Main

Orthobunyaviruses belong to the Bunyaviridae family, which is the largest and most diverse family of RNA viruses, comprising more than 350 named isolates. These negative-sense RNA viruses take their name from Bunyamwera virus (BUNV), which was originally isolated in 1943 from Aedes spp. mosquitoes during an investigation of yellow fever in the Semliki Forest, Uganda. During the following 25 years, several antigenically related viruses were identified in laboratories in South America, Africa and India, mostly owing to work that was funded by The Rockefeller Foundation1. The serological relationships between these isolates initially led to the Bunyamwera supergroup of viruses, until further biochemical and structural information resulted in the formal establishment of the family Bunyaviridae in 1975. The family currently contains five genera: Orthobunyavirus, Hantavirus, Nairovirus, Phlebovirus and Tospovirus. Orthobunyaviruses, nairoviruses and phleboviruses are all arboviruses, which means that they replicate in, and are transmitted by, blood-feeding arthropods (such as mosquitoes or ticks), but they can also infect vertebrates. Tospoviruses are also transmitted by arthropods (in this case, thrips), but they infect plants, such as tomato and watermelon, rather than vertebrates. Hantaviruses are mainly carried by rodents, but humans can be infected via aerosolized rodent excreta2.

At least 30 orthobunyaviruses are responsible for several disease syndromes in humans, including acute but self-limiting febrile illness (for example, Oropouche virus (OROV)), encephalitis (for example, La Crosse virus (LACV)) and haemorrhagic fever (for example, Ngari virus). In addition, some orthobunyaviruses, such as Cache Valley virus, cause abortion and teratogenic effects in ruminants3. However, despite an increasing awareness of the potential detrimental effects of orthobunyaviruses on human and animal health, we know little about the burden of orthobunyavirus disease; for example, it is estimated that there are several hundred thousand asymptomatic LACV infections annually4 in the United States, but the true number of infections is unknown.

In 2011, a novel orthobunyavirus, Schmallenberg virus (SBV), emerged in Europe (Box 1), which emphasizes the potential of these viruses to emerge in new localities and cause serious disease. In addition, the structures of orthobunyavirus nucleocapsid proteins and virus particles have recently been determined, which has generated renewed interest in these viruses. In this Review, I discuss these structural advances, which, together with exploitation of reverse genetics to probe orthobunyavirus replication and interactions with host innate immune responses, are providing new insights into orthobunyavirus biology.

Arthropod transmission

Most orthobunyaviruses are transmitted by mosquitoes or culicoid flies (that is, midges), although a few are vectored by ticks and the Kaeng Khoi virus is transmitted by bed bugs3. In general, only one or a very few arthropod species are competent to transmit a particular orthobunyavirus, and strict vector preference is observed even in regions where different viruses and vectors coexist5. Female mosquitoes acquire the virus when feeding on an infected host, and infected mosquitoes show few adverse signs of infection6, although changes in the feeding behaviour of LACV-infected Aedes spp. mosquitoes have been detected7 and LACV-infected female mosquitoes have been reported to mate more efficiently than their uninfected counterparts8. Mosquitoes become persistently and systemically infected, and in some cases, vertical (that is, transovarial) transmission to offspring has been reported5. Transovarial transmission is an important mechanism for the maintenance of some viruses (such as LACV) during the winter months and can result in infected male mosquitoes that transfer the virus to females during mating9. The mild nature of the infection is clearly advantageous for mosquito transmission, and both horizontal and vertical transmission between mosquitoes obviates the need for a vertebrate host, which suggests that orthobunyaviruses might have initially evolved as insect-only viruses. Indeed, the recent isolation of novel viruses from different mosquito species that form a phylogenetic sister clade to orthobunyaviruses (such as Herbert virus and related viruses) and that seem to be unable to replicate in vertebrate cells, provides support for this hypothesis10.

