Gene splice sites correlate with nucleosome positions

Abstract

Gene sequences in the vicinity of splice sites are found to possess dinucleotide periodicities, especially RR and YY, with the period close to the pitch of nucleosome DNA. This confirms previously reported finding about preferential positioning of splice junctions within the nucleosomes. The RR and YY dinucleotides oscillate counterphase, i.e., their respective preferred positions are shifted about half-period one from another, as it was observed earlier for AA and TT dinucleotides. Species specificity of nucleosome positioning DNA pattern is indicated by predominant use of the periodical GG(CC) dinucleotides in human and mouse genes, as opposed to predominant AA(TT) dinucleotides in Arabidopsis and C.elegans.

Nucleosomes are well positioned in exons and carry characteristic histone modifications

“…However, knowledge of how nucleosomes and histone modifications are distributed relative to other genomic features could result in valuable and novel insights into other biological processes, such as DNA replication and mRNA splicing (Allemand et al 2008;Bres et al 2008;Gottipati and Helleday 2009). It has been previously suggested that intron/exon junctions contain DNA sequences that promote nucleosome positioning (Beckmann and Trifonov 1991;Baldi et al 1996;Kogan and Trifonov 2005). Positioned nucleosomes at these junctions were proposed to protect splice sites at exon starts from mutations (Kogan and Trifonov 2005).…”

“…It has been previously suggested that intron/exon junctions contain DNA sequences that promote nucleosome positioning (Beckmann and Trifonov 1991;Baldi et al 1996;Kogan and Trifonov 2005). Positioned nucleosomes at these junctions were proposed to protect splice sites at exon starts from mutations (Kogan and Trifonov 2005).…”

“…Hence, this positioning seems almost independent of transcriptional activity, suggesting that it can be directed by exon length and/or sequence signals (Baldi et al 1996), but not by RNAPII. Previous sequence analyses indicate that some periodic sequences are more common within exons than in flanking introns (Baldi et al 1996;Kogan and Trifonov 2005). Especially relevant was the observation of dinucleotide periodicities indicative of nucleosome positioning within internal exons.…”

A method for extracting period~10 modulations from DNA sequence correlations applied to the Drosophila melanogaster genome

“…Correlation measures of DNA sequences have been widely studied with the objective of understanding how the genomes of various species are organized. Fourier analysis (Yin and Yau, 2007), dinucleotide frequencies (Kogan and Trifonov, 2005;Cohanim et al, 2006a,b;Kogan et al, 2006), random-walk variances (Peng et al, 1992;Buldyrev et al, 1995;Te Boekhorst et al, 2008), and wavelet analysis (Arneodo et al, 1998) are some examples of related statistical measures. Herzel and co-workers used binary correlation functions and constructed a dependence matrix that would count all the statistical dependences between all nucleotides Groβe, 1995, 1997;Herzel et al, 1998).…”

“…In applying this covariance matrix to human chromosomal regions, for example, long-range correlations in human DNA have been observed, correlated to G + C distributions or isochoric structure of genomes (Carpena et al, 2007;Oliver et al, 2008), as well as chromatin structure (Audit et al, 2001); periodicities of 3 and 10-11 bases have been verified in yeast and bacterial DNA. Many authors have considered dinucleotide frequency distributions rather than nucleotide correlations (Kogan and Trifonov, 2005;Cohanim et al, 2006a,b;Kogan et al, 2006). In a recent study, we were confronted with the problem of calculating correlations along DNA sequences using a representation of the fournucleotide set as a tetrahedron in 3-D (Silverman and Linsker, 1986;Coward, 1997;Licinio and Caligiorne, 2004;Licinio and Guerra, 2007).…”