![[Feedback]](/icons/banner/feedbackACT.gif)
![[For Subscribers]](/icons/banner/subscriberACT.gif)
![[Archive]](/icons/banner/archiveACT.gif)
![[Search]](/icons/banner/searchACT.gif)
![[Contents]](/icons/banner/tocACT.gif)
Previous Article | Next Article 
Molecular and Cellular Biology, November 1999, p. 7447-7460, Vol. 19, No. 11
0270-7306/99/$04.00+0
Copyright © 1999, American Society for Microbiology. All rights reserved.
Rb and Prohibitin Target Distinct Regions of E2F1 for Repression
and Respond to Different Upstream Signals
Sheng
Wang,
Niharika
Nath,
Gina
Fusaro, and
Srikumar
Chellappan*
Department of Pathology, College of
Physicians and Surgeons, Columbia University, New York, New York
10032
Received 19 January 1999/Returned for modification 22 March
1999/Accepted 2 August 1999
 |
ABSTRACT |
E2F transcription factor is subject to stringent regulation by a
variety of molecules. We recently observed that prohibitin, a potential
tumor suppressor protein, binds to the retinoblastoma (Rb) protein and
represses E2F transcriptional activity. Here we demonstrate that
prohibitin requires the marked box region of E2F for repression;
further, prohibitin can effectively inhibit colony formation induced by
overexpression of E2F1 in T47D cells. Prohibitin was also found to
interact with the signaling kinase c-Raf-1, and Raf-1 could effectively
reverse prohibitin-mediated repression of E2F activity. Agents such as
E1A, p38 kinase, and cyclins D and E had no effect on
prohibitin-mediated repression of E2F1, but all of these molecules
could reverse Rb function. Similarly, stimulation of the immunoglobulin
M signaling pathway in Ramos cells could inactivate prohibitin, but
this had no effect on Rb function. Serum stimulation of quiescent Ramos
cells inactivated Rb and prohibitin with different kinetics; further,
while the serum-dependent inactivation of Rb was dependent on
cyclin-dependent kinase activity, the inactivation of prohibitin was
not. We believe that prohibitin is a novel regulator of E2F function
which channels specific signaling cascades to the cell cycle regulatory machinery.
| |
INTRODUCTION |
The E2F family of transcription
factors plays a significant role in the regulation of mammalian cell
cycle progression (33). Studies in recent years have
identified E2F as an important downstream target of the retinoblastoma
(Rb) family of growth regulatory proteins, and it appears that Rb
exerts its growth regulatory function at least in part by regulating
E2F activity (15). Since the E2F family of transcription
factors is capable of inducing different cell fates such as
proliferation, apoptosis, or differentiation, an understanding
of their regulation would throw light on the biochemical pathways
underlying such phenomena (41).
The term E2F generally refers to a family of six proteins named E2F1
through E2F6 (24). Of these, E2Fs 1 to 5 possess a transcriptional activation domain at the carboxy terminal
(50) and can induce transcription from target promoters in
association with dimerization partners 1 or 2 (DP1 or DP2)
(32). In contrast, E2F6 lacks an activation domain and is
repressive in nature; E2F6 has been shown to compete for E2F binding
sites on promoters and repress their activity (7, 19, 58).
Within the five transcriptionally active E2Fs, there are certain
biochemical and functional differences, though it is not yet clear
whether they execute distinct functions in normal cells (10,
50). For example, though overexpression of E2F1 can induce
S-phase entry in quiescent cells (22), E2F1 can also induce
apoptosis under appropriate conditions (1, 18, 28, 45,
67); other E2F family members lack the ability to do so
(27). Similarly, E2F1 is capable of transforming primary cells in association with Ras, but it is not yet clear whether other
E2F family members are capable of doing so (49); in a recent
study, E2F1 was shown to block the differentiation of myeloid cells, but E2F3 could not (53). Thus, despite the fact all
five transcriptionally active E2Fs bind to the same DNA recognition site, they may have different functional niches in the cell
(13).
E2F family members are also under different levels of control by
upstream molecules (15). E2Fs 1, 2, and 3 can all bind to
the Rb protein, but E2Fs 4 and 5 preferentially bind to p107 and p130
proteins (3, 10). It has been demonstrated that the Rb
family proteins bind to a moiety within the transcriptional activation
region of E2Fs, effectively repressing their activity (21, 25,
44). In addition to passive repression of E2F-mediated transcription, Rb has been shown to actively repress transcription from
promoters carrying E2F binding sites by recruiting the histone deacetylase HDAC1 (5, 36, 37, 65, 66). Thus, the presence of
E2F sites on a promoter does not always indicate that it is induced by
E2F but, on the contrary, that it can be repressed through those sites
as well (6, 20, 31). A very good example is the E2F1
promoter itself (23). In addition to the preferential interaction with Rb family members, E2Fs 1, 2, and 3 also possess a
binding site for cyclins at the amino-terminal region, allowing them to
be regulated by the associated cyclin-dependent kinases, mainly cdk2
(26, 69). E2Fs 4 and 5 lack this mode of regulation (10).
Our attempts to characterize additional proteins that bind to Rb family
members led to the identification of prohibitin, a potential tumor
suppressor protein, as an Rb-binding protein (62). Prohibitin could effectively bind to Rb, p107 and p130 and could repress the transcriptional activity of all of the five E2Fs but had no
effect on promoters lacking an E2F binding site. Prohibitin had to
interact with Rb to bring about the transcriptional repression of E2F,
and this correlated with its ability to suppress colony formation in
T47D cells. Interestingly, we also found that adenovirus E1A was unable
to reverse prohibitin-mediated repression of E2F1 (62). This
led us to believe that prohibitin represses E2F activity through
mechanisms different than those used by the Rb protein. The studies
described here demonstrate that prohibitin targets a different region
of E2F1 than Rb, and prohibitin-mediated repression of E2F can be
released by signals which do not target the Rb protein. Our results
suggest that prohibitin-mediated repression of E2F is an additional
mechanism that allows E2F to function in response to specific signaling pathways.
| |
MATERIALS AND METHODS |
Cell lines, vectors, and transfections.
Ramos cells were
maintained in Dulbecco modified Eagle medium, T47D and U937 cells grown
in RPMI medium, both containing 10% fetal bovine serum (FBS).
Transient transfections were conducted on T47D breast carcinoma cells
by using the calcium phosphate precipitation method according to
standard protocols. Ramos cells were transfected by electroporation
with a Bio-Rad Gene Pulser at 250 V, with a 960-µF capacitance. Both
Ramos and T47D cells are Rb positive. For serum stimulation
experiments, transfected cells were maintained in medium containing
10% FBS for 18 h prior to transfer to medium without serum. After
48 h of starvation, the cells were stimulated by using medium
containing 10% FBS for the required periods of time.
A total of 2 µg of plasmids was used in all transfections for
reporter analysis, unless noted otherwise; 8 µg of the expression vectors was used when extracts had to be prepared from the transfected cells for biochemical analysis. A 1-µg amount of a pSV
Gal vector was included as internal control in all transfections, and the -galactosidase value varied only slightly within each experiment. In
all cases, representative chloramphenicol acetyltransferase (CAT) assay
results from multiple experiments are shown. Where data is graphically
represented, CAT activity was assessed by scanning the intensity of
acetylated chloramphenicol. The total amount of DNA used for
transfections was normalized by using salmon sperm DNA in all the lanes.
Expression vectors for GAL4 and VP16 fusions of E2F1 (pCGE2F1 and
pCMVE2F1VP16, respectively), as well as the pG5E1BCAT reporter, were a
kind gift from David Johnson. E2CAT vector contains the CAT gene driven
by the adenovirus E2 promoter carrying two E2F binding sites.
Full-length pDCE2F1 vector was used to induce the E2CAT in all
transfections except those shown in Fig. 2B, where the wild-type and
mutant E2Fs were expressed from the pCR3.1 vector. E2F1 mutants A, B,
and C were generated by a PCR-based overlap-extension protocol, and
their integrity was confirmed by sequencing. pSVRb, pCDNA3Raf-1
(wild-type as well as 
28) and pCDNA3-prohibitin have been described
before (62). Generation of Raf-1 mutants for mapping of
binding domains was described earlier (61). C-terminal deletion mutants of prohibitin were generated in pCR2 vector by using
PCR techniques, and the internal deletion mutant 185-214 was
generated by an overlap-extension strategy. The prohibitin antisense
construct was made by cloning the prohibitin cDNA in the reverse
orientation in pCR3-1 vector. Other expression vectors used in the
study are pRC/CMVCycD1, pCMVCycE, pCMVcdk2D145N (dominant-negative cdk2), pCMVcdk4DN, pCMVcdk6DN, pCMVE1A, and pSR
p38.
