The Process of Degeneration Can be Studied in Newly Evolved Y Chromosomes

There are many examples of intermediate stages in the process of degeneration. (Recall that in Table 22.3, we compared Aedes mosquitoes, which have a single sex determining locus, with Anopheles mosquitoes, which have distinct X and Y chromosomes, as in Drosophila.) Studies of new Y chromosomes, formed by fusion between the X and an autosome, show that the process of degeneration is slow, taking millions of generations (Charlesworth and Charlesworth 2000; Charlesworth et al. 2005). Drosophila americana acquired a neo-Y chromosome at most a few hundred thousand years ago, since it separated from its closest relative, Drosophila texana. All loci studied show similar expression on the neo-Y and on the homologous part of the X (McAllister and Charlesworth 1999). Drosophila miranda has an older neo-Y chromosome, which evolved after separation from Drosophila pseudoobscura, approximately 2 Mya. Divergence between neo-X and neo-Y in D. miranda shows that these chromosomes have been separated by the suppression of recombination for approximately 1.1 Myr. Although homologous loci can be identified, the D. miranda neo-Y shows clear signs of degeneration, and the neo-X has evolved partial dosage compensation (Bachtrog 2003). Finally, the D. obscura group of species shares a neo-Y that evolved approximately 13 Mya. This carries few functional genes and is fully dosage compensated (Charlesworth and Charlesworth 2000).

Bachtrog (2003) has studied the region around seven genes on the neo-X and neo-Y of D. miranda and also homologous regions in its close relatives (Fig. WN23.10A). She found striking evidence of the harmful effects of lack of recombination. The neo-Y shows about 30-fold less variability than on the neo-X at silent sites and at microsatellites—a much smaller level of polymorphism than expected from the threefold lower number of Y chromosomes, relative to X, in the population. The neo-Y has accumulated large numbers of amino acid substitutions, although the rate is still lower than at silent sites, implying that there has been some selective constraint. However, several frameshift mutations have occurred, and transposable elements have been fixed at several sites (Fig. WN23.10B). This contrasts with the usual situation in Drosophila, where transposons are almost always kept at low frequency by selection (p. 594). At five of the genes, there was little amino acid divergence on the neo-X, implying that the substitutions on the homologous Y-linked loci are due to the failure of selection to hold back deleterious mutations. However, at two loci (exu1 and cycB), there is an exceptionally high rate of amino acid substitution on the neo-X. This is accompanied by a high level of divergence relative to polymorphism (i.e., a significant McDonald–Kreitman test; Box 19.1), a significant excess of rare variants compared with that expected under neutrality (i.e., a significant Tajima’s D; see Chapter 19 Web notes) and a reduced level of synonymous divergence in the surrounding region (Fig. WN23.10C). All this suggests that positive selection has caused rapid divergence of the neo-X at these two loci. The neo-Y showed no such pattern, either because selection only acted in females or because positive selection was ineffective in the absence of recombination. Overall, these data give very strong evidence for the action of hitchhiking involving both positive and negative selection.

In humans, the Y chromosome is much older, having emerged approximately 300 Mya, when recombination was suppressed in the region around SRY. It includes a small pseudoautosomal region, in which the X and Y chromosomes recombine freely; the remaining male-specific region consists of 23 Mb of euchromatin, which encodes only 27 distinct proteins or protein families. Several distinct regions can be identified, which differ in the divergence between homologous genes on the X and Y (Fig. WN23.11A). These correspond to the suppression of recombination over ever-larger regions, probably through large inversions. The degree of divergence dates the times when these inversions suppressed recombination (Fig. WN23.11B). Thus, the pseudoautosomal region has become progressively smaller. However, about 80–130 Mya, a translocation from an autosome increased the size of the pseudoautosomal region (Lahn and Page 1999; Fig. WN23.11B).

We saw on page 426 that the level of single-nucleotide polymorphism variation on the Y is about one-third that on the X, which is just as expected under the standard neutral theory. This contrasts with the drastic reduction in polymorphism seen on the D. miranda neo-Y. The likely explanation is that the human Y has lost almost all gene expression, is therefore not influenced by selection, and so is not expected to show the kinds of hitchhiking effect that are seen in the younger Y chromosome of D. miranda.

The slow degeneration of Y chromosomes is caused by their lack of recombination, although we do not know whether their decline is due to Muller’s ratchet, selective sweeps, background selection, or some combination of these processes. The evidence from Y chromosomes, combined with the strong theoretical predictions discussed above, makes it hard to understand how apparently asexual lineages do sometimes survive for long periods. For example, mitochondrial and choloroplast genomes are transmitted maternally (or, sometimes, paternally) and so cannot recombine. (There is some evidence for rare recombination in mtDNA [Eyre-Walker et al. 1999; Ladoukakis and Zouros 2001]. However, it is not clear whether this reflects sequencing errors or genuine recombination [Macaulay et al. 1999; Innan and Nordborg, 2002].) Over very long times, these genomes have degenerated. At the origin of eukaryotes, organelles began as separate organisms with intact genomes, but now almost all their functions have been transferred to the nuclear genome (p. 205). Sequence variation also provides evidence for the slow degeneration of mitochondrial genomes. Mitochondrial transfer RNAs are less stable and more variable than their nuclear counterparts, and mitochondria show a faster rate of amino acid substitution relative to silent changes (Lynch 1997; Lynch and Blanchard 1998; but see Ballard and Whitlock 2004). Both patterns may reflect less effective selection against slightly deleterious mutations. Nevertheless, some functional mitochondrial genes have been retained—for example, mammalian mitochondrial DNA encodes 22 transfer RNAs, 2 ribosomal RNAs, and 13 protein components of the electron transport chain. Examples such as these and the bdelloid rotifers discussed above are puzzling. However, they are exceptional. The overall phylogenetic distribution shows that asexual taxa do almost always soon go extinct on an evolutionary timescale (Fig. WN23.8).