New vascular endothelial growth factor isoform generated by internal ribosome entry site-driven CUG translation initiation.

We recently demonstrated that the very long 5'-untranslated region (5'-UTR) of the vascular endothelial growth factor (VEGF) mRNA contains two independent internal ribosome entry sites (IRES A and B). In the human sequence, four potential CUG translation initiation codons are located in between these IRES and are in frame with the classical AUG start codon. By in vitro translation and COS-7 cell transfections, we demonstrate that a high mol wt VEGF isoform [called large VEGF (L-VEGF)] is generated by an alternative translation initiation process, which occurs at the first of these CUG codons. Using a bicistronic strategy, we show that the upstream IRES B controls the translation initiation of L-VEGF. This isoform is 206 amino acids longer than the classical AUG-initiated form. With a specific antibody raised against this NH2 extension, we show that the L-VEGF is present in different mouse tissues or in transfected COS-7 cells. We also demonstrate that L-VEGF is cleaved into two fragments: a 23-kDa NH2-specific fragment and a fragment with an apparent size similar to that of the classical AUG-initiated form. This cleavage requires the integrity of a hydrophobic sequence located in the central part of the L-VEGF molecule. This sequence actually plays the role of signal peptide in the classical AUG-initiated form. The AUG-initiated form and the COOH cleavage product of the L-VEGF are both secreted. In contrast, the large isoform and its NH2 fragment present an intracellular localization. These data unravel a further level of complexity in the regulation of VEGF expression.     

Initiation of translation by non-AUG codons in human T-cell lymphotropic virus type I mRNA encoding both Rex and Tax regulatory proteins

Human T-cell lymphotropic virus type I (HTLV-I) double-spliced mRNA exhibits two GUG and two CUG codons upstream to, and in frame with, the sequences encoding Rex and Tax regulatory proteins, respectively. To verify whether these GUG and CUG codons could be used as additional initiation codons of translation, two chimeric constructs were built for directing the synthesis of either Rex–CAT or Tax–CAT fusion proteins. In both cases, the CAT reporter sequence was inserted after the Tax AUG codon and in frame with either the Rex or Tax AUG codon. Under transient expression of these constructs, other proteins of higher molecular mass were synthesized in addition to the expected Rex–CAT and Tax–CAT proteins. The potential non-AUG initiation codons were exchanged for either an AUG codon or a non-initiation codon. This allowed us to demonstrate that the two GUG codons in frame with the Rex coding sequence, and only the second CUG in frame with the Tax coding sequence, were used as additional initiation codons. In HTLV-I infected cells, two Rex and one Tax additional proteins were detected that exhibited molecular mass compatible with the use of the two GUG and the second CUG as additional initiation codons of translation. Comparison of the HTLV-I proviral DNA sequence with that of other HTLV-related retroviruses revealed a striking conservation of the three non-AUG initiation codons, strongly suggesting their use for the synthesis of additional Rex and Tax proteins.     

A New 34-Kilodalton Isoform of Human Fibroblast Growth Factor 2 Is Cap Dependently Synthesized by Using a Non-AUG Start Codon and Behaves as a Survival Factor

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.     

Truncated forms of the dual function human ASCT2 neutral amino acid transporter/retroviral receptor are translationally initiated at multiple alternative CUG and GUG codons

The sodium-dependent neutral amino acid transporter type 2 (ASCT2) was recently identified as a cell surface receptor for endogenously inherited retroviruses of cats, baboons, and humans as well as for horizontally transmitted type-D simian retroviruses. By functional cloning, we obtained 10 full-length 2.9-kilobase pair (kbp) cDNAs and two smaller identical 2.1-kbp cDNAs that conferred susceptibility to these viruses. Compared with the 2.9-kbp cDNA, the 2.1-kbp cDNA contains exonic deletions in its 3' noncoding region and a 627-bp 5' truncation that eliminates sequences encoding the amino-terminal portion of the full-length ASCT2 protein. Although expression of the truncated mRNA caused enhanced amino acid transport and viral receptor activities, the AUG codon nearest to its 5' end is flanked by nucleotides that are incompatible with translational initiation and the next in-frame AUG codon is far downstream toward the end of the protein coding sequence. Interestingly, the 5' region of the truncated ASCT2 mRNA contains a closely linked series of CUG(Leu) and GUG(Val) codons in optimal consensus contexts for translational initiation. By deletion and site-directed mutagenesis, cell-free translation, and analyses of epitope-tagged ASCT2 proteins synthesized intracellularly, we determined that the truncated mRNA encodes multiple ASCT2 isoforms with distinct amino termini that are translationally initiated by a leaky scanning mechanism at these CUG and GUG codons. Although the full-length ASCT2 mRNA contains a 5'-situated AUG initiation codon, a significant degree of leaky scanning also occurred in its translation. ASCT2 isoforms with relatively short truncations were active in both amino acid transport and viral reception, whereas an isoform with a 79-amino acid truncation that lacked the first transmembrane sequence was active only in viral reception. We conclude that ASCT2 isoforms with truncated amino termini are synthesized in mammalian cells by a leaky scanning mechanism that employs multiple alternative CUG and GUG initiation codons.     

