Nuclear Localization of DP and E2F Transcription Factors by Heterodimeric Partners and Retinoblastoma Protein Family Members

Home , E2F, Retinoblastoma protein

Journal of Cell Science 109, 1717-1726 (1996) 1717 Printed in Great Britain © The Company of Biologists Limited 1996 JCS7086

Nuclear localization of DP and E2F transcription factors by heterodimeric partners and retinoblastoma protein family members

Junji Magae1, Chin-Lee Wu2, Sharon Illenye1, Ed Harlow2 and Nicholas H. Heintz1,* 1Department of Pathology, University of Vermont, Burlington VT 05405, USA 2Massachusetts General Hospital Cancer Center, Charlestown MA 02129, USA *Author for correspondence

SUMMARY

E2F is a family of transcription factors implicated in the showed that regions of E2F-1 and DP-1 that are required regulation of genes required for progression through G1 for stable association of the two proteins were also required and entry into the S phase. The transcriptionally active for nuclear localization of DP-1. Unlike E2F-1, -2, and -3, forms of E2F are heterodimers composed of one polypep- E2F-4 did not accumulate in the nucleus unless it was coex- tide encoded by the E2F gene family and one polypeptide pressed with DP-2. p107 and p130, but not pRb, stimulated encoded by the DP gene family. The transcriptional activity nuclear localization of E2F-4, either alone or in combina- of E2F/DP heterodimers is influenced by association with tion with DP-2. These results indicate that DP proteins the members of the retinoblastoma tumor suppressor preferentially associate with specific E2F partners, and protein family (pRb, p107, and p130). Here the intracellu- suggest that the ability of specific E2F/DP heterodimers to lar distribution of E2F and DP proteins was investigated in localize in the nucleus contributes to the regulation of E2F transiently transfected Chinese hamster and human cells. activity. In transfected cells, DP-1 did not accumulate in the nucleus unless it was coexpressed with the heterodimeric partners E2F-1, E2F-2, or E2F-3. Domain mapping experiments Key words: Gene expression, Cell cycle, Protein trafficking

INTRODUCTION by cyclin-dependent kinases inhibits E2F DNA binding activity and E2F-dependent transcription (Krek et al., 1994; First identiﬁed as a cellular factor required for the transcrip- Dynlacht et al., 1994; Xu et al., 1994). Disruption of pRb/E2F tional activation of the adenovirus E2 gene promoter (reviewed complexes by viral oncoproteins is essential for cell transfor- by Moran, 1993; Nevins, 1992), E2F has been implicated in mation by DNA tumor viruses, and mutations in pRb that the periodic regulation of cellular genes required for transition disrupt interactions with E2F are implicated in both sporadic through G1 and entry into the S phase (reviewed by Farnham and inherited forms of human cancer (reviewed by Dyson, et al., 1993; Horowitz, 1993; La Thangue, 1994; Lam and 1994; Ewen, 1994; Lam and La Thangue, 1994; Wiman, 1993). LaThangue, 1994; Muller, 1995; Wiman, 1993). The evidence Moreover, deregulated expression of E2F genes leads to mor- that E2F plays a central role in regulating progression through phological changes (Logan et al., 1994; Shan and Lee, 1994; the cell cycle is compelling. E2F forms higher order complexes Xu et al., 1995), entry into the S phase (Almasan et al., 1995; with a number of proteins that regulate progression through the Johnson et al., 1993; Kowalik et al., 1995; Sardet et al., 1995; cell cycle, including the products of the retinoblastoma tumor Shan and Lee, 1994), and cell transformation (Singh et al., suppressor gene family (i.e. pRb, p107 and p130) (Chellappan 1994; Xu et al., 1995; Yang and Sladek, 1995), and, under et al., 1991; Chittenden et al., 1991; Cao et al., 1992; Cobrinik some conditions, apoptosis (Qin et al., 1994; Shan and Lee, et al., 1993; Dyson et al., 1993; Huang et al., 1993; Krek et al., 1994; Wu and Levine, 1994; Kowalik et al., 1995). 1993; Lees et al., 1993; Fagan et al., 1994; Kim et al., 1994; The regulation of E2F by pRb family members and cyclin- Qin et al., 1995) and several cyclin-dependent kinases (Mudryj dependent kinases is exceedingly complex. Although the asso- et al., 1991; Devoto et al., 1992; Lees et al., 1992; Pagano et ciation of regulatory factors with E2F is tightly regulated al., 1992; Ewen et al., 1993; Kato et al., 1993; Dynlacht et al., during the cell cycle (Lees et al., 1992; Pagano et al., 1992; 1994). Association of pRb family members with E2F sup- Shirodkar et al., 1992; Schwarz et al., 1993; Chittenden et al., presses E2F-dependent transcription (Hiebert et al., 1992; 1993; Cobrinik et al., 1993; Hijmans et al., 1995), the sequence Flemington et al., 1993; Helin et al., 1993a; Hiebert, 1993; of these interactions during the cell cycle does not appear to Zamanian and La Thangue, 1993; Qin et al., 1995; Smith and be the same for all cell types. In addition, E2F participates in Nevins, 1995) and cell growth (Hiebert, 1993; Zhu et al., 1993; the transcriptional regulation of genes encoding members of Vairo et al., 1995; Zhu et al., 1995a). Phosphorylation of E2F the E2F/pRb regulatory loop, including the pRb, p107 and 1718 J. Magae and others

