For the second time, I have been burnt by the presence of a previously inserted cassette in a background I was using for linear transformation. In both cases, it was a pro::spec swap. I noticed an abnormally high number of camR transformants the next day. I printed the colonies to spectinomycin medium , and discovered that most were sensitive! What I had succeeded in swapping was the pre-existing cassette for the new one. Very few of the transformants were those which I had intended.
In parallel experiments, I found that the number of transformants increased over the course of a couple of days. Our standard UNI cassette (with the exception of tetAR) is a drug resistance gene embedded in an otherwise constant context, originally derived from the chloramphicol resistance locus of pACYC184. This means that sequence blocks about 200 bp at each end of the cassettes are always the same. Apparently, these regions are being used as recombination sites during the transformation.
I surmise that lambda red is not causing these insertions, but rather the host recombination system, because of the kinetics of their appearance. We incubate our transformations overnight at 42°C, specifically to eliminate pKD46 — this should confine all transformation to the first few minutes after electroporation. In addition, lambda red contains no mechanism for resecting the 3′ end of the transforming DNA back through the 40 bp intended homology block to make available the core of the cassette for strand invasion. This could only be done by the host after the red system had decayed or been diluted by growth, freeing the cell from red repression of the host system.
If I had used tetAR, I would never have noticed this effect. The only homology blocks available are the UNI ends, too small to be used by the host recombination system, but still requiring host resection to be exposed, which could only happen after elimination of red.-- Eric Kofoid