Stable transfections were performed on 35-mm-diameter dishes by using
approximately 10,000 cells and subjected to selection in the
appropriate antibiotic for 14 days. The total amount of DNA transfected
was equalized with salmon sperm DNA in every sample. Cells were fixed
and stained with crystal violet, and colonies with more than 20 cells
were counted.
In vitro binding assays.
Glutathione
S-transferase (GST) fusions of Rb, Raf-1, and prohibitin
were prepared as described earlier (61, 62).
35S-labeled Raf-1 and prohibitin proteins were generated by
in vitro transcription-translation in rabbit reticulocyte lysates by
using standard protocols. First, 8 to 10 µl of synthesized
polypeptide was incubated with glutathione beads carrying equal amounts
of GST fusion proteins in 200 µl of protein binding buffer (20 mM Tris, pH 7.5; 50 mM KCl; 0.5 mM EDTA; 1 mM dithiothreitol [DTT]; 0.5% NP-40; 3 mg of bovine serum albumin [BSA] per ml) at 4°C for
2 h. The beads were then washed six times with 1 ml of protein binding buffer and eluted with 10 mM glutathione. Eluates were separated in an 8% sodium dodecyl sulfate (SDS)-polyacrylamide gel and
visualized by autoradiography. The protein amounts in control input
lanes were approximately one-fifth of the total used in binding assay.
Immunoprecipitation and Western blots.
Polyclonal antibodies
to prohibitin were a kind gift of J. Keith McClung, and monoclonal
antiprohibitin antibodies were obtained from NeoMarkers. Anti-cRaf-1
monoclonal antibody was obtained from Transduction Laboratories;
anti-Rb and anti-c-Myc antibodies were purchased from Oncogene
Science-Calbiochem. Antibodies to p16, p107, p130, and p38 were
obtained from Santa Cruz Biotechnologies; anti-human immunoglobulin M
(IgM) antibody was from Southern Biotechnologies, and the anti-pTyr
antibodies were from UBI.
Whole-cell extracts were prepared by hypotonic shock followed by salt
extraction, as described previously (8). Portions (50 to 200 µg) of whole-cell extracts were treated with 5 µl of the
appropriate primary antibody in a volume of 100 µl at 4°C for
1 h. Then, 3 mg of protein A-Sepharose or protein G-Sepharose in a
100-µl volume was added and incubated for an additional hour. The
binding was performed in a buffer containing 20 mM HEPES (pH 7.9), 40 mM KCl, 1 mM MgCl2, 0.1 mM EGTA, 0.1 mM EDTA, 0.1 mM DTT,
0.1 mM NaF, 0.1 mM Na3VO4, 0.5% NP-40, and 3 mg of BSA per ml. The beads were washed six times with 600 µl of the
same buffer, boiled in 20 µl of SDS sample buffer, and separated on 8 or 10% polyacrylamide gels. After semidry transfer to supported
nitrocellulose membrane, the blots were probed with the appropriate
antibody. The proteins were detected by using an enhanced
chemiluminescence assay system from Amersham.
EMSA.
Electrophoretic mobility shift assay (EMSA) after
immunoprecipitation was performed as previously described (8,
71). An EcoRI-HindIII fragment of the
adenovirus E2 promoter containing two E2F binding sites (TTTCGCGC)
was end labeled by using Klenow fragment and was used as the
probe in all of the assays. Briefly, 8 µg of whole-cell extracts
prepared as described above was incubated with approximately 0.2 ng of
32P-labeled E2F probe in a buffer containing 20 mM HEPES
(pH 7.9); 40 mM KCl; 0.1 mM concentrations each of MgCl2,
EGTA, EDTA, DTT, NaF, and Na3VO4; 1% NP-40, 1 µg of salmon sperm DNA per ml, and 10 µg of BSA per ml. After
incubation at room temperature for 20 min, the reactions were separated
on a 4% polyacrylamide gel in 0.25× TBE at 300 V for 3 h. The
gel was dried, and the bands were detected by autoradiography.
| |
RESULTS |
Prohibitin and Rb target different regions of E2F1 for
repression.
Since we found that prohibitin can specifically
repress all transcriptionally active members of the E2F family
(62), we attempted to identify the domain of E2F that is
targeted by prohibitin. Studies were conducted on E2F1 to this end. As
a first step, we examined the effect of prohibitin on two chimeric E2F1
proteins by transient-transfection experiments in T47D cells. In the
first experiment, a chimeric E2F protein with the carboxy-terminal 153 amino acids of E2F1 (which contains the transcriptional activation domain) fused to a GAL4 DNA-binding domain (11) was tested
for its ability to respond to prohibitin. As shown in Fig.
1A, the GAL4-E2F1 construct, as well as a
control GAL4VP16 vector, could induce transcription from a pG5E1BCAT
reporter; cotransfection of prohibitin or Rb had no effect on the
transcription induced by the GAL4-VP16 but specifically abolished the
transcriptional activity of the GAL4-E2F1 protein. This result suggests
that prohibitin can target E2F1 specifically and that the
carboxy-terminal domain containing the transcriptional activation
domain can respond to prohibitin.
View larger version (40K):
[in this window]
[in a new window]
|
FIG. 1.
Repression of E2F1 fusion proteins by prohibitin and Rb.
(A) A pG5E1BCAT reporter was induced by transfecting the indicated GAL4
fusion proteins. Cotransfection of 2 µg of prohibitin or Rb
specifically represses GAL4E2F1 but not GAL4VP16 protein. (B) An E2CAT
reporter was cotransfected with expression vectors for E2F1 or
E2F1-VP16AD fusion protein. Prohibitin could repress both the
constructs, but Rb does not repress E2F1-VP16 (unlike prohibitin).
|
|
In a second experiment, we examined whether prohibitin can target the
amino-terminal region of E2F1 that contains the DNA-binding domain. An
E2F1-VP16 construct that contains the region from residues 1 to 357 of
E2F1 fused to a VP16 activation domain (11) was tested for
its ability to respond to prohibitin. As shown in Fig. 1B, the
E2F1-VP16 construct could activate transcription from an E2CAT vector
as effectively as E2F1. It was found that cotransfection of prohibitin
could effectively repress both wild-type E2F1 and the E2F1-VP16 fusion
protein. This raised the possibility that prohibitin either targets
multiple regions of E2F1 for repression or targets a region of E2F1
shared between the E2F1-VP16 and GAL4-E2F1 constructs.
In contrast to repression by prohibitin, Rb could repress the activity
of full-length E2F1 but had no effect on the E2F1-VP16 construct (Fig.
1B). The observation that Rb can repress GAL4-E2F1 as well as
full-length E2F1, but not E2F1-VP16 (which lacks an E2F1 activation
domain), is consistent with the fact that Rb specifically targets the
activation domain of E2F1 for repression.
Prohibitin targets the marked box region of E2F1.
Experiments
were designed to further delineate the region of E2F1 targeted by
prohibitin. As shown in Fig. 2A, the
region from residues 284 to 357 of E2F1 is present in GAL4-E2F1 and in the E2F1-VP16 constructs, both of which were repressed by prohibitin. This region corresponds to the highly conserved marked box region of
E2F (50). Attempts were made to create internal deletions of
this region and to assess whether such mutant E2Fs could respond to
prohibitin. A deletion of the entire region from residues 284 to 357 of
E2F1 totally abolished its transcriptional activity (data not shown).
Hence we divided this region into three parts (regions from residues
284 to 304, 304 to 329, and 329 to 357) and generated internal
deletions of each segment by using a PCR-based overlap-extension
protocol. A schematic of the different E2F mutants, as well as of the
E2F1 fusion proteins, is shown in Fig. 2A.
View larger version (55K):
[in this window]
[in a new window]
|
FIG. 2.
Prohibitin targets the marked box region of E2F1 for
repression. (A) Schematic of the different E2F1 fusion proteins, as
well as the E2F1 internal deletion mutants, used in the study. Mutants
A, B, and C lack a 20-amino-acid sequence between the indicated
residues. (B) Repression of E2F1 internal deletion constructs by
prohibitin and Rb in T47D cells. Mutant A had approximately 20-fold
more transcriptional activity than wild-type (WT) E2F1 (lanes 2 and 3).
