la pcr complexes (a network)

Thu Mar 9 10:31:55 EST 1995

Alan Escher wrote:

> We are amplifying a 6 kb fragment from mouse genomic DNA. No problem, but
suddenly from one day to the next all we can see is a high MW complex 
stuck in the well. This phenomenon seems to be the result of the amplifi-
cation reaction. Does anyone know what's happening? 
Wayne Barnes described this as the typical failure mode of la PCR in his
paper on the subject.  I suspect that it's a network made up of your
intended PCR product along with rearranged and truncated versions of 
the same.  You can get the same result on ordinary PCR of small templates
by starting with too much template, overamplifying, or using excessive
extension times.  I think you get it whenever you accumulate partially
extended products en mass.  This happens in long templates because of
misincorporation making mismatched 3' ends that are hard to extend.  
That's the reason behind adding the proofreading enzyme.  It also 
happens if you conduct amplification under conditions where product/
product reannealing effectively competes with extension.

What happens next isn't totally clear to me.  I once thought that
inappropriate priming of partial products leading to strand switching
or other rearranged products was involved in making molecules that
would link the desired product into a network.  However, maybe that
isn't necessary.  If you have enough partially extended stuff, it
will reanneal leaving a sticky end that anneals to another molecule
leaving sticky ends, etc.  That may be enough to suck your entire 
product into a network without any further action.  

There is some reason to think that it's a run-away process.  That is,
it's hard to make the abberrant product in the first place; but once
you  make it, it interferes by annealing to the template in later
cycles and promotes making more abberrant product.  This would explain 
why it often sets in late in the rxn and is sporadic in its

The bottom line is you're trying to avoid conditions that accumulate
partially extended strands.  For long templates, that could include
any of the following:

1) Extension time too short
2) Not enough proofreading activity
3) Amplifying past the point where product/product annealing is getting
in the way of extension [either too much starting template or too
many cycles]
4) Anything suboptimal about the rxn conditions that would make the 
polymerase go slower than expected, leading to problem #1 above.
5) Too little primer (losing the kinetic advantage of priming over
product/product annealing thus reaching the point in #3 above sooner).
6) Seeding the reaction with rearranged or partial templates, say
by reamplifying from a previous rxn that already contained some
aberrant product.

This is mostly theory; I don't know of much data that has been
accumulated about this problem.

Hope this helps.
Steve Hardies, Assoc. Prof. of Biochem., Univ. of Texas HSC at San
Hardies at

More information about the Methods mailing list