In Escherichia coli, the codon AUG of the genetic code is an efficient initiator of translation in protein biosynthesis and is by far the most common one. GUG and UUG are also efficient initiation codons and are not uncommon. AUU is only rarely used, most notably in the gene for translation initiation factor 3.
At initiation of translation, the anticodon of a special initiator transfer RNA, tRNAMef, pairs with the initiation codon bound at the P site of a ribosome. In eubacteria, chloroplasts and mitochondria, the initiating amino acid is formylmethionine (1). The formyl group permits binding by initiation factor 2. This, together with special features of tRNAMef, ensures that formylmethionine is delivered to the ribosomal P site and becomes the first amino-terminal amino acid to be incorporated into a polypeptide chain. The formyl group is removed subsequently, however, so that methionine is the first encoded amino acid. Formylmethionine tRNAMetf is not recognized by the elongation factor
EF-Tu, which delivers aminoacyl tRNA to the ribosomal A site for incorporation at internal positions in polypeptide proteins. Instead, the "elongator" that decode AUG, GUG, UUG, and AUU at internal positions in coding regions are not acylated with formylmethionine, but rather with methionine, valine, leucine, and isoleucine, respectively, as in the normal genetic code.
There is a special context for initiator codons that gives them their distinctive function and meaning. In the overwhelming majority of cases in eubacteria, this special context is a short sequence of bases in the messenger RNA known as a Shine-Dalgarno sequence after its discoverers. It is located approximately 5 or more nucleotides 5′ of the initiator codon. The Shine-Dalgarno sequence of the messenger RNA pairs with a complementary sequence very close to the 3′ end of 16 S rRNA in the small (30 S) ribosomal subunit. The interaction serves to position the initiation codon for interaction with initiator tRNA at the ribosomal P site. In a small number of mRNAs from eubacterial bacteriophage or extreme thermophilic archae, there is no Shine-Dalgarno sequence, and occasionally no 5′ sequence at all. In these cases, some sequence 3′ to the start codon that is only partially defined serves to facilitate initiation.
In eukaryotes, AUG is again by far the most common initiator codon, but CUG, ACG, and GUG are also found. The special initiator tRNA, tRNAMeti, inserts methionine and not formylmethionine. The discriminatory features used to specify when these codons are to function as initiators are different from that utilized in eubacteria. With the aid of initiation factors, the small, 40 S subunit of eukaryotic ribosomes generally associates with mRNA via a -cap structure and, at least in yeast, the poly A tail at the 3′ end of the mRNA also plays a role (2, 3). The recruited subunit then scans until a potential initiator codon flanked by a consensus sequence is encountered (4). The large, 60 S ribosomal subunit subsequently joins to complete assembly of the intact, 80 S ribosome. The great majority of eukaryotic start codons are recognized by ribosomes recruited in this manner, but there are two interesting exceptions.
1. The mRNA for a number of proto-oncogenes and transcription factors have one or more short open reading frames (ORF) preceding the main protein coding sequence. Ribosomes that have translated a particular short ORF may have the potential to reinitiate translation at the main ORF start codon location some distance in the 3′ direction. This mode of initiation has been studied in great depth in yeast GCN4 (5).
2. 2. Special structures known as IRES (internal ribosome entry segments) permit ribosomes to make initial contact with internal regions in mRNA and initiate translation at appropriately positioned start codons (6). IRES-mediated internal initiation was initially discovered in translation of the picornaviruses, polio virus and encephalomyocarditis, where it enables viral protein synthesis to escape the consequences of viral-induced inhibition of cap-mediated initiation (7, 8). This inhibition serves to inactivate host translation. However, the start codons of a small number of chromosomally encoded genes are also recognized by IRES-mediated initiation.
