L ETTERS 387 Copyright © 1998 Elsevier Science Ltd. All rights reserved. 0168-9525/98/$19.00 PII: S0168-9525(98)01529-7 TIG OCTOBER 1998 VOL. 14 NO. 10 at the S–G2 boundary 7 . Thus, if CenpA-containing nucleosomes are deposited on freshly replicated DNA, as are H3-containing nucleosomes, CenpA becomes a marker for very late replicating DNA, and so the exclusive presence of CenpA at mammalian centromeres, regardless of sequence composition, would be evidence that the last region of a human chromosome to replicate becomes the centromere. Although it has been speculated that CenpA is ‘targeted’ to centromeres in a sequence-dependent manner, we note that when CenpA transcription is driven by an S-phase-specific histone H3 promoter, CenpA protein becomes incorporated throughout the chromosomes 7 . This observation suggests to us that CenpA-containing nucleosomes are not targeted, but are simply deposited at freshly replicated DNA. A similar situation might hold for the meiosis-specific histone H1 homolog of lily, which shows accumulation in premeiotic G2 and apparently specific localization to centromeres 8 . Passive incorporation of specific histones at centromeres provides one molecular mechanism whereby the last-to-replicate model can explain centromere identity: as we had proposed, the highest density of bound CenpA might determine the mammalian centromere. We had also pointed out that that last-to-replicate model was ruled out for Saccharomyces cerevisiae centromeres, which replicate early 9 . Additionally, the occurrence of holocentric chromosomes in some higher eukaryotic lineages is similarly inconsistent with this model. Therefore, Schubert’s point, that the last-to-replicate model does not apply to all higher eukaryotic centromeres, is not one that we disagree with. Rather, we had argued that the repeat-based system is ancestral, perhaps deriving from the bacterial replication terminus region. By our model, exceptional examples arose subsequently. For instance, S. cerevisiae might have evolved a specialized centromere sequence that allowed it to dispense with long repeat arrays and evolve a more compact genome. Holocentric chromosomes might have arisen because of difficulties in maintaining regions of very late replication in the face of rampant transposition. Similarly, the existence of apparent exceptions cited by Schubert might imply that other mechanisms for centromere identity have evolved. We look forward to detailed molecular studies on these potentially exceptional systems. References 1 Choo, K.H. (1998) Nat. Genet. 18, 3–4 2 Willard, H.F. (1998) Curr. Opin. Genet. Dev. 8, 219–225 3 Clarke, L. (1998) Curr. Opin. Genet. Dev. 8, 212–218 4 Wiens, G.R. and Sorger, P.K. (1998) Cell 93, 313–316 5 Murphy, T.D. and Karpen, G.H. (1998) Cell 93, 317–320 6 Palmer, D.K. et al. (1987) J. Cell Biol. 104, 805–815 7 Shelby, R.D., Vafa, O. and Sullivan, K.F. (1997) J. Cell Biol. 136, 501–513 8 Suzuki, T., Ide, N. and Tanaka, I. (1997) Chromosoma 106, 435–445 9 McCarroll, R.M. and Fangman, W.L. (1988) Cell 54, 505–513 Steven Henikoff steveh@muller.fhcrc.org Howard Hughes Medical Institute, Fred Hutchinson Cancer Research Center, Seattle, WA 98109-1024, USA. Amy K. Csink csink@andrew.cmu.edu Division of Biological Sciences, Carnegie Mellon University, Pittsburgh, PA 15213-2683, USA. One of the most puzzling features of the commonest form of diabetes (Type 2 or non-insulin dependent diabetes) is the large differences in the prevalence of the disease in different countries and even different parts of the same country. Rates tend to be lower in populations that have retained a traditional lifestyle, for example in rural Africa, where the disease occurs in 1–2% of adults. The prevalence in European and white North American adults is typically around 5% (Ref. 1). However, in populations exposed to rapid Westernization, for example, Asian immigrants to the UK, the prevalence is much higher than in Europeans. This trend is most marked in New World and Pacific communities, where the disease has become epidemic. In the Micronesian island of Nauru, for example, Type 2 diabetes has increased to a level above 30% in adults aged 30–64 years 1 , since the end of the Second World War. Although the high level of Type 2 diabetes in these people is linked to increased obesity, change in diet and adoption of a sedentary lifestyle (caused in Nauru by the sudden affluence brought about by exploitation of mining resources), these factors cannot explain the high disease incidence. To explain the dramatic increase in Type 2 diabetes in these populations, the geneticist James Neel advanced the ‘thrifty genotype’ hypothesis. He suggested that the genotype conferring susceptibility to Type 2 diabetes arose in human evolutionary history as a selective advantage in periods of uncertain food supply. Thus, people with the Does a common mitochondrial DNA polymorphism underlie susceptibility to diabetes and the thrifty genotype?