Does hybrid lethality depend on sex or genotype?

Arecent spate of articles in this and other journals fitness depends solely on genotype: if hybrid males and females have the same genotype, they will both be viable has highlighted a pattern that appears to characterize the genetics of speciation: the genes causing hyor both be inviable. In the second case, a hybrid lethal is expressed in one sex only (say, males). In this case, brid inviability typically kill both sexes, while the genes causing hybrid sterility typically sterilize one sex only hybrid fitness depends solely on somatic sex: if a hybrid of given genotype expresses male genes, it will die; if a (Orr 1993; Wu and Davis 1993; Turelli and Orr 1995; Hollocher and Wu 1996; True et al. 1996). An hybrid of identical genotype expresses female genes, it analogous pattern clearly holds within species: mutagenwill live. esis experiments in Drosophila conclusively show that Here I use genetic tools from Drosophila melanogaster lethal mutations usually kill both sexes, whereas sterile to distinguish these possibilities. In particular, I use a mutations usually sterilize one sex only, although excepmutation at the transformer (tra) locus, part of the Drotions do exist (reviewed in Ashburner 1989). sophila sex determination pathway, to transform the This apparent parallelism between and within species somatic sex of D. melanogaster-D. simulans species hybrids. is important for several reasons. First, it strongly suggests Tra, which acts immediately downstream of the master that speciation typically involves “ordinary” genes having switch gene, Sex-lethal, is the first gene in the sex determinormal within-species functions. (Such parallelism nation cascade that affects sex determination alone, not would be remarkable if speciation routinely involved dosage compensation (Cline 1985). Tra is active in geunusual genetic processes, not normal genes.) Second, netically normal females (X:A ratio 5 2:2) and inactive the fact that hybrid steriles, like those within species, in genetically normal males (X:A 5 1:2), where activity affect one sex only suggests that substitutions affecting, reflects alternative splicing of tra transcripts (for details, say, hybrid males might accumulate faster during evolusee McKeown et al. 1988). The hs-tra mutation is a tion than those affecting hybrid females. Indeed it apP-element construct bearing female-specific tra cDNA pears that such “faster male” evolution, possibly driven fused to an hsp70 promoter (McKeown et al. 1988) that by sexual selection, plays an important role in speciation causes essentially complete somatic transformation of (Wu and Davis 1993; True et al. 1996; Presgraves and genetic males into females (a male-specific muscle reOrr 1998; Turelli 1998). Last, the fact that hybrid mains untransformed). By introducing hs-tra into hylethals affect both sexes implies that Haldane’s rule for brids, we can test the viability of hybrids of identical inviability [the preferential lethality of heterogametic genotype but different somatic sex. Because we must (X/Y) hybrids] cannot be due to faster accumulation transform the sex of species hybrids—which are necesof hybrid male than female lethals. This rule must insarily heterozygous for D. melanogaster mutations—we stead have some other cause, probably that the genes require dominant alleles. Fortunately, hs-tra overrides causing hybrid inviability usually act recessively in hywild-type tra when in single dose (McKeown et al. 1988). brids (Turelli and Orr 1995 and below). (No dominant mutation allowing the reverse transforGiven the importance of this apparent parallelism, mation of genetic females into phenotypic males is availwe require more and better information about the sex able.) specificity of hybrid lethals and steriles. (Surely no more The cross of D. melanogaster females 3 D. simulans information about within-species expression is needed.) males normally produces viable female but inviable To this end, I have performed a simple but novel test male hybrids (Sturtevant 1929), a result that we conof the claim that hybrid lethals typically affect both sexes. firmed (Table 1, line 1). Genetic analysis has shown that To see the logic of the test, consider the simplest case this lethality, which occurs at the larval/pupal transition, in which a single genetic incompatibility causes hybrid does not involve the Y or the cytoplasm (Hutter et inviability. One can imagine two scenarios. In the first, al. 1990; Yamamoto 1992). Instead, it appears that a a hybrid lethal is expressed in both sexes. If so, hybrid gene(s) on the D. melanogaster X is incompatible with autosomal factors from D. simulans (Sturtevant 1929; Hutter et al. 1990). Two explanations of the sex-limited effects of these Author E-mail: [email protected]