Accuracy of natural messenger translation: analysis of codon-anticodon recognition in a simplified cell-free system.

A simplified plasmid-directed coupled system [Robakis, N., Cenatiempo, Y., Meza-Basso, L., Brot, N., & Weissbach, H. (1983) Methods Enzymol. 101, 690-706] was used to study the accuracy of natural messenger translation in vitro. In this system, protein synthesis is limited to the formation of the N-terminal di- or tripeptide of the gene product. Such a control is obtained by restricting the supply of aminoacyl-tRNAs in the assay medium to those corresponding specifically to the first two or three triplets in the mRNA coding sequence. We analyzed comparatively the interaction of 6 different codons with their cognate tRNAs and 18 noncognate tRNAs able to recognize triplets differing from the legitimate sequences by one base only. Special attention was paid to the single base errors occurring at the first and second codon positions during ribosomal selection of aminoacyl-tRNA molecules. The noncognate tRNAs were assayed either in the absence of the legitimate tRNAs or under competition conditions. They were chosen so that all the possibilities for misreading any particular base as each of the other three bases could be studied. First, it was mainly observed that translation mistakes can be equally detected in the first and second codon positions; there is no compelling evidence for a most or least accurate site. Second, pyrimidines seem to be read more accurately than purines. In particular, U cannot be read as either C or G, and C can hardly be mistaken for any other base.(ABSTRACT TRUNCATED AT 250 WORDS)