Genus: Marburgvirus

Genus: Marburgvirus

Distinguishing features

Two marburgviruses (Marburg virus [MARV] and Ravn virus [RAVV]) are highly lethal human pathogens. Marburgviruses are notable for encoding only a single glycoprotein (GP1,2) from their GP genes (Feldmann et al., 1992).



Virions are filamentous in shape, but particles can also be branched, circular, U-shaped, or 6‑shaped. Spherical forms are rare to absent (Figure 1.Marburgvirus). Virions vary greatly in length (>1 μm) but have a uniform diameter of about 80 nm. Peak infectivity has been associated with particles of about 665 nm in length. Virions are composed of a central core, formed by a helical ribonucleoprotein (RNP) complex, surrounded by matrix layer and a lipid envelope, derived from the host cell membranes. Spikes about 7 nm in diameter and spaced at intervals of about 10 nm are seen as globular structures on the surface of virions (Ellis et al., 1979a, Ellis et al., 1979b, Geisbert and Jahrling 1995, Ryabchikova and Price 2004, Welsch et al., 2010).

Figure 1.Marburgvirus. A) Scanning electron micrograph of Marburg virus particles (red) budding from an infected grivet (Chlorocebus aethiops (Linnaeus, 1758)) Vero E6 cell. B) Transmission electron micrograph of Marburg virus particles (red) found both as extracellular particles and budding particles from Vero E6 cells. Images are colorized for clarity. Courtesy of John G. Bernbaum and Jiro Wada, NIH/NIAID/DCR/IRF-Frederick, Fort Detrick, MD, USA.

Physicochemical and physical properties

The buoyant density of Marburg virus (MARV) particles in potassium tartrate is about 1.14 g/ml (Kiley et al., 1988).

Nucleic acid

Marburgvirus genomes are linear non-segmented RNA molecules of negative polarity. These genomes are ≈19.1 kb. Genomic RNAs are not polyadenylated at their 3′‑ends, and there is no evidence of 5′-end terminal cap structures or covalently linked proteins. The Mr of a genomic RNA is about 4.2×106, and the genome represents about 1.1% of the total virion mass (Feldmann et al., 1992, Kiley et al., 1988).


Marburgviruses express seven structural proteins (Table 1.Marburgvirus, nucleoprotein [NP], RNP complex-associated protein 24 [VP24], polymerase cofactor [VP35], transcriptional activator [VP30], the matrix protein [VP40], glycoprotein [GP1,2], and large protein [L]), all of which are homologous to those of cuevaviruses, dianloviruses, and ebolaviruses. The second most abundant structural protein in a marburgvirion is NP, which encapsidates the marburgvirus genome. The least abundant protein is L, which mediates marburgvirus genome replication and transcription (Shi et al., 2018, Feldmann et al., 1992, Kiley et al., 1988). The marburgvirus RNP complex consists of NP, VP24, VP35, VP30, and L. These RNP complexes associate with VP40, which lines the inner side of the virion membrane, and GP1,2, which form globular spikes on the outside of the virion membrane (Kirchdoerfer et al., 2017, Mühlberger et al., 1999, Welsch et al., 2010).

Table 1.Marburgvirus. Location and functions of marburgvirus structural proteins.

