>Variable PCR results - Ideas and Suggestions?

[Mr. T. Weissensteiner]

Dear Netters

Doug Jones (jonesd at telenet.net) wrote:

>>I have been having a problem with an RT-PCR for murine IL-6. After
>> optimization, which went well, I have been unable to obtain consistent
>> results, even in the same run.
>>
>> The PCR is done in a mix using 1.5 mM MgCl, 50mM KCl, 10mM Tris, 0.1%
>> Triton X-100, .1% NP-40, .1% Gelatin, 0.25 mM each dNTP's, 5 U Taq
>> (Promega) and primers in a volume of 50 microliters. The cycle is 1
>> min at 95, 1 minute at 55, and 1.5 minutes at 72 degrees, for 35
>> cycles in a Stratagene RoboCycler.
>>
>> The problem is that there is run-to-run inconsistency, with bands
>> either appearing or not in the spleen cDNA, and within run
>> inconsistency [...]


To the many good points that have been made I would like to add 
the following:


1. Bob Steinberg wrote:

> [ ...] this sounds like sample well variability within the thermal 
> cycler, either because the wells are not uniform or because the 
> tubes are not making equal contact with the wells

This can be a problem especially with small sample volumes and with 
PCR tubes whose tips do not fit the heat block wells as well as the 
sides. Try adding oil into the heat block wells (but leave the outer
wells empty, as the oil will expand when heated, and may spill into
the cycler !)


2. Alexander Dobrovic and Duncan Clark noticed the unusually high 
concentrations of Taq and dNTPs.

I remember a paper in the former "PCR Methods and Applications" claiming 
that increasing both Taq and dNTPs substantially increased the PCR product 
yield. I doubt, however, that this will work for all templates. 

Any comments ?


3. Warren Gallin wrote

> You have too much dNTP. Anything over 200 micromolar inhibits the 
> enzyme. This is not simply due to complexing Mg, it can not be 
> rescued by adding more MgCl2

Is your amplfied IL6 fragment "GC-rich" (> 70%) ? Such amplicons 
can be very sensitive to cation concentrations and dNTP come as salts. 
Adding more salt (MgCl2), of course, won't help.

In this case, however, you should try adding some dsDNA destabilizing
additive, such as glycerol, or betaine.
Moreover, this is a good alternative to a high denaturation 
temperature, as it will be much kinder to the Taq polymerase.


4.) "Allelic dropout" can cause the random failure of PCR reactions with low 
template concentrations.


Here are some examples from the extensive and growing list of 
publications on PCR failure. They contain more details on the above points 
(and if nothing else, show at least that Doug is not alone with
his problem :-) ).

PS Walsh, HA Erlich & R. Higuchi : Preferential amplifiaction of
alleles: mechanisms and solultions
PCR Methods and Applications (1992) 1, 241 - 250

GL Mutter & KA Boynton : PCR bias in amplification of androgen
receptor alleles, a trinucleotide repeat marker used in clonality studies.
Nucleic Acids Research (1995) 23, 1411 - 1418

Weissensteiner,T, Lanchbury,J.S. (1995) How to control preferential amplification in PCR typing: Lessons from an assay for alleles of the HLA-B locus. Europ. J. Immunogenet. 22, 100

Weissensteiner,T, Lanchbury,J.S. (1996) Strategy for controlling 
preferential amplification and avoiding false negatives in PCR typing 
Biotechniques 21, 1102 - 1108

Pai,C.Y., Chou,S.L., Tang,T.K., Wei,Y.H., Yang,C.H. (1996) Prevention
of false results from preferential amplification of VNTR alleles. 
J. Formosan Med. Assoc. 95, 69 - 72

Rosner,A., Maslenin,L., Spiegel,S. (1997) The use of short and long PCR 
products for improved detection of prunus necrotic ringspot virus in 
woody plants. J. Virol. Meth. 67, 135-141

Weissensteiner, T. / Taberlet,P. (1997) Guidelines for microsatellite 
allele typing. Nucleic Acids Research 25(3), Scientific Correspondence


Good luck!


Thomas Weissensteiner
Autoimmunity Group
Jenner Institute for Vaccine Research
Compton / Newbury
Berkshire RG20 7NN

T    :  0044 1635 577920
FAX  :  0044 1625 577901

email : tweissen at RETURNMEhgmp.mrc.ac.uk
(please remove capitalized letters when using this address)

        ---{#$:>  <:$#}-~~