rhodopsin dimers: molecular dynamics simulations using ... · membranes, support a molecular model...

1
Rhodopsin Dimers: Molecular Dynamics Simulations Using Discrete Representations of the Membrane and Water Environment Department of Physiology and Biophysics Weill Medical College of Cornell University, New York, New York 10021 Simon X. Wang, Marta Filizola, Marc Ceruso, Harel Weinstein INTRODUCTION Displacement Displacement Local tilt Local tilt Local tilt Lengthening Lengthening Winding/unwinding Bending Proline Kink Pro G protein-coupled receptors (GPCRs) comprise by far the largest family of cell surface proteins involved in signaling across the plasma membrane and are implicated in numerous diseases. Direct evidences for homo- and/or hetero- dimerization of various GPCRs reinforces the need for the incorporation of these key phenomena into any physiologically relevant functional models of these receptors, especially in light of the recently discovered functional implications of GPCR association, which include pharmacological diversity, G-protein coupling, downstream signaling, and internalization. Both computational and experimental efforts have been made to under- stand the basis of protein-protein interaction in GPCR dimerization. Our computational approaches based on correlated mutation method and three-dimensional molecular models of the transmembrane segments of GPCRs elucidated the likely molecular determinants required for dimer- ization [1, 2]. Recent experimental data coming from an atomic-force mi- croscopy (AFM) map of rhodopsin molecules in their native mouse disk membranes, support a molecular model of rhodopsin monomers orga- nized into two dimensional arrays of dimers [3]. Specifically, transmem- brane (TM) helices 4 and 5 of rhodopsin appear to be involved in intradi- meric contact, whereas helices TM1 and TM2, and the cytoplasmic loop connecting helices TM5 and TM6 are inferred to facilitate the formation of rhodopsin dimer rows. COMPUTATIONAL METHODS System Build Two different hydrated palmitoyl-oleoyl-phosphatidyl cho- line (POPC) bilayer unit cells were constructed and equilibrated for 2 ns. The proteins were then embedded into the unit cells according to major criteria well-established for GPCR family [4]. The resulted systems are the following: Simulation Setup All simulations are performed for an NPT ensemble using GROMACS 3.2.1 package. Lipid parameters were taken from Berger et al. [6] and the water model was SPC. The system was simulated with semiisotropic coupling, with the pressure at 1.0 bar, coupled sepa- rately to the xy plane and z directions. The Particle Mesh Ewald interpola- tion order was set to 6 and the maximum grid spacing for the FFT was set to 0.12 nm. Helix Analysis New algorithms were developed [5] for the analysis of helix properties and their evolution in the dynamics simulations. (Simulaid: http://fulcrum.physbio.mssm.edu/~mezei/simulaid/) 1. Filizola, M., Olmea, O. & Weinstein, H. (2002) Prot. Eng. 15, 881-885. 2. Filizola, M. & Weinstein, H. (2002) Biopolymers (Peptide Science) 66, 317-325. 3. Liang, Y., Fotiadis, D., Filipek, S., Saperstein, D. A., Palczewski, K. & Engel, A. (2003) J. Biol. Chem. 278, 21655-21662. 4. Visiers, I., Ballesteros, J. A. & Weinstein, H. (2002) Meth Enzymol 343, 329-371. 5. Mezei, M., Filizola, M. manuscript in preparation. 6. Berger, O., Edholm, O. & Jahnig, F. (1997) Biophys J 72, 2002-2013. Helix analysis on 20 ns trajectories emphasizes the differences between the rhodopsin monomer and at least one of the interacting subunits of the dimer. Specifically, differences in some of the calculated helix properties were observed thus far for i) the helix rotation of TM4, TM5, TM7, and H8 around their own axes; ii) the displacement of TM4, TM5, and H8 in the x and/or y direction. TM4 TM5 monomer subunit A of dimer subunit B of dimer H8 Since the nature and geometry of the interface of GPCR dimers is essen- tial to the understanding of the role and implications of these phenomena in the functional mechanisms of these receptors, we take advantage of the information available for rhodopsin to carry out a computational study that could provide detailed insights into the determinants of GPCR dimer- ization in a structural context. As a first step, molecular dynamics (MD) simulations are carried out at the systems for which direct structural infor- mation is available, using discrete representations of the membrane and water environment. REFERENCE CONCLUSION Preliminary results from the 20 ns trajectories of rhodopsin dimer and monomer suggest subtle structural changes at the dimer interface. Subtle changes were also observed at distant domains that are critical to receptor activation. Control simulations of various lengths and using different equilibration protocols are ongoing to assess the accuracy of simulations and of the inferences regarding conformational changes. Rotation Helix rotation difference between rhodopsin monomer and dimer from prelimi- nary computational studies. The results are illustrated by dial-type diagram with time on the x- axis and the rotation marked as degrees on the dial. TM5 TM4 TM7 H8 monomer subunit A of dimer subunit B of dimer Helix displacement difference between rhodopsin monomer and dimer from preliminary computational studies. The line represents movement of center of mass (COM) for helix in the x, y direction for different time segment of the simulation identified by the color scale. The current trajectories length is 20 ns for both monomer and dimer. The Left Panel above records the rms deviation of TM bundle of rhodopsin monomer, subunit A and B and whole dimer from the crystallographic structure. Comparison to the crystallographic B-factors (Right Panel) shows that the overall trends of the C-alpha atom rms fluctuations (Middle Right) of rhodopsin dimer subunit A (color red) and B (color blue) are comparable to the B-factors (Top Right). The TM helical regions of rhodopsin exhibit less mobility while the terminal tails and loops show higher fluctuations. The equilibrium structures deviate from the crystal structure among these flexible segments (Bottom Right) and adopt to a new equilibrium conformation. RESULTS Dimer Monomer 16*12.4*9.8nm 3 11.2*11.2*9.8nm 3 Box size 365 hr/ns 241 hr/ns Benchmark Total Atoms Na+ SPC Water POPC Lipid Structural Water Protein 144,638 92,057 2 1 35,852 23,318 567(640) 352(390) 144 72 1N3M 1L9H A serpentine diagram of the rhodopsin receptor. Black circles represent resi- dues that are conserved among class A GPCRs.

