DONALD V. HECK, CAROLE L. YEE, PETER M. HOWLEY, AND KARL MUNGER*. Laboratory of Tumor Virus Biology, National Cancer Institute, Bethesda, MD ...
Proc. Natl. Acad. Sci. USA Vol. 89, pp. 4442-4446, May 1992 Biochemistry
Efficiency of binding the retinoblastoma protein correlates with the transforming capacity of the E7 oncoproteins of the human papillomaviruses DONALD V. HECK, CAROLE L. YEE, PETER M. HOWLEY, AND KARL MUNGER* Laboratory of Tumor Virus Biology, National Cancer Institute, Bethesda, MD 20892
Communicated by Bernard Fields, January 28, 1992
The HPV-16 E7 gene encodes an acidic zinc-binding nuclear phosphoprotein with an apparent molecular size of 18 kDa (4, 14, 15). The high-risk HPV E7 proteins share structural and functional similarities with the adenovirus (Ad) ElA proteins and with the large tumor antigen (TAg) of simian virus 40 (SV40) (16,17). Like Ad ElA and SV40 TAg, E7 can transactivate the Ad E2 promoter (16-18), cooperate with an activated ras oncogene to transform primary baby rat kidney (BRK) cells (16, 17, 19), abrogate the transforming growth factor type j3 (TGF-pB)-induced transcriptional repression of the c-myc promoter (20), and form a complex with pRB (10, 21, 22). The E7 proteins encoded by the low-risk HPVs also share amino acid similarity with Ad ElA and SV40 TAg but differ from the high-risk HPV E7 proteins in a number of biological and biochemical properties. They have a decreased apparent affinity for pRB (23-25), do not efficiently function in cellular transformation (19, 26-28), and do not interfere with the TGF-f-mediated repression of c-myc transcription (26) but are still able to transactivate the Ad E2 promoter (26, 27). The molecular basis for these differences has been mapped to the amino-terminal half of the E7 protein (26). The amino-terminal halves of high-risk and low-risk E7 proteins contain consensus recognition sequences for casein kinase II (CK II) (25, 29). It has been shown that the E7 proteins of the high-risk HPVs are phosphorylated in vitro at a higher rate than the low-risk HPV-encoded E7 proteins (25). Moreover, the HPV-16 E7 protein runs aberrantly on SDS/polyacrylamide gels, whereas the HPV-6 E7 protein runs close to its predicted molecular size (24, 26). The structural basis for these biochemical and biological differences between the E7 proteins encoded by the high-risk and low-risk HPVs has not been extensively studied, and it is not known which of these differences correlate with the different in vitro transformation activities distinguishing the high-risk from the low-risk viruses. Despite the increasing evidence (11, 13) for the importance of the E7/pRB interaction for cellular transformation, additional elements in the amino-terminal half of HPV-16 E7 have been implicated in the transformation function of the E7 protein. Certain mutations within the conserved region 1 (CR1) homology domain of HPV-16 E7 impair transformation, although they do not effect binding to pRB (30, 31). It has also been shown that some mutations in the CK II phosphorylation site result in decreased cellular transformation potential without affecting the apparent pRB binding affinity (25, 32). To investigate the characteristics of the high-risk HPV-16 and low-risk HPV-6-encoded E7 proteins that are responsible for the difference in transformation potential, a series of
The human papillomaviruses (HPVs) associABSTRACT ated with genital tract lesions can be classified as either "high risk" or "low risk" based on their association with human anogenital cancer. The E7 proteins of the high-risk and the low-risk viruses are quite similar in their amino acid composition and structural organization yet differ in their transforming potential and in a number of biochemical properties. A series of chimeric proteins consisting of segments of the highrisk HPV-16 and the low-risk HPV-6 E7 proteins were constructed in order to define which domains within the aminoterminal half of E7 were responsible for the different biological and biochemical properties. The E7 oncogenic capacity, which was determined by assaying transformation of baby rat kidney cells in cooperation with an activated ras oncogene, segregated with the retinoblastoma tumor suppressor protein (pRB) binding domain of the HPV-16 E7 protein. A comparison ofthe pRB binding sites of the sequenced genital tract HPVs revealed a consistent amino acid difference (aspartic acid/glycine) between the high-risk and low-risk viruses. Single amino acid substitution mutations were generated at this position in the HPV-6 and HPV-16 E7 proteins, and this single amino acid residue was shown to be the principal determinant responsible for the differences in the apparent pRB binding affinity and tranormation capacity distinguishing the HPV E7 proteins of the high-risk and low-risk HPVs.
Approximately 20 of the >60 different human papillomavirus types (HPVs) are associated with anogenital tract lesions (1). These genital tract associated HPVs can be subdivided into two groups: the "low-risk" HPVs, such as HPV-6 and HPV-11, which are associated with benign lesions including condyloma acuminata that are at a low risk for malignant progression, and the "high-risk" HPVs, such as HPV-16 and HPV-18, which are associated with cervical intraepithelial neoplasia and cervical carcinoma (2). In cervical cancers and derived cell lines, the HPV DNA is usually integrated into the host genome, and only a subset of the viral genes, E6 and E7, are consistently expressed (3-5). E6 and E7 encode oncoproteins that together are sufficient for efficient immortalization of primary human squamous epithelial cells (6, 7). The E6 proteins of the high-risk HPVs form specific complexes
with p53, a cellular protein with tumor suppressor activity, and can promote the degradation of p53 in vitro (8, 9). The E7 protein can form a complex with the retinoblastoma tumor suppressor protein (pRB) (10). The functional inactivations of p53 and pRB through their interactions with E6 and E7, respectively, are believed to be important steps in cervical carcinogenesis. Recent studies demonstrating that p53 and pRB are mutated in HPV-negative cervical carcinoma cell lines, but are wild-type in HPV-positive cervical carcinoma cell lines, support this concept (11-13).
Abbreviations: HPV, human papillomavirus; pRB, retinoblastoma tumor suppressor protein; BRK, baby rat kidney; CK II, casein kinase II; CR1, conserved region 1; CR2, conserved region 2; Ad, adenovirus; SV40, simian virus 40; GST, glutathione S-transferase; TAg, large tumor antigen. *To whom reprint requests should be addressed.
The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact. 4442
Proc. Natl. Acad. Sci. USA 89 (1992)
Biochemistry: Heck et al.
capable of transactivating the Ad E2 promoter (16, 17, 26, 27). Therefore each of the individual mutant E7 proteins that are depicted in Fig. 1 was tested for its ability to transactivate the Ad E2 promoter in transient transfection assays using CV-1 cells. All mutant E7 proteins were able to efficiently transactivate the Ad E2 promoter (data not shown), indicating that each of these mutants expressed a functionally stable E7 protein. Sequences in the CR1 Homology Domain of HPV-16 E7 Determine Its Aberrant Electrophoretic Mobility. The chimeric E7 proteins were synthesized by in vitro transcription/in vitro translation in the presence of [35S]cysteine and analyzed by SDS/PAGE followed by autofluorography (data not shown). The aberrant slow electrophoretic migration segregated with the HPV-16 E7 CR1 homology domain (see Fig. 2A). Neither the specific amino acid substitutions in the pRB binding region, the CK II site, nor the CR1 point mutation (V6) had any effect on the electrophoretic mobility (see Fig. 2B and data not shown). The pRB Binding Region Is the Major Determinant for the Efficiency of pRB Binding. The relative apparent pRB binding affinity of the individual mutant E7 proteins was next measured using an in vitro coprecipitation assay. Equimolar amounts of [35S]cysteine-labeled E7 proteins were synthesized by in vitro translation and mixed with unlabeled extracts of human ML-1 cells that contain normal pRB. Complex formation between E7 and pRB was measured by coprecipitation using the pRB-specific monoclonal antibody C36. Proteins were resolved by SDS/PAGE, and coprecipitated E7 proteins were visualized by autofluorography. Quantitation of the E7 protein coprecipitating with pRB in this assay, as shown in Fig. 2A, provides a measure of the relative apparent binding affinity of the E7 protein for pRB (23). No coprecipitated E7 proteins were detected using the SV40 TAg-specific monoclonal antibody pAb419 as a control (data not shown). The chimeric proteins that contained the HPV-16 E7 pRB binding domain bound to pRB with a high apparent affinity similar to that of wild-type HPV-16 E7. Conversely, the chimeric E7 proteins that contained the HPV-6 E7 pRB binding domain bound to pRB with the lower apparent affinity characteristic of the HPV-6 E7 protein. A Single Amino Acid Difference Within the pRB Binding Domains of HPV-16 E7 and HPV-6 E7 Is Primarily Responsible for the Differences in Apparent pRB Binding Affinities. Comparison of the amino acid sequences of the high-risk and the low-risk HPV E7 proteins corresponding to the CR2
HPV-16/HPV-6 chimeric E7 proteins that contain all possible combinations of the CR1 homology domain, the pRB binding site, and the CK II site was constructed (Fig. 1). Additional amino acid substitution mutants were engineered in which specific amino acid residues conserved in the high-risk HPV E7 were exchanged with the corresponding conserved residues of the low-risk HPV-6 E7 proteins and vice versa. Finally, HPV-16 E7 CK II phosphorylation site mutants were constructed by substituting serines 31 and 32 with two alanine, two cysteine, or two aspartic acid residues.
MATERIALS AND METHODS Plasmids. The amino-terminal half of the HPV-16 and HPV-6 E7 genes was divided into three regions (the CR1 homology domain, the pRB binding site, and the CK II site) and reconstructed using synthetic oligonucleotides. The oligonucleotides contained the following codons: for HPV-16 E7, amino acids 1-17, 18-28, 29-50; for HPV-6 E7, amino acids 1-17, 18-29, 30-50. Point mutations were generated by altering codons in the appropriate oligonucleotides. The newly constructed amino-terminal halves ofthe E7 gene were cloned between the HindIII and Ban II sites of p858 (16) downstream of the SV40 early promoter and enhancer. Protein Methods. E7/pRB coprecipitation experiments were carried out as described (23). Wild-type and mutant glutathione S-transferase (GST)-E7 fusion proteins were expressed in Escherichia coli HB101 and purified as described (18, 33). CK II reactions were carried out in a final volume of 30 ,ul containing 1 gg of GST-E7 in 20 mM Hepes (pH 7.5), 20 mM MgCl2, and 56 nM [y-32P]ATP (>3000 Ci/mmol; 1 Ci = 37 GBq; Amersham). Purified CK II (from bovine testis) was diluted 1:200 in enzyme dilution buffer (1 mg of bovine serum albumin per ml/5 mM Mops, pH 7.0/200 mM NaCI). Reactions were started by addition of 6 ,ul of CK II; Products were analyzed by SDS/10o PAGE followed by radioanalytic imaging (AMBIS Systems, San Diego). Mammalian Cell Transfections and Chloramphenicol Acetyltransferase (CAT) Assays. Primary BRK cells were cotransfected with 5 ,ug of E7 (mutant) plasmid and 5 ,g of pEJras as described (16). Transfection of CV-1 cells and CAT assays were carried out as described (16). RESULTS Functional Assessment of Stability of the Mutant E7 Proteins. The E7 proteins of the high-risk and low-risk HPVs are
CR2
CR1 _ r|
Ad5E1A
H F E P P T L H E
HPV-16 E7 HPV-6 E7
H H
-
pRB binding site
L Y 0 L ... U P
LL P T LHE D T R IU T LKDIULDL 6 6 6 16 16 16
-
E U
CKII site
DL T C H E A I F P P S D D E
-
D E
P- E TTLVCY Q L D S SEEE-DE... U L HCYE L UISSEEUIE...
CP
6
16 16 6 16 6
16 6 1 6 6 16
16 Dai-G 6 G22-D
16
16 16 16
16 16 16
le Mi~ra-CC 16 SSOas.AA 16 86514-OD
6
6
16 6
4443
6
.6
FIG. 1. Structures of the mutant E7 proteins used in this study. The amino-terminal 37 amino acids of the HPV-16 and HPV-6 E7 proteins are similar in sequence to portions of Ad ElA CR1 and CR2 (16, 17). The sequence similar to Ad ElA CR2 contains the pRB binding site (23) as well as the CK II consensus sequence (25, 29). All chimeric and mutant E7 proteins contain the carboxyl terminus of HPV-16 E7.
Biochemistry: Heck et al.