Alternative splicing within the elk-1 5' untranslated region serves to modulate initiation events downstream of the highly conserved upstream open reading frame 2

The 5' untranslated region (UTR) plays a central role in the regulation of mammalian translation initiation. Key components include RNA structure, upstream AUGs (uAUGs), upstream open reading frames (uORFs), and internal ribosome entry site elements that can interact to modulate the readout. We previously reported the characterization of two alternatively spliced 5' UTR isoforms of the human elk-1 gene. Both contain two uAUGs and a stable RNA stem-loop, but the long form (5' UTR(L)) was more repressive than the short form (5' UTR(S)) for initiation at the ELK-1 AUG. We now demonstrate that ELK-1 expression arises by a combination of leaky scanning and reinitiation, with the latter mediated by the small uORF2 conserved in both spliced isoforms. In HEK293T cells, a considerable fraction of ribosomes scans beyond the ELK-1 AUG in a reinitiation mode. These are sequestered by a series of out-of-frame AUG codons that serve to prevent access to a second in-frame AUG start site used to express short ELK-1 (sELK-1), an N-terminally truncated form of ELK-1 that has been observed only in neuronal cells. We present evidence that all these events are fine-tuned by the nature of the 5' UTR and the activity of the α subunit of eukaryotic initiation factor 2 and provide insights into the neuronal specificity of sELK-1 expression.     

Polyamine transport by mammalian cells and mitochondria: role of antizyme and glycosaminoglycans

The role of antizyme (AZ) and glycosaminoglycans in polyamine uptake by mammalian cells and mitochondria was examined using NIH3T3 and FM3A cells and rat liver mitochondria. AZ is synthesized as two isoforms (29 and 24.5 kDa) due to the existence of two initiation codon AUGs in the AZ mRNA. Most AZ existed as the 24.5-kDa form translatable from the second AUG, but a portion of the 29-kDa AZ from the first AUG was associated with mitochondria because of the presence of a mitochondrial targeting signal between the first and the second methionine. The predominance of the 24.5-kDa isoform was mainly due to the presence of spermidine and a favorable sequence context (Kozak sequence) at the second initiation codon AUG. Spermine uptake by NIH3T3 cells was inhibited by both 29- and 24.5-kDa AZs, but uptake by rat liver mitochondria was not influenced by either form of AZ. Because spermine uptake by mitochondria caused a release of cytochrome c, an enhancer of apoptosis, we looked for inhibitors of mitochondrial spermine uptake other than AZ. Cations such as Na+, K+, and Mg2+ were inhibitors of the mitochondrial uptake. It has been reported that heparan sulfate on glypican-1 plays important roles in spermine uptake by human embryonic lung fibroblasts. Heparin, but not heparan sulfate, slightly inhibited spermine uptake by FM3A cells in the absence of Mg2+ and Ca2+ but had no effect under physiological conditions in the presence of Mg2+ and Ca2+.     

