Ann. Rev. Genet. 1984. 18553412 Copyright D 1984 by Annual Reviews Inc. All rights reserved 'THE MOLECULAR GENETICS OF CELLULAR ONCOGENES J Harold E. Vurmus Department of Microbiology and Immunology, University of California, San Francisco, California 94143 CONTENTS INTRODUCTION ............................................... _.._......... DEFINITIONS OF CELLULAR ONCOGENES ................................................ c-onc's: The Proto-Oncogenic Progenitors of Retroviral Oncogenes . . . . . . . . . . . . . . . . . . . Identification of Cellular Oncogenes by Functional Tests for Dominant Mutations Using Genetic Rearrangements to Identify Oncogenes.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Proviral insertions can activate expression of cellular proto-oncogenes . . . . . . . . . . . Oncogenes at translocation breakpoints.. . . , , . . . . , . . . . . . . . . . . . . . . . . Oncogenes in amplified DNA . , , ... . . ..... , , ... .... , .. .. . .... . .. .. _.. . .._. .... . .. .. .. .. . ... MOLECULAR PROFILES OF ACTIVATED ONCOGENES.. . . . . . . . . Base Substitutions in Mutant ras Alleles ................................ Insertionally Mutated c-myc Loci.. . . . . . . . . . . . , . . , . . . . , , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ~ Structural and Functional Properties of Translocation Bre Oncogenes ............................................... FROM c-onc TO v-onc: COLLUSION OF MUTATIONAL MECHANISMS The Favored Model for Transduction of Oncogenes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Functional Significance of Differences Between c-onc's and v-onc's . . . . . . . . . . . . . . INACTIVATION OF INTEGRATED v-onc GENES: A SAMPLER OF MUTATIONAL MECHANISMS ..... ...................................... NEW DIRECTIONS TOWARD A MOLECULAR DEFINITION OF CANCER .................................. .... ble Hosts . . . . . . . . . . . . . . . . . . . . . . . . . . Identifiing Genes Whose Products Might Interact with Oncogene Proteins ..... .... ........ .... ........... , .... . A CONCLUDING PERSPECTIVE.. . . . 554 555 555 558 562 563 566 567 570 570 574 574 576 578 581 581 583 586 587 587 588 590 592 593 595 595 596 597 597 598 553 0066-4197/84/12 15-0553$02.00 554 VARMUS INTRODUCTION Eventually the techniques of nucleic acid chemistry should allow us to itemize all the differences in nucleotide sequence and gene expression that distinguish a cancer cell from its normal counterpart, and perhaps at that point the steps involved in carcinogenesis will cease to be in doubt. (25) Widespread interest in cellular oncogenes as potential substrates for the somatic mutations believed to underlie neoplastic change has brought students of cancer closer than could reasonably have been expected a mere four years ago to meeting the challenge laid down by Cairns (25). Efforts to define cancerous lesions at the nucleotide level have, moreover, been instructive on a broader front, providing a panoramic view heretofore unavailable of the kinds of mutations that afflict the somatic cells of higher eukaryotes. My objectives in this review are to summarize the nature of the recently defined mutations of cellular oncogenes and to evaluate the hypothesis that they are giving us an accurate picture of both genetic change in somatic cells and the mutational basis of cancer. This evaluation must take into consideration the difficulties inherent in genetic attacks upon oncogenes: the poorly defined complexity of neoplastic phenotypes; the co-existence of multiple genetic changes in a single tumor cell; and the uncertain weight to be given to various lesions that may influence gene expression, alter the amino acid sequence of an oncogene product, or both. It is expressly not my intention to review the structure, origin, and biochemical functions of oncogenes; this has been accomplished in many other places in the recent past (1618, 39, 253, 254). However, it seems appropriate to begin by describing the experimental strategies used to identify normal and mutant cellular oncogenes (Table l), in part because the extent to which these strategies depart from classical genetics underlies current difficulties in correlating phenotypic changes in tumor cells with individual mutations. Table 1 Proto-oncogenes Active oncogenes c-one's v-onc's' Wild-type form of transforming genes Wild-type form of rearranged genes Categories of cellular oncogenes based on mode of discovery Transforming genesb Rearranged genes` a Cell-derived sequences to which oncogenic activity has been ascribed in retroviral Cellular genes competent to transform an appropriate recipient cell, e.g. NIHi3T3 Cellular genes altered by insertion mutations (Table 4). chromosomal transloca- genomes (see Table 2). cells (see Table 3). tions (Table 5). amplification (Table 6). or other rearrangements. CELLULAR ONCOGENES 555 DEFINITIONS OF CELLULAR ONCOGENES Much of the recent work in molecular oncology is motivated by faith in the proposal that a neoplastic cell develops from its normal progenitor as a con- sequence of changes (probably multiple) in some members of a restricted set of cellular genes. Such changes could be epigenetic, but the prevailing belief that cancer has its origins in mutations (24, 25) and the persuasive power of identified alterations in the structure of DNA (the subjects of this review) have focused attention primarily upon genetic events. The mutant genes are known as oncogenes, and corresponding wild-type alleles are called proto-oncogenes (or normal cellular oncogenes). c-onc's: The Proto-Oncogenic Progenitors of Retroviral Oncogenes Members of the largest group of proto-oncogenes have been identified ex- perimentally by the homology of their nucleotide sequences to retroviral oncogenes, those regions of certain retroviral genomes believed to be responsi- ble for swift induction of tumors and neoplastic transformation of cultured cells. The homologous cellular genes, known for convenience as c-om's (39, are deemed proto-oncogenes because they have apparently served as targets for genetic transduction by retroviruses and thus are the normal cellular progenitors of viral oncogenes (v-onc's) (17, 18, 218) (Table 2). Although the dividends of conferring the status of proto-oncogenes upon these cellular genes have been considerable, it must be acknowledged that the basis for doing so, the genetic definition of v-onc's, has not been uniformly rigorous. At one extreme, numerous conditional and non-conditional mutants of the src gene of Rous sarcoma virus (RSV) have convincingly demonstrated that v-src is necessary for both the initiation and maintenance of the trans- formed phenotype (125); other mutants of RSV, as well as manipulation of cloned v-src DNA, have established that the v-src gene is also sufficient for transformation (72, 122, 133a). At the other extreme, host-derived sequences present in the genomes of highly oncogenic viruses have sometimes been granted the provisional designation of v-onc's without supporting genetic evidence, even when the genome is also inhabited by another v-om for which there is genetic confirmation of neo-plastic activity. Belief that such sequences have oncogenic potential of their own or the capacity to enhance the oncogenic- ity of accompanying oncogenes (e.g. by broadening the range of cellular targets for neoplastic effects) can be traced to at least three sources: (a) transduction of each cellular gene is a rare event for which there is likely to be some selective advantage for the virus (e.g. tumorigenic capacity); (b) in some cases of retroviruses bearing components derived from two independent cellu- Table 2 Retroviral oncogenes (v-onc's)" < G Genetic basis Species of Conditional Non-conditioflal In vitro Multiple transductiond 5 Prototype virus(es)k v-om origin mutantsb mutantsb recombinantc RSV Yl3lESV FuSVlPRCII UR2 MC29 MH2 AMV E26 AEV SKV770 REV-T Mo-MSV Ab-MLV MSV-3611 BALB-MSV ~ src yes' fpsf ros myc myc milg mYb mYb ets erbA erbB ski re1 mos ab1 raf Ha-rash chicken chicken chicken chicken chicken chicken + chicken chicken chicken chicken turkey mouse mouse mouse mouse + + + + + + + + + + + + + + + Ha-MSV RaSV Ki-MSV FBJ-MSV FBR-MSV ST-Fe SV, GA-FeSV SM-Fe SV GR-Fe SV PI-FeSV. SSV Ha-rash Ra-rash Ki-ras fos fos fox fes' fms fgr e actin' sis rat rat rat mouse mouse cat cat cat cat, woolly monkey + + + t + + + + + + + + + + + ~~~ 'Further description and bibliography can be found in (1618). bDetails and references can be found in (17, 18, 125. 126). + means that the v-onc sequence has been found to have transforming activity when moved to a new context by in vitro manipulation of cloned DNA (see 19). + means that the onc sequence has been encountered in two or more viruses that appear to have independently transduced the same c-onc from one or more host species (see 17. 18). * v-yes and v-fgr appear to be derived from the same cognate c-onc (144a). 'v-& and v-fes are derived from the same cognate c-onc (206). sv-mil (also known as v-mht) and v-rufare derived from the same cognate c-onc (102, 225). hHa-rus of BALB-MSV (also known as v-bus), Hams of Ha-MSV, and Ra-rus of RaSV are all derived from the same cognate c-onc. called c-Ha-rus (6, 181). 'Part of the mammalian p-actin sequence is found in the genome of GR-FeSV (144a). jA mutant of MH2 that appears not to express v-mil has normal or augmented oncogenic activity (161). 'Abbreviations: RSV. Rous sarcoma virus; Y73, Yamaguchi 73 sarcoma virus; ESV, Esh sarcoma virus; FuSV, Fujinarni sarcomavirus; MC29, myelocytomatosis-29 virus; MH2. Mill Hill-2 virus; AMV, avian myeloblastosis virus; AEV, avian erythroblastosis virus; REV-T, reticulo-endotheliosis virus, strain T; Mo-MSV, Moloney murine sarcoma virus; Ab-MLV, Abelson murine leukemia virus; Ha-MSV, Harveymurinesarcoma virus; RaSV, Rasheed rat sarcoma virus: Ki-MSV. Kirsten murine sarcoma virus; ST-, GA-, SM-, GR-, PI-FeSV, Snyder-Theilen, Gardner-Arnstein, Susan McDonough. Gardner-Rasheed. Parodi-lrgens feline sarcoma viruses; SSV, simian sarcoma virus. 558 VARMUS tar genes (e.g. v-myc and v-mil in MH2 virus), both co~ponents have been found singly in other transforming retroviruses (e.g. MC-29 and MSV-3611) (102,225); and (c) in the case of avian e~throblastosis virus (AEV), mutants constructed in vitro suggest that one v-onc sequence (v-erbA) can potentiate the oncogenic action in the erythropoietic cell lineage of another, more critical sequence (v-erbB) (70,203). For the majority of v-om's there is at least some supportive genetic evidence for oncogenic potential, although usually no direct evidence for a role in maintenance, since conditional mutations have only been isolated in a few cases (Table 2). More commonly the evidence is restricted to multiple isolations of the same oncogene in different highly oncogenic viruses and the retention of transforming activity when the viral oncogene is placed in a new context by the manipulation of cloned DNA. For several years, the contention that the cellular progenitors of retroviral oncogenes might be the instruments of oncogenic change rested upon supposi- tion. Even before the structural differences and similarities between c-onc's and vLonc's were recounted with the high resolution now possible, the two versions of each gene were recognized to be sufficiently alike to propose that partial recapitulation of events that produce active v-onc's from c-one's might occur during tumorigenesis. Moreover, e-onc's are well-conserved during metazoan evolution, commonly expressed in most or all tissues, and sometimes regulated in temporal or lineage-specific fashion (1 6), implying that they might play a central role in growth and development, processes that seem particularly likely to be altered during neoplasia. [The latter idea has received dramatic recent confirmation by the findings that c-sis appears to encode the platelet- derived growth factor (59, 251) and c-erbB the epidermal growth factor receptor (60).] However, new approaches were required to implicate c-UIZC'S directly in neoplasia. Identi~cat~on of Cellular Oncogenes by Functionat Tests for Dominant Mutations The second means to identify cellular oncogenes perceives active oncogenes rather than proto-oncogenes. DNA transformation procedures, used for several years to demonstrate, for example, the oncogenic activity of nuclear DNA from cells containing integrated genomes of DNA and RNA tumor viruses, elicit neoplastic transformation when DNA from a wide variety of tumor cells is introduced into established lines of rodent fibroblasts, most commonly NIW 3T3 cells (39,43,207a, 253). Identified sources of active DNA by now include primary human tumors and cell lines; spontaneous and virus-, chemical-, and radiation-induced tumors of rodents and birds; and chemically-transformed cell lines (Table 3). The interpretation of such findings-an interpretation for which there is now considerable support (see below)-is that neoplastic wansfonna- tion of the recipient cells is dependent upon the acquisition of a mutant allele. CELLULAR ONCOGENES 559 The mutant allele is believed to act in a dominant manner in both recipient and donor cells, and its effect upon the recipient cells is alleged to simulate its biological activity in the donor cells. The presumption that a mutated gene is present in transformation-competent DNA was based initially upon the failure to induce the transformed phenotype in recipient cells with DNA similarly prepared from normal cells. However, the existence of a mutant gene in competent DNA requires more direct proof for several reasons: (a) sheared DNA from normal sources can transform cells at low frequency (43), presum- ably because under these circumstances strong promoters and proto-oncogenes from either donor or recipient cells can recombine to generate an efficiently expressed oncl )gene; (b) some wild-type c-onc proto-oncogenes (e.g. c-mos, c-Ha-rusl, and c-fos) can transform cells in culture when placed under the control of active transcriptional promoters (20, 31, 137, 159); and (c) rare acquisition of multiple copies of proto-oncogenes during transfection could presumably produce neoplastic change (akin to that proposed to occur as a result of gene amplification or transcriptional activation) without mutation in the donor cells. Although the assay for oncogenes in NIW3T3 cells has stimulated remark- able advances in our understanding of oncogenic mutations, it has come under criticism for an obvious reason: it offers a functional test for genetic lesions in a cultured mouse fibroblast that usually differs both in cell lineage and species from the tumor cells whose genes are under scrutiny. This potential weakness may well provide at least part of the e~p~a~a~on for two of the most striking characteristics of collective experience with the NIH/3T3 assay: the failure to observe transforming activity in DNA from 80-90% of spontaneous human tumors (39, 112, 174, 253) and the high proportion of active oncogenes that have proven to be mutant members of the rus gene family (see below). Recent attempts have been made to exploit other types of cell lines (193) or different assays for neoplastic effect (1 19; M. Wigler, personal communication) to seek additional oncogenes by DNA transfection. Although these alterations in methodology still have generally led to isolation of mutant ras genes, M. Wigler and colleagues (personal communication) have tentatively identified novel oncogenes in a human mammary carcinoma line, MCF-7, by treating NIH13T3 cells with tumor cell DNA and a plasmid bearing a dominant genetic marker, then observing for tumor growth after inoculation of mice with pooled, co-transfected cells. The NIH/3T3 cell assay also has been criticized on the grounds that the recipient cells, from an established line, are already abnormai. Inspired by requirements for two adeno- or polyoma virus genes to achieve full transforma- tion of primary rodent embryo cells (62,182), Land et al(114) and Ruley (192) have used embryo cells to show that two activated, molecularly cloned cellular oncogenes coliaborate to produce a mo~holog~cally transfo~ed cell with 560 VARMUS Table 3 Candidate oncogenes detected as transforming genes in the NIH/3T3 cell assay Examples of cell types in which Genes transforming activity is found" c-Ha-ras 1 Bladder carcinoma cell line (human) Urinary tract tumors (human) Lung carcinoma cell line (human) DMBA and NMU-induced mammary carcinomas DMBAmPA-induced skin papillomas, benign and Melanoma cell line (human) (rat) malignant (mice) References 55, 83, 112, 164, 70a 260 220; see footnote c 8, 9 197 below 2 Mammary carcinosarcoma line (human) MC, BP, DEN and MNNG-transformed primary Myeloid tumor cell line (mouse) Colon carcinomas and cell lines (human) See footnote d below 223 246a c-Ki-ras2 -- Lung carcinomas and cell lines (human) 55, 174, 135 55, 174, 135 cells (guinea pig) Pancreatic carcinoma cell line (human) Gall bladder carcinoma cell line (human) Rhabdomyosarcoma cell line (human) Ovarian carcinoma, primary (human) Acute lymphocytic leukemia line (human) hfC-induced fibrosarcomas and MC-transformed y-irradiation induced thymoma (mouse) MC-induced thymic lymphoma and macrophage fibroblasts (mouse) 174 174 174 68 67 65, 165,208 88 246a N-ras tumor lines (mouse) BP-induced fibrosarcoma (mouse) 246a Neuroblastoma cell line (human) 210 Burkitt lymphoma line (human) 77 Fibrosarcoma and rhabdomyosarcoma cell lines Promyelocytic leukemia line (human) 143 T cell leukemia line (human) 213a Melanoma cell lines (human) 2; see footnote e Teratocarcinoma line (human, late passage) 228a Lung carcinoma line (human) 259a Acute and chronic myeloblastic leukemia (human) 67, 77 Lung carcinoma cell line (mouse) 246a 89, 133 (human) below Carcinogen-induced thymoma (mouse) 88 c-mos Plasmacytoma cell lines (mouse) 36, 78, 184 c-erbB-related ENU-related neuroblastoma cell lines (rat) 168, 208, see foot- note f below CELLULAR ONCOGENES 56 1 Blym Tlyrn tx-lb tX-Zb tX-3b tX-4b Unnamed Unnamed ALV-induced bursal lymphomas (chickens) Burkitt lymphoma lines (human) Intermediate T cell lymphoma lines (human) MLV-induced T cell lymphomas (mouse) Mammary carcinoma cell line (human) MMTV and DMBA-induced manima~ carcinomas PfeB tumor cell lines (human) AbMLV-induced pre B tumor cell lines (mouse) Myeloma and plasmacytoma cell lines (human, Mature T cell lymphoma lines (human, mouse) Pre B cell leukemia cell line (human) MNNG-t~ated osteosa~oma cell line (mouse) mouse) 41, 42, 85 56 117, 118 117, 118 116 116 I I7 115 117 117 160 38b * 6bbreviations: MC, methylcholanthrene; DMBA, dimethylbenzanthracene; TPA, tetradecanoyl- 13-phorbol acetate; NMU, nitrosomethylurea; MNNG, N-methyl-N-nitro-N-nitroguanidine BP. benzfafpyrene; DEN, diethylnitrosamine; ENU, ethylnitrosourea. bNomenclature suggested by M.A. Lane and G. Cooper (personal communication). `M. Barbacid, personal communication. dM. Kraus et al. personal communication. `R. A. Padua, personal communication. R. Weinberg, personal communication. tumorigenic potential. These results are provisionally interpreted to mean that one of the genes (an active form of myc, adenovirus ElA, or polyoma virus large T antigen gene) provides a nuclear function akin to that presumed to be active in established lines and required for prolonged growth, whereas the second gene (an active member of the rus gene family, adenovirus ElB, or polyoma middle T antigen gene) provides a cytoplasmic function responsible for the properties conventionally associated with the transformed phenotype (114). This sort of test appears to be particularly helpful for assessing the functional properties of putative oncogenes cloned from tumor cells (rather than for discovering new oncogenes) and for determining the contribution made by each member of a collaborating pair to the oncogenic phenotype in the recipient cell. However, it is uncertain whether two oncogenes are inevitably required (and hence whether they can rightly be called complementary genes): D. Spandidos & N. Wilkie (personal communic~tion) have shown that a mutant c-Hams1 allele expressed under the control of a strong heterologous promoter does not require co-transfection with a myc-like gene for full transformation of primary rodent cells. The l~keli~ood that multiple mut~tions conspire to produce a full-fledged 562 VARMUS cancer cell raises the disconcerting possibility that some of the mutations have phenotypic consequences that cannot be perceived in the absence of the others. One way to deal with this dilemma may be to exploit the traditional strategem of isolating phenotypic revertants of neoplastic cells (259). For example, partial phenotypic revertants of the human fibrosarcoma line HT1080 can be returned to a more advanced neoplastic state by transfection with the mutant N-ras gene formerly isolated from HT1080 cells with the NIW3T3 cell assay (C. J. Marshall, personal communication). In a modest variation on this theme, non-tumorigenic cell lines established from human bladder cancers have been rendered tumorigenic by introduction of a mutant c-Ha-rusl gene from the human bladder cancer line, EJ (C. J. Marshall, personal communication). Again, however, phenotypic revertants or tumor cells with partially trans- formed phenotypes have not yet been used as target cells in DNA transfection studies designed to isolate new oncogenes from unfractionated tumor DNA. The accumulated attempts to identify active oncogenes by transformation of cultured cells with tumor cell DNA have yielded a number of candidate oncogenes (Table 3), but the majority of positive results can be traced to one of three members of the ras gene family, a group of genes that encode closely related proteins generally about 21 kd in size. Two of these genes, known in the human genome as c-Ha-rasl and c-Ki-ras2 (see footnotes to Tables 2 and 3), are members of the c-onc group; the third is called N-ras, a gene recognized as a member of the rus group by immunological cross-reactions of its product with antisera raised against products of other rus genes and, more definitively, by comparison of the deduced amino acid sequence of N-rus protein with corre- sponding sequences from other family members (210, 229). Cooper and Lane and their colleagues have used the NIH/3T3 cell assay to implicate several genes outside the rus family as oncogenes; each oncogene is apparently specific for tumor cells within