4444
Proc. Natl. Acad. Sci. USA 89 (1992)
A CR1
16
pRB bdg site
CKII site
6
16
16
6
6
16
6
16
6
16
6
16
6
6
16
6
16
16
6
16
6
16
6
*! .,..s..tw ~~~~~~ .o,,.,w,,..
% wt
HPV-16 E7 pRB bdg
124
21
74
42
113
26
100
14
6
16
B CR1
pRB bdg
16
site
CKII site
16
(G1)
16
6
6
(D.) 6
16
16
16
16
16
16
6
16
16
16
16
6
16
14
100
(CC.i=) (DDsma) (AAa,.w)
ewt HPV-16 E7 pRB bdg
44
73
118
117
97
FIG. 2. Complex formation of the chimeric (A) and amino acid substitution mutant (B) E7 proteins with the pRB in vitro. Equimolar amounts of the [35S]cysteine-labeled mutant E7 proteins synthesized by in vitro translation were mixed with unlabeled whole cell extracts of human ML-1 cells, which contain normal pRB. E7 proteins coprecipitated with pRB were analyzed by SDS/PAGE followed by autofluorography (top). The E7 pRB binding (bdg) values shown underneath each lane represent average values of several independent experiments and were obtained by radioanalytic imaging. wt,
Wild-type. Molecular size is indicated in kDa.
homology domain containing the pRB binding site revealed a single consistent sequence difference, an aspartic acid residue (D21) in the high-risk E7 sequence and a glycine residue (G22) in the low-risk E7 sequence. Therefore E7 proteins with single amino acid substitutions (HPV-16 E7 G21 and HPV-6 E7 D22) were synthesized and tested for pRB binding (Fig. 2B). The HPV-16 E7 G21 protein had a decreased level of pRB binding, 41% that of wild-type HPV-16 E7 and very similar to the level of binding of the chimeric E7 protein containing the HPV-6 E7 pRB binding region in the context of HPV-16 E7 (42%). The converse substitution mutation in HPV-6 E7 (D22) resulted in an increased level of pRB binding of 73% compared to 14% for wild-type HPV-6 E7. This level of binding was nearly equal to that observed with the chimeric E7 protein containing the entire HPV-16 E7 pRB binding domain in the context of the HPV-6 E7 protein (74%) (Fig. 2 A and B). The mutations generated in the CK II site of HPV-16 E7 had no significant effect on pRB binding (Fig. 2B). Similarly, the HPV-16 E7 V6 mutant protein bound to pRB with wild-type efficiency (data not shown). Sequence Differences in the HPV-16 E7 and HPV-6 E7 CK II Recognition Sites Are Responsible for the Different in Vitro Rates of Phosphorylation by CK IH. To study the phosphorylation of the various E7 mutant proteins by CK II in vitro, some of the chimeric and phosphorylation-defective E7 proteins were synthesized in E. coli as GST fusion proteins. The purified GST-E7 fusion proteins were used as substrates for phosphorylation with CK II in the presence of adenosine 5'-[y-32P]triphosphate in vitro and analyzed by SDS/PAGE and radioanalytic imaging (data not shown). The specificity of the phosphorylation reaction was demonstrated with the CK II phosphorylation mutant HPV-16 E7 AA31/32 as a substrate. Consistent with an earlier report (25), HPV-16 E7 was
phosphorylated at a 2-fold higher rate than the protein containing the HPV-6 E7 amino terminus. This difference in the rates of CK II phosphorylation was solely dependent on the CK II recognition sequences, as demonstrated by the rates of phosphorylation of the mutants containing the CK II phosphorylation sequences exchanged between the two proteins (data not shown). The pRB Binding Site Is the Primary Determinant for the Efficient Transformation Phenotype of HPV-16 E7. To determine the structural and biochemical determinants responsible for the observed differences in cellular transformation between the high-risk and low-risk HPV-encoded E7 proteins, the various E7 mutants were cloned into a eukaryotic expression vector and tested for their ability to transform primary BRK cells in cooperation with an activated ras oncogene. Efficient transformation of BRK cells segregated with the HPV-16 E7 pRB binding site. The HPV-16 G21 E7 mutant, which exhibited a 2.5-fold decreased apparent affinity for pRB binding relative to wild-type HPV-16 E7, was impaired for transformation activity compared to wild-type HPV-16 E7. The converse mutant, HPV-6 D22 E7, which had a 5-fold increased pRB binding efficiency compared to wildtype HPV-6 E7, had significant transformation activity in this assay (Table 1). Therefore a close correlation was observed between the transformation potential of the E7 protein and its apparent pRB binding affinity. The HPV-16 V6 E7 mutant showed transformation properties equal to wild-type HPV-16 E7 (data not shown). Finally, the substitution of the serine residues 31 and 32 in the CK II site by nonphosphorylatable alanine, cysteine, or aspartic acid residues had only a minimal effect on the transformation capacity (Table 1).
DISCUSSION A number of studies have implicated the biologic activities of the E6 and E7 genes of the genital-associated HPVs as the major determinants responsible for the cancer association of the high-risk viruses. There is a significant difference in the transformation potential of HPV-16 E7 and HPV-6 E7 that can be assayed by transformation of BRK cells in cooperation with ras. The molecular determinants for this difference have previously been mapped to the amino-terminal 50 amino acid residues of the E7 proteins (26). The data presented in this study define in more detail the domains within the amino-terminal 50 amino acids responsible for the major biochemical and biological differences between the low-risk and high-risk HPV-encoded E7 proteins. The CR1 homology domain of HPV-16 E7 conferred the property of aberrant electrophoretic migration on SDS/ polyacrylamide gels, although the reason for this aberrant migration is still unknown. It may reflect some posttranslational modification of the protein or, alternatively, it may be indicative of a structure that prevents SDS molecules from binding efficiently to the protein. This second possibility may be more likely since this aberrant migration was also observed for HPV-16 E7 synthesized in wheat germ lysates (Fig. 2) or as a fusion protein in E. coli (data not shown). The determinant for efficient pRB binding localized primarily to the HPV-16 E7 pRB binding domain. The substitution of the HPV-16 E7 pRB binding site into the background of the HPV-6 E7 protein led to a significant increase in pRB binding. The further substitution of either the HPV-16 CR1 homology region or the CK II site resulted in a further slight enhancement in the apparent affinity of the chimeric E7 proteins for pRB. The apparent affinity of the HPV-6 E7 pRB binding site was also somewhat enhanced when flanked by the HPV-16 E7 CR1 homology domain or the CK II site either separately or in combination. Comparison of the amino acid sequences of the high-risk- with those of the low-risk-derived
Biochemistry: Heck et al.
Proc. Natl. Acad. Sci. USA 89 (1992)
4445
Table 1. Transformation properties of the mutant E7 proteins Foci, no. per 10 Ag of DNA ras cooperativity, no. positive/ Exp. 4 Exp. 1 Exp. 2 Exp. 3 CK II site no. of experiments Crl pRB binding site ND 142 67 ND 16 16 6 5/5 0 0 ND 16 0/5 ND 6 6 44 ND 23 ND 6 6 4/5 16 ND 0 0 ND 16 0/5 16 6 ND 28 40 ND 16 5/5 6 16 0 ND 0 ND 16 6 6 0/5 13 ND 8 22 16 16 5/8 16 D21-G 6 ND 8 58 59 6 6 G22-D 6/6 ND 120 13 ND 16 SS31/3r-CC 5/6 16 16 ND 36 12 ND 16 SS31/32-AA 5/6 16 16 ND 94 32 ND 16 SS31/32-DD 5/6 16 16 0 6 0 0 0 6 6 1/10 204 159 16 16 16 22 122 10/10 The indicated E7 plasmids and pEJras were cotransfected by calcium phosphate precipitation into BRK cells. The results are presented in two ways. Under "ras cooperativity," the number of experiments that yielded transformed foci is shown as well as the total number of experiments. The ras cooperativity was assessed positive when more than two foci per 10 ,tg of DNA were counted 3 weeks after transfection. Under "Foci no. per 10 ,g of DNA," the actual number of foci counted in individual experiments 2 weeks after transfection is shown. ND, not done. Negative controls with E7 alone, pEJras alone, and pEJras with p1197 containing a disrupted HPV-16 E7 gene (16) for each experiment yielded no transformants.
