Translational Accuracy

A. Types of Potential Errors
Given the complexity of protein synthesis, there are many steps at which mistakes can occur, and it is remarkable that the error rate is as low as it is. The two principal categories of translational mistakes are missense errors, in which one amino acid is substituted for another or a stop codon is not recognized, and errors of processivity, in which frameshifting or premature termination occur.

Missense errors can take place because of misacylation of a tRNA by an AARS, resulting in insertion of an incorrect amino acid despite a correct codon–anticodon match. This potential problem is counteracted by active editing by AARSs (see earlier), which is critical for maintaining translational accuracy. Errors can also occur at the level of the codon–anticodon interaction, such that a noncognate AA–tRNA is selected at the ribosome, again resulting in amino acid misincorporation. The relative abundance of cognate vs. noncognate AA–tRNA affects the accuracy of amino acid incorporation. In particular when the next amino acid specified by the mRNA is unavailable, the frequency of missense errors increases. A special case of incorrect AA–tRNA selection is that of stop codon readthrough, in which an amino acid is inserted where the protein should actually be terminated. This type of error results in a carboxy-terminal extension (added to the nascent protein) that continues until the ribosome reaches the next stop signal.

Frameshift mutations occur because of a nucleotide insertion or deletion in a protein’s gene. Errors in reading frame also occur when the ribosome “slips” along the mRNA in such a way that the sequence is not read in triplets corresponding to the codons of the message, but the ribosomemoves two or four nucleotides instead. These “slips” are −1 or +1 frameshifts, respectively. Fortunately the ribosome usually encounters a stop codon shortly after the frameshift; this minimizes the effect of the mutation. (Although frameshifting can be the unintended result of translational inaccuracy, “programmed” frameshifts also take place in special situations. These produce alternate polypeptides from a single mRNA.) Premature termination of a protein can occur when a nucleotide substitution produces a stop codon in the middle of the gene sequence. The peptidyl–tRNAmay also dissociate prematurely from the ribosome before the stop codon is reached.

B. Suppressor tRNAs
For several of these error scenarios, faithful translation of the genetic message is accomplished through the action of suppressor tRNAs. These are the result of advantageous mutations in tRNA genes, typically in the anticodon, that allow them to recognize a misplaced stop codon or unintended frameshift. Suppressor tRNAs may be aminoacylated according to the amino acid specificity of their parent sequence, so therefore they insert this amino acid into the polypeptide at the location of the mRNA aberration. Alternatively, they may be misacylated because of the change in the anticodon sequence. As long as the inserted amino acid is not detrimental to the synthesized protein, a suppressor tRNA rescues the gene mutation. For example the E. coli su6 suppressor tRNA inserts leucine into the growing polypeptide chain at the position of a premature UAA stop codon.

Missense suppressors simply substitute one amino acid for another to correct a mutation in the mRNA. Transfer RNAs that suppress +1 frameshifts typically contain an extra base in the anticodon and read a four-nucleotide codon, thereby restoring the correct translational frame.

Suppressor tRNAs are of particular interest as more examples of the genetic basis for specific diseases are found. In many cases, missense or nonsense mutations in genes for essential proteins correlate with diseases such as cancer, cystic fibrosis, and amyotropic lateral sclerosis (ALS). In fact, mutations in the tumor suppressor gene p53 are the single greatest cause of human cancers. One particular p53 hotspot is at codon 213, where nonsense mutations have been identified in human colorectal, gastric, ovarian, and breast cancers. Suppressor tRNAs designed to selectively rescue these mutations by reading through the premature termination codon could be effective therapeutic agents in a gene therapy approach. Several challenges remain, however—the suppressor tRNA must be introduced into the affected cells, must be transcribed at high enough levels to be effective, must also be aminoacylated at high levels, and must not excessively read through stop codons that specify the normal termination of translation for other genes. These are active areas of research in several laboratories.

