replication fidelity and proofreading

Copying three billion letters of a human genome and getting almost none of them wrong sounds impossible, yet cells manage error rates of roughly one mistake per billion or so bases copied. That astonishing accuracy is called replication fidelity, and it is built up in layers rather than coming from any single magic step.

The first layer is base-pairing selectivity: DNA polymerase strongly prefers to add the nucleotide that correctly pairs with the template (A with T, G with C), because correct pairs fit the enzyme's active site far better, weeding out most mistakes as they happen. The second layer is proofreading. Many replicative polymerases carry a separate 3'-to-5' exonuclease activity: right after adding a base, if it senses a mismatch (a wrongly paired, poorly fitting end), it backs up, snips the just-added wrong nucleotide off the 3' end, and tries again. This catch-and-correct step improves accuracy roughly a hundredfold. A third layer, mismatch repair, acts after replication to fix the few errors that still slip through, but that belongs to the repair field.

Fidelity matters because the few errors that survive become permanent mutations passed to all descendant cells. High fidelity keeps the genome stable across the trillions of cell divisions in a lifetime; losing it is dangerous — defects in proofreading or mismatch repair sharply raise mutation rates and predispose to cancer. The honest flip side is that perfect copying would freeze evolution: the rare errors that escape are also the raw material of genetic variation, so cells tune fidelity high but not absolute.