Serological Relationships between Different Strains of Vesicular Stomatis Virus

Transcription

1 J. gen. ViroL (1972), i6, Printed in Great Britain 391 Serological Relationships between Different Strains of Vesicular Stomatis Virus By B. CARTWRIGHT AND F. BROWN Animal Virus Research Institute, Pirbright, Surrey, England (Accepted I2 May I972) SUMMARY The five vesicular stomatitis viruses Indiana, New Jersey, Cocal, Argentina and Brazil have been examined for cross-relationships by complement fixing and neutralization tests. With the exception of the Cocal and Argentina strains, intact virus particles showed little cross-reaction in either test. However, the infective skeleton-like structure produced by uncoating the virus with Tween-ether or Nonidet, and the infective ribonucleoprotein (RNP) produced by deoxycholate disruption of the virus, were neutralized by either homo- or heterotypic antisera and showed considerable cross-reactivity in the complement fixation test. Experiments with the isolated sub-units of the virus indicated that the major part of the cross-reactivity was associated with the RNP. The morphologically similar Piry and Chandipura viruses appeared to be unrelated to the vesicular stomatitis group in both complement fixation and neutralization tests. INTRODUCTION In addition to the classical Indiana and New Jersey serotypes of vesicular stomatitis virus (Cotton, T927) three fairly recent isolates from Trinidad (Cocal virus; Jonkers et al. I964), Argentina (Argentina virus; Garcia Pirazzi, Caggiano & Alonso Fernandez, I963) and Brazil (Brazil virus; Andrade et al I964 cited by Federer, Burrows & Brooksby, I967) have been characterized serologically by Federer et al. (t967) as belonging to the Indiana serotype. However, the new isolates showed marked antigenic difference within the serotype. These relationships are instructive in view of the confusion which exists concerning the cross-reaction between the two classical serotypes. Although there have been several reports describing the cross-reaction between the Indiana and New Jersey serotypes in complement fixation (McClain et al I953; Bankowski & Kummer, I955; Saran, I956; Kang & Prevec, I97o), resin-agglutination (Nieto & Segre, I958 ) and immunodiffusion tests (Myers & Hanson, I96z), the picture is far from clear because there have been equally convincing reports of the failure to detect cross-reactions in these tests (Brooksby, I948; Rice & McKercher, I954; Brown & Crick, I957; Stone & DeLay, I963; Murphy & Fields, I967). With the greater knowledge of the structure of vesicular stomatitis virus which has emerged from studies of the structural sub-units, it is now possible to relate the different serological properties to well-defined structures within the virus. Thus, Cartwright, Smale & Brown (I969) demonstrated that the antibodies produced by different sub-units had different mechanisms for neutralizing the virus. In the light of these results the reactions between the different virus sub-units and antibodies have been studied and the cross-relationships examined between the five isolates of

2 392 B. CARTWRIGHT AND F. BROWN vesicular stomatitis virus. The relationship of the vesicular stomatitis virus group to the recently characterized Piry and Chandipura viruses is also described. METHODS Viruses. The isolates of Argentina, Brazil and Cocal have been described (Federer et ak 1967). These were received from Mr R. Burrows of this Institute as tongue epithelium from infected cattle and were passaged in hamster kidney (BHK 21) monolayers in Eagle's medium for the preparation of virus stocks. Strains IND C (Indiana serotype) and NJ-M (New Jersey serotype) were received from Mr H. H. Skinner, also of this Institute, as egg passaged virus and were grown in BHK 2I monolayers. Piry and Chandipura viruses, received from Dr R. E. Shope, Yale Arbovirus Research Unit, U.S.A., as freeze-dried mouse-brain suspensions, were re-suspended in Eagle's