Evolutionary relationships and systematics of the alphaviruses.

Partial E1 envelope glycoprotein gene sequences and complete structural polyprotein sequences were used to compare divergence and construct phylogenetic trees for the genus Alphavirus. Tree topologies indicated that the mosquito-borne alphaviruses could have arisen in either the Old or the New World, with at least two transoceanic introductions to account for their current distribution. The time frame for alphavirus diversification could not be estimated because maximum-likelihood analyses indicated that the nucleotide substitution rate varies considerably across sites within the genome. While most trees showed evolutionary relationships consistent with current antigenic complexes and species, several changes to the current classification are proposed. The recently identified fish alphaviruses salmon pancreas disease virus and sleeping disease virus appear to be variants or subtypes of a new alphavirus species. Southern elephant seal virus is also a new alphavirus distantly related to all of the others analyzed. Tonate virus and Venezuelan equine encephalitis virus strain 78V3531 also appear to be distinct alphavirus species based on genetic, antigenic, and ecological criteria. Trocara virus, isolated from mosquitoes in Brazil and Peru, also represents a new species and probably a new alphavirus complex.

Suppressor mutations that allow sindbis virus RNA polymerase to function with nonaromatic amino acids at the N-terminus: evidence for interaction between nsP1 and nsP4 in minus-strand RNA synthesis.

The alphavirus RNA polymerase, nsP4, invariably has a Tyr residue at the N-terminus. Previously we reported that the N-terminal Tyr residue of nsP4 of Sindbis virus, the type species of the genus Alphavirus, can be substituted with Phe, Trp, or His without altering the wild-type phenotype in cultured cells but that other substitutions tested, except for Met, were lethal or quasilethal. Here we report the identification of two suppressor mutations in nsP4 (Glu-191 to Leu and Glu-315 to Gly, Val, or Lys) and one in nsP1 (Thr-349 to Lys) that allow nsP4 with nonaromatic amino acids at the N-terminus to function at 30 degrees C. The suppressor mutation at nsP4 Glu-315 occurred most frequently. All three suppressor mutations suppressed the effects of Ala, Arg, or Leu at the N-terminus of nsP4 with almost equal efficiency and thus the effect of the suppressing mutation is independent of the nsP4 N-terminal residue. Reconstructed mutants containing nsP1-T349K or nsP4-E315G combined with Ala-nsP4 had a defect in minus-strand RNA synthesis at 40 degrees C. A double mutant containing nsP4-Q191L combined with Ala-nsP4 was unstable and could not be tested for RNA synthesis because it reverted to temperature-independence too rapidly. Combinations of nsP1-T349K or nsP4-E315G with Leu, Arg, His, or any aromatic amino acid at the N-terminus of nsP4 also made the mutant viruses temperature sensitive. The results from this study and from a previous report on the shutoff of minus-strand RNA synthesis at 40 degrees C with the nsP1-A348T mutation in ts11 suggests that the N-terminus nsP4 interacts with nsP1 during initiation of minus-strand RNA synthesis.

Adaptive mutations in Sindbis virus E2 and Ross River virus E1 that allow efficient budding of chimeric viruses.

Alphavirus glycoproteins E2 and E1 form a heterodimer that is required for virus assembly. We have studied adaptive mutations in E2 of Sindbis virus (SIN) and E1 of Ross River virus (RR) that allow these two glycoproteins to interact more efficiently in a chimeric virus that has SIN E2 but RR E1. These mutations include K129E, K131E, and V237F in SIN E2 and S310F and C433R in RR E1. Although RR E1 and SIN E2 will form a chimeric heterodimer, the chimeric virus is almost nonviable, producing about 10(-7) as much virus as SIN at 24 h and 10(-5) as much after 48 h. Chimeras containing one adaptive change produced 3 to 20 times more virus than did the parental chimera, whereas chimeras with two changes produced 10 to 100 times more virus and chimeras containing three mutations produced yields that were 180 to 250 times better. None of the mutations had significant effects upon the parental wild-type viruses, however. Passage of the triple variants eight or nine times resulted in variants that produced virus rapidly and were capable of producing >10(8) PFU/ml of culture fluid within 24 h. These further-adapted variants possessed one or two additional mutations, including E2-V116K, E2-S110N, or E1-T65S. The RR E1-C433R mutation was studied in more detail. This Cys is located in the putative transmembrane domain of E1 and was shown to be palmitoylated. Mutation to Arg-433 resulted in loss of palmitoylation of E1. The positively charged arginine residue within the putative transmembrane domain of E1 would be expected to alter the conformation of this domain. These results suggest that interactions within the transmembrane region are important for the assembly of the E1/E2 heterodimer, as are regions of the ectodomains possibly identified by the locations of adaptive mutations in these regions. Further, the finding that four or five changes in the chimera allow virus production that approaches the levels seen with the parental SIN and exceeds that of the parental RR illustrates that the structure and function of SIN and RR E1s have been conserved during the 50% divergence in sequence that has occurred.

