Isolation of west nile and sindbis viruses from mosquitoes collected in the Nile Valley of Egypt during an outbreak of Rift Valley fever

As part of an evaluation of potential vectors of arboviruses during a Rift Valley fever (RVF) outbreak in the Nile Valley of Egypt in August 1993, we collected mosquitoes in villages with known RVF viral activity. Mosquitoes were sorted to species, pooled, and processed for virus isolation both by intracerebral inoculation into suckling mice and by inoculation into cell culture. A total of 33 virus isolates was made from 36,024 mosquitoes. Viruses were initially identified by indirect fluorescent antibody testing and consisted of 30 flaviviruses (all members of the Japanese encephalitis complex, most probably West Nile [WN] virus) and three alphaviruses (all members of western equine encephalitis complex, most probably Sindbis). The identity of selected viruses was confirmed by reverse transcriptase-polymerase chain reaction and sequencing. Culex antennatus (Becker) and Culex perexiguus Theobald accounted for five (17%) and 23 (77%) of the WN virus isolations, respectively. Despite isolation of viruses from 32 pools of mosquitoes (both WN and Sindbis viruses were isolated from a single pool), RVF virus was not isolated from these mosquitoes, even though most of them are known competent vectors collected during an ongoing RVF outbreak. Thus, it should be remembered, that even during a known arbovirus outbreak, other arboviruses may still be circulating and causing disease.

Prevalence of sindbis virus neutralizing antibodies among Swedish passerines indicates that thrushes are the main amplifying hosts

The Sindbis virus uses birds as vertebrate hosts in the summer amplification cycle, and the virus is transmitted by ornithophilic Culex species. Previous field and experimental studies have shown that mainly passerine birds are involved in the amplification. To delineate the pattern of Sindbis virus infections among passerines, we collected and sampled birds for blood at five study sites located in northern, central, and southern Sweden. All study sites were lowland forested wetlands and humid forests. The blood samples were assayed for Sindbis neutralizing antibodies, and we tested if the prevalence of Sindbis antibodies varied in relation to bird characteristics (i.e., species, body-mass, sex, and age), and environmental factors (i.e., year, month, and location). We found that Sindbis virus infections occurred in almost all passerine species sampled, but that the infection prevalence was unequally distributed among species. The fieldfare, the redwing, and the songthrush each had significantly higher prevalence than the average for all species. Large passerine species had higher infection prevalence than small species. The infection was less prevalent in hatching-year birds than in older birds during June and July, but not in August. Males and females had the same infection prevalence. The prevalence of Sindbis antibodies was higher in central than in southern Sweden, which coincided with a higher proportion of fieldfare-redwing-songthrush samples in the central region of the country. Thus, it is possible that regional and annual variations in the prevalence of Sindbis antibodies in Swedish passerine species depend on the number of fieldfares, redwings, and songthrushes available for feeding by vector mosquitoes.

Prolonged arthritis associated with sindbis-related (Pogosta) virus infection

OBJECTIVE

A follow-up study of musculoskeletal symptoms after Pogosta virus infection.

METHODS

Twenty-six patients with earlier serologically confirmed Pogosta disease were examined. Ultrasonography of affected joints was performed in patients who had chronic musculoskeletal symptoms. Serum antibodies against Sindbis virus were determined. The patients were typed for HLA-DR and B27. Efforts were made using the polymerase chain reaction to demonstrate the virus.

RESULTS

Only 50% of the patients were symptomless 2.5 yr after onset of Pogosta disease. Three patients had fibromyalgia, six had occasional arthralgia and two had chronic arthritis.

CONCLUSIONS

The epidemiology of Pogosta disease is changing and practitioners should be better aware of it. Pogosta virus infection may lead to chronic musculoskeletal discomfort and arthritis.

Isolation and complete nucleotide sequence of a Chinese Sindbis-like virus

Infection with alphaviruses is common in the Chinese population. Here we report the isolation of a Sindbis-like virus from a pool of Anopheles mosquitoes collected in Xinjiang, China during an arbovirus survey. This virus, designated XJ-160, rapidly produced cytopathic effects on mosquito and hamster cells. In addition, it was lethal to neonatal mice if inoculated intracerebrally. Serologically, XJ-160 reacted with and was neutralized by an anti-Sindbis antibody. Anti-XJ-160 antibodies were found in several cohorts of Chinese subjects. The complete 11626-base nucleotide sequence of XJ-160 was determined. XJ-160 has diverged significantly from the prototype Sindbis virus, with an 18% difference in nucleotide sequence and an 8.6% difference in amino acids; there are 11 deletions and 2 insertions, involving 99 nucleotides in total. XJ-160 is most closely linked to Kyzylagach virus isolated in Azerbaijan. Both belong to the African/European genetic lineage of Sindbis virus, albeit more distantly related to other members.

