Transformation by Simian Virus 40 and Polyoma Virus

0 downloads 0 Views 8MB Size Report
Abortive transformation was first described by Stoker (1968), who infected BHK-21 cells in suspension with polyoma virus and plated them in suspension in ...
4 Transformation by Simian Virus 40 and Polyoma Virus

Cell division is controlled in multicellular organisms. Tumors grow because cancer cells continue to multiply in conditions that regulate the multiplication of normal cells. Polyoma virus and SV40 can effect a change in the growth control of a normal cell; as a result, these viruses can cause primary tumors when injected into susceptible animals. They can also cause changes in the growth control of cells growing in vitro; they transform the cells. Such transformation is believed to be analogous to the induction of primary tumors in animals, because cells transformed in vitro by polyoma virus or SV40 acquire a set of properties (see Table 4.1), some of which (including increased malignancy) are general characteristics of tumor cells. Polyoma virus and SV40 are not oncogenic in their natural host species. They must be injected in large amounts into susceptible rodents to induce tumors, and even then, the tumors do not usually metastasize. All this is also true of the tumors that develop when cells transformed in vitro by these viruses are injected into susceptible host animals. These reservations notwithstanding, transformation of cultivated cells by tumor viruses is the best model system we have for studying, in a quantitative way, at least some of the several cellular events that lead to the development of primary tumors in natural populations of animals and man. In addition, of the transforming viruses available, polyoma viruses and SV40 have the smallest genomes, and the complete nucleotide sequences of their genomes have been determined (see Appendixes A and B). The central question we have to answer is. How does a small amount of new genetic 205 DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

206 Table

DNA Tumor Viruses 4.1

Properties of cells transformed by S V 4 0 or p o l y o m a virus

Growth high or indefinite saturation density" different, usually r e d u c e d , s e r u m requirement" g r o w t h in agar o r M e t h o c e l s u s p e n s i o n — a n c h o r a g e i n d e p e n d e n c e " tumor formation upon injection into susceptible animals not susceptible to contact inhibition of m o v e m e n t g r o w t h in a l e s s - o r i e n t e d m a n n e r " g r o w t h o n m o n o l a y e r s of n o r m a l cells" Surface i n c r e a s e d a g g l u t i n a b i l i t y of p l a n t lectins" c h a n g e s in c o m p o s i t i o n of g l y c o p r o t e i n s a n d g l y c o l i p i d s tight j u n c t i o n s m i s s i n g fetal a n t i g e n s r e v e a l e d virus-specific transplantation antigen different s t a i n i n g p r o p e r t i e s i n c r e a s e d r a t e of t r a n s p o r t of n u t r i e n t s i n c r e a s e d s e c r e t i o n of p r o t e a s e s o r activators" Intracellular d i s r u p t i o n of t h e c y t o s k e l e t o n c h a n g e d a m o u n t s of c y c l i c n u c l e o t i d e s E v i d e n c e of v i r u s virus-specific antigenic proteins detectable viral D N A s e q u e n c e s d e t e c t e d viral m R N A p r e s e n t v i r u s c a n b e r e s c u e d in s o m e c a s e s Transformed cells show many, if not all, of these properties, which are not shared by untransformed parental cells. "Several of these properties have formed the basis of selection procedures for isolating transformants.

information, only part of the infecting viral genome, induce stable changes in the growth control of the host cell?

TUMORIGENICITY OF POLYOMA VIRUS AND SV40 Polyoma Virus

Under normal circumstances, in the natural host species neither polyoma virus nor SV40 behaves as a tumor virus. Polyoma DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

4 / Transformation by SV40 and Polyoma Virus

207

virus is ubiquitous in laboratory mouse colonies (Yabe et al. 1961) and in populations of wild mice (Rowe 1961), but it has no apparent etiological role in the induction of naturally occurring mouse tumors. On the other hand, polyoma virus was so named

(Eddy et al. 1958) because Gross (1953), who discovered the virus, and Stewart and Eddy and their colleagues in the 1950s found that the virus induces in newborn mice of various laboratory strains (and also in newborn hamsters, rats, and rabbits) tumors in a wide variety of organs, including salivary glands, thymus, ovaries, mammary glands, and adrenals. Moreover, polyoma virus can induce tumors when injected at high concentrations into adult animals. After long latent periods, the tumors usually develop at the site of injection (Defendi 1960), but occasionally they occur elsewhere. Polyoma virus DNA also induces tumors if large amounts are injected into newborn

hamsters (Israel et al. 1980).

SV40

SV40 has a much more restricted oncogenic potential than polyoma virus. It is not tumorigenic in the natural host (rhesus monkeys), where it is carried in high concentrations in the kidneys seemingly without deleterious consequences (Sweet and Hilleman 1960). Apparently, it is not oncogenic in humans. About 20 years ago, adenovirus and poliovirus vaccines inadvertently contaminated with SV40 (Sweet and Hilleman 1960) in amounts sufficient to elicit an immune response (Shah 1972) were injected into millions of children in schools and adults in the armed forces, but so far, no tumors attributable to SV40 have resulted (but see Heinonen et al. 1973). To date, the only human disease with which SV40 has been associated is one case of malignant melanoma (Soriano et al. 1974). SV40 is not oncogenic in mice. For example, it was injected into isolated mouse blastocysts, which were then reimplanted into foster mothers and allowed to develop to term. The resulting mice carried SV40 DNA sequences in the cellular chromosomes of all the tissues tested, but these were not associated with an increased incidence of tumors (Jaenisch and Mintz 1974). Indeed, the only animals for which SV40 is DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

208

DNA Tumor Viruses

significantly tumorigenic are hamsters, even under experimental conditions designed to maximize induction (Eddy et al. 1961). As with polyoma virus, newborn hamsters are particularly susceptible, developing gliomas (Gerber and Kirschstein 1962) and subcutaneous fibrosarcomas (Girardi et al. 1963; Girardi and Hilleman 1964) following subcutaneous injection. Adult animals develop tumors at the site of injection, and only after prolonged incubation periods (Allison et al. 1967), unless very high doses of the virus (10 plaque-forming units [pfu]) are injected intravenously. In the latter case, leukemias, lymphomas, osteosarcomas, and reticulum-cell sarcomas develop (Diamandopoulos 1972, 1978; Diamandopoulos and McLane 1975; for review, see Weil 1978). Apparently SV40, like polyoma virus, can induce tumors in a multiplicity of tissues, as long as a high dose reaches the tissues. In summary, polyoma virus and SV40 do not cause malignancies in their natural host species in natural conditions but their oncogenic potential can be revealed by injecting high concentrations of virus into either foreign hosts or natural hosts lacking a functional immune system. However, as we shall see, when tumor cells induced by these viruses are explanted into tissue culture, they closely resemble cells transformed in vitro by these viruses. Conversely, when cells transformed in vitro by SV40 and polyoma viruses are injected into susceptible animals, the tumors that develop resemble those induced in vivo. These similarities nourish the belief that studies of viral transformation of cultivated cells are relevant to tumorigenesis in vivo. 8

IN VITRO TRANSFORMATION Discovery of Transformation

Transformation of cells in vitro was discovered by Earle et al. (1943a), who cultured mouse fibroblasts for prolonged periods in plasma clots on cellophane in the presence not of a virus, but of the potent chemical carcinogen 20-methylcholanthrene. Eventually, cells emerged that could be distinguished either because of their increased retractility or because their shape had changed from that typical of fibroblasts to that of epithelial cells. These DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

4 / Transformation by SV40 and Polyoma Virus

209

"transformed" cells had enhanced tumorigenicity in syngeneic hosts (Earle et al. 1943b). Similar morphological transformations occurred spontaneously in control cultures not exposed to the carcinogen, but only after much longer times, up to a year or more.

Viral Transformation In Vitro

In the mid-1950s, Rous sarcoma virus was found to induce morphological changes in primary chick embryo fibroblasts and to extend the life of the cells in culture (Lo et al. 1955; Manaker and Groupe 1956). This viral transformation occurred much more quickly than the chemical and spontaneous transformations discovered by Earle et al. (1943a), happening within days of infection rather than after months or years. Most cells in the population were affected, and the overall changes in the cells' morphology and growth characteristics were strikingly similar to those observed by Earle. Shortly thereafter, polyoma virus became the first DNA virus shown to produce a transformation in the properties of cultured cells (Dawe and Law 1959). Cells that would ordinarily grow poorly in vitro, or that would normally enter crisis and die, were given extended life in culture (Dawe and Law 1959). This change was accompanied by an increased rate of cell growth and the acquisition of tumorigenicity (Vogt and Dulbecco 1960). These early studies utilized primary fibroblasts derived from embryonic tissue. After transformation, their highly ordered, whorllike pattern of growth became disorganized, with extensive overlapping and piling-up of cells. Cells continued to proliferate until layers of cells accumulated and the normally elongated, wellspread cells adopted a highly refractile, spindle shape. Although it was known that the altered properties of transformants were not due to selection of variants preexisting in a population (Sanford et al. 1954), it was very difficult to quantitate the changes in cellular behavior because primary cultures are a mixed population of several cell types. The development of clonable cell lines, like BHK-21, susceptible to polyoma virus transformation (MacPherson and Stoker 1962; Stoker 1962; Stoker and Abel 1963), and Swiss 3T3, susceptible to transDNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

210

DNA Tumor Viruses

formation by SV40 (Todaro et al. 1964), filled the need for a well-characterized, homogeneous population of host cells. Thus the decade from 1964 to 1974 was spent in careful description of the way in which SV40 and polyoma virus alter the behavior of such clonable, transformable cells.

Phenotype of Transformed Cells

Cell lines are maintained in culture by periodic replacement of medium and serum. Populations of untransformed cells divide while they remain sparse, but as the cell density and the number of cell-to-cell contacts increase, the growth rate of the culture dramatically decreases. The cells of some lines stop dividing altogether once they form a confluent monolayer. If such cells are infected with tumor viruses, some of them may be transformed; the transformants no longer respond to regulatory factors that control the multiplication of untransformed cells in culture. As a result, the transformants continue to multiply under conditions that severely reduce the rate of multiplication of untransformed cells (see Fig. 4.1). The differences between transformed cells and untransformed cells can be classified into three interrelated groups concerned with (1) changes in patterns of cell growth, (2) changes at the cell surface, and (3) changes in the intracellular population of macromolecules resulting from the presence of the viral genome (see Table 4.1). Transformed cells can be isolated from untransformed cells in mixed populations by manipulating the conditions of culture so that only cells possessing one or more of the properties listed in Table 4.1 can survive and/or multiply. Frequently, transformants can be isolated simply because they look different. A transformed cell always exhibits many of these properties, but seldom exhibits the complete set. For example, a cell selected because it can grow in low concentrations of serum may not be able to grow to a high or indefinite saturation density (Smith et al. 1971), and a cell selected because it can grow on top of a monolayer of parental untransformed cells may not be able to grow suspended in agar. Clearly, each of the numerous changes listed in Table 4.1 cannot be caused by a different virus-coded DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

4.1

T r a n s f o r m a t i o n o f B H K - 2 1 c e l l s b y p o l y o m a v i r u s , ( a ) C o l o n y i l l u s t r a t i n g t h e r e g u l a r p a r a l l e l a r r a n g e m e n t of t h e

d

h e m a t o x y l i n a n d e o s i n . ( P h o t o g r a p h s c o u r t e s y o f M . G . P. S t o k e r . )

p o l y o m a v i r u s , i l l u s t r a t i n g t h e r a n d o m o r i e n t a t i o n o f m o r e r o u n d e d c e l l s , (d)

S e c t i o n o f t h e c o l o n y s h o w n in c, s t a i n e d

with

e l o n g a t e d f i b r o b l a s t i c c e l l s , (b) S e c t i o n o f t h e c o l o n y s h o w n i n a, s t a i n e d w i t h h e m a t o x y l i n a n d e o s i n . (c) B H K - 2 1 c e l l s t r a n s f o r m e d b y

Figure

b

212

DNA Tumor Viruses

protein, because polyoma virus and SV40 contain only a small amount of genetic information. Since no more than three virus-coded proteins appear to be involved in polyoma transformation (and no more than two for SV40), most of the observed alterations in transformed cells must be either pleiotropic or indirect responses of the cell to the virus. What we would like to discover is which of the changes concomitant with transformation are direct responses of the cell to the viral gene products and which, if any, of the changes observed in vitro are directly correlated with increased tumorigenicity in vivo.

