An alarming number of synthetic Daf-c mutations James H. Thomas, Kouichi Iwasaki, Elizabeth Malone, and Michael Ailion, Dept. of Genetics, Univ. of Washington, Seattle, WA A small number of genes can mutate to produce a moderately strong dauer constitutive (Daf-c) phenotype. When combined, mutations in certain groups of these genes produce a strongly enhanced Daf-c phenotype, suggesting these groups act in parallel to regulate dauer formation. Over the past several years, there have been scattered reports of a Daf-c phenotype in double mutants comprised of single mutations with little or no effect on dauer formation. These genes include
aex-3,
egl-4,
flr-4 unc-3,
unc-31, and several of the cilium-structure genes including
che-3, osm-l, and
osm-6. We have recently added dec-S (defecation cycle abnormal) and
unc-64 to this list Not all combinations of double mutants among these genes have been analyzed but it appears that the synthetic pattems will be complex. While this list of genes is not yet alarmingly long, two observations suggest that it probably will become so. First, nearly all of these synthetic interactions seem to have been found by chance in the process of mapping or other routine genetic manipulations. For example, we found that
unc-64;
dec-5 is Daf-c in the process of using
unc-64 as a marker in a
dec-5 construction. Such unsystematic analysis surely has missed many interactions among known genes. Second, in a 5,000 genome screen by Katsura, Suzuki, and Ishihara (WBG 13-2), more than forty mutations that cause a synthetic Daf-c phenotype with
unc-31 were found. The naive calculation would suggest about sixteen genes can readily mutate to this phenotype (with
unc-31 alone!). If anyone out there has been surprised to notice some dauers unexpectedly forming in an unstarved stock we'd like to know about them. Why so much redundancy and so much complexity? We offer some observations that suggest why and where much of this complexity occurs. Dauer formation is a complex process that requires the inteagration of many developmental and sensory cues. Among the sensory cues are dauer pheromone, temperature, and food (itself almost certain to be complex). Less is known about developmental cues, but at least heterochronic genes are involved in determining the stage of dauer-formation competence. Integration of complex cues requires convergence of parallel pathways of information, and this implies functional redundancy. Such neuronal integration in larger organisms occurs when information from diverse sensory neurons converges on interneurons (often gradually, over more than one level of cells), whose task is to extract relevant features of the sensory input and orchestrate an appropriate response. Given our poor state of knowledge of the functions of interneurons in C elegans, how can we begin to address whether any of these synthetic Daf-c processes occur in interneurons? We have two simple thoughts at the moment. The first is a consideration of the pleiotropies caused by single mutations in each gene. At least some aspects of the Unc, Aex, Dec, and Flr phenotypes conferred by mutations in the various genes seem unlikely to be caused by sensory neuron defects (in contrast for example, all aspects of the phenotypes of the cilium-structure mutations are easily explicable as sensoly defects). In addition, the synthetic Daf-c phenotype seems unlikely to be caused by motor neuron defects. Thus, for several of these genes it is plausible to suppose that they affect interneurons (though of course they may affect other cells as well). Our second thought is to test epistasis with
daf-3 and
daf-5. Daf-d mutations in these two genes block the Daf-c phenotype caused by killing amphid sensory neurons, suggesting that
daf-3 and
daf-5 function downstream of the sensory neurons. Any neuronal gene that functions downstream of
daf-3 and
daf-5 is thus a candidate for affecting dauer-controlling interneurons. We have just begun to analyze these relations, and have thus far found that in all cases the synthetic Daf-c double mutant is epistatic to
daf-3 and
daf-5 but not
daf-12, consistent with a dauer-related function of both genes in interneurons. We would like to think that these synthetic Daf-c interactions provide a long-awaited avenue into interneuronal function.