I am trying to clone a promoter sequence using a pUC57 plasmid backbone in DH5alpha cells. The finished sequence will contain 6 direct repeats of a ~50 bp sequence I started from complimentary oligo nucleotides with matching overhangs to the cloning site I modified in the vector backbone. Subsequent cloning steps make of two restriction sites with compatible ends to recursively insert another copy of the sequence at the end of the piece already incorporated into the backbone. I have gotten the recombinant plasmid containing 4 repeats very easily. However, on the last cloning step where I ligate an insert containing 2 or the repeats into the backbone containing 4 repeats, I get wildly varying results from colony PCR. Furthermore, when I select the colonies that do show a correct insert size for liquid culture and plasmid extraction, I find by PCR or restriction digest that a deletion always occurs bringing the copy number back down to 1, 2, or 3.

With this cloning strategy I am trying to follow almost exactly the method described for construction of the same sequence in a previous paper, albeit I am using a different vector backbone (pUC57 instead of pBluescript II) and my own restriction sites flanking the sequence. I am even using the same E. coli DH5alpha strain. So I know that cloning and keeping this sequence intact should be possible. I am aware that tandem sequence copy number mutations such as I am experiencing can occur even with E. coli RecA mutants, but I am wondering if there is an E. coli strain out there that would have a chromosomal background more suitable for cloning this tricky sequence. Another concern I have is that after I successfully construct the sequence and after a couple additional cloning steps I will mobilize the finished vector to Agrobacterium tumerfaciens cells. How can I be sure my sequence won't undergo additional mutations in my Agro strain?

One other factor that might be causing a problem is that the repeat sequence I am trying to cloning contains a lac operator sequence. I wonder if it's possible that this is causing some sequence expression problem in the DH5alpha cells where plasmids with 6 copies (but apparently not ones with 4 copies) are at a selection disadvantage. If this is the case is there any techical way to suppress this response in DH5alpha or to use a strain that will be potentially less susceptible to this issue?

Any feedback or recommendations are appreciated. I've run into so many road blocks on this project, I am nearing the end of my rope for coming up with fixes.

Thanks in advance.