“Species are groups of actually or potentially interbreeding natural populations, which are reproductively isolated from other such groups” [1] (Biological Species Concept, BSC). Based on the BSC, speciation researchers have been trying to isolate genes responsible for reproductive isolation. Recent genome-wide analyses in Drosophila accelerate the identification of such genes.

Coyne and Orr [2] classified reproductive isolating barriers into three categories: I. Premating, II. Post-mating/pre-zygotic and III. Post-zygotic. Hybrid sterility and inviability are included in category III and seen in the F1 and later generation hybrids; those seen in the descendants are also called hybrid breakdown [3]. Hybrid sterility and inviability are generally the result of epistatic interaction between genes from different species (e.g., hybrid incompatibility; HI), as it has been suggested by Dobzhansky [3] and Muller [4].

Genomic constructions of the F1 and later generation hybrids (and HI genes) are shown in Figure 1. The F1 genotypes are homogeneous, not differing among individuals of the same sex. The majority of the F1 genomes are heterozygous, carrying heterospecific alleles. Therefore, F1 viability/fertility seems to be generally affected by dominant HI genes (Figure 1i). Exceptionally the sex-linked HI genes can be dominant or recessive, because sex-linked genes may be hemizygous in one sex. On the contrary, the genotype of F2 (produced by the sibling cross) or BC1 (produced by the backcross) varies due to recombination. Some genomic regions may be homozygous for alleles from one parental species, and F2 and BC1 viability/fertility seems to be affected by recessive HI genes (Figure 1ii and 1iii). Such difference of HI genes dominance in between F1 and F2/BC1 has been stressed previously [5].

Figure 1: Genomic constructions and HI genes in hybrids (F1, F2, and BC1). Parental species (sp. 1 and sp. 2) are also indicated. chr., chromosome; dom, dominant; rec, recessive.

Drosophila melanogaster and a sibling species, D. simulans, have been the model system to elucidate HI genes. The cross between D. melanogaster females and D. simulans males produce only a sterile female F1; male F1 is lethal later at the larval stage (Figure 2A) [6]. In the present review we do not consider the reciprocal cross for simplicity, where the viable/lethal sex is reversed [6]. Male F1 is rescued if a D. simulans mutant of the Lethal hybrid rescue (Lhr) gene [7] or a D. melanogaster mutant of the Hybrid male rescue (Hmr) gene [8] is used for the cross (Figure 2B). Genome-wide analyses of HI genes in this cross are reviewed here (Table 1).

Figure 2: Crosses between D. melanogaster females and D. simulans males. A, both parents are wildtype. B, D. simulans has an Lhr mutation. C-F, D. melanogaster is heterozygous for a deficiency (Df) and a balancer (Bal). (D, D. simulans has Lhr.) G, D. simulans has newly induced mutations (including Dfs). chr., chromosome.

Table 1: Genome-wide analyses of hybrid incompatibility genes in the cross between D. melanogaster females and D. simulans males.

Strategy 1

Coyne et al. [9] conducted pioneering work where they made recessive D. simulans HI genes hemizygous over D. melanogaster deficiencies in F1 (Figure 2C). Female F1 viability of carriers vs. non-carriers of the deficiency was compared. The HI partner of D. melanogaster must be dominant, because the F1 genomes are heterozygous. Matute et al. [10] refined the mapping using more deficiencies (and extended the study to a more distantly-related species, D. santomea). These studies resulted in 10 HI gene regions in the 79.4% genome tested.

Strategy 2

Presgraves [11] conducted similar crosses, where he used Lhr instead of the wild type D. simulans to rescue male F1 (Figure 2D) (for a pilot test see Sawamura [12]). Male F1 viability of carriers vs. non-carriers of the deficiency was compared. The HI partner of D. melanogaster can be recessive if it is X-linked, and the screening is more sensitive than Strategy 1. Presgraves [11] detected 20 lethal and 20 semilethal HI gene regions in the 70% autosome tested, and two have been identified by further studies: Nucleoporin 96 (Nup96) and Nucleoporin 160 (Nup160) [13-15].

