Identification of mutations that disrupt phosphorylation-dependent nuclear export of cyclin D1

Article (PDF Available)inOncogene 25(47):6291-303 · November 2006with36 Reads
DOI: 10.1038/sj.onc.1209644 · Source: PubMed
Abstract
Although cyclin D1 is overexpressed in a significant number of human cancers, overexpression alone is insufficient to promote tumorigenesis. In vitro studies have revealed that inhibition of cyclin D1 nuclear export unmasks its neoplastic potential. Cyclin D1 nuclear export depends upon phosphorylation of a C-terminal residue, threonine 286, (Thr-286) which in turn promotes association with the nuclear exportin, CRM1. Mutation of Thr-286 to a non-phosphorylatable residue results in a constitutively nuclear cyclin D1 protein with significantly increased oncogenic potential. To determine whether cyclin D1 is subject to mutations that inhibit its nuclear export in human cancer, we have sequenced exon 5 of cyclin D1 in primary esophageal carcinoma samples and in cell lines derived from esophageal cancer. Our work reveals that cyclin D1 is subject to mutations in primary human cancer. The mutations identified specifically disrupt phosphorylation of cyclin D1 at Thr-286, thereby enforcing nuclear accumulation of cyclin D1. Through characterization of these mutants, we also define an acidic residue within the C-terminus of cyclin D1 that is necessary for recognition and phosphorylation of cyclin D1 by glycogen synthase kinase-3 beta. Finally, through construction of compound mutants, we demonstrate that cell transformation by the cancer-derived cyclin D1 alleles correlates with their ability to associate with and activate CDK4. Our data reveal that cyclin D1 is subject to mutations in primary human cancer that specifically disrupt phosphorylation-dependent nuclear export of cyclin D1 and suggest that such mutations contribute to the genesis and progression of neoplastic growth.
8 Figures

Full-text (PDF)

Available from: John Alan Diehl
ORIGINAL ARTICLE
Identification of mutations that disrupt phosphorylation-dependent nuclear
export of cyclin D1
S Benzeno
1
,FLu
1
, M Guo
2
, O Barbash
1
, F Zhang
1,3
, JG Herman
2
, PS Klein
3
, A Rustgi
1,3
and
JA Diehl
1
1
Department of Cancer Biology, The Leonard and Madlyn Abramson Family Cancer Research Institute and Cancer Center,
University of Pennsylvania, Philadelphia, PA, USA;
2
Department of Hematology/Oncology, John Hopkins Oncology Center,
Baltimore, MD, USA and
3
Department of Medicine, University of Pennsylvania, Philadelphia, PA, USA
Although cyclin D1 is overexpressed in a significant
number of human cancers, overexpression alone is
insufficient to promote tumorigenesis. In vitro studies
have revealed that inhibition of cyclin D1 nuclear export
unmasks its neoplastic potential. Cyclin D1 nuclear
export depends upon phosphorylation of a C-terminal
residue, threonine 286, (Thr-286) which in turn promotes
association with the nuclear exportin, CRM1. Mutation
of Thr-286 to a non-phosphorylatable residue results in a
constitutively nuclear cyclin D1 protein with significantly
increased oncogenic potential. To determine whether
cyclin D1 is subject to mutations that inhibit its nuclear
export in human cancer, we have sequenced exon 5 of
cyclin D1 in primary esophageal carcinoma samples and
in cell lines derived from esophageal cancer. Our work
reveals that cyclin D1 is subject to mutations in primary
human cancer. The mutations identified specifically
disrupt phosphorylation of cyclin D1 at Thr-286, thereby
enforcing nuclear accumulation of cyclin D1. Through
characterization of these mutants, we also define an acidic
residue within the C-terminus of cyclin D1 that is
necessary for recognition and phosphorylation of cyclin
D1 by glycogen synthase kinase-3 beta. Finally, through
construction of compound mutants, we demonstrate that
cell transformation by the cancer-derived cyclin D1 alleles
correlates with their ability to associate with and activate
CDK4. Our data reveal that cyclin D1 is subject to
mutations in primary human cancer that specifically
disrupt phosphorylation-dependent nuclear export of
cyclin D1 and suggest that such mutations contribute to
the genesis and progression of neoplastic growth.
Oncogene (2006) 25, 6291–6303. doi:10.1038/sj.onc.1209644;
published online 29 May 2006
Keywords: cyclin D1; GSK-3b; CDK4; nuclear export;
cancer
Introduction
The cell cycle is a tightly regulated cellular clock
composed of four distinct phases: two gap phases, G1
and G2, separating the DNA synthetic (S) phase from
mitosis (M). Progression through each phase of the cell
cycle is driven by complexes minimally composed of a
regulatory, cyclin, and a catalytic, CDK, subunit.
Progression through G1 phase is initiated by mitogenic
stimulation, which in turn initiates the expression and
assembly of the D-type cyclins (D1, D2, D3) with their
cognate catalytic partners, CDK4/6.
