Ebola virus disease (EVD) is one of the most dangerous viral infections afflicting humans and nonhuman primates. Its first reported outbreak occurred in 1976 in Yambuku, a village in the Democratic Republic of the Congo (former Zaire), and Nzara, a town in South Sudan. That same year, the causative agent of the disease — Ebolavirus, a member of the Filoviridae family — was first isolated from an infected individual who lived in the Ebola river valley that gave its name to the virus [1]. So far, 6 ebolaviruses are known including Bundibugyo ebolavirus (BDBV), Zaire ebolavirus (ZEBOV), Reston ebolavirus (RESTV), Sudan ebolavirus (SUDV), Tai forest ebolavirus (TAFV), and Bombali ebolavirus (BOMV). ZEBOV, SUDV and BDBV are capable of infecting humans and therefore pose a serious threat [2–3]. Since 1976, the world has seen more than 20 outbreaks of EVD caused by ZEBOV, SUDV and BDBV. The largest outbreak that occurred in 2014–2016 in West Africa grew into an epidemic and killed over 12,000 people. The majority of all reported EVD cases have been attributed to ZEBOV (tab. 1) [4].
ZEBOV, SUDV and BDBV cause acute, highly contagious fever in humans and nonhuman primates. RESTV is not known to cause EVD in humans; however, antibodies against this species are detected in the blood serum of individuals who work with monkeys and apes infected with RESTV [4]. The reasons behind such different pathogenicity of RESTV and other pathogenic types of Ebola virus are still unknown.
The epidemic of 2014–2016 urged the researchers all around the globe to put increasing effort into developing a vaccine against EVD. To date, over 10 vaccines have been developed, of which 4 have already been approved for clinical use [5]. Both candidate and approved vaccines confer 100% protection against ZEBOV-associated EBV in nonhuman primates. Their efficacy against other ebolaviruses varies. Because SUDV and BDBV can also cause outbreaks and epidemics and because new ZEBOV strains are emerging, the world is faced with a pressing need for vaccine capable of protecting the human population against all known pathogenic ebolaviruses.

The structure of the virus

Ebolaviruses have a filamentous structure that comes in different shapes and length. The virions consist of an envelope, a nucleocapsid, a polymerase complex and a matrix [6] (fig. 1).
The nucleocapsid core of the virion contains a replication complex composed of a single-stranded RNA genome and a few proteins, including NP, VP35, VP30, and polymerase L. The virus has an outer lipid membrane with glycoprotein (GP) spikes on its surface. The protein matrix formed by proteins VP40 and VP24 lies immediately beneath the outer membrane [6].
The viral genome is represented by negative single-stranded RNA (fig. 2) [6] carrying 7 genes that code for a total of 9 proteins: the nucleoprotein (NP), the viral polymerase cofactor VP35, the major matrix protein VP40, 3 glycoproteins (the secreted sGP, the full-length GP and the small secreted ssGP), the minor nucleoprotein VP30, the membrane-associated protein VP24, and the viral polymerase L [6–7].
The GP glycoprotein of the Ebolavirus is the only protein located on the surface of the virion. It plays a key role in the early stages of infection helping the virion to attach to and enter the cell [7].

Synthesis and proteolytic processing of GP

The ebolavirus glycoprotein gene codes for 3 proteins: pre-sGP, pre-ssGP (they both are precursors of secreted nonstructural glycoproteins) and pre-GP (a precursor of the structural transmembrane glycoprotein). The nucleotide sequence of the glycoprotein gene contains 7 consecutive uracils at positions 880–886, where a hairpin loop is formed. It is difficult for the viral L polymerase to read through the hairpin [7–8]; therefore, this RNA region undergoes editing. As a result, 3 transcripts are produced:

– a transcript containing 7 uracils (~71%), coding for sGP (364 aa);

– a transcript containing 8 uracils (~25%), coding for GP (676 aa);

– a transcript containing 9 uracils (~4%), coding for ssGP (298 aa).

