- Email: [email protected]
Three different human tumor cell lines contain different oncogenes
Cell. Vol. 25, 355-361,
August
1981,
Copyright
0 1981
by MIT
Three Different Human Tumor Cell Lines Contain Different Oncogenes Mark J. Murray, Ben-Zion Deborah Cowing, Ho Wen Robert A. Weinberg Center for Cancer Research and Department of Biology Massachusetts Institute of Cambridge, Massachusetts
Shilo, Chiaho Hsu* and
Shih,
Technology 02139
Summary We have obtained foci of transformed mouse cells after transfection of human DNA from colon and bladder carcinoma cell lines and a promyelocytic leukemia cell line. These foci can be shown to contain a large number of human DNA sequences by use of highly repetitive human DNA sequence probes. Cell DNA from primary foci can be used in a subsequent cycle of transfection resulting in secondary foci that contain relatively little human DNA. Secondary foci appear to contain only the human sequences proximal to those responsible for the transformed phenotype. A set of characteristic DNA restriction fragments is found in common among secondary foci derived from each tumor cell line DNA. Comparison of the common DNA fragments found in secondary foci derived from three different human tumor cell lines indicates that these three cell lines contain three different transforming genes. introduction The ability to transfer a selectable phenotype to a recipient cell by gene-transfer techniques has made it possible to passage a variety of mammalian genes (Wigler et al., 1978, 1979a, 1980; Graf et al., 1979; Lewis et al., 1980). Recently, this approach has been applied to the study of oncogenic transformation. Investigators in our laboratory (Shih et al., 1979, 1981) and others (Cooper et al., 1980; Krontiris and Cooper, 1981) have shown that DNAs from nonvirally transformed cells are able to induce transformation of normal mouse fibroblasts. Initial work in this area involved the transfer of the oncogenic phenotype of 3-methylcholanthrene-transformed mouse fibroblasts to NIH/3T3 mouse cells by DNA transfection (Shih et al., 1979). The successful passaging of these traits via gene-transfer procedures suggests the presence, in these cells, of DNA sequences that behave upon transfection like some well studied viral transforming genes (Klein, 1980). Moreover, this work demonstrated the ability of the transforming DNA to act across tissue and species barriers. For example, DNA extracted from a human carcinoma cell line was able
* Present University,
address: College of Physicians New York, New York
and Surgeons,
Columbia
to transform mouse fibroblasts (Shih et al., 1981; Krontiris and Cooper, 1981). Initial arguments for the authenticity of transformed foci rested on the efficiency of induction of foci in the recipient monolayers. DNAs prepared from these foci also were able to generate foci efficiently in subsequent cycles of transfection. To provide a biochemical verification for the presence of transformed donor ceil DNA in the recipient cells, we added many copies of a cloned DNA sequence (for example, pBR322 plasmid DNA) to the donor DNA prior to transfection. The resulting foci could be shown by Southern blotting to contain copies of the cotransfected plasmid DNA (Shilo et al., 1980). This cotransfected plasmid DNA therefore served as a “sequence tag,” signaling the presence of foreign DNAs in the recipient cell. These data showed that these foci belonged to the small fraction (0.1%) of recipient cells in the transfected culture that takes up and stabilizes donor DNA (Perucho et al., 1980b). It was concluded therefore that the transformation of these cells depended upon the introduction of exogenous DNA sequences. An alternative biochemical verification was made possible when human tumor cell DNAs were found to be able to induce transformation in recipient mouse cells (Shih et al., 1981; Krontiris and Cooper, 1981). Many human genes are punctuated by arrays of highly repeated sequences scattered throughout the human genome (Jelinek et al., 1980; Schmid and Deininger, 1975; Tashima et al., 1981). The most abundant family of these sequences has been termed the “Alu” family by virtue of an Alu I restriction enzyme site found in a high proportion of these repeated sequences (Houck et al., 1979). These “Alu” sequences are 300 base pairs long and are present in as many as 600,000 copies per haploid cell genome (Rinehart et al., 1981). They are found, on average, every few kilobases throughout the genome. These human repetitive sequences can be specifically detected in the mouse cell genetic background (Shih et al., 1981). They therefore serve as naturally occurring “sequence tags” that indicate the presence of human DNA in the mouse genetic background. The human repeat DNA sequence blocs provide a powerful analytic tool in addition to the verification of successful gene transfer. They provide a means to characterize the structural outline of the human tumor genes in the mouse recipient cells. Analysis of these structural outlines has allowed us to draw several conclusions concerning these genes. Perhaps the most salient of these conclusions is that different distinct genes appear to be responsible for the oncogenie transformation observed in different types of human tumor cell lines. Results In a previous report, we demonstrated that the DNA of a human bladder carcinoma cell line was able to
Cell 356
transform NIH/3T3 mouse fibroblasts (Shih et al., 1981). More recently, we have found that the DNAs of a human colon carcinoma and a human myeloid leukemia cell lines can also transform mouse fibroblasts (A. Cassill, D. Cowing and H. W. Hsu, unpublished results). We wanted to demonstrate the presence of human donor DNAs in these putative transfectants by use of sequence probes specific for highly repeated human DNA. Representative, highly repeated “Alu” sequences have been isolated as molecular clones by Schmid and his colleagues, who have called these clones “BLUR” (Rubin et al., 1980). We have used individual cloned Alu DNAs as molecular sequence probes for the presence of human Alu sequences in a mouse genetic background. This procedure depends on the ability of human repeated DNAs to be detected in a rodent genetic background, as originally described by Gusella et al. (1980), and confirmed by us (Shih et al., 1981). This detection can be achieved by use of the Southern transfer procedure and depends upon the fact that human repeated sequence probes do not react with homologous sequences present in the rodent genetic background. Analysis of the DNAs of Primary Foci Foci of transformed cells were picked from recipient monolayers exposed to the DNAs of human colon carcinoma cell line SW-480 (Leibovitz et al., 1976), bladder carcinoma cell line EJ (Marshall et al., 1977) and the myeloid leukemia cell line HL-60 (Collins et al., 1977). DNAs were prepared from the cell lines that grew out from these foci. These DNAs were cleaved with the restriction endonuclease Eco RI and analyzed by Southern transfer. The resulting blots were incubated with the human repeat sequence probe BLURS, which has 300 base pairs of homology to the human Alu sequences (Rubin et al., 1980). Figure 1 demonstrates a representative analysis of the DNAs of such primary transfectants. The foci induced by SW-480 colon carcinoma DNA carry a large number of DNA fragments that are reactive with the human Alu sequence probe BLUR-8. This confirms that these transfectants have taken up human DNA concomitantly with the acquisition of the transformed phenotype. The amount of donor human DNA taken up greatly exceeds that which might be associated with a transformation-inducing sequence. This finding reflects the results of other investigators (Wigler et al., 1979b; Perucho et al., 1980b), who have shown that transfected cells acquiring a selected marker will also display a large array of other donor sequences taken up during the same transfection event. These cotransfected donor sequences may constitute as much as 0.1% of the donor cell genome (Perucho et al., 1980b).
a
bcde
Figure 1. Detection of Human 3T3 Cells following Transfection
DNA Sequences in Transformed with SW-480 Colon Carcinoma
NIH/ DNA
Ten micrograms of the DNA of each focus were digested with the restriction endonuclease Eco RI, followed by agarose gel electrophoresis and Southern transfer as described in Experimental Procedures. The resulting filter was incubated with the BLUR-8 probe (Rubin et al., 1980). (Lane a) Contains normal human cellular DNA; (lane e) contains NIH/3T3 mouse cell DNA: (lanes b, c and d) contain DNA from primary foci of NIH/3T3 cells transformed by SW-480 human colon carcinoma DNA.