The cyclin D/CDK4 holoenzyme has two functions
necessary for cell cycle progression. The first is the
initiation of phosphorylation-dependent inactivation of
the retinoblastoma family of proteins, Rb, p107 and
p130 (Hatakeyama et al., 1994; Harbour et al., 1999;
Calbo et al., 2002; Farkas et al., 2002; Leng et al., 2002),
which is subsequently completed by the cyclin E/CDK2
complex (Harbour et al., 1999). The second involves
the stoichiometric association of cyclin D/CDK4 with
Cip/Kip family proteins. This association facilitates
both cyclin D1/CDK4 activity through increased
nuclear retention and subunit assembly, while simulta-
neously preventing Cip/Kip access to cyclin E/CDK2
complexes (Cheng et al., 1998, 1999; Sherr and Roberts,
1999; Muraoka et al., 2001, 2002; Alt et al., 2002).
The D-type cyclins are unique in responding directly
to mitogenic signaling pathways rather than signaling
intrinsic to cell cycle progression. Expression of cyclin
D1 requires Ras-dependent activation of the Raf–Mek–
Erk kinase module (Albanese et al., 1995; Lavoie et al.,
1996; Winston et al., 1996; Aktas et al., 1997; Kerkhoff
and Rapp, 1997; Cheng et al., 1998). In addition,
maximal accumulation of cyclin D1 depends on
phosphatidylinositol 3
0
-kinase-dependent activation of
Akt and the concomitant inactivation of glycogen
synthase kinase-3 beta (GSK-3b) (Rodriguez-Viciana
et al., 1994; Kauffmann-Zeh et al., 1997). Glycogen
synthase kinase-3 beta-dependent phosphorylation of
cyclin D1 on threonine 286 (Thr-286) triggers the
nuclear export and cytoplasmic proteolysis of cyclin
D1 via the 26S proteasome (Diehl et al., 1998).
Perturbation of the cell cycle machinery is one of the
hallmarks of cancer (Collecchi et al., 2000). Cyclin D1,
Received 7 February 2006; revised 27 March 2006; accepted 27 March
2006; published online 29 May 2006
Correspondence: Dr JA Diehl, The Leonard and Madlyn Abramson
Family Cancer Research Institute and Cancer Center, University
of Pennsylvania, 454 BRB II/III, 421 Curie Blvd., Philadelphia,
PA 19104, USA.
E-mail: adiehl@mail.med.upenn.edu
Oncogene (2006) 25, 6291–6303
&
2006 Nature Publishing Group
All rights reserved 0950-9232/06 $30.00
www.nature.com/onc
the Bcl1 oncogene, is involved in the t(11;14)(q13;32)
chromosomal translocation associated with mantle cell
lymphomas (Lukas et al., 1994; Lovec et al., 1994a, b)
and is amplified in a subset of breast, bladder,
esophageal, lung and squamous cell carcinomas (Hall
and Peters, 1996; Diehl, 2002). Yet, despite its over-
expression in a number of malignancies, overexpression
of wild-type cyclin D1 is not by itself sufficient to induce
a transformed cellular phenotype in cell-based systems
(Quelle et al., 1993; Alt et al., 2000). In contrast, as we
have previously shown, expression of an artificially
engineered cyclin D1-T286A or a naturally occurring
alternative splice variant of cyclin D1 lacking the fifth
exon, cyclin D1b neither D1-T286A nor cyclin D1b is
phosphorylated by GSK-3b and thus, is refractory to
phosphorylation-dependent nuclear export has the
capacity to drive transformation of murine fibroblasts in
the absence of a collaborating oncogene (Alt et al., 2000;
Lu et al., 2003). Cumulatively, our results support a
model wherein dysregulation of cyclin D1 nuclear
export increases its oncogenic potential, emphasizing
the biological relevance of identifying cyclin D1 muta-
tions in human cancer that target residues essential for
these regulatory functions.
Little evidence exists for the identification of cyclin
D1 mutations in human cancer that specifically target its
phosphorylation at Thr-286 and/or nuclear export.
Herein, we report the identification of such cyclin D1
mutations in esophageal cancer. The mutations specifi-
cally disrupt cyclin D1 phosphorylation on Thr-286,
resulting in inhibition of S-phase specific nuclear export.
In addition, through structure–function analysis of
cancer-derived deletion mutants, we have identified
structural requirements in cyclin D1 necessary for its
recognition by GSK-3b that serve to coordinate GSK-
3b-mediated phosphorylation of cyclin D1 on Thr-286
(Benzeno and Diehl, 2004). Indeed, our data demon-
strate the occurrence of deregulated cyclin D1 nucleo-
cytoplasmic shuttling in human cancer.
Results
Identification of cyclin D1 mutations in human cancer
Overexpression of nuclear export-defective cyclin D1
mutants in murine fibroblasts or in transgenic mice
confers a neoplastic phenotype on expressing cells
demonstrating the distinctive oncogenic potential of
such cyclin D1 mutants (Alt et al., 2000; Lu et al., 2003;
Gladden et al., 2005). These data suggest that cyclin D1
may be subject to mutations in human cancer that
specifically disrupt its nuclear export. To address this
issue, we screened a panel of 90 primary human
esophageal tumor-derived patient DNA samples for
cyclin D1 mutations. We chose esophageal cancer owing
to the high prevalence of cyclin D1 overexpression (40–
60%) and its nuclear localization in esophageal adeno-
and squamous cell carcinomas, a striking percentage
of which are not associated with gene amplification
(Adelaide et al., 1995; Arber et al., 1996; Bani-Hani
et al., 2000; Lin and Beerm, 2004). Sequencing exon 5 of
the cyclin D1 gene in this set of tumor samples revealed
a threonine to arginine substitution at codon 286
(T286R) and a deletion encompassing codons 266–295
of cyclin D1 (D266–295). The deletion is reminiscent
of cyclin D1b, which is expressed as a consequence
of cancer-specific alternative splicing (Lu et al., 2003;
Solomon et al., 2003). Importantly, no such mutations
were identified in either matched normal or in 100
unmatched normal esophageal tissue samples. Together,
our results provide strong evidence that these cyclin D1
alterations are not a result of inherent polymorphisms in
the cyclin D1 gene. Supporting evidence stems from the
identification of a second mutation at Thr-286 (T/H)
through a search of the available expressed sequence tag
(EST) database. Unfortunately, we were unable to
obtain the cell line from which this cyclin D1 allele
was sequenced in order to confirm the presence of this
mutation.
