Mismatch repair and mutational bias in microsatellite DNA.

Abstract

Although gene numbers do not seem to vary much even between distantly related species, there are huge differences in total genome size. Known as the C-value paradox, this suggests that mutational or selective forces operating on genome size must be quite different between species. However, despite progress recently made in the sequencing of complete genomes, the enigma of eukaryotic genome size variation remains. This is partiularly so for the class of repetitive DNA formed by simple repeats, microsatellites. These repeats are distributed ubiquitously in eukaryotic genomes, but their numbers and the proportion of the genome that they occupy vary significantly between species. Why?Most microsatellites are thought to be selectively neutral, so the process of mutation might be considered more important for determining the length and distribution of simple repeats. A recent study by Harr et al. [1xMismatch repair-driven mutational bias in D. melanogaster. Harr, B. et al. Mol. Cell. 2002; 10: 199–205Abstract | Full Text | Full Text PDF | PubMed | Scopus (40)See all References][1] now offers a mechanistic explanation of how a mutational bias might operate to govern microsatellite length. They studied microsatellite mutations in Drosophila melanogaster, a species that was previously shown to carry unusually short simple repeat sequences. In the spellchecker mutation accumulation lines deficient for mismatch repair (spel1−/−), microsatellite mutations tended to lead to an increase in repeat number, indicating a slightly upward mutation bias. By contrast, wild-type flies showed a significant downward bias, with deletions involving more than one repeat unit, in particular, being more common in mismatch repair-proficient cells.These observations led to the hypothesis that, although primary microsatellite mutations in D. melanogaster are often in the form of repeat unit insertions, the mismatch-repair machinery preferentially recognizes and/or corrects expansion mutations to give a net loss of repeat units (summed over loci). It is well established that mismatch repair is crucial for correcting replication slippage mutations in microsatellite DNA (the mutation rate is elevated by several orders of magnitude in spel1−/− lines), however, this would suggest that mismatch repair does not treat all primary mutations equally and thereby introduces a mutation bias. Importantly, a downward bias is consistent with microsatellites in D. melanogaster being short. Moreover, an extension of this hypothesis is to postulate that slight differences in the mismatch repair specificity and/or function between species can have profound effects on the genomic length distribution of microsatellite sequences, by affecting the strength and direction of a mutation bias.Further support for a mismatch repair-driven mutational bias on microsatellites in D. melanogaster is provided by a comparison of different types of repeats. The new study found that (AT)n sequences mutate at a higher rate than (GT)n repeats in spel1−/− lines. However, population data on levels of polymorphism suggests that the mutation rate of (GT)n exceeds that of (AT)n. These seemingly contradictory observations can potentially be explained by the idea that although the primary mutation rate of (AT)n is higher than that of (GT)n, mismatch repair might be more efficient in correcting (AT)n mutations than (GT)n mutations. Such a mutational bias would obviously affect the genomic distribution of different types of repeats.It would be of great interest, and importance, to know whether mismatch repair-associated mutation bias also affects features of genome composition such as GC content and the overall molecular evolution of non-repetitive sequences. Although entirely speculative, Harr et al. note that the AT content of D. melanogaster is high and put this in connection with the above-mentioned idea of improved repair of AT sequences. Certainly, this is a question worthy of further investigation.