_
This Southern blot analysis distinguishes these primary foci from other, similarly appearing colonies of a transfected culture that are spontaneous overgrowths not resulting from DNA uptake. Thus we have discarded a series of candidate transfectants obtained in experiments in which human squamous cell and renal carcinoma donor DNAs were used, because the foci picked from these transfections gave no evidence of human repeated sequences (unpublished results) and the DNAs of these foci yielded no secondary foci upon repeated attempts at transfection. The rodent genome contains repeated sequences that are homologous to the highly repeated human sequences (Haynes et al., 1981; Krayev et al., 1980).
Human 357
Tumor
Oncogenes
We do not detect these sequences on our Southern blots (Figure 1, lane e), presumably because the mouse homologs are too mismatched to form a stable hybrid with the BLUR-8 probe. Analysis of the DNAs of the Secondary Foci Previous experiments have shown that transforming genes may be passaged serially by transfection (Shih et al., 1979, 1981; Cooper et al., 1980; Krontiris and Cooper, 1981). The DNA of a primary focus derived from transfection of tumor cell DNA can be prepared and used in a subsequent cycle of transfection. The resulting foci are termed secondary foci. As shown above, the transforming sequence in the primary transfectant is acquired together with a large number of cotransfected human sequences (Figure 1). The transforming gene is probably physically linked to these cotransfected human fragments within the recipient (Perucho et al., 1980b). During the manipulations of the subsequent transfection cycle, however, this array of human sequences will become fragmented and the human fragments will be diluted by the great excess of mouse DNA fragments derived from the chromosomal DNA of the primary transfectant (Shilo et al., 1980). Upon application of primary DNA to a recipient monolayer in the second tranfection cycle, the rare cell that acquires the transforming gene will also acquire a large number of the mouse chromosomal DNA fragments. This secondary transfectant, will acquire however, few if any human fragments besides those closely linked to the transforming gene. Other human fragments, present in the primary transfectant, will be lost because they are distributed to hundreds of other cells of the transfected culture that are discarded because they have not become transformed. Figure 2 shows a Southern blot analysis of secondary transfectants derived from DNAs prepared from the human colon carcinoma primary foci. Six secondary foci were picked after transfection of DNAs from three different primary foci. These foci were grown into mass cultures. Their DNAs were prepared and cleaved with endonuclease Eco RI (Figure 2, lanes a, b, c, d, e and f) prior to Southern analysis as described above. Each of the secondary transfectants carries an array-of DNA fragments detectable with the BLUR-8 probe. It is of central interest that each of these secondary transfectants exhibits a set of seven restriction fragments present in the DNA of all other secondary transfectants of this group. These fragments are indicated by arrows (Figure 2, lanes a-f) and range in size from 3 kb to 7 kb. A discrete set of DNA fragments is therefore associated with the presence of the human colon carcinoma transforming sequence. Most of the secondary foci have additional human DNA fragments besides those they have in common with others of the group. These fragments are other
abcdef
Figure ondary tion
2. Detection Foci Derived
g h i j
of Human DNA Sequences Common to the Secfrom SW-480 Colon Carcinoma DNA Transfec-
DNAs from secondary foci were analyzed with the BLUR-8 probe as in Figure 1. The Eco RI-digested DNAs in lanes a, b, c, d, e and f were independently derived from three primary foci. Secondary foci SW3-1 (lane a) and SW3-2 (lane f) were derived from primary focus SW3; SW2-3, SW2-2 and SW2-1 (lanes b. c and d) were derived from primary focus SW2; SWI-I (lane e) was derived from SW1 Arrows indicate the seven Eco RI-generated restriction fragments found in all six secondary foci. Double arrows indicate bands that are doublets not well resolved on this gel (see Figure 3A). SW3-1 DNA (lane a) and SWZ-3 DNA (lane b) were also analyzed after digestion with endonucleases Xba I (lanes g and h) and Pvu II (lanes i and j). Size markers (in kilobases) are indicated on the vertical ordinate.
human DNA sequences, likely to be found on either side of the transforming sequence, whose presence is not required for the functioning of this sequence. The common fragments seen in Figure 2, lanes af, are conserved in six independently derived secondary foci. Since these secondary foci originated from the DNAs of three independent primary transfectants, we conclude that the association of these fragments with the transforming sequence has survived two cycles of DNA passage and therefore must be intimately associated with this sequence. The data presented in Figure 2 can be used to estimate the size of the transforming sequence. The sum of the sizes of the conserved fragments shared among all the secondary colon carcinoma tranfectants should reflect the size of the functionally essential sequence required for expression of the transformed phenotype. This estimate is compromised because some restriction fragments may lack Alu sequences and thus escape detection in the Southern blot analysis. This problem can be minimized by the use of a
Cell 358
series of restriction endonucleases, because DNA fragments that become unlinked from Alu sequences after cleavage with one endonuclease may remain linked after cleavage with another endonuclease. The sum of the common Eco RI fragments in Figure 2, lanes a-f, gives an estimate of approximately 3.5 kb for the size of the colon carcinoma gene. The sum of the conserved Alu-containing Xba I fragments (lanes g and h) gives an estimate of 35 kb, and the sum of the Pvu II fragments (lanes i and j) gives an estimate of 25 kb. The human fragments observed in Figures 1 and 2 were detected by use of the BLUR-8 probe. We have obtained essentially identical results using as probes other BLUR clones, Cot = 1 selected human DNA and total HeLa cell DNA. The uncloned DNA probes hybridize to the same DNA fragments, and with similar intensity, as the BLUR clones do (data not shown). This indicates that the Alu sequences are the most abundant repetitive DNA family in the human genome (Houck et al., 1979). None of the repetitive DNA sequence probes used in these experiments detects any mouse sequences under the conditions used (see Figure 1, lane e). Moreover, these probes can be used interchangeably. Comparison of Transforming Genes from Other Human Tumor Cell Lines The ability of this approach to detect a characteristic pattern of restriction fragments encompassing the human colon carcinoma transforming sequence demonstrated that it would be a powerful tool for the comparison of human transforming sequences of other origins. The DNAs of secondary foci derived from the transforming sequence of the HL-60 tumor line all share a set of three Alu-containing DNA fragments (Figure 3C, arrows). These fragments are observed as well in a group of tertiary foci derived from the transformation of the DNA of the secondary transfected HL-60 (data not shown). It is noteworthy that the conserved fragments associated with the myeloid leukemia transforming sequence are distinct from those of the colon carcinoma gene. The DNAs of secondary foci derived originally from the EJ bladder carcinoma cell line present still another pattern. In this case, the DNA of each secondary focus contains a single large Eco RI, Alu-containing fragment (approximately 23 kb). The size of the Eco RI fragment varies slightly from focus to focus. This variation appears to reflect the difficulty of preserving intact the entire 23 kb fragment throughout transfection. Such variation was not encountered upon analysis of smaller sized Eco RI fragments (see above). The distinct pattern we observed indicates the existence of a transforming DNA sequence that is different from those of the colon carcinoma and myeloid leukemia cell lines.
-23 - 9.8 -6.6
-2.5
Figure 3. Comparison of Alu-Containing DNA Fragments Characteristic of Transforming Sequences from Three Different Human Tumor Cell Lines The DNAs of secondary foci derived from each human tumor cell line were digested with Eco RI and analyzed as described above. (A) Shows two secondary focus DNAs (SW3-1 and SWZ-3) derived from SW-480 hybridized with the BLUR-8 probe. The seven common colon carcinoma fragments are indicated by arrows. (B) Shows a single common fragment among four secondary focus DNAs derived from the human bladder carcinoma cell line EJ. The probe used in (B) was Cot = 1 DNA. (C) Shows secondary focus DNAs derived from the human promyelocytic leukemia cell line HL-60. Total HeLa cell DNA was used as the probe in (C) and detects three common restriction fragments. Size markers (in kilobases) are indicated to the right of the figure.