medium and passaged in BHK 2I monolayers in Eagle's medium. All virus stocks were stored at -2o in 50 ~ glycerol. Hyperimmune antisera. Guinea pigs were inoculated intradermally in the hind pad with extracts of cattle epithelium or egg-passaged virus for the vesicular stomatitis strains or with BHK 21 passaged material for the Piry and Chandipura viruses. When vesicles appeared the pads were harvested and extracts used to infect the guinea pigs for the production of antisera. A subcutaneous booster inoculation was given 1 month after intradermal inoculation of the guinea-pig-passaged virus. Two weeks later the animals were bled and the sera separated. Antisera against intact virus, trypsin-treated virus or RNP were obtained by inoculating guinea pigs with preparations which had been made non-infective with o'o5 ~o acetylethyleneimine (Brown & Crick, I959 ). The animals were bled after 2I days and the sera separated. Absorbed IgG preparations were made by mixing excess of the RNP with I/2oo antiserum, adding normal guinea-pig serum to a final concentration of 25 ~ and filtering through DEAE-cellulose in o.oi N-phosphate, ph 7"6. The unreacted IgG passed through the column without adsorption, whereas the complex of RNP and its antibody was retained on the column. Estimation of infectivity and neutralizing antibody activity. The infectivity of intact or disrupted virus was titrated by intracerebral inoculation of 7-day-old mice with o'o3 ml. of serial io-fold dilutions of the preparation. Antibody activity was determined by intracerebral inoculation of 7-day-old mice with o'o3 ml. of equal mixtures of antiserum (neat or I/too) with Io-fold dilutions of the appropriate virus. The neutralizing activity indicated by this test with o-oi5 ml. of the serum dilution is shown as the difference between the 50 infectivity end-points of the virus-antiserum mixture and of the virus plus normal serum control. The serum dilutions used for appropriate depressions of virus infectivity are quoted in the Results. Complement fixing activity. Two tests were used which differed only in the conditions of incubation of antigen and antiserum with complement. In the first, the mixture of antigen, antiserum and complement was incubated at 37 for 30 min. before addition of the haemolyric system and further incubation at 37 for 30 min. In the second test the mixture was kept at 4 for i6 hr before addition of the haemolytic system and incubation at 37. Disruption and fractionation of virus. These methods have been described in previous papers on the structure of vesicular stomatitis virus (Cartwright et al. I969).

3 Serological relationships for VSV strains 393 (a) ' Complement-fixing ac[ivity o 320 units/ p- 130 units / /, s, I 1 X q0 3 (b) Complement-fixing activity.4 units '~00 units " Fraction Fig. I. Sucrose gradient centrifugation of [ah]-uridine and p4c]-amino acid labelled virus and trypsin-treated virus after disruption with o.2 ~ sodium deoxycholate. Fractions I3 to I7 and I8 to 23 from each gradient were combined for quantitative complement fixation tests. (a) Virus; (b) trypsin-treated virus. O--O, [nil]. RESULTS Serological reactions of intact virus and virus sub-units with homotypic antiserum Unfractionated harvests of vesicular stomatitis virus contain, in addition to infective particles, interfering component, ribonucleoprotein, rosette-like structures resembling the virus envelope and small particles similar to the surface projections of the intact virus (Brown, Cartwright & Almeida, I966). Sub-virus components can also be produced from virus particles by controlled disruption with detergents and subsequent separation by sucrose gradient centrifugation. By these methods it is possible to obtain almost homogeneous preparations of each virus constituent. When necessary, the virus was labelled specifically in the RNA and protein so that fractionation of the different components could be monitored. These components were used to clarify the serological reactions of the different subunits with homotypic antiserum before examination of their reactions with heterotypic antisera. The virus contains three major proteins conventionally referred to as P~, P3 and P~ (Kang & Prevec, I97o), located on the surface projections, RNP and matrix protein respectively (Cartwright, Talbot & Brown, I97o ). The reaction of hyperimmune serum with intact virus particles is directed towards the surface projections because particles from which