E7 proteins revealed that the pRB binding sites differed consistently at only a single position: an aspartic acid residue present in each of the high-risk-derived E7 proteins (D21 in HPV-16 E7) corresponding to a glycine residue in each of the low-risk HPV-encoded E7 proteins (G22 in HPV-6 E7). Single amino acid substitutions of these residues had a dramatic effect on the pRB binding efficiency and indicated that this single amino acid difference in the E7 pRB binding domains of low-risk and high-risk HPVs is the primary determinant of pRB binding efficiency. The cellular transformation capacities of the individual E7 mutants were assayed in cooperation with ras using BRK cells. All chimeric E7 proteins that contained the pRB binding sequence derived from HPV-16 E7 were transformation competent in this assay, and none of the chimeric E7 proteins containing the HPV-6 E7 pRB binding site demonstrated any transforming activity in this assay. The point mutation of G22 to D within HPV-6 E7, which increased pRB binding to 73% of wild-type HPV-16 E7, by itself converted HPV-6 E7 into a transformation-competent protein. Conversely, the D21 to G mutation in HPV-16 E7, which decreased pRB binding to 41% of wild-type HPV-16 E7, severely impaired, but did not totally abolish, the transformation activity of this protein. Thus for the anogenitalassociated HPVs, the presence of a glycine or an aspartic acid residue at this sequence position of the E7 protein is the major determinant of pRB binding efficiency and of cellular transformation. The retinoblastoma protein is only one of a group of cellular proteins that interact with the sequences of the pRB binding site of HPV E7 (K.M., unpublished). For Ad ElA and SV40 TAg it was shown that the pRB binding site is also involved in complex formation with several additional cellular proteins, including p107, cyclin A, and p130 (34-38). It will be important to determine whether these proteins can also interact with HPV E7 and whether the binding activities of these proteins correlate with cellular transformation as shown here for E7 and pRB. The ability of E7 to cooperate with ras to transform BRK cells provided evidence for the functional similarities between E7 and Ad ElA, and most of our understanding on the mechanism of cellular transformation by the HPV E7 oncoproteins stems from experiments performed with this system (16, 17, 19). For the efficient transformation of primary human keratinocytes, the natural host cell of the HPVs, E6
and E7 are necessary (6, 7). To define the domains of E7 that are important for immortalization of primary human keratinocytes, the HPV E7 chimeric genes will need to be built back into the full-length HPV-16 genome. Recent studies have shown that pRB is mutated in HPV-negative cervical carcinoma cell lines and wild-type in the HPV-positive cervical carcinoma cell lines. These data suggest that inactivation of pRB either by mutation or through association with E7 is an important step in cervical carcinogenesis (11, 13). Thus it seems likely that the higher pRB binding efficiency characteristic of the high-risk E7 proteins would be a determinant involved in the transformation of primary human keratinocytes. All mutant E7 proteins described in this study transactivated the Ad E2 promoter, in agreement with previous reports that the low-risk E7 proteins were also able to transactivate the Ad E2 promoter (26, 27). As recently shown, HPV-16 E7 can alter the interactions of cellular proteins with the E2F transcription factor similar to the 12S Ad ElA protein (18, 39). One of the cellular proteins that is associated with E2F is pRB (40-43). These findings may be difficult to reconcile with the observation that E7 proteins that have low efficiency pRB binding and score negative for cellular transformation can still efficiently transactivate the Ad E2 promoter. One possible explanation is that the levels of E7 produced in the transactivation assays are higher than those in the BRK cells and that high levels of E7 may be sufficient to displace pRB and release free E2F resulting in Ad E2 transactivation. Alternatively, activation of E2F by displacement from the pRB complex may be sufficient for Ad E2 transactivation but may not be sufficient to induce cellular transformation. Indeed, E2F exists in at least two distinct macromolecular complexes, one of which contains cyclin A (44). The transactivation activity of E7 may be reflective of a common function of the E7 proteins of low-risk and high-risk HPVs. The differing rates of CK II phosphorylation previously reported for the high-risk and the low-risk HPV-encoded E7 proteins (25) were localized to sequences containing the corresponding CK II recognition sites. All of the phosphorylation-defective proteins retained relatively high transformation activity (Table 1). The reason for the disparity between these data and a study by Firzlaff et al. (32) is not clear but may reflect differences in the sensitivities in the ras cotransformation assays performed in the two different lab-
4446
Biochemistry: Heck et al.
