Journal of Molecular Biology
Regular articleHow do substrates enter and products exit the buried active site of cytochrome P450cam? 1. Random expulsion molecular dynamics investigation of ligand access channels and mechanisms1
Introduction
Cytochrome P450 enzymes are hemoprotein monooxygenases that play a critical role in the synthesis and degradation of many physiologically important compounds and in the degradation of xenobiotics (Graham-Lorence & Peterson, 1996). The bacterial cytochrome P450cam (P450cam) from Pseudomonas putida catalyses the hydroxylation of camphor, its natural substrate, and also reactions with many other compounds. It has been extensively characterized by biochemical and biophysical techniques, and thus serves as a prototype for P450 structure-function studies. The protein structure has been solved at high resolution in the free form and bound to a number of substrates and inhibitors Raag et al 1993, Poulos et al 1986, Poulos et al 1987, Dmochowski et al 1999. As the active site of P450cam is completely buried in the protein interior and the outer molecular surface is not connected to the inner active-site surface (cf. Figure 1), the protein must undergo structural fluctuations to allow substrate access and product exit.
Examination of the crystal structure Poulos et al 1987, Moeckel et al 1998 indicated two regions where possible channels to and from the active site could form transiently, due to breathing motions of the protein. One of the channels is located between the B′ helix and the F/G loop (cf. Figure 2) and is supported by the following data: (i) in site-directed mutagenesis and stopped-flow kinetic measurements, Tyr29 (N terminus), Phe87 (B/B′ loop) and Phe193 (F/G loop) were found to be key residues for substrate access (Mueller et al., 1995). (ii) The crystal structure of P450cam with a large inhibitor bound shows that this inhibitor, which is larger than the natural substrate, not only occupies the active site, but also part of the putative access channel. In this structure, the side-chains of residues Phe87, Tyr96 and Phe193 are displaced with respect to the side-chains in the camphor bound structure (Raag et al., 1993). (iii) Very recently, the crystal structure of a ruthenium sensitizer-linked adamantane substrate was solved and showed the ligand extending along and opening up the putative access channel (Dmochowski et al., 1999). (iv) In the crystal structures of two structurally similar proteins of the P450 family, cytochrome P450BM-3 and nitric oxide reductase Ravichandran et al 1993, Park et al 1997, this channel is found to be open. As this channel is hydrophobic, it was suggested as an access region for the hydrophobic substrate. Another, more hydrophilic, channel is located near the heme propionic groups and filled with ordered water molecules. This channel was suggested as the exit channel for the product, 5-hydroxy-camphor, which is more hydrophilic than camphor (Poulos et al., 1987).
Adjacent to the proposed substrate entrance channel, there is a tetrad of charged residues, Lys178, Asp182, Arg186 and Asp251, which form a ring of four salt links (cf. Figure 2). Experiments indicate that substrate access to the active site of P450cam is modulated by electrostatic interactions involving Asp251 (Deprez et al., 1994). The two salt links to Asp251 are among the four most electrostatically stable salt links observed in P450cam (Lounnas & Wade, 1997). The proposed role of the salt links is to tether the F-G flap to the I helix, thereby tending to close the active site. It was thus speculated that substrate access may only occur under concomitant perturbation of the salt-link tetrad (Deprez et al., 1994).
As crystallographic B-factors indicate protein atoms with higher levels of flexibility, protein regions with high-level crystallographic thermal factors connecting the active site to the surface may indicate ligand exit/access channels (Carugo & Argos, 1998). We previously analysed the B-factors in the substrate-bound structure of P450cam and found not only one thermal motion pathway (TMP) near the proposed substrate access channel, but two further TMPs elsewhere in the protein (Lüdemann et al., 1997).
The location on the surface of P450cam at which the substrate initially penetrates the protein is unknown. Thus, a defined initial state for the study of substrate access is lacking. This is why, here, ligand escape from the structurally well characterized substrate-bound state is investigated. A successful method for investigating escape pathways of two-atom ligands is locally enhanced sampling (Elber & Karplus, 1990), in which a MD trajectory of a single copy of the protein and several copies of the ligand, which start in the active site, is performed. The ligand molecules do not interact with each other and the protein feels the average field of all ligand copies. It is thus assumed that the protein fluctuations are influenced only to a minor extent by the forces imposed by the ligand. This assumption is less likely to be valid if exit of larger molecules (such as camphor) from a completely buried protein active site is studied. For the systems studied here, successful exit of the ligand can only happen during a short simulation time if a ligand molecule induces conformational changes in the protein, and in this case the classical time-dependent Hartree approximation breaks down. This is why we developed a new molecular dynamics simulation method, which we call random expulsion molecular dynamics (REMD). The probability of spontaneous substrate exit in the time range amenable to MDS is enhanced by an artificial force with random direction imposed upon the substrate in addition to the standard molecular mechanics force field. The REMD method is also expected to be applicable to other problems in which achieving the desired sampling during MD simulations is difficult, such as protein unfolding and large-scale protein dynamics for which a randomly oriented force could be applied to certain domains such as helices. The REMD method has not yet been tested for such other applications.
