Molecular Dynamics Studies of the Activation Mechanism of the Cannabinoid Receptor
Diane L.
Lynch, Judy Barnett-Norris, Rajnish Singh, Dow P. Hurst and Patricia H. Reggio
Kennesaw State University, Department of Chemistry and Biochemistry,
1000 Chastain Rd., Kennesaw, Georgia 30144
| Methods | Results | References | Conclusions | Acknowledgements |
INTRODUCTION:
The cannabinoid (CB) receptors are G protein coupled receptors (GPCRs) which
are integral membrane proteins that serve as an important link through which
cellular signal transduction mechanisms are activated. GPCRs can be thought
to exist in two states that are in equilibrium, i.e. the inactive (R) state
and the active (R*) state [1]. The binding of an endogenous ligand is thought
to induce a conformational change in the TMH bundle (R
R*) that initiates the signaling cascade.
The CB1 receptor sequence contains the highly conserved TMH6 CWXP motif, a motif that has been reported to act as a flexible hinge permitting an agonist-promoted movement of the cytoplasmic part of TMH 6 away from the receptor core and upwards toward the membrane bilayer [2].
In addition, charge neutralizing mutation experiments suggest that a stabilizing salt bridge involving the highly conserved D/ERY motif at the intracellular end of TMH3 and D/E6.30 at the intracellular end of TMH6 is broken upon activation [3]. This is also shown in Fig. 1.
Although the CB1 receptor shares many structure features in common with other GPCRs this receptor has several unusual aspects: (1) Highly lipophiliic endogenous ligands, such as Anandamide (AEA) [4] that are synthesized and degraded (via fatty acid amide hydrolase: FAAH) in the lipid environment (see Figure 2). The partitioning of AEA into lipid suggests that the ligand approaches the receptor via lipid rather than from the extracellular milieu. (2) High levels of constitutive (agonist independent) activity [6].
IMPORTANCE
of the 
motif: Ballesteros has proposed that -branching
residues (Val, Ile, or Thr) located (i,i+3) or (i,i+4) apart on an 
-helix
can form a hydrophobic groove into which an alkyl chain can fit [6]. Both the
crystal structures of adipocyte lipid binding protein (ALBP) complexed with
stearic acid [7] and the recent 2.8 Å structure of FAAH crystallized with the
inhibitor, methoxy arachidonyl phosphonate (MAP) [8], provide support for the
importance of a
motif as a recognition element for lipid-like substrates. In both cases the
acyl chain of the ligand is threaded between two -branched
residues. In addition, using the method of Conformational Memories (CM), we
have recently shown that the interaction of the alkyl tail of AEA (C16-C20)
with the TMH6 V6.43/I6.46 groove (in vacuo) produces a shift in conformer populations
consistent with the TMH 6 hinge-like motion thought to occur in the R to R*
transition [6].
HYPOTHESES:
Based on the above, we have hypothesized that the I6.46/V6.43 groove
on TMH6 in the CB1 receptor is an important recognition element that
1. forms
the initial point of contact for the endogenous agonist anandamide and therefore
the ligand must have both the proper depth and orientation of the alkyl tail
in the lipid bilayer.
2. forms a point of contact between the acyl tails of the lipid itself
and the receptor providing a mechanism for ligand independent activation
(constitutive activity) of the CB1 receptor.
APPROACH:
In order to test these hypotheses, NAMD lipid simulations were undertaken.
In order to
access the likelyhood of the above 2 proposals we have initiated a series of
MD simulations, using the NAMD software package[9]. We have simulated (separately)
TMH6 of CB1 and AEA in a DOPC lipid bilayer. From the resulting trajectories:
1. We can evaluate the probability that acyl chains interact with the
groove, modulate the bend angle associated with activation of the receptor.
2. We can address the location of the agonist in a lipid bilayer and observe
if in fact the alkyl tail of the AEA is at the proper depth for interaction
with the groove.
METHODS: MOLECULAR DYNAMICS
PURE DOPC: We have built, hydrated and equilibrated (2ns) both a low (5.4
waters/lipid) as well as a fully hydrated DOPC bilayer (28 waters/lipid). The
purpose of this phase of the simulations was to generate a reasonable representation
for the lipid bilayer. The calculations were run using an NPT (P = 1atm, T=310K)
ensemble and the CHARMM27 set of all atom parameters designed for lipid simulations
[10].
TMH6 of
CB1 in DOPC: The CB1 transmembrane helix 6 (TMH6) was built as an alpha
helix. The TMH6 initial conformation was taken from earlier in vacuo simulations
of the full receptor in the R state. The simulation protocol is described below:
1. The helix was placed in
the pre-equilibrated DOPC lipid.
2. Overlapping lipids (2 lipids/bilayer) were removed with total number of atoms
remaining nearly constant
3. TMH6 was frozen during a 5000 step energy minimization.
4. The backbone TMH6 atoms only were frozen during a 5000 step energy minimization.
5. TMH6 backbone atoms were constrained during a 5000 step energy minimization.
6. Using an NVT ensemble, the system was heated to to 310K and equilibrated
for 200ps.
