Oncogene (2000) 19, 1857 ± 1867
ã 2000 Macmillan Publishers Ltd All rights reserved 0950 ± 9232/00 $15.00
www.nature.com/onc
Restoration of retinoblastoma mediated signaling to Cdk2 results in cell
cycle arrest
Matthew W Strobeck1, Anne F Fribourg1, Alvaro Puga2 and Erik S Knudsen*,1
1
Department of Cell Biology, University of Cincinnati College of Medicine, Cincinnati, Ohio, OH 45267-0521, USA; 2Department
of Environmental Health, University of Cincinnati College of Medicine, Cincinnati, Ohio, OH 45267-0521, USA
Phosphorylation/inactivation of RB is typically required
for cell cycle progression. However, we have identi®ed a
tumor cell line, C33A, which progresses through the cell
cycle in the presence of an active allele of RB (PSMRB). To determine how C33A cells evade RB-mediated
arrest, we compared RB signaling to downstream
e€ectors in this resistant cell line to that of the RBsensitive SAOS-2 cell line. Although introduction of
PSM-RB repressed E2F-mediated transcription in both
C33A and SAOS-2 cells, PSM-RB failed to repress
Cyclin A promoter activity in C33A. Ectopic expression
of PSM-RB in SAOS-2 cells resulted in a decrease in
both Cyclin A and Cdk2 protein levels without a€ecting
Cyclin E or Cdk4. In contrast, over-expression of PSMRB in C33A cells did not alter endogenous Cyclin A,
Cyclin E, or Cdk2 protein levels or impact Cdk2 kinase
activity, indicating that signaling from RB to downstream targets is abrogated in this cell line. The
importance of Cdk2 activity was demonstrated by
p27Kip1, which attenuated Cdk2 activity and inhibited
cell cycle progression in C33A cells. Since RB signaling
to Cdk2 is disrupted in these tumor cells, we coexpressed two proteins that cooperate with RB in
transcriptional repression, AHR and BRG-1, in an
attempt to correct this signaling dysfunction. Coexpression of AHR/BRG-1 with PSM-RB attenuated
Cyclin A and Cdk2 expression as well as Cdk2associated kinase activity, resulting in cell cycle
inhibition of C33A cells. Importantly, ectopic expression
of Cyclin A was able to reverse the arrest mediated by
co-expression of AHR/BRG-1 with PSM-RB. These
results indicate that down-regulation of Cdk2 activity is
requisite for RB-mediated cell cycle arrest. Thus, this
study reveals a new mechanism through which tumor
cells evade anti-proliferative signals, and provides insight
into how RB-signaling is mediated. Oncogene (2000) 19,
1857 ± 1867.
Keywords: AHR; BRG-1; Cdk; Cyclins; G1
Introduction
The retinoblastoma protein (RB) is the prototypical
tumor suppressor (Bartek et al., 1997; Kaelin, 1997;
Sherr, 1996; Wang et al., 1994). The importance of RBdependent anti-proliferative activity is emphasized by
the observation that RB is inactivated in greater than
60% of human tumors, through both direct and
*Correspondence: ES Knudsen
Received 6 October 1999; revised 28 January 2000; accepted 3
February 2000
indirect mechanisms (Kaelin, 1997; Sellers and Kaelin,
1997; Sherr, 1996). RB-mediated tumor suppression is
believed to involve the cell-cycle inhibitory action of
RB. Through direct binding, RB assembles speci®c
protein complexes that act to inhibit cell cycle
transitions. For example, the E2F family of transcription factors, which activate genes responsible for
initiation of S phase, are critical downstream targets
of RB (Dyson, 1998; Weinberg, 1995). Once bound to
E2F, RB recruits histone deacetylase (HDAC), thus
converting E2F from a transcriptional activator to a
transcriptional repressor (Brehm et al., 1998; DePhino,
1998; Magnaghi-Jaulin et al., 1998; Sidle et al., 1996;
Wang et al., 1994; Weinberg, 1995; Xiao et al., 1995;
Zhang et al., 1999). Since the ability of RB to interact
with speci®c cellular proteins is vital to its role as a cell
cycle inhibitor, it is not surprising that disruption of
RB protein binding capabilities is a common event in
human tumors. Speci®cally, it has been shown that RB
is functionally inactivated via: (i) mutation at the
chromosomal level; (ii) overt phosphorylation due to
the deregulated expression of proteins that phosphorylate RB; and (iii) inactivation by the direct binding of
oncoproteins of DNA tumor viruses (e.g. E7 of the
high risk human papilloma virus (HPV)) (Bartek et al.,
1997; Bartkova et al., 1997; Kaelin, 1997; Reed, 1997;
Sellers and Kaelin, 1997; Weinberg, 1995). In theory,
tumor cells could also abrogate the signaling pathway
to downstream e€ectors of RB as a mechanism to
evade RB-mediated growth arrest.
Cell cycle progression under normal growth conditions requires that RB becomes inactivated by Cyclin
dependent kinase (Cdk) catalyzed phosphorylation,
which occurs in response to mitogenic stimuli (Bartek
et al., 1997; Bartkova et al., 1997; Sherr, 1996).
Mitogens stimulate Cyclin D/Cdk4 activity, which then
initiates phosphorylation and subsequent inactivation
of RB. It is believed that the Cyclin D/Cdk4 complex
acts principally to phosphorylate RB, since the Cyclin
D/Cdk4 inhibitor, p16ink4a, or injection of anti-Cyclin
D antibodies does not inhibit cell cycle progression in
RB de®cient cells (Koh et al., 1995; Lukas et al.,
1995a,b; Medema et al., 1995). In addition to Cyclin
D/Cdk4, Cdk2 activity is implicated in phosphorylating
and inactivating RB in late G1 (Bartek et al., 1997).
Cdk2 associates with both Cyclin A and Cyclin E and
it has been shown that these two complexes are both
required and rate limiting for progression into S phase
(Ohtsubo et al., 1995; Resnitzky et al., 1994).
Phosphorylation inactivates RB, via disruption of its
protein binding abilities. Thus, E2F is derepressed and
activates the transcription of a variety of genes
responsible for S phase (DeGregori et al., 1997; Dyson,
1998; Nevins et al., 1997).
Cdk2 is a functional target of RB
MW Strobeck et al
1858
Oncogene
Reintroduction of wild-type RB into various tumorigenic and non-tumorigenic cells is ine€ective at
inhibiting cellular proliferation, since the protein is
rapidly inactivated by endogenous Cdk activity
(Knudsen and Wang, 1997; Knudsen et al., 1998;
Lukas et al., 1997). To circumvent RB-phosphorylation
and thereby study the in¯uence of active RB, we
generated a phosphorylation site mutant of RB (PSMRB) that renders it resistant to Cdk-mediated inactivation (Knudsen and Wang, 1997). This constitutively
active mutant of RB retains its ability to bind proteins
like the transcription factor E2F, regardless of
phosphorylation status (Knudsen and Wang, 1997).
