Altered Heterochromatin Binding by a Hybrid Sterility Protein in Drosophila Sibling Species

See allHide authors and affiliations

Science  11 Dec 2009:
Vol. 326, Issue 5959, pp. 1538-1541
DOI: 10.1126/science.1181756


Hybrid sterility of the heterogametic sex is one of the first postzygotic reproductive barriers to evolve during speciation, yet the molecular basis of hybrid sterility is poorly understood. We show that the hybrid male sterility gene Odysseus-site homeobox (OdsH) encodes a protein that localizes to evolutionarily dynamic loci within heterochromatin and leads to their decondensation. In Drosophila mauritiana × Drosophila simulans male hybrids, OdsH from D. mauritiana (OdsHmau) acts as a sterilizing factor by associating with the heterochromatic Y chromosome of D. simulans, whereas D. simulans OdsH (OdsHsim) does not. Characterization of sterile hybrid testes revealed that OdsH abundance and localization in the premeiotic phases of spermatogenesis differ between species. These results reveal that rapid heterochromatin evolution affects the onset of hybrid sterility.

One facet of speciation studies is focused on mechanisms by which populations become reproductively isolated (1). Intrinsic postzygotic isolation results in the sterility or lethality of the F1 hybrid offspring following a successful fertilization event and the formation of the zygote (2). The Dobzhansky-Muller model [reviewed in (2)] proposes that such reproductive barriers occur due to incompatibilities between genetic loci arising as a by-product of divergence between two populations. The identification of loci involved in F1 hybrid sterility in the heterogametic sex (XY males or ZW females) is of particular interest, as this defect is postulated to be the earliest postzygotic isolation event to arise between incipient species (2, 3). Yet, the biological basis of the defects that result in hybrid sterility remains largely unknown.

Crosses between D. simulans and D. mauritiana, which separated ~250,000 years ago (4), produce sterile F1 hybrid males and fertile females. A series of introgressions of the D. mauritiana genome into a D. simulans genomic background revealed that the interaction of the D. mauritiana X chromosome–encoded OdsH (OdsHmau) protein with the male D. simulans genome resulted in male hybrid sterility (58). Considerable amino acid divergence was observed between OdsHsim and OdsHmau, especially within the putative DNA-binding homeodomain (16 nonsynonymous and 3 synonymous changes) (5). Homeodomains are characteristic of a well-conserved family of transcription factors regulating early developmental patterning (9). OdsH evolved ~25 million years ago from a gene duplication of unc-4 (10), which encodes a transcription factor that has somatic function in Drosophila (11).

OdsH expression in the testes (6) and its evolutionary descent from unc-4 (10) led to the proposal that OdsH encodes a transcription factor whose introduction into the hybrid background causes misexpression of meiotic genes and, therefore, hybrid sterility (12). However, this model fails to account for how the protein-DNA interaction interface may drive the changes observed in the OdsH homeodomain between Drosophila species. Ablation of the OdsH gene in D. melanogaster had only modest effects on male fertility (6), contrary to expectations that a deletion of OdsH would affect male fertility due to the misregulation of meiotic genes.

An alternative model suggests that evolutionary labile satellite DNAs—found in pericentric, telomeric, and other heterochromatic regions—may result in the divergence of speciation genes (13, 14). Under this model, satellite DNAs and their expansions are perpetuated by female meiotic drive, but affect fitness through reductions in male fertility, which is evident in plant and animal species (15, 16). The evolution of satellite DNA-binding proteins is predicted to be one way to mitigate cost to male fertility and ensure species survival (13, 14). Therefore, we considered the alternative possibility that hybrid sterility genes like OdsH encode proteins that bind to satellite DNA repeats in pericentric or telomeric regions. Under this model, hybrid sterility could result from an inability to correctly package and condense heterochromatin.