Pathogenesis in the vertebrate host

Infection is initiated by the bite of an infected arthropod and the virus spreads to striated muscle, where extensive replication occurs, which results in high viraemia. The virus then spreads to most organs of the body and can cross the blood–brain barrier, where it targets neurons11; infection of the central nervous system (CNS) is age-related and juvenile animals seem to be more susceptible than adults. However, studies of LACV-infected mice have shown that the virus remains neurovirulent (that is, it causes a lethal infection following direct intracranial inoculation), even when neuroinvasiveness (that is, spread to the CNS following peripheral inoculation) in older animals is lost12. Transfer to the CNS might also be mediated by olfactory neurons, which was demonstrated in weanling mice that were infected with LACV12. Examination of malformed lambs and calves that were born from SBV-infected animals has also shown that viral antigens reach high levels in brain tissue13, and similar findings have been reported for other orthobunyaviruses that cause teratogenic effects. In sheep that are experimentally infected with SBV, the period of viraemia is short (3–5 days), but viral RNA is detected in the spleen and lymph nodes for up to 44 days14; the importance of persistence in the lymphoreticular system requires further investigation.

The incubation period of orthobunyaviruses in humans is typically 3–8 days, followed by the sudden onset of fever, headache, joint and muscle aches, which are occasionally accompanied by nausea and rash. Symptoms usually last for around 1 week, and a second bout of fever may occur 1–2 weeks later (particularly with OROV infection). Although most patients recover without long-lasting sequelae, severe LACV infections can result in seizures and convulsions that can last for up to 1 year. In addition, a few fatalities owing to LACV infection have been reported15.

Molecular biology of orthobunyaviruses

There are more than 170 named viruses in the Orthobunyavirus genus, which are grouped into 18 serogroups on the basis of serological relatedness, as determined by complement fixation, neutralization and haemagglutination inhibition assays3. Few of these viruses have been studied at the molecular level, and much of what is known about orthobunyavirus molecular biology is based on the study of only a few viruses, predominately BUNV (which is the prototype of both the genus and family) and LACV. However, as more information becomes available for different viruses, it is clear that there is great diversity among orthobunyaviruses, and it is highly likely that there will be exceptions to the generalizations that are outlined below.

Virus particle. Orthobunyavirus virions are relatively simple in their overall composition and comprise just four structural proteins: two surface glycoproteins, termed Gn and Gc, and two internal proteins, the N (nucleocapsid) protein and the L protein (the viral RNA-dependent RNA polymerase (RdRp)). The three genomic RNA segments are encapsidated by the N protein to form ribonucleoprotein (RNP) complexes that associate with the RdRp and are contained within the lipid envelope of the particle (Fig. 1a). Early electron microscopy analyses of purified virions showed spherical to pleomorphic particles that were 90 nm in diameter, with a double membrane envelope and a fringe of spikes that were predicted to be the glycoproteins16. Recent cryoelectron microscopy studies and subtomographic averaging have revealed much greater detail of the structure of BUNV17. Virions average 108 ± 8 nm in diameter and, although they are pleomorphic, they have a locally ordered lattice of glycoprotein spikes that protrude 18 nm from the membrane (Fig. 1b). The spikes are predicted to comprise trimers of Gn–Gc heterodimers and form a unique tripod-like arrangement on the virus surface (Fig. 1c). This is distinct from the tetrameric square-like lattice of hantavirus particles18 and the icosahedral arrangement of pentamers and hexamers of phlebovirus particles19. Acidification of BUNV particles results in a structural change in the head region of the spike that presumably involves the larger Gc protein, which is the predicted fusion protein (see below).

Figure 1: The orthobunayvirus particle.
figure 1

a | Schematic representation of the orthobunyavirus virion. The three genome segments (small, medium and large) are encapsidated by the N (nucleocapsid) protein to form ribonucleoprotein (RNP) complexes, which associate with the viral RNA-dependent RNA polymerase (RdRp). The RNPs are packaged within a lipid envelope that is derived from the host cell Golgi complex, which is modified by insertion of the viral glycoproteins Gn and Gc. The Gn and Gc proteins form a trimer of heterodimers that create the spikes on the virus particle. b | A cross-section model of a Bunyamwera virus virion based on a tomographic reconstruction. c | The tripodal architecture of the glycoprotein spike can be divided into two major protein contacts: the floor region, which is adjacent to the membrane; and the membrane-distal head region. These two protein contacts are separated from each other by a stalk region. Structural regions and their corresponding dimensions are indicated. Parts b and c of the figure are reproduced from Ref. 17.

PowerPoint slide

Organization of the viral genome. Like all members of the Bunyaviridae family, the genome of orthobunyaviruses comprises three unique segments of single-stranded RNA (ssRNA) that have a common organization that includes a coding region, which is flanked by 3′ and 5′ UTRs. The average size of each genome segment is generally consistent within a genus, and those of orthobunyaviruses are about 6.9 kb for the L (large) segment, 4.5 kb for the M (medium) segment and 1.0 kb for the S (small) segment20 (Fig. 2). The terminal nucleotides at the 3′ and 5′ ends of each segment are complementary, which permits the formation of a 'panhandle' structure that functions as the promoter for both the transcription and replication of each segment21,22,23. Although the lengths of each coding region, and hence the sizes of encoded proteins, are relatively conserved among different viruses, there is much more diversity in the length and sequence of the UTRs. In addition to functioning as promoters, the UTRs are also required for the encapsidation of the genomic RNA by the N protein, the termination of mRNA transcription and packaging of the RNP into virus particles24,25,26,27, and it is remarkable that these common functions are mediated by such disparate sequences. Intriguingly, deletion of internal sequences in the UTRs of BUNV resulted in attenuation of virus replication and loss of cytopathogenicity in mammalian cell culture, but the mechanistic basis of this is unknown28.

Figure 2: Coding strategy of the orthobunyavirus genome.
figure 2

The small, medium and large genomic RNAs are shown and the total length of the RNA is shown above each segment, using Bunyamwera virus as an example. The genomic RNAs are transcribed into mRNAs, which contain host-derived sequences at the 5′ end but are truncated at the 3′ end compared with the template. The coding regions are shown as different coloured boxes. The terminal sequences of each genome segment are complementary (red), which results in base pairing between the 3′ and 5′ ends to form a panhandle that is used as a promoter for transcription and replication. Protein products with their estimated molecular weights are also shown. The L (large) segment encodes the L protein, which is the RNA-dependent RNA polymerase (RdRp). The M (medium) segment encodes a polyprotein that is proteolytically cleaved to generate the viral glycoproteins Gn and Gc and the non-structural protein, NSm. The S (small) segment encodes the N (nucleocapsid) protein and a second non-structural protein (NSs), which is translated from an overlapping reading frame in the same mRNA by leaky ribosomal scanning.

PowerPoint slide

The genomic RNA segments have a negative-sense polarity, which means that the genomic RNA must be transcribed into positive-sense mRNAs for translation, and this occurs immediately after infection. The S segment encodes the N protein, the M segment encodes the two external glycoproteins Gn and Gc, and the L segment encodes the RdRp29. The glycoproteins are encoded as a precursor polyprotein that also contains the non-structural protein NSm. The coding order of the M segment mRNA is Gn–NSm–Gc30, and it is co-translationally processed by host proteases to produce the three proteins. The S segment of most orthobunyaviruses encodes a second non-structural protein, NSs, which is within the N protein-coding sequence and is translated from the same mRNA using an alternative AUG initiation codon31 (in a process known as leaky ribosomal scanning). However, in some orthobunyaviruses, the NSs ORF is severely truncated or even absent32,33. By contrast, a recently characterized orthobunyavirus, Brazoran virus, has a larger S segment (which is 1.7 kb) and encodes the NSs protein in an ORF that precedes the N protein ORF34. To further emphasize the increasing diversity of the bunyavirus family, the newly described mosquito-associated viruses (Herbert-like viruses) that are mentioned above do not encode either NSs or NSm10.

L protein. The L protein that is encoded by the large genomic segment is the RdRp, and it is responsible for catalysing both transcription and replication. Conserved amino acid sequence motifs have been identified in the L protein, and their crucial role in RNA synthesis has been confirmed by mutagenesis35. The amino-terminal domain of the L protein contains an endonuclease domain that cleaves capped oligonucleotides from the 5′ ends of host mRNAs, which are then used to prime viral mRNA synthesis36. This process, which is known as cap snatching, is also a feature of influenza virus transcription, but in contrast to influenza viruses, for orthobunyaviruses it occurs in the cytoplasm rather than in the nucleus37,38.

