Prospecting for novel plant-derived molecules of Rauvolfia serpentina as inhibitors of Aldose Reductase, a potent drug target for diabetes and its complications

Abstract

Aldose Reductase (AR) is implicated in the development of secondary complications of diabetes, providing an interesting target for therapeutic intervention. Extracts of Rauvolfia serpentina, a medicinal plant endemic to the Himalayan mountain range, have been known to be effective in alleviating diabetes and its complications. In this study, we aim to prospect for novel plant-derived inhibitors from R. serpentina and to understand structural basis of their interactions. An extensive library of R. serpentina molecules was compiled and computationally screened for inhibitory action against AR. The stability of complexes, with docked leads, was verified using molecular dynamics simulations. Two structurally distinct plant-derived leads were identified as inhibitors: indobine and indobinine. Further, using these two leads as templates, 16 more leads were identified through ligand-based screening of their structural analogs, from a small molecules database. Thus, we obtained plant-derived indole alkaloids, and their structural analogs, as potential AR inhibitors from a manually curated dataset of R. serpentina molecules. Indole alkaloids reported herein, as a novel structural class unreported hitherto, may provide better insights for designing potential AR inhibitors with improved efficacy and fewer side effects.

Figure 1. The key role of Aldose Reductase in hyperglycemia-induced oxidative stress.

Under normal conditions, glucose is metabolized to release carbon dioxide along with energy. Under hyperglycemic conditions, AR converts glucose to sorbitol, utilizing cofactor NADPH and consequently reduces glutathione level. Further, sorbitol is converted to fructose by NAD⁺ -dependent sorbitol dehydrogenase, leading to production of reactive oxygen species. Intracellular accumulation of sorbitol creates a loss of osmotic integrity and cellular damage, while depletion of NADPH and NAD⁺ cofactors compromises body’s antioxidant defence systems. In addition, high blood levels of fructose may account for increased glycation. These changes result in osmotic and oxidative stresses, ultimately leading to various micro-vascular complications in a number of tissues. Polyol pathway, thus, is involved in various biochemical changes that are relevant to diabetes-induced vascular dysfunction. AR controls the rate-limiting step of polyol pathway, and inhibition of AR provides a possible strategy to prevent complications of chronic diabetes.

Figure 2. Strategy implemented towards prospecting for novel ARIs from R. serpentina.

R. serpentina extracts are reported to be effective against diabetes and its complications. AR controls the rate-limiting step of polyol pathway, and its inhibition is known to prevent complications of diabetes. Founded in these empirical facts, we propose a hypothesis connecting effectiveness of molecular constituents of plant extracts to a regulatory mechanism central to the disorder. Towards our aim of prospecting for novel ARIs, we compiled a structured library of R. serpentina PDMs, and screened them to obtain ‘best PDMs’ (3). The best PDMs were refined to obtain two ‘PDM leads’ on the basis of their structural stability. Further, 16 more ‘ZINC leads’ were identified by screening structural analogs of these plant-derived leads, and representative analogs were assessed for their structural stability. This prospection study presents a repertoire of plant-derived indole alkaloids, and their analogs, as potential AR inhibitors.

Figure 3. Abundance of entries for R. serpentina plant-derived molecules from different plant parts.

Number of PDM entries reflecting the abundance of PDMs from different plant parts: stem, leaves, roots, bark, culture, and unspecified. The PDM entry was classified as ‘Unspecified’, when no specific plant part, from which it was extracted, was reported. The plant part class ‘Culture’ includes following sub-categories: hairy root culture, root culture, hybrid cell culture, cell culture, and cell suspension culture.

Figure 4. Phytochemical composition of R. serpentina plant-derived molecules.

Number of PDMs obtained for different structural classes of phytochemicals.

Figure 5. Validation of the docking protocol.

(A) ROC curve against AR DUD dataset. ROC statistics shows the success of docking protocol implemented in discriminating actives from decoys. AUC of 0.74 was obtained on the basis of binding affinity scores and interactions with critical residues. ROC curve depicts the true positive rate (sensitivity) versus false positive rate (1-specificity). The graph was rendered using ROCR package. (B) Comparison of experimental and computationally predicted docked conformations of the ligand. Overlay of the experimental (orange) and predicted docked conformation (gray) of IDD594 ligand in the binding site of the receptor (AR; PDB ID: 1US0) with RMSD of 0.094 Å. The figure was rendered using PyMol software.

Figure 6. Three best PDMs identified using molecular docking.

2D structures of 3 best PDMs of R. serpentina identified on the basis of binding affinity and interactions with critical residues: (A) RASE0048, (B) RASE0049, and (C) RASE0143.

Figure 7. Details of inhibitory interactions made by best PDMs.

(A) Three best PDMs, RASE0048 (red), RASE0049 (green), and RASE0143 (blue), docked in the binding site of AR, were visualized as cartoons displaying the catalytic center. 2D interaction plots of docked molecules into the binding site: (B) RASE0048, (C) RASE0049, and (D) RASE0143. Dotted green lines represent hydrogen bonds with constraints, while red spoked arcs represent residues making hydrophobic contacts with ligand. Red circles and ellipses indicate protein residues that are in equivalent 3D positions.

Figure 8. Stability evaluation of docked complexes using RMSD.

RMSD profiles of Cα backbone atoms with respect to the starting conformation, as a function of time: (A) Best PDMs and (B) Representative molecules from analogs of PDM leads.

Figure 9. Stability evaluation of docked complexes using RMSF.

Local conformational changes in structure as indicated by RMSF of individual residues: (A) Best PDMs and (B) A representative molecule from analogs of PDM leads.

Figure 10. Stability evaluation of docked complexes using hydrogen bonding pattern.

Time series plot of number of intermolecular hydrogen bonds: (A) Best PDMs and (B) A representative molecule from analogs of PDM leads.

Figure 11. Conformations of PDM leads and ZINC leads, before and after MD simulations.

Comparison of conformations of lead complexes before (blue) and after (orange) MD simulations. PDM leads: (A) RASE0048 and (C) RASE0049; ZINC leads: (B) ZINC04286771 and (D) ZINC49016166. AR-lead complexes were superimposed based on the Cα backbone atoms in the average structure obtained, over the initial docked structure.

Figure 12. ROC curve analysis for lead sets.

To assess the reliability of PDM leads and ZINC leads, ROC curve analysis was performed. For AR DUD actives, corresponding DUD decoys were used (stars; AUC: 0.74). For lead sets, decoys were obtained through DecoyFinder. The DUD actives and decoys were appended with lead sets and their corresponding DecoyFinder decoys, independently (circles and squares) as well as together (triangles). When appended with both the lead sets, AUC improved to 0.85; whereas it improved to 0.76 and 0.84, when appended with PDM leads and ZINC leads, independently.