Multiple effects of the 72-kDa, adenovirus-specified DNA binding protein on the efficiency of cellular transformation

VIROLOGY

156,

366-376

(1987)

Multiple Effects of the 72-kDa, Adenovirus-Specified DNA Binding Protein on the Efficiency of Cellular Transformation STEPHEN A. RICE,*,’ DANIEL F. KLESSIG,*a2 AND JIM WILLIAMSt*3 *Department

of Cellularz

Viral and Molecular Biology, University of Utah, School of Biological Sciences, Carnegie Mellon University, Received

July 28,

1986;

accepted

of Medicine, Salt Lake Pittsburgh, Pennsylvania October

City, Utah 152 13

84 132, and tDepartment

16, 1986

The early region 2A gene (E2A) of adenovirus types 2 and 5 encodes a 72-kDa DNA binding protein (DBP) which contains two physical domains comprising approximately the amino-terminal one-third and carboxyl-terminal twothirds of the protein, respectively. Previous work has shown that some Ad5 mutants containing temperature-sensitive (ts) mutations in the carboxyl-terminal domain of DBP. such as AdSsl25, show a 3- to B-fold enhanced ability to transform rat cells. We have examined the transformation characteristics of a series of Ad5 E2A deletion mutants, Ad5d/801-5, which encode either no functional DBP or encode truncated, defective DBPs. The E2A deletion mutants transformed rat embryo cells at frequencies similar to wild-type (wt) Ad5. These results suggest that the high transformation phenotype of carboxyl-terminal E2A mutants like Ad5ts125 is not due to the simple inactivation of DBP function, but rather results from an activity possessed by an altered DBP. This hypothesis is supported by the fact that the transformation phenotype of Ads&125 and similar mutants is dominant over the wild-type phenotype. A number of additional Ad2 and Ad5 E2A mutants were examined with respect to their ability to transform primary rat embryo cells. It was found that a carboxyl-terminal E2A mutant, Ad2+NDi rs23, also showed the enhanced transformation phenotype. In contrast, several amino-terminal E2A host-range (hr) mutants, originally isolated on the basis of their ability to replicate in monkey cells, transformed rat embryo cells at a frequency similar to wild-type virus. Ad2rs400, and E2A mutant with alterations in both DBP domains, showed a Wilde-type frequency of transformation, while two similar mutants, AdSrsl25X405 and Ad5tsl25X404, showed an enhanced frequency. Last, it was found that coinfection of primary rat embryo cells with the hr mutants plus Ad5t.H 25 or Ad2+NDi ts23 resulted in a wild-type frequency of transformation, demonstrating that the hr mutants are dominant to the rs mutants with regard to transformation phenotype. Thus, DBP can both positively and negatively affect viral transformation in this system. o 1997 Academic

INTRODUCTION

rodent cells when introduced by transfection (Graham et a/., 1974; van der Eb et a/., 1977). At least two additional adenovirus functions can also influence cellular transformation when the ElA and El B genes are introduced into cells by viral infection. One of these is the 140-kDa DNA polymerase encoded in early region 2B (E2B) between viral coordinates 18.5 and 22.0 (Galos et a/., 1979; Stillman et a/., 1982). Temperature-sensitive (6) Ad5 mutants which contain lesions in the polymerase gene are defective at the nonpermissive temperature for the initiation but not the maintenance of the transformed state (Williams et al., 1974, 1979). The second viral function which affects transformation is the 72-kDa DNA binding protein (DBP) encoded in early region 2A (E2A) between viral map units (m.u.) 61.6 and 66.5 (Lewis et a/., 1976). Two independently derived mutants, Ad5tsl25 and Ad5tslO7, contain an identical mutation in E2A which leads to a ts, DNA replication negative phenotype in human cells (Kruijer et a/., 198 1, 1983). In rat cells, the two mutants show a 3- to g-fold enhanced frequency of transformation compared to wild-type (wt) Ad5 at both nonpermissive and semipermissive temperatures

Human adenoviruses transform rodent cells in vitro, and the viral functions necessary and sufficient for cellular transformation are encoded by the early region 1A (El A) and early region 1 B (El B) genes which map in the left 1 1% of the viral genome (for reviews see Graham, 1984; Williams, 1986). This conclusion is based upon the observations that E 1A and E 1 B genes are found integrated into the host genome in transformed cells and are frequently the only viral sequences found (Gallimore et a/., 1974; Sharp et a/., 1974; Flint et a/., 1976) that viral mutations mapping in ElA and El B affect transformation potential (Graham et al., 1978; Jones and Shenk, 1979; Ruben et al., 1982; Ho et a/., 1982; Monte11 et a/., 1984; Babiss et al., 19841, and that purified El A and E 1B genes can fully transform

’ Present address: Department of Microbiology and Molecular Genetics, Harvard Medical School, Boston, MA. * Present address: Waksman Institute, Rutgers University, Piscataway, NJ 08854. 3 To whom requests for reprints should be addressed. 0042-6822187

$3.00

Copyright 0 1987 by Academic Press, Inc. All rights of reproduction in any form reserved.

366

ADENOVIRUS-MEDIATED

(Ginsberg er al., 1974; Williams et al., 1974; Fisher et a/., 1982). Revertants of Ad5tsl25 and Ad5fslO7 have been isolated based on their ability to replicate in human cells at the nonpermissive temperature (Carter and Blanton, 1978b; Nicolas eta/., 1981). Genetic mapping and DNA sequence analysis have shown that the revertants are either true revertants or contain second site point mutations in the E2A gene (Kruijer era/., 1983; Nicolas et a/., 1983). One class of intragenic second site revertants retains the enhanced transformation phenotype (Logan et a/., 1981; Carter et al., 1982) suggesting that the DBP activity required for DNA replication and the DBP activity which modulates cellular transformation may be functionally distinct. In this study we have examined the transformation phenotypes of a number of Ad2 and Ad5 E2A mutants, including recently isolated deletion mutants which either fail to make detectable DBP or make truncated, defective DBPs (Rice and Klessig, 1985). The results of this analysis demonstrate that the efficiency of adenovirus transformation can be influenced both positively and negatively by the DBP. MATERIALS

AND

CELLUCAR

TRANSFORMATION

367

mutants, which were isolated on the basis of their ability to grow in monkey cells have been described by us in a number of previous publications, as has Ad2ts400 (Klessig, 1977; Klessig and Grodzicker, 1979; Rice and Klessig, 1984). Physical maps of the mutations borne by these various mutants are shown in Fig. 1. Measurement

of plating-efficiency

To measure colony-forming ability or plating efficiency in virro, appropriate numbers (usually 1O3 cells) of live infected or uninfected cells were seeded on feeder cultures (Puck et a/., 1956). To prepare feeder cultures, monolayers of rat embryo cells first irradiated with 3000 rads of gamma rays (we thank Dr. Richard Galas, Department of Otolaryngology, Eye and Ear Hospital, Pittsburgh, for carrying this out). These were then seeded at lo5 cells per 60-mm dish in DMEM + 109/o fetal calf serum, and grown at 37’ for 24 hr prior to addition of live cells. Afterward the cultures were incubated at 37” for 2 weeks, and at this time they were fixed with 3% formaldehyde in PBS and stained with Geimsa to make colonies visible for counting.

