A slippery boundary

The Y chromosome has provided one of the greatest challenges in finalizing complete mammalian genome sequences in part because of its unusual relationship with the X chromosome. Part of the Y chromosome, known as the pseudoautosomal region, must pair with the complementary region on the X chromosome and undergo recombination, so that the resulting crossovers stabilize the sex chromosomes for proper separation during meiosis. The Y chromosome also bears at least one gene that is male-determining, and this region of the Y chromosome must not recombine with the X chromosome or sterility or intersexuality may result. Apart from these two rules, the gene content of the pseudoautosomal and nonrecombing parts of the Y chromosome are subject to relatively weak evolutionary forces. Iwase et al. (1), in this issue of PNAS, describe a remarkable finding that the boundary between these two portions of the Y chromosome moved relatively recently, and that there appears to be considerable opportunity for chance to play a large role in gene content of these two very different segments of the Y chromosome.

Iwase et al. (1) make a compelling case that the pseudoautosomal boundary (PAB) previously resided in the second intron of the gene encoding amelogenin. To see how this inference could be made solely based on DNA sequence comparisons, it will help to refer to Fig. 1. Any region of the X and Y chromosomes that freely recombines would be expected to show divergence levels that are equal to the level of polymorphism on the X chromosome, or ≈1 bp per 1,000. This is the state of the current pseudoautosomal region, which falls on the right end of the diagram. The far left portion of the diagram indicates the region that has been nonrecombining. Amelogenin arrived onto the sex chromosomes ≈100 million years ago (2), and remained active on both sex chromosomes. Iwase et al. (1) show that sequences from a variety of mammals in this region form a monophyletic clade for the X chromosome and a distinct monophyletic clade for the Y chromosome, suggesting that mammalian species have diversified since this region became a nonrecombining part of the sex chromosomes. This is so because the tree indicates that comparisons among mammals in genes on the nonrecombining Y chromosome indicate greater similarity than any comparison between the X and Y chromosomes. As one moves rightwards in Fig. 1, the genealogy changes, such that some X- and Y-linked genes are closest neighbors on the tree. In this region, the divergence between the X and Y chromosomes drops from ≈30% to ≈10%, and this drop occurs at a transposable element insertion into the second intron of amelogenin. Because of this relatively lower X–Y divergence, Iwase et al. propose that the 3′ region of amelogenin used to freely recombine between the X and Y, but later the PAB moved to the right, leaving all of amelogenin in the nonrecombining region where it is today. To date the time of movement of the pseudoautosomal boundary, Iwase et al. (1) make use of the X vs. Y divergence, and arrive at an estimate of 27–70 million years ago. This is after the mammalian radiation, implying that there may have been more than one change in the pseudoautosomal boundary. Consistent with this, the location of the PAB, and gene content of nonrecombining vs. pseudoautosomal regions, are widely different among mammals (3).

Figure 1.

The location of the PAB is inferred by Iwase et al. to have been in the amelogenin second intron at approximately the time of radiation of mammals. At this time, the 5′ end of amelogenin was nonrecombining, and so the X and Y copies became strongly divergent in sequence (≈30%). The gene genealogy for the 5′ end of amelogenin shows a monophyletic clustering of the X-linked copies among mammals, and a separate monophyletic clustering of the Y-linked copies. The 3′ end of amelogenin recombined up until ≈26–70 million years ago. Before this time, the X and Y chromosomes were recombining, and their sequences were assumed to be homogenized. However, after the PAB moved, recombination ceased, and the X and Y copies of the 3′ end of amelogenin diverged as well. The gene genealogy of the 3′ end of amelogenin thus shows greater similarity of the X and Y copies within each mammalian species. Presently, the PAB is much further to the right, and all of amelogenin is in the nonrecombining part of the sex chromosomes.

