Chacterization of Influenza Antibody-Antigen Complexes Using Hydrogen-Deuterium Exchange Mass Spectrometry and Computational Modeling

Abstract

Influenza viruses pose a major threat to public health due to their diversity and rapid evolution. Recently, a novel class of Abs was discovered that targets the trimeric interface (TI) of the HA head with remarkable breadth and affinity, making the HA TI a promising target for immunogen design. In this thesis, we investigated the interactions of protective H7-specific monoclonal Abs targeting the TI-2 site—the second major TI antigenic site in the HA head—to identify shared structural determinants essential for eliciting these broadly reactive Abs, ultimately informing rational influenza vaccine design. Using hydrogen-deuterium exchange mass spectrometry (HDX-MS) and computational modeling, our study characterized the binding of three TI-2 Abs (H7-214, H7-241, and H7-247). We identified key structural determinants critical for TI-2 site recognition: the N208–S216 β-strand and the G196–V202 loop. Our epitope mapping studies also suggested that the HA TI-2 antigenic landscape comprises multiple distinct yet overlapping epitopes rather than a single fixed binding site. This work advanced our understanding of TI regions as a valuable focus for influenza immunity and identified key structural features within the HA TI-2 site as targets for developing a broadly protective pan-H7 vaccine. Additionally, my project proposed a hybridized approach that combines computational docking with HDX-MS epitope mapping for structural prediction of Ab-Ag complexes, known as RosettaHDX. Sparse data from HDX-MS can restrict the conformational space to relevant structures, while computational docking provides atomic-level resolution models. By incorporating HDX data as both distance restraints and a scoring term in the RosettaDock algorithm, RosettaHDX successfully generated near-native models (interface root-mean-square deviation ≤ 4 Å) for all nine benchmark complexes, producing an average of 3.6 times more near-native models than Rosetta alone. To our knowledge, no other platform has benchmarked HDX-MS data in docking. Additionally, we are the first to devise a predictive metric based on docking results with HDX to identify allosteric peptides. With HDX-MS data acquisition taking as little as one week per epitope mapping experiment, our established method enables more reliable predictions of Ab-Ag complexes in relatively short timeframes. This method could be a valuable tool for structure-based therapeutic and vaccine development, particularly in high-pressure, time-sensitive situations such as pandemics or public health emergencies.