E2F-1 genes (Hsiao et al., 1994; Johnson et al., 1994b; al., 1981), and the human cell lines U2OS and HeLa were cultured in Neuman et al., 1994; Zhu et al., 1995b). a 5% CO2 atmosphere at 37¡C using D-MEM supplemented with 5% The transcriptionally active forms of E2F are a collection of FBS (Gibco). For DNA transfections, 1×106 CHOC 400 cells were heterodimeric protein complexes (Girling et al., 1993; Helin et plated in 85 mm culture dishes in 10 ml DMEM with 5% fetal bovine serum (FBS), incubated overnight, changed into fresh culture al., 1993b; Huber et al., 1993; Wu et al., 1995; Zhang and Chel- µ lappan, 1995), each composed of one E2F protein family subunit medium, and 6 hours later were transfected with 24 g of DNA using the calcium phosphate precipitation method as described previously and one DP protein family subunit. (Here we use E2F to refer (Helin et al., 1993b; Wu et al., 1995). The E2F and DP pCMV to the collection of heterodimeric complexes formed by the asso- expression vectors (Helin et al., 1993b; Wu et al., 1995) were used at ciation of specific E2F and DP family members, with the indi- 8 µg/culture dish with pBSK as carrier DNA. After incubation vidual components of each complex identified by the name of overnight, cells were washed twice with phosphate buffered saline the cloned cDNA.) To date cDNA clones for seven members of (PBS, pH 7.5), incubated in fresh medium for 4 hours, trypsinized, the mammalian E2F and DP gene families have been isolated. plated on coverslips, and incubated for an additional 20 hours in Interaction with pRb was used to clone E2F-1 (Helin et al., 1992; culture medium. Kaelin et al., 1992; Shan et al., 1992; Li et al., 1994), and Immunostaining homology with E2F-1 was then used to isolate E2F-2, E2F-3, and E2F-4 (Ivey-Hoyle et al., 1993; Lees et al., 1993; Ginsberg Cells on coverslips were washed with PBS, fixed with 4% paraformaldehyde in PBS for 15 minutes, permeabilized with 0.2% et al., 1994). E2F-4 was also cloned by virtue of its specific inter- Triton-X in PBS for 15 minutes, and blocked in PBS containing 0.1% action with p107 (Beijersbergen et al., 1994), and both E2F-4 azide, 0.1% Tween-20 and 2% FBS for 60 minutes. Coverslips were and E2F-5 were isolated in a screen for proteins that interact with then incubated with primary murine antibodies at a 1:5 dilution of p130 (Hijmans et al., 1995; Sardet et al., 1995). DP-1 was cloned culture supernatant, or 1:200 dilution of polyclonal mouse serum, in as a protein component of DRTF, a developmentally-regulated blocking buffer for 60 minutes. The coverslips were then washed three E2F-like activity from mouse embryonal F9 cells (Girling et al., times in Tris-buffered saline (pH 7.5) containing 0.1% Tween-20 1993). Recently, homology with DP-1 was used to isolate DP- (TBS-T), and incubated with FITC-conjugated anti-mouse 2, a second member of the mammalian DP protein family (Wu immunoglobulin diluted 1:100 in blocking buffer for 60 minutes. After washing three times in TBS-T, coverslips were incubated in 1 et al., 1995; Zhang and Chellappan, 1995). Comparison of the µ amino acid sequences of mammalian E2F family members g/ml propidium iodide (PI) in TBS-T for 60 minutes at 50¡C, washed with TBS-T, mounted and photographed with an Olympus BX50 suggests E2F-4 and E2F-5 are highly related to one another, and microscope using a WIBA filter for FITC, WG filter for PI, and MT represent a subclass of factors distinct from E2F-1, -2 and -3 filter for PI/FITC. (Sardet et al., 1995). Both DP-1 and DP-2 dimerize with several E2F family members to form an array of transcriptionally active Antibodies forms of E2F/DP heterodimers (Bandara et al., 1993; Girling et The primary antibodies were monoclonal antibody KH20 for E2F-1 al., 1993; Helin et al., 1993b; Huber et al., 1993; Wu et al., 1995; (Helin et al., 1993b), monoclonal antibody WTH10 for DP-1 (Wu et Zhang and Chellappan, 1995). al., 1995), polyclonal antibodies for E2F-2, E2F-3, or E2F-4, and The function of individual E2F/DP heterodimers is not either a polyclonal antibody against DP-2 or monoclonal antibody known. While all E2F/DP complexes stimulate transcription of 12CA5 for HA-tagged DP-2 and HA-tagged E2F-3. Polyclonal anti- reporter genes from consensus E2F binding sites, different E2F bodies for E2F-2, E2F-3, E2F-4 and DP-2 were made against His- tagged full length proteins in mice and characterized as described (Wu complexes may regulate E2F-dependent cellular promoters et al., 1995). FITC-conjugated anti-mouse immunoglobulin was through different E2F binding sites (Zhu et al., 1995b). Func- purchased from Amersham (Arlington Heights, IL). tional distinctions between heterodimeric E2F/DP transcription factor complexes include tissue-specific patterns of expression (Lees et al., 1993; Wu et al., 1995), differences in RESULTS the time of expression during the cell cycle (Sardet et al., 1995), and preferential association with different members of Localization of E2F and DP proteins the retinoblastoma tumor suppressor protein family or other The intracellular localization of E2F and DP proteins was cell cycle regulators (Cao et al., 1992; Chellappan et al., 1991; evaluated in transiently transfected cells by immunofluor- Chittenden et al., 1991; Cobrinik et al., 1993; Cress et al., 1993; escence microscopy. At 48 hours after transfection with pCMV Dyson et al., 1993; Fagan et al., 1994; Kim et al., 1994; Krek expression plasmids, cells were stained with antibodies against et al., 1993; Lees et al., 1993; Beijersbergen et al., 1994; Sardet E2F or DP proteins (or a hemagglutinin (HA) epitope tag), and et al., 1995; Vairo et al., 1995). To address the role of intra- the intracellular distribution of the protein was assessed by cellular localization in E2F function, we examined the intra- immunofluorescence microscopy. Both western blotting and cellular location of DP and E2F family members in transiently- immunostaining confirmed that each of the E2F and DP anti- transfected Chinese hamster ovary (CHO) or human cells. Our bodies reacted only with its cognate antigen (data not shown). results indicate that E2F and DP proteins have preferred het- When DP and E2F family members were expressed alone, erodimeric partners, and that nuclear localization may provide several distinct patterns of cell staining were observed (Fig. 1). an additional level of control over E2F activity. E2F-1, E2F-2 and E2F-3 were found predominantly in the cell nucleus in 50-70% of the transfected cell population (Table 1A). The remainder of the cells transfected with E2F-1, -2 or MATERIALS AND METHODS -3 displayed faint staining throughout the cytoplasm in addition to prominent nuclear staining (Fig. 1; data not shown). Cell culture and DNA transfection E2F-1, -2, and -3 were not found in the cytoplasm in the The methotrexate-resistant CHO cell strain, CHOC 400 (Milbrandt et absence of nuclear staining. In contrast, E2F-4 was evenly dis- Nuclear localization of E2F 1719

