Impact of Glycosylation on the Local Backbone Flexibility of Well-Defined IgG1-Fc Glycoforms Using Hydrogen Exchange-Mass Spectrometry

Abstract

We have used hydrogen exchange-mass spectrometry to characterize local backbone flexibility of 4 well-defined IgG1-Fc glycoforms expressed and purified from Pichia pastoris, 2 of which were prepared using subsequent in vitro enzymatic treatments. Progressively decreasing the size of the N-linked N297 oligosaccharide from high mannose (Man8-Man12), to Man5, to GlcNAc, to nonglycosylated N297Q resulted in progressive increases in backbone flexibility. Comparison of these results with recently published physicochemical stability and Fcγ receptor binding data with the same set of glycoproteins provide improved insights into correlations between glycan structure and these pharmaceutical properties. Flexibility significantly increased upon glycan truncation in 2 potential aggregation-prone regions. In addition, a correlation was established between increased local backbone flexibility and increased deamidation at asparagine 315. Interestingly, the opposite trend was observed for oxidation of tryptophan 277 where faster oxidation correlated with decreased local backbone flexibility. Finally, a trend of increasing C'E glycopeptide loop flexibility with decreasing glycan size was observed that correlates with their FcγRIIIa receptor binding properties. These well-defined IgG1-Fc glycoforms serve as a useful model system to identify physicochemical stability and local backbone flexibility data sets potentially discriminating between various IgG glycoforms for potential applicability to future comparability or biosimilarity assessments.

Keywords: antibody; glycosylation; hydrogen exchange; mass spectrometry; stability.

Figure 1:

HX kinetics of representative peptides found in the four IgG1-Fc glycoforms. The extent of HX has been corrected for back-exchange using deuteration controls as defined by equation (1). The error bars represent sample standard deviation from three independent HX measurements.

Figure 2:

Effects of glycan truncation on the average HX kinetics of the IgG1-Fc for all peptides relative to the high mannose IgG1-Fc glycoform. A) Regions of interest are mapped onto a homology model of the IgG1-Fc. B) HX kinetics of Man5-Fc (blue), GlcNAc-Fc (green), and N297Q-Fc (red) with respect to HM-Fc. $\Delta \overline{HX}$ represents the average of the deuterium differences at all HX times, normalized for peptide length and corrected for back-exchange, as defined by equation (3). The dashed lines represent the threshold for statistical significance as detailed in the Supporting Information. The arrows indicate the regions of interest. The peptides are indexed along the horizontal axis in order from N- to C-terminus as indicated in Supporting Table S1. C) Regions with statistically significant $\Delta \overline{HX}$ are shown mapped onto a homology model. Strong deprotection $\left(\Delta \overline{HX}\ge 5\%\right)$ is shown as dark blue, moderate de-protection $\left(0\%<\Delta \overline{HX}<5\%\right)$ as cyan, and protection $\left(\Delta \overline{HX}<0\%\right)$ as pink compared to the high mannose glycoform. The homology model is based on PDB 1HZH.