• EbolaGP
• References
• Monogram
• Slides

Ebola glycoprotein (PDB ID: 5JQ3) from Zaire ebolavirus

Created by: Natalie Kessler

    Ebola virus (EBOV) belongs to the family Filoviridae, members of which cause severe hemorrhagic fever in humans and other primates (1). EBOV and other Filoviruses have extremely high morbidity and mortality rates. Both EBOV and the Angola strain of Marburg virus, the filoviruses with the most significant mortality, have case fatality rates approaching almost 90% (1). High morbidity and mortality rates, significant infectivity and ease of transmission, and their weaponization during the Cold War has led to Filoviruses being deemed global health and biosecurity risks (1). Despite its classification as a Category A priority pathogen, no vaccine or anti-viral therapy has been successfully developed and approved at this time (1). The most recent outbreak in West Africa with 28,000 people infected and 11,000 deaths has demonstrated the urgent need for anti-viral therapies against EBOV (2). The EBOV glycoprotein (GP) (PDB ID: 5JQ3), located on the surface of EBOV, is exclusively responsible for initial attachment to the host cell, endosomal entry, and membrane fusion leading to the release of the viral genome (2, 3). EBOV GP’s essential role in infection makes the EBOV GP a clear target for proposed antiviral therapies (2). Understanding exactly how EBOV GP initiates infection could help lead to the development of therapies against EBOV disease.

EBOV GP exists as a homotrimer on the envelope of EBOV (1). EBOV uses this trimer to bind to the target cell membrane, and is subsequently internalized by a macropinocytosis-like mechanism and trafficked to the late endosome (1). EBOV GP’s molecular weight is 54.6 kDa without including the mass of attached sugar ligands, and its isoelectric point is 6.21, so it exists in an overall acidic form in the human body (4). During biosynthesis in the Golgi body, EBOV GP is cleaved into two subunits, GP1 and GP2 (1). The two subunits remain bound through non-covalent interactions and a disulfide bond (1). GP1 is deemed the attachment or receptor binding subunit, while GP2 is called the fusion subunit (2). Further cleavage of GP1 within the late endosome is required for virions to successfully bind their endosomal receptors (including Niemann-Pick C1 (NPC1) (PDB ID: 5I31) and two-pore channels (TPCs), like TPC1 (PDB ID: 5DQQ)) and infect target cells (1, 3, 5). GP1 is cleaved by cathepsin L into a 20kD form and then by cathepsin B into a 19kD form. These cleavage events expose the receptor binding domain of GP1 by eliminating the mucin-like domain, glycan cap, and outermost β-strand of the GP1 head group (1, 6). Additionally, cathepsins cut the β13-β14 loop that likely confines GP2’s fusion loop in the unprimed structure (1). The exposed fusion loop mediates fusion of the endosomal and viral membranes, resulting in release of the viral genome into the target cell’s cytoplasm where it can replicate. GP2 is completely protected during this priming (6).

EBOV GP contains α-helices, random coil, and β-strands. GP1 is mostly composed of β-strands that secure an α-helix and β-strands of GP2 (2). β-strands play an important role in maintaining the stability of the protein and preventing premature release of the fusion loop, increasing the probability of infection (1). GP2 includes three heptad repeat regions that form α-helices, with a Cys-X6-Cys-Cys motif (2). Cys-601 and Cys-608 form an intrasubunit disulfide bond, while Cys-53 and Cys-609 form intersubunit disulfide bonds that keep GP1 and GP2 the fusion loop from springing prematurely (2). Additionally, glycerol and several sugar ligands (N-acetyl-D-glucosamine, β-D-mannose, and α-D-mannose) are attached to native EBOV GP, especially in the mucin-like domain and glycan cap. The exact function of these sugars is unknown, but it is believed that the oligosaccharide coating of GP helps to stabilize its structure in certain environments and protect the virus from humoral immune response (7). The mucin-like domain may promote initial viral attachment, but is not required for entry. In fact, removal of the mucin-like domain improves EBOV binding and therefore its ability to infect cells (6).

    Four lysine residues (Lys-114, Lys-115, Lys-140, and Lys-95) in the binding pocket (or chalice) of EBOV GP are critical for interaction with host cell receptors (6). Lys-114, Lys-115, and Lys-140 all lie in the receptor binding region (RBR) of GP1 and have direct roles in binding GP1 to NPC1 due to their polar, basic nature (1, 6). Lys-95 lies deeper in the binding chalice outside of the RBR, so it is likely that it has an indirect role in maintaining a favorable conformation of GP1 for binding (6). Substitution of these four residues does not result in the complete inhibition of infection, suggesting that other residues also play an important role in binding (6, 8). A group of hydrophobic residues between Gly-524 and Ala-540 is part of the GP2 fusion loop, which is indispensable for infection (9). This collection of hydrophobic residues allows the fusion loop to be inserted into the nonpolar endosomal membrane and force fusion with the viral membrane, which releases the viral genome for replication (1, 9).