oratories. It is clear, however, that although the CK II sites of the low-risk and high-risk HPV-derived E7 proteins differ in their rate of phosphorylation by CK II, they do not contribute in a significant manner to the pRB binding or cellular transformation properties. Several mutagenesis studies of high-risk HPV E7 proteins have suggested that there are amino acid sequences in the CR1 homology domain that are required for cellular transformation that themselves do not effect pRB binding (30, 31). This additional function must be common to HPV-6 E7 and HPV-16 E7 since this domain is interchangeable between the two proteins with no effect on transformation. It is interesting to speculate that the E7 CR1 homology domain may interact with another cellular protein or a set of cellular proteins and that these interactions may be important for biological functions that are common to the E7 proteins of the low-risk and the high-risk HPVs. This study establishes that the E7 protein is a modular protein with domains that can be exchanged between HPV types with no significant effects on the functional stability of the protein, as measured by the Ad E2 transactivation assay. The presence of a high-affinity pRB binding site, which is likely to be involved in the binding of cellular proteins in addition to pRB, segregates as the major determinant for efficient cellular transformation. Moreover, a single consistent amino acid change in the pRB binding sites of the low-risk- and the high-risk-derived E7 proteins is largely responsible for the observed differences in pRB binding and cellular transformation. We are grateful to W. C. Phelps, M. Scheffner, and J. M. Huibregtse for a critical reading of this manuscript. We thank A. McBride for help with the CK II assays, D. Litchfield for the gift of purified CK II, E. Harlow for the ML-1 cell line, and J. Byrne for oligonucleotide synthesis. D.V.H. was supported by the Howard Hughes Medical Institute-National Institutes of Health Research Scholars Program. 1. DeVilliers, E. M. (1989) J. Virol. 63, 4898-4903. 2. zurHausen, H. & Schneider, A. (1987) in The Papovaviridae, eds. Salzmann, N. P. & Howley, P. M. (Plenum, New York), Vol. 2, pp. 245-263. 3. Schwarz, E., Freese, U. K., Gissmann, L., Mayer, W., Roggenbuck, B., Stremlau, A. & zurHausen, H. (1985) Nature (London) 314, 111-114. 4. Smotkin, D. & Wettstein, F. 0. (1986) Proc. Nat!. Acad. Sci. USA 83, 4680-4684. 5. Baker, C. C., Phelps, W. C., Lindgren, V., Braun, M. J., Gonda, M. A. & Howley, P. M. (1987) J. Virol. 61, 962-971. 6. Munger, K., Phelps, W. C., Bubb, V., Howley, P. M. & Schlegel, R. (1989) J. Virol. 63, 4417-4421. 7. Hawley-Nelson, P., Vousden, K. H., Hubbert, N. L., Lowy, D. R. & Schiller, J. T. (1989) EMBO J. 8, 3905-3910. 8. Werness, B. A., Levine, A. J. & Howley, P. M. (1990) Science 248, 76-79. 9. Scheffner, M., Werness, B. A., Huibregtse, J. M., Levine, A. J. & Howley, P. M. (1990) Cell 63, 1129-1136. 10. Dyson, N., Howley, P. M., Monger, K. & Harlow, E. (1989) Science 234, 934-937. 11. Scheffner, M., Monger, K., Byrne, J. C. & Howley, P. M. (1991) Proc. Natl. Acad. Sci. USA 88, 5523-5527. 12. Crook, T., Wrede, D. & Vousden, K. H. (1991) Oncogene 6, 873-875.
Proc. Natl. Acad. Sci. USA 89 (1992) 13. Wrede, D., Tidy, J. A., Crook, T., Lane, D. & Vousden, K. H.