To distinguish between euchromatic versus heterochromatic localization, we expressed OdsHsim fused to a 3xFLAG epitope in a D. simulans embryonic cell culture line (Fig. 1, A and B). We observed a punctate localization pattern of OdsHsim in interphase cells—reminiscent of the D1 satellite-binding protein (17). In D. simulans, D1 predominantly localizes to repetitive satellite sequences on the Y and 4th chromosomes (fig. S1), providing a cytological marker relative to OdsH localization. On this basis, OdsHsim localized adjacent to D1 in D. simulans cells (Fig. 1B). However, the localization of OdsHmau protein (fused to Venus, yellow fluorescence protein) partially overlapped with D1 (Fig. 1, A and C). Coexpression of the OdsHsim and OdsHmau fusion proteins revealed that the two proteins localize to common sites, but that OdsHmau has additional localization (Fig. 1D).

Fig. 1

OdsH proteins differ in their localization to D. simulans heterochromatin. We use D1 staining as a marker for 4th and Y chromosome heterochromatin (fig. S1). (A) FLAG-OdsHsim or Venus-OdsHmau epitope-tagged proteins were expressed in transiently transfected D. simulans cultured cells (B to D) or in transgenic D. simulans larval neuroblasts (E to H) under the control of a heat-shock promoter. (B and C) D1 staining (red) is a cytological landmark for localization of OdsHsim (B) and OdsHmau (C) proteins (both shown in green) to D. simulans heterochromatin. DNA staining by DAPI (4′,6-diamidino-2-phenylindole) is shown in blue in merge. (D) Coexpression of OdsHsim (red) and OdsHmau (green). (E) OdsHsim (red) localizes to the X chromosome and adjacent to D1 staining (green) on the 4th chromosome but not to the Y chromosome. (F) OdsHmau (red) localization has additional localization to the Y chromosome. (G) Coexpression of OdsHsim (green) and OdsHmau (red) on male mitotic chromosomes versus (H) female mitotic chromosomes. Arrows in (E) to (G) highlight Y chromosomes. (I) GFP–Unc-4 (red) staining in interphase larval neuroblast cells of D. simulans is diffuse and does not overlap specifically with D1 (green). Scale bars represent 2 μm.

We mapped the chromosomal localization of tagged OdsHsim and OdsHmau proteins with N-terminal 1xFLAG or Venus, respectively. Expression and immunofluorescent detection of OdsHsim in mitotic larval neuroblast cells confirmed that OdsHsim was associated with repeat-rich regions of the D. simulans genome, namely, the X pericentric region and the 4th chromosome (Fig. 1E). OdsHmau localized similarly on the D. simulans X and 4th chromosomes but showed gross localization to the D. simulans Y chromosome (Fig. 1F). Coexpression of OdsHsim and OdsHmau showed the additional localization of OdsHmau to the Y chromosome in male cells (Fig. 1G), whereas these two proteins localized identically in female cells (Fig. 1H). In Drosophila, both the Y and 4th chromosomes are principally heterochromatic, gene-poor, and repeat-rich (18). On the basis of these results, we conclude that both OdsHsim and OdsHmau encode heterochromatin-binding proteins but that they have different localization specificities, resulting in altered localization to the D. simulans Y chromosome.

Because OdsH shares ancestry with a transcription factor, unc-4, we expressed a green fluorescence protein (GFP)–tagged unc-4 protein to contrast its localization with that of the OdsH protein. As expected for a transcription factor, unc-4 showed diffuse staining in interphase neuroblast cells (Fig. 1I). Hence, the OdsH protein may have gained specificity for localization to heterochromatin since it diverged from unc-4.

We investigated whether divergence of the underlying satellite DNAs was associated with changes in OdsH binding specificity. We identified targets of OdsH binding in sibling species D. mauritiana or D. sechellia (Fig. 2A and fig. S2) by crossing transgenic D. simulans lines expressing fusion proteins of either OdsHsim or OdsHmau to these species. By examining localization in male and female hybrids, we found altered localization of OdsHsim and OdsHmau on the Y and 4th chromosomes of sister species (Fig. 2, B to F, and fig. S3). The D. sechellia and D. simulans Y chromosomes were enriched for OdsHmau binding, whereas OdsHmau localization to its own Y chromosome was restricted (Fig. 2, E and F). In contrast, OdsHsim did not associate with any of the three Y chromosomes (Fig. 2, C and D). Furthermore, OdsHsim and OdsHmau bound to the 4th chromosome from D. simulans, but not from D. sechellia or D. mauritiana (Fig. 2, B to F). Thus, the localization of OdsH proteins differed on homologous chromosomes from different Drosophila species, suggesting that there has been a reorganization or wholesale replacement of OdsH-binding sites within heterochromatin on both the Y and 4th chromosomes within the past 250,000 years (summarized in Fig. 2B). These rapid changes in heterochromatin mirror the rapid evolution of OdsH’s homeodomain during this span (5).