a cell lineage or at a certain stage in a cell lineage, as often deduced from the lists of restriction endonucleases that do and do not inactivate transformation competence (41,42, 115-1 17) (Table 3). However, among these putative oncogenes only chicken and human Blym have been molecularly cloned and sequenced, and for these genes only transforma- tion-competent alleles are available (56, 85). Hence, in no instance outside of the rus group has a mutation that renders a proto-oncogene competent to transform NIH/3T3 cells been proven by delinition of the nucleotide sequence. Using Genetic Rearrangements to Identify Oncogenes Some circumstantial evidence is very strong, as when you find a trout in the milk. (Thoreau) Identification of genes that are structurally altered in tumors has proven to be an important if difficult means to register candidate oncogenes. Implicit in this endeavor is the assumption that a gene found to be altered in most cells in a CELLULAR ONCOGENES 563 tumor is likely to be selected for and therefore contributory to the neoplastic phenotype. Clearly this is a hazardous simplification, particularly if, as sus- pected from some experiments (e.g. 32), cancerous cells have an altered mutation rate or if some extraneous mutations are highly favored to occur. To achieve a credible identification of a putative oncogene through a purely structural analysis, multiple sightings of the same gene in mutant form would seem to be a minimal requirement, some effect upon the expression or com- position of its product is expected, and a detectable phenotypic consequence of introducing the rearranged allele into another cell is desirable, although infre- quently available to date. A structural approach to the discovery of oncogenes is, of course, further circumscribed by the complexity of the eukaryotic genome. Even the limited number of suspected targets for oncogenic mutations (e.g. the c-onc group of proto-oncogenes) presents too large a challenge to propose to seek mutations detectable only by nucleotide sequencing procedures. [There is a notable exception: base substitutions that alter restriction endonuclease recognition sites in certain ras genes occur with sufficient frequency that it is feasible to survey tumor DNAs for these mutations (c.f. 195).] Efforts to find mutant oncogenes, particularly novel ones, by a direct inspection of the genotype have been restricted to lesions that alter the mutant gene or its environs sufficiently to affect the mobility or abundance of its restriction fragments. Three kinds of genetic rearrangement have been found to affect c-onc proto-oncogenes, and each kind offers useful if laborious means to identify new oncogenes. PROVIRAL INSERTIONS CAN ACTIVATE EXPiiESSION OF CELLULAR PROTO- ONCOGENES Most retroviruses lack v-onc genes and fail to transform cul- tured cells but are nevertheless competent to induce tumors, generally after a protracted latent period (233). There is now broad support for the idea that many tumors that arise following infection by v-onc- retroviruses contain mutant cellular oncogenes activated by proviral insertions (243). It is generally held that such insertion mutations are primary events in tumorigenesis and that their effect is to stimulate expression of the target gene by the provision of a strong viral promoter or enhancer element within the viral long terminal repeat (LTR) (48a, 74, 94, 145a, 155, 166a, 167, 170a, 184). Insertion mutations caused by the proviruses of horizontally transmitted retroviruses or of endogenous retrovirus-like elements have been found to afflict several cellular genes, some previously known as c-onc's (c-myc, c-erb B, c-myb, c-mos, and c-Ha-ras) (Table 4). In most of these situations, the target gene for insertion mutation can be identified by examining restriction digests of tumor DNA for rearrangements using molecular probes for c-onc genes (74, 94, 147, 151a, 184, 217). Other candidate oncogenes (e.g. int-1, int-2, pim) have come to view as repeated targets for insertion mutation only by 563 VARMUS applying methods akin to what has been called transposon tagging by students of mobile elements in Drosophila (14) (Table 4). To exploit this approach optimally requires that a tumor contain only a single copy of newly acquired proviral DNA. On the assumption that the single copy is acting as an insertional mutagen for an adjacent cellular gene, the viral DNA is molecularly cloned with some of the flanking DNA. Part of the retrieved cellular DNA is then used as a molecular probe to ask whether the same region of the host genome is occupied by insertions in other tumors (156). It has also been possible to isolate cellular sequences that appear to be targets for oncogenic insertion mutations from tumors bearing two or more proviruses (48a, 170, 236). The conclusion that commonly interrupted cellular domains are sites of functionally important insertion mutations rests, in part, upon the assumption that retroviral proviruses integrate into the host genome without significant regional specificity. Although the available evidence argues that many sites in host chromosomes can accommodate proviral DNA (244a), it is premature to claim that integration occurs randomly. Other kinds of evidence are therefore used to sustain the view that the identification of candidate oncogenes by transposon tagging reflects the selective advantage conferred upon rare cells in which the insertion mutations occur, rather than integration at highly preferred sites. For example, the vast majority of the multiple MMTV proviruses in tumors with insertions in znt-1 or int-2 are located outside these domains, and each tumor appears to carry an insertion in only one int-1 or one int-2 locus (156, 170). Evidence for an adjacent coding domain that is expressed at enhanced levels in the tumors and is evolutionarily conserved also fosters belief that the common insertion site harbors an oncogene (155). In some instances-x.g. B cell leukemias induced by bovine leukemia virus, (86a, IOS), T cell leukemias associated with T cell leukemia virus (204a), and nephroblastomas induced by myeloblastosis-associated virus (MAV) (157)- persistent efforts have failed to identify genetic targets for insertion mutation by probing for c-onc rearrangements or by transposon tagging, even though the tumors appear clonally derived. Such negative findings could imply that insertion mutations are not involved in these tumors, that several genetic targets are used, or that the insertions are remarkably far from the activated gene. Under certain circumstances, the existence of abundant transcripts from the mutated gene can assist in the identification and molecular cloning of the gene, e.g. when the transcripts contain viral sequences because the LTR serves as promoter for adjacent cellular sequences. For example, cloning of cDNA containing LTR sequences from a MAV-induced chicken nephroblastoma permitted the identification of the target gene for proviral insertion mutation as c-Ha-ras, despite the failure to perceive rearrangements by restriction mapping (D. Westaway, unpublished data). Although heightened expression of an oncogene is believed to be the primary Table 4 Candidate oncogenes affected by insertion mutations Gene Insertion mutagen" c-myc ALV, CSV, RPV csv RPV MAV Mo-MLV, MCF-MLV FeLV Documented effect upon Tumor type Species expression References B cell lymphoma B cell lymphoma B cell lymphoma B cell lymphoma T cell lymphoma T cell lymphomas chicken chicken chicken quail rat, mouse cat 71, 73, 94, 167 151a see footnote c below see footnote d below 43a, 217; see footnote e 147 below c-erb B ALV erythroblastosis chicken -t 74 c-Ha-ras MAV . nephroblastoma chicken + see footnote d below c-mos IAP plasmacytomas mouse + 36, 78, 184 c-myb Mo-MLV plasmacytoid mouse lymphosarcomas 144; see footnote f below int-1 MMTV mammary carcinoma mouse + 156 inr-2 MMTV mammary carcinoma mouse + 170 Mlvi- 1 Mo-MLV T cell lymphoma rat 236 Mlvi-2 Mo-MLV T cell lymphoma rat 235 Mlvi-3 Mo-MLV RMO-int-1 Mo-MLV T cell lymphoma rat T cell lymphoma rat see footnote g below 120 pim-1 MCF-MLV, MO-MLV T cell lymphoma mouse + 48a c-raf MLV-LTR~ transformed fibroblast mouse 14c `Abbreviations: ALV, avian leukosis virus; CSV, chicken syncytial virus; RPV, ring-necked pheasant virus; MAV, myeloblastosis-associated virus; Mo-MLV, Moloney murine bCloned DNA containing an MLV LTR was used as an insertion mutagen via transfection. leukemia virus; FeLV, feline leukemia virus; IAP, intracisternal A particle; MMTV, mouse mammary tumor virus; MCF-MLV, mink cell focus-forming murine leukemia virus. C. Simon, W. Hayward, personal communication. D. Westaway, unpublished data. eG. Selton, A. Berns, Y. Li. N. Hopkins, P. O'Donnell, E. Fleissner, personal communication. `G. Shen-Ong et al, personal communication. g P. Tsichlis, personal communication. 566 VARMUS consequence of tumor-associated proviral insertion mutants, in some situations listed in Table 4, expression of the disrupted cellular domain has yet to be assessed. Furthermore, even in those instances (e.g. ALV insertion mutations of c-myc or MMTV mutations of int-1) in which tumor cells contain relatively high concentrations of RNA transcribed from the target gene, it is difficult to know whether the appropriate normal cell has been chosen for comparison. The insertions may also affect the efficiency of translation (e.g. by forming a shortened hybrid leader region in mRNA) or truncate the gene product (e.g. in ALV insertions within c-erbB). Lastly, it is now apparent that potentially significant alterations other than the insertion mutation may occur within affected loci; these will be described in greater detail below. ONCOGENES AT TRANSLOCATION BREAKPOINTS It has been recognized for many years that chromosomal translocations a e common in tumor cells and that certain combinations of chromosomes occur repeatedly in certain types of tumors (189-191, 261). Proto-oncogenes are sufficiently numerous and dis- persed to inhabit virtually all chromosomes, so it is not credible to implicate a proto-oncogene in neoplasia merely because it is located somewhere within a partner in a frequent translocation event. Emphasis has instead been placed upon more stringent criteria: that a gene involved in the neoplastic process must be sufficiently close to the recombination site to have undergone detectable structural or functional alterations or must be within reach of the breakpoint as gauged, say, by restriction mapping with probes for adjacent sequences. Given the enormity of eukaryotic chromosomes, it is daunting to consider the routine identification of novel oncogenes by direct isolation of breakpoint regions. Sensibly, greater attention has been directed instead to the issue of whether proto-oncogenes known to be residents of translocated chromosomes might be positioned close to the breakpoints and whether their structure or expression is deranged as a consequence. To date, two known proto-oncogenes, c-myc and c-abl, have met the criteria that implicate them as oncogenic products of translocations (Table 5). A large number of B cell tumor cell lines of both human and murine origin harbor translocations that join the c-my gene on one chromosome to one of the three immunoglobulin loci on another (107, 169, 191) (Table 5), in the manner originally predicted by Klein (106). The Phi- ladelphia (Ph') chromosome in chronic myelogenous leukemia (CML) links c-ab1 from one end of chromosome 9 to a restricted region of chromosome 22 (54, 87, 95). Novel putative oncogenes have been tentatively identified at chromosomal breakpoints in B cell tumors because they, like c-myc, are joined to immunoglobulin genes from the partner chromosome (Table 5). The translocations of c-myc have been most closely studied and appear to CELLULAR ONCOGENES 567 perturb expression of the oncogene by either transcriptional or translational mechanisms, with ancillary mutations within c-myc perhaps further augment- ing the oncogenicity of the translocated allele; these issues will be addressed in a later section. ONCOGENES IN AMPLIFIED DNA The initial sightings of ampli~ed proto- oncogenes in human tumor cell lines grew out of attempts to explain unex- pectedly high levels of expression of c-onc genes or to test the possibility that karyological markers of gene amplification in tumor cell lines-homogeneous- ly staining regions (HSRs) or double minute chromosomes (DMs) (45Gmight harbor c-onc's (3, 37, 51, 199). Although a search for amplified oncogenes is inherently biased toward those proto-oncogenes already in hand, at least one novel oncogene (known as N-myc) has been discovered as an amplified gene in many neuroblastoma cell lines and tumors by virtue of limited homology with the second exon of c-myc (111, 198). Since methods for enriching genomic DNA for its amplified components are at hand4.g. rate zonal sedimentation of DMs (80), gradient fractionation to prepare chromosomes enriched for markers bearing HSRs (104), isolation of middle repetitive DNA based upon reannealing kinetics (138), differential screening of cDNA libraries (1 1 I), and reannealing of restriction fragments within agarose gels (188)--candidate oncogenes may be advanced on the basis of their frequent amplification in certain types of tumors, even without any overt relationship to known proto- oncogenes. Other clues to amplified oncogenes have recently been exploited. The human tumor line A43 1, long known to express high levels of epidermal growth factor receptor (67a, 88a), was successfully tested for amplified num- bers of c-erbB genes (124a, 136,237a) when c-erbB was proposed to encode the receptor (60). In all cases in which expression of amplified oncogenes has been examined, augmented levels of RNA have been encountered, although not always in direct proportion to the extent of amplification (201; F. Alt, personal communica- tion). However, the effects of oncogene amplification may not represent a simple gene dosage phenomenon: preceding or consequent mutations within amplified genes, translocation to new chromosomal contexts, or co- amplification of adjacent genes may have additional influences (see below). Amplifications of cellular oncogenes are proposed to influence tumor pro- gression to advanced stages of malignancy. Two recent findings support this view: amplification of N-myc has been observed only in those neuroblastomas clinically assigned to Stages III or IV (22), and amplification of c-myc is regularly detected in those variants of small cell lung carcinoma cell lines with highly malignant growth potential (127). Table 5 Translocated cellular oncogenes .- Known or Locus on candidate Partner partner Tumor oncogenes Chromosome chromosome chromosome References Mouse plasmacytoma cell lines c-myc 15(D2/3) 12(F1) IgH I, 44, 46, 131, Rat immunocytoma cell lines c-my 7 6 ? 224 205 wl cn W Human Burkitt lymphoma cell c-myc S(q24) Wq32) IgH lines 2(p 13) IEK 22(q11) w Human chronic myelogenous c-ab1 9(q34-+pter) 22(ql1) bcr Human B cell leukemia and "bcl- 1 "(?)" 1 m13) 14(q32) leukemia cell lines lymphoma Murine plasmacytoma "TKNS-l"(?)" 10 6(C2) CK I, 47, 49, 50, 52, 98, 231 87, 95 237a 168 Murine plasmacytomas "p\Jt"( ?)" 15 6(C2) CK see footnote b below "The oncogenic potential of these sequences is speculative. bJ. Adam, personal communication. Table 6 Amplification of cellular oncogenes Oncogene Tumor Degree of Source amplification ~- Reference ~ ~__ ~~~~ c-myc Promyelocytic leukemia cell line, HL60 and Human 20 x a 51, 37, I53 c-myc APUDoma cell line, COL0320 Human 40 x .. +I+ 3 primary tumoIg c-myc Small cell lung carcinoma cell lines (variants) Human 5-30X ? 127 c-myc CSV-induced bursal Iymphoma Chicken 5-1OX ? 151a N-myc Primary neuroblastomas (Stage 111 and IV) Human 5- 1000 X +/+ 22, Ill, 198, 201 N-myc Retinobiastoma cell line, Y79, and primary Human 10-200 x +I+ 111, 119a, 200 N-~Yc Small ceil lung carcinoma cell lines and Human 50 X t- i? See footnote c below c-ab1 Chronic myeloid leukemia cell line, K562 Human 5-1ox ? 38, 95 c-myb Acute myeloid leukemia Human 5-IOX ? 167a c-Ki-ra~2~ Primary carcinomas of lung, colon, bladder Human 4-20 x ? See foo!:mte d below % 0 z Colon carcinoma cell line, SW480 ? 135 n 0 and neuroblastoma cell lines tumors tumor I n c-myb Colon carcinoma cell lines, COL0201/205 Human lox ? 4 rn c-erbB Epidermoid carcinoma cell line, A431 Human 30 X ? 124a, 136, 237a $ @ and rectum c-Ki-ras Adrenocortical carcinoma cell line, Y1 Mouse 30-60x +!f 199 N-ras Mammary carcinoma line, MCF7 Human 5-1OX , ? see footnote e below t: PDMs were observed in early passages of HL60 cells and an abnormal banding pattern in 8q was later observed by Wowkll et ai (153); both appear to coincide with C-J~~C ampli~cation. some of these tumors, rearrangements of c-Ki-ms2, with and without amplification, were also observed (M. Barbacid, personal communication) 1. Minna, personal communication. dM. Barbacid, personal communication. e M. Wigler, personal communication. 570 VARMUS MOLECULAR PROFILES OF ACTIVATED ONCOGENES Strategies for the identification of mutant cellular oncogenes are currently limited by their ability to detect only certain structural lesions (mainly rear- rangements of various kinds) or lesions with certain functional consequences (transformation of cultured fibroblasts). Nevertheless, as summarized in the following sections, the observed genetic alterations seem to represent a diverse sampling of the possibilities for somatic mutation in vertebrate cells. Base Substitutions in Mutant ras Alleles Nucleotide substitution and concomitant amino acid change in members of the ras gene family is the genetic alteration that most commonly explains the transforming activity of tumor cell DNA (Tables 3, 7). The findings have particular force because it has been possible to use in vitro recombinants of wild type and mutant ras genes to validate the functional significance of the identi- fied mutations (e.g. 28, 227, 260) and because changes affecting the same codons, 12 and 61, have been encountered in multiple mutant alleles of c-Ha-rasl, c-Ki-ras2, and N-ras (Table 7). The reiterative pattern of these lesions raises two immediate questions: (a) what is the effect of amino acid changes at the observed positions on the behavior of ras gene products? (b) does the finding of repeated lesions at the same sites imply a hot spot for mutation or a selection for random mutations that confer transforming activity upon a proto-oncogene? The prevailing hypothesis holds that mutations may occur at various sites in ras genes (and presumably throughout the genome), but that strong selective pressures favor the outgrowth of cells that undergo changes in a restricted number of sites within coding regions. It has been conjectured, for example, that replacement of the glycine residue normally encoded by the twelfth codon of c-Ha-rasl by amino acids with side chains will significantly alter the conformation of p2 I~-~~-~~~~ or the nucleotide binding site predicted to reside in this domain (17 1, 196,258). Support for the functional importance of positions 12, 59, and 61 in p21c~Ha~"" and for the likelihood that random mutations at these sites are selectable has come from three genres of in vitro mutagenesis experiments. First, A. Levinson & W. Colby (personal communication) have used synthetic oligonucleotide primers to generate mutants of c-Ha-rasl encod- ing 18 of the possible amino acids other than glycine at the twelfth position. All of these, save pro-12, are capable of transforming NIH/3T3 cells, although the phenotypes vary from highly transformed to relatively subtle ones. Second, Fasano et a1 (67b) have used sodium bisulfite to produce C to T transitions at many positions in c-Ha-rasl and then have tested cloned mutants for their ability to transform NIH/3T3 cells. Oncogenic mutations occurred only at a restricted number of sites, including codons 12, 13, 59, 61, and 63. Three of CELLULAR ONCOGENES 57 1 these positions conform precisely to the positions of mutations encountered in spontaneous tumors and in viral rus oncogenes (Table 7), and all are close to such sites. The change observed at position 59, ala -+. thr, is particularly interesting because it is one of the three differences in amino acid sequence between c-Ha-rusl and v-Ha-rus, implying that the viral oncogene has ac- quired two alterations independently sufficient to induce cellular transforma- tion. This prediction has been confirmed by showing that a virus bearing a v-Ha-rus gene with gly-12 and thr-59 is as oncogenic as wild-type Ha-MSV (P. Tambourin, D. Lowy, personal communication). Third, K. C. Vousden andC. Marshall (personal communication) have mutated c-Ha-rus1 in vitro with a diol epoxide and found that five of twenty-one resultant transforming genes have restriction site changes in the eleventh or twelfth codon. The identification of single nucleotide substitutions, both transitions and transversions, that account for the activation of rus proto-oncogenes in spon- taneous human tumors (Table 7) is consonant with the widely disseminated notion that environmental mutagens serve as carcinogens (5). Further support for this idea has emerged from the isolation and structural analysis of mutant ras alleles from tumors induced experimentally in rodents with a variety of chemical carcinogens (Tables 3 and 7). For example, with strlking reproduci- bility, both dimethylbenzathracene (DMBA) and nitrosomethylurea (NMU) induce rat mammary tumors bearing activated c-Ha-rasl alleles, the lesions in all examined instances being G to A transitions in the twelfth exon (22; M. Barbacid, personal communication). In the classical two-stage model for induction of skin carcinomas in the mouse with an initiator (e.g. DMBA) and tumor promoter, Balmain & Pragnell(8) have shown that the c-Ha-rus gene is activated, even at a relatively early stage (benign papillomas) (9). Similarly, mouse cells transformed in culture with methylcholanthrene (MCA) and MCA- induced murine fibrosarcomas carry activated c-Ki-ras genes (65, l65,206a, 246a). Some central issues and assumptions about ras point mutations remain uncertain, if not controversial: 1. Are ras mutations dominant? Although most evidence favors the idea that these mutations act in a simple dominant fashion, it is not uncommon to encounter results that undermine confidence in that assumption. Some tumor lines appear to carry only the mutant allele and hence are homo- or hemizygous for it, as judged by analysis of restriction sites that overlap the altered codon (195); in one instance, the human cell line SW480, the same twelfth codon mutation in c-Ki-rus2 is present in two alleles distinguished by an apparent polymorphism in a non-coding region (29). Furthermore, mutant alleles are sometimes amplified [e.g. a four- to eight-fold increase in c-Ki-ras2 in SW480 cells (135)], duplicated by tetraploidy [e.g. in the EJ bladder carcinoma line (U. Franke, personal communication)], or expressed more frequently than the Table 7 Mutations in transforming ras genes ul -4 h) Codons rus allelc 12 I 59 61 References < P Source of allele Normal Normal Ha-MSV RaSV Bladder Ca lines Lung Ca line Mammary Ca Normal Ki-MSV Lung Ca line Colon Ca line Human c-Ha-rusl Rat c-Ha-rasl v-Ha-ras v-Ra-ras" Human EJIT24-Ha-rust Human HS242-Ha-rasl Rat NMU-Ha-raJ I Human c-Ki-rus2 v-Ki-rus Human Calu-Ki-ras2 Human SW480-Ki-ms2 GGC GCC CAG GGA g'Y GGT gIY ala ? ACA rzq GCA ala GCC ala GCC ala ? GCA - - - ala GCA ala GCA ala 28 222 55a 181 227. 