Molecular dynamics (MD) simulations

JW Jason Wallach
AC Andrew B. Cao
MC Maggie M. Calkins
AH Andrew J. Heim
JL Janelle K. Lanham
EB Emma M. Bonniwell
JH Joseph J. Hennessey
HB Hailey A. Bock
EA Emilie I. Anderson
AS Alexander M. Sherwood
HM Hamilton Morris
RK Robbin de Klein
AK Adam K. Klein
BC Bruna Cuccurazzu
JG James Gamrat
TF Tilka Fannana
RZ Randy Zauhar
AH Adam L. Halberstadt
JM John D. McCorvy
request Request a Protocol
ask Ask a question
Favorite

Simulations were performed for each of two ligands (25CN-NBOH or 25N-N1-Nap (16) in complex with the human 5-HT2A serotonin receptor in a lipid bilayer of 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) and SPC water with enough Cl- and Na+ ions to neutralize the system at biologic salinity (0.15 M ionic strength). For the 25CN-NBOH system a model of the protein-ligand part of the system was made from chain A of the 6WHA PDB entry, downloaded from the OPM server88. Missing extracellular loops were filled in with those from chain A of PDB entry 6WGT. Schrödinger’s Protein Preparation Wizard89 was employed to use Prime90 to fill in missing side chains, and to use Epik91 to select the appropriate tautomers and protonation states of the protein and ligand. Due to anticipated difficulty modeling the long ICL3, the helices of TM5 and TM6 were capped, along with the N and C termini, with polar terminating residues: COOH for residue Q2625.66 and NH2 for I3156.27. During pilot simulations, entanglement of polar sidechains in the membrane occurred. To prevent this, it was found necessary that the rotamers for the loops and capped termini whose sidechains would otherwise interact with the polar heads of the lipid molecules be manually adjusted to ensure adequate water solvation at the beginning of the simulation. This was achieved by maximizing the magnitude of atomic coordinate z of terminal side-chain atoms, where the z-axis is perpendicular to the membrane and z = 0 locates the middle of the bilayer. Another issue identified during preliminary simulations was that DPPC tails would intrude into the orthosteric pocket between helices TM4 and TM5, causing the nitrile end of the ligand to project up toward the extracellular side of the pocket. On careful inspection, it was discovered that there is a stable position for a water molecule among residues D1202.50, S1623.39, and N3767.49. Manual introduction of a solvent molecule into this void volume successfully prevented lipid intrusion.

A disulfide bond between C1483.25 and C22745.50 18 was added to the topology, as well one between C3496.61 and C353ECL3. Each ligand was protonated at the basic nitrogen (as confirmed by Epik) and topology parameter files needed for subsequent dynamics simulation were determined using the ATB server [https://atb.uq.edu.au/]. Partial atomic charges for the ligands were computed via Schrödinger’s Jaguar tool using the density-functional method (B3LYP-D3 functional, 6-31 G** basis), with discrete charges derived from the geometry-optimized wavefunction using the Hirshfeld approach (Supplementary Table 4). The total protein comprised 278 residues with 2849 atoms. The remaining system consisted of 6933 water molecules, 29 Cl- ions, 20 Na+ ions, and 88 DPPC molecules. 25CN-NBOH has 44 atoms. The total number of atoms in the system is 28,137. The orthorhombic initial system dimensions were 61.319 × 60.911 × 97.412 Å3.

The protein-ligand complex for 25N-N1-Nap (16) starts with the structure from the top-ranked pose determined by induced fit docking (IFD). A chimeric homology model was created in Maestro using the IFD structure as the template everywhere except at the missing loops; these are taken from the 25CN-NBOH-system. This resulting protein-25N-N1-Nap (16) homology model was superposed (using the positions of the TM Cαs) onto the protein-ligand complex in the full (water-ion-lipid-protein-ligand) unrelaxed construct built for the 25CN-NBOH simulation. The protein-25CN-NBOH complex was then removed and the 25N-N1-Nap (16)-protein complex substituted in its place. As the IFD docking involved minimal changes to backbone conformation, the resulting model for the 25N-N1-Nap (16) ligand had helix and loop positions nearly identical to those of the 25CN-NBOH model, permitting a direct structural substitution. That said, prior to any further computations, a restrained minimization was carried out with only the membrane/solvent system fully mobile, to remove any high-energy interatomic overlaps that might have been inadvertently produced by the alignment and substitution. The 25N-N1-Nap (16) simulation comprised a total of 28,143 atoms (50 atoms for the ligand and otherwise identical to the 25CN-NBOH system). The orthorhombic initial system dimensions were 61.864 × 61.453 × 96.12 Å3.

Do you have any questions about this protocol?

Post your question to gather feedback from the community. We will also invite the authors of this article to respond.

post Post a Question
0 Q&A