PCR of supercoils-observations

[Dima Klenchin]

Jim.Hartley at invitrogen.com ("Hartley, Jim") wrote:
>I recently posted a question about improving the efficiency of PCR of a
>plasmid template.  The only treatment that really worked well was to preheat
>the DNA in TE, 98 C for 5 min, then cool and add the rest of the components
>and go straight to cycling.  The same heating in amplification buffer was
>not effective.  I estimate the yield was about 1.8 fold per cycle, 400 fold
>in ten cycles.  This yield was approximately the same as the same plasmid
>linearized.  Pretreatment with topoisomerase (in amplification buffer) did
>not help.  Several different enzyme cocktails did approximately the same
>(YieldAce and Pfu Turbo from Stratagene, ThermalAce from Invitrogen).  Added
>magnesium did not help.  Several amounts of the enhancer PCRx did not help.
>We'll do a little more work on timing and cooling, if we see anything
>striking I'll post.

It's an interesting observation and something that makes sense  (supercoiles
being topologically not a good substrate), but estimates liker 1.8X/cycle, 
400X/10 cycles simply do not make any sense because after first couple of 
cycles your template is effectively not a supercoil at all. 

IMHO, if you are to make something resembling science about this issue, 
you'd have to present real-time amplification graphs together with 
their dependence on initial template concentration and some important 
implications of it. (Say, if this difference, with regard to a final product
yield,  only exists between 1/10000th  and 1/1000th part of miniprep - 
then I doubt anyone cares). 

        - Dima