Proofreading x engineering terminal sites

[Petr GRUZ]

        I am trying to introduce terminal restriction enzyme sites into my 
PCR product by primers which have about 15 bp homology with the template at 
their 3'-end and some mismatches with the template at their 5'-end (the 
region of hopefully engineered restriction site).  I am using the 
concatamerization of the product prior to digestion to make sure it is cut 
(NAR 18, p. 6156).  If I amplify it with Taq it ligates well.  However, if I 
use Pfu I cannot get any expected ligation product.  I prefer the Pfu enzyme 
because of its high fidelity.  I have a suspicion that the other side of the 
high fidelity of Pfu is that the proofreading activity of Pfu causes it to 
chew back into the primer and reach the mismatches between the 5'-end and 
the template.  That may correct the sequence of restriction site according 
to the sequence of template and that may be why the PCR product does not cut 
and does not ligate.  Is anybody there on the net who has experienced 
similar problems or knows more about this matter?
        My second question is about for me (sorry if this is some commonly 
used buffer) mysterious compound abbreviated TAPS.  The authors of an 
article (Mutation Research 303(1993)171-175) apparently used this compound 
to increase the fidelity of Taq polymerase which was in their hands higher 
than the fidelity of Pfu.   Unfortunately I could not find what TAPS was.
        Thanks in advance for any and all responses.

Petr Gruz
National Institute of Hygienic Sciences
Division of Genetics and Mutagenesis
1-18-1 Kamiyoga, Setagaya-ku, Tokyo 158
Fax +81-3-3707-6950
e-mail gruz at nihs.nihs.go.jp