Upload: others

Post on 22-Aug-2021

6 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Rhodopsin Dimers: Molecular Dynamics Simulations Using ... · membranes, support a molecular model of rhodopsin monomers orga-nized into two dimensional arrays of dimers [3]. Specifically,

Rhodopsin Dimers: Molecular Dynamics Simulations Using Discrete Representationsof the Membrane and Water Environment

Department of Physiology and Biophysics Weill Medical College of Cornell University, New York, New York 10021Simon X. Wang, Marta Filizola, Marc Ceruso, Harel Weinstein

INTRODUCTIONDisplacementDisplacement

Local tiltLocal tiltLocal tilt

LengtheningLengthening

Winding/unwinding

Bending

Proline KinkPro

G protein-coupled receptors (GPCRs) comprise by far the largest family of cell surface proteins involved in signaling across the plasma membrane and are implicated in numerous diseases. Direct evidences for homo- and/or hetero- dimerization of various GPCRs reinforces the need for the incorporation of these key phenomena into any physiologically relevant functional models of these receptors, especially in light of the recently discovered functional implications of GPCR association, which include pharmacological diversity, G-protein coupling, downstream signaling, and internalization.

Both computational and experimental efforts have been made to under-stand the basis of protein-protein interaction in GPCR dimerization. Our computational approaches based on correlated mutation method and three-dimensional molecular models of the transmembrane segments of GPCRs elucidated the likely molecular determinants required for dimer-ization [1, 2]. Recent experimental data coming from an atomic-force mi-croscopy (AFM) map of rhodopsin molecules in their native mouse disk membranes, support a molecular model of rhodopsin monomers orga-nized into two dimensional arrays of dimers [3]. Specifically, transmem-brane (TM) helices 4 and 5 of rhodopsin appear to be involved in intradi-meric contact, whereas helices TM1 and TM2, and the cytoplasmic loop connecting helices TM5 and TM6 are inferred to facilitate the formation of rhodopsin dimer rows.

