Host Restriction of Friend Leukemia Virus. Role of the Viral Outer Coat (mice/fv-1 locus/vesicular stomatitis virus)

Transcription

1 Proc. Nat. Acad. Sci. USA Vol. 70, No. 9, pp , September 1973 Host Restriction of Friend Leukemia Virus. Role of the Viral Outer Coat (mice/fv-1 locus/vesicular stomatitis virus) THEODORE G. KRONTIRIS, RUY SOEIRO, AND BERNARD N. FIELDS Albert Einstein College of Medicine, Departments of Cell Biology and Medicine, Bronx, New York Communicated by Harry Eagle, June 8, 1973 ABSTRACT Host restriction of oncogenesis of RNA tumor viruses in vivo is associated with several gene loci. One of these genes, the Fv-1 locus in mice, is expressed in vitro and may be studied in mouse-embryo cultures that are restrictive or permissive for replication of Friend leukemia virus. Two strains of Friend leukemia virus, N- or B-tropic, show reciprocal ability to replicate successfully in either NIH Swiss (N-type) or BALB/c (B-type) cells that differ at the Fv-1 locus. These two strains of virus and two cell lines form a system to measure host restriction in Vitro. Measurement of adsorption of Friend leukemia virus to permissive or restrictive cells reveals no difference in rate or total amount of virus bound. Furthermore, studies with virions of vesicular stomatitis virus phenotypically mixed within an envelope containing Friend leukemia virus protein show no differences in penetration or replication of vesicular stomatitis virus. These results strongly suggest that host restriction of Friend leukemia virus is due to an intracellular event in the viral replication cycle. The complexity of virus-host interactions in the ultimate production of leukemic disease is well illustrated by the genetic aspects of the host response to murine RNA tumor viruses. Although host response is a multigenic effect, one gene in particular, the Fv-1 locus in mice, determines susceptibility or resistance to exogenous Friend or other murine leukemia virus infections (1-4). In vivo, this gene determines the susceptibility of mice to Friend disease and, in vitro, controls the ability of Friend leukemia virus (FLV) to replicate successfully (2, 3, 5). This gene influences the cell-tocell spread of endogenous virus produced in response to chemical induction in mouse-embryo cells in culture (6, 7) and the spontaneous appearance of endogenous virus in mice (7, 8). An understanding, therefore, of the molecular basis for host restriction of this gene would be of major importance in the overall study of the relationship of host cell and viral genes in viral oncogenesis. Strains of murine leukemia virus exist, termed N- or B- tropic, defined by their ability to replicate successfully in either NIH Swiss (N-type) or BALB/c (B-type) strains of mice that differ at the Fv-1 locus (2-5). Certain strains (e.g., Rauscher and Moloney leukemia virus) replicate equally well in either cell type and are termed NB-tropic. Both N- and B-tropic strains of Friend virus exist as defined in vivo by Lilly and Steeves (10). In vitro, the reciprocal inability of the N-strain of FLV (N-FLV) to grow in BALB/c cells or of Abbreviations: N-FLV, N-tropic Friend leukemia virus; B-FLV, B-tropic FLV; MLV, murine leukemia virus; VSV, vesicular stomatitis virus; MOI, multiplicity of infection; S +L- sarcomapositive, leukemia-negative, PFU, plaque-forming units the B virus (B-FLV) to grow in NIH cells has been demonstrated (Krontiris et al., unpublished results), and forms the basis of a system in which host restriction may be studied in vitro. Studies of the antigenicity of these two virus strains suggest that they cannot be distinguished by immunological techniques (10). This finding has suggested that host restriction by nonpermissive cells is a consequence of an intracellular event. However, no direct evidence is available which would define the step in FLV replication at which restriction occurs. Experiments to be presented here attempt to define whether differences in adsorption and penetration of virus, steps of replication usually associated with the virus envelope, are involved in restriction of FLV by nonpermissive cells. Since the ratio of virus particle to plaque-forming units (PFU) for FLV is not known, a study of the binding of radiolabeled virus to host cell might measure mainly biologically inert material. Therefore, absorption of virus was studied by means of infectivity titer rather than binding of radiolabeled virions. Furthermore, in order