coronavirus attenuated Elaine Anjeanette School for revision Ksiazek et al 2003 a positive stranded enveloped RNA virus and the 29 7 kb genome sequence was determined viruses mouse hepatitis virus MHV transmissible gastro enteritis virus TGEV and infectious bronchitis virus IBV generates protective immunity against these viruses 1992 Ignjatovic Virology 340 2005 174 In the fall of 2002 an emerging human disease termed SARS severe acute respiratory syndrome drew global attention The disease was characterized as an atypical pneumonia accompanied by high fever According the World Health Organization there were 8096 cases of SARS 21 of which were in health care workers Approximately 10 of the infected individuals died as a result of the illness and the death rate exceeded 50 for those over 60 years old http www who int csr sars en WHOconsensus pdf and Peiris et al 2004 The causative agent was identified as a coronavirus Drosten et al 2003 rapidly Like other coronaviruses the SARS coronavirus SARS CoV has 6 major open reading frames two encoding the polymerase and four encoding the structural proteins spike S membrane envelope and nucleocapsid Marra et al 2003 Rota et al 2003 The SARS S protein is a type I transmembrane glycoprotein It is responsible for viral binding to host cellular receptors identified as angiotensin converting enzyme 2 Li et al 2003 and CD209L Jeffers et al 2004 followed by viral fusion to host cells Bosch et al 2004 Liu et al 2004 Yuan et al 2004 Immunization with the S proteins of other corona Although the recent SARS coronavirus SARS CoV that appeared in 2002 has now been contained the possibility of re emergence of SARS CoV remains Due to the threat of re emergence the overall fatality rate of 10 and the rapid dispersion of the virus via international travel viable vaccine candidates providing protection from SARS are clearly needed We developed an attenuated VSV recombinant VSV S expressing the SARS coronavirus SARS CoV spike S protein In cells infected with this recombinant S protein was synthesized glycosylated at approximately 17 Asn residues and transported via the Golgi to the cell surface Mice vaccinated with VSV S developed SARS neutralizing antibody and were able to control a challenge with SARS CoV performed at 1 month or 4 months after a single vaccination We also demonstrated by passive antibody transfer that the antibody response induced by the vaccine was sufficient for controlling SARS CoV infection A VSV vectored SARS vaccine could have significant advantages over other SARS vaccine candidates described to date D 2005 Published by Elsevier Inc Keywords Intranasal Neutralizing antibodies Protective immunity Spike Abstract Available online 25 July 2005 Long term protection from SARS single immunization with an Sagar U Kapadia a b John K Rose a Kanta Subbarao c a Department of Pathology Yale University School of Medicine b Department of Pharmacology Yale University c Laboratory of Infectious Diseases National Institute of Allergy and Infectious Received 21 April 2005 returned to author 0042 6822 see front matter D 2005 Published by Elsevier Inc doi 10 1016 j virol 2005 06 016 Corresponding author E mail address john rose yale edu J K Rose infection conferred by a VSV based vaccine Lamirande c Leatrice Vogel c Roberts c 310 Cedar Street LH 315 New Haven CT 06510 USA of Medicine New Haven CT 06510 USA Disease National Institutes of Health Bethesda MD 20892 USA 1 June 2005 accepted 10 June 2005 182 in various animal models De Diego et al and Galli 1994 Wesseling et al 1993 Expression of the S protein of SARS CoV from either in vitro expression a re emergence still exists and in such an event a SARS vaccine would be valuable especially to protect the most and VSV L genes Fig 1A Appropriate VSV transcription start and stop sites flanked the S gene We recovered a recombinant VSV designated VSV S from this plasmid using published methods Lawson et al 1995 To determine if the recombinant virus expressed the SARS S protein we infected BHK cells with the recovered virus or with wt VSV and metabolically labeled cells for 1 h with 35 S methionine Lysates of radiolabeled cells were then analyzed by SDS PAGE Since VSV shuts off host protein synthesis wt VSV infected cells express predominantly the five viral proteins L G N P and M VSV S infected cells expressed the five VSV proteins Fig 1B and also a protein of the expected size for S calculated as 197 kDa assuming N glycosylation at all 23 potential sites To verify the identity of the S protein S was immunoprecipitated from the cell lysates using a rabbit polyclonal antibody to the S cytoplasmic domain The immunoprecipitated protein irology 340 2005 174 182 175 vulnerable groups the elderly and health care workers In this study an experimental vesicular stomatitis virus VSV based SARS vaccine was developed and tested VSV is a negative strand RNA virus with a non segmented genome that encodes five structural proteins nucleocapsid N