Start and stop codons

The start codon and the three stop codons

In the standard genetic code, AUG is the start codon and UAA, UAG and UGA are the stop codons. In DNA letters they are ATG and TAA, TAG and TGA, read on the coding strand. Translation begins at a start codon and ends at the first stop codon in the same reading frame. The start codon codes for methionine; a stop codon adds no amino acid (NCBI table 1).

AUG has two jobs. At the start of a coding sequence it begins the protein; inside one it simply adds another methionine. The codon chart marks the start and the three stops among all 64 codons.

How a start codon begins translation

The ribosome begins where an initiator tRNA pairs with the start codon and brings in methionine (formylmethionine in bacteria). Not every AUG is a start: most AUGs in an mRNA lie inside the coding sequence. Which one starts the protein depends on the bases around it. In eukaryotes the ribosome scans from the mRNA's 5′ end and favors an AUG in a good Kozak context. In bacteria a Shine-Dalgarno sequence a few bases upstream pairs with the ribosome and places the start codon.

Other start codons

NCBI's standard code also lists UUG and CUG as possible starts, and its bacterial, archaeal and plant plastid code (table 11) adds GUG, AUU, AUC and AUA. Whichever codon a protein starts at, that first codon is read as methionine by default (NCBI genetic codes). How often genes use them, counted from the first codon of the annotated genes in this site's start-codon references (pseudogenes and annotated exceptions left out):

In E. coli, 418 of 4,294 genes (9.7%) start with GTG or TTG. In human MANE Select transcripts, all but 27 of 19,248 start with ATG: 25 start with CTG and 2 with TTG. A scan that accepts only ATG therefore misses some bacterial genes; the ORF inspector can search with the starts NCBI lists for the code you choose.

How a stop codon ends translation

Normally no tRNA reads UAA, UAG or UGA. Instead a protein called a release factor recognizes the stop codon in the ribosome and frees the finished chain. Bacteria have two: RF1 responds to UAA and UAG, and RF2 to UAA and UGA (Scolnick et al., 1968). In eukaryotes a single release factor, eRF1, recognizes all three (Seit-Nebi, Frolova and Kisselev, 2002).

How often genes use each stop codon

All three stops end real genes, but organisms prefer different ones. Counted from the last codon of every complete annotated coding sequence in the codon usage tables:

E. coli ends 64% of its genes with UAA, while human genes end with UGA about half the time. UAG is the least used stop in all three.

How the three stops were found

UAG and UAA were identified as chain-terminating "nonsense" triplets in 1965, through the amber and ochre mutants of bacteriophage T4 (Brenner, Stretton and Kaplan, 1965). UGA was shown to be the third in 1967 (Brenner, Barnett, Katz and Crick, 1967). Geneticists still call UAG amber, UAA ochre and UGA opal.

When a stop codon codes for an amino acid

A stop codon can carry an amino acid in a few special cases. In some proteins UGA inserts selenocysteine, as in the mouse glutathione peroxidase gene (Chambers et al., 1986), and some archaea read UAG as pyrrolysine (Srinivasan, James and Krzycki, 2002). The codon alone doesn't say when this happens, so a sequence translator, including this site's, still shows those codons as stops.

Whole genetic codes also change the stops. NCBI's vertebrate mitochondrial code reads UGA as tryptophan and lists AGA and AGG as stops instead, and the ciliate nuclear code reads UAA and UAG as glutamine, leaving UGA as its only stop.

Start and stop codons in all 27 genetic codes

Each NCBI genetic code's stop codons and the codons it lists as possible starts, in mRNA letters (read T for U in DNA). A listed start is a possibility, not a promise that translation begins there. To see a whole code, choose it under Options in the codon chart.

† Stop or amino acid, depending on where the codon sits in the mRNA. Source: NCBI genetic codes, last updated September 23, 2024.

Finding start and stop codons in a sequence

A codon means something only in its reading frame. To find a protein's start and stop, read the coding strand 5′ to 3′ in steps of three from a start codon until the first stop in the same frame. Shift the frame by one base and the same letters make different codons; the genetic-code guide shows an example.

On the template strand the same signals are reverse complements. Written 5′ to 3′, ATG appears there as CAT, and TAA, TAG and TGA appear as TTA, CTA and TCA. If your sequence is the template strand, reverse-complement it first. To list every start-to-stop stretch in all six frames, paste the sequence into the ORF inspector.