You'd probably need to run separate simulations of the two proteins isolated from each other, and include the solvation energies in the calculations at the least. Because of the likely conformational changes of surface groups (and more) involved in the binding, the entropy effects could be substantial.

Rick Venable
computational chemist