HBImolbiol

I am making a transgenic mouse targetting vector and I am currently trying to insert the 3' ROSA26 homology arm+DTA marker into a KpnI site on the construct.

I am getting tonnes of colonies which I screen by colony PCR and mini-preps and I am just getting a lot of religated vector.

I have tried the following:

A) 1 ug plasmid, digest with 1 ul KpnI-HF with 1 ul CIP for 90 min @ 37 followed by PCR column purification of which 10 ng (based on gel) goes into my ligation.

I first digested for 60 min with KpnI-HF, performed a clean up with a PCR column, then the resulting eluate was incubated with antarctic phosphatase for 60 min at 37, heat inactivated for 5 min at 65 and used 10 ng of this directly in the ligation.

In both cases the insert was digested for 60 min with KpnI-HF, cleaned up with a PCR column prior to ligation.  I am using a 2:1 molar ratio of insert to plasmid.

In both cases I am just getting hundreds of religated plasmids.  Both my CIP and antarctic phosphatase are fairly new from NEB.

I understand that 3' overhangs are more difficult to dephosphorylate but I really want to crack this problem and finish up the construct.

Any tips for:
-How to reduce the self-religation problem?  Using 2 different enzymes for cloning doesn't seem to be an option right now as I'm running out of unique enzymes?
-Any tips to improve success with ligation of large fragments?

Thanks in advance...