Marilyn Kozak

Point mutations define a sequence flanking the AUG initiator codon that modulates translation by eukaryotic ribosomes

By analyzing the effects of single base substitutions around the ATG initiator codon in a cloned preproinsulin gene, I have identified ACCATGG as the optimal sequence for initiation by eukaryotic ribosomes. Mutations within that sequence modulate the yield of proinsulin over a 20-fold range. A purine in position -3 (i.e., 3 nucleotides upstream from the ATG codon) has a dominant effect; when a pyrimidine replaces the purine in position -3, translation becomes more sensitive to changes in positions -1, -2, and +4. Single base substitutions around an upstream, out-of-frame ATG codon affect the efficiency with which it acts as a barrier to initiating at the downstream start site for preproinsulin. The optimal sequence for initiation defined by mutagenesis is identical to the consensus sequence that emerged previously from surveys of translational start sites in eukaryotic mRNAs. The mechanism by which nucleotides flanking the ATG codon might exert their effect is discussed.

At least six nucleotides preceding the AUG initiator codon enhance translation in mammalian cells

Sequences flanking the AUG initiator codon influence its recognition by eukaryotic ribosomes. From a comparison of several hundred mRNA sequences, CCA/GCCAUGG emerged as the consensus sequence for initiation in higher eukaryotes. Systematic mutagenesis of a cloned preproinsulin gene confirmed the facilitating effect of A or G in position -3 (i.e. 3 nucleotides upstream from the AUG codon), C in positions -1 and -2, and G immediately following the AUG codon. The analysis of a new set of mutants now reveals that sequences slightly farther upstream are also influential, the optimal context for initiation being (GCC)GCCA/GCCAUGG. Possible mechanistic implications of the repeating GCC motif are discussed.

Interpreting cDNA sequences: some insights from studies on translation.

This review discusses some rules for assessing the completeness of a cDNA sequence and identifying the start site for translation. Features commonly invoked—such as an ATG codon in a favorable context for initiation, or the presence of an upstream in-frame terminator codon, or the prediction of a signal peptide-like sequence at the amino terminus—have some validity; but examples drawn from the literature illustrate limitations to each of these criteria. The best advice is to inspect a cDNA sequence not only for these positive features but also for the absence of certain negative indicators. Three specific warning signs are discussed and documented: (i) The presence of numerous ATG codons upstream from the presumptive start site for translation often indicates an aberration (sometimes a retained intron) at the 5′ end of the cDNA. (ii) Even one strong, upstream, out-of-frame ATG codon poses a problem if the reading frame set by the upstream ATG overlaps the presumptive start of the major open reading frame. Many cDNAs that display this arrangement turn out to be incomplete; that is, the out-of-frame ATG codon is within, rather than upstream from, the protein coding domain. (iii) A very weak context at the putative start site for translation often means that the cDNA lacks the authentic initiator codon. In addition to presenting some criteria that may aid in recognizing incomplete cDNA sequences, the review includes some advice for using in vitro translation systems for the expression of cDNAs. Some unresolved questions about translational regulation are discussed by way of illustrating the importance of verifying mRNA structures before making deductions about translation.

Pushing the limits of the scanning mechanism for initiation of translation

Abstract Selection of the translational initiation site in most eukaryotic mRNAs appears to occur via a scanning mechanism which predicts that proximity to the 5′ end plays a dominant role in identifying the start codon. This ‘position effect’ is seen in cases where a mutation creates an AUG codon upstream from the normal start site and translation shifts to the upstream site. The position effect is evident also in cases where a silent internal AUG codon is activated upon being relocated closer to the 5′ end. Two mechanisms for escaping the first-AUG rule – reinitiation and context-dependent leaky scanning – enable downstream AUG codons to be accessed in some mRNAs. Although these mechanisms are not new, many new examples of their use have emerged. Via these escape pathways, the scanning mechanism operates even in extreme cases, such as a plant virus mRNA in which translation initiates from three start sites over a distance of 900 nt. This depends on careful structural arrangements, however, which are rarely present in cellular mRNAs. Understanding the rules for initiation of translation enables understanding of human diseases in which the expression of a critical gene is reduced by mutations that add upstream AUG codons or change the context around the AUGSTART codon. The opposite problem occurs in the case of hereditary thrombocythemia: translational efficiency is increased by mutations that remove or restructure a small upstream open reading frame in thrombopoietin mRNA, and the resulting overproduction of the cytokine causes the disease. This and other examples support the idea that 5′ leader sequences are sometimes structured deliberately in a way that constrains scanning in order to prevent harmful overproduction of potent regulatory proteins. The accumulated evidence reveals how the scanning mechanism dictates the pattern of transcription – forcing production of monocistronic mRNAs – and the pattern of translation of eukaryotic cellular and viral genes.

