Chapter contents

Posted May 2018

Togaviridae: The family

Member taxa

Supporting information

  • Authors - corresponding authors: Rubing Chen (
  • References - sequence alignments and tree files


A summary of this ICTV Report chapter has been published as an ICTV Virus Taxonomy Profile article in the Journal of General Virology, and should be cited when referencing this online chapter as follows: 

Chen, R., Mukhopadhyay, S., Merits, A., Bolling, B., Nasar, F., Coffey, L.L., Powers, A., Weaver, S.C., and ICTV Report Consortium. 2018, ICTV Virus Taxonomy Profile: Togaviridae, Journal of General Virology, 99, 761–762.


The Togaviridae is a family of small, enveloped viruses with single-stranded positive-sense RNA genomes of 10–12 kb. Within the family, the Alphavirus genus includes a large number of species, while the Rubivirus genus includes the single species Rubella virus.  Most alphaviruses are mosquito-borne and are pathogenic in their vertebrate hosts. Many are important human and veterinary pathogens (e.g. chikungunya virus, eastern equine encephalitis virus). Rubella virus is transmitted by respiratory routes among humans.

Table 1.Togaviridae. Characteristics of the family Togaviridae.



Typical member

Sindbis virus (J02363), species Sindbis virus, genus Alphavirus


Enveloped, 65–70 nm spherical virions for alphaviruses or 50–90 nm pleomorphic virions for rubella virus with a single capsid protein and 3 or 2 envelope glycoproteins, respectively


10–12 kb of positive-sense, non-segmented RNA


Cytoplasmic, in vesicles derived from the plasma membrane/endosomal compartment. Assembled virions bud from plasma membrane (alphaviruses) or into the lumen of Golgi apparatus (rubella virus)


Non-structural proteins are translated from genomic RNA, and structural proteins from subgenomic RNA, both as polyprotein precursors

Host range

Humans and nonhuman primates, equids, birds, amphibians, reptiles, rodents, pigs, sea mammals, salmonids, mosquitoes and some other arthropods; most members of genus Alphavirus are mosquito-borne


Two genera (Alphavirus, Rubivirus) including more than 30 species



Togaviruses are enveloped viruses that contain a single-stranded, positive-sense RNA genome. The morphology and structures of members of the two genera, Alphavirus and Rubivirus, are very different (Figure 1.Togaviridae) and are described in their respective sections.

Figure 1.TogaviridaeImages and structures of Togaviradae particles. (A) Chikungunya virus particles purified from BHK cells and flash frozen in vitreous ice (courtesy of J. C-Y Wang and S. Mukhopadhyay).(B) Three dimensional cryo-electron reconstruction of chikungunya virus at 10.2 Å resolution. The triangle outlines one icosahedral unit with the symmetry axes labeled (courtesy of J. C-Y Wang). (C) Purified rubella virus flash frozen in vitreous ice (courtesy of AJ Battisti and V. Magala Prasad). (D) Representation of three different rubella virions, each determined using cryo-electron tomography and no averaging procedures. The virions are different sizes and shapes but all show surface glycoproteins arranged in rows around the particle. The resolution of the reconstructions are not absolute but are estimated to be better than 50 Å, (reproduced and modified from (Mangala Prasad et al., 2017)).

Physicochemical and physical properties

Alphavirus particles have a buoyant density in sucrose gradients of between 1.15 and 1.20 g cm-3 (Kuhn 2013, Sokoloski et al., 2013). The buoyant density of rubella virions in sucrose gradients is similar at 1.18–1.19 g cm-3 (Hobman 2013, Parkman 1965). Togaviruses can be physically denatured by treating with various chemicals including urea, formaldehyde, beta-propiolactone, detergents, and acids. Infectivity of togaviruses is also decreased by heat inactivation and exposure to ultraviolet light (Kuhn 2013, Hobman 2013, Parkman 1965, Park et al., 2016).

