Activity of the her-l gene is required for normal male development in C. elegans. Loss-of-function her-l mutations cause feminization of XO animals (normally male) to produce fertile hermaphrodites, while gain-of-function her-l mutations cause masculinization of XX animals ( normally hermaphrodite) to produce pseudo-males. Genetic analysis indicated that her-l functions early in the hierarchy of regulatory interactions among the major sex determination genes (1). We have analyzed her-l genetic mosaics to determine which cells must express her-l for wild-type male development. If her-l functions cell- autonomously, then in XO animals her-l (+) cells will follow male fates, and
her-1(-) cells will follow hermaphrodite fates. If her-l or a her-l regulated activity functions non-autonomously, then the her-l genotype of a cell may not determine its sexual phenotype. The phenotypes of XO her-l genetic mosaics are variable, but clearly show that both
her-1(-) and her-l (+) cells can follow either hermaphrodite or male fates. We conclude that her-l is neither necessary nor sufficient to cell-autonomously determine male sexual phenotypes and, therefore, that her-l or a her-l regulated activity must be able to function non-autonomously. The observed patterns of non-autonomous effects suggest that many or all cells express and respond to her-l activity. These findings implicate cell interactions in C. elegans sex determination. Earlier genetic analysis indicated that her-l might directly control
tra-2 activity (1). Molecular characterization of the gene products of her-l and
tra-2 (see abstracts by M. Perry et al. and P. Kuwabara et al) is consistent with a model in which a secreted ligand encoded by her-l interacts with a cell-surface receptor encoded by
tra-2. We will discuss the possible role of such an interaction in coordinating the sex determination decision.