Data from: All-atom molecular dynamics simulation and rate calculation for norepinephrine binding beta adrenergic receptor

This dataset is for the manuscript "A multiscale predictive digital
twin for neurocardiac modulation". We utilized all-atom molecular
dynamics (MD) simulations to predict the affinities as well as association
and dissociation rates of norepinephrine (NE) with the beta-2 adrenergic
receptor (β2AR). Calculations were done for both cationic (+) and neutral
(0) forms of NE. An active-state β2AR model based on its crystal structure
PDB ID: 4LDO was used with the flexible intracellular loop 3 (ICL3)
truncated to 3 residues. ROSETTALIGAND was used for docking NE(+) and
NE(0) to the receptor's orthosteric ligand binding site with one of
the top scoring models used for MD simulations. MD simulation systems of
the NE(+) or NE(0) bound β2AR receptor embedded in a
palmitoyloleoylphosphatidylcholine (POPC) bilayer solvated by aqueous 0.15
M NaCl was generated using CHARMM-GUI membrane builder. MD simulations
were run with NAMD in the NPT ensemble at 310 K and 1 atm pressure using
CHARMM36m force field for protein, C36 for lipids, CGENFF based parameters
for NE(+) and NE(0), and TIP3P water model. Initial equilibration MD
simulations were run for ~52 ns with the first ~2 ns using gradually
reduced harmonic restraints on the protein and lipid tail carbon atoms
followed by unrestrained 50 ns long MD simulation runs. They were used to
start 200 ns long well-tempered metadynamics MD simulations to estimate
ligand binding affinities and rates. Distance between centers of mass
(COM) of protein alpha-helical core and NE was used as reaction coordinate
R extending to 50 Å in bulk, whereas angle θ between vectors connecting
these points and ligand orthosteric binding site was used to restrain
sampling within θ≤30° cone. Affinities in the form of dissociation
constants, Kd, were computed as was done previously (Provasi D et al
Biochemistry 2009, 48:10020) NE Kd value at pH=7.4 was computed from
corresponding estimates for NE(+) and NE(0) using Henderson-Hasselbach
equation as was done previously (Yang PC et al Circ Res 2020, 126:947)
NE(+) dissociation or “off” rate was estimated using Kramer’s rate
formalism (Dorairaj S & Allen TW. PNAS 2007, 104:4943) Free energy
profile from metadynamics MD simulation of the β2AR – NE(+) system was
used to estimate activation free energy barrier height and corresponding
curvatures at free energy maxima and minima. Diffusion coefficient at the
free energy maximum was estimated using position autocorrelation function
from last 20 ns of separate 30 ns long MD simulations with harmonic
restraints on the NE(+) position as was done previously (Vorobyov I et al
Biophys. J. 2014, 106:586) The z component of a distance between COM of
protein α-helical core and NE(+) was used as a reaction coordinate for
diffusion coefficient calculations. NE(+) association or “on” rate was
then computed as a ratio of the "off" rate and Kd.