change upon binding

Dear CCLers,
 I have a question regarding Autodock. I had a crystalised
 protein-ligand complex, in which I took out the ligand and docked it
 with Autodock. THe results have been consistent with the crystal
 structure. But when I used a different ligand for docking, the results
 are not quite as expected. I believe that the conformation of the
 binding site has been shifted favourable to the original ligand and so
 therefore, as autodock treats the protein rigid, the prediction of the
 interactions with a different ligand is not a good. Has anybody been
 through this sort of problem? I'd love to hear some opinions on it.
 Actually, I'd like to find out how the binding of the another ligand
 affects the conformation of the binding site, and I can't quite do it
 with autodock. So, I am giving molecular dynamics a try. If there's
 somebody who encountered something like this before, would be glad to
 hear your input.
 I've learned a lot from the mailing list so I thank all of you for your
 Rowyna K.