(a) Ribbon diagram of the GP1 subunit. The base (I) subdomain (green) contains four discontinuous sections (residues 33-69, 95-104, 158-167 and 176-189), which form two mixed β sheets with strands β3 and β13 shared between the two β sheets. The head (II) subdomain (purple) is composed of residues from four discontinuous segments, 70-94, 105-157, 168-175 and 214-226, and forms a four-stranded, mixed β sheet supported by an α helix and a smaller, two-stranded, antiparallel β sheet. The glycan cap (III) region (cyan) is composed of a continuous polypeptide chain, residues 227-310, and forms an α helix packed against a four-stranded, mixed β sheet. Intramolecular disulfide bridges between Cys108-Cys135 and Cys121-Cys147 are coloured red. (b) Ribbon diagram of the Ebola virus GP trimer. Each GP1 subdomain is coloured according to panel (a) and the three GP2 subunits are shown in grey. The GP1 base subdomain forms a clamp onto the GP2 internal fusion loop and HR1_A helix through interactions with hydrophobic residues (inset box). The GP1-GP2 disulfide bridge (Cys53-Cys609), CX₆CC motif and HR2 region are disordered in the structure and are marked and indicated by dashed lines. (c) The pre-fusion conformation of EBOV GP2. The EBOV GP2 internal fusion loop contains a disulfide bond at its base (red), antiparallel β strand (grey) and a hydrophobic fusion peptide. The EBOV GP2 HR1 region is segmented into four parts (HR1_A-HR1_D). Note that the HR2 and the CX₆CC motif are disordered in the pre-fusion conformation.

N-linked bi-antennary complex-type glycans (Gal₂Man₃GlcNAc₄) were modeled onto the GP1 glycan cap subdomain. Oligosaccharides are shown as yellow space-filling spheres and for figure clarity, only those glycans belonging to the purple monomer are labeled. Note that the glycans on N228 and N563 reside on the back of the purple monomer and are partially obscured. The glycans at N204 and N268 are found in regions that are poorly ordered in the structure and as a result, their tentative locations are shown as orange ovals. The C termini of the last ordered residues of GP1, to which mucin-like domains are linked, are marked with ’C’ (top of the chalice) and coloured spheres (beige, pink and purple) outline the predicted positions of the mucin-like domains attached in each of these regions. Surface residues previously identified to be critical for viral entry, recessed in the chalice bowl and RBS, are coloured green. Fab KZ52 (coloured grey) recognizes a non-glycosylated, predominantly GP2-containing epitope at the base of the chalice.

(a) Domain schematic of GP. Domains observed in the crystal structure are coloured and numbered according to the description in the text. White and hash-marked regions designate crystallographically disordered and construct-deleted regions, respectively. Abbreviations are as follows: SP, signal peptide; I, GP1 base; II, GP1 head; III, GP1 glycan cap; mucin, mucin-like domain; IFL, internal fusion loop; HR1, heptad repeat 1; HR2, heptad repeat 2; MPER, membrane-proximal external region; and TM, transmembrane domain. Red Y-shaped symbols designate the predicted N-linked glycosylation sites; those sites marked with an asterisk were mutated. The final model includes EBOV GP residues 33-189, 214-278, 299-310 and 502-599. No electron density is observed for residues 190-213, 311-312, 464-501 and 600-632. Weak or discontinuous electron density is seen in the loop containing the GP1-GP2 disulfide bridge (residues 49-56) and the outer regions of the GP1 glycan cap (residues 268-278 and 299-310); these regions are modeled as poly-alanine fragments. (b) Molecular surface of the GP trimer viewed on its side (left) and top (right), as viewed down the three-fold axis. Monomer A is coloured according to the scheme in panel (a), and monomers B and C of the trimer are shown in white and grey, respectively. Predicted N-linked oligosaccharides (N228, N238, N257 and N268) belonging to the glycan cap of monomer A are shown as grey ovals. The location of the N-linked glycan at N268 is tentative, as the sequence assignment in this region is ambiguous. (c) Molecular surface of the EBOV GP chalice and cradle. The three GP1 subunits that form the chalice are shown in various shades of grey and GP2 subunits forming the cradle are shown as ribbons, in various shades of orange, underneath the transparent molecular surface. The putative receptor-binding sites (RBS) are recessed in the inner bowl of the GP trimer.

KZ52 recognizes residues 505-514 (red) and 549-556 (orange) of GP2 and residues 42-43 (green) of GP1. The β17-β18 loop (residues 279-298) is disordered, but may interact with KZ52. One EBOV GP monomer is coloured and labeled according to Fig. 1b and Fab heavy and light chains are shown in black and light grey, respectively. Side-chain interactions at the GP-KZ52 interface are magnified in the inset box. The putative receptor-binding site (RBS) is outlined in a red circle on the EBOV GP ribbon diagram. Note that only selected residues from GP and KZ52 are shown for figure clarity and residue 42 is a threonine in the wild-type Zaire ebolavirus sequence.

Residues coloured in cyan, green, royal blue and red were previously identified by mutagenesis to be important for viral entry48-50. Residues coloured in cyan (D55, L57, L63 and R64) reside at the base of the chalice, near the GP1-GP2 disulfide bond and HR1_D, and are likely important for fusion-mediated conformational changes rather than receptor binding. Residues coloured in red (F159, F160, Y162 and I170) are primarily buried hydrophobic amino acids that help to maintain the structural stability of GP1. Residues coloured in royal blue (G87, F88, F153 and H154) are in proximity to the putative receptor-binding site (RBS) and pack against the hydrophobic residues from a neighbouring internal fusion loop (coloured orange). Mutations to these residues may affect viral entry by altering the structural integrity of the RBS and/or by affecting packing of the fusion loop. Residues coloured in green (K114, K115, K140, G143, P146 and C147) reside on or near the GP surface and may contribute to receptor binding. A molecular surface representation of a GP monomer, coloured and oriented according to the ribbon diagram, is presented in the inset box. Indicated in this view: cleavage at a site on the loop (shown as purple dots) between residues 190-213 would remove the entire glycan cap (coloured in purple) and the mucin-like domain (not shown), leaving GP2 and an ∼18 kDa fragment of GP1.