Ligation problems

Paul Shinn pshinn at force.stwing.upenn.edu
Wed Feb 18 19:02:22 EST 2004


   We have run into some problems with our ligations.  Our transformants 
are either non-existent, non-abundant, incorrectly cloned, or vector 
only.

   We do directional cloning of PCR products using Sfi I and all our
transformants are sequence verified by us.  The vector has 2 Sfi sites 
and is digested with Sfi I and then gel purified.  Because of how Sfi
cuts, there should be no self ligation of the vector and the PCR products 
(ends are also digested with Sfi) should clone in only one direction.

GGCCN NNN|NGGCC
CCGGN|NNN NCCGG

But like I said, we are getting vector only clones or the transformants
have backward inserts or there is truncation of the PCR product ends
during the ligation.  And sometimes (rarely) they actually work.  Most of
the clones look like they are blunt end ligated rather than sticky end
because the Sfi site is gone.  Sometimes the Sfi is gone on the insert
side and sometimes on the vector side.  There is no rhyme or reason to
this.

For the most part the PCR is using Taq.  Then sodium acetate/ethanol 
precipitated.  Digested with Sfi overnight.  Precipitated again and then 
ligated overnight at 16.  For larger inserts we use a proof reader and 
then do phenol/chloroform and then sodium acetate precip.  We had some 
problems with the proof reader chewing back the ends of the insert.

I synthesize the PCR primers for cloning.  The forward primer is 56mer; 
the reverse primer is 43mer.  If the oligos are degraded, the 3' end would 
usually go first, right?  If so, the restriction sites at the 5' end 
should be intact.  PCRs are still working, too.

Each person has their own stock of vector and generates their own clones.  
It's hard to believe we're all doing something wrong at the same time.  
I worry that we're seeing the result of multiple problems like vector is 
degraded and PCR ends are degraded independently.  Can a sticky end blunt
ligate to a blunt end?

We're out of ideas.  Thanks for your help.

Paul



More information about the Methods mailing list