The RNA Tumor Viruses Background and Foreground

Proc. Nat. Acad. Sci. USA Vol. 69, No. 4, pp. 1016-1020, April 1972

HOWARD M. TEMIN McArdle Laboratory, Madison, Wisconsin 53706

ABSTRACT The members of the RNA tumor virus (or cells at mitosis. The structure containing the viral information leukovirus) group of animal viruses replicate via a DNA intermediate and transmit their information stably in is called the provirus. In rat cells infected with Rous sarcoma cells as DNA. Although some of these viruses are capable of virus, there is no production of virus, and virus production inducing neoplastic transformation, others are not. These cannot be induced by superinfection with another virus. viruses may be related to Spontaneous neoplasia, but the However, the entire viral genome remains in the infected relationship is that of analogy rather than etiology. The cell, and infectious virus appears after relationship of these viruses to human disease is unknown. fusion of the cell with Two models of spontaneous neoplasia that involve either chicken cells. [As will be discussed later, it appears that some activation of the information of a transforming RNA abnormality in viral RNA synthesis in the infected rat cells tumor virus or the creation of information for neoplastic prevents virus production (2) ]. transformation by recombination involving information Much of the recent excitement involving RNA tumor transfer from RNA to DNA are being tested in model systems. viruses relates to the novel mode of information transfer that appears to be a part of their replication. A model for In this paper, I shall discuss the known properties of RNA this replication is summarized in Fig. 1. Information is tumor viruses and some of the problems requiring further transferred from viral RNA to DNA, the provirus. This study. RNA tumor virus is the name of a group of animal DNA is then a template for the production of progeny viral viruses, also called leukoviruses, that were discovered in RNA. When an infected cell divides, the DNA provirus also 1908 by Ellerman and Bang (who found a virus causing a divides. This division explains the maintenance of viral leukemia in chickens) and in 1911 by Peyton Rous (who information in dividing, infected cells. The evidence for the found a virus causing a sarcoma in chickens). Leukoviruses are existence of the DNA provirus is still indirect, and is also widely distributed in nature. They are associated with summarized in Fig. 1. Inhibitor experiments show that tumors in various animals, not yet shown to include man, virus production is blocked by actinomycin D, an inhibitor and are also very widely distributed in normal chickens and of DNA-directed RNA synthesis, and that infection is pre- mice. A virion of a leukovirus is defined as being medium vented by inhibitors of DNA synthesis, but not by inhibitors sized (80-100 nm), with an envelope containing a nucleo- of protein synthesis. After a provirus has been formed, capsid with no clearly observable symmetry and single- further DNA synthesis is not required for virus production. stranded RNA of 60-70S size. Additional characteristics are: The most convincing evidence for the physical existence of the the replication of the virus requires early DNA synthesis and DNA provirus are experiments, reported by Boettiger and is sensitive to actinomycin D, the sedimentation coefficient Temin (3) and by Balduzzi and Morgan (4), indicating that in of the RNA changes to about 35 S after a brief treatment cultures of stationary cells a provirus can be labeled with with denaturing conditions, and the virion contains a DNA bromodeoxyuridine and thereby rendered sensitive to in- polymerase. [I have recently written a review of these activation with visible light. Under such conditions, cell DNA viruses, which contains references to most work published is not labeled with the bromodeoxyuridine. Therefore, before 1971 (1).] The virions of the RNA tumor viruses are somewhat similar to the virions of myxoviruses. Myxovirus virions are also medium sized, enveloped, and contain single- stranded RNA. However, the nucleocapsids of myxoviruses Inhibition of Inhibition of are clearly of helical symmetry, and the RNA is either about Protein Synthesis DNA, Synthesis 55 S (paramyxoviruses or myxovirus II) or disperse (ortho- myxovirus or myxovirus I). RNAage&FM b DNA,- I *RNA..,RV Replication of RNA tumor viruses

The replication of the classical RNA tumor viruses does not Inhibitors of Actinomycin D lead to cell death, as does the replication of most other DNA Synthesis animal viruses. Instead, the infected cell divides. After infection by the prototype RNA tumor virus, Rous sarcoma virus, all infected cells appear to become transformed into neoplastic Infecting Provirus Progeny cells. These cells may or may not produce virus. However, Virus Virus both in cells producing virus and in cells not producing FIG. 1. Replication of Rous sarcoma virus (RSV). The viral virus, the information of the virus is transmitted to daughter replication may or may not involve neoplastic transformation. 1016 Downloaded by guest on September 26, 2021 Proc. Nat. Acad. Sci. USA 69 (1972) RNA Tumor Viruses 1017