Nucleic acid

Alphaviruses and rubiviruses have single-stranded, positive-sense RNA genomes. The viral genomes contain two open reading frames encoding non-structural proteins and structural proteins, respectively. In addition to the genome RNA, a subgenomic RNA containing the structural protein genes and 3ʹ-non-coding region is synthesized during replication. Both the genomic and subgenomic RNAs have a viral type 0 7meGpppA cap at their 5ʹ-terminus (Ahola and Kaariainen 1995, Oker-Blom et al., 1984and a poly-A tail at their 3ʹ-terminus. The negative-strand RNA of the replication intermediate is neither capped nor contains a poly-A tail. The structural proteins are translated from the subgenomic mRNA (Kuhn 2013)


Togavirus virions, like other enveloped viruses, have their genome protected by capsid protein and have glycoprotein spikes on the surface of the particle that are responsible for cell entry, including receptor binding and membrane fusion (Kuhn 2013, Hobman 2013). The E1 and E2 glycoproteins of togaviruses are both type I membrane proteins that are glycosylated and dimerize to form the functional spike complexes found on the virion surfaces. The roles of specific proteins in the alphavirus and rubivirus virions are described in their individual sections.


Togaviruses acquire their lipid membrane from the host during budding. In general the lipid membrane accounts for close to 30% of the virion mass. See individual genus sections for lipid composition and sites of budding. 


Three of the alphavirus proteins (E3, E2, and E1) and two rubivirus proteins (E2, and E1) are glycosylated. For alphaviruses the carbohydrates are N-linked (Rice and Strauss 1982, Sefton 1977while rubella virus has both N- and O-linked modifications (Qiu et al., 1992, Lundstrom et al., 1991, Frey 1994). More detail about the glycosylation of these two groups is found in the genus sections.

Genome organization and replication

Two thirds of the genome of alphaviruses encodes the non-structural polyprotein(s) in a single ORF immediately after a 5′-non-coding region. The proteins are oriented as 5′ – nsP1 (methyltransferase and guanyltransferase) – nsP2 (helicase and protease) – nsP3 (phosphoprotein, (ADP-ribosyl) hydrolase and a key protein for interaction with host factors) – nsP4 (RNA dependent RNA polymerase, terminal adenylyltransferase) – 3′ (Kuhn 2013). There is a stop codon present between the nsP3 and nsP4 genes in the majority of alphaviruses resulting in a limited amount of polyprotein P1234 generated by inefficient read-through. Overlapping with the 3ʹ-end of the non-structural ORF, there is a promoter for transcription of the subgenomic mRNA from which the structural polyprotein is translated (Ou et al., 1982). The structural proteins include the capsid (C/CP), E3, E2, 6K, TF and E1 proteins. The structural ORF is followed by a non-coding region of varying length (range from 77 to > 980 nucleotides) and finally, a poly-A tail (Figure 2.Togaviridae). 

Figure 2.TogaviridaeTogavirus genomic coding strategies. Shown are comparative schematic representations of the alphavirus and rubivirus genomic RNAs with non-coding regions represented as solid black lines and ORFs as open boxes (NS-ORF=non-structural protein ORF; S-ORF=structural protein ORF). Within each ORF, the coding sequences for the proteins processed from the translation product of the ORF are delineated. The asterisk between nsP3 and nsP4 in the alphavirus NS-ORF indicates the stop codon present in some alphaviruses that must be translationally-read through to produce a precursor containing nsP4. Additionally, within the NS-ORFs, the locations of motifs associated with the following activities are indicated: (Mtr) methyl transferase, (Pro) protease, (Hel) helicase, (X) unknown function, and (RdRP) RNA-dependent RNA polymerase. The sequences encompassed by the sgRNA are also shown. (Courtesy of T. K. Frey.)

While rubella virus (RUBV) and alphaviruses share similar genome organization, there are major differences in their replication processes (Figure 3.Togaviridae). Alphavirus replication complexes (termed spherules) are initially formed at the plasma membrane of infected vertebrate cells (Frolova et al., 2010, Spuul et al., 2010). The initial P1234 polyprotein is cleaved by the viral protease activity of the nsP2 protein into P123 precursor protein and nsP4; these proteins form the replication complex associated with negative-sense RNA replication (Lemm et al., 1994). This negative-sense RNA forms a duplex with the positive-sense (genomic) strand and serves as the template in the synthesis of a full-length, positive-sense RNA that will eventually be encapsidated, as well as a subgenomic 26S mRNA that encodes the viral structural proteins (Pietilä et al., 2017). Cleavage of the P123 polyprotein is highly regulated and generates nsP1, nsP2 and nsP3 individual proteins, which are involved in the synthesis of positive-sense RNA (Lemm et al., 1994, Vasiljeva et al., 2003). Both alphavirus non-structural proteins and RNAs interact with multiple cellular proteins; some of these interactions are essential for replication (Sokoloski et al., 2010, Kim et al., 2016). In RUBV, the polyprotein precursor is cleaved either in cis or in trans into two products, P150 and P90, by a protease located near the C-terminus of P150. The order of the non-structural protein functions in the RUBV genome is different from that of the alphaviruses with P90 protein containing both helicase and RNA dependent RNA polymerase motifs (Kuhn 2013).

