SmaI/XmaI digests

Song Tan tan at
Tue Nov 15 09:34:15 EST 1994

In article <9411142223.AA04169 at>,
P_Norton at CALVIN.JCI.TJU.EDU (Dr. Pamela Norton) wrote:

>         As the subject line says, we have been having terrible problems
> getting  a SmaI fragment re-subcloned; essentially we are trying to turn a
> Sma fragment around. First attempt: cut plasmid with Sma, isolate both
> bands, ligate back together, look for inserts in the right orientation.
> Only one or two - but these don't cut with Sma, and sequencing says that a
> single base is missing at both sites. Start again, with XmaI this time. Got
> about four in the right orientation, these now cut only once with Sma or
> Xma. Sequencing reveals that one of the sites (the same in all) has a base
> missing. The cursed site is within a very GC-rich region, perhaps this is
> the problem?
>         I am literally tearing my hair out over this one. Does anyone know
> whether Sma tends to nibble 5' or 3' ends? In other words, would it help if
> we try to refill any nibbled ends before ligation? I'm thinking of
> mutagenizing the site back into the plasmid with a single site present
> (both are in coding sequence, so deletions are a major problem). Before I
> resort to that, I thought that I would draw on the accumulated net wisdom.
> All suggestions and relevant anecdotes are welcome. 

Others on the net have reported problems using SmaI digested vector to
subclone blunt PCR product, and the consensus seems to be that SmaI does
have an exonuclease problem/contamination.  EcoRV appears to be the enzyme
of choice for subcloning blunt ends, though this won't be applicable in
your case.  Afraid that I don't know exactly what the exonuclease activity
of SmaI is like (i.e. nibbling 5' vs 3' ends).

You could try filling in the SmaI ends before ligation as you mentioned. 
Shame that using XmaI didn't help.  One would have thought that should
have worked very well, given the nice sticky ends.  I do note that Biolabs
suggests long incubation times (o/n) for complete cutting of plasmid DNA
by XmaI.  Such long incubation times might aggravate slight exonuclease
contaminations, of course.

My only new idea to this is to use AvaI (cuts C/PyCGPuG) instead of SmaI
or XmaI.  I believe it's a relatively good enzyme (it was a long time ago
since I last used AvaI.  perhaps someone else with more extensive or more
recent experience could post any good or bad comments about AvaI).  It's
relatively inexpensive, at least.  The one obvious drawback is that AvaI
will recognize more sites than SmaI or XmaI because of its degeneracy.  So
I hope that you don't have any other AvaI sites in your plasmid.  

Good luck!!

Song Tan
Institute for Molecular Biology and Biophysics
ETH-Honggerberg (Swiss Federal Institute of Technology)
8093 Zurich, Switzerland
email:  tan at

More information about the Methods mailing list