Heredity 78 (1997)354—362 Received/8 December 1995 Genetic basis of sperm and testis length differences and epistatic effect on hybrid inviability and sperm motility between Drosophila simulans and D. secheiia DOMINIQUE JOLY*t, CLAUDE BAZINt, LING-WEN ZENG & RAMA S. SINGH tLaboratofre Populations, Genétique et Evolution, CNRS 91198 Gifsur Yvette Cedex, France, lDepartment of Ecology and Evolution, 1101 E. 57th St., The University of Chicago, Chicago, IL 60637, USA and §Department of Biology, MacMaster University, 1280 Main Street West, Hamilton, Ontario, Canada L8S 4K1 Resultsare reported from a genetic study of hybrid inviability and three 'fertilization traits' (sperm motility and length, and testis size) that affect hybrid sterility between the sibling species Drosophila simulans and D. sechellia. The main findings are as follows. (i) For sperm length there was a dominant effect of the D. simulans genome over that of D. sechellia, and the Y chromosome of D. sechellia in the background of D. simulans reduced the sperm length. (ii) In contrast, testis length, in spite of its generally high correlation with sperm length, showed an additive effect. (iii) We found a strong asymmetric incompatibility between the D. sechellia X chromosome and D. simulans autosomes: D. sechellia X chromosome with D. simulans auto- somes, but not the reverse, showed a significant reduction in testis length as well as in hybrid inviability compared to the parental species. (iv) Between the two autosomes, chromosome 3 had a greater effect on these traits than chromosome 2, and there was additionally an epistatic effect between these chromosomes with respect to their parental vs. recombinant status: recombinant chromosomes 2 and 3, together, had lower viability than any other combination. (v) The testis size in the backcross generation was greater than the parental species, suggesting that some modifier genes are being released from their species-specific genetic control. (vi) The species-specific homogeneity of the genome was important for all three traits—offspring viability, hybrid male fertility and testis length. These results are discussed with respect to the role of sexual selection and genetic divergence during speciation. Keywords:Drosophila,epistatic effect, speciation, species differences, sperm length, testis length. Introduction Zouros et al., 1988; Zeng & Singh, 1993a,b, 1995). Previous studies have generally focused on hybrid Thegenetic mechanisms of species formation is one male sterility and relatively little attention has been of the most challenging problems in evolutionaty paid to hybrid inviability; this approach was intended biology and we have seen the resurgence of a to identify genes affecting hybrid sterility. Species renewed interest in these mechanisms (see Orr, differ in more than reproductive isolation, so a 1991; Coyne, 1992; Pantazidis et a!., 1993). One way broader approach, including all traits affecting to approach this problem is to study the genetic reproduction, is required. This report combines the basis of reproductive isolation between related two approaches. Sperm length is an excellent trait species and this has currently become the favoured for studying species differences as it is species- experimental approach (Dobzhansky, 1936; Zouros, specific and is the most rapidly evolving trait in 1981; Wu & Beckenbach, 1983; Coyne, 1984, 1985, insects (Joly et al., 1989, 1991). Drosophila species 1993; Coyne & Kreitman, 1986; Orr, 1987, 1989; are good candidates on which to perform such an investigation because they encompass the extreme *Correspondence E-mail: joly@sunbge,bge.cnrs-gif.fr range of sperm length variation in all invertebrates, 354 1997The Genetical Society of Great Britain. GENETICS OF SPECIATION 355 ranging from 33 em in D. subobscura (Joly et al., was not successful. All stocks and crosses were 1989) to nearly 6 cm in D. bifurca (Joly et a!., 1995). reared in 8-dram vials on banana medium at This study reports the genetic analysis of hybrid 21—23°C. inviability and three strongly related fertilization traits, sperm length and motility and testis size, in Y chromosome introgression the sibling species D. simulans and D. sechellia from Interspecific the melanogaster complex. These species are closely Toexamine the effect of the Y chromosome on related (Coyne, 1984; Lachaise et a!., 1986). Their sperm length between D. simulans and D. sechellia, chromosomes are entirely homosequential and we introgressed the Y chromosome of D. sechellia crosses (D. simulans female by D. sechellia males) into the background of D. simulans. The method can be easily made giving fertile, hybrid females and used is the same as described in Zeng & Singh sterile, hybrid males (Lemeunier & Ashburner, (1993b). F hybrid females were backcrossed to 1984). Drosophila sechellia sperm length has strongly males of D. sechellia and the resulting backcross diverged from that of the mauritiana-simulans pair, males were crossed to D. simulans females. The being almost one and a half times