the start codon

A messenger RNA is a long ribbon of letters, but a protein begins at one exact spot, not somewhere in the middle. How does the cell know where the very first word starts? It looks for a special signal — the start codon — that marks 'begin reading the protein here'.

The start codon is almost always AUG, the three bases adenine-uracil-guanine. AUG is special in two ways at once. First, it sets the reading frame: by committing to begin a codon at this A, the cell automatically fixes where every following triplet falls. Second, AUG also codes for the amino acid methionine, so every freshly made protein chain starts life with a methionine at its front (in bacteria, a slightly modified version, formylmethionine). The ribosome does not just grab the first AUG it sees, though — it needs help finding the right one, guided in bacteria by a nearby Shine-Dalgarno sequence and in eukaryotes by scanning along from the capped 5' end until it hits a suitable AUG.

Because the start codon both says 'go' and inserts the first amino acid, it ties the abstract code to the physical machine. That starting methionine is frequently trimmed off afterward, so you should not expect every mature protein to literally begin with Met. And because the choice of which AUG to start at fixes the reading frame for the entire downstream message, a mutation that destroys the start codon, or creates a tempting new AUG upstream, can silence a gene or produce a useless protein.