Protein (abbreviation)
Encoding gene
Nucleoprotein (NP)
1 (NP)
RNP complex component; second‑most abundant protein in infected cells and in virons; consists of two distinct functional modules; homo‑oligomerizes to form helical polymers; binds to genomic and antigenomic RNA, VP35, VP40, VP30, and VP24; phosphorylated
Nucleocapsid and cellular inclusion body formation; encapsidation of marburgvirus genome and antigenome; genome replication and transcription
(Becker et al., 1994, Becker et al., 1998, Mühlberger et al., 1998, Sanchez et al., 1992, Lötfering et al., 1999, Kolesnikova et al., 2000, Liu et al., 2017, Zhu et al., 2017)
Polymerase cofactor (VP35)
2 (VP35)
RNP complex component; homo‑oligomer; phosphorylated; binds to double‑stranded RNA, NP, and L
Replicase‑transcriptase cofactor; inhibits innate immune response by interfering with IRF3, IRF7, and DDX58; inhibits protein kinase R and PRKRA activity
(Becker et al., 1998, Mühlberger et al., 1998, Möller et al., 2005, Bale et al., 2012, Ramanan et al., 2012, Edwards et al., 2016, Bruhn et al., 2017, Hume and Mühlberger 2018)
Matrix protein (VP40)
3 (VP40)
Most abundant protein in infected cells and in virons; consists of two distinct functional modules; homo‑oligomerizes to form dimers and circular hexamers and octamers; binds single‑stranded RNA, VP35; hydrophobic; membrane‑associated; contains one late‑budding motif; binds NEDD4 and TSG101
Matrix component; regulation of genome transcription and replication; regulation of virion morphogenesis and egress; inhibits JAK‑STAT pathway
(Kolesnikova et al., 2002, Kolesnikova et al., 2007, Urata et al., 2007, Urata and Yasuda 2010, Valmas and Basler 2011, Kolesnikova et al., 2012, Koehler et al., 2016, Koehler et al., 2018, Kolesnikova et al., 2004a, Kolesnikova et al., 2004b, Dolnik et al., 2015, Valmas et al., 2010)
Glycoprotein (GP1,2)
4 (GP)
Type I transmembrane and class I fusion protein; cleaved to GP1 and GP2 subunits that heterodimerize; mature protein is a trimer of GP1,2 heterodimers; inserts into membranes; heavily N- and O‑glycosylated, acylated, phosphorylated. ADAM17 converts GP1,2 into a soluble form (GP1,2Δ)
Virion adsorption to marburgvirus‑susceptible cells via cellular attachment factors; determines marburgvirus cell and tissue tropism; induction of virus‑cell membrane fusion subsequent to endolysosomal binding to NPC1; inhibits innate immune response by interfering with BST2. Function of GP1,2Δ is unknown
(Funke et al., 1995, Carette et al., 2011, Côté et al., 2011, Jouvenet et al., 2009, Dolnik et al., 2004, Misasi et al., 2012, Geyer et al., 1992, Volchkov et al., 2000, Will et al., 1993, Sakuma et al., 2009)
Transcriptional activator (VP30)
5 (VP30)
RNP complex component; hexameric zinc finger protein; binds single-stranded RNA, NP, and L; phosphorylated
Transcription initiation, reinitiaition, and antitermination; enhancement of transcripion (but to a lesser extent than ebolavirus VP30)
(Becker et al., 1998, Tigabu et al., 2018, Modrof et al., 2001, Enterlein et al., 2006, Wenigenrath et al., 2010, Albariño et al., 2013)
RNP complex-associated protein (VP24)
6 (VP24)
RNP complex component; homo‑tetramerizes; hydrophobic and membrane‑associated
Regulation of genome transcription and replication; regulation of virion morphogenesis and egress; targets KEAP1 to activate NFE2L2-induced cytoprotective responses
(Wan et al., 2017, Edwards et al., 2014, Edwards and Basler 2015, Lee et al., 2009, Zhang et al., 2014, Bamberg et al., 2005, Johnson et al., 2016, Page et al., 2014)
Large protein (L)
7 (L)
RNP complex component; homo‑dimerizes; binds to genomic and antigenomic RNA, VP35, and VP30; mRNA capping enzyme
Genome replication and mRNA transcription
(Kirchdoerfer et al., 2017, Mühlberger et al., 1999, Becker et al., 1998, Mühlberger et al., 1998, Koehler et al., 2016, Mühlberger et al., 1992)
ADAM17, ADAM metallopeptidase domain 17; BST2, bone marrow stromal antigen 2; DDX58, DExD/H-box helicase 58; JAK-STAT, Janus kinase-signal transducer and activator of transcription; KEAP1, kelch‑like ECH‑associated protein 1; IRF, interferon regulatory factor; NPC1, NPC intracellular cholesterol transporter 1; NEDD4, E3 ubiquitin protein ligase; NFE2L2, nuclear factor, erythroid 2 like 2; PRKRA, protein activator of interferon induced protein kinase EIF2AK2; RNP, ribonucleoprotein; TSG101, tumor susceptibility 101; VP, virus protein


The viral envelope is derived from host cell membranes and is considered to have a lipid composition similar to that of the host cell membranes (Bavari et al., 2002). Marburgvirus glycoproteins are acylated (Funke et al., 1995).


The glycoproteins of marburgviruses are oligomannosidic contain hybrid-type N‑glycans and neutral fucosylated bi-, tri-, and tetra-antennary species, most of which carry an additional bisecting N-acetylglucosamine. In addition, the glycoproteins contain O-linked glycans of the neutral mucin type. In contrast to ebolavirions, sialylation is rare or absent. Glycans constitute >50% of the GP1,2 total mass (Geyer et al., 1992, Feldmann et al., 1991, Hashiguchi et al., 2015).

Genome organization and replication

The marburgvirus genome has the gene order 3′-NP-VP35-VP40-GP-VP30-VP24-L-5′ (Figure 2.Marburgvirus). The extragenic sequences at the extreme 3′-end (leader) and 5′-end (trailer) of the genome are conserved and partially complementary. Genes are flanked by conserved transcriptional initiation and termination (polyadenylation) sites. Most genes are separated by non‑conserved intergenic sequences, but two genes overlap. In addition, most genes possess relatively long 3′-end and 5′-end noncoding regions. Contrary to cuevaviruses and ebolaviruses, but similar to dianloviruses, the GP genes of marburgviruses encode only a single protein, GP1,2 (Feldmann et al., 1992, Bukreyev et al., 1995). 

Figure 2.Marburgvirus. Schematic representation of the marburgvirus genome organization. Genomes are drawn to scale. Courtesy of Jiro Wada, IRF-Frederick, Fort Detrick, MD, USA.