In previous studies, it has been shown that expression
of PSM-RB invokes cell cycle arrest in almost all cell
lines, including those which are not in¯uenced by wildtype RB (Knudsen and Wang, 1997; Knudsen et al.,
1998; Lukas et al., 1997, 1999). In PSM-RB arrested
cells, both Cyclin A expression and Cdk2 associated
kinase activity are attenuated (Alevizopoulos et al.,
1997; Knudsen et al., 1998, 1999a; Lukas et al., 1999).
Interestingly, Cyclin E expression is never attenuated in
such arrested cells, which is surprising since Cyclin E is
an E2F-regulated gene (Alevizopoulos et al., 1997;
Knudsen et al., 1998, 1999a; Lukas et al., 1999). The
down-regulation of Cyclin A is of functional importance, as ectopic expression of Cyclin A can overcome
PSM-RB induced cell cycle arrest (Knudsen et al.,
1999a; Meraldi et al., 1999).
The idea that phosphorylation of RB is an absolute
requirement for cell cycle progression was challenged,
however, by the recent ®nding that expression of
constitutively active RB does not universally inhibit cell
cycle progression (Knudsen et al., 1999b). We have
previously shown that ectopic expression of PSM-RB
does not inhibit cell cycle progression in the RB-
de®cient, HPV negative, cervical carcinoma cell line,
C33A (Knudsen et al., 1999b; Sche€ner et al., 1991).
We hypothesized that a novel mechanism must exist
that renders these cells resistant to RB-mediated cell
cycle arrest. One likely hypothesis was that RB
signaling to downstream targets is abrogated in C33A
cells. Such a signaling defect could represent an
additional mechanism through which RB is inactivated
during tumorigenesis. Unlike p16ink4a loss or Cyclin
D1 over-expression which act upstream of RB, such a
defect would e€ectively inactivate RB by uncoupling it
from its downstream signaling pathway. This idea is
supported by the prior ®ndings that certain proteins
act in cooperation with RB, indicating that loss of such
proteins may compromise RB-signaling. For example,
it has been reported that the brahma-related gene
(BRG-1) can enhance RB-mediated transcriptional
repression and perhaps co-mediate RB-induced cell
cycle arrest (Dunaief et al., 1994; Kadonaga, 1998;
Kingston et al., 1996; Strober et al., 1996; Trouche et
al., 1997). Similarly, the aromatic hydrocarbon receptor (AHR) can also enhance the function of RB as a
transcriptional repressor and synergize with RB to
induce cell cycle arrest (Ge and Elferink, 1998; Puga et
al., 2000).
In this report, we investigated the mechanism
through which tumor cells evade PSM-RB mediated
cell cycle inhibition. We found that expression of PSMRB in both C33A cells, which are resistant to RB, and
in SAOS-2 (RB-de®cient osteogenic sarcoma) cells, that
are arrested by RB, E2F transcriptional activity was
repressed. However, PSM-RB failed to down-regulate
Cyclin A promoter activity only in C33A cells. In
addition, ectopic expression of active RB in SAOS-2
cells resulted in the down regulation of endogenous
Cyclin A and Cdk2 protein while there was no change
Cdk2 is a functional target of RB
MW Strobeck et al
in the expression of these downstream targets in C33A
cells. We show that Cdk2 activity is required for cell
cycle progression using the Cdk-inhibitor p27Kip1,
suggesting that if RB signaling to Cdk2 could be
restored, C33A cells should arrest. Strikingly, coexpression of PSM-RB with AHR and BRG-1 caused
the attenuation of Cyclin A and Cdk2 expression along
with a reduction in Cdk2 kinase activity resulting in
cell cycle arrest. This arrest could be partially alleviated
through ectopic co-expression of Cyclin A, suggesting
that the down-regulation of Cdk2 activity is causative
for the cell cycle inhibition. These data provide the ®rst
evidence that disruption of RB signaling to downstream e€ectors occurs in human tumor cells, and
provides a unique system to probe the mechanism
through which RB signals to mediate cell cycle arrest.
Results
C33A cells are resistant to constitutively active RB
although E2F-activity is repressed
In the majority of cell lines studied, ectopic expression
of PSM-RB inhibits cell cycle progression (Knudsen et
al., 1998, 1999b; Lukas et al., 1997). This ®nding is
consistent with the idea that phosphorylation of RB is
required for transition through the cell cycle. However,
in the RB-de®cient C33A cervical carcinoma cell line,
over-expression of PSM-RB does not inhibit cell cycle
progression (Knudsen et al., 1999b). Consistent with
this observation, when PSM-RB and H2B-GFP were
co-transfected into the C33A cervical carcinoma cell
line, the cells proved to be resistant to the constitutively active allele of RB, as evidenced by their ability
to eciently incorporate BrdU (Figure 1a). The per
cent of PSM-RB transfected cells incorporating BrdU
was similar to that of vector transfected (GFP positive)
and untransfected (GFP negative) cells (Figure 1a).
Consistent with the BrdU data, analysis of the cell
cycle distribution by ¯ow cytometry revealed that
PSM-RB had no signi®cant e€ect on the cell cycle
distribution (data not shown). However, in the SAOS-2
cell line, over-expression of PSM-RB was sucient to
cause the inhibition of BrdU incorporation (Figure 1b),
arresting these cells in G1 (data not shown). These
results indicate that the C33A tumor cell line does not
require the phosphorylation of RB for cell cycle
progression. As such, C33A cells provide a unique
system with which to study the e€ects of RB on cell
1859
Figure 1 The C33A tumor cell line bypasses RB mediated cell cycle control. (a) C33A and (b) SAOS-2 cells were co-transfected
with an H2B-GFP-expression plasmid (0.125 mg) and vector (3.875 mg) or PSM-RB (3.875 mg). The percentage of BrdU positive
cells was determined from at least two independent experiments with at least 150 transfected (GFP positive) cells and 150
untransfected (GFP negative cells) scored per experiment. Photographs of the immuno¯uorescent cells were taken at 636
magni®cation and the arrows denote GFP positive (transfected) cells. (c, d) SAOS-2 and C33A cells were co-transfected with CMVbgal (1 mg) and luciferase reporter plasmids containing either 3XE2F (1 mg), or DHFR (1 mg) promoters, and the indicated
expression plasmids: PSM-RB (3 mg), or p27kip1 (3 mg). All transfections were brought to 8 mg total plasmid DNA with the
parental vector (CMVNeoBam). Data shown are from at least three independent experiments and were normalized to b-gal activity
for transfection eciency. Relative luciferase activity was set to 100%
Oncogene
Cdk2 is a functional target of RB
MW Strobeck et al
1860
cycle and provide insight into how tumor cells can
speci®cally evade the anti-proliferative e€ects of RB.
