Ebola Virus Glycoprotein: From Trimeric Structure to the Precise Mechanism of Membrane Fusion

The glycoprotein (GP) of Ebola virus (Ebolavirus) is a key molecule in viral infection. This type I envelope glycoprotein not only mediates viral entry into host cells but also forms a "glycan shield" through complex glycosylation modifications to evade host immune surveillance.

Ebola virus GP exists as homotrimers on the viral envelope surface. Each monomer consists of two subunits connected by disulfide bonds:

- GP1 subunit: Responsible for receptor binding, containing three subdomains—base, head, and glycan cap

- GP2 subunit: Mediates membrane fusion, containing internal fusion loop (IFL) and multiple heptad repeat regions (HR)

The complete GP trimer exhibits a unique chalice-like morphology, with three GP1 subunits forming the bowl structure and three GP2 subunits surrounding the base in a cradle configuration. This precise conformation ensures GP remains in a stable pre-fusion state before invasion.

GP is a highly glycosylated protein, forming a protective "glycan shield":

- N-glycosylation: Up to 17 N-linked glycosylation sites

- O-glycosylation: Mucin-like domain (MLD) contains 16 unique O-linked glycosylation sites

- C-mannosylation: At least 3 C-linked mannosylation sites in GP1 subunit

Mass spectrometry reveals most N-glycosylation sites exhibit complex glycan structures, while conserved sites N257 and N563 are enriched with unprocessed glycans that may facilitate viral attachment through DC-SIGN/L-SIGN receptor binding. Extensive glycosylation coats GP surface with thick glycan layers, effectively shielding neutralizing antibody epitopes, explaining the rare production of effective neutralizing antibodies during natural infection.

Structural and mutational studies show the lysine cluster (Lys114, Lys115, Lys140) in GP1 base subdomain is crucial for viral attachment. Key differences in receptor affinity among viral strains are determined by four amino acid positions (79, 141, 142, 148):

- Position 141 has most significant impact, with SUDV's Ala141 showing better binding than EBOV's Val141

- Position 148 shows Pro148 superior to Ala148

- These differences determine pathogenicity and tissue tropism among strains

Ebola virus entry is a spatiotemporally precise cascade process, believed to occur via endocytosis:

(A) Initially, metastable pre-fusion Zaire ebolavirus GP may bind to cell-surface lectins or unidentified attachment factors (green ovals) through its mucin-like domain (gray spheres) or other GP sites.

(B) The virus is then endocytosed and transported to endosomes. Whether lectins remain bound depends on their properties. In endosomes, host cathepsins cleave GP, removing glycan cap and mucin-like domain to form a ~19 kDa GP1 core region connected to GP2 via disulfide bonds.

(C) Newly exposed surfaces may enhance binding of receptors transported from cell surfaces or alternative molecules in endosomes. Binding of these molecules or further cathepsin cleavage may trigger GP2 fusion subunit conformational changes.

(D) GP2 structural rearrangement promotes HR1 formation into a single 44-residue helix, with internal fusion loop (IFL) inserting into host endosomal membrane. After insertion, IFL forms 310 helices, constituting extended pre-hairpin intermediates.

(E) Based on influenza studies, multiple GP2 trimers may be required to complete membrane fusion.

(F) HR2 and membrane-proximal external region (MPER) swing from viral membrane toward host membrane and HR1. Initial HR2 folding toward HR1 distorts and brings viral and host bilayers into contact, forming hemifusion stalks.

(G) Hemifusion stalks expand into fusion pores, and when three HR2 helices pack into HR1 trimeric bundles, low-energy postfusion six-helix bundles (6HB) form.

As the primary target of neutralizing antibodies, GP is central to vaccine and drug development. However, the glycan shield significantly hinders exposure of effective neutralizing epitopes. Recent studies using optical tweezers reveal that even without NPC1 receptor, cleaved GP can strongly bind membranes via fusion loops, providing new insights for developing inhibitors that block early membrane interactions.

Understanding molecular details of GP-NPC1 interaction not only explains why certain strains (e.g., RESTV) are non-pathogenic to humans but also provides theoretical basis for predicting cross-species transmission risks and developing broad-spectrum inhibitors. Future research should further resolve GP's dynamic structural changes under different pH conditions and how glycosylation precisely regulates its function.

```