The first 295 amino acids bases in GP, sGP and ssGP are identical; however, the proteins differ in their C-terminus, which naturally affects their function. A newly synthesized pre-sGP is processed by cell proteases, leading to the formation of secreted sGP, that reduces the efficacy of humoral response by misdirecting antibodies and Δ-peptide responsible for pore formation in the cell membrane (fig. 2) [8–9].
Рre-GP is also cleaved by cell proteases into two subunits: GP1 and GP2. The subunits form heterodimers that are trimerized and constitute spikes on the surface of the viral particle. GP1 contains a receptor-binding domain, a glycan cap and a mucin-like domain required for the interaction with cell surface receptors. GP2 is a transmembrane domain, anchoring the complex in the membrane (fig. 2). GP2 has a binding site for the TACE protease; in proteolytic cleavage, the glycoprotein is cut off from the membrane and another type of GP is formed: the shed GP [8].
Mature surface GP exerts one of the most crucial functions in the lifecycle of the virus: it interacts with cell receptors, promoting fusion of the virion with the membrane. The virus is taken up into the endocytic/macropinocytic pathway; then, the mucin-like domain and the glycan cap of the glycoprotein are cut off by furin and cathepsins in the cell endosome. The truncated GP binds to Niemann-Pick C1 (NCP1) cholesterol transporter, initiating fusion of the endosomal and viral membranes and allowing the nucleocapsid to enter the cytoplasm [10–11].

Structural and immunogenic features of different GP forms

All secretory forms of GP (sGP, ssGP and shed GP) serve to protect the virus from being neutralized by the host’s natural defenses. Cells infected with Ebolavirus secrete these proteins thereby guiding the humoral response against the limited number of epitopes [12–13]. Produced in abundance, sGP, ssGP and shed GP misdirect the majority of IgG, reducing the efficacy of the host’s humoral response [14]. These glycoproteins (especially sGP and ssGP) trigger production of antibodies that have zero or weak virus-neutralizing potential, causing the phenomenon of antibody-dependent enhancement of the infection: the antibodies recognize the virus and interact with Fc-receptors of phagocytes, “ordering” the latter to take up the virus-antibody complexes via FcγR-mediated phagocytosis [15]. Importantly, although secreted forms of GP have binding sites for the protective antibodies, not all animals vaccinated with truncated forms of proteins develop protective immunity against the virus. A strong immune response against EVD can be achieved in all vaccinated animals only when a full-length GP is used [16–18]. This is probably due to the presence of additional neutralization sites and T-cell epitopes in the structure of the full-length protein. This hypothesis is supported by a few observations. Firstly, there are reports that apart from GP1 (glycan cap)- recognizing antibodies isolated from convalescent patients, those that specifically bind to the submembrane domain of GP2 also have protective potential [18–21]. Secondly, a study of CD8+-memory cells in convalescent patients has identified glycoprotein epitopes crucial for provoking a protective T-cell response, among which are regions of the receptor-binding domain and the glycan cap of GP [22].
Because a full-fledged protective immune response can be induced by using a full-length glycoprotein or structures expressing the gp gene, the majority of candidate and approved vaccines against ebolaviruses are based on the GP glycoprotein [23–24].

Analysis of cross-reactive immunity in vaccinated individuals and patients recovered from EVD

An ideal EVD vaccine must ensure protection against all variants of Ebolavirus that infect humans. Therefore, it is important to understand whether immune response can be induced against both homologous and phylogenetically distant species. The vaccine based on the recombinant vesicular stomatitis virus (rVSV-ZEBOV) that expresses glycoprotein GP of the ebolavirus isolated in 1995 has been reported to protect non-human primates against infection with any of known ZEBOV strains (isolated in 1976, 1995 and 2014) [25]. Studies of sera samples obtained from patients recovering from ZEBOV confirm those findings: IgG antibodies detected in the sera of the patients were capable of cross-reacting with glycoproteins of heterologous species SUDV and BDBV [26–27]. The studies of cross-protective immunity in non-human primates demonstrate that the use of ZEBOV glycoprotein (as a component of the rVSV vaccine) confers protection against the lethal BDBV infection in 100% of animals; in contrast, the rVSV-SUDV vaccine does not protect all animals against ZEBOV and BDBV [28–30].
Research into cross-protective immunity against EVD in animals has revealed that postvaccination immunity is cross-protective against ZEBOV and BDBV but not against these species and SUDV.