4 394 B. CARTWRIGHT AND F. BROWN Table I. Reactions of intact and Tween-ether d&rupted (skeleton) vesicular stomatitis virus (Indiana c) in neutralization tests with antiserum against intact virus, trypsin-treated virus or ribonucleoprotein Virus Log Skeleton Log infectivity: depression of infectivity depression of log infectivity due log infectivity due Antibody (ID5o/o'o3 ml.) to antiserum* (ID5o/o'o3 ml.) to antiserum* None 9" I t -- 4"7:~ -- Anti-virus 48 4"3 I "8 2"9 Anti-virus absorbed with RNP 5"4 3"7 3"8 o'9 Anti-trypsin virus antibody 8-8 0"3 I'2 3"5 Anti-trypsin absorbed with RNP "8 o'9 Anti-RNP antibody 8.8 o'3 2q z.6 Anti-RNP absorbed with RNP "4 0"3 * The antisera were the o.oi M-phosphate, ph 7"6, eluates from DEAE-cellulose columns and contained the equivalent of I/4oo of the initial antisera. t Activity in complement fixation against anti-virus serum but not against anti-trypsin virus serum. :~ Equal activity in complement fixation against anti-virus or anti-trypsin virus sera. the projections have been removed with trypsin no longer fix complement with the antiserum (Cartwright et al. I969). Disruption of the trypsin-treated virus with o.i ~ sodium deoxycholate released the RNP and matrix protein and the complement fixed by this mixture was about 3 ~ of that fixed by untrypsinized virus particles after similar disruption. This indicates that the surface projections account for 70 ~, and P3 and P4 together for 3o ~, of the total complement fixing activity of disrupted virus. The matrix protein P~ appears to possess very little complement fixing activity because fractionation of deoxycholate disrupted virus in sucrose gradients gave a peak of RNP which contained 30 ~ of the total activity of disrupted virus and a peak (containing P2 and P4) with 7o ~ of the activity (Fig. ~ a). Since P2 accounted for 7o ~ of the total activity (see above) very little activity can be associated with P4. This was confirmed by fractionation of deoxycholate disrupted trypsin-treated virus in a sucrose gradient. Most of the complement fixing activity of the preparation was present in the ribonucleoprotein peak and less than 5 ~ was found at the position in the gradient corresponding to P4 (Fig. I b). This indicated that less than 1.5 ~o of the complement fixing activity of the disrupted virus was associated with this P4 protein. Reactions of neutralizing antibodies with different sites in the virus particle We have shown (Cartwright et al. I969 ) that antiserum produced by inoculating guinea pigs with virus particles from which the surface projections had been removed with trypsin, did not reduce the infectivity of virus suspensions and failed to react in complement fixation tests with either intact virus particles or particles treated with trypsin. However, the same antiserum fixed complement with both skeletons and isolated RNP, and reduced the infectivity of the skeleton-like structures produced by Tween-ether disruption of the virus as effectively as did antiserum against intact virus (Table 0. Since the antiserum produced by inoculation of trypsin-treated virus particles does not contain antibody against surface projections, the neutralization reaction with the skeletons must be with the RNP or matrix protein. This suggests the presence of at least two kinds of neutralizing antibodies directed against different proteins of the virus. Neutralization of intact virus particles involves the reaction of antibody with the surface projections only, whereas neutralization of the skeletons involves the internal proteins P3 and P4-

