• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of pnasPNASInfo for AuthorsSubscriptionsAboutThis Article
Proc Natl Acad Sci U S A. Apr 29, 2003; 100(9): 5302–5307.
Published online Apr 14, 2003. doi:  10.1073/pnas.0836927100
PMCID: PMC154340
Genetics

A rapidly evolving MYB-related protein causes species isolation in Drosophila

Abstract

Matings among different species of animals or plants often result in sterile or lethal hybrids. Identifying the evolutionary forces that create hybrid incompatibility alleles is fundamental to understanding the process of speciation, but very few such alleles have been identified, particularly in model organisms that are amenable to experimental manipulation. We report here the cloning of the first, to our knowledge, Drosophila melanogaster gene involved in hybrid incompatibilities, Hybrid male rescue (Hmr). Hmr causes lethality and female sterility in hybrids among D. melanogaster and its sibling species. We have found that Hmr encodes a protein with homology to a family of MYB-related DNA-binding transcriptional regulators. The HMR protein has evolved both amino acid substitutions and insertions and deletions at an extraordinarily high rate between D. melanogaster and its sibling species, including in its predicted DNA-binding domain. Our results suggest that hybrid lethality may result from disruptions in gene regulation, and we also propose that rapid evolution may be a hallmark of speciation genes in general.

The process of speciation requires that organisms establish and maintain reproductive isolation from other species. Postzygotic reproductive isolation is widely observed in plants and animals as the sterility and lethality of species hybrids (1). The theory of how such hybrid incompatibilities evolve was proposed independently >60 years ago by Dobzhansky and Muller (2, 3). Their theory envisions that during or after speciation different alleles will evolve in different species. These alleles can reach fixation by either positive selection or neutral drift, but during this period, there is no selection against potentially deleterious interactions with derived alleles that are evolving in other species. These deleterious interactions will be observed only in hybrids between the species.

The power of the Dobzhansky/Muller model derives from its generality. The model can be easily extended to include complex multigene interactions, and it makes no assumptions about the types of genes or allelic changes required to produce incompatibilities. The model is well supported by both theoretical (4) and experimental (1) genetic studies, but only a handful of specific hybrid incompatibility genes have been identified. Without molecular and functional studies of hybrid incompatibility genes, it remains unknown whether particular types of genes or pathways are preferentially involved and what evolutionary processes drive the divergence of these speciation genes.

The fruitfly Drosophila melanogaster can hybridize with the closely related species Drosophila simulans, Drosophila mauritiana, and Drosophila sechellia. These three species are estimated to have diverged from D. melanogaster ≈2–3 million years ago (5). The cross of D. melanogaster females to males of any of these three sibling species produces F1 hybrid daughters that are viable but sterile at low temperatures and lethal at high temperatures, whereas F1 hybrid males are invariably lethal (6, 7).

D. melanogaster hybrids have served as a model for investigating species incompatibilities for >80 yr, often using creative and novel genetic techniques (8). By inducing high rates of nondisjunctional progeny, Sturtevant (6) demonstrated that lethality in D. melanogaster hybrids is not truly sex-specific but rather correlates with the presence of the D. melanogaster X chromosome (6). Muller and Pontecorvo developed the technique of creating pseudobackcross hybrid progeny from matings between heavily irradiated triploid D. melanogaster and normal diploid D. simulans, and further studies (9) suggested that several autosomal genes may also be involved in hybrid lethality.

A limitation of these classic studies is that genetic effects could not be localized beyond the chromosome level, because the complete sterility of F1 hybrids precludes genetic mapping through backcrosses. Candidate hybrid incompatibility genes have instead been found by searching for mutations that suppress F1 hybrid lethality (8). The D. melanogaster gene Hybrid male rescue (Hmr) was discovered based on loss-of-function mutant alleles that suppress hybrid male lethality (10, 11) and that were subsequently shown to also suppress high-temperature hybrid female lethality (12) and to partially suppress female sterility (13). Increasing the dosage of the wild-type Hmr+ gene has the reciprocal property of reducing hybrid viability (12, 14). These results demonstrate that Hmr has a major effect on D. melanogaster hybrid viability and female fertility and is therefore a prime candidate to initiate molecular studies of hybrid incompatibilities in this model organism.

