Interactions of Anesthetics with the Water-Hexane Interface. A Molecular Dynamics Study Christophe Chipot, †,§ Michael A. Wilson, †,‡ and Andrew Pohorille* ,†,‡ Exobiology Branch, NASAsAmes Research Center, MS 239-4, Moffett Field, California 94035-1000, and Department of Pharmaceutical Chemistry, UniVersity of California, San Francisco, San Francisco, California 94143 ReceiVed: May 23, 1996; In Final Form: August 26, 1996 X The free energy profiles characterizing the transfer of nine solutes across the liquid-vapor interfaces of water and hexane and across the water-hexane interface were calculated from molecular dynamics simulations. Among the solutes were n-butane and three of its halogenated derivatives, as well as three halogenated cyclobutanes. The two remaining molecules, dichlorodifluoromethane and 1,2-dichloroperfluoroethane, belong to series of halo-substituted methanes and ethanes, described in previous studies (J. Chem. Phys. 1996, 104, 3760; Chem. Phys. 1996, 204, 337). Each series of molecules contains structurally similar compounds that differ greatly in anesthetic potency. The accuracy of the simulations was tested by comparing the calculated and the experimental free energies of solvation of all nine compounds in water and in hexane. In addition, the calculated and the measured surface excess concentrations of n-butane at the water liquid-vapor interface were compared. In all cases, good agreement with experimental results was found. At the water-hexane interface, the free energy profiles for polar molecules exhibited significant interfacial minima, whereas the profiles for nonpolar molecules did not. The existence of these minima was interpreted in terms of a balance between the free energy contribution arising from solute-solvent interactions and the work to form a cavity that accommodates the solute. These two contributions change monotonically, but oppositely, across the interface. The interfacial solubilities of the solutes, obtained from the free energy profiles, correlate very well with their anesthetic potencies. This is the case even when the Meyer-Overton hypothesis, which predicts a correlation between anesthetic potency and solubility in oil, fails. Introduction Although the phenomenon of general anesthesia has been known for over a century, the site and mechanism of anesthetic action remain unknown. For many years, the thinking about this phenomenon has been significantly influenced by the Meyer-Overton hypothesis. This hypothesis predicts a cor- relation between the potency of inhaled anesthetics and their lipophilicity, i.e., solubility in a hydrophobic phase (e.g., olive oil), considered as a model for the interior of the membrane. 1,2 Since anesthetics rapidly equilibrate across all cellular compart- ments, 3 their equilibrium properties, such as solubility, are relevant to the mechanism of their action. For conventional anesthetics, the observed correlation is remarkably accurate over several orders of magnitude of anesthetic potencies and for different animals. 4 The success of the Meyer-Overton hy- pothesis led to the suggestion that the site of anesthetic action might be located inside the lipid bilayer of neuronal tissue. 5 In this environment, anesthetics would act either by disrupting the lipid bilayer or by interacting with receptor sites buried in the membrane. If, however, the Meyer-Overton hypothesis is tested on a broader range of potentially anesthetic compounds, the correla- tion between potency and lipophilicity is markedly less convinc- ing. In particular, some halogenated hydrocarbons that are predicted to be good anesthetics from the Meyer-Overton hypothesis exhibit no anesthetic activity (nonanesthetics), 6-9 and several other compounds show an anesthetic potency that is much weaker than predicted (transitional compounds). 7 All these compounds have very low solubilities in water. In contrast, alkanols, which are highly soluble in water, appear to be more potent anesthetics than expected from the Meyer- Overton hypothesis. 10 These results indicate that anesthetic potencies cannot be fully predicted from solubilities in oil alone. In particular, some level of solubility in water appears to be required for anesthetic action. These recent discoveries suggest that the anesthetic potency might correlate better with the solubility at the interface between water and a nonpolar phase than with the solubility in oil. This paper is devoted to testing this idea. Since interfacial solubility is very difficult to measure experimentally, molecular dynamics simulations were employed to calculate this quantity at the water-hexane interface. In this work, we study two series of molecules of structurally similar halogenated compounds, each ranging from potent anesthetics to nonanesthetics. In the first series, n-butane, a weak anesthetic, and 1,1,2,2,3,3,4,4-octafluo- robutane, a known anesthetic, are compared with the nonanes- thetic 2,3-dichloroperfluorobutane [(R,R)-enantiomer] and the transitional 1,2,3,4-tetrachloroperfluorobutane [(S,S)-enanti- omer]. The second series consists of cyclobutane derivatives: 1-chloro-1,2,2-trifluorocyclobutane, 1,2-dichloroperfluorocy- clobutane, 11 and perfluorocyclobutane. The first of these compounds is an anesthetic, whereas the remaining are not. In addition, we study the anesthetic dichlorodifluoromethane and the transitional 1,2-dichloroperfluoroethane. These two mol- ecules are structurally closely related to fluorinated methanes and ethanes, the transfer of which across the water-hexane and water-membrane interfaces has been studied previously. 12,13 † NASAsAmes Research Center. ‡ University of California. § On leave from: Laboratoire de Chimie The ´orique, Unite ´ de Recherche Associe ´e au CNRS No. 510, Universite ´ Henri Poincare ´-Nancy I, BP. 239, 54506 Vandoeuvre-le ´s-Nancy Cedex, France. X Abstract published in AdVance ACS Abstracts, December 1, 1996. 782 J. Phys. Chem. B 1997, 101, 782-791 S1089-5647(96)01513-1 CCC: $14.00 © 1997 American Chemical Society