METHODS

Cells and viruses HeLa cells used for growth and titration of wild-type, fs, and host-range (hr) viruses were derived from stocks previously used in studies of ts and hr mutants (Williams et a/., 1971; Harrison et al., 1977). The E2A-complementing cell lines gmDBP1 and gmDBP2 used for propagation of the E2Adl mutants were those described by Klessig et al. (1984). Monolayer cultures of Hela and gmDBP cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM) supplemented by either 7 or 10% calf serum. The 293 cells (Graham et a/., 1977) used in transfections for construction of recombinants (see below) were those we have previously used for isolation and growth of El hr mutants (Harrison et al., 1977) and the CVl cells used to select for recombinants able to grow in monkey cells were originally obtained from Dr. J. Mertz. The adenovirus 5 wild-type strain was derived from the same seed stock previously used in all of our genetic and biochemical studies (Williams et al., 1971, 1974; Harrison et a/., 1977; Rice and Klessig, 1985). The type 2 virus was originally obtained from Dr. U. Pettersson and has also been used in some of our previous studies (Rice and Klessig, 1985). Ad5ts125 was originally obtained from Dr. H. Ginsberg (Ensinger and Ginsberg, 1972). Ad2+NDlrs23 was isolated by Lesley Fraser and J. Williams (Fraser and Williams, unpublished results). The E2A dl mutants were recently described in Rice and Klessig (1985) while the E2A hr

Transformation of rat embryo cells The transformation assay is essentially the same as those we have previously used (Williams and Ustacelebi, 197 1; Graham et a/., 1978; Williams et al., 1979). Primary rat embryo cultures were prepared by trypsinization of 16- to 18-day-old embryos taken from pregnant inbred Fisher rats, and were grown in 60-mm petri dishes in DMEM supplemented with 7% fetal calf serum. When confluent (3-4 days after seeding) these cells were infected with virus at an input multiplicity of 10 PFU/cell unless otherwise stated (see dose-response experiments, for example). Immediately after infection, the monolayers were split by trypsinization, and replicate dishes were seeded with lo6 cells in DMEM + 7% calf serum and incubated at the appropriate temperature. At 4 days after seeding the medium was removed and replaced with calcium-free DMEM (Freeman et al., 1967) + 7% CS, and for the duration of the experiment this medium was replaced at 3- to 4-day intervals. In Ad5tsl25 and Ad2+NDl rs23-infected cultures at 37 or 38.5”, foci first appeared around 14 days postinfection, while with wild-type and all other mutants, foci were usually visible 2-4 days later. Final counts of transformed foci were made at 22-25 days at 37 and 38.5”, and at 33-35 days at 32.5”. Derivation

of recombinant

viruses

AdEitsl25X405 and AdStsl25X404 were constructed by a marker transfer protocol in which purified

368

RICE, KLESSIG, AND WILLIAMS

DNAs containing DBP hr mutations were crossed into the Ad5ts125 genome. For construction of Ad5ts125x405, 1 pg of Ad5tsl25 DNA plus 20 M equivalents of a restriction fragment (viral m.u. 63.665.9) derived from Ad2ts400 which contains both Ad2hr405 hr mutations (Brough et al., 1985; Rice and Klessig, 1984) were transfected into 293 cell monolayers (60-mm plates) by the calcium phosphate technique (Graham and van der Eb, 1973). For construction of Ad5tsl25X404, 1 pg of AdStsl25 plus 3 M equivalents of HindIll-digested pAd5d/804 plasmid (Rice and Klessig, 1985; pAd5d1804 is a deletion derivative of the Hindill A fragment of Ad5hr404 cloned into pBR322, lacking sequences between m.u. 62.2 and 64.4) were used in the transfection. Transfected cells were incubated at 33” for 4 days and then frozen. The cell lysates were then passaged in monkey CV, cells for two infectious cycles to select for recombinants which had acquired the ability to grow in monkey cells. For Ad5ts125X405, the selections were at 33’, but for Ad5tsl25X404 they were at 35.5” since the growth of Ad5hr404 is cold sensitive in CV, cells. Virus from the second passage was then isolated as plaques on monkey cells, plaque purified again, and grown into large stocks. Both stocks were characterized by plaque assay on HeLa and CV, cells and were found to have the expected phenotypes, i.e., both were ts but were able to grow efficiently in monkey cells at the permissive temperature (33” for Ad5ts125X405, 35.5” for Ad5ts125X404). Several other plaque-purified isolates of each of the two recombinants were grown into smaller stocks consisting of crude cell lysates, and were also tested in the transformation assay.

RESULTS Ad5 E2A deletion mutants transform frequencies similar to wild-type Ad5

rat cells at

The temperature-sensitive adenovirus E2A mutants Ad5ts125 and Ad5ts107 transform rat cells at a 3- to 8-fold higher frequency than wild-type virus (Ginsberg et a/., 1974; Williams et a/., 1974; Fisher et a/., 1982) suggesting that the wild-type DBP carries out an activity in rat cells which suppresses transformation. We have recently isolated and characterized five additional Ad5 mutants which contain deletions in the E2A coding region (Rice and Klessig, 1985; see Fig. 1). All five mutants are absolutely defective for growth in human cells, and one of the mutants, Ad5d/802, makes no detectable DBP or DBP fragment and thus is a true DBPnegative mutant. The four other mutants, Ad5d/801 and Ad5d/803 through Ad5d/805, make DBP fragments which in some cases contain a carboxyl-terminal seg-

ment encoded by an alternate E2A reading frame. If DBP normally suppresses transformation in rat cells, then one would expect that Ad5d1802 and possibly the four other deletion mutants should express the hightransformation phenotype. To test this possibility, primary rat embryo cells were infected with wild-type Ad5 or various Ad5 mutants at a multiplicity of infection of 10 PFU per cell and incubated at 38.5, 37, or 32.5”. As expected, at nonpermissive (38.5”) and semipermissive (37”) temperatures, Ad5tsl25 transformed rat embryo cells at a 3-to 5-fold enhanced frequency compared to Ad5 (Table 1). Furthermore, Ad5tsl25-induced foci appeared 3-4 days earlier than other foci at these temperatures. An enhanced frequency of transformation by Ad5rsl25 was also observed at 32.5” in some experiments, although the difference between Ad5ts125 and Ad5 was less dramatic (2-to 3-fold). In contrast, the five E2A deletion mutants transformed rat cells at all three temperatures at frequencies that were very similar to those induced by wild-type Ad5. In some experiments, transformation by the deletion mutants was slightly elevated at 38.5” (about 1.5- to 2-fold) compared to wt Ad5. However, since the frequency of transformation by Ad5tsl25 was 4- to 5-fold higher than wild type at this temperature, the significance of this observation is unclear. Adenovirus-induced transformation of rat cells is dependent on the multiplicity of viral infection (Gallimore, 1974; Logan et al., 1981; Williams et a/., 1974). The transformation frequencies of wild-type and mutant viruses were therefore compared at 37” as a function of input viral multiplicity (Fig. 2). This experiment showed that at the multiplicities tested, 0.3-30 PFU per cell, Ad5tsl25 transformed at a 3- to 5-fold enhanced frequency while AD5d/802 transformed at wild-type frequency. We therefore conclude that lack of a functional DBP does not per se cause an elevation in the frequency of rat cell transformation.

The effects of DBP point mutations transformation efficiency

on viral

The transformation phenotypes of several previously untested Ad2 and Ad5 E2A point mutants were also examined. Figure 1 summarizes the E2A lesions in each mutant. The results of the transformation assays on primary rat embryo cells demonstrated that the E2A mutants fell into two general categories with respect to transformation (Table 2). The first class consists of ts mutants which are ts for DNA replication in human cells, and which transform at an elevated frequency compared to wild type. This class includes Ad5tsl25 and Ad2+NDlts23, a ts derivative of the adenovirussimian virus 40 hybrid Ad2+NDl (L. Fraser and J. Williams, unpublished results; Kruijer et a/., 1982). In these

ADENOVIRUS-MEDIATED

CELLULAR

TRANSFORMATION

ala toval148

413 pro to se, X

369

130histotyr

x x 148 130

<

x

<

2 hr405

x 130

5hr404

I

X

2hr400 282 leu to phe

2+NOlts23

< X

5ts125 66-7

64.4 w----------f-,

5dl805

64-4

62.2 m-------_-----_-----I

130 I

5dl804

X 62.2 ------------------------,-~-,

65-7 65.4 azQg&----a 65.3 Q&------a

5dl803 66-l 5dl802 66.2 5dl801

c-

Ad2 Ad5

---

1

62

1 I

63

I I

64

I I

I

66

66

---

map

units

FIG. 1. Physical map of mutations which alter E2A expression, and schematic representation of DBPs and DBP-related polypeptides predicted to be encoded by the mutants used in the transformation assays. The portion of the adenovirus genome which encodes the DBP in Ad2 and Ad5 is shown by the bottom two bars, with viral map coordinates indicated. The bars above represent the various DBP polypeptides, with carboxyl termini at the left (pointed) ends. Open bars denote Ad5 DBP protein sequence, and since the carboxyl-terminal sequence of the Ad2DBP (residues 181-529) is identical to the Ad5 sequence, these segments are also shown by open bars. The black bars denote the Ad2specific amino-terminal segment of the DBP (residues l-l 80) while cross-hatched bars indicate protein segments derived from an alternate reading frame. The dotted lines connecting bars indicate fusion of protein segments due to deletion of DNA in the genomes of the d/ mutants. The crosses below bars indicate the positions of DBP point mutations which give rise to mutants with ts or hr phenotypes; the amino acid changes resulting from the mutations are shown above the crosses. Note that although Ad5d1801 and Ad5c/802 are serotypicallytype 5, their DBP genes were derived from type 2 (Rice and Klessig, 1985).

experiments, Ad2+ ND1 ts23 transformed rat cells at a 5- to g-fold enhanced rate compared to its phenotypically wt parent, Ad2+NDl. This phenotype was also quite evident at the permissive temperature of 32.5” where Ad2+NDlts23 showed a 6-fold enhanced frequency of transformation. The remaining E2A point mutants transformed rat cells at all three temperatures at frequencies that were approximately equal to, or slightly lower than, Ad2 and Ad5. This group included E2A hr mutants, which have been isolated based on their ability to replicate efficiently in monkey cells (Klessig, 1977; Klessig and Grodzicker, 1979; Rice and Klessig, 1984). The mutations in these hr mutants all are found in the aminoterminal segment of the DBP gene (Brough eta/., 1985).

Ad2ts400, a DBP mutant with both a carboxyl-terminal ts mutation, and two amino-terminal hr mutations (Rice and Klessig, 1984) also transformed at wild-type or slightly higher frequency of transformation. This result was surprising given that this mutant contains the same carboxyl-terminal E2A mutation as Ad5tsl25 (Brough et al., 1985) and displays a similar phenotype at the nonpermissive temperature (Rice and Klessig, 1984). The lower frequencies shown by Ad2rs400 were not dependent on the multiplicity of infection, since Ad2ts400 transformation frequencies were similar to (but slightly higher than) Ad2 over a wide range of multiplicities (0.3 to 30 PFU per cell; see Fig. 3). It should be noted that hr 405 transformed at lower than wildtype frequencies over this multiplicity range.

370

RICE, KLESSIG. AND WILLIAMS TABLE 1

al., 1981). Ad5ts125X405

TRANSFORMATION OF RAT EMBRYO CELLS BY Ad!%1 AND Ad5 E2A DELETION MUTANTS

25

No. of foci per dish (1 O6 cells seeded) Expt

Virus

32.5’

37”

1.

38.5”

None Ad5 Ad5fsl25 Ad5-81 Oc Ad5d/80 1 Ad5d1803 Ad5dl804

0, 0, 0

0, 0, 0

0, 0, 0

6, 6, 5 14, 15, 17 8, 7, 6 8, 8, 6 6, 5, 6 5, 3, 4

4, 4, 7 20, 21, 18’ 4, 3, 3 6, 5, 7 4, 4, 3 4,4, 3

5, 5, 6 22, 18, 20* 4, 6, 7 6, 7, 9 5, 6, 6 7, 8, 7

2.

None Ad5 Ad5tsl25 Ad581 Oc Ad5d/80 1 Ad5dl802 Ad5d1803 Ad5dl804

0, 0, 0 6, 5, 5 10,9, 12 6, 5, 5 6, 7, 9 7, 5, 6 6, 8, 6 6, 7, 7

0, 0, 0 5, 5, 4 21,19,206 3, 4, 5 6, 5, 4 5, 4, 6 6, 5, 4 4, 4, 6

0, 0, 0 4, 4, 5 22, 20, 17b 6, 4, 3 7, 9, 12 7, 7, 8 6, 8, 11 4, 5, 6

3.