Materials and Methods

Fly Stocks.

Deficiency and mutant stocks have been described (12, 15) or are available on FlyBase (http://flybase.org) (16). Hmr was sequenced from the sibling stocks D. simulans w, D. mauritiana iso-female line 207, and D. sechellia Robertson Line 1.

Deficiency Mapping.

The location of deficiency breakpoints was determined by Southern blot analysis of genomic DNA from heterozygous adults or by PCR analysis of hemizygous deficiency embryos identified as GFP-negative progeny from deficiency/FM7i, P{w+mC = ActGFP}JMR3 stocks.

Transgenic Constructs.

Inserts for P element constructs were subcloned from a genomic cosmid clone (17). Ligations of DNA fragments with incompatible ends were performed after filling in the 3′ recessed ends with Klenow enzyme. The insert of p59 was cloned into the NotI and StuI sites of CaSpeR-4, the insert of p60 was cloned into the XhoI site of CaSpeR-4, the insert of p72 was cloned into the BamHI and SacII sites of CaSpeR-2, and the insert of p74 was cloned into the StuI site of CaSpeR-2. The 4-bp frameshift insertion of p73 was made by digesting p72 with SpeI, filling in with Klenow enzyme, and religating the plasmid. p83 was made by digesting p72 with Acc65I and BamHI, filling in with Klenow enzyme, and religating the plasmid. The structures of p73 and p83 were confirmed by DNA sequencing. Flies were transformed by standard methods; only autosomal insertions were used.

Molecular Biology.

Total RNA was prepared by using TRI Reagent (Molecular Research Center, Cincinnati). For RT-PCR, 1 μg of DNase-treated RNA was reverse-transcribed with 0.5 μg of oligo(dT)12–18 and 200 units of SuperScript II (GIBCO/BRL). Thirty cycles of PCR were performed on 1/20th of the reverse transcription reaction with the primers 5′-TACGTTGGGCACAAACTTTAGGGGTAT-3′ and 5′-AGTGCCATTTCAAACGAAGTCCAT-3′ (58°C annealing temperature).

Northern blot analysis was performed with 15 μg of total RNA by using standard methods (18). rRNA was detected with methylene blue after transfer to nylon membranes. The template for the D. simulans Hmr probe in Fig. Fig.33A was a PCR product amplified from D. simulans DNA by using the primers 5′-GAACTGCTTAGGCCACAAAATC-3′ and 5′-CTGGTATTTTAGCGCTATCTCGTC-3′. This probe includes parts of exons 2 and 3.

Figure 3
Hmr evolves rapidly in Drosophila. (A) Northern blot analysis of CG1619 expression in the sibling species. Note that the levels of CG1619 cannot be accurately compared between the species because the D. simulans probe used has incomplete homology to the ...

DNA fragments from the rescue alleles and the sibling species were amplified by PCR and sequenced directly on both strands by using Big Dye chemistry (Applied Biosystems, version 3).

Results

Mapping and Cloning of Hmr.

We obtained complete clone coverage across the Hmr region (12) by chromosome walking and defined its boundaries by locating the breakpoints of deficiencies previously typed as Hmr+ or Hmr by assays of high-temperature rescue of hybrid females (12, 15) (Fig. (Fig.1).1). We focused on the region bounded by the distal breakpoint of Df(1)ras203 and the proximal breakpoints of Df(1)CH6 and Df(1)AC2LABR. This mapping is consistent with all deficiencies except Df(1)EP307-1-2 (15); the relationship of this rescuing mutation to Hmr function is unclear (see below). The Hmr region contains the complete transcription units for five predicted genes (Fig. (Fig.1).1).

Figure 1
Physical and genetic maps of the Hmr region. A chromosome walk of cosmid (28) and P1 (29) clones with their cytological locations on the X chromosome is shown above a map of deficiency breakpoints. Rescue status refers to assays for high-temperature hybrid ...

