are also involved in establishing the embryonic axis through translational repression. In the case of Drosophila, this occurs through localised nanos acting (via Pumilio) on nanos response elements in hunchback mRNA. No such regulatory pair has yet been demonstrated in C. elegans or X. laevis, but each contains a candidate for one half of the interaction: glp-1 could be a target for an unidentified nanos-like protein; Xcat-2 may control translation of an unknown NRE-containing mRNA. Another common feature of mRNA localisation is that in every case where the targeting signal has been determined, it has been mapped to a region of the 3' UTR capable of forming an extensive secondary structure (e.g., David and Ish-Horowicz, 1991; Dalby and Glover, 1992; Gavis and Lehmann, 1992; Kim-Ha et al., 1993; Kislauskis et al., 1993, 1994; Lantz and Schedl, 1994). In several cases, translational control and transcript stability signals have also been mapped to these regions (Jackson and Standart, 1990; Standart et al., 1990; Standart and Hunt, 1990; Davis and Ish-Horowicz, 1991; Wharton and Struhl, 1991; Dalby and Glover, 1993; Evans et al., 1994; Kim-Ha et al., 1995). The large secondary structures may provide a means for stably exposing sequence-specific regions of RNA to proteins. Due to the ease with which RNA forms base pairs, it is likely that rather than remaining single-stranded, RNA within the cell forms some sort of secondary structure. The geometry of purely double-stranded RNA does not permit sequence specific interactions between proteins and the bases because the major groove is inaccessible to amino acid side chains (Weeks and Crothers, 1993). However, the distortions to the dsRNA helix provided by bulges, pseudoknots, and the single-strand loop regions in stem-loop structures do present sequence information that can be "read" by proteins. The extensive 3'UTRs may produce a stable secondary structure which ensures that regulatory elements remain exposed in such regions rather than hidden in double-stranded stems. (ABSTRACT TRUNCATED)