None Ad5 Ad5fsl25 Ad5d/80 1 Ad5dI802 Ad5dl805

0, 0, 0 8, 5, 5 10, 10, 8 7. 10,9 11, 6, 8 8, 9, 10

0, 0, 0 4, 5, 4 19,20, 21b 6, 4, 4 4, 4, 5 5, 5, 6

0, 0, 0 4, 5, 4 15,22, 18’ 7, 9, 9 8, 9, 5 8, 6, 7

a Transformation assays were performed as described under Materials and Methods. The three numbers at each point in the table correspond to the number of foci in each of the three replicate plates, b Foci appeared 3-4 days earlier than foci induced by wild type. ’ Ad5-810 is a phenotypically wild-type virus which contains Ad2 DNA sequences between viral coordinates 61.3 and 7 1.4 It serves as a control for the mutants Ad5d/801 and Ad5d1802 which contain the same Ad2 substitution.

The effects of the host-range transformation efficiency

DBP mutations

on

The unexpected transformation phenotype of Ad2ts400 suggested that the N-terminal DBP hr mutations may act to negatively modulate the enhanced transformation frequency displayed by the TSDBP mutants. To test this possibility, two recombinant virus stocks were constructed by a marker transfer technique (see Materials and Methods for details) in which hr mutations were introduced into an Ad5ts125 genetic background. One virus, designated Ad5tsl25X405, contains Ad2hr405 DNA recombined into the aminoterminal segment of the E2A gene between viral map units 63.6 and 65.9. The second virus, designated Ad5tsl25X404, contains a DBP amino-terminal segment derived from Ad5hr404. This virus differs from Ad5ts125 only in containing the histidine to tyrosine alteration at amino acid 130 of DBP which confers on the virus the ability to grow in monkey cells (Kruijer et

and Ad5ts125X404 were tested in the transformation assay, as well as several other independent isolates of these recombinants which were obtained during the process of their construction (Table 3). Two isolates of Ad5tsl25X405, both retained the high transformation phenotype of Ad5fsl25, showing a 6-fold enhanced ability to transform the rat cells compared to wild-type Ad5. Likewise, the three isolates of AdStsl25X404 stock also showed an enhanced transformation frequency compared to Ad5, although the difference was slightly less (about 3-to 4-fold). Thus, the presence of the hr mutations in cis to a C-terminal DBP ts mutation does not necessarily decrease the efficiency of viral transformation. The effect of the hr mutants in tram was also studied, by coinfecting primary rat embryo cells with different combinations of ts and hr mutants. In this experiment, coinfections were also performed with ts mutants and Ad5 wild-type virus. The results of this experiment (Table 3) clearly demonstrate two points. First, the transformation phenotypes of Ad5tsl25 and Ad2+NDl ts23 are dominant over the wild-type phenotype. A second and somewhat surprising finding is that the transformation phenotype of the hr mutants is dominant to the phenotype of Ad5tsl25 and Ad2+NDl ts23. Three independently derived hr mutants, Ad2hr401, Ad2hr405, and Ad5hr404, when added at equal input multiplicity, all were able to decrease the transformation frequencies of the ts mutants to the wild-type frequency. Plating efficiencies of infected and uninfected primary rat embryo cells One possible explanation for the lower efficiency of transformation in cells infected by the hr mutants is that those mutants grow more efficiently in the cells of primary rat embryo cultures, killing more of them and leaving fewer available for transformation. To test this hypothesis, the plating efficiencies of infected and uninfected primary rat embryo cells were compared (Table 4). In all cases, including uninfected cells, the plating efficiency of the rat cells was between 8.3 and 10.7%. Thus, the ability of the hr mutants to decrease the frequency of transformation in Vans does not appear to be due to an enhanced ability of the hr mutants to grow lytically in the primary rat embryo cells, and kill them. These tests do not, however, rule out the possibility that a small, specific subset of cells, which are targets for transformation, are preferentially killed by these mutants. DISCUSSION These experiments confirm previous studies (Ginsberg et a/., 1974; Williams et a/., 1974; Logan et a/.,

ADENOVIRUS-MEDIATED

INPUT

MULTIPLICITY

CELLULAR TRANSFORMATION

371

I P W./CELL1

FIG. 2. Relationship between input multiplicity and transformation frequency in primary rat embryo cells infected by Ad5 (0) rs 125 (O), and Ad5d/802 (A). The transformation procedure is described under Materials and Methods. Incubation was at 37”; final counts of transformed foci were made at 25 days postinfection.

1981; Carter et al., 1982; Fisher et al., 1982) in demonstrating that certain Ad2 and Ad5 mutants with alterations in early region 2A, the gene encoding the major viral DNA binding protein, show an enhanced ability to transform rat cells. This phenomenon could be most simply explained by presuming that the wild-type DBP normally carries out an activity in infected rat cells which suppresses transformation. Our experiments rule out this hypothesis, however, since deletion mutants which make no detectable DBP or make truncated, defective DBPs transform rat cells at frequencies that are similar to those of wild-type virus. Since primary rat cells are semipermissive for Ad2 and Ad5 growth (Gallimore, 1974) it is possible that enhanced levels of transformation by Ad5ts125 and similar E2A mutants is a result of the decreased level of cytotoxicity caused by these defective viruses. Work by Fisher et a/. suggested that this was an unlikely explanation since Ad5&125 transformed 4- to 1 O-fold more efficiently than wild-type Ad5 in CREF cells, a cloned line of rat embryo fibroblasts which is completely nonpermissive for Ad5 DNA replication and growth (Fisher et al., 1982). Our results also argue against this hypothesis, since the DBP deletion mutants are cer-

tainly as defective as Ad5rsl25 in primary rat embryo cells, yet did not transform at an elevated frequency. Furthermore, we found no differences in plating efficiency of rat embryo cells infected with either high- or low-transforming DBP mutants. Since the high transformation phenotype cannot be explained by the simple inactivation of DBP function, it is likely that mutants such as Ad5rs125 and Ad2+ NDlrs23 encode altered DBPs which carry out an activity in infected rat cells leading to an enhancement of transformation. Consistent with this idea is our observation that the transformation phenotypes of Ad5t.H 25 and Ad2+NDrs23 are dominant over the wildtype phenotype. What DBP function is altered in the high transforming mutants and how does this alteration enhance transformation frequency? One hypothesis consistent with the present data is that these mutants overproduce adenovirus early gene products in infected rat cells, thus leading to higher numbers of transformants. The relevant early products could be the ElA and/or El B proteins, which directly mediate transformation, or the E2B-encoded viral DNA polymerase which is essential for transformation by viral particles (Williams et a/.,

372

RICE. TABLE

KLESSIG,

2

TRANSFORMATION OF RAT EMBRYO CELLS BY Ad2 AND Ad5 TEMPERATURE-SENSITIVE AND HOST-RANGE MUTANTS No. of foci Expt 1.