The Hmr1 and In(1)AB (11) rescue alleles behave as loss-of-function mutations in Hmr+ (11, 12, 14). Consistent with this proposal is the fact that the Hmr+ duplication Dp(1;2)v+75d suppresses the hybrid male rescue activity of these alleles (ref. 11; D.A.B. and J.R., unpublished data). We therefore reasoned that a cloned copy of Hmr+ introduced into the fly on a P element transgene should mimic this behavior. Transgenic constructs covering four candidate genes were made and assayed in hybrid males for suppression of rescue by Hmr1 in D. melanogaster/D. mauritiana hybrids and by In(1)AB in D. melanogaster/D. simulans hybrids (Fig. (Fig.22A; Table Table11 and Table 2, which is published as supporting information on the PNAS web site, www.pnas.org).

Figure 2
Identification of CG1619 as Hmr. (A) P element transgene constructs assayed for Hmr activity shown with the restriction sites used for cloning. Transgenes were classified as Hmr+ if they suppressed Hmr1- and In(1)AB-dependent hybrid male rescue ...
Table 1
An Hmr+ P element transgene reduces viability and suppresses female fertility rescue

The p72 construct containing Rab9D, CG1619, and CG2124 had Hmr+ activity by these assays. Hmr+ activity remained in derivative constructs that deleted Rab9D (p83) or the majority of the predicted CG2124 coding region (p74). In contrast, a construct identical to p72 except containing a frame-shift mutation in the CG1619 coding region (p73) had no Hmr+ activity. The results from these four constructs demonstrate that CG1619 is necessary and sufficient to supply Hmr+ function.

In addition to its effects on hybrid males, the Hmr+ duplication Dp(1;2)v+75d reduces hybrid female viability (12) and suppresses the rescue of female sterility by In (1)AB (13). Both of these properties are mimicked by the p72 transgene containing CG1619 (Table (Table1).1). Adding one copy of the transgene to Hmr+/+ female hybrids reduced their viability >10-fold. Unrescued hybrid females are normally agametic and In (1)AB rescues this phenotype (13). Egg counts of hybrid females confirmed this rescuing activity in In(1)AB/+ animals, whereas their sisters carrying the transgene were almost completely suppressed for rescue.

The Structure and Sequence of Hmr+ and Rescue Alleles.

The structure of CG1619 was defined by the 3.4-kb LD22117 cDNA clone (19) (Fig. (Fig.22C). Northern blot analysis, however, identified a single band of ≈5 kb at all developmental stages (Fig. (Fig.22B), including the critical early larval period defined by temperature-shift experiments (10, 12). Searches of the Drosophila EST database identified the 5′ end of cDNA clone RE54143 beginning 884 bp upstream of LD22117. Sequencing of this 4.7-kb cDNA revealed that besides having an additional 5′ exon, the region defined as being intron 1 by LD22117 is present in RE54143 as an exon (Fig. (Fig.22C). RT-PCR analysis provided no evidence for the LD22117-derived gene structure in two different wild-type backgrounds (Fig. (Fig.22D). These results indicate that the LD22117 cDNA clone either corresponds to a rare minor form or is an artifact, and that RE54143 represents the true structure of CG1619.

To determine whether the rescue alleles contain mutations affecting CG1619, we examined the region corresponding to the p72 transgene by genomic Southern analysis (data not shown) and by DNA sequencing. Hmr1 was isolated from a wild stock and previously shown (11) by in situ hybridization to contain P elements at several locations, including in the Hmr region at 9D3.4 or 9E1.2. We found a 652-bp internally truncated element 32 bp 5′ to CG1619 (Fig. (Fig.22C). Northern analysis showed that Hmr1 produces a high molecular weight transcript whose size is consistent with the possibility that it initiates within or just upstream of the truncated P element (Fig. (Fig.22B, lane 7). The wild-type transcript is also present but apparently reduced in level, which is consistent with genetic data suggesting that Hmr1 is a partial loss-of-function (hypomorphic) mutation (12). The extra-high molecular weight band is lost in Hmr1r1, a derivative of Hmr1 that we generated by transposase-mediated excision (Fig. (Fig.22B, lane 8). Sequencing of Hmr1r1 showed that the P element is completely excised and the sequence reverted back to the wild-type. Hmr1r1 has also completely lost hybrid rescue activity, as shown in crosses to D. mauritiana, where only females were produced (n = 350), whereas control crosses with nonexcision stocks produced 162 rescued males and 192 females. We conclude that the Hmr1 hybrid rescue phenotype is caused by the P element insertion upstream of CG1619.

