DNA is a twisted ladder, and each rung is a pair of letters that only click together one way: A with T, C with G. To copy itself, the ladder unzips down the middle, and each lonely half grabs new matching letters floating in the cell. Because the letters only fit their partners, each half rebuilds the exact strand it lost — one ladder becomes two identical ones.
Most people imagine DNA being copied like a photocopy. In fact the ladder unzips and each lonely half grabs new letters — and because A only fits T and C only fits G, each half rebuilds the exact strand it lost.
What's actually happening
DNA stores its information as a sequence of four chemical letters (A, T, G and C) strung along a twisted double ladder. The genius of the structure, spotted by Watson and Crick in 1953, is that the rungs are picky: an A on one side will only bond to a T on the other, and a C only to a G. So the two strands aren't a random pair; each is the exact negative of the other. Tell me one strand and I can write the second without seeing it.
That pickiness is the whole copying mechanism. When a cell needs to divide, an enzyme unzips the ladder down its middle, breaking the rungs and leaving two single strands with their letters exposed. The cell is full of loose A, T, G and C units, and they drift in and snap onto the open strands — but only into their matching slots, because no other pairing fits. An A on the template can only catch a T; a G can only catch a C. Each half-ladder thus rebuilds the partner it just lost, and where there was one double helix there are now two, each identical to the original. The drawing shows it directly: drag the zip open and watch each strand grow back its missing side.
The fidelity is staggering. Your genome is about three billion letters, and it gets copied every time one of your ~30 trillion cells divides — yet the copying enzyme, with a built-in proofreader that backs up and fixes its own mistakes, errs only about once per billion letters. Those rare errors that slip through are mutations: almost always harmless, occasionally harmful, and very occasionally the raw material that, filtered by natural selection over billions of years, turned single cells into everything alive.
The pairing rule makes each strand the exact negative of the other, so one ladder becomes two identical ones with an error only about once per billion letters.
- 1Write a short strand of random letters using only A, T, G, C — say, ATGCCGTA. That's your template.
- 2Underneath, write each letter's partner using the rule A↔T, C↔G. You'll get TACGGCAT — the complementary strand, built with no guesswork.
- 3Now cover the original and rebuild it from the complement using the same rule. You get the original back exactly. That guaranteed reversibility is why one strand can perfectly dictate the other.
Common questions
Because an A only bonds to a T and a C only to a G, each strand is the precise negative of the other. Tell me one strand and I can write the second without seeing it — which is the whole basis of copying.
Staggering. The roughly three-billion-letter genome is copied with a built-in proofreader that fixes its own mistakes, erring only about once per billion letters.
The rare copying errors that slip past the proofreader. Almost all are harmless, occasionally harmful, and very occasionally the raw material that natural selection turned into the diversity of life.