Prohibitin could fully repress mutant A but had no effect on mutants B
and C (lanes 7 and 8). Rb could repress all three (lanes 9 to 11). (C)
Western blot analysis of extracts from T47D cells transiently
transfected with 8 µg of pCR3.1 E2F1 vectors expressing wild-type or
mutant E2F1 proteins. (D) EMSA for E2F in the same extracts. There was
no significant difference in the DNA binding activities of the three
proteins.
|
|
The ability of each internal deletion to induce transcription from an
E2CAT vector was tested (Fig. 2B, lanes 2 to 5). Mutants B and C (which
had internal deletions of regions from residues 304 to 329 and 329 to
357) had amounts of transcriptional activity comparable to that of
wild-type E2F1; surprisingly, deletion mutant A, which lacked residues
284 to 304, had 20-fold more transcriptional activity compared to the
wild-type E2F1 (lane 2). The ability of Rb and prohibitin to repress
the three mutant E2Fs was next examined by a cotransfection experiment.
As shown in lane 6, prohibitin could effectively repress mutant A; in
contrast, prohibitin had no effect on the transcriptional activity of
mutants B and C (lanes 7 and 8), suggesting that the region of E2F1
spanning residues 304 to 357 is essential for prohibitin-mediated
repression. It was found that Rb could effectively repress all three
mutants (Fig. 2B, lanes 9 to 11), since all had an intact activation
domain. This result suggests that prohibitin targets a region of E2F1 at the carboxy terminal of the marked box domain and not the activation domain itself.
A Western blot analysis of T47D cells transiently transfected with the
mutant E2F constructs showed that all are expressed equally well and
produce stable proteins (Fig. 2C). In addition, there was no
significant difference in the DNA-binding activity of the mutants (Fig.
2D), and it is not clear at this moment why mutant A has such a high
transcriptional activity.
Prohibitin can bind to E2F in vitro and in vivo.
Since we
found that prohibitin can efficiently repress E2F activity through the
marked box domain, we examined whether prohibitin can physically
interact with E2F1. As a first step, we examined whether E2F1 can bind
to prohibitin in an in vitro GST binding assay. E2F1 was synthesized by
in vitro transcription-translation in rabbit reticulocyte lysates, and
its binding to GST fusions of prohibitin or Rb was tested. As shown in
Fig. 3A, top panel, the full-length E2F1
could efficiently bind to both the fusion proteins; there was no
binding to control unprimed GST beads. This suggested that prohibitin
might be interacting with E2F1 to bring about the repression. Since we
found that the marked box of E2F1 responds to prohibitin, we tested the
mutants of this region for their ability to bind to prohibitin. As
shown in Fig. 3A, E2F1 mutant A could bind efficiently to prohibitin as
well as to Rb; in contrast, mutants B and C could bind only to GST-Rb and not to prohibitin. This correlates with their ability to respond to
prohibitin and Rb, since we find that mutant A responds to Rb, as well
as prohibitin, but that mutants B and C can be repressed only by Rb.
View larger version (49K):
[in this window]
[in a new window]
|
FIG. 3.
Prohibitin binds to the marked box region of E2F1. (A)
In vitro binding assay showing the interaction of full-length E2F1 or
mutants A, B, or C to GST fusions of prohibitin or Rb. RL indicates
one-fifth of the loading material, and GST indicates unprimed beads.
(B) Association of prohibitin with E2F1 in Ramos cells. Ramos
whole-cell extracts were immunoprecipitated with the indicated
antibodies, and the presence of prohibitin was examined by Western
blotting. WCE indicates an equivalent amount of whole-cell extract. (C)
Endogenous prohibitin levels affect E2F activity. ZR751 cells were
transfected with 12 µg of E2CAT vector alone or with increasing
amounts of an antisense prohibitin construct. Transcriptional activity
from the E2CAT reporter increases with the amount of antisense
prohibitin construct used. (D) The effect of antisense prohibitin
construct on 12 µg of E2CAT or 4 µg of c-FosCAT vectors in ZR751
and BT549 cells. Transcription from the E2CAT vector is increased in
response to the antisense construct in ZR751 cells but not in BT549;
there was no increase in c-FosCAT in either cell line. (E) Western blot
analysis of E2F1 and prohibitin in BT549 cells, as well as ZR751 cells,
transiently transfected with the indicated amounts of antisense
prohibitin construct.
|
|
It was next examined whether the prohibitin-E2F1 interaction can be
detected in vivo; an immunoprecipitation-Western blot analysis was used
for this purpose. Whole-cell extract prepared from the human B-cell
line Ramos was immunoprecipitated with antibodies to p16, E2F1, and Rb;
the presence of prohibitin was examined by Western blotting by using a
monoclonal antiprohibitin antibody. As shown in Fig. 3B, there was no
prohibitin detected in the p16 immunoprecipitate, but prohibitin was
found to be associated with E2F1, as well as Rb. This suggests that
prohibitin and E2F1 can associate in vivo, and this association can be
detected without overexpressing any component.
We previously reported that the level of endogenous prohibitin in human
breast cancer cells could affect the activity of an E2CAT reporter, but
not a c-fos promoter or AP1CAT reporter. To assess whether
endogenous E2F activity was affected by altering the amounts of
prohibitin already present in cells, we used an antisense strategy. The
cell line ZR751, which has abundant levels of endogenous prohibitin,
was used for this purpose. Transfection of 12 µg of an E2CAT reporter
resulted in a low level of CAT activity in ZR751 cells, but
cotransfection of increasing levels of an antisense prohibitin
construct led to an increase in the observed CAT conversion. This
reflects an increase in the transcriptional activity of the endogenous
E2F present in the cell when prohibitin level is reduced (Fig. 3C).
Two separate control experiments were performed to verify the above
result. First, cotransfection of 6 or 12 µg of antisense prohibitin
did not lead to an increase in E2F activity in BT549 cells, which have
no detectable amount of prohibitin (Fig. 3D, striped bars). Second, to
rule out that the repression of E2F by prohibitin is through a
nonspecific effect on components of the general transcriptional
machinery, a similar antisense experiment was conducted with a c-FosCAT
reporter. As shown in Fig. 3D, whereas increasing amounts of antisense
prohibitin construct elevated the transcriptional activity of the E2CAT
in ZR751 cells, it did not induce c-FosCAT activity, suggesting that
the effect on E2F is a specific event. The antisense prohibitin had no
significant effect on either reporter in BT549 cells, which lack
endogenous prohibitin (Fig. 3D, striped bars).
A Western blot experiment was conducted to examine whether transfection
of antisense prohibitin did reduce the levels of prohibitin protein. As
shown in Fig. 3E, there is a comparable amount of E2F1 present in BT549
and ZR751 cells, and the levels of E2F1 do not change by transfecting
antisense prohibitin. In contrast, BT549 cells have no detectable
prohibitin (Fig. 3E, lane 1); further, the level of prohibitin in ZR751
cells is reduced upon transfecting the antisense construct, correlating
with the increase in E2F activity. Taken together, these results
strongly suggest that the level of prohibitin in cells has a direct
bearing on E2F mediated transcription.
Prohibitin can repress E2F1-mediated induction of cell
proliferation.
Overexpression of E2F1 has been demonstrated to
induce cell proliferation in many mammalian cell lines, and the Rb
protein could repress this efficiently (1, 14, 22). Since
prohibitin has strong antiproliferative activity that correlated with
its ability to bind to the Rb protein, we designed experiments to examine whether prohibitin could repress E2F1-mediated induction of
cell proliferation. Toward this purpose, a colony formation assay was
performed on T47D cells by using standard protocols as described
previously (62).
As shown in Table 1, transfection of a
pSV-NEO vector gave rise to ca. 119 colonies after 14 days of
selection; in contrast, transfection of 2 µg of a prohibitin
expression vector reduced the number of colonies to ca. 43. All the
E2F1 constructs could stimulate the proliferation of T47D cells at the
levels transfected (3 µg); the induction was most pronounced in the
case of mutant A, which doubled the number of colonies. Prohibitin
could effectively suppress the colonies induced by the transfection of
wild-type E2F1 and mutant A; surprisingly, prohibitin could not
suppress the colony formation when mutants B and C were transfected.
This experiment suggests that prohibitin and E2F1 have antagonistic effects on cell proliferation, and there is a correlation between the
ability of prohibitin to repress the transcriptional activity of E2F1
and its ability to reverse E2F1-mediated cell proliferation.
Raf-1 can bind to prohibitin and regulate its function.