To characterize the properties of the cancer-derived
mutants, we engineered NIH-3T3 cell lines that express
either D1-T286R isolated from esophageal cancer or
cyclin D1-D289–292, a deletion mutant recently identi-
fied in endometrial cancer (Moreno-Bueno et al., 2003).
Neither D1-T286R nor D1-D289–292 was phosphory-
lated at Thr-286 in vivo (data not shown). Consistent
with inhibition of Thr-286 phosphorylation, both
D1-T286R and D1-D289–292 exhibited a significantly
increased half-life relative to wild-type cyclin D1
(Figure 1a). In addition, the patient-derived cyclin D1
mutants featured constitutive nuclear localization as
assessed by indirect immunofluorescence (Figure 1b and
c) and retained their ability to support CDK4 catalytic
activity as assessed by in vitro Rb kinase assays
(Figure 1d).
To expand our analysis, we screened a panel of 20
independently derived human esophageal carcinoma cell
lines for mutations in exon 5 of cyclin D1. Three of
these, TE3, TE7 and TE12, harbored a mutation in
cyclin D1 resulting in a proline to alanine substitution at
codon 287. The proline-287 (P287)A mutation was
confirmed in TE3 and TE7 cell lines obtained from an
independent source. The expression of the alternatively
spliced cyclin D1b was not detected in either TE3 or
TE7, but could be detected in TE12 (Lu et al., 2003).
The capacity of GSK-3b to function as a proline-
directed kinase (Doble and Woodgett, 2003) implies that
P287 is likely to be required for Thr-286 phosphoryla-
tion (Diehl et al., 1997). To test this notion, NIH-3T3
cell lines were engineered to express the cyclin D1-
P287A mutant. Immunoblot with a phosphospecific
antibody revealed the absence of Thr-286 phosphoryla-
tion in the D1-P287A mutant (Figure 2a). Consistent
with inhibition of Thr-286 phosphorylation, D1-P287A
accumulated predominantly in the nucleus of expressing
cells (Figure 2b and c). We also examined the subcellular
localization of D1-P287A in TE3 (squamous cell
carcinoma) and TE7 (adenocarcinoma), which harbor
the endogenous cyclin D1-P287A mutant. D1-P287A
was almost entirely nuclear in both TE3 and TE7,
whereas wild-type cyclin D1, which is expressed in
KYSE520 cells, exhibited both nuclear and cytoplasmic
Constitutively nuclear cyclin D1 in human cancer
S Benzeno et al
6292
Oncogene
staining (Figure 3a and b). As the P287A mutation only
occurs in one cyclin D1 allele in TE3/T37 cell lines, the
absence of cytoplasmic staining suggests that accumula-
tion of cyclin D1-P287A greatly exceeds that of cyclin
D1 in these cells.
Phosphorylation of Thr-286 not only directs cyclin
D1 nuclear export but also promotes rapid proteasome-
dependent destruction of cyclin D1 (Diehl et al., 1997).
We therefore predicted that cyclin D1-P287A would
exhibit an extended half-life relative to wild-type cyclin
D1. We examined turnover of cyclin D1 versus D1-
P287A in both the esophageal carcinoma-derived cell
lines and NIH-3T3 cell engineered to express D1-P287A
ectopically. Consistent with loss of Thr-286 phosphor-
ylation and increased nuclear retention, cyclin
D1-P287A exhibited decreased kinetics of turnover in
both NIH-3T3 cells and the esophageal cancer cell lines,
TE3 and TE7, harboring the constitutively nuclear
D1-P287A mutant (Figure 3c). However, endogenous
wild-type cyclin D1, in NIH-3T3 and KYSE520 cells,
exhibited the expected short half-life of approximately
30 min (Figure 3c). Importantly, cyclin D1-P287A
retained the ability to support CDK4 activation as
demonstrated by phosphorylation of recombinant Rb
using in vitro kinase assays (Figure 3d).
C-terminal acidic residues within cyclin D1 coordinate
glycogen synthase kinase-3 beta-dependent
phosphorylation
Although our data demonstrate that the deletion of
amino acids 289–292 disrupts Thr-286 phosphorylation,
the underlying mechanism is not inherently obvious.