Discussion The application of DNA transfection procedures to the study of oncogenic transformation has demonstrated that the transformed phenotype is encoded in genomic DNA. This experimental approach also allows the transforming sequence to be dissociated from all other irrelevant DNA sequences in the tumor cell line and to be studied directly. The affiliation of human tumor transforming sequences with highly repeated DNA sequences makes possible a number of conclusions concerning the structure of these sequences. The results show that the transforming activity of each of these DNAs is associated with a specific, definable set of DNA sequences. The transformation of cells following transfection is therefore not a consequence of the uptake and fixation of exogenous DNA, as such, but rather a reflection of the presence of a discrete array of DNA sequences that probably encompasses a transforming gene. An analogous conclusion was drawn in earlier work on the DNA of 3-methylcholanthrene-transformed mouse fibroblasts (Shilo et al., 1980). In that
Human 359
Tumor
Oncogenes
case, defined sequence tags were interspersed experimentally throughout the genomic DNA of the transformed donor cell prior to transfection. Subsequent analysis revealed the presence of the same sequencetag-containing DNA fragments in the DNAs of all secondary foci, thus implying the necessity for transfer of a unique DNA fragment to elicit transformation. In a subsequent study, we demonstrated that the transforming sequences of four different, 3-methylcholanthrene-transformed mouse fibroblast lines were susceptible or resistant to the same set of restriction enzyme treatments (Shilo and Weinberg, 1981). This suggested that the transforming genes in these four, independently induced lines represent four different, allelic variants of the same cellular gene. Less complete analysis of the transforming genes of three different ethylnitrosourea-induced rat neuroblastomas indicated that the transforming sequences of these three tumor lines behaved identically to one another after restriction endonuclease treatment (C. Shih, unpublished data). Thus, within a given tumor cell type (for example, 3-methylcholanthrene-transformed fibroblasts), we suggested that perhaps only one transforming gene may be activated. These earlier studies, however, could not indicate whether the same or different transforming sequences were present in different tumor types because the two groups of donor tumor cell lines derived from two different host species. The present data allow a preliminary resolution of this issue, since three different transforming genes from the same species are being analyzed. We conclude that different tumor types appear to carry different transforming sequences. Because of the absence of additional biologically active DNAs of these three types of human tumors, we do not yet know whether all human tumors of a given type (for example, colon carcinoma) carry the same active transforming gene in their DNA. Nevertheless, we can already consider it unlikely that activation of a single cellular oncogene is responsible for all types of human tumors. The genes transferred in these experiments derive from tumor cell lines. Generally, these cell lines are difficult to establish and require long-term culture to become stable (Fogh and Trempe, 1975). We have assumed that the transforming genes originally activated in the tumor are responsible for the continued transformation of the resulting cell lines, and for the transformation of mouse fibroblasts following transfection. This issue will be clarified by isolation of these genes directly from the DNA of tumor samples. The characterization of human tumor genes described here is applicable to the analysis of any transfectable human gene for which a selection is available. Thus one might use parallel procedures to analyze DNAs of secondary transfectants that have acquired
selectable enzyme markers or drug resistance markers. In principle, the molecular cloning of such genes might follow, by introducing the DNA of a secondary tranfectant into lambda phage vectors and screening the resulting genomic library for those components reactive with human repeat sequence probes. Others have reported the successful cloning of the cellular thymidine kinase and adenosine phosphoribosyl transferase genes by experimentally linking bacterial plasmids or plasmid fragments to these marker genes prior to serial transfections (Perucho et al., 1980a; Lowy et al., 1980). The resulting juxtaposition of these foreign sequences with the gene of interest allows the cloning of these genes by use of plasmidmediated rescue (Perucho et al., 1980a) or by the isolation of chimeric lambda phage carrying the identifiable plasmid sequences linked to the marker gene (Lowy et al., 1980). Exogenous sequence tags are of limited use when cloning large genes such as the transforming genes identified here, because the tag only allows isolation of a closely linked terminal fragment of the gene. Isolation of the remainder of the gene requires use of a chromosome “walking” protocol. In contrast, the presence of interspersed repeated sequences provides a naturally occurring sequence tag whose linkage to a marker gene need not be established by difficult experimental procedures. Moreover, these interspersed sequences should make possible identification and isolation of fragments spanning the entirety of the gene of interest. This strategy should be applicable to analysis of transfected marker genes that are introduced into a recipient cell of a heterologous species. These highly repeated gene sequences appear to evolve far more rapidly than do unique DNA sequences (Gusella et al., 1980). Thus one should be able to detect, specifically, repeated DNA sequences of one species introduced into a recipient cell of another species of sufficient evolutionary distance. Experimental
Procedures
Human Tumor Cell Lines The human bladder carcinoma cell line SW-480 (Leibowitz et al., 1976) was obtained from J. Fogh, the human promyelocytic leukemia cell line HL-60 (Collins et al., 1977) was obtained from R. C. Gallo and the human bladder carcinoma cell line EJ (Marshall et al., 1977) was obtained from I. Summerhayes and L. 6. Chen. DNA Transfection Assays DNA transfection procedures were essentially as described elsewhere (Andersson et al., 1979). In each transfection. 75 pg of DNA were applied to 1.5 x 1 O6 NIH/3T3 cells, and the appearance of foci was scored 14 days later. Transfection of chromatin of the SW-480 cell line was performed as described by Miller and Ruddle (1978), with modifications described by Shih et al. (1979). This transfection yielded three primary foci whose morphology conformed to that described previously (Shih et al., 1979) and whose DNA contained human Alu sequences (Figure 1). Transfection of DNA extracted from the HL-60 cell line was performed either with or without an admixture of copies of a molecular cDNA clone of dihydrofolate reductase (M.
Cell 360
Murray, Fi. Kaufman, S. Latt, A. Cassill and R. A. Weinberg, manuscript in preparation). One primary focus was obtained from the cotransfection of HL-60 DNA and the dihydrofolate reductase cDNA clone. This focus was also methotrexate-resistant, and its DNA is analyzed here. Transfection of HL-60 DNA alone yielded 12 primary foci. Cells of the HL-60 primary foci exhibited an extremely hyperrefractile. spindle-shaped morphology. Derivation of foci by transfection of EJ cell DNA was previously reported (Shih et al., 1981). Analysis of Transformed Focus DNAs Whole cell DNA was prepared from transformed cells grown from foci obtained after transfection with tumor cell line DNAs. Following restriction endonuclease digestion, 10 pg of the DNA of each focus were resolved by electrophoresis through a 1% agarose gel in 40 mM Tris (pH 7.91, 50 mM sodium acetate and 1 mM EDTA. After electrophoresis. the DNA was transferred to nitrocellulose by the method of Southern, and the resulting blots hybridized with probes for human repetitive DNA. The BLUR probes (Rubin et al., 1980) were obtained from D. Jolly and P. Deininger. The Cot = 1 probe was prepared by heat-denaturing sonicated human DNA at 1OO’C and reannealing it at 68°C to a Cot value of 1. This was followed by Sl nuclease treatment for 1 hr at 37°C and hydroxyapatite chromatograpy to isolate repetitive DNA duplexes. Total HeLa cell DNA was sheared through a 20 gauge needle prior to use as a probe. Each probe was “P-labeled by nick translation (Rigby et al., 1977). We used 50 pCi of each of the four 3ZP-labeled deoxyribonucleotides. The labeled probes (5 X 1 O6 cpm) were incubated with the Southern blots for 24 hr in the presence of 10% dextran sulfate (Wahl et al., 1979) at 4O’C. When HeLa DNA was used as a probe, sonicated NIH/3T3 mouse cell DNA was added to the prehybridization solution at 50 pg/ml and to the hybridization solution at 20 AS/ml.
segments from specific USA 77,2829-2833.
human
chromosomes.
Proc.
Nat. Acad.
Sci.
