Methodology:
Predicted Complex Analysis: Generate AlphaFold2 models of PDH2-antibody interfaces to identify key binding residues (e.g., CDR regions) .
Mutational Validation: Introduce point mutations in predicted interaction sites (e.g., VH CDR2) and test binding affinity via surface plasmon resonance (SPR) .
Case Study: A humanized PDH2 antibody showed a 7x improvement in KD after AlphaFold2-guided framework optimization .
Integrated Approach:
Multi-Method Validation: Combine IF with cellular fractionation followed by Western blot (e.g., nuclear vs. mitochondrial fractions) .
Proteomic Overlay: Use proximity ligation assays (PLA) to confirm PDH2 interactions in situ .
Example: Jurkat cell studies showed 80% mitochondrial localization via IF but only 60% in fractionation assays, resolved by accounting for antibody cross-reactivity with mitochondrial matrix proteins .
Workflow:
Enrichment: Treat cells with biotinylated PDH2 antibody, perform streptavidin pull-down, and analyze via LC-MS/MS .
Bioinformatics: Compare enriched proteins against CRAPome (Contaminant Repository for Affinity Purification) to filter nonspecific binders .
Validation: Use competitive displacement assays with excess PDH2 peptide to confirm specificity (e.g., 90% reduction in off-target signals) .
Generative AI Models:
De Novo Design: Train models on HER2/trastuzumab structural data to generate PDH2-specific HCDR sequences with low-nM affinity .
Affinity Maturation: Combine RosettaAntibody with molecular dynamics simulations to optimize electrostatic complementarity .
Outcome: AI-designed antibodies achieved EC50 values of 5–10 nM in cell-based assays, comparable to trastuzumab benchmarks .
Decision Tree: