Frequency and coverage of trinucleotide repeats in eukaryotes

In the aim to assess whether the tri-repeat shortage reported in vertebrates affects specific motifs, such as those causing neuromuscular diseases in man, we detected approximate di-, tri- and tetra-repeats (STR) longer than 25 bases in human chromosomes 21 and 22, and in some model organisms (M. musculus, D. melanogaster, C elegans, A. thaliana and S. cerevisiae). We found that overall STR are more represented in mouse and in man than in the other organisms. However, tri-repeats are less represented than di- and tetra- in man and mouse, but show intermediate values between di- and tetra- in the other organisms. In man, ACG shows the lowest both frequency and coverage, ATC the highest coverage and AAT the highest frequency. In general, coverage and frequency of tri-repeats are linearly related, except for ACC, ATC, AAG, AGG motifs in man and AAG, AGG in mouse, which exhibit unexpectedly long repeats. Often their copy numbers exceed that found responsible for the dynamic mutations, set at around 40. The shortage in frequency and coverage of tri- vs. di- and tetra-repeats observed in man and mouse can be ascribed to a subset of the remaining tri-repeat motifs, but among them those recognized as dynamically mutable (AAG, AGC and CCG) are not the least represented. Possible constraints in tri-repeat expansion seem to be structural and conserved along the evolutionary scale: a motif-specific relaxation of the relevant controls may be responsible for the occasional expansions found in mouse and man. (C) 2003 Elsevier B.V. All rights reserved.