phosphoprotein P matrix M glycoprotein G and an RNA dependent RNA polymerase L VSV can elicit strong humoral and cellular immune responses in a variety of animals and VSV based vaccines have been shown to confer immunity in animal models of respiratory syncytial virus Kahn et al 2001 influenza virus Roberts et al 1998 SHIV AIDS Rose et al 2001 measles virus Schlereth et al 2000 and human papilloma virus Roberts et al 2004 VSV has significant advantages over many other vaccine vectors It has a relatively small RNA genome but can accommodate insertion of large foreign genes It replicates exclusively in the cytoplasm via RNA intermedi ates only and does not undergo recombination VSV seropositivity is also extremely low in the general popula tion Rose and Whitt 2001 Recombinant VSV generated by the plasmid DNA system employed here have been shown to be less pathogenic attenuated when compared with naturally occurring strains of VSV Roberts et al 1998 and cause no disease symptoms in non human primates Rose et al 2001 when given by intranasal oral or intramuscular routes In the present study we developed a recombinant attenuated VSV expressing the SARS CoV S protein and tested it as a SARS vaccine in a mouse model Subbarao et al 2004 In this model SARS CoV replicates to high titers in the respiratory tract allowing protection from SARS infection to be assayed readily in vaccinated animals following SARS CoV challenge Results Construction and characterization of a VSV recombinant VSV S expressing SARS CoV S protein To determine if the SARS CoV S protein could be expressed from a VSV recombinant we first generated a cDNA clone of the SARS S gene by reverse transcription and PCR amplification of RNA purified from SARS CoV systems or from live viral vectors has recently been reported to generate protection against SARS CoV infection in animal models Bisht et al 2004 Bukreyev et al 2004 Chen et al 2005 Gao et al 2003 Yang et al 2004 The spread of SARS was curtailed by public health measures and the outbreak was successfully contained in July 2003 Although several animals are susceptible to SARS CoV infection the source of the outbreak and host reservoir are uncertain Guan et al 2003 The possibility of S U Kapadia et al V infected cells We next constructed a plasmid with the SARS CoV S coding sequence inserted between the VSV G Fig 1 VSV recombinant expressing the SARS CoV S protein A The diagram illustrates the genome of the VSV S recombinant expressing the S gene and shows the terminal coding and flanking VSV transcription start and stop signals B BHK 21 cells were infected with wt VSV lanes 1 and 2 or VSV S lanes 3 and 4 labeled with 35 S methionine and lysates were analyzed by SDS PAGE without lanes 1 and 3 or with immunoprecipitation with anti S antibody lanes 2 and 4 C Immuno precipitated S was left undigested lane 1 digested with PNGase F lane 2 or Endo H lane 3 appeared as two bands Fig 1C lane 1 potentially representing two different glycosylation states of S To examine this possibility we treated the protein with PNGase F Fig 1C lane 2 which cleaves all N linked glycans PNGase F digestion yielded a single band of approximately 156 kDa confirming that variable glycosylation states of the S protein were responsible for the observed size hetero geneity Digestion of immunoprecipitated S protein with Endo H an endoglycosidase that cleaves only unprocessed or partially processed glycans from proteins yielded two bands Fig 1C lane 3 The Endo H resistant band of original size accounted for 15 of total S protein The lower band representing protein which had not yet under gone processing of its carbohydrate side chains accounted for 85 of total S protein In a pulse chase experiment with a 20 min pulse labeling we observed that 50 of S expressed by VSV was Endo H resistant within 30 min after the chase was begun data not shown From these experi ments we concluded that the S protein is transported through the exocytic pathway at least as far as the Golgi apparatus where Endo H resistance is acquired To determine if the S protein was transported from the Golgi to the cell surface we examined VSV S infected cells using indirect immunofluorescence microscopy BHK 21 cells were infected with VSV S or wt VSV fixed and then incubated with serum from a person who had recovered from SARS CoV infection A secondary Alexa Fluor 488 conjugated anti human antibody was used for visualization with fluorescence microscopy We found that the SARS S protein was clearly expressed on the cell surface as indicated by the strong fluorescent signal visible in VSV S but not wt VSV infected cells Fig 2 Immunization with VSV S elicits humoral immunity Initially we immunized four groups of eight mice with either wt VSV VSV S or SARS CoV Four weeks later serum was collected from each mouse To determine if the VSV and VSV S inoculations were effective we performed an assay to measure VSV neutralizing antibodies Pooled sera from mice infected with wt VSVor