Independent Transformation Activity by Adenovirus-5 El A-Conserved Regions 1 or 2 Mutants CATHARINA SVENSSON,*,’ MARIA BONDESSON,* ELISABETH NYBERG,* STIG LINDER,t NICHOLAS JONES,+ AND GdRAN AKUSJARVI* *Department of Microbial Genetics, Karolinska Institute, 104 01 Stockholm, Sweden; tDepartment of Oncology, Division of Experimental Oncology, Radiumhemmet, Karolinska Hospital, 104 01 Stockholm, Sweden; and +ICRF Laboratories, Lincoln’s Inns Fields, London WC2A 3PX, England Received November
27, 199 1; accepted
15, 199 1
Two conserved regions (CR1 and CR2) on the adenovirus ElA proteins have previously been shown to be required for of primary rodent cells. Sequences within these regions are cooperation with the ras oncogene in the transformation essential for the ability of ElA to associate with the 105K product of the retinoblastoma susceptibility gene, plO&RB, as well as with other cellular proteins, including a 107K (~107) and a 300K (~300) species. In this paper, we show that CR1 mutants deficient in ~300 binding and CR2 mutants with lost or reduced binding of plO&RB and/or ~107 have a low, but not abolished focus formation activity. In contrast, CR1 /CR2 double mutants were deficient in focus formation, suggesting that the transformation activities displayed by the single CR1 or CR2 mutants were due to an independent transformation activity by both CR1 and CR2. No strict correlation between plOBRB binding and ElA-mediated transformation was observed. The ElA enhancer repression function was found to correlate with the binding of ~300 but not with El A-mediated transformation. Complex formation between El A and ~107, similar to the pl05-RB binding, required sequences within both CR1 and CR2. The CR2 sequences required for binding of p107K or pl05-RB were overlapping, but not identical. Finally, a larger segment of CR2 was required for stable complex formation between ElA o 1991 Academic and phosphorylated forms of pl05-RB or ~107 compared to corresponding unphosphotylated species. Press.
The ElA gene encodes a minimum of five alternatively spliced mRNAs (9S-13s) during a lytic infection (Chow et a/., 1979; Stephens and Harlow, 1987; Ulfendahl et a/., 1987). In transformed cells, only the 13s and 12s mRNA are expressed (Berk and Sharp, 1978). Highly related polypeptides of 289 and 243 amino acids (289R and 243R) are translated from the 13s and 12s mRNAs, respectively. These have identical aminoand carboxy-terminal ends and differ from each other by the presence of 46 unique amino acids in the 289R protein (Perricaudet et al,, 1979). Comparison of ElA genes from both human and simian adenovirus serotypes have identified three well conserved regions (CRl, CR2, and CR3) at the protein level (Kimelman, 1985; Fig. 1). Mutational analysis of the ElA protein coding region has shown that region CR3 encodes for a transcriptional activator domain which is not essential for transformation or immortalization of primary cells (Lillie eta/., 1986; Moran eta/., 1986; Schneider et al., 1987) although it is needed to obtain the complete morphologically transformed phenotype (Monte1 et a/., 1984; Adami and Babiss, 1990). In contrast, mutants which delete CR1 or CR2 have been reported to be deficient in both the El A-mediated immortalization and transformation functions. Most likely, the transforming domain previously mapped to CR1 is not limited by the
The ElA gene products of human adenoviruses have been shown to be multifunctional, playing central roles in processes like control of viral and cellular gene expression and transformation. The El A gene alone can immortalize and partially transform primary cells in vitro (Houweling et a/., 1980). However, for full transformation a second oncogene such as El B (van den Elsen et al., 1983) polyoma middle T-antigen, or activated ras oncogenes is required (Ruley, 1983). The E 1A gene encodes proteins with both positive and negative transcription regulatory properties. Thus, early viral genes (Berk et al., 1979; Jones and Shenk, 1979) as well as certain cellular genes (Nevins, 1982; Stein and Ziff, 1984) are transcriptionally activated by E 1A. In contrast, ElA inhibits enhancer-dependent gene expression (Borelli et al., 1984; Hen et al., 1985; Velcich and Ziff, 1985; Stein and Ziff 1987; Sogawa et al., 1989) as well as transcription of several cellular genes where the targets for repression are not conventional enhancers (Timmers et a/., 1988; Webster eta/., 1988; Offringa et a/., 1988; Young et a/., 1989).
’ To whom correspondence
should be addressed. 553
Copynght 0 1991 by Academic Press, Inc. All rights of reproduction 10 any form reserved.
289 aa 243 aa
A CR2a A CR2b A CR1.2a -
and ~300, ~107, and pl05-RB are required for full transformation efficiency. In contrast to several previous studies, we report a substantial transformation activity of CR1 or CR2 mutants which have lost, or reduced to below detection level, their ability to associate with two of the three major cellular proteins. No strict correlation between pl05-RB binding and ElAmediated transformation was observed.
FIG. 1. Schematic representation of adenovirus ElA deletion mutants used in this study. Numbers above arrows indicate amino and residues bordering the deletions. For more details see Schneider et al. (1987).
region of amino acid homology, but also includes the N-terminal part of El A (Subramanian et a/., 1988). A close link between enhancer repression and transformation has previously been suggested by the observation that mutations in either CR1 or CR2, which abolish the transformation activity, severely reduce (CR2 mutants) or inhibit (CR1 mutants) the enhancer repression function of El A (Lillie et a/., 1986, 1987; Schneider eT al., 1987). However, transformation negative CR2 mutants which do not significantly reduce the capacity of El A to inhibit SV40 enhancer-dependent transcription have been described (Kuppuswamy and Chinnadurai, 1987). In addition, there are mutants affecting sequences located outside CR1 and CR2 which only affect the transformation capacity (Subramanian et al., 1988) or enhancer repression function (Velcich and Ziff, 1988) of El A. A number of cellular proteins have been found to associate with the El A proteins in viva (Yee and Branton, 1985; Harlow et a/., 1986); the three most prominent being of sizes 300K (p300), 107K (p107), and 105K. The 105K product has been shown to be identical to the product of the retinoblastoma-susceptibility gene, pl05-RB (Whyte et al., 1988). The complex formation between ElA and the pl05K-RB or ~107 has been shown to require sequences within CR1 and/or CR2 (Egan et a/., 1988; Whyte et a/., 1989). Binding of the ~300 has been shown to require sequences close to the N-terminus and within CR1 (Egan ef a/., 1988; Whyte et al., 1989). Based on the close correlation between sequences required for protein binding and biological activities displayed by El A, it has been suggested that physical interactions between ElA and these cellular proteins are crucial for El A-mediated transformation (Egan et a/., 1989; Whyte et a/., 1989). We have investigated the ability of several first exon E 1 A mutants to cooperate with T24 H-ras in a transformation assay on primary rat embryo fibroblast cells. Our results suggest that associations between ElA
Plasmid DNA Plasmid pKGO-007 SVRI (referred to in the text as wild type) contains 1 to 2800 base pairs of genomic Ad2 sequences (Svensson et al., 1983). Deletion mutants G5/3, GCX, and G3/2 (Schneider et al., 1987) (referred to in text as ACRl, ACR2a, and ACR2b) lack amino acids 38-65, 121-l 25, and 125-l 33, respectively. Mutants ACR1,2a and ACR1.2b were constructed by combining the mutation in G5/3 with the mutation in GCX or G3/2, respectively. The constructions were made by a three-fragment ligation using restriction enzyme fragments Sacll(356)-BspMll(825) from plasmid G5/3, BspMll(825)-Xbal(l336) from GCX or G3/2 and Xbal(l336)-Sacll(356) from pKGO-007 SVRI. Plasmid pSV2/NEOT24 was constructed by recloning the 6.6-kb genomic T24-Ha-ras BarnHI fragment into the BamHlsite of pSV2NEO. Cell culture
Primary cells were prepared from Fisher rat embryos of different gestation ages. The 12- to 14-day embryos were minced and then incubated in 0.259/o trypsin, 1 mg/ml collagenase, 0.1 mg/ml DNase 1 at 37°C for 45 min. Cells were plated in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 5% fetal calf serum and allowed to attach overnight. The following day, cells were detached by trypsin treatment and frozen in aliquots in liquid nitrogen. Cells were thawed and transfected with 5 pg of plasmid DNA per 60-mm dish essentially as described by Wigler et al. (1978). For immunoprecipitation of ElA protein complexes 10 pg of El A plasmid was transfected into HeLa cells. To measure enhancer repression 5 pg of ElA plasmid and 3 pg of pSXp + DNA (Banerji et al., 198 1) encoding the rabbit P-globin gene under the transcriptional control of the SV40 enhancer were used. Radioactive
labeling and immunoprecipitation
Approximately 44 hr post-transfection cells were starved in methionine-free DMEM for 45 min followed by a 3.5 hr incubation in the presence of 200 &i [35S]methionine/ml. After the labeling period cells were
I A CRlJa+T24ms
BY Ad5 ElA
FIG. 2. Transformation assay on primary REF cells with T24-H-ras alone or in combination with the different ElA deletion mutants. Transformation was carried out as described in Table 1.
washed in PBS and fractioned into nuclei and cytoplasm by isoBNP40 (IO mM Tris, pH 7.9, 150 ml\/l NaCI, 1.5 rnn/l MgCI, 0.5% Nonidet-P40) treatment. The cytoplasmic extract was precleared twice using an inactivated suspension of Staphylococcus aureus (Sigma Chemicals). Extracts were then incubated with the mouse monoclonal ElA antibody M73 (Harlow et a/., 1985) or the mouse monoclonal pl05-RB antibody C36 (Whyte et al., 1988). Bound complexes were isolated by binding to protein A-Sepharose (Pharmacia Fine Chemicals), washed extensively in IsoB-NP40, and analyzed on an 8% SDS-polyacrylamide gel.
CR1 OR CR2 MUTANTS
the sensitivity of the transformation assay to allow for detection of weakly transforming mutants, we compared different preparations of REF cells with respect to size and number of transformed foci induced by cotransfection with wtE1A and T24 H-ras. We were able to select a batch of REF cells which grew to low saturation densities in 5% serum and were particularily susceptible to E 1A-mediated transformation. A typical result from one ElA + T24 H-ras cotransformation experiment is shown in Fig. 2 and the data from several experiments are summarized in Table 1. As shown, mutants ACRl, ACR2a, and ACR2b were all able to produce foci on REF cells (Fig. 2). However, the number of foci was significantly reduced compared to wild type. The most severe effect was seen with mutant ACRl which produced only around 10% of the number of foci observed with wild type. The two mutants affecting different parts of CR2 (ACR2a and ACR2b; Fig. 1) transformed with somewhat different frequencies; ACR2b being reproducibly better than ACR2a (Table 1). Mutants ACRl, ACR2a, and ACR2b transformed cells were also atypical in morphology compared to wild type; they were less condensed and less prone to grow on top of each other (Fig. 2 and Table 2). Collectively, these data suggest that CR1 and CR2, independently of each other, are able to cooperate with ras in a focus formation assay. To investigate this hypothesis further we constructed two double mutants (ACRl,2a and ACR1,2b, Fig. 1) by combining the CR1 mutation with either of the two CR2 mutations. As shown in Fig. 2 and summarized in Table 1, neither of these two mutants retained any transformation capacity, suggesting that the observed transformation activity with the single CR mutants was due to an independent transformation activity of CR1 or CR2.