to isolate effects of the virus coat in the study of penetration, we used vesicular stomatitis virus (VSV) virions phenotypically mixed within an envelope containing FLV protein. The results of these studies suggest strongly that host restriction of FLV is due to a limiting intracellular step in viral replication. MATERIALS AND METHODS Cells. Primary Swiss (N-type) and BALB/c (B-type) mouse-embryo cells were obtained from Microbiological Associates, Rockville, Md. The cells were used at the second or third passage for all experiments and maintained the in vitro characteristics of restriction described by Pincus et al. (7, 8). S+L- (sarcoma-positive, leukemia-negative) cells, kindly provided by Dr. Robert Bases, are Swiss 3T3 cells transformed with the Moloney strain of murine sarcoma virus as described by Bassin (11). All cell lines were maintained in Eagle's minimal essential medium supplemented with 10% fetal-calf serum (Grand Island Biologicals), iglutamine, and penicillin-streptomycin. Viruses. Friend leukemia virus strains F-S (N-tropic) and F-T (B-tropic) (10) were obtained as spleen extracts of infected mice from Dr. Frank Lilly. Each strain was passaged in susceptible mouse-embryo cell cultures. Infected cell supernatants were harvested and frozen at -85 to provide tissue-culture passage stocks of "N" and "B" virus which were then used for all subsequent experiments. VSV, Indiana serotype, was

2 2550 Microbiology: Krontiris et al. TABLE 1. Titration of FLV infectivity in S+L- cells Dilution Virus type N-FLV No. of plaques 37, 45 18, 6 4, 4 PFU per ml (X 1Q-6) 1.3 ± ±0 B-FLV No. of plaques 59, 51 17, 20 7, 6 PFU per ml (X 10-6) 1. 8 ± ± ±0. 1 Serial dilutions of N-FLV and B-FLV stocks were assayed in S+L- cells. Numbers of plaques from duplicate plates and the resultant calculated titers (PFU/ml) are presented for each dilution. Dilution is in logarithmic units. obtained from Dr. Donald Summers. A temperature-sensitive mutant of VSV (ts 45) was kindly supplied by Dr. C. R. Pringle (12, 13). Plaque Assays. Infectivity of FLV was titered both by the S+L- plaque assay (14) and the XC assay (15), with the appropriate N- or B-type mouse-embryo cells. Preliminary studies demonstrated that titers of N-FLV and B-FLV, as determined by the S+L- assay, were comparable to those obtained with the XC assay. This S+L- cell line originated from a Swiss 3T3 cell transformed by murine sarcoma virus. Although this line was originally B-tropic (Bassin, personal commaunication), our studies indicated no loss of infectivity titer of N-tropic FLV when assayed in this cell. Furthermore, the multiple-hit curve normally demonstrated by N-FLV when titered in BALB/c cultures (Hartley, J. W., Pincus, T. & Rowe, W. P., personal communication) was not seen in the S+L- assay. That is, the number of plaques of FLV decreases appropriately by 10-fold per 10-fold dilution of virus (Table 1). Moreover, titers of N- or B-FLV calculated from dilutions over a 10-fold range are constant. If host restriction occurred in these cells, the plaque number would decrease to a greater degree than predicted by the given dilution, that is, the calculated titer for each dilution would not be constant, but would decrease with virrus dilution (multiple-hit kinetics). Our findings indicate that in S+L- Swiss 3T3 cells, host restriction of N-FLV no longer occurs. VSV plaque assays were done in 60 X 15-mm tissue culture dishes (Falcon Plastics) seeded with 2 X 106 mouse-embryq or 2.5 X 106 HeLa or L cells. After 24 hr, plates were inoculated and overlayed with 1% agarose (Sargeant-Welch). For plaque enumeration, the plates were subsequently overlayed with neutral red agarose (1%) at 36 hr (370 incubation) or at hr (310 incubation) and examined for plaques 24 hr later, after which time there was no significant increase in plaque number. Adsorption of FLV to Mouse-Embryo Cells. Confluent monolayers of N- and B-type mouse-embryo cells in 60 X 15-mm dishes were inoculated with 0.3 ml of N-FLV or B-FLV in Eagle's medium at a multiplicity of infection (MOI) of The plates were agitated by tilting periodically at 370 and, at stated times, the entire inoculum volume was withdrawn and titrated on S+L- cells. All experiments were done with duplicate plates. As a control for nonspecific surface adsorption and heat inactivation, empty dishes were inoculated and the medium was titered. Isolation of PhenotypicaUy Mixed Virions of VSV. Both Swiss and BALB/c mouse-embryo