ligating into vector: sequencing trouble - vector sequences before RE but not after (Oct/26/2007 )

I am trying to clone a 2.0kb gene into a 4.8kb vector. And I’m running into a bit of trouble. For some reason, the vector sequences ok with our sequencing primers when uncut, but once it’s cut (EcoRI-EcoRI) and the insert appears ligated, it no longer works with our sequencing primers. Any thoughts?

The vector is pKH3, given to us by another lab. We use our own primers that we ordered from IDT for sequencing, and our sequencing primers do work with this vector when it's uncut. We get the sequence we expect after sequencing the vector (of what we can see... we haven’t sequenced the entire vector).

Any suggestions why sequencing fails after RE/ligation?

Details…

I used PCR to amplify my gene with primers designed to include EcoRI on either end, I put the Eco-Eco gene into TOPO vector, and sequenced it. It looked good. So then I cut it out with EcoRI and gel purified, both with and without EtBr present in the gel. (Because of what we know about the vector and because the gene has a BamHI site within the gene, we chose to do just EcoRI cut.) I then took pKH3, EcoRI cut, CIAP or SAP treat, heat inactivate, gel purify both with and without EtBr. It seems I need to ligate this particular combo for two days, one day at RT and the second at 15C, in order to have a successful transformation (some other folks and I were doing other ligations at the same time that were successful in much less time! so the ligase and buffer should be ok). I get colonies, grow those up, do mini-preps, digest with EcoRI. Most of the colonies appear to be vector only but a small number appear to have the insert. The vector band and insert band appear to be the correct size. So I send them off to sequencing.
I send three primers for each mini-prep. The two vector primers (that have previously worked with the exact same prep of pKH3) and one gene primer in the middle of the gene. The middle gene primer comes back ok. But BOTH vector primers fail… zip, zero, zilch. No sequencing whatsoever.

Any thoughts as to why this vector is being so finicky?

Thank you!!