It has
been reported that prohibitin is associated with the IgM receptor
(55), and recently it was shown to associate with the mixed
lineage kinase-2 (MLK2) (46). Since these observations indicated that prohibitin could be involved in mediating signal transduction cascades and since prohibitin had a potential tyrosine phosphorylation site, we felt it was prudent to examine whether prohibitin was tyrosine phosphorylated. To address this, we
immunoprecipitated whole-cell extracts from proliferating or
tetradecanoyl phorbol acetate (TPA)-treated U937 cells with an
antiprohibitin antibody and probed the immunoprecipitate with an
anti-phosphotyrosine antibody in a Western blot. There was no
detectable tyrosine phosphorylated band corresponding to the size of
prohibitin; surprisingly, the major band recognized by the
phosphotyrosine antibody in the prohibitin immuneprecipitate was at ca.
72 to 76 kDa (Fig. 4A). The 76-kDa band
was more pronounced in the U937 extracts that were treated with TPA and
the prohibitin immuneprecipitate of this TPA-treated extract. It is
known that Raf-1 is ca. 72 to 76 kDa, that it is tyrosine
phosphorylated (16, 40), and that Raf-1 tyrosine phosphorylation increases after TPA treatment (2). Further, Raf-1 is known to be involved in IgM-mediated signaling cascades (57). Hence, we decided to examine whether the protein
associated with prohibitin was Raf-1. This was verified by a
coimmunoprecipitation-Western blot experiment, with whole-cell extracts
from dividing U937 cells. As shown in Fig. 4B, all the three Rb family
members could be detected in association with prohibitin; we observed
that in two other cell lines as well: Ramos and Daudi (62).
Interestingly, Raf-1 could be detected in the prohibitin immune
precipitate but not in an unrelated kinase (p38). These interactions
were specific, since neither Raf-1 nor Rb family proteins could be
detected in a c-Myc immune precipitate.
View larger version (31K):
[in this window]
[in a new window]
|
FIG. 4.
Prohibitin associates with Raf-1 in vivo and in vitro.
(A) Western blot analysis with an anti-P-Tyr antibody on whole-cell
extracts or prohibitin immunoprecipitates from dividing or TPA-treated
U937 cells. The major 72-kDa protein band comigrates with Raf-1. (B)
Association of Rb family proteins and Raf-1 with prohibitin in U937
cells. Extracts from dividing U937 cells were immunoprecipitated with
an anti-c-Myc antibody as a control or an antiprohibitin antibody. The
immune precipitates were tested for the presence of the indicated
proteins by Western blotting. Raf-1 can be detected in prohibitin
immune precipitates but not p38 kinase. (C) GST-binding assay with in
vitro-synthesized Raf-1 and GST fusions of Rb and prohibitin. RL
indicates one-fifth of the amount of rabbit reticulocyte lysate used in
the binding reaction. (D) In a similar experiment, in vitro-synthesized
prohibitin was tested for binding to GST-Raf-1 beads or beads primed
with GST protein. Prohibitin specifically binds to GST-Raf-1.
|
|
Additional experiments were conducted to verify whether Raf-1 and
prohibitin could interact directly. GST binding assays were conducted
toward this purpose, and the results are shown in Fig. 4C and D. In the
first experiment, Raf-1 was synthesized in vitro in rabbit reticulocyte
lysates, and its binding to beads carrying GST as control, or to GST
fusions of Rb or prohibitin, was assessed. As shown in Fig. 4C, Raf-1
could bind to Rb and prohibitin very efficiently. Under identical
conditions, Raf-1 had no binding to a variety of other GST fusion
proteins, indicating that its binding to Rb and prohibitin is a
specific interaction.
Similarly, the binding of in vitro-synthesized prohibitin to
GST-Raf-1 was assessed. As shown in Fig. 4D, prohibitin
could bind efficiently to GST-Raf-1 beads but not to control GST
beads. These experiments suggest that prohibitin and Raf-1 can bind to each other, and this interaction is probably direct.
Different regions of Raf-1 are involved in binding to Rb and
prohibitin.
Attempts were made to identify the region of Raf-1
involved in binding to Rb and prohibitin. GST-binding assays using a
panel of Raf-1 deletion mutants had shown that the extreme
amino-terminal 28 amino acids of Raf-1 are involved in binding to Rb
(61). The ability of the same set of Raf-1 deletion mutants
to bind to prohibitin was tested similarly. As shown in Fig.
5A, full-length Raf-1 protein, as well as
one spanning residues 1 to 335, could bind to both Rb and prohibitin.
Deletion of the amino-terminal 132 amino acids abolished the ability of
Raf-1 to bind to Rb, but it could bind to prohibitin very well. A
similar binding pattern was observed with a Raf-1 molecule lacking the
amino-terminal 254 amino acids. In contrast, the amino-terminal 254 amino acids could bind to Rb but not to prohibitin. A constitutively
active form of Raf-1, Raf-1 BXB, whose amino-terminal 26 amino acids were fused to the carboxy-terminal residues 297 to 648, could bind to
both Rb and prohibitin. This indicated that the prohibitin-binding region of Raf-1 spans residues 297 to 335. Interestingly, the Raf-128 construct, which could not bind to Rb, was very efficient in
binding to prohibitin.
View larger version (26K):
[in this window]
[in a new window]
|
FIG. 5.
Raf-1 binds to Rb and prohibitin through distinct
domains. (A) Binding assay results with different deletion mutants of
Raf-1 synthesized in vitro and GST-Rb and GST-prohibitin. Filled box
represents the region used for binding, and the line indicates the
deleted region. RL represents one-fifth of the lysate used for binding.
The regions of Raf-1 involved in binding to Rb and prohibitin are also
shown. (B) Reversal of prohibitin-mediated repression of E2F1 by Raf-1.
T47D cells were transfected with an E2CAT reporter and E2F1.
Transfection of Rb (lanes 3 to 5) or prohibitin (lanes 6 to 8) could
repress transcription. Cotransfection of wild-type Raf-1 could reverse
both Rb-mediated (lane 4) and prohibitin-mediated (lane 7) repression;
Raf-128 had no effect on Rb (lane 5) but could fully reverse
prohibitin-mediated repression (lane 8).
|
|
Since our earlier experiments had shown that Raf-1 could effectively
reverse Rb-mediated repression of E2F activity, we next examined
whether Raf-1 could reverse prohibitin-mediated repression of E2F1
activity as well. A transient-transfection experiment in T47D cells
showed that wild-type Raf-1 could reverse Rb-mediated repression of
E2F1 activity, (Fig. 5B, lane 4), whereas Raf-128 had no effect. In
contrast, prohibitin-mediated repression of E2F1 could be effectively
reversed by Raf-1 wild type (lane 7), as well as Raf-128. This
experiment suggests that Raf-1 can reverse prohibitin-mediated
repression of E2F activity and that residues 1 to 28 of Raf-1 are
essential only for targeting Rb function.
A colony formation assay was performed on T47D cells to verify whether
Raf-1 could affect prohibitin-mediated suppression of cell
proliferation. As shown in Table 2,
transfection of prohibitin reduced the number of colonies about
threefold, whereas transfection of Raf-1 could increase the number of
colonies formed. Cotransfection of Raf-1 could effectively reverse
prohibitin-mediated repression of cell proliferation, but the effect
was not as pronounced as that on Rb.
Raf-1 targets the carboxy-terminal region of prohibitin.
We
had earlier found that the Rb-binding region (residues 74 to 116) is
necessary for prohibitin to repress the transcriptional activity of
E2F1. Experiments were designed to examine whether additional
regions of prohibitin were necessary for the repressive activity of
prohibitin. As a first step, we made deletion mutants of prohibitin
that lacked various numbers of residues from the C-terminal end and
examined their ability to repress E2F activity. A prohibitin construct
expressing residues 1 to 157 was unable to repress E2F activity (Fig.
6A); similarly, a construct spanning residues 1 to 185 had no repressive effects on E2F1. Prohibitin constructs having residues 1 to 214, as well as residues 1 to 243, could repress E2F activity as effectively as the full-length prohibitin, suggesting that the region spanning residues 185 to 214 is
required for repression of E2F activity, in addition to the Rb-binding
domain (residues 74 to 116). This was confirmed by making a prohibitin
construct with an internal deletion of residues 185 to 214 (as shown in
Fig. 6B), whereas wild-type prohibitin could repress E2F1 activity, the
deletion construct was unable to do so.
View larger version (33K):
[in this window]
[in a new window]
|
FIG. 6.
Raf-1 targets the carboxy-terminal end of prohibitin.
(A) Mapping of the regions of prohibitin required for repression of E2F
activity. C-terminal deletion mutants of prohibitin were tested for
their ability to repress E2F1-mediated transcription in T47D cells.
