Allostery Inhibition of BACE1 by Psychotic and Meroterpenoid Drugs in Alzheimer's Disease Therapy

Abstract

In over a century since its discovery, Alzheimer's disease (AD) has continued to be a global health concern due to its incurable nature and overwhelming increase among older people. In this paper, we give an overview of the efforts of researchers towards identifying potent BACE1 exosite-binding antibodies and allosteric inhibitors. Herein, we apply computer-aided drug design (CADD) methods to unravel the interactions of some proposed psychotic and meroterpenoid BACE1 allosteric site inhibitors. This study is aimed at validating the allosteric potentials of these selected compounds targeted at BACE1 inhibition. Molecular docking, molecular dynamic (MD) simulations, and post-MD analyses are carried out on these selected compounds, which have been experimentally proven to exhibit allosteric inhibition on BACE1. The SwissDock software enabled us to identify more than five druggable pockets on the BACE1 structural surface using docking. Besides the active site region, a melatonin derivative (compound 1) previously proposed as a BACE1 allostery inhibitor showed appreciable stability at eight different subsites on BACE1. Refinement with molecular dynamic (MD) simulations shows that the identified non-catalytic sites are potential allostery sites for compound 1. The allostery and binding mechanism of the selected potent inhibitors show that the smaller the molecule, the easier the attachment to several enzyme regions. This finding hereby establishes that most of these selected compounds failed to exhibit strong allosteric binding with BACE1 except for compound 1. We hereby suggest that further studies and additional identification/validation of other BACE1 allosteric compounds be done. Furthermore, this additional allosteric site investigation will help in reducing the associated challenges with designing BACE1 inhibitors while exploring the opportunities in the design of allosteric BACE1 inhibitors.

Keywords: Alzheimer’s disease; BACE1; allosteric inhibitor; molecular docking; molecular dynamics (MD) simulations; multisite targeting.

Figure 1

Illustration of APP amyloidogenic and non-amyloidogenic pathways [1,11]. The cleavage mechanism by γ-secretase is non-amyloidogenic, while β-secretase or α-secretase cleavage results in the amyloidogenic pathway. CTFα and CTFβ are alpha and beta carboxy-terminal fragments, AICD represents the amyloid precursor protein intracellular domain.

Figure 2

Illustration of BACE1 active site (primary binding site), binding subsites (secondary sites), and anti-BACE1 attachment [20]. The characteristic catalytic Asp32/228 dyad hydrolyzes the substrate or reversible inhibitor; other non-catalytic residues are involved if an inhibitor binds at secondary sites.

Figure 3

Zoomed-in 3D snapshot of BACE1 crystal structure complexed with compound 12 (PDB code: 5MXD) devoid of Asp32/228 interaction [32]. The compound binds at the flap region of BACE1 and we generated this image with the Discovery Studio R2017 [35].

Figure 4

The selected potential allosteric inhibitors of BACE1 with reported IC₅₀ values (no data for anisoperidone and * indicates extrapolated from 88% inhibition of BACE1 at 5 μM). Besides LY2811376, all the molecules are naturally occurring compounds or their derivatives. LY2811376 0.9 nM [38,39]. Gamma-linolenic acid 76 μM [31]. Sargahydroquinoic acid 4.4 μM [40]. Anisoperidone [41]. Compound 1 2.84* μM [37].

Figure 5

Binding modes of the allostery inhibitors on BACE1 through docking. LY2811376 converges mainly at the active site and partially on top of the flap and gamma-linolenic acid binds favorably at the primary site and two other secondary sites as predicted previously [31]. Sargahydroquinoic acid and anisoperidone interact at the active site and extend to the flap tip, while compound 1 binds at several regions.

Figure 6

The interaction profile of compound 1 for its various poses on the BACE1 scaffold. The first and second values below each pose are full fitness and binding energy in kcal/mol, respectively. Allos. site signifies allosteric site, the dashed green and blue lines represent classical hydrogen bond and bridging water HB, respectively.

Figure 7

Evolution of the alpha carbon backbone atom root-mean-square deviation (RMSD/angstrom) over 120 ns molecular simulations for compound 1 binding to various regions on BACE1 structure. Allos. site represents the allosteric site and the definition for each site including the interaction residues is available in Figure 6.

Figure 8

Protein secondary structure prediction per 385 residues over time 120 ns (120,000 frames) for compound 1 binding to BACE1 active site and allosteric site 1.

Figure 9

Per residue (a total of 385) fluctuation through all-atom root-mean-square fluctuation (RMSF/angstrom) scoring for compound 1 binding to several regions on BACE1. Allos. site represents the allosteric site and the definition for each site including the interacting residues is available in Figure 6.

Figure 10

Time (in nanoseconds) evolution of the mass-weighted RoG—radius of gyration (in angstrom) induced by coupling compound 1 to the various regions on BACE1 structure. Allos. site represents the allosteric site and the definition for each site including the interacting residues is available in Figure 6.