Inhibition of virus-induced neuronal apoptosis by Bax

The Bax protein is widely known as a pro-apoptotic Bcl-2 family member that when overexpressed can trigger apoptosis in multiple cell types and is important for the developmental cell death of neurons. However, Bax was found here to be a potent inhibitor of neuronal cell death in mice infected with Sindbis virus. Newborn mice, which are highly susceptible to a fatal infection with neurotropic Sindbis virus, were significantly protected from neuronal apoptosis and fatal disease when infected with a recombinant Sindbis virus encoding Bax. Deletion of the N terminus of Bax, which mimics cleaved Bax, converted Bax into a pro-apoptotic factor in vivo. As mice mature during the first week after birth, they acquire resistance to a fatal Sindbis virus infection. However, Bax-deficient mice remained very sensitive to fatal disease compared with their control littermates, indicating that endogenous Bax functions as a survival factor and contributes to age-dependent resistance to Sindbis virus-induced mortality. The protective effects of Bax were reproduced in cultured hippocampal neurons but not in cultured dorsal root ganglia neurons. These findings indicate that cell-specific factors determine the anti-apoptotic versus pro-apoptotic function of Bax.

Virus reduction in the preparation of intravenous immune globulin: in vitro experiments

BACKGROUND

While immune globulins for intravenous administration (IGIV) have an excellent record with respect to virus safety, concern regarding these preparations has been raised by reports of transmission of hepatitis C virus (HCV) to patients treated with IGIV and the presence of genetic material for HCV in IGIV preparations.

STUDY DESIGN AND METHODS

This in vitro study evaluated the effectiveness of several manufacturing steps, including ethanol precipitation and pasteurization, in reducing HIV and model viruses including encephalomyocarditis (EMC) virus, pseudorabies virus (PRV), bovine viral diarrhea virus (BVDV), Sindbis virus, vaccinia virus, and vesicular stomatitis virus (VSV), as well as HCV RNA, in IGIV.

RESULTS

Ethanol precipitation carried out after pasteurization resulted in virus reductions (log10) of >3.97 for HIV, 1.95 for EMC virus, >5.39 for PRV, and 3.52 for BVDV. Pasteurization inactivated EMC virus by 4.52 log10 and resulted in a log10 reduction of >6.54 for HIV, >5.39 for PRV, >6.64 for BVDV, >7.78 for Sindbis virus, >5.84 for vaccinia virus, and >6.99 for VSV. All viruses except EMC virus were reduced below the limit of detection within 6 hours of the beginning of pasteurization. Cohn processing of Fraction II + III paste and the 4.5-percent alcohol precipitation step prior to pasteurization provided additional virus removal. Studies using the polymerase chain reaction technique found that HCV RNA was detectable in the starting fraction of Cohn Fraction II paste, but not in the final IGIV preparation.

CONCLUSION

These findings strongly support the viral safety of IGIV prepared by this method and show a significant added measure of virus safety associated with pasteurization of this preparation.

Efficient removal of viruses by a novel polyvinylidene fluoride membrane filter

Virus removal by a novel filter (Ultipor VF DV50), comprising three layers of PVDF membrane, has been evaluated by infectivity studies using a range of viruses and conditions. The filter was able to remove at least 6 log of various viruses, i.e. Sindbis and Semliki Forest (40-70 nm), herpes simplex (120-200 nm) and vaccinia (200 x 350 nm), from cell-culture medium or phosphate buffered saline pH 6.8 containing 0.5% albumin. However, the removal of polio virus (25-30 nm) under these conditions was only limited, i.e. about 1 log. This filter is thus effective for removing viruses of about 50 nm or larger. Proteins as large as immunoglobulins (MW 160,000), were able to pass through the filter with recoveries of at least 85%. Due to its ability to remove viruses of medium to large size, this filter shows potential for increasing the safety of biological products where viruses such as hepatitis B, C, herpes and retroviruses are of concern.