TABLE 1 TRANSFORMINGACTIVITYBY AMINO-TERMINAL E 1A MUTANTS
Number of foci per plate
Sl endonuclease analysis of rabbit P-globin mRNA was as described previously (Svensson and AkusjaG, 1984). RESULTS Independent T24-ras
ability of CR1 or CR2 to cooperate
A series of ElA mutants lacking parts of the conserved region 1 or 2, or both (Fig. l), were analyzed with respect to their ability to cooperate with T24 H-ras in the transformation of REF cells. In order to improve
WT ACRl ACR2a ACR2b ACRl,2a ACR1,2b
50 5 20 N.D.a 0 0
60 4 8 5 N.D. N.D.
34 1 4 13 0 0
81 10 18 25 0 0
100 8 21 26 0 0
Note. Primary rat embryo fibroblasts were cotransfected with plasmids encoding deletion mutants of the adenovirus ElA gene and T24-H-ras. At 2 weeks post-transfection cells were fixed and stained with Giemsa and foci were counted. a Not done.
TABLE 2 SUMMARY OF AMINO-TERMINAL ElA MUTANT PHENOTYPES Cooperation
with ras Association
WT ACRl ACR2a ACR2b ACRl,2a ACRl,2b
Number of foci
Cell density in foci
+++ + + + -
High Low Low Low -
105 kDa +++ + ++ +
105 kDa +++ (+) -
with cellular proteins 107 kDa +++ ++ (+I ++ -
Repression of SV40 enhancer
+++ +++ -
++ +++ -
form of pl05K-RB.
The ElA proteins have been shown to form stable complexes with several cellular proteins in extracts from both adenovirus-infected and transformed cells (Fig. 3, lane 3) (Yee and Branton, 1985; Harlow et a/., 1986; Egan et al., 1987, 1988). The binding sites on ElA for the three most prominent cellular proteins (pl05-RB, ~107, and ~300) have been extensively mapped (Egan et a/., 1988; Whyte et al., 1989). Some minor differences in defining the borders of sequences necessary for protein binding exist. Basically, pl05-RB has been shown to bind to sequences within CR1 and CR2, ~107 to CR2, and ~300 to the N-terminus and CR1 The ability of the El A mutants used in this study to bind the cellular proteins was investigated in order to correlate protein complex formation to transformation capacity. For this experiment HeLa cells were transfected with ElA mutants and the [35S]methioninelabeled cell extract immunoprecipitated with the El Aspecific monoclonal antibody M73 (Harlow et al., 1985). The three single mutants affecting CR1 or CR2, which principally showed similar properties in their abilities to cooperate with ras in a focus formation assay (Table l), varied significantly with respect to their ability to associate with the three cellular proteins (Fig. 3, lanes 6-9). As expected from previous reports (Egan et a/., 1988; Harlow et a/., 1988), deletion of amino acids 41-65 within CR1 (ACRl; lane 7) reduced ~300 and pl05-RB protein binding without grossly affecting ~107 binding. Mutant ACR2a- and ACR2b-encoded proteins, with deleted amino- and carboxy-terminal parts of CR2, respectively, both bound ~300 as efficiently as wild type (lanes 8 and 9). They differed from each other with respect to binding of pl05-RB and ~107. ACR2a showed a drastically reduced association to both pl05-RB and ~107. In fact, with this mu-
tant, pl05-RB binding was reduced to below detection levels (note that the band at about 105K migrated somewhat faster than authentic pl05-RB). In contrast, ACR2b showed only a slightly reduced ~107 and pl05-RB-binding capacity (lane 9). Consistent with the hypothesis that binding of ~300, ~107, and pl05-RB is required for efficient transformation the double mutant, ACR1,2a, was unable to form a complex with any of these proteins (lane 10) (as with ACR2a the band of about 105K migrated somewhat faster than correct pl05-RB). Surprisingly, ACRl,2b retained a low, residual ability to associate with pl05-RB (lane 11). Thus, binding of small amounts of pl05-RB is not sufficient to cause focus formation. Taken together, these results indicate that full transformation requires binding of all three host proteins. A limited transformation activity was obtained when efficient binding of only one or two of the cellular proteins was detected. Binding of ~107 to ElA is mediated sequences in both CR1 and CR2
The association of the ~107 to ElA has been mapped to sequences within CR2 which overlap those required for binding of the pl05-RB protein. Binding of both pl05-RB and ~107 requires sequences between 121 and 127 (Whyte et a/., 1989). plO5-RB binding shows stronger dependence on sequences immediately upstream of CR2 (11 1 to 121) compared to ~107, which instead also requires the carboxy-terminal part of CR2 (128 to 138) for maximal binding (Egan et a/., 1988). However, whereas pl05-RB has been shown to bind to both CR1 and CR2, binding of ~107 has been suggested to directly require only CR2 (Whyte et a/., 1989) and to be only marginally influenced by sequences within CR1 and the N-terminal region (Egan et a/., 1989). Our experiments demonstrate that the interaction between ElA and ~107 indeed requires se-
BY Ad5 ElA
FIG. 3. The ability of ElA deletion mutants to associate with pl05-RB, ~107. and ~300. Extracts prepared from HeLa and 293 cells were rmmunoprecipitated with mouse monoclonal antibodies directed against ElA (M73; lanes 3 and 5 to 11) or ~105.RB (C36; lane 2). Lanes 1 to 3; immunoprecipitation of extracts prepared from 293 cells; -ab, unspecific antibody; M, marker; lanes 5 to 11; immunoprecipitation of extracts prepared from transfected HeLa cells. Mock, mocktransfection; WT, transfectron with pKG0007 SVRI encoding the wild-type ElA gene (Svensson er a/., 1983). 105* indicates phosphorylated forms of ~105.RB.