VSV S had a VSV neutralizing antibody titer of 1 2560 while mice infected with SARS CoV did not have any detectable VSV neutralizing antibody We then examined the individual serum samples for neutralizing antibodies to SARS CoV Table 1 Infection of mice with VSV S generated a stronger neutralizing antibody response average SARS neutralizing titer of 1 32 to SARS CoV than did infection with SARS CoV average SARS neutralizing titer of 1 12 Sera from control mice cells S U Kapadia et al Virology 340 2005 174 182176 Fig 2 SARS CoV S protein is expressed on the surface of VSV S infected Cells were fixed and stained for SARS S as described in Materials and methods The Microphot FX microscope using a 60C2 objective BHK 21 cells were infected with VSV S A and B or wt VSV C and D fluorescence left and DIC right images were generated with a Nikon immunized with wt VSV did not have detectable neutraliz ing antibody titers to SARS CoV Immunization with VSV S protects against SARS CoV infection Mice inoculated intranasally with SARS CoV do not show any clinical symptoms of infection however the virus replicates to high titers in the lungs and nasal turbinates NTs SARS CoV replication in respiratory tissues peaks by 2 days post infection and virus is cleared within a week Subbarao et al 2004 We examined SARS CoV repli cation in vaccinated and control animals to evaluate the effectiveness of the VSV S vaccine in protecting against SARS CoV infection Three groups of four mice each were immunized with either wt VSV VSV S or SARS CoV Four weeks later the mice were challenged with SARS CoV and 2 days after the challenge the lungs and nasal turbinates were collected and the corresponding viral titers were determined Fig 3A Control mice immunized with wt VSV had high SARS viral titers in both lungs and nasal turbinates following SARS CoV challenge In contrast mice immu nized with VSV S or SARS CoV were completely pro tected and controlled the SARS CoV challenge as indicated by SARS CoV titers at or below the detection limits in both the lungs and nasal turbinates These data indicate that vaccination with VSV S is as effective as a primary infectionwithSARS CoVinpreventingsubsequent SARS CoV infections To determine if VSV S vaccination could provide sustained protection from SARS CoV infection a similar experiment was carried out with a 4 month interval between immunizations and challenge Fig 3B Two days after challenge SARS CoV titers in the lungs of VSV S immu nized mice were at the limit of detection 1 of 4 mice or Table 1 Neutralizing antibody titers to SARS CoV Primary immunogen VSV VSV S SARS Individual SARS CoV neutralizing antibody titer 1 16 1 13 1 8 1 8 1 25 1 20 1 8 1 20 1 10 1 8 1 32 1 25 1 4 1 81 1 11 1 4 1 16 1 11 1 8 1 40 1 8 1 8 Average 1 32 1 12 S U Kapadia et al Virology 340 2005 174 182 177 Fig 3 Protection from SARS CoVinfection in immunized mice Balb c mice were challenged 4 weeks later with 10 5 TCID 50 SARS CoV Two days later lungs and Additional mice were challenged 4 months after immunization with wt VSV VSV turbinates were collected and viral titers were determined B Limits of detection log 10 TCID 50 g tissue in 5 nasal turbinate homogenates inoculated intranasally with wt VSV VSV S or SARS CoV The mice were nasal turbinates were collected and the viral titers were determined A SARS S and SARS CoV Two days after the challenge lungs and nasal were 1 5 log 10 TCID 50 g tissue in 10 lung homogenates and 1 8 was then infected ology undetectable 3 of 4 mice while the titer in the lungs of SARS CoV immunized mice was slightly above the limit of detection 1 of 4 mice or undetectable 3 of 4 mice Viral titers in the nasal turbinates were slightly above the limit of detection for both VSV S 3 of 4 mice and SARS CoV 2 of 4 mice immunized mice The mean SARS viral titers in nasal turbinates of both VSV S and SARS CoV inoculated mice were nearly 10 000 fold lower than the viral titers in nasal turbinates from wt VSV immunized mice These findings demonstrate that VSV S vaccination provides sustained protection from SARS CoV challenge and may provide protective immunity equal to or better than primary SARS CoV infection Fig 4 Inhibition of SARS CoV replication by transfer of serum from immunized CoV diluted 1 10 in PBS or wt VSV was injected into na ve mice Serum The limit of detection was a reciprocal dilution of 8 The mice were then SARS CoV titer The limit of detection was 1 5 log 10 TCID 50 g S U Kapadia et al Vir178 Serum from mice immunized with VSV S can confer passive protection VSV infection can elicit strong humoral and cellular immune responses To determine if antibody alone could provide protection against SARS CoV infection serum from wt VSV VSV S or SARS CoV infected mice was administered intraperitoneally into na ve mice Fig 4 Mice were bled 24 h after administration of antisera and SARS CoV neutralizing antibody titers were measured Mice were subsequently challenged with SARS CoV and 2 days later lungs