Through careful examination of the C-terminus of all
three D-type cyclins, we noted the presence of aspartic
acid residues positioned at (n þ 3) and (n þ 6) relative to
Thr-286 (n) that are spatially conserved in other D-type
cyclins (Figure 4a). Importantly, the deletion of amino
acids 289–292 of cyclin D1 in endometrial cancer
encompasses both aspartic acids D289 and D292. The
crystal structure of GSK-3b revealed a unique priming
mechanism that regulates GSK-3b-mediated phosphor-
ylation of its substrates (Bax et al., 2001; Dajani et al.,
2001; Frame et al., 2001). Priming phosphorylation
occurs at position (n þ 4) relative to the bona fide
residue of GSK-3b phosphorylation (n). Because cyclin
D1 is not a recipient of a priming phosphorylation, we
considered the possibility that a proximal negatively
charged acidic residue (aspartic acid) within cyclin
D1 might facilitate substrate binding within the dense
positively charged substrate-binding groove of GSK-3b.
We generated cell lines expressing either double
Figure 1 Cancer-derived cyclin D1 mutants are constitutively nuclear and retain kinase activity. (a) Cyclin D1 cancer-derived mutants
are stabilized. NIH-3T3 cells stably overexpressing the indicated cyclin D1 mutants were treated with cycloheximide for the indicated
intervals. Lysates prepared from the respective cell lines were processed for Western analysis using the cyclin D1 antibody. (b) NIH-
3T3 cells stably overexpressing wild-type cyclin D1, D1-T286R and D1-D289–292 were fixed and the localization of either cyclin D1
protein was determined by immunofluorescence. Corresponding Hoechst staining is shown. (c) Quantification of immunofluorescence
shown in (b). (d) D1, D1-T286R, and D1-D289–292 were precipitated from NIH-3T3 cells stably overexpressing the respective Flag-
tagged protein and subjected to in vitro kinase assays using recombinant glutathione-S-transferase-Rb. Phosphorylated Rb protein was
visualized by autoradiography following transfer to nitrocellulose membrane. The same membrane was immunoblotted for cyclin D1
and co-precipitated p21
Cip1
protein. IP with normal rabbit immunoglobulin G was used as a control.
Constitutively nuclear cyclin D1 in human cancer
S Benzeno et al
6293
Oncogene
D1-D289/D292A, or the single D1-D289A and D1-
D292A, alanine substitution mutants to assess their role
in the regulation of Thr-286 phosphorylation. In
contrast to D1-D292A, neither D1-D289/D292A nor
D1-D289A was detectably phosphorylated on Thr-286
(Figure 4b, lanes 3–5), demonstrating that D289
positioned (n þ 3) from Thr-286 contributes to GSK-
3b-dependent phosphorylation of cyclin D1. In addi-
tion, to ensure these results were not owing to epitope
masking, we tested the ability of GSK-3b to phosphor-
ylate wild-type cyclin D1 and D1-D289A in vitro. Cyclin
D1/CDK4 and D1-D289A/CDK4 complexes purified
from Sf9 lysates were used as substrates. Because a
majority of wild-type cyclin D1 molecules are phos-
phorylated at Thr-286 in Sf9 cells, we first subjected
purified cyclin D1 and D1-D289A to phosphatase
treatment. Following extensive washing to remove
phosphatase, equivalent concentrations of the respective
cyclin substrates were incubated with increasing con-
centrations of recombinant GSK-3b. Wild-type cyclin
D1 was efficiently phosphorylated by GSK-3b, whereas
phosphorylation of D1-D289A was significantly atte-
nuated (Figure 4c and d). Similar results were obtained
using the phospho-286 antibody to assess phosphoryla-
tion (data not shown). Importantly, all mutants retained
the ability to associate with GSK-3b (Figure 4e). Thus,
abrogated cyclin D1 phosphorylation at Thr-286 is not
the result of a disrupted cyclin D1–GSK-3b complex.
Our results identify residue D289, positioned (n þ 3)
relative to Thr-286, as critical for efficient GSK-3b-
mediated phosphorylation of cyclin D1.
These data demonstrate that Asp-289 is required for
Thr-286 phosphorylation and suggest that the negative
character of the aspartic acid residue might mimic that
associated with a phosphate group and direct GSK-3b
phosphorylation of cyclin D1. To test whether cyclin D1
is recognized by GSK-3b as a primed substrate, we
assessed the ability of the GID polypeptide (Gsk-3b
interacting domain of axin) to inhibit GSK-3b phos-
phorylation. GSK-3b-interacting domain of axin pre-
ferentially inhibits GSK-3b activity toward unprimed
substrates (Yost et al., 1998; Hedgepeth et al., 1999;
Fraser et al., 2002; Zhang et al., 2003). For these
experiments, the C-terminal 41 residues of either cyclin
D1 or D1-T286A were fused to glutathione-S-transfer-
ase (GST), expressed in bacteria and used as substrates
(Diehl et al., 1998), whereas protein phosphatase-1
inhibitor 2 (I-2) protein was used as an unprimed
substrate control (Park et al., 1994; Zhang et al., 2003).
As expected, GSK-3b phosphorylation of I-2 was
inhibited by GID in a concentration-dependent manner
Figure 4f (lanes 7–9) and g. In contrast, GID did not
alter GSK-3b-mediated phosphorylation of cyclin D1
Figure 4f (lanes 2–4) and g, demonstrating that
phosphorylation of cyclin D1 by GSK-3b differs from
that of a true unprimed substrate.
