Template-Derived Engineering of Peptide Inhibitors Targeting HSP90–CDC37 Interface: A Theoretical Framework for Cancer Therapeutics

Abstract

The HSP90–CDC37 complex plays a key role in stabilizing multiple oncogenic client proteins, thereby sustaining cancer cell survival. However, targeting these dynamic interfaces remains challenging due to the large, shallow binding surfaces, which limit the effectiveness of conventional small-molecule inhibitors. Peptide-based therapeutics offer a promising alternative owing to their ability to mimic native motifs. Two continuous interfacial segments from CDC37 were extracted to construct template peptides, which were used to generate peptide libraries. A multi-level hierarchical screening strategy was applied to evaluate 4050 and 843,750 sequences derived from Template-1 and Template-2, respectively, thereby prioritizing top candidates based on binding profiles and physicochemical suitability. The shortlisted peptides were further validated using the Boltz-2 deep learning framework to enhance confidence in the predicted peptide structures and protein–peptide complexes. Binding free-energy and molecular dynamics simulation analyses were subsequently performed to assess their interaction with HSP90. The lead peptide was then benchmarked against positive and negative controls to provide a relative assessment of its binding performance. Collectively, these findings identify Pep14 derived from Template-2 as a lead peptide candidate for disrupting the HSP90–CDC37 interaction. However, further experimental validation is required to confirm its therapeutic potential in cancer treatment.