(Received for publication, September 19, 1996, and in revised form, November 20, 1996)
From the ¶ Laboratory of Chemoprevention, Division of Basic
Sciences, National Cancer Institute, Bethesda, Maryland 20892-5055, the Laboratory of Leukocyte Biology, Division of Basic
Sciences, and the § Intramural Research Support Program,
SAIC Frederick, National Cancer Institute-Frederick Cancer Research
and Development Center, Frederick, Maryland 21702-1201
The E2F family of transcription factors regulates cell cycle progression, and deregulated expression of E2F-1 can lead to neoplastic transformation. In myeloid cells, introduction and expression of the Abelson leukemia virus causes growth factor independence. Here, the p120 v-Abl protein activates E2F-1-mediated transcription through a physical interaction with the E2F-1 transcription factor. BCR-Abl and c-Abl also stimulate E2F-1-mediated transcription. Our results suggest a new mechanism by which v-Abl leads to factor-independent myeloid cell proliferation: the activation of E2F-1-mediated transcription.
Protein kinases play a central role in the regulation of cellular growth and differentiation. The cellular homolog of viral abl (v-abl), which was first isolated from Abelson murine leukemia virus (A-MuLV)1 (1, 2), is a member of the tyrosine kinase family of proto-oncogenes (3, 4). Tyrosine kinase activity has been shown to be essential to the transforming activity of the v-abl oncogene (5). As in the Src family of tyrosine kinases, the Src homology 2 and 3 (SH2 and SH3) domains are located NH2-terminal to the catalytic domain of Abl (6). The SH2 and SH3 domains of nonreceptor tyrosine kinases may be involved in regulation of kinase activity in vivo (7).
The v-Abl tyrosine kinase can stimulate cell proliferation. A-MuLV induces pre-B lymphomas in mice and transforms lymphoid, myeloid, and fibroblastic cells in vitro (8). A-MuLV transformed hematopoietic cells grown without interleukin-3 (9, 10) or interleukin-2 (11) and transformed NIH3T3 cells grown without serum (12). The abrogation of growth factor requirement in A-MuLV-transformed cells can occur through a non-autocrine mechanism, in which the v-Abl tyrosine kinase is essential for the maintenance of factor-independent proliferation (13). These findings suggest that v-Abl activates the mitogenic program.
Recent results indicate that the tyrosine kinase activity of nuclear c-Abl is also regulated during cell cycle progression through an interaction with the retinoblastoma (RB) gene product (14, 15). Inactivation of RB, but not its related genes p107 and p130, has been implicated in the etiology of a subset of human tumors (6, 16-19). RB, a negative regulator of cell proliferation, binds to a number of cellular proteins, and this binding is disrupted by viral oncoproteins (20-23). One of the RB-associated proteins is the transcription factor E2F-1 (20). E2F-1, one of five E2F family members, was originally identified as a cellular DNA-binding protein required for activation of the adenovirus E2A promoter (24). E2F binding sites were found subsequently in the promoters of many cellular genes whose products regulate cell proliferation (4, 25-28). For DNA binding, E2F must form heterodimers with a member of the DP family (29). In addition to RB family members (RB, p107, and p130; 30, 31), E2F interacts with cyclins and their associated kinases (32-35). Expression of the proteins that form complexes with E2F are, in part, cell cycle-dependent. In this study, we investigated whether v-Abl regulates the activity of the transcription factor E2F-1.
Introduction of the v-abl oncogene into an interleukin-3-dependent murine myeloid cell line (32D-123) leads to growth factor independence and unregulated cell proliferation (10, 36). 32D-123 cells and 32D-abl cells were grown as described previously (36). CCL-64 (mink lung epithelial) cells and C33A (human cervical) cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum.
CCL-64 and C33A cells were transfected by the calcium phosphate coprecipitation method, using either 10 µg of control plasmid or 10 µg of v-Abl expression plasmid. Cells were harvested 48 h after the addition of DNA, and extracts were assayed for chloramphenicol acetyltransferase (CAT) activity. The murine myeloid 32D-123 and 32D-abl cells were transfected by electroporation as described previously (37). The DNA concentration was equalized in each case by the addition of pUC when necessary. After 24 h at 37 °C, cells were harvested, and the protein concentration of cell lysates was determined using the Bio-Rad protein assay. Equal amounts of protein were used to assay for the CAT enzyme. For normalization of transfection efficiencies, a human growth hormone expression plasmid (pSVGH) was included in the cotransfections. The level of growth hormone expression was determined using a growth hormone detection kit (Nichols Institute, San Juan Capistrano, CA). All experiments were repeated at least three times.
