How proofreading work Re: How to avoid PCR errors?

Bob DeLisio delisior at
Thu Feb 17 17:59:22 EST 1994

"chai_z at" "Zhonglin Chai" writes:

>Does proofreading activity correct the mutant primers used in the mutagenesis
>PCR reaction? I meant, that the bases of mutant primers which don't match the
>template might be cut off and replaced by matching nucleotide bases when Vent
>or Pfu being used. I was told that they wouldn't correct the primers, but only
>the sequences being synthesized in between the primers. If so, could any one
>tell me how the proofreading works?
>Zhonglin Chai

Vent polymerase chews back the primers to 15 bases under normal PCR conditions.
Thus, if your primers are 30mers with the mutation(s) somewhere in the middle
chances are very good that you won't get the mutant you're looking for. As a
matter of fact, all polymerases with 3'-> 5' exonuclease activity can and will
potentially degrade primers used in PCR mutagenesis. The question becomes one
of severity.

For mutagenesis, I prefer to use good ol' Taq and adjust the conditions to
favor a fidelitious synthesis. In general, you should:

        1.) use balanced dNTP's (40-50 uM each)
        2.) use as low a Mg++ concentration as possible
        3.) increase the annealing temperature
        4.) decrease the amount of enzyme
        5.) decrease the extention time
        6.) decrease the number of cycles to the fewest possible
        7.) use hot start (optional)

I should be noted that Taq tends to add non-template directed bases to the
synthesizing strand and that the addition is preferential to a single A. Thus,
when synthesizing a large fragment from several shorter ones, I always take
this into account by engineering the overlap as if the A is placed there
automatically because an A:N mismatch will almost always fail to extend in the
next round. Since the final base in the synthesizing strand would be a A,
the template therefore must have a T compliment:

3' -------------------------T---------------------------------------------- 5'
5' ----------------------T------------------------------------------------- 3'

You should have no problems using Taq for your mutagenesis if you follow the
above guidelines. It works for me all the time.

Bob DeLisio
 _/   _/     Robert DeLisio
  _/ _/      Roche Institute of Molecular Biology
    _/       Roche Research Center
   _/ _/     Nutley, New Jersey 07110-1199
  _/   _/
   _/ _/     Voice: (201) 235-3728
     _/      Fax:   (201) 235-2318
    _/ _/    Internet: delisior at
   _/   _/   CompuServe: 73517,2616

More information about the Methods mailing list