In vitro systems for protein synthesis have been in wide use for about 10 years. In most of the early work protein synthesis was measured by following the incorporation of radioactive amino acids into acid precipitable material. This test cannot distinguish between the synthesis of complete, active proteins and various sorts of aberrant products which might result from incorrect initiation, translation in the wrong reading frame, amino acid substitutions, improper termination, or other defects in in vitro translation. The first unambiguous demonstrations of biologically meaningful in vitro protein synthesis were obtained with hemoglobin and the coat proteins of the RNA bacteriophages, where it was possible to show good correspondence between the peptide patterns from in vitro and natural products. These results were very encouraging but did not establish whether the fidelity of the in vitro translation process could be great enough to permit the synthesis of active enzymes in vitro.
Recent progress in elucidation of 5' stimulatory elements for translational recoding is reviewed. A 5' Shine-Dalgarno sequence increases both +1 and -1 frameshift efficiency in several genes; examples cited include the E. coli prfB gene encoding release factor 2 and the dnaX gene encoding the gamma and tau subunits of DNA polymerase III holoenzyme. The spacing between the Shine-Dalgarno sequence and the shift site is critical in both the +1 and -1 frameshift cassettes; however, the optimal spacing is quite different in the two cases. A frameshift in a mammalian chromosomal gene, ornithine decarboxylase antizyme, has recently been reported; 5' sequences have been shown to be vital for this frameshift event. Escherichia coli bacteriophage T4 gene 60 encodes a subunit of its type II DNA topoisomerase. The mature gene 60 mRNA contains an internal 50 nucleotide region that appears to be bypassed during translation. A 16 amino acid domain of the nascent peptide is necessary for this bypass to occur.
Previous work of others reported an untranslated stretch of 12 nucleotides in the 5' coding sequence of carA from Pseudomonas aeruginosa. However, N-terminal protein sequencing of carA-lacZ translational fusions shows that these 12 nucleotides are normally translated in a continuous triplet manner, both in P. aeruginosa and in Escherichia coli.
Gene 10 of bacteriophage T7, which encodes the most abundant capsid protein, has two products: a major product, 10A (36 kDa), and a minor product, 10B (41 kDa). 10B is produced by frameshifting into the -1 frame near the end of the 10A coding frame and is incorporated into the capsid. The frameshift occurs at a frequency of about 10% and is conserved in bacteriophage T3. This study shows that sequences important to frameshifting include the originally proposed frameshift site, consisting of overlapping phenylalanine codons and the 3' noncoding region that includes the transcriptional terminator over 200 bases downstream of the frameshift site. The frameshift occurs at the overlapping phenylalanine codons as determined from peptide sequencing data. Complementation studies show that there is only a very weak phenotype associated with phage infections in which there is no 10A frameshifting. Capsids from such infections are devoid of 10B and are as stable as wild-type capsids.