Plasmid ConstructsPlasmids containing a deletion mutant or mutations in the enhancer of the adenovirus E2 promoter linked to the CAT gene were a generous gift of Dr. John Brady (NCI, Bethesda, MD) and were described previously (4, 38). c-abl and BCR-abl expression vectors were kindly provided by Dr. Ann Marie Pendergast (Duke University Medical Center, Durham, NC). v-Abl expression plasmids (Abl-FL, -M7, and -M9) used for the expression of native and mutant v-Abl DNA fragments were described previously (63). The GAL4-E2F-1, cytomegalovirus-E2F-1 and GST-E2F-1 constructs (a generous gift of William G. Kaelin, Harvard Medical School, Boston) were described previously (39). For the construction of GST fusion plasmids, polymerase chain reaction products were ligated into pGEX-2T using standard methods to generate GST-E2F-1 (89-190, 89-250, 89-400, 89-437, 188-250, 251-400, and 251-437).
All GAL4-E2F-1 fusion plasmids were constructed by inserting the
appropriate E2F-1 DNA fragment in-frame to the GAL4 (1-147) sequence
in the vector pSG424 (40). E2F-1 DNA fragments were produced by
polymerase chain reaction. The 5-oligonucleotide contained an
EcoRI site, and the 3
-oligonucleotide contained an
XbaI site. Using these oligonucleotides, fragments were
amplified according to the standard protocol of the GeneAmp kit
(Perkin-Elmer). The junctions of all GAL4 fusion plasmids were
confirmed by DNA sequencing.
For in vitro transcription and translation in rabbit reticulocyte extracts, various portions of the v-abl coding regions were placed in-frame behind the phage T7 promoter in the pGEM4 plasmid (Promega). v-abl DNA fragments with terminal BamHI and EcoRI sites were produced by polymerase chain reaction.
Immunoprecipitations and Western Blot AnalysisAntibodies used in this study were obtained from Santa Cruz Biotechnology, Inc. unless otherwise specified. Normal control IgG was obtained from Sigma. For coimmunoprecipitations, 2 × 107 cells were lysed with 100 µl of RIPA buffer (150 mM NaCl, 1.0% Nonidet P-40, 0.5% deoxycholic acid, 0.1% SDS, 50 mM Tris (pH 8.0)) containing 1% bovine serum albumin. Antibody (1 µg) was incubated with the extracts for 1 h at 4 °C. The antigen-antibody complexes were precipitated with protein G-Plus Sepharose (Santa Cruz Biotechnology), washed three times with RIPA containing 1% bovine serum albumin, three times with RIPA lacking bovine serum albumin, and then separated by SDS-PAGE. The separated proteins were transferred to reinforced nitrocellulose membranes (Schleicher & Schuell) and probed by Western blot analysis. The antigen-antibody complexes were detected by enhanced chemiluminescence following the manufacturer's instructions (Amersham Corp.). In vivo phosphorylation analysis was carried out by labeling the 32D cells with inorganic 32P for 2 h at 37 °C. 32P-Labeled cell extracts were prepared in RIPA containing phosphatase inhibitors (1 mM sodium orthovanadate, 30 mM NaF, 30 mM NaPPi), and proteins were immunoprecipitated with the specific antisera and analyzed by electrophoresis as described in the figures. Proteins were visualized by autoradiography and/or Western blot analysis.
GST-Pull-down assayThe GST-pull-down assay was performed as described previously (41) with minor modifications. The concentration of the GST fusion proteins was determined by Coomassie Blue staining. The beads were washed twice with NETN buffer (20 mM Tris-HCl (pH 8.0), 100 mM NaCl, 1 mM EDTA, 0.5% Nonidet P-40) and once with binding buffer (50 mM Tris-HCl (pH 8.0), 140 mM NaCl, 0.5% Nonidet P-40, 100 mM NaF, 200 µM sodium orthovanadate, and 1% bovine serum albumin). The beads were rocked for 1 h at room temperature with 2-10 µl of in vitro synthesized [35S]v-Abl protein in a final volume of 200 µl of binding buffer. The beads were then washed three times in 1 ml of NETN buffer, pelleted at 500 × g for 2 min, boiled in sample buffer, and the bound 35S-labeled v-Abl proteins were resolved by SDS-PAGE.
For the pull-down assay of v-Abl from 32D-abl nuclear extracts by GST-E2F-1 deletion fusion proteins, the affinity matrix was prewashed in NETN with 500 mM NaCl, subsequently equilibrated in NETN buffer, and incubated with nuclear extracts. Bound proteins were detected by immunoblotting with specific antiserum to Abl.
