Y Chromosome of D. pseudoobscura Is Not Homologous to the Ancestral Drosophila Y

See allHide authors and affiliations

Science  07 Jan 2005:
Vol. 307, Issue 5706, pp. 108-110
DOI: 10.1126/science.1101675


We report a genome-wide search of Y-linked genes in Drosophila pseudoobscura. All six identifiable orthologs of the D. melanogaster Y-linked genes have autosomal inheritance in D. pseudoobscura. Four orthologs were investigated in detail and proved to be Y-linked in D. guanche and D. bifasciata, which shows that less than 18 million years ago the ancestral Drosophila Y chromosome was translocated to an autosome in the D. pseudoobscura lineage. We found 15 genes and pseudogenes in the current Y of D. pseudoobscura, and none are shared with the D. melanogaster Y. Hence, the Y chromosome in the D. pseudoobscura lineage appears to have arisen de novo and is not homologous to the D. melanogaster Y.

The origin and evolution of the Drosophila Y seem to be different from that of the canonical (e.g., mammalian) Y chromosomes. Mammalian sex chromosomes originated from an ordinary pair of autosomes, so that when one of the homologs acquired a strong male-determining gene and became a Y, the other homolog became the X. Progressive gene loss from the Y resulted in a mostly degenerated chromosome (1, 2). The main evidence for this common origin of the sex chromosomes is that most of the mammalian Y-linked genes and pseudogenes are shared with the X (3). On the other hand, in D. melanogaster none of the nine known single-copy Y-linked genes have homologs on the X. Instead, their closest homologs are autosomal (4, 5), which strongly suggests that they were acquired from the autosomes by gene duplications, as has been shown for two mammalian Y genes (3). D. melanogaster X and Y chromosomes do share two multicopy genes, but it is doubtful that they represent common ancestry: Ste-Su(Ste) genes were recently acquired from an autosome (6), whereas rDNA genes are present in nonhomologous chromosomes in different Drosophila species (5, 7, 8). The lack of any clear sign of X-Y homology is consistent with the hypothesis that the Drosophila Y did not originate from the degeneration of an X-homolog, but rather from a supernumerary (“B”) chromosome that acquired X-pairing and male-related genes, though it is also possible that the degeneration went so far that all signs of homology were erased (5, 9). Whatever its true origin, the melanogaster-like Drosophila Y is at least 63 million years (My) old, dating back to the separation of the Sophophora and Drosophila subgenera (10), because at least three genes (kl-2, kl-3, and kl-5) are shared between the Y chromosomes of D. melanogaster (4, 5) and D. hydei (11, 12).

The assembled D. pseudoobscura genome sequence (13) now makes possible a genome-wide study of Y-linked genes in a second Drosophila species. Given that this species also belongs to the subgenus Sophophora, we expected it to share the ancestral Y chromosome with D. melanogaster, as does the more distant D. hydei. In fact, we found in the D. pseudoobscura genome orthologs for most D. melanogaster Y-linked genes (kl-2, kl-3, kl-5, ORY, PPr-Y, and ARY), as well as the orthologs of their autosomal parental genes (14). The orthology is strongly supported by phylogenetic analysis (Fig. 1 and fig. S1) and by the reciprocal best match criteria (14). However, when we tested for Y linkage by polymerase chain reaction (PCR) with genomic DNA from males and females, we found that all six genes are present in both sexes, ruling out Y linkage. A formal genetic analysis by means of single-nucleotide polymorphisms that exist between the subspecies D. p. pseudoobscura (the sequenced one) and D. p. bogotana demonstrated that all six genes have autosomal inheritance in D. pseudoobscura (fig. S2). Thus, the ancestral Drosophila Y chromosome suffered at least one translocation to an autosome in the D. pseudoobscura lineage (supporting online text). In this study, we analyzed in detail the orthologs of kl-2, kl-3, ORY, and PPr-Y genes; ARY and kl-5 will be dealt with in a separate paper. We failed to find the orthologs of four genes (14).

Fig. 1.

Phylogeny of PPr-Y and its parental gene CG13125. Orthology between the PPr-Y genes of D. melanogaster and D. pseudoobscura is indicated by their statistically significant grouping (note the bootstrap test). The same result was obtained with ORY, kl-2, kl-3, kl-5, and ARY (fig. S1). The bar indicates the number of amino acid substitutions per site. D. melanogaster PPr-Y is Y-linked, and all other genes are autosomal. Accession numbers: mel CG13125, NP_609325; mel PPr-Y, AAL25119; and Anopheles CG13125 ortholog, XP_319283. The sequences of D. melanogaster ARY and all D. pseudoobscura genes are available in the Third Party Annotation Section of the DDBJ/EMBL/GenBank databases under the accession numbers TPA: BK005622–BK005633 and BK005606.