(1991) Mol. Carcinog. 4, 171-175. 14. Barbosa, M., Lowy, D. R. & Schiller, J. T. (1989) J. Virol. 63, 1404-1407. 15. Sato, H., Watanabe, S., Furuno, A. & Yoshiike, K. (1989) Virology 170, 311-315. 16. Phelps, W. C., Yee, C. L., Monger, K. & Howley, P. M. (1988) Cell 53, 539-547. 17. Phelps, W. C., Yee, C. L., Monger, K. & Howley, P. M. (1988) Curr. Top. Microbiol. Immunol. 144, 153-166. 18. Phelps, W., Bagchi, S., Barnes, J., Raychaudhuri, P., Krause, V., Monger, K., Howley, P. M. & Nevins, J. R. (1991) J. Virol. 65, 6922-6930. 19. Storey, A., Pim, D., Murray, A., Osborn, K., Banks, L. & Crawford, L. (1988) EMBO J. 7, 1815-1820. 20. Pietenpol, J. A., Stein, R. W., Moran, E., Yaciuk, P., Schlegel, R., Lyons, R. M., Pittelkow, M. R., Monger, K., Howley, P. M. & Moses, H. L. (1990) Cell 61, 777-785. 21. Whyte, P., Buchkovich, K. J., Horowitz, J. M., Friend, S. H., Raybuck, M., Weinberg, R. A. & Harlow, E. (1988) Nature (London) 334, 124-129. 22. DeCaprio, J. A., Ludlow, J. W., Figge, J., Shew, J.-Y., Huang, C.-M., Lee, W.-H., Marsilio, E., Paucha, E. & Livingston, D. M. (1988) Cell 54, 275-283. 23. Munger, K., Werness, B. A., Dyson, N., Phelps, W. C., Harlow, E. & Howley, P. M. (1989) EMBO J. 8, 4099-4105. 24. Gage, J. R., Meyers, C. & Wettstein, F. 0. (1990) J. Virol. 46, 723-730. 25. Barbosa, M. S., Edmonds, C., Fisher, C., Schiller, J. T., Lowy, D. R. & Vousden, K. H. (1990) EMBO J. 9, 153-160. 26. Monger, K., Yee, C. L., Phelps, W. C., Pietenpol, J. A., Moses, H. L. & Howley, P. M. (1991) J. Virol. 65, 3943-3948. 27. Storey, A., Osborn, K. & Crawford, L. (1990) J. Gen. Virol. 71, 165-171. 28. Barbosa, M. S., Vass, W. C., Lowy, D. R. & Schiller, J. T.
(1991) J. Virol. 65, 292-298.
29. Firzlaff, J. M., Galloway, D. A., Eisenman, R. N. & Loscher, B. (1989) New Biol. 1, 44-53. 30. Edmonds, C. & Vousden, K. H. (1989) J. Virol. 63, 2650-2656. 31. Phelps, W. C., Munger, K., Yee, C. L., Barnes, J. A. & Howley, P. M. (1992) J. Virol. 66, 2418-2427. 32. Firzlaff, J. M., Loscher, B. & Eisenman, R. N. (1991) Proc. Natl. Acad. Sci. USA 88, 571-574. 33. Smith, D. B. & Johnson, K. S. (1988) Gene 67, 31-40. 34. Dyson, N., Buchkovich, K., Whyte, P. & Harlow, E. (1989) Cell 58, 249-255. 35. Ewen, M. E., Ludlow, J. W., DeCaprio, J. A., Millikan, R. C., Cheng, S. H., Paucha, E. & Livingston, D. M. (1989) Cell 58, 257-267. 36. Whyte, P., Williamson, N. M. & Harlow, E. (1989) Cell 56, 67-75. 37. Giordano, A., Whyte, P., Harlow, E., Franza, B. R., Beach, D. & Draetta, G. (1989) Cell 58, 981-990. 38. Pines, J. & Hunter, T. (1990) Nature (London) 346, 760-763. 39. Bagchi, S., Raychaudhury, P. & Nevins, J. R. (1990) Cell 62, 659-669. 40. Chellappan, S. P., Hiebert, S., Mudryj, M., Horowitz, J. M. & Nevins, J. R. (1991) Cell 65, 1053-1061. 41. Bagchi, S., Weinmann, R. & Raychaudhuri, P. (1991) Cell 65, 1063-1072. 42. Bandara, L. R. & La Thangue, N. B. (1991) Nature (London) 351, 494-497. 43. Chittenden, T., Livingston, D. M. & Kaelin, W. G. (1991) Cell 65, 1073-1082. 44. Mudryj, M., Devoto, S. H., Hiebert, S. W., Hunter, T., Pines, J. & Nevins, J. R. (1991) Cell 65, 1243-1253.