Northern analysis showed no changes in transcript level in RNA from In(1)AB or Df(1)EP307-1-2 animals (Fig. (Fig.22B, lanes 6 and 9). Sequencing of In(1)AB identified the amino acid replacement changes E371K and G527A. In the absence of a parental allele, future work will be required to determine whether these substitutions impair CG1619 function. Df(1)EP307-1-2 contains the substitutions D191A and D288G; however, both substitutions are also found in the nonrescuing parental allele EP307 and thus cannot be causing the rescue phenotype. Df(1)EP307-1-2 could conceivably rescue by affecting CG1619 expression specifically in hybrids. Alternatively, it might have a mutation in a different unidentified gene that affects hybrid viability; whether such a hypothetical gene is likely to be within the region deleted by Df(1)EP307-1-2 is unclear, because this region is also deleted by the nonrescuing Df(1)C52 (Fig. (Fig.11).

Database queries using the predicted HMR protein identified two regions with homology to the MADF protein domain (Fig. (Fig.22 C and E), which was first identified in the Drosophila sequence-specific transcription factor ADF1 (20). The MADF domain of ADF1 contains similarities to the helix–loop–helix motif of the DNA-binding domain of the MYB oncoproteins (20), as well as the SANT domain (21), which is found in DNA-binding proteins throughout the eukaryotes. Both domains of HMR contain conserved amino acids found in these protein families (Fig. (Fig.22E), suggesting that it also may have DNA-binding and transcriptional regulatory activities. Outside of the MADF domains, however, no homology was found, including in the recently completed Anopheles gambiae (mosquito) genome, suggesting that Hmr may be a relatively young gene.

High Level of Divergence of Hmr.

Genetic experiments, such as our transgene analysis, demonstrate that the wild-type D. melanogaster Hmr+ is deleterious to hybrids. One explanation for the rescuing activity of Hmr mutations is that they are second-site suppressors that reduce or eliminate the effects of deleterious interactions among other genes that have diverged between D. melanogaster and its sibling species. If so, then Hmr would not necessarily be functionally diverged between the species. Alternatively, functional divergence at the Hmr locus itself might directly cause the observed hybrid incompatibilities. The molecular evolution of Hmr might suggest which scenario is more likely and, more importantly, will guide future experiments necessary to distinguish these possibilities.

Northern analysis identified a single transcript in the sibling species of comparable size to that seen in D. melanogaster (Fig. (Fig.33A). DNA sequence analysis revealed that Hmr from each species encodes a protein of similar length with two MADF domains (Fig. (Fig.33C and Fig. 4, which is published as supporting information on the PNAS web site), but with an unusually high average rate of 0.077 for nonsynonymous (replacement) divergence (Dn) between D. melanogaster and the sibling species (Fig. (Fig.33B). In addition to the extraordinary number of amino acid replacements, D. melanogaster Hmr also contains between 8 and 11 in-frame insertions or deletions in its coding region compared with the sibling species (Fig. 4). Divergence is also unusually high among the sibling species, with a Dn value (0.02) slightly greater than the average synonymous divergence (0.015).

Divergence rates, however, are not homogeneous across Hmr (Fig. (Fig.33B). The beginning of exon 4 has reduced replacement divergence, suggesting that it may be under functional constraint. Several regions, such as the MADF1 domain and surrounding position ≈2,800 in exon 2, have Dn values greater than synonymous divergence both between D. melanogaster and the sibling species and among the sibling species, which may indicate adaptive evolution.

Discussion

Several pieces of evidence indicate that Hmr corresponds to the MADF-containing gene CG1619. First, CG1619 is found in region 9D where Hmr was previously mapped by meiotic and cytological mapping. Second, P element constructs containing CG1619 have the properties of Hmr+ described for hybrid viability and female fertility. Third, we have shown that the original Hmr rescue allele, Hmr1, is caused by a P element insertion very close to the 5′ end of CG1619.