Fig. 1. Immunolocalization of E2F and DP family members in transiently transfected Chinese hamster cells. CHOC 400 cells were transfected with pCMV plasmid constructs that express the indicated E2F or DP proteins, and 48 hours after the addition of DNA cells were fixed and stained by indirect immunofluorescence for the corresponding E2F or DP antigen or an HA epitope tag. Representative fields of cells that were photographed to show cell morphology (phase contrast), the red signal for nuclear DNA content (PI), the green FITC signal for E2F or DP proteins (FITC), or both the DNA and FITC signals (PI/FITC). Colocalization of the red PI and green FITC signals produces a yellow nuclear signal. Bar, 5 µm. tributed throughout the cell cytoplasm (Fig. 1), and prominent erodimers, we then studied the intracellular localization of E2F nuclear staining of E2F-4 was evident in less than 5% of the and DP proteins in cotransfection experiments (Fig. 2). In 3-6 transfected cells (Table 1A). Immunostaining also showed that independent experiments, the fraction of transfected cells dis- the intracellular distribution of DP-1 and DP-2 differed in tran- playing prominent nuclear staining of each DP or E2F antigen siently transfected cells. At 12 hours after transfection, DP-1 was determined 48 hours after transfection (Table 1A). Coex- staining was faint and diffuse throughout the cytoplasm (data pression of E2F-1, E2F-2, or E2F-3 with DP-1 dramatically not shown). At later times, DP-1 was observed in bright cyto- increased the fraction of cells displaying nuclear staining of plasmic speckles. By 48 hours, DP-1 was observed in DP-1 from 15% to 90% (Table 1A). Coincident with the prominent perinuclear inclusion bodies in the cytoplasm in induction of nuclear staining of DP-1, coexpression of E2F-1, 95% of the transfected cells (Fig. 1), and only 5-15% of the E2F-2 and E2F-3 also induced a concomitant loss of DP-1 transfected cells displayed significant nuclear staining of DP- staining in cytoplasmic inclusion bodies (Fig. 2). Coexpression 1 (Tables 1A and 2). The perinuclear inclusion bodies con- with E2F-4 reduced the accumulation of DP-1 in perinuclear taining DP-1 were evident by phase contrast microscopy (Fig. inclusion bodies (Fig. 2), but did not result in the nuclear local- 1). Although DP-2 also accumulated in the cytoplasm in a ization of DP-1 (Table 1). Rather, in the presence of E2F-4, fraction of transfected cells (for example, see Fig. 1), by 48 DP-1 was found distributed throughout the cytoplasm in a hours after transfection at least 40% of the cells contained DP- staining pattern similar to E2F-4 (Figs 2 and 3). In contrast to 2 predominantly in the nucleus (Fig. 1 and Table 1A). Thus, its effects on DP-1, E2F-4 significantly increased the fraction when expressed at high levels in transiently transfected CHO of transfected cells that showed nuclear staining of DP-2 from cells, DP-1 and E2F-4 were localized mainly in the cytoplasm, 40% to 80% (Table 1). Cotransfection with DP-1 tended to whereas E2F-1, E2F-2, E2F-3, and DP-2 were found predom- increase the nuclear staining of E2F-1, E2F-2 and E2F-3, inantly in the nucleus. Similar results were obtained with trans- although this effect did not appear to be statistically significant fected human HeLa or U2OS cells (data not shown). (Table 1A). Immunofluorescence microscopy revealed a significance E2F-1, E2F-2 and E2F-3 stimulate nuclear difference in the ability of DP-1 and DP-2 to influence the localization of DP-1 intracellular distribution of E2F-4 (Fig. 3). When coexpressed Given that the transcriptionally-active forms of E2F are het- with DP-1, both DP-1 and E2F-4 were observed in the nucleus 1720 J. Magae and others

Table 1. Effect of E2F proteins on the nuclear localization of DP-1 and DP-2

A % Cells with nuclear signal

E2F DP-1+E2F HA-DP-2+E2F Plasmids Structure vector alone DP-1 E2F DP-2 E2F Vector 15.2±6.4 40.9±9.8 437 409 426 284 68 108 120 191 1 pCMV-E2F-1 53.0±9.7 90.3±6.5 76.2±16.4 87.1±7.3 51.3±9.9 406 423 433 285 71 111 118 1 188 pCMV-E2F-2 66.3±7.4 92.2±4.5 74.9±9.9 68.9±9.7 83.3±12.3 432 449 465 241 337 1 102 153 171 pCMV-HA-E2F-3 60.4±19.3 85.7±14.7 77.9±1.4 87.0±9.9 84.6±7.9 308 328 382 413 1 10 82 187 pCMV-E2F-4 2.5±3.1 4.3±6.1 0.6±1.0 81.1±8.7 88.2±0.4

B 409 426 437 284 68 191 108 120 1 pCMV-E2F-1 53.0±9.7 90.3±6.5 76.2±16.4 87.1±7.3 51.3±9.9 437

1 Y C pCMV-Y411C 411 39.1±5.1 78.2±23.6 63.7±16.8 68.9±12.1 69.8±14.2 1 pCMV-1-191 191 20.2±13.1 4.6±2.9 51.2±23.5 19.4±11.9 34.8±21.6

pCMV-223-437 223 437 73.0±10.1 86.5±11.6 416 pCMV-∆417-437 1 23.4±12.9 84.0±3.2 74.1±12.5 80.0±10.7 73.6±13.6 1 pCMV-1-284 284 17.7±13.8 74.4±13.8 71.5±10.1 43.2±15.7 45.1±14.8 1 pCMV-1-374 374 25.6±16.7 90.9±4.1 89.5±7.8 27.6±11.5 62.7±7.4

: cyclinA binding; : DNA binding; : Dimerization; : Serine repeat; : Rb/p107 binding

CHOC 400 cells were transfected with pCMV expression vectors for DP-1 and HA-tagged DP-2 alone, or in combination with the indicated E2F expression plasmid. After 48 hours, cells were stained for DP-1 with monoclonal antibody WTH10 and HA-DP-2 with monoclonal antibody 12CA5. Replicate coverslips were stained for the coexpressed E2F moiety as in Fig. 1. At least 100 cells displaying an FITC signal were scored for nuclear fluorescence in each experimental group. The percentage of cells displaying nuclear fluorescence is presented as an average of 3-6 independent experiments ± the standard deviation of the mean. (A) Nuclear localization of DP-1 and DP-2 in the presence of E2F-1, -2, -3 and -4. (B) Specific domains of E2F-1 are required to stimulate nuclear localization of DP-1 and DP-2. The structural domains of E2F-1, E2F-2, E2F-3, and E2F-4 are derived from Beijersbergen et al., 1994; Helin et al., 1993b; Ginsberg et al., 1994; Sardet et al., 1995. in about 5% of the transfected cells (Table 1). In contrast, coex- pCMV-1-191 reduced rather than increased the fraction of cells pression with DP-2 increased the fraction of cells showing that displayed nuclear fluorescence of either DP-1 or DP-2 nuclear localization of E2F-4 from 4% to about 90% of the (Table 1B). Although truncated forms of E2F-1 containing transfected cell population (Table 1). Thus, E2F-4 differed amino acid residues 1-284 (pCMV 1-284) or 1-374 (pCMV 1- from E2F-1, E2F-2 and E2F-3 in that it did not induce nuclear 374) stimulated nuclear localization of DP-1 (Table 1B), these localization of DP-1. Rather, only DP-2 was able to induce forms of E2F-1 did not eliminate the accumulation of DP-1 in nuclear localization of E2F-4. cytoplasmic inclusion bodies (Fig. 2). In contrast, the carboxy terminal residues of E2F-1 encoded by pCMV 223-437 Dimerization between E2F-1 and DP-1 is required for increased the nuclear localization of both DP-1 and DP-2 nuclear localization of DP-1 (Table 1), and greatly reduced the accumulation of DP-1 in the To determine if specific regions of E2F-1 stimulate nuclear cytoplasm (Fig. 2). Based on previous studies (Helin et al., localization of DP-1, plasmids that encode truncated versions 1993b), these results indicate C-terminal E2F-1 sequences that of E2F-1 were cotransfected with the pCMV-DP-1 expression include (but are not limited to) the region required for dimer- vector (Table 1B, and data not shown). Cotransfection with ization with DP-1 (encoded by residues 191-284) stimulate Nuclear localization of E2F 1721

Table 2. Regions of DP-1 required for nuclear localization by E2F-1

% Cells with nuclear signal

+ E2F-1 E2F-1 Binding Plasmids Structure HA HA E2F-1 DNA E2F-1 204 277 410 84 1 pCMV-HA-DP-1 HA 5.6±0.5 51.7±17.3 52.6±20.2 + +