Molecular genetic study of the interaction of Sindbis virus E2 with Ross River virus E1 for virus budding.

Glycoprotein PE2 of Sindbis virus will form a heterodimer with glycoprotein E1 of Ross River virus that is cleaved to an E2/E1 heterodimer and transported to the cell plasma membrane, but this chimeric heterodimer fails to interact with Sindbis virus nucleocapsids, and very little budding to produce mature virus occurs upon infection with chimeric viruses. We have isolated in both Sindbis virus E2 and in Ross River virus E1 a series of suppressing mutations that adapt these two proteins to one another and allow increased levels of chimeric virus production. Two adaptive E1 changes in an ectodomain immediately adjacent to the membrane anchor and five adaptive E2 changes in a 12-residue ectodomain centered on Asp-242 have been identified. One change in Ross River virus E1 (Gln-411-->Leu) and one change in Sindbis virus E2 (Asp-248-->Tyr) were investigated in detail. Each change individually leads to about a 10-fold increase in virus production, and combined the two changes lead to a 100-fold increase in virus. During passage of a chimeric virus containing Ross River virus E1 and Sindbis virus E2, the E2 change was first selected, followed by the E1 change. Heterodimers containing these two adaptive mutations have a demonstrably increased degree of interaction with Sindbis virus nucleocapsids. In the parental chimera, no interaction between heterodimers and capsids was visible at the plasma membrane in electron microscopic studies, whereas alignment of nucleocapsids along the plasma membrane, indicating interaction of heterodimers with nucleocapsids, was readily seen in the adapted chimera. The significance of these findings in light of our current understanding of alphavirus budding is discussed.

Superinfection exclusion of alphaviruses in three mosquito cell lines persistently infected with Sindbis virus.

Three Aedes albopictus (mosquito) cell lines persistently infected with Sindbis virus excluded the replication of both homologous (various strains of Sindbis) and heterologous (Aura, Semliki Forest, and Ross River) alphaviruses. In contrast, an unrelated flavivirus, yellow fever virus, replicated equally well in uninfected and persistently infected cells of each line. Sindbis virus and Semliki Forest virus are among the most distantly related alphaviruses, and our results thus indicate that mosquito cells persistently infected with Sindbis virus are broadly able to exclude other alphaviruses but that exclusion is restricted to members of the alphavirus genus. Superinfection exclusion occurred to the same extent in three biologically distinct cell clones, indicating that the expression of superinfection exclusion is conserved among A. albopictus cell types. Superinfection of persistently infected C7-10 cells, which show a severe cytopathic effect during primary Sindbis virus infection, by homologous virus does not produce cytopathology, consistent with the idea that cytopathology requires significant levels of viral replication. A possible model for the molecular basis of superinfection exclusion, which suggests a central role for the alphavirus trans-acting protease that processes the nonstructural proteins, is discussed in light of these results.

Recombinational history and molecular evolution of western equine encephalomyelitis complex alphaviruses.