196, 186. 28. 230 260 222 13% 237 209, 29 29 Lung Ca Lung Ca line Bladder Ca line Lung Ca line Lung Ca line Normal Neuroblastoma line Teratocarcinoma line, late passage Fibrosarcoma line Malignant melanoma line Lung carcinoma line Human LL- I 0-KI-ras2 Human A2182-Ki-rus2 Human A1698-Ki-rus2 Human PR371-Ki-ras2 Human PR3IO-Ki-ras2 Human N-rus Human SK-N-rus Human PA I -N-rus Human HT1080-N-rus Human Me1 N-ras Human SW1271 N-rus ? ? ? ? GCA ala GCT ala GCT ala ? - ? - CCT ala - GCT ala ? ? ? ? his CAA gln ? I95 I95 I95 145 see footnote b below 229 229 8 5 r r see footnote c below P 228a ;d 0 0 see footnote d below 259a % 'Amino acid positions .aligned with those of other ras proteins bM. Perucho, personal communication. dR. A. Padua. personal communication. A. Hall, C. Marshall, personal communication. ul 4 W 574 YARMUS normal allele [e.g. in the CaLu tumor line (29)l. And when introduced into the NIH13T3 cell containing normal ras alleles, it is usual to observe several copies of the mutant gene. On the other hand, it has been possible to demonstrate heterozygosity at the mutant locus in some instances [e.g. a primary human lung carcinoma (195)l. On balance, it seems likely that the ras mutations are dominant but dose-dependent, with consequent selection for cells in which the mutant allele is augmented by chromosomal duplication, gene amplification, gene conversion, or preferential expression. It is still uncertain how the find- ings with ras alleles are to be reconciled with evidence that at least some aspects of the oncogenic genotype behave as recessive properties in somatic cell hybrids (e.g. 214). 2. When are ras mutations acquired? Although it is generally held that ras mutations arise somatically during oncogenesis, only in rare instances has this hypothesis been directly validated: e.g. in chemically induced tumors in inbred mouse strains (222) and in a patient whose lung carcinoma contained a muta- tion that-created a Sac1 site not present in multiple samples of normal tissue (195). Greater debate, however, surrounds the question of whether ras muta- tions are early or late events in tumor cell development, a question that can be extended to consider the possibility that the mutations are merely products of the exaggerated mutation rates observed in tumor cells 02). There are now numerous instances of mutant ras alleles in primary tumors of man and animals (Tables 3 and 7), countering earlier proposals that such mutations occurred only in cultured cells. Moreover, mutations in c-Ha-ras appear to be early events in chemically induced skin tumors, as is evident by transformation assays when the tumors are still benign by traditional criteria (9). Based on studies of human melanoma (2) and teratocarcinoma (228) lines, however, it has been argued that ras mutations may be late events in some settings; most tellingly, only one of five lines derived from a single melanoma patient exhibited a mutant gene, implying either that the mutation arose in culture or that the tumor was heterogeneous with respect to the lesion. At present, it appears that ras mutations are somatic events that can occur at various stages of the neoplastic process, but better functional tests will be required to establish the contribution the lesions make to the cell types in which they are encoun- tered. Insertionally Mutated c-myc Loci LOCATIONS OF PROVIRUSES A large number of insertion mutations caused by ALV, CSV, and MLV proviral DNA introduced into c-myc loci have now been analyzed structurally in DNA from avian bursal and rodent T cell lympho- mas. The c-myc genes of chickens, mice, and'man appear to consist of three exons (239), the first of which is non-coding and presumed to have a regulatory function (e.g. 119). Mapping with restriction endonucleases indicates that most CELLULAR ONCOGENES 575 of the insertions in bursal lymphomas reside within the first intron of c-myc between the non-coding 5' exon and the exon that contains the apparent initiation site for translation of c-my; furthermore, most of the proviral mutagens are oriented to allow their LTRs to serve as promoters for c-myc (7 1, 94, 167; R. Swift, H. Robinson, personal communication). Since there is little reason to believe that retroviral integration occurs preferentially at certain sites in host genomes or with preferred orientation, it seems likely that the common- ly encountered configurations of viral DNA in c-myc loci imply a high prob- ability of efficient expression of c-myc. Proviruses within the first intron appear to be clustered in a fashion that conforms to the idea that insertions at some sites are selectively advantageous over insertions at other sites (207; R. Swift, H. J. Kung, personal communication). Nucleotide seqyencing of a few of the host- proviral junctions (146,256; R. Swift, personal communication) shows that the insertions are at different positions within the intron without apparent homolo- gies between insertion sites or between host and viral sequences at the recom- bination sites, findings consistent with studies of host-viral junctions from proviruses not known to affect the behavior of adjacent cellular genes (244a). The clustering of proviruses within the c-rnyc intron is particularly likely to be affected by the presence of nearby sequences that can serve as splice donor sites to join with the splice acceptor site at the 5' end of the second exon of c-myc. Insertions have been found within the splice donor site from the first exon of c-myc (146; C. Nottenburg; unpublished data), just upstream of a cryptic splice donor site that was also used during the transduction of c-myc to generate the v-myc gene in MC-29 virus (207,252,256), and on the 5 ' side of another intron sequence that conforms to the splice donor consensus (146). Other arrangements have been encountered within insertionally activated c-myc loci in bursal lymphomas: proviral DNA in the first intron (or further upstream) in the opposite transcriptional orientation [( 167); M. Linial, cited in (90)], on the 3 ' side of the coding domain in the same transcriptional onenta- tion (167); and on the 5' side of the first exon in the same orientation (162; H. Robinson, personal communication). In the former two situations the LTRs cannot serve as promoters for c-myc and activation of either a novel c-myc promoter or the normal promoter must occur, presumably under the inflnence of the enhancer element described in restriction fragments containing avian virus LTRs (129). The initiation sites for transcription under these circum- stances have not yet been mapped. In MLV-induced T cell lymphomas of mice and rats, c-myc insertion mutations that forbid the use of the MLV LTR as a promoter for c-myc seem to be the norm. Almost all of these insertions are located within or on the 5 side of the first c-myc exon and directed away from the gene (43a, 217; Y. Li, N. Hopkins, G. Selton, A. Berns, personal communication), although a few are positioned downstream of c-myc, in one case at least 24 kb beyond the third 576 VARMUS exon (43a). Proviral insertions with a transcriptional orientation directed away from the activated oncogenes are also commonly observed in MLV mutations of pim (48a) and in MMTV mutations of inf-1 and int-2 (see below). ADDITIONAL MUTATIONS IN PROVIRALLY MUTATED LOCI In the several instances in which host-viral junctions have been examined at the nucleotide level in c-myc insertion mutations, it appears that the accepted rules of retrovi- ral integration have been followed: two base pairs have been removed from the ends of LTRs during insertion, and six bp at the host insertion site have been duplicated (146, 256). However, as first recognized from restriction mapping of the proviral mutagens (71, 73, 145a, 166a, 167), deletions and possibly other secondary rearrangements have altered normal proviral structure. Most commonly, these deletions remove regions of viral (and perhaps cellular DNA) encompassing the 5' LTR in proviruses located on the 5' side of the c-rnyc coding sequence in the same transcriptional orientation. There is experimental support for the claim that one effect of such deletion mutations is more efficient transcription of c-myc from the 3' LTR (48). However, other sorts of deletions have also been observed: proviruses on the 5' side of c-myc in the same transcriptional orientation have sustained a loss of their 3' ends, so that transcription of c-myc proceeds from the 5 LTR through viral coding domains (162) in the manner predicted to precede transduction of proto-oncogenes (see below), and proviruses in the opposite transcriptional orientation have incurred deletions that prevent expression of viral genes, e.g. by removing viral signals required in cis for synthesis, processing, or translation of RNA (166a, 167). Two proviral deletion mutants have been examined by nucleotide sequenc- ing. One retains only a single copy of the ALV LTR that lacks two bp from both ends and is flanked by a six-base duplication of cell DNA, indicating that it arose by homologous recombination between the LTRs of an initially intact provirus (256). Since the provirus in this case was inserted in the same transcriptional orientation as c-myc, the residual LTR is positioned to serve as promoter for a hybrid transcript that includes the coding exons of c-myc. The other sequenced deletion mutation occurred in a provirus (from tumor LL3) positioned in the opposite transcriptional orientation to c-myc (256). Although the LTRs were spared, the approximately 1 kb deletion removed the tRNA primer binding site, the RNA packaging site, the splice donor site, and the translational initiation codon for all the viral proteins. Both deletion mutations must have occurred during or after infection, since the proviruses lack se- quences required in cis for transmission through the virus life cycle. Nucleotide sequence alterations in c-myc, as well as deletion mutations in proviruses, are present in some insertionally mutated loci. Comparison of a partial sequence of the mutant c-rnyc locus from tumor LL3 with the homolo- gous region of normal chicken c-myc (252) revealed three additional G residues CELLULAR ONCOGENES 577 in the short portion of intron between the insertion site and the second exon (D. Westaway, unpublished data) and ?hree nucleotide substitutions in the first 200 nucleotides of the second exon (256). One of these, a silent C -+ A transver- sion, was later judged to be an inherited polymorphism because the same change was observed in two other c-myc alleles from the same flock (D. Westaway, C. Nottenburg, unpublished data). However, a second change that does not alter the coding sequence, a G -+ A transition, must represent a somatic mutation because it inactivates an SstI recognition site shown to exist in both c-myc alleles in normal tissue from the bird carrying tumor LL3. A third nearby change, a C 3 A transversion that converts a proline codon to a threonine codon, appears to be a somatic mutation as well, since the change is not found in alleles from the same flock that harbor the polymorphic marker. A further genetic alteration of c-myc has been observed in bursal lymphomas induced by CSV (15 la). Some insertionally mutated alleles are five- to ten-fold more abundant than single copy genes and the unaffected alleles are not detected. Apparent loss of normal c-myc alleles, presumably a result of loss of one member of a chromosomal pair, also seems to occur in some ALV-induced bursal lymphomas, perhaps accompanied by low level amplification of mutant c-myc, since only the rearranged allele can be identified by the Southern transfer procedure (71, 73, 94, 167). The discovery of additional mutationsdeletions, base substitutions, and amplification-at the site of proviral insertion mutations raises further ques- tions. When do the additional mutations occur? Do they contribute to the cellular phenotype, providing a selective advantage? Since cells bearing multi- ple mutations within c-myc dominate the tumor mass, either the additional mutations must be generated at the time of (or before) the proviral insertions or they must produce a selectable change. Proviruses with deletions of the type encountered in c-myc loci have not been observed among the many chosen randomly for structural scrutiny (244a); however, under selective conditions (see below), various deletion mutants, including those resulting from homolo- gous recombination between LTRs, can be recovered at low frequency (245). Thus, barring some marked stimulus to deletion formation in lymphoma cells, a selective mechanism is likely to operate. Thehming of nucleotide changes in c-myc seems less certain, since tl-ere is insufficient information about the effect of proviral integration upon hc ' sequences surrounding those proviruses chosen randomly for study. There i - rears to be concerted mutation, however, since some of the changes do not affect the coding potential of c-myc and seem unlikely to affect the efficiency of expression. The selectable properties that might be conferred upon bursal cells by the additional mutations have not yet been rigorously defined. Deletion mutations within proviral DNA could mitigate the host's immune response against tumor cells by preventing synthesis of virus-specific proteins that serve as antigens on 578 VARMUS the cell surface. Proviral deletions might also have the more intriguing effect of augmenting expression of c-myc (48). Amplification of insertionally mutated alleles could provide selective advantage on quantitative grounds. Lastly, nucleotide changes that alter the amino acid sequence of c-myc protein could render the protein more effective as an oncogenic factor. What can be done to discern the phenotypic effects (if any) of these multiple mutations within c-myc alleles? Removal of the mutant alleles from the com- plex environment of the tumor cells to a new context would seem to be necessary to measure their effects, but to date the only convenient, direct assay for activated myc alleles is the co-transformation assay in rat embryo cells (1 14). In this assay, the crucial determinant of activity appears to be an efficient transcriptional promoter upstream from c-myc coding sequence in the correct orientation. Thus, normal c-rnyc exons transcribed from LTRs linked to c-myc in vitro are competent in the co-transformation assay, whereas c-myc alleles from tumors with proviral insertions in atypical arrangements are not, even when the e-rnyc gene has undergone nucleotide substitutions (W. Lee et al, unpublished data). More quantitative means to assess transformation efficien- cies or the use of more appropriate cell types (e.g. B lymphocytes) may be required to elicit the consequences of the mutational events recorded in the c-myc alleles described above. Characteristics of Other Proviral Insertion Mutations in Proto-Oncogenes The several other cellular loci found afflicted by insertion mutations (Table 4) have been subjected to less scrutiny thus far than the c-myc alleles from bursal lymphomas. Nevertheless, a number of interesting contrasts with the activated c-myc alleles have come to light. 1. The insertion mutations in c-erbB described in ALV-induced erythro- blastosis (74) seem likely to result from proviruses that are flanked on both sides by parts of c-erbB. This inference is based on the idea that c-erbB encodes the EGF receptor, a protein of approximately 138-175,000 Mr, whereas the abundant c-erbB protein in erythroleukemia cells is similar in size (approx- imately 62-75,000 Mr) and peptide composition to the pfoduct of v-erbB (H. Beug, H. J. Kung, personal communication), apparently representing the carboxy terminal half of the EGF receptor [(60) and references therein]. This finding suggests that the oncogenic activation of the c-erbB gene may depend on truncation of the protein product in both viral transductants and insertionally mutated loci. A limited analysis of an LTR-c-erbB junction from one erythro- blastic tumor shows that the LTR is in the same transcriptional orientation as c-erbB, just upstream from an exon homologous to v-erbB, and that two bp are missing from the LTR as anticipated in a normally integrated provirus (74). CELLULAR ONCOGENES 579 2. MMTV proviruses are about equally distributed on the 3' and 5' sides of the int-1 locus, but all save two are directed transcriptionally away from the gene (155; R. Nusse, personal communication). A similar situation obtains in MMTV-activated int-2 loci (56a, 170). Hence, the observed transcriptional activation of these genes normally depends on the effect of a viral enhancer rather than of a viral promoter. Two other features of int-1 and int-2 distinguish them from some of the other targets of oncogenic insertion mutations: first, the genes are not expressed at detectable levels in any of the normal tissues examined to date, including mammary glands from pregnant and lactating females (56a, 156; Y. K. T. Fung, unpublished data), implying that the genes may be converted from transcriptionally inactive to active genes rather than merely stimulated to higher levels of expression; second, these two genes have not been encountered among retroviral oncogenes (i.e. they are not c-onc's). The int-1 gene is composed of four exons that contain the entire coding domain, with a relatively long (approximately 800 bp) non-coding region at the 3' end of the final exon (A. van Ooyen, R. Nusse, personal communication; G. Shackleford, T. K. T. Fung, unpublished data). Most of the MMTV insertions on the 5' side of int-1 appear to reside upstream from the transcriptional initiation site, with the exception of two insertions in the same transcriptional orientation as int-1 (R. Nusse, personal communication). MMTV insertions on the 3' side of int-1 are sometinics positioned within the final exon, between the end of the coding sequence and the polyadenylation site (155); as a result, int-1 transcripts conclude with the U3 sequence copied from the MMTV LTR. Since the U3 region is approximately 1200 bp in length (57), insertions of this type produce a transcript longer than the usual 2.6 kb mRNA. The absence of the 2.6 kb int-1 RNA species in tumors containing the longer, allele-specific RNA confirms that activation of int-1 functions in cis but not in trans. Few lesions other than the MMTV insertions in mutant int4 and int-2 alleles have been implicated in the activation process. Restriction mapping has not revealed overtly mutated proviruses in most cases, although a mutant int-2 allele has been found to contain a solitary LTR, the presumptive product of homologous recombination between MMTV LTRs (G. Peters, C. Dickson, personal communication). The few host-viral junctions that have been se- quenced show only the expected absence of two bp from the end of the LTR (R. Nusse, personal communication). The mutant alleles do not appear to be amplified in number, but structural or functional tests for additional mutations within the int-1 locus have not been performed. 