Mar 15, 1983 - were selected and pooled for these experiments to provide an immunoreagent which would react with all the viral early antigens of region 1.
Vol. 46, No. 3
JOURNAL OF VIROLOGY, June 1983, p. 1039-1044
0022-538X/83/061039-06$02.00/0 Copyright © 1983, American Society for Microbiology
Transformation of Rodent Cells by DNA Extracted from Transformation-Defective Adenovirus Mutants DAVID T. ROWE* AND FRANK L. GRAHAM Departments of Pathology and Biology, McMaster University, Hamilton, Ontario, Canada L8S 4K1
Received 22 November 1982/Accepted 15 March 1983
Complementation group II host range mutants of adenovirus type 5 which map in early region 1B (E1B, 4.5 to 11.0 map units) have been shown to be defective for the synthesis of the E1B 58,000-dalton (58K) antigen in infections of HeLa or KB cells (Lassam et al., Cell 18:781-791, 1979) and unable to transform cultured rodent cells (Graham et al., Virology 86:10-21, 1978). In this report we show that DNA extracted from group II mutants hr6 and hr5O can transform rat cells with the same efficiency as wild-type DNA. Furthermore, group II mutant-transformed hamster cells were shown to contain no detectable E1B 58K tumor antigen but were capable of inducing tumors in newborn hamsters. Hamster cell lines 1019-3 and 1019-C3, transformed by hr5O DNA, produced no detectable quantities of either the E1B 58K or 19K antigen but nonetheless exhibited a fully transformed oncogenic phenotype. Our results show that the E1B 58K antigen is not absolutely required for oncogenic transformation and suggest that even cells lacking the 19K protein can be oncogenic. Early region I (El) of adenovirus type 5 (Ad5), located within the left 12% of the genome, contains the viral genes necessary and sufficient for transformation of cultured rodent cells (12, 14, 20). Host range (hr) mutations in this region have been isolated (22) by screening mutants for differential plaquing efficiency on HeLa cells and 293, a transformed human embryonic kidney cell line which contains the left 12.5% of the AdS genome and expresses viral mRNA from EI (1, 17). These hr mutants define two complementation groups (I and II) (22) which map in the nonoverlapping transcriptional units ElA (1.5 to 4.5 map units) and E1B (4.5 to 11.0 map units), respectively (3, 5, 11, 13). The complementation group I class of AdS hr mutants is phenotypically deficient in viral DNA and late protein synthesis (26). The mutations appear to affect one or more gene products which stimulate synthesis of viral mRNA from other early regions, including E1B, and thus this region may not be directly involved in DNA replication or control of late gene expression (2, 32). Group I mutants are defective in the transformation of most rodent cells but will induce a semiabortive or abnormal transformation of primary rat kidney cells (16, 35). In contrast, the hr group II mutants are not defective in viral DNA synthesis or late gene expression (26) but are defective in the synthesis of the 58,000-dalton (58K) major tumor antigen as determined by immunoprecipitation and polyacrylamide gel electrophoresis (27; our unpublished data). The hr mutations, most of which
were generated by chemical mutagenesis (22) and probably represent single base changes, have been mapped to the sequences of the E1B 22S message (13) which are spliced out of the E1B 13S mRNA. The 13S and 22S messages both direct the synthesis of the 19K major tumor antigen, but the 22S message also encodes the 58K antigen by utilizing an internal translation initiation site for a different reading frame which partially overlaps the coding sequences for the 19K antigen (5, 21, 31, 33, 37, 38). For hr7, the one mutant which has been studied by S1 nuclease gel analysis, the pattern of early cytoplasmic RNA expressed from region E1B in nonpermissive cells is indistinguishable from that of wildtype (wt) virus (2). Thus, the reason for the failure of group II mutants to synthesize detectable 58K protein is not immediately clear. It remains possible that during group II mutant infections a form of the 58K protein or a truncated version of this antigen is synthesized which does not react with the battery of sera we have used to detect it. Another possibility is that this antigen is rapidly degraded in group II mutantinfected nonpermissive cells. In any case, defective 58K synthesis together with the failure of the group II mutants to transform any of the cell types tested so far (16) suggest a role for the 58K antigen in adenovirus-induced transformation. In contrast to these results with the group II mutant virions, studies on the biological activity of purified viral DNA fragments have shown that only the left-hand 8% (Hindlll G fragment)
