Chris Paszty pasz at MH1.LBL.GOV
Tue Dec 5 16:51:33 EST 1995

								Dec 1995
Gene targeters:

A couple of years ago I sent out the sequence compilations of the PGKneo
and PGKtk cassettes that make up pPNT [Tybulewicz et al. Cell vol 65
p1153-1163 (1991)] with the goal of making it easier for people to
construct targeting vectors using these 2 cassettes and to design screening
strategies for identifying correctly targeted ES cell clones.  There were a
number of requests for copies of this compilation so I ended up e-mailing
it out to a bunch of knock-out labs.  The purpose of this follow up note is
to address the issue of DNA rearrangement that can be encountered
during vector construction.

A good way to build one's targeting vector using pPNT has been to clone
the short arm into the polylinker that is located between the PGKneo and
PGKtk cassettes and to then clone the long arm into the unique XhoI site
that lies on the other side of the PGKneo cassette.  The final targeting
vector can then be linearized at the unique NotI site.  With regards to the
effectiveness of negative selection by PGKtk in mammalian cells this
configuration of the short and long arms as well as the linearization by
NotI is ideal for two reasons.  First, the distance between the PGKneo
and PGKtk cassettes is minimized thereby making it more unlikely for
events (such as endonuclease attack) other than homologous recombination
to separate the two markers.  Second, by linearizing with NotI the
pUC/Bluescript region of the vector is left flanking the PGKtk cassette
thereby decreasing the probability of PGKtk being inactivated by

When building my first targeting vector I found that most (99%+) of the
clones that were recovered after the last ligation step (cloning the long arm
into the XhoI site) contained deletions and rearrangements.  From talking
to a number of people this seemed to be a fairly common occurrence when
including both PGKneo and PGKtk in ones targeting construct.  The
unrearranged clone that I finally isolated turned out to be totally stable
even when grown for many "passages".  When miniprep DNA from
this clone was retransformed into bacteria no rearrangement was observed.
This seemed to indicate that the rearrangement problem was not due to
the particular murine sequences that I had included in my targeting vector
but was more likely due to the topology of the transforming DNA.
Since I had used calf intestinal phosphatase (CIP) to dephosphorylate the
XhoI ends of my XhoI cut short arm vector (to eliminate vector religation)
the topology of DNA in my ligation mix was OC (open circular) containing
dephosphorylated nicks.  From the subsequent construction of a number
of targeting vectors based on pPNT and a 2 tk version of pPNT (pPN2T)
it is now clear that by not CIP treating the short arm vector during the
cloning in of the long arm, the incidence of rearrangement is reduced to
almost zero, thereby making it much easier to obtain the final targeting

With regards to vector religation without the CIP treatment one can still
use the XhoI site for cloning in the long arm by playing with the
insert-vector ratio to minimize vector religation.  Alternatively, one
could eliminate vector religation without using CIP and also get
directional cloning of the long arm using the NotI and XhoI sites in
pPNT, but because these two sites overlap in pPNT, achieving
effective double digestion was thought to not be possible (93' NEB
catalogue p 180).  Recently the people at NEB (Moreira and Noren,
Biotechniques 1995 vol 19 p56-59; Minimum Duplex Requirements for
Restriction Enzyme Cleavage Near the Termini of Linear DNA
Fragments) have reexamined the overlapping restriction enzyme site
issue and their new results indicate that it should be possible to achieve
effective double digestion of the NotI and XhoI sites in pPNT and pPNT
derivatives by first cutting with NotI and then subsequently with XhoI.
So although we haven't tried it , using NotI and XhoI may be a realistic
option for the cloning in of the long arm.

Another approach which we've been using successfully and prevents
vector religation as well as allowing for forced directional cloning of the
long arm is as follows:

The recognition site for a fairly new 8 cutter, Sse8387I (CCTGCAGG,
available from Amersham), is located between the NotI site and the
PGKneo cassette in both pPNT and pPN2T.  By double digesting the
short arm-containing targeting vector with NotI and Sse8387I and then
gel purifying it one can eliminate vector religation.  The different cohesive
ends also allow for the directional cloning of the long arm.

There are various straightforward approaches for creating a NotI site and
an Sse8387I site at either end of ones long arm:

- if one is using PCR to obtain the long arm then it is easy to just include
a NotI site on the 5' end of one of the primers and an Sse8387I site on
the 5' end of the other primer.  TA cloning of the product leaves one with
a long arm that is ready to clone into one's short arm-containing vector.

- another approach involves creating a blue/white vector that has a blunt
end cloning site in-between a NotI site and an Sse8387I site.  The long
arm can then be blunt end cloned in both orientations into this
intermediate vector.  The proper orientation clone leaves one with a long
arm that is ready to clone into one's short arm-containing vector.
To make a blue/white NotI-Sse8387I intermediate vector one can modify
the polylinker of any common blue/white vector using custom oligos.
Alternatively one can combine complementary fragments from two
different blue/white vectors...the 1892 bp XbaI-AlwNI fragment from
pUC18 (polylinker contains an Sse8387I site) and the 833 bp XbaI-AlwNI
fragment from pBluescript SK +/- (polylinker contains a NotI site) can be
combined to create such a vector (pUCNotSse).  This hybrid blue/white
vector has a HincII site (blunt end cutter) in-between the NotI site and the
Sse8387I site.

pUCNotSse polylinker:

Over the past year, using the sequence compilations of PGKneo and
PGKtk as well as the NotI-Sse8387I long arm cloning strategy, we've
significantly increased the speed and ease with which we build our targeting

Hope this information is useful to others.

Chris Paszty
Lawrence Berkeley Laboratory
C_Paszty at

More information about the Methods mailing list