Molecular dynamics simulation of the crystallizable fragment of IgG1-insights for the design of Fcabs

Abstract

An interesting format in the development of therapeutic monoclonal antibodies uses the crystallizable fragment of IgG1 as starting scaffold. Engineering of its structural loops allows generation of an antigen binding site. However, this might impair the molecule's conformational stability, which can be overcome by introducing stabilizing point mutations in the CH3 domains. These point mutations often affect the stability and unfolding behavior of both the CH2 and CH3 domains. In order to understand this cross-talk, molecular dynamics simulations of the domains of the Fc fragment of human IgG1 are reported. The structure of human IgG1-Fc obtained from X-ray crystallography is used as a starting point for simulations of the wild-type protein at two different pH values. The stabilizing effect of a single point mutation in the CH3 domain as well as the impact of the hinge region and the glycan tree structure connected to the CH2 domains is investigated. Regions of high local flexibility were identified as potential sites for engineering antigen binding sites. Obtained data are discussed with respect to the available X-ray structure of IgG1-Fc, directed evolution approaches that screen for stability and use of the scaffold IgG1-Fc in the design of antigen binding Fc proteins.

Figure 1.

Graphical representation of the crystallizable fragment (Fc) of human IgG1 consisting of two N-glycosylated CH2 domains and two CH3 domains (individual domains are shown in different colours). Coordinates were taken from the crystal structure 1OQO. The glycosylation is shown in ball and stick representation and schematically in the top-right panel. Y407 located at the interface of the two CH3 domains forms a favorable stacking interaction. The position of the point mutation Q347 and adjacent residues affected by the exchange of glutamine by glutamate (Q347E) are shown in stick representation. Glycan numbering reflects the order in which they are attached to the topologies used for simulations (GlcNac: 1, 2, 5, 8; Mannose: 3, 4, 7; Fucose: 6). The hinge region and disulphide bonds connecting the two CH2 domains are shown as dotted lines.

Figure 2.

Atom-positional root-mean-square fluctuations (RMSF) for backbone atoms in various domains and glycans in simulations B4 (black), B7 (red), H4 (green), H7 (blue) and Q347E (magenta). CH2 and CH3 domain secondary structures are shown: helices (blue circle), 3/10 helices (green diamonds), β-sheets (red triangle) and coils (black line). C-terminal loop regions are indicated by corresponding letters and a solid black bar, which specifies the name and range of the loop.

Figure 3.

Relative orientation of individual domains as defined through the (dihedral) angles (in degree) as a function of time for simulation H7. (A) angle between ↛_CH2A, ↛_CH3A; (B) angle between ↛_CH2B, ↛_CH3B; (C) angle between ↛_CH2A, ↛_CH2B; (D) angle between ↛_CH3A, ↛_CH3B; (E) dihedral angle defined by ↛_CH2A, ⇉_CH3A, – ⇉_CH2A, ↛_CH3A; (F) dihedral angle defined by ↛_CH2B, ⇉_CH3B, – ⇉_CH2B, ↛_CH3B; (G) dihedral angle defined by ↛_CH2A, ⇉_CH2B, – ⇉_CH2A, ↛_CH2B; and (H) dihedral angle defined by ↛_CH3A, ⇉_CH3B, – ⇉_CH3A, ↛_CH3B. See Experimental Section and Supplementary Information for details concerning the definition of the vectors.

Figure 4.

Top panel: Average number of hydrogen bonds (excluding backbone-backbone interactions) per residue of the CH3 domain, as observed in simulation H7. The green fraction of the bars indicate hydrogen bonds between chains A and B. Bottom panel: Average surface area per residue of the CH3 domain, as observed in simulation H7. Orange bars indicate solvent accessible surface area; blue bars indicate interfacial area between chains A and B.