Serum Growth Factors

As mentioned above, when most types of untransformed cells are plated in vitro, they grow to a certain density and then either stop dividing or divide at a greatly reduced rate. In the same conditions, the corresponding transformed cells continue to multiply and may reach saturation densities 10-25 times greater than those of the untransformed cells, when nutritional factors probably become limiting. At one time it was thought that the growth of cells in tissue culture was controlled solely by "contact inhibition" (Stoker and Rubin 1967), the individual cells responding to the proximity of other cells by ceasing to multiply. However, it is now clear that although cell-to-cell contact plays a role in inhibiting cell division (Dulbecco 1970; Burger and Noonan 1970) as well as cell movement (Abercrombie and Heaysman 1954), the supply of factors in serum also has a part in regulating the extent to which cells in culture multiply. The addition of a fresh dose of serum to a nondividing monolayer of untransformed cells usually results in further rounds of division, so that the cells pile on top of one another (Fig. 4.1) (Kruse and Miedema 1965; Temin 1967; Todaro et al. 1967; Holley and Kiernan 1968). Viral transformation changes, and usually reduces, a cell's dependence on serum. Characteristically transformed cells require much less serum than untransformed cells to initiate their division cycle (Temin 1967; Biirk 1966; Holley and Kiernan DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

4 / Transformation by SV40 and Polyoma Virus

213

1968; Eagle et al. 1970; Dulbecco 1970; Clark et al. 1972; Paul et al. 1971). Early attempts to fractionate the various survival and growth-promoting factors in calf serum (Paul et al. 1971; Lipton et al. 1971, 1972; Holley and Kiernan 1971) confirmed that untransformed and transformed cells in culture differ in their response to serum factors (Holley and Kiernan 1968; Clarke et al. 1970). They also showed that serum contains several substances that can be loosely classified together as growth factors; these include factors required by cells for survival, for multiplication, and for migration. Furthermore, we now know that at least some transformed cells secrete into the surrounding medium substances that can satisfy the serum requirement of other cells (Austin et al. 1971; Alfred and Pumper 1960; Rubin 1966; Stoker et al. 1971). Shodell (1972) detected in the serumfree medium, in which cells of a mouse L-cell tumor line had grown, material that satisfies the serum requirement of baby hamster kidney (BHK) cells and causes them to proliferate in the absence of serum. More recently, Nishikawa et al. (1975) have identified a cell survival factor and a fibroblast growth factor (FGF), the latter originating in the pituitary, and Stiles et al. (1979) have shown that, in the absence of serum, repeated division of 3T3 cells is sustained by a combination of two factors: insulin and platelet-derived growth factor (PDGF). Even though the mechanism of action of growth factors is unknown, their discovery clearly represents a significant advance. We have reason to hope that we may someday be able to decipher the signals that guide cells through their growth cycles and learn how viral transformation alters a cell's dependence on, or response to, external growth factors.

Assays for Transformation

Four chief methods have been developed for obtaining and assaying cells transformed by polyoma virus or SV40. All of them are selective for one or more of the phenotypic characteristics of transformation. Although the methods select different characters, they all yield cells that usually grow to high saturation densities, contain viral nucleic acids and proteins, and DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

214

DNA Tumor Viruses

possess many of the other traits listed in Table 4.1. The four methods can be summarized as follows: 1. Subconfluent cultures of cells are exposed to the virus and then plated at low cell densities. This method imposes no selection other than for the ability to grow on a plastic or glass surface; both the untransformed and the transformed cells form colonies that can be distinguished by their density and morphology (Stoker and MacPherson 1961; Todaro and Green 1964a). The method was originally devised for obtaining BHK cells transformed by polyoma virus, but it has since become the procedure most commonly used for obtaining 3T3 cells transformed by SV40. 2. Cells in exponential growth are infected with the virus and then plated in 0.33% agar or 1.2% Methocel on a nonadhesive surface (e.g., culture dishes coated with 0.9% agar). Transformed cells are able to form large colonies, the viscosity of the medium causing the clones to remain isolated, whereas untransformed cells are unable to multiply. This technique is now the method of choice for the selection of BHK or NIL-2 hamster cells transformed by polyoma virus. For these two virus-cell combinations, both methods detect transformants with the same efficiency, but the second method is much simpler (MacPherson and Montagnier 1964). 3. Subconfluent monolayers of cells are exposed to the virus. After 2-3 weeks without replating, transformed clones can be selected as dense clones of cells growing out from the (by now) monolayer of untransformed cells. 4. Subconfluent cultures of cells are infected with virus and plated at low concentrations in medium containing very low concentrations of serum or growth factors insufficient to support the growth of untransformed cells. Many cells are induced by viral infection to multiply for a few generations (abortive transformation), but they do not survive transfer into fresh medium containing only low concentrations of serum. The cells that are stably transformed by the virus acquire the ability to multiply continuously in factor-free medium. Most of these cells show most if not all of the properties typical of virus-transformed cells (see Table 4.1). However, a minority of transformants selected in this way, while retaining the ability to multiply in medium depleted of DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

4 / Transformation by S V 4 0 a n d P o l y o m a Virus

215

factors, grow only to low saturation densities in complete medium. Growth in serum-free medium therefore allows the selection of a class of viral transformants that would not be selected in the other assays. To date, this method has been used only with Balb/c-3T3 cells infected with SV40 (Smith et al. 1970). All these assays depend in one way or another on an increase in the cell's nutritional autonomy following transformation, but the agar-suspension assay imposes in addition selection for the ability to divide without anchorage to a solid support.

Parameters of the Transformation Event

Although the multiplicity of infection, the genetic properties of the virus, and the physiological state of the cells can influence the frequency of transformation, the most important factor is the genetic properties of the cell, including, of course, its species. Genetic Properties of the Host Cell: Permissive, Semipermissive, and Nonpermissive Hosts Polyoma virus and SV40 transform cells of many different species. Mouse cells are fully permissive to polyoma virus, and African green monkey cells are permissive to SV40 (see Table 4.2). In such cells the genome of the virus is fully expressed, the viral DNA is replicated, progeny viral particles are released, and the cell is killed. In contrast, completely nonpermissive cells survive an infection and progeny viral particles are never released, but some viral genes are expressed at least transiently. Between these two extremes lie cells that are semipermissive for polyoma virus or SV40. When populations of such semipermissive cells are infected, some of the cells support the replication of the virus and die, whereas others provide a nonpermissive environment and so survive the infection. At present, we do not know what causes a cell to be nonpermissive or permissive. However, Basilico et al. (1970) showed that the yield of polyoma virus particles obtained from hybrids of infected mouse (permissive) and hamster (semipermissive) cells was positively correlated with the number of murine chromosomes in these hybrids. This experiment strongly sugDNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

216

Table

DNA Tumor Viruses

4.2

R e s p o n s e o f c e l l s of d i f f e r e n t s p e c i e s t o i n f e c t i o n b y

S V 4 0 a n d p o l y o m a virus Infection by Species

Infection by S V 4 0

p o l y o m a virus

Human

semipermissive

nonpermissive

Mouse

nonpermissive

permissive

African green permissive

nonpermissive

Rat

nonpermissive or semipermissive

semipermissive

G u i n e a pig

nonpermissive or semipermissive

n.t.

Rabbit

nonpermissive or semipermissive

n.t.

monkey

a

Cow

nonpermissive or semipermissive

n.t.

Hamster

semipermissive

semipermissive

Lizard

semipermissive

n.t.

a

n.t. = not tested.

gests that the permissivity of mouse cells for polyoma virus depends on the expression of a mouse chromosomal gene(s) that specifies a factor(s) essential for replication of the virus. If this is the case, it seems reasonable to suggest that cells of nonpermissive species do not have a gene(s) for such a factor(s), while the differences in the extent to which differently differentiated mouse cells support the replication of polyoma virus may reflect differences in the extent to which the gene(s) for the factor is expressed. 3T3 cells are totally nonpermissive for SV40, but after they are infected by this virus, they acquire SV40-specific antigens. Furthermore, after high-multiplicity infection by SV40, quiescent 3T3 cells replicate their DNA and go through one or more rounds of mitosis (Smith et al. 1971). SV40 early proteins are made, but the viral genome is not fully replicated and neither capsid antigen (V antigen) nor progeny particles can be detected. These data suggest that the factor lacking in nonpermissive cells is required for the complete replication of viral DNA. The block to late gene expression can be overcome by microinjecting about 3000 viral genomes directly into the nuclei of 3T3 cells (Graessmann et al. 1976), in which case V antigen, as well as T antigen, can be detected. Progeny virions, however, are not released by the cells even after such drastic manipulation. Presumably, semipermissive cells for one reason or another DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

4 / Transformation by SV40 and Polyoma Virus

217

do not continuously supply adequate amounts of the factor(s) required for replication of these viruses. We can further speculate that at some low frequency, and at random, individual nonpermissive cells in a population make enough of the factor(s) to support replication of the virus and, as a result, if they are infected, eventually die. It now seems clear that populations of BHK cells are semipermissive rather than absolutely nonpermissive for polyoma virus because (1) Fraser and Gharpure (1963) found that a small proportion of the cells in populations of BHK cells infected with polyoma virus stained with antibody against polyoma virus; (2) Bourgaux (1964) showed that some BHK cells infected with polyoma virus incorporated [ P]orthophosphate into material that cosedimented with polyoma virus; and (3) Folk (1973) showed that as many as 1% of the cells in some clones of BHK cells transformed by polyoma virus yielded progeny virus when they were chilled from 39°C to 31°C. Rabbit cells (Black and Rowe 1963; Watkins 1975) and Syrian hamster cells are semipermissive for SV40. Transformants can be obtained which yield SV40 on fusion with monkey cells. These transformants also can be induced to yield SV40 after exposure to agents such as mitomycin C (Burns and Black 1968, 1969), bromodeoxyuridine (BrdU), caffeine and ultraviolet light or X-irradiation (Rothschild and Black 1970), and cycloheximide, or after amino acid deprivation (Kaplan et al. 1972). Rat cells (Diderholm et al. 1966), bovine cells (Diderholm et al. 1965), some guinea pig cells (Diderholm et al. 1966), and even some cells of a reptilian species (Clark et al. 1972; Michalski et al. 1974) also can be transformed by SV40. Although they have not been fully characterized, most of these species are probably semipermissive. In summary, the outcome of infection by wild-type SV40 and polyoma virus depends first and foremost on the genetic properties (and usually, therefore, the species) of the host cell. Permissive cells are killed; nonpermissive cells survive, but they may become stably transformed and acquire a subset of the set of properties listed in Table 4.1. Clearly, if permissive cells are infected with defective mutant viruses lacking functions essential for replication, they may survive and be transformed. So as a general rule, it is safe to say that permissive cells are transformed by defective viruses, whereas nonpermissive cells may be 32

DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

218

DNA Tumor Viruses

transformed by wild-type or defective viruses. All the species of cells commonly used to study polyoma virus are semipermissive. Therefore, it seems reasonable that most polyoma virus transformants should harbor defective rather than wild-type viral genomes. This is presumably the reason why polyoma virus cannot be rescued from most transformants by fusing them with permissive mouse cells. SV40 Transformation of Human Cells The interaction between SV40 and human cells is unusual. Shein and Enders (1962) and Koprowski et al. (1962) reported that human cells transformed by SV40, like rodent cells transformed by polyoma virus, grew rapidly, had an epithelial morphology, and formed dense colonies. These SV40-transformed human cells, however, continued to release SV40 after months in culture (Ponten et al. 1963). After about 50-75 cell generations, uninfected human fibroblasts in culture enter a crisis from which they do not recover. Following transformation by SV40 (full transformation as judged by cellular morphology, rate of growth, and saturation density achieved), this crisis is greatly delayed (Todaro et al. 1968). Eventually, after prolonged periods of culture, the transformed human cells enter a crisis (Shein et al. 1964), from which, only by careful nursing, is it occasionally possible to recover fully transformed, stable clones (Girardi et al. 1965). In other words, transformation by SV40 does not make it significantly easier to establish permanent lines of human fibroblasts, and in general, stable lines of SV40 transformants can only readily be obtained from species whose cells, even without viral transformation, can be established as permanent lines. Genetic Variation in Susceptibility of Human Cells to SV40 Transformation Human cells obtained from some persons with a high risk of developing cancer are 10-50 times more susceptible to transformation in vitro by SV40 than cells obtained from normal persons. Cells with increased transformation rates have been obtained from a family with a history of multiple cases of sarcoma, from patients with Fanconi's anemia, from heterozygous relatives of patients with Fanconi's anemia, and from DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

4 / Transformation by SV40 and Polyoma Virus

219

patients with Down's syndrome (Todaro et al. 1966; Todaro and Martin 1967; Potter et al. 1970). Moreover, Mukerjee et al. (1970) found that cells taken from people suffering from Klinefelter's syndrome are also very susceptible to transformation by SV40. All these syndromes are characterized not only by a high incidence of chromosomal abnormalities, but also by a high incidence of cancer. Skin fibroblasts from patients with primary immunodeficiencies, who also have a greater than normal risk of developing cancer, are not, however, more susceptible to transformation by SV40 than skin fibroblasts from healthy persons (Kersey et al. 1972). For this reason, initial hopes that by assaying human cells for susceptibility to transformation one might be able to screen for individuals with a high risk of cancer were not fulfilled. The basis of the increased susceptibility to transformation is unknown. It seems to be concerned with an event after the adsorption of the virus, but before T antigen becomes detectable (Aaronson and Todaro 1968). Aaronson (1970) has shown that the differences in transformation rates among human cell strains infected with intact SV40 particles are eliminated when SV40 DNA is used as a transforming agent. Apparently, normal human cells are resistant to transformation by SV40 because of some block in either the penetration or uncoating of the virus, and this block, which is lacking in susceptible cells, seems to be specific for SV40. For example, there is no such block against the SV40 genome in adenovirus-SV40 hybrid particles. Multiplicity of Infection Transformation is an inefficient process. The frequency of transformation by the DNA tumor viruses varies greatly, but usually 10 -10 infectious units (i.e., 10 —10 particles) are needed per transforming event. At low input multiplicities, the number of cells transformed is directly proportional to the multiplicity of infection, which suggests that a single viral particle or DNA molecule is sufficient to initiate transformation (Stoker and Abel 1963; MacPherson and Montagnier 1964; Todaro and Green 1966a). With polyoma virus and BHK cells, a maximum of only about 5% of the cells in a population ever become transformed at very high input multiplicities of virus. This low frequency of trans2

4

4

6

DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

220

DNA Tumor Viruses

formation is not the result of genetic inhomogeneity in the cell population; freshly isolated subclones show the same behavior (MacPherson and Stoker 1962; Black 1964), although the genotype of a BHK cell has some influence on the probability of its being transformed by polyoma virus (Stoker and Smith 1964). In the SV40-3T3 system, up to 40% of the cells can be transformed (Todaro and Green 1966a) if very high multiplicities of virus (10 infectious particles/cell) are used. The frequency of integration of the viral DNA into host chromosomes may well be the rate-limiting step to transformation. 6

Physiological State of the Cells There is no evidence that cells have to be in any particular physiological state in order to be transformed. Experiments involving the infection of synchronized populations of cells have revealed only small differences in the rates of transformation of cells infected at different states of their growth cycle (Basilico and Marin 1966). Cells in log phase, however, are more susceptible to transformation than resting cells (Todaro and Green 1967), and cells surviving X-irradiation (Stoker 1964) or exposure to the base analogs 5-bromodeoxyuridine and 5iododeoxyuridine (Todaro and Green 1964b) are transformed at increased rates. Furthermore, at least in the polyoma-virusBHK cell system, the probability of transformation may be affected by the pH and the concentration of magnesium ions (Stoker and Abel 1963; Kirsch and Fraser 1964). Although the physiological state of the cells does not seem to affect transformation in a quantitative sense, it does seem to influence the phenotype of the transformants subsequently derived. Cuzin and his coworkers (Seif and Cuzin 1977; Rassoulzadegan et al. 1978a,b; Gaudray et al. 1978) have shown that transformation of rat fibroblasts with either SV40 or polyoma virus may lead to two distinct types of transformants, which they refer to as type A and type N. The two types can be distinguished operationally by using cells transformed with the early temperature-sensitive (ts) mutants tsA of polyoma virus and tsA30 of SV40. Type-A transformants are exclusively selected by cloning in agar, and they are fully transformed at both low and high temperatures. Type-N transformants, on the other hand, are derived from foci that overgrow monolayers of normal DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

4 / Transformation by SV40 and Polyoma Virus

221

cells and revert at high temperature to the normal phenotype for many of the characters that define the transformed state. The state of growth of the cell during the first 4 days after infection determines whether it is transformed into the A or the N state. Cells arrested in the G state are converted into type-A transformants, whereas actively growing cells develop into type-N transformants. The molecular events that underlie these phenomena are unknown, but see below. 0

EVENTS DURING TRANSFORMATION Abortive Transformation of Nonpermissive Cells

Although stable transformation of nonpermissive or semipermissive cells by polyoma virus or SV40 is an inefficient process, for every transformed cell that emerges, 10-100 cells in the infected culture transiently acquire the transformed phenotype; such cells are said to be abortively transformed. Abortive transformation was first described by Stoker (1968), who infected BHK-21 cells in suspension with polyoma virus and plated them in suspension in semisolid media. Whereas uninfected cells did not divide, many of the infected cells multiplied for about five or six generations and then stopped dividing. When these clones were picked and examined, they were normal by the two criteria used, i.e., the orientation of cell growth and the ability to grow in suspension. Only a small proportion of the initially infected cells went on to be stably transformed. Subsequently, Smith et al. (1970, 1971) reported "abortive transformation" of 3T3 cells adhering to a plastic petri dish following infection by SV40 in medium containing a low concentration of serum. Like stably transformed cells, those transformed abortively have a reduced serum requirement (Dulbecco 1970), appear morphologically transformed (Dulbecco 1970; Taylor-Papadimitriou et al. 1971), have an increased susceptibility to agglutination by lectins (Benjamin and Burger 1970), and are not dependent on anchorage for division (Stoker 1968). In short, to the extent that it has been tested, the phenotype of abortive transformants is identical to that of comparable stable transformants, except that it is ephemeral. In some cases, however, DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

222

DNA Tumor Viruses

the abortively transformed cells retain cryptic viral DNA sequences after they have reverted to the untransformed state (M. Fried, unpubl.). The duration of abortive transformation coincides with the duration of synthesis in the infected cells of viral T antigens (Oxman and Black 1966; Stoker and Dulbecco 1969), and at least one aspect of abortive transformation results solely from the action of these viral gene products. Graessmann and Graessmann (1976) have shown that, in cells microinjected with RNA complementary to SV40 DNA, T antigen is synthesized and host-cell DNA synthesis is stimulated. More directly, Tjian et al. (1978) have been able to stimulate host-cell DNA synthesis by microinjection of a purified protein containing SV40 T antigen linked to an adenoviral polypeptide (see Chapter 11). This novel form of abortive transformation could provide a valuable method for analyzing the dependence of other aspects of transformation on the gene products of SV40 and polyoma virus. The role of the products of the early genes of polyoma virus during abortive transformation has been further studied with two useful classes of temperature-sensitive mutants: tsA mutants, which are defective for production of functional large T antigen; host-range transformation (hr-t) mutants, which are defective for production of middle and small T antigens; and dl23 mutants, which make small T antigen but not functional middle and large T antigens (see Chapter 5). No mutant defective in the production of polyoma virus middle T antigen can fully transform or abortively transform cells (see below and Chapter 5). We do not yet understand why, at any one time in a clonal population of cells, only a small proportion of the cells are susceptible to stable transformation, while a majority are susceptible to abortive transformation. The phenomenon of abortive transformation, however, indicates that there is a link between continued expression of the early genes of the virus and the maintenance of transformation.

Biochemical Changes

Infection of 3T3 cells with SV40 results in a three- to eightfold increase in the specific activity of several enzymes involved in DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

4 / Transformation by SV40 and Polyoma Virus

223

DNA synthesis (e.g., thymidine kinase, DNA polymerase, dCMP deaminase, and dTMP kinase) (Kit et al. 1967a), and in this respect, infections of nonpermissive cells resemble productive infections (see Chapter 3). Because no differences have been detected between these induced enzymes and those present in uninfected cells (Hatanaka and Dulbecco 1966; Kit 1966), it seems probable that most of the enzymes induced in infected nonpermissive cells and stably transformed cells are coded by the cell. The mechanism of this enzyme induction is obscure; neither DNA synthesis nor cell division is required (Kit 1966; Kit et al. 1966). However, protein synthesis is necessary (Frearson et al. 1966), and therefore it seems likely that the enzymes are synthesized de novo. Cellular DNA synthesis is induced both in 3T3 cells infected with SV40 and in rat embryo cells infected with polyoma virus (Gershon et al. 1965, 1966; Sheinin 1966; Henry et al. 1966; Kit et al. 1967b; May et al. 1971). Attempts to detect viral DNA synthesis and capsid proteins have been unsuccessful (Gershon et al. 1966; Henry et al. 1966); however, viral RNA can be detected in abortively infected cells (Khoury et al. 1972).

