upper limit to Inverse PCR products?

Henry Broekhuyse broekh at unixg.ubc.ca
Fri Jan 19 00:29:01 EST 1996

Any comments on what the upper limit is for products from inverse PCR?

I have amplified the middle of a conserved gene from a fungus using
primers to conserved residues.  Sequencing the 0.7 kb PCR fragment and
comparison of its translation to homologs from other organisms tell me I
have the real thing, and now I want to retrieve the flanking DNA to get
the rest of the gene (the conserved regions are only in the middle of the
gene).  I know there are a number of methods based on linker ligations and
so forth, but I thought I would give inverse PCR a try. I made primers
(22-24 mers, no mismatches) for inverse PCR but have not managed to get
any product yet.  I recently read that 3 or 4 kb is about tops for
recovering things by PCR from a complex (genomic) DNA prep, so I'm
wondering if the product I want may be too big for this technique.  Any
way to tell?

For those of you with more time/interest to brainstorm, here's what I did:

I "simulated" the desired end product by cloning the 0.7 kb PCR product
into a 4 kb vector and I verified that the inverse PCR primers, the
program, etc. worked to make the 4 kb product.

I digested a number of aliquots of genomic DNA with 6-cutter enzymes (-> a
variety of ends 5', 3' overhangs and flush) that don't cut within the 0.7
kb fragment cloned.  I expect the whole gene will be about 2 or 3 kb
(based on comparison with homologs in other organisms), but I don't know
anything else about the gene (forgot to say I cannot do helpful southerns
first because we don't have much DNA). I stopped the digestion by heating
at 70C for 10+ min, and then ligated.  I checked a small amount of the
digestion and ligation reactions on a gel and saw they were ok.  I used a
couple of microliters of the ligation mix (=about 5 ng of ligated genomic
DNA) in PCR reactions with the program as set before for the 4 kb
simulation.  No product.  I have tried a longer extension time (twice as
long as before, 6 min).  No product.  I spiked one of the reactions with
the plasmid containing the cloned 0.7 kb product to be sure that the short
cuts (e.g. heat denaturation, etc.) didn't leave behind something that
would inhibit the PCR (at least in trans) - that worked fine.  Previous
dilution expts say that 1 pg template is enough for PCR, so it doesn't
seem like the problem.  Also, the very same DNA that I couldn't make the
inverse PCR product from was still plenty enough to template the 0.7 kb
internal product, so it doesn't seem like a general PCR or concentration
problem.  My top 2 guesses are that I happened to chose enzymes that give
very big fragments, or that maybe the ligations ended up intermolecular
instead of intramolecular.  Determining if the second one is the problem
seems like it will be a pain, but the fixing should be easy.  I guess. 
Any relevant references, comments, ideas, general consolation, offers of
other employment, etc. gratefully accepted!  :-)

More information about the Methods mailing list