1-204 HA 0.7±1.2 5.9±5.1 33.4±4.7 + 1-277 HA 0.0±0.0 5.1±2.3 24.7±17.1 + ± 1-346 HA 1.3±2.3 59.8±18.4 51.6±3.4 + + 102-410 HA 1.9±2.0 63.1±6.9 46.3±13.5 + +

127-410 HA 0.0±0.0 55.9±16.7 60.2±4.7 + 182-410 HA 2.3±2.0 13.3±8.3 29.4±13.2 + 232-410 HA 0.0±0.0 4.3±2.6 36.1±19.3

103-126 HA 0.6±0.0 47.5±13.4 67.7±10.7 +

233-272 HA 0.6±1.1 12.8±13.6 31.8±9.7 +

: DNA binding; : Dimerization

pCMV constructs that express the indicated HA-tagged DP-1 proteins were transfected alone or with pCMV-E2F-1 into CHOC 400 cells, and 48 hours later cells were stained for the HA tag with monoclonal antibody 12CA5 or for E2F-1 with monoclonal antibody KH20. The percentage of cells displaying nuclear fluorescence is presented as an average of 3 independent experiments ± the standard deviation of the mean. Data for DNA binding and association of HA-DP-1 with E2F-1 are from Wu et al. (1995) and C.-L. Wu, M. Classon, N. Dyson and E. Harlow (unpublished). nuclear localization of DP-1 in transiently transfected cells. the DP-1 DNA binding domain (HA-DP-1 D108-126) did not Since cotransfection with a pRB expression vector containing prevent nuclear localization by E2F-1. In contrast, targeted a mutation (pCMV-E2F-1 Y411C) or deletion (pCMV D417- deletion of amino acids 233-272 (HA-DP-1 D233-272) within 437) within the pRB binding domain of E2F-1 did not signif- the dimerization domain abolished nuclear localization of DP- icantly diminish the ability of E2F-1 to induce nuclear local- 1 when coexpressed with E2F-1 (Table 2). In experiments to ization of DP-1 (Fig. 2 and Table 1B), the C-terminal be published elsewhere, western blotting of E2F-1 immuno- stimulatory activity does not appear to reside in the pRb precipitates with antibody 12CA5 was used to assess the asso- binding pocket. ciation of E2F-1 with truncated or mutant forms of HA-DP-1 We then sought to identify the regions of DP-1 required for in vivo (C.-L. Wu, M. Classon, N. Dyson, and E. Harlow, nuclear localization by E2F-1. Truncated forms of HA-tagged unpublished). Here we have observed a direct correlation DP-1 were coexpressed with full length E2F-1, and the per- between the ability of E2F-1 to associate with DP-1 in vivo centage of cells displaying nuclear localization of the HA with the ability of E2F-1 to induce nuclear localization of HA- epitope were scored (Table 2). None of the truncated forms of DP-1. Together with the data from the experiments with HA-DP-1 accumulated in the nucleus when expressed alone truncated forms of E2F-1, these results strongly suggest that (Table 2). When coexpressed with E2F-1, HA-DP-1 1-204, a stable protein-protein interactions between E2F-1 and DP-1 are form of DP-1 that contains the DNA binding domain but lacks required for E2F-1 to stimulate the nuclear localization of DP- sequences previously shown to be required for dimerization 1. with E2F-1 (Wu et al., 1995), did not localize in the nucleus. Surprisingly, HA-DP 1-277, which contains the amino acids Effect of pRb family members on nuclear 204 to 277 that have been shown previously to mediate dimer- localization E2F and DP proteins ization with E2F-1 (Helin et al., 1993b), also did not localize E2F/DP heterodimers form higher order complexes with in the nucleus when coexpressed with E2F-1 (Table 2). A sig- proteins encoded by members of the retinoblastoma (pRb) nificant increase in nuclear localization of HA-DP-1 was tumor suppressor gene family. The ability to bind pRb, or the observed when sequences from 277 to 346 were included in related factors p107 and p130, resides in the E2F subunit the expression construct (HA-DP-1 1-346, Table 2). This (Hiebert, 1993; Huang et al., 1993; Kim et al., 1994). E2F-1, portion of DP-1 contains a region downstream of the dimer- E2F-2 and E2F-3 preferentially bind pRb (Helin et al., 1992; ization domain that is conserved between DP-1 and DP-2 (Wu Kaelin et al., 1992; Helin et al., 1993a; Krek et al., 1993; Lees et al., 1995). Deletion of the first 127 amino acids of HA-DP- et al., 1993), while E2F-4 preferentially binds either p107 or 1 (HA-DP-1 127-410) did not diminish the ability of E2F-1 to p130 (Cobrinik et al., 1993; Dyson et al., 1993; Beijersbergen induce nuclear localization of HA-DP-1, whereas deletion of et al., 1994; Sardet et al., 1995). To determine if association the first 232 amino acids did (Table 2). Targeted deletion of with pRb family members influences the intracellular dis- 1722 J. Magae and others

Fig. 2. Effect of E2F proteins on the intracellular localization DP-1 and DP-2. Cells were cotransfected with expression vectors for either DP-1 or DP-2 and the indicated E2F protein, and 48 hours after transfection, cells were ﬁxed, stained for DNA with propidium iodide, and stained for either DP-1 or HA-DP-2 with FITC as in Fig. 1. Presented are the effects of the indicated E2F protein on the FITC staining patterns of cells expressing either DP-1 (left column) or HA- DP-2 (right column). Bar, 5 µm.

tribution of E2F and DP proteins, the effect of coexpression of pRb, p107, and p130 on nuclear localization of E2F-4 was assessed, both in the presence and absence of DP-2. As shown in Fig. 4 and Table 3, p107 increased the fraction of transfected cells displaying nuclear staining of E2F-4 from about 2% to over 40%. p107 also stimulated nuclear localization of E2F-4 in the presence of DP-2, from about 80% to 94% of the transfected cell population. While p130 also stimulated nuclear localization of E2F-4 in the presence and absence of DP-2, it was not as active as p107 (Fig. 4 and Table 3). Coexpression of pRb with E2F-4 did not stimulate nuclear localization of E2F-4, either with or without DP-2 (Table 3). Although p107 increased the fraction of cells displaying nuclear staining of DP-1, the other pRb family members were not active in stim- ulating nuclear localization of DP-1, either with or without E2F-1 (Table 4).

Table 3. Nuclear localization of E2F-4 is stimulated by p107 and p130 % Cells with Plasmids nuclear signal pCMV-E2F-4 1.8±0.2 + pCMV-p107 41.7±7.6 + pCMV-p130 18.4±3.4 + pCMV-pRb 4.6±5.0 pCMV-E2F-4 + pCMV-DP-2 79.0±4.8 + pCMV-p107 93.5±1.4 + pCMV-p130 87.3±1.7 + pCMV-pRb 71.2±10.1

CHOC 400 cells were transfected with pCMV-E2F-4, or pCMV-E2F-4 and pCMV-DP-2, along with either pCVM alone (vector) or pCMV-p107, pCMV-p130, or pCMV-pRb. At 48 hours after the addition of DNA, cells were ﬁxed and stained for E2F-4 with FITC as in Fig. 4. The percentage of transfected cells displaying nuclear staining of E2F-4 in these experiments is presented ± the standard deviation of the mean.