RECOGNITION OF AUG AND ALTERNATIVE INITIATOR CODONS IS AUGMENTED BY G IN POSITION +4 BUT IS NOT GENERALLY AFFECTED BY THE NUCLEOTIDES IN POSITIONS +5 AND +6

A primer extension (toeprinting) assay was used to monitor selection by ribosomes of the first versus the second AUG codon as a function of introducing mutations on the 3′ side (positions +4, +5 and +6) of the first AUG codon. Six different flanking codons starting with G (GCG, GCU, GCC, GCA, GAU and GGA) strongly augmented selection of AUG#1 when compared with matched mRNAs that had A or C instead of G in position +4. Augmentation by G in position +4 failed only when it was combined with U in position +5, as in the sequence augGUA. In contrast with the usual enhancing effect of introducing G in position +4, most mutations in position +5 had no discernible effect, as shown with the series augANA (where N = C, A, G or U) and the series augCNA. AUG codon recognition was also unaffected by mutations in position +6, as shown by testing four mRNAs that had augCCN as the start site. Thus the primary sequence context that augments the recognition of AUG start codons does not appear generally to extend beyond G in position +4. When the toeprinting assay was used with mRNAs that initiate translation at CUG instead of AUG, cugGAU was not recognized better than cugGGU, contradicting the hypothesis that initiation at non‐AUG codons might be favored by A instead of G in position +5.

Bifunctional messenger RNAs in eukaryotes

Most eukaryotic mflNAs have a single open reading frame and a single functional initiation site, which is usually the AUG codon that lies closest to the 5’-end (Kozak, Cell 15, 1109-1123, 1978; Microbial. Rev. 47, l-45, 1983). The mechanisms that enforce this monocistronic rule have been elucidated in part by studying the small subset of mRNAs that break the rule: mRNAs, constituting some 50/o-10% of the total, in which spurious AUG codons occur upstream from the start of the long open reading frame, and the even rarer mRNAs, all of viral origin, that direct the synthesis of two separately initiated polypeptides. My primary objective in reviewing these bifunctional viral mRNAs and the mechanisms that underlie their misbehavior is to call attention to certain cellular mRNAs that may also be able to produce two proteins. This exercise in forecasting seems justified by past experience: with every bifunctional viral message, the second protein went undetected until it was realized, based on the nucleotide sequence, that a second protein should be produced. Synthesis of Two Proteins from Overlapping Reading Frames The monocistronic character of most eukaryotic mRNAs, and the occasional deviations therefrom, can be rationalized by ascanning mechanism in which both position (i.e., proximity to the Y-terminus) and flanking sequences dictate which AUG codon will initiate translation. The scanning model postulates that a 40s ribosomal subunit binds initially at the capped 5’-end of a message and migrates linearly until it reaches the first AUG codon. If the first AUG codon lies in an optimal context, which in higher eukaryotes is CCACCAUGG (Kozak, Cell 44,283-292,1986), the 40s subunit migrates no farther; it couples with a 60s subunit and protein synthesis initiates uniquely at that site. When the context around the first AUG codon is less favorable (positions -3 and +4 are especially critical), some 40s subunits nevertheless stop and initiate there, while some bypass the first site and initiate at the next AUG codon downstream. This “leaky scanning” mechanism rationalizes most of the known instances of dual initiation in animal virus systems, as summarized in Table 1. By analogy with these well-studied viral mRNAs, it seems reasonable to expect certain cellular mRNAs to be bifunctional. When the V-proximal AUG codon lies in an unfavorable context and in the same reading frame as the second AUG codon, ribosomes should initiate at both sites, producing “long” and “short” versions of the encoded polypeptide. This is the predicted pattern of expression for the gene that is abnormal in human chronic granulomatous disease (Royer-Pokora et al., Nature 322, 32-38,1986), as well as for some oncogenes (Rosson and Reddy, Nature 379, 604-606, 1986; Rao et al., PNAS 83, 2392-2396, 1986), hormone receptors (Meriino et al., MC6 5, 1722-1734, 1985; Miesfeld et al., Ceil 46,389-399, 1986) and growth factors (Gray et ai.: Nature 312,721-724, 1984; Derynck et al., Nature 376, 701-705, 1985). Since N-terminal amino acids often determine the intracellular distribution, and sometimes the activity of proteins (e.g., oncogene products), the predicted synthesis of long and short forms might have regulatory implications. The consequences are even more interesting when the first and second AUG codons lie in different reading frames, and an unfavorable context around the first AUG codon allows some ribosomes to reach the second. The prediction that such mRNAs can synthesize two proteins from different, overlapping reading frames has again been verified with viral mRNAs, as indicated in Table 1. Although no cellular mRNA has yet been proven to work this way, one promising candidate is the message that encodes a murine lymphokine-a 20,000 dalton protein that initiates at the AUG codon nearest the 5’-end of the message (Noma et al., Nature 379,640-646, 1986). Because the first AUG codon lies in an unfavorable context for initiation (UUGAUGG), some 40s subunits should bypass that

