Strange you should mention that, because just yesterday someone showed me a paper that claims 58-97% mutagenesis with a single primer, no subcloning, no special vectors, and no nucleotide analogues.
The reference is Lai et al, (1993) Nuc. Acids Res. Vol.21, No.7, pp3977-3980
The method involves a very clever trick using 3 pre-existing restriction sites (none of which are lost) and a varient of the gapped-duplex method. I personally have not tried this (yet), but I have second-hand information that it works as well as it claims.
> >I am going run the gel retardation to show the band shifting using my protein and a small fragment DNA. The small fragment is about 20 bases, which is planned to be synthesized. For sure, the synthesis of a 20-mer of oligonucleotide is much cheaper than that of two strands to make a ds DNA.
My question is if some methods can be used to synthesis such small ds DNA fragment (20bp) from the single strand oligonucleotide.
I used to label my probe for gelshift in this way:
Synthesis of the 20mer of interest
Synthesis of a 4 (or 6mer)complementary to the 20mer.
Klenow reaction using the 20mer as a template and the 4mer as a primer.
Usually I use a 3-4 fold molar excess of primer. The radiolabeled nucleotide can be changed according to the base composition of the template to obtain the desired specific activity of the probe. I used this method to obtain high specific activity probes.
The method is simple and even if you need two oligos, one of them is only a 4 or 6mer and therefore cheaper and if you want to save you can design several oligos with the same end and use the same primer for all of them. Check that you don't create an artificial binding site by gelshift.