Since it is well established that dephosphorylated RB
can bind to E2F and repress transcription of E2F
target genes, we examined whether expression of PSMRB in C33A was defective in signaling to an E2F
dependent promoter. Previous studies have shown that
in cells arrested by PSM-RB, E2F transcriptional
activity is repressed (Knudsen et al., 1999a,b; Lukas
et al., 1997). To test this in both RB-resistant cells
(C33A) and RB-sensitive cells (SAOS-2) the 3XE2Fluc
reporter was utilized, which harbors three binding sites
for the E2F transcription factor (Whittaker et al.,
1998). When PSM-RB was transfected into C33A and
SAOS-2 cells, transcription from the E2F dependent
promoter was repressed by 77% (Figure 1c) and 73%
(Figure 1d) respectively. In C33A, ectopic expression of
the Cdk-inhibitor p27kip1 had no e€ect on 3XE2F-Luc
activity (Figure 1c). To extend these ®ndings, the
dihydrofolate reductase promoter (DHFRLuc) was
utilized which is a physiological E2F-dependent
promoter. These experiments showed that expression
of PSM-RB in both C33A and SAOS-2 cells was also
sucient to inhibit DHFR-promoter activity (Figure
1c,d). Expression of p27kip1 in C33A cells had no
e€ect on DHFRLuc and had values similar to that of
vector (Figure 1c). Thus in the RB-resistant C33A cell
line, RB is still e€ective in inhibiting E2F dependent
promoter activity, but this signaling is not sucient for
halting cell cycle progression.
PSM-RB fails to inhibit Cdk2 activity in C33A cells
In all cell lines studied, dephosphorylation of RB
results in reduced Cdk2 activity and reduced Cyclin A
promoter activity (Alevizopoulos et al., 1997; Knudsen
et al., 1998; Lukas et al., 1999). To determine if
activated RB expressed in C33A retains these capabilities, we co-transfected the PSM-RB expression
plasmid with a human Cyclin A reporter (-608cycALuc) plasmid. In contrast to what we previously
observed in PSM-RB-responsive cell lines (Knudsen
et al., 1999a), transfection of PSM-RB into C33A cells
had no e€ect on Cyclin A promoter activity, while
p27kip1 was capable of inhibiting Cyclin A dependent
promoter activity (Figure 2a). This result is consistent
with the prior ®nding that p27kip1 can bind to and
directly inhibit Cyclin A promoter activity (Shulze et
al., 1996). In SAOS-2 cells, ectopic expression of PSMRB was e€ective at inhibiting Cyclin A dependent
promoter activity when compared to vector (Figure
2b). Therefore, in C33A cells there is a speci®c defect in
transcriptional repression of Cyclin A, but no general
defect in E2F activity.
To investigate the in¯uence of RB on endogenous
Cdk/Cyclins; C33A and SAOS-2 cells were cotransfected with vector, PSM-RB, or p27kip1 along
with a puromycin resistance plasmid (pBABE-puro).
After puromycin selection the cells were harvested and
subjected to immunoblot analysis. In SAOS-2 cells,
ectopic expression of PSM-RB caused a decrease in
both endogenous Cyclin A and Cdk2 protein while not
a€ecting Cyclin E or Cdk4 (Figure 2, lanes 1 and 2).
As shown in Figure 2d, expression of PSM-RB in
C33A cells had no e€ect on the levels of endogenous
Cyclin A protein, which was similar to vector (lanes 2
Oncogene
and 3). However, transfection of p27kip1 was sucient
to inhibit Cyclin A protein expression (Figure 2d, lane
1). C33A cells over expressing vector, PSM-RB or
p27kip1 did not demonstrate altered Cdk4 or Cdk2
protein expression. Interestingly, p27kip1 caused an
increase in endogenous Cyclin E protein while vector
and PSM-RB did not cause this augmentation (Figure
2d, lane 1). The ability of p27kip1 to increase Cyclin E
is consistent with previous ®ndings (Ciaparrone et al.,
1998; Won and Reed, 1996). Lastly, to determine if RB
in¯uenced Cdk2 kinase activity, in vitro kinase
reactions were performed. Cdk2 was immunoprecipitated from vector, PSM-RB and p27kip1 transfected
C33A cells that had undergone puromycin selection.
The Cdk2 complex immunoprecipitated from vector
and PSM-RB transfected cell was active as evidenced
by its ability to phosphorylate a histone H1 substrate
(Figure 2e, lanes 2 and 4). Cdk2 immunoprecipitated
from vector and PSM-RB transfected cells has
approximately equal activity when normalized to
Cdk2 protein (Figure 2e). However, Cdk2 immunoprecipitated from p27kip1 expressing cells was inactive, as
demonstrated by an inability to eciently phosphorylate histone H1 (Figure 2e, lane 3). These data suggest
that RB signaling to downstream targets such as Cyclin
A or Cdk2 is disrupted in C33A cells.
Cdk2 kinase activity is required for cell cycle progression
in C33A cells
Since RB does not downregulate Cdk2 kinase activity
in C33A cells, we sought to determine if this activity is
vital for cell cycle progression. Since p27kip1 is
e€ective at inhibiting Cyclin A as well as Cdk2 kinase
activity, we hypothesized that p27kip1 over-expression
would inhibit cell cycle progression. To test this
hypothesis, C33A cells were co-transfected with H2BGFP and vector, PSM-RB, or p27kip1 and scored for
BrdU incorporation. Interestingly, over expression of
p27kip1 was e€ective at reducing BrdU incorporation
to 25.3% when compared with vector alone (76.8%) or
untransfected (GFP negative) cells (Figure 3a,b). Thus,
down regulation of Cdk2 is sucient to inhibit cell
cycle progression and these data suggest that the
inability of RB to signal to Cdk2 contributes to the
ability of C33A cells to evade RB-mediated cell cycle
arrest.