Gn and Gc proteins. Orthobunyavirus glycoproteins typically differ in size: Gn is 32–35 kDa, whereas Gc is 110–120 kDa. Both proteins are type I integral membrane proteins that are modified by N-linked glycosylation and are rich in cysteine residues (which constitute about 5% of the total number of amino acids). Conservation of cysteine localization in the glycoproteins of different viruses suggests that they are involved in the formation of intermolecular and intramolecular disulphide bonds, which is important for protein structure and function. Gn and Gc form a heterodimer in the ER, and their transport to the Golgi complex depends on a targeting signal that has been mapped to the transmembrane domain of Gn in BUNV39. When Gc is expressed alone, it is retained in the ER, which suggests that Gn functions as a chaperone for the correct trafficking of Gc to the Golgi complex before viral budding39. Gc is predicted to contain a class II fusion domain40, and mutation of this motif in LACV supports this prediction41. Mutation of residues around the fusion peptide in BUNV Gc led to a severe reduction in membrane fusion, whereas deletion of the N-terminal domain of the Gc ectodomain had only minimal effects on virus replication42. Indeed, it has been possible to generate a recombinant virus that expresses a Gc protein with an N-terminal-truncation that is fused to GFP using reverse genetics43 (Box 2).

N protein. The N protein, which is around 25–30 kDa in size, is the main protein that is produced in infected cells; it is highly immunogenic and elicits complement-fixing antibodies. The N protein encapsidates both genomic and antigenomic RNA (but not viral mRNA) to form RNPs that are the templates for the viral RdRp. The signal for encapsidation is located in the 5′ end of both the genome and antigenome, and it is likely that the N protein binds to ssRNA rather than double-stranded RNA (dsRNA; such as that formed in the panhandle structure) (Fig. 2) in a structure-dependent manner24,25,44. In addition to interacting with RNA, the N protein interacts with itself and also with RdRp, Gn and Gc45, although the molecular details of these other interactions are currently lacking.

The atomic structures of four orthobunyavirus N proteins (from BUNV, LACV, Leanyer virus and SBV) were recently solved, either as apoproteins or in complex with RNA46,47,48,49,50,51 (Fig. 3). These structures show that the N protein has a unique fold that is not seen in other negative-sense viruses or in other Bunyaviridae genera, such as phleboviruses (for example, Rift Valley fever virus) or nairoviruses (for example, Crimean Congo haemorrhagic fever virus)52. The N protein protomer consists of two major domains (an N-terminal domain and a carboxy-terminal domain) that have flexible N-terminal and C-terminal arms and a central positively charged cleft that binds the viral RNA (Fig. 3a,b). The protein forms a tetramer, in which each monomer interacts with neighbouring molecules in a head-to-tail manner that is mediated by the N- and C-terminal arms, and the central cleft of each monomer is aligned with that of its neighbour to create a continuous RNA-binding groove in the interior of the N protein ring, which binds the RNA in a cross-shape configuration (Fig. 3c). Each protomer binds to 11 bases: 8 bases in the cleft and 3 bases near the N-terminal arm. The bound RNA is thus protected, and a local rearrangement of the N protomers would be required for the RdRp to access it for transcription and replication. By comparing the apoN and N protein–RNA complex structures from SBV, conformational changes in both the N- and C-terminal arms and in some of the RNA-binding residues in the RNA-binding groove were observed, which suggests how such reorganization might occur46,48 — the challenge now is to show this experimentally. Electron microscopy has shown that the orthobunyavirus RNP has a highly ordered helical appearance48,50,51, which correlates with the structure of a particular crystallographic form of the LACV N protein51. In addition to these structural studies, mutational studies of the N proteins from BUNV and SBV have provided functional data that are consistent with the structural information47,53,54. Furthermore, the strict requirement of the N- and C-terminal arms for the oligomerization of the N protein provides a target for the development of small molecules that inhibit these interactions.

Figure 3: Structure of the orthobunyavirus N protein.
figure 3

a | The crystal structure of the N (nucleocapsid) protein protomer from Schmallenberg virus (SBV) shows that it consists of an amino-terminal arm (NTA), an N-terminal domain (NTD), a carboxy-terminal domain (CTD) and a C-terminal arm (CTA). The NTA and CTA are important for mediating interactions between N protein monomers. b | Electrostatic surface potential map of the N protomer from SBV. Positively charged residues are shown in blue and negatively charged residues are shown in red. The central cleft that binds to viral RNA is indicated. c | Structure of the SBV N protein–RNA complex. The N protein forms a tetramer that binds the RNA in a cross-like shape on the inside of the tetramer. The N protein tetramer is shown in ribbon form, with each protomer in a different colour. The RNA is shown in stick form in orange. Parts a and b of the figure from Journal of Virology, 2013, 87, 5593–5601, http://dx.doi.org/10.1128/JVI.00223-13, reproduced with permission from American Society for Microbiology. Part c is modified, with permission, from Ref. 46, Cold Spring Harbor Laboratory Press.

