CODON USAGE IN HIGHLY EXPRESSED GENES OF HAEMOPHILLUS-INFLUENZAE AND MYCOBACTERIUM-TUBERCULOSIS - TRANSLATIONAL SELECTION VERSUS MUTATIONALBIAS

Biases in the codon usage and base compositions at three codon sites i n different genes of A + T-rich Gram-negative bacterium Haemophillus i nfluenzae and G + C-rich Gram-positive bacterium Mycobacterium tubercu losis have been examined to address the following questions: (1) wheth er the synonymous codon usage in organisms having highly skewed base c ompositions is totally dictated by the mutational bias as reported pre viously (Sharp, P.M., Devine, K.M., 1989. Codon usage and gene express ion level in Dictyostelium discoideum: highly expressed genes do 'pref er' optimal codons. Nucleic Acids Res. 17, 5029-5039), or is also cont rolled by translational selection; (2) whether preference of G in the first codon positions by highly expressed genes, as reported in Escher ichia coli (Gutierrez, G., Marquez, L., Marin, A., 1996. Preference fo r guanosine at first codon position in highly expressed Escherichia co li genes. A relationship with translational efficiency. Nucleic Acids Res. 24, 2525-2527), is true in other bacteria; and (3) whether the us age of bases in three codon positions is species-specific. Result pres ented here show that even in organisms with high mutational bias, tran slational selection plays an important role in dictating the synonymou s codon usage, though the set of optimal codons is chosen in accordanc e with the mutational pressure. The frequencies of G-starting codons a re positively correlated to the level of expression of genes, as estim ated by their Codon Adaptation Index (CAI) values, in M. tuberculosis as well as in ii influenzae in spite of having an A + T-rich genome. T he present study on the codon preferences of two organisms with opposi tely skewed base compositions thus suggests that the preference of G-s tarting codons by highly expressed genes might be a general feature of bacteria, irrespective of their overall G + C contents. The ranges of variations in the frequencies of individual bases at the first and se cond codon positions of genes of both H. influenzae and M. tuberculosi s are similar to those of E. coli, implying that though the compositio n of all three codon positions is governed by a selection-mutation bal ance, the mutational pressure has little influence in the choice of ba ses at the first two codon positions, even in organisms with highly bi ased base compositions. (C) 1998 Elsevier Science B.V. All rights rese rved.