A major model system for the study of evolutionary divergence between closely related species has been the unisexual lethality resulting from reciprocal crosses of Drosophila melanogaster and D. simulans. Sex-lethal (Sxl), a critical gene for sex determination, is misregulated in these hybrids. In hybrid males from D. melanogaster mothers, there is an abnormal expression of Sxl and a failure of localization of the male-specific lethal (MSL) complex to the X chromosome, which causes changes in gene expression. Introduction of a Sxl mutation into this hybrid genotype will allow expression of the MSL complex but there is no sequestration to the X chromosome. Lethal hybrid rescue (Lhr), which allows hybrid males from this cross to survive, corrects the SXL and MSL defects. The reciprocal cross of D. simulans mothers by D. melanogaster males exhibits underexpression of Sxl in embryos.
MANY closely related species exhibit sterility or lethality of hybrids, indicating divergence of gene interactions during speciation. Despite the description of many cases of such postzygotic isolation mechanisms, little is known about their molecular basis. Crosses between the sibling species Drosophila melanogaster and D. simulans, which diverged ∼0.8 million years ago, result in unisexual lethality and provide a model system to investigate these evolutionary processes (Eisses et al. 1979; Lemeunier et al. 1984; Hutter et al. 1990; Hey and Kliman 1993; Sawamura et al. 1993a,b; Jeffs et al. 1994). The sex that is inviable in the hybrid progeny is determined by the direction of the cross: males usually die at the larval/pupal transition when D. melanogaster is the maternal parent, while females are embryonic lethal from the reciprocal cross (Sturtevant 1920; Sawamura et al. 1993c). Several analyses have shown that lethality of hybrid males is associated neither with cytoplasmic factors nor with the Y chromosome (Hutter and Ashburner 1987), but rather with a gene or genes on the D. melanogaster X chromosome (Sturtevant 1920; Yamamoto 1992). Exceptional flies resulting from nondisjunction of the sex chromosomes have a switched sex-specific viability, indicating an interaction between the X chromosome and the hybrid autosomal background (Sturtevant 1920). The search for mutations from natural populations or laboratory stocks that rescue inviable males identified two X chromosomal loci in D. melanogaster (Hutter and Ashburner 1987; Hutter et al. 1990; Barbash et al. 2000, 2003) and the autosomal Lethal hybrid rescue (Lhr) (Watanabe 1979) in D. simulans. Two other genes (Sawamura et al. 1993b) were found to suppress the embryonic lethality of females generated from the reciprocal cross. Several regions of the genome interact with Lhr to reverse the rescue (Presgraves 2003), of which one region has been identified to encode as a nuclear pore protein (Presgraves et al. 2003).
Only 5–6% of hybrid males carrying a D. melanogaster X survive to the pupal stage (Sturtevant 1920). Their chromosomal morphology is altered significantly (Takamura and Watanabe 1980). The chromosomal changes are similar to the mutational effect of the maleless (mle) gene of the D. melanogaster parent (Belote and Lucchesi 1980). The functional product of mle and four other male-specific lethal (msl) genes (msl1, msl 2, msl3, mof) and two noncoding RNAs (roX1, roX2) form a heteromeric complex (Kelley and Kuroda 2000) that enriches a histone acetylase males absent on the first (MOF) and kinase (JIL-1) in the male X chromosome (Hilfiker et al. 1997; Bhadra et al. 1999, 2000; Jin et al. 1999). Together, they increase the level of H4 Lys16 acetylation and H3 phosphorylation on the male X at the expense of such residues on the autosomes (Bhadra et al. 1999, 2000). The loss-of-function mutations of the msl genes dissociate the complex, and as a result, the histone acetylase and kinase become uniformly distributed to all chromosomes (Bhadra et al. 1999, 2000; Pal Bhadra et al. 2005). The formation of a complete male-specific lethal (MSL) complex is inhibited by the product of the Sex-lethal (Sxl) gene, which is normally expressed in females but not in males (Bopp et al. 1991).