quences within both CR1 and CR2. ACR2a, lacking the amino-terminal part of CR2 (residues 121-l 25) demonstrated a very poor ~107 binding compared to wild type. In contrast, ACR2b, which lacks the second half of CR2 (residues 125-l 33), was not severely affected in ~107 binding. In striking contrast, the double mutant ACR1,2b was completely deficient in its ability to associate with ~107 (Fig. 3, lane 11). We conclude that, similarly to pl OSRB, efficient binding of ~107 depends on sequences located in both CR1 and CR2. However, the exact amino acid contact points in CR2 appear to differ. Whereas residues 121 to 125 will suffice for efficient pl05-RB binding, ~107 binding requires a larger region; residues 121 to 133. Phosphorylated forms of pl05-RB and ~107 have a lower affinity to mutants ElA proteins The ~300, ~107, and 105K-RB proteins have previously been shown to be phosphorylated in viva.
CR1 OR CR2 MUTANTS
As shown in Fig. 3 the pl05-RB band immunoprecipitated from 293 cell extracts with either C36 (lane 2) or M73 (lane 3) monoclonal antibodies appeared as a sharp band with a smear of slower migrating bands. Phosphatase treatment of the immunoprecipitated material resulted in a disappearance of the slower migrating bands (data not shown). The multiple bands therefore represent different phosphorylated forms of pl05-RB (Ludlow eta/., 1989). Densitometric scanning of autoradiograms (data not shown) revealed that the phosphorylated pl05-RB species were efficiently precipitated only from wtElA-transfected cells (Fig. 3, lanes 6 to 1 1). Most strikingly, ACR2b, which was only slightly reduced in its ability to associate with faster migrating forms of pl05-RB, did not coprecipitate any detectable amounts of phosphorylated pl05-RB (Fig. 3, lane 9). To confirm this result, transfected HeLa cells were labeled with [32P]orthophosphate prior to immunoprecipitation with the C36 or the M73 monoclonal antibodies (Fig. 4). As shown in Fig. 4, deletion of resi-
FIG. 4. Requirement of CR2 for stable associatron between ElA proteins and overphosphorylated forms of pl05-RB and ~107. The experiment was as described in Fig. 3 with the exception that either [35S]methionine or [32P]orthophosphate was used for labeling of the transfected cells.
dues 125 to 133 (ACR2b) resulted in an almost completely abolished binding of phosphorylated forms of pl05-RB and pl07K (lane 7). The contribution of CR1 in binding the phosphorylated forms of both proteins was less obvious, but could not be excluded. Densitometric scanning of [35S]methionine-labelled pl05-RB coprecipitated by the M73 antibody failed to detect any slower migrating forms of pl05-RB (Fig. 3, lane 7 and data not shown). Since the level of bound pl05-RB in these cells was much reduced, the presence of low levels of phosphorylated pl05-RB was difficult to determine. In fact, small amounts of phosphorylated forms of pl05-RB and pl07K were detected in 32P-labeled cells (Fig. 4, lane 5). Collectively, our results suggest that residues 125 to 133, which play a minor role in the binding of underphosphorylated forms of pl05-RB, are vital for stable binding of the phosphorylated species. Residues 125 to 133 are important for binding of both phosphorylated and unphosphorylated forms of ~107, although probably more essential in the formation of stable complexes between ElA and the phosphorylated forms. The SV40 enhancer repression with binding of ~300
The ability of the ElA proteins to repress enhancerdriven transcription has been investigated in great detail in several laboratories (Borelli et a/., 1984; Hen et a/., 1985; Velcich and Ziff, 1985; Lillie et al., 1986, 1987; Schneider et al., 1987; Rochette-Egly et al., 1990). Since both CR1 and CR2 appear to be required for repression and oncogenic transformation, it has been suggested that transcription repression is an important parameter in El A-mediated transformation. However, yet other reports have presented evidence that El A-mediated repression requires sequences outside CR1 and CR2 which are not essential for the transformation process (Velcich and Ziff, 1988; Jelsma et a/., 1989). We have used a HeLa cell cotransfection assay to analyze the capacity of CR1 and CR2 mutants to repress transcription from a rabbit p-globin reporter gene linked to the SV40 enhancer ([email protected]
+, Banerji et al., 1981). As shown in Fig. 5, ACRl completely failed to repress @-globin gene expression. In contrast, ACR2b showed full repression of ,&globin expression whereas ACR2a was slightly reduced. As could be predicted, both double mutants failed to repress transcription. Thus, the E 1A enhancer repression function showed a strong correlation with the presence of CR1 and the ability of El A to associate with ~300. DISCUSSION A number of studies have identified conserved regions 1 and 2 as essential elements for El A-mediated
FIG. 5. Repression of SV40 enhancer-driven transcription requires CR1 HeLa cells were cotransfected with the ElA expressing plasmids and plasmid pSXfl+ containing the rabbit o-globin gene linked to the SV40 enhancer (Banerji et al., 1981). Cytoplasmic RNA was prepared at 48 hr post-transfection and subjected to Sl nuclease analysis. End-labeled DNA fragments specific for p-globin mRNA and ElA mRNAs were used as probes (Svensson and Akusjarvi, 1984). Protected fragments were separated on a 8% sequencing gel.
