Open circular DNA does not run faster than its linear equivalent, it runs slower. It is the supercoiled circular form of the DNA plasmid molecule which runs faster. The ligation reaction only produces open circular DNA, with little if any supercoiling.
Given that you can see high molecular weight bands, at least we know the ligase is working. I am assuming that the ligation mix was tested just prior to transformation (no further manipulation were done on the ligation mixture) and thus not a case of losing DNA after desalting (assuming electroporation was the transformation method used)
This leaves two places to look for trouble. The restriction digest and the transformation.
First can you tell us the conditions that you used to digest the vector and insert. In particular the SmaI digest. Could you tell us how the insert was made. Was it by PCR? Was the insert gel purified. How far apart are the restriction sites BamHI and EcoRV in the vector?
Can you tell us about the how the transformation was done. Is this electroporation, chemical transformation? Did you desalt the DNA before hand. Are there any trouble with the cells?
You are correct in your assumption, the picture I attached features the digested and undigested ligation product prior to transformation.
Both digestions of the vector and plasmid containing the insert were performed in a 40 ul reaction volume. For the vector digestion, I added ~1ug (~11 ul) of DNA to 4 ul of NEBuffer4 and 20-25U of EcoRV-HF and filled the rest of the volume with dH2O. I digested for 30mins at 37C at which point I added 25U of BamHI-HF to the reaction to digest at 37C for another 30mins. After the EcoRV blunt cut, there are only 3 base pairs between the end of the DNA fragment and the recognition site for BamHI, however this should not be an issue as BamHI has 97% cleavage efficiency with only 1 bp from the end. When I performed the phosphatase step, I added 10ul of phosphatase buffer and 1ul of antarctic phosphatase directly to the digestion mix and incubated at 37C for 15 minutes for 5' sticky ends and blunt ends as directed by the protocol.
The digestion of the Origene plasmid was performed in a similar fashion, starting with the SmaI digestion at 25C for 30mins and adding BamHI afterwards for another 30mins digestion at 37C.
Following the digestion times, I split up the reaction volumes into 4 wells per reaction on a 0.9% agarose gel and gel extracted after a sufficient amount of time for running the gel.
After gel extraction, I tried a 8:2 and 8:1 insert:vector ratio (in terms of volume, as far as molecular weight it should be 50ng of vector + ~43ng of insert) in 10ul of ligation buffer and 1ul of ligase from the Quick Ligation Kit from NEB. The DNA was not desalted as I am not using electroporation. I have used DB3.1 non-competent cells from Invitrogen for transformations before and have not had any issues. I make the cells competent with CaCl2 and carry out a normal transformation protocol before plating on LB/Amp agar plates. After overnight incubation at 37C, I see anywhere from 0-2 colonies per plate...
One thing I failed to mention is that I have done this exact ligation in the past with similar results - low transformation efficiency of only two colonies on the LB/Amp agar plate - and LUCKILY both of these colonies contained the insert ligated successfully into the vector backbone. I don't know why the efficiency is so low and why I'm not having the luck this time of obtaining the correct construct!
Thanks for any suggestions you may be able to offer.