The absence of Thr-286 phosphorylation in the D1-
D289/292A and D1-D289A mutants suggested that both
might accumulate in the nuclear compartment and be
refractory to proteasomal degradation. Indeed, both
mutants exhibited exclusively nuclear localization pat-
terns (Figure 5a and b) as determined by immunofluor-
escent staining. Consistent with the absence of Thr-286
phosphorylation and their nuclear accumulation, both
D1-D289A and D1-D289/292A exhibit a half-life in
excess of 3 h, which is comparable to cyclin D1-T286A
(Figure 5c). Although phosphorylation at Thr-286 was
absent in both D1-D289A and D1-D289/292A mutants,
both retained the ability to bind to Cip/Kip proteins and
to support Rb kinase activity (Figure 5d), emphasizing
the retained structural and functional integrity of these
mutants. These data support a model wherein D289
functions to coordinate GSK-3b-mediated phosphoryla-
tion of cyclin D1.
Figure 2 The tumor-derived cyclin D1-P287A mutant is consti-
tutively nuclear. (a) Abrogated D1-P287A phosphorylation at
threonine 286. Sf9 cells were infected with D1 or D1-P287A. Flag-
tagged protein was precipitated with M2 antibody and assayed for
phosphorylation at threonine 286 by Western blot using phos-
phospecific threonine 286 antibody. Total immunoprecipitated
cyclin D1 was confirmed using cyclin D1 antibody. (b) D1–3T3 or
D1-P287A-3T3 cells were fixed and the localization of cyclin D1
protein was determined by immunofluorescence. Corresponding
Hoechst staining is shown. (c) Quantification of immunofluores-
cence shown in (b).
Constitutively nuclear cyclin D1 in human cancer
S Benzeno et al
6294
Oncogene
Cancer-derived cyclin D1 mutants exhibit oncogenic
properties relative to wild-type cyclin D1
Cells engineered to overexpress cyclin D1 display a
contracted G1 interval (Quelle et al., 1993) but do not
exhibit a transformed phenotype. In contrast, a cyclin
D1 mutant that cannot be phosphorylated at Thr-286
(D1-T286A) and is constitutively nuclear promotes
cellular transformation (Alt et al., 2000; Gladden
et al., 2005), demonstrating that disruption of cyclin
D1 nuclear export is an oncogenic event. Thus, we
hypothesized that cancer-derived cyclin D1 mutants that
cannot be phosphorylated at Thr-286 and are thus
constitutively nuclear should exhibit an increased
transforming potential relative to wild-type cyclin D1.
We tested this hypothesis using NIH-3T3 or NIH-3T3
derivatives that overexpress Flag-tagged mutant cyclin
D1 isoforms and analysed these cells for characteristics
of transformation. As published previously (Alt et al.,
2000), early passage cell lines displayed normal
growth control (data not shown). In contrast, over
the course of five independent experiments utilizing
cell lines of passage p10 or later, cyclin D1 mutants,
D1-T286R and D1-D289–292, reproducibly formed
foci (Figure 6a). Consistent with these results, identical
results were obtained in experiments analysing ancho-
rage-independent growth in soft agar, a more stringent
criterion for oncogenic potential (Figure 6b and c).
These data demonstrate that tumor-derived cyclin D1
mutations targeting Thr-286 phosphorylation and nu-
clear export confer an oncogenic advantage to cells
expressing mutant forms of cyclin D1 that exhibit an
increased capacity to drive neoplastic growth in the
absence of a collaborating oncogene.
Transformation by constitutively nuclear cyclin D1
correlates with CDK4 activation
Cyclin D1 was initially characterized as the allosteric
regulatory subunit for CDK4/6 kinases. However,
accumulating evidence suggests that cyclin D1 can also
function as a transcriptional regulator independent
of its capacity to regulate CDK activity (Fu et al.,
2004). These observations prompted us to determine
whether transformation by cyclin D1 is kinase depen-
dent. To do so, we generated a cyclin D1-K114E (and
compound T286A/R point mutants) that was previously
characterized in vitro as ‘kinase-dead’ (Hinds et al.,
1994; Coqueret, 2002). This single amino-acid change,
Figure 3 Identification of esophageal carcinoma-derived cell lines harboring cyclin D1-P287A. (a) The subcellular localization of
endogenous cyclin D1 protein in KYSE520 and D1-P287A in TE3 and TE7 cell lines was determined by immunofluorescence using a
cyclin D1 antibody (Ab3, Calbiochem, San Diego, CA, USA) followed by fluorescein isothiocyanate-conjugated anti-mouse secondary
antibody (green). Corresponding Hoechst DNA staining is shown (blue) and a merged view (Merge) of both channels (green/blue).
(b) Quantification of immunofluorescence is shown in (a). (c) NIH-3T3 cells stably overexpressing D1-P287A and patient-derived
KYSE520, TE3 and TE7 cell lines were treated with cycloheximide for the indicated intervals. Lysates from the indicated cell lines were
assayed for cyclin D1 expression by Western analysis using the cyclin D1 antibody. (d) Following co-infection of D1, D1-T286A and
D1-P287A with CDK4 in Sf9 cells, cyclin D1 protein was immunoprecipitated using either the M2 antibody or cyclin D1 antibody and
assayed for its ability to support Rb kinase activity, visualized by autoradiography following transfer onto nitrocellulose membrane.
