Photo-Cross-Linking Mass Spectrometry and Integrative Modeling Enables Rapid Screening of Antigen Interactions Involving Bacterial Transferrin Receptors.

Abstract

Structure-based approaches to the delineation of immunogens for vaccine development have a throughput requirement that is difficult to meet in practice with conventional methods of structure determination. Here we present a strategy for rapid and accurate structure generation in support of antigen engineering programs. The approach is developed around the modeling of interactions between host transferrin (Tf) and the bacterial vaccine target transferrin binding protein B (TbpB) from Gram-negative pathogens such as Neisseria meningitidis. Using an approach based solely on cross-linking mass spectrometry (XL-MS) data, monomeric structural models, and the Integrative Modeling Platform (IMP), we demonstrate that converged representations of the Tf:TbpB interactions can be returned that accurately reflect the binding interface and the relative orientation of the monomeric units, with the capacity to scale to the analysis of interactions from any number of additional strains. We show that a key element to accurate modeling involves the application of hetero-bifunctional cross-linkers incorporating fast-acting photoactivatable diazirines coupled with conventional amine-targeting N-hydroxysuccinimide esters, and we demonstrate that conventional homo-bifunctional reagents used in cross-linking kinetically trap dynamic states in the ensemble. Therefore, the application of both classes of cross-linker provides an opportunity to empirically detect protein dynamics during integrative structural modeling.