Table 4. Nuclear localization of DP-1 is not stimulated by pRb family members % Cells with Plasmids nuclear signal pCMV-DP-1 3.1 + pCMV-p107 16.9 + pCMV-p130 5.0 + pCMV-pRb 4.9 pCMV-DP-1 + pCMV-E2F-1 82.3 + pCMV-p107 87.7 + pCMV-p130 76.2 + pCMV-pRb 65.5

CHOC 400 cells were transfected with pCMV-DP-1, or pCMV-DP-1 and pCMV-E2F-1, with pCMV-p107, pCMV-p130, and pCMV-pRb. The percentage of cells displaying nuclear staining of DP-1 in a single experiment is presented. Nuclear localization of E2F 1723

Fig. 3. DP-2 stimulates nuclear localization of E2F- 4. CHOC 400 cells were transfected with pCMV- E2F-1 or pCMV-E2F-4 alone (left column), or with the E2F expression vector and either pCMV-DP-1 (middle column) or pCMV-HA-DP-2 (right column). At 48 hours after the addition of DNA, cells were ﬁxed and stained for E2F-1 or E2F-4 as in Fig. 1. Shown are the FITC signals for E2F-1 (top row) and E2F-4 (bottom row). Bar, 5 µm.

DISCUSSION ferential regulation of these related transcription factor complexes. DP-1 is expressed as a single mRNA in most Considerable evidence suggests E2F is a central regulator of tissues and cell lines, while DP-2 is expressed at variable levels the cell cycle (reviewed by Nevins, 1992; Farnham et al., 1993; as multiple mRNAs in a tissue- and cell-specific manner Ewen, 1994; Lam and La Thangue, 1994; Muller, 1995). (Girling et al., 1993; Wu et al., 1995). Variation is also Different forms of E2F are known to form physiologically observed in the expression of E2F family members in different important complexes with critical regulators of cell growth, cells and tissues (Lees et al., 1993), and during the cell cycle and deregulated expression of certain E2F family members (Li et al., 1994; Sardet et al., 1995). For example, both E2F-4 induces premature entry into the S phase, changes in cell mor- and E2F-5 mRNA are expressed earlier than E2F-1 during the phology, and alterations in cell growth control. Since a number G1 to S phase transition (Sardet et al., 1995). Thus, regulation of different E2F/DP heterodimers are able to activate reporter of E2F and DP gene transcription represents the first level of gene expression from identical E2F DNA binding sites control over the formation of different E2F heterodimers. (Girling et al., 1993; Helin et al., 1993b; Huber et al., 1993; Clearly an additional level of control is the preferential asso- Wu et al., 1995; Zhang and Chellappan, 1995), mechanisms ciation of different E2F/DP protein complexes with pRb, p107 other than DNA binding are certain to be important for the dif- and p130 or other cell cycle regulators. Here we have examined the intracellular localization of E2F and DP proteins in cells transiently transfected with pCMV expression vectors. Several factors influence interpretation of these studies. First, the levels of protein produced by pCMV expression vectors exceed those produced from endogenous cellular promoters. Thus, it is not possible to definitively conclude that E2F and DP proteins made during transfection behave identically to the homologous cellular factors. Nonetheless, transient transfection has been used extensively to study complex formation and transcriptional activation by E2F and DP proteins. Thus, the conditions of our study are analagous to those that others have used in transient transfection for examining E2F-dependent gene expression. Second, we have not rigorously assessed the effect of the cell cycle on nuclear localization after transient transfection. In other studies, we have generated stable cell lines that express DP-1 and E2F-1 under the control of tetracycline-responsive promoters (J. Magae, S. Illenye, J. Wells, and N. H. Heintz, unpublished). In these cells, DP-1 is predominantly cytoplasmic and E2F-1 is nuclear during G1 (data not shown). Thus, use of transient transfection to express DP-1 does not account for the cytoplasmic location of this protein. Since E2F activity is regulated during the cell cycle, it is possible that heterogen- ity in immunostaining patterns may be related to the position Fig. 4. Nuclear localization of E2F-4 by p107. CHOC 400 cells were transfected with pCMV-E2F-4 (left column) or pCMV-E2F-4 and of the cell in the cell cycle. pCMV-DP-2 (right column) along with either pCVM alone (vector) Given these considerations, our results provide several con- or pCMV-p107, pCMV-p130, or pCMV-pRb. At 48 hours after the clusions concerning E2F and DP proteins. As previously addition of DNA, cells were fixed and stained for E2F-4 with FITC suggested (Sardet et al., 1995), E2F-1, E2F-2, and E2F-3 as in Fig. 3. Bar, 5 µm. appear to be functionally related, and represent a class of 1724 J. Magae and others factors distinct from E2F-4 (and perhaps E2F-5). Here E2F-1, or DP proteins may be subject to the availability of the appro- E2F-2, and E2F-3 all stimulated the nuclear localization of DP- priate heterodimeric partner, a process that would provide an 1, whereas E2F-4 was less effective (Fig. 1 and Table 1). While additional level of regulation during the cell cycle, or in E2F-4 forms complexes with DP-1 in cotransfection experi- response to extracellular regulators of cell growth. In contrast ments (Beijersbergen et al., 1994; Ginsberg et al., 1994; Sardet to DP-1, p107 and p130 clearly stimulated nuclear localization et al., 1995), cotransfection of E2F-4 with DP-1 only stimu- of E2F-4, either alone or in combination with DP-2 (Fig. 4 and lates reporter gene expression from consensus E2F sites about Table 3). Although the reason for this effect is not known, it 10-fold (Beijersbergen et al., 1994; Ginsberg et al., 1994; data is possible that p107 and p130 stabilize interactions of E2F-4 not shown). In contrast, cotransfection of E2F-1, E2F-2, or with endogenous DP proteins, or faciliate the retention of E2F- E2F-3 with DP-1 stimulates reporter gene expression 200- to 4 within the nucleus. 