C. Selenocysteine Insertion at UGA Codons
The trace element selenium is present in a number of proteins in the form of cotranslationally inserted selenocysteine (Sec). This insertion occurs when a unique selenocysteinyl–tRNASec recognizes an internal UGA codon. Selenocysteine has therefore been called the “21st amino acid” and its insertion at a nucleotide triplet that is normally a termination signal represents an expansion of the genetic code.

Insertion of selenocysteine requires several adaptations of the translational machinery. First, tRNASec is aminoacylated with serine by seryl–tRNA synthetase. The serine attached to this misacylated species is then converted to selenocysteine by the enzyme selenocysteine synthase, which uses selenophosphate as a donor. Elongation factor Tu does not bind Sec–tRNASec as it does other elongator AA–tRNAs; instead, a unique protein SELB transports Sec-tRNASec to the ribosomal A-site. SELB is specific for Sec–tRNASec, rejecting other AA–tRNAs including Ser–tRNASec. The final novel feature of selenocysteine insertion is the mechanism ofmRNArecognition. The UGA triplet can be used in the same organism as either a selenocysteine or a stop codon. The sequence context determines its recognition as a selenocysteine codon. The ternary complex SELB:GTP:Sec–tRNASec recognizes a stem-loop structure immediately 3´ (downstream) from a UGA codon that is read as selenocysteine. This structural feature of the mRNA is specifically bound by the carboxyl-terminal portion of SELB, while other regions of the protein are highly homologous to EF-Tu as expected. Insertion of selenocysteine into polypeptides therefore requires formation of a quaternary SELB:GTP:Sec–tRNASec:mRNA complex. This is in contrast to all other elongation steps in protein synthesis, which proceed through ternary complexes.

D. Degradation of Incomplete Polypeptides
One consequence of the necessary accuracy in protein synthesis is the release of peptidyl–tRNA molecules representing incomplete translation products. These products can be the result of ribosome stalling, a premature stop codon that is not suppressed, or detection by the ribosome of noncognate tRNA present in the decoding center. It has been estimated that this “drop-off” might result in a truncated polypeptide chain approximately 10% as often as the full-length protein. Not only is this a waste of amino acids resulting in useless products but also tRNA molecules are sequestered and unavailable for translation of other genes.

The incomplete peptidyl–tRNAs are substrates in bacteria and yeast for the enzyme peptidyl–tRNA hydrolase, which cleaves the ester bond between the tRNA and its attached polypeptide. Peptidyl–tRNA hydrolase therefore removes the useless (and potentially harmful) protein fragments and recycles the tRNAs. Interestingly, although the initiator tRNA(fMet–tRNAfMet) mimics a peptidyl–tRNA by virtue of its N-formyl group, peptidyl–tRNAhydrolase does not recognize the initiator as a substrate. The hydrolase bypasses fMet–tRNAfMet because of the presence of structural features unique to fMet–tRNAfMet.

E. A Dual-Function RNA
A particular challenge arises when an mRNA lacks inframe stop codons due to deletion or degradation. Bacteria use a unique tRNA–mRNA hybrid (tmRNA, also called 10Sa RNA) to remove the resulting partially synthesized proteins and free the ribosomal subunits. The 5´- and 3´-ends of this molecule resemble alanyl–tRNA, while the central portion encodes a peptide “tag.” Alanyl–tRNA synthetase aminoacylates the tmRNA, which is then transported to the stalled elongation complex by EF-Tu. After attaching alanine to the end of the growing nascent polypeptide, the ribosome switches from the truncated mRNA to the tmRNA in a mechanism called transtranslation. The ribosome adds 10 additional amino acids to the end of the protein according to the tmRNA sequence. The resulting tagged protein is released and degraded, as the tag is a recognition signal for several proteases.