Strategy 3

Cuykendall et al., [16] using crosses similar to as Strategy 1, mapped dominant HI gene regions that rescued F1 males when the regions are deleted (Figure 2E). The D. melanogaster Hmr is an example of such dominant HI genes; a loss-of-function of the gene rescues male F1 [17]. Of note, Cuykendall et al. [16] preferentially used D. mauritiana instead of D. simulans, because hybrids can be rescued easier [8,18]. Cuykendall et al. [16] did not detect major HI genes but detected multiple minoreffect HI genes in the 89% autosome tested.

Strategy 4

Female F1 is viable at low temperature (e.g., 18C) but die at the late pupal stage or just after eclosion at high temperature (e.g., 25C) (Figure 2A) [19,20]. Deficiencies of dominant HI genes are expected to rescue the female F1 (Figure 2F). Although the effect of the D. melanogaster Lhr was not detected in male F1 (e.g., not rescued by the deficiency) (Strategy 3; see also Brideau et al. [21]), it was detected in female F1 (e.g., rescued by the deficiency) [22]. The data presented in Coyne et al., [9] Matute et al., [10] and Cuykendall et al. [16] can be reanalyzed for female F1 viability rescue by deficiencies. The genome-wide analysis of the dominant HI genes is in progress (KS, T. Hayashi, K. Miura, and Q. Araye, unpublished).

Strategy 5

Until now, it was difficult to screen D. simulans mutations because elegant genetic tools like balancer chromosomes were not available in this species. Phadnis et al. [23] overcame this difficulty by inducing point mutations (and deficiencies) in D. simulans males and crossing them with D. melanogaster females (Figure 2G). Rescued male F1 must have mutations on dominant HI genes. The D. simulans Lhr is an example of such genes; a loss-of-function of the gene rescues male F1 [21]. Phadnis et al. [23] used next-generation sequencing of the recovered D. simulans mutations and discovered the third gene involved in the F1 inviability: Suppressor of Killerof- prune (Su(Kpn))=glutathione-S-transferase-containing FLYWCH zinc finger protein (gfzf). More genes will be discovered, because this screening has not been saturated [23].

Strategy 6

If D. melanogaster has attached-X chromosomes, the cross between XXY D. melanogaster females and D. simulans males produce only sterile male F1; female F1 is lethal at the late larval stage (Figure 3A) [19]. Matute and Gavin-Smyth [24] used such D. melanogaster females who carry a series of X duplications on the Y chromosome (Figure 3B), (and extended the study to D. mauritiana and D. santomea). The male F1 would be lethal if the duplicated region contains dominant HI genes. Matute and Gavin-Smyth [24] detected two HI gene regions in the 72% X chromosome tested and one seems to be the Hmr locus.

Figure 3: Crosses between XXY D. melanogaster females and D. simulans males. A, chromosomes are normal other than the attached-X chromosomes in females. B, the D. melanogaster Y chromosome carries a duplication (Dp) of genes from the X chromosome; Dp (1;Y). chr., chromosome.

The genome-wide analyses of HI genes in the cross between D. melanogaster females and D. simulans males are productive. Number of the HI genes was roughly estimated and some of the HI genes have been molecularly identified. As it can be seen from Table 1, recessive HI genes of D. simulans have been well documented; more than 20 such genes exist and two of them (Nup96 and Nup160) have been identified. Dominant HI genes of D. melanogaster have also been investigated (e.g., Hmr), and more will be discovered by strategy 4. One of dominant HI genes of D. simulans is known since classic studies (Lhr), and more will be discovered by strategy 5 (e.g., gfzf). On the contrary, recessive HI genes of D. melanogaster have not been investigated at all. New strategies identifying such genes are awaited. In the near future, more HI genes will be identified not only in this crossing system but also from diverse organisms thanks to the advance of the genomic sequencing technology.

I am grateful to Dr. T. L. Karr and Mrs. T. Hayashi, K. Miura, and Q. Araye for discussion. The author’s current study is supported by Grants-in-Aid for Scientific Research (21570001 and 24570001) to KS from the Japan Society for the Promotion of Science.