2.

3.

Virus None Ad5 Ad5tsl25 Ad2+NDl Ad2+NDl Ad2 Ad2hr400 Ad2hr405 Ad2ts400

per dish

32.5”

(10’

37”

cells

seeded) 38.5”

0, 0, 0 8, 5, 5 10, 10, 8 3, 3, 4 19, 18, 21 4, 5, 6 2, 3, 2 5, 3, 4 5, 5, 4

0, 0, 0 4, 5, 4 19,20,21b 2. 3, 2 24,21,21b 3, 2, 4 2, 1, 1 4, 2, 2 5, 5, 6

0, 0, 0 4, 5, 4 15,22, 18’ 4, 3, 3 14, 17, 186 2. 3, 1 2, 3, 1 3, 2. 2 3, 2, 1

None Ad5 Ad5tsl25 Ad2fs400 Ad2ts400’

0, 0, 0 5, 6, 4 12, 13, 11 7, 6, 6 6, 4, 3

0, 0, 0 3, 4, 5 27,25, 28’ 8, 9, 10 7,9,9

0, 0, 0 6, 5, 4 23, 22, 17’ 4, 4, 6 4, 5, 5

Ad5fsl25

12, 9, 10 4, 3, 5 3, 2, 2 3, 4, 5

33,29, 34b 5, 4, 5 3, 2, 2 7, 9, 10

27. 24, 25’ 3, 5, 4 2, 1, 1 3, 4, 4

Ad2 Ad2hr405 Ad2rs400

ts23

a The transformation assays were carried out as described under Materials and Methods. The three numbers at each point in the table correspond to the number of foci in each of three replicate plates. b Foci appeared 3-4 days earlier than foci induced by wild type.

’ A second stock of Ad2ts400 was used.

1974, 1979) as suggested by Logan et al. (198 1). DBP has been implicated in the negative regulation of early viral gene expression in human cells since Ad5tsl25infected cells have been found to accumulate abnormally high levels of early mRNAs at the nonpermissive temperature (Carter and Blanton, 1978a, b; Nevins and Jensen-Winkler, 1980; Babich and Nevins, 1981). An essential role for the DBP in early gene regulation, however, is now in doubt since we have recently shown that in HeLa cells infected with Ad5d/802, which does not make any functional DBP, the expression of early region 1A, 1 B, 3, and 4 mRNAs is apparently normal (Rice and Klessig, 1985). Thus Ad5fsl25 and Ad5d/802 appear to be phenotypically different with respect to both early gene expression in human cells and transformation in rat cells. To determine if the overproduction of early gene products by Ad5rs.125 is related to, or responsible for its high transformation phenotype, it will be necessary to examine wild-type and Ad5tsl25 viral early gene expression in infected rat embryo cells. A second reasonable hypothesis is that the altered DBP of the high transforming mutants affects the intracellular structure of the viral chromosome, thus leading to a higher frequency of integration into the

AND

WILLIAMS

host genome. This model would explain why both the viral DNA polymerase and the DBP can influence the frequency of cellular transformation. It is known that the viral DNA polymerase and the 80-kDa viral preterminal protein form a complex at the ends of the viral DNA molecule which is essential for the initiation reaction of viral DNA replication (Enomoto et a/., 1981; Challberg et al., 1982; Lichy et al., 1982; Stillman et al., 1982; Friefeld et al., 1984). In addition, the DBP can enhance the initiation reaction in vitro when purified viral and cellular components are used (Friefeld et a/., 1984). Perhaps the absence of a functional DNA polymerase or the presence of an altered DBP in vivo can affect the structure or stability of the initiation complex or some other viral DNA structure which recombines with the host genome, leading to integration. In this regard, it is interesting to note that adenovirus DNA is believed to integrate via a process which involves the ends of the viral DNA molecule (Graham, 1984). Furthermore, rat cells transformed by Ad5ts125 and Ad5tslO7 have been found to contain a larger proportion of the viral genome integrated than do cells transformed by wild-type Ad5 (Mayer and Ginsberg, 1977; Doersch-Hasler et al., 1980). This latter observation may reflect an actual difference in the mechanism of integration. Alternatively, it may simply reflect the more nonpermissive nature of the Ad5tsl25/Ad5ts107 interaction with rat cells since other nonpermissive interactions, such as Ad12 infection of hamster cells, also give rise to transformants containing most or all of the viral genome (Graham, 1984). The DBP is thought to contain two physical and functional domains which are separated by a hinge region that is particularly susceptible to proteolytic cleavage (Klein et al., 1979; Schechter et a/., 1980; Klessig and Quinlan, 1982). These domains correspond roughly to the amino-terminal one-third and carboxyl-terminal two-thirds of the protein. The amino-terminal domain has been implicated in the regulation of late viral functions (Klessig and Quinlan, 1982; Rice and Klessig, 1984) while the carboxyl-terminal domain has been implicated in viral DNA replication, early viral gene regulation, and the modulation of cellular transformation (Ariga et a/., 1980; Klessig and Quinlan, 1982; Kruijer et a/., 1983). The results obtained in this study are broadly consistent with this model. The high-transforming mutants, Ad5tsl25 and Ad2+NDl ts23, contain alterations in the carboxyl-terminal segment of the E2A gene. In contrast, Ad2hr400, Ad2hr401, Ad2hr405, and Ad5hr404 all contain amino-terminal alterations (Brough et a/., 1985; Kruijer et al., 1981) and transform at essentially wild-type levels. The notion of two completely separate functional domains of DBP is challenged, however, by the finding

ADENOVIRUS-MEDIATED

373

CELLULAR TRANSFORMATION

0

l 0 0 0

ts400

0

Ad2 /

hr 405

/

0.3

1.0

3.0

10

30

INPUT MULTIPLICITY IREU./CELLI FIG. 3. Relationship between input multiplicity and transformation frequency in primary rat embryo cells infected by Ad2 (A), ts 125 (O), Ad2fs400 (O), and Ad2hr405 (A). Procedures as described under Materials and Methods and in legend to Fig. 2.