400-fold (Helin et al., 1993b; Wu et al., 1995; data not shown). The studies presented here suggest that transport to the Together these data suggest that E2F and DP proteins may nucleus of different E2F/DP heterodimers represents an addi- have preferred heterodimeric partners. Our results suggest DP- tional level of control over E2F activity. While the mechanism 1 may preferentially associate with E2F-1, E2F-2 and E2F-3, by which E2F and DP proteins stimulate the accumulation of while DP-2 may preferentially associate with E2F-4. We have their heterodimeric partners in the nucleus remains to be deter- recently shown that certain species of endogenous E2F mined, our results raise the possibility that nuclear entry of complexes can be distinguished on the basis of their affinity E2F/DP complexes is an active process that may be regulated for E2F sites containing either TTTCGCGC or TTTGGCGC during the cell cycle. Since very low concentrations of the sequences (Wells et al., 1996). Functional differences in E2F peptide aldehyde MG-132 blocks nuclear entry of E2F-1, and sites containing either GG and CG sequences also have been DP-1 in the presence of E2F-1 (J. Magae, unpublished data), noted at the human p107 promoter (Zhu et al., 1995b). Thus, we consider it unlikely that these proteins equilibrate passively it remains important to determine which E2F/DP complexes between the cytoplasm and nucleus. Once cellular genes that form in living cells, and how these complexes distinguish are regulated by specific E2F/DP heterodimers are identified, between different E2F sequences. understanding the functional relationship between preferential Nuclear localization studies suggest that sequences of DP-1 association of E2F and DP subunits, nuclear localization of in addition to the previously mapped dimerization domain E2F/DP heterodimeric complexes, and transcriptional activa- from amino acids 204-277 are important for association tion can be addressed in further detail. between E2F-1 and DP-1 (Table 2). Comparison of DP-1 and DP-2 show that these proteins are 85% identical in a down- We thank R. Bernards for pCMV-E2F4, Merck Research Labora- stream region located between amino acids 277-335 of DP-1 tories for pCMV-E2F-2, Kevin Murakami for assistance with the and amino acids 231-289 of DP-2 (Wu et al., 1995; Zhang and tables and figures, and members of the Harlow laboratory for critical Chellappan, 1995). The conserved downstream region from discussions and technical support. C.-L.W. was supported by an NRSA from the NIH. N.H.H. is the recipient of an American Cancer 277-335 of DP-1 coincides with the DP-1 sequences from Society Research Faculty Award and the J. Walter Juckett Scholar- amino acids 277-346 that mediate a 10-fold increase in nuclear ship from the Lake Champlain Cancer Research Organization and the localization by E2F-1 (Table 2). Since the association of HA- Vermont Cancer Center. E. H. is an American Cancer Society DP-1 1-277 with E2F-1 in vivo is weak (C.-L. Wu, M. Classon, Research Professor. This work was supported by grants to N.H.H. and N. Dyson, and E. Harlow, unpublished), the downstream E.H. from the NIH. region of DP-1 from 277-346 likely contributes to the stability of DP-1/E2F-1 heterodimers, and thereby stimulates nuclear localization of DP-1. Clearly the previously mapped dimeriza- REFERENCES tion domain of DP-1 is critical for nuclear localization, for targeted deletion of amino acids 233-272 of DP-1 abolishes Almasan, A., Yin, Y., Kelly, R. E., Lee, E. Y., Bradley, A., Li, W., Bertino, nuclear localization by E2F-1 (Table 2). The notion that direct J. R. and Wahl, G. M. (1995). Deficiency of retinoblastoma protein leads to protein-protein interactions between DP-1 and E2F-1 are inappropriate S-phase entry, activation of E2F-responsive genes, and required for nuclear localization of DP-1 is also supported by apoptosis. Proc. Nat. Acad. Sci. USA 92, 5436-5440. Bandara, L. R., Buck, V. M., Zamanian, M., Johnston, L. H. and La the experiments with truncated forms of E2F-1 (Table 1). Thus, Thangue, N. B. (1993). Functional synergy between DP-1 and E2F-1 in the we conclude from our studies that stable interactions between cell cycle-regulating transcription factor DRTF1/E2F. EMBO J. 12, 4317- E2F-1 and DP-1 are required for E2F-1 to stimulate nuclear 4324. localization of DP-1. Beijersbergen, R. L., Kerkhoven, R. M., Zhu, L., Carlee, L., Voorhoeve, P. M. and Bernards, R. (1994). E2F-4, a new member of the E2F gene family, Although E2F-1, E2F-2 and E2F-3 preferentially interact has oncogenic activity and associates with p107 in vivo. Genes Dev. 8, 2680- with pRb, pRb did not appear to influence nuclear localization 2690. of DP-1, either alone or in association with E2F-1 (Table 4). Cao, L., Faha, B., Dembski, M., Tsai, L. H., Harlow, E. and Dyson, N. In the absence of pRb, accumulation of DP-1 in the nucleus by (1992). Independent binding of the retinoblastoma protein and p107 to the E2F-1, E2F-2, or E2F-3 did not appear to be soley the result transcription factor E2F. Nature 355, 176-179. Chellappan, S. P., Hiebert, S., Mudryj, M., Horowitz, J. M. and Nevins, J. of retention by nuclear targets, for neither DNA binding R. (1991). The E2F transcription factor is a cellular target for the RB protein. activity nor association with pRb was necessary for E2F-1 to Cell 65, 1053-1061. stimulate localization of DP-1 in the nucleus. Perhaps E2F- Chittenden, T., Livingston, D. M. and Kaelin, W. G. Jr (1991). RB 1/DP-1 resembles other heterodimeric helix-loop-helix tran- associates with an E2F-like, sequence-specific DNA-binding protein. Cold Spring Harbor Symp. Quant. Biol. 56, 187-195. scription factors in which the signal for nuclear entry for one Chittenden, T., Livingston, D. M. and DeCaprio, J. A. (1993). Cell cycle protein is provided by its heterodimeric partner (Goldfarb and analysis of E2F in primary human T cells reveals novel E2F complexes and Lewandowska, 1994). If so, nuclear entry of endogenous E2F biochemically distinct forms of free E2F. Mol. Cell. Biol. 13, 3975-3983. Nuclear localization of E2F 1725