3. Insertions of intracisternal A particle (IAP) proviruses into the mouse c-mos gene have been described in cell lines derived from two BALB/c plasmacytomas, one tumor induced with mineral oil (MOPC 21) and one induced with pristane (XRPL24). In the latter case, the 4.7 kb IAP insert is 580 VARMUS positioned within the coding sequence of the only known exon of c-mas, 66 codons downstream from the start of homology with one of the v-mos alleles in the transcriptional orientation opposite to that of c-mos (27, 184). The conse- quences of the insertion include the transcriptional activation of an oncogenic sequence that is normally silent (139), producing a 1.2 kb RNA species presumably initiated from an occult promoter and generating a transforming gene competent to cause neoplastic change in NIH/3T3 fibroblasts (184). In cell lines derived from MOPC 21 (including a popular partner in the formation of hybridomas, P3 X63 Ag8), the IAP DNA is located 58 codons closer to the 5' end of c-mos in the same transcriptional direction (36). In this arrangement, c-mos RNA is at low or undetectable levels despite the presence of a 3' LTR positioned to serve as a promoter for downstream transcription (36, 78); it is not known whether the relatively poor expression of c-mos in derivatives of MOPC 21 is still adequate to affect the phenotype of the plasmacytoma cell. A striking peculiarity of the IAP insertion in XRPC24 cells warrants further comment. It is generally held that IAP proviruses are defective for production of infectious extracellular virus particles, but the intracisternal particles are thought to synthesize new IAP DNA molecules from IAP RNA and to direct its integration by retroviral mechanisms. Analysis of IAP-host junctions supports this view: two bp from the ends of LTRs appear to be missing at the junctions and six bp of host sequence are duplicated to flank the IAP DNA (27, 113). However, in the XRPC24 insertion mutation, the two IAP LTRs exhibit considerable sequence divergence, a finding inconsistant with the generation of the LTRs from sequences present only once in the RNA template for reverse transcription (244a). The marked differences between the LTRs could be explained by a high mutation rate following the insertion, by synthesis of the insertional mutagen from an IAP with heterodimeric RNA, by recombination between heterologous IAP DNAs prior to integration, or, most provocatively, by transposition of IAP DNA without passage through an RNA intermediate. More examples of newly integrated IAP DNAs or recovery of the progenitor of the XRPC24 IAP mutagen will need examining to judge among these possibili- ties. 4. The c-rnyb rearrangements in several plasmacytoid lymphosarcoma lines derived from Ab-MLV-infected animals occur in the absence of Ab-MLV proviruses and were originally proposed to result from insertion and deletion of proviral DNA (144). However, cloning of the affected loci has revealed the presence of helper MLV proviruses, with small internal deletion mutations, in an intron upstream from known c-myb exons and in the same transcriptional orientation as c-myb (G. Shen-Ong, D. Mushinski, M. Potter, personal com- munication). It seems likely that these MLV insertional mutagens, rather than Ab-MLV, were the primary determinants of tumorigenesis. CELLULAR ONCOGENES 58 1 Structural and Functional Properties of Translocation Breakpoints near Cellular Oncogenes GENERAL FEATURES OF c-myc; IMMUNOGLOBULIN GENE TRANSLOCA- TIONS Efforts in several laboratories to characterize the translocations in murine plasmacytoma (PC) cell lines and human Burkitt lymphoma (BL) lines that join immunoglobulin (Ig) genes to c-myc have yielded a rich but complex harvest [see reviews in (107, 119, 169, 191)]. Although each translocation event appears to have its interesting peculiarities best viewed by descriptions of individual cases (see below), common features have emerged from a census of the translocations examined to date. 1. Whenever t(8;14) is observed cytogenetically in a BL line, c-myc with distal portions of chromosome 8 is transferred next to the Ig heavy chain constant region locus on 14qf, and when t(12;15) is observed in a PC line, c-myc from chromosome 15 is joined to the heavy chain locus on chromosome 12. In both species, the heavy chain and c-myc genes are joined head-to-head in divergent transcriptional orientations. The less frequent unions of human chromosome 8 with a chromosome carrying a light chain gene [t(8;22) and t(2;8)] have only been examined in a few BL lines, but again the patterns seem constant; with t(2;8) or with t(8;22), the light chain constant regions, some- timrs preceded by at least part of variable regions, appear on 8qt on the 3' side of c-myc, with the joined loci in the same transcriptional orientation (47, 52, 62a, 63, 230b). The role of c-myc in the relatively rare t(6;15) in plasmacytoma lines (96) has yet to be established at the molecular level (J. Adarns, personal communication). 2. In BL lines with the less common translocations, there is concordance between the type of light chain encoded at the site to which c-myc is translo- cated and the type of light chain produced, even though the translocation chromosome contains a non-functional (allelically excluded) locus (12 1). There is, however, at least one exception to this rule, a BL line in which t(8;22) is accompanied by production of kappa light chains (129a). 3. In the PC lines, most breakpoints within c-myc map to a region of 1-2 kb that includes the first (non-coding) exon and part of the fist intron. Breakpoints within the heavy chain locus have generally fallen in a switch (S) region. Most commonly, the latter sites are within a domain containing an Sa region, usually S, recombined with S, (implying that heavy class switching might have occurred prior to the translocation) or within an Sa region not directly joined to 4. In the BL lines with t(8;14), the c-myc breakpoints appear to be scattered on either side of the first (non-coding) exon, far enough from c-myc in several instances to fail to produce altered c-myc restriction fragments. Most of the heavy chain gene breakpoints in BL lines are within the Sp region, although significant exceptions exist [e.g. the breakpoint in the Daudi line appears SF (44). 582 VARMUS to be within the VH3 region (64), and in the Raji line it is within an S, region joined to S, (176)l. 5. Although breakpoints within the Ig and c-myc loci are sometimes clus- tered within a few kbp, there is no evidence for homology between recombined sequences or for strict sequence specificity. The Ig breakpoints within S regions generally lie near consensus S sequences, implying that the enzymes involved in heavy chain class switches participate in translocations. Fortuitous S-like sequences may also influence the sites of recombination within c-myc in some cases @Ob), and the sequence GAGG has been noted to lie 10-20 bp from the breakpoint in c-myc in five PC lines (17Oc). 6. The reciprocal products of translocation events can be found by molecular (if not by karyological) techniques, implying that reciprocal exchange is the norm. 7. Reciprocal translocations are imprecise, with loss of a few to a few hundred nucleotide pairs from one or both chromosomal partners. Unexpected nucleotide pairs (1-5bp) are sometimes found at junctions in the recombina- tional products, and in one case thus far a duplication was generated at a breakpoint. These findings suggest a general model for translocation involving staggered single-stranded cuts at the breakpoints, with polymerization or nu- cleolytic activity prior to ligation (80b). 8. Alterations in the nucleotide sequence of c-myc are common accompani- ments of translocations in BL lines. Generally these are confined to the first exon and hence do not affect the amino acid sequence of the c-myc protein (175a, 230b, 232), but in some cases changes are also found in the second exon (176). Thus far, a limited search has not revealed significant nucleotide sequence changes in translocated c-myc genes in PC lines (L. Stanton, C. Croce, M. Cole, personal communication). 9. Some features of the rearranged chromosome carrying the coding region of c-myc suggest that c-myc should not be expressed: recombination inevitably occurs with the allelically excluded Ig locus, the c-myc gene is often deprived of its normal promoter, and the Ig components are not suitably positioned to provide a promoter for c-myc. Nevertheless, production of c-myc RNA occurs from the translocation chromosome, with little or no expression of the c-myc allele that remains on the normal chromosome, sometimes producing what appear to be overtly high levels of c-myc RNA (7, 63, 131, 150) and at other times more subtle differences [e.g. changes in differential usage of the two transcriptional initiators in the first exon (119, 232).] In a few instances (e.g. 93) such effects may be mediated by cell-specific enhancer elements identified within immunoglobulin loci (10, Sl), but normally the known enhancers are relegated to the translocation product lacking c-myc coding exons (e.g. 176). Even in the absence of known enhancers, the expression of translocated c-myc genes seems dependent upon the cellular environment, since it is repressed in CELLULAR ONCOGENES 583 hybrids between BL cells and mouse fibroblasts (149, 150). Effects of trans- location upon the translation efficiency of c-myc mRNA have been proposed to result from deletions or base changes affecting the first exon (175a, 232), particularly since a portion of the first exon appears to be inversely repeated in the second exon (194); formation of a stable duplex might normally impede translation of c-myc mRNA, and an inability to form it could augment the yield of c-myc protein. However, direct comparisons of amounts of c-myc protein in BL or PC lines have yet to be published. Many of the features of these translocations may reflect the specialized properties of genetic partners that normally undergo programmed rearrange- ments (the Ig genes) and of the type of cell (early or late B cells) in which the events occur. The location of chromosomal breakpoints within S regions, the loss of nucleotides at recombination sites, and the presence of somatic muta- tions in c-myc (moved to the position normally assumed by the hypermutable VH genes) are characteristics that mirror events in the ontogeny of Ig genes (234). It remains to be determined whether translocations involving many sites occur during the development of such tumors, with selection for those that activate c-myc, or whether there is a strong predisposition to the kinds of translocations observed. There is little reason to suppose that external initiating factors in tumoI igenesis affect the translocation process: there is no evident correlation with the agents used to induce PC tumors in mice (pristane or mineral oil) or with the presence or absence of Epstein-Barr virus in BL lines. Regrettably, translocations have yet to be described in PC or BL removed directly from the hosts. SPECIFIC EXAMPLES OF C-myC; IG TRANSLOCATIONS IN PC AND BL LINES Properties of the best studied translocation products illustrate some of the general features of c-myc-Ig translocations. 1. The PC cell line X63 Ag8 (sometimes known as P3 and commonly used to generate mouse hybridomas) was derived from the tumor MOPC 21 induced in a BALB/c mouse by peritoneal injections of mineral oil. c-myc is joined to the heavy chain locus at sites within the first exon of c-myc and within an S, region directly joined to Sy2b in a manner that forbids distinction between a transloca- tion that preceded and one that followed heavy chain switching. Comparison of the sequences of both recombinational products with the sequences of germ line c-myc and IgH genes reveals that sevenbp have been lost from chromosome 15 during the reciprocal translocation events; a single A:T nucleotide pair of unknown origin is present at one junction point (61,80b, 148). In addition, 491 bp have been deleted by homologous recombination between short direct repeats on the 3' side of the switch region, although it has not been unequivo- cally demonstrated that this deletion occurred in the tumor cell rather than during molecular cloning. Notably, X63 Ag8 and sibling lines contain three 584 VARMUS other candidate oncogenes described elsewhere in this essay: a c-nios gene activated by an IAP insertion mutation (36,78), a transcribed gene on chromo- some 10 joined to a kappa light chain locus by translocation (168), and a transforming gene unrelated to rus genes (1 17). 2. A second BALB/c-derived PC line, MPC 11, has undergone a transloca- tion involving the first exon of c-myc and the Sy2b region; in this case, nuc- leotides have been lost from both chromosomes: 11 bp from chromosome 15 and approximately 300 bp from chromosome 12 (216). Short inverted repeats are present near the breakpoints in c-myc, but these have not been directly implicated in the translocation mechanism. 3. The PC line 5558 has lost about 1 kb of c-myc on the 5 ' side of the site in the first exon that has been joined to an S, sequence in a S, S, domain (44), and the same or a second deletion has removed at least part of S, (80b). The chromosome containing the c-myc coding region from the 3' side of the breakpoint has an extra A:T base pair at the junction, as in X63 Ag8, where2,s the reciprocal product contains an extraneous four bp sequence (AACC) at the junction (80b). Sequencing of c-myc exons on this chromosome reveals only a single non-coding change from the germ line c-myc sequence (L. Stanton, personal communication). 4. The situation in the PC line W267 is similar to that in 5558. Twelve bp are missing from the translocation breakpoint in the first intron of c-myc and about 400 bp from the breakpoint in S,. In addition, three extraneous bp (CTT) are present at the fusion site in the reciprocal translocation product (80b). 5. In the PC line HOPC 11, 106 bp from the breakpoint region in the first intron of c-myc are duplicated and present in both products of the translocation event (80b). This finding provides the best support for the claim that the translocation mechanism involves staggered cutting at the breakpoints. In this cell line, 1.6 kb have been lost from the partner site in the S, region of chromosome 12, and an extraneous 5 bp sequence (CCTAT) is present at the joint on the reciprocal product of translocation (80b). 6. In the cell line BL22, a site about 1 kb on the 5 ' side of the first c-myc exon is joined to S, (12). Although the second and third c-myc exons do not deviate from the germ line sequence (12), the first exon displays small deletions and several nucleotide substitutions (232). Similar changes confined to the first exon-7% base changes and a 35 bp duplication-have been encountered in the translocated c-myc allele from the Daudi BL line, t(8;14), with a breakpoint at least 12 kb 5' of c-myc (175a). The amount of c-myc RNA is not appreciably greater in BL22 than in lymphoblastoid cell lines, but the first of the two transcriptional initiation sites is used 4-5 times as frequently as in lymphoblas- toid cell lines (232). 7. The BL line, Raji, resembles BL22 in that a site about 2 kb on the 5' side of c-myc is joined to S,, but the S, region is fused with S, and changes in the CELLULAR ONCOGENES 585 c-myc sequence are not confined to the first exon (176). In addition to several substitutions and a 10 bp deletion in the first exon, there are three short deletions in the first intron (including one that removes the splice donor site), 25 nucleotide substitutions dictating 16 amino acid changes and a three'bp insertion in the second exon, and 9 nucleotide substitutions in the second intron. The third exon does not differ from published germ line sequence. In the Raji line, unlike most BL lines, the normal (untranslocated) c-myc locus is also transcribed (175a), prompting the suggestion that the product of the translo- cated locus is defective in a trans-active repressor function (175a, 119). It is possible that the striking changes in Raji c-myc are consequent to concerted mutagenic activity during heavy gene class switching or di Iing translocation. Although normal tissue is not available from the patient in which the tumor arose, it is unlikely that these many differences are inherited polymorphisms. 8. In the BL line JBL2 with t(2;8), alterations in the structure and expression of c-myc are associated with translocation breakpoints at least 20 kb down- stream from c-myc and upstream of the kappa chain constant gene. A 2.5 kb region that includes the first c-myc exon has been duplicated on the recombined chromosome, and the duplicated exon contains several nucleotide substitutions and a small internal duplication. The concentration of c-myc RNA is elevated three-five-fold over levels in lymphoblastoid lines, with one RNA species retaining sequences from the first intron and another appearing to be spliced normally (230b). 9. Nucleotide substitutions in c-rnyc have also been observed in BLlines, LY 67 and MAKU, with t(8;22) (175a). The changes are most numerous in first exon and intron, but cDNA cloned from the LY 67 line reveals T-C transition in the second exon that dictates a change of ser-62 to pro-62 in c-myc protein (175a). 10. An 8;22 translocation in line BL37 has joined a site 5 kb upstream of Ck to a site 4 kb downstream of c-myc (98). During recombination, 21 bp were lost from chromosome 22. In this variant translocation, the joining sites bear no relationship to S regions of heavy chain loci, Despite the distance between the c-myc promoter and the Ig locus (> 10 kb), a marked effect on expression of c-myc was observed; c-myc RNA is fifteen-fold more abundant than in lym- phoblastoid lines and the first of the two initiation sites is preferred over the normally favored second site. 11. The BL line, Manca, differs from most BL lines with t(8;14) in two respects: the breakpoint in c-myc is within the first intron and, more exception- al, the breakpoint in the Ig heavy chain locus is on the 5' side of Sp,, placing a known enhancer element next to the coding exons of c-myc (93). Thus, transcription of c-myc may be affected by an enhancer element and translation may be augmented by the absence of sequences from the first exon in c-myc mRNA (194). 586 VARMUS 12. In another BL line with t(8;14), ST486, the Ig heavy chain enhancer is joined to the first exon of c-myc on the reciprocal translocation product, 8q- (46a, 79); short transcripts containing sequences from the first exon of c-myc are abundant, whereas the c-myc coding exons are expressed at near normal (79) or augmented (46a) levels, presumably from 14q+. Differences in regula- tory mechanisms are implied by findings that expression of the first exon on the reciprocal product is suppressed in cell hybrids formed with human lympho- blastoid cells, whereas expression of the coding c-myc exons is suppressed in hybrids formed with mouse PC cells (46a). Both translocation products have been isolated from this line: although a simple, non-homologous joining to S, was found 280 bp downstream from the first c-myc exon, restriction mapping of the reciprocal product suggested that additional undefined rearrangements had occurred (79). THE PH~ CHROMOSOME The piece of clwomosome 9 transferred to chromo- some 22 during the reciprocal translocation that generates the Ph' chromo- some, a common marker in chronic myelogenous leukemia (CML) in man, is karyologically invisible and hence estimated to be less than 5000 kb in size. Detection of DNA from chromosome 9 on Ph' was initially accomplished by using a probe"for the c-ab1 gene (54), and chromosomal walking later identified the breakpoint on chromosome 9 in one CML cell line to lie within 14 kb upstream of a region homologous to v-abl(95). The DNA specific for chromo- some 22 from the 9-22 chimeric clone was subsequently used to identify a region of about 5 kb, called bcr, that encompasses all of the chromosome 22 breakpoints in leukemic cells obtained from 17 of 19 CML patients (87). (The two negative cases did not have karyological evidence of the 9;22 transloca- tion.) The chromosome 22 breakpoints appear to be scattered throughout the bcr region, though nucleotide sequencing of the junctions has not been re- ported. The sites of recombination on chromosome 9 seem to be less restricted and are usually at least 40 kb upstream from known exons of c-abl. These findings focus attention on the possibility that the bcr region may have some role in leukemogenesis; although bcr is transcribed in at least some CML lines (J. Groffen, personal communication), there is no information about its coding potential. Similarly, the effect of the translocation upon the behavior of c-ab1 has not been fully defined. c-ab1 mFWA of atypically large size and increased abundance (27a, 38a) and a larger than normal form of the c-ab1 protein with an active tyrosine kinase (0. Witte, personal communication) have been recently observed in the CML line K562. Moreover, in the same line, the Ph' chromo- some is accompanied by a modest amplification and an incompletely elucidated rearrangement of c-abl, with co-amplification of a resident marker for chromo- some 22, the h Ig gene (38, 95). Other Ph'-positive CML lines that do not display amplification or local rearrangement of c-abl also contain augmented levels of an enlarged c-ab1 RNA (38a, 27a). Neither the X Ig gene [which is at CELLULAR ONCOGENES 587 least 45 kb distant from bcr and retained on Ph' (175)] nor c-sis [which is far from bcr on chromosome 22 and transferred to 9q+ during formation of Ph' (ll)] is considered likely to play a direct role in CML. OTHER TRANSLOCATIONS AND REARRANGEMENTS IN E4 CELL TUMOR LINES Perlmutter et a1 (168) have described a 6;10 translocation in the NS-1 line, a derivative of MOPC21, that joins sequences about 1 kb on the 5 ' side of the Ig kappa constant locus to an unidentified region of chromosome 10, callcd TKNS-1. Although there is no sequence homology at the breakpoints, there is an int~guing similarity between the involved region of the IgK locus and a nearby region of the Igrc locus that is joined to the S, region from chromosome 12 in a bizarre t(6;12;15) rearrangement in the PC line 7183 (242). In the latter line, the other end of the S, domain is joined to an unidenti~ed region of chromosome 15 (Le. not c-myc). There are now some circumstantial reasons to believe that the TKNS-1 region may be functionally significant: it is rearranged in another mouse tumor cell line, the Abelson MLV-induced B cell line, 18-81; modestly amplified in X63 Ag8, another line derived from MOPC 21; ex- pressed in the NSl cells but not in several normal organs; and evolutionarily conserved (R. Perlmutter, personal communication). It is possible that the effects of translocations upon gene expression can be mimicked by deletions or inversions. Weiner et a1 have described interstitial deletions of band D of mouse chromosome 15 that appear to augment the expression of e-myc (assigned to 15 D2/3), with or without rearrangements detectable by restriction mapping of c-myc (255). These deletions are found in Ab-MLV producing PC lines, but it is not known whether the rearrangements are directly related to virus infection. Structural Features of Amplified Oncogenes The current picture of the molecular events that amplify proto-oncogenes is at least as fuzzy as our view of amplifications affecting other genes. Because a phenotype as simple and selectable as drug resistance cannot as yet be ascribed to oncogenic amplifications, it has not been possible to define experimental conditions to identify factors, such as inhibitors of DNA synthesis (23), that might influence the frequency with which such amplifications occur. Nor has a full structural analysis of any amplified units bearing oncogenes been attempted. Nevertheless, a few instructive features of amplified oncogenes have been described. 1. Amplified oncogenes may exist as part of either DM chromosomes or HSRs, even within sibling lines derived from a single tumor (3, 199). 2. The crudely approximated sizes of some amplified domains bearing oncogenes-I00 to 1000 kb or more-are similar to the sizes of amplified 588 VARMUS domains bearing other recognizable genes (197a). Accordingly, additional presumably adjacent genes may be co-amplified with the implicated oncogenes, at least one other in the mouse adrenocortical tumor with amplified c-Ki-ras (80a) and several others in the IMR 32 line with ampli- fied N-myc (1 11). 