The HMR protein is related to the transcriptional regulatory proteins ADF1 and MYB (Fig. (Fig.22E), and Hmr is expressed at all stages of the life cycle (Fig. (Fig.22B). These findings suggest that the deleterious effect of Hmr+ in hybrids may be caused by misregulation of gene expression throughout development. This proposal is consistent with previous observations that have revealed widespread defects in hybrids. F1 hybrid males die as larvae lacking imaginal discs, which is a phenotype characteristic of zygotic cell cycle mutations. In conjunction with observed defects in chromosome condensation in dying hybrids, these findings have led to the proposal that hybrid lethality may be caused by aberrant cell division (22). Semiviable F1 hybrid females suffer from various morphological defects (12), and recent work has shown that Hmr has a major effect on ovarian morphology and function (13).

We have found that Hmr is one of the most rapidly evolving genes in Drosophila. Such a high Dn value for Hmr of 0.077 between D. melanogaster, and its sibling species has generally been observed only for the short accessory gland proteins (23) and contrasts with a mean Dn value of 0.014 between D. melanogaster and D. simulans obtained in a sample of 21 X-linked genes (24). Population genetic analysis will be required to test the hypothesis that the divergence of Hmr reflects positive selection, but our present results do strongly suggest that this divergence cannot be explained as nonfunctionalization of Hmr caused by a loss of selective constraint. We have shown that despite the many substitutions and insertions and deletions in Hmr, the gene is expressed in all four species analyzed (Fig. (Fig.33A) and maintains an intact ORF (Fig. 4). We have also found that the divergence of Hmr is not homogeneous across the gene (Fig. (Fig.33B), as one would expect if it is evolving neutrally. It is also notable that, whereas there are many replacement substitutions within the MADF domains, most have occurred at residues that are not among the most conserved in the MADF domain family (Fig. (Fig.33C). This last observation suggests that HMR is a functional DNA-binding protein in both D. melanogaster and its sibling species, but that the different species forms may have altered DNA-binding preferences and/or regulatory properties.

Although other genes likely contribute to D. melanogaster hybrid lethality (12), we have shown here that altering the activity of a single gene, Hmr, has a major effect on hybrid viability. The relatively simple genetic basis of hybrid lethality in D. melanogaster appears very different from the complex multigenic basis of hybrid male sterility in the D. melanogaster sibling species (25, 26). It is therefore especially striking that the Odysseus gene, which causes male sterility in D. simulans/D. mauritiana hybrids and is the only other hybrid incompatibility gene identified in Drosophila, also exhibits rapid evolutionary divergence with a high rate of replacement substitutions in its putative DNA-binding domain (27). Many proteins involved in fertilization also evolve rapidly and may cause prezygotic species isolation, for example through sperm-egg incompatibilities (23). Our first glimpses of speciation genes therefore suggest that they may have the special property of rapid sequence divergence.

These findings raise important questions about the mechanism of species incompatibilities. If the rapid evolution of speciation genes is found to be typical, it might suggest that multiple substitutions with synergistic effects are required before a gene becomes functionally diverged enough between species to cause a hybrid phenotype. Alternatively, a few substitutions, or even a single substitution, might be sufficient to create a hybrid incompatibility allele, with their high divergence rates reflecting the fact that many substitutions must occur before one has functional significance in hybrids. These possibilities can be addressed by investigating the mechanism of hybrid lethality by using the powerful genetic resources of D. melanogaster. Identifying the genes that interact with or are downstream of Hmr will also help to answer whether hybrid incompatibility genes are in general evolving rapidly.

Supplementary Material

Supporting Information:

Acknowledgments

We gratefully acknowledge M. Ashburner and C. H. Langley, in whose laboratories this research was done, for advice and support. We also thank P. Awadalla for helpful discussions, P. Hutter for communicating unpublished results, C. Jones for debugging, D. J. Begun and S. Russell for comments on the manuscript, and the United Kingdom Human Genome Mapping Project Resource Centre for clones. This work was supported by grants from the National Science Foundation (to D.A.B. and C. H. Langley) and the Medical Research Council and Leverhulme Trust (to M. Ashburner).

Abbreviation

Dn
nonsynonymous divergence

Footnotes

This paper was submitted directly (Track II) to the PNAS office.

Data deposition: The sequences reported in this paper have been deposited in the GenBank database (accession nos. AY260051AY260057).

