Genes VII

The lack of detectable variation when the sequence of a protein is analyzed demonstrates that protein synthesis must be extremely accurate. Very few mistakes are apparent in the form of substitutions of one amino acid for another. There are two stages in protein synthesis at which errors might be made:

• Charging a tRNA only with its correct amino acid clearly is critical. This is a function of the aminoacyl-tRNA synthetase. Probably the error rate varies with the particular enzyme, but generally mistakes occur in <1/10⁵ aminoacylations.
• The specificity of codon-anticodon recognition is crucial, but puzzling. Although binding constants vary with the individual codon-anticodon reaction, the specificity is always much too low to provide an error rate of <10 ^V5. When free in solution, tRNAs bind to their trinucleotide codon sequences only relatively weakly. Related, but erroneous triplets (with two correct bases out of three) are recognized 10 ^V1 V10 ^V2 times as efficiently as the correct triplets.

Codon-anticodon base pairing therefore seems to be a weak point in the accuracy of translation. The ribosome has an important role in controlling the specificity of this interaction, functioning directly or indirectly as a "proofreader," to distinguish correct and incorrect codon-anticodon pairs, and thus amplifying the rather modest intrinsic difference. And in addition to the role of the ribosome itself, the factors that place initiator- and aminoacyl-tRNAs in the ribosome also may influence the pairing reaction.

So there must be some mechanism for stabilizing the correct aminoacyl-tRNA, allowing its amino acid to be accepted as a substrate for receipt of the polypeptide chain; contacts with an incorrect aminoacyl-tRNA must be rapidly broken, so that the complex leaves without reacting. Suppose that there is no specificity in the initial collision between the aminoacyl-tRNA PEF-Tu PGTP complex and the ribosome. If any complex, irrespective of its tRNA, can enter the A site, the number of incorrect entries must far exceed the number of correct entries. How does a ribosome assess the codon-anticodon reaction in the A site to determine whether a proper fit has been achieved?

The ability of the ribosome to influence the accuracy of translation was first shown by the effects of mutations that confer resistance to streptomycin. Streptomycin inhibits protein synthesis by binding to the 16S rRNA and inhibiting the ability of EG-G to catalyze translocation. Another effect of streptomycin is to increase the level of misreading of the pyrimidines U and C (usually one is mistaken for the other, occasionally for A). The site at which streptomycin acts is influenced by the S12 protein; the sequence of this protein is altered in resistant mutants. Ribosomes with an S12 protein derived from resistant bacteria show a reduction in the level of misreading compared with wild-type ribosomes. This compensates for the effect of streptomycin on misreading. S12 stabilizes the structure of the 16S rRNA in the region that is bound by streptomycin. The important point to note here is that the P/A site region influences the accuracy of translation: translation can be made more or less accurate by changing the structure of 16S rRNA. The combination of the effects of the S12 protein and streptomycin on the rRNA structure explains the behavior of different mutants in S12, some of which even make the ribosome dependent on the presence of streptomycin for correct translation (1185).

How does ribosome structure influence accuracy? Proofreading processes scrutinize the aminoacyl-tRNA in the A site. The overall error rate in protein synthesis is ~5 10 ^V4 per codon, and the majority of errors probably occur by recognition of mistaken aminoacyl-tRNAs in the A site. Mismatched aminoacyl-tRNA dissociates more rapidly than correctly matched aminoacyl-tRNA, probably by a factor of ~5 . So Increasing time spent in the A site gives a mismatched aminoacyl-tRNA more time to escape before peptide bond formation occurs. This kinetic proofreading increases the probability that the correct aminoacyl-tRNA will be utilized. The effect of the ribosome on accuracy may be an indirect result of an effect on the speed of peptide bond formation. Slowing this step gives more time to correct errors (for review see 437).

The cost of protein synthesis in terms of high-energy bonds may be increased by proofreading processes. An important question in calculating the cost of protein synthesis is the stage at which the decision is taken on whether to accept a tRNA. If a decision occurs immediately to release an aminoacyl-tRNA PEF-Tu PGTP complex, there is little extra cost for rejecting the large number of incorrect tRNAs that are likely (statistically) to enter the A site before the correct tRNA is recognized. But if the GTP is hydrolyzed when the complex binds, an additional high-energy bond will be cleaved for every incorrectly associating tRNA. This would increase the cost of protein synthesis well above the three high-energy bonds that are used in adding every (correct) amino acid to the chain. There is some evidence that the use of GTP in vivo is greater than had been expected, possibly involving an extra 3 GTP cleavages per amino acid.

The specificity of decoding has been assumed to reside with the ribosome itself, but some recent results suggest that translation factors influence the process at both the P site and A site.

A striking case concerns initiation. Mutation of the AUG initiation codon to UUG in the yeast gene HIS4 prevents initiation. Extragenic suppressor mutations can be found that allow protein synthesis to be initiated at the mutant UUG codon. Two of these suppressors prove to be in genes coding for the α and β subunits of eIF2, the factor that binds Met-tRNA_i to the P site. The mutation in eIFβ2 resides in a part of the protein that is almost certainly involved in binding nucleic acid. It seems likely that its target is either the initiation sequence of mRNA as such or the base-paired association between the mRNA codon and tRNA_i^Met anticodon. This suggests that eIF2 participates in the discrimination of initiation codons as well as bringing the initiator tRNA to the P site.

An indication that EF-Tu is involved in maintaining the reading frame is provided by mutants of the factor that suppress frameshifting (see next section). This probably means that EF-Tu does not merely bring aminoacyl-tRNA to the A site, but also is involved in positioning the incoming aminoacyl-tRNA relative to the peptidyl-tRNA in the P site.

This section updated 10-4-2000