Page 117 - ITPS-7-2

INNOSC Theranostics and
Pharmacological Sciences PI3K-α inhibitors for cancer immunotherapy

Epik wizard of Maestro to improve virtual screening the predicted poses. However, in Schrödinger Maestro, the
enrichment. Moreover, the ligand was desalted to remove binding affinity was calculated as (Equation III):
extra molecules, such as water molecules and counter- Binding affinity = ϑ + ϑ + ϑ + ϑ + ϑ + ϑ (III)
ions, present in the ligand file from certain structure L/VdW E Hb SM π Halogen
databases. To define the stereoisomers, the computation where ϑ L/vdW is the energy contribution associated with
was restrained to determine chirality from the 3D lipophilic pair concerning total Van der Waals force of
structure of P5J to account for keto-enol tautomerization, interaction, ϑ is the energy contribution associated with
E
analogous sulfur and nitrogen tautomerizations, as well electrostatic interactions, ϑ is the energy contribution
Hb
as histidine- and DNA-base tautomerizations. However, associated with hydrogen bond interaction, ϑ is the
SM
the generation of tautomeric forms of P5J was avoided to energy contribution associated with site map interactions,
prevent tautomeric duplicates while maintaining accuracy, ϑ is the energy contribution associated with pi-cation
π
computational efficiency, quality experimental validation, interaction, and ϑ Halogen is the energy contribution associated
reliable structure-activity relationship, consistency, and with halogen bond interactions.
simplifying analysis. Similarly, refined PI3K-α ligand In this study, a receptor grid was initially generated
molecules obtained from the binding database (bindingdb. around the region occupied by P5J in the 6PYS protein,
org) were refined for docking using a similar 6PYS protein- aiming to map the properties of the binding site onto a grid.
ligand complex LigPrep approach. Next, the refined ligands were docked into the minimized
6PYS protein through Schrödinger’s Virtual Screening
The LigPrep settings for refining the inhibitory
molecules excluded tautomer generation. However, Workflow panel, with customized settings that automated a
virtual screening-molecular docking workflow. The virtual
stereoisomer computation was carried out to determine screening workflow was automated to screen the ligands
chirality based on the 3D structure with the objective of through successive stages, starting from high-throughput
having the internally produced stereoisomers filtered to virtual screening mode (HTVS) to standard precision
remove any structures, fused ring systems, or chirality (SP) mode, and lastly, through extra-precision (XP) mode.
that were incompatible with that of natural products to The submission ratio of screened ligands at each stage
generate the desired enantiomers. Ligand alignment was was set to 70%, 60%, and 8%, respectively, as depicted in
performed on all refined inhibitory molecules, including Figure 3. A trade-off between the speed and accuracy of
the P5J cocrystallized ligand. This option aligned structures the virtual screening served as the basis for selecting this
with similar orientations, facilitating the identification of ratio. By filtering out a large fraction of compounds in the
the ligand pose that maximizes beneficial interactions, early stages, the workflow could save time and resources,
such as hydrogen bonding, hydrophilic interactions, and considering that docking and post-processing are
electrostatic interactions, while minimizing detrimental computationally intensive and time-consuming compared
interactions or clashes. to ligand preparation. However, by retaining a sufficient

2.6. Ligand virtual screening-molecular docking
A structure-based, in silico virtual screening approach using
Schrödinger Maestro was applied to predict the interaction,
favorable binding orientations, and conformation of the
refined ligands within the active site of the target PI3K-α
protein. Essentially, for a molecule to tightly bind to a
receptor, both geometric (shape) and electrostatic (charge)
complementarities must exist. 47,48 These complimentary
aspects define the molecular dynamics of the ligand-receptor
relationship by incorporating interaction maximization
while minimizing the total energy of the complex.
Typically, most in silico docking programs are built to
predict binding mode and binding affinity between protein
and ligand using a hybrid search algorithm and scoring
function. While the search algorithm robustly generates
multiple poses for a ligand in the binding site of the receptor, Figure 3. Schematics of submission ratio for virtual screening workflow
for docking in Glide.
the scoring function ranks or orders the conformations to Abbreviations: HTVS: High-throughput virtual screening; SP: Standard
distinguish the experimental binding pose from the rest of precision; XP: Extra-precision.

Volume 7 Issue 2 (2024) 7 doi: 10.36922/itps.2340