To date the Y → autosome translocation event(s) we examined progressively more distantly related species (15), by doing PCR in males and females with the same primers used for D. pseudoobscura. As shown in Fig. 2, we found that the four genes (kl-2, kl-3, ORY, and PPr-Y) are present in both sexes of D. persimilis and D. miranda (pseudoobscura subgroup) and of D. affinis and D. azteca [affinis subgroup; 2- to 3-My ago divergence from D. pseudoobscura (16)], and are male-specific in D. guanche and D. bifasciata [obscura subgroup; 18-My ago divergence (10)]. This shows that the Y-autosome translocation(s) occurred between 2 and 18 My ago [these numbers are approximations, because the divergence times are not precisely known (10)]. These results also help explain the observation that, in contrast to almost all Drosophila species, D. affinis males devoid of Y chromosomes are fertile (17).

Fig. 2.

Phylogenetic origin of the Y-autosome translocation. (A) Simplified phylogeny of the genus Drosophila (15). (B) PCR amplification of the kl-3 gene in males and females of D. pseudoobscura (pse, lanes 2 and 3), D. azteca (azt, lanes 4 and 5), D. affinis (aff, lanes 6 and 7), D. bifasciata (bif, lanes 8 and 9), and D. guanche (gua, lanes 10 and 11). Y-linkage only occurs in the last two species, which belong to the more distant obscura subgroup. Amplification of kl-3 in both sexes also occurred in D. persimilis and D. miranda (pseudoobscura subgroup; not shown). The same pattern of kl-3 was observed for kl-2, PPr-Y, and ORY. These data show that the Y-autosome translocation occurred after the split of the obscura subgroup, and before the affinis-pseudoobscura split [arrow in (A)]. The identity of the PCR products was verified by sequencing (GenBank accession numbers: AY808044 to AY808058). Primer sequences and PCR conditions are available upon request.

The D. pseudoobscura genes described here almost certainly are functional, because (i) the coding regions have no premature stop codons, (ii) the putative splice junctions have the correct sequence, and (iii) reverse transcriptase (RT)–PCR shows that all genes are transcribed and are correctly spliced (fig. S3). Most of these genes have testis-restricted expression, as do their Y-linked orthologs in D. melanogaster (fig. S3) (4, 5).

The Y-autosome translocation(s) seems to have led to a profound genomic reorganization in the former Y-linked genes. In D. melanogaster and D. hydei the Y-linked genes are unusually large [∼1 to 5 megabase pairs (Mbp)], owing to megabase-sized introns (11, 18). Intergenic distances also seem to be in the Mbp range. We infer that in the ancestral state, these genes had huge introns and large intergenic spacing. In D. pseudoobscura, kl-2 and kl-3 span a mere 28 and 105 kb (the cDNAs themselves each span 13 kb), which strongly suggests that they shrank at least to 1/10th of their original size after the Y-autosome translocation (fig. S4). Furthermore, kl-2 and ORY are separated by 20 kb, and ORY and PPr-Y by 15 kb, whereas in no case was any pair of Y-linked genes in D. melanogaster even on the same scaffold. These observations strongly suggest that intergenic distances were also reduced. At least in genomic structure, the Y genes appear to have become nearly euchromatic after the translocation to the autosome.

D. pseudoobscura does have a free Y chromosome, and given the randomness of the whole-genome shotgun sequencing strategy, portions of its sequence must be among the unmapped scaffolds (19). We searched for Y-linked genes using the fact that the whole-genome shotgun libraries were prepared from unsexed embryos, and hence Y-linked DNA should have about one-fourth as many reads as autosomal DNA (19). This and other computational methods yielded “candidates” that were then tested for Y linkage by PCR (fig. S5). Among 37 tested scaffolds, we identified 14 Y-linked ones (∼150 kb) that contain 15 genes and pseudogenes (Table 1; tables S1 and S2). All scaffolds had a close euchromatic counterpart (>90% nucleotide identity); they arose through three segmental duplications involving several genes, one small genomic duplication, and one retrotransposition (table S1). None of these are shared with the D. melanogaster Y (Table 1 and table S1), which further supports the conclusion that these chromosomes are not homologous. However, it remains possible that the large uncharacterized parts of the two Y chromosomes are homologous. Efforts to identify more Y-chromosomal genes in both species and infer their functions are under way. The evidence that the 15 Y-linked genes and pseudogenes of D. pseudoobscura are not Y-linked in D. melanogaster is the following. BlastN searches using the D. melanogaster coding sequences of these genes and pseudogenes as the query and the D. melanogaster genome as the database always detected only the euchromatic copy, whereas the same search against the D. pseudoobscura genome readily detected two or more copies (one euchromatic and at least one Y-linked copy). Sequence gaps in the D. melanogaster Y chromosome cannot explain this result: Although the assembly of the Y chromosome is disrupted in the highly repetitive sequences (19), the coding sequences are fairly well represented (table S1).

Table 1.

Gene content of the D. melanogaster and D. pseudoobscura Y chromosomes.

