Conservation of non-consensus nucleotides in transcription factor binding sites

Abstract

Transcription factors (TFs) affect gene expression by binding to their sites (TFBSs) in the genome. As regulation of transcription is essential for the optimization of metabolism, evolution of TFBSs is thought to be restricted by the TF binding preferences. This means that nucleotides important for binding (consensus nucleotides, CNs) experience the pressure of negative selection and, therefore, are highly conserved. However, for some regulated genes, the strongest possible repression or activation may not necessarily be optimal (i.e. provide the highest fitness). Here, we show that, along with CNs, nucleotides causing relatively weaker binding (non-consensus nucleotides, NCNs) are sometimes preserved by negative selection. We then consider several possible reasons for the conservation of NCNs and demonstrate that NCNs depend epistatically on other loci of the same TFBS for a substantial fraction of TFBSs.