Fixation of Transformation

At least one round of cell division is necessary to "fix" the transformed state. Todaro and Green (1966b) infected 3T3 cells with SV40 and seeded the cells at different concentrations so that they underwent varying numbers of divisions before reaching their saturation density. They found that at least one cell division is required for the transformed state to become irreversibly fixed; infected cells that did not undergo one division after infection were not transformed. Using interferon, the fixation event in the SV40-3T3 system has been further localized. Todaro and Green (1967) synchronized cultures of 3T3 cells, infected them with SV40 at different stages of the cell cycle, and then added interferon to block the expression of functions coded by the virus. They found that interferon added before the S period prevented transformation and that there was no effect if interferon was added after the S period. Interferon also blocks abortive transformation (Dulbecco and Johnson 1970; TaylorPapadimitriou and Stoker 1971). The mechanism by which inDNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

224

DNA Tumor Viruses

terferon prevents transformation is not known, but the fact that, after one cell division, transformation is insensitive to interferon (Todaro and Green 1966b) suggests that the fixation of the transformed state occurs in either the S or G phase of the cell cycle (Basilico and Marin 1966). The nature of the fixation event remains unknown, but may very well be the integration of the viral DNA into cellular chromosomes in such a way as to allow the continued expression of the early gene region. 2

Transformation of Permissive Ceils

As a general rule, permissive cells are transformed by defective viruses. Such transformants are not easy to isolate without manipulating the experimental conditions in ways that either increase the proportion of defective viral particles in the stock used or prevent the spread of progeny wild-type virus throughout the culture (Fernandes and Moorhead 1965; Jensen and Koprowski 1969; Rapp and Trulock 1970; BarbantiBrodano et al. 1970; Sauer and Hahn 1970; Defendi 1968; Shiroki and Shimojo 1971; Gluzman et al. 1977). Furthermore, because transformation of permissive cells is a rare event, it cannot be quantitated. Using a stock of SV40 that had been heavily irradiated with ultraviolet light, Shiroki and Shimojo (1971) isolated lines of African green monkey kidney (AGMK) cells transformed by SV40 (SV-AGMK). These cells had the following properties: (1) They contained SV40 T antigen and SV40 DNA; (2) they did not contain V antigen; (3) they did not yield SV40 when fused with uninfected AGMK cells, but they were susceptible to superinfection by SV40 and were killed as a consequence of it. This set of properties is exactly that expected of permissive cells transformed by a defective virus. More recently, Gluzman et al. (1977) isolated SV-AGMK cells by essentially the same procedure and showed that they contained T antigen and an average of 1-2 genome equivalents of SV40 DNA, which was associated with cellular DNA. The growth properties of the cells were typical of nonpermissive transformed cells, serum dependence was reduced, the saturation density was increased, and so forth. The cells were fully susceptible to superinfection by DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

4 / Transformation by SV40 and Polyoma Virus

225

wild-type SV40. Interestingly, however, these SV-AGMK cells failed to support the replication of SV40 tsA mutants at the restrictive temperature. The discovery of SV40 small T antigen has made it possible to interpret these observations as indicating that the SV-AGMK cells were transformed by a virus unable to specify functional SV40 large T antigen but able to specify normal small T antigen. Knowles et al. (1968) transformed AGMK cells with SV40 without taking any special measures to ensure a high proportion of defectives in the stock of SV40 they used. Their SV-AGMK transformants had curious properties. When cells from each of these lines were fused with fresh AGMK cells, SV40 was not rescued. However, when two transformants from different lines were fused with an AGMK cell, apparently wild-type SV40 was often rescued. Knowles et al. argued that this indicated that cells of each transformant line contained defective SV40 and that each line was transformed by a different defective SV40 genome. Rescue of virus by fusion of two different transformants with a fresh AGMK cell was, they suggested, the result of complementation and recombination of defective SV40 genomes within the heterokaryon. Curiously, however, when pairs of transformants of different lines were fused together, virus was not rescued, even though the transformants, and presumably therefore the heterokaryons, were susceptible to superinfection by wild-type SV40. More recently, Wilson et al. (1976) selected those AGMK cells that survived in a population infected by wild-type SV40. Unlike the cells obtained by Knowles et al. (1968), these cells were not phenotypically transformed. They contained no viral T antigen, and they proved to be much less susceptible to SV40 infection than the parental cells; however, none of them were absolutely resistant. Such resistant cells appear to be selected for, rather than induced, by SV40 infection, and the block to infection appears to be at some stage in viral uncoating, a situation reminiscent of the resistance of human cells to infection by SV40. Wilson (1977) has subsequently partially characterized host-range mutants of SV40, isolated by Wilson et al. (1976), which overcome this block to infection by wild-type virus. Clearly, if in populations of permissive cells nonpermissive variants arise which allow the virus to uncoat but can no longer DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

226

DNA Tumor Viruses

support viral replication, these cells should be susceptible to transformation by wild-type virus. Shiroki and Shimojo (1971) have apparently isolated SV40 transformants of such variant, nonpermissive AGMK cells. These transformed cells contain SV40 T antigen and SV40 RNA; at least some yield SV40 virus when fused with normal AGMK cells, and as expected they cannot be superinfected by wild-type SV40. Additionally, Koprowski et al. (1967) and Swetly et al. (1969) have reported isolating clones of AGMK cells transformed by SV40 that can be superinfected and killed by wild-type SV40 and that yield wild-type SV40 upon fusion with uninfected AGMK cells. In summary, we can tentatively identify three classes of SV40transformed monkey cells: (1) cells transformed by defective virus; (2) variant, nonpermissive cells transformed by wild-type virus; and (3) a rare class of permissive cells apparently transformed by wild-type virus. Lines of mouse cells transformed by polyoma virus have been isolated on several occasions (Vogt and Dulbecco 1960; Hellstrom et al. 1962; Todaro and Green 1965; Benjamin 1970). Many of these cells resist superinfection by polyoma virus particles but can be superinfected by polyoma virus DNA, and the transforming viral genome cannot be rescued by fusion (Watkins and Dulbecco 1968; Watkins 1971). These properties strongly suggest that such transformants arise when permissive cells are infected by defective viral particles.

Double Transformation of Cells

Polyoma virus and SV40 have similar biologies, and it is perhaps surprising to find that an appropriate cell transformed by one of these viruses can be subsequently transformed by the other. Using colony morphology as a marker, Todaro and Green (1965) detected the transformation by SV40 of 3T3 cells already transformed by polyoma virus, and these double transformants contained both SV40-specific and polyoma-virus-specific T antigens. In addition, Takemoto and Habel (1966) showed that cells of hamster tumors induced by SV40 could be transformed by polyoma virus to yield cells with the specific T antigens and tumor-specific transplantation antigens (TSTAs) of both viruses. DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

4 / Transformation by SV40 and Polyoma Virus

227

Cells can also be doubly transformed by SV40 and a human adenovirus. Some hamster cells transformed in vitro or in vivo by stocks of PARA (defective SV40-adenovirus hybrids; see Chapter 11) contain both SV40-specific and adenovirus-7-specific T antigens (Rapp et al. 1966,1969; Butel et al. 1971; Richardson and Butel 1971; Duff and Rapp 1970) and both SV40-specific and adenovirus-7-specific RNAs (Levin et al. 1969). Furthermore, some cells transformed by stocks of PARA transcapsidated by adenovirus 12 (see Chapter 11) contain SV40-specific RNA, adenovirus-7-specific RNA, and adenovirus-12-specific RNA. Transformants isolated as dense foci after SV40 infection can be further transformed to anchorage independence by the RNA tumor virus, murine sarcoma virus (Renger 1972); and rat cells infected, but not transformed, by Rauscher leukemia virus have an enhanced susceptibility to transformation by SV40 (Rhim et al. 1971). All these data clearly indicate that different DNAtransforming viruses and DNA- and RNA-transforming viruses do not exclude one another by competition, as would be expected if the DNAs of different transforming viruses were integrated at different sites in the host genome.

Transformation of Differentiated Cells

Most experimental work on transformation has involved relatively undifferentiated cells of fibroblastic origin. However, a number of investigators have infected highly differentiated cells with SV40 or polyoma virus either to obtain information about the effect of transformation on differentiated functions or to establish permanent lines of the cells themselves. The catalog of differentiated cell types transformed by SV40 or polyoma virus in vitro is now quite long (see Table 4.3). In most cases, where enzymatic markers were compared before and after transformation, they were conserved. Hamster prostate cell transformants possess a tartrate-sensitive acid phosphatase similar to that of the prostate gland; pineal cell transformants have hydroxyindole-O-methyltransferase activity; and parathyroid and thyroid cell transformants secrete parathyroid hormone and thyrocalcitonin and prostaglandins, respectively. On the other hand, hamster salivary gland transformants show no amylase DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

228

DNA Tumor Viruses

Table

4.3

Differentiated cells that h a v e b e e n transformed by S V 4 0

or p o l y o m a virus Cell

Origin

Virus

Reference D a w e and L a w (1959)

Salivary gland

Mouse

Polyoma

Salivary gland

Hamster

P o l y o m a a n d S V 4 0 W e l l s e t al. ( 1 9 6 6 c )

Thyroid gland

Hamster

SV40

W e l l s e t al. ( 1 9 6 6 b )

Thyroid gland

Human

SV40

G r i m l e y e t al. ( 1 9 6 9 )

Pineal gland

Hamster

P o l y o m a a n d S V 4 0 W e l l s e t al. ( 1 9 6 6 a )

Pineal gland

Hamster

P o l y o m a a n d S V 4 0 O r m e e t al. ( 1 9 6 8 )

Prostate gland

Hamster

SV40

P a u l s o n e t al. ( 1 9 6 8 a , b )

Lens

Hamster

SV40

A l b e r t e t al. ( 1 9 6 9 )

Lung

Hamster

SV40

D i a m a n d o p o u l o s and

Liver

Hamster

SV40

D i a m a n d o p o u l o s and

Uvea

Hamster

SV40

A l b e r t e t al. ( 1 9 6 8 )

Retina

Hamster

SV40

A l b e r t e t al. ( 1 9 6 8 )

Parathyroid gland

Human

SV40

D e f t o s e t al. ( 1 9 6 8 )

S p l e e n cells

Rabbit

SV40

C o l l i n s e t al. ( 1 9 7 4 )

epithelium

Enders (1965) Enders (1965)

activity characteristic of the parent tissue. In a particularly elegant experiment, Black's group (Collins et al. 1974; Strosberg et al. 1974) infected with SV40 spleen cells from a rabbit hyperimmunized with pneumococcal vaccine. The resulting transformants, which contained T antigen and from which SV40 could be rescued, synthesized and secreted into the culture medium IgG molecules directed against the immunizing antigen. Thus it seems that, although transformed cells have substantially altered surface, growth, and enzymatic (e.g., protease activation) properties, these changes do not necessarily preclude the continued expression of the differentiated functions of the cell. Recently, Topp et al. (1976, 1977) infected cells of a teratocarcinoma line with SV40. In culture, these cells undergo differentiation to a wide variety of mature cell types. Topp et al. hoped to obtain sublines that might represent partially differentiated states stabilized by the virus-cell interaction. The experiment was only partially successful. Although clearly transformed, the resulting cells retained the high levels of enzymatic activities characteristic of the untransformed cell population. These included the brain or muscle isozymes of DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

4 / Transformation by SV40 and Polyoma Virus

229

creatine phosphokinase, which are markers of central nervous system or muscular differentiation, and the secretion of plasminogen activator, a characteristic of trophoblastic, vascular endothelial, and other tissues. Such enzymatic markers of different differentiation pathways were not mixed in any one clone, but unfortunately very few of the transformed cell lines, unlike the initial population, proved to be tumorigenic. The few that were gave rise to relatively undifferentiated fibrosarcomas rather than tumors with a well-defined and differentiated histology.

PROPERTIES OF TRANSFORMED CELLS

Of the distinguishing properties of transformed cells listed in Table 4.1, we have already discussed changes in the pattern of growth and how these have been exploited in assays for transformation. The many changes at the cell's surface that accompany transformation and include the appearance of new antigens (virus-specific transplantation antigens and so-called surface antigens) are discussed in detail in Tooze (1980) and will only briefly be mentioned here.