AHR and BRG-1 restore RB signaling to Cdk2 and
Cyclin A
The data presented indicate that RB fails to arrest
C33A cells because a disruption exists in the signal
from RB to Cdk2. In an attempt to reverse this
dysfunction, we employed two proteins that are known
to cooperate with RB, AHR and BRG-1. Since C33A
cells have reduced BRG-1 expression (Dunaief et al.,
1994) and lack endogenous AHR protein (data not
shown), loss of such cooperating proteins might be one
mechanism by which these cells confer resistance to
PSM-RB. Therefore, we tested whether either of these
two proteins may be able to restore RB-mediated
signaling in C33A cells. To do so, C33A cells were
transfected with vector, AHR, BRG-1, AHR+PSMRB (3 : 1 ratio) or BRG-1+PSM-RB (1 : 1 ratio) along
with a puromycin resistance plasmid. AHR was
Cdk2 is a functional target of RB
MW Strobeck et al
transfected at a 3 : 1 ratio with PSM-RB based on the
report of Puga et al. (2000). who showed this to be the
optimal concentration for cooperation in the arrest in
SAOS-2 cells. Co-expression of AHR+PSM-RB or
BRG-1+PSM-RB decreased endogenous Cyclin A and
Cdk2 expression but not Cyclin E or Cdk4 (Figure 4a,
lane 3 and Figure 4b, lane 3). Expression of AHR,
BRG-1 or vector alone had no e€ect on Cyclin E,
Cyclin A, Cdk2 or Cdk4 protein expression (Figure 4a
lanes 1 and 2 and Figure 4b lanes 1 and 2). To
demonstrate that the e€ects observed upon co-expression of AHR/BRG-1 with PSM-RB could not be
attributed to the ability of AHR or BRG-1 to increase
PSM-RB expression, PSM-RB protein levels were
monitored. As shown in Figure 4c, co-expression of
AHR/BRG-1 with PSM-RB caused no change in PSMRB protein levels, suggesting that the reduction in
Cyclin A and Cdk2 protein was due to the cooperation
1861
Figure 2 RB is ine€ective in signaling to Cyclin A in C33A cells. C33A and SAOS-2 cells were transfected with 0.5 mg of the
Cyclin A promoter reporter plasmid -608Luc, 1 mg of CMV-bgal, and the following expression plasmids: p27kip1 (3 mg) or PSM-RB
(3 mg) as indicated. All transfections were brought to 8 mg total plasmid DNA using parental vector (CMVNeoBam). Data shown
are from four independent experiments and were normalized to b-gal activity for transfection eciency (a, b). (c) SAOS-2 cells were
transfected with (1 mg) pBabe-Puro along with (15 mg) CMVNeoBam (lane 1) and (15 mg) PSM-RB (lane 2). (d) C33A cells were
transfected with (1 mg) pBabe-Puro along with (15 mg) CMVNeoBam (lane 1), (15 mg) PSM-RB (lane 2) and (15 mg) p27kip1 (lane
1). (c, d) Following puromycin selection, equal total protein was resolved by SDS ± PAGE, and then immunoblotted for Cyclin A,
Cyclin E, Cdk2, Cdk4, pRB or p27kip1. (e) Using the same transfection and selection procedure used in (d), Cdk2 was
immunoprecipitated from the various transfected C33A cells and examined for its ability to phosphorylate histone H1 (upper panel).
Lower panel: Immunoblot of Cdk2 from the in vitro kinase reaction
Oncogene
Cdk2 is a functional target of RB
MW Strobeck et al
1862
Figure 3 p27kip1 inhibits cell cycle progression. (a) C33A cells were co-transfected with an H2B-GFP-expression plasmid
(0.125 mg) and either vector (3.875 mg), PSM-RB (3.875 mg), p16ink4a (3.875 mg), or p27kip1 (3.875 mg). The percentage of BrdU
positive cells was determined from at least two independent experiments with at least 150 GFP positive cells scored per experiment.
(b) Photographs were taken at 636 magni®cation and the arrows denote GFP positive cells
of AHR/BRG-1 with PSM-RB. Since introduction of
active RB in combination with either AHR or BRG-1
restored the ability of RB to inhibit Cyclin A
expression, we hypothesized that AHR and/or BRG-1
may restore RB signaling to inhibit Cdk2 kinase
activity. To test this, Cdk2 was immunoprecipitated
from cells that were co-transfected with a puromycin
resistance plasmid and either vector, AHR, BRG-1,
AHR+PSM-RB or BRG-1+PSM-RB and rapidly
selected. In vitro kinase assays demonstrated that
expression of AHR and BRG-1 alone was not able
to modulate Cdk2 kinase activity as compared to
vector alone (Figure 4d lanes 2 and 3 and 4e lanes 2
and 3). However, Cdk2 immunoprecipitated from
AHR+PSM-RB or BRG-1+PSM-RB transfected cells
was signi®cantly inhibited in its ability to phosphorylate histone H1 (Figure 4d lane 4 and Figure 4e lane
4). Therefore, AHR and BRG-1 cooperate with RB to
inhibit Cdk2. To determine if AHR or BRG-1 could
restore PSM-RB signaling to the Cyclin A promoter,
C33A cells were co-transfected with the human Cyclin
A reporter construct, 608cycALuc and AHR or BRG-1
with PSM-RB. Consistent with the decrease in
endogenous Cyclin A protein noted in Figure 4a,b,
AHR/BRG-1 also facilitated the ability of PSM-RB to
attenuate Cyclin A promoter activity (Figure 4f). Thus,
AHR and BRG-1 can restore RB signaling to
attenuate Cyclin A promoter activity and endogenous
Cyclin A/Cdk2 protein.
AHR and BRG-1 cooperate with RB to halt progression
from G1-S
To determine whether active RB mediated reduction of
Cdk2 activity is important for cell cycle inhibition,
C33A cells were co-transfected with H2B-GFP and
either vector, AHR, BRG-1, PSM-RB, AHR+PSMRB (3 : 1 ratio) or BRG-1+PSM-RB (1 : 1 ratio) and
scored for BrdU incorporation. Cells transfected with
vector, AHR, BRG-1 and PSM-RB incorporated BrdU
at approximately equal rates (Figure 5a). However, the
cells co-transfected with AHR+PSM-RB and BRG1+PSM-RB demonstrated a signi®cant decrease in
Oncogene
BrdU incorporation when compared to vector (Figure
5a,b). It should be noted that approximately 25% of
AHR+PSM-RB and BRG-1+PSM-RB transfected
cells maintain a sub-maximal level of BrdU incorporation (Figure 5c). Such punctate staining is consistent
with restoration of the previously described S phase
inhibitory action of RB and these cells were scored as
negative (Knudsen et al., 1998; Lukas et al., 1999).
Consistent with this e€ect, ¯ow cytometry data
revealed a decrease in 4N DNA content with a modest
increase in 2N (data not shown). Thus RB-mediated
cell cycle arrest can be achieved but only via
cooperation with AHR or BRG-1.
To determine whether Cyclin A was a functional
target in the AHR/BRG-1 and PSM-RB-mediated
arrest, Cyclin A was ectopically expressed in the
presence of AHR/BRG-1 and PSM-RB (Figure 6).
