Correlations Between mRNA Expression Levels and GC Contents of Coding and Untranslated Regions of Genes in Rodents

Abstract.

Gene expression is regulated by a highly coordinated network of events whose efficiency may constrain the level of expression. Among other factors, natural selection for increased translational efficiency and/or fidelity may shape nucleotide composition and, hence, codon usage during evolution. Previous studies have shown that highly expressed genes in Saccharomyces cerevisiae, Caenorhabditis elegans, and Drosophila melanogaster have relatively higher codon usage biases. However, in the case of mammals, results have been equivocal. In this study, we assessed the correlation between nucleotide composition and mRNA expression levels of rodent genes measured by cDNA microarray and serial analysis of gene expression (SAGE) techniques. We found that mRNA expression levels were correlated with the third nucleotide position GC (GC3) content for both Rattus norvegicus (r= 0.246, p= 0.01; N= 110) and Mus musculus (r= 0.21, p= 0.0026; N= 203) genes. However, no significant correlation was evident between mRNA expression level and GC contents of 5′- and 3′-untranslated regions (UTRs) for either species. This suggests that, in rodents, nucleotide composition of coding sequences and UTRs might evolve differentially when considered along an expression gradient. Accordingly, it is possible that higher GC levels may present the rodent genes with a selective advantage for translational efficiency. However, the increase in GC3 content seems to level off above an expressional threshold (e.g., ≥threefold the median expression for R. norvegicus), suggesting that conflicting demands posed by different aspects of transcriptional and translational machineries (e.g., efficiency versus fidelity) may set an upper limit for GC3.