    Although neither a vaccine nor a therapeutic drug against EBOV have been discovered yet, some repurposed drugs have had moderate success in inhibiting EBOV infection. Toremifene and ibuprofen both inhibit EBOV pseudovirus infection, with Toremifene being a more potent inhibitor (2). Both ibuprofen (PDB ID: 5JQB) and toremifene (PDB ID: 5JQ7) bind in a pocket between GP1 and GP2 (2, 9). Tyr-517 of EBOV GP interacts with the all three phenyl groups of toremifene, while other nonpolar residues (for example, Val-66, Leu-68, Leu-515, and Leu-558 of EBOV GP with phenyl ring A of toremifene) simultaneously bind the phenyl groups of toremifene via hydrophobic interactions (2). Ibuprofen binds EBOV GP in roughly the same place, but through a different variety of interactions (2). Ibuprofen’s phenyl ring interacts predominantly with Met-548, another non-polar residue (2). Additionally, ibuprofen’s propanoic acid group hydrogen bonds with Arg-64’s guanidino group (2). Zhao, et al, showed through a thermal-shift assay that toremifene and ibuprofen binding destabilize the GP1-GP2 heterodimer, causing the release of GP2 prior to endosomal binding and therefore leads to the deactivation of the virus (2). Destabilization caused by these drugs is more pronounced at high concentrations and low pH values (2). This renders the virus incapable of fusion with the late endosome and unable to replicate (2). More potent drugs can be developed to target this pocket using the structural knowledge gained from experiments with toremifene and ibuprofen (2).

    PSI-BLAST is a database that compares and evaluates the similarity of different proteins’ primary structure in terms of a parameter called the E value (10). E values below 0.05 indicate a high degree of sequence similarity between two proteins, with decreasing E values signifying higher degrees of likeness. Inputting the sequence of a specific protein will return a list of other proteins that have the most similar sequence. A PSI-BLAST search of EBOV GP yielded several GPs from other viruses, the most notable of which was the Marburg virus GP (PDB ID: 3X2D) (3, 10).

    The Dali Server yields a different but complementary set of information to PSI-BLAST; the purpose of a Dali Server search is to find proteins with similar tertiary structure to the protein of interest (11). Dali Server calculates the difference in molecular distances by using a “sum of pairs” method and returns a list of proteins with comparable tertiary structures that are evaluated by a Z value parameter (10). A Z value above 2 indicates that the protein has similar folds to the protein of interest (10). A Dali Server search for EBOV GP returned only EBOV GP complexed with different ligands, so the search did not yield any significant results. It can be inferred that the tertiary structure of EBOV GP is relatively unique compared to other similar glycoproteins and other proteins. Since Dali Server failed to produce a unique protein, Marburg GP will continue to be used as a comparison to EBOV GP. Both are filovirus GPs, and hence are evolutionarily related and function in the attachment and entry of their respective filoviruses into the target cell (12). Both virus GPs have two subunits (GP1 and GP2) that are attached via a disulfide linkage, a mucin-like domain, and two heptad repeat domains and a transmembrane domain in GP2 (12). Biologically, EBOV GP and Marburg GP have analogous functions in mediating viral entry into host cells (12). Additionally, not only do EBOV and Marburg GPs function to accomplish the same result of viral entry, but they also use strikingly similar entry mechanisms to achieve this objective (12). Marburg virus and EBOV both enter via macropinocytosis and are introduced to the endocytic pathway where their GPs are cleaved by endosomal cathepsins (12). The entry mechanisms of EBOV and Marburg virus differ in their requirements of cathepsin B; while EBOV GP must be cleaved by cathepsin B in order to achieve infection, Marburg GP does not share this requirement (12). The structural basis for this difference is unknown (12). Both EBOV GP and Marburg GP interact with the critical domain of NPC1 through a highly conserved hydrophobic trough on GP1 (12). The side chains of Phe-72 in Marburg and Phe-88 in EBOV both form sides of the hydrophobic trough (12). The importance of this trough is highlighted by the fact that it is 85% conserved among all Filoviruses, even though as little as 30% of the rest of the protein sequence is conserved (12).