DNA replication

Each parental strand serves as a template for a new complementary strand, so each daughter DNA molecule contains one old and one new strand (semiconservative, Meselson–Stahl 1958). Replication begins at many origins (tens of thousands per human cell cycle). Licensing in G1 — loading of the MCM helicase onto origins by ORC, Cdc6 and Cdt1 — and firing in S phase by S-phase CDK and DDK kinases ensure every origin fires once and only once per cycle.

At each bidirectional replication fork, the helicase (CMG complex) unwinds DNA, single-strand binding protein (RPA) stabilises the exposed strands, and topoisomerases relieve the supercoiling that builds up ahead (topoisomerase I cuts one strand; topoisomerase II cuts both). DNA polymerases cannot start a chain de novo and synthesise only 5′ → 3′, so primase (with DNA polymerase α) lays down short RNA primers. The leading strand is synthesised continuously, mainly by polymerase ε; the lagging strand discontinuously as Okazaki fragments (≈100–200 nucleotides in eukaryotes), mainly by polymerase δ, held on DNA by the sliding clamp PCNA. Primers are removed (RNase H, FEN1), gaps filled and nicks sealed by DNA ligase. Polymerases δ and ε proofread with 3′ → 5′ exonuclease activity, and post-replicative mismatch repair corrects remaining errors.

Because the lagging strand cannot be completed at the very end of a linear chromosome, telomeres shorten with each division (the "end-replication problem"), eventually triggering senescence (Hayflick limit). Telomerase — a reverse transcriptase (TERT) carrying its own RNA template — extends telomeres in germ cells, stem cells and ≈85–90 % of cancers. New histones are deposited behind the fork, and epigenetic marks are copied to maintain cell identity.