5 Serological relationships for VSV strains 395 Table 2. Complement fixation reactions at 4 for 18 hr of deoxycholate-disrupted viruses Antiserum Virus New Jersey Indiana C Brazil Argentina Cocal Piry Chandipura Complement fixing activity as ~ of homologous activity New Jersey IOO 3 lo 7 < I < I < I Indiana C Io ioo < I < 1 Brazil 3 i i ioo 3o 16 < i < I Argentina I0O > 5 O < I <~ I Cocal 2 2i 20 > 70 I00 < 2 < 2 Piry <.2 < z < 2 < a < < z Chandipura < I < I < I < I < I < I I00 The matrix protein P~ appears to have no role in the neutralization of the skeletons because isolated RNP absorbed the neutralizing activity of anti-trypsin virus antiserum (Table I). The same preparation of RNP also absorbed the activity to neutralize skeletons of anti-virus serum but did not reduce its activity to neutralize intact virus (Table I). This means that we have not detected antibody to P4 either in neutralization or in complement fixation tests. Direct evidence that antibody to P3 is involved in the neutralization of the skeletons was obtained in the following experiment. Ribonucleoprotein was isolated from virus particles by mixing with o'5 ~ sodium deoxycholate and centrifuging through a sucrose gradient (Cartwright, Smale & Brown, I97o ). The RNP was then inactivated with o'o5 ~ acetylethyleneimine (Brown & Crick, I959 ) and inoculated into guinea pigs. The antiserum so produced did not reduce the infectivity of virus suspensions but neutralized the infectivity of the skeletons produced by disruption with Tween-ether (Table I). Reactions of virus and virus sub-units with heterotypic antisera Unfractionated virus harvests showed no cross-reaction in complement fixation tests at 37 for 3o rain. except in the case of the Cocal and Argentina strains, in agreement with the experiments of Federer et al. (I967). Even after deoxycholate treatment, the unfractionated viruses showed no cross-reaction except with the same two strains. The Cocal strain gave 82 ~ of the homologous fixation in tests with Argentina antiserum and the Argentina strain showed 73 ~ of the homologous fixation with Cocal antiserum. In tests of I8 hr duration at 4, some cross-fixation was observed with all the strains but there were only trace crossreactions, when purified virus particles were used. However, disruption of the particle with o-i ~o sodium deoxycholate to expose the RNP and matrix protein gave significant cross-reactivity (Table 2). Whereas virus particles were not neutralized by heterotypic antisera, skeleton-like structures obtained by Tween-ether treatment were neutralized by the same sera. For example, the extent of cross-neutralization of the infectivity of intact Indiana or New Jersey viruses by heterotypic sera was less than o'5 log units, but as much as 2. 4 log units of infectivity were neutralized when the skeletons were used. Similar results were obtained with the skeleton-like particles prepared from the other viruses (Table 3). Comparable results were also obtained with skeletons produced by Nonidet treatment of the virus and with the RNP prepared by deoxycholate disruption. Federer et al. (1967) considered that the Argentina, Brazil and Cocal viruses were more closely related to the Indiana serotype than 26 v~ 16

6 396 B. CARTWRIGHT AND F. BROWN Table 3. Neutralization of intact virus and Tween-ether disrupted virus by homologous and heterologous antisera Proteins Antiserum involved in New Chandi- Virus reaction Jersey Indiana Brazil Cocal Argentina Piry pura Log depression of infectivity in test with o'oi5 ml. of I/IOO antiserum New Jersey P~ 5" "4 0"9 0"3 -o'3 P3P "I I'9 -o'2 1"7 --0'3 o'i Indiana C P~ 0'3 5o 1.6 I "7 I "7 o'z 0'7 PaP4 2"4 3"6 3' I 3"o 3"4 o.i 0"5 Brazil P2 o'4 o5 5'3 1.8 I'7 -o'3 -o.i PaP~ 2"9 2"9 ~ 3'7 2"3 2'3 0"5 0'7 Cocal P2 0'9 14 0"7 4"9 3"2 -o'3 -o.i PaP4 -~ I'7 ~ I5 ~ 1-7 ~ 1"9 ~ i. 9 --o.i 0"7 Argentina P2 0'7 l 9 I '9 5" I 4"4 0"3 o" I P3P4 I'3 ~. 