DNA purification loss, large insert

Frank O. Fackelmayer Frank at Fackelmayer.de
Mon Apr 15 05:02:56 EST 2002

Gr8Guy wrote:
> Hey all - I've been beating myself to death trying to manipulate a couple of
> inserts I pulled from a genomic library.  The insert sizes are 5 and 6 kb in
> a pBluescript backbone (total ~10-11 kb), which isn't extremely large, but I
> have yet to find an efficient and reproducible way to purify them from
> agarose.  I've tried the Qiagen MinElute (which seems best, although it's
> "not recommended" for fragments over 4 kb), phenol extraction with low
> melting agarose, and the QiaEx gel purification kit with the resin (this is
> the WORST, even though it's what they recommend for large DNA).  Every
> method gives unacceptable loss of DNA.  I don't have enough left to do a
> ligation!  You can only start with so a couple of micrograms...
> The idea is to take a pBluescript + insert (9kb), open it, and insert
> another 6 kb fragment (this is a genomic construct for transgenic animal
> construction).  There's a 5' overhang on one end and a blunt on the other
> end.  When I take the DNA loss into account, and the low ligation
> efficiency, I never have success with transformations.

Maybe I´m not getting your problem right, but why do you have to gel
purify your construct? 
To clarify things:
1. you have an insert A in pBluescript (construct A)
2. you have an insert B in pBluescript (construct B)
3. you want to clone insert A into construct B


So you cut out the insert from construct A, try to purify it, and then
ligate it into linearized construct B?

In such a case, I would not purify the insert A at all. Why not cut it
out from construct A, add to the linearized construct B, and ligate?
You´ll get a mixture of your two original constructs and the new one.
With this approach, the wanted new construct will possibly be only a few
percent of the clones you get. 
To improve things, you may want to first treat the digested construct A
with shrip alkaline phosphatase (so you won´t get back lots of construct
A), and use very little of the linearized construct B for the ligation
(molar excess of insert >10:1). (Frankly, I personally wouldn´t do the
SAP step, because it often creates more problems than it solves...)

Maybe, you are lucky and can positively select fo your wanted construct
by digesting the ligation product that linearizes construct A and
construct B before you transform bacteria. For that, you´d need a
restriction site that is present in construct A (e.g. in the polylinker)
but not in the wanted construct. That very often works when you clone
with compatible ends that do not recreate either site (e.g. SalI into
XhoI, or BglII into BamHI).

If all that doesn´t do the job, you might consider the following:
cut out insert A, subclone into a vector with a different antibiotic
resistance (e.g. kanamycin); let´s call this one construct A2. Then cut
out again, ligate it to linearized construct B and plate onto ampicillin
plates. You won´t get back any construct A or A2. 

As said, I´d simply try first ligating insert A into linearized
construct B without any purification. It usually gives enough positive
clones, even though you´ll have to look for them.


More information about the Methods mailing list