Western equine encephalomyelitis (WEE) virus (Togaviridae: Alphavirus) was shown previously to have arisen by recombination between eastern equine encephalomyelitis (EEE)- and Sindbis-like viruses (C. S. Hahn, S. Lustig, E. G. Strauss, and J. H. Strauss, Proc. Natl. Acad. Sci. USA 85:5997-6001, 1988). We have now examined the recombinational history and evolution of all viruses belonging to the WEE antigenic complex, including the Buggy Creek, Fort Morgan, Highlands J, Sindbis, Babanki, Ockelbo, Kyzylagach, Whataroa, and Aura viruses, using nucleotide sequences derived from representative strains. Two regions of the genome were examined: sequences of 477 nucleotides from the C terminus of the E1 envelope glycoprotein gene which in WEE virus was derived from the Sindbis-like virus parent, and 517 nucleotide sequences at the C terminus of the nsP4 gene which in WEE virus was derived from the EEE-like virus parent. Trees based on the E1 region indicated that all members of the WEE virus complex comprise a monophyletic group. Most closely related to WEE viruses are other New World members of the complex: the Highlands J, Buggy Creek, and Fort Morgan viruses. More distantly related WEE complex viruses included the Old World Sindbis, Babanki, Ockelbo, Kyzylagach, and Whataroa viruses, as well as the New World Aura virus. Detailed analyses of 38 strains of WEE virus revealed at least 4 major lineages; two were represented by isolates from Argentina, one was from Brazil, and a fourth contained isolates from many locations in South and North America as well as Cuba. Trees based on the nsP4 gene indicated that all New World WEE complex viruses except Aura virus are recombinants derived from EEE- and Sindbis-like virus ancestors. In contrast, the Old World members of the WEE complex, as well as Aura virus, did not appear to have recombinant genomes. Using an evolutionary rate estimate (2.8 x 10(-4) substitutions per nucleotide per year) obtained from E1-3' sequences of WEE viruses, we estimated that the recombination event occurred in the New World 1,300 to 1,900 years ago. This suggests that the alphaviruses originated in the New World a few thousand years ago.

Interactions between PE2, E1, and 6K required for assembly of alphaviruses studied with chimeric viruses.

During the assembly of alphaviruses, a preassembled nucleocapsid buds through the cell plasma membrane to acquire an envelope containing two virally encoded glycoproteins, E2 and E1. Using two chimeric viruses, we have studied interactions between E1, E2, and a viral peptide called 6K, which are required for budding. A chimeric Sindbis virus (SIN) in which the 6K gene had been replaced with that from Ross River virus (RR) produced wild-type levels of nucleocapsids and abundant PE2/E1 heterodimers that were processed and transported to the cell surface. However, only about 10% as much chimeric virus as wild-type virus was assembled, demonstrating that there is a sequence-specific interaction between 6K and the glycoproteins required for efficient virus assembly. In addition, the conformation of E1 in the E2/E1 heterodimer on the cell surface was different for the chimeric virus from that for the wild type, suggesting that one function of 6K is to promote proper folding of E1 in the heterodimer. A second chimeric SIN, in which both the 6K and E1 genes, as well as the 3' nontranslated region, were replaced with the corresponding regions of RR also resulted in the production of large numbers of intracellular nucleocapsids and of PE2/E1 heterodimers that were cleaved and transported to the cell surface. Budding of this chimera was severely impaired, however, and the yield of the chimera was only approximately 10(-7) of the SIN yield in a parallel infection. The conformation of the SIN E2/RR E1 heterodimer on the cell surface was different from that of the SIN E2/SIN E1 heterodimer, and no interaction between viral glycoproteins and nucleocapsids at the cell plasma membrane could be detected in the electron microscope. We suggest that proper folding of the E2/E1 heterodimer must occur before the E2 tail is positioned properly in the cytoplasm for budding and before heterodimer trimerization can occur to drive virus budding.

Characterization of the rubella virus nonstructural protease domain and its cleavage site.

The region of the rubella virus nonstructural open reading frame that contains the papain-like cysteine protease domain and its cleavage site was expressed with a Sindbis virus vector. Cys-1151 has previously been shown to be required for the activity of the protease (L. D. Marr, C.-Y. Wang, and T. K Frey, Virology 198:586-592, 1994). Here we show that His-1272 is also necessary for protease activity, consistent with the active site of the enzyme being composed of a catalytic dyad consisting of Cys-1151 and His-1272. By means of radiochemical amino acid sequencing, the site in the polyprotein cleaved by the nonstructural protease was found to follow Gly-1300 in the sequence Gly-1299-Gly-1300-Gly-1301. Mutagenesis studies demonstrated that change of Gly-1300 to alanine or valine abrogated cleavage. In contrast, Gly-1299 and Gly-1301 could be changed to alanine with retention of cleavage, but a change to valine abrogated cleavage. Coexpression of a construct that contains a cleavage site mutation (to serve as a protease) together with a construct that contains a protease mutation (to serve as a substrate) failed to reveal trans cleavage. Coexpression of wild-type constructs with protease-mutant constructs also failed to reveal trans cleavage, even after extended in vitro incubation following lysis. These results indicate that the protease functions only in cis, at least under the conditions tested.

Budding of alphaviruses.

The icosahedral structures of alphaviruses and of the external shell of the viral nucleocapsid have been defined to very high resolutions, revealing details of the interactions between the glycoproteins to form trimeric spikes and the nucleocapsid. The structural studies complement biochemical and molecular genetic studies showing that a sequence-specific interaction between the cytoplasmic domains of the glycoproteins and the nucleocapsid drives budding.