cultures were inoculated, respectively, with N-FLV and B-FLV (MOI 2.0). After 6 days, when maximum FLV yields occur, such cultures were superinfected with a temperature-sensitive mutant of VSV (ts 45) at a MOI of 5-10 PFU per cell. Control uninfected mouse-embryo cultures were infected with VSV as a source of stock VSV. The VSV-superinfected cultures were incubated at 310 for a further 20 hr; the cells were suspended mechanically into their medium, and sonicated. The sonicate was clarified by centrifugation at 2000 X g for 10 min, and aliquots were frozen at Phenotypically mixed virions of VSV were selected from the mixed progeny of such an infection by heating a 1:10 dilution of virus sample at 450 for 1 hr. This treatment removes the coat from the ts mutant, producing viral cores, and decreases the titer of stock VSV (ts 45) by 4-5 logarithmic units (16). The demonstration of residual phenotypically mixed virions [VSV(FLV) ] was determined as described (17). Surface specificity of the phenotypically mixed virions was examined by virus neutralization and host range, by formation of plaques on HeLa, L, or FLV-infected mouse-embryo cells. Neutralization was done by mixing one volume of virus sample with one volume of a 1: 5 dilution of antibody against VSV prepared in mouse (hyperimmune ascitic fluid) (18). Where indicated, an additional 1 volume of antiserum against FLV prepared in mouse was added. (This antiserum, produced by infection of BALB/c mice with both N- and NBtropic FLV, was kindly provided by Dr. Frank Lilly.) Virus appropriately diluted with medium was a neutralization control. These mixtures were incubated overnight at 40. Plates were inoculated with suitable dilutions of virus, incubated 1 hr at 25, washed several times with complete medium, and overlayed with agar. TABLE 2. Proc. Nat. Acad. Sci. USA 70 (1973) Adsorption of FLV to N- or B-type cells Mouse- % Re- Virus embryo Min- Titer main- (inoculum) cell type utes (X 10-5) ing Exp. I. N-FLV N (5 X 105PFU) B B-FLV B (5 x 105PFU) N Exp. II. N-FLV No cells (2 X 105 PFU) Confluent monolayers of either N-type or B-type mouseembryo cells were inoculated in duplicate with 0.3 ml of either N-FLV or B-FLV at an MOI of Inocula were withdrawn at the times indicated and titered for infectivity on S+L- cells. Unadsorbed virus is expressed as the percent remaining virus relative to the amount recovered at 0 min.

3 Proc. Nat. Acad. Sci. USA 70 (1973) RESULTS Adsorption of FLV. Initial experiments were designed to determine whether significant differences existed between susceptible and resistant cells as to their ability to adsorb a given virus type. In each case, both the kinetics and final amount of virus adsorbed were studied. At all times, the ability to perform the experiment with two virus and two cell types served as an internal control. A series of confluent plates of BALB/c or Swiss mouseembryo cells were inoculated with either N-FLV or B-FLV. At intervals, the entire volume of medium was withdrawn and the virus remaining unadsorbed was titrated by means of the S+L- cell assay. Although initiation of replication of RNA tumor viruses is optimal during exponential cell growth and appears to require cellular DNA synthesis (19, 20), preliminary studies indicated that the rate of uptake of virus from the medium was not significantly influenced by the degree of confluence of the monolayer. The absolute amount of virus adsorbed increased only slightly for a defined viral inoculum as cell density increased from about half-confluent (7 X 105 cells per dish) to a fully-confluent (2 X 106 cells per dish) monolayer. Results of these studies are shown in Table 2. Virus recovered immediately on inoculation was designated as 100% unadsorbed; each subsequent figure is represented as the percent virus remaining in medium after each adsorption period. These results indicate clearly that virus of either N or B tropism adsorbs equally well to either resistant or susceptible cells. Furthermore, no significant differences can be observed in the rate of uptake. The maximal amount of adsorption has occurred within the first 15 min after inoculation of viruses onto either cell type. In this experiment, uptake of N-FLV was nearly complete by 15 min and represented 75-80% of the input virus. We can conclude that host restriction is not associated with failure to bind to the restrictive cell type. Cell Penetration of Phenotypically Mixed Virions of VSV. Experiments