Virus-specific Transplantation Antigens

Animals immunized with either polyoma virus or SV40 are resistant to subsequent challenges with transplantable tumors induced by that virus (Habel 1961; Sjogren et al. 1961; Khera et al. 1963; Habel and Eddy 1963; Koch and Sabin 1963; Defendi 1963). The tumor-rejection mechanism appears to be mediated by sensitized immune lymphocytes and is highly specific. There is no cross-reaction between cells transformed in vitro with SV40 or polyoma virus or between cells from tumors induced by these two viruses (Sjogren 1965). The same transplantation antigen is present in cells of different species transformed by the same virus (Sjogren 1965). This was demonstrated by the rejection of transplantable tumors within a species if the animals had previously been immunized with DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

230

DNA Tumor Viruses

transformed cells or tumor cells produced by the same virus but from a different species. Two pieces of evidence suggest that TSTA is present on the surface of transformed cells. First, the rejection of tumor cells by immune animals depends on contact between sensitized lymphocytes and TSTA (Coggin et al. 1967). Second, newborn hamsters injected with cell membranes from SV40-induced tumor cells become immunologically tolerant to SV40 TSTA (Tevethia and Rapp 1966), and adult hamsters can be immunized against SV40 tumor cells by injections of membranes from SV40 tumor cells (Coggin et al. 1967). SV40 TSTA is also apparently synthesized as the virus replicates in monkey cells, and its appearance is blocked by inhibitors of protein synthesis and actinomycin D, but not by inhibitors of DNA synthesis (Girardi and Defendi 1970). Recent experiments indicate that SV40 TSTA and T antigen are closely related (Tenen et al. 1975; Anderson et al. 1977a,b; Tevethia and Tevethia 1977; Chang et al. 1977a,b,c). The amount of both antigens is modulated coordinately in cells transformed by tsA mutants. Partially purified, large T antigen and TSTA have similar sedimentation properties, both bind to DNA, and they copurify during DEAE-cellulose or phosphocellulose chromatography. In addition, animals immunized with purified T antigen reject SV40-induced tumors, and T antigen and TSTA serologically cross-react. Probably TSTA is large T antigen, or some derivative of it, inserted in the transformed-cell membrane (Tevethia et al. 1980).

Surface Antigen

A surface antigen not detectable on the surface of normal cells can be detected on the surface of transformed cells by immunofluorescence, mixed hemagglutination tests (Hayry and Defendi 1968, 1969, 1970; Metzgar and Oleinick 1968), cytotoxic antibodies (Hellstrom and Sjogren 1965; Tevethia and Rapp 1965), and colony-inhibition tests (Hellstrom and Sjogren 1965, 1967). The chemical identity of the surface antigen(s) is unknown, but is not coded by the virus (Hayry and Defendi 1970).

DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

4 / Transformation by SV40 and Polyoma Virus

231

Enhanced Agglutination by Lectins

Aub et al. (1963) noticed that a contaminant of wheat germ lipase caused the selective clumping or agglutination of tumor cells in a mixed population of tumor cells and normal cells. Subsequently, the contaminant, wheat germ agglutinin (WGA), was purified and shown to be a glycoprotein with two binding sites specific for N-acetylglucosamine residues (Burger and Goldberg 1967) that occur in cell surface glycoproteins. Agglutination results when the divalent WGA binds to glycoproteins on pairs of cells. Several other plant lectins, each binding specifically to particular glycosyl groups, have since been purified and have been found to discrimate between normal cells and tumor cells (for review, see Burger 1973). There was considerable excitement when Burger (1969) and Inbar et al. (1969) reported that lectins were able to discriminate between cells transformed by polyoma virus, SV40, or adenovirus 12 and their untransformed counterparts. However, the initial idea that transformation directly changes the number of lectin-binding sites at the cell surface proved to be false, and it is now believed that the increased susceptibility of transformed cells to agglutination by lectins is a secondary consequence of an increased fluidity of the cell membrane which allows more extensive clustering of the lectin receptors (see, e.g., Nicholson 1974a,b).

Enhanced Uptake of Metabolites

To survive, cells in culture must transport, across the plasma membrane from their medium, low-molecular-weight metabolites such as glucose and amino acids. Many cells in culture must be supplied with particular amino acids; otherwise, they cease dividing and remain quiescent until the nutrient is restored (Holley and Kiernan 1974a). Hormones in the culture medium, usually supplied by the serum, stimulate cells to concentrate nutrients and thereby enhance growth (Hershko et al. 1971). Insulin, for example, stimulates cell division in stationary monolayers of 3T3 cells (Rozengurt and DeAsua 1973; Holley

DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

232

DNA Tumor Viruses

and Kiernan 1974b) and is probably a requirement of virtually all cells in culture. In the light of both the critical dependence of cells in culture on an adequate supply of nutrients and the role of hormones in stimulating their uptake, the fact that transformation of cells by polyoma virus, SV40, and other tumor viruses stimulates the selective uptake of nutrients assumes considerable significance. Following transformation of BHK-21 cells by polyoma virus or 3T3 cells by SV40, for example, the uptake of inorganic phosphate, aminoisobutyric acid, glutamine, and 2-deoxyglucose increases two- to threefold (Cunningham and Pardee 1969; Foster and Pardee 1969; Isselbacher 1972). This enhanced transport is not simply a reflection of some generalized increase in permeability, since it is specific for particular nutrients (Foster and Pardee 1969; Cunningham and Pardee 1969). Experiments with cells transformed by SV40 containing a temperature-sensitive lesion in the early (A) gene (see Chapter 5) indicate that enhanced transport of nutrients depends on a functional A-gene product, i.e., large T antigen. Within 6 hours of transfer from the nonpermissive temperature, the rate of transport of nutrients rises to the enhanced level characteristic of transformants (Martin et al. 1971; Brugge and Butel 1975). These observations suggest an explanation of the ability of transformed cells to continue to multiply in media containing only low concentrations of serum, insufficient to support the multiplication of untransformed cells. Perhaps the increased efficiency of the transport of nutrients, resulting from the presence of large T antigen, compensates for the low concentration of those serum hormones that stimulate the scavenging of nutrients from the medium. The transformed cells, less dependent on exogenous hormones, can therefore maintain a sufficient intracellular concentration of essential nutrients to allow cell division. In short, transformation increases the nutritional self-sufficiency of the cells.

Changes in Fibronectin Distribution

One of the chief protein components of the outer surface of the plasma membrane of fibroblastic cells is a glycoprotein with a DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

4 / Transformation by SV40 and Polyoma Virus

233

molecular weight of 220,000-230,000 (Wickus and Robbins 1973; Hynes 1973). Now known as fibronectin, this glycoprotein was formerly called LETS (large, external, transformation-sensitive) protein (Hynes and Bye 1974). In some viral transformation systems, the distribution of fibronectin is changed following transformation. For example, when chick fibroblasts are transformed by Rous sarcoma virus (Blumberg and Robbins 1975; Hynes et al. 1975) or when rat fibroblasts are transformed by human adenoviruses (Chen et al. 1976), fibronectin disappears from the outer surface of the cell. However, transformation by SV40 does not inevitably lead to the disappearance of fibronectin. Many SV40 transformants that are highly tumorigenic have on their cell surface normal amounts of this protein (Steinberg et al. 1979). An explanation for this difference between SV40 transformants and adenoviral transformants has yet to be found.

Proteases and Transformation

In the immediate vicinity of invasive tumors, proteolytic activity is unusually high, and this probably reflects the fact that invasive tumors are able to break down the structure of adjacent normal tissue (Busch 1962). In contrast, the few data available indicate that proteolytic activity is not enhanced in benign tumors. The discovery in 1973 that cultures of chick fibroblasts transformed by Rous sarcoma virus have much higher fibrinolytic activity than cultures of untransformed cells and, furthermore, that this enhanced activity is temperature-sensitive in cells transformed by temperature-sensitive viral mutants caused great interest (Unkeless et al. 1973). Ossowski et al. (1973b) reported that transformation of hamster or mouse embryo cells by SV40 results in similar increases in fibrinolysis attributable to activation of serum plasminogen by a factor associated with transformed cells, which proved to be similar to the plasminogen activator urokinase, a normal product of the kidney (Ossowski et al. 1973a; Quigley et al. 1974). Previously, Burger (1969) had shown that limited proteolysis of untransformed 3T3 cells by a variety of enzymes caused the cells to become more susceptible to agglutination by lectins and also stimulated quiescent cells to synthesize DNA and divide DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

234

DNA Tumor Viruses

(Burger 1970; Sefton and Rubin 1970). Subeulturing of cells, of course, routinely involves exposure to proteases. Furthermore, exposure of untransformed cells to trypsin stimulated glucose transport (Sefton and Rubin 1970), although not to the extent characteristic of transformation (Hale and Weber 1975). These observations suggested that, following transformation, stimulation of the secretion of protease or a protease activator leading to increase autoproteolysis of the cell's surface could induce several of the phenotypic changes that characterize transformation. Further support for this suggestion included the following. Hamster fibroblasts transformed by SV40 or polyoma virus reverted to an untransformed morphology when cultured in a medium containing serum depleted of plasminogen, the precursor of the protease plasmin (Ossowski et al. 1973b). The cytoskeleton does not re-form, however, and the reversion of saturation density was not observed when plasminogen activator alone was specifically inhibited (Chou et al. 1974). Exposure to plasmin or trypsin caused dissociation of the well-developed intracytoplasmic networks of actin in untransformed cells (Pollack and Rifkin 1975), which is a change often associated with transformation (see below). Even the ability of viral transformants to divide when suspended in soft agar was found to appear to depend, in part, on the activation of serum plasminogen (Ossowski et al. 1973a; Pollack et al. 1974). In short, many of the phenotypic traits of transformed cells can be explained as a consequence of autoproteolysis of the cell surface. On the other hand, in the absence of enhanced proteolysis, many transformed cells retain the growth characteristic of transformants (Unkeless et al. 1973; Topp et al. 1976, 1977), and there are numerous examples of cells that secrete copious amounts of protease or protease activators but have few if any of the phenotypic traits of transformation (Beers et al. 1975; Astrup 1975; Granelli-Piperno et al. 1977; Vassalli et al. 1977).

Changes in the Cytoskeletal Architecture

Cells contain a variety of cytoplasmic proteins, which can reversibly polymerize to form tubular or filamentous structures that DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

4 / Transformation by SV40 and Polyoma Virus

235

assume highly ordered patterns. Tubulin, for example, a globular protein with a molecular weight of about 55,000, can polymerize into microtubules approximately 250 A in diameter. These microtubules constitute the mitotic spindle and have been implicated in a wide range of cellular functions from phagocytosis to the movement of organelles (for reviews, see Goldman and Knipe 1973; Porter 1976). Fibroblasts also contain actin and myosin molecules which are very similar, but not identical, to the actin and myosin of skeletal muscle (Pollard and Weihing 1974). In the phase-contrast microscope, fibers or cablelike structures, which were called stress fibers because they usually occur oriented along an axis of movement of the cell, had repeatedly been seen in fibroblasts (Buckley and Porter 1967). Ishikawa et al. (1969) showed that these stress fibers bind heavy meromyosin, suggesting that they might contain actin; subsequently, Lazarides and coworkers, using immunofluorescencelabeled antibodies against actin and myosin, were able to show in a series of elegant studies that the stress fibers consist of bundles of contractile microfilaments of polymerized actin, myosin, tropomyosin, and a-actinin (Lazarides and Weber 1974; Weber and Groeschel-Stewart 1974; Lazarides and Burridge 1975; Lazarides and Hubbard 1976). The stress fibers, or actin cables, are concentrated within the cell, close to sites of cell-to-cell or cell-to-substrate contact (McNutt et al. 1971), and there is evidence to suggest that they may be involved in cellular motility (Wessells et al. 1971; Goldman et al. 1975b), contact inhibition of motility (Heaysman and Pegrum 1973), maintenance of cell shape (McNutt et al. 1971), cell adhesion (Abercrombie et al. 1971), and the distribution of lectin-binding sites at the cell surface (Nicholson 1974a,b; Ukena et al. 1974) (see Fig. 4.2). Because precisely these properties are changed by transformation, it seemed possible that disruption of actin cables following transformation might underlie several of the phenotypic traits of transformation. A series of observations are consonant with this possibility. In a variety of (but not all) mouse, hamster, and rat cells transformed by either SV40, adenovirus 5, or Rous sarcoma virus, actin cables are disrupted, although polymerized actin microfilaments are present in disarray (McNutt et al. 1971, 1973; Goldman et al. 1975a; Dermer et al. 1974). The loss of actin DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