As shown above, the combination of PSM-RB with
AHR or BRG-1 was capable of signi®cantly inhibiting
BrdU incorporation. The ectopic expression of Cyclin
A was able to restore BrdU incorporation, indicating
that the down-regulation of Cyclin A contributes to the
observed cell cycle inhibition.
Discussion
Abrogation of RB signaling to downstream effectors
In this study, we describe how the C33A tumor cell line
evades RB-mediated cell cycle control. We had
previously observed that ectopic expression of PSMRB is not sucient to arrest C33A cells (Knudsen et
al., 1999b), and hypothesized that in this tumor line
RB has lost the ability to signal to downstream
e€ectors. E2F is a well-studied target of RB, and it is
thought that the ability of RB to repress E2Ftranscription is the underlying mechanism through
which RB inhibits entry into S phase (DeGregori et
al., 1997; Dyson, 1998; Nevins et al., 1997; Zhang et
al., 1999). Accordingly, we examined whether ectopic
expression of PSM-RB in C33A cells was capable of
signaling to E2F-dependent promoters. Interestingly,
Cdk2 is a functional target of RB
MW Strobeck et al
1863
Figure 4 AHR and BRG-1 cooperate with RB to inhibit Cyclin A expression. (a, b) C33A cells were co-transfected with 1 mg
pBabe-Puro and either (15 mg) CMVNeoBam (lane 1), (15 mg) AHR (a, lane 2), (12 mg) AHR+(4 mg) PSM-RB (a, lane 3), (15 mg)
BRG-1 (b, lane 2) or (7.5 mg) BRG-1+(7.5 mg) PSM-RB (b, lane 3) as indicated. Following puromycin selection, equal total protein
was resolved by SDS ± PAGE and then immunoblotted for Cyclin A, Cdk2, Cyclin E and Cdk4. (c) Immunoblot analysis of ectopic
PSM-RB expression was performed from lysates collected from C33A cells transfected with pBabe-Puro along with PSM-RB (lane
1), BRG-1 (lane 2), AHR+PSM-RB (lane 3) and BRG-1+PSM-RB (lane 4) following puromycin selection. (d, e) Using the same
transfection and selection procedure used in (a, b) Cdk2 was immunoprecipitated from the various transfected cells and examined
for its ability to phosphorylate histone H1. (f) C33A cells were transfected with 1 mg of CMV-bgal, 0.5 mg of -608Luc, and the
following expression plasmids: PSM-RB (3 mg), AHR (4.8 mg)+PSM-RB (1.6 mg), and BRG-1 (4.8 mg)+PSM-RB (1.6 mg) as
indicated. All transfections were brought to 8 mg total plasmid DNA using blank vector (CMVNeoBam). Data shown are from four
independent experiments and were normalized to b-gal activity for transfection eciency
expression of PSM-RB did result in down regulation of
two distinct E2F dependent promoter activities, 3XE2F
and DHFR. In fact, fold inhibition of the 3XE2F
promoter was approximately the same in cells which
failed to arrest by PSM-RB (C33A), and in cells which
were eciently cell cycle inhibited (SAOS-2). A
previous report has shown that wild-type RB is a
more potent repressor of E2F activity in SAOS-2 cells
rather than in C33A, but the basis behind this
di€erence is likely attributed to the fact that wild-type
RB is phosphorylated in C33A but not SAOS-2 cells
(Trouche et al., 1997; Harbour et al., 1999; Bremner et
al., 1995). Consistent with other reports, our data with
PSM-RB indicate that ectopic over-expression of active
Oncogene
Cdk2 is a functional target of RB
MW Strobeck et al
1864
Figure 5 AHR and BRG-1 cooperate with RB to induce cell cycle arrest. (a) C33A cells were co-transfected with a H2B-GFPexpression plasmid (0.125 mg) and vector (3.87 mg), PSM-RB (3.87 mg), AHR (3.875 mg), BRG-1 (3.875 mg), AHR (1 mg)+PSM-RB
(3 mg) or BRG-1 (2 mg)+PSM-RB (2 mg)+PSM-RB (2 mg). BrdU was added 48 h post-transfection and cells were ®xed following a
16-h label period. After staining with an anti-BrdU antibody (red) the cells were analysed by indirect immuno-¯uorescence. The
displayed values were determined from two independent experiments with at least 150 GFP positive cells scored per experiment. (b)
Photographs were taken at 636magni®cation and the arrows represent GFP positive cells. (c) These panels were further magni®ed
to show visualization of `punctate' staining
Figure 6 Ectopic expression of Cyclin A reverses the cell cycle
arrest induced by PSM-RB with AHR/BRG-1. C33A cells were
seeded onto glass coverslips and were co-transfected with a H2BGFP-expression plasmid (0.125 mg) and vector (3.875 mg), AHR
(1 mg)+PSM-RB (3 mg), BRG-1 (2 mg)+PSM-RB (2 mg), AHR
(0.75 mg)+PSM-RB (2.25 mg)+Cyclin A (1.5 mg), BRG-1
(1.3 mg)+PSM-RB (1.3 mg)+Cyclin A (1.3 mg). BrdU was added
48 h post-transfection and cells were ®xed following a 16-h label
period. After staining with an anti-BrdU antibody (red) the cells
were analysed by indirect immuno-¯uorescence. The displayed
values were determined from two or three independent experiments with at least 150 transfected (GFP positive) and
untransfected (GFP negative) cells scored per experiment
Oncogene
RB does eciently repress E2F-dependent transcription in C33A cells (Whittaker et al., 1998; Harbour et
al., 1999; Bremner et al., 1995). Therefore, in C33A
cells PSM-RB is able to signal to inhibit E2F promoter
activity but this inhibition is not sucient to halt cell
cycle progression.
Since E2F is known to regulate the expression of the
cell cycle regulatory genes Cyclin A, Cyclin E, and
Cdk2, we assessed whether over-expression of active
RB could alter the expression of these downstream
targets in SAOS-2 and C33A cells (Dyson, 1998;
Nevins et al., 1997). Initially, we found that in C33A
cells, ectopic expression of PSM-RB does not signal to
down-regulate Cyclin A promoter activity. This is in
contrast to what we observed in SAOS-2 cells, wherein
PSM-RB caused a signi®cant reduction in Cyclin A
promoter activity. Furthermore, in all settings where
RB-mediates cell cycle arrest, Cyclin A promoter
activity was inhibited (Phillips et al., 1998; Knudsen
et al., 1999a; Lukas et al., 1999). These ®ndings suggest
that while RB eciently signals to E2F, it is
compromised for signaling to other transcriptional
targets in C33A cells. To verify these ®ndings and
examine additional targets of RB, we analysed the level
of endogenous proteins in C33A and SAOS-2 cells
following the expression of PSM-RB. In C33A cells
there was no detectable change in the level of Cdk4,
Cdk2 is a functional target of RB
MW Strobeck et al
Cdk2, Cyclin E, or Cyclin A proteins following PSMRB transfection. This contrasts with p27Kip1 expression in C33A cells (using the identical experimental
protocol), after which an upregulation of Cyclin E
protein levels and a signi®cant decrease in Cyclin A
protein levels was observed. These ®ndings are
consistent with known e€ects of p27Kip1 on Cyclin
expression, since ectopic expression of p27Kip1
stabilizes Cyclin E protein, thereby leading to enhanced
accumulation, while p27Kip1 binds to and directly
inhibits Cyclin A promoter activity (Won and Reed,
1996; Schulze et al., 1996). In contrast with either
PSM-RB or p27Kip1 expression in C33A cells,
expression of PSM-RB in SAOS-2 cells led to a
speci®c reduction in Cyclin A and Cdk2 protein levels.
