We are examining the requirements for maternally provided versus embryonically transcribed gene products in early developmental decisions. Past studies have shown that C. elegans embryos initiate transcription prior to gastrulation (1, 2). Indeed, the embryonic pes-10 transcript is detectably expressed by the 4-cell stage (3). Devitellinized embryos cultured in the presence of a-amanitin, which inhibits RNA polymerase II, cease dividing at approximately 100 cells. The lineage-specific timing of early cleavages in these embryos is not dramatically altered (2). a-amanitin experiments cannot address the potential role of embryonic transcription in directing gastrulation, since perturbations of the vitelline membrane disrupt gastrulation (4). To block the translation of maternally provided RNA polymerase II large subunit in otherwise intact embryos, we injected capped antisense ama-1 RNA (50 ng / ml) into the gonads of wild-type hermaphrodites. (The ama-1 cDNA clones were generously provided by David Bird.) We scored the progeny of these animals 20 hrs. to 48 hrs. postinjection, and 95% of the embryos arrested early in development (n = 850). We examined 40 of these embryos under Nomarski optics, and all arrested at 80 to 100 cells and showed no signs of morphogenesis or differentiation. To further assay embryonic transcription, we injected antisense ama-1 into a strain carrying vet-6::lacZ, which is strongly expressed by the 28-cell stage (5). The resulting arrested embryos did not detectably express the reporter gene, indicating that the antisense ama-1 RNA had effectively blocked the function of RNA polymerase II. Moreover, preliminary results show that the cleavage pattern of ama-1 antisense embryos is normal through the first four cell cycles, and P-granules are segregated appropriately; thus, these maternally-controlled events are not detectably disrupted. To control for possible artifacts caused by RNA injection, we injected hermaphrodites with RNA antisense to her-1, an embryonically transcribed sex-determining gene. Of 2609 progeny examined, 1.1% failed to hatch, and only 0.3% arrested with fewer than 300 cells. Further controls for ama-1 antisense specificity are underway. We have recorded the development of ama-1 antisense embryos past the point of cell division arrest using the 4-D microscope, and all of the embryos (n = 6) failed to gastrulate. The Ea and Ep cells and their daughters divided precociously and did not migrate into the blastocoel, reminiscent of the maternal-effect emb-16 mutant phenotype (6). These data indicate that embryonically transcribed gene products are necessary for initiating gastrulation. Further analysis of ama-1 antisense embryos will be instructive as to the role of embryonically transcribed genes in early development. 1. Schauer, I. and W. B.Wood (1990) Development 113: 1303-1317 2. Edgar, L. G., Wolf, N. and W. B. Wood (1994) Development 120: 443 - 451 3. Seydoux, G. and A. Fire (1994) Development 120: 2823 - 2834 4. Schierenberg, E. and B. Junkersdorf (1992) Roux's Arch. Dev. Biol. 202: 10-16 5. Pettitt, J. and W. B. Wood. (1993) Worm Breeder's Gazette. 13 53 6. Denich, K.T.R., Schierenberg, E., Isnenghi, E. and R. Cassada. (1984) Roux's Arch. Dev. Biol. 193: 164 - 179