Figure 3.Togaviridae. Model for the processing of the alphavirus nonstructural polyprotein during replication. When low levels of P123 are present, cis-cleavage of P1234 generates the negative-sense RNA replicase of the virus. This results in primarily negative-sense RNA being transcribed from the incoming genomic RNA of the virus (upper panel). As the level of the trans-acting protease P123 rises in the infected cell, cleavage in trans generates other RNA replicase complexes. This results in a shift from the production of primarily negative-sense RNA to primarily positive-sense RNA. Eventually, replicase complexes capable of producing negative-sense RNA will no longer be present in the infected cell resulting in the complete cessation of negative-sense RNA synthesis (lower panel). The presence of a leaky opal termination codon (indicated by an asterisk) in the virus genome is believed to lead to a more rapid build-up of P123 in the infected cell, and thus a more rapid conversion to the production of positive-sense RNA.

The structural polyproteins for both alpha- and rubiviruses are translated from a subgenomic mRNA. For alphaviruses the predominant translation product is CP/E3/E2/6K/E1, but, at a low frequency, there is a (-1) translational frameshifting event that produces CP/E3/E2/TF. Alphavirus polyproteins are then cleaved by both viral and cellular proteases to produce individual structural proteins (Firth et al., 2008). Rubella virus (RUBV) CP lacks the viral autoprotease activity of the alphavirus C protein and thus relies completely upon cellular endopeptidases for structural protein cleavage. The glycoproteins that are produced are inserted into the endoplasmic reticulum during translation and are translocated to the plasma membrane for alphaviruses or to intracellular membranes for RUBV (Kuhn 2013). Upon generation of a sufficient amount of C protein, this protein assembles with the viral RNA to form the viral nucleocapsids. This process occurs in the cytosol for alphaviruses and during the budding process associated with the Golgi apparatus for RUBV. Budding through the plasma membrane (alphaviruses) or the Golgi and plasma membrane (RUBV) leads to the acquisition of a lipid envelope containing the two main membrane glycoproteins (Frey 1994, Jose et al., 2009).


The alphaviruses were originally described as Group A arboviruses based upon their antigenic cross-relationships. Using specific serological testing, antigenic complexes were proposed where all members of a particular complex were closely related to each other. Eight (eleven including the fish, seal and mosquito-specific alphaviruses) such complexes are described whose members, for the most part, are also genetically clustered (Kuhn 2013). The members of the two genera of the Togaviridae are antigenically quite distinct from each other with no detectable serological cross-reactivity.


Host Range

The genus Alphavirus mainly consists of mosquito-borne viruses although other hematophagous insects, including ticks, lice, and cliff swallow bugs, have been implicated in transmission (Lwande et al., 2013, Hayes et al., 1977). Vertebrate hosts include humans, non-human primates, equids, birds, amphibians, reptiles, rodents, and pigs (Kuhn 2013). There are two aquatic alphaviruses, southern elephant seal virus and salmon pancreas disease virus, infecting sea mammals and fish respectively. Furthermore, representing a new host-restricted complex within the alphaviruses, Eilat virus (EILV), has been shown to only infect insect cells and is incapable of replicating in vertebrate cells (Nasar et al., 2012).  RUBV is also host-restricted, as it appears to infect only humans.


With a few exceptions, most alphaviruses are transmitted alternately between their vertebrate and invertebrate hosts. For many alphaviruses, humans are dead-end hosts, incapable of developing sufficient viremia to infect mosquitoes, although human-mosquito-human transmission has been implicated in chikungunya (CHIKV) outbreaks. The aquatic alphaviruses, which include southern elephant seal virus (La Linn et al., 2001and salmon pancreas disease virus (Weston et al., 1999), appear to be transmitted horizontally, although both viruses have been isolated from lice, suggesting arthropod-borne transmission may also play a role. RUBV is transmitted from person-to-person, primarily through respiratory routes (Strauss and Strauss 2008). The insect-specific Eilat virus is defective for replication in vertebrates, and therefore is probably maintained via vertical transmission in mosquitoes (Nasar et al., 2012). 

