Self ligation of PCR products

Hiranya Roychowdhury hroychow at NMSU.EDU
Thu Dec 18 12:57:24 EST 1997

At 10:36 AM 12/18/97 -0600, Timur Yarovinsky wrote:
>I've got some unexpected results and need comments.
>While trying to get blunt ends in PCR products I used
>unpolished PCR products as a negative control. Briefly, the protocol was:
>1. PCR with Taq: got 1.3kb products with supposed A overhang
>2. Polishing of my experimental with T4 DNA polymerase to generate blunt
>ends, negative control was not polished
>3. Run ligation with T4 DNA polymerase expecting self-ligation of polished
>DNA and no ligation of unpolished DNA.
>Result: >20 kb bands  in both control and experimental DNA.
>1. Why did unpolished DNA with overhangs self ligate?
>2. Is there any way to check out polishing step?

I don't see how. Why do you need to check out polishing step anyway? The
proof should be in the pudding...:). You have to make sure that there is
enough free dNTP's available. I usually add a uL of 2.5mM dNTP mix to the
PCR soup before adding the polymerase. 

>Any comments are encouraged.
>Timur Yarovinsky, Ph.D., M.D.

Actually, the frequency of  dA overhangs depends on the terminal ntd. of the
product (C/G's?) and on the enzyme being used.  Not all products from a
reaction would have dA overhang. Secondly, even without polishing, the PCR
products are ligatable, albeit at greatly reduced efficiencies. I believe
that the amplicons with single ntd. overhangs, if any, essentially act as
"blunt" molecules in a ligation rxn.  (not all agree with me on this) or
that there may be enough products without the "overhang" to afford some
blunt ligations. I have been able to ligate amplicons directly following
PCR, without polishing, into SmaI-digested vectors.


Dr. Hiranya Sankar Roychowdhury
Plant Genetic Engineering Lab.
New Mexico State University
Las Cruces, NM 88003
Ph. (505) 646-5785
hroychow at

More information about the Methods mailing list