Letters

Speciation and Centromere Evolution

See allHide authors and affiliations

Science  21 Dec 2001:
Vol. 294, Issue 5551, pp. 2478-2480
DOI: 10.1126/science.294.5551.2478

A new model for the origin of species is proposed by S. Henikoff and co-authors in their review “The centromere paradox: Stable inheritance with rapidly evolving DNA” (special issue on Epigenetics, 10 Aug., p. 1098). Referring to the research in Drosophila to illustrate their idea, Henikoff et al. suggest that concerted evolution of centromeric satellite DNA and the centromeric histone protein centromere identifier (Cid) in isolated populations should result in a loss of compatibility between these elements in the hybrids. This should lead to chromosome nondisjunction (the failure of homologous chromosomes to segregate properly during meiosis) in the hybrids, and their sterility. Therefore, “speciation is an inevitable consequence of centromere evolution.”

The authors suggest several tests for their model, but there is a simple test that should be definitive. If the model is correct, then the genes for hybrid sterility must be located predominantly at centromeres or the sites of Cids (or both). Alas, the mapping data from Drosophila and mouse indicate that they are not.

In formulating their model, the authors appear to have overestimated the role of chromosome nondisjunction in hybrid sterility. This type of malfunction of the chromosome segregation machinery accounts for only a minor part of gametogenic failure in hybrid animals. The major causes are the inability of hybrid germ cells to enter meiosis or to proceed through meiotic prophase. Another cause of hybrid sterility is the formation of nonfunctional gametes resulting from disruptions of the process of postmeiotic modifications (1). None of these events is associated with chromosome nondisjunction. Thus, apparently DNA satellite/Cid divergence takes place after the speciation event and is a consequence rather than a cause of speciation.

References and Notes

Response

Borodin raises the important issue of the connection between hybrid sterility and sex chromosome nondisjunction that we discussed in our review. He points out that nondisjunction is not expected to result in spermatogenic failure. However, the actual situation is more complex. In Drosophila, nondisjunction of the XY pair caused by alterations in centromeric X heterochromatin is correlated with failure of sperm development (1), and failure occurs whether or not the X and Y chromosomes had been paired at meiotic prophase (2). To explain these puzzling observations, McKee and colleagues (3) have proposed the existence of a meiotic checkpoint that aborts spermatogenesis when proper orientation of the centromeres is not achieved. Metaphase checkpoints that monitor tension on spermatocyte kinetochores (the attachment point of microtubules to the centromere) are well known (4), and unequal pulling on X and Y centromeres should trigger such a checkpoint. The operation of an early checkpoint is consistent with postmeiotic defects, because failures in the early steps of meiosis can cause sperm dysfunction at much later stages of gamete development (5). Therefore, both nondisjunction and sperm dysfunction would result from unequal pulling on X and Y centromeres.

Borodin also questions the agreement of our model with current mapping data. However, in established species, such as those cited by Borodin, mapping of a locus involved in an initial speciation event is complicated by the accumulation of complex secondary hybrid incompatibilities (6). These secondary incompatibilities would be indistinguishable from the primary one. To map primary speciation loci, one should use incipient species, such as the 150,000-year-old Bogota population of Drosophila pseudoobscura (6). In this case, very few incompatibilities were detected. As Borodin suggests, we expect that an initial hybrid sterility determinant should map to the cid gene encoding the centromeric histone, or to another gene that alleviates unequal pulling on centromeres. Candidate proteins that might alleviate this imbalance include the Drosophila DNA binding protein Prod, which binds to a species-specific satellite DNA during mitosis (7). Another is the protein encoded by the Drosophila mauritiana hybrid sterility gene, Odysseus, which might affect centromere imbalances by binding to satellite DNA through its adaptively evolving homeobox sequence (8). Because such proteins from one species will become incompatible with multiple centromeres from the other species, we do not necessarily expect that hybrid sterility will map to any particular centromere. Therefore, the simplest test of our model is the examination of Cid and other candidate proteins in incipient species.

References and Notes

Stay Connected to Science

Navigate This Article