Genetic relatedness of Sindbis virus strains from Europe, Middle East, and Africa

The relatedness of 40 strains of Sindbis virus (SIN) from Europe, the Middle East, and Africa was investigated by limited sequencing within the gene encoding the E2 glycoprotein corresponding to amino acid residues 117 to 229 and encompassing one of the major neutralization epitopes. Phylogenetic analyses using distance matrix and parsimonious methods identified two major genetic clusters of western SIN strains, although the variability was less than that of the corresponding region for Venezuelan equine encephalomyelitis virus with a maximum divergence of 12.4% versus 28.5%, respectively. One cluster comprising 19 strains included the HR derivate of the Egypt SIN prototype, AR339, and strains from Israel, Saudi-Arabia, Italy, Slovak Republic, Azerbaijan, as well as three Swedish strains. Another cluster of 17 strains included the Ockelbo virus (OCK) prototype, Edsbyn 5/82, and the majority of SIN strains from northern Europe including strains from Sweden, Norway, and Karelia, as well as two strains from South Africa. A third cluster, supported by the Neighbor joining method, was made up of four strains from South Africa, Uganda, and Cameroon. Residue 212, either Ser or Thr, previously appointed important for the differences in neutralization assays between SIN and Edsbyn 5/82, respectively, correlated with the two major genetic clusters, but was a Thr for two of the three Swedish strains in the SIN prototype cluster, and a Ser in one Swedish and one Karelian strain in the OCK cluster. The finding of strains similar to prototype SIN in Middle Sweden and of strains in South Africa relating to the northern cluster of SIN strains supports the notion of the dispersal of SIN by migrating birds as previously suggested for New World alphaviruses.

An eruption associated with alphavirus infection

Some alphaviruses, e.g. Sindbis, cause an acute febrile illness associated with papular rashes and arthralgia. The diagnosis is usually serological and, hence, the histopathology of the rashes has been poorly elucidated. We report on two patients with rapidly healing eruptions associated with Sindbis virus infection. The histopathology of the rashes showed large, pronounced lymphohistiocytic infiltrates with atypical lymphoid cells around the hair follicles, changes not usually seen in rapidly-healing dermatoses.

[Sindbis viruses of various geographic origin and differentiation of them from Western equine encephalomyelitis viruses using the polymerase chain reaction]

Comparison of Sindbis virus strains isolated in different regions of the world (in Africa, Australia, and Europe, including Russia and its nearest neighbors) in the polymerase chain reaction (PCR) by the primary gene structure of proteins NSP1 and E1 and in the neutralization test showed the greatest similarity between geographically close strains isolated in Northern Europe (KFL, Karelia, 1381 and 1388, Estonia). Sindbis strains AR339 and Babanki isolated in Africa were similar to each other and to strains from Northern Europe by the examined gene sites but different from the Northern variants in the neutralization test. Geographically remote strains F-720 (Armenia and Southern Europe) and Whataroa (New Zealand) were close to Sindbis virus from Africa and Northern Europe by only one of the genes examined (F-720 by NSP1 and Whataroa by E1). PCR was carried out using oligonucleotide primers containing nucleotide sequences identical to genes NSP1 and E1 sites of Sindbis strains HRSP, Okelbo, and KFL, but different from gene sites of other known representatives of alphaviruses by at least 5 positions. PCR analysis showed that the appurtenance of the geographic variants to Sindbis group can be ascertained only after investigating the homology of at least two genes coding for the replicative and structural proteins. Such a procedure of PCR permits the detection of Sindbis viruses of different geographic origin with changes in their primary structure and allows the differentiation between Sindbis viruses and Western equine encephalomyelitis viruses within the serological complex.

Vector-borne viral diseases in Sweden--a short review

Ockelbo disease, caused by a Sindbis-related virus transmitted to man by mosquitoes, was first described in the central part of Sweden in the 1960s as clusters of patients with fever, arthralgia and rash. An average annual rate of 30 cases was recorded in the 1980s but no cases have been diagnosed during the last few years. Nephropathia epidemica (NE) characterized by fever, abdominal pain and renal dysfunction has been known to cause considerable morbidity in Sweden during the last 60 years but the etiologic agent (Puumala virus) was not isolated until 1983. This virus's main reservoir is the bank vole (Clethrionomys glareolus). NE is endemic in the northern two thirds of Sweden where more than a hundred cases are diagnosed each year. Tick-borne encephalitis transmitted by Ixodes ricinus ticks is restricted to the archipelago and Lake M-alaren on the east coast close to Stockholm. Between 30 and 110 cases are diagnosed every year. Inkoo virus, a California encephalitis group virus, has been isolated from mosquitoes in Sweden. The antibody prevalence to Inkoo virus is very high in a normal population, but no disease has as yet been associated with this virus in Sweden. Among the vector-borne virus diseases imported to Sweden, dengue is the most important, with approximately 50 cases recorded every year.

