Crystallographic Studies on the Complex Behavior of Nicotine Binding to P450cam (CYP101) ² Michael Strickler, § Barry M. Goldstein, § Kathleen Maxfield, ‡ Laura Shireman, ‡ Gyungyoun Kim, ‡ Donald S. Matteson, ‡ and Jeffrey P. Jones* ,‡ Department of Biophysics, UniVersity of Rochester, Rochester, New York 14627, and Department of Chemistry, Washington State UniVersity, Pullman, Washington 99164-4630 ReceiVed May 20, 2003; ReVised Manuscript ReceiVed August 8, 2003 ABSTRACT: Crystallographic and spectroscopic studies have been undertaken to characterize the binding behavior of the non-native substrate nicotine in the active site of the monooxygenase hemoprotein cytochrome P450cam. Despite the existence of a theoretical model that is consistent with the observed distribution of monooxygenation products, the crystal structure of the complex indicates that the primary binding mode of nicotine is unproductive. The structure is confirmed by spectral data that indicate direct coordination of substrate pyridine nitrogen with the heme iron. This would be the proper structure for evaluating binding affinity and inhibition. Reduction of the heme from Fe(III) to Fe(II) and introduction of carbon monoxide into crystals of the nicotine-P450cam complex, to simulate molecular oxygen binding, produces reorientation of the nicotine. This orientation is the appropriate one for predicting regioselectivity and the kinetic features of substrate oxidation. While it is not clear that such complicated behavior will be exhibited for other enzyme-substrate interactions, it is clear that a single crystal structure for a given substrate-enzyme interaction may not provide a good description of the binding mode responsible for product formation. Crystal structures have been available for P450 enzymes for almost 20 years. Numerous studies have been done to predict the binding affinity and important features of the bacterial P450 enzymes (1-11). More recently, structures have become available for mammalian enzymes (12, 13). The P450 family is rather unique among enzymes in its ability to bind compounds of widely varying structure in the same active site. These enzymes can also give multiple products that must arise from multiple binding modes for the same compound. Given this diversity of binding modes, it remains to be seen how valuable a single crystal structure will be in drug design. Herein we describe the multiple binding modes that are available to a single substrate throughout the catalytic cycle for P450cam, a P450 enzyme with a cornucopia of structural information. To best perform rational modeling of the binding of novel substrates to enzymes, experimental structural information is needed. The structure of cytochrome P450cam with its natural substrate (R)-camphor bound has previously been determined to a resolution of 1.63 Å (14). Crystal structures have also been determined for the substrate-free enzyme, for the binding of camphor-like substrate analogues (15), and for the inhibitors metyrapone and phenylimidazole (16). Structures with carbon monoxide bound have been used as an analogue for molecular oxygen in the crystallographic study of the oxygen-binding step in the P450 reaction pathway (17). Many of the crucial residues for P450cam action have also been studied by mutagenesis (18). The mutagenesis studies have underscored the importance of tyrosine 96 in camphor binding; the hydrogen bond between tyrosine 96 and camphor observed in the crystal structure is important for camphor metabolism. It was subsequently discovered that substrates for cytochrome P450cam need not be similar to camphor; indeed, many mammalian P450 substrates are also P450cam substrates (19). One such mammalian substrate is nicotine, or 3-(2-(N- methylpyrrolidinyl))pyridine (10). As the name indicates, nicotine is composed of a pyridine ring joined to N- methylpyrrolidine by a carbon-carbon single bond. Binding, labeling, and kinetic studies have been performed on the metabolism of nicotine by P450cam as well as for mam- malian enzymes (10, 20-22); these have indicated that nicotine is hydroxylated by P450cam with high regiospeci- ficity, and that there are differences in binding and kinetics between the enantiomers. Molecular dynamics simulations and calculations of free energy differences were performed for the binding of (R)- and (S)-nicotine to the active site of P450cam (10). These simulations were able to predict the binding free energy difference between (R)-nicotine and its enantiomer (S)-nicotine, as experimentally determined by spectral dissociation constants, and resulted in the configu- ration shown in Figure 1. ² This work was funded by NIEHS 009122, GM 32165, and NSF BES 9710129. This work is based upon research conducted at the Cornell High Energy Synchrotron Source (CHESS), which is supported by the National Science Foundation under award DMR 97-13424, using the Macromolecular Diffraction at CHESS (MacCHESS) facility, which is supported by award RR-01646 from the National Institutes of Health, through its National Center for Research Resources. * To whom correspondence should be addressed. Tel.: (509) 335- 5983. Fax: (509) 335-8867. E-mail: jpj@wsu.edu. § University of Rochester. ‡ Washington State University. 11943 Biochemistry 2003, 42, 11943-11950 10.1021/bi034833o CCC: $25.00 © 2003 American Chemical Society Published on Web 09/26/2003