3. Oncogene-containing HSRs may be located in chromosomal contexts other than those in which the oncogene is normally situated. For instance, N-myc, a normal resident of chromosome 2p23-24, is amplified within HSRs found on a different chromosome in each of several neuroblastoma cell lines (1 1 1, 201). In a subline of the COL0320 tumor line, amplified c-myc (derived from 8q24) is present in an HSR situated on both sides of the centromere in a marker chromosome that also contains part of the X chromosome (K. Alitalo, C. C. Lin, unpublished data). It is possible that the identified sites are predisposed to amplify transposed DNA (e.g. after integration of one or more DM chromosomes) in a sense analogous to that observed for metaboli- cally-selectable genes inserted at different sites in cells subjected to DNA transformation (247). 4. Amplified oncogenes may exhibit other kinds of potentially oncogenic mutations that presumably preceded an amplification step. For example, the transformation-competent, point-mutated c-Ki-ras2 allele in the human colon carcinoma line SW480 and the translocated c-ab1 locus on a Ph' chromosome in the human CML line K562 are both modestly amplified (38, 95, 135). In a COL0320 subline bearing DM chromosomes, many of the amplified copies of c-myc have lost the first (non-coding) exon as a result of a rearrangement that joins a site in the first intron to DNA of unknown provenance (3; M. Schwab, unpublished data); an atypical c-myc RNA species presumed to represent the rearranged locus is at least as abundant as RNA from the unrearranged gene. A rearrangement of c-erb-B also appears to have occurred during its amplification in A43 1 cells (1 36). On the other hand, no coding changes in codons 12,59, and 61 have been encountered in c-Ki-ras cDNA clones from the mouse tumor line Y 1 (D. George, personal communication). 5. Although oncogene amplification is assumed to be a somatic event, direct evidence is generally lacking. In at least one patient with neuroblastoma, it has been formally proven that N-myc amplification is a somatic event, since a normal complement of N-myc genes is present in normal tissue (200). FROM C-OIZC TO V-OIZC: COLLUSION OF MUTATIONAL MECHANISMS The most flagrant examples of mutant oncogenes are to be found among the retroviral oncogenes, the transduced, multiply mutated, and highly tumorigen- CELLULAR ONCOGENES 589 ic forms of cellular proto-oncogenes. What molecular events underwrite the transduction mechanism? What changes in the oncogenes ensue? And which changes are significant for the biological activity of the v-onc's? The complete process of transduction of proto-oncogenes by retroviruses occurs too infre- quently to studied under controlled conditions. Thus, the events must be reconstructed instead from those artefacts presently available-the sequences of v-onc's and c-onc's-and conjectures tested against further sequence data and against attempts to recapitulate portions of the scheme under more favor- able conditions than those existing naturally. Several important differences between v-onc's and c-onc's have been en- countered in the many comparisons now available (239). 1. v-onr's appear to be truncated versions of c-onc's, frequently lacking coding sequence from both ends of the transduced domain. When missing -the portion encoding the amino terminus of the c-onc product, the trans- duced domain is generally expressed under viral translational signals as a fusion protein encoded in part by viral replicative genes (gag or env). Occasionally, v-om's retain the coding sequence from one or both ends of the c-onc progenitor (e.g. in the cases of v-SYC, FBJ-fos, or v-Ha-rus), but regulatory signals for synthesis and processing of RNA are invariably provided by retroviral sequences (including the LTRs) present on both 5' and 3` sides of oncogcnic sequences. 2. All of the introns that lie between transduced exons in c-onc's are absent from v-onc's; but in some instances [e.g. v-src (226, 228), v-myb (108, 1 10) and MC29-myc (207, 252; C. Nottenburg, unpublished data)] there remain short regions of intron-derived sequences from the 5' side of the first transduced exon. These intronic sequences may be contiguous with the exonic sequences in the host genome and thus include an intact splice acceptor site that is used to generate both c-onc and v-onc mRNAs (as is true for myb and src genes); or the intronic sequences may be non-contiguous with exonic sequences, having been positioned upstream from a cryptic splice donor site activated during the molding of the viral oncogene (e.g. in the generation of MC29-myc). 3. Comparisons of the sequences of transduced genes and transducing retrovi- ral genomes frequently display short (e.g. 4-6 bp) stretches of homology at the recombination sites on either side of the onc region (239). 4. Several changes in the sequence of the transduced domain are present in each v-onc. These commonly take the form of multiple base substitutions, including silent, conservative, and non-conservative changes; in addition, deletions of various sizes are often encountered [e.g. in the PRCII--s gene (29a, 99)], and some of these shift the reading frame [e.g. in the FBJ-fos gene (24)l. 590 VARMUS Two central concerns arise from these comparisons: (a) what mechanisms for transduction are compatible with observed differences between c-onc's and v-onc's, and (b) which of the differences between c-om's and v-onc's are significant factors in oncogenic potency? The Favored Model for Transduction of Oncogenes Any transduction mechanism must recombine retroviral sequences with both 5 ' and 3' sites in or near the oncogene, eliminate all introns between transduced exons, and permit retention of some intronic sequences on the 5' side of the first transduced exon. Base substitutions and deletions within the onc domain could be explained by events that occur before, during, or after transduction and are hence less problematic. The prevailing model for conversion of proto-oncogenes to viral transfonn- ing genes (16,18,226,244,249) recapitulates several of the mutational themes of this essay. The initial event is held to be an insertion mutation. During retroviral infection, the transducing provirus is integrated on the 5' side of the c-onc to be captured in the same transcriptional orientation. Although direct evi- dence is lacking, it is attractive to presume that the insertion mutation either initiates tumorigenesis as proposed above for several retrovirus-induced tumors or at least stimulates growth; clonal expansion of the cell containing an insertion mutation appropriate to instigate transduction would markedly improve the probability that the subsequent rare events also necessary for transduction will occur. The second step is proposed to be a deletion mutation that extends from a site within the provirus to a site within the oncogene. The latter site can be within either an intron or an exon, accounting for the variety of 5' end points for host-derived sequences in retroviral genomes. (Smaller deletions or none are also compatible with the model provided that the polyadenylation signal in the 3' LTR is occasionally overridden; aberrant splicing or later deletions during reverse transcription would then have to be invoked to explain the observed viral genomes.) The proposed deletion creates a hybrid transcriptional unit, one that should generate a primary transcript with sequences normally found at the 5' end of viral RNA joined to c-onc sequences that extend to the 3' end of the gene. 3. The hybrid RNA will be recognized as a substrate both for splicing (to remove the introns that lie between c-onc exons) and for packaging into virus particles. 4. To join viral sequences necessary for viral replication to the 3 ' side of onc sequences, the mechanism that facilitates high frequency recombination between retroviral genomes is invoked. Although the details of that CELLULAR ONCOGENES 59 1 mechanism are still in dispute, the central events are agreed to be the formation of heterodimers during virus assembly and recombination during reverse transcription, the latter step favorably influenced by homology between the genomic subunits (33, 100, 103, 244a). Heterodimers com- posed of a viral-onc hybrid transcript and a wild-type viral RNA from another provirus in the same cell can be expected to recombine at relatively low frequency, since any homology at the 3' ends of the RNAs will be fortuitous. Occasional and irregular recombinational products have been obtained in cell culture using model substrates (82,84), however, providing one of the few pieces of experimental support for the transduction mecha- nism. When partial deletion mutants of v-onc-containing viruses are used to induce tumors or infect cells in cultures, the missing domain is recovered from the c-onc at a modest frequency, implying that homology accelerates the process of transduction (177,204,249,250). Such results are consistent with the model discussed here and with others as well. It might be possible to challenge aspects of the model more forcefully by using the few situations in which the transduction of oncogenes appears to occur in cell culture without obvious homology (178-1 SO), by attempting to capture metaboli- cally selectable genes, or by exploiting the recent discovery that myc- containing FeLV proviruses are produced in a substantial number of FeLV- associated lymphomas arising in domestic cats (124, 141, 147). The mechanism proposed for transduction of oncogenes requires further embroidery to account for a number of retroviral genomes. 1. There are now five instances in which sequences from two apparently unlinked cellular genes have been found in single viral genomes (Table 2). This situation could arise if transduction occurred at a site of chromosomal translocation that had previously fused the two-cellular domains to be transduced; alternatively, the retrovirus that arises from the capture of one host sequence might then direct transduction of a second. 2. Two peculiarities at the 3' end of v-src in the genomes of RSV (123, 202, 226, 228) also require more elaborate explanation. First, the sequence that terminates the v-src coding region originates outside the c-src coding sequence, about 900 bp downstream from the stop codon in c-src. This situation probably results from deletion formation in the host chromosome or during reverse transcription of hybrid RNA; either deletion would be facilitated by an eight bp direct repeat shown to exist at the joining sites (228). Alternatively, an aberrant splicing event could have joined the downstream sequence to the penultimate coding exon. Second, v-src alleles are flanked by direct repeats of viral sequences (approximately 100 nu- cleotides in length). These are proposed to result from recombinational events during reverse transcription: the 3' end of the onc region appears to 592 VARMUS have been joined to a viral sequence in wild-type RNA that was also present upstream of the oncogene in the viral-host hybrid RNA. 3. The several nucleotide substitutions and deletions within the transduced domains of virtually all oncogenes (18,239) can be most simply explained as the products of the error-prone process of reverse transcription (34), with selection for any alterations that create a more effective transforming gene. However, it is conceivable that some of the observed differences in nu- cleotide sequence arose in the host chromosome in the early stages of transduction, augmenting the proposed neoplastic effects of the primary insertion mutations. The Functional Significance of Differences between c-onc's and v-onc's Despite considerable experimental effort, it remains difficult to assign func- tional attributes to the differences that distinguish v-om's from c-onc's. This is so in part because it is often difficult to control the intracellular concentration of gene products and hence to distinguish qualitative from quantitative effects, and in part because the cell types most readily employed in the laboratory are not necessarily the target cells for oncogenesis in the animal. Several c-om's [including c-Ha-rusl (31), c-mos (20, 159), c-fos (137), and c-myc (114; W. Lee et al, unpublished data)] have oncogenic effects on cultured fibroblasts when expressed more efficiently than usual under the influence of viral promoters, even though their protein coding sequence is unaltered. Furthermore, some of these genes (and others) have been implicated in rearrangements (Tables 4-6) that appear primarily to affect gene expression, implying an oncogenic effect based on dosage. But all of the v-om derivatives of these genes have multiple alteration in the coding sequence that may also be significant. Most obviously, the differences between c-Ha-rusl and v-Ha-rus include amino acid changes sufficient to render c-Ha-rasl transforming with- out the replacement of its native promoter (Table 7). Evidence for the trans- forming activity of normal c-myc under the control of an LTR has been obtained in primary rodent embryo cells in the company of a mutant c-Ha-rasl allele; the various v-myc alleles have yet to be compared to c-myc in a systematic fashion in the several target cells for oncogenesis by myc genes. There are gross differences between v-fos and c-fos alleles [the v-fos sequences are either fused with sequences from two other genes, gag andfox, or altered by a frameshifting deletion of 104 bp (240,241)], and these could enhance the neoplastic potential of the genes in ways too subtle to appreciate in cultured fibroblasts. As an illustration of this possibility, the v-ab1 sequence, expressed as a gag-ab1 fusion gene in the genome of Ab-MLV, appears to require at least a portion of the CELLULAR ONCOGENES 593 preceding gag domain for the transformation of lymphocytes but not for the transformation of fibroblasts (173). The transforming activity of the v-src gene seems to depend primarily upon qualitative changesaither the scattered base substitutions throughout the gene or the concerted change at its 3' terminus. Expression of c-src at a level above that required for transformation of the same cells by v-src does not induce transformation (166; H. Hanafusa, personal communication), although it is possible that some high threshold for transformation by c-src exists. In con- trast, an amount of v-src protein only slightly higher than that of c-src protein is transforming, although again the dose is important: the cellular phenotype can be radically altered with a four-fold change in v-src protein achieved by manipulating expression with a hormone-responsive promoter (E. Jakobovits, unpublished data). In the aggregate, these findings suggest that full activation of proto- oncogenes through retroviral transduction may depend upon multiple changes that affect both the structure and expression of the oncogr ne product. Further- more, these changes mimic many of the individual alterations described in cellular oncogenes during tumorigenesis. It is likely that retroviral oncogenes, particularly those with a long passage history, are repositories of a number of mutations that enhance their neoplastic potential and thus favor their selection under laboratory conditions. A completely satisfying analysis of the phenotypic consequences of these iesions has yet to be performed. INACTIVATION OF INTEGRATED V-OY~C GENES A SAMPLER OF MUTATIONAL MECHANISMS Most of this essay has been devoted to those mutations that activate proto- oncogenes; what can be said of mutations that cripple active oncogenes? Numerous mutations of the latter sort have been isolated during the study of viral oncogenes (l25,234a), but the vast majority of these have been generated during virus life cycles, with or without the aid of mutagens, or by manipula- tion of viral DNA in vitro. Hence, although crucial to an understandii g of the functions of oncogenes, they are not directly informative about the mutations that occur in eukaryotic chromosomes and fall beyond our purview here. In a few instances, however, mutant viral oncogenes have been isolated by seeking phenotypic revertants of virally transformed cell lines that carry only integrated viral genomes (58, 134, 219, 246). In one particularly large collec- tion of mutants, assembled from non-transformed derivatives of a rat cell (called B31) carrying a single integrated RSV provirus, several kinds of lesions have been shown to.inactivate the v-src gene (132, 158, 245, 246). Among 594 VARMUS over 100 morphological revertants of B3 1 cells, approximately one-third have lost the entire RSV provirus, presumably by either chromosomal non- disjunction or a large deletion; about 5% have suffered deletions (1.5 kb to over 20 kb in length) that encompass the 5' LTR but spare the 3' portion of the provir,is; two incuried insertion mutations following superinfection with another retrovirus lacking an oncogene (MLV); at least two mutations involved the loss or gain of a single base pair in the v-src coding sequence; and the rest, over half, appear to be due to nucleotide substitutions in v-src. The insertion and frameshift mutations in the B31 cell line have been examined in greatest detail, in part because some of them undergo secondary mutations that restore aspects of the transformed phenotype. Both of the insertional mutagens are apparently intact MLV proviruses located between the splice donor and acceptor sites for production of src mRNA in the same transcriptional orientation as the RSV provirus (245). Mature transcripts from the mutant loci appear to begin in the RSV 5' LTR and end in the MLV 3' LTR; thus, v-src is not expressed in mRNA. One of the two insertion mutants reverts to a wild-type phenotype by homologous recombination between the MLV LTRs at a frequency of 10-6-10-7 per cell per generation, leaving behind a single MLV LTR that has little apparent effect on the production of src mRNA. The residual MLV LTR does not appear to serve as a promoter for v-src; however, when retrieved from the cell and dissociated from the RSV LTRs by molecular cloning, it is competent to direct the expression of a heterologous gene lacking its native promoter (S. Ortiz, unpublished data). Two frameshift mutations of v-src have been defined by nucleotide sequenc- ing, and both conform to the dictum that loss or gain of a nucleotide pair is most likely to occur by slippage of DNA polymerase in regions containing multiple residues of a single base (221). In one case, the sequence GAZG was converted to GA3G (132); in the other, the sequence CGSA was converted to CG4A (G. Mardon, unpublished data). Both frameshift mutants revert to phenotypes resembling wild type at moderate frequency. One revertant of the +1 frameshift mutant was found to make a v-src protein approximately 8 kd larger than normal as a result of a spontaneous 242 bp duplication: since the dupli- cated region included the primary mutation and created an additional + 1 shift of frame at the boundary between the two copies, three successive i-1 frameshifts ultimately restored the correct reading frame (132). Two revertants of the - 1 frameshift mutant result from insertion of one bp or the deletion of two bp downstream from the primary mutation, with a portion of v-src trans- lated from a fortuitously open reading frame between the primary and secon- dary mutations (G. Mardon, unpublished data). Neither of these secondary mutations involves expansion or contraction of pre-existing multiples of a single nucleotide; in one case, the sequence ACTGAA was converted to ACAA, and in the other the sequence GGCTG was converted to GGCTTG. I CELLULAR ONCOGENES 595 NEW DIRECTIONS TOWARD A MOLECULAR DEFINITION OF CANCER Our main business is not to see what lies dimly at a distance, but to do what lies clearly at hand (Carlyle) Attending to the main business at hand, the biochemical characterization of mutations affecting cellular oncogenes in turnor cells, has provided the fuel for this essay. However, the described oncogenic mutations are no doubt limited in scope by the procedures available for discovering them; further technical and theoretical innovation may now be required to provide a satisfactory picture of oncogenesis, of the normal functions of proto-oncogenes, and of the nature of somatic mutations in vertebrate oncogenes. This penultimate section considers some prospects for the more complete view that may lie dimly at a distance. Recessive Oncogenic Mutations All of the lesions thought to influence neoplasia and summarized to this point are presumed or documented to act in a dominant manner. However, there is abundant rationale for believing that recessive mutations contribute directly to oncogenesis. On purely speculative grounds, it is possible that cells could progress to neoplasia because they are defective in functions required for entry into an irreversible differentiation pathway or required for responsiveness to external signals that normally retard cell cycling. Such vague possibilities take on weight in view of genetic evidence for inherited heterozygous states that strongly predispose to neoplasia in defined organs, appear to depend on a second mutation for expression (172), and have been dissociated from some of the known proto-oncogenes by linkage studies (loa). Direct support for the idea that the second mutation might involve the normal allele at the same locus on the companion chromosome has recently emerged from studies of retinoblastomas (13, 30, 142). Individuals with karyological or enzymatic signs of a small deletion mutation involving 13q14 in germinal cells develop retinoblastomas in which enzymatic assays for esterase D (a marker closely linked to 13q14) and tests for restriction site polymorphisms suggest conversion of a heterozygous to a homozygous or hemizygous state. The studies to date indicate that a number of mechanisms- including mitotic recombination, chromosomal nondisjunction with or without duplication, gene conversion, and point mutation-could account for one or another of the observed cases, although mitotic recombination seems particu- larly prevalent (30). These findings should stimulate efforts to isolate the posited retinoblastoma gene (Rb), to define the lesions that produce Rb-/Rb+, Rb-/Rb-, and Rb-/O genotypes, and to seek better definition of other loci at which recessive mutations are believed to contribute directly to cancer [e.g. the Wilms's tumor-aniridia locus on llp13 (69)l. 596 VARMUS Mutations of Proto-Oncogenes in Genetically Malleable flosts The only mutations of proto-oncogenes in vertebrate cells that seem presently accessible are dominant and apparently oncogenic. Such mutations seem in- herently unlikely to provide much insight into the normal functions of proto- oncogenes; the phenotypic requirements may in addition impose a bias toward certain types of molecular changes (e.g. genetic rearrangements). The discov- cry of homologues of vertebrate proto-oncogenes, mainly c-onc's, in the genomes of Saccharomyces cerevisiae and Drosophila melanogajter, tradi- tionally favored for the induction and isolation of mutants, has enormously increa: ed the opportunity for obtaining natural, induced, and engineered mutants in such genes. Thus far, most of the reported studies of c-onc homologues in yeast and flies have been confined to structural analysis of such genes [e.g. ras-like genes in yeast (76, 153, 172a) and src- and ubl-like genes in Drosophila (97,212)], but genetic approaches are near or at fruition as well. 1. Homologous recombination during DNA transformation of yeast has been used to inactivate the two genes (RAS1 and RAS2) homologous to Ha- and Ki-uas genes; disruptive lesions in either locus do not affect viability of haploids or diploids, but lesions in both loci prevent resumption of vegeta- tive growth by haploid spores (104a, 230a). Similar mutation of another yeast gene less closely related to vertebrate ras genes (76) is also a recessive lethal (N. Segev, D. Botstein, personal communication). 