The replication strategy of marburgviruses is assumed to be highly similar to that of dianloviruses and reminiscent of that of cuevaviruses and ebolaviruses (Brauburger et al., 2015, Manhart et al., 2018). Marburgvirions enter host cells by endocytosis (Bhattacharyya et al., 2011) after marburgvirus GP1,2 mediates cell surface attachment factor (e.g., C-type lectins, integrins, hepatitis A virus cellular receptor 1 [HAVCR1]) binding and subsequent low‑pH‑dependent fusion into endosomes (Davey et al., 2017). Cathepsin L cleavage is required for GP1,2 binding to the endosomal receptor NPC intracellular cholesterol transporter 1 (NPC1), which is also used by cuevaviruses, dianloviruses, and ebolaviruses (Carette et al., 2011, Côté et al., 2011, Ng et al., 2014, Misasi et al., 2012). Uncoating is presumed to occur in a manner analogous to that of other mononegaviruses. Marburgvirus transcription and genome replication take place in the cytoplasm and, in general, follow the models for members of the families Paramyxoviridae and Pneumoviridae. Transcription starts at the conserved transcriptional initiation site, and polyadenylation occurs at a stretch of uridine residues within the transcriptional termination site. The 5′-terminal non-coding sequences likely form hairpin‑like structures in all (capped and polyadenylated) mRNAs. Replication involves the synthesis of full‑length positive-sense genome copies (antigenomes) (Feldmann et al., 1992, Brauburger et al., 2015, Mühlberger et al., 1999, Becker et al., 1998, Mühlberger et al., 1998). During infection, massive amounts of nucleocapsids accumulate intracellularly and form intracytoplasmic inclusion bodies, which are the sites of marburgvirus transcription, replication, and nucleocapsid assembly. Mature nucleocapsids are transported for envelopment to the plasma membrane, where budding occurs in a VP40-mediated process (Kolesnikova et al., 2002, Kolesnikova et al., 2007, Kolesnikova et al., 2012, Kolesnikova et al., 2004a, Kolesnikova et al., 2004b, Dolnik et al., 2015, Schudt et al., 2013, Takamatsu et al., 2019) (Figure 3.Filoviridae).


Marburgviruses appear to be endemic in Western Africa (MARV), Middle Africa (MARV), Eastern Africa (MARV, RAVV), and Southern Africa (MARV). Egyptian rousettes (Rousettus aegyptiacus (Geoffroy, 1810)) are naturally infected by marburgviruses. The route of initial human infection is unknown (Amman et al., 2017, Amman et al., 2012, Amman et al., 2014, Towner et al., 2009, Pawęska et al., 2018). The major route of human-to-human transmission of marburgviruses requires direct contact with blood, bodily fluids, or injured skin. Both marburgviruses (MARV and RAVV) are highly lethal human pathogens (Kuhn et al., 2020). In the laboratory, rodents (laboratory mice, domesticated guinea pigs, golden hamsters) and nonhuman primates (common marmosets, crab-eating macaques, grivets, hamadryas baboons, rhesus monkeys, squirrel monkeys) can be infected experimentally with various marburgviruses, but lethal infection of rodents requires sequential adaptation (Siragam et al., 2018, St Claire et al., 2017, Kuhn 2008). In contrast to ebolaviruses, marburgviruses do not cause disease in domestic ferrets (Cross et al., 2018, Wong et al., 2018).


The antigencity of marburgvirions is primarily due to their glycoproteins. Several anti‑GP1,2 monoclonal antibodies are specific for marburgviruses. The ability of a monoclonal antibody to neutralize marburgvirus infection in vitro may not necessarily be predictive of protective efficacy in vivo; conversely, antibodies that are non-neutralizing in vitro may be protective in vivo. There is only limited antigenic relatedness between the virions of marburgviruses and other filoviruses (Hashiguchi et al., 2015, Flyak et al., 2015, Froude et al., 2017, King et al., 2018, Marzi et al., 2018, Mire et al., 2017, Sangha et al., 2017).

Derivation of names

Marburgvirus: from Marburg an der Lahn, the city in West Germany where the first registered outbreak of Marburg virus disease occurred (Siegert et al., 1967). 

Species demarcation criteria

The genus currently includes only a single species.

Phylogenetic relationships across the genus have been established from maximum‑likelihood trees generated using coding-complete or complete genome sequences (Figure 4.Filoviridae) or by phylogenetic analysis of RNA-directed RNA polymerase (RdRP) sequences (Wolf et al., 2018). MARV and RAVV are relatively distinct in genomic sequence similarity analyses but occupy the same ecological niche and cause the same disease in humans and nonhuman primates. Hence, they are currently considered distinct viruses belonging to the same species.

Member species

Exemplar isolate of the species
SpeciesVirus nameIsolateAccession numberRefSeq numberAvailable sequenceVirus Abbrev.
Marburg marburgvirusMarburg virusMarburg virus/H.sapiens-tc/KEN/1980/Mt. Elgon-MusokeDQ217792NC_001608Complete genomeMARV
Marburg marburgvirusRavn virusRavn virus/H.sapiens-tc/KEN/1987/Kitum Cave-810040DQ447649Complete coding genomeRAVV

Virus names, the choice of exemplar isolates, and virus abbreviations, are not official ICTV designations.