New ways of initiating translation in eukaryotes.

Three interesting ideas about the initiation of translation in eukaryotes have recently emerged in the literature. One is the possibility that internal initiation of translation might occur not only with viral but also with some cellular mRNAs. The second is the possibility of initiating translation without Met-tRNA. The third concerns circumstances under which base pairing might occur between mRNA and rRNA. These ideas, if upheld, would significantly expand our understanding of how eukaryotic ribosomes function. As detailed below, however, there are serious deficiencies in the supporting evidence. Understanding how the present studies fall short might facilitate the design of experiments that are more convincing.

A second look at cellular mRNA sequences said to function as internal ribosome entry sites

This review takes a second look at a set of mRNAs that purportedly employ an alternative mechanism of initiation when cap-dependent translation is reduced during mitosis or stress conditions. A closer look is necessary because evidence cited in support of the internal initiation hypothesis is often flawed. When putative internal ribosome entry sequences (IRESs) are examined more carefully, they often turn out to harbor cryptic promoters or splice sites. This undermines the dicistronic assay, wherein IRES activity is measured by the ability to support translation of the 3′ cistron. Most putative IRESs still have not been checked carefully to determine whether the dicistronic vector produces only the intended dicistronic mRNA. The widespread use of the pRF vector is a major problem because this vector, which has Renilla luciferase as the 5′ cistron and firefly luciferase as the 3′ cistron, has been found to generate spliced transcripts. RNA transfection assays could theoretically circumvent these problems, but most candidate IRESs score very weakly in that test. The practice of calling even very weak results ‘positive’ is one of the problems discussed herein. The extremely low efficiency of putative IRESs is inconsistent with their postulated biological roles. ‘…if it is a Miracle, any sort of evidence will answer, but if it is a Fact, proof is necessary’ —Mark Twain, Letters from the Earth

Regulation of Protein Synthesis in Virus-Infected Animal Cells

Publisher Summary This chapter summarizes the structural features that govern the translation of viral mRNAs: where the synthesis of a protein starts and ends, how many proteins can be produced from one mRNA, and how efficiently. It focuses on the interplay between viral and cellular mRNAs and the translational machinery. That interplay, together with the intrinsic structure of viral mRNAs, determines the patterns of translation in infected cells. It also points out some possibilities for translational regulation that can only be glimpsed at present, but are likely to come into focus in the future. The mechanism of selecting the initiation site for protein synthesis appears to follow a single formula. The translational machinery displays a certain flexibility that is exploited more frequently by viral than by cellular mRNAs. Although some of the parameters that determine efficiency have been identified, how efficiently a given mRNA will be translated cannot be predicted by summing the known parameters.

Emerging links between initiation of translation and human diseases.

Some diseases are caused by mutations that perturb the initiation step of translation by changing the context around the AUG(START) codon or introducing upstream AUG codons. The scanning mechanism provides a framework for understanding the effects of these and other structural changes in mRNAs derived from oncogenes, tumor suppressor genes, and other key regulatory genes. In mRNAs from mutated as well as normal genes, translation sometimes initiates from an internal AUG codon. Sanctioned mechanisms that allow this, including leaky scanning and reinitiation, are discussed. Thrombopoietin mRNA is an example in which translation normally initiates from an internal position via an inefficient reinitiation mechanism. Mutations that restructure this mRNA in ways that elevate production of thrombopoietin cause hereditary thrombocythemia, demonstrating that some mRNAs are designed deliberately with upstream AUG codons to preclude efficient translation and thus to prevent harmful overproduction of potent proteins. While upstream AUG codons in certain mRNAs thus play an important regulatory role, the frequency of upstream AUG codons tends to be exaggerated when cDNA sequences are compiled and analyzed. Because the discovery of mutations that perturb translation usually begins with cDNA analysis, some misunderstandings vis-a-vis the interpretation of cDNA sequences are discussed.