2. Introduction into normal cells of a mutant yeast ras gene with a gly + Val change in the position that corresponds to amino acid 12 in human c-Ha- rusl interferes with efficient sporulation (1 04a). 3. Existing mutants that affect progression through the yeast cell cycle seem likely in some cases to be assigned to the homologues of vertebrate proto- oncogenes. The product of the cdc28 locus, for example, shows about 25-30% amino acid homology to several c-onc proteins and to cyclic AMP-dependent protein kinase (128), and the cdc4 and cdc36 loci show modest similarity to the host-derived sequence, ets, in an avian retrovirus (170b). 4. By using known deficiencies that encompass the chromosomal positions of oncogene homologues in Drosophila, it is possible to screen for new recessive lethal mutants likely to map within loci of interest (M. Simon, M. Hoffman, personal communication). 5. Striking variations in the mode and level of expression of the src and ab1 homologues occur during development of Drosophila (123a, M. Simon, B. Drees , personal communication), offering promise of insight into multiple effects of proto-oncogenes upon ontogeny. CELLULAR ONCOGENES 597 Identijying Genes Whose Products Might Interact with Oncogene Proteins In other experimental systems it has proven instructive to harness genetic as well as biochemical methods to identify protei is that interact in functionally important ways (21). This is generally done by seeking extragenic suppressors of primary mutations. At least two sorts of approaches seem plausible exten- sions of this general strategy to oncogenes. First, viewing active oncogenes as mutant proto-oncogenes, it should be possible t9 isolate cells with suppressor mutations in other genes, e.g. host mutants that prevent transformation by certain retroviral oncogenes, perhaps by altering targets for critical enzymatic activities such as protein phosphorylation on tyrosine residues (40). There are accounts of mutants that might meet these criteria for papovaviral or retroviral oncogenes (e.g. 86, 101, 152, 187, 220), but the most thoroughly studied are rat lines resistant to transformation by several, but not all, retroviral oncogenes (151). Products of the restricted oncogenes-v-Ki-rus, v-Hams, v-fes, and v-src-are not similar on biochemical grounds, suggesting that the mutation might affect some relatively late function in neoplastic transformation. Further- more, analysis of somatic cell hybrids suggests, surprisingly, that resistance is dominar,t. These features may diminish the chances of using the mutants to identify a protein that normally comes in direct contact with a product of a restricted v-onc gene. A second strategy is to isolate mutant cellular genes that suppress conditional or non-conditional mutations in active oncogenes. To date, no cellular mutants of this type have been described, although there exist mutant viral oncogenes whose behavior is host dependent. One of the mutants of the B31 v-src gene, for example, produces no apparent phenotypic effect in rat cells (in which the v-src protein displays no tyrosine kinase activity), but the same gene trans- forms chicken cells where its product is active as a protein kinase (158, 246). DNA transformation and molecular cloning procedures theoretically could permit the isolation of cellular genes that restore the activity of such mutant proteins. Assessing Epigenetic Changes The findings summarized in this review further the argument that cancer has its origins in genetic rather than epigenetic change. Still, important non- mutational events seem likely to be intimately involved in oncogenic mecha- nisms and may be unfairly neglected (191a). Measurement of polyadenylated c-onc RNA is perhaps the most obvious way to begin a search for relevant epigenetic change, but even when relatively dramatic, specific, and consistent increases in c-onc RNA are found in the absence of overt mutations (66,213), interpretation is uncertain. In particular, it is difficult to distinguish between effects that directly influence the oncogenic phenotype and those that simply 598 VARMUS reflect either the transformed state or the position of the tumor cell within a cellular lineage. As experience broadens with cultured cells in many differenti- ated states and as the potential increases for transforming such cells with mutant oncogenes, it may be possible to distinguish between events that are critical for achieving the transformed state and those that are only consequences of it. A CONCLUDING PERSPECTIVE Prior to the intensive pursuit of oncogenes, information about the structural forms of mutations in eukzryotic chromosomes was derived largely from germ line mutations of globin genes (130) and from the genetic alterations that accompany the specialized process of immunoglobulin gene development (234). Attempts to use bacterial genes introduced into cultured animal cells as easily recovered targets for somatic mutations have been compromised by the extremely high frequency of mutational events (26, 183). With molecular cloning of several cellular genes that confer metabolically selectable pheno- types in culture, somatic mutations affecting a number of genes in their normal chromosomal context have been described in molecular terms, if not usually at the level of nucleotide sequence. Among these are nucleotide substitutions and deletions affecting the uprt gene (2 1 1); deletions, possible translocations, amplification, and presumptive nucleotide substitutions affecting the hprt locus (75); and amplification and deletions affecting the dkfr locus (197, 238). However, the sheer number of scrutinized somatic mutations that affect proto- oncogenes cannot be rivalled in other quarters of the eukaryotic genome, and the analysis to date is certain to affect predictions about forms of mutational change throughout the genome. Although the types of mutations observed cannot be considered novel, the drama for geneticist and oncologist alike is to be found in the startling patterns: the recurrence of certain types of mutant ras alleles in a wide variety of tumors or the strong association of certain rearrange- ments with certain tumors (e.g. N-myc amplifications in human neuroblasto- mas, int-1 or int-2 insertion mutations in mouse mammary tumors, or c-myc translocations in Burkitt lymphomas). What do such observations imply about mutational mechanisms? Do some lesions occur with higher frequency than expected from random mutation? Or can all be explained by selective pressures during the growth of tumor cells? What can be said about oncogenic mutations at present is largely descriptive; with the possible exception of proviral insertion mutations, little is known about the mechanisms by which such somatic mutations arise. It has been argued that cells en route to malignancy may be mutators and more liable to incur either nucleotide substitutions or genetic rearrangements than are normal cells or cells less advanced on a neoplastic pathway (32). Whether the observa- tions that sustain this argument are dependent on some specialized mutational CELLULAR ONCOGENES 599 apparatus or on more frequent action of mechanisms shdred with normal cells is beyond our power to decide at present. More importantly for an understanding of cancer, we lack clear insight into phenotypic consequences of mutations affecting proto-oncogenes. This de- ficiency reflects in part our imprecise definition of neoplastic change and present constraints on the ways in which we assay for it. But, as must be evident from this survey, the difficulty of assigning biological significance to mutations in tumor cells is implicit in the complexity of the neoplastic genotype, a? well as the phenotype. Wc have seen several examples of single alleles affected by multiple mutations and of cells affected by mutations at multiple loci, although there is little reason to believe that the search for genetic change in neoplastic cells has been exhaustive. To learn which lesions initiate tumorigenic growth, which maintain it, and which lead to higher states of malignancy are tasks likely to occupy students of oncogenic mutations for many years to come. ACKNOWLEDGEMENTS I thank many colleagues here and elsewhere for helpful discussions, J. Marinos for artful preparation of the manuscript, and the NIH and ACS for support. Literature Cited 1. Adams, J. M., Gerondakis, S., Webb, E., Corcoran, L. M., Cory, S. 1983. Cellular myc oncogene is altered by chromosome translocation in murine plasmacytomas and is rearranged similar- ly in human Burkitt lymphomas. Proc. Natl. Acad. Sci. USA 80:1982-86 2. Albino, A. P., LeStrange, R., Oliff, A. I., Furth, M. E., Old, L. J. 1984. Trans- forming ras genes from human melano- ma: A manifestation of tumour heter- ogeneity? Nature 308:69-72 3. Alitalo, K., Schwab, M., Lin, C. C., Varmus, H. E., Bishop, J. M. 1983. Homogeneously staining chromosomal regions contain amplified copies of an abundantly expressed cellular oncogene (c-myc) in malignant neuroendocrine cells from a human colon carcinoma. Proc. Natl. Acad. Sci. USA 80:1707-11 4. Alitalo, K., Winqvist, R., Lin, C. C., de la Chapelle, A,, Schwab, M., Bishop, J. M. 1984. Aberrant expression of an am- plified c-myb oncogene in two cell lines from a colon carcinoma. Proc. Natl. Acad. Sci. USA 81: 453438 5. Ames, B. N. 1979. Identifying environ- mental chemicals causing mutations and cancer. Science 204:587-93 6. Andersen, P. R., Devare, S. G., Tro- nick, S. R., Ellis, R. w., Aaronson, S. A., Scolnick, E. M. 1981. Generation of BALB-MuSV and Ha-MuSV by type C virus transduction of homologous trans- forming genes from different species. Cell 26: 129-34 7. ar-Rushdi, A,, Nishikura, K., Erikson, J., Watt, R., Rovera, G., Croce, C. M. 1983. Differential expression of the translocated and the untranslocated c- myc oncogene in Burkitt lymphoma. Sci- ence 222:39&93 8. Balmain, A,, Pragnell, I. B. 1983. Mouse skin carcinomas induced in vivo by chemical carcinogens have a trans- forming Harvey-ras oncogene. Nature 9. Balmain, A,, Ramsden, M , Bowden, G. T., Smith, J. 1984. Activation of the mouse cellular Harvey-ras gene in chem- ically induced benign skin papillomas. Nature 307:658-60 10. Banerji, J., Olson, L., Schaffner, W. 1983. A lymphocyte-specific cellular en- hancer is located downstream of the join- ing region in immunoglobulin heavy chain genes. Cell 33:729-40 loa. Barkee, D., McCoy, M., Weinberg, R. A,, Goldfarb, M., Wigler, M., et al. 1983. A test of the role of two oncogenes in inherited predisposition to colon can- cer. Mol. Cell Med. 1:199-206 303 172-74 600 VARMUS 11. Bartram, C. R., de Klein, A,, Hagemeij- er, A,, Grosveld, G., Heisterkamp, N., Groffen, J. 1984. Localization of the hu- man c-sis oncogene in Ph'-positive and Ph'-neeative chronic mvelocvtic leuke- mia by in situ hybri&atibn. Blood 63:223-25 12. Battey, J., Moulding, C., Taub, R., Murphy, W., St-wart, T., Potter, H., Lenoir. G.. Ledcr, P. 1983. The human .. c-myc oncogene: Structural conse- quences of translocation into the IgH locus in Burkitt lymphoma. Cell 34:779- 87 13. Benedict, W. F., Murphree, A. L., Banerjee, A,, Spina, C. A,, Sparkes, M. C., Sparkes, R. S. 1983. Parient with 13 chromosome deletion: Evidence that the retinoblastoma gene is a recessive cancer gene. Science 219:973-74 14. Bingham, P. M., Levis, R., Rubin, G. M. 1981.-Cloning of DNA sequences form the white locus of D. melanogaster by a novel and general method Cell 25:693-704 15. Deleted in proof 16. Bishop, J. hl. 1983. Cellular oncogenes and retroviruses. Ann. Rev. Biochem. 52:301-54 17. Bishou. J. M.. Varmus. H. E. 1982. Functions and origins of retroviral trans- forming genes. In Molecular Biology of Tumor Viruses, Part 111 RNA Tumor Viruses, ed. R. Weiss, N. Teich, H. Var- mus, J. Coffin., pp. 999-1108. Cold Spring Harbor, New York: Cold Spring Harbor Press 18. Bishop, J. M., Varmus, H. E. 1984. Functions and origins of retroviral trans- forming genes. In Molecular Biology of Tumor Viruses, RNA Tumor Viruses, ed. R. Weiss, N. Teich, H. Varmus, J. Cof- fin, Cold Spring Harbor Press, New York: Cold Spring Harbor Press. Rev. ed. 19. Blair, D. G., Cooper, C. S., Oskarrson, M. K., Eader, L. A., Vande Woude, G. F. 1982. New method for detecting cellu- lar transforming genes. Science 218: 1122-24 20. Blair, D. G., Oskarsson, M., Wood, T. G., McClements, W. L., Fischinger, P. J., Vande Woude, G. G. 1981. Activa- tion of the transforming potential of a normal cell sequence: A molecular model for oncogenesis. Science 212:941-43 21. Botstein, D., Maurer, R. 1982. Genetic approaches to the analysis of microbial development. Ann. Rev. Gener. 16:61- 84 22. Brodeur, G., Seeger, C., Schwab, M., Varmus, H. E., Bishop, J. M. 1984. Am- plification of N-myc in untreated human neuroblastomas correlates with advanced disease stage. Science. 224:1121-24 23. Brown, P. C., Tlsty, T. D., Schimke, R. T. 1983. Enhancement of methotrexate resistance and dihydrofolate reductase gene amplification by treatment of mouse 3T6 cells with hydroxyurea. Mol. Cell. Biol. 3:1097-1107 24. Cairns, J. 1975. Mutation selection and the natural history of cancer. Nafure 25. Cairns, J. 1981. The origin of human cancers. Nature 289:353-57 26. Calos, M. P., Lebkowski, J. S., Botch- an, M. R. 1983. High mutation frequen- cy in DNA transfected into mammalian cells. Proc. Natl. Acad. Sci. USA 80: 27. Canaani, E., Dreazen, O., KIzr, A., Rechavi, G., Ram, D., et al. 1983. Activation of the c-mos oncogene in a mouse plasmacytoma by insertion of an endogenous intracisternal A-particle genome. Proc. Natl. Acad. Sci. USA 27a. Canaani, E., Gale, R. P., Steiner- Staltz, D., Berrebi, A,, AghairE., Janus- zewicz, E. 1984. Aliered transcription of an oncogene in chronic myeloid leukaemia. Lancet 1593-94 28. Capon, D. J., Chen, E. Y., Levinson, A. D., Seeburg, P. H., Goeddel, D. B. 1983. Complete nucleotide sequences of the T24 human bladder carcinoma oncogene and its normal homologue. Na- ture 302:33-37 29. Capon, D. J., Seeburg. P. H., McGrath, J. P., Hayflick, J. S., Edman, U., et al. 1983. Activation of Ki-rus2 gene in human colon and lung carcinomas by two different point mutations. Nature 29a. Carlberg, K., Chamberlin, M. E., Beemon, K. 1984. The avian sarcoma virus PRCII lacks 1020 neucleotides of the fps transforming gene. Virology 135: 157-67 30. Cavenee, W. K., Dryja, T. P., Phillips, R. A,, Benedict, W. F., Godbout, R., et al. 1983. Expression of recessive alleles by chromosomal mechanisms in retino- blastoma. Nature 305:779-84 31. Chang, E. H., Furth, M. E., Scolnick, E. M., Lowy, D. R. 1982. Tumorigenic transformation of mammalian cells in- duced by a normal human gene homolo- gous to the oncogene of Harvey murine sarcoma virus. Nature 297:479-83 32. Cifone, M. A,, Fidler, I. J. 1981. In- creasing metastatic potential is associated with increasing instability of clones iso- lated from murine neoplasms. Proc. Natl. Acad. Sci. USA 78:6949-52 255:197-202 3015-19 80~7118-22 304:507-12 CELLULAR ONCOGENES 60 1 33. Coffin, J. M. 1979. Structure, replica- tion, and recombination of retrovirus genomes. Some unifying hypotheses. J. Gen. Virol. 42:l-26 34. Coffin, J. M. 1982. Retroviral genomes. See Ref. 17, pp. 261-368 35. Coffin, J. M., Varmus, H. E., Bishop, J. M., Essex, M., Hardy, W. D., et al. 1981. A proposal for naming host cell- derived inserts in retrovirus genomes. J. Virol. 40:953-57 36. Cohen, J. B., Unger, T., Rechavi, G., Canaani, E., Givol, D. 1983. Rearrange- ment of the oncogene c-mos in mouse myeloma NS 1 and hybridomas. Nature 37. Collins, S. J., Groudine, M. 1982. Am- plification of endogenous myc-related DNA sequences in a human myeloid leukaemia cell line. Nature 298:679-81 38. Collins, S. J., Groudine, M. 1983. Rear- rangement and amplification of c-ab1 sequences in the human chronic myelogenous leukaemia cell line K-562. Proc. Nail. Acad. Sci. USA 80:4813-17 38a. Collins, S. J., Kubonishi, I., Miyoshi, I., Groudine, M. 1984. Altered transcrip- tion of the c-ab1 oncogene in K-562 and other chronic myelogenous leukemia cells. Science 225:72-74 38b. Cooper, C. S., Park, M., Blair, D. G., Tainsky, M. A,, Huebner, K., et al. 1984. Molecular cloning of a new trans- forming gene from a chemically trans- formed human cell line. Nature In press. 39. Cooper, G. M. 1982. Cellular transform- ing genes. Science 217:801-06 40. Cooper, G. M., Hunter, T. 1983. Identi- fication and characterization of cellular 3061797-99 targets for tyrosine protein kinases. J. Biol. Chem. 258:1108-15 41. Cooper, J., Neiman, P. E. 1980. Trans- fodng genes of neoplasms induced by avian lymphoid leukosis viruses. Nature 287:65&59 42. Cooper, G. M., Neiman, P. E. 1981. Two distinct candidate transforming genes of lymphoid-leukosis virus- induced neoplasms. Nature 2925357-58 43. Cooper, G. M., Okenquist, S., Silver- man, L. 1980. Transforming activity of DNA of chemically transformed and nor- mal cells. Nature 284:418-21 43a. Corcoran, L. M., Adams, J. M., Dunn, A. R., Cory, S. 1984. MurineTlympho- mas in which the cellular myc oncogene has been activated by retroviral insertion. Cell 37: 113-22 44. Cory, S., Gerondakis, S., Adams, J. M. 1983. Interchromosomal recombination of the cellular oncogene c-myc with the immunoglobulin heavy chain locus in murine plasmacytomas is a reciprocal ex- change. EMBO J. 2:697-703 45. Cowell, J. K. 1982. Double minutes and homogeneously staining regions: Gene amplification in mammalian cells. Ann. Rev. Genet. 16:21-59 46. Crews, S., Barth, R., Hood, L., Prehn, J., Calame, K. 1982. Mouse c-myc oncogene is located on chromosome 15 and translocated to chromosome 12 in plasmacytomas. Science 218:1319-21 46a. Croce, C. M., Erikson, J., ar-Rushdi, A., Aden, D., Nishikura, K. 1984. Translocated c-myc oncogene of Burkitt lymphoma is transcribed in plasma cells and repressed in lymphoblastoid cells. Proc. Natl. Acad. Sci. USA 81:3170-74 47. Croce, C. M., Thierfelder, W., Erikson, J., Nishikura, K., Finan, J., et al. 1983. Transcriptional activation of an unrear- ranged and untrarislocated c-myc oncogene by translocation of a C locus in Burkitt lymphoma cells. Proc. Nd. Acad. Sci. USA 80:6922-26 48. Cullen, B. R., Lomedico, P. T., Ju, G. 1984. Transcriptional interference in avian retroviruses: Implications for the promoter insertions model of leuke- mogenesis. Nature 307:241-44 48a. Cuypers, H. T., Selten, G., Quint, W., Zijlstra, M., Maandag, E. R., et al. 1984. Murine leukemia virus-induced T- cell lymphomagenesis: Integration of proviruses in a distinct c!-.romosomal re- gion. Cell 37:141-50 49. Dalla-Favera, R., Bregni, hi,, Erikson, J., Patterson, D., Gallo, R. C., et al. 1982. Human c-myc oncogene is located on the region of chromosome 8 that is translocated in Burkitt lymphoma cells. Proc. Natl. Acad. Sci. USA 79:7824-27 50. Dalla-Favera, R., Martinotti, S., Gallo, R. C., Erikson, J., Croce, C. M. 1983. Translocation and rearrangements of the c-myc oncogene locus in human undiffer- entiated B-cell lymphomas. Science 51. C. Dalla-Favera, R., Wong-Staal, F., Gallo, R. C. 1982. Onc gene amplifica- tion in promyelocytic leukaemia cell line HL-60 and primary leukaemic cells of the same patient. Nature 299:61-63 52. Davis, M., Malcolm, S., Rabbitts, T. H. 1984. Chromosomal translocation can occur on either side of the c-myc oncogene in Burkitt lymphoma cells. Na- ture 308:28688 53. DeFeo-Jones, D., Scolnick, E. M., Kol- ler, R., Dhar, R. 1983. ras-Related gene sequences identified and isolated from Saccharomyces cerevisiae. Nature 306: 707-09 54. de Klein, A,, Guerts van Kessel, A., Grosveld, G., Bartram, C. R., 2 19:963-67 602 VARMUS Hagemeijer, A,, et al. 1982. A cellular oncogene is translocated to the Ptiil- adelphia chromosome in chronic myeloid leukaemia. Nature 300:765-67 55. Der, C. J., Krontiris, T. G., Cooper, G. M. 1982. Transforming genes of human bladder and lung carcinoma cell lines are homologous to the ras genes of Harvey and Kirsten sarcoma viruses. Proc. Natl. Acad. Sei. USA 79:3637-40 55a. Dhar, R., Ellis, R., Shih, T. Y., Orosz- Ian, S., Shapiro, B., et a]. 1982. Nucle- otide sequence of the p21 transforming protein of Harvey murine sarcoma virus. Science 217:934-37 56. Diamond, A., Cooper, G. M., Ritz, J., Lane, M. A. 1983. Identification and molecular cloning of the human Blym transforming gene activated in Burkitt's lymphomas. Nature 305:112-16 56a. Dickson, C., Smith, R., Brookes, S., Pews, G. 1984. Tumorigenesis by mouse manimarv tumor virus: Proviral activatiin of a ckllular gene in the com- mon integration region int-2. Cell: 37.529-36 -. - - 57. Donehover, L. A,, Huang, A. L., Ha- ger, G. L. 1981. Regulatory and coding potential of the mouse mammary tumor virus long terminal redundancy. J. Virol. 