Haynes, S., Toomey, T., Leinwand, L. and Jelinek, W. (1981). The Chinese hamster Alu-equivalent sequence: a conserved highly repetitious interspersed DNA sequence in mammals has the structure suggestive of a transposable element. Mol. Cell. Biol. 7, 573-583. Houck, C. M.. Rinehart, F. P. and Schmid. C. W. (1979). A ubiquitous family of repeated DNA sequences in the human genome. J. Mol. Biol. 132, 289-306. Jelinek, W. R., Toomey, T. P., Leinwand, L., Duncan, C. H., Biro, P. A., Choudary, P. V., Weissman, S. M., Rubin, C. M., Houck, C. M., Deininger. P. L. and Schmid, C. W. (1980). Ubiquitous, interspersed repeated sequences in mammalian genomes. Proc. Nat. Acad. Sci. USA 77, 1398-1402. Klein, G., ed. (1980).
Viral Oncology.
(New
Krayev. A. S., Krameriv, D. A., Skryabin, A. A. and Georgiev, G. P. (1980). The ubiquitous repetitive DNA sequence 61 abundant class of mouse fold-back RNA. 1215.
York:
Raven
Press).
K. G., Ryskov, A. P.. Bayer, nucleotide sequence of the complementary to the most Nucl. Acids Res. 8, 1201-
Krontiris, T. G. and Cooper, G. M. (1981). Transforming activity human tumor DNAs. Proc. Nat. Acad. Sci. USA 78, 1181-l 184.
of
Leibovitz, A., Stinson, J. C., McCombs, W. B., McCoy, C. E.. Mazur, K. C. and Mabry, N. D. (1976). Classification of human colorectal adenocarcinoma cell lines. Cancer Res. 36, 4562-4569. Lewis, W. H., Srinivasan, P. R., Stokoe. N. and Siminovitch, L. (I 980). Parameters governing the transfer of the genes for thymidine kinase and dihydrofolate reductase into mouse cells using metaphase chromosomes or DNA. Somatic Cell Genet. 6, 333-347. Lowy, I., Pellicer, A., Jackson, J. F., Sim. G.-K., Silverstein, S. and Axel, R. (I 980). Isolation of transforming DNA: cloning the hamster aprt gene. Cell 22, 817-823.
Acknowledgments We would like to thank R. Gallo, J. Fogh, I. Summerhayes and L. B. Chen for providing cell lines: D. Jolly and P. Deininger for providing the BLUR clones; Aaron Cassill for excellent technical assistance: J. Toole, C. Schmid and W. Jelinek for helpful discussions. M. M. was a fellow of the Damon Runyon-Walter Winchell Cancer Fund. B. S. was a Chaim Weizmann postdoctoral fellow. This work was supported by grants from the National Cancer Institute to R. A. W. and to S. E. Luria. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Marshall, C. J., Franks, L. M. and Carbonell, A. W. (1977). Markers of neoplastic transformation in epithelial cell lines derived from human carcinomas. J. Nat. Cancer Inst. 58, 1743-l 747.
Received
Rinehart, F. P., Ritch, T. G., Deininger, P. L. and Schmid, C. W. (1981). Renaturation rate studies of a single family of interspersed repeated sequences in human DNA. Biochemistry 20, 3003-3010.
May 8, 1981;
revised
May 28, 1981
References Andersson. P., Goldfarb, M. P. and Weinberg, R. A. (1979). A defined subgenomic fragment of in vitro synthesized Moloney sarcoma virus DNA can induce cell transformation upon transfection. Cell 76, 6375. Collins, S. J., Gallo, R. C. and Gallagher, R. E. (1977). growth and differentiation of human myeloid leukemic pension culture. Nature 270, 347-349.
Continuous cells in sus-
Cooper, G. M., Okenquist. S. and Silverman, L. (1980). Transforming activity of DNA of chemically transformed and normal cells. Nature 284, 418-421. Fogh, J. and Trempe, G. (1975). New human tumor cell lines. In Human Tumor Cells in Vitro. J. Fogh, ed. (New York: Plenum Press), pp. 115-l 59. Graf. L. H., Urlab, G. and Chasin, L. (1979). gene for hypoxanthine phosphoribosyltransferase. net. 5, 1031-1044.
Transformation Somatic
of the Cell Ge-
Gusella, J. F., Keys, C., Varsanyi-Breiner, A., Kao, F., Jones, C., Puck, T. T. and Housman, D. (1980). Isolation and localization of DNA
Miller, C. L. and Ruddle, F. H. (I 978). Co-transfer of human X-linked markers into murine somatic cells via isolated metaphase chromosomes Proc. Nat. Acad. Sci. USA 75, 3346-3350. Perucho, M., Hanahan, D., Lipsich, L. and Wigler, M. (1980a). Isolation of the chicken thymidine kinase gene by plasmid rescue. Nature 285, 207-210. Perucho, M., Hanahan. D. and Wigler, physical linkage of exogenous sequences 22, 309-317.
M. (1980b). Genetic in transformed cells.
and Cell
Rubin, C. M., Houck, C. M., Deininger. P. L., Friedmann, T. and Schmid, C. W. (1980). Partial nucleotide sequence of the 300-nucleotide interspersed repeated human DNA sequences. Nature 284, 372-374. Schmid, C. W. and Deininger, P. L. (1975). the human genome. Cell 6, 345-358.
Sequence
organization
of
Shih, C., Shilo. B.. Goldfarb, M. P., Dannenberg, A. and Weinberg, R. A. (I 979). Passage of phenotypes of chemically transformed cells via transfection of DNA and chromatin. Proc. Nat. Acad. Sci. USA 76, 5714-5718. Shih, C., Padhy, L. C., Murray, M. J. and Weinberg, Transforming genes of carcinomas and neuroblastomas into mouse fibroblasts. Nature 290, 261-264. Shilo, B. and Weinberg, carcinogen-transformed
R. A. (1981). introduced
R. A. (1981). Unique transforming mouse cells. Nature 289, 607-609.
gene
in
Shilo, B., Shih, C., Merrick, M., Goldfarb, M. and Weinberg, R. A. (1980). Initiation of molecular cloning of a transforming allele of chemically transformed mouse cells. In Mobilization and Reassembly of Genetic Information, Miami Winter Symposium, W. A. Scott, R.
Human 361
Tumor
Oncogenes
Werner, D. Ft. Joseph Press), pp. 275-284.
and J. Schultz,
eds.
(New
York:
Academic
Tashima. M., Calabretta, B., Tore& G., Scofield, M., Maizel, A. and Saunders, G. F. (1981). Presence of a highly repetitive and widely dispersed DNA sequence in the human genome. Proc. Nat. Acad. Sci. USA. 78, 1508-l 512. Wahl, G. M., Stern, M. and Stark, G. Ft. (1979). Efficient transfer of large DNA fragments from agarose gels to diazobenzyloxymethylpaper and rapid hybridization by using dextran sulfate. Proc. Nat. Acad. Sci. 76, 3683-3687. Wigler, M., Pellicer, A., Silverstein, S. and Axel, R. (1978). Biochemical transfer of single-copy eucaryotic genes using total cellular DNA as donor. Cell 74, 725-731. Wigler, M., Pellicer, A., Silverstein, S., Axel, R., Urlab, G. and Chasin, L. (1979a). DNA-mediated transfer of the adenine phosphoribosyltransferase locus into mammalian cells. Proc. Nat. Acad. Sci. USA 76, 1373-i 376. Wigler, M.. Sweet, R., Sim, G. K., Wold, B., Pellicer, A., Lacy, E., Maniatis, T.. Silverstein, S. and Axel. R. (1979b). Transformation of mammalian cells with genes from procaryotes and eucaryotes. Cell 16, 777-785. Wigler, M., Perucho, M., Kurtz, D.. Dana, S., Pellicer, A., Axel, Ft. and Silverstein. S. (1980). Transformation of mammalian cells with an amplifiable dominant-acting gene. Proc. Nat. Acad. Sci. USA 77, 3567-3570.
