Transcript A Multistep Approach to Structure

A Multistep Approach to Structure-Based Drug
Design: Studying Ligand Binding at the
Human Neutrophil Elastase
(T. Steinbrecher, D. A. Case, and A. Labahn)
Richard S. L. Stein
CS 379a
March 14, 2006
Modeling of HNE Inhibition
• Overabundance of human neutrophile elastase (HNE)
degrades healthy tissue and results in various diseases
• Inhibitory action on HNE modeled by ligand docking
calculations and molecular dynamics (MD) simulations of
several known inhibitors of HNE; structure dynamics
analyzed by MM-PBSA method
• MD runs of 2 ns exhibited instability of many ligandprotein complexes
• Binding free energies from MD simulations disagreed
with experiment
• The two most effective binding ligands, bornyl caffeate &
fukinolic acid, were each given 30 docking placements
divided into groups according to docking formation
Bornyl Caffeate & Fukinolic Acid
(color = distinct placement within HNE)
Docking Configuration Analysis
• MD simulation runs of 1 ns were performed on placement groups for
bornyl caffeate and fukinolic acid, and closely related compounds
• Most stable ligand-protein complexes were further analyzed by
thermodynamic integration (TI) calculations
• Binding free energies agreed with experiment better with bornyl
caffeate derivatives than with fukinolic acid derivatives
• MD simulation analysis of bornyl caffeate—HNE complex found a
prospective binding mode
• Bornyl ferulate (a bornyl caffeate derivative) was found to bind more
strongly to HNE with the transfer of a ring OH group
• Caffeic acid was found to bind significantly more weakly to HNE
than bornyl caffeate, indicating importance of ligand hydrophobicity
Advantages & Drawbacks
• Advantages:
– Combination of techniques (ligand docking, MD simulation and
TI calculations) discovered a possible binding mode within HNE
for bornyl caffeate
– Ability to propose binding modes for bornyl caffeate derivatives
(ex. one conformation of bornyl ferulate)
• Drawbacks:
– Failure to find binding mode for a good HNE inhibitor (fukinolic
acid and its derivatives)
– MM-PBSA has statistical error (1 kcal/mol) equal to differences in
binding free energies of similar placements for differing ligands
(1—2 kcal/mol)
– Length of simulations: could more reliable ligand-protein
complex stability data be obtained for runs longer than 1 ns?