Duplicate regions in site-directed mutagenesis

I'm fixing several cDNA clones with point mutations using
site directed mutagenesis. I'm using the Stratagene
QuickChange protocol but using KOD HiFi polymerase
rather than Pfu Ultra because I can get almost 100%
product every time with KOD. After PCR I do a DpnI digest.
All the clones I sequence have the point mutation fixed.
However, a high percentage of the clones have the mutated
region duplicated--exact copy of the primers used. It
looks like a tandem region.

I've seen only a little talk of this happening and why, but
no real solutions to the problem. I synthesize the oligos
myself and desalt them prior to use. Of the 4 clones I pick
for each "fixit" reaction, all of them will have this duplication.

KOD is very fast and for a 6kb plasmid, my extension time
is only 40 seconds. After 23 cycles I get tons of product.
However, I see the same problem with Pfu, too.

Any ideas? Thanks, Paul

Paul Shinn wrote:

I have never seen this problem, and I have done numerous mutagenesis
using the Quikchange protocol. Are there sequences on the oligos that
may prime to each other, and do you do a hot start? Check the
duplicated sequence to see if the sequences are completely duplicated
or only partially duplicated, if it is the latter, then I'd think that
the oligos are annealing to each other.

[Helen Troilo]

I realise the question is a million years our of date- but since this is one of the first things to come up when you google this problem I'll answer since I've had it happen and I think I've just figured out why.

If they are exact duplicates like mine (non-overlapping) are your primers 5' phosphorylated? This is required for some systems such as KAPA mutagenesis. If they are and your PCR cleanup wasn't 100% effective it may be that when you ligate the vector you are also ligating a few primers into it.

This has just happened to me and is my working theory.