1039
1040
J. VIROL.
NOTES
of the Ad5 genome is needed to transform rat kidney cells (14). These fragment-transformed cells do not synthesize an immunoprecipitable 58K antigen (37; our unpublished data) but are otherwise similar in growth and morphology to cells transformed by larger fragments or by wt virus. In particular, HindIlI G fragment-transformed hamster cells, like hamster cells transformed by larger DNA fragments, are capable of inducing tumors after injection into newborn hamsters (D. T. Rowe and F. L. Graham, unpublished data). These results would suggest that the sequences which code for the carboxyl portion of the 58K antigen are not necessary for transformation in vitro, in apparent contradiction to the conclusions derived from studies with group II hr mutants. To resolve this apparent paradox we have postulated that initiation of transformation by viral DNA proceeds through a mechanism slightly different from that operating during virion-mediated transformation (34). Evidently the latter process requires the 58K function, whereas DNA-mediated transformation is independent of 58K. An obvious prediction of this model is that DNA extracted from group II mutant virions should be as efficient in transformation as DNA derived from wt virions. The results shown in Fig. 1 indicate that this is indeed the case. After transfection of baby rat kidney cells with DNA from wt or group II mutant hr6 or hrSO, the frequency of transformation increased in direct proportion to the amount of DNA applied to the cells, and DNA from both hr6 and hr5O mutants transformed rat cells with the same efficiency as wt DNA. Thus the block to transformation by group II mutant virions disappears when transformation is assayed with purified DNA. Group II mutant DNA was also capable of converting primary baby hamster kidney cells to morphologically fully transformed cells, a finding which made it possible to study the oncogenicity of group II mutant-transformed hamster cells in newborn hamsters. Table 1 shows the tumorigenicity of six lines measured in terms of the fraction of positive animals and the time elapsed between injection and appearance of palpable tumors. After injection of some cell lines, only a few of the animals produced tumors and then only after long latent periods. One cell line, 1019-3, and another line, 1019-1T, derived from a 1019-1 tumor, appeared to be relatively tumorigenic, as most of the animals injected bore tumors in 3 weeks or less, though the tumors were smaller and grew more slowly than the tumors of animals injected with wttransformed cells. Thus, although the tumorigenicity of some group II mutant-transformed cells was low, all lines tested were capable of inducing tumors, and since not all wt-transformed
/
cn) 40 _-
,hr5o
LF
z~0z 20
O or
l
l
0
2
4
-
{o l
6
8
/ug DNA/dish FIG. 1. Transformation of primary baby rat kidney cells by Ad5 DNA purified from wt, hr6, and h5 virions. Baby rat kidney cells in 60-mm petri dishes were transformed with DNA by using the calcium technique (18, 19) with modifications (15). Specifically, viral DNA which had been digested with restriction endonuclease Xhol to eliminate infectivity was mixed with high-molecular-weight carrier DNA (rat embryo DNA) at 10 I,g/ml in 250 mM CaCI2. The DNA in CaCF2 was then added slowly to 2x concentrated HEPES (N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid)-buffered saline (18) with bubbling (39). After 30 min, the DNA-calcium phosphate coprecipitate was added to baby rat kidney cell cultures at 0.5 ml of suspension per culture dish containing 5 ml of medium. After 20 h of incubation at 37°C, the medium was changed (a minimal essential medium plus 10% fetal bovine serum), and the cells were incubated a further 2 days at which time the medium was changed to selective medium (Joklik modified minimal essential medium plus 5% horse serum) as described previously (19). Negative controls (cultures treated with carrier DNA and untreated cultures) gave no colonies. The results are expressed as colonies per dish after 12 days of incubation in selective medium.
lines are tumorigenic (c.f. 268-C3), group II mutant-transformed cells may possess the same range of oncogenic potential as wt-transformed hamster cell tines. Though it is difficult to draw firm conclusions from in vivo studies involving a limited number of animals, these results suggest a role for the 58K coding sequences chiefly in initiation of transformation by viruses (but not by viral DNA). Thereafter, expression of 58K may not play any major role in determining the transformed phenotype and in particular may
NOTES
VOL . 46, 1983 TABLE 1. Tumorigenicity of wt- and hrtransformed hamster cells Tumorigenicitya Latent Fraction Transformed by: Cell line (days) positive
1019-Cl 1019-1 1019-2
1019-iT 1019-C3 1019-3
hr6 DNA hr6 DNA hr6 DNA (derived from tumor induced by line 1019-1) hr5O DNA hrSO DNA
14b AdS sl4, virusb 972-1 AdS, XhoI C DNA 983-2 AdS, XhoI C DNA 268-C3 AdS, sheared DNA 983-4 AdS, XhoI C DNA 297-C43 AdS, sheared DNA a Subcutaneous injection of 5 born hamsters. b See reference 40.
x
1/19 4/20 3/10 7/11
100 25 110 16
4/24 16/20
90 16
20/20 2/9 9/12 0/13
12 40 27
9/12 13/22
27 25
106 cells into
new-
not be needed for tumorigenicity of cells transformed in vitro by DNA. Characterization of serum from tumor-bearing hamsters by immunoprecipitation of AdS antigens from 35S-labeled KB cells early after wt infection (Table 2) revealed that none of the sera raised against hr mutant-transformed cells contained antibodies to the 58K antigen. Most of the sera reacted with the 19K El protein or the 14K E4 protein (8), and antisera from hamsters injected with 1019-3 cells appeared to be specific for the 44K ElA protein. The possibility that sera from group II mutant tumor-bearing hamsters contain antibodies to a 58K-like antigen but do not react with the wt 58K seems remote. When these sera were used to immunoprecipitate early antigens from hr6-infected cells, no novel antigen was observed upon polyacrylamide gel electrophoresis and autoradiography (data not shown), suggesting that at least this mutant does not induce the synthesis of a stable truncated form of 58K. The absence of a 58K antigen from group II mutant-infected cells (27) and the absence of antibodies to the 58K antigen in group II mutanttransformed cell anti-tumor sera suggests that the phenotype of the group II mutants in infected cells (lack of 58K protein) was likely expressed in the transformed hamster cells. To test this, and to ensure that the cells had indeed been transformed by group II mutants and not by wt revertants, six mutant-transformed lines were assayed for the presence of tumor antigens by immunoprecipitation with anti-tumor serum ca-
1041
pable of reacting with 58K and other El antiThe autoradiogram in Fig. 2 shows that cells transformed by hr DNA fragments or by fragments of wt DNA which exclude the 3' part of E1B (954-1, 945-Cl) fail to synthesize the E1B 58K antigen which is readily detectable in cells transformed by wt DNA fragments containing all of El (983-2, 972-3). hr6-transformed hamster cell lines expressed the E1B 19K antigen, which has been detected in all cell lines transformed by wt virus or DNA fragments and infected cells (Fig. 2; our unpublished results). On high-resolution polyacrylamide gels, the 19K band has been shown to be composed of two proteins which share a common amino acid sequence (D. T. Rowe, F. L. Graham, and P. E. Branton, submitted for publication). Somewhat surprisingly, the hrSO-transformed cell lines 1019-3 and 1019-C3 produced no detectable quantities of either the 58K or 19K antigen, consistent with the specificity of tumor sera obtained from hamsters injected with these transformed cells. The present study does not rule out the possibility that E1B antigens are made and then rapidly degraded in the group II mutant-transformed cells. Although the 58K antigen has never been detected in hr6- or hr5O-infected KB cells, even with very short labeling times, both of these mutants induce the synthesis of an immunoprecipitable 19K antigen in nonpermissive infections (27; our unpublished data). Thus, hrSO-. transformed cell lines 1019-3 and 1019-C3 exhibited the unique phenotype of being fully transformed and relatively tumorigenic in the absence of any detectable E1B antigens. Our results are not in agreement with the findings of Jochemsen et al. (24), which show that Adl2 HindIII G (0 to 7.2%) fragmenttransformed rat cells are unable to induce tumors in nude mice. It is possible that the
gens.
TABLE 2. Specificity of hamster anti-tumor seraa Cell line
Transformed by:
Early antigen 58
44
29
19
14b
++ + 1019-1 hr6 ++ + + 1019-1T hr6 1019-Cl hr6 ++ 1019-2 hr6 ++ 1019-3 hr5O hrSO ++ 1019-C3 ++ + 983-2 wt a AdS early antigens present in 10' KB cells labeled with 50 ,uCi of [355]methionine from 8 to 10 h postinfection were detected with 30 R1 of anti-tumor scrum by using the Sepharose-protein A immunoabsorption technique (36). Major avidity is indicated by ++. Lesser avidity is indicated by +. bAn E4 encoded protein (8).
