Translation begins when a small ribosomal subunit, along with a special initiator transfer RNA and several protein factors, assembles at a specific site on a messenger RNA molecule and locates the start codon. This process differs significantly between bacteria and eukaryotes (animals, plants, fungi), but in both cases, the goal is the same: position the ribosome precisely at the right spot so protein synthesis can begin.
The Two Core Steps in All Translation Initiation
Regardless of the organism, initiation follows two universal requirements. First, the small ribosomal subunit must find the correct start codon on the mRNA. In nearly all cases, this is the triplet AUG, which codes for the amino acid methionine. Second, a special initiator tRNA carrying methionine must be placed in the ribosome’s P site, the pocket where the first amino acid sits before the ribosome starts reading the genetic code.
The initiator tRNA is structurally distinct from the regular methionine tRNAs used during the rest of protein synthesis. In bacteria, the methionine attached to the initiator tRNA is chemically modified with a formyl group, producing formyl-methionine. Two features make an initiator tRNA functional: it can be formylated, and it binds directly to the ribosomal P site. Regular tRNAs enter a different pocket (the A site) and lack both of these properties. Experiments in E. coli showed that engineering these two features into an unrelated tRNA was enough to make it work as an initiator.
How Bacteria Find the Start Codon
Bacterial ribosomes locate the start codon through a direct base-pairing interaction between the mRNA and the ribosome itself. Upstream of the start codon, most bacterial mRNAs contain a short purine-rich sequence called the Shine-Dalgarno (SD) sequence. The small ribosomal subunit (the 30S subunit) carries a complementary sequence at the tail end of its 16S ribosomal RNA. These two sequences pair together, forming a short RNA duplex that anchors the ribosome in the correct position on the mRNA so the start codon lands directly in the P site.
Three initiation factors assist this process. IF1 and IF3 help ensure the ribosome selects the correct start codon. IF3 is specifically involved in verifying that the right codon sits in the P site. IF2 facilitates recognition of the start codon and helps recruit the initiator tRNA. Once the small subunit, initiator tRNA, and mRNA are properly assembled, the large ribosomal subunit (50S) joins, the initiation factors leave, and the ribosome is ready to begin adding amino acids.
This direct-recruitment system means bacterial ribosomes can land in the middle of an mRNA, not just at the beginning. Bacterial mRNAs often encode multiple proteins in a row, and each coding region has its own Shine-Dalgarno sequence that independently recruits ribosomes.
Eukaryotic Initiation: Cap Recognition and Scanning
Eukaryotic cells use a fundamentally different strategy. Instead of base-pairing to find the start codon, their ribosomes bind to the very beginning of the mRNA and slide along it until they encounter the first AUG. This is called the scanning model, and it involves far more protein factors than the bacterial system.
The process begins at the 5′ cap, a modified guanosine nucleotide added to the front of every eukaryotic mRNA. A protein complex called eIF4F recognizes this cap. eIF4F has three components, each with a distinct job: eIF4E physically grabs the cap, eIF4G acts as a scaffold connecting the other players, and eIF4A is a helicase that unwinds secondary structures in the mRNA so the ribosome can move through it. The helicase activity of eIF4A is boosted by both eIF4G and another factor called eIF4B.
Structural studies have revealed that two separate copies of the eIF4A helicase work during initiation. One sits within the eIF4F complex near the mRNA exit channel of the ribosome, while a second, independent copy binds at the mRNA entry channel. This arrangement means that RNA structures are unwound both ahead of and behind the ribosome as it moves along the mRNA.
Meanwhile, eIF4G also interacts with a protein bound to the poly(A) tail at the opposite end of the mRNA. This effectively loops the mRNA into a circle, bringing the front and back ends close together, which is thought to promote efficient recycling of ribosomes after they finish translating.
Assembling the Scanning Complex
Before the ribosome can begin scanning, a pre-initiation complex must be built. The small ribosomal subunit (40S in eukaryotes) associates with the initiator tRNA and several initiation factors to form what’s called the 43S complex. The initiator tRNA is delivered by a factor called eIF2, which carries it in a GTP-bound state.