The chromosomal morphology changes noted above might suggest involvement of a misregulation of the Sex-lethal gene. The lethal phases of the reciprocal hybrids parallel those in different types of Sex-lethal mutants as noted previously (Orr 1989). The loss-of-function alleles of Sex-lethal are embryonic lethal in females, but are inconsequential to males (Cline 1978). The ectopic expression of SXL in males via a Sxl[M] mutation results in male lethality at the larval–pupal transition (Cline 1978; Bhadra et al. 2000). The SXL protein triggers the female sex determination pathway and acts to inhibit the translation of the MSL-2 protein (Bell et al. 1988; Bashaw and Baker 1997; Kelley et al. 1997). In males, the MSL-2 protein is expressed and catalyzes the sequestration of the MSL complex to the male X chromosome (Bashaw and Baker 1997; Kelley et al. 1997). In Sxl[M] mutants, the MSL sequestration is disrupted (Bhadra et al. 2000). Given these observations, we examined whether the MSL complex was absent in hybrid males carrying a D. melanogaster X chromosome and in general the expression of Sxl in hybrids.
MATERIALS AND METHODS
Flies were maintained on standard culture media at 25° with the exception of crosses to test rescue, which were performed at 18°. Genetic mutations are described in FlyBase (). To generate hybrids from reciprocal crosses, D. melanogaster females were mated to D. simulans males or vice versa. To generate hybrids carrying the Lhr mutation, D. melanogaster females were crossed to D. simulans Lhr/Lhr males or vice versa. The normal strain of D. melanogaster used was Canton-S. The D. simulans strain used was f or, where specifically noted, y1 v1 f1.
To examine the effect of the Sex-lethal gene in hybrids, y Sxlf#1/Basc females were crossed to D. simulans males, carrying the above-mentioned y marker on the X chromosome. For examination of larvae, the male hybrids carrying y Sxlf#1 on their X chromosomes were distinguished by their brown mouthparts conditioned by the y mutation compared to their Basc brothers with black mouthparts (y+ in Basc).
In other crosses, the roX transgenes were introduced into Sxl stocks (w Sxlf10,M1 sn/Basc; y w cm SxlfP7B0 ct/Basc; w Sxlf#1 oc ptg v/Basc) and used in crosses to males of f D. simulans. Two types of crosses were made. In one case, a stock with homozygous transgenes (y Sxlf#1/Basc; hsp70-roX2/hsp70-roX2; hsp83-roX1/hsp83-roX1) was crossed to D. simulans males. In a second cross, to examine the effect of independent assortment of Sxl and the two roX transgenes on hybrid male survival, balancer chromosomes for chromosomes 2 and 3 were present in the female D. melanogaster parent (y Sxl/Basc; hsp70-roX2/SM6a, Cy; hsp83-roX1/TM3, Ser).
To examine the effects of attached X D. melanogaster hybrid females, C(1)RM, y f females were mated with D. simulans males.
Fixation of the polytene chromosomes from the salivary glands of larvae and the staining of the chromosomes with antibodies for MSL proteins and H4Ac16 was performed as described (Bhadra et al. 1999, 2000). For confocal microscopy, Cy-5-conjugated goat anti-rabbit secondary antibody (Jackson ImmunoResearch, West Grove, PA) was used. After incubation with secondary antibody, chromosomes were washed with phosphate buffer saline (PBS) for 20 min. The chromosomal RNAs were digested with dilute RNase for 30 min at a concentration of 20 μg/ml. The slides were mounted with Vectashield mounting media–propidium iodide mixture and examined with a Bio-Rad (Hercules, CA) 1024 confocal microscope using a ×100 oil lens.
The developing embryos were collected, dechorionated, fixed, and processed for antibody labeling as described (Bhadra et al. 1999). Embryos were fixed in 4% formaldehyde, 0.1% Tween-20, and 0.1% sodium deoxycholate in PBS for 30 min, rinsed in PBS containing 0.1% Tween-20 and 0.1% Triton X-100, and digested with 20 μg/ml Proteinase K. Embryos were incubated with mouse anti-SXL antibodies (University of Iowa Hybridoma Bank) (1:50) in the same blocking solution. For fluorescence color detection, we used goat anti-mouse Cy-5 conjugates for SXL antibodies at a 1:200 dilution. Embryos were mounted in propidium iodide with Vectashield antifade mounting media. Photographs were taken with a BioRad 1024 confocal microscope and processed using Adobe Photoshop 8.0 software.