Cobrinik, D., Whyte, P., Peeper, D. S., Jacks, T. and Weinberg, R. A. Transcription factor E2F binds DNA as a heterodimer. Proc. Nat. Acad. Sci. (1993). Cell cycle-specific association of E2F with the p130 E1A-binding USA 90, 3525-3529. protein. Genes Dev. 7, 2392-2404. Ivey-Hoyle, M., Conroy, R., Huber, H. E., Goodhart, P. J., Oliff, A. and Cress, W. D., Johnson, D. G. and Nevins, J. R. (1993). A genetic analysis of Heimbrook, D. C. (1993). Cloning and characterization of E2F-2, a novel the E2F1 gene distinguishes regulation by Rb, p107, and adenovirus E4. Mol. protein with the biochemical properties of transcription factor E2F. Mol. Cell. Biol. 13, 6314-6325. Cell. Biol. 13, 7802-7812. Devoto, S. H., Mudryj, M., Pines, J., Hunter, T. and Nevins, J. R. (1992). A Johnson, D. G., Schwarz, J. K., Cress, W. D. and Nevins, J. R. (1993). cyclin A-protein kinase complex possesses sequence-specific DNA binding Expression of transcription factor E2F1 induces quiescent cells to enter S activity: p33cdk2 is a component of the E2F-cyclin A complex. Cell 68, 167- phase. Nature 365, 349-352. 176. Johnson, D. G., Cress, W. D., Jakoi, L. and Nevins, J. R. (1994a). Oncogenic Dynlacht, B. D., Flores, O., Lees, J. A. and Harlow, E. (1994). Differential capacity of the E2F1 gene. Proc. Nat. Acad. Sci. USA 91, 12823-12827. regulation of E2F transactivation by cyclin/cdk2 complexes. Genes Dev. 8, Johnson, D. G., Ohtani, K. and Nevins, J. R. (1994b). Autoregulatory control 1772-1786. of E2F1 expression in response to positive and negative regulators of cell Dyson, N., Dembski, M., Fattaey, A., Ngwu, C., Ewen, M. and Helin, K. cycle progression. Genes Dev. 8, 1514-1525. (1993). Analysis of p107-associated proteins: p107 associates with a form of Kaelin, W. G. Jr, Krek, W., Sellers, W. R., DeCaprio, J. A., Ajchenbaum, E2F that differs from pRB-associated E2F-1. J. Virol. 67, 7641-7647. F., Fuchs, C. S., Chittenden, T., Li, Y., Farnham, P. J., Blanar, M. A. and Dyson, N. (1994). pRB, p107 and the regulation of the E2F transcription factor. et al. (1992). Expression cloning of a cDNA encoding a retinoblastoma- J. Cell Sci. 18, 81-87. binding protein with E2F-like properties. Cell 70, 351-364. Ewen, M. E., Sluss, H. K., Sherr, C. J., Matsushime, H., Kato, J. and Kato, J., Matsushime, H., Hiebert, S. W., Ewen, M. E. and Sherr, C. J. Livingston, D. M. (1993). Functional interactions of the retinoblastoma (1993). Direct binding of cyclin D to the retinoblastoma gene product (pRb) protein with mammalian D-type cyclins. Cell 73, 487-497. and pRb phosphorylation by the cyclin D-dependent kinase CDK4. Genes Ewen, M. E. (1994). The cell cycle and the retinoblastoma protein family. Dev. 7, 331-342. Cancer Metast. Rev. 13, 45-66. Kim, Y. W., Otterson, G. A., Kratzke, R. A., Coxon, A. B. and Kaye, F.J. Fagan, R., Flint, K. J. and Jones, N. (1994). Phosphorylation of E2F-1 (1994). Differential specificity for binding of retinoblastoma binding protein modulates its interaction with the retinoblastoma gene product and the 2 to RB, p107, and TATA-binding protein. Mol. Cell. Biol. 14, 7256-7264. adenoviral E4 19 kDa protein. Cell 78, 799-811. Kowalik, T. F., DeGregori, J., Schwarz J. K. and Nevins, J. R. (1995). E2F1 Farnham, P. J., Slansky, J. E. and Kollmar, R. (1993). The role of E2F in the overexpression in quiescent fibroblasts leads to induction of cellular DNA mammalian cell cycle. Biochim. Biophys. Acta 1155, 125-131. synthesis and apoptosis. J. Virol. 69, 2491-2500. Flemington, E. K., Speck, S. H. and Kaelin, W. G. Jr. (1993). E2F-1- Krek, W., Livingston, D. M. and Shirodkar, S. (1993). Binding to DNA and mediated transactivation is inhibited by complex formation with the the retinoblastoma gene product promoted by complex formation of different retinoblastoma susceptibility gene product. Proc. Nat. Acad. Sci. USA 90, E2F family members. Science 262, 1557-1560. 6914-6918. Krek, W., Ewen, M. E., Shirodkar, S., Arany, Z., Kaelin, W. G. Jr and Ginsberg, D., Vairo, G., Chittenden, T., Xiao, Z. X., Xu, G., Wydner, K. L., Livingston, D. M. (1994). Negative regulation of the growth-promoting DeCaprio, J. A., Lawrence J. B. and Livingston, D. M. (1994). E2F-4, a transcription factor E2F-1 by a stably bound cyclin A-dependent protein new member of the E2F transcription factor family, interacts with p107. kinase. Cell 78, 161-172. Genes Dev. 8, 2665-2679. La Thangue, N. B. (1994). DP and E2F proteins: components of a Girling, R., Partridge, J. F., Bandara, L. R., Burden, N., Totty, N. F., heterodimeric transcription factor implicated in cell cycle control. Curr. Hsuan, J. J. and La Thangue, N. B. (1993). A new component of the Opin. Cell Biol. 6, 443-450. transcription factor DRTF1/E2F. Nature 365, 468. Lam, E. W. and La Thangue, N. B. (1994). DP and E2F proteins: coordinating Goldfarb, A. N. and Lewandowska, K. (1994). Nuclear redirection of a transcription with cell cycle progression. Curr. Opin. Cell Biol. 6, 859-866. cytoplasmic helix-loop-helix protein via heterodimerization with a nuclear Lees, E., Faha, B., Dulic, V., Reed, S. I. and Harlow, E. (1992). Cyclin localizing partner. Exp. Cell Res. 214, 481-485. E/cdk2 and cyclin A/cdk2 kinases associate with p107 and E2F in a Helin, K., Lees, J. A., Vidal, M., Dyson, N., Harlow, E. and Fattaey, A. temporally distinct manner. Genes Dev. 6, 1874-1885. (1992). A cDNA encoding a pRB-binding protein with properties of the Lees, J. A., Saito, M., Vidal, M., Valentine, M., Look, T., Harlow, E., transcription factor E2F. Cell 70, 337-350. Dyson, N. and Helin, K. (1993). The retinoblastoma protein binds to a Helin, K., Harlow E. and Fattaey, A. (1993a). Inhibition of E2F-1 family of E2F transcription factors. Mol. Cell. Biol. 13, 7813-7825. transactivation by direct binding of the retinoblastoma protein. Mol. Cell. Li, Y., Slansky, J. E., Myers, D. J., Drinkwater, N. R., Kaelin, W. G. and Biol. 13, 6501-6508. Farnham, P. J. (1994). Cloning, chromosomal location, and Helin, K., Wu, C. L., Fattaey, A. R., Lees, J. A., Dynlacht, B. D., Ngwu, C. characterization of mouse E2F1. Mol. Cell. Biol. 14, 1861-1869. and Harlow, E. (1993b). Heterodimerization of the transcription factors Logan, T. J., Jordan, K. L. and Hall, D. (1994). Altered shape and cell cycle E2F-1 and DP-1 leads to cooperative trans-activation. Genes Dev. 7, 1850- characteristics of fibroblasts expressing the E2F1 transcription factor. Mol. 1861. Biol. Cell 5, 667-678. Hiebert, S. W., Chellappan, S. P., Horowitz, J. M. and Nevins, J. R. (1992). Milbrandt, J. D., Heintz, N. H., White, W. C. Rothman, S. M. and Hamlin, The interaction of RB with E2F coincides with an inhibition of the J. L. (1981). Methotrexate-resistant Chinese hamster ovary cells have transcriptional activity of E2F. Genes Dev. 6, 177-185. amplified a 135-kilobase-pair region that includes the gene for dihydrofolate Hiebert, S. W. (1993). Regions of the retinoblastoma gene product required for reductase. Proc. Nat. Acad. Sci. USA 78, 6042-6047. its interaction with the E2F transcription factor are necessary for E2 promoter Moran, E. (1993). DNA tumor virus transforming proteins and the cell cycle. repression and pRb-mediated growth suppression. Mol. Cell. Biol. 13, 3384- Curr. Opin. Genet. Dev. 3, 63-70. 3391. Mudryj, M., Devoto, S. H., Hiebert, S. W., Hunter, T., Pines, J. and Nevins, Hijmans, E. M., Voorhoeve, P. M., Beijersbergen, R. L., van ‘t Veer, L. J. J. R. (1991). Cell cycle regulation of the E2F transcription factor involves an and Bernards, R. (1995). E2F-5, a new E2F family member that interacts interaction with cyclin A. Cell 65, 1243-1253. with p130 in vivo. Mol. Cell. Biol. 15, 3082-3089. Muller, R. (1995). Transcriptional regulation during the mammalian cell cycle. Horowitz, J. M. (1993). Regulation of transcription by the retinoblastoma Trends Genet. 11, 173-178. protein. Genes Chromosom. Cancer 6, 124-131. Neuman, E., Flemington, E. K., Sellers, W. R. and Kaelin, W. G. Jr (1994). Hsiao, K. M., McMahon, S. L. and Farnham, P. J. (1994). Multiple DNA Transcription of the E2F-1 gene is rendered cell cycle dependent by E2F elements are required for the growth regulation of the mouse E2F1 promoter. DNA-binding sites within its promoter. Mol. Cell. Biol. 14, 6607-6615. Genes Dev. 8, 1526-1537. Nevins, J. R. (1992). E2F: a link between the Rb tumor suppressor protein and Huang, P. S., Patrick, D. R., Edwards, G., Goodhart, P. J., Huber, H. E., viral oncoproteins. Science 258, 424-429. Miles, L., Garsky, V. M., Oliff A. and Heimbrook, D. C. (1993). Protein Pagano, M., Draetta, G. and Jansen-Durr, P. (1992). Association of cdk2 domains governing interactions between E2F, the retinoblastoma gene kinase with the transcription factor E2F during S phase. Science 255, 1144- product, and human papillomavirus type 16 E7 protein. Mol. Cell. Biol. 13, 1147. 953-960. Qin, X. Q., Livingston, D. M., Kaelin, W. G. Jr and Adams, P. D. (1994). Huber, H. E., Edwards, G., Goodhart, P. J., Patrick, D. R., Huang, H. S., Deregulated transcription factor E2F-1 expression leads to S-phase entry and Ivey-Hoyle, M., Barnett, S. F., Oliff, A. and Heimbrook, D. C. (1993). p53-mediated apoptosis. Proc. Nat. Acad. Sci. USA 91, 10918-10922. 1726 J. Magae and others