This 43S complex attaches to the capped 5′ end of the mRNA (with the help of the eIF4F complex) and begins scanning downstream through the 5′ untranslated region. During scanning, the mRNA sits loosely in a widened channel within the 40S subunit, with minimal contacts along its length. The initiator tRNA’s anticodon inspects each codon as the complex moves forward.
The large initiation factor eIF3 plays a key structural role throughout this process. Cryo-electron microscopy studies show that during scanning, eIF3 wraps almost entirely around the 40S subunit. Part of it anchors to the outer surface, while a mobile module swings to the inner face where it contacts the initiator tRNA and other factors near the decoding center. When the complex shifts from an open, scanning-competent state to a closed state that locks onto a start codon, eIF3 reorganizes its position but maintains this encircling architecture.
Recognizing the Start Codon
The scanning complex typically stops at the first AUG codon it encounters. However, the surrounding nucleotide sequence strongly influences how efficiently a given AUG is recognized. In vertebrates, the optimal context is a purine (A or G) at position -3 (three nucleotides before the AUG) and a G immediately after it, within a broader consensus sequence of GCCGCC(A/G)CCAUGG. In human genes, 39% of transcripts have an A at position -3 and 35% have a G, both substantially enriched compared to what you’d expect by chance.
When the context around an AUG is weak, some scanning ribosomes skip past it and initiate at a downstream start codon instead. This “leaky scanning” can produce multiple protein variants from the same mRNA. The difference in initiation efficiency between an optimal and a weak AUG context can be up to 12-fold. Codons other than AUG can also serve as start codons, but they are inherently leaky, meaning downstream AUG-initiated proteins are always produced as well.
Joining the Large Subunit
Once the scanning complex locks onto the start codon, the resulting structure (now called the 48S complex) needs to recruit the large ribosomal subunit (60S) to form a complete, functional ribosome. This step requires the initiation factor eIF5B loaded with GTP.
The 60S subunit docks onto the 40S in the presence of the initiator tRNA, eIF5B, and several remaining factors. Correct positioning of the initiator tRNA triggers eIF5B to hydrolyze its GTP. A specific loop in the 60S ribosomal RNA physically repositions a key amino acid in eIF5B’s active site, pushing it closer to the GTP molecule and catalyzing the reaction. Once GTP is split into GDP and phosphate, eIF5B changes shape, its grip on the ribosome weakens, and it falls off along with the remaining initiation factors. The ribosome is now an 80S complex, and it transitions into the elongation phase, where amino acids are added one by one.
How Cells Regulate Initiation
Because initiation is the slowest and most controlled step in translation, it serves as a major point of regulation. One of the best-understood control mechanisms targets eIF2, the factor that delivers the initiator tRNA.
After each round of initiation, eIF2 is left holding GDP instead of GTP, making it inactive. A recycling factor called eIF2B swaps the GDP for a fresh GTP, recharging eIF2 for another round. Under stress conditions like viral infection, nutrient deprivation, or the accumulation of misfolded proteins, cells activate kinases that add a phosphate group to eIF2’s alpha subunit. This phosphorylated eIF2 binds to eIF2B so tightly that eIF2B cannot release it. Trapped in this unproductive complex, eIF2B can no longer recycle any eIF2 molecules, and global protein synthesis drops sharply.
Paradoxically, this shutdown selectively increases translation of certain stress-response genes. In mammals, the transcription factor ATF4 is translated more efficiently when eIF2 activity is low, because its mRNA contains upstream open reading frames that normally divert ribosomes away from the main coding sequence. When scanning ribosomes are scarce, they’re more likely to bypass these decoy sequences and reach the correct start codon. This pathway is called the integrated stress response.
Cap-Independent Initiation Through IRES Elements
Not all translation begins at the 5′ cap. Some mRNAs contain internal ribosome entry sites (IRES), structured RNA elements within the 5′ untranslated region that recruit the 40S subunit directly to the vicinity of the start codon, skipping cap recognition and scanning entirely. Viruses were the first systems where IRES elements were discovered, but many cellular mRNAs also contain them.
IRES-mediated translation becomes especially important during stress, when cap-dependent initiation is shut down through eIF2 phosphorylation or other mechanisms. Certain proteins called IRES trans-acting factors (ITAFs) help stabilize the IRES into the right three-dimensional shape for ribosome binding. This allows cells to continue producing critical stress-response proteins even when the normal translation machinery is largely inactive.