 |
|
|
|
Aquaporin Inhibition: Identification of the binding site for TEA in hAQP1
|
|
E. Matthias Müller, Bert L. de Groot and Helmut Grubmüller
A prerequisite of performing a structure based approach in rational drug design is to know the binding site of the protein that should be targeted. In case of human Aquaporin 1 (hAQP1) we started with a homology model, built and refined by B. L. de Groot. Furthermore, from oocyte swelling experiments it is known that TEA (tetraethylammonium) inhibits hAQP1.
|
| |
|
To identify the binding site for TEA in hAQP1 'blind docking' in the extracellular part of the pore was performed using DOCK 4.0. About 100 equally ranked poses for TEA were obtained. To refine these results, 4 representative poses were used as starting positions for molecular dynamics (MD) simulations. The MD simulations function as a 'stability test' to sort out poses which do not represent probable binding sites.
In a 15ns simulation TEA stays only bound at one position.
|
| |
|
As a test, we performed a second simulation. Here, in each pore TEA was placed at the stable position specified in the 'stability test'.
In this second simulation all 4 TEA remained bound. This simulation was used to characterize the binding region for TEA in hAQP1.
|
| |
|
For the first time a probable binding site for TEA was identified! Hence, we showed that refining docking results with molecular dynamics simulations is a valuable method for identifying binding sites.
|
| |
|
'First Principles Docking'
|
|
It is a challenge to identify the binding site of a protein if 'direct' experimental data like an x-ray-, nmr- or an electron microscope structure of the protein with a bound inhibitor or an analog are not available. One way to solve this problem is to combine 'blind docking' and molecular dynamics simulations (shown here).
Here, an alternative approach is presented. 20 TEA molecules were randomly placed in the bulk water of the hAQP1 simulation system (full atomistic description of the protein, the membrane, the ions, and water molecules. See Figure, for clarity, only the protein as cartoon and TEA as cpk are shown.)
|
| |
|
Subsequently, a conventional molecular dynamics simulation was performed. During the 80ns simulation 3 binding and 2 unbinding events of TEA have been observed.
All binding events started after about 5-10ns. Two binding events lasted for about 15-20ns (snapshot: 3 TEA bound, left side). The third one lasted until the end of the 80ns. Here, TEA bound at the same binding site that already has been identified previously.
|
| |
|
Structure-based approach to find a better binder than TEA
|
|
Here is a brief sketch of the steps that I use in my virtual screening approach:
-
Derive pharmacophor model from prior informations (e.g. intuition, theoretical approaches, experiments, ...)
-
do a pharmacophore search on database
-
dock molecules found via pharmacophore search to binding site (e.g. AUTODOCK, ...)
-
take best 10 docking solutions of each docked molecule
-
re-apply pharmacophore search on docking results
-
use results for a consensus scoring procedure
-
test results experimentally
-
refine pharmacophore model and restart from the beginning unless you found a 'good candidate' for a drug
|
| |
|
References
|
|
-
F. J. M. Detmers, B. L. de Groot, E. M. Müller, A. Hinton, S. L. Flitsch, H. Grubmüller, and P. M. T. Deen.
Quaternary ammonium compounds as aquaporin water channel blocker: specificity, potency and putative site of action ,
J. Biol. Chem. 281, 14207–14214 (2006)
[pdf]
|
| |
|
Top