Gene expression analysis in embryos:
Gene expression in embryos at approximately stage 13 was examined as previously described (Pal Bhadra et al. 2005). The means were calculated from >20 embryos/genotype/gene. Male and female genotypes were distinguished by the different intensity of SXL expression. In the hybrid genotypes, the expression of SXL is abnormal, but still distinctive between the sexes. The respective patterns for the D. simulans females by the D. melanogaster cross were confirmed by using a lacZ-marked Y chromosome stock (Lu et al. 1996) and crossing it with the same D. simulans stock, followed by staining for SXL and lacZ.
MSL binding on hybrid chromosomes:
To test for binding of MSL proteins and the distribution of H4Ac16 residues, we immunostained chromosomes of parental and hybrid males generated from reciprocal crosses. The parental species and hybrids were examined separately as well as in mixtures on the same slide to allow a direct comparison. The binding of MSL-1, MSL-2, and MOF (Hilfiker et al. 1997) proteins were examined on the D. simulans chromosomes. All are preferentially sequestered to the male X chromosome, as in D. melanogaster, indicating that sequestration of the MSL complex is also present in D. simulans as expected (Figure 1A). Double staining with MSL-1 and MOF antibodies in hybrid simulans X or melanogaster X males (from melanogaster mothers by simulans fathers) showed that both proteins are strongly associated with the X chromosome in the simulans X hybrid larvae, whereas no trace of the MSL-1 protein was bound in the melanogaster X hybrids. Only MOF is associated with all chromosomes at a low level in the melanogaster X males (Figure 1B), as is also the case in hybrid females (and normal females of both species) (Bhadra et al. 1999, 2000; data not shown for D. simulans females). We also examined the enrichment of histone 4 Lys16Ac residues on the X chromosomes in the same genotypes. The H4 Lys16Ac distribution overlaps with MSL localization on the X chromosome in both species. In male hybrids with a simulans X, acetylation is enriched on the X, while a uniform genomewide distribution occurs in nuclei from hybrid males with a melanogaster X (not shown, but see Figure 1D). These results indicate that sequestration of the MSL complex and enrichment of H4 Lys16Ac residues is disrupted in hybrid males carrying a melanogaster X.
Effect of the Lethal hybrid rescue mutation in hybrids:
In D. simulans, a dominant autosomal mutation, Lhr (Watanabe 1979), conditions the recovery of 50–60% of otherwise lethal male hybrids to the adult stage. The Lhr mutation has no effect on chromosomal morphology in the D. simulans parents, but the chromosomes in hybrid males with a melanogaster X and Lhr are changed in appearance compared to crosses with D. simulans males without Lhr (Takamura and Watanabe 1980).
The binding of MSL proteins and accumulation of H4 Lys16Ac in +/+ or Lhr/+ hybrid males and females was compared. The chromosomes of D. simulans individuals that are heterozygous for the Lhr mutation were immunostained with anti-MSL-2 and anti-H4Ac16 antibodies. The antibodies bound all chromosomes at a low level in females but associated with only the X chromosome in males, exhibiting a conserved MSL sequestration pattern (Figure 1C). This finding shows that the Lhr mutation does not affect MSL sequestration in D. simulans. The chromosomes of Lhr/+ and +/+ hybrid males with a melanogaster X chromosome were examined as a mixture in the same microscopic field using antibody probes for MSL-1, MSL-2, and H4Ac16. H4Ac16 is uniformly distributed on all chromosomes of the male escapers without Lhr, but neither MSL is detectable. In contrast, all three antibodies are bound to many sites on the X chromosome of the rescued males carrying a copy of the Lhr mutation (Figure 1D; MSL1 not shown). Thus, the Lhr mutation appears to rescue larval lethality and restores sequestration of the MSL complex, including the histone acetylase, to the male X chromosome in hybrids.
Crosses of attached X D. melanogaster by D. simulans males result in female lethality and male viability. The Lhr mutation will rescue the female lethality. When the simulans X chromosomes were examined in the male progeny of crosses to attached X D. melanogaster females, the association of the MSL complex with the X chromosome appeared normal (Figure 2). Interestingly, when the same cross was made using the Lhr D. simulans strain, the male X chromosome exhibited a bloated and diffuse appearance (Figure 2) in contrast to the same direction of cross without an attached X (Figure 1D). This bloated type of chromosome morphology has also been observed in numerous mutant male larvae, for example, with ISWI (Corona et al. 2002), HP1, and Su(var)3-7 (Delattre et al. 2004; Spierer et al. 2005) among others (see discussion in Pal Bhadra et al. 2005). Whether there is a relationship to these defects is not known.