37:22638 55. Donner, L., Turek, L. P., Ruscetti, S. K., Fedelc, L. A,, Sherr, C. J. 1980. Transformation-defective mutants of feline sarcoma virus which express a product of the viral src gene. J. Virol. 35: 129-40 59. Doolittle,. R. F., Hunkapiller, M. W., Hood, L. E., DeVare, S. G., Robbins, K. C., et al. 1983. Simian sarcoma virus onc gene, v-sis, is derived from the gene (or genes) encoding a platelet-derived growth factor. Science 221:275-76 60. Downward, J., Yarden, Y., Mayes, E., Scrace, G., Totty, N., et al. 1984. Close similarity of epidermal growth factor re- ceptor and v-erb-B oncogene protein sequences. Nature 307521-27 61. Dunnick, W., Shell, B. E., Dery, C. 1983. DNA sequences near the site of reciprocal recombination between a c- myc oncogene and an immunoglobulin switch region. Proc. Natl. Acad. Sei. 62. Eisen, P. V. D., Houweling, A,, van der Eb, A. 1983. Expression of region Elb of human adenoviruses in the absence of region Ela is not sufficient for complete transformation. Virology 128:377-90 62a. Emanuel, B. S., Selden, J. R., Chagan- ti, R. S. K., Jhanwar, S., Nowell, P. C., et al. 1984. The 2p breakpoint of a 2;8 translocation in Burkitt lymphoma inter- ' USA 80:7269-73 rupts the V, locus. Proc. Natl. Acad. Sci. USA 8 1 :2444-46 63. Frikson, J., ar-Rushdi, A., Drowinga, H. L., Nowell, P. C., Croce, C. hI. 1983. Transcriptional activation of the translocated c-myc oncogene in Burkitt lymphoma. Proc. Natl. Acad. Sei. USA 80: 820-24 64. Erikson, J., Finan, J., Nowell, P. C., Croce, C. M. 1982. Translocation of im- munoglobulin V, genes in Burkitt lym- phoma. Proc. Natl. Acad. Sci. USA 65. Eva, A,, Aaronson, S. A. 1983. Frequent activation of c-kis as a transforming gene in fibrosarcomas induced by niethylcho- lanthrene. Science 220:955-56 66. Eva, R., Robbins, K. C., Andersen, P. R., Srinivasan, A,, Tronick, S. R., et al. 1982. Cellular genes analogous to retro- viral onc genes are transcribed in human tumor cells. Nature 295:116-19 67. Eva, A., Tronick, S. R., Gol, R. A., Pierce, J. H., Aaronson, S. A. 1983. Transforming genes of human hemato- poietic tumors: Frequent detection of ras- related oncogenes whose activation appears to be independent of tumor phe- notype. Proc. Natl. Acad. Sci. USA 80:492&30 67a. Fabricant, R. N., DeLarco, J. E., Todaro, G. J. 1977. Nervegrowth factors on human melanoma cells in culture. Proc. Natl. Acad. Sci. USA 14:565- 69 67b. Fasano, O., Aldrich, T., Tamanoi, F., Taparowsky, E., Furth, M., et al. 1984. Analysis of the transforging potential of the human H-ras gene by random mutagenesis. Proc. Natl. Acad. Sei. USA 8 1 :4008-12 68. Feig, L. A,, Bast, R. C. Jr., Knapp, R. C., Cooper, G. M. 1984. Somatic activa- tion of rasK gene in a human ovarian carcinoma. Science 223:698-700 69. Prancke, U., Holmes, L. B., Alkins, I., Riccardi, V. W. 1979. Aniridia-Wilms tumor association: Evidence for specific deletion of llp13. Cytogenet. Cell Genet. 24: 185-92 70. Frykberg, F., Palmieri, S., Beug, H., Graf, T., Hayman, M. J., et al. 1983. Transforming capacities of avian ery- throblastosis virus mutants deleted in the erbA or erbB onocogenes. Cell 32: 70a. Fujita, J., Yoshida, O., Yuasa, Y., Rhim, J. S., Hatanaka, M., et al. 1984. Ha-ras oncogenes are activated by so- matic alterations in human urinary tract turnours. Nature 309:46466 71. Fung, Y. K., Crittenden, L. B., Kung, H. J. 1982. Orientation and position of 79:5611-15 227-38 72. 73. 74, 75 avian leukosis virus DNA relative to the cellular oncogene c-myc in B-lymphoma tumors of highly susceptible 151 X 7* chickens. J. Virol. 44:74246 Fung, Y. K. T., Crittenden, L. B., Fad- ly, A. M., Kung, H. J. 1983. Tumor induction by direct injection of cloned v-src DNA into chickens. Proc. Null. Acad. Sci. USA 80:353-57 Fung, Y. K. T., Fadly, A. M., Critten- den, L. B., Kung, H. J. 1981. On the mechanism of retrovirus-induced avian lymphoid leukosis: Deletion and integra- tion of the proviruses. Proc. Natl. Acad. Sci. USA 78:3418-22 Fung, Y. K. T., Lewis, W. G., Kung,H. J., Crittenden, L. B. 1983. Activation of the cellular oncogene c-erbB by LTR in- sertion: Molecular basis for induction of erythroblastosis by avian leukosis virus. Cell 3 3: 3 5 7-68 Fuscoe, J. C., Fenwick, R. G. Jr., Ledbetter, D. H., Caskey, C. T. 1983. Deletion and amplification of the HGPRT locus in Chinese hamster cells. Mol. Cell Biol. 3:1086-96 76. Gallwitz, D., Donath, C., Sander, C. 1983. A yeast gene encoding a protein homologous to the human c-haslbas pro- to-oncogene product. Nuture 306:704- 07 77. Gambke, C., Signer, E., Moroni, C. 1984. Activation of N-ras in bone mar- row cells from a patient with acute myeloblastic leukemia. Nature 307:47& 77 78. Gattoni-Celli, S., Hsiao, W. L. W., Weinstein, I. B. 1983. Rearrangedc-mos locus in a MOPC 2 1 murine myeloma cell line and its persistence in hybridomas. Nature 306:795-97 79. Gelmann, E. P., Psallidopoulos, M. C., Papas, T. S., Dalla-Favera, R. 1983. Identification of reciprocal translocation sites within the c-myc oncogene and im- munoglobulin p, locus in a Burkitt lym- phoma. Nature 306:799-802 80. George, D. L., Powers, V. E. 1981. Cloning of DNA from double minutes of Y 1 mouse adrenocortical tumor cells: Evidence of gene amplification. Cell 24: 1 17-24 80a. George, D. L., Scott, A. F., de Martin- ville, B., Francke, U. 1984. Amplified DNA in Y 1 mouse adrenal tumor cells: Isolation of cDNAs complementary to an amplified c-Ki-ras gene and localization of homologous sequences to mouse chromosome 6. Nucl. Acids Res. 12: 273143 80b. Gerondakis, S., Cory, S., Adams, J. M. 1984. Translocation of the myc cellular oncogene to the immunoglobulin heavy CELLULAR ONCOGENES 603 chain locus in murine plasmacytomas is an imprecise reciprocal exchange. Cell 81. Gillies, S. D., Morrison, S. L., Oi, V. T., Tonegawa, S. 1983. A tissue-specific transcription enhancer element is located in the major intron of a rearranged im- munoglobulin heavy chain gene. Cell 82. Goff, S. P., Tabin, C. J., Wang, J. Y. J., Weinberg, R., Baltimore, D. 1982. Transfection of fibroblasts by cloned Abelson murine leukemia virus DNA and recovery of transmissible virus by recom- bination with helper virus. J. Virol. 41:271-85 83. Goldfarb, M., Shimizu, K., Perucho, M., Wigler, M. 1982. Isolation and preliminary characterization of a hu- man transforming gene from T24 blad- der carcinoma cells. Nature 296:404 09 84. Goldfarb, M. P., Weinberg, R. A. 1981. Generation of novel, biologically active Harvey sarcoma virus via apparent ille- gitimate recombination. J. Virol. 38: 136-50 85. Goubin, G., Goldman, D. S., Luce, J., Neiman, P. E., Cooper, G. M. !983. Molecular cloning and nucleotide se- quences of a transforming gene detected by transfection of chicken B-cell lympho- ma DNA. Nature 302:114-18 86. Graf, T., Beug, H. 1976. A novel type of cellular variant with altered expression of virus-induced cell transformation. Virol- 86a. Gregoire, D., Couez, D., Deschamps, J., Heuertz, S., Hors-Cayla, M. C., et al. 1984. Different bovine leukemia virus- induced tumors harbor the provirus in different chromosomes. J. Virol. 50: 87. Groffen, J., Stephenson, J. R., Heister- kamp, N., de Klein, A,, Bartram, C. R., et al. 1984. Philadelphia chromosomal breakpoints are clustered within a limited region, bcr, on chromosome 22. Cell 88. Guerrero, I., Calzada, P., Mayer, A., Pellicer, A. 1984. A molecular approach to leukemogenesis: Mouse lymphomas contain an activated c-ras oncogene. Proc. Natl. Acad. Sci. USA 81:202- 05 88a. Haigler, H., Ash, J. F., Singer, S. J., Cohen, S. 1978. Visualization by fluorescence of the binding and intemal- ization of epidermal growth factor in hu- man carcinoma cells A-43 1. Proc. Narl. Acad. Sci. USA 75:3317-21 89. Hall, A,, Marshall, C. J., Spun, N., Weiss, R. A. 1983. The transforming 36:973-82 33 :7 17-28 ogy 72:283-86 275-79 36:93-99 604 VARMUS gene in two human sarcoma cell lines is a new member of the ras gene family lo- cated on chromosome one. Nature 304: 90. Hann, S. R., Abrams, H. D., Rohr- schneider, L. R., Eisenman, R. N. 1983. Proteins encoded by the v-myc and c-myc oncogenes: Identification and iocalisa- tion in acute leukemia virus transform- ants and bursal lymphoma cell line. Cell 135-39 34~789-98 91.. Deleted in proof 92. Deleted in proof 93. Hayday, A. C., Gillies, S. D., Saito, €I., Wood, C., Kiman, W., et al. 1984. Activation of a translocated human c-myc gene by an enhancer in the irnmuno- globulin heavy-chain locus. Nature 307:33440 94. Hayward, W. S., Neel, B. G., Astrin, S. M. 1981. Activation of a cellular onc gene by promoter insertion in ALV- induced lymphoid leukosis. Nature 95. Heisterkamp, N., Stephenson, J. R., Groffen, J., Hansen, P. F., de Klein, A,, et al. 1983. Localization of the c-abl oncogene adjacent to a translocation breakpoint in chronic myelocytic leukaemia. Nature 306:239-42 96. Hengartner, H., Meo, T., Muller, E. 1978. Assignment of genes for im- munoglobulin Kappa and heavy chains to chromosome 6 and 12 in mouse. Proc. Natl. Acnd. Sei. USA 75:4494-98 97. Hoffman-Falk, H., Einat, P., Shilo, B. Z., Hoffman, F. M. 1983. Drosophila melanogaster DNA clones homologous to vertebrate oncogenes: Evidence for a common ancestor to src and ab1 cellular genes. Cell 32589-98 98. Hollis, G. F., Mitchell, K. F., Battey, J., Potter, H., Taub, R., et al. 1984. A variant translocation places the im- munoglobulin genes 3' to the c-myc oncogene in Burkitt's lymphoma. Nature 307:752-54 99. Huang, C. C., Hammond, C., Bishop, J. M. 1984. Nucleotide seadence of v-fk in 209:475-79 the PRC I1 strain of avi& sarcoma &us. J. Virol. 50:125-31 100. Hunter, E. 1978. The mechanism for genetic recombination in the avian retro- viruses. Curr. Top. Microbiol. Immunol. 101. Inoue, H., Yutsudo, M., Hakura, A. 1983. Rat mutant cells showing tempera- ture sensitivity for transformation by wild-type Moloney murine sarcoma virus. Virology 125:242-45 102. Jansen, H. W., Lurz, R., Bister, K. 1984. Homologous cell-derived onco- 791295-309 genes in avian carcinoma virus MH2 and murine sarcoma virus 3611-MSV. 103. Junghans, R. P., Boone, L. R., Skalka, A. M. 1982. Retroviral DNA H struc- tures: Displacement-assimilation model of recombination. Cell 3053-62 104. Kanda, N., Schreck, R., Alt, F., Bruns, G., Baltimore, D., Latt, S. 1983. Isola- tion of amplified DNA sequences from IMR-32 human neuroblastoma cells: Facilitation by fluorescence-activated flow sorting of metaphase chromosome. Proc. Natl. Acad. Sei. USA 80:4069-73 104a. Kataoka, T., Powers, S., McGill, C., Fasano, O., Strathem, J., et al. l'~d4. Genetic analysis of yeast RASl and RAS2 genes. Cell 37:437-45 105. Keitman, R., Deschamps, J., Couez, D., Claustriaux, J. J., Palm, R., Burny, A. 1983. Chromosome integration domain for bovine leukemia provirus in tuniors. J. Virol. 47:146-50 106. Klein, G. 1981. The ra 'le of gene dosage and genetic transpositions in carci- nogenesis. Nature 294:313-18 107. Klein, G. 1983. Specific chromosomal translocations and the genesis of B cell derived tumors in mice and men. Cell 108. Klempnauer, K. H., Bishop, 3. M. 1983. The transduction of c-rnyb into avian myeloblastosis virus: Locating the points of recombination within the cellular gene. J. Virol. 48565-72 321311-15 109. Deleted in proof 110. Klempnauer, K. H., Gonda, T. J., Bishop, J. M. 1982. Nucleotide sequence of the retrovirus leukemia gene v-myb and its cellular progenitor c-myb: The architecture of a transduced oncogene. Cell 31:453-63 11 1. Kohl, N. E., Kanda, N., Schreck, R. R., Bums, G., Latt, S. A., et al. 1984. Transposition and amplification of oncogene-related sequences in human neuroblastomas. Cell 35:359-67 112. Krontiris, T., Cooper, G. M. 1981. Transforming activity of human tumor DNAs. Proc. Natl. Acad. Sci. USA 113. Kuff, E. L., Feenstra, A., Lueders, K., Rechavi, G., Givol, D., Canaani, E. 1983. Homology between an endogenous viral LTR and sequences inserted in an activated cellular oncogene. Nature 302547-48 114. Land, H., Parada, L. F., Weinberg, R. A. 1983. Tumorigenic conversion of primary embryo fibroblasts requires at least two cooperating oncogenes. Nature 304:596-601 115. Lane, M. A., Neary, D., Cooper, G. M. 78: 1 18 1-84 CELLULAR ONCOGENES 6G.5 1982. Activation of a cellular transfonn- ing gene in tuniors induced by Abetson murine leukaemia virus. Nature 300: 116. Lane, M. A,, Sainten, A,, Cooper, G. M. 1981. Activation of related trans- forming genes in mouse and human carci- nomas. Proc. Natl. Acad. Sci. USA 785 185-89 117. Lane, M. A,, Sainten, A,, Cooper, G. M. 1982. Stage-specific transforming genes of human and mouse B- and T- lymphocyte neoplasms. Cell 28:87>80 118. Lane, M. A,, Sainten, A., Doherty, K. M., Cooper, G. M. 1984. Isolation and characterization of a stage-specific trans- forming gene, Tlym-I, from T cell lym- phomas. Proc. Natl. Acad. Sci. USA 81:2227-31 119. Leder, P., Battey, J., Lenoir, G., Mould- .- ing, C., Murphy, W., et al. 1984. Trans- locations among antibody genes in hu- man cancer. Science 222:765-71 1193. Lee, W.-H., Murphree, A. L., Ben- edict, W. F. 1984. Expression and ampli- fication of the N-myc gene in primary retinoblastoma. Nature 309:458-59 120. Lemay, G., Jolicoeur, P. 1984. Rear- rangement of a DNA sequence homolo- gous to a cell-virus junction fragment in several Moloney murine leukemia virus- induced rat thymomas. Proc. Natl. Acud. Sci. USA 81:3842 121. Lenoir, G. M., Preud-Homme, J. L., Bemhcim, A,, Berger, R. 1982. Correla- tion between immunoglobulin light chain expression and variant translocation in Burkitt's lymphoma. Nature 298:474-76 122. Deleted in proofs 123. Lerner, T. L., Hanafusa, H. 1984. DNA sequence of the Bryan high-titer strain of Rous sarcoma virus: Extent of env dele- tion and possible genealogical rela- tionship with other viral strains. J. Virol. 49:549-56 123a. Lev, Z., Leibovitz, N., Segev, O., Shi- lo, B. 2. 1984. Expression of the src and ab1 cellular oncogenes during develop- ment of Drosophila melanogaster. Mol. Cell Biol. 4:982-84 124. Levy, L. S., Gardner, M. B., Casey, J. W. 1984. Isolation of a feline leukemia provirus containing the oncogene myc from a feline lymphosarcoma. Nature 308:853-56 124a. Lin, C. R., Chen, W. S., Kruiger, W., Stolarsky, L. S., Weber, W., et al. 1984. Expression cloning of human EGF recep- tor complementary DNA: Gene ampli- fication and three related messenger RNA products in A431 cells. Science 224S43-48 659-60 125. Linial, M., Blair, D. 1982. Genetics of retroviruses. See Ref. 17, pp. 649-783 126. Linial, M., Blair, D. 1984. Genetics of retroviruses. See Ref. 18, In press 127. Little, C. D., Nau, M. h!., Carney, D. N., Gazdar, A. F., Minna, J. D. 1983. Amplification and expression of the c- myc oncogene in human lung cancer cell lines. Nature 306: 194-96 128. Lorincz, A. T., Reed. S. I. 1984. Pri- mary structure homology betv,zen the product of yeast cell division conirol gene CDC28 and vertebrate oncogenes. Na- ture 307:183-85 129. Luciw, P., Bishop, J. M., Varmus, H. E. 1983. Locati m and function of retroviral and SV40 sequences that enhance biochemical transformation after micro- injection of DNA. Cell 33:705-16 129a. Luciw, P. A,, Oppermann, H., Bishop, J. M., Varmus, H. E. 1984. Integration and expression of several molecular forms of Rous sarcoma virus DN.4 used for transfection of mouse cells. Mol. Cell Bid. 4:126&69 129b. Magrath, I., Erikson, J., Whang-Peng, J., Sieverts, H., Armstrong, G., et al. 1983. Synthesis of kappa light chains by cell lines containing an 8;22 chromosorn- al translocation derived from a male homosexual with Rurkitt's lymphoma. Science 222: 1094-98 130. Maniatis, T., Fritsch, E. F., Lauer, J., Lawn, R. M. 1980. The molecular genet- ics of human hemoglobins. Ann. Rev. Genet. 14:145-78 131. Marcu, K. B., Hams, L. J., Stanton, L. W., Erikson, J., Watt, R., Croce, C. M. 1983. Transcriptionally active c-myc oncogene is contained within NIARD, a DNA sequence associated with chromo- some translocations in B-cell neoplasia. Proc. Natl. Acad. Sci. USA 80519-23 132. Mardon, G., Varmus, H. E. 1983. Frameshift and intragenic suppressor mutations in a Rous sarcoma provirus suggest src encodes two proteins. Cell 133. Marshall, C. J., Hall, A,, Weiss, R. A. 1982. A transforming gene present in hu- man sarcoma cell lines. Nature 299: 171- 73 133a. Martin, G. S., Radke, K., Hughes, S., Quintrell, N., Bishop, J. M., et al. 1979. Mutants of Rous sarcoma virus contain- ing extensive deletions of the viral genome. Virology 9653046 134. Mathey-Prevot, B., Shibuya, M., Samarut, J., Hanafusa. H. 1984. Rever- tants and partial transformants of rat fi- broblasts infected with Fujinami sarcoma virus. J. Virol. 50:325-34 32:871-79 606 VARMUS 135. McCoy, M. S., Took, J. J., Cunning- ham, J. M., Chang, E. H., Lowy, D. R., Weinberg, R. A. 1983. Characterization of a human colonilung carcinoma oncogene. Nature 302:79-82 136. Merlino, G. T., Xu, Y. H., Ishii, S., Clark, A. J. L., Semba, K., et al. 1984. Amplification and enhanced expression of the epidermal growth factor receptor gene in A431 human carcinoma cells. Science 224:417-19 137. Miller, A. D., Curran, T., Verma, I. M. 1984. c-fos protein can induce cellular transformation: A novel mechanism of activation of a cellular oncogene. Cell 365 1-60 138. Montgomery, IC. T., Biedler, J. L., Spengler, 8. A., Melera, P. W. 1983. Specific DNA sequence amplification in human neuroblastoma cells. Proc. Natl. Acad. Sci. USA 805724-28 139. Muller, R., Slamon, D. J., Tremblay, J. M., Cline, M. J., Verma, I. M. 1982. Differential expression of cellular oncogenes during pre- and postnatal development of the mouse. Nature 299:640-43 140. Muller, R., Muller, D. 1984. Co- transfection of normal NIHi3T3 DNA and retroviral LTR sequences: a novel strategy for the detection of potential c- one genes. EMBO J. 3.1 121-27 141. Mullins, J. J., Brody, D. S., Binari, R. C. Jr., Cotter, S. M. 1984. Viral trans- duction of c-myc gene in naturally occur- ring feline leukemia. Nature 308:856-58 142. Murphree, A. L., Benedict, W. F. 1984. Retinoblastoma: Clues to human onco- genesis. Science 223:1028-33 143. Murray, M. J., Cunningham, J. M., Pa- rada, L. F., Dautry, F., Lebowitz, p., Weinberg, R. A. 1983. The HL-60 trans- forming sequence: A rus oncogene coex- isting with altered myc genes in hemato- poietic tumors. Cell 33:749-57 144. Mushinski, J. F., Potter, M., Bauer, S. R., Reddy, E. P. 1983. DNA rearrange- ment and altered RNA expression of the c-myb oncogene in mouse plasmacytoid lymphosarcomas. Science 220:795-98 144a. Naharro, G., Robbins, K. C., Reddy, E. P. 1984. Gene product of v-fgr onc: Hybrid protein containing a portion of actin and a tyrosine-specific protein kinase. Science 223:63-66 145. Nakano, H., Yamamoto, F., Neville, C., Evans, D., Mizuno, T., Perucho, M. 1984. Isolation of transforming se- quences of two human lung carcinomas: Structural and functional analysis of the activated c-K-ras oncogenes. Proc. Natl. Acad. Sci. USA 81:71-75 145a. Neel, B. G., Hayward, W. S., Robin- son, H. L., Fang, J., Astrin, s. M. 1981. Avian leukosis virus-induced tuniors have common proviral integration sites and svnthesize disxete new RNAs: ~~~ ~ . Oncogknesis by promote: insertion. Cell 23: 323-34 146. Neel, B. G., Gasic, G. P., Rogler, C. E., Skalka, A. M., Ju, G., et al. 1982. Molecular analysis of the.c-myc locus in normal tissue and in avian leukosis virus- induced lymphomas. J. Virol. 44: 158--66 147. Neil, J. C., Hughes, I)., McFarlane, R., Wilkie, N. M., Onions, D. E., et al. 1984. Transduction and rearrangement of the myc gene by feline leukemia virus in naturally occurring T-cell lymphomas. Nature 308:S14-20 148. Neuberger, M. S., Calabi, F. 1983. Re- ciprocal chromosome translocation be- tween c-myc and immunoglobulin y 2b genes. Nature 305:24&42 149. Nishikura, K., ar-Rushdi, A,, Erikson, J., DeJr.sus, E., Dugan, D., Croce, C. M. 1981. Repression of rearranged p. gene and translocated c-myc in mouse 3T3 cells X Burkitt lymphoma cell hy- brids. Science 224:399-402 150. Nishikura, K., ar-Rushdi, A,, Erickson, J., Watt, R., Rovera, G., Croce, C. M. 1983. Differential expression of the nor- mal and of the translocated human c-myc oncogenes in B cells. Proc. Natl. Acad. Sci. USA 80:4822-26 15 1. Noda, M., Selinger, Z., Scolnick, E. M., Bassin, R. H. 1983. Flat revertants iso- lated from Kirsten sarcoma virus- transformed cells are resistant to the ac- tion of specific oncogenes. Proc. Nutl. Acad. Sci. USA 805602-06 151a. Noori-Daloii, M. R., Swift, R. A,, Kung, H.-J., Crittenden, L. B., Witter, R. L. 1981. Specific integration of REV proviruses in avian bursal lymphomas. Nature 29457476 152. Norton, J. D., Cook, F., Roberts, P. C., Clewley, J. P., Avery, R. J. 1984. Ex- pression of Kirsten murine sarcoma virus in transformed nonproducer and revertant NIW3T3 cells: Evidence for cell- mediated resistance to a viral oncogene in phenotypic reversion. J. Virol. 50:43!3- 44 153. Nowell, P. C., Finan, J., Dalla-Favera, R., Gallo, R. C., ar-Rushdi, A,, et al. 1983. Association of amplified oncogene c-myc with an abnormally banded chromosome 8 in a human leukaemia cell line. Nature 306:494-97 154. Nunn, M. F., Seeberg. P. H., Moscovi- ci, C., Duesberg, P. H. 1983. Tripartite structure of the avian erythroblastosis CELLULAR ONCOGENES 607 virus E26 transfoming gene. Nature 15.5. Nusse, R., van Ooyen, A,, Cox, D., Fung, Y. K., Varmus, H. E. 1984. Mode of proviral activation of a putative mmi- mary oncogene (int-I) on mouse chromo- some 15. Nature 307:131-36 156. Nusse, R., Varmus, H. E. 1982. Many tumors induced by the mouse mammary tumor virlis contain a provirus integrated in the same region of the host genome. Cell 3 1 :99-109 157. Nusse, R., Westaway, D., Fung, Y. K., van Ooyen, A., Moscovici, C., Varmus, H. E. 1984. Oncogene activation by pro- viral insertion. In The Cancer Cell, Vol. I I, Cell Proliferation Series, Cold Spring Harbor, NY: Cold Spring Harbor Biol. Lab. 158. Oppermanii, H., Levinson, A. D., Var- mus, H. E. 1981. The structure and pro- tein kinase activity of proteins encoded by non-conditional mutants and back -- mutants in the src gene of avian sarcoma virus. Virology 108:47-70 159. Oskarsson, M., McClements, W. L., Blair, D. G., Maizel, J. V., Vande Woude, G. F. 1980. Properties of a nor- mal mouse cell DNA sequence (sarc) homologous to the src sequence of Moloney sarcoma virus. Science 207: 1222-24 160. Ozanne, B., Wheeler, T., Zack, J., Smith, G., Dale, B. 1982. Transfonning gene of a human leukaemia cell is unre- lated to the expressed tumor virus related gene of the cell. Nature 299:74447 161. Pachl, C., Biegalke, B., Linial, M. 1983. RNA and protein encoded by MH2 virus: Evidence for subgenomic expres- sion of v-myc. J. Virol. 