that Ad2rs400 transforms rat cells at a frequency similar to wild-type virus. This result was unexpected since Ad2Ls400 contains the same nucleotide alteration as Ad5rsl25 and Ad5tslO7 @rough et a/., 1985) and displays a similar LS,DNA replication negative phenotype (Rice and Klessig, 1984). Since the DBPs encoded by Ad2 and Ad5 are homologous in their carboxyl-terminal domains (no amino acid differences between amino acid 181 and the carboxyl terminus at amino acid 529; Kruijer era/., 1981; Kruijer et al., 1982) it was expected that Ad2ts400 would behave like Ad5ts125 with respect to transformation. Differences in transformation behavior cannot simply be explained by Ad2/Ad5 serotypic differences, since Ad2’NDl ts23 demonstrates that the high transformation phenotype can be expressed in an Ad2 serotypic background. The unexpectedly low transformation efficiency of Adrs400 thus raised the possibility that the DBP amino-terminal hr mutations counteract the effect of the ts mutation in the carboxy terminus. To test this hypothesis, two re-

combinant viruses were constructed in which aminoterminal DBP hr mutations were introduced into the Ad5tsl25 genetic background. One of these viruses, Adtjtsl25X405, encodes a DBP very similar to that of Ad2rs400, sharing the same hr alterations at amino acids 130 and 148 and the same ts mutation at amino acid 413. Both recombinant% however, behaved like Ad5tsl25 and transformed at higher frequency, demonstrating that the hr mutations located in cis to the ts alteration at amino acid 413 do not necessarily revert the high transformation phenotype. With this in mind, we can propose two possible explanations for the low transformation behavior of Ad2fs400. First, perhaps the effect of DBP on transformation is dependent either on the small number of differences between Adz and Ads which exist in the amino terminal region of the protein, and/or to the genetic background (i.e., Ad2 vs Ad5) in which a particular DBP allele is found. In theory, only these two factors distinguish Ad2fs400 and Ad5ts125X405. A second explanation is that, since

374

RICE, TABLE

3

None Ad5 Ad5rsl25 Ad2+NDl ts23 Ad2ts400 Ad2hr405 Ad2hr401 Ad5hr404 Ad5tsl25X405 Ad5fsl25X405-2’ Ad5tsl25X404 Ad5ts125X404-2c Ad5&125X404-3’ Ad5 + AdSfsl25 Ad5 + Ad2+NDl fs23 Ad5tsl25 + Ad2hr405 Ad5t.125 + Ad2hr401 Ad5tsl25 + Ad5hr404 Ad2+NDl fs23 + Ad2hr405 Ad2+NDlfs23 + Ad2hr401 Ad2+NDl fs23 + Ad5hr404

Number of foci per dish” 0, 0, 0 3, 6, 4 26, 28. 27, 25, 6, 4. 4 1, 1, 1 1, 3. 2 3,4, 1 28, 24, 27, 25, 12, 13, 15, 16, 16, 21, i9,20, 20, 21, 6, 5, 6 7, 7, 9 4, 5, 3 4, 4, 2 5, 7,4 4, 5, 4

AND

WILLIAMS

in a multimeric

EFFECT OF DBP HOST-RANGE MUTATIONS TRANSFORMATION EFFICIENCY

Virus

KLESSIG,

ON

Relative transformation efficiencyb -

24 24

23 24 9 17 18

ia 19

complex, and multimers consisting

of

ts and hr proteins were inactive. If this model is correct,

1.0 6.0 5.9 1.1 0.2 0.5 0.6 5.8 5.9 2.6 3.7 4.3 4.4 4.7 1.3

1.8 0.9

0.8 1.2 1 .o

a Transformation assays were performed at 37” as described under Materials and Methods. The three numbers at each point in the table correspond to the number of foci in each of three replicate plates in which 1 O6 cells were seeded. b Compared to transformation efficiency of Ad5 wild type, defined as 1 .O. c A crude stock was used, derived from a different plaque isolate.

Ad2ts400 was derived after chemical mutagenesis (Rice and Klessig, 1984) its decreased transformation frequency is a result of another mutation in its genome, unrelated to the E2A alterations. Our data do indicate, however, that the DBP hr mutations do decrease the efficiency of viral transformation when present in tram to a DBP fs mutation. Primary rat cells coinfected with both an hr mutant and either AdSrsl25 or Ad2+NDl ts23 show the low-transformation frequency typical of the hr mutants alone. Thus, although the ts mutants are dominant over wild-type virus in their transformation phenotype, the hr mutants are dominant over the fs mutants. This dominance relationship might be explained in one of several different ways. Perhaps the hr DBP carries out an activity which negatively affects transformation and thus counteracts the ts DBP. Alternatively, the hr DBP may itself have no direct role in affecting transformation but may simply interfere with the activity of the ts DBP which enhances transformation. This might occur if the ts DBP functions

however, it is not clear why multimers consisting of wild-type and fs proteins would remain active. Another possibility is that the hr mutants are dominant to the ts mutants because they grow more lytically in those cells which otherwise would be transformed. Since we have shown that the hr mutants are not measurably cytotoxic to primary rat embryo cultures, this model must include the assumptions that only a very small percentage of cells in the culture are competent to be transformed, and that the hr mutants grow preferentially in these few cells. Transformation by adenovirus is a complex phenomenon, and since it occurs at very low frequency it is difficult, if not impossible, to study the early events of the process biochemically. Thus, viral genetics has been a powerful tool in our understanding of adenovirus-mediated transformation. Although the viral E 1A and El B genes are sufficient for transformation, our studies with several classes of E2A mutants indicate that the DBP can significantly influence the efficiency of transformation, both positively and negatively. Further understanding of the role of DBP in viral transformation may depend on the development of a system which will allow a study of the transformation process at the biochemical level.

TABLE

4

PLATING EFFICIENCY OF INFECTED AND UNIFECTED

Virus None Ad5 Ad2hr40 1 Ad2hr405 Ad5hr404 Ad5tsl25 Ad2+NDl fs23 Ad2fs400 Ad5d1801 Ad5dl802 AdStsl25x405 Ad5tsl25X404 Ad5tsl25 +Ad2hr405 Ad5rsl25 + Ad5hr404 Ad2+NDl fs23 + Ad2hr405

No. of coloniesa per dish

88,95, 100 103.110,93 98. 104, 94 98, 99, 90 80, 83, 87 90, 95, 101 102, 107.96 aa,97, 103 103, 106, 113 105, 98, 108 96, 92, 106 86,102, 112 94, 102. 98 91. 95, a9 103, 105, 110

RAT EMBRYO CELLS % Plating efficiency6 9.4 10.2 9.9 9.6 a.3 9.5 10.2 9.6 10.7 10.4 9.8 10.0 9.8 9.2 10.6

a 1 O3 infected or uninfected rat embryo cells were seeded on feeder dishes (prepared as described under Materials and Methods), and grown at 37”. Cultures were fixed and stained 2 weeks later. The three numbers at each point in the table correspond to the number of colonies in each of three replicate dishes. * Number of colonies/number of cells seeded.