1042
J. VIROL.
NOTES INF
9 5 4 9 "4 1019 1019 1019 1019 1019 - T C 1 -2 -3 Ir-D
3-li __
. -v1
c.
VW. .>
-
.a v -Vc
%70
o. -Vnt
19-a
-Vlii -ix 14-
-
FIG. 2. Autoradiograph of a 14% polyacrylamide gel (25) showing the separation of Ad5 early viral antigens from infected and transformed cells. To detect antigens early during lytic infection, approximately 10' Spinnergrown KB cells were infected with 50 PFU of AdS per cell and labeled with 100 jLCi of [35S]methionine at 10 to 12 h postinfection. For transformed cells, approximately 108 cells in petri dishes were labeled for 8 h with 150 FCi of [35S]methionine in 3 ml of methionine-deficient Eagle medium. The immunoprecipitation procedure, using the protein A-Sepharose technique (36) with modifications, has been described (27). A number of tumor antisera were selected and pooled for these experiments to provide an immunoreagent which would react with all the viral early antigens of region 1. The molecular weights of the early antigens detected in the infected cells (INF) appear on the left, and structural proteins of the virion (V) are indicated by Roman numerals on the right. Hamster cell lines 954-1 and 954-5 were transformed by the HindIII G (0 to 7.8%) fragment of Ad5 DNA; 1019-1, 1019-Cl, and 1019-2 are hamster lines transformed with purified hr6DNA; and 1019-1T is a line derived from a 1019-1-induced tumor. Lines 1019-3 and 1019-C3 were transformed by hr5O DNA. Lines 972-3 and 983-2 were transformed with the XhoI C (O to 15%) fragment of wt DNA.
different conclusions are due to inherent differences in the oncogenicity testing systems employed in the two studies. Bernards et al. (4) also examined the oncogenicity in nude mice of rat cells transformed by Ad5-Adl2 hybrid El plasmids and have suggested that the oncogenicity of adenovirus-transformed cells is determined by products specified by region E1B. Although a role for the E1B proteins in determining the degree of tumorigenicity cannot be ruled out, our studies indicate that the 58K E1B antigen is not absolutely required for oncogenic transformation and suggest that even cells lacking the 19K protein can be oncogenic. There are superficial similarities between the results we have obtained here and biochemical and genetic studies of polyoma DNA transformation. Recent evidence suggests that polyoma DNA fragments encoding only the small and
middle T antigens are sufficient for transformation in vitro (23, 29) and for tumorigenicity (28), although it has been established that tsA mutants of polyoma virus, which affect only large T antigens, are greatly impaired in their ability to transform rat and hamster cells at nonpermissive temperatures (7, 9, 10). Large T antigen appears to promote integration of viral DNA in tandem arrays, a mode of insertion which increases the efficiency of transformation (6). The formation of tandem head-to-tail integration is suggested to occur through aberrant replication intermediates present in infected cells (6), and since large T antigen is required to initiate DNA replication (9), mutants of large T antigen which fail to synthesize DNA would ultimately be transformation deficient. This particular model does not appear to be applicable to adenovirus 58K antigen involvement in cell transformation since the
NOTES
VOL. 46, 1983
58K protein does not appear to be required for viral DNA replication. Nevertheless, a role for the 58K antigen in the initial step of integration cannot be ruled out, and indeed such a role seems likely. The group II mutant hr6 supports adeno-associated virus DNA replication but is unable to rescue the integrated form of the adeno-associated viral genome in human Detroit 6 cells latently infected with adeno-associated virus (30). This finding suggests that the mechanisms by which viral information is integrated into or excised from the cellular genome during adenovirus infection may be influenced by viral gene products from E1B, a model which is consistent with the results presented here. We are grateful to J. Rudy and J. Ng for excellent technical assistance. This work was supported by grants from the National Cancer Institute of Canada. D.T.R. is a research student and F.L.G. is a research associate of the National Cancer Institute of Canada.
1.
2.
3. 4.
5.
6.
7.
8.
9.
10. 11.
12.
