Zr. Wasserman et Cn. Hodge, FITTING AN INHIBITOR INTO THE ACTIVE-SITE OF THERMOLYSIN - A MOLECULAR-DYNAMICS CASE-STUDY, Proteins, 24(2), 1996, pp. 227-237
We used molecular dynamics computer simulation to ''fly'' a small flex
ible ligand, L-leucine hydroxamic acid, into the active site of thermo
lysin. The potential, which imposed no constraints between protein and
ligand, produced conformations close to the crystallographically dete
rmined one. The calculations made use of the combined molecular mechan
ics/grid method of Luty et al. (J. Comp. Chem. 16:454-464, 1995), in w
hich atoms of the active site are free to move whereas the rest of the
protein, assumed to be rigid, is represented as points of a grid, and
which also includes an implicit solvation model, The method is suffic
iently fast that large sets of simulations could be carried out, enabl
ing statistical sampling and exploration of the effect of initial posi
tion and conformation of the ligand on the probability of successful d
ocking. In a charged catalytic Glu/uncharged ligand regime, when the i
nitial position of the ligand was determined by random translations ar
id rotations that kept the center of mass within 8.0 Angstrom Of the c
rystal one, none of the 20 runs placed the ligand correctly, In a seco
nd set with uncharged Glu and zwitterionic ligand, 3 of 24 similarly p
laced random structures flew the ligand in sucessfully. In a third set
with the same protonation scheme as the second the starting positions
had randomly determined conformations but kept the hydroxamate oxygen
s within 4.0 Angstrom of the zinc; in this case 22 of 25 runs reorient
ed correctly, A diverse set of energetic, structural, and dynamic crit
eria was used for evaluation of the calculations. The results indicate
the method to be a promising tool for the rational drug design proces
s. (C) 1996 Wiley-Liss, Inc.