DNA Replication

How the double helix unzips and copies itself with near-perfect fidelity

Biology Genetics intermediate 7 min
Step 1 of 6

Helicase opens the replication fork

Watch the yellow Hel disc near the top of the canvas: helicase sits exactly where the blue and red strands split apart. It uses energy from ATP to break the hydrogen bonds (the faint rungs) between paired bases, opening a Y-shaped replication fork. Helper proteins hold the separated single strands apart, and topoisomerase relieves the twisting strain that builds up ahead of the fork.

Primase lays down a short RNA primer

DNA polymerase cannot start a strand from nothing; it can only extend one. Look for the green Pri disc and the short green bar beside the fork: primase writes an RNA primer about 10 nucleotides long, matched to the template. That primer gives polymerase a free 3′ end to build from. Every new strand, and every lagging-strand piece, begins on a primer like this.

The leading strand grows in one piece

Set the 'Strand' menu to Leading so the other new strand fades. The teal line growing down the right is the leading strand: the purple Pol disc reads its template 3′ → 5′ and adds matching nucleotides (A with T, G with C), building 5′ → 3′ toward the fork without stopping. Drag 'Polymerase speed' from 1000 down to 200 nt/s and watch the teal line grow five times more slowly.

The lagging strand is built in short pieces

Switch 'Strand' to Lagging. This template runs the opposite way, so polymerase can only copy it moving away from the fork, in short bursts. The pink segments on the left are Okazaki fragments, each started from its own primer and roughly 1,000 to 2,000 nucleotides long in bacteria. Set 'Strand' back to Both to compare one smooth teal line with a row of separate pink pieces.

Worked Example: copying a bacterial genome

Polymerase backs up to check the red mismatched base on the canvas. Set 'Error checks' to Proofreading + repair with 'Polymerase speed' at 1000: the readout shows 1 error per 10⁹ bases. A 4.6 million base-pair E. coli genome, copied by two forks at 1000 nucleotides per second each, takes about 2,300 s, or 38 minutes, with roughly 0.0046 expected errors (a model estimate). With None, about 46 errors remain.

Ligase seals two half-old, half-new copies

The last stage shows two finished double helices with the teal Lig disc between them. An enzyme removes the RNA primers, polymerase fills those gaps with DNA, and ligase seals the remaining nicks between Okazaki fragments. Each copy keeps one original strand (blue or red) paired with one new teal strand, which is called semi-conservative replication, and both carry the same sequence.

How It Works — Step by Step

Step 1 of 6
  1. 1

    Helicase opens the replication fork

    Watch the yellow Hel disc near the top of the canvas: helicase sits exactly where the blue and red strands split apart. It uses energy from ATP to break the hydrogen bonds (the faint rungs) between paired bases, opening a Y-shaped replication fork. Helper proteins hold the separated single strands apart, and topoisomerase relieves the twisting strain that builds up ahead of the fork.

  2. 2

    Primase lays down a short RNA primer

    DNA polymerase cannot start a strand from nothing; it can only extend one. Look for the green Pri disc and the short green bar beside the fork: primase writes an RNA primer about 10 nucleotides long, matched to the template. That primer gives polymerase a free 3′ end to build from. Every new strand, and every lagging-strand piece, begins on a primer like this.

  3. 3

    The leading strand grows in one piece

    Set the 'Strand' menu to Leading so the other new strand fades. The teal line growing down the right is the leading strand: the purple Pol disc reads its template 3′ → 5′ and adds matching nucleotides (A with T, G with C), building 5′ → 3′ toward the fork without stopping. Drag 'Polymerase speed' from 1000 down to 200 nt/s and watch the teal line grow five times more slowly.

  4. 4

    The lagging strand is built in short pieces

    Switch 'Strand' to Lagging. This template runs the opposite way, so polymerase can only copy it moving away from the fork, in short bursts. The pink segments on the left are Okazaki fragments, each started from its own primer and roughly 1,000 to 2,000 nucleotides long in bacteria. Set 'Strand' back to Both to compare one smooth teal line with a row of separate pink pieces.

  5. 5

    Worked Example: copying a bacterial genome

    Polymerase backs up to check the red mismatched base on the canvas. Set 'Error checks' to Proofreading + repair with 'Polymerase speed' at 1000: the readout shows 1 error per 10⁹ bases. A 4.6 million base-pair E. coli genome, copied by two forks at 1000 nucleotides per second each, takes about 2,300 s, or 38 minutes, with roughly 0.0046 expected errors (a model estimate). With None, about 46 errors remain.

  6. 6

    Ligase seals two half-old, half-new copies

    The last stage shows two finished double helices with the teal Lig disc between them. An enzyme removes the RNA primers, polymerase fills those gaps with DNA, and ligase seals the remaining nicks between Okazaki fragments. Each copy keeps one original strand (blue or red) paired with one new teal strand, which is called semi-conservative replication, and both carry the same sequence.

Key Takeaways

Semi-conservative, not from scratch

Each new DNA molecule keeps one strand from the parent and builds one new complementary strand. This was elegantly proven by the Meselson–Stahl experiment in 1958 using heavy nitrogen isotopes.

The asymmetry of the two strands

Because DNA polymerase can only work 5′ → 3′, the two template strands are copied differently. The leading strand is synthesised in one go; the lagging strand requires repeated priming and joining of Okazaki fragments.

Extraordinary fidelity

Base selection alone gets about 1 in 10⁵ bases wrong; proofreading by polymerase brings this to ~1 in 10⁷, and mismatch repair to ~1 in 10⁹. In a human cell that means copying ~6 billion base pairs with only a handful of uncorrected errors.

Added
Updated
  • — New connections to Epigenetics and Gene Regulation, which control which copied genes get read.
  • — Added strand, polymerase speed and error-check controls with a fidelity readout; six steps and a genome-copy example.