1Semiconservative replication

When the double helix is opened, each old strand still carries the full sequence, written in complementary form. Free nucleotides pair with the exposed bases of each old strand, A with T and G with C, and are joined into a new strand. The result is two DNA molecules, each made of one old (parental) strand and one new strand. That is what semiconservative means: each new molecule conserves half of the old one.

The two molecules are identical to each other and to the original, because the pairing rule allows only one partner for each base. So the information is copied, not just the material.

The evidence. If you grow bacteria on a heavy form of nitrogen until all their DNA is heavy, then move them to light nitrogen, the DNA can be separated by density. An invented set of results:

After one round every molecule is a hybrid: one heavy (old) strand and one light (new) strand. That rules out conservative replication, which would give one wholly heavy molecule and one wholly light one (two bands, not one). After two rounds the hybrids have each given one hybrid and one light molecule, so half are hybrid and half light: exactly what semiconservative copying predicts. Notice that the hybrid fraction halves every generation, because the two original strands are never lost; they just end up in a smaller and smaller share of the molecules.

2The direction rule: 5′ to 3′

Each strand of DNA has two different ends. The 3′ end finishes with the sugar's free –OH group; the 5′ end finishes with a phosphate. DNA polymerase can only add a new nucleotide to a free 3′ end. So every new strand grows from its 5′ end towards its 3′ end, which we write as "synthesised 5′ to 3′".

Since the new strand is antiparallel to its template, DNA polymerase reads the template 3′ to 5′ while it builds the new strand 5′ to 3′. Hold on to that pair of directions: it explains the whole shape of the fork.

3The enzymes at the fork

Figure 1 shows a replication fork, the Y-shaped region where the helix is being opened and copied. Old strands are teal, new strands amber, RNA primers white.

• Helicase unwinds the double helix and separates the two strands by breaking the hydrogen bonds between paired bases. It moves along the DNA, so the fork moves with it.
• Topoisomerase works ahead of the fork. Unwinding a twisted rope at one point over-twists the part beyond it; in DNA this tension is called supercoiling. Topoisomerase cuts the DNA, lets it relax, and rejoins it, so that helicase can keep going.
• DNA polymerase builds each new strand, adding nucleotides that pair with the template, always to the 3′ end. It cannot start a strand from nothing: it needs an existing 3′ end to add to.
• RNA primers provide that start. A short stretch of RNA, paired to the template, gives DNA polymerase its first 3′ end. The primers are later replaced by DNA.
• Ligase joins pieces of DNA together by sealing the gap in the backbone between them. Its main job here is joining the fragments of the lagging strand.

The CED names these five, and only these: helicase, topoisomerase, DNA polymerase, ligase, and (for transcription) RNA polymerase. The names of the other proteins at the fork are not examined.

4Leading strand and lagging strand

Helicase opens both templates in the same direction, the direction the fork is moving. But DNA polymerase can only build 5′ to 3′, and the templates are antiparallel. So:

• On one template, building 5′ to 3′ means moving towards the fork. DNA polymerase can follow helicase continuously, from one primer. This is the leading strand, made continuously.
• On the other template, building 5′ to 3′ means moving away from the fork. DNA polymerase can only build a short piece before it runs into the DNA it made earlier. As the fork opens more template, a new primer is laid down near the fork and another short piece is built. This is the lagging strand, made discontinuously, as a series of fragments (you may see them called Okazaki fragments; the name is not needed).
• Ligase joins the fragments of the lagging strand into one continuous strand.

In Figure 1, find the leading strand at the top, with one primer at its far 5′ end, and the lagging strand below it, with the newest fragment just started from a primer near the fork and ligase sealing two older fragments.

DNA polymerase only adds to a 3′ end, so it needs a primer to start and builds 5′ to 3′; that is why one strand is continuous and the other is made in fragments.

5How fast, and how accurate

Replication starts at places called origins of replication and runs in both directions from each one, so each origin makes two forks. A typical bacterial chromosome has one origin; a eukaryotic chromosome has many, which is how very long linear chromosomes are copied in a few hours.

Worked example (invented numbers). A bacterium has a circular chromosome of 4.2 million base pairs and one origin. Each fork copies 700 base pairs per second. How long does replication take?

• Two forks work at once, so the rate is 2 × 700 = 1,400 base pairs per second.
• Time = 4,200,000 ÷ 1,400 = 3,000 seconds = 50 minutes.

A question can also turn this round: if the same genome were copied in 25 minutes, the rate per fork must have been doubled, or there must be twice as many forks.

DNA polymerase checks each base as it adds it and removes most wrong ones, and other repair systems fix most of the rest. The few errors that remain become mutations (topic 6.7), which is one source of new genetic variation.