Geographical Distribution

The alphaviruses have a worldwide distribution, inhabiting all continents except Antarctica (Powers et al., 2001). The alphaviruses are geographically restricted based on preferred ecological conditions, reservoir hosts and vector species, but continue to move around the globe and colonize new areas. Based on their distribution, alphaviruses have classically been described as Old World or New World viruses. However, CHIKV, belonging to the Old World Semliki Forest complex, was introduced in 2013 from the South Pacific via the Caribbean to North and South America, and in 2014 from Angola to Brazil, in both cases causing severe outbreaks and local transmission (Leparc-Goffart et al., 2014, Nunes et al., 2015).  RUBV also has a global distribution, but is much less prevalent in developed countries where vaccination programs are in place (Strauss and Strauss 2008).


RUBV infects humans via the respiratory route and typically produces persistent infection in children and adults characterized by fever, sore throat, and rash. Congenital rubella syndrome can cause stillbirth or severe defects in babies born to women infected in the first trimester of pregnancy. By contrast, most alphaviruses are cytopathic to vertebrate cells and cause a short febrile illness that can lead to prolonged arthritis or encephalitis, but are rarely fatal. 

Genus demarcation criteria

Distinctions in virus genome characteristics, transmission dynamics and routes, vector associations, and ecological traits separate the Alphavirus genus from the Rubivirus genus.

Derivation of names 

Alpha: from Greek letter α., originally group A arboviruses.

Rubi: from Latin rubeus, “reddish”.

Toga: from Latin toga, “cloak”.

Phylogenetic relationships

RUBV is genetically distinct from the alphaviruses, with possible amino acid sequence homology for only a short fragment of the non-structural protein ORF (Dominguez et al., 1990). The taxonomic relationship of Rubivirus to Alphavirus is under active assessment because of significant differences in their members' virion structures and because their genome sequences are not monophyletic with respect to those of other positive-sense ssRNA viruses. Among alphaviruses, salmon pancreatic disease virus (SPDV) is the most divergent, with sequence similarity only in parts of the structural and non-structural proteins. However, this may not necessarily mean that this virus is ancestral, but could instead reflect its adaptation to the fish host and a lack of evolutionary constraints for other alphaviruses associated with alternating transmission between a mammalian host and arthropod vector.

An alignment of alphavirus sequences (excluding SPDV), demonstrates a high level of heterogeneity in the hypervariable region (HVR) of the nsP3 gene, the capsid gene, and a few short regions scattered throughout the genome where accurate alignment cannot be made. In addition, a recombination event was involved in the origination of western equine encephalitis, Highlands J and Fort Morgan viruses, where the parental sequences were derived from both the eastern equine encephalitis virus complex (donating non-structural proteins genes and capsid genes) and the Whataroa/Sindbis lineage ancestor (donating the envelope protein genes). A phylogenetic tree based on the conserved regions of envelope genes is shown in Figure 4.Togaviridae.

Figure 4.TogaviridaePhylogenetic tree of representative isolates of all alphavirus species generated from a conserved region of envelope protein gene nucleotide sequences (2184 nt) using the GTR+I+Γ substitution model and Maximum likelihood method. The tree is mid-point rooted. Bootstrap values above 70 generated by 1000 replicates of neighbour-joining tree are indicated next to the main branches. This phylogenetic tree and corresponding sequence alignment are available to download from the Resources page.

There are 3 major clades in the Alphavirus genus tree, all supported by high bootstrap values. The first major clade further diverges into Venezuelan equine encephalitis and eastern equine encephalitis complexes. The second bifurcates into Trocara virus, Eilat virus complex, and western equine encephalitis complex, sequentially. The last one contains Barmah Forest virus, salmon pancreas disease virus, southern elephant seal virus, Ndumu virus, Middleburg virus, and Semliki Forest virus complex. Notably, despite Middleburg virus being phylogenetically located within the Semliki Forest complex, it is considered a distinct antigenic complex. 

Similarity with other taxa

Flaviviruses were previously included in the family Togaviridae due to some biological and virological similarities to the current members of this family. An atomic resolution crystal structure of an alphavirus E1 protein shows a folding pattern related to the E protein of flaviviruses, suggesting homology of at least some genes between these families (Jose et al., 2009). However, a distinct genetic organization and replication led to the separation of these groups. Similarities in replication proteins among members of RNA viruses from several plant families implies that an alphavirus-like superfamily may also exist (Rozanov et al., 1992).