DNA Replication

Contents

DNA replication

Replication of DNA is semiconservative: each daughter helix contains one strand of the parental helix. (Conservative replication would be where the entire parental double helix was used as a template for the formation of an entirely new daughter double helix). The classical experiment by Meselson and Stahl that proved this used isotopes of nitrogen: Bacteria were grown on medium containing only the 'heavy' isotope of nitrogen, 15N, until all their DNA was 'heavy'. The bacteria were then transferred to 14N-medium for several generations. In each generation, buoyant density centrifugation (in CsCl) was performed on a DNA extract from the bacteria. Heavier DNA molecules spin down further in the gradient tube.

In the zeroeth generation, all the DNA was heavy. After one generation, all the DNA was intermediate in mass ('hybrid', having one heavy chain and one light). After two generations, half the DNA was hybrid, and the other half light. This is consistent with semiconservative replication.

Meselson Stahl experiment.

S phase

In eukaryotic cells, DNA is replicated exclusively during S-phase. At the start of S-phase, each chromosome is a single double helix. By the end of S-phase, each chromosome contains two double helices (chromatids) bound together at the centromere.

S-phase takes 3 min to 8 hr (!) to complete in eukaryotes, depending on the number of origins of replication being used. C-phase (the equivalent in prokaryotes) is usually 40 min. This can be determined using 3H-thymidine autoradiography: cells in S-phase will readily incorporate radioactive thymidine into their DNA. Thymidine is specific for DNA synthesis as thymidine is not required for the RNA synthesis that occurs during the rest of the cell cycle.

The induction of S phase is under the control of soluble inducers. If a G1-phase nucleus and S-phase cytoplasm are fused, DNA replication begins in the G1 nucleus. However, if a G2-phase nucleus and S-phase cytoplasm fused, no replication occurs in the G2 nucleus. This indicates that the S-phase cell contains inducers of DNA replication and that the G2 nucleus is not susceptible to these inducers (whereas the G1 nucleus is). These inducers are cyclins and cyclin dependent kinases.

Origins of replication

The origin of replication is a length of DNA into which proteins can insert and form a 'replication bubble' by loading in DNA polymerase. Two replication forks migrate from the bubble outwards, synthesising DNA as they go. These forks can be readily visualised using 3H-thymidine autoradiography in Escherichia coli: increasingly large bubbles are seen with time. The entire genome (5 × 106 bp) is copied in about 40 min, which means a speed of about  5 × 106 ⁄ ( 40 × 60 × 2 ) ≈ 1000 bp s−1. Prokaryotic chromosomes only have one origin of replication.

Prokaryotic replication bubbles.
Prokaryotic chromosomes only have one origin of replication. If autoradiographs are used to monitor the replication of a circular chromosome, ever-larger radiolabelled 'bubbles' are seen as the forks migrate outwards from the origin.

Eukaryotic origins of replication can be viewed under EM with negative staining: a consistent speed of replication bubble expansion of c. 14 µm min−1 is observed. Given that a base pair is about 3.38 nm 'tall', the speed is about 14 × 103 ⁄ 3.38 × 60 ≈ 70 bp s−1, which is much slower than in prokaryotes.

With only one origin, this indicates replication of a single average chromosome (150 × 106 bp) would take c. 800 hours! Eukaryotes must have more than one origin per chromosome: in fact, they have up to 60 (on average one per looped 300 kbp domain).

Origins sequences can be studied by adding potential origin sequences to origin-free plasmids in yeast containing the HIS gene (mutants for this locus are unable to synthesise histidine). These plasmids are then transformed into HIS mutant yeast on histidine-free medium. Significant growth of transformants is only observed if the sequences allows the plasmid to replicate independently.

HIS mutants can be used to investigate origins of replication.

Yeast have an origin of replication protein complex, that binds a well characterised 150 bp DNA sequence. There are e.g. nine origins on chromosome III (which contains only 180 genes; making it one of the smallest chromosomes known). Human origins of replication and replication complexes have been very difficult to identify. Human origin sequences are longer and 'fuzzier'.