August
1981,
Copyright
0 1981
by MIT
Three Different Human Tumor Cell Lines Contain Different Oncogenes Mark J. Murray, Ben-Zion Deborah Cowing, Ho Wen Robert A. Weinberg Center for Cancer Research and Department of Biology Massachusetts Institute of Cambridge, Massachusetts
Shilo, Chiaho Hsu* and
Shih,
Technology 02139
Summary We have obtained foci of transformed mouse cells after transfection of human DNA from colon and bladder carcinoma cell lines and a promyelocytic leukemia cell line. These foci can be shown to contain a large number of human DNA sequences by use of highly repetitive human DNA sequence probes. Cell DNA from primary foci can be used in a subsequent cycle of transfection resulting in secondary foci that contain relatively little human DNA. Secondary foci appear to contain only the human sequences proximal to those responsible for the transformed phenotype. A set of characteristic DNA restriction fragments is found in common among secondary foci derived from each tumor cell line DNA. Comparison of the common DNA fragments found in secondary foci derived from three different human tumor cell lines indicates that these three cell lines contain three different transforming genes. introduction The ability to transfer a selectable phenotype to a recipient cell by gene-transfer techniques has made it possible to passage a variety of mammalian genes (Wigler et al., 1978, 1979a, 1980; Graf et al., 1979; Lewis et al., 1980). Recently, this approach has been applied to the study of oncogenic transformation. Investigators in our laboratory (Shih et al., 1979, 1981) and others (Cooper et al., 1980; Krontiris and Cooper, 1981) have shown that DNAs from nonvirally transformed cells are able to induce transformation of normal mouse fibroblasts. Initial work in this area involved the transfer of the oncogenic phenotype of 3-methylcholanthrene-transformed mouse fibroblasts to NIH/3T3 mouse cells by DNA transfection (Shih et al., 1979). The successful passaging of these traits via gene-transfer procedures suggests the presence, in these cells, of DNA sequences that behave upon transfection like some well studied viral transforming genes (Klein, 1980). Moreover, this work demonstrated the ability of the transforming DNA to act across tissue and species barriers. For example, DNA extracted from a human carcinoma cell line was able
* Present University,
address: College of Physicians New York, New York
and Surgeons,
Columbia
to transform mouse fibroblasts (Shih et al., 1981; Krontiris and Cooper, 1981). Initial arguments for the authenticity of transformed foci rested on the efficiency of induction of foci in the recipient monolayers. DNAs prepared from these foci also were able to generate foci efficiently in subsequent cycles of transfection. To provide a biochemical verification for the presence of transformed donor ceil DNA in the recipient cells, we added many copies of a cloned DNA sequence (for example, pBR322 plasmid DNA) to the donor DNA prior to transfection. The resulting foci could be shown by Southern blotting to contain copies of the cotransfected plasmid DNA (Shilo et al., 1980). This cotransfected plasmid DNA therefore served as a “sequence tag,” signaling the presence of foreign DNAs in the recipient cell. These data showed that these foci belonged to the small fraction (0.1%) of recipient cells in the transfected culture that takes up and stabilizes donor DNA (Perucho et al., 1980b). It was concluded therefore that the transformation of these cells depended upon the introduction of exogenous DNA sequences. An alternative biochemical verification was made possible when human tumor cell DNAs were found to be able to induce transformation in recipient mouse cells (Shih et al., 1981; Krontiris and Cooper, 1981). Many human genes are punctuated by arrays of highly repeated sequences scattered throughout the human genome (Jelinek et al., 1980; Schmid and Deininger, 1975; Tashima et al., 1981). The most abundant family of these sequences has been termed the “Alu” family by virtue of an Alu I restriction enzyme site found in a high proportion of these repeated sequences (Houck et al., 1979). These “Alu” sequences are 300 base pairs long and are present in as many as 600,000 copies per haploid cell genome (Rinehart et al., 1981). They are found, on average, every few kilobases throughout the genome. These human repetitive sequences can be specifically detected in the mouse cell genetic background (Shih et al., 1981). They therefore serve as naturally occurring “sequence tags” that indicate the presence of human DNA in the mouse genetic background. The human repeat DNA sequence blocs provide a powerful analytic tool in addition to the verification of successful gene transfer. They provide a means to characterize the structural outline of the human tumor genes in the mouse recipient cells. Analysis of these structural outlines has allowed us to draw several conclusions concerning these genes. Perhaps the most salient of these conclusions is that different distinct genes appear to be responsible for the oncogenie transformation observed in different types of human tumor cell lines. Results In a previous report, we demonstrated that the DNA of a human bladder carcinoma cell line was able to
Cell 356
transform NIH/3T3 mouse fibroblasts (Shih et al., 1981). More recently, we have found that the DNAs of a human colon carcinoma and a human myeloid leukemia cell lines can also transform mouse fibroblasts (A. Cassill, D. Cowing and H. W. Hsu, unpublished results). We wanted to demonstrate the presence of human donor DNAs in these putative transfectants by use of sequence probes specific for highly repeated human DNA. Representative, highly repeated “Alu” sequences have been isolated as molecular clones by Schmid and his colleagues, who have called these clones “BLUR” (Rubin et al., 1980). We have used individual cloned Alu DNAs as molecular sequence probes for the presence of human Alu sequences in a mouse genetic background. This procedure depends on the ability of human repeated DNAs to be detected in a rodent genetic background, as originally described by Gusella et al. (1980), and confirmed by us (Shih et al., 1981). This detection can be achieved by use of the Southern transfer procedure and depends upon the fact that human repeated sequence probes do not react with homologous sequences present in the rodent genetic background. Analysis of the DNAs of Primary Foci Foci of transformed cells were picked from recipient monolayers exposed to the DNAs of human colon carcinoma cell line SW-480 (Leibovitz et al., 1976), bladder carcinoma cell line EJ (Marshall et al., 1977) and the myeloid leukemia cell line HL-60 (Collins et al., 1977). DNAs were prepared from the cell lines that grew out from these foci. These DNAs were cleaved with the restriction endonuclease Eco RI and analyzed by Southern transfer. The resulting blots were incubated with the human repeat sequence probe BLURS, which has 300 base pairs of homology to the human Alu sequences (Rubin et al., 1980). Figure 1 demonstrates a representative analysis of the DNAs of such primary transfectants. The foci induced by SW-480 colon carcinoma DNA carry a large number of DNA fragments that are reactive with the human Alu sequence probe BLUR-8. This confirms that these transfectants have taken up human DNA concomitantly with the acquisition of the transformed phenotype. The amount of donor human DNA taken up greatly exceeds that which might be associated with a transformation-inducing sequence. This finding reflects the results of other investigators (Wigler et al., 1979b; Perucho et al., 1980b), who have shown that transfected cells acquiring a selected marker will also display a large array of other donor sequences taken up during the same transfection event. These cotransfected donor sequences may constitute as much as 0.1% of the donor cell genome (Perucho et al., 1980b).
a
bcde
Figure 1. Detection of Human 3T3 Cells following Transfection
DNA Sequences in Transformed with SW-480 Colon Carcinoma
NIH/ DNA
Ten micrograms of the DNA of each focus were digested with the restriction endonuclease Eco RI, followed by agarose gel electrophoresis and Southern transfer as described in Experimental Procedures. The resulting filter was incubated with the BLUR-8 probe (Rubin et al., 1980). (Lane a) Contains normal human cellular DNA; (lane e) contains NIH/3T3 mouse cell DNA: (lanes b, c and d) contain DNA from primary foci of NIH/3T3 cells transformed by SW-480 human colon carcinoma DNA.