LITERATURE CITED Aiello, L., R. Guilfoyle, K. Huebner, and R. Weimnann. 1979. Adenovirus 5 DNA sequences present and RNA sequences transcribed in transformed human embryo kidney cells. Virology 96:460-469. Berk, A. J., F. Lee, T. Harrison, J. Williams, and P. A. Sharp. 1979. Pre-early adenovirus 5 gene product regulates synthesis of early viral messenger RNAs. Cell 17:935-944. Berk, A. J., and P. A. Sharp. 1978. Structure of the adenovirus 2 early mRNAs. Cell 14:695-711. Bernards, R. A., A. Houweling, P. Schrier, J. L. Bos, and A. J. van der Eb. 1982. Characterization of cells transformed by AdS/Adl2 hybrid early region 1 plasmids. Virology 120:422-432. Bos, J. L., L. J. Polder, R. Bernards, P. I. Schrier, P. J. van den Elsen, A. J. van der Eb, and H. van Ormondt. 1982. The 2.2Kb mRNA of the E1B region of human adenovirus type 5 and 12 directs the synthesis of two major antigens from different AUG triplets. Cell 27:121131. Della Valle, G., R. G. Fenton, and C. Baslico. 1981. Polyoma large T antigen regulates the integration of viral DNA sequences into the genome of transformed cells. Cell 23:347-355. Di Mayorca, G., J. Callender, G. Marin, and R. Giordano. 1%9. Temperature-sensitive mutants of polyoma virus. Virology 38:126-133. Downey, J. F., D. T. Rowe, S. Bacchetti, F. L. Graham, and S. T. Bayley. 1983. Mapping of a 14,000-dalton antigen to early region 4 of the human adenovirus 5 genome. J. Virol. 45:514-523. Eckhart, W. 1975. Properties of temperature sensitive mutants of polyoma virus. Cold Spring Harbor Symp. Quant. Biol. 39:37-40. Fried, M. 1%5. Cell transformation ability of a temperature-sensitive mutant of polyoma virus. Proc. Natl. Acad. Sci. U.S.A. 53:486-491. Frost, E., and J. Williams. 1978. Mapping temperature sensitive and host range mutations of adenovirus type 5 by marker rescue. Virology 91:39-50. Gallimore, P. H., P. A. Sharp, and J. Sambrook. 1974. Viral DNA in transformed cells II. A study of the sequences of Ad2 DNA in nine lines of transformed rat cells using specific fragments of the viral genome. J. Mol. Biol. 89:49-72.
13. Galos, R. S., J. Williams, T. Shenk, and N. Jones. 1980. Physical location of host-range mutations of adenovirus
14.
15.
16.
17.
18. 19.
20.
21. 22.
23.
24.
1043
type 5: deletion and marker-rescue mapping. Virology 104:510-513. Graham, F. L., P. J. Abrahams, C. Mulder, H. L. HeUneker, S. 0. Warmaar, F. A. J. de Vries, W. Flers, and A. J. van der Eb. 1974. Studies on in vitro transformation by DNA and DNA fragments of human adenoviruses and SV40. Cold Spring Harbor Symp. Quant. Biol. 39:637650. Graham, F. L., S. Bacchettl, R. McKinnon, C. Stanners, B. Cordell, and H. M. Goodman. 1980. Transformation of mammalian cells with DNA using the calcium technique, p. 3-25. In R. Baserga, C. Croce, and G. Rovera (ed.), Introduction of macromolecules into viable mammalian cells. Alan R. Liss, Inc., New York. Graham, F. L., T. Harrison, and J. Williams. 1978. Defective transforming capacity of adenovirus 5 hostrange mutants. Virology 86:10-21. Graham, F. L., J. Smiley, W. C. Russell, and R. Nairn. 1977. Characteristics of a human cell line transformed by DNA from human adenovirus 5. J. Gen. Virol. 36:59-72. Graham, F. L., and A. J. van der Eb. 1973. A new technique for the assay of infectivity of human adenovirus 5 DNA. Virology 52:456-467. Graham, F. L., and A. J. van der Eb. 1973. Transformation of rat cells by DNA of human adenovirus 5. Virology 54:536-539. Graham, F. L., A. J. van der Eb, and H. L. HeiJneker. 1974. Size and location of the transforming region in human adenovirus DNA. Nature (London) 251:687-691. Halbert, D. N., D. J. Spector, and H. J. Raskas. 