Periodic oscillations of the genomic nucleotide sequences disclose major differences in the way of constructing homologous proteins from different procaryotic species

Abstract

In this study, a set of 80 completely sequenced procaryotic genomes has been analysed by an alignment-free method, namely the expectancy-rectified frequency of bigrams or 2-tuples, representing the 16 combinations of A, T, G, C. It demonstrates that all genomes exhibit periodic oscillations of their nucleotide sequence, with a period close to 11 phosphodiester bonds, and resembling in shape an exponentially dampened sinusoid at the distance from 5 to 49 bonds. Interestingly, the amplitude of nucleotide oscillation (but not the period) can differ drastically from one species to another. I show that these differences are due neither to the (G + C) content, nor to the size of the genome. They are not directly related to phylogeny, since specific genomes from Archaea and Bacteria can display large as well as small amplitudes. I have compared also a set of genes coding for proteins rich in alpha helical structures (as determined by X-ray diffraction) with a set of genes coding for proteins devoid of alpha helices. The first set has periodic oscillations of large amplitude, with an 11-bond period, while the second has none. Furthermore, I analysed a large number of sets of homologous genes from several different species. They exhibit very different amplitudes of oscillations. Altogether, the data with their statistical analyses strongly suggest that the nucleotide oscillations are due to the ‘genomic style of proteins’, which means that homologous proteins, having the same biochemical function in different organisms, may have different secondary structures or may use different ways to be constructed. I realize that this idea is a heterodox one, but I believe that it can shed a new light both on phylogenies and on constraints between proteins and their coding sequences. To cite this article: P.P. Slonimski, C. R. Biologies 330 (2007).