Figure

4.2

(a) A c t i n c a b l e s r e v e a l e d b y i n d i r e c t i m m u n o f l u o r e s c e n t m i c r o s c o p y in u n t r a n s f o r m e d R a t - 1 c e l l s , (b) A c t i n d i s t r i b u t i o n in

T a n t i g e n . A l l t h e m i c r o g r a p h s w e r e t a k e n at t h e s a m e m a g n i f i c a t i o n . ( M i c r o g r a p h s c o u r t e s y o f W . C . T o p p . )

s e r u m , (d) S V 4 0 - t r a n s f o r m e d R a t - 1 c e l l s ( 1 4 B l i n e ) s t a i n e d w i t h t h e s a m e s e r u m a n d r e v e a l i n g t h e n u c l e a r l o c a l i z a t i o n o f S V 4 0 l a r g e

an S V 4 0 - t r a n s f o r m e d c e l l l i n e ( 1 4 B ) d e r i v e d f r o m R a t - 1 c e l l s , ( c ) U n t r a n s f o r m e d R a t - 1 c e l l s s t a i n e d w i t h h a m s t e r a n t i - S V 4 0 - t u m o r

4 / Transformation by SV40 and Polyoma Virus

237

cables was found to correlate well with anchorage independence of cell division of many transformants (Shin et al. 1975), and even during abortive transformation of 3T3 cells by SV40, actin cables disappear (R. Pollack and W. Topp, unpubl.). Conversely, cells reverted from transformation regain cables (McNutt et al. 1971, 1973; Pollack et al. 1975a; Steinberg et al. 1978), as do cells transformed by tsA mutants of SV40 cultured at the restrictive temperature (Osborn and Weber 1975; Altenburg et al. 1976; Pollack et al. 1975a; Vollet et al. 1977). Disruption of the cytoskeleton is another example of a phenotypic change that often results from transformation and may cause or amplify further changes, but it is not ubiquitous. Transformation and Intracellular Cyclic Nucleotides

A small amount of viral genetic information brings about such a multiplicity of physiological changes in the transformed cell, presumably by altering some fundamental regulatory pathways. Because in normal cells the regulation of such pathways by peptide hormones involves the relay of signals within the cell by the so-called second messengers, cyclic adenosine and guanosine 3',5'-monophosphates (cAMP and cGMP) (Rail et al. 1957, 1958; Robinson et al. 1971), it became fashionable to investigate the changes in metabolism of cyclic nucleotides following transformation. In general, rapidly dividing cells were found to contain less cAMP than quiescent, nondividing cells (Sheppard 1972; Otten et al. 1972; Seifert and Paul 1972; Burger et al. 1972), and the stimulation of quiescent cells to divide was accompanied by a fall in the concentration of cAMP and an increase in the concentration of cGMP (Siefert and Rudland 1974; Rudland et al. 1974). Although Oey et al. (1974) showed that the increase in concentration of cAMP during serum starvation was less marked in transformants or revertants able to grow in low concentrations of serum, the addition of dibutyl cAMP to the medium of transformed cells brought about partial reversion of cellular morphology (Hsie and Puck 1971; Sheppard 1971) and apparently restored partial susceptibility to contact inhibition of growth (Sheppard 1971). On the other hand, Minton et al. (1976) found that, contrary DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

238

DNA Tumor Viruses

to expectations, tumor tissue contains unusually high concentrations of cAMP, a result that casts some doubt on the relevance of the findings with transformed cells in vitro. Differences in the cell-division cycle between untransformed cells and their transformed counterparts must also be kept in mind. The celldivision cycle of untransformed cells involves a quiescent phase, G , immediately after division and a protracted G i phase preceding the phase of DNA synthesis (S), the G phase, and actual division. The division cycle of corresponding transformed cells has no G phase, and the G j phase is shortened. It is precisely during G and G ) that concentrations of cAMP reach their maximum in untransformed cells. Perhaps the lower concentrations of cAMP in transformed cells reflect nothing more than the shortening of the cell-cycle time. Only by studying homogeneous and synchronized populations of cells are we likely to discover the specific relationship, if any, between cyclic nucleotide metabolism and transformation. 0

2

0

0

TRANSFORMATION AND TUMORIGENICITY

When a population of susceptible fibroblasts of a stable cell line is infected with polyoma virus or SV40 and transformants are selected using one of the standard assays, the cells that emerge usually exhibit most of the complete set of phenotypic traits listed in Table 4.1, even though the selection imposed was for only one or a limited number of traits. It appears that transformation is an all-or-nothing phenomenon. Such a conclusion would not, after all, be surprising, because no more than two or three viral gene products are necessary and sufficient for transformation (see below) and the plethora of phenotypic changes presumably result from a sequence of primary and secondary cellular responses to the few viral proteins. Such a conclusion would nonetheless be incorrect. Cells of stable lines are already fully adapted to indefinite growth in culture. They have survived rigorous selection over countless generations and have acquired, by genetic changes quite unrelated to tumor virus infection, several of the phenotypic traits of transformation that are amplified or modified by the tumor virus. In comparison with cells of fresh primary DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

4 / Transformation by SV40 and Polyoma Virus

239

cultures, cells of stable lines are already partially transformed. Both 3T3 cells (Boone et al. 1976) and BHK-21 cells (Stoker and Abel 1963), for example, not only grow indefinitely in culture and have a high plating efficiency, but also are tumorigenic under certain circumstances. Moreover, the standard assays for transformation impose a selection at the very least for the ability to plate on glass or plastic surfaces and usually for other changes. Cells that, as a result of viral infection, suffer only a small subset of the changes characteristic of transformation may not survive the selection procedure. Any attempt to determine whether or not particular combinations of phenotypic traits are acquired coordinately following transformation and to correlate individual traits or different subsets of traits with enhanced tumorigenicity demands the use of primary cultures and a method for isolating transformants that exerts the minimal possible selection. In 1974, Pollack's group (Risser and Pollack 1974; Pollack et al. 1974) embarked on such experiments. Initially they infected 3T3 cells with SV40, plated the infected cells at a low density on plastic dishes, and selected at random 100 colonies (Risser and Pollack 1974). Tfiese clonal lines proved to fall into the following three classes: (1) untransformed cells that contained no viral antigens and were identical to uninfected 3T3 cells; (2) "minimal" transformants that contained SV40 T antigen and were able to grow in media with only low concentrations of serum, but were unable to grow when suspended at low cell densities in methylcellulose; and (3) "full" transformants that were able to grow at low cell densities in suspension and possessed all the phenotypic characteristics of transformation that were assayed. The conclusions of this experiment were (1) that transformation of 3T3 cells by SV40 in the absence of any selection other than the ability to grow on a plastic surface can result in the acquisition of virtually all the phenotypic traits of transformation or only a subset of them, and (2) that the ability to grow in suspension, i.e., anchorage independence, is the most stringent criterion of transformation in this system. Pollack et al. (1974, 1975b) decided to repeat the experiments with primary cells; they chose primary rat embryo cultures, to be able to correlate tumorigenicity with "minimal" and "full" transformation. Primary rat embryo cultures infected with SV40 DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

240

DNA Tumor Viruses

were sparsely plated in plastic dishes. All the clonal colonies that grew had SV40 T antigen and were either minimally or fully transformed. Unlike uninfected or abortively transformed 3T3 cells, primary rat embryo fibroblasts do not survive plating at low cell densities unless they are transformed. When the transformants were assayed for their tumorigenicity by injection into nude mice (this assay has been shown by Stiles et al. [1976a] to differentiate reliably between cells of normal and neoplastic origin), only the fully transformed cells, which at low densities were able to grow in suspension cultures, were tumorigenic (Pollack et al. 1957b; Freedman and Shin 1974; Shin et al. 1975). The minimal transformants were not tumorigenic when 10 cells were injected into nude mice. If we could discover the mechanism by which SV40 induces rat embryo fibroblasts to grow at low cell densities without anchorage, we should shed light on one of the pathways leading to tumorigenicity. We say "one of the pathways" advisedly, because the close correlation between anchorage independence and tumorigenicity, which holds for rat embryo fibroblasts transformed by SV40, does not hold for some cells of other origins transformed by SV40 (Stiles et al. 1976b; Topp et al. 1976, 1977) or by other viruses (Wolf and Goldberg 1976) or, indeed, for rat embryo fibroblasts transformed by adenovirus 2 (Gallimore et al. 1977). In these systems, not all anchorageindependent cells are tumorigenic, but so far, all tumorigenic cells tested have proved in vitro to be anchorage-independent (Kahn et al. 1980). In short, there is no single property conferred by viral transformation that universally correlates with enhanced tumorigenicity. 6

VIRAL GENES INVOLVED IN TRANSFORMATION

The entire SV40 genome is not required to bring about transformation. Infections with fragments of SV40 DNA as small as that produced by digestion with restriction endonucleases Baml and Hpall (0.14-0.72 map units) have produced transformation of cells in culture (Graham and van der Eb 1973), although fragments slightly smaller (0.14-0.67 map units or 0.28-0.72 map DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

4 / Transformation by SV40 and Polyoma Virus

241

units) do not. Furthermore, conditional mutations in genes coding for late proteins cause levels of transformation in nonpermissive conditions indistinguishable from that of wild-type virus (Eckhart 1969; di Mayorca et al. 1969; Kimura and Itagaki 1975; Tegtmeyer 1975; Martin and Chou 1975; Anderson and Martin 1976; Noonan et al. 1976), as do deletion mutations mapping in the late region. From these data it is clear that it is the early region of polyoma virus and SV40 that is required, in all or in part, for transformation. To analyze the contribution that the several early viral gene products (two for SV40, three for polyoma virus) make to the establishment and maintenance of the transformed state, extensive use has been made of deletion (dl) mutants and temperature-sensitive mutants, tsA or d/54/59 or hr-t (see Chapters 3 and 5). There have been indications that the method used to select transformants may influence whether the presence of large T antigen is required to maintain cells in a transformed state. As we have mentioned, Seif and Cuzin (1977) analyzed rat cells transformed by the tsA mutant of polyoma virus and identified two types of transformants, depending on the selection pressure used. Type-N transformants, isolated on the basis of their ability to form dense foci on plastic surfaces, were found to exhibit a temperature-sensitive phenotype; type-A transformants, selected for their ability to grow in agar, displayed the transformed phenotype at both restrictive and permissive temperatures. The frequencies of appearance of the two types of clones were approximately equal. In type-N cells the transformed phenotype can be considered as primarily, if not exclusively, under the control of the viral A gene. In type-A cells, however, it would seem that a functional A-gene product is not required for maintaining the transformed state. Seif and Cuzin (1977) suggest that type-A cells may have been modified, under the influence of the viral genome, to express permanently the transformed phenotype. There is dispute about the role of the A gene of SV40 in the maintenance of transformation. Several authors have carried out experiments which, in aggregate, strongly suggest that cells transformed by ts mutants in the SV40 A gene display a temperature-sensitive phenotype for at least some of the marDNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

