Hi, This is an intriguing question :) It's quite hard to test for 'stray' nucleases when you potentially have mg/ml amounts of 'real' nuclease.
Typically, the DNAse contamination of proteins purified from E. coli is not massive - especially after several purification steps. Otherwise, the companies who make recombinant site-specific nucleases would be in huge trouble :) In fact, when we did test a few things for nuclease contamination a while back, we found RNAse contamination (not surprising) but not DNAse contamination to any reasonable degree. Some of the things that come to mind immediately include: a) trying the cleavage experiment in presence of DNAse inhibitors, EDTA, and so forth. b) testing whether the protein of interest binds to DNA somehow (start by looking at pI - if the protein has a pI of 10 then Bob's your uncle) c) check if this protein works on non-methylated as well as methylated DNA There are enzymes out there that can perform DNA hydrolysis without themselves being 'classical' DNAses (for instance, under some condition polymerases, topoisomerases, or helicases can exhibit DNAse activity). In addition you may have a chemical effect in place (irradiation of certain metal complexes with visible light can generate radicals, causing breakdown of nucleic acids to which these complexes bind - see e.g. DNA and RNA footprinting). It is not a huge stretch of imagination to say that a strongly positively charged enzyme with an appropriate metal center *could* also degrade DNA in a process that is completely unrelated to the normal function of the enzyme. Could you describe in more detail what do you see when you refer to DNA being cleaved as well as with a commercial nuclease? What products are formed (by gel)? Does this happen in darkness as well as under light? Can you repeat this in an oxygen-free environment? Does this work with linear DNA or synthetic oligos? Can you cleave single-stranded DNA (i.e. just one oligo without any self-complementarity). Artem
