The mRNA of vascular endothelial growth factor (VEGF), the major angiogenic growth factor, contains an unusually long (1,038 nucleotides) and structured 5 untranslated region (UTR). According to the classical translation initiation model of ribosome scanning, such a 5 UTR is expected to be a strong translation inhibitor. In vitro and bicistronic strategies were used to show that the VEGF mRNA translation was cap independent and occurred by an internal ribosome entry process. For the first time, we demonstrate that two independent internal ribosome entry sites (IRESs) are present in this 5 UTR. IRES A is located within the 300 nucleotides upstream from the AUG start codon. RNA secondary structure prediction and site-directed mutagenesis allowed the identification of a 49-nucleotide structural domain (D4) essential to IRES A activity. UV cross-linking experiments revealed that IRES A activity was correlated with binding of a 100-kDa protein to the D4 domain. IRES B is located in the first half of the 5 UTR. An element between nucleotides 379 and 483 is required for its activity. Immunoprecipitation experiments demonstrated that a main IRES B-bound protein was the polypyrimidine tract binding protein (PTB), a well-known regulator of picornavirus IRESs. However, we showed that binding of the PTB on IRES B does not seem to be correlated with its activity. Evidence is provided of an original cumulative effect of two IRESs, probably controlled by different factors, to promote an efficient initiation of translation at the same AUG codon.

Alternative initiations of translation of the human fibroblast growth factor 2 (FGF-2) mRNA, at three CUG start codons and one AUG start codon, result in the synthesis of four isoforms of FGF-2. This process has important consequences on the fate of FGF-2: the CUG-initiated products are nuclear and their constitutive expression is able to induce cell immortalization, whereas the AUG-initiated product, mostly cytoplasmic, can generate cell transformation. Thus, the different isoforms probably have distinct targets in the cell. We show here that translation initiation of the FGF-2 mRNA breaks the rule of the cap-dependent ribosome scanning mechanism. First, translation of the FGF-2 mRNA was shown to be cap independent in vitro. This cap-independent translation required a sequence located between nucleotides (nt) 192 and 256 from the 5' end of the 318-nt-long 5' untranslated region. Second, expression of bicistronic vectors in COS-7 cells indicated that the FGF-2 mRNA is translated through a process of internal ribosome entry mediated by the mRNA leader sequence. By introducing additional AUG codons into the RNA leader sequence, we localized an internal ribosome entry site to between nt 154 and 318 of the 5' untranslated region, just upstream of the first CUG. The presence of an internal ribosome entry site in the FGF-2 mRNA suggests that the process of internal translation initiation, by controlling the expression of a growth factor, could have a crucial role in the control of cell proliferation and differentiation.

The human proto-oncogene c-myc encodes two proteins, c-Myc1 and c-Myc2, from two initiation codons, CUG and AUG, respectively. It is also transcribed from four alternative promoters (P0, P1, P2, and P3), giving rise to different RNA 5-leader sequences, the long sizes of which suggest that they must be inefficiently translated by the classical ribosome scanning mechanism. Here we have examined the influence of three c-myc mRNA 5-leaders on the translation of chimeric myc-CAT mRNAs. We observed that in the reticulocyte rabbit lysate, these 5-leaders lead to cap-independent translation initiation. To determine whether this kind of initiation resulted from the presence of an internal ribosome entry site (IRES), COS-7 cells were transfected with bicistronic vectors containing the different c-myc 5-leaders in the intercistronic region. An IRES was identified, requiring elements located within the P2 leader, between nucleotides ؊363 and ؊94 upstream from the CUG start codon. This is the first demonstration of the existence of IRES-dependent translation for a proto-oncogene. This IRES could be a translation enhancer, allowing activation of c-myc expression under the control of trans-acting factors and in response to specific cell stimuli.

Four isoforms of human fibroblast growth factor 2 (FGF-2) result from alternative initiations of translation at three CUG start codons and one AUG start codon. Here we characterize a new 34-kDa FGF-2 isoform whose expression is initiated at a fifth initiation codon. This 34-kDa FGF-2 was identified in HeLa cells by using an N-terminal directed antibody. Its initiation codon was identified by site-directed mutagenesis as being a CUG codon located at 86 nucleotides (nt) from the FGF-2 mRNA 5' end. Both in vitro translation and COS-7 cell transfection using bicistronic RNAs demonstrated that the 34-kDa FGF-2 was exclusively expressed in a cap-dependent manner. This contrasted with the expression of the other FGF-2 isoforms of 18, 22, 22.5, and 24 kDa, which is controlled by an internal ribosome entry site (IRES). Strikingly, expression of the other FGF-2 isoforms became partly cap dependent in vitro in the presence of the 5,823-nt-long 3' untranslated region of FGF-2 mRNA. Thus, the FGF-2 mRNA can be translated both by cap-dependent and IRES-driven mechanisms, the balance between these two mechanisms modulating the ratio of the different FGF-2 isoforms. The function of the new FGF-2 was also investigated. We found that the 34-kDa FGF-2, in contrast to the other isoforms, permitted NIH 3T3 cell survival in low-serum conditions. A new arginine-rich nuclear localization sequence (NLS) in the N-terminal region of the 34-kDa FGF-2 was characterized and found to be similar to the NLS of human immunodeficiency virus type 1 Rev protein. These data suggest that the function of the 34-kDa FGF-2 is mediated by nuclear targets.