The same membrane was processed for Western blot analysis of phospho-serine 780 Rb, cyclin D1 and co-precipitating CDK4.
Constitutively nuclear cyclin D1 in human cancer
S Benzeno et al
6295
Oncogene
Lys114-Glu (K114E), is thought to disrupt cyclin
D1-CDK4 binding. We generated NIH-3T3 cells that
stably express cyclin D1-T286R-KE, D1-KE and
T286A-KE. All three KE mutants exhibited localization
and degradation kinetics identical to their counterparts:
D1, D1-T286A, D1-T286R (Figure 7a and b). To
confirm whether the K114E mutation disrupts cyclin
CDK interactions, lysates prepared from NIH-3T3 cell
lines overexpressing the respective K114E cyclin D1
mutant were precipitated with the M2 monoclonal
Figure 4 GSK-3b phosphorylation of cyclin D1 depends on the integrity of aspartic acid D289. (a) Schematic of spatially conserved
acidic aspartic acid residues within the C-terminus of D-type cyclins. (x) ¼ Non-conserved amino acids, (n) ¼ GSK-3b phosphorylation
site and (n þ 3) ¼ conserved proximal aspartic acid 289. (b) threonine 286 phosphorylation status of cyclin D1 mutants. Cyclin D1 was
precipitated using M2 antibody, and threonine 286 phosphorylation was assayed by immunoblot using the phospho-threonine 286
antibody. Total cyclin D1 was confirmed using an antibody that recognizes both phosphorylated and unphosphorylated species.
(c) Purified cyclin D1 or D1-D289A were incubated with increasing concentrations of recombinant GSK-3b and [g
32
-P]ATP. D1-
T286A, as a negative control, was incubated with the highest concentration of GSK-3b ( þ ). Cyclin D1 phosphorylation was visualized
by autoradiography. (d) Quantification of g
32
-P-incorporation is shown in (c). (e) GSK-3b co-precipitates with wild-type and mutant
cyclin D1 isoforms. Sf9 lysates co-expressing kinase-defective GSK-3b and the indicated Flag-D1 isoforms were precipitated with the
M2 antibody. Cyclin D1 and co-precipitating GSK-3b levels were confirmed by immunoblot. IP with immunoglobulin G was used as a
control. (f)
GSK-3b-interacting-domain of axin (GID) fails to inhibit GSK-3b phosphorylation of cyclin D1. Purified C-terminal
glutathione-S-transferase-cyclin D1 or recombinant inhibitor 2 (I-2) was added to in vitro kinase reactions in the presence of GSK-3b
and [g
32
-P]ATP. Phosphorylated proteins were separated on a denaturing polyacrylamide gel and visualized by autoradiography.
Synthesized GID
380–404
peptide (GID) was added at increasing concentrations (black triangle). Cyclin D1 and I-2 protein levels were
visualized by Coomassie staining. () ¼ No GSK-3b and control ¼ reactions in the absence of GID. D1-T286A (T286A) is a negative
control for cyclin D1 phosphorylation. (g) Quantification of cyclin D1 and I-2 phosphorylation. Y-axis ¼ normalized phosphorylation
relative to control reaction. Results are representative of at least four independent experiments.
Constitutively nuclear cyclin D1 in human cancer
S Benzeno et al
6296
Oncogene
antibody followed by immunoblot for CDK4. Surpris-
ingly, endogenous CDK4 efficiently co-precipitated with
all K114E derivative proteins, demonstrating that cyclin
D1-CDK4 binding is not abolished in vivo (Figure 7c).
Although mutant cyclins appear to associate effectively
with CDK4 in cultured cells, immune complex kinase
assays using the respective cyclin–CDK complexes
revealed that all K114E mutant holoenzymes exhibited
reduced kinase activity relative to wild-type counter-
parts, as assessed by total g
32
-P incorporation and
immunoblotting for Rb phosphorylation at serine 780
(Ser780) (Kitagawa et al., 1996) (Figure 7c). Densito-
metric scanning revealed KE mutants retain approxi-
mately 20% of wild-type cyclin D1 activity towards Rb.
Although the K114E mutants retain 20% of wild-type
kinase activity, we reasoned that if kinase activation
were critical for transformation of mouse fibroblasts,
K114E mutant activity would be compromised. We
assessed the ability of K114E mutant cyclin D1
derivatives to induce growth of NIH-3T3 cells in
semisolid medium. Neither wild-type cyclin D1 nor the
D1-KE triggered significant growth (Figure 6d).
Although cell lines of passage 10 or greater expressing
constitutively nuclear TA-KE and TR-KE were capable
of growth in soft agar, growth, as assessed by colony
number, was reduced by greater than 50% relative to
both D1-T286A and D1-T286R, respectively (Figure 6d
and e). As the KE-T286A mutant retained the ability to
accelerate G1 progression (Figure 7d), their reduced
transforming capacity cannot be attributed to defects in
G1 progression. These data demonstrate that cyclin D1
mutants with compromised kinase activity are also
compromised with respect to transforming potential.