transformation. The aim of the present investigation was to assess the relative importance of the two regions in the transformation process and to correlate El A protein complex formation to cell transformation. For these experiments we selected a batch of REF cells highly susceptible to El A-mediated transformation. By cotransfecting these cells with El A + T24-Hras we were able to demonstrate that CR1 or CR2 mutants are independently capable of inducing focus formation at a frequency of approximately 10 to 25% of wild type (Table 1). To this end it is interesting to note that CR1 and CR2, when presented on separate protein molecules, have been shown to cooperate with ras in transformation (Moran and Zerler, 1988). The foci induced by the El A CR1 and CR2 mutants were morphologically altered compared to wild-type transformed cells. In general, foci were smaller and possibly also unable to grow to the same cell saturation density (Table 2). The foci appearing after transformation with mutant ACRl were most atypical and differed from wild-type ACR2a- and ACR2b-transformed foci by being very easily detached from the plastic support. This
BY Ad5 ElA
phenotype may possibly be explained by the inability of CR1 mutants to repress expression of extracellular proteases, such as stromelysin and collagenase (Offringa et al., 1988, 1990; van Dam et al., 1989) thus resulting in a degradation of the extracellular matrix of mutant-transformed cells. The initial characterization of the CR1 and CR2 mutants used in this investigation failed to demonstrate any transforming activities on BRK cells (Schneider et al., 1987). The differences in results using primary REF cells could be due to the longer replicative life span of REF compared to BRK cells, making them not as strictly senescent. In this paper we demonstrate a correlation between ElA-mediated enhancer repression and the ability of E 1A proteins to bind ~300. This conclusion is based on the observations that mutant ACRl, which failed to repress the SV40 enhancer (Fig. 5), associated with pl05-RB and ~107, but not with ~300 (Fig. 3, lane 7) and that ACR2a, which was only slightly reduced in SV40 enhancer repression, associated efficiently only with ~300 (Fig. 3, lane 8). ElA transrepression therefore appears to correlate with the presence of CR1 and the association between El A and ~300. We therefore find it unlikely that the ability to repress enhancerdriven gene expression is obligatory for the focus formation process. This is based on the results that the three single mutants with deletions in CR1 or CR2, showing very different abilities in transrepression function (Fig. 5) all retained a significant transformation capacity. The threefold lower transformation frequency by mutant ACRl compared to the CR2 mutants suggests a modulating role of El A-mediated repression in transformation. The contribution made by CR2 may, however, vary between different experimental systems. For example, Jelsma et al. (1989) have shown that deletion of sequences in the extreme C-terminus of CR2 (128-l 38) reduced transrepression of the SV40 enhancer four- to fivefold whereas deletions of the remainder of CR2 had a minor effect on transrepression. In the original characterization of the ElA mutants used in this study, CR2 were shown to have a major impact on repression of the polyoma virus enhancer (Schneider et al., 1987). It is therefore possible that the relative contribution of CR2 in transrepression varies between different enhancers. No simple relationship was found between E 1A-mediated transformation and the ability of ElA to associate with specific cellular proteins. A mutant which efficiently associated only with ~300 (ACR2a; Fig. 3, lane 8) showed similar focus formation activity as a mutant which associated efficiently with all three cellular proteins (ACR2b Fig. 3, lane 9, Table 1). Therefore, in a focus formation assay, El A binding to either ~300 or p105-RB/p107 may be sufficient. Since ElA asso-
CR1 OR CR2 MUTANTS
ciates with a number of cellular proteins in addition to pl05-RB, ~107, and ~300 (Fig. 3, lane 3; Yee and Branton, 1985; Harlow et al., 1986) it is possible that the transformation potential correlates better with other parameters as, for example, binding of some minor cellular proteins found complexed to E 1A or binding to the phosphorylated forms of pl05-RB and ~107. Of specific interest is that we do not observe a strict correlation between ElA association to pl05-RB and El A-mediated transformation; the amount of pl05-RB found in association with mutant ACR1,2b was similar to that found complexed to ACRl but still only the latter mutant was transformation competent. A parallel situation was recently described for SV40 large T transformation (Chen and Paucha, 1990). Mutants defective in pl05-RB binding were still able to immortalize primary rodent fibroblasts in vitro. The ElA region required for efficient binding of ~107 to ElA has previously been mapped to the first half of CR2, with only minor contributions to binding made by CR1 (Egan et al., 1988; Whyte et a/., 1988). Our experiments show that ~107, similar to pl05-RB, makes contact with sequences located within both CR1 and CR2, and furthermore, that stable complex formation between ElA and pl05-RB or ~107 requires different segments of CR2. Thus, underphosphorylated plO5RB requires primarily the residues located between 121 and 125 for stable binding whereas underphosphorylated and phosphorylated forms of ~107 interact with sequences located overa much larger region; residues 121-l 33. ~107 makes essential contacts with the amino-terminal part of CR2 (residues 12 l-l 25) but also interacts with residues 125 to 133, although this region seems less crucial for binding of the underphosphor-ylated species. We could only manifest the contribution of the carboxy-terminal end of CR2 for binding of underphosphorylated forms of ~107 in combination with a deletion in CR1 (Fig. 3, lane 1 1). The contribution of the carboxy-terminal of CR2 (residues 121-l 33) in binding of underphosphorylated forms of pl05-RB is relatively small since ACRl (Fig. 3, lane 7) and ACRl,2b (Fig. 3, lane 11) retained approximately the same binding affinity for pl05-RB. In contrast, these residues were essential for binding of the phosphorylated forms of pl05-RB (Fig. 4, lanes 7 and 9). Thus, underphosphorylated pl05-RB interacts with the same CR2 sequences as ~107, i.e., residues 121 to 133. The contribution of CR1 in binding of the phosphorylated species could not be unequivocally determined because of the poor binding of pl05-RB to ACRl It seems likely from our results that the unphosphorylated forms of pl05-RB also interact with residues 125 to 133. However, since the main interaction point between ElA and pl05-RB most likely is located be-
tween residues 121 and 125, the contribution of the carboxy-terminus of CR2 is only minor for complex formation. This is particularly interesting considering the fact that SV40 T binds only underphosphorylated forms of pl05-RB (Ludlow eta/., 1989). In normal cells, pl05-RB becomes phosphotylated in a cell cycle regulated fashion (Buchkovich and Harlow, 1989; DeCaprio et al., 1989). SV40 T associates with pl05-RB in G, cells and then dissociates in S-phase when pl05-RB becomes phosphorylated (Ludlow et al., 1990). ElA wild type, on the other hand, binds both unphosphorylated and phosphorylated forms of pl05-RB (Dyson et a/., 1989). It is therefore possible that ElA polypeptides are capable of associating with pl05-RB throughout or at least further into the cell cycle compared to SV40 T. Such an association past the GJG, boundary may infer additional alterations on the plO5RB function. Since the pl05-RB binding domain of SV40 T can substitute for CR2 in E 1A-mediated transformation (Moran, 1988), it would be of interest to investigate whether such a chimeric gene would be able to bind phosphorylated forms of pl05-RB. There is a significant sequence homology between the pl05-RB binding domain of the SV40 T antigen and the second half of CR2 (Figge et a/., 1988). Still, the difference between ElA and SV40 T complex formation with pl05-RB is possibly due to differences in the extent of the protein-protein interaction. El A probably has a larger region of interaction; a primary interaction point shared with SV40 T consisting of residues 121125 and a unique secondary interaction point, which is essential for binding of phosphorylated forms of plO5RB, making contact with residues 125 to 133. ACKNOWLEDGMENTS We are grateful to MS Christina Stromberg for competent technical assistance and to MS Elvira Meterus for excellent secretarial help. We thank Dr. Ed Harlow for the generous gift of antisera. This work was supported by grants from the Swedish Cancer Society, the Swedish Medical Research Council, and King Gustav V:s Jubilee foundation. Note added in proof. After completion of this work Howe et al. (Proc. Nat/. Acad. Sci. USA 87,5883-5887,199O) showed that ~107 interacts with sequences within both CR1 and CR2. Furthermore, Stein et a/. u. viral. 64, 442 l-4427, 1990) demonstrated a link between ~300 binding and the ElA enhancer repression function.
REFERENCES G. R., and BABISS, L. E. (1990). The efficiency of adenovirus transformation of odent cells is inversely related to the rate of viral ElA gene expression. J. viral. 64, 3427-3436. BANERJI,J., RUSCONI, S., and SCHAFFNER,W. (1981). Expression of a @globin gene is enhanced by remote SV40 DNA sequences. Cell 27,299-308.