Gene expression in hybrids:
To examine whether the failure of MSL sequestration has any significant impact on gene expression, we estimated expression for a sample of three selected X-linked and three autosomal genes in embryos of D. melanogaster, D. simulans, and the two reciprocal crosses (Figure 3). Some genes showed different levels of expression between the two species. Gene expression in hybrid males carrying a simulans X chromosome fell between the parental values or was elevated. Hybrid gene expression in which a melanogaster X was present showed no loss of X chromosome dosage compensation, but two of the three autosomal genes were elevated. The autosomal gene, β-tubulin, was not changed in expression, which has previously been found not to change in mutants with disrupted MSL localization (Hiebert and Birchler 1994; Bhadra et al. 1999, 2000) and which can therefore serve as a control. Thus, the behavior of this sample matches the pattern of expression as found in the male-specific mutations mle and SxlM in D. melanogaster, where an increased autosomal acetylation and autosomal gene expression is correlated (Bhadra et al. 1999, 2000; Pal Bhadra et al. 2005). Note that this pattern of expression is distinct from that observed in genotypes that eliminate both the MSL complex and the histone acetylase, MOF, which lose X chromosome dosage compensation and have no autosomal increases (Pal Bhadra et al. 2005). Thus, the overall pattern of gene expression in melanogaster X hybrid males is consistent with previous analyses of gene expression when the MSL sequestration is disrupted (Hiebert and Birchler 1994; Bhadra et al. 1999, 2000; Birchler et al. 2001, 2003; Pal Bhadra et al. 2005).
In the female hybrid embryos from the cross in which D. simulans was the mother, the X chromosomal genes exhibited normal expression, but the autosomal loci showed reduced levels. This pattern of gene expression is typical of situations in which the MSL complex is ectopically expressed in females, causing the abnormal sequestration of the histone modifiers to the X chromosomes with concomitant depletion of histone acetylation in the autosomes (Bhadra et al. 1999, 2000; Pal Bhadra et al. 2005). A reduced expression of Sxl is the likely explanation for this situation (see below).
Misexpression of Sex-lethal in hybrids:
Because the above data suggested an altered expression of Sxl, we probed embryos of each species and reciprocal crosses using SXL antibodies. Similar to D. melanogaster, Sxl was strongly expressed in D. simulans females but not in males (Figure 4A). The SXL staining of embryos from the cross between D. melanogaster females and D. simulans males resulted in two classes of embryos. One class showed a strong expression of SXL, which is typical of females, starting from blastoderm and persisting into later stages of development (Figure 4A). In the second class, Sxl was expressed at a lesser but considerable level at the same stages especially as development proceeded, suggesting the abnormal presence of the SXL in males. When D. melanogaster females were crossed by Lhr D. simulans males, the progeny exhibited the normal SXL expression (Figure 4A).
In the reciprocal cross of D. simulans females by D. melanogaster males, male embryos continue development to the adult stage and lack SXL staining as expected. The female embryos, however, exhibited SXL labeling at a low level, which remained low at later stages (Figure 4A). As noted above, the cross of C(1)RM D. melanogaster attached-X mothers by D. simulans males results in hybrid female lethality. These female embryos also have reduced SXL expression (Figure 4A). To confirm that this expression level was below normal, a mixture was made with wild-type D. simulans and probed for SXL. The normal female embryos in the mixture had much greater labeling (Figure 4B).