(0 w U 60. w

40- -I (10 20 (jQ4) titer - (104(104) -viru~s

0- 20 40 60 80 100 120 140 160 180 HOURS AFTER INFECTION FIG. 2. Activation of Rous sarcoma virus synthesis. Stationary chicken cells were exposed to Schmidt-Ruppin strain of Rous sarcoma virus at a multiplicity of infection of 1 focus-forming unit/cell. At different times after infection, the cells were examined for the presence of avian tumor virus antigens by the use of fluorescent antibodies. At the same times, the medium was assayed for the amounts of virus present. At the times indicated by arrows, serum was added to some cultures to cause the cells to multiply (Humphries and Temin, un- published).

irradiation with light does not kill the cells, but it does One of the most striking aspects of this transformation is prevent the infection. In Boettiger's experiments (3), it was the difference in multiplication between transformed and further shown that the shape of the killing curve depended normal cells. This difference in multiplication appears to be upon the initial multiplicity of infection, indicating that the related to an increased efficiency of the transformed cells in the input virus controls the number of DNA proviruses. utilization of specific multiplication-stimulating factors in Because of the failure of inhibitors of protein synthesis to serum (19, 20). One of these factors has been purified about block DNA provirus formation, we looked for and found 7000-fold from calf serum (21) (Fig. 3). RNA-directed DNA polymerase activity in virions of Rous Many of these features of the replication of Rous sarcoma sarcoma virus (5). Independently, Baltimore found a similar virus, including such properties of the virion as the presence polymerase activity in virions of a murine leukemia virus (6). of a DNA polymerase and a DNA intermediate for replication, The DNA polymerase of the leukovirus virion requires for are shared by all the other members of the classical RNA full activity a detergent or some other means of disruption of tumor virus group. However, other leukoviruses differ from the virion, a divalent cation (such as magnesium), four Rous sarcoma virus in the efficiency of transformation. For deoxyribonucleoside triphosphates, and a buffer. The polymerase activity is sensitive to treatment with ribonuclease, indicating that the template is RNA. The polymerase is 5 a located in the virion core (7, 8). The product of the polymerase 4 reaction is DNA, as shown by its chemical and physical 3 behavior. This DNA product is complementary to viral 2 RNA, as shown by nucleic acid hybridization experiments 1 IlII (9, 10). Polynucleotide ligase and other nucleic acid-related F R@L ii enzymes are also present in the virion (11, 12). These results z suggest that the virion transfers information from RNA to > 4 - integrated DNA (13).

After exposure of cells to Rous sarcoma virus, some cellular i I -| processes are needed before virus production begins (Fig. 2). If the cells are in the G1-b phase of the cell cycle (14), exposure 0I in to virus will not lead to virus production or transformation. Nt 5 If serum is added, so that the cells enter into mitosis, viral 5-F- c antigens appear in the cells, virus production starts, and the 4 cells become transformed. This requirement for processes in 3 the mitotic cell cycle has not been found for replication of 2 most other groups of animal viruses. 1 One result of the activated infection is to transform infected I- cells into neoplastic cells, which differ in their morphology 0 1 2 3 4 5 6 7 8 and growth properties from normal cells (see ref. 1 for refer- MIGRATION DISTANCE cm ences). This neoplastic transformation appears to be con- FIG. 3. Tracing of gel electrophoregrams of (a) whole serum, trolled by genes in the virus, as shown by morphological and (b) partially purified multiplication-stimulating activity, and temperature-sensitive mutants for this conversion (15-18). (c) insulin (from ref. 21). Downloaded by guest on September 26, 2021 1018 ,Temin Proc. Nat. Acad. Sci. USA 69 (1972)

TABLE 1. Viruses with RNA and a DNA polymerase virus, such as rat cells infected by Rous sarcoma virus, an in their virions RNA-directed DNA polymerase system can also be isolated. This preparation shares certain properties with precursors of Nonlethal to cells the Rous sarcoma virus virion core, but, by nucleic acid hybrid- C-type ization experiments, does not appear to use virion RNA as strongly transforming (Rous sarcoma virus) a template (2). There is some relationship between the RNA weakly transforming (lymphoid leukosis virus) nontransforming (avian erythroblastosis strain R-associated of these particles from rat cells infected with Rous sarcoma virus) virus and RNA from uninfected rat cells, as shown by an- B-type (mouse mammary tumor virus) nealing experiments with the DNA product of the cellular Lethal to cells in usual host RNA-directed DNA polymerase system. In addition, there is Visna an RNase-sensitive DNA polymerase system in uninfected rat Syncytium-forming cells. This DNA polymerase system in normal rat cells is similar to a virion DNA polymerase system in its requirement for all four deoxyribonucleoside triphosphates, its sensitivity to ribonuclease, and its synthesis of a DNA product (2). instance, avian myeloblastosis virus does not usually trans- Unfortunately, it is not easy to interpret the RNA-directed form fibroblasts, but can transform precursor cells from the DNA polymerase systems that are apparently not related to reticuloendothelial system. Avian lymphoid leukosis virus virus and are in uninfected or virus-infected cells. Two types does not transform any cells in culture; in animals, it causes of interpretation can be imagined. One is that these polymerase tumors only at low frequencies and after long latent periods. particles are protoviruses or other precursors of induced leuko- Avian erythroblastosis strain R-associated virus does not viruses (27), the other is that they are some kind of artifact seem to cause any tumors in infected animals (22). Therefore, caused by the binding of a cellular DNA polymerase to RNA. one may speak of strongly transforming, weakly transforming, The reason the latter hypothesis has been given some credence and nontransforming leukoviruses (13). The induced leuko- is that recent work [see Mizutani and Temin (12)] has shown viruses reported by Weiss et al. (23) and discussed by W. P. that virions of RNA tumor viruses contain various enzymes, Rowe and G. J. Todaro (29, 31) in this symposium are probably for example hexokinase, that do not appear to be related to other examples of nontransforming leukoviruses. viral replication, but may be bound to virion components. A similar binding might occur in cells and result in formation RNA-directed DNA polymerases of a particulate RNA-directed DNA polymerase system. After a DNA polymerase was found in virions of Rous Etiology of "spontaneous" neoplasia sarcoma virus and Rauscher murine leukemia virus, the How do these experimental results relate to the question virions of many other RNA viruses were examined for the of the etiology of human neoplasia? We do not know the presence of a DNA polymerase. RNA viruses containing a etiology of most human neoplasias. A number of hypotheses DNA polymerase in their virion can be separated into the to and some not. classes listed in Table 1. Some of these viruses, like Rous can be proposed, some related RNA viruses, Those hypotheses not related to RNA viruses include classical sarcoma virus, are able to induce tumors; others are notable to induce tumors but have been isolated from tumors (many Still others appear not to induce murine C-type viruses). Protovirus Rous Sarcoma Virus tumors and have not been isolated from tumors. This dis- m' a a x tribution suggests that the presence of a DNA polymerase in m~~~~~~~~~~~~~ a virion of an RNA virus does not necessarily mean that the virus containing it causes neoplastic transformation. A virion DNA polymerase might have a selective advantage RNA-directed-- as a means for a nontransforming virus to establish a latent ago DNA Synthesis infection (24). The presence of a virion DNA polymerase in an RNA virus does not mean that the virus does not kill cells; both visna and the syncytium-forming viruses kill EDNAvDNI cells in their normal hosts. RNA After the finding of DNA polymerases in the virions of RNA tumor viruses, a search was started for related DNA polymerases in infected and in uninfected cells. This search 01 1 was both for soluble enzymes and for particulate polymerases. a ox ay KLa0oAVVVV The work with soluble enzymes is difficult to interpret because, as first noted by Lee-Huang and Cavalieri (25), all DNA Oncogene Derepression Tumor Cell polymerases appear to have the capacity to use RNA as a FIG. 4. of transformation. A tumor A virus has been from two Mechanisms neoplastic template. polymerase separated cell is assumed to have regions of DNA (a to x) specifying syn- cellular DNA polymerases in virus-producing cells from thesis of RNA that leads to neoplastic transformation. In trans- mice (26). formation by Rous sarcoma virus, this DNA is synthesized The work with particulate enzymes is also somewhat diffi- from the viral RNA. In the oncogene hypothesis, this DNA cult to interpret. RNA-directed DNA polymerase systems that exists in all cells and is derepressed during neoplastic transforma- appear to be precursors of virions are found in infected, virus- tion. In the protovirus hypothesis, this information is created by producing cells (8). In infected cells that do not produce successive RNA to DNA to RNA information transfers. Downloaded by guest on September 26, 2021 Proc. Nat. Acad. Sci. USA 69 (1972) RNA Tumor Viruses 1019