were designed to test the penetration of FLV Host Restriction of Friend Leukemia Virus 2551 into restrictive or permissive cells. Since virus at the cell surface may be neutralized by antiserum, penetration may be measured by protection of adsorbed virus from neutralization. The results, although consistent with equal rates of penetration, were difficult to reproduce, and we turned to other means to demonstrate this point. It has recently been reported (17) that superinfection of cells infected with either avian myeloblastosis or murine leukemia virus with VSV yields a small proportion of progeny VSV that could not be neutralized with VSV antiserum. This fraction of VSV possessed antibody neutralization, host range, and interference properties of the leukemia virus. Furthermore, the phenomenon appeared to be due to phenotypic mixing, for after a single passage of uninfected cells, all the progeny reverted to wild-type VSV. The percentage of phenotypically mixed VSV in the mixture of virus types produced by such an infection can be increased by the use of a VSV temperature-sensitive mutant (ts 45) (16). This mutant is thermolabile; unlike the wildtype virus, its titer of infectivity can be reduced by 4-5 logarithmic units by heating at 450 for 1 hr. Thermolability in ts 45 is a property of an abnormal VSV coat protein, and the thermal loss of infectivity is due to production of noninfectious viral cores. Superinfection with ts 45 of cells infected with murine leukemia virus (MLV) at the permissive temperature (310) results in progeny virus, the bulk of which is thermolabile. Most of the heat-resistant progeny are VSV cores phenotypically mixed within a MLV coat [VSV(MLV) ]. Since such a preparation of VSV may be essentially freed of particles with VSV-specific coat proteins by heat treatment, we reasoned that the resultant, essentially pure, VSV(MLV) could be used to examine the role of the FLV coat in host restriction. Preliminary studies showed that VSV replicated equally well on either N or B mouse-embryo cultures; therefore we postulated that VSV(FLV), once virus penetration had occurred, should be able to replicate normally. However, if the FLV envelope proteins were involved in the host-restriction mechanism, then a VSV virion whose envelope was TABLE 3. Characterization of wild-type and ts 45 VSV Virus grown Antiserum Exp. Virus in treatment Heat NME BME HeLa L N/NME I VSV NME (wild type) Anti-VSV Anti-FLV anti-flv II VSV NME (wild type) VSV NME (ts 45) III VSV NME (ts 45) Anti-VSV Anti-FLV VSV (wild type or ts 45) grown in mouse-embryo cells, was titered for infectivity in the cell lines indicated. Before inoculation, virus was treated with antiserum or heat or both. NME and BME, N- and B-type mouse-embryo cells, respectively. NINME, N-type mouseembryo cells infected with N-FLV.

4 2552 Microbiology: Krontiris et al. Proc. Nat. Acad. Sci. USA 70 (1973) TABLE 4. Phenotypically mixed virions of VSV Virus Antiserum Log titer in Virus grown in treatment Heat NME BME HeLa L N/NME VSV NME (ts 45) N-FLV anti-flv <2.0 <2.0 VSV BME (ts 45) B-FLV anti-flv N- or B-type mouse-embryo cells, infected, respectively, with N-FLV or B-FLV were superinfected with ts 45 VSV. The progeny of this double infection were characterized as to thermal stability, neutralization by antiserum, and host range as in Table 3. Results are expressed as logarithm of infectivity titer remaining on each cell type with or without antiserum or heat treatment. specified either by N- or B-type FLV might demonstrate reduced titers when assayed on the nonpermissive cell. In order to isolate phenotypically mixed virions of VSV [VSV(FLV)1, N- and B-type mouse-embryo cells, infected with N-FLV and B-FLV, respectively, were superinfected with ts 45 VSV. Analysis of progeny VSV produced in uninfected mouse-embryo cells is seen in Table 3. Stock VSV, wild type or ts 45, grown in N-type mouse-embryo cells when assayed directly on either N- or B-type mouse-embryo cells shows the same titer (Exps. I and II). When assayed on HeLa or L cells, the titer of ts 45 VSV was consistently 0.5 logarithmic units greater (Exp. II). No interference with VSV titer was observed when assayed on mouse-embryo cells infected with N-FLV. Heat treatment reduces the titer 5 logs, and demonstrates clearly the effect of the ts mutation. The same extent of neutralization with antiserum made against Indiana serotype VSV was observed on treatment of ts 45 VSV (Exps. I