Computational Screening of AI-Derived Cyclotides as Putative VEGFR2 Binders for Wound-Site Angiogenesis
Summary: Published March 13, 2026 in Scientific Reports (Nature), this computational study from Üsküdar University (Istanbul, Turkey) and the University of Central Punjab / Rashid Latif Khan University (Lahore, Pakistan) screens a curated library of 25 cyclotides — ultra-stable, disulfide-rich cyclic peptides of plant origin — for their potential to modulate vascular endothelial growth factor receptor 2 (VEGFR2), the primary driver of angiogenesis, as a strategy to address the impaired blood vessel formation that underlies chronic and diabetic wound non-healing. Insufficient VEGFR2 activation is a well-established pathological feature of chronic wounds, and while recombinant VEGF and PDGF therapies (e.g., becaplermin) exist, they carry tumorigenic risks and limited efficacy profiles. Cyclotides — whose cystine knot core and head-to-tail cyclisation confer extraordinary resistance to heat, proteolysis, and chemical degradation — have been studied for antimicrobial, anticancer, and wound-healing properties, and their engineered scaffolds have previously been used to graft pro-angiogenic peptides. In this study, the 25 cyclotides were modelled using AlphaFold, and all were docked into the predicted VEGFR2 binding pocket using HADDOCK. Cycloviolacin O13 (from Viola odorata) yielded the best interaction score (HADDOCK score −84.7; ligand RMSD 0.8 nm). A 500-nanosecond molecular dynamics simulation confirmed complex stability (RMSD 0.25–0.45 nm, 200–260 persistent hydrogen bonds, compact radius of gyration). Dynamic cross-correlation analysis supported coordinated binding motions, and normal mode analysis indicated low deformation and high mechanical resilience. Immuno-informatics confirmed cycloviolacin O13 is non-antigenic, non-allergenic, and non-toxic, with no predicted adverse B- or T-cell immune responses. The authors explicitly note that computational docking cannot determine whether O13’s VEGFR2 binding would be agonistic, antagonistic, or functionally neutral — in vitro VEGFR2 phosphorylation and downstream signalling assays in endothelial cells are required before any therapeutic inference can be made.
Key Highlights:
- 25 plant cyclotides screened via AlphaFold modelling + HADDOCK protein-peptide docking; cycloviolacin O13 (Viola odorata) identified as best VEGFR2 binder (HADDOCK score −84.7; RMSD 0.8 nm)
- 500 ns molecular dynamics: stable complex (RMSD 0.25–0.45 nm), 200–260 persistent hydrogen bonds, compact radius of gyration — indicates robust structural persistence under simulation conditions
- Immuno-informatics: non-antigenic, non-allergenic, non-toxic; no adverse B- or T-cell responses predicted — supporting a low immunological risk profile as a lead candidate
- Cyclotide rationale: cystine knot core and head-to-tail cyclisation confer extreme protease resistance, thermal stability, and oral bioavailability potential — key advantages over linear peptide scaffolds
- VEGFR2 context: primary mediator of angiogenesis; its insufficient activation drives impaired wound healing in diabetes and chronic wound states; represents a validated therapeutic target in wound care
- Critical limitation: computational study only — whether O13 binding is agonistic, antagonistic, or neutral is unknown; VEGFR2 phosphorylation, ERK/AKT signalling, and endothelial tube formation assays are required before translational relevance can be established
Keywords: VEGFR2 wound angiogenesis, cyclotide wound healing, impaired angiogenesis diabetic wound, computational drug discovery wound, AlphaFold wound care research, plant peptide wound healing
Özlem Karaca Ocak, Nouman Ali