DNA binding/EMSA question - question about controls in DNA binding/EMSA (Oct/25/2009 )
Hi all, I have a question about proper controls for a protein-DNA binding experiment (EMSA). I have purified "Protein A" which binds a P32-labeled consensus DNA sequence, "Sequence A." I have purified "Protein B" which is not supposed to bind "Sequence A." I show using EMSA that "Protein A" binds "Sequence A" but "Protein B" does not bind "Sequence A." Is that an acceptable control for the experiment? Others have suggested that I, in addition to what I've already done, try and compete away the P32-labeled "Sequence A" with non-radioactive "Sequence A" as well as show "Protein A" does not bind a P32-labeled random sequence (scrambled) of DNA of a similar length. I agree that would be good to do, but I wonder if it's aboslutely necessary. What is the minimum I need to do in terms of showing that "Protein A" binds "Sequence A"? Thanks!
You've established that protein A binds DNA and protein B does not; you have not, however, established that this is in any way sequence specific. It would be most valuable to test the binding of protein A against one or a few randomly permuted sequences to verify that it does not bind. The competitive binding experiment would also be good, but again you should be using both sequence A and a randomized sequence as competitors.
phage434 on Oct 25 2009, 02:50 PM said:
Thanks for the reply, it was very informative. I see the distinction now. I have have a followup question on the randomized/scramble sequences. Is there a logic to choosing the scramble sequence? Some papers I've read only change a couple of bases in the consensus region while others make it completely different.
Depends on your goal. If you are establishing that specific binding happens to that sequence, then any old control sequence which is not that one will work. If instead, you are trying to locate where the binding motif is located, and how sensitive it is to point mutations, then you would want one which was similar with a few changes.
What about using an antibody against your protein A in the reaction proteinA+sequence A (you will see an upshift) and using an antibody against your protein B in the same reaction you get no upshift