242

DNA Tumor Viruses

kers of transformation (Kimura and Itagaki 1975; Martin and Chou 1975; Tegtmeyer 1975; Brugge and Butel 1975; Osborn and Weber 1975). All these experiments confirm that mutants are temperature-sensitive for the initiation of transformation. The cells studied include primary rat (Osborn and Weber), a continuous rat line (Kimura and Itagaki), the continuous mouse cell lines Swiss 3T3 (Tegtmeyer) and Balb/3T3 (Brugge and Butel), Syrian hamster embryo (Tegtmeyer; Brugge and Butel), Syrian hamster tumor cells (Brugge and Butel), Chinese hamster lung cells from a single animal (Martin and Chou), rabbit kidney (Tegtmeyer; Brugge and Butel), and human cells from a patient with Franconi anemia (Brugge and Butel). The Chinese hamster lung, rabbit kidney, and human cells are semipermissive for SV40, whereas all the other cells are nonpermissive. As if this variability were not enough, each of the authors used a different procedure to transform the cells. For example, most of the transformed lines were obtained by incubating the infected cells at 33°C, the permissive temperature for the ts mutants. Brugge and Butel, however, used a temperature of 37°C when transforming hamster embryo and human cells by an A-gene mutant and its corresponding wild type. Individual transformed clones isolated at the permissive temperature (except for the wild-type clones of Kimura and Itagaki) were used by Martin and Chou, Osborn and Weber, and Kimura and Itagaki. Tegtmeyer and Brugge and Butel used uncloned populations of cells which, after infection with virus and subculture at the permissive temperature, were deemed to be transformed by their morphology and the presence of T antigen. Thus, these two groups looked at mixed populations of transformed cells. Different criteria for transformation were analyzed by the different groups. These included morphology (Brugge and Butel; Tegtmeyer; Osborn and Weber); saturation density (Osborn and Weber; Brugge and Butel); colony formation at low cell density in high serum (Tegtmeyer; Brugge and Butel), in low serum (Martin and Chou), in agar (Kimura and Itagaki; Brugge and Butel), and on top of a monolayer of untransformed cells (Brugge and Butel; Martin and Chou); growth rate in low serum (Kimura and Itagaki); presence of T antigen (Osborn and Weber; Tegtmeyer; Brugge and Butel); and increased uptake of hexose (Brugge and Butel). DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

4 / Transformation by SV40 and Polyoma Virus

243

Using six different A-gene mutants, Martin and Chou found eight out of nine of their clonal lines transformed by tsA mutants to be temperature-sensitive for colony formation at low cell density in low serum and for their ability to form morphologically distinct colonies on top of untransformed monolayers. Kimura and Itagaki found that three independently transformed clones, transformed by one A-gene mutant, were temperature-sensitive for colony formation in agar and on top of untransformed cell monolayers. They also reported that three clones grew less well in low serum (but not in high serum) at the nonpermissive temperature. Osborn and Weber presented data for only one clone transformed by an A-gene mutant and one wild-type-transformed clone derived from primary rat kidney cells. They observed that the clone transformed by the A-gene mutant was temperaturesensitive for morphology, alignment of actin fibers, saturation density, and the presence of T antigen. Moreover, the number of cells transformed by A-gene mutants and containing T antigen decreased (15% positive) after 3 days at the nonpermissive temperature. Tegtmeyer, and Brugge and Butel, however, did not see any decrease in such cells at the nonpermissive temperature. Brugge and Butel used a variety of parameters in their study of cell lines transformed by A-gene mutants, including morphology, saturation density, cloning in agar and on top of untransformed monolayers, the uptake of hexose, and the presence of T antigen. They found that all lines tested were temperature-sensitive for these parameters, except for one human line that was not temperature-sensitive for the increased uptake of hexose; furthermore, all the lines retained T antigen at the high temperature. Tegtmeyer's results were somewhat in disagreement with those of Brugge and Butel, although the same A-gene mutants, and in some cases the same cell types and parameter of transformation, were used. Tegtmeyer found that none of his mouse lines transformed by A-gene mutants were temperature-sensitive for colony formation in high serum (rabbit kidney cells transformed by five other A-gene mutants were not temperature-sensitive). The temperature of 41.5°C used by Tegtmeyer for his rabbit kidney cell experiments was the highest DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

244

DNA Tumor Viruses

nonpermissive temperature used by any of these workers. Tegtmeyer did find that hamster cell lines transformed by five of six A-gene mutants were temperature-sensitive for colony formation at low cell density, but not at high cell density. All of Tegtmeyer's lines transformed by A-gene mutants, retained T antigen at the high temperature. All mouse, rabbit, or hamster cells that were transformed by A-gene mutants and grew at high temperature retained their transformed morphology. Tegtmeyer concluded that the temperature sensitivity of the different transformed phenotypes induced by A-gene mutants may well depend on the particular combination of virus and cell under study. More recently, Brockman (1978) has investigated the role of the A gene of SV40 in transformation of Balb/c-3T3 cells by examining the temperature sensitivity of the transformed phenotype of clones initially transformed at the permissive temperature with each of six tsA mutants whose sites of mutation lie at different positions in the early region of the SV40 genome. Almost all of the clones examined were unable to grow at nonpermissive temperature in soft agar, or in the presence of low concentrations of serum, nor could they form colonies on a monolayer of untransformed cells. These results, together with the observation that cells transformed by tsA mutants can be retransformed by wild-type virus at nonpermissive temperature, indicate strongly that the temperature sensitivity of Balb/c-3T3 tsA transformants is due to a viral defect and that the continued expression of the SV40 A gene is required for maintenance of the transformed state (Brockman 1978). No temperature-sensitive mutations have been obtained in the other portions of the early region, encoding the small T antigen of SV40 and the small and middle T antigens of polyoma virus. Therefore, no evidence is available to indicate whether these polypeptides are required to maintain the transformed phenotype. It is clear, however, that they are required for initiation of transformation, at least in some assays with certain cell types. For example, when rat cells are infected by deletion mutants of SV40, which contain mutations that map in the region of the genome coding for small T antigen, no transformants are obtained that can grow in Methocel (Bouck et al. 1978; Sleigh et al. 1978). However, if the same cells are plated on culture dishes, dense DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

4 / Transformation by SV40 and Polyoma Virus

245

colonies of transformed cells appear at a frequency not significantly different from that obtained with cells infected with wild-type virus (Shenk et al. 1976; Sleigh et al. 1978). These cells (transformed by the deletion mutants), when isolated from the culture dishes, fail to grow in Methocel (Sleigh et al. 1978). Exactly how small T antigen acts is unknown, but it presumably carries out functions distinct from those of large T antigen, because complementation can be observed between the viable deletion and tsA mutants (Bouck et al. 1978) when transformation is assessed by growth in agar. From these data, conventional wisdom would dictate that small T antigen is required for expression of at least some aspects of the transformed phenotype. Unfortunately, however, this conclusion appears to be true only for certain types of cells. If hamster cells are used in place of rat cells, the differences in transformation frequency between the deletion mutants and wildtype virus largely disappear, and the resulting lines of transformed cells are virtually indistinguishable in their properties. From all these results, no simple pattern emerges. No single viral gene product can be shown to be causally involved in the establishment or maintenance of transformation. Instead, the phenomenon appears to result from the complex interplay between host gene products and at least two viral gene products whose relative importance to the transformation process appears to differ from cell line to cell line and from assay to assay.

VIRAL GENES IN STABLY TRANSFORMED CELLS

Once cloned lines of transformed cells have been established, it is usually impossible to find in them any evidence of infectious SV40 or polyoma virus. However, viral DNA is invariably present and can be detected by a variety of procedures involving nucleic acid hybridization. The amount of viral DNA and its physical state depend on the particular combination of cell and virus, the simplest of which is rodent cells transformed by SV40. Such cells, except when physically injected with massive quantities of SV40 DNA (Graessmann et al. 1976), are not permissive for SV40 infection, and their transformed derivatives do not produce infectious virus unless they are fused with permissive DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

246

DNA Tumor Viruses

simian cells. Viral DNA is present in only small quantity (1-10 genome equivalents per diploid quantity of cellular DNA) and is covalently integrated into high-molecular-weight cellular DNA. Only early viral gene products are expressed. In contrast, the interaction between polyoma virus and rat cells is more complex. The cells do not respond uniformly to the virus, and a small proportion of them synthesize capsid antigen after infection at high multiplicity (Prasad et al. 1976), The transformants that arise subsequently commonly contain many copies (10-50) of viral DNA per diploid quantity of cellular DNA. The majority of this viral DNA exists in a nonintegrated state, although integrated viral sequences also are present. A few, exceptional, transformed rat cell lines have been described which spontaneously produce small amounts of polyoma virus (Fogel and Sachs 1969, 1970; Zouzias et al. 1977). In nearly every case, however, fusion with permissive murine cells is required to induce production of virus (Prasad et al. 1976). The molecular events associated with these phenomena have been under investigation for about 10 years. In the remainder of this chapter, we will summarize the major findings that have emerged.

Amount of Viral DNA in Transformed Cells

The amount of viral DNA in cells transformed by polyoma virus and SV40 has been measured by two techniques. The first method, worked out by Westphal and Dulbecco (1968), involves synthesis in vitro of highly radioactive RNA from purified viral DNA using Escherichia coli RNA polymerase. The RNA is hybridized to DNA extracted from transformed cells or control DNAs immobilized on nitrocellulose filters, and the number of copies of viral DNA per transformed cell is normally calculated from reconstruction experiments using mixtures of DNA from untransformed cells and known quantities of viral DNA. The method is extremely sensitive and is capable of detecting as little as one part of SV40 DNA in 5 x 10 parts of cellular DNA (equivalent to about 0.2 copies/cell). Westphal and Dulbecco (1968) showed that different lines of cells transformed by SV40 contained different numbers of viral genome equivalents, rang6

DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

4 / Transformation by SV40 and Polyoma Virus

247 12

ing from about 10 to 30 copies per diploid quantity (3.9 x 10 daltons) of mammalian cell DNA. Cells transformed by polyoma virus contained rather less viral DNA—from 4 to 10 copies per diploid quantity of cellular DNA. Westphal and Dulbecco's (1968) experiment was extremely influential in that it provided the first direct demonstration that viral sequences could be detected in DNA extracted from transformed cells. However, for reasons that became clear only later, their estimates of the number of viral copies per cell were wrong. First, Westphal and Dulbecco always found hybridization between the probe RNA and DNA extracted from untransformed cells. Why this hybridization should occur remained unknown until 1972, when Lavi and Winocour (1972) showed that the growth of SV40 at high multiplicities results in the production of viral particles containing covalently linked sequences of host DNA and viral DNA. Unwittingly, therefore, Westphal and Dulbecco used as template preparations of viral DNA that almost certainly contained host sequences; in retrospect, it is in no way surprising that the RNA hybridized to DNA from untransformed cells. Clearly, the problem can be circumvented by using, as probe, RNA synthesized off viral DNA from stocks of virions from cells infected at low multiplicity. A further criticism of the filterhybridization experiment concerned the fidelity of the reconstruction experiment. It has been claimed that under certain conditions DNA-RNA hybrids are lost selectively from nitrocellulose filters (Haas et al. 1972), an effect that conceivably could lead to an inflated estimate of the number of copies of viral DNA per transformed cell. The potential problem can be solved by using as probe saturating quantities of RNA whose specific activity is accurately known (Botchan et al. 1974). From the number of counts bound, it is possible to calculate directly the amount of viral DNA present in the genome of transformed cells without resorting to reconstruction experiments. In this way it has been shown that cell lines transformed by SV40 contain between 1 and 10 copies of viral DNA per diploid equivalent of cellular DNA. Despite the important results obtained, the filterhydridization method fell into disuse when Gelb et al. (1971) showed that the number of copies of viral DNA in the genomes of transformed cells could be determined by measuring the rate of reannealing of small amounts of radioactive DNA in the DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

