Biophysical Features of Non-Coding Genome Sequences

Abstract

Non coding sequences of DNA represent most of the genome in eukaryotes. Such sequences are complex and probably include features which are essential for chromosome structure, such as chromosome condensation, axis formation, homologous chromosome pairing in meiosis, etc. It is unlikely that coding regions (exons) play a significant role in chromosome structure, since they have evolved to optimize protein synthesis. We are trying to determine by computational methods the presence of frequent sequences, which might play a role in chromosome structure. Such frequent sequences should also be considered as potential targets for drug binding. We have analyzed the occurrence of short words (2-4 nucleotides), as well as the distribution of frequent longer words (9-14 nucleotides) and of microsatellites (long repeats of short words). We have found that some sequences, such as (AT)n and (AG)n, have a distribution in most eukaryotic genomes which suggests a structural role. Analysis of longer words shows the presence of many frequent sequences which contain clusters of purines/pyrimidines such as GGAA, TTT, CCC, etc. We have studied in more detail the genome of Caenorhabditis elegans: we have found words with a similar sequence which punctuate the whole genome and provide structural marks. Surprisingly very few structural data are available on such frequent sequences, as obtained by biophysical methods (x-ray crystallography, NMR): further work is required. Their eventual influence on nucleosome structure should also be established. As an example we present the structure of some (AT)n sequences which are polymorphic and frequently present Hoogsteen instead of the standard Watson-Crick base pairing. Interaction with pentamidine, for example, presents novel features when compared with highly studied sequences such as d(CGCGAATTCGCG). It is worth noting that the latter sequence is not frequent in most genomes.