Site-directed mutagenesis

Site-directed mutagenesis is a procedure used to induce a specific mutation in a cell. It may be used for a host of reasons, including the generation of restriction sites, investigating the role of a gene or regulatory element by knock-out, understanding the role a particular amino acid has in a protein, or the creation of new and 'better' proteins with, for example, greater thermal stability or more efficient catalytic ability.

The method quite simply involves template DNA - the DNA to be mutated, usually bacterial DNA - and an oligonucleotide carrying the reverse complement of the desired mutation which can anneal to the template and be used as a primer for DNA synthesis. For instance, if a TGG codon is present in the bacterial DNA, and the desired mutation is AAT, then the oligonucleotide primer should read ATT (all read from 5' to 3'). Once the mutagenic primer is annealed to its template, the complete structure is called a heteroduplex, owing to the differences between the strands. The heteroduplex is used to transform a cell - most often E. coli - and it is left there overnight.

In theory, both strands of the heteroduplex should be replicated at equal frequency to give a 50/50 mixture of mutant to template DNA in the cell. In practice, mutant recovery in this way is poor for two reasons: firstly, because of the cell's intrinsic mismatch repair system, and secondly, because the template DNA is methylated, and methylated DNA is preferentially replicated by the host cell machinery. Consequently, higher-efficiency mutagenesis approaches have been developed to raise the percentage mutant recovery from around 0.1% to as high as 50%. These approaches work on the principle that once the template DNA has been used to copy the mutant strand it is of no further use, and can only hinder mutant recovery.


Quikchange is one high-efficiency mutagenesis approach that has been developed.

The plasmid template is denatured and mutagenic primers are annealed to each strand. The primers are de-phosphorylated so that although they can be extended, there cannot be ligation between the end of the synthesised strand and the start of the primer. Once DNA synthesis is complete, the template DNA and mutagenic DNA are denatured in each of the two PCR-like products. The mutant strands cannot be reused for DNA synthesis because when the primers anneal to them, they have no template material to copy. Instead the parental strands are reused: one new mutagenic primer is added to each, DNA is synthesised, the products are denatured and then the parental strands used again. Unlike conventional PCR, which makes products exponentially, this is a linear amplification: two new mutated strands are made in each 'cycle'.

At the end of this amplification period, the parental templates are recognised by a methylation-specific restriction enzyme called Dpn I. Because the mutated DNA is unmethylated, it goes unrecognised by this enzyme and remains intact. The consequence is that parental DNA cannot be used to transform E. coli, while mutant DNA can.

Although currently the mutant products are all linearised, their lengthy (approx 40 nt) complementary primers can anneal to result in a double-stranded circular plasmid containing a homoduplex of only mutant DNA. The E. coli repair system will complete ligation where the dephosphorylated primers fail to do this, to form complete plasmids ready for host cell replication.

A Quikchange protocol might look something like this:

1. Design 2 long (25-40nt) complementary primers containing the mutations
- complementarity is important for circularisation of the mutant plasmid later on
- the melt temperature of the primers should be around 78C
- there is no need for either primer to have a 5' phosphate as there is no ligation step

2. Mix the template plasmid, primers, dNTPs and a thermostable polymerase and run for 16-25 thermal cycles

3. Digest (methylated) DNA with Dpn I

4. Transform E. coli with the remaining DNA and leave overnight

5. Pick four colonies and isolate plasmid DNA

6. Sequence the plasmid to ensure that the mutation has been correctly inserted

Uracil-containing DNA method

This approach is based on the simple notion that (deoxy)uracil is not a usual component of DNA. It involves the following protocol:

1. Grow the template DNA to contain a high proportion of deoxyuracil (dU) by growing it in an E. coli mutant:

- dut- (which lacks dUTPase; an enzyme which normally prevents the incorporation of uracil into DNA)
- ung- (which lacks uracil glycosylase; an enzyme which normally removes uracil from DNA)

2. Anneal the mutagenic primers, as usual, and begin DNA synthesis

3. The next step can either be performed in vivo or in vitro:

In vivo: transform the heteroduplex DNA into a wild-type E. coli which retains its uracil glycosylase function (ung+). The template DNA, which is rich in dU, will then be repaired using the newly-incorporated mutant strand as a template. The product is a homoduplex DNA containing only mutant strands.

In vitro: extract the heteroduplex DNA from E. coli and treat with uracil glycosylase to remove the parental DNA. Then synthesise a new strand with DNA polymerase and dNTPs, using the mutant strand as a template. Both of these are performed in the test tube. The product, again, is a homoduplex DNA containing only mutant strands.

Cassette mutagenesis

Cassette mutagenesis is a technique employed to introduce multiple mutations to the same region of DNA. A cassette (block of DNA) is designed to contain all of the desired mutations and then given ligatable ends to facilitate its insertion into the wild-type DNA. Quikchange, described above, can be used to generate suitable restriction sites for its insertion, and the cassette should have both 5' phosphorylation and 4-base 'sticky' overhangs at each end in order to encourage its insertion into the host molecule.

PCR mutagenesis

PCR mutagenesis is similar in principle to Quikchange. The target plasmid is heated to denature, mutagenic primers (forward and reverse) are added to each strand, and roughly 8 cycles are performed to amplify the mutant plasmid (fewer cycles are ideal to minimise the risk of error). The methylated (parental) DNA is then treated with Dpn I and E. coli is transformed with the mutant homoduplexes and left to grow overnight. Again, the plasmid DNA is isolated from selected colonies and sequenced to ensure that the desired mutation has been incorporated.

Sticky feet PCR

Sticky feet PCR is used to generate insertional mutations in the wild-type DNA. The mutagenic primer contains a series of bases (the desired insertion) which is not present in the template DNA. Because it cannot form complementary pairs with the template DNA upon annealing, the desired insertion 'loops out'. When the primer is extended, to generate heteroduplex DNA, the parental strand is digested, as usual, using Dpn I. This leaves a single-stranded mutant strand, containing the insertion, which then itself acts as a template for new DNA synthesis; the nascent DNA strand will contain the complement of the insertion. The product is homoduplex DNA containing both strands with the insertional mutation.

The size of insertion that can be generated by sticky feet is limited, however, by the size of oligonucleotide primer that can be accurately synthesised (certainly no more than 80 nucleotides in length).

Deletions are performed in a similar manner, except the mutagenic primer lacks the bases which need to be deleted (it contains only the flanking sequences). Upon annealing, this causes the deletion bases in the template DNA to 'loop out' because they have nothing to anneal to. Unlike with insertions, there is no size limitation to deletions because the oligonucleotide only need be big enough to correspond to the flanking regions of the desired site of deletion.