COMPUTATIONAL METHODS

System Build Two different hydrated palmitoyl-oleoyl-phosphatidyl cho-line (POPC) bilayer unit cells were constructed and equilibrated for 2 ns. The proteins were then embedded into the unit cells according to major criteria well-established for GPCR family [4]. The resulted systems are the following:

Simulation Setup All simulations are performed for an NPT ensemble using GROMACS 3.2.1 package. Lipid parameters were taken from Berger et al. [6] and the water model was SPC. The system was simulated with semiisotropic coupling, with the pressure at 1.0 bar, coupled sepa-rately to the xy plane and z directions. The Particle Mesh Ewald interpola-tion order was set to 6 and the maximum grid spacing for the FFT was set to 0.12 nm.

Helix Analysis New algorithms were developed [5] for the analysis of helix properties and their evolution in the dynamics simulations. (Simulaid: http://fulcrum.physbio.mssm.edu/~mezei/simulaid/)

1. Filizola, M., Olmea, O. & Weinstein, H. (2002) Prot. Eng. 15, 881-885.2. Filizola, M. & Weinstein, H. (2002) Biopolymers (Peptide Science) 66, 317-325.3. Liang, Y., Fotiadis, D., Filipek, S., Saperstein, D. A., Palczewski, K. & Engel, A. (2003) J. Biol. Chem. 278, 21655-21662.4. Visiers, I., Ballesteros, J. A. & Weinstein, H. (2002) Meth Enzymol 343, 329-371.5. Mezei, M., Filizola, M. manuscript in preparation.6. Berger, O., Edholm, O. & Jahnig, F. (1997) Biophys J 72, 2002-2013.

Helix analysis on 20 ns trajectories emphasizes the differences between the rhodopsin monomer and at least one of the interacting subunits of the dimer. Specifically, differences in some of the calculated helix properties were observed thus far for i) the helix rotation of TM4, TM5, TM7, and H8 around their own axes; ii) the displacement of TM4, TM5, and H8 in the x and/or y direction.

TM4

TM5

monomer subunit A of dimer subunit B of dimer

H8

Since the nature and geometry of the interface of GPCR dimers is essen-tial to the understanding of the role and implications of these phenomena in the functional mechanisms of these receptors, we take advantage of the information available for rhodopsin to carry out a computational study that could provide detailed insights into the determinants of GPCR dimer-ization in a structural context. As a first step, molecular dynamics (MD) simulations are carried out at the systems for which direct structural infor-mation is available, using discrete representations of the membrane and water environment.

REFERENCE

CONCLUSIONPreliminary results from the 20 ns trajectories of rhodopsin dimer and monomer suggest subtle structural changes at the dimer interface. Subtle changes were also observed at distant domains that are critical to receptor activation. Control simulations of various lengths and using different equilibration protocols are ongoing to assess the accuracy of simulations and of the inferences regarding conformational changes.

Rotation

Helix rotation difference between rhodopsin monomer and dimer from prelimi-nary computational studies. The results are illustrated by dial-type diagram with time on the x- axis and the rotation marked as degrees on the dial.

TM5

TM4

TM7

H8

monomer subunit A of dimer subunit B of dimer

Helix displacement difference between rhodopsin monomer and dimer from preliminary computational studies. The line represents movement of center of mass (COM) for helix in the x, y direction for different time segment of the simulation identified by the color scale.

The current trajectories length is 20 ns for both monomer and dimer. The Left Panel above records the rms deviation of TM bundle of rhodopsin monomer, subunit A and B and whole dimer from the crystallographic structure. Comparison to the crystallographic B-factors (Right Panel) shows that the overall trends of the C-alpha atom rms fluctuations (Middle Right) of rhodopsin dimer subunit A (color red) and B (color blue) are comparable to the B-factors (Top Right). The TM helical regions of rhodopsin exhibit less mobility while the terminal tails and loops show higher fluctuations. The equilibrium structures deviate from the crystal structure among these flexible segments (Bottom Right) and adopt to a new equilibrium conformation.

RESULTS

DimerMonomer

16*12.4*9.8nm311.2*11.2*9.8nm3Box size

365 hr/ns241 hr/nsBenchmark

Total Atoms

Na+

SPC Water

POPC Lipid

Structural Water

Protein

144,63892,057

21

35,85223,318

567(640)352(390)

14472

1N3M1L9H

A serpentine diagram of the rhodopsin receptor. Black circles represent resi-dues that are conserved among class A GPCRs.