2-7 ~ 2. 7 ~ ~ Pity P2-0"3 o -0'3 --o.i o.i 3"6 o.i PaP4 --0"7 --0"9 --0"9 -- I'I - I"3 ~ 3"5 0"4 Chandipura P " '8 -o'5 5"3 P~P4 o'3 -o5 -o'3 -o'7 -o.i o ) 3'4 The upper figure in each pair shows the neutralization of infectivity of intact virus and the lower figure that of the Tween-ether disrupted virus skeleton. Each value is the mean of at least two determinations. to the New Jersey serotype and the results in Table 3 show that the Indiana virus skeleton was neutralized by antisera to these viruses to about the same extent as by the homologous antiserum but less by New Jersey antiserum. Relationship of Piry and Chandipura viruses to vesicular stomatitis virus The recently described Piry and Chandipura viruses have a morphology similar to that of the vesicular stomatitis viruses but quite different from that of the FLURY strain of rabies virus (Fig. 2; Murphy, Harrison & Whitfield, I969; Bergold & Munz, 197o ). Murphy & Shope (197I) demonstrated a small but distinct cross-neutralization of Piry and Chandipura viruses by the Indiana serotype when hyperimmune mouse ascitic fluid was used as antibody. Using I/IOO hyperimmune guinea pig serum, which neutralized the homologous viruses by 5 log units, we detected no cross-neutralization with any of the five vesicular stomatitis viruses and (Table 3) and only slight neutralization with undiluted serum. More significantly, the vesicular stomatitis viruses disrupted with Tween-ether were not neutralized by Piry or Chandipura antisera, and the skeletons and RNP from the Pity and Chandipura viruses were not neutralized by the vesicular stomatitis virus antisera. DISCUSSION Vesicular stomatitis virus contains three major proteins, P2 located in the surface projections, P8 in the RNP and P4 in the matrix. We suggest that the type and strain specificity are determined by an antigen on the surface projections because when these are removed with Tween-ether or Nonidet an infective substructure is left which can be neutralized with heterotypie antisera. For the Indiana and New Jersey viruses, the extent of this crossneutralization is less than that for homotypic neutralization. However, the skeleton of

7 Serological relationships for V S V strains 397 Fig. 2. Comparison of the morphology of the five vesicular stomatitis viruses, Piry and Chandipura viru3es and the FLURYstrain of rabies virus. Indiana virus is similarly neutralized by homologous antiserum and by Cocal, Brazil or Argentina antisera. This supports the provisional classification by Federer et al. U967) of the Cocal, Brazil and Argentina viruses as subtypes of the Indiana serotype. Our results show clearly that the cross-reactions described are a property of the RNP, and that the matrix protein has no detectable role in the complement fixation and neutralization reactions. The cross-reactivity of the RNP is in accord with the observation of Kang & Prevec (x97o) that the 2o s antigen separated from suspensions of New Jersey virus showed some cross-reaction in complement fixation tests with antiserum produced by inoculation of the core antigens of Indiana virus. The discrepancy between the reports of previous workers (see Introduction) concerning the specifity of the serological reactions between the two serotypes of vesicular stomatitis virus may thus be due to the free internal antigen present in the different virus preparations used. Despite several attempts to confirm the observation by Murphy & Shope (~ 97 ~) that Piry and Chandipura viruses have a significant reaction with Indiana antiserum and with each other, we detected no relationship in complement fixation or neutralization tests using either intact virus or infective skeletons. This difference may be due to the use by Murphy & Shope (I97I) of hyperimmune mouse ascitic fluid as the source of antibody, whereas we used hyperimmune guinea pig serum. It is prudent to delay judgement on the serological classification of these two viruses until definitive tests are made. We thank Mr C. J. Smale for providing the electron micrographs. Z6-z

