September 22, 1999 ECK
Primer design is an art. Here are rules which guide me in designing primers.
I nearly alway do this in conjunction with a program such as "Prime" (GCG).
Primers should have melting temperatures (Tm) between 65 and 70íC
(calculated over the region of specific hybridization -- that is, ignoring
any extra sequence in the 5' tail).
The annealing temperature used during PCR should be 10 to 15íC less than
Primers should be at least 18 bases long.
Strive to use approximately equal numbers of each nucleotide type, and
avoid long runs of the same nucleotide.
The 3' end is the end which extends. Try to make it terminate in C or A,
as neither nucleotide wobbles. This increases specificity.
It is better if the 3' end is slightly AT rich, as it will be less prone
Avoid internal hairpin structures, especially those which engulf the 3'
Never design a primer with 3' self-complementarity, such as
(5')...GATC(3'). Such primers will anneal to each other, forming primer-dimers,
which become useless end products of the reaction.
If you are forced to have a small amount of self-complementarity at this
end, such as (5')...GT(3'), where the T can wobble pair with the G, by all
means avoid any 5' self-complementarity. The primer-dimers which form can
then reanneal to each other at their new 3' ends, which are complements of
the original self-complementary 5' ends. These form ever increasing primer
multimers with each cycle!
Finally, remember this: If a random 20-mer from the sequence of the
Salmonella typhimurium genome is used as a PCR primer, the odds are very
good that it will work, given a little optimization in annealing
temperature and magnesium concentration. Don't despair!