ET AL. BERK, A. J., and SHARP, P. A. (1978). Structure of the adenovirus 2 early mRNAs. Cell 14, 695-711. BERK, A., LEE, F., HARRISON,T., WILLIAMS, J., and SHARP, P. A. (1979). A pre-early adenovirus 5 gene product regulates synthesis of early viral messenger RNAs. Cell 17, 935-944. BORRELLI, E., HEN, R., and CHAMBON, P. (1984). Adenovirus-2 ElA products repress enhancer-induced stimulation of transcription. Nature 312, 608-612. BUCHKOVICH,K.. and HARLOW, E. (1989). The retionoblastoma protein is phosphorylated during specific phases of the cell cycle. Cell 58, 1097-l 105. CHEN, S., and PAUCHA, E. (1990). Identification of a region of simian virus 40 large T antigen required for cell transformation. /. Viral. 64,3350-3357. CHOW, L. T., BROKER,T. R., and LEWIS,J. B. (1979). Complex splicing patterns of RNAs from the early regions of adenovirus 2. J. Mol. Biol. 134, 265-303. DECAPRIO, 1. A., LUDLOW, 1. W., LYNCH, D., FURUKAWA,Y., GRIFFIN, J., PIWNICA-WORMS,H., HUANG, C-M., and LIVINGSTON,D. M. (1989). The product of the retinoblastoma susceptibility gene has properties of a cell cycle regulatory element. Ce// 58, 1085-l 095. DYSON, N., BUCHKOVCH, K., WHYTE, P., and HARLOW, E. (1989). The cellular 107K protein that binds to adenovirus ElA also associates with the large T antigens of SV40 and JC virus. Cell 58, 249-255. EGAN, C., BAYLEY, S. T., and BRANTON, P. E. (1989). Binding of the Rbl protein to ElA products is required for adenovirus transformation. Oncogene 4, 383-388. EGAN. C., JELSMA, T. N., HOWE, 1. A., BAYLEY, S. T., and BRANTON, P. E. (1988). Mapping of cellular protein-binding sites on the products of early-region 1A of human adenovirus type 5. Mol. Cell. Biol. 8,3955-3959. EGAN, C., YEE, S. P., FERGUSON,B., ROSENBERG,M., and BRANTON, P. E. (1987). Binding of cellular polypeptides to human adenovirus type 5 El A proteins produced in Escherichia co/i. virology 160, 292-296. FIGGE,J., WEBSTER,T., SMITH, T. F., and PAUCHA, E. (1988). Prediction of similar transforming regions in simian virus 40 large T, adenovirus E 1A, and myc oncoproteins. 1. Viral. 82, 18 14-l 8 18. HARLOW, E., FRANZA. B. R., JR., and SCHLEY, C. (1985). Monoclonal antibodies specific for adenovirus early region 1A proteins; extensive heterogenety in early region 1A products. J. Viral. 55, 533546. HARLOW, E., WHME, P., FRANZA, JR., B. R., and SCHLEY, C. (1986). Association of adenovirus early-region 1A proteins with cellular polypeptides. Mol. Cell. Biol. 6, 1579-l 589. HEN, R., BORRELLI, E., and CHAMBON, P. (1985). Repression of the immunoglobulin heavy chain enhancer by the adenovirus-2 Ela products. Science 230, 1391-l 394. HOUWELING,A., VAN DEN ELSEN, P. J., and VAN DER EB, A. J. (1980). Partial transformation of primary rat cells by the leftmost 4.5% fragment of adenovirus 5 DNA. virology 105, 537-550. JELSMA, T. N., HOWE, J. A., MYMRYK. J. S., EVELEGH,C. M., CUNNIFF, N. F., and BAYLEY, S. T. (1989). Sequences in ElA proteins of adenovirus-5 required for cell transformation, repression of a transcriptional enhancer, and induction of proliferating cell nuclear antigen. Virology 171, 120-l 30. JONES, N.. and SHENK, T. (1979). An adenovirus type 5 early gene function regulates expression of other early viral genes. Proc. Nafl. Acad. Sci USA 76,3665-3669. KIMELMAN, D.. MILLER, J. S., PORTER,D., and ROBERTS,B. E. (1985). Ela regions of the human adenoviruses and of the highly oncogenie simian adenovirus 7 are closely related. J. viral. 53, 399409. KUPPUSWAMY,M. N., and CHINNADURAI, G. (1987). Relationship be-
BY Ad5 ElA
tween the transforming and transcriptional regulatory functions of adenovirus 2 El a oncogene. \/iro/ogy 159, 3 l-38. LILLIE, J. W., GREEN, M., and GREEN, M. R. (1986). An adenovirus El a protein region required for transformation and transcriptional repression. Cell 46, 1043-l 051. LILLIE, 1. W., LOEWENSTEIN, P. M., GREEN, M. R., and GREEN, M. (1987). Functional domains of adenovirus type 5 El a proteins. Cell 50,1091-1100. LUDLOW, J. W., DECAPRIO, J. A., HUANG, C.-M., LEE, W.-H., PAUCHA, E., and LIVINGSTON,D. M. (1989). SV40 large T antigen binds preferentially to an underphosphorylated member of the retinoblastoma susceptibility gene product family. Cell 56, 57-65. LUDLOW, J. W.. SHON, J., PIPAS, J. M., LIVINGSTON, D. M., and DECAPRIO, 1. A. (1990). The retinoblastoma susceptibility gene product undergoes cellcycle dependent dephosphorylation and binding to and release from SV40 large T. Ce// 60, 387-396. MONTEL. G., COURTOIS, G., ENG, C., and BERK. A. (1984). Complete transformation by adenovirus 2 requires both ElA proteins. Cell 36, 951-961. MORAN, E. (1988). A region of S 40 large T antigen can substitute for a transforming domain of the adenovirus ElA products. Nature 334, 168-l 70. MORAN, E., and ZERLER, B. (1988). Interaction between cell growthregulating domains in the products of the adenovirus ElA oncogene. Mol. Cell. Biol. 8, 1756-1764. MORAN, E.. ZERLER,B., HARRISON,T. M., and MATHEWS, M. B. (1986). Identification of separate domains in the adenovirus ElA gene for immortalization activity and