8 398 B. CARTWRIGHT AND F. BROWN REFERENCES BANKOWSKI, R. A. & KUMMER, M. B. (1955). Vesicular stomatitis and vesicular exanthema differentiation by complement fixation. American Journal of Veterinary Research I6, 374. BER~OLD, G. ~. & MUNZ, K. 0970)- Characterization of Piry virus. Archivfiir die gesamte Virusforschung 3t, I52. ~ROOKSBY, s.b. (1948). Vesicular stomatitis and foot-and-mouth disease differentiation by complement fixation. Proceedings of the Society for Experimental Biology and Medicine 67, 254. BROWN, V., CARTWRlCHT, B. ~, ALMEIDA, J. D. (1966). The antigens of vesicular stomafitis virus. I. Separation and immunogenicity of three complement fixing components. Journal of Immunology 96, 537. BgowY, v. & C~ I(, J. (1957). Specific precipitin reactions with the viruses of foot-and-mouth disease and vesicular stomatitis. Nature, London x79, 316. BROWN, r. & C~CK, J. (I959). Application of agar-gel diffusion analysis to a study of the antigenic structure of inactivated vaccines prepared from the virus of foot-and-mouth disease. Journal of Immunology 82, 444. CARTWRIGHT, B., SMALE, C. J. & BROWN, F. (1969). Surface structure of vesicular stomatitis virus. Journal of General Virology 5, I. CARTWRmHT, B., SMALE, C. J. & BROWN, F. (1970). Dissection of vesicular stomatitis virus into the infective ribonucleoprotein and immunizing components. Journal of General Virology 7, I9. CARTWRmUT, B., TALBOT, e. & BROWN, F. (1970). The proteins of biologically active sub-units of vesicular stomatitis virus. Journal of General Virology 7, 267. COTTON, W. E. (1927). Vesicular stomatitis. Veterinary ~VIedicine 22, I69. FEDERER, K. E., BURROWS, R. & BROOKSBY, J. B. (1967). Vesicular stomatitis virus - the relationship between some strains of the Indiana serotype. Research in Veterinary Science 8, ~o3. CARCIA PIRAZZl, A. J., CA~IANO, C. ~. & ALONSO fernandez, A. (1963). Publication Tecnica No. 2, CANEFA, Ministry of Agriculture, Buenos Aires, Argentina. JONKERS, A. H., SHOPE, R. E., AITKEN, r. u. ~. & SPENCE, L. (1964). Cocal virus, a new agent in Trinidad related to vesicular stomatitis virus. American Journal of Veterinary Research 25, 236. KANG, C. V. & ~'RrWC, L. (197O). Proteins of vesicular stomatitis virus. 2. hnmunological comparisons of viral antigens. Journal of Virology 6, 20. MCCLAIN, M.E., MADIN, S.H., SEELY, J., ANDRIESE, ~'., CLINGER, D. & SMITIJ, H. (1953)- Studies on equine vesicular stomafitis. Report from Naval Biological Laboratory, California. Section 8, 31. MURPI~Y, F. A. &FIELDS, B. N. 0967). Kern Canyon virus: electron microscopic and immunological studies Virology 33, 6z5. MURPHY, r. A., HARRISON, A. K. & WHITFIELD, S. G. 0969). The infection of cells by bullet-shaped viruses. Proceedings of the Electron Microscopic Society of America, 27th meeting, p MURPIqY, F. A. & SHOPE, R. E. 097I)- Proceedings of the Second International Congress for Virology, Budapest. Workshop on ' Bridging Groups of Viruses', p. 26 I. MYERS, W.L. & HANSON, R. 1'. (t962). Immunodiffusion studies of the antigenic relationships within and between serotypes of vesicular stomatitis virus. American Journal of Veterinary Research 23, 896. ~ETO, F.M. & StoRE, D. 0958). The resin-agglutination test for vesicular stomatitis. American Journal of Veterinary Research XD, 761. RICE, c. E. & MCKERCHER, P. D. (I954). Studies of the complement fixation reaction in virus systems. 6. In vesicular stomatitis in horses, cattle and swine. Journal of Immunology 43, 309. SAVAN, M. (1956). Studies on bovine serum and its use in complement fixation reactions. Ph.D. Thesis, University of Wisconsin, Madison, U.S.A. STONE, S. S. & DELAY, P. D. 0963)- A rapid complement-fixation test for identification of vesicular stomatitis virus in cattle. American Journal of Veterinary Research 24, IO6O. (Received 6 March I97Z)