PCR Problems

Dr S T May lsrei at csv.warwick.ac.uk
Sun Feb 2 08:57:36 EST 1997

I'd just like to comment on the statement that 1min extension
is not enough time to amplify a 3-10 KB fragment in PCR.

Frequently, people use cycles of 1min; 1min; 1min or even higher
with the slightly confused (not to say 'old fashioned') idea that
as taq extends at approx. 72n/s at 72C, one should calculate the
extension time as a multiple of bp(n)/72 seconds thereby obtaining
enormously long cycle times.


Taq does not work by synthesizing the entire length at once,
it falls off frequently (this is the processivity of the enzyme)
and the same, or another molecule of taq then binds and continues

On each cycle, the previously synthesised partial fragments have
another opportunity to anneal and be continued or 'finished off'.

In effect, this means that the extension time is not as important
as many people assume.

A 15s extension time is quite sufficient to amplify fragments of
at least 4-5KB in practice.

In addition, extended denaturing times do nothing to improve
denaturation in the overwhelming majority of PCRs. Even with 
genomic PCRs, after a small number of cycles, the more easily
denatured 'new' fragments are the predominant template, so an
argument of genomic DNA being hard to denature doesn't wash.
The only thing that long denaturing times achieves is the more
rapid depletion of viable enzyme.

Again 15s is easily sufficient (it takes many cyclers this long
to get to 94/95C)

Lastly, an extended annealing time gives a false annealing temperature.
Consider, a primers **efficiency** of binding is affected by temperature,
it is not 'all or none'. By increasing the time at a given annealing
temperature, one is allowing less efficiently binding primers a longer time
to anneal. In other words, it is as if you have effectively REDUCED the
annealing temperature.

To make annealing temperatures MEAN anything, they must be short, if the
temperature is correct for efficient annealing, it will happen in
milliseconds given the huge quantity of primer available in PCR reactions.

Again 15s (a realistic time to get to temperature) is sufficient.

So, a cycle of 15s 15s 15s (preferably with hot start) is suitable
for almost any PCR (with the possible exception of PCRs incorporating
the slower proof-reading enzymes such as pwo).
AND they take approx 2.5 hours instead of tying up machines and
wasting time like the 5-8hr monsters people often do.

This is based, not on theory, but on practical application, and the 
personal use of PCR for nearly a decade. Try it, you will be pleasantly
surprised. Your 4KB PCR WILL work with these parameters (although it is
'possible' you may need to adjust your annealing temp by a degree or so)

I hope this helps someone.


(BTW: another issue would be the other wasteful use of 50-100ul PCRs
they are NOT necessary and are promoted normally by companies that
wish you to use more of their enzymes, or so it would seem. 
10-20ul PCRs, when efficient i.e. with sensible primers, produce
ugs of DNA, if you need more than that - what on earth are you doing
with all that DNA. In many cases a larger PCR volume gives less
efficient product simply because the reaction has a significant
gradient of temperature across the liquid volume. The reaction is
thereby NOT proceeding as intended.. the annealing temperature in the
middle is lower ... etc..etc..) 
Sean May  [sean.may at warwick.ac.uk]
Warwick University, Coventry, U.K.

More information about the Mol-evol mailing list