Aura virus is a New World representative of Sindbis-like viruses.

Aura virus is an alphavirus present in Brazil and Argentina that is serologically related to Sindbis virus (present throughout the Old World) and to Western equine encephalitis (WEE) virus (present in the Americas). We have previously shown that WEE is a recombinant virus whose glycoproteins and part of whose 3' nontranslated region (NTR) are derived from a Sindbis-like virus, but the remainder of whose genome is derived from Eastern equine encephalitis (EEE) virus. We show here that Aura virus is a Sindbis-like virus that shares considerable organizational and sequence identity with Sindbis virus. Certain nucleotide sequence elements present in Aura RNA that are believed to function as promoters are almost identical to their Sindbis counterparts, repeated elements in the 3' nontranslated region are shared with Sindbis virus, and important antigenic epitopes are conserved between the two viruses. Despite their close relationship, the two viruses have diverged significantly, sharing 73% amino acid sequence identity in the nonstructural proteins and 62% identity in the structural proteins. This is about the same as the identities between EEE and Venezuelan equine encephalitis virus, whose promoter elements, 3' NTRs, and antigenic epitopes have diverged more radically, such that these two viruses are considered to belong to different subgroups. Importantly, the glycoproteins of WEE are more closely related to those of Sindbis than to those of Aura virus. From this we propose that an ancestral Sindbis-like virus present in the Americas (probably South America) diverged 1000-2000 years ago into a lineage that gave rise to Aura virus and a lineage that gave rise to Sindbis virus and to the Sindbis-like parent of WEE. At some time after this divergence, a Sindbis-like virus belonging to the latter lineage was transferred to the Old World where it gave rise to Sindbis viruses distributed throughout the Old World, and in a separate event a Sindbis-like virus belonging to the same lineage underwent recombination with EEE to give rise to WEE.

Aura alphavirus subgenomic RNA is packaged into virions of two sizes.

The alphavirus genome is 11.8 kb in size. During infection, a 4.2-kb subgenomic RNA is also produced. Most alphaviruses package only the genomic RNA into virions, which are enveloped particles with icosahedral symmetry, having a triangulation number (T) = 4. Aura virus, however, packages both the genomic RNA and the subgenomic RNA into virions. The genomic RNA is primarily packaged into a virion that has a diameter of 72 nm and which appears to be identical to the virions produced by other alphaviruses. The subgenomic RNA is packaged into two major, regular particles with diameters of 72 and 62 nm. The 72-nm-diameter particle appears to be identical in construction to virions containing genomic RNA. The 62-nm-diameter particle probably has T = 3. The large and small Aura virions can be partially separated in sucrose gradients. In addition to these two major classes of particles, there are other particles produced that appear to arise from abortive assembly. From these results and from previous studies of alphavirus assembly, we suggest that during assembly of alphavirus nucleocapsids in the infected cell there is a specific initiation event followed by recruitment of additional capsid subunits into the complex, that the triangulation number of the complex is not predetermined but depends upon the size of the RNA and interactions that occur during assembly, and that budding of assembled nucleocapsids results in the acquisition of an envelope containing glycoproteins arranged in a manner determined by the nucleocapsid.

The alphaviruses: gene expression, replication, and evolution.

The alphaviruses are a genus of 26 enveloped viruses that cause disease in humans and domestic animals. Mosquitoes or other hematophagous arthropods serve as vectors for these viruses. The complete sequences of the +/- 11.7-kb plus-strand RNA genomes of eight alphaviruses have been determined, and partial sequences are known for several others; this has made possible evolutionary comparisons between different alphaviruses as well as comparisons of this group of viruses with other animal and plant viruses. Full-length cDNA clones from which infectious RNA can be recovered have been constructed for four alphaviruses; these clones have facilitated many molecular genetic studies as well as the development of these viruses as expression vectors. From these and studies involving biochemical approaches, many details of the replication cycle of the alphaviruses are known. The interactions of the viruses with host cells and host organisms have been exclusively studied, and the molecular basis of virulence and recovery from viral infection have been addressed in a large number of recent papers. The structure of the viruses has been determined to about 2.5 nm, making them the best-characterized enveloped virus to date. Because of the wealth of data that has appeared, these viruses represent a well-characterized system that tell us much about the evolution of RNA viruses, their replication, and their interactions with their hosts. This review summarizes our current knowledge of this group of viruses.

