In molecular biology, there are often occasions where you might want to introduce a specific mutation into a particular gene. Let's say you are examining a particular protein, and you want to find out how crucial a certain cystine residue is. You might want to change that cystine residue into, perhaps, a tyrosine residue, and observe how the protein behaves. You could do this by random mutagenesis and sifting through thousands of mutants until you find the one that you need, or you could do it a much quicker way.
The Kunkel method is one such way, and it is actually quite simple, both practically and theoretically. The first step is to clone the gene you want to mutate into whatever plasmid you choose to use. You can use whatever cloning procedure you wish. It doesn't really matter how you clone, just as long as you have a plasmid with your gene in it.
Next, the plasmid must be transformed into an E.coli strain that is ung- dut-. The dut gene encodes for dUTPase, an enzyme that prevents the bacteria from incorporating uracil during DNA replication (remember that uracil pairs with adenine but only in RNA. It is dUTPase that prevents uracil from being used in DNA by destroying all the cell's reserves of uracil during replication). A strain that lacks ung (dut-) will then randomly add uracil to your plasmid when it replicates. E.coli, however, have a backup mechanism in case dUTPase fails. This is uracil deglycosidase, encoded by the ung gene. Uracil deglycosidase cleaves out any uracil that has been added to DNA if dUTPase has been slacking on the job. Transforming into a strain that is also dut- will make sure the uracil in your plasmid stays there.
The next step is to design a primer that contains the region of the gene which you wish to mutate, along with the mutation you want to introduce. If you wanted to change a cystine to a tyrosine, then your primer would span the codon for cystine, but contain the necessary base pair changes to turn that cystine into a tyrosine. This primer will anneal to your plasmid when it denatures, even if the primer does not match it's target 100%. Once you isolate your uracil-containing plasmid, you can do PCR using your mutated primers to create hybrid plasmids: each plasmid will now contain one strand without the mutation and uracil bases, and another strand with the mutation and lacking uracil.
The final step is to isolate this hybrid plasmid and transform it into a different strain that does contain the ung gene. The uracil deglycosidase will destroy the strands that contain uracil, leaving only the strands with your mutation. When the bacteria replicate, the resulting plasmids will contain your mutation on both strands. In essence, you have completely replaced the original gene with your mutated version! You are then free to do as you wish with your new mutated gene.
The Kunkel method, however, is a bit outdated. Many bioscience companies offer kits that allow you to do site-directed mutagenesis even easier. Stratagene, for example, provides a kit that works as follows:
1) Transform your plasmid with gene of interest into any regular laboratory strain. Most strains will be dam+. The dam gene allows the bacteria to methylate the DNA in the plasmid (that is, they will add methyl groups to spots on the DNA. This is a mechanism that the bacteria has to allow it to determine what DNA is its own, and what DNA might be from an invading virus, since viral DNA will not be methylated).
2) Isolate the methylated plasmid, and do PCR with primers containing the mutation you wish to introduce. This will create hemimethylated plasmids: one strand (the one with the original gene) will be methylated and one (the one with your mutated gene) will not be methylated.
3) Add a small amount of the enzyme DpnI to the reaction mix. DpnI actively cleaves up methylated DNA. This will destroy the strands from the original plasmid and leave only your mutated strands.
4) PCR the remaining fragments to produce complete plasmids that contain your mutated gene.
And there you have it. Now you can mutate any gene you want in any manner you wish, and only take a day to do it. No more screening thousands of mutants! Huzzah!