PowerPoint slide

Non-structural proteins. Orthobunyaviruses encode one or, more commonly, two non-structural proteins. NSm is a 16–18 kDa protein that is encoded between Gn and Gc in the glycoprotein precursor (Fig. 2). The BUNV NSm protein was shown to localize to the Golgi complex independently of other viral proteins55,56 and to interact with the C terminus of Gn. The localization of NSm to the site of virus maturation suggests that NSm may be involved in virus assembly and budding56. Consistent with this, high-resolution microscopy of infected cells has revealed that virus factories contain novel tubular structures that comprise cellular proteins and the NSm protein57. However, the intact NSm protein is not essential for virus viability, as recombinant viruses that contain deletions within NSm or that express GFP-tagged NSm are viable, although they grow slower than the wild-type virus and to lower titres in mammalian cells45. In addition, insertion of β-galactosidase sequences into NSm of LACV did not affect the processing of the glycoprotein precursor or intracellular targeting of the mature glycoproteins to the Golgi complex58. The N terminus of NSm functions as an internal signal sequence for NSm itself, but it can also substitute for the signal sequence of Gc in viruses that have extensive deletions at the NSm–Gc gene boundary43.

Most orthobunyaviruses encode another non-structural protein called NSs, which is around 10 kDa in size. In contrast to the N protein, the amino acid sequence of NSs tends to be more variable between different viruses, and the protein is mostly hydrophobic and localizes to both the cytoplasm and the nucleus59. The NSs protein has multiple functions in the viral replication cycle and is the major virulence factor: recombinant viruses in which NSs expression is silenced are severely attenuated in animal models60. However, NSs is not essential for growth, as mutant viruses that lack NSs can replicate in mice and in cell culture13,61,62,63. In mammalian cell culture, NSs inhibits translation, and deletion mutants have an impaired ability to block protein synthesis; however, in mosquito cells, NSs has no effect on protein production61,63,64. NSs proteins are also involved in apoptosis, but the effects of the BUNV and LACV proteins seem to differ. For BUNV, NSs delays apoptosis65, whereas the NSs protein of LACV and other California serogroup viruses seems to induce apoptosis62,66. These NSs proteins have amino acid similarity to a region in Drosophila spp. Reaper protein, which is a known inducer of apoptosis67.

In minigenome assays, NSs inhibits RNA synthesis68,69, although the importance of this during natural infection has not yet been explored; however, it could be a mechanism to reduce the production of viral RNA transcripts and thereby block immune detection (see below).

Replication cycle

Cell entry involves interactions between the glycoproteins Gn and/or Gc and cell surface receptors (Fig. 4); the receptor (or receptors) are unknown, although it has been reported that Germiston orthobunyavirus entry is promoted by DC-SIGN70. Although DC-SIGN is highly expressed on dermal dendritic cells and is a good candidate for entry following transmission by an arthropod bite, other receptors are probably involved, as these viruses infect a wide range of tissues. Studies with OROV and LACV have shown that entry is likely to occur by clathrin-mediated endocytosis71,72, and for LACV, neither macropinocytosis nor caveolar endocytosis is required for entry. Furthermore, LACV entry is dependent on Rab5 but not on Rab7 (Rab5 and Rab7 are GTPases that are key regulators for the transport of endocytic vesicles to early and late endosomes, respectively), which indicates that trafficking to early endosomes is required for uncoating (Fig. 4). Several studies have suggested that Gc is the primary protein involved in attachment to both vertebrate and invertebrate cells73, but it has also been suggested that the smaller protease-resistant Gn protein of LACV functions as the attachment protein for mosquito cells74,75 — these conflicting results need to be revisited.

Figure 4: The orthobunyavirus replication cycle.
figure 4

Replication begins with viral attachment to host receptors (step 1), which is mediated by either or both of the viral glycoproteins (Gc and Gn), followed by virus internalization via receptor-mediated endocytosis (step 2). Acidification of endocytic vesicles leads to virion uncoating (step 3) and fusion of the viral membrane with the endosomal membrane. The viral RNA-dependent RNA polymerase (RdRp) catalyses primary transcription of viral mRNAs (step 4), which are primed by host cell-derived primers (not shown). Following translation of the viral mRNAs (step 5), the glycoproteins Gn and Gc dimerize in the ER and localize to the Golgi complex using a signal in the transmembrane domain in Gn. The three negative-sense viral genome segments (genomic RNA (gRNA)) are converted into positive-sense antigenomes (antigenomic RNA (agRNA)) for genome replication in the virus factory (step 6). The ribonucleoproteins (RNPs) are then transported to membranes of the Golgi complex that have been modified by insertion of Gn and Gc, and virus particles bud into Golgi membrane-derived vesicles (step 7). Golgi vesicles that contain virus particles are trafficked to the cell surface (step 8) and fusion of the vesicular membranes with the plasma membrane leads to the release of infectious virions, which may be facilitated by actin filaments (step 9). Nascent viral particles are subsequently released from the cell (step 10).