E2F was
functionally defined as a transcription factor that mediates
transcriptional activation of the adenovirus E2 promoter (24, 42).
Therefore, we tested the ability of v-Abl to transactivate either the
wild type E2 promoter or a series of E2F promoters containing
5-10-base pair mutations from 85 to
28 (Fig.
1A). The E2-CAT reporter constructs were
transfected into both 32D-123 cells and CCL-64 cells with either the
v-Abl expression vector or the control vector. Fig. 1, B and
C, shows data from a representative transfection. Expression
of the v-abl gene resulted in a 2.5-6-fold increase in the
level of expression arising from the wild type E2 reporter construct
(pEC79). Single-site mutation of the ATF binding site between
80
and
70 or a single mutation of the E2F binding sites,
64 to
60 or
45 to
36, had minimal or no effect on v-Abl-mediated stimulation of
E2 promoter activity (Fig. 1C). Mutation of the ATF binding
site,
75 to
71, or the second E2F binding site,
65 to
61, also
had no effect on stimulation of the E2 promoter by v-Abl. We also
tested the effect of v-Abl on an E2 reporter construct containing a
double mutation ATF (
80 to
70) and E2F (
45 to
36; Fig.
1C, lanes 11 and 12). Similar to the
80 to
70 mutation, an increase in CAT expression was observed in
the presence of v-Abl. In contrast, mutation of both E2F binding sites
at
64 to
60 and
45 to
36 abrogated induction of E2 promoter
activity by v-Abl (Fig. 1, B and C, lanes
7 and 8 or 9 and 10,
respectively), demonstrating that v-Abl activates E2 promoter activity
through the E2F binding sites.
E2F-1 Physically Associates with v-Abl
Previous studies by
our laboratories have shown that the p120-v-Abl protein is localized in
the nucleus of 32D-abl cells and binds and activates the transcription
factor CREB (37). To identify a possible association between E2F-1 and
v-Abl leading to increased E2F transactivation, coimmunoprecipitations
using antibodies to the E2F-1 protein followed by Western blot analysis
using antibodies to the Abl protein were performed (Fig.
2A). An association between v-Abl and E2F-1
was demonstrated in 32D-abl cells by detection of the v-Abl protein in
the anti-E2F-1 immunoprecipitates. Immunoprecipitation with antibodies
against Fos, Myc, or normal mouse IgG failed to precipitate the v-Abl
protein. Next, the reciprocal experiments were performed. Using lysates
from 32Pi-labeled 32D-abl cells,
immunoprecipitation experiments using antibodies to Abl or E2F-1
followed by Western blot with the indicated antibodies (Fig. 2,
B-D) were conducted. The phosphorylated 120-kDa v-Abl
protein (Fig. 2B) and the 65-kDa E2F-1 protein (Fig.
2C) were observed in anti-Abl and anti-E2F
immunoprecipitates. As a control, immunoprecipitation with antibodies
against Src and IgG failed to precipitate v-Abl or E2F-1 protein (Fig.
2, B-D). The identity of the E2F-1 protein was verified by
Western blotting the same membrane as in Fig. 2C with E2F-1
antibodies (Fig. 2D). Immunoprecipitation and Western
analysis of the v-Abl protein as well as Western analysis of the E2F-1
protein showed that 32D-abl cells express the v-Abl and E2F-1 proteins
abundantly (Fig. 2E). These results show that the
immunoprecipitation of E2F-1 by the E2F-1 antibody is less than that of
E2F-1 by the Abl antibody (Fig. 2D). This could be a result
of an innate difference in the antibodies such as lower antibody
affinity. Overall, the data in Fig. 2 show that a significant portion
of E2F-1 (but not all of E2F-1) is associated with v-Abl.
Localization of the v-Abl Binding Site in E2F-1
To map the
v-Abl binding site in E2F-1, we assayed the ability of the v-Abl
protein from 32D-abl nuclear extracts to bind the GST-E2F-1 fusion
protein (Fig. 3). v-Abl in the 32D-abl cell nuclear
extracts bound GST fusion proteins containing either the RB binding
domain (COOH-terminal, 38 amino acids from 400 to 437) or the DNA
binding domain (amino acids 89-190) of the E2F-1 protein (Fig. 3,
A and C). A protein consisting of the amino acids
188-400 showed no significant binding to v-Abl (data not shown). Also, two different constructions (188-250 and 251-400), which covered the
region 188-400, failed to interact with v-Abl (Fig. 3B).
