Cell, Vol. 42, 769-777, October 1985, Copyright 0 1985 by MIT Identification and Cloning of Localized Maternal RNAs from Xenopus Eggs 0092-8674/85/l 00769-09 $02.0010 M. R. Rebagliati, D. L. Weeks, FL R Harvey, and D. A. Melton Department of Biochemistry and Molecular Biology Harvard University 7 Divinity Avenue Cambridge, Massachusetts 02138 Summary A central question in developmental biology is to ex- plain how cells in different regions of an embryo ac- quire different developmental fates. We have begun to address this question by investigating whether spe- cific RNAs are localized within a frog egg. Differential screening of a cDNA library shows that most mater- nal RNAs are uniformly distributed along the animal- vegetal axis. However, we find that a rare class of maternal RNAs is localized. cDNA clones of four lo- calized RNAs have been characterized. Three of these cDNAs are derived from maternal RNAs that are concentrated in the animal hemisphere of unfertilized eggs and remain localized through the early blastula stage. One cDNA is derived from a maternal RNA found almost exclusively in the vegetal hemisphere at both stages. These studies show that some informa- tional molecules, specifically RNAs, are localized in eggs and are inherited by particular blastomeres. Introduction Early in embryogenesis, cells acquire specific develop- mental fates with the consequence that their descendants will express characteristic sets of genes. One possible ex- planation for this phenomenon, proposed decades ago (Wilson, 1925) is that maternal factors are differentially distributed among cleavage cells and that these factors somehow determine the fate (affect gene expression) of the recipient cells. A variety of studies have provided in- direct support for the existence of such maternal factors, sometimes called cytoplasmic determinants. For example, in frog embryos, the formation of ciliated epithelia (Chung and Malacinski, 1983; Kageura and Yamana, 1984), em- bryonic muscle (Gurdon et al., 1985), and germ cells (Smith, 1966; Buehr and Blackler, 1970) all depend to some extent on the inheritance of materials concentrated in different regions of the uncleaved egg. In addition, several studies suggest that localized maternal compo- nents may be involved in establishing the dorsal-ventral axis of Xenopus embryos (Nieuwkwoop, 1977; Gerhart et al., 1983). The localization of maternal factors, either in a discrete or in a graded fashion, provides the most attrac- tive explanation for how cell fate is determined in other animals as well. Notable examples include the formation of muscle in ascidians (Whittaker, 1982) comb plates and photocytes in ctenophores (Freeman, 1976), and body axis (Kalthoff, 1983; Nusslein-Volhard, 1979) and germ cells (Illmensee and Mahowald, 1974; Okada et al., 1974) in insects. The localized maternal factors implicated in determin- ing cell fate in these various embryos have not been iden- tified, but it has often been suggested that maternal pro- teins and/or mRNAs could serve such a function. One way to test this suggestion is to determine whether localized maternal mRNAs exist in eggs. If so, one can isolate cDNA clones for these RNAs and use these clones to determine the function of the RNAs and investigate the mechanism for RNA localization. With this in mind we have under- taken acomprehensive search for cDNA clones coding for maternal RNAs that are localized to specific regions in Xenopus eggs. Our results show that most RNAs in an egg are uniformly distributed throughout the cytoplasm. However, we do find a special class of localized maternal RNAs. Four different localized RNAs and their corre- sponding cDNA clones have been examined in some de- tail. Our studies show that these RNAs are localized be- fore fertilization and remain unevenly distributed during cleavage. Previous studies indicated that it might be possible to obtain cDNA clones for localized RNAs. In particular, RNA:cDNA hybridization experiments suggested that there may be regional differences in the complexity of egg RNAs in sea urchins (Rodgers and Gross, 1978; Ernst et al., 1980) and in frogs (Carpenter and Klein, 1982). The isolation of specific cDNA clones as reported here pro- vides a direct demonstration of the existence of localized RNAs and makes it possible to ask questions about their developmental significance. Results Screening Strategy The unfertilized Xenopus egg is radially symmetric about the animal-vegetal axis with respect to morphological and biochemical properties. For example, it has been shown that the prospective dorsal mesoderm can be formed any- where around the equator of the egg, the actual location being dependent on the site of sperm entry (Roux, 1887; Gerhart et al., 1981). In contrast, there are distinct differ- ences in morphological, biochemical, and developmental properties along the animal-vegetal axis. Fate mapping has shown that the animal and vegetal regions are des- tined to form, respectively, ectodermal and endodermal tissues in the embryo. In addition, subequatorial regions are critically involved in inducing the dorsal-ventral axis (reviewed by Gerhart et al., 1983). These and other studies suggest that it is along the animal-vegetal axis that one might expect to find differences in the distribution of maternal factors. To test for the presence of and to determine the amount of maternal RNAs localized in eggs, we first constructed a cDNA library sufficiently large to contain representatives of most transcripts present in unfertilized eggs. Total poly- adenylated RNA from defolliculated oocytes was used to