Agarose gel electrophoresis on PCR product

Brian Foley brianf at med.uvm.edu
Thu Jun 2 08:42:09 EST 1994


Michael Joiner (joiner at pobox.upenn.edu) wrote:
: I am using a fairly standard PCR protocol to produce DNA which I am then
: checking by running the reaction product on a 0.5X TBE 2% agarose gel,
: identifying the band of interest, then cutting this out and extracting the
: DNA with the GeneClean kit. I have two problems.
: First, before running on the gel, I reduce the volume of the PCR reaction
: product from 40 microliters by adding NaOAC to 300mM, 2 volumes of EtOH,
: precipitating and spinning, then resuspending the "pellet" in 10
: microliters of water to which I add 2 microliters of "Manniatis type II"
: loading buffer. However, many times this has produced a gel which looks as
: if much of the DNA has remained in the well and with considerable smearing
: rather than nice discreet bands. Sometimes though, it does work OK.

	1) There should be no need to concentrate the DNA prior to running
the gel.  In fact, it is probably detrimental.  If you are using a 2% gel,
I assume you have a very small PCR product, like less than 400 base pairs,
and this will not precipitate as eficiently as larger DNA.  A good PCR
reaction should produce enough DNA that 1/10 th of the reaction should
give you a nice strong band on a gel.  If you are not getting that, then
there is something wrong with the PCR reaction and you should fix that,
rather than trying to load the whole reaction in one lane.
	2) I doubt it is the precipitation step that causes some samples
to look bad, while others look good.  I bet that if you run a fraction
of the reactions prior to precipitation, you will see that the same
ones that look good after precipitation, also look fine before 
precipitation.

: The second problem is that a further round of reamplification on the
: GeneCleaned product has never produced decent results. I can't understand
: why this should be.

	1) Why are you re-amplifying?  In my experience this is a very 
bad idea.  Taq and other polymerases can mis-incorporate and lead to
"PCR artifact" mutations.  You should try to get the first round of
PCR to give a good strong product, rather than trying to re-amplify.
	2) GeneClean is quite poor at recovering very small fragments
of DNA.  Expect somewhere between 20% and 50% recovery.
	3) What does your re-amplification look like?  If you are
getting a smear of larger size DNA than expected, it is because you
are adding too much template and going too many cycles.  I find that
1/100 th or less of the first reaction product is plenty.  If you
use too much template, the second reaction will give you a good product
band after about 5 cycles, but this will dissapear by 10 or 15 cycles.
I discovered this by taking 1/10 th aliquots out of the reaction tube
after 5, 10, 15, 20 and 25 cycles and running each on a gel.  If I uses
picomole quantities of template the peak ocurred at about 15-20 cycles,
but if I used micromole quantities of starting template the peak was at
5-10 cycles and there was no product of the correct size remaining after
15-20 cycles.
	This also applies to amplification from plasmid DNA.  Too much is
not a good thing.  Too much template, too many cycles, or too much
polymerase can all result in poor yeild of the desired product.
Titrate your template, your polymerase, and the number of cycles.  
You will find that a little is good, and more does not make it better.

: Any advice on these two problems would be greatly appreciated.

	Please let us know what you discover.  We are all here to
learn from each other.  The purpose of this list is to hand out
handy little tips that may not be worth a full publication.

: -- 
: Mike Joiner
: Radiation Oncology
: University of Pennsylvania                   joiner at pobox.upenn.edu

--
********************************************************************
*  Brian Foley               *     If we knew what we were doing   *
*  Molecular Genetics Dept.  *     it wouldn't be called research  *
*  University of Vermont     *                                     *
********************************************************************



More information about the Methods mailing list