Plasmid mutation in e. coli?

Dr. Duncan Clark Duncan at
Thu Aug 17 04:51:54 EST 2000

In article <399ac587.963593 at>, nvahsen at hot.[nospam] writes
>There were about 10 to 12 errors compared to the GenBank sequence in
>the 2kB. I think this is too much for Taq/Pwo errors. Also for RT
>(MMLV enzyme) errors.

Even with a Taq/pwo mix, 40 cycles of 2kb will give errors. Add on the
errors from RT and maybe.

For instance. Many moons ago in a Boehringer Biochemica (April 1995)
they showed how they measured the fidelity of their enzymes. They PCR'd
around a 4kb pUC plasmid containing lacI for a calculated 9 duplications
(for Expand). Ligate the PCR product and stick it into DH5alpha plated
on XGAL. Whites are correct, blues are mutations in lacI from the PCR.
For Expand 2.6% were blue, for Taq 7.2% and for Pfu 0.5%. That was for
just 9 duplications. For 40 cycles maybe 35 duplications (PCR is not
100% efficient each cycle) you will have a lot of errors.  

Can you sequence the PCR product directly? Are the errors still in that
rather than the individual PCR molecule you cloned and subsequently
sequenced. If they are then you have to find a way of reducing that and
cloning makes no difference. Of course the published sequence could be

>Is anyone aware how well (in terms of error free) a plasmid is
>maintained in e. coli cells? Other ideas what might have happened?

This is the next problem. If you are cloning this cDNA downstream of a
strong promoter i.e. lac, maybe in reading frame such that you generate
some expression and your gene product is giving E.coli a problem then
the plasmid will re-arrange, delete or the gene will be mutated (not
sure of the mechanism but it happens!) until it can be maintained.

The problem with being on the cutting edge is that you occasionally get 
sliced from time to time....

Duncan Clark
DNAmp Ltd.
Tel: +44(0)1252376288
FAX: +44(0)8701640382

More information about the Methods mailing list