The same three-act shape, now run by the ribosome
By now the cast is familiar. You know the genetic code reads the message three letters at a time, that a tRNA is the adapter carrying the right amino acid for each codon, and that the ribosome is the two-part machine that does the reading. This guide stops naming the players and watches them work — it follows one ribosome through a single, complete act of building a protein, from the moment it commits to a message to the moment it sets the finished chain free. That whole performance breaks into the same three acts you saw in transcription: initiation (assembling at the start), elongation (the steady middle), and termination (knowing when to stop).
The shape rhymes with transcription, but the parts and the stakes are different. There the machine read DNA and wrote RNA; here the ribosome reads RNA and writes a chain of amino acids, switching alphabets entirely — nucleotides going in, a polypeptide coming out. And the rhythm is tied to that three-letter beat: every codon must be matched, bonded, and stepped past in turn. Throughout, I will lean on a running picture — an assembly line where each station reads a three-letter instruction, snaps on one bead, and ratchets the belt forward by exactly one instruction.
Act one — initiation: assembling on the right starting line
Initiation solves the hardest problem in the whole process: where exactly to begin. This matters far more than it might sound, because the message is read in non-overlapping triplets, so the choice of starting point fixes the reading frame for everything that follows. Pick the wrong AUG, or start one letter off, and every downstream codon is mis-grouped — the protein comes out as gibberish. So the ribosome cannot simply grab the message and go; it must find the genuine start codon, the AUG that opens the real coding sequence, and lock onto it precisely.
Bacteria and eukaryotes find that AUG in strikingly different ways, and it is the cleanest place to see the two domains diverge. In bacteria the message carries a short landing strip a few bases upstream of the start, the [[shine-dalgarno-sequence|Shine-Dalgarno sequence]]. It is complementary to the tail of the small subunit's own rRNA, so the two simply base-pair — the ribosome's small subunit clicks onto the message like a key into a labeled slot, parking the start codon right where it needs to be. Because the bases pair directly, a bacterial ribosome can dock anywhere along a message that carries this signal, which is how one bacterial mRNA can carry several genes that are each translated independently.
Eukaryotes have no Shine-Dalgarno signal, so they use cap-scanning instead. The small subunit, loaded with initiation factors and the special initiator tRNA-Met, latches onto the 5' cap at the very front of the mRNA (the cap you met when we processed pre-mRNA) and then *scans* downstream, sliding letter by letter until it reaches the first AUG in a suitable context. Picture a reader who must always start at page one and skim forward to find the chapter heading, rather than flipping straight to a bookmark. Either way, the act ends the same: the initiator tRNA settles into the ribosome's P site paired with the start codon, the large subunit joins on, and the assembled ribosome stands ready with its first amino acid in place and its A site empty, waiting.
Act two — elongation: read, bond, shift, repeat
Now the steady middle begins. Elongation is a cycle that repeats once per codon, and the key to seeing it clearly is the ribosome's three tRNA slots, named A, P, and E. Think of them as three adjacent chairs: the A site (Arrival) is where a fresh tRNA delivers its amino acid, the P site (Peptide) holds the tRNA carrying the chain built so far, and the E site (Exit) is the doorway through which a spent tRNA leaves. The whole cycle is just the orderly procession of tRNAs through these three chairs, one codon at a time.
- Arrive and check: an elongation factor (EF-Tu in bacteria) ushers the next tRNA into the empty A site. Its anticodon is test-fitted against the A-site codon; only a correct three-letter match lets the factor spend its GTP and lock the tRNA in. A wrong tRNA falls out before it can commit — this proofreading step is most of why translation is accurate.
- Bond: with two amino acids now side by side (the growing chain in the P site, the new one in the A site), the ribosome's own catalytic core, the [[peptidyl-transferase|peptidyl transferase]] centre, joins them with a peptide bond — transferring the whole chain onto the A-site amino acid. Strikingly, this catalyst is made of rRNA, not protein: the ribosome is a ribozyme.
- Shift: a second elongation factor (EF-G), spending another GTP, drives translocation — the ribosome ratchets forward by exactly one codon. The tRNA that just handed off its chain slides from A to P, the now-empty tRNA slides from P to E and is ejected, and a fresh codon rolls into the open A site. The cycle is ready to repeat.
E site P site A site
[ exit ] [ chain ] [ arrival ]
| | |
5'... C C C A U G G C A U U U ...3' mRNA
(Met) (Ala) (Phe-tRNA arriving)
read 3 letters -> bond chain onto A-site aa -> shift one codon ->
P-tRNA slides to E (ejected), A-tRNA slides to P, new codon enters A
growing peptide: H2N-Met-Ala- ... -[next residue added at A site]A few honest details keep this from being a fairy tale. The chain always grows from its amino end toward its carboxyl end, and the message is always read 5'-to-3', so synthesis runs in a single fixed direction. It is fast — a bacterial ribosome adds on the order of 15 to 20 amino acids per second — and it is accurate, mis-incorporating only about one amino acid in 10,000, thanks to that A-site checking step. And the ribosome rarely works alone: many ribosomes usually load onto one mRNA at once, following nose-to-tail like cars on a single road, so a single message is translated into many protein copies in parallel. That convoy is called a [[polysome|polysome]].
Act three — termination: a stop codon with no tRNA
The cycle would run forever if nothing told it to stop — so the message ends with one of three stop codons (UAA, UAG, or UGA). Here is the elegant trick: there is no tRNA for these codons. Nothing in the normal pool of adapters can read them, so when a stop codon rolls into the A site, no tRNA arrives and the assembly line stalls with an empty chair. That gap is the signal. Into the vacant A site steps a [[translation-termination-release-factor|release factor]] — a protein, not a tRNA, that is shaped to recognize a stop codon directly and fit where a tRNA would have gone.
Once the release factor is in place, it does something clever with the same catalytic centre that built every peptide bond. Instead of letting the chain be transferred to another amino acid, it lets a water molecule attack the bond holding the finished protein to the last tRNA. That hydrolysis snips the chain free — the completed polypeptide floats away to fold and do its job. Then the rest of the cast disassembles: with help from a recycling factor (and another spent GTP), the two ribosomal subunits separate, the last tRNA and the mRNA are released, and the parts are free to find a new message and start the whole performance over.
Stepping back — the shape of the whole process
Notice how cleanly the three acts mirror what you saw in transcription, and how the ribosome's accuracy is built differently from a polymerase's. Transcription proofreads by backing up and clipping a wrong base it already added; translation does most of its checking *before* committing — a wrong tRNA is rejected at the A site, before any peptide bond is forged. Notice too where the energy goes: nearly every step that matters (loading a tRNA, translocating, recycling) is paid for in GTP, which is why building protein is one of the most energy-hungry things a cell does. The machine is not free to run; it is bought, one codon at a time.
Two honest pointers before you climb on. First, the line that the chain emerging from the ribosome is a finished, working protein is a useful simplification, not the truth: a fresh polypeptide is a floppy thread that still has to fold, and often be modified, before it works — which is exactly where the next rung of this ladder picks up. Second, resist the tidy fiction that one gene means one protein. The same coding message can be read into more than one product (through alternative splicing upstream, and choices at the start codon), and the chain can be cut, decorated, and edited afterward. The three-act story you now hold is how *one* protein gets built; it is the spine of gene expression, but the body around it is richer than a single straight line.