ADENOVIRUS-MEDIATED

CELLULAR TRANSFORMATION

ACKNOWLEDGMENTS We thank Merlyn Williams for excellent technical assistance. This study was supported by Public Health Service Grant Al 17315 from the National Institutes of Health to D.F.K., and Grant CA-21 375 from the National Cancer Institute to J.W. S.A.R. was supported by predoctoral Training Grant T32GM07464-08 from NIH.

REFERENCES ARIGA,H., KLEIN, H., LEVINE,A., and HORWITZ,M. (1980). A cleavage product of the adenoviral DNA-binding protein is active in DNA replication in vitro. Virology 101, 307-3 10. BABICH,A., and NEVINS,J. R. (1981). The stability of early adenovirus mRNA is controlled by the viral 72 kd DNA-binding protein. Cell 26,37 l-379. BABISS,L. E., FISHER,P. B., and GINSBERG,H. S. (1984). Deletion and insertion mutations in early region 1A of type 5 adenovirus producing cold-sensitive or defective phenotypes for transformation. J. Viral. 49, 731-740. BROUGH,D. E., RICE, S. A., SELL, S., and KLESSIG,D. F. (1985). Restricted changes in the adenovirus DNA binding protein that lead to extended host range or temperature-sensitive phenotypes. I. Viral. 55, 206-212. CARTER,T. H., and BLANTON,R. A. (1978a). Possible role of the 72,000 dalton DNA binding protein in regulation of adenovirus type 5 early gene expression. I. Vkol. 25, 664-674. CARTER,T. H., and BLANTON,R. A. (1978b). Autoregulation of adenovirus type 5 early gene expression II. Effect of temperature sensitive early mutations on virus RNA accumulation. J. !/ho/. 28,450456. CARTER,T. H., NICOLAS,J. C., YOUNG, C. S. H., and FISHER,P. B. (1982). Multiple transformation phenotypes among revertants of temperature-sensitive mutants in the type 5 adenovirus DNA-binding protein. CHALLEERG,M. D., OSTROVE,J. M., and KELLY,T. J., JR.(1982). Initiation of adenovirus DNA replication: Detection of covalent complexes between nucleotide and the 80-kilodalton terminal protein. J. Viral. 41, 265-270. DORSCH-HASLER,K., FISHER,P. B., WEINSTEIN,I. B., and GINSBERG, H. S. (1980). Patterns of viral DNA integration in cells transformed by wild type or DNA-binding protein mutants of adenovirus type 5 and effect of chemical carcinogens on integration. I. viral. 34, 305314. ENOMOTO,T., LICHY,J. H., IKEDA,1. E., and HURWIT~,J. (1981). Adenovirus DNA replication in vitro: Purification of the terminal protein in a functional form. Proc. Nat/. Acad. Sci. USA 78, 6779-6783. ENSINGER,M. J., and GINSBERG,H. S. (1972). Selection and preliminary characterization of temperature-sensitive mutants of type 5 adenovirus. 1. Viral. 10, 328-339. FISHER,P. B., BABISS,L. E., WEINSTEIN,B., and GINSBERG,H. S. (1982). Analysis of type 5 adenovirus transformation with a cloned rat embryo cell line (CREF). Proc. Nat/. Acad. SC;. USA 79, 35273531. FLINT, S. J., SAMBROOK,J., WILLIAMS,J., and SHARP,P. A. (1976). Viral nucleic acid sequences in transformed cells. IV. A study of the sequences of adenovirus 5 DNA and RNA in four lines of adenovirus 5-transformed rodent cells using specific fragments of the viral genome. Virology 72, 456-470. FREEMAN,A. E., BLACK,P. H., VANDERPOOL,E. A., HENRY,P. H., AUSTIN, 1. B., and HUESNER,R. J. (1967). Transformation of primaty rat embryo cells by adenovirus type 2. Proc. Nat/. Acad. Sci. USA 58, 1205-1212. FRIEFELD,B. R., LICHY,J. H., FIELD,J., GRONOSTAJSKI, R. M., GUGGENHEIMER,R. A., KREVOLIN,M. D., NAGATA,K., HURWIT~,J., and HORWIT~,

375

M. S. (1984). The in vitro replication of adenovirus DNA. Cuff. Top. Microbial. Immunol. 110, 22 l-225. GALLIMORE,P. H. (1974). Interactions of adenovirus type 2 with rat embryo cells. Permissiveness, transformation and in vitro characteristics of adenovirus transformed rat embryo cells. J. Gen. Virol. 25,263-273. GALLIMORE.P. H., SHARP, P. A., and SAMBROOK.J. (1974). Viral DNA in transformed cells. Il. A study of the sequences of adenovirus 2 DNA in nine lines of transformed rat cells using specific fragments of the viral genome. I. Mol. Biol. 89, 49-72. GALOS, R. S., WILLIAMS,J., BINGER,M. H., and FLINT, S. J. (1979). Location of additional early gene sequences in the adenoviral chromosome. Cell 17, 945-956. GINSBERG,H. S., ENSINGER,M. J., KAUFFMAN,R. S., MAYER,A. J., and LUNDHOLM,U. (1974). Cell transformation: A study of regulation with types 5 and 12 sdenovirus temperature-sensitive mutants. Co/d Spring Harbor Symp. Quanf. Biol. 39, 419-426. GRAHAM,F. L. (1984). Transformation by and oncogenicity of human adenoviruses. In “The Adenoviruses” (H. Ginsberg, Ed.), pp. 339398. Plenum, New York. GRAHAM,F. L., HARRISON,T., and WILLIAMS,J. (1978). Defective transforming capacity of adenovirus type 5 host range mutants. Virology 122.109-124. GRAHAM,F. L., and VAN DEREB, A. J. (1973). A new technique for the assay of infectivity of human adenovirus 5 DNA. Virology 52,456467. GRAHAM,F. L., VANDEREB, A. J., and HEIJNEKER,H. L. (1974). Size and location of the transforming region in human adenovirus type 5 DNA. Nature (London) 251, 687-691, HARRISON, T., GRAHAM,F., and WILLIAMS,J. (1977). Host-range mutants of adenovirus type 5 defective for growth in HeLa cells. Virology 80, 111-126. Ho, Y. S., GALOS, R., and WILLIAMS,1. (1982). Isolation of type 5 adenovirus mutants with a cold-sensitive host range phenotype: Genetic evidence of an adenovirus transformation maintenancefunction. Virology 122, 109-l 24. HUEBNER,R. 1. (1967). Transformation of primary rat embryo cells by adenovirus type 2. Proc. Nat/. Acad. Sci. USA 58, 1205-l 212. JONES,N., and SHENK,T. (1979). Isolation of adenovirus type 5 host range deletion mutants defective for transformation of rat embryo cells. Cell 17, 683-689. KLEIN,H., MALTZMANN,W., and LEVINE,A. J. (1979). Structure-function relationships of the adenovirus DNA-binding protein. J. Biol. Cbem. 254,11051-l 1060. KLESSIG,D. F. (1977). Isolation of a variant of human adenovirus serotype 2 that multiplies efficiently on monkey cells. J. Viral. 21, 1243-l 246. KLESSIG,D. F., and GRODZICKER, T. (1979). Mutations that allow human Ad2 and Ad5 to express late genes in monkey cells map in the viral gene encoding the 72K DNA binding protein. Cell 17, 957966. KLESSIG,D. F., and QUINLAN,M. P. (1982). Genetic evidence for separate functional domains on the human adenovirus specified, 72 kd, DNA binding protein. 1. Mol. Appl. Genet 1, 263-272. KRUIJER,W., NICOLAS,J. C., VAN SCHAIK, F. M. A., and SUSSENBACH, 1. S. (1983). Structure and function of DNA binding proteins from revertants of adenovirus type 5 mutants with a temperature-sensitive DNA replication. virology 124, 425-433. KRUIJER, W., VAN SCHAIK,F. M. A., and SUSSENBACH,J. S. (1981). Structure and organization of the gene coding for the DNA binding protein of adenovirus type 5. Nucleic Acids Res. 9, 4439-4457. KRUIJER,W., VAN SCHAIK,F. M. A., and SUSSENBACH.J. S. (1982). Nucleotide sequence of the gene encoding adenovirus type 2 DNA binding protein. Nucleic Acids Res. 10, 4493-4500.