Evidence of vertical transmission of Ross River and Sindbis viruses (Togaviridae: Alphavirus) by mosquitoes (Diptera: Culicidae) in southeastern Australia

Ross River and Sindbis viruses were isolated from Aedes camptorhynchus adults reared from immatures collected from a salt marsh in coastal Victoria, indicating the existence of field vertical transmission. These first isolations of an arbovirus from adult mosquitoes reared from field-collected immatures in Australia indicates one mechanism for arbovirus maintenance in temperate regions.

Characterization of revertants of a Sindbis virus 6K gene mutant that affects proteolytic processing and virus assembly

Alphaviruses of the Togaviridae encode a small hydrophobic polypeptide of 55 amino acids, noted as the 6K protein, that is synthesized as part of a polyprotein containing the sequences of the two major transmembranal viral structural glycoproteins. Mutations, insertions and deletions in the 6K appear to selectively interfere with the final stages of virus assembly and budding, producing aberrant, multi-cored infectious viruses. In addition, some of these mutations were pleiotropic and much more inhibitory to virus formation. One of the latter, a substitution of alanine in the wild-type Sindbis virus 6K gene by arginine, has been studied further and shown to interfere with normal proteolytic processing of the polyprotein. Cells infected with this mutant but not the wild-type virus also displayed viral antigens in nuclear membranes and released fragments of membranes into the cell culture media. A revertant, obtained by enrichment for a faster growing strain, 'suppressed' these defects and genetic mapping showed that the arginine codon had been modified to encode a methionine. However, the sequence of the 6K protein in this revertant was not wild-type and the revertant was still defective in assembly as demonstrated by formation of aberrant particles. A complete restoration of wild-type particle formation for this revertant could be effected by modifying the E2 glycoprotein sequence.

Iodine-mediated inactivation of lipid- and nonlipid-enveloped viruses in human antithrombin III concentrate

Human plasma-derived protein concentrates intended for clinical use must be treated for viral inactivation to ensure patient safety. This study explored the use of liquid iodine for inactivation of several lipid- and nonlipid-enveloped viruses in an antithrombin III (AT-III) concentrate. Iodine at levels of 0.01% to 0.02% caused between 43% and 94% loss of AT-III activity, as well as degradation of AT-III as shown by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and Western blot analysis. However, addition of up to 0.1% human albumin protected the AT-III against both inactivation and fragmentation. At albumin levels sufficient to retain greater than 75% of AT-III activity, greater than 6 logs of sindbis, encephalomyocarditis, and vesicular stomatitis viruses, greater than 4 logs of pseudorabies, and greater than 3 logs of human immunodeficiency virus were inactivated. Except with sindbis virus, this represented complete inactivation of all the viruses spiked into the AT-III concentrate.

[Concentrate of coagulation factor IX: viral validation of the thermoinactivation process]

The coagulation factor IX obtained at the Institute of Haematology and Transfusion in Warsaw was subjected to virological validation. Two lipide enveloped viruses were used: HSV-1 and Sindbis which may be a model for the viruses of B and C types in hepatitis. It was shown that heating of the freeze-drying preparations at 80 degrees C during 72 hrs leads to full inactivation of the added model viruses.

[Viral validation of the bio-globulin production process]

Intravenous human immunoglobulin of domestic production was subjected to validation studies. Tests were performed in the system of model viruses for pathogenic factors of the B and C types of hepatitis. For simulated immunoglobulin infections two lipid-enveloped viruses were chosen: the first DNA virus pseudorabies PR-75 (model for HBV) and the second-RNA virus Sindbis (model for HCV). The survival of viruses in particular stages of bio-globulin production was checked, showing that pepsin digestion led to full inactivation of both viruses. The total reduction of infective titers of model viruses during the whole process exceeded 10 logs. Thus we can conclude that the process of bio-globulin manufacturing ensures the elimination and inactivation of lipid-enveloped viruses.