To further investigate the mechanism underlying the
retained kinase activity and assembly of the K114E cyclin
D1 mutants with CDK4, we considered the possibility
that association of D1-KE mutants with CDK4 in vivo
was facilitated by a cellular cofactor. Indeed, p21
Cip1
and
p27
Kip1
have been shown to promote stable assembly of
the cyclin D1/CDK4 kinase (LaBaer et al., 1997; Cheng
et al., 1999). We used insect Sf9 cells, which express no
detectable endogenous p21
Cip1
, to determine whether it
facilitates the association of mutant cyclin D1 with
CDK4. Sf9 were co-infected with baculovirus encoding
wild-type cyclin D1 or D1-KE along with CDK4 and
increasing concentrations of p21
Cip1
. Western analysis of
the cyclin D1 immune complexes revealed markedly
reduced levels of CDK4 in D1-KE relative to wild-type
cyclin D1 precipitates in the absence of p21
Cip1
(Figure 8a,
compare lanes 4 and 10). Serial titration of the p21
Cip1
baculovirus increased co-precipitation of CDK4 with
either immunoprecipitated D1 or D1-KE.
We subsequently investigated whether cyclin D1
KE–CDK4 complexes assembled by p21
Cip1
in Sf9 cells
were functionally active (Figure 8b and c). In the
absence of p21
Cip1
, D1-KE retained severely reduced
kinase activity (Figure 8b, lanes 4, 5 and 9, 10) relative
to wild-type D1 (Figure 8b, lanes 2 and 7) but above
that of controls (Figure 8b, lanes 1 and 6). Immunoblot
for phospho-Ser780 Rb confirmed reduced Rb phos-
phorylation by mutant D1-KE. Co-expression of p21
Cip1
increased CDK4 association with D1-KE (Figure 8a)
Figure 5 Mutation of D289 results in a constitutively nuclear and stable cyclin D1. (a) D1-, D1-D289/292A- and D1-D289A-3T3 cells
were fixed and cyclin D1 localization determined by indirect immunofluorescence. (b) Quantification of cyclin D1 immunofluorescence
shown in (a). (c) Stabilization of cyclin D1 mutants. NIH-3T3 cells stably overexpressing the indicated cyclin D1 proteins were treated
with cycloheximide for the indicated intervals. Lysates were subjected to Western analysis using the cyclin D1 antibody. (d) Cyclin D1
mutants are kinase competent. Proteins from NIH-3T3 cells stably overexpressing the indicated Flag-tagged cyclin D1 were
precipitated using M2 antibody and assayed for the ability to support Rb kinase activity. Phosphorylated proteins were separated by
SDS–polyacrylamide gel electrophoresis (PAGE), transferred onto a nitrocellulose membrane and visualized by autoradiography. The
same nitrocellulose membrane was processed for Western blot using antibodies for cyclin D1 and co-precipitating p21
Cip1
and p27
Kip1
.
Constitutively nuclear cyclin D1 in human cancer
S Benzeno et al
6297
Oncogene
and increased the kinase activity of the D1-KE
holoenzyme (Figure 8c). Strikingly, whereas high con-
centrations of p21
Cip1
rapidly inhibit wild-cyclin D1
activity, inhibition of D1-KE kinase activity by p21
Cip1
is achieved only at the highest p21
Cip1
concentrations
(Figure 8a and c). These results indicate that in the
absence of a physiologic assembly factor in vitro,a
K114E mutant fails to associate with CDK4. However,
expression of p21
Cip1
in vivo or its supplementation
in vitro is sufficient to maintain assembly of a KE cyclin
D1–CDK4 complex, providing for attenuated yet
sustained catalytic potential. As the K114E mutation
does not result in differential association of cyclin D1
with the general transcriptional apparatus, our results
suggest that reduced cellular transformation in the
presence of a K114E mutation results from reduced
nuclear CDK4 activation.
Discussion
Previous work has suggested that phosphorylation-
dependent nuclear export of cyclin D1 is critical for
Figure 6 Tumor-derived cyclin D1 mutants transform murine fibroblasts. (a) Cells were plated in six-well dishes in medium
supplemented with 5% fetal calf serum and stained with Giemsa to visualize foci following 21 days of growth. (b) NIH-3T3 cells stably
overexpressing the indicated cyclin D1 proteins were plated in semisolid medium and allowed to proliferate for 21 days. Colonies were
visualized by 0.01% neutral red stain. (c) Quantification of colonies scored in (b). (d) NIH-3T3 cells stably expressing the indicated
cyclin D1 proteins were plated in semisolid medium and allowed to proliferate for 21 days. Colonies were visualized by 0.01% neutral
red stain. (e) Quantification of colonies scored is shown in (d).
Constitutively nuclear cyclin D1 in human cancer
S Benzeno et al
6298
Oncogene
the prevention of aberrant cell growth in vitro and in
mouse models (Alt et al., 2000; Gladden et al., 2005).