F. mRNA Surveillance
Eukaryotes use a surveillance mechanism to identify mRNAs with mutations (typically premature termination codons) or processing errors (such as incorrect splicing). Once detected, the aberrant messages are degraded to prevent the synthesis of truncated proteins. Recent evidence suggests that these mRNAs must be at least partially translated to determine whether a stop codon is in its proper context. In mammalian systems, the translating ribosome is proposed to measure the distance between the final splicing junction and the termination signal—if they are within 50 nucleotides of one another, termination is allowed to proceed. If they are further apart, either because of a misplaced stop codon within the mRNA reading frame or a splicing error, the mRNA is targeted for rapid degradation. How the ribosome recognizes this distance is not yet known. This mRNA surveillance is also called nonsense-mediated decay because the majority of mutational errors result in premature termination codons.

G. Accuracy Mechanisms
Despite the types of translational errors described above, mRNA-directed protein synthesis is remarkably accurate. How is it that the ribosome and translational factors are able to achieve such faithful transmission of genetic information? One way to describe the specificity of cognate over noncognate AA–tRNAs is termed the “kinetic proofreading” mechanism. One can imagine that selection of an EF-Tu:GTP:AA–tRNA ternary complex by the ribosome during elongation can be considered a “scanning” step. Depending on the codon–anticodon interaction, the AA–tRNA will either bind irreversibly in the ribosomal A-site (in the case of the cognate AA–tRNA), or dissociate from the ribosome before or after GTP hydrolysis (noncognate AA–tRNA). In this model, the rate of EF-Tu-triggered GTP hydrolysis is the same for cognate and noncognate substrates. However, because the cognate AA–tRNA spends more time in complex with the ribosome, GTP hydrolysis is likely to occur prior to dissociation of the tRNA, promoting tight binding of this tRNA in the A-site prior to peptide bond formation.

In addition to passive kinetic proofreading, selection of the correct codon–anticodon interaction is proposed to trigger a conformational change in the ribosome that is transmitted to EF-Tu. Hydrolysis of GTP is then accelerated for the cognate compared to the noncognate substrate. Recent structural evidence shows that structural rearrangements do occur upon binding of the cognate AA–tRNA, suggesting that selection of the correct substrate depends on an induced fit mechanism. How the codon–anticodon interaction is detected by the ribosome is not currently understood.

H. Ribosomal Contributions to Accuracy
Naturally occurring mutations have been identified in ribosomal proteins and rRNA that increase or decrease translational accuracy. These have provided clues as to which regions of the ribosome are involved in proofreading, although detailed mechanisms are not known. The small subunit, which contains the decoding site, has several proteins that are likely involved in controlling missense and processivity errors. For example, mutations have been identified in proteins S4 and S5 that reduce the level of translational fidelity (these are called ribosomal ambiguity mutations). In contrast, S12 mutants increase the accuracy of translation by conferring resistance to streptomycin, an error-causing antibiotic.

The small subunit 16S rRNA also contains regions that appear to be involved in translational accuracy. In particular, substitutions at nucleotides in the so-called 530 loop lead to changes in the error rate of protein synthesis. This rRNA loop is known to be spatially near the proteins S4, S5, and S12, and effects of nucleotide changes in the 530 loop parallel those observed for the proteins. Thus, the rRNA and proteins in this part of the small subunit together act as a proofreading domain. Substitutions at some of these nucleotides increase the rate of missense or frameshift errors, while others are detrimental because they prevent binding of the EF-Tu:GTP:AA–tRNAternary complex. Still other mutations actually increase the accuracy of translation, in that they make the ribosomes resistant to error-inducing antibiotics.

The toxins α-sarcin and ricin inactivate ribosomes by altering a highly conserved sequence in 23S-like rRNAs (the sarcin–ricin loop). Elongation factors Tu and G bind near this loop, suggesting that it is involved in AA–tRNA selection and translocation. Protein L6 and the sarcin– ricin loop are involved in translational accuracy control, as suggested by the mutations in these components that increase fidelity, apparently by slowing down translation to allow more thorough proofreading.