A cell line that expresses a reporter gene in response to infection by Sindbis virus: a prototype for detection of positive strand RNA viruses

We describe a stably transformed cell line (BHKSINLuc2) that contains a defective Sindbis virus genome under the control of a Rous sarcoma virus promoter and the luciferase gene downstream of the viral subgenomic RNA promoter. This cell line expresses high levels of luciferase activity following infection with Sindbis virus and provides a sensitive assay for titering variants of Sindbis virus that lack the structural protein genes, in particular, Sindbis virus replicons that express heterologous proteins. Cell lines such as this may be of value for detection of positive-strand RNA viruses.

Mutations in the nsP1 coding sequence of Sindbis virus which restrict viral replication in secondary cultures of chick embryo fibroblasts prepared from aged primary cultures

SVMPA, a mutant of Sindbis virus (SV), which is able to replicate in Aedes albopictus cells treated with mycophenolic acid (MPA) or ribavirin, also has a host range phenotype. This phenotype is most clearly demonstrated by means of efficiency of plaquing (EOP) assays on secondary chick embryo fibroblasts (CEF) prepared from aged primary CEF. For example, in one experiment in which standard SV (SVSTD) had a relative EOP (EOP on primary CEF divided by EOP on secondary CEF) of 1.6 the corresponding value for SVMPA was 2340. The host restriction of SVMPA was also seen with similarly prepared secondary cultures of duck embryo fibroblasts, but not with established lines of quail cells. The finding that the accumulation of viral RNA was much lower in SVMPA-infected secondary CEF than in SVSTD-infected CEF indicated that the replication of SVMPA in these cultures was blocked at an early step. Revertants of SVMPA were isolated which were no longer host-restricted but had retained their resistance to MPA. Of the three mutations [nucleotide (nt) 120, 127, and 963] in the nsP1 coding sequence of SVMPA, which lead to amino acid changes, these revertants had retained the nt 127 and nt 963 mutations but had lost the nt 120 mutation. This, along with earlier findings, indicated that only the nt 127 and nt 963 mutations are required for resistance to MPA. This result also associated the nt 120 mutation with the host restriction phenotype. In other experiments derivatives of pToto1101 (a plasmid from which infectious Sindbis virus RNA can be transcribed) were constructed by site-directed mutagenesis and used to test the effect of specific mutations on the viral phenotype. Although we were unable to obtain viable virus with the nt 120 mutation alone, virus with the nt 120 and nt 127 mutations was viable and host-restricted. We suggest that the nt 120 mutation by itself is lethal and that the nt 127 mutation suppresses the lethal effect of the nt 120 mutation. The SVMPA mutation at nt 120, which is associated with the host range phenotype, changes Gln21 of nsP1 to Lys. When a more conservative change was engineered, i.e., to Asn, the virus was not host-restricted. Although the reason for the restriction of SVMPA replication in secondary CEF is not known, some possible explanations are discussed.

Size exclusion removal of model mammalian viruses using a unique membrane system, Part II: Module qualification and process simulation

The performance of a fabricated device may be influenced by the characteristics of the fluid management within the device and reproducibility with which the device is manufactured. The performance of the Viresolve/70 membrane is not diminished when incorporated into a fabricated module. The Viresolve/70 fabricated modules are shown to reproducibly retain viruses via a sieving mechanism, independent of virus type or character, in excellent agreement with the base membrane retention coefficients reported previously. The retention coefficients measured for the Viresolve/70 modules are shown to increase with increased recirculation flow rate within the module. Mammalian virus spiked protein solutions processed through the Viresolve/70 system show that mammalian viruses can be removed from solution in accordance with the apparent membrane retention coefficients. The retained virus is recoverable on the upstream side of the membrane. Process clearance factors for murine leukemia virus is in excess of 6.7 LRV and that of human immunodeficiency virus I is 8.5 LRV.