The reduction of Cyclin A levels are consistent with
other instances of RB-mediated cell cycle inhibition
(Knudsen et al., 1998, 1999a; Alevizopoulos et al.,
1997; Lukas et al., 1999). Although Cdk2 is an E2F
regulated gene, Cdk2 protein levels have not been
previously reported to be down regulated by RB
(Knudsen et al., 1998; Lukas et al., 1999). However,
we have recently observed that Cdk2 expression is
inhibited in immortalized rat cells expressing PSM-RB,
albeit with later kinetics than Cyclin A (not shown).
Thus Cdk2 likely represents an additional RB target.
Together, the data presented here suggest that the
ability of RB to attenuate Cyclin A and Cdk2
expression can be separated from basic E2F-mediated
repression. Thus in SAOS-2 cells RB signals to E2F,
Cyclin A, and Cdk2 resulting in cell cycle arrest; while
in C33A cells RB signals to E2F, but not Cyclin A or
Cdk2 and these cells do not arrest.
Reconnecting RB to downstream effectors
RB carries out its growth suppressing function by
recruiting cellular proteins and assembling protein
complexes (Bartek et al., 1997; Kaelin, 1997; Sidle et
al., 1996; Wang et al., 1994). Among the greater than
50 identi®ed RB binding proteins, only BRG-1 and
AHR have been shown to be important for cooperating with RB to induce cell cycle arrest (Dunaief et al.,
1994; Puga et al., 2000; Strober et al., 1996). Strikingly,
we found that co-expression of PSM-RB with either
AHR or BRG-1 resulted in the reduction of Cyclin A
and Cdk2 expression but not Cyclin E or Cdk4
proteins in C33A cells. The ®nding that co-expression
of BRG-1 or AHR with PSM-RB enabled PSM-RB to
inhibit the expression/transcription of Cyclin A, is
consistent with what is observed in cell lines arrested
by PSM-RB (Alevizopoulos et al., 1997; Knudsen et
al., 1998, 1999a; Philips et al., 1998). Interestingly, coexpression of BRG-1 or AHR with PSM-RB also led
to RB-mediated inhibition of Cdk2 expression. These
®ndings are consistent with what was observed in
SAOS-2 cells, and suggest that the expression of BRG1 or AHR e€ectively converts C33A cells into cells
properly regulated by RB. It is interesting that in all
cells which are sensitive to RB mediated arrest, Cyclin
E expression is never attenuated, despite the fact that
Cyclin E is an E2F-regulated gene (Knudsen et al.,
1998, 1999; Lukas et al., 1999). As such, the
mechanism through which these speci®c E2F-regulated
genes are targeted for transcriptional repression by RB
remains unclear. However, our ®ndings would indicate
that the presence of RB alone does not mandate
repression, and at least the repression of Cyclin A can
be achieved via the co-expression of BRG-1 and AHR.
The mechanism by which AHR and BRG-1 allow
for this RB-dependent repression of speci®c targets is
yet unresolved. One possibility is that AHR or BRG-1
could enhance RB signaling by bringing RB to
promoters that are not accessible to E2F. Consistent
with this notion, Murphy et al. (1999) discovered a role
for BRG-1 in negatively regulating the c-fos promoter.
Speci®cally, it was shown in C33A cells that repression
of the c-fos gene by BRG-1 was RB-dependent but
E2F-independent (Murphy et al., 1999). Another
possibility is that AHR or BRG-1 can increase the
association between RB and E2F, making it a more
potent repressor of physiological E2F target genes.
Supporting this idea is the observation that both AHR
and BRG-1 enhance RB-mediated repression of E2F
transcription on synthetic reporters (Puga et al., 2000;
Trouche et al., 1997).
1865
Restoration of RB-mediated arrest
Regardless of RB status, Cdk2 kinase activity is
essential for cell cycle progression (Ohstubo et al.,
1995). It has also been shown in previous reports that a
RB-mediated G1 arrest results in the reduction of
Cdk2 activity (Knudsen et al., 1998, 1999a; Lukas et
al., 1999). However, signaling to reduce Cdk2 activity
is abrogated in C33A cells. Apparently the downregulation of Cdk2 is both sucient and required for
cell cycle arrest in C33A cells as indicated by the
observations that: (i) C33A cells do not undergo an
arrest or down regulate Cdk2 in the presence of active
RB; (ii) ectopic expression of p27kip1, which downregulates Cdk2 kinase activity, inhibits cell cycle
progression; and (iii) over-expression of Cyclin A is
sucient to overcome the arrest achieved by RB and
BRG-1/AHR co-expression. Together these results
show that connecting RB signaling to downstream
targets is required and sucient to induce a G1 arrest.
Thus, C33A cells provide a novel system in which to
study the mechanism through which RB must function
to induce cell cycle arrest, since RB phosphorylation in
these cells is not required for cell cycle progression. In
C33A cells the dependence of RB on AHR and BRG-1
co-expression to induce cell cycle arrest makes these
proteins possible targets for inactivation in cancer.
Since AHR is not expressed in SAOS-2 cells (Puga et
al., 2000) and these cells are eciently arrested by
PSM-RB, apparently AHR is not required for PSMRB mediated arrest. Importantly, AHR is a tool that
can be utilized to enhance RB-mediated transcriptional
repression and promote cell cycle inhibition in C33A
cells. Our results do not address the requirement for
BRG-1, but clearly show that BRG-1 is sucient to
cooperate with PSM-RB to induce cell cycle arrest in
normally resistant tumor cells. Interestingly, it has been
reported that loss of hSNF5/INI1, a gene that encodes
a BRG-1 like protein, occurs in malignant rhabdoid
tumors (Versteege et al., 1998). Although numerous
mechanisms act upstream to inactivate RB (i.e.
ampli®cations of Cyclin D1 or loss of p16ink4a) this
report shows that the failure of down-stream signaling
can also lead to functional inactivation of RB-mediated
cell cycle inhibition.
