Computational Drug Discovery and Design

(backadmin) #1
treatment of hepatotoxicity relies on timely diagnosis via laboratory
testing. Diagnostic assays and biosensors designed to quantify the
concentration of acetaminophen in a biological sample are often
based on aryl acylamidase [9], which cleaves the amide bond of
acetaminophen to yield p-aminophenol and acetate (Fig.2). The
concentration of acetaminophen can be inferred from changes in
absorbance following reaction of the product p-aminophenol with
a color-forming compound [10, 11], or amperometrically follow-
ing electrochemical oxidation of p-aminophenol [12].
A recent crystallographic structure characterized the binding
interaction of a bacterial aryl acylamidase with acetaminophen at
atomic detail, revealing key structural features that underlie its
enzymatic activity [13]. MD simulations of the crystallographic
complex can be used to probe the structure–function relationship
of aryl acylamidase, which represents a key component in diagnostic
assays and biosensors used to detect acetaminophen toxicity
[9]. The following section uses the aryl acylamidase–acetamino-
phen system as an example to describe a general protocol for
preparing and running simulations of protein–drug complexes.
Step-by-step instructions are provided for the example project
based on using the psfgen plugin from VMD [14] for system
setup and the NAMD [15] software for MD simulation.

3 Computational Protocol


3.1 Preparing a
Robust Structural
Model of the
Protein–Drug Complex


The critical prerequisite for any computational study is a robust
initial model. In the case of the example project for this chapter, a
crystallographic structure is available from the RCSB Protein Data
Bank (https://www.rcsb.org) (PDB ID: 4YJI) [13]. Importantly,
the structure describes the system at all-atom detail and contains no
missing regions (see Note 1). Experimental structures that are
incomplete require computational modeling to fill in missing struc-
tural features. Software like Modeller [16] and Rosetta [17] include
routines to perform de novo folding of protein loops and terminal
domains, regions that are often difficult to resolve experimentally
due to their flexible nature. Models based on cryo-electron micros-
copy or tomography data may require flexible fitting of structural

Fig. 2Schematic of the chemical reaction catalyzed by aryl acylamidase. The enzyme cleaves the amide bond
of acetaminophen to yield p-aminophenol and acetate


Molecular Dynamics Simulations of Protein-Drug Complexes 247
Free download pdf