and III). An additional control of our antiserum reagents showed no neutralization of either wildtype or ts 45 VSV with FLV antiserum. These data show that no immunological differences could be observed between the neutralization antigen of wild-type and ts 45 VSV by this antiserum, and that no VSV neutralizing antibodies could be detected in FLV antiserum. VSV(FLV) should possess the following characteristics: (i) because of the FLV envelope, they should be neutralized by antiserum directed against FLV, (ii) they should demonstrate the host range of FLV, that is, they should show reduced efficiency of plating on either HeLa or L cells, and (iii) their replication in mouse-embryo cells previously infected with FLV should be inhibited due to viral interference (16, 17). The progeny of virus produced by superinfection of mouseembryo cells infected with N-FLV were tested for these traits (Table 4). Unheated virus forms plaques equally on N- and B-type mouse-embryo cells, and, as above, demonstrates a slightly greater titer on HeLa and L cells. These titers mainly represent VSV encapsidated within a VSVspecific coat protein produced at 310. Heat treatment of this mixture reduces the titer on mouse-embryo cells about 3 logarithmic units, and now the heat-resistant virions have the host range of FLV, that is, they have a lower titer on HeLa or L cells, or on mouse-embryo cells infected with FLV. These facts are consistent with the results expected of a VSV(FLV) phenotypically mixed virion. As further evidence that FLV envelope protein is present, no neutralization of infectivity occurs on exposure of heated virions to antiserum against VSV; but a further drop of 2.5 logarithmic units of infectivity is seen when this heat-resistant fraction is also exposed to FLV neutralizing antiserum. These findings clearly demonstrate that VSV virions bearing FLV envelope protein are responsible for the major heat resistant infectivity titer found here on mouse-embryo cells. The point to be stressed, however, is that the titer of VSV(N- FLV) on either N- or B-type mouse-embryo cells is exactly the same. These results (Table 4) are verified when a similar analysis of VSV(B-FLV) is done. Again, no restriction of VSV phenotypically mixed virus is demonstrated by the two types of mouse-embryo cells. These results demonstrate that phenotypically mixed VSV(FLV) virions can be obtained which by neutralization, host range, and interference pattern possess the envelope specificities of FLV. Results of plaque assays on mouseembryo cells demonstrate that VSV(N-FLV) and VSV(B- FLV) are not restricted by either B- or N-type cells, respectively. This finding suggests strongly that the outer FLV envelope protein is not involved in host restriction. DISCUSSION Host restriction of RNA tumor viruses has been described clearly in avian cells with the Rous sarcoma complex (21) and in murine cells with the Friend complex. The biological and genetic characteristics of host restriction in these two systems differ markedly. The allele for susceptibility (permissiveness) to Rous sarcoma virus in chick cells is dominant (22), whereas in mice the allele for resistance (nonpermissiveness) to FLV is dominant (1, 7, 8). Host range of avian viruses is strictly correlated with the serotype of the viral envelope. This is not the case for host restriction of FLV, since no serological differences between the envelope antigens of N-FLV and B-FLV have been demonstrated (9). Host restriction in chicks is virtually absolute (21, 23), whereas restriction of FLV is relative (7, 8, 24). The correlate of this latter fact is the phenomenon of a multiple-hit curve seen on titration of

5 Proc. Nat. Acad. Sci. USA 70 (1973) FLV in restrictive cells compared with the single-hit curve seen in permissive cells. Host restriction in avian cells appears to act at the step of viral penetration of host cells (23, 25, 26). As in chick cells with Rous sarcoma virus, the results presented in this paper demonstrate that FLV adsorbs equally well to permissive and restrictive cell types. We have further shown that VSV(FLV) mixed virions can be prepared as characterized by neutralization, host range, and interference specificity of FLV. The result that these mixed virions form plaques equally well on both N- and B-type mouse-embryo cells suggests strongly that host restriction of FLV is due to an intracellular step in virus replication and does not involve an interaction of FLV envelope