Polypeptide requirements for assembly of functional Sindbis virus replication complexes: a model for the temporal regulation of minus- and plus-strand RNA synthesis.

Proteolytic processing of the Sindbis virus non-structural polyproteins (P123 and P1234) and synthesis of minus- and plus-strand RNAs are highly regulated during virus infection. Although their precise roles have not been defined, these polyproteins, processing intermediates or mature cleavage products (nsP1-4) are believed to be essential components of viral replication and transcription complexes. In this study, we have shown that nsP4 can function as the polymerase for both minus- and plus-strand RNA synthesis. Mutations inactivating the nsP2 proteinase, resulting in uncleaved P123, led to enhanced accumulation of minus-strand RNAs and reduced accumulation of genomic and subgenomic plus-strand RNAs. In contrast, no RNA synthesis was observed with a mutation which increased the efficiency of P123 processing. Inclusion of this mutation in a P123 polyprotein with cleavage sites 1/2 and 2/3 blocked allowed synthesis of both minus- and plus-strand RNAs. We conclude that nsP4 and uncleaved P123 normally function as the minus-strand replication complex, and propose that processing of P123 switches the template preference of the complex to minus-strands, resulting in efficient synthesis of plus-strand genomic and subgenomic RNAs and shut-off of minus-strand RNA synthesis.

Nucleocapsid-glycoprotein interactions required for assembly of alphaviruses.

We have studied interactions between nucleocapsids and glycoproteins required for budding of alphaviruses, using Ross River virus-Sindbis virus chimeras in which the nucleocapsid protein is derived from one virus and the envelope glycoproteins are derived from the second virus. A virus containing the Ross River virus genome in which the capsid protein had been replaced with that from Sindbis virus was almost nonviable. Nucleocapsids formed in normal numbers in the infected cell, but very little virus was released from the cell. There are 11 amino acid differences between Ross River virus and Sindbis virus in their 33-residue E2 cytoplasmic domains. Site-specific mutagenesis was used to change 9 of these 11 amino acids in the chimera from the Ross River virus to the Sindbis virus sequence in an attempt to adapt the E2 of the chimera to the nucleocapsid. The resulting mutant chimera grew 4 orders of magnitude better than the parental chimeric virus. This finding provides direct evidence for a sequence-specific interaction between the nucleocapsid and the E2 cytoplasmic domain during virus budding. The mutated chimeric virus readily gave rise to large-plaque variants that grew almost as well as Ross River virus, suggesting that additional single amino acid substitutions in the structural proteins can further enhance the interactions between the disparate capsid and the glycoproteins. Unexpectedly, change of E2 residue 394 from lysine (Ross River virus) to glutamic acid (Sindbis virus) was deleterious for the chimera, suggesting that in addition to its role in nucleocapsid-E2 interactions, the N-terminal part of the E2 cytoplasmic domain may be involved in glycoprotein-glycoprotein interactions required to assemble the glycoprotein spikes. The reciprocal chimera, Sindbis virus containing the Ross River virus capsid, also grew poorly. Suppressor mutations arose readily in this chimera, producing a virus that grew moderately well and that formed larger plaques.

Host-cell receptors for Sindbis virus.

Sindbis virus has a very wide host range, infecting many species of mosquitoes and other hematophagous insects and infecting many species of higher vertebrates. We have used two approaches to study host cell receptors used by Sindbis virus to enter cells. Anti-idiotype antibodies to neutralizing antibodies directed against glycoprotein E2 of the virus identified a 63-kDa protein as a putative receptor in chicken cells. In a second approach, monoclonal antibodies identified a 67 kDa protein, believed to be a high affinity laminin receptor, as a putative receptor in mammalian cells and in mosquito cells. We conclude that the virus attains its very wide host range by two mechanisms. In one mechanism, the virus is able to use more than one protein as a receptor. In a second mechanism, the virus utilizes proteins as receptors that are highly conserved across the animal kingdom.

Subgenomic mRNA of Aura alphavirus is packaged into virions.

