Evaluation of structural changes in protein therapeutics by HX-MS

Abstract

Function of protein therapeutics is governed by its high order structure (HOS). Changes in HOS can provoke changes in protein structure, stability, safety, and protein activity among others. Thus, monitoring HOS of protein therapeutics and identification of conformational changes is essential for development, production, and usage of protein therapeutics. Hydrogen exchange-mass spectrometry (HX-MS) is a well-established technique to probe HOS of proteins and is considered an intermediate resolution technique capable to point structural changes at a peptide level. Here we described experiments to: (1) probe conformational changes in a scFv used as a protein switch, (2) recognize allosteric effects in an IgG1 during interaction with a commercially available protein A ligand and (3) identify structural changes in the NIST mAb after a change in the pH formulation buffer.We used HX-MS experiments to identify conformational changes in a single chain variable fragment used as a protein switch. Previous results showed that introducing a cavity forming mutation into the scFv (W-to-G) decreased the protein activity. Later, by the addition of indole to the solution the protein activity is partially restored. Our results support the model of chemical rescue of the structure, whereby introducing a cavity forming mutation, the activity of the protein gets reduced due to critical conformation changes and later by the addition of a small molecule, the structure and activity of the protein is restored. Statistical analysis of HDX data showed that WT and rescued scFv protein (after the addition of indole) have a very similar structure, with minor differences that could explain why the rescued protein does not have the same binding affinity as the wild-type. To identify allosteric effects in the NIST mAb during interaction with a commercially available protein A ligand (ProA), we used ProA ligand both free in solution and attached onto a resin. When in-solution protein A was used at a 2:1 ProA:mAb ratio, allosteric effects in the Fab region were identified. However, at lower concentrations (1:1 ratio) or when ProA resin was employed, only peptides right at the binding site showed strong protection against hydrogen exchange. With these experiments, we were able to rule out all reversible allosteric effects in the interaction between the NIST mAb and protein A attached to a resin. Finally, we used HX-MS to identify conformational changes in a model IgG1 mAb after changing the formulation pH. Results showed that decreasing the pH of the solution induce a general increase in flexibility across the whole mAb, and stronger effects were seen in the CH2 domain and parts of the CH2-CH3 interface of the model mAb, most probably due to introduction of protonated histidine residues in the CH2-CH3 interface that induce new and affect critical electrostatic interactions in the region. An increase in pH showed the contrary, strong rigidity in the CH2 domain close to the hinge region and in the CH3 domain close to the CH2-CH3 interface. Additionally, changes in the pH caused changes in flexibility in the Fab domain of the antibody, involving CDRH3 when the pH is decreased and CDRL2 when the pH is increased. These effects are in critical regions for monoclonal antibodies that could affect effector functions or antigen binding.