Translation stops when a ribosome encounters one of three specific sequences in the messenger RNA called stop codons: UAA, UAG, or UGA. Unlike every other three-letter codon in the genetic code, no transfer RNA molecule matches these sequences. Instead, proteins called release factors recognize the stop codon and trigger the release of the newly built protein from the ribosome. The entire termination process unfolds in a precise series of steps, from stop codon recognition to ribosome disassembly.
The Three Stop Codons
Every stop codon begins with uracil (U) in the first position. The three sequences, UAA, UAG, and UGA, are universal across nearly all life. They carry historical nicknames from early genetics research: UAG is “amber,” UAA is “ochre,” and UGA is “opal.” Because no transfer RNA is designed to pair with these codons, the ribosome essentially hits a wall. Instead of adding another amino acid, the machinery shifts into termination mode.
UAA is often considered the most “reliable” stop codon because it is recognized by the widest range of release factors across species. UGA is the most frequently reassigned codon in nature, sometimes coding for an amino acid rather than stopping translation (more on that below).
How Release Factors End Translation
In bacteria, two release factors split the work. RF1 recognizes UAA and UAG, while RF2 recognizes UAA and UGA. A third factor, RF3, uses the energy molecule GTP to help the other release factors do their job and then eject from the ribosome afterward.
Eukaryotic cells (including human cells) simplify this with a single release factor, eRF1, that recognizes all three stop codons. It does this through a set of protein motifs that read the stop codon almost like a lock and key. The first position must be uracil. The second and third positions must each be a purine (adenine or guanine), but two guanines in a row are excluded because their chemical groups would repel each other and clash with a critical part of eRF1. Researchers have described this logic as functioning like a NAND gate, a term from computer science meaning “anything except both.” A partner factor, eRF3, acts as a GTPase: in its GTP-bound state it helps position eRF1 correctly on the stop codon, and after GTP is broken down to GDP, it promotes the release of both factors from the ribosome.
How the Protein Is Actually Released
Once a release factor locks onto the stop codon, the finished protein chain needs to be physically cut free from the last transfer RNA holding it. For decades, scientists believed a water molecule performed this cut. A 2024 study published in Science overturned that model. High-resolution structures of bacterial ribosomes caught just before the cleavage moment showed no water molecule positioned for the reaction.
Instead, the release factor reshapes the final nucleotide of the transfer RNA (called A76), flipping its sugar ring into a different configuration. This repositioning aims a chemical group on the RNA itself, a hydroxyl group, directly at the bond connecting the protein chain to the transfer RNA. The RNA essentially cuts itself. As the researchers put it, nature used RNA’s most fundamental property, its tendency to self-cleave when the hydroxyl group is in the right position, rather than evolving a more elaborate enzymatic mechanism. This self-cleavage releases the completed protein into the cell.
What Happens to the Ribosome Afterward
Releasing the protein is not the final step. The ribosome is still sitting on the mRNA with a now-empty transfer RNA inside it. This “post-termination complex” needs to be taken apart so the ribosome’s subunits can be recycled for future rounds of translation.
In bacteria, a protein called ribosome recycling factor (RRF) teams up with elongation factor G and GTP energy to pry the ribosome off the mRNA. This initially releases the ribosome as an intact unit (both subunits still joined). A separate factor, IF3, then splits it into its large and small subunits, which are now free to find another mRNA and start translating again. In eukaryotes, a protein called ABCE1 fills a similar recycling role, using ATP energy to split the ribosome apart after termination.
When Stop Codons Appear Too Early
Mutations can introduce a stop codon in the middle of a gene, creating what’s called a premature termination codon (PTC). The ribosome obeys the stop signal just as it would a normal one, producing a truncated, usually nonfunctional protein. Roughly 10 to 15 percent of all inherited genetic diseases are caused by these premature stop mutations, including cystic fibrosis, Duchenne muscular dystrophy, hemophilia, and some cancers.
Cells have a surveillance system to catch these mistakes called nonsense-mediated mRNA decay, or NMD. The key trigger involves markers called exon junction complexes (EJCs) that are deposited on the mRNA during processing in the nucleus, about 20 to 24 nucleotides upstream of each spot where two exons were spliced together. During normal translation, the ribosome physically pushes these markers off as it reads through. But if a premature stop codon halts the ribosome before it can clear an EJC located more than about 50 to 55 nucleotides downstream, the remaining EJC recruits degradation machinery. A central protein called UPF1 is recruited to the stalled ribosome, gets activated by other factors, and marks the entire mRNA for destruction. This prevents the cell from continuing to churn out defective proteins.
NMD can also be triggered without EJCs. Some mRNAs with unusually long untranslated regions at their tail end are targeted for decay, likely because the physical distance between the stop codon and certain stabilizing proteins at the mRNA’s tail is too great for normal termination signals to work properly.
When Translation Gets Stuck Without a Stop Codon
Sometimes mRNA is damaged or incomplete and simply lacks a stop codon altogether. The ribosome reads to the end of the message and stalls, unable to proceed or terminate. Bacteria solve this with a remarkable molecule called tmRNA, which functions as both a transfer RNA and a messenger RNA. It enters the stalled ribosome like a normal tRNA, then redirects the ribosome onto a short built-in coding sequence that adds a peptide tag to the end of the incomplete protein. This tag marks the defective protein for immediate destruction by the cell’s protein-recycling machinery. The process also frees the ribosome and promotes degradation of the broken mRNA that caused the stall in the first place.
When a Stop Codon Doesn’t Mean Stop
UGA doesn’t always end translation. In both bacteria and eukaryotes, UGA sometimes codes for selenocysteine, a rare amino acid containing selenium. This recoding depends on a specific RNA structure called a SECIS element (selenocysteine insertion sequence) located in the mRNA. In bacteria, the SECIS element sits in the coding region about 20 to 25 nucleotides downstream of the UGA codon. In eukaryotes and archaea, it’s found in the untranslated region at the tail end of the mRNA, and can be located much farther from the UGA codon.
The SECIS element recruits specialized translation factors that deliver selenocysteine-loaded tRNA to the ribosome before the normal termination machinery can act. This is how cells produce selenoproteins, enzymes critical for antioxidant defense and thyroid hormone metabolism. Without the SECIS element, the same UGA codon simply terminates translation as usual.
Drugs That Override Premature Stops
Because premature stop codons cause so many genetic diseases, researchers have developed drugs that coax the ribosome to read through them. Ataluren (sold as Translarna in the European Union for Duchenne muscular dystrophy caused by nonsense mutations) works by interacting with the ribosome and helping it accept a near-match transfer RNA at the premature stop codon instead of engaging the termination machinery. The result is a full-length protein with a single amino acid substitution at the site of the original mutation, often functional enough to provide therapeutic benefit.
At UAA and UAG premature stops, ataluren tends to insert glutamine, lysine, or tyrosine. At UGA stops, it favors tryptophan, arginine, or cysteine. Older readthrough-promoting compounds like gentamicin also force readthrough, but they insert less predictable amino acids, producing proteins more likely to malfunction or trigger immune reactions. Ataluren’s selectivity for specific amino acid insertions is considered one of its advantages.