Oncogene
Cdk2 is a functional target of RB
MW Strobeck et al
1866
Materials and methods
Cells, plasmids and transfection
SAOS-2 and C33A cells were maintained in Dulbecco's
modi®ed Eagle's medium supplemented with 10% fetal
bovine serum (FBS), 100 units/ml penicillin-streptomycin
and 2 mM L-glutamine at 378C in 5% CO2. C33A and
SAOS-2 cells were passaged less than ten times. Plasmids
were transfected using the calcium phosphate method
previously described (Chen and Okayama, 1987). The
following plasmids: pH2B-GFP, pBabe-Puro, PSM-RB,
p27kip1, Cyclin A, 3XE2FLuc, DHFRLuc, and -608CycALuc have also been previously described (Knudsen and
Wang, 1997; Knudsen et al., 1998, 1999a). The aryl
hydrocarbon receptor (AHR) plasmid was provided by Dr
Alvaro Puga (University of Cincinnati), and the BRG-1
expression plasmid was supplied by Dr Stephen P Go€
(Columbia University).
BrdU incorporation
Approximately 56105 C33A cells and 2.56105 SAOS-2 cells
were seeded onto coverslips in each well of a 6-well dish.
Twenty-four hours later the cells were transfected with 4 mg
total DNA (as indicated). Forty-eight hours post-transfection
the cells were labeled with BrdU to detect DNA synthesis.
After 16 h of labeling the cells were ®xed in 3.7%
formaldehyde in PBS for 15 min and then permeabilized in
0.3% Triton X-100 in PBS for 15 min. The cells were then
probed with an anti-BrdU antibody (Accurate Scienti®c) at a
1 : 500 dilution in an Immuno¯uorescence bu€er (PBS, 5 mg/
ml BSA, 0.5% NP-40, 10 mM MgCl, 100 U/ml DNAse 1) for
1 h at 378C in a humidi®ed chamber. After washing the
coverslips with PBS the cells were incubated with a
rhodamine-conjugated secondary antibody in PBS supplemented with 5 mg/ml BSA and 0.5% NP-40, along with
Hoechst for visualization of the nuclei. After 1 h of
incubation the cells were washed in PBS and mounted onto
slides. At least 100 transfected (GFP positive) and untransfected (GFP negative) cells were analysed from two to
three independent experiments.
Immunoblotting and kinase reactions
Approximately 26106 C33A cells and 16106 SAOS-2 cells
were plated in 10-cm dishes 24 h before transfection.
Transfected C33A and SAOS-2 cells were selected posttransfection with 2.5 mg/ml puromycin (Sigma) for 48 h and
then harvested for protein analysis or in vitro kinase assay.
For immunoblotting, cells were trypsinized and then washed
twice with PBS. The cell pellets were resuspended in RIPA
bu€er containing protease inhibitors (10 mg/ml, 1,10 phenanthroline, 10 mg/ml aprotinin, 10 mg/ml leupeptin, and 1 mM
phenylmethylsulfonyl ¯uoride) and phosphatase inhibitors
(10 mM sodium ¯uoride, 0.1 mM sodium vanadate, and
60 mM b-glycerophosphate) for 15 min on ice. The lysates
were brie¯y sonicated and centrifuged at 20 000 g for 10 min
at 48C. The lysates were fractionated by SDS-polyacrylamide
gel electrophoresis and then transferred onto Immobilon-P
(Millipore). The membranes were then incubated with
primary polyclonal antibodies: anti-Cdk2 (clone H-298, Santa
Cruz), anti-Cyclin A (clone H-432, Santa Cruz), anti-Cyclin E
(clone C-19, Santa Cruz), anti-Cdk4 (clone H-22, Santa
Cruz), anti-p27kip1 (M-197, Santa Cruz), and anti-pRb (851,
polyclonal antibody) (Welch and Wang, 1993). The blots
were then incubated with horseradish peroxidase conjugated
goat anti-rabbit IgG (Biorad) for 1 h at room temperature.
The antibody-antigen complex was detected by enhanced
chemiluminescence procedure (ECL, Amersham Pharmacia
Biotech). The kinase reactions were performed as previously
described (Knudsen and Wang, 1996), using 50 ± 100 mg of
total cell lysate. The Cdk2 kinase complex was immunoprecipitated using anti-Cdk2 (clone H-298, Santa Cruz) antibodies and protein A-Sepharose (Pharmacia). The kinase
complexes were then used to phosphorylate a histone H1
substrate. Reactions were separated by SDS-polyacrylamide
gel electrophoresis and then transferred onto Immobilon-P
membranes (Millipore). Incorporation of [g-32P]ATP into
histone H1 was detected by autoradiography; blots were then
probed for Cdk2 using an anti-Cdk2 antibody (clone H-298,
Santa Cruz). For antibody detection, protein A conjugated
horseradish peroxidase (Biorad) was utilized at 1 : 2000 and
visualized by ECL.
Reporter assays
Approximately 56105 C33A cells and 2.56105 SAOS-2 cells
were plated into 6-cm dishes. Twenty-four hours later the
cells were transfected (as indicated) with 8 mg of total plasmid
DNA. Seventy-two hours post-transfection the cells were
harvested. The Luciferase Assay System (Promega) was used
to determine the luciferase activity of the transfected cells.
Transfection eciency was determined by b-galactosidase
activity and the relative luciferase activity shown was
determined after its normalization to b-galactosidase activity.
The data re¯ects the average of at least three independent
experiments.
Acknowledgments
We thank Drs Karen E Knudsen and Zvjezdana SeverChroneos for critical reading of the manuscript, and Dr
Jerry Lingrel for luminometer use. We are also grateful to
Dr Steven Go€ for supplying us with the BRG-1 plasmid.
This work was supported by grants from the NCI,
National Institutes of Health (RO1-CA82525-01) and the
Sidney Kimmel Cancer Foundation (to Dr ES Knudsen)
and by NIH ES06273 (to Dr A Puga). Matthew W
Strobeck is supported in part by a training grant from
the National Institutes of Health.
References
Alevizopoulos K, Vlach J, Hennecke S and Amati B. (1997).
EMBO J., 16, 5322 ± 5333.
Bartek J, Bartkova J and Lukas J. (1997). Exp. Cell Res.,
237, 1 ± 6.
Bartkova J, Lukas J and Bartek J. (1997). Prog. Cell Cycle
Res., 3, 211 ± 220.
Brehm A, Miska EA, McCance DJ, Reid JL, Bannister AJ
and Kouzarides T. (1998). Nature, 391, 597 ± 601.