PowerPoint slide

Transcription. After release of the RNPs into the cytoplasm, the first event is primary transcription (Fig. 4), which involves primer-dependent synthesis of viral mRNAs and results in mRNAs that contain 12–18 host-derived nucleotides at their 5′ ends37. Although this sequence is generally heterogeneous, a preference for a certain motif at the junction of the primer and the viral mRNA has been observed, whereby the nucleotides at the 3′ end of the scavenged primer are similar to those at the 5′ terminus of the viral RNA. It has been suggested that, after the initial transcription of two or three nucleotides, the nascent mRNA could realign backwards on the template as a consequence of the terminal sequence repeats, before further elongation resumes, which results in partial reiteration of the 5′ sequence76.

Orthobunyavirus mRNAs do not have detectable poly(A) tracts at their 3′ ends, and no U-rich tract (which is used to signal polyadenylation by other negative-sense viruses) is present in the genome. The 3′ ends of the viral mRNAs have the potential to form stem–loop structures that may be involved in promoting efficient translation in a poly(A)-binding protein-independent manner77. Detailed mapping of transcription termination of the S segment of BUNV showed that the pentanucleotide sequence 5′-UGUCG-3′ signalled termination26,27. Although this sequence is present in the BUNV L segment and in the S segments of some other orthobunyaviruses, it does not occur in the BUNV M segment, which indicates that transcription termination signals are not universally conserved.

Bunyavirus transcription requires on-going translation in the host cell, which is a unique feature among negative-sense RNA viruses78. When infected cells are treated with protein synthesis inhibitors, such as cycloheximide, essentially no full-length viral mRNAs are produced79, and in vitro transcriptase reactions yield only short RNA products, unless they are supplemented with rabbit reticulocyte lysate80,81. Further experiments have shown that ongoing protein synthesis is needed to prevent premature termination of transcription, and it was postulated that in the absence of functional ribosomes the nascent mRNA and the template RNA hybridize, which prevents progression of the RdRp82. This model has been tested using BUNV genome analogues that contain premature stop codons, which showed that translation of the nascent mRNA prevented premature termination83. This coupling of translation and transcription is also a feature of bacteria and archaea but is rare in eukaryotes; thus, it is unclear why this family of viruses has adopted such a mechanism.

Genome replication. The template for genome replication is full-length positive-sense RNA (known as the antigenome), which is produced in a primer-independent manner from the negative-sense genome (Fig. 4). The molecular basis for the switch from transcription to replication is unclear, but the level of N protein in the cell is important, as sufficient N protein must be available to encapsidate nascent antigenomes and genomes during replication. Only the viral RdRp and N proteins are required for transcription and replication84, and whether different forms of the polymerase — for example, those that have been modified by association with different cellular factors — carry out these different activities has not been determined. The antigenome is always found encapsidated by the N protein in the cell, which might prevent interactions with the template for transcription, as well as minimizing the formation of dsRNA in the cell, which is important to avoid detection by the host immune response (see below). In cells that are co-infected with two different orthobunyaviruses, progeny virions may contain genome segments that are derived from both parents; this process is known as genome reassortment and can lead to reassortant viruses that have altered biological properties (Box 3).

Assembly and budding at the Golgi complex. The viral glycoproteins Gn and Gc are targeted to, and accumulate in, the Golgi complex, where assembly occurs. Viral RNPs are transported to regions of the Golgi membrane that have been modified by the insertion of the glycoproteins, which promotes budding into Golgi-derived vesicles (Fig. 4). Using a virus-like particle production assay, direct interactions between the cytoplasmic tails of the glycoproteins and the RNP were found to be necessary for budding85. Inside the Golgi stacks, annular particles are observed, which are defined as immature precursors that transform into dense intermediate particles86,87. This maturation is accompanied by a change in the sensitivity of Gn and Gc to endoglycosidase H (endo H), from complete sensitivity to partial resistance87,88, which indicates that the terminal sugar residues on the glycoproteins are processed during transport through the Golgi stacks to the trans-Golgi.

