degenerate PCR

Tim Chipman tchipman at
Fri Feb 28 09:04:52 EST 1997

Disclaimer: I don't pretend to be an expert.  There are many opinions on
this stuff.  I'll contribute what I can, and hopefully won't be too
misleading :-)

I've done some of what you describe in a different context so I'll answer
a bit as best I can.

: 1. How much redundancy is still acceptable in the primers?

Your mileage will vary :-) - I tend to use Inosine for any position which
has all 4 base degeneracy instead of using a g/a/t/c mixture in that
position.  Alas, Inosine tends to cost much more per base.  However,
*never* use inosine in 3 bases of the 3'clamp region which ideally should
be a perfect match, but failing that, should have low degeneracy (and
ideally GC-rich of course) 

: 2. Should I use all possible codons, or allow for mismatches, to 
: minimise redundancy?

Not sure I follow you.  For what I did, might be applicable to you, is to
generate a protein alignment of the gene from the nearest 6 related
species in the on-line DataBases.  Also, if there are any particular
species you KNOW you want to exclude that might be a contaminant, ie,
human, E.coli, whatever, it might be wise to include those seqs in the
alignment and try to select against those seqs in your oligo design

: 3. Should I increase primer concentration in the PCR?

I did try this for a while, then gave up, and it made no difference.
Probably the bottom line: Try a "normal" primer conc. first and see if it
works.  (Always good to try against a positive control at this point :-)

: 4. Any tips on annealing conditions?

Again, mileage varies greatly.  I've gone as low as 35 degrees for
annealing - only in extreme cases - and as high as 65 degrees in other
extreme cases.  Probably is a good idea to start with a middle-ish anneal
temp and if needed, dropping back gradually.  (alternately, you can try it
from the other end of the scale - start low, ie, 40degrees, and then
gradually increase anneal by 5deg. increments until all the undesirable
bands go away)

One regieme that proved somewhat useful was a "2-step" annealing regieme,

10 cycles of:  35degree anneal / 50% ramp up to 72

followed by

25 cycles of:  45degree anneal / 50% ramp up to 72

<note that you can twiddle ramping values if you want to be "gentle" with
your oligos..>

Note2: Not sure if I have to mention this or not, but.. When doing
degenerate PCR you've always got to include ALL single primer controls (at
least while you are in the "troubleshooting" stages.)

Thus, if I had a layout thus:

[---------------GENE OF INTEREST--------------------------]
   *--->     *--->                   <---*         <---*
   Oligo1    Oligo2                  Oligo3        Oligo4

Reactions to do in a single setup would include these oligo combos:

 1 by itself
 2 by itself
 3 by itself
 4 by itself
 1&3 pair 
 1&4 pair
 2&3 pair
 2&4 pair

In this manner you can easily identify single-primer products that may be
present in your desirable paired-oligo PCR reactions.

Also note that by doing a bunch of oligo combinations at once, the ratio
of {single control} to {interesting pairs} is better.. thus you "waste" 
(proportionally) less resources on single primer reactions

: 5. I intend to use these degenerate oligos as sequencing primers as 
: well (which works fine for specific oligos), should I use more primer 
: than in standard reacions?

I've done this too, and the bottom line was this:  You can get sequence,
but it is kind of cruddy.  Of 5 seq. reactions I attempted, one was good
enough to verify the PCR product was the thing I thought it should be. 
Two were hopeless. The remaining two were obviously the right sort of
gene, but the genome of origin was unclear (mito/cp/nuclear SSU
discrimination issues :-) 

Subsequently, I broke down and used a TA vector, pT7Blue (Novagen) to
clone these PCR products, and then sequenced using the non-degenerate
m13 and T7 primer sites in the vector.  This seq. data was FAR cleaner, 
and I was able to identify all the products which successfully cloned.

(Of course..cloning PCR products is always great fun :-)

Hope this is the kind of stuff you were after,

Tim Chipman

More information about the Methods mailing list