During embryonic cleavages, the maternal factor PIE-1 is segregated to the germ lineage where it is required to repress RNA polymerase II-dependent gene expression and maintain germline fate (1). The mechanisms that localize PIE-1 to the germ lineage are not known, but may be related to those that localize P granules. Like P granules, PIE-1 is uniformly distributed in oocytes and newly fertilized eggs and becomes localized asymmetrically prior to each unequal germline cleavage. Furthermore, mutants that mislocalize P granules also mislocalize PIE-1 (2). To begin to identify the mechanisms used to localize PIE-1, we have started to define
pie-1 sequences required for the localization of a PIE-1:GFP fusion. We used an 8kb genomic clone capable of rescuing a
pie-1 mutation (see Abstract by Dunn et al.) to generate an in frame PIE-1:GFP fusion. Using the Fire lab's complex array injection method, we detected GFP fluorescence in oocytes and early embryos in a pattern remarkably similar to that described for endogenous PIE-1. For example, in dividing 1-cell embryos, PIE-1:GFP could be detected throughout the posterior cytoplasm, on the posterior centrosome, and on posterior granules (presumably P granules), as is seen for endogenous PIE-1 by immunofluorescence. To determine whether the
pie-1 3 UTR is required for this localization, we replaced 400 bases downstream of the PIE-1 stop codon with the 3 UTR from
let-858, a ubiquitously expressed gene (3). We found that PIE-1:GFP localization was unaffected by this change, indicating that the
pie-1 3 UTR is not essential. We also replaced all
pie-1 coding sequences with
his-11 coding sequences (histone H2B). This fusion was expressed in all embryonic blastomeres, suggesting that PIE-1 coding sequences are required for localization. Consistent with this interpretation, a deletion of 133 amino-acids from the carboxy-terminal end of PIE-1 caused PIE-1:GFP to remain uniformly distributed in 1 and 2-cell embryos. These results suggest that PIE-1 localization depends primarily on mechanisms acting on the PIE-1 protein rather than the
pie-1 RNA. This is in contrast to what was found for GLP-1 and PAL-1, two maternal factors that rely on their 3 UTR for localization (4). We are continuing to delete
pie-1 coding sequences with the goal of defining a minimal PIE-1 localization domain. 1. Mello et al., 1992, Cell 70, 163-176; Mello et al., 1996, Nature 382,
p710-712; Seydoux, et al., 1996, Nature 382, 713-716; Seydoux and Dunn, 1997. Development 124, 2191-2201. 2. Tenenhaus et al., 1998, Developmental Biology, in press. 3. Kelly et al., 1997, Genetics 146: 227-238 4. Evans et al., 1994, Cell 77, 183-194; Hunter and Kenyon, 1996, Cell 87, 217-226.