Thus, p120-v-Abl binds two distinct regions of the E2F-1 protein, the DNA binding domain and the RB binding domain.
v-Abl Stimulates GAL4-E2F-1-mediated Transcription
To examine
whether the ability of v-Abl to bind E2F-1 leads to activation of
E2F-1-mediated transcription, we expressed various GAL4-E2F-1
derivatives. These plasmids were cotransfected into the 32D-123 or
32D-abl cells with a CAT reporter (G5BCAT) containing five GAL4 binding
sites upstream from the E1B TATA box (Fig. 4). The
p120-v-Abl expression plasmid (Fig. 4C) or the control
vector pUC (Fig. 4B) was also added to the cotransfection
mixtures for the 32D-123 cells. Transcriptional activation directed by
v-Abl was diminished with deletion of the COOH-terminal region of E2F-1 (Fig. 4C, lane 9). v-Abl, however, had no effect
on the transcriptional activation of the GAL4-DNA binding domain (see
Figs. 6 and 7). Using the same protocol, the mutant construct 10-200
was not susceptible to v-Abl activation (data not shown). This region
is not involved in v-Abl-mediated E2F-1-transactivation.
To determine whether v-Abl directly transactivates E2F-1 or merely
releases E2F-1 from RB-mediated suppression, we cotransfected GAL4-E2F-1, both in the presence and absence of a v-abl expression construct, into C33A cells that are Rb/
. v-Abl
continued to demonstrate transactivation of E2F-1 in this Rb-lacking
system, indicating that the effect of v-Abl on E2F-1-activated transcription occurs independent of the presence of RB (data not shown).
To define the
domain(s) of the v-Abl protein interacting with E2F-1, the ability of
various in vitro translated v-Abl fragments to bind the
GST-E2F-1 fusion protein was assayed (Fig. 5). The COOH-terminal 368 amino acids (611-981) or the
NH2-terminal 355 amino acids of v-Abl did not bind E2F-1
via these regions (Fig. 5B). Expression of various internal
fusion proteins suggested that the E2F-1 binding domain in v-Abl is
located in the kinase domain between amino acids 355 and 613 (Fig. 5,
A and B). We therefore examined the activity of
kinase domain deletion mutants of v-Abl to transactivate GAL4-E2F-1.
Kinase-inactive v-Abl mutants containing either a point mutation
(Abl-M1) or deletions (Abl-M7 and Abl-M9) in the ATP binding domain
only minimally transactivated GAL4-E2F-1 (10-18% of wild type v-Abl;
Fig. 6B). Furthermore, a mutant form of v-Abl
containing a deletion of the nuclear localization sequence (Abl-M6)
similarly abolished the transactivation potential of the v-Abl protein
(data not shown), indicating that nuclear expression is needed for the
transactivation potential of v-Abl.
v-Abl, c-Abl, and BCR-abl Stimulate GAL4-E2F-1-mediated Transcription
Transfection of 32D-123 cells with plasmids encoding other members of the abl gene family, c-abl or the naturally occurring leukemogenic protein, BCR-abl, also stimulated E2F-1-mediated transcription (Fig. 7B, lanes 3 and 4), whereas v-Abl, c-Abl or BCR-Abl did not stimulate G5BCAT expression in the presence of the GAL4 DNA binding domain alone (pSG147; lanes 5-8, Fig. 7B). Since the v-Abl kinase domain is common to c-Abl and BCR-Abl, it is possible that c-Abl and BCR-Abl also interact with E2F-1 and stimulate E2F-1-mediated transcription through a similar mechanism.
In this study, we found that v-Abl overexpression can stimulate transcription of the adenovirus E2 promoter through the two E2F binding sequences. To implicate E2F-1 directly in v-Abl-stimulated transcription, we showed that v-Abl significantly stimulates transcription mediated by GAL4-E2F-1. Our studies indicate that the RB binding domain of E2F-1 is sufficient to confer positive transcriptional regulation by v-Abl. Previous studies by Helin et al. (43) using an E2F-1 protein truncated by a 20-amino acid deletion (417-437) showed no significant effect on E2F-1 transactivation. In contrast, we found that v-Abl does not induce transcription mediated by GAL4-E2F-1 with a larger COOH-terminal 37-amino acid deletion (401-437). This 17-amino acid difference could be critical for E2F-1 transactivation ability, explaining the difference in our results from those of Helin et al.