genetic hypotheses, epigenetic hypotheses, and hypotheses stationary infected chicken cells, there is some suggestion involving DNA viruses. Those hypotheses related to RNA that there may be changes in transport of viral-specific RNA viruses include hypotheses of vertically transmitted virions from the cell nucleus. Study of these two systems suggests requiring a sensitive target cell for neoplastic transformation, that we need to understand more about the control of RNA and the oncogene and protovirus hypotheses (Fig. 4). synthesis in animal cells, especially whether the control For neoplastic transformation by Rous sarcoma virus, there of RNA synthesis by a provirus of an RNA tumor virus must be formation in the cell, by means of RNA-directed or by a related element is the same as that of other RNA DNA synthesis, of the genes for viral replication and of the synthesis. The other approach we have been taking is to genes for neoplastic transformation. The expression of these search for protoviruses in uninfected cells (2). two sets of genes can be completely separate, as shown by It might be worth while to consider briefly the induction virus-producing untransformed cells and by cells that do not phenomenon discussed by Rowe and Todaro (23, 29-31). produce virus. In the oncogene hypothesis, the entire in- This phenomenon has several noteworthy features: the cells formation of a transforming leukovirus, including that for that are induced may not be immune to the induced virus, the neoplastic transformation, is transmitted in the germ cells. induced cells are killed, and the frequency of induction seems In the protovirus hypothesis, only some information related to increase with long passage of cells, i.e., in cell lines. These to the information for neoplastic transformation is trans- features are more compatible with a genetic hypothesis of mitted in the germ cells. The oncogene and protovirus virus induction, like the protovirus hypothesis, than with a hypotheses also differ in the nature of the changes that lead derepression hypothesis, like the oncogene hypothesis. How- to neoplasia. In the oncogene hypothesis, the change is ever, since the virus induction does not lead to concomitant derepression of pre-existing information. In the protovirus transformation, this phenomenon is only a model for neo- hypothesis, the change is a series of successive RNA to DNA plastic transformation. to RNA information transfers that lead to modification of the We have been discussing the status of the scientific study original information. This modification is usually for physio- of RNA tumor viruses. It is clear that much more work can logical processes, but the influence of carcinogenic agents can be done and needs to be done before we understand the lead to misevolution and the production of the information replication of these viruses and their relationship to human for neoplastic transformation. (Other evolutions could lead neoplasia. Now, we should also ask the question whether to formation of leukoviruses.) there should be an expansion of cancer research. I think the How does the work with leukoviruses relate to the testing answer to this question depends upon the source of the money of these hypotheses? A virus like Rous sarcoma virus seems to to support this expansion. If this money is taken from the contain separate sets of genes for neoplastic transformation defense budget, fine. An expansion of cancer research which and for viral replication. Infection by Rous sarcoma virus causes a diminution of the money spent on defense would be a leads to neoplastic transformation as a result of the formation double benefit to our society. However, if the monev for an in cells of the genes for viral replication and for neoplastic expansion of cancer research comes at the expense of the vast transformation. The processes that virologists know how to domestic social needs that are tearing our society apart, I study are those related to viral replication. Therefore, most oppose an expansion of cancer research. Other problems of studies of models of the etiology of human neoplasia relate to our society, I feel, are more important at present. Even in the viral replication and not to neoplastic transformation. How- field of biological research, work on the pollution of the bio- ever, we must remember that viral replication is not the sphere and on population control is more important than an same as neoplastic transformation. As later talks at this expansion of cancer research. symposium will point out, there are many approaches: looking for viral nucleic acids, for viral-related enzymes, and for the appearance of viruses from normal cells. These approaches The work from my laboratory was supported by Public Health viral replication and Service Grant CA-07175 from the National Cancer Institute, are important, but they are looking at and by American Cancer Society Grant VC7. I hold Research information for replication, rather than at information for Career Development Award 1OK3-CA-3182 from the National neoplastic transformation. Cancer Institute. Our experiments are subject to the same criticism. We have been studying two processes related to the start of viral replication. These involve the induction of virus production 1. Temin, H. M. (1971) Annu. Rev. Microbiol. 25, 609-648. from rat cells infected and transformed by Rous sarcoma 2. Coffin, J. & Temin, H. M. (1971) J. Virol. 8, 630-642. virus and the production of virus from stationary chicken 3. Boettiger, D. & Temin, H. M. (1970) Nature 228, 622-624. 4. Balduzzi, P. & Morgan, H. R. (1970) J. Virol. 5, 470-477. cells. A rat cell infected with Rous sarcoma virus does not 5. Temin, H. M. & Mizutani, S. (1970) Nature 226, 1211-1213. contain virus, nor precursors of the virus, but it does contain 6. Baltimore, D. (1970) Nature 226, 1209-1211. small amounts of the virion group-specific antigen (about 7. Gerwin, B. I., Todaro, G. J., Zeve, V., Scolnick, E. M. & 2% as much as is present in virus-producing chicken cells) and Aaronson, S. A. (1970) Nature 228, 435-438. study of the nucleic 8. Coffin, J. & Temin, H. M. (1971) J. Virol. 7, 625-634. of virion nucleic acids. Quantitative 9. Spiegelman, S., Burny, A., Das, M. R., Keydar, J., Scholm, acids in these cells by the use of nucleic acid hybridization J., Travnicek, M. & Watson, K. (1970) Nature 227, 563-567. suggests that viral RNA is produced in much smaller amounts 10. Duesberg, P. H. & Canaani, E. (1970) Virology 42, 783-788. in these cells than in virus-producing chicken cells (2). 11. Mizutani, S., Temin, H. M., Kodama, M. & Wells, R. D. Virus production after fusion can be explained by the pas- (1971) Nature New Biol. 230, 232-235. that 12. Mizutani, S. & Temin, H. M. (1971) J. Virol. 8, 409-416. sage from chicken cells to the rat cells of something 13. Temin, H. M. (1972) in The RNA Viruses and Host Genome enables the rat cells to synthesize this missing class of in Oncogenesis, ed. Emmelot, P. & Bentvelzen, P. (North- RNA. In the case of activation of virus production from HoUand Publ. Co., Amsterdam), 351-362. Downloaded by guest on September 26, 2021 1020 Temin Proc. Nat. Acad. Sci. USA 69 (1972)

14. Temin, H. M. (1971) J. Cell. Physiol. 78, 161-170. 24. Temin, H. M., "Triangle," The Sandoz Journal of Medical 15. Temin, H. M. (1960) Virology 10, 182-197. Science, in press. 16. Martin, G. S. (1970) Nature 227, 1021-1023. 25. Led-Huang, S. & Cavalieri, L. F. (1963) Proc. Nat. Acad. 17. Biquard, J. M. & Vigier, P. (1970) C. R. H. Acad. Sci. 271, Sci. USA 50, 1116-1122. 2430-2433. 26. Ross, J., Scolnick, E. M., Todaro, G. J. & Aaronson, S. A. 18. Friis, R., Toyosbina, R. K. & Vogt, P. K. (1971) Virology (1971) Nature New Biol. 231, 163-167. 43, 375-389. 27. Temin, 11. M. (1971) J. Nat. Cancer Inst. 46, III-VII. 19. Temin, H. M. (1969) J. Cell. Physiol. 74, 9-15. 28. Altaner, C. & Temin, H. M. (1970) Virology 40, 118-134. 20. Temin, H. M. (1970) J. Cell. Physiol.-75, 107-120. 29. Lowy, D. R., Rowe, W. P., Teich, N. & Hartley, J. W. 21. Pierson, R. W., Jr. & Temin, H. M., J. Cell. Physiol., in (1971) Science 174, 155-156. press. 30. Aaronson, S. A., Todaro, G. J. & Scolnick, E. M. (1971) 22. Ishizaki, R. & Shimizu, T. (1970) Cancer Res. 30, 2827-2831. Science 174, 157-159. 23. Weiss, R. A., Friis, R. R., Katz, E. & Vogt, P. K., (1971) 31. Klement, V., Nicolson, M. 0. & Huebner, R. J. (1971) Virology, 46, 920-938. Nature New Biol. 234, 12-14. Downloaded by guest on September 26, 2021