Prohibitins 1-157 and 1-185 were unable to repress transcription, but
prohibitin 1-214 and 1-243 could. While both the 1-214 and 1-243 mutants of prohibitin could repress E2F1, the repression was not
affected by cotransfected Raf-1. (B) The involvement of residues 185 to
214 in transcriptional repression was tested by using an internal
deletion mutant of prohibitin. Deletion of the residues 185 to 214 impaired the ability of prohibitin to repress E2F1 activity. (C) The
carboxy-terminal end of prohibitin interacts with Raf-1. A GST-binding
assay showing the binding of prohibitin to GST fusions of Raf-1 and Rb;
while full-length prohibitin could bind to both the beads, prohibitin
1-243 could bind only to Rb. (D) Residues 185 to 214 of prohibitin is
involved in binding to E2F1. E2F1 synthesized in vitro was checked for
binding to GST fusions of full-length prohibitin or prohibitin
185-214, or Rb. E2F1 could bind only to full-length prohibitin and
Rb (upper panel), whereas Raf-1 could bind to all the three GST fusion
proteins. (E) Schematic showing the different regions of prohibitin
involved in binding to Rb, E2F1, and Raf-1.
|
|
The ability of Raf-1 to reverse the repression mediated by the
C-terminal truncated prohibitin molecules was examined by a cotransfection experiment in T47D cells. As shown in Fig. 6A, Raf-1 was
unable to reverse the repression mediated by prohibitin 1-243; this
result suggested that Raf-1 targets the C-terminal region of prohibitin
spanning residues 243 to 275. It was examined by using a GST-binding
assay whether the region of prohibitin residues 243 to 275 that was
targeted by Raf-1 was involved in the binding to Raf-1 also. As shown
in Fig. 6C, left panel, in vitro-synthesized full-length prohibitin was
able to bind to GST-Rb as well as GST-Raf-1 but not to control GST
beads. In contrast, prohibitin 1-243 could bind to GST-Rb but not to
GST-Raf-1 (right panel). This suggests that the extreme C-terminal
region of prohibitin is involved in binding to Raf-1, as well as
responding to Raf-1-mediated signaling events. Similar experiments were
conducted to examine whether there was any correlation between the
abilities of prohibitin to bind E2F and to repress transcription. E2F1
made in rabbit reticulocyte lysates could bind very well to GST fusions
of full-length prohibitin and Rb, but it could not bind to the
truncated prohibitin 1-185 or the internal deletion mutant 185-214
(Fig. 6D, top panel). This suggests that prohibitin has to associate
with E2F1 to repress its transcription. The binding of Raf-1
synthesized in vitro to the same beads was also tested; it was found
that Raf-1 could not bind to the truncated prohibitin 1-185, but it
could bind to the internal deletion mutant 185-214 (Fig. 6D, bottom
panel). This result confirms the earlier findings that Raf-1 targets
the C-terminal region of prohibitin. These experiments help delineate three distinct functional domains of prohibitin: the region from residues 74 to 116, which is involved in Rb binding; the region from
residues 185 to 214, which is necessary for transcriptional repression;
and the carboxy-terminal 243-275 region involved in responding to Raf-1
(Fig. 6E).
The contribution of the various domains of prohibitin to growth
suppression was assessed by a colony formation assay on T47D cells
(Table 3). It was found that full-length
prohibitin could effectively inhibit colony formation after 14 days of
antibiotic selection; interestingly, prohibitin mutants 1-157 and 1-185 had little effect on the number of colonies. In addition, prohibitin constructs possessing an internal deletion of the region from residues
74 to 116 or 185 to 214 were not able to suppress colony formation.
Thus, it appears that the ability of prohibitin to exert growth
regulation coincides with its ability to repress transcription mediated
by E2F, as in the case of Rb.
In contrast, the constructs lacking the carboxy-terminal end
(constructs 1-214 and 1-243) were quite efficient in repressing proliferation. Since these two mutants of prohibitin could not respond
to Raf-1 in transient-transfection assays, we examined whether Raf-1
could affect the growth suppression mediated by these constructs. It
was found that Raf-1 could reverse the suppression of colony formation
mediated by wild-type prohibitin; similarly, the presence of Raf-1
increased the number of colonies when the truncation mutants 1-157 and
1-185 were present. Raf-1 had no effect, however, on the suppression of
colony formation mediated by the prohibitin mutants 1-214 and 1-243;
this again suggests that the carboxy terminus of prohibitin is the
functional target of Raf-1. Further, it appears that the ability of
prohibitin to suppress cell proliferation correlates with its ability
to repress transcription and that agents that can reverse the
transcriptional repression can also affect the growth suppression
mediated by prohibitin.
E2F regulation by Rb and prohibitin respond to different
signals.
We had observed that E1A was unable to reverse
prohibitin-mediated repression of E2F1. Since the results described
above demonstrate that Raf-1 was quite effective in releasing
prohibitin-mediated repression of E2F1, we performed a series of
transient-transfection experiments to assess the effect of different
agents on Rb- and prohibitin-mediated repression of E2F1. First, we
compared the effects of E1A, Raf-1, and p38 kinase on Rb and prohibitin
(Fig. 7A). As we have observed earlier,
E1A, Raf-1, and p38 all could effectively inactivate Rb, reversing the
repression of E2F transcriptional activity (63). In
contrast, only Raf-1 was able to reverse prohibitin-mediated inhibition
of E2F activity (Fig. 7A), indicating that Rb and prohibitin probably
respond to different signal transduction cascades. A summary of these
results is presented in Table 4.
View larger version (38K):
[in this window]
[in a new window]
|
FIG. 7.
Differential regulation of Rb and prohibitin by upstream
molecules. (A) Transient-transfection experiment in T47D cells with an
E2CAT reporter, which is induced by E2F1. Both Rb and prohibitin could
repress E2F1 activity. Whereas Raf-1 could reverse both Rb- and
prohibitin-mediated repression of E2F1, p38 and E1A could only reverse
Rb but not prohibitin repression. Similarly, cotransfection of cyclins
D and E can reverse Rb-mediated repression of E2F1 but had no effect on
prohibitin. (B) Control experiment showing that the different agents
used do not directly affect E2F1 activity in T47D cells.
|
|
Since cyclins D and E, along with their dependent kinases, are known to
be the major regulators of Rb function during the course of
G1 progression (35, 64), we decided to examine
whether these molecules had any effect on prohibitin function. In a
similar transient transfection experiment on T47D cells, cotransfection of either cyclin D or cyclin E could effectively repress Rb-mediated repression of E2F transcriptional activity (Fig. 7A). Both of these
cyclins had no effect on prohibitin-mediated repression of E2F1
activity, suggesting that prohibitin is not regulated by the same
cyclin dependent kinases that regulate Rb. Cotransfection of E1A, p38,
Raf-1, or cyclins D and E had only minimal effect on E2F1 directly in
T47D cells (Fig. 7B).
Prohibitin-mediated repression of E2F is reversed by IgM receptor
stimulation.
Prohibitin and a prohibitin-related protein have been
reported to be associated with IgM receptors in B-cell lines
(55), and stimulation of IgM receptors has been shown to
activate Raf-1 kinase (57). Since prohibitin can reverse E2F
activity and since it associates with IgM receptors as well as Raf-1,
we examined whether prohibitin-mediated repression of E2F activity is
affected by stimulating the IgM receptors. The human B-cell line Ramos was transiently transfected with E2CAT reporter whose activity was
stimulated by cotransfecting E2F-1. Interestingly, treating the
transfected cells with 1 µg of an anti-IgM antibody per ml could
reverse prohibitin-mediated repression of E2F (Fig.
8A). The reversal was reflected in an
increase in CAT activity, which could be observed as early as 45 min
after treatment; treatment of the cells for additional periods of time
(up to 48 h) did not result in any further increase in E2F
activity. Identical results were obtained in two other B-cell lines,
Daudi and Raji (data not shown).
View larger version (32K):
[in this window]
[in a new window]
|
FIG. 8.
IgM stimulation can modulate prohibitin function. (A)
Ramos cells were transiently transfected with E2CAT and E2F1, along
with prohibitin or Rb. Cells were stimulated with 1 µg of an anti-IgM
antibody per ml for the indicated periods of time; prohibitin-mediated
repression of E2F1 is reversed within 45 min of stimulation. IgM
stimulation, even for prolonged periods of time, had no effect on Rb.
