In-Silico Docking of Indian Spice Phytoconstituents against DPP-4 (PDB 1X70) with ADMET and Drug-Likeness Screening

India has over ten crore adults living with diabetes, and DPP-4 (dipeptidyl peptidase-4) inhibitors such as sitagliptin, vildagliptin and teneligliptin are now common second-line oral drugs. DPP-4 rapidly degrades the incretin hormones GLP-1 and GIP; blocking it prolongs their glucose-dependent insulin release. Many Indian kitchen spices (turmeric, black pepper, cinnamon, clove, fenugreek, ginger, kalonji) have traditional and experimental antidiabetic reports, which makes their constituents natural candidates for an in-silico screening study.

This B.Pharm project performs structure-based virtual screening of about 20 phytoconstituents from seven spices against human DPP-4, using the crystal structure PDB 1X70 (DPP-4 co-crystallised with sitagliptin). The docking protocol is validated by redocking sitagliptin and checking that the pose reproduces the crystal pose (RMSD ≤ 2 Å). Each compound is docked with AutoDock Vina; binding affinity, key interactions with the catalytic and S1/S2 pocket residues (such as Glu205, Glu206, Tyr662, Ser630) and pose quality are compared with sitagliptin.

Top-ranked compounds are then filtered for drug-likeness and ADMET using SwissADME (Lipinski, Veber, bioavailability radar, PAINS alerts) and pkCSM (absorption, CYP inhibition, hERG, hepatotoxicity predictions). The study requires no animals, no ethics approval and no wet lab, so any college can run it on ordinary computers. It fits PCI BP813PW Project Work linked to the elective BP807ET Computer Aided Drug Design.

Abstract

Twenty phytoconstituents from seven Indian culinary spices were screened in silico against human dipeptidyl peptidase-4 (PDB 1X70). After validating the docking protocol by redocking the co-crystallised sitagliptin, ligands were docked with AutoDock Vina and ranked by predicted binding affinity and interaction with key active-site residues. Top candidates were evaluated for drug-likeness and ADMET properties using SwissADME and pkCSM. The study identifies spice-derived scaffolds worth further in-vitro enzyme assay and illustrates a reproducible, low-cost CADD workflow.

Introduction

DPP-4 inhibitors improve glycaemic control with low hypoglycaemia risk and are widely prescribed in India, including low-cost generics such as teneligliptin. Natural products remain a major source of lead compounds, and spices are attractive because they are consumed daily and have long safety records at dietary levels. Computer-aided drug design (CADD) allows rapid, inexpensive prioritisation of compounds before laboratory work.

Literature gap

Many published docking studies of natural compounds against DPP-4 exist, but quality varies: some skip protocol validation, use unspecified grid boxes, or report scores without interaction analysis or ADMET filtering. Few focus specifically on a curated set of Indian kitchen-spice constituents with a fully documented, reproducible workflow suitable for undergraduate laboratories.

Proposed work

• Prepare the protein and ligands with documented settings.
• Validate docking by redocking sitagliptin.
• Dock 20 phytoconstituents and the standard; analyse interactions.
• Screen top hits for drug-likeness and ADMET.
• Recommend candidates for in-vitro DPP-4 inhibition testing.

Feasibility

• Technical: all software is free for academic use or web-based; a laptop with 8 GB RAM is sufficient.
• Economic: no reagents or animals; cost is limited to printing.
• Ethical: no human or animal subjects; no ethics committee approval needed.
• Schedule: fits within the Semester-8 project period with time for documentation.
• Reproducibility: every configuration file, PDB ID, PubChem CID and software version is recorded so another group can repeat the work exactly.

DPP-4 inhibitors are effective but are patented or recently off-patent synthetic molecules, and there is continuing interest in new scaffolds from natural sources. Indian spices contain many bioactive constituents with reported antidiabetic effects, but which of them could plausibly bind DPP-4, and whether they have acceptable drug-like and pharmacokinetic properties, is not established in a consistent, validated way. Laboratory screening of every compound is costly. A validated computational screen can narrow the list to the few compounds that justify an enzyme assay. The problem addressed is to prioritise spice-derived phytoconstituents as potential DPP-4 inhibitors using a validated molecular docking protocol and in-silico ADMET screening, and to identify a small set of candidates for future in-vitro assays.

The study is purely computational. It uses one DPP-4 crystal structure, rigid-receptor docking and predicted ADMET values. It does not include molecular dynamics, free-energy calculations or laboratory validation, which are listed as future work. Docking scores are predictions of binding, not proof of inhibition or antidiabetic effect, and the study makes no therapeutic claims about spices.

Research design

In-silico, structure-based virtual screening study with protocol validation and ADMET filtering.

Step 1 — Ligand selection (literature-based)

Inclusion: reported presence in the spice and some antidiabetic evidence in literature; structure available in PubChem. Standard: sitagliptin (co-crystallised ligand).

Step 2 — Protein preparation

Download PDB 1X70 from RCSB; keep one chain; remove water and heteroatoms except the reference ligand (saved separately); add polar hydrogens and Kollman charges in AutoDock Tools; save as PDBQT.

Step 3 — Ligand preparation

Download 3D SDF from PubChem (or draw in ChemDraw/MarvinSketch); minimise with Open Babel (MMFF94); assign Gasteiger charges and torsions in AutoDock Tools; save PDBQT.

Step 4 — Grid and validation

Grid box centred on the co-crystallised sitagliptin with margins covering the S1, S2 and catalytic regions; exhaustiveness 16. Redock sitagliptin; accept the protocol if the top pose RMSD from the crystal pose is ≤ 2.0 Å.

Step 5 — Docking and analysis

Dock each ligand (three independent runs, record best and mean affinity); visualise top poses in PyMOL or Discovery Studio; tabulate hydrogen bonds, salt bridges and π interactions with key residues (Glu205, Glu206, Arg125, Ser630, Tyr547, Tyr662, Tyr666 and the S1 pocket); compare with sitagliptin's interaction fingerprint.

Step 6 — ADMET and drug-likeness

SwissADME: Lipinski, Veber, Egan rules, TPSA, consensus log P, GI absorption, BBB permeation, PAINS and Brenk alerts, bioavailability score. pkCSM: Caco-2 permeability, CYP2D6/3A4 inhibition, hERG I/II inhibition, AMES toxicity, hepatotoxicity, oral rat LD50 prediction.

Step 7 — Ranking

Composite ranking using affinity (relative to sitagliptin), interaction similarity, and ADMET flags; Spearman correlation between affinity and molecular weight to check for size bias.

Timeline (14 weeks)