Fig. 2

OdsH-binding sites are evolutionary labile in sibling species. (A) A schematic phylogeny of the recently diverged D. simulans, D. sechellia, and D. mauritiana species. mya, millions of years ago. (B) Summary of localization studies (C to F) highlighting the differences in genomic localization of OdsHsim and OdsHmau to the Y and 4th chromosomes. (C) Localization of FLAG-OdsHsim fusion protein on neuroblast mitotic chromosomes from D. mauritiana/D. simulans and (D) D. sechellia/D. simulans hybrid larvae (fig. S2). (E) Localization of Venus-OdsHmau fusion protein on neuroblast mitotic chromosomes from D. mauritiana/D. simulans and (F) D. sechellia/D. simulans hybrid larvae. In all cases, OdsH staining is red and D1 staining green. Only relevant X, Y, and 4th chromosomes are shown.

Localization of OdsH appears to markedly affect local chromosome condensation. For instance, in D. simulans neuroblast nuclei expressing OdsHmau, the general decondensation and intercalation of the Y and 4th chromosomes obscured our ability to distinguish these chromosomes from each other (Fig. 1, F and G, and fig. S4). This is in contrast to nuclei expressing only OdsHsim, in which the Y chromosome is condensed (Fig. 1E). Decondensation of the X and 4th chromosomes was also associated with the localization of either OdsHsim or OdsHmau (Fig. 2, D to F). Decondensation was seen for both OdsHsim and OdsHmau and in all Drosophila genomes assayed (fig. S4). It appears that the retention of an ancestral unc-4–like transcriptional function, coupled with its localization to heterochromatic regions that are otherwise condensed, has resulted in OdsH-mediated chromosome decondensation in hybrids.

To confirm the localization of OdsH in D. simulans cell culture and mitotic chromosomes, we characterized endogenous OdsH localization in the D. simulans testis with an antibody raised to the C terminus of OdsHsim (fig. S5). Whole-mount immunohistochemistry (IHC) of OdsH in wild-type D. simulans testes showed that OdsH is restricted to developing postmitotic primary spermatocytes in the G2 phase (Fig. 3A and figs. S7 and S8). Within these spermatocytes, immunofluorescence revealed two punctate dots, consistent with OdsHsim localization to only two major loci within the D. simulans genome (Fig. 3G). At this stage of spermatogenesis, homologous chromosomes arrange into individual chromatin domains associated with the nuclear membrane (19). Typically, two large hazy domains represent the associated 2nd and 3rd chromosomes, respectively, while the other two domains represent the associated X-Y and 4th chromosomes (19). The presence of OdsH in the latter chromatin domains suggests that endogenous OdsH binds to loci on the X and 4th chromosomes, consistent with our observations that OdsHsim binds to the X and 4th chromosomes of D. simulans.

Fig. 3

Endogenous OdsH localization to G2 spermatocyte nuclei. (A) Whole-mount immunohistochemistry with an antibody to OdsH on adult male testes from D. simulans wild type, (B) D. simulans fertile introgression, (C) D. simulans sterile introgression, and (D) D. mauritiana. Scale bars represent 50 μm. An 8× magnification of the regions identified by the red dashed boxes is shown in the insets, along with arrows pointing to OdsH protein. (E) OdsH protein abundance was significantly expanded in the sterile line, measured by the length of detectable protein within the testes (P < 0.001). Error bars represent one standard deviation. (F) Western blot from D. simulans wild type, fertile introgression, and sterile introgression testes immunoblotted for OdsH (top) and β-tubulin control (bottom). (G) OdsH (red) detected by immunofluorescence in G2 primary spermatocyte nuclei from the adult testes of D. simulans wild type and (H) D. simulans sterile introgression. DNA staining is in blue. Scale bars represent 5 μm.

