The eptinezumab:CGRP complex structure – the role of conformational changes in binding stabilization

Figures & data

Table 1. X-ray diffraction statistics

	Fab:CGRP	Fab
Wavelength (Å)	0.95372	0.97917
Space group	C2	P21221
Unit cell parameters (a, b, c in Å; α, β, γ in °)	a = 68.235, b = 207.104, c = 68.204; α = γ = 90.00, β = 91.08	a = 62.781, b = 83.485, c = 90.190; α = β = γ = 90.00
Resolution (Å)	48.53–3.10 (3.31–3.10)	50.00–1.54 (1.60–1.54)*
Measured reflections	66,115 (11,076)	423,952
Unique reflections	17,052 (2,996)	70,440 (6,946)
Rsym	17.4 (64.6)	0.054 (0.823)
Completeness (%)	99.2 (96.7)	99.8 (99.9)
Average I/σ	6.0 (1.7)	21.7 (1.4)
Redundancy	3.9 (3.7)	6.0 (5.7)
CC1/2	0.984 (0.665)	0.999 (0.639)
Fab molecules per asymmetric unit	2	1

*Highest resolution shell of 3.31–3.10 Å (Fab:CGRP) and 1.60–1.54 Å (Fab).

Abbreviations: CC_1/2, correlation coefficient determined by two random half-sets of data; CGRP, calcitonin gene-related peptide; Fab, antigen-binding fragment; Rsym, agreement between symmetry equivalent reflections.

Table 2. Model refinement statistics

	Fab:CGRP	Fab
Refinement resolution (Å)	48.57–3.10	45.14–1.54
Rcryst (%)	18.05	15.9
Rfree (%)	26.58	19.0
Protomer details (per asymmetric unit) Eptinezumab Fab CGRP	2 2	1 ‒
Bond lengths, rms (Å)	0.005	0.016
Bond angles, rms (°)	1.467	1.99
Ramachandran plot (%) Preferred Allowed Outliers	86.9 9.9 3.2	97.4 2.6 0.0

Abbreviations: CGRP, calcitonin gene-related peptide; Fab, antigen-binding fragment

Table 3. Main epitope/paratope interactions

Eptinezumab	Atom	Chain	CGRP	Atom
# H-bond
Ser92	OG	L3	Ser116	O	P
Tyr37	OH	L	Phe119	O	P
IIe53	N	H2	Asn113	O	P
Asn54	N	H2	Asn113	OD1	P
Gln47	NE2	L	Ala118	O	P
# Water-mediated H-bond (putative)
Ser92	OG	L3	Phe119Cter	N	P
Tyr93	O	L3	Ser116	OG	P
Cys100	O	L3	Ser116	OG	P
Asp51	O	L2	Lys117	O	P
# Main hydrophobic contacts
Cys95	SG	L3	Val114	CG2	P
Cys95	SG	L3	Phe109	CE1	P
Tyr58	CD2	H	Val114	CG1	P
Val50	CG1	H	Val114	CB	P
Val37	CG2	H	Phe119	CZ	P
Leu90	CD2	L3	Phe119	CG	P
Tyr33	CE1	L1	Lys117	CD	P
Tyr29	CE1	L1	Pro111	CG	P
Tyr33	CE2	L1	Pro111	CB	P
# Main pi-pi interactions
Asn54		H2	Asn113		P
Asn35		H	Phe119		P
Phe101		L2	Phe119		P
Tyr33		H1	Val114		P
Arg97		H3	Phe119		P
Tyr29		L1	Phe109		P
Asp94		L3	Phe109		P

Abbreviations: CGRP, calcitonin gene-related peptide; Cter, C-terminus.

Figure 1. Crystal structure of eptinezumab in its free (unbound) state, in which the heavy chain (cyan) and the light chain (green) are shown from the side and top views. The CDRs are colored using PyMOL color names, where light chain regions are in shades of green (L1 [pale green], L2 [lemon], L3 [lime]) and heavy chain regions are in shades of blue (H1 [light blue], H2 [marine], H3 [blue])

Figure 2. Crystal structure of the eptinezumab:CGRP complex. CGRP (yellow) is shown bound in the deep pocket formed by the light (green) and heavy (cyan) chains of the Fab. The CDRs are colored using PyMOL color names, where light chain regions are in shades of green (L1 [pale green], L2 [lemon], L3 [lime]) and heavy chain regions are in shades of blue (H1 [light blue], H2 [marine], H3 [blue])

Figure 3. Detailed interactions between CGRP and eptinezumab. Hydrogen bond in yellow, hydrophobic contact in pale green, and pi-pi interactions in raspberry. The CDRs are colored such that light chain regions are in shades of green (L1 [pale green], L2 [lemon], L3 [lime]) and heavy chain regions are in shades of blue (H1 [light blue], H2 [marine], H3 [blue]). All colors are PyMOL color names

Figure 4. Evidence for structural water molecule in eptinezumab:CGRP interface. CGRP is colored in yellow and CDR L2 in Limon, using PyMOL color names. The locations of water molecules from the X-ray crystallography structure of the uncomplexed eptinezumab (orange spheres), of water molecules obtained by Watermap from the complex structure (light blue spheres 1–3), and of explicit clusters of water molecules obtained by Watermap from the eptinezumab:CGRP complex (HOH clusters). Watermap places 2 water molecules (light blue spheres 1 and 2) in a cavity within the bound structure, forming a network of hydrogen bond between Asp51 NH (CDR L2) to water1, water1 to water2, water2 to Leu90 C = O and Ser92 OG (CDR L3), and finally Ser92 OGH to Ser116 C = O (CGRP); CGRP Ser116 OGH does not make any direct interaction with eptinezumab but faces another small cavity. Watermap placed one water molecule (light blue sphere 3) in this cavity, indicating a hydrogen bond network with O from water3 and water3 to Cys100 and Tyr93 (CDR L3) carbonyls

Figure 5. Comparison of bound and unbound structures indicate conformational changes during eptinezumab binding to CGRP. Eptinezumab is in yellow. The CDRs are colored using PyMOL color names, where light-chain regions are in shades of green (L1 [pale green], L2 [lemon], L3 [lime]) and heavy-chain regions are in shades of blue (H1 [light blue], H2 [marine], H3 [blue]). Superposed structure using the FV domain as template a) of the unbound (translucent) and bound (opaque) structure. CDR H3 (blue) clearly goes through a structural reorganization where an internal H-bond between Asp99 and Arg97 in the bound form goes to an H-bond between Asp99 and light chain Lys46 inducing a movement of Asp99 CA by ~3 Å (from 2.5 Å to 3.5 Å following the alignment); b) of the unbound (translucent) and bound (opaque). Asn54 rotates and makes a direct contact with CGRP Asn113. Although the angles between both amides is not right for a proper hydrogen bond, the inter distance between N-Asn113 and O from Asn54 CDR H2 is 2.8 Å

Figure 6. Computational alanine scanning of CGRP identifies key interface residues. a) The CDRs are colored using PyMOL color names, where light-chain regions are in shades of green (L1 [pale green], L2 [lemon], L3 [lime]) and heavy-chain regions are in shades of blue (H1 [light blue], H2 [marine], H3 [blue]). CGRP is in gray. Four positions in CGRP are important for binding: Phe109 (red), Val114 (Orange), Gly115 (yellow), Phe119 (green). b) Table showing loss in binding energetics observed when these key positions mutated to the corresponding amino acid in related neuropeptides based on sequence alignment