Here, the REMD method was first tested by applying it to cytochrome P450BM-3 (P450BM-3) and then it was applied to different substrates, camphor (CAM) and endo-borneol allyl ether (CAL), and the product of P450cam, 5-hydroxy-camphor (5OH-CAM). CAL, which was designed to fill an empty region in the active site unoccupied by CAM (Helms et al., 1996), was chosen as it is larger than CAM. It also binds more favorably to the active site than CAM by 2.5 kJ/mol. 5OH-CAM is more hydrophilic than CAM. Application of REMD permitted investigation of possible exit routes and mechanisms and also whether different routes and mechanisms would be used by different ligands.
Section snippets
Substrate pathway classes in P450cam and P450BM-3
The trajectories obtained by REMD as detailed in Methods were clustered into ligand-exit pathway classes by graphical inspection. A pathway class comprises all trajectories connecting the active site and a particular sector of the protein surface (see Figure 3). If within a pathway class, the ligand trajectory endpoints show further clustering into sub-clusters, pathway subclasses (denoted by lower-case letters) were introduced for further classification. Three pathway classes, pw1, pw2 and
Method
The time-range amenable to MDS is generally not sufficient for observation of spontaneous ligand access to or exit from the buried active site. In order to enhance the probability of such an event, an artificial force is imposed on the ligand in addition to the standard MD force field. In order to keep the bias introduced by the artificial force as small as possible, the following procedure, referred to as random expulsion molecular dynamics, was adopted. It is expected to be generally
Acknowledgements
We thank T. P. Straatsma for providing us with the ARGOS program and for helping us with technical problems. The authors also thank R. Gabdoulline, R. Abseher and P. Winn for critical reading of the manuscript. This work was partially supported by the EU Biotechnology Programme (BIO2-CT94-2060). S.K.L aknowledges an Erwin-Schrödinger Fellowship granted by the Austrian Fonds zür Förderung der Wissenschaftlichen Forschung (JO1379-CHE) and a Marie-Curie Fellowship by the EC (BIO4-CT97-5036).
References (38)
- et al.
Molecular dynamics study of unbinding of the avidin-biotin complex
Biophys. J.
(1997) - et al.
Temperature dependence of the structure and dynamics of myoglobin. A simulation approach
J. Mol. Biol.
(1990) - et al.
Towards molecular dynamics simulation of large proteins with a hydration shell at constant pressure
Biophys. Chem.
(1999) - et al.
How do substrates enter and products exit the buried active site of Cytochrome P450cam? 2. Steered molecular dynamics and adiabatic mapping of substrate pathways
J. Mol. Biol.
(2000) - et al.
A structure-based model for cytochrome P450cam-putidaredoxin interactions
Biochimie
(1996) - et al.
High-resolution crystal structure of cytochrome P450cam
J. Mol. Biol.
(1987) - et al.
Numerical integration of the Cartesian equations of motions in a system with constraintsmolec ular dynamics of n-alkanes
J. Comp. Phys.
(1977) - et al.
Mammalian microsomal cytochrome P450 monoxygenasestructural adaptations for membrane binding and functional diversity
Mol. Cell
(2000) - et al.
CHARMMa program for macromolecular energy, minimization, and dynamics calculations
J. Comp. Chem.
(1983) - et al.
Accessibility to internal cavities and ligand binding sites monitored by protein crystallographic thermal factors
Proteins: Struct. Funct. Genet.
(1998)
Molecular dynamics simulations of P450BM3 - examination of substrate-induced conformational change
J. Biomol. Struct. Dynam.
Electrostatic control of the substrate access channel in cytochrome P450cam
Biochemistry
Optical detection of cytochrome P450 by sensitizer-linked substrates
Proc. Natl Acad. Sci. USA
Enhanced sampling in molecular dynamicsuse of the time-dependent Hartree approximation for a simulation of carbon monoxide diffusion through myoglobin
J. Am. Chem. Soc.
Bioorganic activation of cytochrome P450cam
A computational procedure for determining energetically favorable binding sites on biologically important macromolecules
J. Med. Chem.
P450sstructural similarities and functional differences (1)
FASEB J.
Ligand binding and molecular mechanics calculation of the streptavidin-biotin rupture force
Science
Improved binding of cytochrome P450cam substrate analogues designed to fill extra space in the substrate binding pocket
Biochemistry
Cited by (297)
Revealing the conformational dynamics of UDP-GlcNAc recognition by O-GlcNAc transferase via Markov state model
2024, International Journal of Biological MacromoleculesRational engineering of cofactor specificity of glutamate dehydrogenase for poly-γ-glutamic acid synthesis in Bacillus licheniformis
2022, Enzyme and Microbial TechnologyPrediction of Drug Metabolism: Use of Structural Biology and In Silico Tools
2022, Comprehensive PharmacologyDevelopment of enhanced conformational sampling methods to probe the activation landscape of GPCRs
2022, Advances in Protein Chemistry and Structural BiologyMolecular Dynamics Study on the Binding Characteristics and Transport Mechanism of Polysaccharides with Different Molecular Weights in Camellia Oleifera Abel.
2024, Journal of Computational Biophysics and Chemistry
- 1
Edited by J. Thornton
- 2
Present address: V. Lounnas, Molecular Design and Informatics, N.V. Organon, Molenstraat 110, PO Box 20, 5340 BH Oss, The Netherlands; S. K. Lüdemann, Bioinformatics Unit, Institute of Molecular Pathology, Dr. Bohrg. 7, 1030 Vienna, Austria