(B) The carboxy-terminal region of prohibitin is necessary for
responding to IgM. In a cotransfection experiment, full-length
prohibitin or prohibitin 1-243 was used to repress E2F1. IgM
stimulation for 2 h reversed the repression mediated by the
full-length prohibitin but not prohibitin 1-243. (C) Association of
E2F1 and Rb with prohibitin is disrupted upon IgM signaling. Extracts
from Ramos cells stimulated with an anti-IgM antibody for the indicated
period of time were immunoprecipitated with an anti-E2F1 antibody
(lanes 5 to 8) or anti Rb antibody (lanes 9 to 12). Prohibitin was
detected by Western blotting. Lanes 1 to 4 show a prohibitin level in
an equivalent amount of the extracts. (D) Similar experiment showing
the association of Rb with E2F1 (lanes 5 to 8). Rb remains associated
with E2F1 during the course of IgM stimulation. (E) EMSA showing the
E2F binding activity in Ramos cells stimulated with anti-IgM antibody
for the indicated periods of time.
|
|
It was next examined whether the effect of IgM receptor stimulation was
restricted to prohibitin. In a similar cotransfection experiment on
Ramos cells, Rb was used to repress E2F1 activity instead of
prohibitin. Stimulation of the IgM receptor with 1 µg of an anti-IgM
antibody per ml had no effect on Rb-mediated repression of E2F1
activity even after 48 h (Fig. 8A). This experiment suggests that
the IgM receptor stimulation specifically targets prohibitin and that
prohibitin can respond to signals from this receptor. It was next
examined whether the C-terminal deletion mutant of prohibitin that
could not respond to Raf-1 was inactivated by IgM-mediated signaling.
As shown in Fig. 8B, IgM stimulation could reverse repression mediated
by full-length but not the C-terminal truncated prohibitin, indicating
that Raf-1 and IgM target the same domain of prohibitin.
Since prohibitin was found to bind to E2F1 in vivo, we attempted to
assess whether the physical interaction between prohibitin and E2F1 is
disrupted upon IgM signaling. Whole-cell extracts prepared from Ramos
cells stimulated with an anti-IgM antibody for various periods of time
were immunoprecipitated with antibodies to E2F1 or Rb (Fig. 8C). The
presence of prohibitin in the immunoprecipitates was examined by
Western blot analysis. The level of prohibitin in the extracts did not
change upon IgM stimulation. There was a considerable amount of
prohibitin associated with E2F1 and Rb in unstimulated cells and cells
stimulated for 30 min with the IgM antibody but, interestingly, there
was no detectable amount of prohibitin associated with E2F1 or Rb after
2 or 4 h of stimulation. Thus, it appears that the reversal of
prohibitin-mediated repression of E2F1 by IgM correlates with a
disruption of its interaction with Rb and/or E2F1. In a parallel
experiment, we tested whether there is any change in the amount of Rb
associated with E2F1 when cells are stimulated with IgM. As shown in
Fig. 8D, the amount of Rb associated with E2F1 remains constant during
the course of stimulation. It appears that IgM receptor stimulation
leads to a specific dissociation of prohibitin from Rb and E2F.
There was no significant change in the DNA binding activity of E2F and
E2F-containing complexes in whole-cell extracts of Ramos cells treated
with IgM (Fig. 8E). This might suggest that the release of repression
from prohibitin may not involve changes in the DNA binding activity of
E2F1 but might be occurring through the inactivation of additional
repressors tethered by prohibitin onto E2F1. Though it has been
reported that p107-E2F complexes are altered in response to IgM
stimulation, such changes occur only after prolonged periods of
stimulation (30).
IgM-mediated inactivation of prohibitin does not require
cyclin-dependent kinase activity.
Attempts were made to assess
which signaling cascades are involved in mediating the effects of IgM
receptor stimulation or prohibition. As a first step, IgM stimulation
was conducted in the presence of chemical inhibitors of different
kinases, and the results are shown in Fig.
9A. IgM stimulation for 2 h in the presence of PD98059, a specific inhibitor of MEK1 kinase
(39), totally blocked IgM-mediated reversal of prohibitin
function. Olomoucine, which inhibits cdk2 and cdc2, had no effect on
prohibitin function, ruling out a role for these kinases in the
process. Further, a p38 kinase inhibitor, SB203580 (56, 70),
could also not block IgM-mediated inactivation of prohibitin.
This experiment suggested that IgM-mediated reversal of prohibitin
occurs mainly through the mediation of the mitogen-activated
protein (MAP) kinase cascade; the direct role of Raf-1 in this process
is not clear.
View larger version (23K):
[in this window]
[in a new window]
|
FIG. 9.
Cyclin-dependent kinases are not involved in the
regulation of prohibitin function. (A) Ramos cells were transiently
transfected with E2CAT, E2F1, and prohibitin as in Fig. 8A. Stimulation
with anti-IgM antibody for 2 h released prohibitin-mediated
repression of E2F1; MEK inhibitor (PD98059, 100 µM) could block
IgM-mediated reversal of prohibitin function, but inhibitors of p38
kinase (SB203580, 10 µM) or cdk2 (Olomoucine, 200 µM) had no
effect. A similar experiment was conducted to check whether
dominant-negative cyclin-dependent kinases (4 µg each of
dominant-negative cdk4 and cdk6 and 4 µg of dominant-negative cdk2)
could block IgM-mediated inactivation of prohibitin. Stimulation with
anti-IgM could still reverse prohibitin-mediated repression of E2F1.
(B) Serum stimulation of quiescent Ramos cells reverses prohibitin and
Rb function. Ramos cells were transiently transfected with E2CAT and
E2F1, along with Rb and prohibitin; serum stimulation reversed
Rb-mediated repression of E2F1 within 12 h; repression by
prohibitin was released only after 20 h. (C)
Cyclin-cyclin-dependent kinase activity is required for reversing Rb,
but not prohibitin, function. Ramos cells were transiently transfected
with E2CAT, E2F1, Rb, or prohibitin as in panel B. Serum stimulation
for 24 h reversed both Rb and prohibitin, but cotransfected
dominant-negative cdk4 and cdk6 blocked inactivation of Rb not
prohibitin. Similar results were obtained with dominant-negative
cdk2.
|
|
The possibility that cyclin-dependent kinases are not involved in
IgM-mediated regulation of prohibitin was examined by a direct
experiment. A transient-transfection experiment was conducted, where a
dominant-negative cdk4 and cdk6, or cdk2 was included. As shown in Fig.
9A, cotransfection of a combination of dominant-negative cdk4 and cdk6
or a dominant-negative cdk2 had no significant effect on IgM-mediated
reversal of prohibitin function. This finding is consistent with the
observation that cyclins D and E had no direct effect on
prohibitin-mediated repression of E2F activity and supports the results
obtained with Olomoucine.
Since it was found that cyclins and cyclin-dependent kinases had no
significant effect on prohibitin function, it was examined whether
prohibitin-mediated repression of E2F1 activity responds to serum
stimulation (Fig. 9B). Ramos cells were transiently transfected with
E2CAT and E2F1, and E2F activity was repressed by cotransfecting Rb or
prohibitin. Serum stimulation of the transfected cells showed that
Rb-mediated repression of E2F1 activity was completely reversed within
12 h; there was no change in prohibitin-mediated repression of
E2F1 at the same time point. It was found that prohibitin-mediated inhibition of E2F1 activity was reversed only after 20 h of serum stimulation. Similarly, serum stimulation resulted in a delayed release
of prohibitin-mediated repression of E2F1 activity in T47D cells also.
This experiment suggests that Rb- as well as prohibitin-mediated
repression of E2F1 activity can respond to serum stimulation but that
they follow different kinetics.
G1 progression after serum stimulation of quiescent cells
is known to involve the inactivation of the Rb protein by cyclins and
cyclin-dependent kinases (35, 54, 64). Since Rb and prohibitin responded differently to IgM stimulation and serum stimulation, attempts were made to evaluate whether dominant-negative cyclin-dependent kinases had any differential effects on Rb and prohibitin. Ramos cells were transiently transfected with E2CAT and
E2F1, along with Rb or prohibitin. The transfected cells were serum
stimulated for 24 h, and both Rb- and prohibitin-mediated repression of E2F1 were released by this time (Fig. 9C). Cotransfection of a combination of cdk4 and cdk6 totally prevented the release of
Rb-mediated repression of E2F1 in response to serum; a
dominant-negative cdk2 construct also had a similar effect.
Interestingly, neither cdk4 and cdk6 nor cdk2 dominant-negative
constructs could block serum-mediated inactivation of prohibitin. This
experiment shows that though prohibitin and Rb can respond to serum
stimulation, they are inactivated by different signaling pathways.
| |
DISCUSSION |
The transcriptional activity of E2F1 is regulated at multiple
levels, mainly through its interaction with a variety of cellular proteins (15, 41). The interaction of E2F1 with Rb has
attracted considerable attention since E2F appears to be the major
downstream target of Rb in modulating G1/S transition. In
addition, E2F1 can interact with positive modulators such as DP1 and
DP2, as well as the adenovirus E4 protein (11, 12, 33).