_
This Southern blot analysis distinguishes these primary foci from other, similarly appearing colonies of a transfected culture that are spontaneous overgrowths not resulting from DNA uptake. Thus we have discarded a series of candidate transfectants obtained in experiments in which human squamous cell and renal carcinoma donor DNAs were used, because the foci picked from these transfections gave no evidence of human repeated sequences (unpublished results) and the DNAs of these foci yielded no secondary foci upon repeated attempts at transfection. The rodent genome contains repeated sequences that are homologous to the highly repeated human sequences (Haynes et al., 1981; Krayev et al., 1980).
Human 357
Tumor
Oncogenes
We do not detect these sequences on our Southern blots (Figure 1, lane e), presumably because the mouse homologs are too mismatched to form a stable hybrid with the BLUR-8 probe. Analysis of the DNAs of the Secondary Foci Previous experiments have shown that transforming genes may be passaged serially by transfection (Shih et al., 1979, 1981; Cooper et al., 1980; Krontiris and Cooper, 1981). The DNA of a primary focus derived from transfection of tumor cell DNA can be prepared and used in a subsequent cycle of transfection. The resulting foci are termed secondary foci. As shown above, the transforming sequence in the primary transfectant is acquired together with a large number of cotransfected human sequences (Figure 1). The transforming gene is probably physically linked to these cotransfected human fragments within the recipient (Perucho et al., 1980b). During the manipulations of the subsequent transfection cycle, however, this array of human sequences will become fragmented and the human fragments will be diluted by the great excess of mouse DNA fragments derived from the chromosomal DNA of the primary transfectant (Shilo et al., 1980). Upon application of primary DNA to a recipient monolayer in the second tranfection cycle, the rare cell that acquires the transforming gene will also acquire a large number of the mouse chromosomal DNA fragments. This secondary transfectant, will acquire however, few if any human fragments besides those closely linked to the transforming gene. Other human fragments, present in the primary transfectant, will be lost because they are distributed to hundreds of other cells of the transfected culture that are discarded because they have not become transformed. Figure 2 shows a Southern blot analysis of secondary transfectants derived from DNAs prepared from the human colon carcinoma primary foci. Six secondary foci were picked after transfection of DNAs from three different primary foci. These foci were grown into mass cultures. Their DNAs were prepared and cleaved with endonuclease Eco RI (Figure 2, lanes a, b, c, d, e and f) prior to Southern analysis as described above. Each of the secondary transfectants carries an array-of DNA fragments detectable with the BLUR-8 probe. It is of central interest that each of these secondary transfectants exhibits a set of seven restriction fragments present in the DNA of all other secondary transfectants of this group. These fragments are indicated by arrows (Figure 2, lanes a-f) and range in size from 3 kb to 7 kb. A discrete set of DNA fragments is therefore associated with the presence of the human colon carcinoma transforming sequence. Most of the secondary foci have additional human DNA fragments besides those they have in common with others of the group. These fragments are other
abcdef
Figure ondary tion
2. Detection Foci Derived
g h i j
of Human DNA Sequences Common to the Secfrom SW-480 Colon Carcinoma DNA Transfec-
DNAs from secondary foci were analyzed with the BLUR-8 probe as in Figure 1. The Eco RI-digested DNAs in lanes a, b, c, d, e and f were independently derived from three primary foci. Secondary foci SW3-1 (lane a) and SW3-2 (lane f) were derived from primary focus SW3; SW2-3, SW2-2 and SW2-1 (lanes b. c and d) were derived from primary focus SW2; SWI-I (lane e) was derived from SW1 Arrows indicate the seven Eco RI-generated restriction fragments found in all six secondary foci. Double arrows indicate bands that are doublets not well resolved on this gel (see Figure 3A). SW3-1 DNA (lane a) and SWZ-3 DNA (lane b) were also analyzed after digestion with endonucleases Xba I (lanes g and h) and Pvu II (lanes i and j). Size markers (in kilobases) are indicated on the vertical ordinate.
human DNA sequences, likely to be found on either side of the transforming sequence, whose presence is not required for the functioning of this sequence. The common fragments seen in Figure 2, lanes af, are conserved in six independently derived secondary foci. Since these secondary foci originated from the DNAs of three independent primary transfectants, we conclude that the association of these fragments with the transforming sequence has survived two cycles of DNA passage and therefore must be intimately associated with this sequence. The data presented in Figure 2 can be used to estimate the size of the transforming sequence. The sum of the sizes of the conserved fragments shared among all the secondary colon carcinoma tranfectants should reflect the size of the functionally essential sequence required for expression of the transformed phenotype. This estimate is compromised because some restriction fragments may lack Alu sequences and thus escape detection in the Southern blot analysis. This problem can be minimized by the use of a
Cell 358
series of restriction endonucleases, because DNA fragments that become unlinked from Alu sequences after cleavage with one endonuclease may remain linked after cleavage with another endonuclease. The sum of the common Eco RI fragments in Figure 2, lanes a-f, gives an estimate of approximately 3.5 kb for the size of the colon carcinoma gene. The sum of the conserved Alu-containing Xba I fragments (lanes g and h) gives an estimate of 35 kb, and the sum of the Pvu II fragments (lanes i and j) gives an estimate of 25 kb. The human fragments observed in Figures 1 and 2 were detected by use of the BLUR-8 probe. We have obtained essentially identical results using as probes other BLUR clones, Cot = 1 selected human DNA and total HeLa cell DNA. The uncloned DNA probes hybridize to the same DNA fragments, and with similar intensity, as the BLUR clones do (data not shown). This indicates that the Alu sequences are the most abundant repetitive DNA family in the human genome (Houck et al., 1979). None of the repetitive DNA sequence probes used in these experiments detects any mouse sequences under the conditions used (see Figure 1, lane e). Moreover, these probes can be used interchangeably. Comparison of Transforming Genes from Other Human Tumor Cell Lines The ability of this approach to detect a characteristic pattern of restriction fragments encompassing the human colon carcinoma transforming sequence demonstrated that it would be a powerful tool for the comparison of human transforming sequences of other origins. The DNAs of secondary foci derived from the transforming sequence of the HL-60 tumor line all share a set of three Alu-containing DNA fragments (Figure 3C, arrows). These fragments are observed as well in a group of tertiary foci derived from the transformation of the DNA of the secondary transfected HL-60 (data not shown). It is noteworthy that the conserved fragments associated with the myeloid leukemia transforming sequence are distinct from those of the colon carcinoma gene. The DNAs of secondary foci derived originally from the EJ bladder carcinoma cell line present still another pattern. In this case, the DNA of each secondary focus contains a single large Eco RI, Alu-containing fragment (approximately 23 kb). The size of the Eco RI fragment varies slightly from focus to focus. This variation appears to reflect the difficulty of preserving intact the entire 23 kb fragment throughout transfection. Such variation was not encountered upon analysis of smaller sized Eco RI fragments (see above). The distinct pattern we observed indicates the existence of a transforming DNA sequence that is different from those of the colon carcinoma and myeloid leukemia cell lines.
-23 - 9.8 -6.6
-2.5
Figure 3. Comparison of Alu-Containing DNA Fragments Characteristic of Transforming Sequences from Three Different Human Tumor Cell Lines The DNAs of secondary foci derived from each human tumor cell line were digested with Eco RI and analyzed as described above. (A) Shows two secondary focus DNAs (SW3-1 and SWZ-3) derived from SW-480 hybridized with the BLUR-8 probe. The seven common colon carcinoma fragments are indicated by arrows. (B) Shows a single common fragment among four secondary focus DNAs derived from the human bladder carcinoma cell line EJ. The probe used in (B) was Cot = 1 DNA. (C) Shows secondary focus DNAs derived from the human promyelocytic leukemia cell line HL-60. Total HeLa cell DNA was used as the probe in (C) and detects three common restriction fragments. Size markers (in kilobases) are indicated to the right of the figure.
