Universal constraint on evolution of all coding sequences

Abstract

What often escapes notice is the inherent paradox in the role of natural selection in evolution. Inasmuch as a coding sequence encoding a functionless protein shall be ignored by natural selection, the initial acquisition of a function by a protein has to be intrinsic in the construction principle of coding sequences. The very fact that many proteins of very divergent functions tend to have rather similar amino acid compositions suggests that there must be a universal rule in the construction of all coding sequences, and this rule was previously defined as the TA/CG-deficiency—TG/CT-excess rule. When manifesting itself in coding sequences of rather balanced base composition, this rule establishes C T G as one of the most numerous base trimers. While Leu is the major residue of most proteins; C T G is the most frequently utilized of 6 Leu codons in E. coli to man. Yet, the two most common base tetramers, G C T G and C C T G, need not be translated in their second reading frame to yield Leu. Their preferential translation in the first reading frame increases Ala and Pro at the expense of Leu, while that in the third reading frame yielded extremely Cys-rich metallothionein.