were collected and SARS CoV titers were deter mined SARS CoV neutralizing antibody titers were detect able in the groups of mice that received VSV S antisera and SARS CoV antisera diluted 1 10 Only these mice were able to protect against SARS CoV infection upon challenge Mice receiving wt VSV antisera or normal mouse sera did not have measurable neutralizing antibodies to SARS CoV nor were they protected from SARS CoV infection upon challenge These results demonstrate that VSV S can elicit a strong antibody response that is sufficient for controlling SARS CoV infection Discussion In this study we used an attenuated VSV vector to express the SARS CoV S protein gene Because both SARS CoV and VSV replicate in the cytoplasm of infected cells we anticipated efficient expression of SARS CoV S protein by a recombinant VSV without problems associated with nuclear transcription including mRNA modification and mRNA export from the nucleus The level of S protein expression obtained from the VSV S recombinant was clearly sufficient to generate long term protective immunity to SARS CoV in this animal model The levels of SARS CoV neutralizing antibody titers following VSV S immunization were equal to or better than mice Sera from uninfected mice or mice immunized with VSV S SARS collected from mice to measure SARS CoV neutralizing antibody titer with SARS CoV and 2 days later the lungs were harvested to measure 340 2005 174 182 those elicited following primary infection with SARS CoV Although passive transfer of immune sera demonstrated that antibody induced by either VSV S or SARS CoV inocu lations is sufficient to prevent SARS CoV infection cellular immune responses may also contribute to protection Similarly it has been shown with other coronaviruses such as mouse hepatitis virus type 3 that antibody mediated immunity is sufficient for protection Pope et al 1996 Furthermore the complete protection from SARS viral replication in the lower respiratory tissues of VSV S immu nized mice observed when challenge was administered 4 months post immunization suggests that prolonged protec tion may be provided from a single immunization with VSV S Ours is the first study to show such long term protection after immunization with an experimental SARS CoV vaccine Inpreviousstudiesonvaccinationofmicewith VSV expressing an influenza hemagglutinin we found that protective influenza neutralizing antibody titers were very long lived decliningonly2 to4 foldoverthecourseofmore than one year Roberts and Rose unpublished data Our VSV based SARS vaccine could have significant advantages over previously described SARS vaccines Bisht AY278741 with the exception of a silent mutation at nucleotide 1173 which was G instead of A irology et al 2004 Bukreyev et al 2004 Chen et al 2005 Gao et al 2003 Yang et al 2004 Pre existing immunity to VSV is rare in the human population Effectiveness of other vaccine candidates based on adenovirus or parainfluenza virus could be limited by a high prevalence of pre existing immunity to the vectors in the human population Vaccines based on modified vaccinia Ankara MVA could face the same problem because the parental vaccinia virus has been used worldwide as the smallpox vaccine and there is extensive immunity to vaccinia especially in the older population most susceptible to SARS Crotty et al 2003 DNA vaccines have been effective especially in small animal models however efficacy in humans has yet to be demonstrated VSV has the additional advantages of growing to very high titers in cell lines such as Vero that are approved for human vaccine production VSV is also effective at very low doses Roberts et al 1998 1999 and is effective as an intranasal vaccine Egan et al 2004 Inactivated SARS CoV itself has also been shown to elicit an immune response He et al 2004 Qu et al 2005 Takasuka et al 2004 Tang et al 2004 Xiong et al 2004 Zhang et al 2004 however the production of such a vaccine would carry the inherent risks of exposure to SARS CoV from handling large volumes of infectious material that could result in accidental infection and secondary spread Additionally there might be concerns of proper and complete inactivation of the virus The SARS CoV S protein expressed by VSV was estimated from gel mobility to be approximately 200 kDa and was converted to about 156 kDa upon the removal of all N linked glycans Based on this calculation we estimate that at least 17 of the 23 potential N linked glycosylation sites are used assuming 2 5 kDa per N linked glycan We also demonstrated that S expressed from the VSV recombi nant moves to the Golgi apparatus where it acquires Endo H resistance About 15 of S expressed during a 1 h labeling period contained Endo H resistant glycans In a pulse chase experiment with a 20 min pulse labeling we observed that 50 of S expressed by VSV was Endo H resistant within 30 min after the chase was begun data not shown Finally we readily observed S protein on the cell surface Our results indic