Discussion The application of DNA transfection procedures to the study of oncogenic transformation has demonstrated that the transformed phenotype is encoded in genomic DNA. This experimental approach also allows the transforming sequence to be dissociated from all other irrelevant DNA sequences in the tumor cell line and to be studied directly. The affiliation of human tumor transforming sequences with highly repeated DNA sequences makes possible a number of conclusions concerning the structure of these sequences. The results show that the transforming activity of each of these DNAs is associated with a specific, definable set of DNA sequences. The transformation of cells following transfection is therefore not a consequence of the uptake and fixation of exogenous DNA, as such, but rather a reflection of the presence of a discrete array of DNA sequences that probably encompasses a transforming gene. An analogous conclusion was drawn in earlier work on the DNA of 3-methylcholanthrene-transformed mouse fibroblasts (Shilo et al., 1980). In that
Human 359
Tumor
Oncogenes
case, defined sequence tags were interspersed experimentally throughout the genomic DNA of the transformed donor cell prior to transfection. Subsequent analysis revealed the presence of the same sequencetag-containing DNA fragments in the DNAs of all secondary foci, thus implying the necessity for transfer of a unique DNA fragment to elicit transformation. In a subsequent study, we demonstrated that the transforming sequences of four different, 3-methylcholanthrene-transformed mouse fibroblast lines were susceptible or resistant to the same set of restriction enzyme treatments (Shilo and Weinberg, 1981). This suggested that the transforming genes in these four, independently induced lines represent four different, allelic variants of the same cellular gene. Less complete analysis of the transforming genes of three different ethylnitrosourea-induced rat neuroblastomas indicated that the transforming sequences of these three tumor lines behaved identically to one another after restriction endonuclease treatment (C. Shih, unpublished data). Thus, within a given tumor cell type (for example, 3-methylcholanthrene-transformed fibroblasts), we suggested that perhaps only one transforming gene may be activated. These earlier studies, however, could not indicate whether the same or different transforming sequences were present in different tumor types because the two groups of donor tumor cell lines derived from two different host species. The present data allow a preliminary resolution of this issue, since three different transforming genes from the same species are being analyzed. We conclude that different tumor types appear to carry different transforming sequences. Because of the absence of additional biologically active DNAs of these three types of human tumors, we do not yet know whether all human tumors of a given type (for example, colon carcinoma) carry the same active transforming gene in their DNA. Nevertheless, we can already consider it unlikely that activation of a single cellular oncogene is responsible for all types of human tumors. The genes transferred in these experiments derive from tumor cell lines. Generally, these cell lines are difficult to establish and require long-term culture to become stable (Fogh and Trempe, 1975). We have assumed that the transforming genes originally activated in the tumor are responsible for the continued transformation of the resulting cell lines, and for the transformation of mouse fibroblasts following transfection. This issue will be clarified by isolation of these genes directly from the DNA of tumor samples. The characterization of human tumor genes described here is applicable to the analysis of any transfectable human gene for which a selection is available. Thus one might use parallel procedures to analyze DNAs of secondary transfectants that have acquired
selectable enzyme markers or drug resistance markers. In principle, the molecular cloning of such genes might follow, by introducing the DNA of a secondary tranfectant into lambda phage vectors and screening the resulting genomic library for those components reactive with human repeat sequence probes. Others have reported the successful cloning of the cellular thymidine kinase and adenosine phosphoribosyl transferase genes by experimentally linking bacterial plasmids or plasmid fragments to these marker genes prior to serial transfections (Perucho et al., 1980a; Lowy et al., 1980). The resulting juxtaposition of these foreign sequences with the gene of interest allows the cloning of these genes by use of plasmidmediated rescue (Perucho et al., 1980a) or by the isolation of chimeric lambda phage carrying the identifiable plasmid sequences linked to the marker gene (Lowy et al., 1980). Exogenous sequence tags are of limited use when cloning large genes such as the transforming genes identified here, because the tag only allows isolation of a closely linked terminal fragment of the gene. Isolation of the remainder of the gene requires use of a chromosome “walking” protocol. In contrast, the presence of interspersed repeated sequences provides a naturally occurring sequence tag whose linkage to a marker gene need not be established by difficult experimental procedures. Moreover, these interspersed sequences should make possible identification and isolation of fragments spanning the entirety of the gene of interest. This strategy should be applicable to analysis of transfected marker genes that are introduced into a recipient cell of a heterologous species. These highly repeated gene sequences appear to evolve far more rapidly than do unique DNA sequences (Gusella et al., 1980). Thus one should be able to detect, specifically, repeated DNA sequences of one species introduced into a recipient cell of another species of sufficient evolutionary distance. Experimental
Procedures
Human Tumor Cell Lines The human bladder carcinoma cell line SW-480 (Leibowitz et al., 1976) was obtained from J. Fogh, the human promyelocytic leukemia cell line HL-60 (Collins et al., 1977) was obtained from R. C. Gallo and the human bladder carcinoma cell line EJ (Marshall et al., 1977) was obtained from I. Summerhayes and L. 6. Chen. DNA Transfection Assays DNA transfection procedures were essentially as described elsewhere (Andersson et al., 1979). In each transfection. 75 pg of DNA were applied to 1.5 x 1 O6 NIH/3T3 cells, and the appearance of foci was scored 14 days later. Transfection of chromatin of the SW-480 cell line was performed as described by Miller and Ruddle (1978), with modifications described by Shih et al. (1979). This transfection yielded three primary foci whose morphology conformed to that described previously (Shih et al., 1979) and whose DNA contained human Alu sequences (Figure 1). Transfection of DNA extracted from the HL-60 cell line was performed either with or without an admixture of copies of a molecular cDNA clone of dihydrofolate reductase (M.
Cell 360
Murray, Fi. Kaufman, S. Latt, A. Cassill and R. A. Weinberg, manuscript in preparation). One primary focus was obtained from the cotransfection of HL-60 DNA and the dihydrofolate reductase cDNA clone. This focus was also methotrexate-resistant, and its DNA is analyzed here. Transfection of HL-60 DNA alone yielded 12 primary foci. Cells of the HL-60 primary foci exhibited an extremely hyperrefractile. spindle-shaped morphology. Derivation of foci by transfection of EJ cell DNA was previously reported (Shih et al., 1981). Analysis of Transformed Focus DNAs Whole cell DNA was prepared from transformed cells grown from foci obtained after transfection with tumor cell line DNAs. Following restriction endonuclease digestion, 10 pg of the DNA of each focus were resolved by electrophoresis through a 1% agarose gel in 40 mM Tris (pH 7.91, 50 mM sodium acetate and 1 mM EDTA. After electrophoresis. the DNA was transferred to nitrocellulose by the method of Southern, and the resulting blots hybridized with probes for human repetitive DNA. The BLUR probes (Rubin et al., 1980) were obtained from D. Jolly and P. Deininger. The Cot = 1 probe was prepared by heat-denaturing sonicated human DNA at 1OO’C and reannealing it at 68°C to a Cot value of 1. This was followed by Sl nuclease treatment for 1 hr at 37°C and hydroxyapatite chromatograpy to isolate repetitive DNA duplexes. Total HeLa cell DNA was sheared through a 20 gauge needle prior to use as a probe. Each probe was “P-labeled by nick translation (Rigby et al., 1977). We used 50 pCi of each of the four 3ZP-labeled deoxyribonucleotides. The labeled probes (5 X 1 O6 cpm) were incubated with the Southern blots for 24 hr in the presence of 10% dextran sulfate (Wahl et al., 1979) at 4O’C. When HeLa DNA was used as a probe, sonicated NIH/3T3 mouse cell DNA was added to the prehybridization solution at 50 pg/ml and to the hybridization solution at 20 AS/ml.
segments from specific USA 77,2829-2833.
human
chromosomes.
Proc.
Nat. Acad.
Sci.
