Which residues are most involved in the process of binding/unbinding? How large are the conformational changes induced in the protein by the binding/unbinding process? Can the hormone use one pathway for binding and another pathway for unbinding? To answer these questions, the unbinding of retinoic acid from its receptor was induced by applying an external force on the hormone.

The crystal structure of the human retinoic acid receptor (hRAR)-gamma bound to all-trans retinoic acid (t-RA) suggests that the possible entry point for the hormone is closed by helix 12. Renaud et al. proposed that when t-RA binds to the receptor, helix H12 changes its conformation from extended into solvent, as seen in the crystal structure of retinoid X receptor (RXR) , to a conformation closing the entrance to the binding pocket as in the crystal structure of retinoic acid receptor (RAR) . Analysis of the surface of the protein identified another possible entry point: the only ``window'' in the surface of the protein that allows the hormone to be seen. Three possible unbinding pathways were studied. Path 1 and path 2 were chosen in the proximity of helix H11 and H12. Path 3 was chosen to study the unbinding of the hormone through the ``window'' .

Simulation results indicated that it is possible to unbind the hormone along path 1 and path 3 without greatly affecting the structure of the protein. Particular characteristics of the simulated pathways suggest path 1 as the binding pathway for the hormone and path 3 as the unbinding pathway . Using the reversed order of events observed in the simulated unbinding along path 1 one can describe the binding mechanism of the hormone. First, Arg413, and then Arg396 attract and orient the carboxylate end of the hormone towards the binding pocket. When the hormone is within hydrogen bonding distance of these two residues, it experiences also the influence of the charged and polar residues Arg278, Lys236, Arg274, and Ser289 located at the opposite end of the binding pocket. The two steps described above take place almost simultaneously, i.e., the strong electrostatic attraction between the carboxylate end of the hormone and residues Arg278, Lys236, Ser289, and Arg 274 helps the beta-ionone ring to pass between protein residues, thus, leading to the penetration of the hormone into the binding pocket. The entrance of path 1 is surrounded by highly fluctuating residues. These residues may become more ordered upon binding of the hormone, making contacts with the hormone or with other protein residues. Induced ordering of the protein side chains may be favorable for the hormone entry without requiring a motion of helix H12. Path 3 is an unlikely candidate for a binding pathway for the following reasons: 1) the carboxylate end of the hormone should be the one to be attracted by the point of entry into the protein since it furnishes stronger and more specific interactions than the beta-ionone ring; 2) the ``window'' would need to be sufficiently large to let the bulky ring pass through first. Therefore, one may then consider path 3 as a possible unbinding pathway, either as the sole unbinding pathway or an alternative to path 1.

Movies for the SMD simulations


[1] D. Kosztin, S. Izrailev and K. Schulten. Unbinding of retinoic acid from its receptor studied by Steered Molecular Dynamics. Biophysical J., 76:188--197, 1999