the activation of virus early genes. Mol. Cell. Biol. 6, 3470-3480. NEVINS, J. R. (1982). Induction of the synthesis of a 70,000 dalton mammalian heat shock protein by adenovirus El a gene product. Cell 29, 913-919. OFFRINGA. R.. GEBEL, S.. VAN DAM, H., TIMMERS, M., SMITS, A., ZWART, R., STEIN, B., Bos, J. L.. VAN DER EB, A., and HERRLICH, P. (1990). A novel function of the transforming domain of El a: Repression of AP-1 activity. Cell 62, 527-538. OFFRINGA, R., SMITS, A. M. M., HOUWELLING,A., Bos, J. L.. and VAN DER Ee, A. 1. (1988). Similar effects of adenovlrus ElA and glucocorticoid hormones on the expression of the metalloprotease stromelysin. Nucleic Acids Res. 16, 10,973-l 0.984. PERRICAUDET,M.. AKUSJARVI,G., VIRTANEN, A., and PETTERSSON,U. (1979). Structure of two spliced mRNAs from the transforming region of human subgroup C adenoviruses. Nature (London) 281, 694-696. ROCHETTE-EGLY.C.. FROMENTAL, C., and CHAMBON, P. (1990). General repression of enhancon activity by the adenovirus-2 ElA proteins. Genes Dev. 4, 137-l 50. RULEY, H. E. (1983). Adenovirus early region 1H enables viral and cellular transforming genes to transform primary cells in culture. Nature 304, 602-606. SCHNEIDER,J. F., FISHER, F., GODING, C. R., and JONES, N. C. (1987). Mutational analysis of the adenovirus ElA gene: The role of transcriptional regulation in transformation. EM50 J. 6, 2053-2060. SOGAWA. K., HANDA, H., FUJISAWA-SEHARA,A., HIROMASA,T., YAMANE, M., and FUJII-KURIYAMA, Y. (1989). Repression of cytochrome P-450~ gene expression by cotransfection with adenovirus ElA DNA. Eur. J. Biochem. 181, 539-544. STEIN, R., and ZIFF, E. B. (1984). HeLa cells P-tubulin gene transcrip-
CR1 OR CR2 MUTANTS
tion is stimulated by adenovirus 5 in parallel with viral early genes by an El a-dependent mechanism. Mol. Cell. Biol. 4, 2792-2801. STEIN, R. W.. and ZIFF. E. B. (1987). Repression of insulin gene expression by adenovirus type 5 ElA proteins. Mol. Gel/. Biol. 7, 1164-1170. STEPHENS, C., and HARLOW, E. (1987). Differential splicing yields novel adenovirus 5 ElA mRNAs that encode 30kD and 35kD proteins. (1987). EM50 J. 6, 2027-2035. SUBRAMANIAN,T., KUPPUSWAMY,M., NASR, R. J., and CHINNADURAI, G. (1988). An N-terminal region of adenovlrus El a essential for cell transformation and induction of an epithelial cell growth factor. Oncogene2, 105-112. SVENSSON,C., and AKUSJWRVI,G. (1984). Adenovirus 2 early region 1A stimulates expression of both viral and cellular genes. EILIBOI. 3, 789-794. SVENSSON,C., PEITERSSON, U., and AKUSJ~~RVI, G. (1983). Splicing of adenovirus 2 early region 1A mRNAs is non-sequential. J. Mol. Biol. 165, 475-499. TIMMERS. H. T. M., DE WIT, D., Bos, J. L., and VAN DEREB. A. J. (1988). ElA products of adenoviruses reduce the expression of cellular proliferation-associated genes. Oncogene Res. 3, 67-76. TIMMERS, H. T. M., VAN DAM, H., PRONK, G. J., Bos. J. L., and VAN DER EB, A. J. (1989). Adenovirus ElA represses transcription of the cellular JE gene. /. Viral. 63, 1470-l 473. ULFENDAHL, P. J.. LINDER, S.. KREIVI,J.-P., NORDQVIST, K., SVENSSON, C., HULTBERG, H., and AKUSJ~~RVI, G. (1987). A novel adenovirus-2 ElA mRNAencoding a protein with transcription activation properties. EM50 J. 6, 2037-2044. VAN DAM, H.. OFFRINGA, R., SMITS, A. M. M., Bos, J. L., and VAN DER EB, A. J. (1989). The repression of the growth factor-inducible genes JE, c-myc and stromelysin by adenovirus ElA is mediated by conserved region 1. Oncogene 4, 1207-1212. VAN DEN ELSEN, P., HOUWELING,A., and VAN DEREB, A. (1983). Expression of region El b of human adenoviruses in the absence of region Ela is not sufficient for complete transformation. Virology 128, 377-390. VELCICH. A., and ZIFF, E. B. (1985). Adenovirus El a proteins repress transcription from the SV40 early promoter. Cell 40, 705-716. VELCICH, A., and ZIFF, E. B. (1988). Adenovirus ElA ras cooperation activity is separate from its positive and negative transcription regulatory functions. Mol. Cell. Biol. 8, 2177-2183. WEBSTER,K. A., MUSCAT, G. E. O., and KEDES. L. (1988). Adenovirus ElA products suppress myogenic differentiation and inhibit transcription from muscle-specific promoters. Nature 322, 553-557. WHME, P., BUCHKOVICH, K. J., HOROWITZ,J. M., FRIEND, S. H.. RAYBUCK, M., WEINBERG, R. A., and HARLOW, E. (1988). Association between an oncogene and an antioncogene: The adenovirus ElA proteins bind to the retinoblastoma gene product. Nature 334, 124-129. WHYTE. P., WILLIAMS, N. M., and HARLOW, E. (1989). Cellular targets for transformation by ElA proteins. Cell 56, 67-75. WIGLER, M., PELLICER,A., SILVERSTEIN,S., and AXEL, R. (1978). Biochemical transfer of single copy eukaryotic genes using total cellular DNA as clonor. Cell 14, 725-731, YEE, S.. and BRANTON, P. E. (1985). Detection of cellular proteins associated with human adenovirus type 5 early region 1A polypeptides. Virology 147, 142-153. YOUNG, K. S., WEIGEL, R.. HIEBERT,S., and NEVINS,1. R. (1989). Adenovirus El A-mediated negative control of genes activated during F9 differentiation. Mol. Cell. Biol. 9, 3109-31 13.