These results beg the question what exactly defines the PAB? It is a rather remarkable phenomenon that a pair of chromosomes can freely recombine up to this point, and then beyond this point recombination is absolutely prohibited. It seems to be a common feature of the PAB that a transposable element has inserted there. Although it is plausible that a transposable element insertion could disrupt pairing, and that once pairing is disrupted so is recombination, this is not a very satisfying answer because transposable elements insert in autosomes all of the time, and chromosome pairing and recombination is disrupted in only a minor and local way. It may be useful to consider that it is not only recombination that breaks at the PAB, but that well before recombination occurs, the pairing between the X and Y chromosomes may change drastically at the PAB. In particular, the PAB almost certainly marks the end of the region of synapsis between the sex chromosomes, a condition necessary for normal recombination. Despite our ignorance of what exactly are the features that make a particular part of a chromosome become a PAB, the sequence analysis of Iwase et al. (1) does demonstrate very clearly that it does not require anything particularly unusual. The fact that different mammals have such a diversity of PABs, and that they have moved more than once in our evolutionary history, suggests that there is a large component of chance in the setting of the PAB boundary. This chance aspect leaves open the possibility that it is determined by some form of chromatin remodeling, determined by the proteins and RNAs complexed with the DNA, as well as their methylation and acetylation states.

Although it seems that chance may play a role in the location of the PAB, there are evolutionary principles that do have an impact on the nature of recombination between the X and the Y. In particular, we know that at least some of the X and Y chromosomes must recombine for normal meiosis, and we know that genes involved in male sex determination must not recombine onto the X. These genes include SRY, SOX9, and perhaps several others. At the very least, this implies that the nonrecombining portion of the Y must span these genes. In addition, there is a growing literature on the chromosome distribution of genes with sex-specific effects. In mammals, it appears that testis-specific ESTs tend to be on the X. In Drosophila, testis ESTs exhibit a remarkable deficit on the X chromosome (4). A male-specific favorable mutation would actually go to fixation faster on the X than the Y, leading Rice (5) to predict that male-specific factors ought to cluster on the X rather than avoid the X. On the other hand, genes with male-specific advantageous effects are totally protected from selective effects in females if they occur on the Y chromosome. Consistent with this prediction, the Y chromosome of Drosophila melanogaster shows a remarkably strong bias toward genes that are necessary for male fertility (6, 7). Overall, the simple evolutionary predictions of what sorts of genes ought to be on which sex chromosome do not provide a very satisfying explanation for observed patterns. The reasons probably lie in the fact that there are strong historical effects, that we rarely if ever know all of the pleiotropic effects of mutations (so our assumptions about sex-specific effects are not always accurate), and that chance plays a big role in sex chromosome composition.

The absence of recombination in a segment of the Y chromosome makes it vulnerable to degeneration due to operation of Muller's ratchet. Without recombination serving as a source of template for correcting errors on the nonrecombining Y, mutations accumulate and may go to fixation. The end result is a Y chromosome that has lost most of its initial genetic functions (8, 9). One of the clearest demonstrations of this comes from Drosophila, where fusion of an autosome to the Y chromosome gave rise to a neo-Y chromosome. Its homolog, which segregates with the X chromosome, becomes a neo-X chromosome. Because males do not undergo recombination in Drosophila the neo-Y is immediately thrust into a position of not ever recombining. Recent work by Bachtrog and Charlesworth (10) demonstrate the rapid degeneration of the Drosophila miranda neo-Y chromosome at the molecular sequence level. Y chromosome degeneration is relevant to amelogenin and the wandering PAB because it shows that whatever genes are in the nonrecombing portion of the Y chromosome are placed in jeopardy. Whenever the PAB moves, there must be a period in which the population is segregating for the old and new PAB location. If the new location goes to fixation in the population, by either drift or selective mechanisms, there may follow a time in which the genes that are newly in the nonrecombining portion remain active. The species may run a risk at this time of losing function of these Y-linked genes. If such a gene is vital for survival or fertility, and if loss of the Y-linked expression is not compensated by the X-linked copy, then this arrangement could be strongly deleterious. Dosage compensation could retain the balance of expression of genes in the newly nonrecombing region, and make it less likely that the PAB would wander back to its original location.