Effect of a loss-of-function Sex-lethal mutation on male hybrid larval viability:
We next examined the effect of a loss-of-function Sex-lethal mutation on viability in D. melanogaster/D. simulans hybrids. Previous data showed that Sxl mutations had no influence on adult hybrid viability (Orr 1989). Therefore, we examined the effect on hybrid larvae. If the hybrid lethality results from inappropriate SXL expression in males, then the loss-of-function Sxlf#1 allele (Maine et al. 1985) on the melanogaster X chromosome might improve viability, because the SXL protein could no longer be ectopically expressed under this circumstance. Females carrying the Sxlf#1 null allele heterozygous with the X chromosome balancer, Basc, were crossed by D. simulans males. The yellow mutation is linked to Sxlf#1 to distinguish the two classes of larvae on the basis of the color of the mouthparts. Hybrid males carrying Sxlf#1 survive to the third instar larval stage at a seven- to eightfold greater rate compared to hybrids with Sxl+ (Table 1). To compare Sxlf#1 against a chromosome other than Basc, females with Sxlf#1 heterozygous with a normal X chromosome were crossed to D. simulans males. While recombination will separate the y and Sxlf#1 mutations in this case (typically ∼19 cM separate the two loci), the yellow larvae were again preferentially recovered, suggesting an impact by the Sxl chromosome on viability.
Misregulation of Sex-lethal in male lethal hybrids:
We also examined the impact of the Sxlf#1 mutation on MSL sequestration in hybrid larvae. First, we estimated the expression of the Sxl gene, using probes that detect male- and female-specific SXL RNAs on Northern gels. SXL is a splicing factor whose expression becomes established in females to generate a correctly spliced RNA that produces functional protein. In males, a portion of the primary transcript remains in the RNA, blocking the production of protein (Bell et al. 1988). The expression of all three male transcripts was maintained in viable simulans X hybrids at levels nearly equal to those in D. melanogaster or D. simulans males. However, in hybrid melanogaster X males, the female-specific transcripts were also present (Figure 5A). The existence of both male and female transcripts in these hybrids indicates that the Sxl gene on the melanogaster X chromosome is misexpressed in the hybrid background at the larval stage as well as the embryonic stage described above. The Lhr mutation corrects the misexpression in that it restores the situation in which only male transcripts are present (Figure 5A).
To examine the relationship of SXL presence in hybrid nuclei and MSL sequestration, we double stained a mixture of the polytene chromosomes of the two types of male hybrids generated from reciprocal crosses using anti-SXL and anti-MSL1 antibodies. In hybrid males from D. simulans mothers, anti-MSL1 was associated with only the X chromosome, as described above, coincident with no anti-SXL binding (Figure 5B). However, the SXL antibody was present in the nuclei of the reciprocal hybrid male melanogaster X larvae. In parallel, no binding of MSL1 proteins was detected (Figure 5B). The binding pattern in the two types of nuclei indicates that hybrid males that carry the melanogaster X produce functional SXL protein to a level that is sufficient to prevent formation of a complete MSL complex and sequestration of histone modifiers to the X chromosome in males.
We also compared the binding of the same proteins in hybrid male nuclei carrying a melanogaster X with a wild-type or mutant Sxlf#1 allele. This mutation would eliminate ectopic expression of SXL in the hybrid males and indeed, as noted above, improved larval viability. Anti-SXL is associated with chromosomes of hybrid male nuclei that carry a melanogaster X (Figure 5C). Such binding acts as a marker to distinguish them from the other set of nuclei in the mixture carrying the Sxlf#1 mutation, because Sxlf#1 does not produce a functional protein (Maine et al. 1985). In these nuclei, MSL staining occurs at a high level on all chromosomes of the male nuclei in the presence of the Sxlf#1 mutation (Figure 5C) with the exception of MSL2, which is weakly present on all chromosomes (Figure 5D). The presence of the complex throughout the genome but without sequestration to the X chromosome may explain the partial but not complete restoration of viability by Sxlf#1. This result also suggests the requirement of an additional component for X localization in this situation or insufficient amounts of MSL2 for proper targeting to the X chromosome.
The distribution of the MSL complex throughout the nucleus in the Sxlf#1 male larvae is reminiscent of, but not identical to, the more genomewide MSL distribution when the RNA components (Amrein and Axel 1997; Meller et al. 1997) of the complex, encoded by the roX genes, are absent (Meller and Rattner 2002). In the latter case, there is a greater concentration at heterochromatic regions. We considered the possibility that although the protein components of the complex are present in the Sxlf#1 hybrid genotype, the roX genes were not sufficiently expressed to allow localization to the X chromosome. To examine this issue, autosomally inserted transgenes of roX1 and roX2 under the control of different constitutive heat-shock promoters (hsp70-roX2; hsp83-roX1) (Meller et al. 2000) were combined with the Sxlf#1, Sxlf10,M1, and SxlfP7B0 null mutations as described in materials and methods. When these D. melanogaster genotypes were crossed as females by D. simulans males, the hybrid male genotypes that inherit the loss-of-function Sxl mutations and the roX transgenes still do not condition adult rescue (data not shown).
