Gilda - 

If you have a relatively small and homogenous DNA fragment that you are adding, you can do an EMSA using EtBr staining.  You could just run your samples on a native (no SDS) acrylamide gel.  It should be easy enough to see if this will work by just trying with your wild type sample - have simple DNA only lane and lanes where you titrate your protein to the same DNA concentration to see if you can resolve the complexes reasonably well.  Just run the gel and soak in EtBr afterwards followed by a few washes to remove residual stain.  An issue here is that you will need more DNA sample than if you were doing more sensitive techniques so you will only be able to get a non-quantitative answer for DNA binding. (really this will only provide a binary yes it binds or no it does not bind at all result).  A better way to do the EMSA would be to have low input amounts of labeled DNA (you can use FAM labeled DNA fragments or something inexpensive) and titrate your protein sample and even consider adding unlabeled competitor to get rid of nonspecific binding.  EMSAs are very tricky and extremely difficult to do quantitatively.

Hi, you can also use fluorescence anisotropy, assuming your DNA is relatively small.

Fluorescence-based EMSA can work as well.

Gilda - 

If you have a relatively small and homogenous DNA fragment that you are adding, you can do an EMSA using EtBr staining.  You could just run your samples on a native (no SDS) acrylamide gel.  It should be easy enough to see if this will work by just trying with your wild type sample - have simple DNA only lane and lanes where you titrate your protein to the same DNA concentration to see if you can resolve the complexes reasonably well.  Just run the gel and soak in EtBr afterwards followed by a few washes to remove residual stain.  An issue here is that you will need more DNA sample than if you were doing more sensitive techniques so you will only be able to get a non-quantitative answer for DNA binding. (really this will only provide a binary yes it binds or no it does not bind at all result).  A better way to do the EMSA would be to have low input amounts of labeled DNA (you can use FAM labeled DNA fragments or something inexpensive) and titrate your protein sample and even consider adding unlabeled competitor to get rid of nonspecific binding.  EMSAs are very tricky and extremely difficult to do quantitatively.