References

1. Coyne J A, Orr H A. Philos Trans R Soc London B. 1998;353:287–305. [PMC free article] [PubMed]
2. Dobzhansky T. Genetics and the Origin of Species. New York: Columbia Univ. Press; 1937.
3. Muller H J. In: The New Systematics. Huxley J, editor. Oxford: Clarendon Press; 1940. pp. 185–268.
4. Turelli M, Orr H A. Genetics. 2000;154:1663–1679. [PMC free article] [PubMed]
5. Lachaise D, Cariou M L, David J R, Lemeunier F, Tsacas L, Ashburner M. Evol Biol. 1988;22:159–225.
6. Sturtevant A H. Genetics. 1920;5:488–500. [PMC free article] [PubMed]
7. Lachaise D, David J R, Lemeunier F, Tsacas L, Ashburner M. Evolution (Lawrence, KS) 1986;40:262–271.
8. Sawamura K. Plant Species Biol. 2000;15:237–247.
9. Pontecorvo G. J Genet. 1943;45:51–66.
10. Hutter P, Ashburner M. Nature. 1987;327:331–333. [PubMed]
11. Hutter P, Roote J, Ashburner M. Genetics. 1990;124:909–920. [PMC free article] [PubMed]
12. Barbash D A, Roote J, Ashburner M. Genetics. 2000;154:1747–1771. [PMC free article] [PubMed]
13. Barbash D A, Ashburner M. Genetics. 2003;163:217–226. [PMC free article] [PubMed]
14. Orr H A, Irving S. Genetics. 2000;155:225–231. [PMC free article] [PubMed]
15. Barbash, D. A., Roote, J., Johnson, G. & Ashburner, M. (2003) Genetica, in press.
16. FlyBase. Nucleic Acids Res. 1999;27:85–88. [PMC free article] [PubMed]
17. Tamkun J W, Deuring R, Scott M P, Kissinger M, Pattatucci A M, Kaufman T C, Kennison J A. Cell. 1992;68:561–572. [PubMed]
18. Sagerström C G, Sive H L. In: A Laboratory Guide to RNA: Isolation, Analysis, and Synthesis. Krieg P A, editor. New York: Wiley–Liss; 1996. pp. 83–103.
19. Adams M D, Celniker S E, Holt R A, Evans C A, Gocayne J D, Amanatides P G, Scherer S E, Li P W, Hoskins R A, Galle R F, et al. Science. 2000;287:2185–2195. [PubMed]
20. England B P, Admon A, Tjian R. Proc Natl Acad Sci USA. 1992;89:683–687. [PMC free article] [PubMed]
21. Aasland R, Stewart A F, Gibson T. Trends Biochem Sci. 1996;21:87–88. [PubMed]
22. Orr H A, Madden L D, Coyne J A, Goodwin R, Hawley R S. Genetics. 1997;145:1031–1040. [PMC free article] [PubMed]
23. Swanson W J, Vacquier V D. Nat Rev Genet. 2002;3:137–144. [PubMed]
24. Begun D J. Mol Biol Evol. 2002;19:201–203. [PubMed]
25. Palopoli M F, Wu C I. Genetics. 1994;138:329–341. [PMC free article] [PubMed]
26. True J R, Weir B S, Laurie C C. Genetics. 1996;142:819–837. [PMC free article] [PubMed]
27. Ting C-T, Tsaur S-C, Wu M-L, Wu C-I. Science. 1998;282:1501–1504. [PubMed]
28. Siden-Kiamos I, Saunders R D C, Spanos L, Majerus T, Treanear J, Savakis C, Louis C, Glover D M, Ashburner M, Kafatos F C. Nucleic Acids Res. 1990;18:6261–6270. [PMC free article] [PubMed]
29. Hartl D L, Nurminsky D I, Jones R W, Lozovskaya E R. Proc Natl Acad Sci USA. 1994;91:6824–6829. [PMC free article] [PubMed]
30. Hutter P. Dev Genes Evol. 2002;212:504–512. [PubMed]
31. Szabó K, Jékely G, Rørth P. Mech Dev. 2001;101:259–262. [PubMed]
32. Nei M, Gojobori T. Mol Biol Evol. 1986;3:418–426. [PubMed]

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences
PubReader format: click here to try

Formats:

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...

Links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...