salvador is one of many genes that Hariharan isolated in a screen for cell growth mutants. Mitotic recombination in fly eyes rendered possible mutations homozygous. Then Hariharan's group (four postdocs working for two years) looked for mutant patches that grew larger than the corresponding wild-type patch. “One of the lessons is that there are many pathways that we know nothing about,” says Hariharan. “That's why we did a phenotype-based screen, because that assumes nothing.”
salvador appears to be part of a new pathway, but it can be tied to certain known cellular events. In late larval stages it induces cell cycle exit by down-regulating cyclin E, with cells lacking salvador undergoing one or more extra divisions. Then, in the pupal stage salvador is needed to down-regulate Diap1 (an apoptosis inhibitor) and thus induce apoptosis in the eye. This apoptosis eliminates extra cells that have not taken on a specific cell fate.
Both actions of salvador reduce cell numbers, but the logic for putting both functions in a single gene remains elusive. Clues may come from studies of the worm, mouse, and human homologues, or from inspection of human cancer cell lines, at least two of which have mutations in salvador. ▪