7. After 200ps, the simulation was switched to NPT ensemble conditions and run
for 2.5ns.
RESULTS:
A snapshot of TMH6 in a hydrated DOPC bilayer, along with a definition of the
z direction, is displayed in Figure 3. The kink angle about the P6.50 was computed
using ProKink [11]. Figure 4 illustrates the time evolution of the proline kink
angle. Clearly, TMH6 undergoes bending/flexing about P6.50 with the most bent
structure having a proline kink of about 130o.
Interaction
of lipid chains with the groove residues: A typical example of an acyl tail
in the groove is displayed in Figure 5. The acyl
tail was defined to be in the groove if a carbon atom in the acyl chain was
within 4.0 Å of either the I6.46 or V6.43 C atom
and simultaneously a 2nd acyl carbon atom was within 4.0 Å of these same residue
sidechain atoms. These distance criteria were based upon relative distances
between acyl chain atoms and groove residue C’s
in the ALBP [7] and FAAH [8] crystal structures.
Our preliminary results suggest that the percentage of helices with an acyl tail in the groove increases when the helix is straight, i.e. the presence of an alkyl tail in the groove straightens the helix, leading to a conformation of the helix that is representative of the activated state of the receptor.
However, whether this trend has any statistical significance needs to be explored. We are currently running the simulation to longer times to observe the frequency with which the helix bends, as well as to collect the data necessary to establish the likelihood that the acyl chains are interacting with the TMH6 groove and this interaction is correlated with the P6.50 bend angle.
REFERENCES:
1. Leff, P. Trends Pharmacol. Sci. 16: 89-97(1995).
2. Jensen, A.D. et al., J. Biol. Chem., 276:9279 (2001).
3. Ballesteros, J.A. et. al. J. Biol. Chem., 276:29171 (2001).
4. Reggio, P. Curr. Med. Chem. 6:665(1999).
5. Kearn, C.S. et. al. J. Neurochem. 72:2379 (1999).
6. Barnett-Norris, J. et al. Int. J. Quantum Chem. 88:76 (2002).
7. Xu, Z. et. al. Biochem. 31:3484 (1992) and J. Biol. Chem. 268:7874
(1993).
8. Bracey, M.H. et. al. Science 298: 1793 (2002).
9. Kale, M., et. al. J. Comp. Phys., 151:283 (1999).
10. Feller, S.E., et. al. J. Am. Chem. Soc., 74:2419 (2002).
11. Visiers I. et al. Protein Eng 13(9):603-606 (2000).
12. Isele, J. et al. Biophys. J. 79: 3063-3071 (2000).
13. Weiss, T. et al. Biophys. J. 84: 379-385 (2003).
Anandamide (AEA) in DOPC: In a separate MD simulation AEA was placed in the DOPC bilayer. The simulation was run using the same protocol as described previously for TMH6 in DOPC. As illustrated in Figure 6, AEA was found to adopt an extended conformation in DOPC. A plot of the density distribution of the AEA terminal methyl group along with the density distribution of the I6.46 and V6.43 groove residues shows that the tail of AEA reaches the depth in the bilayer necessary to interact with the I6.46 groove residue (see Fig. 6, right).
It should be
noted that there are several uncertainties inherent in these initial calculations:
1) Bilayer thickness has been reported to vary for a single TM helix vs. a TMH
bundle [12,13]. Modification of the membrane thickness will alter the depth
that the AEA achieves in the bilayer.
2) Presence of the full receptor may alter the location of the groove itself
due to interhelical interactions. These uncertainties will be addressed in future
calculations that will include the entire receptor in the simulation.
CONCLUSIONS: The simulations described here suggest the following:
The I6.46/V6.43 groove residues can interact with lipid acyl chains. Preliminary statistical analysis suggests that the bend angle in TMH6 can be correlated to the presence of lipid tails in the groove area. We will continue to access the statistical significance of these early results.
Anandamide, the endogenous ligand for the CB1 receptor, has the depth necessary in order to interact with the groove residues. The headgroup of AEA resides in the glycerol region of the phospholipid bilayer, in agreement with recent NMR data.
The precise location of the ligand with respect to the groove will be explored with further MD studies wherein the entire receptor is included in the simulation. Not only will this take into account adjustment of TMH6 when the entire receptor is present, but also the modification of the lipid thickness (change in curvature) when multiple transmembrane helices are present.
ACKNOWLEDGEMENTS:
This work was supported by a starter grant from the Pittsburgh Super Computer
Center; by the Theoretical Biophysics group, an NIH Resource for Macromolecular
Modeling and Bioinformatics, at the Beckman Institute, University of Illinois
at Urbana-Champaign through the use of the Molecular Dynamics package, NAMD,
and the affiliated visualization package, VMD, made freely available at http://www.ks.uiuc.edu;
and, by NIDA grant RO1 DA03934.