Similarly to other viruses that have segmented genomes, orthobunyaviruses have the problem of packaging a single copy of each of the three genomic segments in order to generate an infectious particle. Genome packaging has been extensively studied in influenza virus89, but comparatively little is known about this process in orthobunyaviruses. However, packaging seems to require only the UTR sequences, and the inclusion of N protein-coding sequences did not improve packaging efficiency24. In addition, the generation of a viable recombinant BUNV that contains a hybrid M–L segment (in which M segment UTRs flank the L segment ORF) would argue against the presence of packaging signals that extend into the coding sequence90. These findings differ from those in influenza virus, in which both non-coding and coding sequences are used as packaging signals.

Virus release. The progeny virions are transported to the cell surface via the exocytic pathway, which has been visualized using a recombinant BUNV that expresses GFP-tagged Gc43. Following virus release, a second maturation step occurs, during which the viral spikes (that is, the Gn and Gc trimers) become more defined86,87. Correlative light and electron microscopy has shown that multilamellar structures and extracellular membrane-deficient filament bundles are generated in infected cells at times of maximal virus release91. Progeny virions were found to be associated with the filament bundles (which contain actin), and cytochalasin D treatment of infected cells, which disrupts actin filaments, reduced released virus titres by 100-fold. These data suggest that actin filaments are involved in BUNV egress and could also be involved in transport of the virus to uninfected neighbouring cells.

Persistent infection of arthropod cells

In contrast to the usually lytic infection of vertebrate cells, orthobunyavirus infection of arthropod cells results in the establishment of a persistent infection. The cellular interactions between the virus and the host that enable efficient viral replication without harming the vector are unknown. Some clues come from studies of persistent infection of cultured mosquito cells, which have been described for several orthobunyaviruses29,92. The production of L segment-derived defective-interfering particles93 and the self-limiting production of N protein by encapsidation of S segment-derived mRNA94 have been suggested to contribute to persistence. Studies of BUNV mutants that lack NSs have shown that there is little impairment in their ability to replicate in certain Aedes albopictus cells (such as C6/36 and C7-10, which lack a functional Dicer 2-mediated RNAi response) but that NSs is important for infection of the A. albopictus U4.4 cell line (which has an active Dicer 2 enzyme) and also for efficient infection and spread out of the midgut of Aedes aegypti mosquitoes95. This suggests that a balance between the innate defence that is mounted by the RNAi pathway and virus replication must be achieved in the invertebrate host, but whether NSs directly disrupts the RNAi system requires further investigation. Although infected mosquito cells are not killed by the infection, transient morphological changes are observed95,96, and extensive filopodial-like structures occur in cells during the early phases of infection. The presence of virus particles in these projections suggests a mechanism for cell-to-cell spread of virus without rupture of the cell membrane.

Evasion of innate immunity in vertebrate cells

Pattern recognition receptors (PRRs) recognize foreign molecules, known as pathogen-associated molecular patterns (PAMPs), and in the case of viruses, these include various forms of nucleic acid, such as dsRNA and 5′-triphosphorylated RNA, which are recognized intracellularly by two cytoplasmic RNA helicases, retinoic acid-inducible gene I (RIGI) and melanoma differentiation-associated protein 5 (MDA5)97. Binding of an RNA ligand to these PRRs activates a signalling cascade that induces the expression of several transcription factors (including IFN regulatory factor 3 (IRF3), nuclear factor-κB (NF-κB) and cyclic AMP-dependent transcription factor (ATF2)–c-Jun), which promote transcription of interferon-β (IFNβ), leading to the upregulation of hundreds of IFN-stimulated genes (ISGs), some of which have antiviral activity98. It was recently shown that RIGI recognizes the nucleocapsid of LACV, which contains the terminal panhandle with a 5′ triphosphate, and recognition occurs in the absence of transcription or replication99. This suggests that the innate immune response is triggered at the earliest stage of infection, just after viral entry. Little is known about the antiviral activities of most ISG products, although they probably inhibit viral replication at many stages of the virus life cycle. The protein MxA inhibits LACV and BUNV replication by sequestering newly synthesized N protein in cytoplasmic inclusions, which blocks replication100. Other ISGs that inhibit BUNV include PKR, RSAD2 (commonly known as viperin), MTAP44, IFITM3, IRF1, HPSE and UBE2L6 (Refs 101, 102, 103). Individually, the degree of inhibition by these ISGs is relatively modest, but they presumably function together such that their combined effects are highly restrictive.