Further, the transcriptional coactivator p300/CBP has been suggested to be a positive modulator of E2F1 activity (34, 59, 60). E2F1 activity has been shown to be modulated by cyclin A and cdk2 after direct binding to its amino-terminal domain (69). This
results in the phosphorylation of E2F1 or DP1, which affects the
transcriptional activity of the former (29). Further,
proteins like Mdm2 and p53 have been suggested to bind to E2F1 or to
affect its transcriptional activity indirectly (43, 51, 68).
Our results suggest that prohibitin is a novel regulator of E2F
function and that prohibitin is probably acting as a link between
certain specific signal transduction pathways and the cell cycle machinery.
We have previously shown that the repression mediated by Rb and
prohibitin showed qualitative differences, mainly based on the
observation that adenovirus E1A protein could not affect
prohibitin-mediated repression of E2F activity. Further, Rb could
repress only E2Fs 1, 2, and 3, whereas prohibitin could inhibit all
five transcriptionally active E2Fs. We show here that Rb and prohibitin
target different regions of the E2F1 protein for repression, since
prohibitin requires the marked box region for exerting its effects. The
amino-terminal end of the marked box region is involved in binding to
DP1 (50), but the region targeted by prohibitin has no known
function. Interestingly, this region is fairly conserved between the
different E2F family members, probably enabling prohibitin to repress
all of them, unlike Rb family members. Our experiments show that
prohibitin can physically interact with E2F1 through this region of the
marked box. Nevertheless, the finding that the Rb binding region is
also necessary for repressing E2F activity raises the possibility that prohibitin has to tether to Rb or an Rb family member to effectively bind E2F and repress its activity. The precise mechanism by which prohibitin represses the transcriptional activity of E2F is not yet
clear, and in preliminary experiments we do not find prohibitin affecting the DNA binding activity of E2F1 or its interaction with DP1
(data not shown). Hence, we believe that prohibitin could be recruiting
additional corepressor molecules onto E2F (62); a good
candidate appears to be HDAC1, which has been shown to be involved in
mediating Rb-mediated transcriptional repression (5, 36,
37). As we had proposed earlier, it may be imagined that
prohibitin is acting as an adaptor between such repressors and Rb/E2F.
The identity of the repressors, as well as the stoichiometry of the
interactions, remains to be elucidated.
One common feature of the repression of E2F by Rb and prohibitin is the
correlation between their ability to inhibit E2F activity and their
ability to suppress cell proliferation. It has been established in the
case of Rb that E2F can antagonize Rb-mediated growth suppression and
that an E2F molecule that could not bind to Rb could induce cell
proliferation irrespective of the presence of Rb (48). We
find that a similar situation exists in the case of prohibitin as well,
since prohibitin could block induction of colony formation by wild-type
E2F, as well as E2F1 mutant A, both of which it could repress.
Prohibitin had no effect on the induction of colony formation by
mutants B or C, which it could not repress. This is analogous to our
finding that a prohibitin mutant that could not bind to Rb and repress
E2F activity could not suppress colony formation in T47D cells. Though
growth-suppressive properties of Rb correlated with its ability to
repress E2F, other phenomena, such as induction of a senescent
phenotype in Saos-2 osteosarcoma cells, did not require the binding or
repression of E2F activity (47). It would be interesting to
see whether prohibitin has similar functions related to growth control
which are separable from its E2F regulation.
We had reported previously that Raf-1 could interact with the Rb
protein and reverse its growth-suppressive properties (61). We find that Raf-1 is capable of interacting with prohibitin as well;
in fact, our studies on the Rb-Raf-1 interaction stemmed from
the observation that Raf-1 can bind to prohibitin, which in turn could
bind to Rb. So far, we have no evidence to suggest that
prohibitin, Rb, and Raf-1 exist in a ternary complex, since all three
proteins can bind to each other in a GST binding assay. The domains of
Raf-1 and prohibitin involved in interacting with each other are
distinct from the domains involved in Rb binding. Unlike the
interaction of Raf-1 with Rb, which is serum inducible, we do not find
any change in the status of the prohibitin-Raf-1 interaction.
Prohibitin protein has a certain degree of similarity to the 14-3-3 class of adapter proteins and may be functioning as an adapter itself.
This raises the possibility that prohibitin functions as a specific
adapter that can stably interact with a variety of molecules,
facilitating their modulation by signaling cascades. Interestingly, it
has been reported recently that prohibitin physically interacts with
mixed lineage kinase-2 (MLK2), which is a MAP kinase-kinase-kinase like
Raf-1, that functions in the p38/JNK signaling pathway (17,
46).
A role for prohibitin in replicative senescence, as well as
mitochondrial functions, has been reported (4, 9, 52), and a
putative amino-terminal domain is believed to facilitate these
functions. Deletion of the amino-terminal 34 amino acids did not affect
prohibitin-mediated repression of E2F activity, suggesting that this
region plays a distinct role in growth regulation separate from the
potential mitochondrial functions. Further, prohibitin, as well as a
prohibitin-related protein, has been reported to be associated with the
IgM receptor in B cells, but the functional consequence of this
interaction is not known. We find that IgM stimulation can specifically
release prohibitin-mediated repression of E2F activity, which
corresponds with a dissociation of prohibitin from E2F1 as well
as Rb. It is possible that the release of prohibitin causes the
inactivation or dissociation of a corepressor leading to the
transcriptional activity. So far, we do not find any change in the
DNA-binding activity of E2F in response to IgM stimulation. It has been
reported that IgM stimulation can affect the Rb-related p107 protein
and its regulation of E2F activity, but this does not occur until
48 h after stimulation (30). Hence, we believe that the
regulation of prohibitin by the IgM receptor is a distinct event which
occurs early in the signaling cascade. It may be that
prohibitin-mediated repression of E2F is released first in response to
IgM signaling and, after a significant lapse of time, the
p107-E2F complex is disrupted by secondary events taking place in
the cell.
The release of prohibitin-mediated repression of E2F activity
apparently involves the MAP kinase pathway, since the chemical inhibitor of MEK1 could inhibit this; we are not sure whether the
concentration of the inhibitor used can affect Raf-1 activity. It has
been reported that IgM receptor stimulation leads to an activation of
Raf-1 kinase. Interestingly, IgM stimulation cannot affect a truncated
prohibitin protein that could not respond to Raf-1. Based on these
observations, it is tempting to speculate that IgM-mediated
inactivation of prohibitin occurs through the recruitment of the Raf-1
protein, but so far we have not found an increased binding of Raf-1 to
prohibitin upon IgM stimulation. We had observed earlier that serum
stimulation of human diploid fibroblasts leads to the nuclear
translocation of a small fraction of Raf-1, allowing it to colocalize
with Rb (59). Hence, although it may be that a similar event
is taking place upon IgM stimulation, it is not yet clear whether Raf-1
has to physically interact with prohibitin to bring about this effect.
This has been difficult to determine, since the prohibitin-binding
domain of Raf-1 is immediately adjacent to its kinase domain, and
mutations of this region might alter the kinase activity of Raf-1. The
actual mechanisms of prohibitin inactivation remain unclear, since in
the preliminary in vitro kinase reaction prohibitin could not be
phosphorylated either by Raf-1 or ERK2.
One of the intriguing aspects of prohibitin-mediated repression of E2F
is that the inhibitory effects of prohibitin do not respond to many
agents that reverse Rb-mediated repression of E2F. A case in point that
is notable is the adenovirus E1A protein. Similarly, cyclins D and E,
which inactivate the Rb protein by phosphorylation, had no effect on
prohibitin-mediated repression of E2F1. This was observed in direct
overexpression experiments, where cyclins D and E could reverse Rb, but
not prohibitin, function. Further, we find that though
prohibitin-mediated repression of E2F1 is negated upon prolonged serum
stimulation of quiescent T47D cells, this reversal could not be blocked
by dominant-negative cyclin-dependent kinases. This observation implies
that the inactivation of prohibitin during cell cycle progression is
mediated by a different set of kinases or by other means that are
independent of phosphorylation. Similarly, while we find that
overexpression of Raf-1 can inactivate prohibitin, p38 kinase has no
effect; in contrast, both Raf-1 and p38 can target the Rb protein. This
scenario further supports the notion that prohibitin-mediated
regulation of E2F responds to a different set of signals than those
regulating Rb.