Size exclusion removal of model mammalian viruses using a unique membrane system, Part I: Membrane qualification

We have previously described a new class of composite membrane that has the capability to efficiently remove particles, including viruses, from protein solution. The qualification of this membrane requires that it reproducibly and predictably remove model mammalian viruses. Using the Viresolve/70 membrane, the mammalian viruses polio, Simian virus-40, Sindbis, reovirus type 3 and Rauscher murine leukemia virus are shown to be reproducibly removed via a sieving mechanism. Mammalian virus retention increases constantly with virus diameter independent of virus class or type, increasing from 3.5 logs with polio virus to greater than 6.8 logs with murine leukemia virus. Consistent with a sieving mechanism, mammalian virus retention with the Viresolve/70 membrane is independent of virus concentration. These results are shown both in the presence and absence of protein in solution. The presence of protein in solution is shown to increase the virus retention coefficient of each virus above that measured in phosphate buffered saline. The model virus retention is shown to be well predicted by hard particle retention reported previously for this membrane. In addition, the hard particle retention is shown to predict the worse case performance expected of the membrane in the presence of protein.

Insect-transmitted vertebrate viruses: alphatogaviruses

Alphatogaviruses, of which Sindbis virus (SV) is the prototype, replicate to high titer in the laboratory both in mosquito cells and in vertebrate cells. By studying the replication of SV in mosquito cells as well as in vertebrate cells, we were able to obtain several viral mutants which have novel phenotypes and have contributed to our basic knowledge of this virus family. These include three host range mutants: SVAP15/21 which replicates normally in mosquito cells but is restricted in vertebrate cells and SVCL35 and SVCL58, which are restricted in mosquito cells but replicate normally in vertebrate cells. As well, two other mutants are described here: SVLM21, which can replicate in methionine-starved mosquito cells and SVMPA, which can replicate in mosquito cells treated with mycophenolic acid or ribavirin. The causal mutations of both SVLM21 and SVMPA are within the sequence encoding the nonstructural protein nsPl; these and other findings have enabled us to associate the capping and methylation of the viral mRNAs with the nsPl protein. Our work serves to emphasize that it is both worthwhile and important to study the replication of arthropod-borne viruses in cells derived from the arthropod host as well as in cells derived from the vertebrate host.

The distribution of Sindbis virus proteins in mosquito cells as determined by immunofluorescence and immunoelectron microscopy

Two Aedes albopictus (mosquito) subclones, C7-10 and C6/36, were examined by immunofluorescence and immunoelectron microscopy for the distribution of Sindbis virus structural and non-structural proteins. Both the viral glycoproteins, E1 and E2, and the non-structural proteins, nsP1 and nsP2, were found within vesicles and electron-dense, amorphous matrices associated with Sindbis virus infection. The labelling patterns indicated that both replication of viral RNA and production of virus particles were localized within the same structures in the infected cell. The data support previous reports that alphavirus infection is contained within specific structures in the cytoplasm and provide additional evidence that the C6/36 and C7-10 subclones may represent different tissue types in the adult insect.

Persistence of viral RNA in mouse brains after recovery from acute alphavirus encephalitis

Little is known about the relationship between recovery from acute viral encephalitis and the clearance of viral genetic material from the central nervous system. In a mouse model of Sindbis virus encephalitis, we have previously shown that clearance of infectious virus is mediated by antibody-induced restriction of viral gene expression rather than by cytotoxic destruction of virally infected cells. To explore whether Sindbis virus genomes persist in mouse brain after the clearance of infectious virus, we used reverse transcriptase-polymerase chain reaction amplification methods to detect Sindbis virus RNA in brain samples from immunocompetent BALB/c and antibody-treated immunodeficient scid/CB17 mice. RNA sequences from both the nonstructural region (NSP1 gene) and structural regions (E2 gene) of Sindbis virus were detected in the brains of all BALB/c and antibody-treated scid mice examined at 1, 2, and 3 months after infection. Additional BALB/c mouse brains were also positive at 8, 12, and 17 months after infection. To determine whether persistent RNA was capable of resuming unrestricted replication in the absence of the continuous presence of antiviral antibodies, viral titers were measured in the brains of scid mice at 1, 2, 3, and 6 months after antibody treatment. Viral reactivation was seen in scid mice treated with hyperimmune serum or a low dose of monoclonal antibody to the E2 envelope glycoprotein, but not in mice treated with a high dose of monoclonal antibody to E2. Replication of infectious virus isolated from scid mouse brain could be restricted by repeat treatment with immune serum, indicating that viral reactivation is not due to antibody-escape mutations. These results demonstrate that Sindbis virus can persist long term in a nonproductive form in mouse brain and suggest that the humoral immune response plays an important role in preventing viral reactivation.