One single in-frame AUG codon is responsible for a diversity of subcellular localizations of glutaredoxin 2 in Saccharomyces cerevisiae

Glutaredoxins belong to a family of small proteins with glutathione-dependent disulfide oxidoreductase activity involved in cellular defense against oxidative stress. The product of the yeast GRX2 gene is a protein that is localized both in the cytosol and mitochondria. To throw light onto the mechanism responsible for the dual subcellular distribution of Grx2 we analyzed mutant constructs containing different targeting information. By altering amino acid residues around the two in-frame translation initiation start sites of the GRX2 gene, we could demonstrate that the cytosolic isoform of Grx2 was synthesized from the second AUG, lacking an N-terminal extension. Translation from the first AUG resulted in a long isoform carrying a mitochondrial targeting presequence. The mitochondrial targeting properties of the presequence and the influence of the mature part of Grx2 were analyzed by the characterization of the import kinetics of specific fusion proteins. Import of the mitochondrial isoform is relatively inefficient and results in the accumulation of a substantial amount of unprocessed form in the mitochondrial outer membrane. Substitution of Met(35), the second translation start site, to Val resulted in an exclusive targeting to the mitochondrial matrix. Our results show that a plethora of Grx2 subcellular localizations could spread its antioxidant functions all over the cell, but one single A to G [corrected] mutation converts Grx2 into a typical protein of the mitochondrial matrix. The "A" denotes adenine, rather than alanine, and the "G" refers to guanine, not glycine [corrected]     

Translation of psbC mRNAs starts from the downstream GUG, not the upstream AUG, and requires the extended Shine-Dalgarno sequence in tobacco chloroplasts

The plastid gene psbC encodes the CP43 subunit of PSII. Most psbC mRNAs of many organisms possess two possible initiation codons, AUG and GUG, and their coding regions are generally annotated from the upstream AUG. Using a chloroplast in vitro translation system, we show here that translation of the tobacco plastid psbC mRNA initiates from the GUG. This mRNA possesses a long Shine-Dalgarno (SD)-like sequence, GAGGAGGU, nine nucleotides upstream of the GUG. Point mutations in this sequence abolished translation, suggesting that a strong interaction between this extended SD-like sequence and the 3' end of 16S rRNA facilitates translation initiation from the GUG.     

Context effects and inefficient initiation at non-AUG codons in eucaryotic cell-free translation systems.

The context requirements for recognition of an initiator codon were evaluated in vitro by monitoring the relative use of two AUG codons that were strategically positioned to produce long (pre-chloramphenicol acetyl transferase [CAT]) and short versions of CAT protein. The yield of pre-CAT initiated from the 5'-proximal AUG codon increased, and synthesis of CAT from the second AUG codon decreased, as sequences flanking the first AUG codon increasingly resembled the eucaryotic consensus sequence. Thus, under prescribed conditions, the fidelity of initiation in extracts from animal as well as plant cells closely mimics what has been observed in vivo. Unexpectedly, recognition of an AUG codon in a suboptimal context was higher when the adjacent downstream sequence was capable of assuming a hairpin structure than when the downstream region was unstructured. This finding adds a new, positive dimension to regulation by mRNA secondary structure, which has been recognized previously as a negative regulator of initiation. Translation of pre-CAT from an AUG codon in a weak context was not preferentially inhibited under conditions of mRNA competition. That result is consistent with the scanning model, which predicts that recognition of the AUG codon is a late event that occurs after the competition-sensitive binding of a 40S ribosome-factor complex to the 5' end of mRNA. Initiation at non-AUG codons was evaluated in vitro and in vivo by introducing appropriate mutations in the CAT and preproinsulin genes. GUG was the most efficient of the six alternative initiator codons tested, but GUG in the optimal context for initiation functioned only 3 to 5% as efficiently as AUG. Initiation at non-AUG codons was artifactually enhanced in vitro at supraoptimal concentrations of magnesium.     

The AUG start codon of the Saccharomyces cerevisiae NFS1 gene can be substituted for by UUG without increased initiation of translation at downstream codons.

The selection of the site for initiation of translation for the Saccharomyces cerevisiae NFS1 gene was examined using mutated AUG1, AUG2 and AUG3 codons. When AUG1 of the yeast NFS1 gene was mutated to UUG and the resulting mRNA was translated in vitro using a reticulocyte system, initiation from the mutated codon was abolished and occurred instead at downstream codons at increased rates. When the same mRNA was translated using a yeast extract, translation initiated at the mutated codon, albeit at a reduced rate, and there was no increased translation at downstream AUG codons. The NFS1 gene in which AUG1 was replaced by UUG was also able to substitute for the wild-type gene in vivo in yeast. Western blots confirmed that the encoded protein was the same size as that encoded by the wild-type gene and that both the wild-type and mutated proteins localized to mitochondria. This is apparently the first example of a yeast protein where mutagenesis of AUG1 does not lead to alternate use of a downstream AUG.