248

DNA Tumor Viruses

presence of large quantities of unlabeled DNA from transformed cells. This technique depends on the fact that the rate of reannealing of a given DNA sequence in solution is proportional to its initial concentration (Wetmur and Davidson 1968; Britten and Kohne 1968). The number of viral genomes per transformed cell is generally calculated from the magnitude of the increase in the rate of hybridization of labeled probe DNA in the presence of DNA from transformed cells, compared with control DNA. By using this type of analysis, it has been shown that different lines of mouse cells transformed by SV40 contain viral sequences that range in amount from 1.1 copies to as many as 8-10 copies of SV40 DNA per diploid equivalent of cellular DNA (Gelb et al. 1971; Ozanne et al. 1973). Mouse cells transformed by polyoma virus contain between 0.6 and 2.9 copies of the viral genome per diploid equivalent of cellular DNA (Kamen et al. 1974); rat cells contain many more (up to 50 copies) of the viral DNA (Zouzias et al. 1977). In calculating these figures, it was assumed that all sequences of the labeled probe are present in the transformed cells at equal frequencies. However, if transformed cells were to contain only a part of the viral genome or if some parts of the viral genome were present at much higher frequency than others, the observed rate of reannealing would be an average of the rates of each independent segment of DNA. The best way to resolve this problem is to use as probes in kinetic hybridization experiments defined segments of viral DNAs produced by cleavage of the intact viral genomes with restriction enzymes. When this sort of experiment was carried out with the SV40-transformed mouse cell line, SVT2, which by conventional analysis had been shown to contain between 1.56 and 2.2 genome equivalents of viral DNA per diploid quantity of cellular DNA (Gelb et al. 1971; Ozanne et al. 1973), it was found that the early region of the SV40 genome was present in the transformed cells at a frequency of about 6.0 copies per diploid quantity of cellular DNA and that the late region was represented only once (Botchan et al. 1974). By the end of 1974, it had been established that there is a variable quantity of viral DNA present in cells transformed by polyoma virus or SV40 and that the amount of viral DNA could range quite widely from cell line to cell line. Furthermore, there DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

4 / Transformation by SV40 and Polyoma Virus

249

was no requirement that all segments of the viral DNA be present in equimolar quantities. The next step was to come 2 years later, when methods became available to study the arrangement of viral sequences in transformed cells.

Arrangement of Viral Sequences in Transformed Cells

Nonpermissive Cells The physical state of viral DNA in nonpermissive cells transformed by SV40 was established many years ago. That highmolecular-weight, transformed cellular DNA continues to hybridize to viral probes after zonal sedimentation in alkaline gradients and equilibrium centrifugation in cesium chloride gradients containing ethidium bromide (Sambrook et al. 1968) argues strongly that the viral and cellular DNA sequences are linked by covalent bonds. A similar conclusion was reached later for polyoma virus DNA sequences in transformed hamster cells (Shani et al. 1972; Folk 1973; Manor et al. 1973). For two technical reasons, however, details of the arrangement of the integrated viral sequences were very much more difficult to obtain. On one hand, the detection of small quantities of viral DNA in transformed cells demands the application of stringent and sensitive hybridization techniques; on the other hand, the size of the mammalian genome is so vast, and its organization so complex, that few reliable methods have been available for its fractionation. Recently, both of these problems have been at least partially overcome. Ffigh-moleeular-weight cellular DNA can be cleaved at specific sites with restriction endonucleases, and the resulting fragments can be fractionated by electrophoresis through agarose gels, denatured in situ, and transferred to nitrocellulose sheets (Southern 1975); highly purified viral DNA can be labeled with P to specific activities greater than 10 cpm/jtxg (Rigby et al. 1977). Thus, the presence of as little as 10" g of viral DNA can be detected among the different-sized fragments of DNA from transformed cells by hybridization and autoradiography of the nitrocellulose sheet. By analyzing 5 fig of cellular DNA, it is possible to recognize as few as 0.02 copies of SV40 DNA per diploid equivalent of cellular DNA (Ketner and Kelly 1976; Botchan et al. 1976). 32

8

13

DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

250

DNA Tumor Viruses

A variety of restriction endonucleases have been used and can be classified into three groups for the purposes of these experiments (see Fig. 4.3). First, nucleases such as Ball and Smal find no recognition site within the SV40 genome and therefore cleave transformed-cell DNA only at sites within cellular sequences. By counting the number of bands that contain SV40 sequences after digestion of transformed-cell DNA with these enzymes, minimum estimates can be obtained of the number of separate Isolate h i g h - m o l e c u l a r - w e i g h t

cellular

DNA

Cleave DNA w i t h a r e s t r i c t i o n enzyme

Class I

Class 2

Separate f r a g m e n t s by

a

b

a

c

Class 3

electrophoresis

b

a

c

b

c

Blot; hybridize w i t h l a b e l e d S V 4 0 ; autoradiogram •

Figure

4.3

S c h e m e f o r d e t e r m i n i n g t h e a r r a n g e m e n t o f S V 4 0 s e q u e n c e s in t h e

D N A of transformed cells. T h e diagram s h o w s the results e x p e c t e d w h e n the DNA

from

transformed

c e l l s is a n a l y z e d

using three

types of

restriction

e n z y m e s . Class-1 e n z y m e s d o not cleave S V 4 0 D N A , class-2 e n z y m e s cleave t h e viral D N A o n c e , a n d c l a s s - 3 e n z y m e s c l e a v e it s e v e r a l t i m e s , ( a ) T r a n s f o r m e d - c e l l D N A ; (b) c o n t r o l D N A ; (c) r e c o n s t r u c t i o n . ( M o d i f i e d f r o m B o t c h a n e t al. 1 9 7 6 . ) DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

4 / Transformation by SV40 and Polyoma Virus

251

insertions of viral DNA into the genomes of transformed cells. Enzymes of the second class (EcoRl, Bgll, and Bam I), each of which cleaves SV40 DNA once, are most useful for detecting the presence in transformed cells of partially duplicated or tandem copies of SV40 DNA. These enzymes should cleave within repetitious viral sequences to free from them a copy of the viral genome that will migrate through agarose gels at the same rate as linear (form-Ill) SV40 DNA. Finally, enzymes such as Hpal and HmdIII cleave SV40 D N A at several sites to yield fragments whose size and location on the map of the viral genome are well established. The results obtained with two enzymes, Ball and Hpal, are shown in Figure 4.4.

Figure

4.4

D e t e c t i o n o f D N A f r a g m e n t s (a, b, c) c o n t a i n i n g S V 4 0

after h y d r o l y s i s

of S V 4 0 - t r a n s f o r m e d

sequences

rat c e l l D N A b y e n d o n u c l e a s e

( R e d r a w n f r o m B o t c h a n e t al. 1976.)

DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

Hpal.

252

DNA Tumor Viruses

Using endonuclease Ball, which does not cleave SV40 DNA, 7 of 11 isogenic lines of SV40-transformed rat cells were found to contain more than one insertion of viral DNA; in the case of clone 5, there were at least six separate insertions. The molecular weights of the fragments of cellular DNA containing SV40 DNA differed from cell line to cell line and ranged from 2 x 10 (equivalent in length to 60% of the SV40 genome) to a maximum of several times the size of the viral genome. The DNAs of the remaining four cell lines of transformed cells (clones 14B, 17, 12, and 19) yield, after cleavage by Ball, only one detectable band that hybridizes to SV40 DNA. Thus, the genomes of these cells contain a single insertion of SV40 DNA. In no case, however, are the SV40 sequences present in DNA fragments of the same size. From these and similar analyses carried out with other restriction endonucleases, the following observations and conclusion emerge. The sum of the molecular weights of all SV40-containing fragments in any given digest of transformed-cell DNA is always greater than that of unit-length SV40 DNA. Different restriction endonucleases yield from any one cell line sets of SV40-containing fragments that add up to different molecular weights. These data prove that SV40 sequences in transformed rat cells are contained within a segment of DNA that is considerably larger than the SV40 genome, and thereby provide independent confirmation that the viral sequences are carried in an integrated state. The complex patterns of hybridization found with the DNAs of many transformed cell lines are not a consequence of heterogeneity in the cell population, because recloning the cells does not result in any alteration of the patterns of hybridization. These data provide strong evidence that SV40 DNA can integrate at several sites into the genome of a single transformed cell. Because the SV40 sequences in "single-copy" cell lines are not present in DNA fragments of the same size, integration of viral DNA occurs at different chromosomal sites in different cell lines. The chromosomal locations of SV40 DNA within any one cell are stable; no change in hybridization pattern is detected during growth of the cells for more than 200 generations. 6

DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

4 / Transformation by SV40 and Polyoma Virus

253

The approximate positions of the integration sites on the SV40 genome have been determined for the four single-copy cell lines by using restriction enzymes that cleave SV40 DNA several times at known locations (Fig, 4.5). Three of the four single-copy lines contain a segment of viral DNA that is of unit length or close to it, with junctions between cellular and viral DNA sequences mapping at different sites in the "late" region of the viral genome. The fourth of these cell lines contains about 1.5 copies of viral DNA. One end of the viral array maps near the site at which DNA synthesis originates; the other is located in the "early" region of the SV40 genome. This partial duplication of viral DNA sequences provides an intact copy of the early gene, whose continued expression is necessary for maintenance of the transformed state. Thus, in each of four independently transformed cell lines, the viral integration sites are not identical. It seems that there is no single position on the viral genome that serves as an obligatory site of connection of SV40 DNA to the cellular chromosome. It becomes obvious from these experiments and from those of Ketner and Kelly (1976), who carried out a similar analysis of the viral DNA sequences integrated into transformed murine cells, that integration of SV40 DNA into the genome of mammalian cells is quite different from that of bacteriophages like A into JE. coll Instead of the tidy insertion of unit-length viral DNA molecules into a defined chromosomal site, mediated by specific enzymes, integration of SV40 DNA sequences of various lengths can occur at many places in the cellular DNA. It is a random or quasi-random process. The conclusion that there are many possible sites of integration of SV40 DNA in the cellular genome is borne out by analysis of somatic-cell hybrids. Several groups of workers have exploited the fact that human chromosomes are shed preferentially from hybrids made by fusing mouse cells with SV40transformed human cells. In early experiments of this type, Weiss (Weiss et al. 1968; Weiss 1970) was unable to correlate the loss of T antigen with the loss of any specific chromosome and concluded that the viral DNA might integrate at more than one site. On the other hand, Croce and his coworkers, who examined hybrids between mouse cells and two lines of SV40transformed human cells, were able to correlate the expression of viral functions with the presence of a specific human DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

DNA Tumor Viruses: The Molecular Biology of Tumor Viruses, 2nd Ed. © 1980 Cold Spring Harbor Laboratory Press 0-87969-141-7 For conditions see www.cshlpress.com/copyright.

\ 14 \

/ 4 0 /38

50

80

60

20

90

70

0

80

40

30

A •

6

60

0

6

6

20

(046 xiO )

10

(3.2 x l O )

90

O

50

30

10

70

c3

111

•• A

(7.1 x I 0 !

50

6

40

III

a

(2,1 x l O )

I I 1

IDA

30

A

1I

• A •

10

o

40

20

A

80

I



60

50

30

90

I 70

60

40

6 6 V 76

50

76V86