and the cell surface. Zavada (16) has reported that only neutralization antigen of MLV is exchanged with VSV in formation of the mixed virion. However, McSharry et al. (27) have demonstrated that two glycoproteins are exchanged in formation of phenotypically mixed virions of VSV and simian virus-5. Since MLV appears to contain at least two envelope glycopeptides (28), it is possible that more than one of the surface proteins is necessary to demonstrate host restriction. Furthermore, it could be argued that VSV mixed virions therefore penetrate cells by a mechanism fundamentally different from that of FLV alone. This conclusion is not supported by our results showing the decreased titer of VSV(FLV) when assayed on mouse-embryo cells infected with FLV. Rather, the viral interference, which presumably reflects interference with penetration of the superinfecting virus, demonstrated by VSV(FLV), implies a similarity of adsorption and uptake with native FLV. These data suggest another interesting fact, that viral interference is associated with the neutralization antigen of the leukemia virus. The observation that both N-FLV and B-FLV form plaques with a one-hit curve on S+L- cells deserves some comment. Since S +L- cells represent B-tropic Swiss 3T3 cells transformed by Moloney sarcoma virus, one might expect that N-FLV would exhibit a reduced efficiency of plating on this cell line. It is possible that the presence of the Moloney sarcoma virus genome somehow alters the cell to permit normal replication. These data are consistent with two possibilities; either that Moloney sarcoma virus genome provides a direct helper effect for intracellular replication of otherwise restricted FLV, or that in these transformed cells expression of the Fv-1 gene is suppressed. These possibilities are testable, and may allow further exploration of the intracellular step that is limiting in the FLV restriction phenomenon. NOTE ADDED IN PROOF Corroborating the result that host restriction is not due to Host Restriction of Friend Leukemia Virus 2553 the virus envelope is the data of Yoshikura, H. (1973) J. Gen. Virol. 19, , with pseudotypes of murine sarcoma virus, and that of Huang, A. S., Besmer, P., Chu, L. & Baltimore, D. (1973) J. Virol., in press, who have independently performed similar experiments with VSV and N-FLV. We thank Drs. Frank Lilly and C. R. Pringle for donating the virus strains used in these experiments, and Dr. Donald Summers who suggested the use of phenotypically mixed virions. Special thanks to Mr. James Hsiang for technical assistance. This work was supported by Contract NIH NCI CP of the Special Virus Cancer Program. We were supported by grants as follows: T.K. is supported by a training grant to the M.D.- Ph.D. Program (5T5-GM ); R.S., National Institutes of Health, AI-TO1-Al , Health Research Council of the City of New York no. I-755, and an NIH Career Development Award, 1-K04-CA ; B.F., National Institutes of Health, AI-TO1-A , American Cancer Society Faculty Research Award No. PRA Lilly, F. (1970) J. Nat. Cancer Inst. 45, Pincus, T., Hartley, J. W. & Rowe, W. P. (1971) J. Exp. Med. 133, Pincus, T., Rowe, W. P. & Lilly, F. (1971) J. Exp. Med. 133, Odaka, T. (1969) J. Virol. 3, Ware, L. M. & Axelrad, A. A. (1972) Virology 50, Stephenson, J. R. & Aaronson, S. A. (1972) Proc. Nat. Acad. Sci. USA 69, Rowe, W. P. & Hartley, J. W. (1972) J. Exp. Med. 136, Rowe, W. P. (1972) J. Exp. Med. 136, Steeves, R. A. & Eckner, R. J. (1970) J. Nat. Cancer Inst. 44, Lilly, F. & Steeves, R. A. (1973), Virology, in press. 11. Bassin, R. H., Tuttle, N. & Fischinger, P. J. (1970) Int. J. Cancer 6, Flamand, A. (1970) J. Gen. Virol. 8, Flamand, A. & Pringle, C. R. (1971) J. Gen. Virol. 11, Bassin, R. H., Tuttle, N. & Fischinger, P. J. (1971) Nature 229, Rowe, W. P., Pugh, W. E. & Hartley, J. W. (1971) Virology 42, Zavada, J. & Zavadova, Z. (1972) Nature New Biol. 240, Zavada, J. (1972) J. Gen. Virol. 15, Murphy, F. A. & Fields, B. N. (1967) Virology 33, Temin, H. M. (1967) J. Cell. Physiol. 69, Bader, J. P. (1972) Virology 48, Hanafusa, H. (1965) Virology 25, Payne, L. N. & Biggs, P. M. (1966) Virology 29, Piraino, F. (1967) Virology 32, Hartley, J. W., Rowe, W. P. & Huebner, R. J. (1970) J. Virol. 5, Steck, F. T. & Rubin, H. (1966) Virology 29, Steck, F. T. & Rubin, H. (1966) Virology 29, McSharry, J. J., Compans, R. W. & Choppin, P. W. (1971) J. Virol. 8, Duesberg, P. H., Martin, G. S. & Vogt, P. K. (1970) Virology 41,