as long as that of males produced in this cross, which carried the Y D. simulans (1.6 and 1.2 mm, respectively; Joly, chromosome of D. sechellia, were then crossed to D. 1987). simulans females repeatedly to replace other The results show interesting differences in the chromosomes by those of D. simulans. effects of sex chromosomes, autosomes and their interactions on hybrid inviability and sterility, sperm Testisdissection, s/ide preparation and sperm length and motility and testis size. length measurement Brokensperm are indistinguishable from intact Materialand methods sperm, so it is problematic to measure individual sperm longer than 1 mm directly. Instead, the length Species stocks, genetic markers and crossing of bundles of fully elongated mature cysts were scheme measured in the testis. Sperm and mature cysts of AD.simulansstrain homozygous for a recessive one species have similar lengths and the latter could mutant marker on each of the five major chromo- be identified as a fully elongated cyst by the charac- some arms, and a D. sechellia strain (Robertson) teristic aspects of the extremities (one is occupied by were supplied courtesy of Dr Jerry Coyne. The five the sperm head which begins to spread and the mutant markers are forked-2 on the X chromosome other by what will form the waste bag at the sperm individualization stage). This method allows rapid (f2:1-60),net and plum on the second chromosome (nt: 11-2, 2L; pm: 11-103, 2R), and scarlet and ebony measurement of large numbers of samples. Pure- on the third chromosome (st: 111-44, 3L; e: 111-71, species males and backcross males with different 3R). The tiny fourth chromosome was not studied phenotypes were dissected in saline solution because no marker was available. The crossing (Ringer) on microscope slides using a pair of forceps scheme is the same as used by Dobzhansky (1937) and a dissecting needle. Testes were separated from and others (Coyne, 1985; Coyne & Kreitman, 1986; seminal vesicles and the rest of the male reproduc- Orr, 1989). Females of the marker strain were tive tract. Cysts were spread out of the testes, and crossed to D. sechellia males and the F1 hybrid separated from each other by using a very fine females were backcrossed to D. simulans (f2;ntpm; needle. The slides with the cysts or the testis (only St e) males producing offspring with different one testis per male) were allowed to dry at room combinations of D. simulans and D. sechellia temperature, fixed with 95 per cent ethanol for 4 chromosomes. The identity of these chromosomes or mm, dried again, and then stained with 2 per cent chromosomal regions was read from the individual's aceto-orcein solution. They were finally washed with phenotype. By examining a trait of each backcross distilled water, dried and sealed with permanent phenotype, one can determine the effect of each mounting medium. chromosomal combination on that trait. We exam- The cyst and testis length measurements were ined four different traits between species or between aided by using the public domain program NIH species and their hybrids: sperm and testis lengths, IMAGE. This program (written by W. Rasband at the hybrid male sperm motility, and viability of back- US National Institutes of Health) can be obtained cross offsprings. The reciprocal F1 (the cross through anonymous ftp from zippy.nimh.nih.gov and between D. sechellia females and D. simulans males) runs on a Macintosh computer which was connected The Genetical Society of Great Britain, Heredily, 78, 354—362. 356 D. JOLY ETAL. to a dissecting microscope through a video camera Viability effects (Hitachi, model VK-C150). Because previous studies have shown little interindividual variation (Joly, TheX chromosomal effect on hybrid viability can be 1987), 50 cysts for one male from each parental determined by comparing the same autosomal geno- species and F1 hybrids were measured. In the back- types segregating with the D. simulans or D. sechellia cross progeny, 15 cysts per male and 25 males per X chromosome. In this experiment the Y chromo- phenotype were analysed. In the following sections, some comes from D. simulans. In Table 1 we can see the measures are called sperm length but are made that the X chromosomes, regardless of their origin on the cyst. The mean testis length was based on the (from D. sechellia or D. simulans), have no effect mean of 25 independent testes for each phenotype. when autosomes were heterozygous (1/2 D. sechellia and 1/2 D. simulans, 617 and 608, respectively). On the contrary, when autosomes were homozygous Results from D. simulans (iv' pm; st e) the X chromosome of D. sechellia had a significant reduction in the F1male hybrids between D. simulans females and D. number of progeny (440) compared to the X sechellia are perfectly viable but sterile. To inves- chromosome of D. simulans progeny (626). tigate the autosomal or sex chromosomal contribu- The viability of males from the 16 backcross tion to male viability and sterility we have analysed phenotypes having the X chromosome from D.