The IFN response is extremely powerful at controlling virus replication and, in cells that have been pretreated with IFN (to induce the antiviral state), orthobunyavirus replication is restricted, but the addition of IFN after the infection is established has little effect on virus replication102. In mammalian cells, BUNV, LACV and SBV NSs proteins inhibit the induction of IFN61,63,104,105, and for BUNV and LACV, this occurs downstream of transcription factor activation. BUNV NSs prevents phosphorylation of serine 2 in the heptad repeat of the C-terminal domain (CTD) of RNA polymerase II, which leads to an arrest in mRNA elongation and 3′ processing, thereby causing a general inhibition of cellular transcription59. This inhibition seems to be mediated by interactions between the C terminus of NSs and the Mediator component MED8 (Ref. 106). MED8 is in the head module of Mediator, which interacts directly with the CTD of RNA polymerase II (Ref. 107). The N terminus of BUNV NSs is also involved in transcriptional inhibition, although the mechanism involved remains to be elucidated108. Similarly, NSs of LACV targets RNA polymerase II, but in this case, the protein triggers proteasomal degradation of the hyperphosphorylated polymerase (that is, the transcriptionally active form) by exploiting early events in the cellular DNA damage response pathway109.

Outlook

Important progress has been made in recent years on several aspects of orthobunyavirus biology, particularly advances in determining the structures of individual viruses and virus proteins, which are providing new insights into the infection and replication processes, and the interactions between viruses and both vertebrate and arthropod hosts. However, there is much more to learn about this fascinating group of viruses. We are only just beginning to appreciate the evolutionary relationships between orthobunyaviruses: complete genome sequencing on a larger scale would be a major advance for this area, as less than one-quarter of the 170 recognized orthobunyavirus genomes have been fully sequenced. Analysis of the available genomes indicates that reassortment (Box 3) is more common in nature than perhaps previously appreciated, which has led to the suggestion that “most if not all bunyaviruses currently recognized may represent reassortants, some of which may be reassortants of existing viruses, and some of which may be reassortants of extinct viruses” (Ref. 110). The propensity for genome reassortment, the emergence of SBV in Europe, the spread of OROV in South America and the potential for climate change to aid the emergence of other novel variants highlight that increased global surveillance, together with improved diagnostic methods, are required to monitor orthobunyavirus activity; for example, the establishment of sentinel herds and active vector trapping at international ports and airports would aid early detection efforts111.

Besides vaccines against Simbu serogroup viruses112, including the recent development of two inactivated vaccines that have been approved for SBV113, no vaccines or antiviral agents are available against orthobunyaviruses. However, vaccine manufacturers have little incentive to develop vaccines against human orthobunyaviruses owing to a perceived limited market. Similarly, there has been little interest in research on antiviral drugs. The availability of efficient reverse genetics systems (Box 2) could be exploited for vaccine development, and several studies have shown that attenuated viruses can be engineered by the introduction of deletions within non-coding sequences, the rearrangement of coding and non-coding sequences, the deletion of the NSs protein and mutations in the N protein28,54,61,90,105. The construction of such attenuated strains could form the basis for live attenuated vaccines, and by combining such alterations, it might be possible to generate a virus that is unable to revert to virulence via back-mutation. In addition, it should be possible to generate viruses that are unable to reassort (Box 3) with circulating wild-type strains, which would limit their propensity to evolve. Second, the determination of orthobunyavirus N protein structures provides opportunities for the rational design of small-molecule inhibitors to interfere with N protein function. Understanding how individual ISGs inhibit orthobunyavirus replication could also be exploited to develop effective inhibitory compounds98.

Another area of orthobunyavirus research that is likely to expand in the coming years is focused on developing a detailed understanding of the interaction with the arthropod vector, such as the molecular basis for vector competence, and how this knowledge might also be exploited for infection control. This will require further advances in genomics and reagent development for vector species. Arthropods have their own innate immune defences, which are mostly based on RNAi for protection against viruses114, and how the viruses can survive these defences and replicate without harming the vector requires further investigation.