The misregulation of Sxl in melanogaster–simulans hybrids provides a model system in which to examine the molecular consequences of postzygotic speciation mechanisms. The misregulation of Sxl causes a dissociation of the MSL complex in lethal melanogaster X males with a corresponding change in gene expression typical of MSL mislocalization. The changes in gene expression in lethal females from crosses of D. simulans females by D. melanogaster males are consistent with abnormal expression of MSL2, which would cause sequestration of the MSL complex to the X chromosomes. The introduction of a Sxl null into lethal hybrid males caused a restoration of MSL2 expression, but not X chromosome localization. The reason for failure to localize to the X chromosome is unknown. Although misregulation of Sxl might contribute to the hybrid lethality, other factors also appear likely to be involved.
Other potential candidates might include nucleoporins. As noted above, mutations in Nup98-Nup96 can reverse the hybrid rescue of Lhr (Presgraves et al. 2003). Interestingly, other nucleoporin genes, nup153 and Mtor, are required for the proper X localization of the MSL complex (Mendjan et al. 2006) in SL-2 cells. Knockdowns of additional nucleoporins, including Nup98, did not affect MSL localization, although, from SF4 cells, Nup98 was found in protein complexes retrieved using tandem affinity purification-tagged MOF (Mendjan et al. 2006). These results, together with those reported above, suggest an involvement of nucleoporins and MSL sequestration in the hybrid incompatibility of D. melanogaster and D. simulans, although additional influences are likely, given the misregulation of Sxl.
The nature of the misregulation of Sex-lethal appears to involve an aspect of expression that is operative in both sexes. Interestingly, the misregulation occurs oppositely in the reciprocal crosses, causing a SXL reduction in females and ectopic expression in males. An explanation for the misregulation might come from the observation of nonadditive gene expression in hybrid situations (Hammerle and Ferrus 2003; Michalak and Noor 2003; Gibson et al. 2004; Ranz et al. 2004; Birchler et al. 2005). Given the intricate balance of X and autosomal factors for the proper initiation of SXL transcription and proper splicing (Erickson and Cline 1993), such nonadditive expression could easily upset this balance, which would cause altered expression of Sxl.
Sawamura et al. (1993a,b,c) have suggested that a different basis exists for the lethality of the reciprocal hybrid genotypes. Our results are not inconsistent with their idea, given that Sxl misregulation might not be the ultimate basis of the lethality. As for the misregulation of Sxl being different in the reciprocal crosses, this observation might reflect only a commonality in the phenomenon of misregulation of gene expression in hybrid genotypes.
Within D. melanogaster or D. simulans, the regulation of Sxl is obviously appropriate. It is apparent, however, that divergence has occurred such that the regulatory components from different species are incompatible in hybrids for proper Sxl expression. This incompatibility might arise from different regulatory components from one species contributed from the X chromosome vs. those from the hybrid autosomes. Alternatively, maternally derived components all from one species, might interact incorrectly with a zygotic hybrid genotype. In the case of female lethality, this is the most likely explanation, given that the same female genotype from reciprocal crosses has unidirectional lethal effects.
We thank J. Roote for Lhr and D. simulans stocks, M. Kuroda and J. C. Lucchesi for antibodies, T. Cline for Sxl alleles, and V. Meller for the roX transgenes. This work was supported by a National Science Foundation grant (MCB 0211376) and by a National Institutes of Health grant (R01 GM068042). U.B. and M.P.B. were supported by a Welcome Trust Senior International Research Fellowship (GRA070065MA and GRA076395AIA) and a Department of Science and Technology grant of India (SP/SO/D-22/2001).
↵1 These authors contributed equally to this work.
Communicating editor: G. Gibson
- Received May 9, 2006.
- Accepted August 7, 2006.
- Copyright © 2006 by the Genetics Society of America