Haynes, S., Toomey, T., Leinwand, L. and Jelinek, W. (1981). The Chinese hamster Alu-equivalent sequence: a conserved highly repetitious interspersed DNA sequence in mammals has the structure suggestive of a transposable element. Mol. Cell. Biol. 7, 573-583. Houck, C. M.. Rinehart, F. P. and Schmid. C. W. (1979). A ubiquitous family of repeated DNA sequences in the human genome. J. Mol. Biol. 132, 289-306. Jelinek, W. R., Toomey, T. P., Leinwand, L., Duncan, C. H., Biro, P. A., Choudary, P. V., Weissman, S. M., Rubin, C. M., Houck, C. M., Deininger. P. L. and Schmid, C. W. (1980). Ubiquitous, interspersed repeated sequences in mammalian genomes. Proc. Nat. Acad. Sci. USA 77, 1398-1402. Klein, G., ed. (1980).
Viral Oncology.
(New
Krayev. A. S., Krameriv, D. A., Skryabin, A. A. and Georgiev, G. P. (1980). The ubiquitous repetitive DNA sequence 61 abundant class of mouse fold-back RNA. 1215.
York:
Raven
Press).
K. G., Ryskov, A. P.. Bayer, nucleotide sequence of the complementary to the most Nucl. Acids Res. 8, 1201-
Krontiris, T. G. and Cooper, G. M. (1981). Transforming activity human tumor DNAs. Proc. Nat. Acad. Sci. USA 78, 1181-l 184.
of
Leibovitz, A., Stinson, J. C., McCombs, W. B., McCoy, C. E.. Mazur, K. C. and Mabry, N. D. (1976). Classification of human colorectal adenocarcinoma cell lines. Cancer Res. 36, 4562-4569. Lewis, W. H., Srinivasan, P. R., Stokoe. N. and Siminovitch, L. (I 980). Parameters governing the transfer of the genes for thymidine kinase and dihydrofolate reductase into mouse cells using metaphase chromosomes or DNA. Somatic Cell Genet. 6, 333-347. Lowy, I., Pellicer, A., Jackson, J. F., Sim. G.-K., Silverstein, S. and Axel, R. (I 980). Isolation of transforming DNA: cloning the hamster aprt gene. Cell 22, 817-823.
Acknowledgments We would like to thank R. Gallo, J. Fogh, I. Summerhayes and L. B. Chen for providing cell lines: D. Jolly and P. Deininger for providing the BLUR clones; Aaron Cassill for excellent technical assistance: J. Toole, C. Schmid and W. Jelinek for helpful discussions. M. M. was a fellow of the Damon Runyon-Walter Winchell Cancer Fund. B. S. was a Chaim Weizmann postdoctoral fellow. This work was supported by grants from the National Cancer Institute to R. A. W. and to S. E. Luria. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Marshall, C. J., Franks, L. M. and Carbonell, A. W. (1977). Markers of neoplastic transformation in epithelial cell lines derived from human carcinomas. J. Nat. Cancer Inst. 58, 1743-l 747.
Received
Rinehart, F. P., Ritch, T. G., Deininger, P. L. and Schmid, C. W. (1981). Renaturation rate studies of a single family of interspersed repeated sequences in human DNA. Biochemistry 20, 3003-3010.
May 8, 1981;
revised
May 28, 1981
References Andersson. P., Goldfarb, M. P. and Weinberg, R. A. (1979). A defined subgenomic fragment of in vitro synthesized Moloney sarcoma virus DNA can induce cell transformation upon transfection. Cell 76, 6375. Collins, S. J., Gallo, R. C. and Gallagher, R. E. (1977). growth and differentiation of human myeloid leukemic pension culture. Nature 270, 347-349.
Continuous cells in sus-
Cooper, G. M., Okenquist. S. and Silverman, L. (1980). Transforming activity of DNA of chemically transformed and normal cells. Nature 284, 418-421. Fogh, J. and Trempe, G. (1975). New human tumor cell lines. In Human Tumor Cells in Vitro. J. Fogh, ed. (New York: Plenum Press), pp. 115-l 59. Graf. L. H., Urlab, G. and Chasin, L. (1979). gene for hypoxanthine phosphoribosyltransferase. net. 5, 1031-1044.
Transformation Somatic
of the Cell Ge-
Gusella, J. F., Keys, C., Varsanyi-Breiner, A., Kao, F., Jones, C., Puck, T. T. and Housman, D. (1980). Isolation and localization of DNA
Miller, C. L. and Ruddle, F. H. (I 978). Co-transfer of human X-linked markers into murine somatic cells via isolated metaphase chromosomes Proc. Nat. Acad. Sci. USA 75, 3346-3350. Perucho, M., Hanahan, D., Lipsich, L. and Wigler, M. (1980a). Isolation of the chicken thymidine kinase gene by plasmid rescue. Nature 285, 207-210. Perucho, M., Hanahan. D. and Wigler, physical linkage of exogenous sequences 22, 309-317.
M. (1980b). Genetic in transformed cells.
and Cell
Rubin, C. M., Houck, C. M., Deininger. P. L., Friedmann, T. and Schmid, C. W. (1980). Partial nucleotide sequence of the 300-nucleotide interspersed repeated human DNA sequences. Nature 284, 372-374. Schmid, C. W. and Deininger, P. L. (1975). the human genome. Cell 6, 345-358.
Sequence
organization
of
Shih, C., Shilo. B.. Goldfarb, M. P., Dannenberg, A. and Weinberg, R. A. (I 979). Passage of phenotypes of chemically transformed cells via transfection of DNA and chromatin. Proc. Nat. Acad. Sci. USA 76, 5714-5718. Shih, C., Padhy, L. C., Murray, M. J. and Weinberg, Transforming genes of carcinomas and neuroblastomas into mouse fibroblasts. Nature 290, 261-264. Shilo, B. and Weinberg, carcinogen-transformed
R. A. (1981). introduced
R. A. (1981). Unique transforming mouse cells. Nature 289, 607-609.
gene
in
Shilo, B., Shih, C., Merrick, M., Goldfarb, M. and Weinberg, R. A. (1980). Initiation of molecular cloning of a transforming allele of chemically transformed mouse cells. In Mobilization and Reassembly of Genetic Information, Miami Winter Symposium, W. A. Scott, R.
Human 361
Tumor
Oncogenes
Werner, D. Ft. Joseph Press), pp. 275-284.
and J. Schultz,
eds.
(New
York:
Academic
Tashima. M., Calabretta, B., Tore& G., Scofield, M., Maizel, A. and Saunders, G. F. (1981). Presence of a highly repetitive and widely dispersed DNA sequence in the human genome. Proc. Nat. Acad. Sci. USA. 78, 1508-l 512. Wahl, G. M., Stern, M. and Stark, G. Ft. (1979). Efficient transfer of large DNA fragments from agarose gels to diazobenzyloxymethylpaper and rapid hybridization by using dextran sulfate. Proc. Nat. Acad. Sci. 76, 3683-3687. Wigler, M., Pellicer, A., Silverstein, S. and Axel, R. (1978). Biochemical transfer of single-copy eucaryotic genes using total cellular DNA as donor. Cell 74, 725-731. Wigler, M., Pellicer, A., Silverstein, S., Axel, R., Urlab, G. and Chasin, L. (1979a). DNA-mediated transfer of the adenine phosphoribosyltransferase locus into mammalian cells. Proc. Nat. Acad. Sci. USA 76, 1373-i 376. Wigler, M.. Sweet, R., Sim, G. K., Wold, B., Pellicer, A., Lacy, E., Maniatis, T.. Silverstein, S. and Axel. R. (1979b). Transformation of mammalian cells with genes from procaryotes and eucaryotes. Cell 16, 777-785. Wigler, M., Perucho, M., Kurtz, D.. Dana, S., Pellicer, A., Axel, Ft. and Silverstein. S. (1980). Transformation of mammalian cells with an amplifiable dominant-acting gene. Proc. Nat. Acad. Sci. USA 77, 3567-3570.