Qin, X. Q., Livingston, D. M., Ewen, M., Sellers, W. R., Arany, Z. and Wiman, K. G. (1993). The retinoblastoma gene: role in cell cycle control and Kaelin, W. G. Jr (1995). The transcription factor E2F-1 is a downstream cell differentiation. FASEB J. 7, 841-845. target of RB action. Mol. Cell. Biol. 15, 742-755. Wu, C. L., Zukerberg, L. R., Ngwu, C., Harlow, E. and Lees, J. A. (1995). In Sardet, C., Vidal, M., Cobrinik, D., Geng, Y., Onufryk, O., Chen, A. and vivo association of E2F and DP family proteins. Mol. Cell. Biol. 10, 2536- Weinberg, R. A. (1995). E2F-4 and E2F-5, two members of the E2F family, 2546. are expressed in the early phases of the cell cycle. Proc. Nat. Acad. Sci. USA Wu, X. and Levine, A. J. (1994). p53 and E2F-1 cooperate to mediate 92, 2403-2407. apoptosis. Proc. Nat. Acad. Sci. USA 91, 3602-3606. Schwarz, J. K., Devoto, S. H., Smith, E. J., Chellappan, S. P., Jakoi, L. and Xu, G., Livingston, D. M. and Krek, W. (1995). Multiple members of the E2F Nevins, J. R. (1993). Interactions of the p107 and Rb proteins with E2F transcription factor family are the products of oncogenes. Proc. Nat. Acad. during the cell proliferation response. EMBO J. 12, 1013-1020. Sci. USA 92, 1357-1361. Shan, B., Zhu, X., Chen, P. L., Durfee, T., Yang, Y., Sharp, D. and Lee, W. Xu, M., Sheppard, K. A., Peng, C. Y., Yee, A. S. and Piwnica-Worms, H. H. (1992). Molecular cloning of cellular genes encoding retinoblastoma- (1994). Cyclin A/CDK2 binds directly to E2F-1 and inhibits the DNA- associated proteins: identiﬁcation of a gene with properties of the binding activity of E2F-1/DP-1 by phosphorylation. Mol. Cell. Biol. 14, transcription factor E2F. Mol. Cell. Biol. 12, 5620-5631. 8420-8431. Shan, B. and Lee, W. H. (1994). Deregulated expression of E2F-1 induces S- Yang, X. H. and Sladek, T. L. (1995). Overexpression of the E2F-1 phase entry and leads to apoptosis. Mol. Cell. Biol. 14, 8166-8173. transcription factor gene mediates cell transformation. Gene Express. 4, 195- Shirodkar, S., Ewen, M., DeCaprio, J. A., Morgan, J., Livingston, D. M. 204. and Chittenden, T. (1992). The transcription factor E2F interacts with the Zamanian, M. and La Thangue, N. B. (1993). Transcriptional repression by retinoblastoma product and a p107-cyclin A complex in a cell cycle- the Rb-related protein p107. Mol. Biol. Cell 4, 389-396. regulated manner. Cell 68, 157-166. Zhang, Y. and Chellappan, S. P. (1995). Cloning and characterization of Singh, P., Wong S. H. and Hong, W. (1994). Overexpression of E2F-1 in rat human DP2, a novel dimerization partner of E2F. Oncogene 10, 2085-2093. embryo ﬁbroblasts leads to neoplastic transformation. EMBO J. 13, 3329- Zhu, L., van den Heuvel, S., Helin, K., Fattaey, A., Ewen, M., Livingston, 3338. D., Dyson, N. and Harlow, E. (1993). Inhibition of cell proliferation by Smith, E. J. and Nevins, J. R. (1995). The Rb-related p107 protein can p107, a relative of the retinoblastoma protein. Genes Dev. 7, 1111-1125. suppress E2F function independently of binding to cyclin A/cdk2. Mol. Cell. Zhu, L., Enders, G., Lees, J. A., Beijersbergen, R. L., Bernards, R. and Biol. 15, 338-344. Harlow, E. (1995a). The pRB-related protein p107 contains two growth Vairo, G., Livingston, D. M. and Ginsberg, D. (1995). Functional interaction suppression domains: independent interactions with E2F and cyclin/cdk between E2F-4 and p130: evidence for distinct mechanisms underlying complexes. EMBO J. 14, 1904-1913. growth suppression by different retinoblastoma protein family members. Zhu, L., Xie, E. and Chang, L. S. (1995b). Differential roles of two tandem Genes Dev. 9, 869-881. E2F sites in repression of the human p107 promoter by retinoblastoma and Wells, J., Held, P., Illenye, S. and Heintz, N. H. (1996). Protein-DNA p107 proteins. Mol. Cell. Biol. 15, 3552-3562. interactions at the major and minor promoters of the divergently transcribed dhfr and rep3 genes during the Chinese hamster ovary cell cycle. Mol. Cell. Biol. 16, 634-647. (Received 22 January 1996 - Accepted 16 April 1996)