45:133-39 162. Pachl, C., Schubach, W., Eisenman, R., Linial, M. 1983. Expression of c-myc RNA in bursal lymphoma cell lines: Identification of c-myc-encoded proteins by hybrid-selected translation. Cell 33:335-44 163. Padhy, L. C., Shih, C., Cowing, D., Finkelstein, R., Weinberg, R. A. 1982. Identification of a phosphoprotein specif- ically induced by the transforming DNA of rat neuroblastomas. Cell 285365-71 164. Parada, L. F., Tabin, C. J., Shih, C., Weinberg, R. A. 1982. Human EJ blad- der carcinoma oncogene is a homologue of Harvey sarcomavirus ras gene. Nature 297:47478 165. Parada, L. F., Weinberg, R. A. 1983. Presence of a Kirsten murine sarcoma virus ras oncogene in cells transformed by 3-methylcholanthrene. Mol. Cell Biol. 3:2298-01 306:391-95 166. Parker, R. C., Varmus, €I. E., Bishop, J. M. 1984. Expression of v-src and chick- en c-src in rat cells demonstrates qualita- tive differences between pp60"-"" arid pp60'~"". Cell 37:131-39 166a. Payne, G. S., Courtneidge, S. A,, Crit- tenden, L. B., Fadly, A. M., Bishop, J. M., et al. 1981. Analysis of avian leuko- sis virus DNA and RNA in bursal tumors: Viral gene expression is not required for maintenance of the tumor state. Cell 23:311-22 167. Payne, G. S., Bishop, J. M., Varmus, H. E. 1982. Multiple arrangements of viral DNA and an activated host oncogene (c-myc) in bursal lymphomas.. Nature 295:209-I7 167a. Pelicci, P. G.. Lankancone, L., Brath- waite, M. D., \Volman, S..R.; Dalla- Favera, R. 1984. Amplification of the c-myb oncogene in a case of human acute myelogenous leukemia, Science 224: 168. Perlmutter, R. M., Klotz, J. L.; Pravt- cheva, D., Ruddle, F., Hood, L. 1984. A novel 6: IO chromosomal translocation in the murine plasmacytoma NS-1. Nature 169. Perry, R. P. 1983. Consequences of myc invasion of immunoglobulin loci: Facts and speculations. Cell 33:64749 170. Peters, G., Brookes, S., Smith, R., Dickson, C. 1983. Tumorigenesis by mouse mammary tumor virus: Evidence for a common region for provirus integra- tion in mammary tumors. Cell 33:369- 71 170a. Peters, G., Lee, A. E., Dickson, C. 1984. Activation of cellular gene by mouse mammary tumour virus may occur early in mammary tumour development. Nature 309:273-75 170b. Peterson, T. A,, Yochem, J., Byers, B., Nunn, M. F., Duesberg, P. H., et al. 1984. A relationship between the yeast cell cycle genes CDC4 and CDC36 and the ets sequence of oncogenic virus E26. Nature 309:556-57 70c. Piccoli, S. P., Caimi, P. G., Cole, M. D. 1984. A conserved sequence at c-myc oncogene chromosomal translocation breakpoints in plasmacytomas. Nature 3 10:327-30 71. Pincus, M. R., van Renswoude, J., Har- ford, J. B., Chang, E. H., Carty, R. p., Klausner, R. D. 1983. Prediction of the three-dimensional structure of the trans- forming region of the EJ/T24 human bladder oncogene product and its normal cellular homologue. Proc. Natl. Acad. Sci. USA 80:5253-57 172. Ponder, P. A. J. 1980. Genetics andcan- 11 17-20 307:473-75 608 VARMUS cer. Biochim. Biophys. Acra 605:369- 410 172a. Powers, S., Kataoka, T., Fasano, O., Goldfarb, M., Strathem, J. et al. 1984. Genes in S. cerevisiae encoding proteins with domains homologous to the mammalian ras proteins. Cell 36:607-12 173. Prywes, R., Foulkes, J. G., Rosenberg, N., Baltimore, D. 1983. Sequences of the A-MuLV protein needed for fibro- blast and lymphoid cell transformation. Cell 34:569-79 174. hlciani, S., Santos, E., Lauver, A. V., Long, L. K., Barbacid, M. 1982. Oncogenes in solid human tumors. Nu- ture 300:539-42 174a. Rabbitts, T. H., Forster, A,, Hamlyn, P., Baer, R. 1984. Effect of somatic mutation within translocated c-myc genes in Burkitt's lymphoma. Nature 309:592- 97 175. Rabbitts, T. H., Forster, A,, Matthews, J. G. 1983. The breakpoint of the Phil- adelphia chromosome 22 in chronic myeloid leukaemia is distinct to the im- munoglobulin lambda light chain con- stant region genes. Mol. Biol. Med. 1: 11-19 176. Rabbitts, T. H., Hamlyn, P. H., Baer, R. 1983. Altered nucleotide sequences of a translocated c-myc gene in Burkitt lym- phoma. Nature 306:760-65 177. Ramsay, G. M., Enrietto, P. J., Graf, T., Hayman, M. J. 1982. Recovery of myc- specific sequences by a partially trans- formation-defective mutant of avian myelocytomatosis virus, MC29, corre- lates with the restoration of transforming activity. Proc. Natl. Acad. Sci. USA 178. Rapp, U. R., Todaro, G. J. 1978. Gen- eration of new mouse sarcoma viruses in cell culture. Science 201:821-23 179. Rapp, U. R., Todaro, G. J. 1980. Gen- eration of oncogenic mouse type-C viruses-in vitro selection of carcinoma- inducing variants. Proc. Natl. Acad. Sei. USA 77:624-28 180. Rasheed, S., Gardner, M. R., Huebner, R. J. 1978. In vitro isolation of stable rat sarcoma viruses. Proc. Natl. Acad. Sci. 181. Rasheed, S., Norman, G. L., Heidecker, G. 1983. Nucleotide sequence of the Rasheed rat sarcoma virus oncogene: New mutations. Science 221: 155-57 182. Rassoulzadegan, M., Cowie, A,, Can, A., Glaichenhous, N., Kamen, R. 1982. The role of individual polyoma virus ear- ly proteins in oncogenic transformation. Nature 300:713-18 183. Razzaque, A,, Mizusawa, H., Seidman, 79:6885-89 USA 7512972-76 M. M. 1983. Rearrangement and mutagenesis of a shuttle vector plasmid after passage in mammalian cells. Proc. Natl. Acad. Sci. USA 80:3010-14 184. Rechavi, G., Givol, D., Canaani, E. 1982. Activation of a cellular oncogene by DNA rearrangement: Possible in- volvement of an IS-like element. Nature 300:607-10 185. Reddy, E. P. 1983. Nucleotide sequence analysis of the T24 human bladder carci- noma oncogene. Science 220: 1060-63 186. Reddy, E. P., Reynolds, R. K., Santos, E., Barbacid, M. 1982. A point mutation is responsible for the acquisition of trans- forming properties of the T24 human bladder carcinoma oncogene. Nature 187. Renger, H. C., Basilico, C. 1972. Muta- tion causing temperature sensitive ex- pression of cell transformation by a tumor virus. Proc. Natl. Acad. Sci. USA 69: 188. Roninson, I. B. 1983. Detection and mapping of homologous, repeated, and amdified DNA seauences bv DNA re- 3001149-52 109-16 nakration in agaroie gels. kucl. Acids Res. 11:5413-31 189. Rowley, J. D. 1980. Chromosome abnormalities in human leukemia. Ann. Rev. Genet. 14:17-39 190. Rowley, J. D. 1982. Identification of the constant chromosome regions involved in human hematologic malignant disease. Science 216:749-51 191. Rowley, J. D. 1983. Human oncogene locations and chromosome aberrations. Nature 301:29&91 191a. Rubin, H. 1984. Mutations and oncogenes--causc or effect. Nature 309:5 18 192. Ruley, H. E. 1983. Adenovirus early re- gion 1A enables viral and cellular trans- forming genes to transform primary cells in culture. Nature 304:602-06 193. Sager, R., Tanaka, K., Lau, C. C., Ebi- na, Y., Anisowicz, A. 1983. Resistance of human cells to tumorigenesis induced by cloned transforming genes. Proc. Natl. Acad. Sci. USA 80:7601-05 194. Saito, H., Hayday, A. C., Wiman, K., Hayward, W. S., Tonegawa, S. 1983. Activation of the c-myc gene by trans- location: A model for translational con- trol. Proc. Natl. Acad. Sci. USA 80: 7476-80 195. Santos, E., Martin-Zanca, D., Reddy, E. P., Pierotti, M. A., Della Porta, G., Bar- bacid, M. 1984. Malignant activation of a K-ras oncogene in lung carcinoma but not in normal tissue of the same patient. Science 223:661-64 CELLIJLAR ONCOGENES GO9 196. Santos, E., Reddy, E. P., Pulciani, S., Feldman, R. J.. Barbacid, M. 1983. Spontaneous activation of a human proto- oncogene. Proc. Natl. Acad. Sei. USA 80:4679-83 197. Santos, E., Tronick, S. R., Aaronson, S. A,, Pulciani, S., Barbacid, M. 1982. T24 human bladder carcinoma oncogene is an activated form of the normal human homologue of BALB and Harvey-MSV transforming genes. Nature 298:343- 47 197a Schimke, R. T., ed. 1982. GeneAmpli- )cation. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory 198. Schwab, M., Alitalo, K., Klempnauer, K. H., Varmus, H. E., Bishop, J. M., et al. 1983. Amplified DNA with limited homology to the myc cellular oncogene. Nature 305:245-48 199. Schwab, M., Alitalo, K., Vamus, H. E., Bishop, J. M., George, D. 1983. A -cellular oncogene (c-Ki-ras) is ampli- fied, over-expressed, and located within karyotypic abnormalities in mouse adre- nocortical tumor cells. Nature 303:497- 50 1 200. Schwab, M., Ellison, J., Busch, M., Rosenau, W., Varmus, H. E., Bishop, J. M. 1984. Enhanced expression of the hu- man gene N-myc consequent to ampli- fication of DNA may contribute to malig- nant progression of neuroblastoma. Proc. Natl. Acad. Sci. USA 81:494044 201. Schwab, M., Varmus, H. E., Bishop, J. hf., Grzeschik, K. H., Naylor, S. L., et al. 1984. Chromosome localization in normal human cells and neuroblastomas of a gene related to c-myc. Nature 308:288 91 202. Schwartz, D. E., Tizzard, R., Gilbert, W. 1983. Nucleotide sequence of Rous sarcoma virus. Cell 32:853-69 203. Sealy, L., Moscovici, G., Moscovici, C., Bishop, J. M. 1983. Site-specific mutagenesis of avian erythroblastosis virus: v-erb-A is not required for trans- formation of fibroblasts. Virology 130: 179-94 204. Sealy, L., Moscovici, G., Moscovici, C., Privalsky, M., Bishop, J. M. 1983. Site-specific mutagenesis of avian erythroblastosis virus: erb-B is required for oncogenicity. Virology 130: 155-78 204a. Seiki, M., Eddy, R., Shows, T. B., Yoshida, M. 1984. Nonspecific integra- tion of the HTLV provirus genome into adult T-cell leukemia cells. Nature 309:64&42 205. Shen-Ong, G. L., Keath, E. J., Piccoli, S. P., Cole, M. D. 1982. Novel myc oncogene RNA from abortive immuno- globulin-gene recombination in mouse plasmacytomas. Cell 31:443-52 206. Shibuya, M., Hanafusa, T., Hanafiusa, H., Stephenson, R. 1980. Homology ex- ists among the transforming sequences of avian and feline sarcoma viruses. Proc. Natl. Apad. Sei. USA 77:6536-40 207. Shih, C. K., Linial, M., Goodenow, M. M., Hayward, W. S. 1984. Nucleotide sequence 5' of the chicken c-myc coding region: Localization of a non-coding exon that is absent from myc transcripts in mod ALV-induced lymphomas. Proc. Natl. Acad. Sci. 81: 4697-701 207a. Shih, C., Shilo, B.-Z., Goldfarb, M. P., Dannenberg, A,, Weinberg, R. A. 1979. Pkssage of phenotypes of chemi- cally transformed cells via transfection of DNA and chromatin. Proc. Nafl. Acad. Sei. USA 76:5714-18 208. Shih, C., Padhy, L. C., Murray, M., Weinberg, R. 1981. Transforming genes of carcinomas and neuroblastomas intro- duced into mouse fibroblasts. Narkre 209. Shimizu, K., Birnbaum, D., Ruley, M. A,, Fasano, O., Suard, Y., Edlund, L., et al. 1983. Structure of the Ki-ras gene of the human lung carcinoma cell linc Calu-1 . Nature 304:497-500 210. Shimizu, K., Goldfarb, M., Suard, Y., Pemcho, M., Li, Y., et al. 1983. Three human transforming eenes are related to 290:261-63 YY the >iral ras oncogenes. Proc. Natl. Acad. Sci. USA 80:2112-16 2 11. Simon, A. E., Taylor, M. W., Bradley, W. E. C. 1983. Mechanism of mutation at the aprt locus in Chinese hamster ovary cells: Analysis of heterozygotes and hemizygotes. Mol. Cell Bioi. 3:1703-10 212. Simon, M. A,, Kornberg, T. B., Bishop, J. M. 1983. Three loci related to the src oncogene and tyrosine-specific protein kinase activity in Drosophila. Nature 213. Slamon, D. J., deKernion, J. B., Verma, I. M., Cline, M. J. 1984. Expression of cellular oncogenes in human malignan- cies. Science 224:256-62 213a. Souyri, M., Fleissner, F. 1983. Identi- fication by transfection of transforming sequences of human T-cell leukemias. Proc. Natl. Acad. Sei. USA 80:6676- 79 214. Stanbridge, E. J., Der, C. J., Doersen, C. J., Nishimi, R. Y., Peehl, D. M., et al. 1982. Human cell hybrids: Analysis of transformation and tumorigenicity. Science 215:252-60 215. Stanton, L. W., Watt, R., Marcu, K. B. 1983. Translocation, breakage and mn- cated transcripts of c-myc oncogene in 302x837-39 610 216. 217. 218. 219. 220 22 1 VARMUS murine plasmacytomas. Narure 303: 401-05 Stanton, L. W., Yang, J. Q., Eckhardt, L. A., Harris, L. J., Birshtein, B. K., Marcu, K. B. 1984. Products of a rccip- rocal chromosome translocation involv- ing the c-myc gene in a murine plasmacy- toma. Proc. Natl. Acad. Sci. USA 81: Steffen, D. 1984. Proviruses are adjacent to c-myc in some murine leukemia virus- induced lymphomas. Proc. Natl. Acad. Sci. USA 81:2097-01 Stehclin, D., Varmus, H. E., Bishop, J. M., Vogt, P. K. 1976. DNA related to the transforming gene(s) of avian sarco- ma viruses is present in normal avian DNA. Nature 260:170-73 Steinberg, B., Pollack, R., Topp, W., Eotchan, M. 1978. Isolation and charac- terization of T antigen-negative rever- tants from a line of transformed rat cells containing one copy of the SV40 genome. Cell 13:19-32 Stephenson, J. R., Reynolds, R. K., Aaronson, S. A. 1973. Characterization of morphological revertants of murine and avian sarcoma virus-transformed cells. J. Virol. 11:218-22 Streisinger, G., Okada, Y., Enrich, J., Newton, J., Tsugita, A., Terzaght, E., Inouye, M. 1966. Frameshift mutations and the genetic code. Cold Spring Har- bor Symp. Quant. Bid. 31:77-84 829-33 222. SukumG, S., Notario, V., Martin- Zanca, D., Barbacid, M. 1983. Induction of mammary carcinomas in rats by nitro- so-methylurea involves malignant activa- tion of H-ras-1 locus by single point mutations. Nature 306:658-61 223. Sukumar, S., hlciani, S., Doniger, J., DiPaolo, J. A,, Evans, C. H., et al. 1984. A transforming ras gene in tumorigenic guinea pig cell lines initiated by diverse chemical carcinogens. Science 223: 1 197-99 224. Sumegi, J., Spira, J., Bazin, H., Szpirer, J., Levan, G., Klein, G. 1983. Rat c-myc oncogene is located on chromosome 7 and rearranges in immunocytomas with t(6;7) chromosomal translocation. Na- ture 306:497-98 225. Sutrave, P., Bonner, T. I., Rapp, U. R., Jansen, H. W., Patschinsky, T., Bister, K. 1984. Nucleotide sequence of avian retroviral oncogene v-mil: Homologue of murine retroviral oncogene v-raf. Nature 309:85-88 226. Swanstrom, R., Parker, R. C., Varmus, H. E., Bishop, J. M. 1983. Transduction of a cellular oncogene: The genesis of Rous sarcoma virus. Proc. Natl. Acad. Sci. USA 80:2519-23 227. Tabin, C. J., Bradley, S. M., Bargmann, C. I., Weinbcrg, R. A,, Papagcorge, A. G., et al. 1982. Mechanism of activation of a human oncogene. Nature 300:143- 49 228. Takeya, T., Hanafusa. H. 1983. Struc- ture and sequence of the cellular gene homologous to the RSV src gere and the mechanism for generating the transfomi- ing virus. Cell 32:881-90 228a. Tainsky, M. A,, Cooper, C. S., Giovanella, B. C., Vande Woude, G. F. 1984. An activated rasN gene is detected in late but not early passage human PA1 teratocarcinoma cells. Nature In press 229. Taparowsky, E., Shimizu, K., Goldfarb, M., Wigler, M. 1983. Structure and activation of the human N-ras gene. Cell 34581-86 230. Taparowsky, E., Suard, Y., Fasano, O., Shimizu, K., Goldfarb, M. P., Wigler, M. P. 1982. Activation of T24 bladder carcinoma transforming gene is linked to a single amino acid change. Nature 300:762-65 230a. Tatchell, K., Chaleff, D. T., DeFeo- Jones, D., Scolnick, E. M. 1984. Re- quirement of either of a pair of ras-related genes of Saccharomyces cerevisiae for spore viability. Nature 309523-27 230b. Taub, R., Kelly, K., Battey, J., Latt, S., Lenoir, G. M., et al. 1984. A novel alteration in the structure of an activated c-myc gene in a variant t(2;8) Burkitt lymphoma. Cell 3751 1-20 231. Taub, R., Kirsch, I., Morton, C., Lenoir, G., Swan, D., Tronick, S., Aaronson, S., Leder, P. 1982. Trans- location of the c-myc gene into the im- munoglobulin heavy chain locus in hu- man Burkitt lymphoma and murine plas- macytoma cells. Proc. Natl. Acad. Sci. USA 79:783741 232. Taub, R., Moulding, C., Battey, J., Murphy, W., Vasicek, T., Lenoir, G. M., Leder, P. 1984. Activation and somatic mutation of the translocated c- myc gene in Burkitt lymphomacells. Cell 36:339-48 233. Teich, N., Wyke, J., Mak, T., Bem- stein, A,, Hardy, W. 1982. Pathogenesis of retrovirus-induced disease. See Ref. 17, pp. 785-998 234. Tonegawa, S. ed. 1983. Somatic genera- tion of antibody diversity. Narure 302:575-81 234a. Tooze, J. 1980. Molecular Biology of Tumor Viruses. DNA Tumor Viruses. Cold Spring Harbor, NY: Cold Spring Harbor Biol. Lab. 235. Tsichlis, P. N., Hu, L. F., Strauss, P. G. CELLIJLAK ONCOGENES 61 1 1983. Two common regions for proviral DNA integration in MoMuLV induced rat thymic lymphomas. Implications for oncogenesis. UCLA Symp. Mol. Celi. Biol. 9:399-415 236. Tsichlis, P. N., Strauss, P. G., Hu, L. F. 1983. A common region for proviral DNA integration in MoMuLV-induced rat thymic lymphomas. Nature 302:445- 48 236a. Tsichlis, P. N., Strauss, P. G., Koz,ak, C. A. 1984. Cellular DNA region in- volved in induction of thymic lympho- mas (Mlvi-2) maps to mouse chromo- some 15. Mol. Cell. Bid. 4:997-1000 237. Tsuchida, N., Ryder, T., Ohtsubo, E. 1982. Nucleotide sequence of the oncogene encoding the p21 transforming protein of Kirsten murine sarcoma virus. Science 217:937-38 237a. Tsujimoto, Y., Yunis, I., Onorato- .. Shoew, L., Erikson, J., Nowell, P. C. et al. 1984. Molecular cloning of the chro- mosomal breakpoint of B-cell lympho- mas and leukemias with the t(11;14) chromosomal translocation. Science 224: 237b. Ullrich, A,, Coussens, L., Hayflick, J. S., Dull, T. J., Gray, A. et al. 1984. Human epidermal growth factor receptor cDNA sequence and aberrant expression of the amplified gene in A431 epidemioid carcinoma cells. Nature 309:418-24 238. Urlaub, G., Kas, E., Carothers, A. M., Chasin, L. A. 1983. Deletion of the dip- loid dihydrofolate reductase locus from cultured mammalian cells. Cell 33:405- 12 1403-06 242. 243. 239. Van Beveren, C. 1984. Sequences of retroviral genomes. See Ref. 18: In press 240. Van Beveren, C., Enami, S., Curran, T., Verma, I. M. 1984. FBR murine osteosarcoma virus. 11. Nucleotide se- quence of the provirus reveals that the genome contains sequences from two cellular genes. Virology 135:229-43 241. Van Beveren, C., van Straaten, F., Cur- ran, T., Muller, R., Verma, I. M. 1983. Analysis of FBJ-MuSV provirus and c- fos (mouse) gene reveals that viral and cellular fos gene products have different carboxy termini. Cell 32:1241-55 Van Ness, B. G., Shapiro, M., Kelley, D. E.. Pew. R. P.. Weimt. M.. D'Eus- tachio, P.,. Ruddle, F.-1983. Aberrant rearrangement of the K light chain locus involving the heavy chain locus and chromosome 15 in a mouse plasmacyto- ma. Nature 301:425-27 Varmus, H. E. 1982. Recent evidence for oncogenesis by insertion mutagenesis and gene activation. CancerSurv. 2:301- 19 244. Varmus, H. E. 1982. Form and function of rctroviral proviruses. Science 216: 8 12-20 244a. Varmus, H. E. 1983. Retroviruses. In Mobile GeneticElements, ed. J. Shapiro, pp. 411-503. New York: Academic 245. Varmus, X. E., Quinrrell, N., Ortiz, S. 1981. Retroviruses as mutagens: Inszr- tion and excision of a non-transforming provirus alters expression of a resident transforming provirus. Cell 25:23-36 246. Varmus, H. E., Quintrell, N., Wyke, J. 1981. Revertants of an ASV-transformed rat cell line have lost the complete pro- virus or sustained mutations in src. Virol- ogy 108:2%46 246a. Vousden, K. H., Marshall, C. J. 1984. Three different activated rus genes in mouse tumors; evidence for oncogene activation during progression of a niouse lymphoma. EMBO J. 3:913-17 247. Wahl, G. M., de St. Vincent, B. R., DeRose, M. L. 1984. Effect of chromo- somal position on amplification of trans- fected genes in animal cells. Nafure 307:s 16-21 248. Wang, J. Y. J. 1983. Fromc-abltov-abl. Nature 304:400 249. Wadg, L. H., Reck,:on, M., Anderson, S. M., Hanafusa, H. 1984. Identification of the viral sequence required for the gen- eration of recovered avian sarcoma viruses and characterization of a series of replication-defective recovered avian sarcoma viruses. J. Virol. 49:881-91 250. Wang, L. H., Halpem, C. C., Nadel, M., Hanafusa, H. 1978. Recombination between viral and cellular sequence generates transforming sarcoma virus. Proc. Natl. Acad. Sci. USA 755812-16 251. Waterfield, M. D., Scrace, G. J., Whit- tle, N., Stroobant, P., Johnson, A. et al. 1983. Platelet-derived growth factor is structurally related to the putative trans- forming protein ~28''' of simian sarcoma virus. Nature 304:35-39 252. Watson, D. K., Reddy, E. P., Duesberg, P. H., Papas, T. S. 1983. Nucleotide sequence analysis of the chicken c-myc gene reveals homologous and unique coding regions by comparison with the transforming gene of avian myelocyto- matosis virus MC29, delta-gag-myc. Proc. NatI. Acad. Sci. USA 80:2146- 50 253. Weinberg, R. A. 1982. Oncogenes of spontaneous and chemically induced tumors. In Advances in Cancer Re- search, 36:149-64. New York: Aca- demic 612 VARMUS 254. Weinberg, R. A. 1983. A molecular basis of cancer. Sci. Am. 249:12&43 255. Weiner, F., Ohno, S., Babonits, M., Sumegi, J., Wirschubsky, Z., et al. 1984. Hemizygous interstitial deletion of chromosome 15 (band D) in three trans- location-ncgative murine plasmacyto- mas. Proc. Natl. Acad. Sci. USA 256. Westaway, D., Payne, G., Varmus, H. E. 1984. Deletions and base substitutions in provirally mutated c-myc alleles may contribute to the progression of B-cell tumors. Proc. Natl. Acad. Sci.. USA 51:843-47 257. Westin, E. H., Wong-Staal, F., Gel- mann, E. P., Favera, R. D., Papas, T. s., et al. 1982. Expression of cellular honio- logues of retroviral onc genes in human hematopoietic cells. Proc. Natl. Acad. Sei. USA 79:2490-94 258. Wierenga, R. K., Hol, W. G. J. 1983. Predicted nucleotide-binding properties 81 : 1159-63 of p21 protein and its cancer-acsochted variant. Nature 302:842-44 259. Wyke, J. A,, Beamand, J. A,, Vannus, H E. 1980. Factors affecting phenotypic reversion of rat cells transfonned by avian sarcoma virus. Coldspring Harbor Symp. Quant. Biol. 44:1065-75 259a. Yuasa, Y., Gol, R. A,, Chang, A., Chin, I.-M., Reddy, E. P. et al. 1984. Mechanism of activation of an N-ras oncogene of SW-1271 human lung carci- noma cells. Proc. Nail. Acad. Sei. USA 81:3670-74 260. Yuasa, Y., Srivastava, S. K., Dunn, C. Y., Rhim, J. S., Reddy, E. P., Aaron- son, S. A. 1983. Acquisition of trans- forming properties by alternative point mutations within c-baslhas human neo- plasia. Nature 303:775-79 261. Yunis, J. J. 1983. The chromosomal basis of human neoplasia. Science 221:227-35