In search of explanation for this phenomenon, we compared the structure of GP in different species of the ebolavirus. We aligned amino acid sequences of ZEBOV, SUDV and BDBV and mapped immunodominant GP epitopes (i. e., those with the highest immunogenicity; IE) in 1,548 Ebola virus isolates; 10 BDBV, 23 SUDV and 1,515 ZEBOV sequences were taken from a public database [31].
The detailed analysis of immunodominant regions carried out in T Cell Epitope Prediction Tools in the deimmunization mode [32] allowed us to identify 22 IE (tab. 2 and tab. 3). The vastest diversity was observed for the mucin domain of GP1; the lowest, for GP2. Paired comparison of IE revealed that homology between ZEBOV and BDBV immunodominant glycoprotein epitopes was 75.8%; between ZEBOV and SUDV, 63.2%; and between SUDV and BDBV, 61.5% (tab. 2). It should be noted that glycoproteins representing different of ZEBOV isolates dating back to 1976, 1995, 2014 and 2018 are almost identical and only have minor differences in the region of the glycan cap and the mucin domain (tab. 3). On average, IE homology was 98.7–100%.

The obtained data suggest closer phylogenetic relationship between ZEBOV and BDBV, in comparison with SUDV, but do not explain the difference in their ability to induce immune response in animals immunized with the corresponding variants of GP. So, we decided to analyze the sites that bind antibodies conferring cross-protective immunity against the lethal infection caused by various ebolavirus species. Such antibodies are specific to both GP1 and the regions adjacent to the transmembrane domain of GP2 [19–21, 33]. A few recent works point to the fact that protective antibodies bind to the conformational epitopes of GP and not to the linear ones [34–36]. Our analysis of GP sequences has revealed that positions of key amino acids essential for antibody binding are quite conserved. The analysis of sites for binding antibodies with protective potential shows that the positions of key amino acids (i.e., those whose substitution fully blocks the ability of the antibodies to bind to GP) in ZEBOV epitopes are absolutely identical to the positions of amino acids BDBV GP epitopes. Homology between these amino acids and those found in SUDV glycoprotein varies from 30 to 60% (fig. 3). In our opinion, mutations at such amino acid positions inhibit the protective potential of the antibodies. It seems that homology of the sites that bind the protective antibodies to the glycoproteins representing different Ebolavirus species is the factor that ensures cross-protective immunity against ZEBOV and BDBV and the lack of cross-protective immune response against these two species and SUDV.
The discovery of universal antibodies capable of protecting humans against pathogenic ebolaviruses [19–21, 33–34, 36] will boost the development of effective EBD therapies and inspire new approaches to the design vaccines against this virus. Studies of cross-protective immunity and antibodies isolated from convalescent patients with EVD give us hope that a ZEBOV GP-based vaccine inducing immunity against both ZEBOV и BDBV is not just wishful thinking. Adding SUDV glycoprotein to the vaccine would make it effective against SUDV species, as well.

CONCLUSIONS

The comparative analysis of GP in 1,548 ZEBOV, SUDV and BDBV isolates has demonstrated a high variability of amino acid sequences in the glycoproteins representing different ebolaviruses (~60–65% homology). Further analysis of epitope homology in the glycoproteins of ZEBOV, SUDV and BDBV, accounting for the tertiary protein structure, has established that ZEBOV and BDBV glycoproteins have identical amino acids capable of binding to the protective antibodies and thus neutralizing these viral species characterized by low homology of linear amino acid sequences. These findings are fully consistent with the reports of the ability of candidate and approved vaccines against Ebolavirus to induce cross-immunity against ZEBOV and BDBV. Protection against the lethal infection caused by SUDV can be ensured only by vaccines based on SUDV GP.
We believe that the facts listed above clearly establish that development of effective vaccine protecting humans against pathogenic Ebolavirus species should focus on vectors expressing at least two glycoprotein types: of ZEBOV and SUDV.