1979. In vitro translation products specified by the transforming region of adenovirus type 2. J. Virol. 31:621-629. Harrison, T., F. L. Graham, and J. Williams. 1978. Hostrange mutants of adenovirus type 5 defective for growth in HeLa cells. Virology 77:319-329. Hassell, J. A., W. C. Topp, D. B. Rifkin, and P. E. Moreau. 1980. Transformation of rat embryo fibroblasts by cloned polyoma virus DNA fragments containing only part of the early region. Proc. Natl. Acad. Sci. U.S.A. 77:3978-3982. Jochenisen, H., G. S. G. Daniels, J. J. Hertoghs, P. I. Schrier, P. J. van den Elsen, and A. J. van der Eb. 1982. Identification of adenovirus type 12 gene products involved in transformation and oncogenesis. Virology
122:15-28. 25. Laemmli, U. K. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature (London) 227:680-685. 26. Lassam, N. J., S. T. Bayley, and F. L. Graham. 1978. Synthesis of DNA, late polypeptides, and infectious virus by host-range mutants of Ad5 in non-permissive cells. Virology 87:463-467. 27. Lassam, N. J., S. T. Bayley, and F. L. Graham. 1979. Tumor antigens of human Ad5 in transformed cells and in cells infected with transformation defective host-range mutants. Cell 18:781-791. 28. Moore, J. L., K. Chowdhury, M. A. Martin, and M. Israel. 1980. Polyoma large tumor antigen is not required for tumorigenesis mediated by viral DNA. Proc. Natl. Acad.
Sci. U.S.A. 77:1336-1340. 29. Novak, U., S. M. Dilworth, and B. E. Griffin. 1980. Coding capacity of a 35% fragment of the polyoma virus genome is sufficient to initiate and maintain cellular transformation. Proc. Natl. Acad. Sci. U.S.A. 77:3278-3282. 30. Ostrove, J. M., and K. I. Berns. 1980. Adenovirus early region Ib gene function required for rescue of latent adeno-associated virus. Virology 104:502-505. 31. Perricaudet, M., J. M. Le Moullec, and U. Petterson. 1980. Predicted structure of two adenovirus tumor antigens. Proc. Natl. Acad. Sci. U.S.A. 77:3778-3782. 32. Persson, H., H. Monstein, G. Akusjarvi, and L. Philipson. 1981. Adenovirus early gene products may control viral mRNA accumulation and translation in vivo. Cell 23:4854%. 33. Ross, S. R., S. J. Flint, and A. J. Levine. 1980. Identifica-
1044
NOTES
tion of the adenovirus early proteins and their genomic map positions. Virology 100:419-432. 34. Rowe, D. T., M. Ruben, S. Bacchetti, and F. L. Graham. 1982. Identification of an adenovirus function required for initiation of cell transformation possibly at the level of DNA integration. In M. Pearson and N. Steinberg (ed.), Proceedings of the 4th Technical Transfer Workshop. Raven Press, New York. 35. Ruben, M., S. Bacchetti, and F. L. Graham. 1982. Integration and expression of viral DNA in cells transformed by host range mutants of adenovirus type 5. J. Virol. 41:674685. 36. Schaffhausen, B. S., J. E. Silver, and T. L. Benjamin. 1978. Tumor antigen(s) in cells productively infected by wildtype polyoma virus and mutant NG-18. Proc. Natl. Acad.
J. VIROL. Sci. U.S.A. 75:79-83. 37. Schrier, P. I., P. J. Van den Elsen, J. T. Hertoghs, and A. J. van der Eb. 1979. Characterization of tumor antigens in cells transformed by fragments of adenovirus type 5 DNA. Virology 99:372-385. 38. van Ormondt, H., J. Maat, and C. P. van Beveren. 1980. The nucleotide sequence of the transforming early region El of adenovirus type 5 DNA. Gene 11:299-309. 39. Wigler, M., A. Pellicer, S. Silverstein, R. Axel, G. Urlaub, and L. Chasin. 1979. DNA-mediated gene transfer of the APRT locus into mammalian cells. Proc. Nati. Acad. Sci. U.S.A. 76:1373-1376. 40. Williams, J. 1973. Oncogenic transformation of hamster embryo cells in vitro by adenovirus type 5. Nature (London) 243:162-163.
