Mutation and selection explain why eukaryotic centromeric DNA is often A+T rich

The reasons why centromeric DNA is often A+T rich are not understood. We have used chromosome engineering to replace native centromeric DNA with different test sequences at native centromeres in two different strains of the fission yeast Schizosaccharomyces pombe and have discovered that A+T rich DNA, whether synthetic or of bacterial origin, will function as a centromere and that the functionality of the A+T rich DNA does not differ substantially from that of native centromeric DNA. Using genome size as a surrogate for the inverse of effective population size (Ne) we also show that the relative A+T content of centromeric DNA scales with Ne across 45 eukaryotic species. This suggests that in most eukaryotes the A+T content of the centromeric DNA is adaptive and that the A+T content of centromeric DNA is determined by a balance between selection and mutation. Combing the experimental results and the evolutionary analyses allows us to conclude that A+T rich DNA of almost any sequence will function as a centromere in most eukaryotes. The fact that many G/C to A/T substitutions are unlikely to be selected against may contribute to the rapid evolution of centromeric DNA. We also show that neo-centromeres are not simply weak versions of native centromeres but that their establishment or inheritance may require factors in addition to those required by native centromeres. ChIP~seq analysis of GFP-labelled centromeric histone CENP-ACnp1 in different strains of S. pombe that have been modified by genome engineering techniques to investigate the sequence requirments for centromere function.