Complementary Structural Information for Antibody-Antigen Complexes from Hydrogen-Deuterium Exchange and Covalent Labeling Mass Spectrometry

Abstract

Characterizing antibody-antigen interactions is necessary for properly developing therapeutic antibodies, understanding their mechanisms of action, and patenting new drug molecules. Here, we demonstrate that hydrogen-deuterium exchange (HDX) mass spectrometry (MS) measurements together with diethylpyrocarbonate (DEPC) covalent labeling (CL) MS measurements provide higher order structural information about antibody-antigen interactions that is not available from either technique alone. Using the well-characterized model system of tumor necrosis factor α (TNFα) in complex with three different monoclonal antibodies (mAbs), we show that two techniques offer a more complete overall picture of TNFα's structural changes upon binding different mAbs, sometimes providing synergistic information about binding sites and changes in protein dynamics upon binding. Labeling decreases in CL generally occur near the TNFα epitope, whereas decreases in HDX can span the entire protein due to substantial stabilization that occurs when mAbs bind TNFα. Considering both data sets together clarifies the TNFα regions that undergo a decrease in solvent exposure due to mAb binding and that undergo a change in dynamics due to mAb binding. Moreover, the single-residue level resolution of DEPC-CL/MS can clarify HDX/MS data for long peptides. We feel that the two techniques should be used together when studying the mAb-antigen interactions because of the complementary information they provide.

Figure 1:

HDX/MS and DEPC-CL/MS results upon comparing the unbound and adalimumab-bound forms of TNFα. (A) Primary sequence of TNFα with peptides that decrease significantly in HDX indicated with blue horizontal lines. Residues involved in the epitope are indicated with magenta boxes. The vertical lines indicate residues that are labeled by DEPC but show no change in labeling extent (gray), increases in labeling extent (red), and decreases in labeling extent (blue). The β-strands of TNFα are indicated with arrows above the sequence. The sequence coverage during the HDX/MS experiments was 96% (Figure S1A), and the sequence coverage for the DEPC-CL/MS experiments was 100%. (B) Representative HDX/MS kinetic plots for select regions of TNFα. Residues in magenta are epitope residues. (C) DEPC-CL/MS results for residues that undergo decreased labeling upon binding to adalimumab. (D) DEPC-CL/MS results for residues that increase in labeling extent upon adalimumab binding (E) DEPC-CL/MS results for residues that do not significantly change in labeling extent upon adalimumab binding, according to a t-test at a 95% confidence interval.

Figure 2:

HDX/MS and DEPC-CL/MS results upon comparing the unbound and infliximab-bound fonns of TNFα. (A) Primary sequence of TNFα with peptides that decrease significantly in HDX indicated with blue horizontal lines. Residues involved in the epitope are indicated with magenta boxes. The vertical lines indicate residues that are labeled by DEPC but show no change in labeling extent (gray), increases in labeling extent (red), and decreases in labeling extent (blue). The β-strands of TNFα are indicated with arrows above the sequence. The sequence coverage during the HDX/MS experiments was 92% (Figure S1B), and the sequence coverage for the CL/MS experiments was 95%. (B) Representative HDX/MS kinetic plots for select regions of TNFα. Residues in magenta are epitope residues. (C) CL/MS results for residues that undergo decreased labeling upon binding to infliximab. (D) CL/MS results for residues that increase in labeling extent upon infliximab binding (E) CL/MS results for residues that do not significantly change in labeling extent upon infliximab binding, according to a t-test at a 95% confidence interval.

Figure 3:

HDX/MS and DEPC-CL/MS results upon comparing the unbound and golimumab-bound fonns of TNFα. (A) Primary sequence of TNFα with peptides that decrease significantly in HDX indicated with blue horizontal lines. Residues involved in the epitope are indicated with magenta boxes. The vertical lines indicate residues that are labeled by DEPC but show no change in labeling extent (gray), increases in labeling extent (red), and decreases in labeling extent (blue). The β-strands of TNFα are indicated with arrows above the sequence. The sequence coverage during the HDX/MS experiments was 92% (Figure S1C), and the sequence coverage for the CL/MS experiments was 95%. (B) Representative HDX/MS kinetic plots for select regions of TNFα. Residues in magenta are epitope residues. (C) CL/MS results for residues that undergo decreased labeling upon binding to golimumab. (D) CL/MS results for residues that increase in labeling extent upon golimumab binding (E) CL/MS results for residues that do not significantly change in labeling extent upon golimumab binding, according to a t-test at a 95% confidence interval.

Figure 4:

Decreases in HDX and CL upon adalimumab binding mapped on the TNFα structure. (A) Protein regions that undergo decreased HDX are shown in cyan and span the majority of the protein. (B) Protein residues that undergo decreased CL are indicated in blue. (C) Protein residues that are considered part of the adalimumab epitope are show in magenta.