Bromodeoxyuridine (BrdU) can be used to label portions of a chromosome being replicated. BrdU is added for brief periods to synchronised cells during S-phase, where it is incorporated into DNA replicating at that point. We can then look at chromosomes during M-phase using anti-BrdU antibodies. These studies reveal the order of replication to be reproducible and coordinated:

Early origins in yeast can be moved about, and sometimes this makes the DNA around them replicate earlier, but not always.

Origins of replication are only activated once per cell cycle. In eukaryotes, blocker proteins (geminin) are added after the bubble has been blown, and these are only removed at beginning of a new cycle. Prokaryotes use a more elegant system based on methylation. In E. coli, a DNA methylase adds methyl groups to the GATC sequence in the ori sequence. Recently replicated DNA is only methylated on the parental strand, and this hemimethylated DNA is resistant to initiation because it won't bind initiator proteins. There is a sufficient lag before the methylation of the daughter strand to prevent the origin being reactivated too soon.

Methylation of DNA regulates origins in prokaryotes.
The hemimethylated DNA produced after DNA has been replicated (but before DNA methylase can acts on it) is resistant to initiation.

The replication fork

What must DNA replication achieve?

Most work on eukaryotic replication forks is done in SV40 (a monkey oncovirus - simian virus 40). Most prokaryotic work done in Escherichia coli.

DNA molecules are tightly bound by hydrogen bonds into a double helix. The two stands of the helix must be separated ('melted') to allow formation of a replication bubble. The 'melting' temperature is how hot dsDNA needs to be to separate the two strands. This is also essential to know for performing PCR reactions in the lab:

Tm ≈ 4×(G+C) + 2×(T+A)

The Tm is also useful as a gauge of the GC to AT ratio, and the Tm of hybrid DNA can be used to judge the similarity of two genomes (chimpanzee and human c. 98%) without actually sequencing it.

The melting point of DNA strands is c. 90°C in vitro; this must be made to happen at c. 37°C in vivo. DNA is somewhat like a piece of tightly twisted string, and proteins are needed to alter its shape (topology) and destabilise the helix. There are many Enzymes that alter DNA's topology:

In SV40, the T-antigen (large tumour antigen) protein (a double hexamer) serves as both replicase and helicase. Each half of the molecule comprises three domains: a helicase, a DNA binding domain, and a regulatory region. Replication bubbles are initiated by this multifunctional protein by its binding to AT-rich sequences in the ori. This both initiates synthesis, and acts as helicase, parting the two strands.

In E. coli DnaA and DnaB proteins play equivalent roles to T-antigen. DnaA (the initiator) binds dsDNA and warps it enough to allow helicase (DnaB) to load.

Prokaryotic origins are initiated by DnaA, which loads in the DnaB helicase.
Prokaryotic origins are initiated by DnaA, which loads in the DnaB helicase.

DNA-helicase (DnaB) parts the helices: DNA-binding tightens when dTTP (the deoxythymidine analogue of ATP) binds to helicase. It is thought that helicase rotates as the dTTP is hydrolysed, screwing itself along the helix, and parting the strands.

DNA helicase parts the dsDNA.
DNA helicase uses the energy of dTTP to screw along a dsDNA molecule, melting the helix and parting the strands.

Single strand binding proteins (SSBPs) stabilise the parted strands, which would otherwise anneal to themselves (or other strands), forming hairpin loops and other unwanted tangles. Human SSBP have two fused homologues of E. coli  SSBP.

SSBPs stabilise the ssDNA.
SSBPs stabilise ssDNA and prevent it from re-annealing inappropriately.

Topoisomerases stop tangles. Topoisomerase I nicks a single strand ahead of helicase, allowing strain-relief, and preventing snarling up of the strands as they are forced apart.

DNA topoisomerase I (before cission)  DNA topoisomerase I (after cission)
DNA helicase nicks a single strand of dsDNA, alowing rotation of one strand around the other, which prevents snarling up (supercoiling) of the DNA ahead of helicase.
DNA topoisomerase I.
Topoisomerase I allows strain relief ahead of the replication fork.