Only in D. melanogaster YView inline Shared genes Only in D. pseudoobscura YView inline
kl-2 None found CG12218Y-Ψ (X)
kl-3 CG12848Y (2R)
kl-5 CG3894Y-Ψ (2R)
ORY CG3880Y-Ψ (2R)
PPr-Y CG10274Y (3L)
ARY CG10289Y-Ψ (3L)
Pp1-Y1 CG7376Y-Ψ (3L)
Pp1-Y2 CG10191Y (3L)
PRY CG17687Y-Ψ (3L)
CCY CG32120Y-Ψ (3L)
CG10171Y (3L)
CG14110Y (3L)
CG14111Y (3L)
CG14112Y (3L)
CG31175Y-Ψ (3R)
  • View inline* Orthologs in the D. pseudoobscura genome were detected for the first six genes, but these are located in an autosome. No ortholog was found for the last four genes.

  • View inline In D. pseudoobscura all 15 genes are present as duplications from euchromatic genes. In D. melanogaster only the euchromatic copy was found, whose location is shown in parentheses. Pseudogenes were assigned with the standard Ψ symbol. See also table S1.

  • An explanation for this exceptional Y-chromosome turnover may lie in the X chromosome of the D. pseudoobscura lineage. The D. pseudoobscura X originated from a fusion of the ancestral X and an autosome [the homolog of the D. melanogaster 3L chromosome, also called the Muller D element (Fig. 3) (13, 20)]. This fusion and the Y-autosome translocation(s) co-occur along the phylogeny: Both are present in the subgroups pseudoobscura and affinis, and both are absent in the obscura subgroup. In these X-autosome fusion systems, one member of the autosomal pair became part of the X and its homolog remained independent. However, owing to the mechanics of meiotic divisions, the free homolog behaves as a Y chromosome (20). Thus, the X-autosome fusion effectively produced two Y chromosomes: the ancestral one and a “neo-Y” [formerly a Muller D autosome (20)]. Theory (1, 2, 21) and empirical data (2123) indicate that neo-Y chromosomes degenerate, and the observation of only one Y chromosome in D. pseudoobscura had earlier led to the conclusion that the neo-Y was lost or fused to the Y (20) and that the current Y is the ancestral one. However, as we show here, the ancestral Y is now part of an autosome in D. pseudoobscura, leaving unresolved the origin of the current Y. An intriguing possibility is that the current Y is a degenerated neo-Y (Fig. 3). This hypothesis predicts that 3L-derived genes will be found on the current D. pseudoobscura Y, and indeed most of the genes we found are 3L-derived (Table 1). It remains to be seen whether they really trace back to the Y-A fusion or are more recent duplications. Generally, fusions between an autosome and a sex chromosome do not cause severe meiotic problems (20), but a detailed examination of one such case (D. miranda) revealed a low frequency (1 to 3%) of aneuploid gametes (24). We suggest that a second fusion, involving the other sex chromosome, could ameliorate the meiotic problems; this would explain the Y-autosome “fusion” reported here. Such a pair of fusions has occurred in D. albomicans (25); however, in that case the original Y chromosome retained its Y-chromosome status, whereas in D. pseudoobscura the formerly Y-linked genes became autosomal, and the current Y shares no homology with the ancestral one.

    Fig. 3.

    A proposed model for the origin of D. pseudoobscura sex chromosomes. Autosomes are shown as solid bars, sex chromosomes as striped bars, and centromeres as open circles. (A) Ancestral state (D. melanogaster and obscura subgroup species), with autosomal pairs (shown three: A1, A2, and A3) and the sex chromosomes. A centric fusion between the X and one autosome (Muller element D, corresponding to the 3L in D. melanogaster) generated an X Y1 Y2 sex-determination system, shown in (B). This system was transient in the D. pseudoobscura lineage, but exists in many species (20). A second fusion (or large translocation) between the ancestral Y (Y1) and another autosome (still unidentified) led to the current state of D. pseudoobscura and closely related species [pseudoobscura and affinis subgroups; step (C)]. The order of the fusions and the origin of the current Y (shown here as the neo-Y) are hypothetical, but it is certain that formerly Y-linked genes now reside on an autosome. For the sake of simplicity we represented this event as a centric fusion, but it may instead be one or more translocation events. The proposed origin of the current Y (from Muller element D) predicts that it pairs during meiosis with the autosomally derived arm of the X (XR), rather than with the ancestral X (XL).

    Y chromosomes are known to accumulate autosomal male-related genes both in Drosophila and in mammals (35). The finding that the ancestral Drosophila Y chromosome is now part of an autosome in the D. pseudoobscura lineage is thus unexpected, because the movement was in the opposite direction. The D. pseudoobscura Y has all the features of a typical Drosophila Y chromosome (including X pairing and being essential for male fertility), and yet it is less than 18 My old. In addition to the origin of Y chromosomes, the present findings have a direct bearing on the evolution of heterochromatin and intron sizes, because heterochromatin expanded in introns and intergenic regions of Y-linked genes and contracted when those Y-linked genes moved back to autosomes.

    Supporting Online Material

    Materials and Methods

    SOM Text

    Figs. S1 to S5

    Tables S1 and S2


    References and Notes

    View Abstract

    Stay Connected to Science

    Navigate This Article