RFD: Plasmid re-ligation without ligase?

Rafael Maldonado rafael at howard.genetics.utah.edu
Thu Apr 24 17:00:59 EST 1997

Klaus Salger wrote:

> I think you're right - supercoiled plasmids will transform more efficiently
> than relaxed circles 

Actually, this one of those thing going around for long time, but
Hanahan says: "... there is no significant difference between relaxed
and supercoiled plasmids" (Hanahan and Bloom 1996,in "E. coli and
Salmonella cellular and molecular biology", F.D. Neidhardt, Ed. ASM

> I guess some of the cut plasmids formed a circle where the two compatible ends
> hybridized leaving a plasmid with 2 nicks. In vivo ligation of 2 nicks is also
> neccessary if you ligate an insert to a dephosphorylated vector. And it
> works!

I do not know how efficient will be this porcess, but knowing that the
melting temperature of a four bases annealing region is quite low, this
will be very difficult. But, of course, we do not have data to prove any
either way!

About my last statement that for in vivo cloning, a exonuclease
deficient (recBC) strain must be used, this procedure is used by Oliner
et al (Oliner, Kinzler and Vogelstein 1993, NAR 21, 5192-97), where
ligation of an insert is mediated by homologous recombination to a
linearizated vector. But I have to say that, surprisingly, other
non-recBC strains may work as well. Martin et al (1995), NAR 23,
1642-43, use DH5a for transforming a linear plasmid produces by PCR with
a 25 base repetition in each end as homologous recombination site. An I
say surprisingly not only because the presence of the exonuclease, but
also the absence of recA in that strain. So I guess that the homologous
recombination which allows the recircularization of the plasmid must
come through the recF or recE pathway (?). Bubeck et al (1993) NAR 21,
3601-2, also use DH5a for cloning insertss using homologous

Martin et al say that they get >400 transformants using from 15 to 75 ng
of linear plasmid. Oliner et al say that their method is 25-fold more
efficient than Bubeck's. So probably the trick is in adding a lot of
lienar plasmid and expecting to get some molecules that escape the
exonuclease, recombine and start replication.

I do not have any data about in vivo ligation efficiency, but my guess
is that is enough DNA is given, it will be as factible as recombination
in a recBC+ strain is. So IT IS possible, like Klaus said, an in vivo
ligation. But I do not know if with the plasmid concetration in a
ligation the effects would be observable.

If one could get a linear plasmid, free of circular contaminats (and I
have proof that agarose gel purification is not clean enough to make
such statement) we would have the answer. :-)

rafael at howard.genetics.utah.edu

More information about the Methods mailing list