ИСТИНА

Two major and principally different mechanisms of translation initiation exist in eukaryotes: cap-dependent scanning and internal ribosome entry. The latter mechanism requires specific structures known as internal ribosome entry site (IRES) elements located in the 5’ untranslated regions (UTRs) of the mRNAs. IRES elements were initially discovered in the 5’ UTRs of uncapped mRNAs from viruses that replicate in the cytoplasm. Later, the presence of IRES elements within the 5’UTRs of many cellular mRNAs which encode proteins with regulatory functions was also suggested. As a rule, these 5’ UTRs have a complex secondary structure and bind specific mRNA-binding proteins. However, the existence of cellular IRESs is still a matter of debate. Here, we have compared the translation initiation characteristics of several well- characterised viral IRESs (HCV, EMCV and HRV) with those 5’ UTRs from cellular mRNAs reported to direct internal ribosome entry (c-Myc, Apaf-1, hsp70). Two mRNAs, LINE-1 (with a long and highly structured 5’-UTR) and beta-actin (with a short 5’UTR) were used as controls, as both are known to employ an exclusively cap-dependent mechanism of initiation. On transfecting cells with dicistronic DNA constructs we found that translation of the second cistron driven by the cellular IRESs listed above could be accounted for either by the presence of a cryptic promoter or alternative splicing. Therefore, all subsequent transfections were performed with polyadenylated mRNAs (both capped and uncapped versions). These experiments were recapitulated in vitro, in non-nuclease treated extracts from Krebs-2 cells, and the results obtained were in excellent agreement with the ex vivo data. In contrast, the data obtained from translation in RRL proved to be inconsistent with the other systems employed. In order to evaluate the actual mechanisms of mRNA translation we employed a number of assays: 1) comparison of translation initiation from a 5’UTR in a monocistronic context versus a dicistronic context; 2) comparison of the efficiency of translation from m7G capped monocistronic transcripts vs. the corresponding A-capped mRNAs; and 3) estimation of the extent of inhibition of translation initiation from the different 5’ UTRs with the addition of increasing concentrations of eIF4-BP1 and m7GTP to Krebs-2 extracts. In contrast to the viral IRESs, translation directed by the putative cellular IRESs was much more efficient in the context of a monocistronic mRNA than in the context of a dicistronic mRNA; indeed, in a dicistronic context IRES activities were extremely low. Translation initiation on these IRESs was strongly dependent on the presence of the cap and was dramatically inhibited by the addition of eIF4E-BP1 or m7GTP. As expected, no inhibition of viral IRES activity was observed on the addition of eIF4E-BP or m7GTP. In addition, we have confirmed our previous observations that there is no direct correlation between the length and complexity of a 5’ UTR and the efficiency of translation of the corresponding mRNA. For instance, efficiency of translation initiation from the Apaf-1 5’ UTR (578 nt) was more than twice that of the beta-actin 5’UTR (84 nt). This data leads us to conclude that, at least under normal cellular conditions, mRNAs with long, structured 5’UTRs are translated efficiently by the cap-dependent mechanism.