Folding of the Apolipoprotein A1 Driven by the Salt Concentration as a Possible Mechanism to Improve Cholesterol Trapping - Selected Topics of Computational and Experimental Fluid Mechanics - page 271

Computer simulations have come to play an increasingly important role in the
understanding of protein folding, protein-protein interactions and the surface activ-
ity of proteins, to name but a few. Some of the advantages of atomistic simulations
are the complete control over the physicochemical variables of the problem and the
essentially exact solution of the equations of motion that govern the nature of the
system, not tomention the vivid representations of the spatial conformations of mole-
cules and their evolution with time. In this work we focus mainly on predicting how
APOA1 folds as the salt content is increased, using atomistically detailed computer
simulations. In particular, we study the contraction followed by re-expansion of some
residues of APOA1 induced by the increasing concentration of monovalent salt ions.
This phenomenon has been found to occur in other polyelectrolytes in aqueous solu-
tion using a variety of methods (Hsiao and Luijten 2006 ; Feng et al. 2009 ; Alarcón
et al. 2013 ; Pollard et al. 2013 ; Frank and Marcel 2000 ). In this context, APOA1
is a much more complex system because there are positive, negative and neutral
sequences along its structure, as well as hydrophilic and hydrophobic residues. We
expect that the spatial conformation is a crucial factor in determining the ability of
APOA1 to capture cholesterol and other fatty acids, which is the functionality we
would like to optimize so that a mechanism can be proposed to help design drugs
and treatments that improve the quality of life.
2 Model and Methodology
We start by taking the fragment of APOA1 known as 1GW3 from the Protein Data
Bank (PDB, http://www.rcsb.org ; Wang et al. 1997 ), which has 142-187 aminoacid
residues from the complete APOA1. It is computationally very demanding to
model the entire protein, while only a fraction of its residues is responsible for
the folding we are interested in. We take into account only a 20% of the orig-
inal protein. Therefore, we work only with the above mentioned fragment and
set up molecular dynamics simulations with it and varying NaCl concentrations,
ranging from 0.01M up to 2.0M. For the interatomic interactions we used the
Lennard-Jones model (Allen and Tildesley 1987 ), while for the electrostatic inter-
actions we used the so called Particle Mesh Ewald (PME) method (Darden et al.
1993 ). To conserve the bonds between the atoms that make up the protein we used
LINCS (Linear Constraint Solver, Hess et al. 1997 ). The force field parameters
for the protein were taken from OPLS (Optimized Potential for Liquid Simula-
tions) (Jorgensen and Tirado-Rives 1988 ), and the water model used was SPCE
(Berendsen et al. 1987 ). Then all interactions are solved using GROMACS 4.6.4
(Spoel et al. 2005 ), where the interactions are calculated at every time step using the
Verlet scheme (Pall and Hess 2013 ) with a grid scheme for GPU's, which allow us to
perform large simulations. The cut off distance for the Lennard-Jones and electrosta-
tic interactions was equal to 1.0nm. A leapfrog algorithm (Snyman 2000 ) was used
for the calculation of the atoms positions and velocities. The energy minimization
was performed using the steepest descends method (Chaichian and Demichev 2001 ).