Prohibitin was originally cloned based on its ability to bring about a
very potent G1/S arrest (38, 42), and our
earlier results suggest that prohibitin brings about the growth arrest by repressing E2F activity. The results presented here suggest that
prohibitin is a novel regulator of E2F function and that prohibitin
facilitates the cell cycle machinery to respond to extracellular
signals that cannot target Rb and its family members. These findings,
we believe, will help us further understand the molecular mechanisms by
which prohibitin regulates cell proliferation, as well as how
extracellular signals contact the cell cycle machinery.
| |
ACKNOWLEDGMENTS |
We thank David Johnson for the kind gift of E2F1 fusion proteins
and for helpful suggestions and J. Keith McClung for anti-prohibitin antibodies.
This work was supported by a grant from the National Cancer Institute
(RO-1 CA63136). S.P.C. is a recipient of the Irma-Hirschl Trust
Research Award.
| |
FOOTNOTES |
*
Corresponding author. Mailing address: Department of
Pathology, College of Physicians and Surgeons, Columbia University,
630W 168th St., New York, NY 10032. Phone: (212) 305-3736. Fax: (212) 305-5498. E-mail: spc10{at}columbia.edu.
| |
REFERENCES |
| 1.
|
Adams, P. D., and W. J. Kaelin.
1996.
The cellular effects of E2F overexpression.
Curr. Top. Microbiol. Immunol.
208:79-93[Medline].
|
| 2.
|
Barnard, D.,
B. Diaz,
D. Clawson, and M. Marshall.
1998.
Oncogenes, growth factors and phorbol esters regulate Raf-1 through common mechanisms.
Oncogene
17:1539-1547[Medline].
|
| 3.
|
Beijersbergen, R. L., and R. Bernards.
1996.
Cell cycle regulation by the retinoblastoma family of growth inhibitory proteins.
Biochim. Biophys. Acta
1287:103-120[Medline].
|
| 4.
|
Berger, K. H., and M. P. Yaffe.
1998.
Prohibitin family members interact genetically with mitochondrial inheritance components in Saccharomyces cerevisiae.
Mol. Cell. Biol.
18:4043-4052[Abstract/Free Full Text].
|
| 5.
|
Brehm, A.,
E. A. Miska,
D. J. McCance,
J. L. Reid,
A. J. Bannister, and T. Kouzarides.
1998.
Retinoblastoma protein recruits histone deacetylase to repress transcription.
Nature
391:597-601[Medline].
|
| 6.
|
Bremner, R.,
B. L. Cohen,
M. Sopta,
P. A. Hamel,
C. J. Ingles,
B. L. Gallie, and R. A. Phillips.
1995.
Direct transcriptional repression by pRB and its reversal by specific cyclins.
Mol. Cell. Biol.
15:3256-3265[Abstract].
|
| 7.
|
Cartwright, P.,
H. Muller,
C. Wagener,
K. Holm, and K. Helin.
1998.
E2F-6: a novel member of the E2F family is an inhibitor of E2F-dependent transcription.
Oncogene
17:611-623[Medline].
|
| 8.
|
Chellappan, S. P.,
S. Hiebert,
M. Mudryj,
J. M. Horowitz, and J. R. Nevins.
1991.
The E2F transcription factor is a cellular target for the RB protein.
Cell
65:1053-1061[Medline].
|
| 9.
|
Coates, P. J.,
D. J. Jamieson,
K. Smart,
A. R. Prescott, and P. A. Hall.
1997.
The prohibitin family of mitochondrial proteins regulate replicative lifespan.
Curr. Biol.
7:607-610[Medline].
|
| 10.
|
Cobrinik, D.
1996.
Regulatory interactions among E2Fs and cell cycle control proteins.
Curr. Top. Microbiol. Immunol.
208:32-63.
|
| 11.
|
Cress, W. D., and J. R. Nevins.
1994.
Interacting domains of E2F1, DP1, and the adenovirus E4 protein.
J. Virol.
68:4213-4219[Abstract/Free Full Text].
|
| 12.
|
Cress, W. D., and J. R. Nevins.
1996.
Use of the E2F transcription factor by DNA tumor virus regulatory proteins.
Curr. Top. Microbiol. Immunol.
208:63-78[Medline].
|
| 13.
|
DeGregori, J.,
G. Leone,
A. Miron,
L. Jakoi, and J. R. Nevins.
1997.
Distinct roles for E2F proteins in cell growth control and apoptosis.
Proc. Natl. Acad. Sci. USA
94:7245-7250[Abstract/Free Full Text].
|
| 14.
|
DeGregori, J.,
G. Leone,
K. Ohtani,
A. Miron, and J. R. Nevins.
1995.
E2F-1 accumulation bypasses a G1 arrest resulting from the inhibition of G1 cyclin-dependent kinase activity.
Genes Dev.
9:2873-2887[Abstract/Free Full Text].
|
| 15.
|
Dyson, N.
1998.
The regulation of E2F by pRB-family proteins.
Genes Dev.
12:2245-2262[Free Full Text].
|
| 16.
|
Fabian, J. R.,
I. O. Daar, and D. K. Morrison.
1993.
Critical tyrosine residues regulate the enzymatic and biological activity of Raf-1 kinase.
Mol. Cell. Biol.
13:7170-7179[Abstract/Free Full Text].
|
| 17.
|
Fanger, G. R.,
P. Gerwins,
C. Widmann,
M. B. Jarpe, and G. L. Johnson.
1997.
MEKKs, GCKs, MLKs, PAKs, TAKs, and tpls: upstream regulators of the c-Jun amino-terminal kinases?
Curr. Opin. Genet. Dev.
7:67-74[Medline].
|
| 18.
|
Field, S. J.,
F. Y. Tsai,
F. Kuo,
A. M. Zubiaga,
W. J. Kaelin,
D. M. Livingston,
S. H. Orkin, and M. E. Greenberg.
1996.
E2F-1 functions in mice to promote apoptosis and suppress proliferation.
Cell
85:549-561[Medline].
|
| 19.
|
Gaubatz, S.,
J. G. Wood, and D. M. Livingston.
1998.
Unusual proliferation arrest and transcriptional control properties of a newly discovered E2F family member, E2F-6.
Proc. Natl. Acad. Sci. USA
95:9190-9195[Abstract/Free Full Text].
|
| 20.
|
Hamel, P. A.,
R. M. Gill,
R. A. Phillips, and B. L. Gallie.
1992.
Transcriptional repression of the E2-containing promoters EIIaE, c-myc, and RB1 by the product of the RB1 gene.
Mol. Cell. Biol.
12:3431-3438[Abstract/Free Full Text].
|
| 21.
|
Helin, K.,
J. A. Lees,
M. Vidal,
N. Dyson,
H. Harlow, and A. Fattaey.
1992.
A cDNA encoding a pRB-binding protein with properties of the transcription factor E2F.
Cell
70:337-350[Medline].
|
| 22.
|
Johnson, D. G.,
J. K. Schwarz,
W. D. Cress, and J. R. Nevins.
1993.
Expression of transcription factor E2F1 induces quiescent cells to enter S phase.
Nature
365:349-352[Medline].
|
| 23.
|
Johnson, D. G.,
K. Ohtani, and J. R. Nevins.
1994.
Autoregulatory control of E2F1 expression in response to positive and negative regulators of cell cycle progression.
Genes Dev.
8:1514-1525[Abstract/Free Full Text].
|
| 24.
|
Johnson, D. G., and R. Schneider-Broussard.
1998.
Role of E2F in cell cycle control and cancer.
Front. Biosci.
3:447-448.
|
| 25.
|
Kaelin, W. G.,
W. Krek,
W. R. Sellers,
J. A. DeCaprio,
F. Ajchenbaum,
C. S. Fuchs,
T. Chittenden,
Y. Li,
P. J. Farnham, and M. A. E. A. Blanar.
1992.
Expression cloning of a cDNA encoding a retinoblastoma-binding protein with E2F-like properties.
Cell
70:351-364[Medline].
|
| 26.
|
Kitagawa, M.,
H. Higashi,
T. I. Suzuki,
K. Segawa,
S. K. Hanks,
Y. Taya,
S. Nishimura, and A. Okuyama.
1995.
Phosphorylation of E2F-1 by cyclin A-cdk2.
Oncogene
10:229-236[Medline].
|
| 27.
|
Kowalik, T. F.,
J. DeGregori,
G. Leone,
L. Jakoi, and J. R. Nevins.
1998.
E2F1-specific induction of apoptosis and p53 accumulation, which is blocked by Mdm2.
Cell Growth Diff |
Top
Abstract
Introduction
