Pushing the boundaries: towards single-residue resolution hydrogen/deuterium exchange mass spectrometry

Hydrogen/deuterium exchange (HDX) is a spontaneous process observed in proteins exposed to deuterated solvent. The observed rate of exchange of a residue is the ratio between its ‘intrinsic exchange rate’, reflecting the chemical properties of its environment, and the ‘protection factor’, accounting for structural and dynamic properties of the protein. HDX kinetics provide useful insights into protein conformational dynamics, complementing static structures from X-ray crystallography, NMR or cryo-EM.
HDX is primarily studied using NMR and mass spectrometry (MS). HDX-NMR offers single-residue resolution but is limited to smaller proteins, while HDX-MS can study larger systems, though it captures deuteration at peptide-level resolution. While spatial resolution in HDX-MS can be enhanced (e.g. with alternative fragmentation techniques), achieving NMR-like detail remains challenging. As a result, HDX-MS is typically used for differential studies, comparing protein states to qualitatively localize perturbations (e.g. binding sites), rather than quantify absolute biophysical properties such as protection factors.
This thesis explores whether single-residue resolution can be inferred from peptide-level HDX-MS data, despite the inherent underdetermination of the data.
We critically reviewed the strengths and limitations of various strategies used by software tools for analysing HDX-MS data. Building on this, we demonstrated the advantages of a method that exploits self-consistency to identify those alternative protection factor sets which are compatible with experimental data, cluster them into a finite number of solutions, and reduce the degeneracy by incorporating the information encoded in isotopic envelopes. Since our method assumed intrinsic exchange rates as known constants (under fixed environmental conditions), we tested their accuracy by performing millisecond HDX-MS experiments on a mixture of unstructured peptides. Finally, we proposed and tested against NMR data a protection factor analysis that addresses the ambiguity of peptide-level data by either performing a random search to broadly explore the rugged cost function landscape or incorporating an informed initial guess to focus on physically meaningful solutions.
Overall, our findings show that protection factors can be inferred from peptide-level HDX-MS data with sufficient redundancy and temporal sampling, providing insights comparable to NMR. These results push the boundaries of HDX-MS toward becoming an absolute quantitative experiment rather than a differential qualitative tool.