Bremner R, Cohen BL, Sopta M, Hamel PA, Ingles CJ,
Gallie BL and Phillips RA. (1995). Mol. Cell. Biol., 15,
3256 ± 3265.
Chen C and Okayama H. (1987). Mol. Cell. Biol., 7, 2745 ±
2752.
Oncogene
Ciaparrone M, Yamamoto H, Yao Y, Sgambato A,
Cattoretti G, Tomita N, Monden T, Rotterdam H and
Weinstein IB. (1998). Cancer Res., 58, 114 ± 122.
DeGregori J, Leone G, Miron A, Jakoi L and Nevins JR.
(1997). Proc. Natl. Acad. Sci. USA, 94, 7245 ± 7250.
DePhino RA. (1998). Nature, 391, 535 ± 536.
Dunaief JL, Strober BE, Guha S, Khavari PA, Alin K,
Luban J, Begemann M, Crabtree GR and Go€ SP. (1994).
Cell, 79, 119 ± 130.
Dyson N. (1998). Genes Dev., 12, 2245 ± 2262.
Ge NL and Elferink CJ. (1998). J. Biol. Chem., 273, 22708 ±
22713.
Cdk2 is a functional target of RB
MW Strobeck et al
Harbour JW, Luo RX, Dei Santi A, Postigo AA and Dean
DC. (1999). Cell, 98, 859 ± 869.
Kadonaga JT. (1998). Cell, 92, 307 ± 313.
Kaelin Jr WG. (1997). Canc. Invest., 15, 243 ± 254.
Kingston RE, Bunker CA and Imbalzano AN. (1996). Genes
Dev., 10, 905 ± 920.
Knudsen ES and Wang JY. (1996). J. Biol. Chem., 271,
8313 ± 8320.
Knudsen ES and Wang JY. (1997). Mol. Cell. Biol., 17,
5771 ± 5783.
Knudsen ES, Buckmaster C, Chen TT, Feramisco JR and
Wang JY. (1998). Genes Dev., 12, 2278 ± 2292.
Knudsen KE, Fribourg AF, Strobeck MW, Blanchard JM
and Knudsen ES. (1999a). J. Biol. Chem., 274, 27632 ±
27641.
Knudsen KE, Weber E, Arden KC, Cavenee WK, Feramisco
JR and Knudsen ES. (1999b). Oncogene, 18, 5239 ± 5245.
Koh J, Enders GH, Dynlacht BD and Harlow E. (1995).
Nature, 375, 506 ± 510.
Lukas J, Bartkova J, Rohde M, Strauss M and Bartek J.
(1995a). Mol. Cell. Biol., 15, 2600 ± 2611.
Lukas J, Parry D, Aagaard L, Mann DJ, Bartkova J, Strauss
M, Peters G and Bartek J. (1995b). Nature, 375, 503 ± 506.
Lukas J, Herzinger T, Hansen K, Moroni MC, Resnitzky D,
Helin K, Reed SI and Bartek J. (1997). Genes Dev., 11,
1479 ± 1492.
Lukas J, Sorensen CS, Lukas C, Santoni-Rugiu E and Bartek
J. (1999). Oncogene, 18, 3930 ± 3935.
Magnaghi-Jaulin L, Groisman R, Naguibneva I, Robin P,
Lorain S, Le Villain JP, Troalen F, Trouche D and HarelBellan A. (1998). Nature, 391, 601 ± 605.
Medema RH, Herrera RE, Lam F and Weinberg RA. (1995).
Proc. Natl. Acad. Sci. USA, 92, 6289 ± 6293.
Meraldi P, Lukas J, Fry AM, Bartek J and Nigg EA. (1999).
Nature Cell Biol., 2, 88 ± 93.
Murphy DJ, Hardy S and Engel DA. (1999). Mol. Cell. Biol.,
19, 2724 ± 2733.
Nevins JR, Leone G, DeGregori J and Jakoi L. (1997). J.
Cell. Phys., 173, 233 ± 236.
Ohtsubo M, Theodoras AM, Schumacher J, Roberts JM and
Pagano M. (1995). Mol. Cell. Biol., 15, 2612 ± 2624.
Philips A, Huet X, Plet A, Le Cam L, Viae A and Blanchard
JM. (1998). Oncogene, 16, 1373 ± 1381.
Puga A, Barnes SJ, Chang CY, Knudsen ES, Maier MA and
Schumann BL. (2000). J. Biol. Chem., 275, 2943 ± 2950.
Reed SI. (1997). Canc. Surv., 29, 7 ± 23.
Resnitzky D, Gossen M, Bujard H and Reed SI. (1994). Mol.
Cell. Biol., 14, 1669 ± 1679.
Sche€ner M, Munger K, Byrne JC and Howley PM. (1991).
Proc. Natl. Acad. Sci. USA, 88, 5523 ± 5527.
Sellers WR and Kaelin Jr WG. (1997). J. Clin. Onc., 15,
3301 ± 3312.
Sherr CJ. (1996). Science, 274, 1672 ± 1677.
Shulze A, Zerfass-Thome K, Berges J, Middendorp S,
Jansen-Durr P and Henglein B. (1996). Mol. Cell. Biol.,
16, 4632 ± 4638.
Sidle A, Palaty C, Dirks P, Wiggan O, Kiess M, Gill RM,
Wong AK and Hamel PA. (1996). Crit. Rev. Biochem.
Molec. Biol., 31, 237 ± 271.
Strober BE, Dunaief JL, Guha S and Go€ SP. (1996). Mol.
Cell. Biol., 16, 1576 ± 1583.
Trouche D, Le Chalony C, Muchardt C, Yaniv M and
Kouzarides T. (1997). Proc. Natl. Acad. Sci. USA, 94,
11268 ± 11273.
Versteege I, Sevenet N, Lange J, Rousseau-Merck MF,
Ambros P, Handgretinger R, Aurias A and Delattre O.
(1998). Nature, 394, 203 ± 206.
Wang JY, Knudsen ES and Welch PJ. (1994). Adv. Canc.
Res., 64, 25 ± 85.
Weinberg RA. (1995). Cell, 81, 323 ± 330.
Welch PJ and Wang JY. (1993). Cell, 75, 779 ± 790.
Whittaker LL, Su H, Baskaran R, Knudsen ES and Wang
JY. (1998). Mol. Cell. Biol., 18, 4032 ± 4042.
Won KA and Reed SI. (1996). EMBO J., 15, 4182 ± 4193.
Xiao ZX, Chen J, Levine AJ, Modjtahedi N, Xing J, Sellers
WR and Livingston DM. (1995). Nature, 375, 694 ± 698.
Zhang HS, Postigo AA and Dean DC. (1999). Cell, 97, 53 ±
61.
1867
Oncogene