Prokaryotic chromosomes are Möbius strips. Topoisomerase II creates gates by nicking both strands of the dsDNA, allowing helices to cross one another. Tangles and linked loops can therefore be separated. It also forms part of the eukaryotic nuclear scaffold, for similar reasons.

DNA topoisomerase II.
Topoisomerase II allows DNA molecules to pass through one another.

There is a common misconception about DNA replication, illustrated in the diagram below:

This does not happen!
DNA is not synthesised in the 3′ to 5′ direction, as erroneously shown on the left-hand strand.

The thing happening on the left hand strand does not happen in nature. This is because DNA has a direction. It consists of two antiparallel strands with distinguishable 5′ and 3′ ends - the numbers refer to the number of the carbon in the ribose sugar.

With the naļve model shown above, you'd need two different DNA polymerases: one for 5′ → 3′; and one for 3′ → 5′. However, experimentally, we find that all DNA polymerases synthesise the new strand 5′ → 3′ (as shown below, note that the numbers refer to the direction of synthesis of the new strand, not the direction along the template strand that the DNA polymerase runs). Why?

DNA synthesis proceeds from five to three prime.
DNA synthesis only occurs in the 5′ and 3′, so DNA polymerases must move in antiparallel directions to synthesise the two daughter helices.

The reason is that DNA is synthesised from dNTPs. Hydrolysis of (two) phosphate bonds in dNTP drives this reduction in entropy:

DNA replication is driven by the hydrolysis of pyrophosphate.
DNa replication requires dNTPs, whose incorporation into DNA produces pyrophosphate. The enzyme pyrophosphatase ensures that this is rapidly hydrolysed to two phosphate ions. This hydrolysis drives the synthesis of DNA: the decrease in the entropy of the DNA is driven by the pyrophosphate hydrolysis, which is energetically favourable, spontaneous and extremely rapid (Le Chatelier's principle).

Binding of nucleotides has error of c. 10−4, due to extremely short-lived imino and enol tautomery. However, the lesion rate in DNA is only 10−9. This increased accuracy is due to the fact that DNA polymerase can chew back mismatched pairs to a clean 3′ end using its built-in 3′→5′ 'proof-reading' exonuclease activity.

Imino-cytosine pairs with adenine, but exonuclease activity detects and corrects this mis-pair.
Imino-cytosine pairs (incorrectly, as far as the cell is concerned) with adenine. Nanoseconds later, the iC converts back to the normal amino form of cytosine, and no longer pairs with adenine. The mismatched base sticks out and is cleaved off by the exonuclease activity of the enzyme, leaving a fresh OH group to try again.

In prokaryotes, other enzymes scout unmethylated (new) DNA for lesions, and correct even more such errors.

So why does DNA polymerase always make DNA 5′→3′ If DNA were synthesised 3′ → 5′, the 5′ end of the strand would carry a triphosphate group. Chewing back would not restore the triphosphate group, and this would stop DNA synthesis. This is why DNA polymerases are always unidirectional.

Exonuclease activity could not occur if DNA polymerase could make DNA in the 3 to 5 prime direction.
If 3′ → 5′ synthesis were possible, the DNA molecule would have a triphosphate group at the end. If for some reason the last nucleotide had to be removed, this would leave no triphosphate group at the end, and DNA synthesis would stop.

All organisms produce a number of different DNA polymerases, ranging from small monomers to enormous (1 MDa or more) multimers. The smaller enzymes are often involved in DNA repair, the larger in actual replication:

Prokaryotes:

Eukaryotes:

The enzymes often have 5′→3′ exonuclease activity as well as proofreading 3′→5′ to remove primers. The Klenow fragment is a 5′→3′ exonuclease-free hydrolysis product of DNA polymerase I that is widely studied as a model DNA polymerase system. It is shaped approximately like a right hand.

DNA polymerase Klenow fragment - click for Jmol version
The Klenow fragment of DNA polymerase III has a shape somewhat like a right hand.