The antibody that we raised to OdsHsim also recognizes OdsHmau with high specificity (fig. S5). Therefore, we investigated OdsH localization using whole-mount IHC on the testes of a D. simulans male sterile line, containing an introgression of the OdsH-containing region from the D. mauritiana X chromosome (Fig. 3C and fig. S6) (5). Immunofluorescence of OdsH confirmed the localization of OdsHmau to three loci corresponding to the X, Y, and 4th chromosomes within these primary spermatocytes (Fig. 3H). We also observed an expansion of OdsH protein localization in late-G2 primary spermatocytes of the sterile introgression line as compared to wild-type lines (Fig. 3E), suggesting an increase in protein amounts or stability. Western blot analysis confirmed that OdsH protein abundance was increased in the sterile introgression testes (Fig. 3F). Despite the presence of OdsH protein at this late stage of primary spermatocyte development, we did not observe OdsH protein in spermatocytes that have progressed from G2 to meiotic prophase or a delay in the onset of meiosis (fig. S8).

A fertile introgression line differing from the sterile line in that exons 3 and 4 of OdsH are derived from D. simulans (fig. S6) (4) showed no detectable OdsH protein, both by IHC and by Western analyses (Fig. 3, B and F). In addition, we observed no detectable OdsHmau protein in the testes of D. mauritiana males (Fig. 3D). Our characterization of endogenous OdsH abundance and localization reveals differences between both sibling Drosophila species and the sterile and fertile D. simulans introgression lines. Although it is still unclear how the additional binding capacity of OdsHmau adversely affects premeiotic stages of sperm development to cause male sterility, there is precedent for premeiotic defects manifesting in postmeiotic dysfunction in Drosophila and mice (2022).

Because the studied introgression lines differ only at their OdsH locus, OdsHmau is unambiguously linked to the hybrid male sterility phenotype (5, 6). Our findings reveal the genetic architecture underlying OdsH-mediated hybrid sterility. Both D. mauritiana and fertile introgressed D. simulans males lack detectable OdsH protein, yet are completely fertile, consistent with the fact that OdsH function is not required for male fertility in D. melanogaster (6). Thus, OdsHmau-mediated hybrid sterility involves both the unusual abundance and retention of OdsHmau protein in the D. simulans testis, as well as an unusual localization and possibly decondensation of the D. simulans Y chromosome. We conclude on the basis of these data that hybrid male sterility is caused by a gain-of-function interaction between OdsHmau and some component of the D. simulans Y chromosome heterochromatin, with this protein-DNA interaction representing the Dobzhansky-Muller incompatibility.

OdsH shares similarities with the hybrid sterility genes Prdm9 (or Meisetz) in mouse (23) and Overdrive (Ovd) in Drosophila (24), all of which encode proteins with putative DNA-binding domains. Satellite DNAs have also been implicated in hybrid inviability, including a pericentric satellite locus (Zhr) (25, 26) and a gene encoding a heterochromatin-binding protein (Lhr) (27). Thus, rapidly evolving repetitive DNA elements driven by genetic conflict may represent a major evolutionary force driving sequence divergence of speciation genes that would ultimately result in hybrid incompatibilities (13, 14, 28).

Supporting Online Material

Materials and Methods

Figs. S1 to S8


  • * Present address: Dernburg Lab, MCB, University of California–Berkeley, Berkeley, CA 94720, USA.

References and Notes

  1. We thank C-I. Wu for the D. simulans fertile and sterile introgression lines; C. Ting for scientific discussions and sharing data; G. Findlay for initial observations on OdsH cytology; and K. Ahmad, S. Biggins, N. Elde, S. Henikoff, N. Phadnis, T. Tsukiyama, and D. Vermaak for comments on the manuscript. Supported by NIH training grant PHS NRSA T32 GM07270 (J.J.B.), and grants from the Mathers foundation and NIH R01-GM74108 (H.S.M.). H.S.M. is an Early-Career Scientist of the Howard Hughes Medical Institute.
View Abstract

Stay Connected to Science

Navigate This Article