A previous study has shown that deletion of the COOH-terminal region of E2F-1 does not abolish function as a transcriptional activator, suggesting that the NH2-terminal region is not a component of the E2F-1 transactivation domain (43). In contrast, this study demonstrates that the E2F-1 COOH-terminal region alone does possess weak transcriptional activation properties which, in the presence of v-Abl, become further activated (Fig. 4). It is possible that this activity occurs in specific cell types and requires the expression of unidentified transactivators that interact with the E2F-1 COOH terminus. In this regard, the ability of v-Abl to recruit binding proteins to E2F complexes has recently been reported (44). Moreover, we found that a GAL4-Abl fusion protein has no transcriptional activation properties.2 These results indicate that E2F-mediated transcription is activated by v-Abl. Recently, we have shown that the oncoprotein v-Abl can also activate CREB-mediated transcriptional events (63), indicating that E2F-1 is not the sole target for v-Abl regulation.
Studies to determine whether v-Abl increases E2F binding activity or changes the mobility of E2F complexes were attempted and proved to be difficult to perform. E2F binds DNA tightly so that supershifts with specific E2F-1 antibodies have not been reported, consequently gel shift assays will not distinguish E2F-1 complexes from complexes containing other E2F family members. However, supershift analysis has been used to show the ability of v-Abl to recruit binding proteins to E2F complexes (44).
The E2F family of transcription factors plays a crucial role in cell cycle progression by linking the cell cycle machinery with the transcription apparatus (45, 46). E2F-1 is one of the molecules that can drive quiescent cells to enter S phase (48). This activity depends on its ability to bind DNA and activate transcription. E2F transcription factor family members can act as oncogene products (48-52). Singh et al. (48) showed that overexpression of E2F-1 leads to uncontrolled cellular proliferation and causes neoplastic transformation in rat embryo fibroblasts. Morishita et al. (53) showed a transcription factor decoy of the E2F binding site to inhibit smooth muscle proliferation in vivo. Our results suggest that activation of genes responsive to E2F-1 by p120-v-Abl contributes to v-Abl-mediated cell transformation.
Previous studies by Renshaw et al. (54) support a dual
function for v-Abl, as an activator and as a suppressor of cell growth, depending on the cellular context. Inhibition of cell proliferation by
c-Abl overexpression in fibroblasts has also been reported (55). Our
data show that c-Abl induced E2F-1 transactivation. These seemingly
contradictory data, c-Abl-mediated activation of E2F with c-Abl being
growth inhibitory (55) and, in contrast, v-Abl-mediated activation of
E2F with v-Abl being growth stimulatory (presented here and by Wong
et al. (44)), could be explained by as yet unidentified
differences that are intrinsic to the nature of the oncogenic v-Abl
proteins as well as the cellular environment that interacts with these
proteins. Interestingly, the E2F-1 knock-out mice have led to similar
conundrums at the organismal level. In the E2F-1/
mice
expressing phenotype (56, 57), E2F-1 functions to regulate apoptosis
and suppress cell proliferation, whereas overexpression studies
indicate that E2F-1 stimulates cell proliferation. These results
suggest that the actions of a single gene (like E2F-1) are either (i)
deterministically dependent upon the cellular environment or (ii)
inherently different in its mode of action, perhaps relying on its
concentration or other factors that allow for it to behave both as a
simulator or as an inhibitor of cellular functions (58).
The v-Abl-E2F-1 interaction and transactivation events shown here
suggest that a complex system exists for controlling the activity of
the E2F transcription factor. This interaction can occur in the absence
of RB as demonstrated by our studies in RB/
cells.
Another study with Rb
/
murine fibroblasts showed
abnormal S phase entry and activation of E2F-responsive genes (59)
despite the presence of p107 and p130, which seem to bind distinctly to
other members of the E2F family (31, 60). Inactivation of RB, but not
the p107 and p130 genes, has been implicated in the etiology of a
subset of human tumors (16-19, 61). Recently, c-Abl has been shown to
form a complex with the hypophosphorylated form of RB (14, 15). Hypophosphorylated RB also binds to a domain in the carboxyl-terminal region of E2F-1 and decreases transcriptional activation of E2F target
genes like c-myc (43, 62) and c-myb (38).
Phosphorylation of RB disrupts the interaction between RB and c-Abl, as
well as RB and E2F-1.
In 32D-abl cells, in addition to binding to E2F-1 as shown here, we have reported recently that the p120-v-Abl protein can bind to RB (37). Both events increase E2F-1 transactivation ability. These findings suggest that these two regulatory pathways are interrelated and may implicate v-Abl-mediated transcriptional activation in v-Abl-driven cellular transformation events. v-Abl transformation leads to significant overexpression of p120-v-Abl, which not only binds E2F-1 but also binds RB, resulting in constitutive activation of E2F-mediated transcription.
We thank D. Kim and A. Roberts for comments on the manuscript and helpful discussion.