UC Riverside

Protein-protein interactions (PPIs) represent critical therapeutic targets in cancer and neurodegeneration, yet their large, flat, and dynamic interfaces pose fundamental challenges for drug discovery. This dissertation develops biophysics-guided inhibitors for two therapeutically relevant PPI systems: EphA4-ephrin, which is implicated in amyotrophic lateral sclerosis, Alzheimer's disease, and cancer, and KRAS-effector, which drives oncogenesis in pancreatic, colorectal, and lung cancers. An integrated pipeline combining nuclear magnetic resonance (NMR) spectroscopy, isothermal titration calorimetry (ITC), denaturation thermal shift assays (TSA), dissociation-enhanced lanthanide fluorescence immunoassay (DELFIA), sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), mass spectrometry (MS), and X-ray crystallography (XRC) was employed to drive rational inhibitor design and validate binding mechanisms. To target the EphA4-ephrin PPI, starting from the disulfide-stabilized β-hairpin peptide APY-d3 (Kd = 138 ± 8 nM), structure-activity relationship studies replaced the flexible β-Ala8 residue with 4-amino-5-methylthiophene-3-carboxylic acid, yielding disulfide-free peptides that adopt stable β-hairpin conformations in solution. Lead compound 22 (151G10) exhibited near-zero binding entropy by ITC, indicating conformational pre-organization, while maintaining nanomolar affinity (Kd = 137 ± 2 nM). The 2.5 Å crystal structure (PDB 9CY8) confirmed the thiophene moiety occupies the β-turn position, preserving the binding mode of APY-d3. Compound 23 (151H3), incorporating a biphenyl substituent at position 6, induced larger conformational changes in the EphA4 JK loop as evidenced by ¹³Cε-Met NMR chemical shift perturbations. Both compounds demonstrated EphA4-selective agonism, inducing receptor phosphorylation in primary cortical neurons, with compound 23 showing significantly greater activity than compound 22. Selectivity profiling confirmed minimal binding to EphA2 and weak binding to EphA3. This work establishes thiophene-based constraints as a novel β-hairpin stabilizing strategy, expanding the toolkit for conformationally restricted peptide design. To target the KRAS-effector PPI, cyclic peptides derived from KRpep-2 were engineered to position aryl-fluorosulfate warheads for covalent modification of His95 in the KRAS switch-II pocket. The lead compound 26 demonstrated pan-KRAS engagement across G12C, G12D, and G12V mutants, confirmed by gel shift assays, mass spectrometry (+1973 Da adduct), and time-dependent DELFIA IC50 improvements (159.0 ± 30.3 nM to 12.1 ± 0.3 nM for G12D after 18-hour incubation). Thermal shift assays revealed dramatic stabilization (ΔTm > 20 °C) for covalent adducts versus modest shifts (ΔTm ~4–6 °C) for reversible binders. The crystal structure of compound 26 bound to KRAS G12V displayed continuous electron density from the aryl-fluorosulfate to the His95 Nɛ̝2 atom, providing definitive structural evidence of covalent bond formation. ITC measurements confirmed that fluorosulfate incorporation did not disrupt reversible binding affinity (Kd values: 314 ± 163 nM for G12C, 100 ± 16 nM for G12D, 294 ± 69 nM for G12V). The partial covalent modification observed for wild-type, Q61H, and Q61K mutants is consistent with the reported selectivity of KRpep-2 for oncogenic KRAS variants. This strategy demonstrates that His95 covalent targeting enables pan-KRAS inhibition beyond G12C-specific drugs, addressing a critical gap in KRAS-targeted therapeutics. This dissertation establishes that conformational pre-organization reduces entropic penalties in PPI inhibitors, which was effectively quantified by ITC entropy measurements approaching zero for constrained peptides. Site-specific covalent modification of histidine residues using aryl-fluorosulfates provides a generalizable strategy for expanding the druggable proteome beyond cysteine-targeting approaches. The orthogonal biophysical validation pipeline combining NMR spectroscopy for conformational analysis, ITC for thermodynamic profiling, DELFIA for time-resolved potency assessment, and X-ray crystallography for structural confirmation, provides a robust framework for PPI inhibitor development. Lead compounds identified in both chapters 2 and 3 exhibit favorable inhibitory and binding properties, supporting their potential as chemical probes and therapeutic starting points. The thiophene-based β-hairpin constraint and His95-covalent targeting strategies are broadly applicable to other PPI systems, demonstrating how biophysics-guided design can transform "undruggable" targets into tractable therapeutic opportunities.