We are using a genetic mosaic screen for mutations in zygotically acting genes that are essential for early development. This strategy allows us to distinguish gene products whose functions are required in specific lineages from gene products required for general cell viability (housekeeping functions). In contrast to other approaches, mosaic analysis does not rely upon lethal phenotypes, but upon focus of gene activity, and therefore may allow us to identify mutations in genes whose lethal phenotypes we may not anticipate. We hypothesized that this approach is reasonable since we can make the following predictions about the properties that different classes of essential, zygotically active genes may exhibit: Class I. genes whose functions are required in all lineages (housekeeping); Class II. genes whose functions are required everywhere but expression in part of the lineage is sufficient (diffusible housekeeping); Class III. genes whose functions are involved in determination and differentiation and required only in specific lineages; Class IV. genes having essential functions in certain lineages and nonessential functions in others. We have generated 31 independent lethal mutations in the strain ncl- 1
glp-1; d by Judith Austin). The EMS induced lethal mutations are linked to these recessive markers and complemented by the free duplication qDp3. The visible markers
unc-36 and
glp-1 have specific foci of activity in ABp and P4, respectively, allowing a rapid visual screen for rare somatic losses of qDp3 during early cell divisions. Losses of qDp3 are determined precisely using the cell autonomous nucleolar marker
ncl-1 ( E. Hedgecock) visualized under Nomarski optics. The pattern of rare viable mosaics should reflect the requirement for gene activity. Mosaic analysis among a set of the 31 lethal strains that we have generated has distinguished mutations consistent with the predicted classes: some lethal strains never generate mosaics (Class I); a few strains generate viable mosaics in a restricted part of the lineage ( Class III); and two strains generate viable mosaics in a restricted part of the lineage and show an additional specific phenotype (Class IV); (see below). Although we have not yet seen any mosaic patterns consistent with Class II gene products, Bob Herman (personal communication) and Wightman and Ruvkun (WBG (1989) 11: 48) have observed mosaic patterns that we consider to be of this class. Strain GS192 appears to contain a mutation in a Class IV gene that has an essential function in the P1 lineage and a nonessential function in the AB lineage. Rare viable mosaics (picked as Uncs or semi Uncs) in GS192 result from losses in AB, ABp, ABpl, and ABpr . In contrast, we have never isolated a viable mosaic resulting from loss in the P1 lineage, suggesting that the expression of this gene product is absolutely required in these cells. However, a nonessential function was revealed by the abnormal morphology of neurons derived from AB in mosaic animals lacking qDp3 in these AB derived cells. Neurons derived from AB do not exhibit their characteristic granular appearance and do not have visible Nc1 nucleoli. GS161 also appears to encode a Class IV product. Unc and semi-Unc mosaic animals result from losses in either ABp, ABpl, or ABpr. This pattern suggests that expression of the gene product is absolutely required in the P1 lineage as well as in the ABa lineage. A nonessential function was revealed by abnormal vulval formation in mosaic animals. This observation is especially interesting since cells giving rise to the vulva descend from ABp. All mosaics are Egl-; some mosaics have no vulva, whereas other mosaics have an abnormal vulva. Preliminary mapping data place this allele left of
ncl-1 where, to our knowledge, no genes affecting vulval development have been previously identified. The variable vulval phenotype may indicate either that this is not a null allele or, alternatively, may reflect some aspect of this gene product's requirement in the vulval lineages. These data argue that this genetic mosaic screen may indeed allow us to identify lethal mutations in zygotically expressed genes that are required for basic developmental processes. Furthermore, mosaic analysis not only allows us to distinguish among genes products having different foci of activity, but may also allow us to identify mutations in genes which may be too pleiotropic to recognize by criteria such as terminal phenotype. Some gene products of interest to us may have multiple, but specific roles in development (Cline, T.W. (1989) Cell 59: 231-234.) and thus may be classified as pleiotropic by classical criteria but may be identified in this genetic mosaic strategy as interesting. (Note: we have detected recombination between qDp3 and the chromosomes with some growth advantage for recombinants.)