DNA polymerase is held onto the DNA by a doughnut-shaped sliding clamp (processivity factor).

Sliding clamp shaped like a doughnut.
DNA passes through the hole in the 'doughnut'.

This is loaded onto the DNA by a clamp loader:

The DNA clamp is loaded onto DNA by a clamp loader.
The DNA clamp is loaded onto DNA by a clamp loader. The clamps increases the binding and processivity of the DNA polymerase.

Because the synthesis of DNA only occurs in one direction, different processes must occur on the two strands. These two strands are termed the leading and lagging strands. The leading strand is synthesised continuously 5′→3′. However, the other, 'lagging' strand is still synthesised 5′→3′ but in discrete chunks called Okazaki fragments, from the replication fork back towards the origin.

Okazaki fragments are produced on the lagging strand.
The leading and lagging strands are topologically distinguishable and have different suites of enzymes to replicate them.

DNA polymerase will not synthesise dsDNA from ssDNA: it needs a short double stranded section. RNA primers are therefore added by DNA primase. The primase plus the helicase are often called the lagging strand primosome. In eukaryotes, DNA polymerase α has a primase subunit, rather than a separate enzyme.

Since priming is the most likely point for mismatch to occur because the short helix is unstable, DNA polymerase has been evolutionarily tuned to refuse to bind to ssDNA, and will only bind to dsDNA or a DNA/RNA hybrid. RNA is distinguishable from DNA and can be removed later, whereas a DNA primer would not allow this erasibility. Hence RNA is used.

DNA polymerase produces Okazaki fragments on the lagging strand:

DNA primase adds RNA primers to ssDNA.
DNA primase adds RNA primers to ssDNA. These are extended to Okazaki fragments by DNA polymerase.

DNA polymerase drops off when it hits a primer on the lagging strand. The RNAse-H complex then begins the process of tidying up by removing the primer (in prokrayotes; in eukaryotes, DNA polymerase-α has an inherent primer-removing RNAase). DNA polymerase then fills in the hole and DNA ligase joins the two fragments of DNA together. This palaver also needs to be done when DNA polymerase reaches the next origin along a eukaryotic chromosome (or completes the full-circle in prokaryotes).

Clean-up after DNA replication involves RNAse, DNA polymerase and DNA ligase.
RNA primers are removed using RNAse, DNA is synthesised in the gaps by DNA polymerase and DNA ligase joins the fragments together.

The whole complex of two polymerases, topology-altering enzymes and primases is termed the replisome, and can be viewed directly uder electron microscopy.

A replisome.
Replisomes require the co-ordinated action of two DNA polymerase molecules and their clamps, helicases, primase and SSBPs.

Eukaryotes have one further complication: it is impossible to directly replicate the very end of the lagging strand at the end of linear chromosomes (the telomeres). Consequently, telomeres get shorter and shorter with each generation in somatic cells. However, something must be done to stop this in germ-line cells.

Human telomeres contain 5′-GGGTTA-3′ repeats. In germ-line cells, telomerase (a reverse transcriptase) contains a 3′-AUCCCAAU-5′ oligo-RNA which allows extension of the parental strand. A new strand is then synthesised on this template by normal means. Note that this means there will always be a ssDNA section at the end of a linear chromosome (in eukaryotes this is probably annealed back on itself as a hairpin).

Telomerase extends ssDNA at the telomere.
Telomerase is a reverse transcriptase with its own integral RNA template. It pairs with the telomere sequence and extends the ssDNA of the telomere.

The telomerase RNA pairs with the telomere sequence and the enzyme then extends the telomere by RNA-templated DNA synthesis. DNA polymerase can then make (all but the end) double stranded.

Summary

Test yourself

  1. What adaptations does the replisome possess that increase the fidelity of DNA replication? How are these adaptations better than their alternatives?
  2. Why is the DNA replication fork showing two continuously synthesised strands, so beloved of A-level text books, so wrong?

Answers

Bibliography

Peer Review.
This page has been peer reviewed by 1 person.