Purified virions of Aura virus, a South American alphavirus related to Sindbis virus, were found to contain two RNA species, one of 12 kb and the other of 4.2 kb. Northern (RNA) blot analysis, primer extension analysis, and limited sequencing showed that the 12-kb RNA was the viral genomic RNA, whereas the 4.2-kb RNA present in virus preparations was identical to the 26S subgenomic RNA present in infected cells. The subgenomic RNA is the messenger for translation of the viral structural proteins, and its synthesis is absolutely required for replication of the virus. Although 26S RNA is present in the cytosol of all cells infected by alphaviruses, this is the first report of incorporation of the subgenomic RNA into alphavirus particles. Packaging of the Aura virus subgenomic mRNA occurred following infection of mosquito (Aedes albopictus C6/36), hamster (BHK-21), or monkey (Vero) cells. Quantitation of the amounts of genomic and subgenomic RNA both in virions and in infected cells showed that the ratio of genomic to subgenomic RNA was 3- to 10-fold higher in Aura virions than in infected cells. Thus, although the subgenomic RNA is packaged efficiently, the genomic RNA has a selective advantage during packaging. In contrast, in parallel experiments with Sindbis virus, packaging of subgenomic RNA was not detectable. We also found that subgenomic RNA was present in about threefold-greater amounts relative to genomic RNA in cells infected by Aura virus than in cells infected by Sindbis virus. Packaging of the Aura virus subgenomic RNA, but not those of other alphaviruses, suggests that Aura virus 26S RNA contains a packaging signal for incorporation into virions. The importance of the packaging of this RNA into virions in the natural history of the virus remains to be determined.

Viral determinants of age-dependent virulence of Sindbis virus for mice.

Many alphaviruses cause more severe disease in young animals than in older animals. The age-dependent resistance to severe disease is determined primarily by maturation of the host, but strains of virus can be selected that overcome the increased resistance of mature animals. Sindbis virus (SV) strain AR339 causes fatal encephalitis in newborn mice and nonfatal encephalitis in weanling mice, whereas NSV, a neuroadapted strain of SV, causes fatal encephalitis in weanling as well as newborn mice. We have previously shown that the E2 glycoprotein of NSV contained His-55, whereas AR339 E2 had Gln-55 (S. Lustig, A. C. Jackson, C. S. Hahn, D. E. Griffin, E. G. Strauss, and J. H. Strauss, J. Virol. 62:2329-2336, 1988) and that SV with E2 containing Gly-172 was more virulent for newborn mice than SV with E2 containing Arg-172 (P. C. Tucker and D. E. Griffin, J. Virol. 65:1551-1557, 1991). Here we tested the virulence for both newborn and older mice of SV containing a number of different amino acids at E2 position 55 (His, Gln, Lys, Arg, Glu, Gly) in combination with both Gly-172 and Arg-172. All the viruses were virulent for newborn mice, but the residues at both 55 and 172 influenced the virulence of the virus, and there were differences in virulence observed among the various viruses. However, only viruses with His-55 were fully virulent for 14-day-old mice, and this virulence was independent of the residue at position 172. Virus with Lys-55 was virulent for 7-day-old mice, although slightly attenuated relative to His-55. Viruses with His-55 grew more rapidly and to higher titer in the brains of 7- and 14-day-old mice, in N18 neuroblastoma cells, and in BHK cells. Our data suggest that His-55 is important for neurovirulence in older mice and acts by increasing the efficiency of virus replication.

Identification of the active site residues in the nsP2 proteinase of Sindbis virus.

The nonstructural polyproteins of Sindbis virus are processed by a virus-encoded proteinase which is located in the C-terminal domain of nsP2. Here we have performed a mutagenic analysis to identify the active site residues of this proteinase. Substitution of other amino acids for either Cys-481 or His-558 completely abolished proteolytic processing of Sindbis virus polyproteins in vitro. Substitutions within this domain for a second cysteine conserved among alphaviruses, for four other conserved histidines, or for a conserved serine did not affect the activity of the enzyme. These results suggest that nsP2 is a papain-like proteinase whose catalytic dyad is composed of Cys-481 and His-558. Since an asparagine residue has been implicated in the active site of papain, we changed the four conserved asparagine residues in the C-terminal half of nsP2 and found that all could be substituted without total loss of activity. Among papain-like proteinases, the residue following the catalytic histidine is alanine or glycine in the plant and animal enzymes, and the presence of Trp-559 in alphaviruses is unusual. A mutant enzyme containing Ala-559 was completely inactive, implying that Trp-559 is essential for a functional proteinase. All of these mutations were introduced into a full-length clone of Sindbis virus from which infectious RNA could be transcribed in vitro, and the effects of these changes on viability were tested. In all cases it was found that mutations which abolished proteolytic activity were lethal, whether or not these mutations were in the catalytic residues, indicating that proteolysis of the nonstructural polyprotein is essential for Sindbis replication.