376

RICE, KLESSIG, AND WILLIAMS

LEWIS,1. B., ATKINS,J. F., BAUM, P. R., SOLEM, R., GESTELAND,R. F., and ANDERSON,C. W. (1976). Location and identification of the genes for adenovirus type 2 early polypeptides. Cell 7, 14 l-l 5 1. LICHY,1. H., FIELD,J., HORWITZ,M. S., and HURWITZ,J. (1982). Separation of the adenovirus terminal protein precursor from its associated DNA polymerase: Role of both proteins in the initiation of adenovirus DNA replication. froc. Nat/. Acad. Sci. USA 79, 5225-5229. LOGAN, J., NICOLAS,J. D., TOPP, W. C., GIRARD,M., SHENK, T., and LEVINE,A. J. (1981). Transformation by adenovirus early region 2A temperature-sensitive mutants and their revertants. virology 115, 419-422. MAYER,A. J., and GINSBERG,H. S. (1977). Persistence of type 5 adenovirus DNA in cells transformed by a temperature-sensitive mutant, H5ts125. Proc. Natl. Acad. Sci. USA 74, 785-788. MONTELL,C., COURTOIS,G., ENG, C., and BERK,A. J. (1984). Complete transformation by adenovirus 2 requires both El A proteins. Cell 36,951-961. NEVINS,J. R., and JENSEN-WINKLER,1. J. (1980). Regulation of early adenovirus transcription: A protein product of early region 2 specifically represses region 4 transcription. Proc. Nat/. Acad. Sci. USA 77, 1893-1897. NICOLAS,J. C., SUAREZ,F., LEVINE,A. J., and GIRARD,M. (1981). Temperature-independent revertants of adenovirus H5ts125 and H5ts107 mutants in the DNA binding protein: Isolation of a new class of host range temperature conditional revertants. Virology 108,521-524. NICOLAS,J. C.. YOUNG,C. S. H., SUAREZ,F., GIRARD,M., and LEVINE, A. 1. (1983). Detection, rescue, and mapping of mutations in the adenovirus DNA-binding protein. Proc. Nat/. Acad. Sci. USA 80, 1674-1677. PUCK,T. T., MARCUS,P. I., and CIECIURA,S. J. (1956). Clonal growth of mammalian cells in vitro. Growth characteristics of colonies from single HeLa cells with and without a “feeder” layer. f. Exp. Med. 103,273-284. RICE,S. A., and KLESSIG,D. F. (1984). The function(s) provided by the adenovirus-specified DNA-binding protein required for late viral gene expression is independent of the role of the protein in viral DNA replication. J. l&o/. 49, 35-49.

RICE, S. A., and KLESSIG,D. F. (1985). Isolation and analysis of adenovirus type 5 mutant containing deletions in the gene encoding the DNA binding protein. 1. \/ire/. 56, 767-778. RUBEN,J., BACHET~, S., and GRAHAM, F. L. (1982). Integration and expression of viral DNA in cells transformed by host range mutants of adenovirus type 5. J. Viral. 41, 674-685. SCHECHTER,N. M., DAVIES,W., and ANDERSON,C. W. (1980). Adenovirus coded deoxyribonucleic acid binding protein. Isolation, physical properties, and effects of proteolytic digestion. Biochemistry 19, 2802-2810. SHARP,P. A., PESTERSSON, U., and SAMBROOK,J. (1974). Viral DNA in transformed cells. I. A study of the sequences of adenovirus 2 DNA in a line of transformed rat cells using specific fragments of the viral genome. J. Mol. Biol. 86, 709-726. STILLMAN,B. W., TAMANOI,F., and MATHEWS, M. B. (1982). Purification of an adenovirus coded DNA polymerase that is required for initiation of DNA replication. Cell 31, 613-623. VAN DEREB, A. J., MULDER, C., GRAHAM, F. L., and HOUWELING,A. (1977). Transformation with specific fragments of adenovirus DNAs. I. Isolation of a specific fragment with transforming activity of adenovirus 2 and 5 DNA. Gene 2, 115-l 32. WILLIAMS,1. (1986). Adenovirus genetics. ln “Adenovirus DNA” (W. Doerfler, Ed.), pp. 247-309. Nijhoff, Boston. WILLIAMS,J., GHARPURE,M., USTACELEBI,S., and MCDONALD,S. (1971). Isolation of temperature-sensitive mutants of adenovirus type 5.1. Gem Viol. 11, 95-l 0 1. WILLIAMS,J., and USTACELEBI,S. (1971). Temperature-restricted mutants of human adenovirus type 5. ln “Ciba Foundation Symposium on Strategy of the Viral Genome” (G. E. W. Wolstenholme and M. O’Conner, Eds.), pp. 275-289. Churchill Livingston, London. WILLIAMS,J. F., GALOS, R. S., BINGER,M. H., and FLINT, S. J. (1979). Location of additional early regions within the left quarter of the adenovirus genome. Cold Spring Harbor Symp. &ant. Biol. 44, 353-366. WILLIAMS,J. F., YOUNG,C. S. H., and AUSTIN, P. E. (1974). Genetic analysis of human adenovirus type 5 in permissive and nonpermissive cells. Cold Spring Harbor Symp. &ant. Biol. 39,427-437.