However, there is little data demonstrating the existence
of cyclin D1 mutations that disrupt this regulatory
process in human cancer. With the identification of
critical regulatory residues within the C-terminus of
cyclin D1 that maps to exon 5 of the cyclin D1 gene, we
have identified and characterized mutations in cyclin D1
isolated from primary human cancer. We have identified
an arginine for threonine substitution at position 286 in
primary esophageal cancers, a histidine for threonine
substitution at position 286 through an EST tag search,
a deletion of amino acids 266–295 in primary esophageal
cancer and a proline to alanine substitution at position
287, D1-P287A, identified in three independently
derived esophageal cancer cell lines, TE3, TE7 and
TE12. These mutations target residues responsible for
Thr-286 phosphorylation and therefore, directly impact
Figure 7 K114E mutants confer a hypomorphic phenotype. (a) NIH-3T3 cells stably expressing D1-KE, TA-KE and TR-KE were
fixed, and cyclin D1 localization determined by indirect immunofluorescence. (b) NIH-3T3 cells stably expressing the indicated proteins
were treated with cycloheximide. Lysates from the respective cell lines were processed for Western blot analysis using cyclin D1
antibody. (c) Lysates from NIH-3T3 cell lines stably expressing the indicated Flag-tagged proteins were processed for precipitation
with M2 antibody and the resulting precipitates were assayed for the ability to support Rb kinase activity. Phosphorylated Rb was
visualized by autoradiography, following SDS–polyacrylamide gel electrophoresis (PAGE) and transfer to nitrocellulose membrane.
The same membrane was probed with antibodies specific for phospho-serine 780 Rb. (d) NIH-3T3 (yellow), D1-3T3 (blue), D1-T286A
(green) and TA-KE-3T3 (maroon) cell lines synchronized by culturing in media containing 0.1% serum for 24 h were stimulated to re-
enter the cell cycle by the addition of serum-derived growth factors and 5-bromo-2-
0
deoxyuridine. S-phase entry was monitored by flow
cytometry.
Constitutively nuclear cyclin D1 in human cancer
S Benzeno et al
6299
Oncogene
cyclin D1 nuclear export and ubiquitin-dependent
proteolysis. Characterization of the identified cyclin
D1 mutants revealed wild-type properties with regard to
association and activation of CDK4, suggesting that any
oncogenic properties do not reflect an increased
propensity to induce CDK activity. Rather, the impact
of the mutations lies exclusively upon the constitutive
nuclear accumulation of active cyclin D1/CDK com-
plexes.
Several conclusions can be drawn regarding cyclin D1
alterations in cancer. First, mutations in the cyclin D1
coding region likely occur at a relatively low frequency,
B4% (four mutations identified in 110 samples includ-
ing cell lines and primary tissue) in the current study.
Although this figure is less than the estimated frequency
of cyclin D1 overexpression in esophageal cancer, it is
important to consider that B-Raf mutations have been
estimated to occur only occur at an estimated frequency
of 8% in human cancer (Davies et al., 2002; Garnett and
Marais, 2004). Second, the mutations specifically disrupt
CRM1 nuclear export resulting in the constitutive
nuclear accumulation of an active cyclin D1/CDK4
kinase that is refractory to rapid degradation via the 26S
proteasome. Finally, whereas this might argue against a
generalized model wherein inactivation of cyclin D1
nuclear export contributes to cancer initiation, it is likely
that this mutation frequency dramatically under-repre-
sents the actual frequency with which cyclin D1 nuclear
export is targeted in human cancer. In point of fact, we
have previously demonstrated that cancer-specific alter-
native splicing of cyclin D1 contributes to the expression
of a constitutively nuclear cyclin D1 isoform; our work
suggests that this isoform is expressed in approximately
40% of primary esophageal carcinomas (Lu et al., 2003).
Similar to our results, C-terminal cyclin D1 mutants
were also identified in endometrial cancer (Moreno-
Bueno et al., 2003) that map to residues adjacent to
Thr-286. The in-frame deletion of residues 289–292 in
D1-D289–292 deletes both a portion of the cyclin D1
nuclear export signal (Benzeno and Diehl, 2004) and a
critical aspartic acid, D289, that is essential for
subsequent phosphorylation of Thr-286 by GSK-3b.
As predicted from loss of both Thr-286 phosphorylation
and deletion of the NES, D1-D289–292 accumulates in
the nucleus and drives cellular transformation. Thus,
cyclin D1 mutations isolated from both esophageal and
endometrial cancer target the same regulatory sites
important for nuclear export. Given that deregulation of
cyclin D1 is proposed to contribute to the genesis and
progression of additional cancers, it will be critical to
expand this analysis to ascertain the potential contribu-
tion of cyclin D1 mutations to additional malignancies.
The GSK-3b kinase targets two distinct sets of
substrates, primed and unprimed. Canonically primed
GSK-3b substrates are characterized by a preceding
phosphorylation event that is (n þ 4) residues from the
target GSK-3b phosphorylation site (n). This substrate
priming event potentiates the alignment of b- and a-
helical domains within the positively charged binding
groove of GSK-3b and provides optimal coordination
of the target phosphorylation residue within the
catalytic cleft of this kinase (Dajani et al., 2001; ter
Haar et al., 2001). Examples of primed GSK-3b
substrates include glycogen synthase, cyclin E, c-myc
and NFAT (Fiol et al., 1988, 1990; Beals et al., 1997;
Sears et al., 2000; Welcker et al., 2003, 2004). Unprimed
GSK-3b substrates include the I-2 inhibitor of PP1 and
b-catenin (Hagen et al., 2002; Zhang et al., 2003).
Figure 8 D1-K114E binds CDK4 and retains kinase activity in the presence of p21. (a) Cyclin D1 complexes were
immunoprecipitated from Sf9 lysates prepared following co-infection with either wild-type cyclin D1 or D1-KE along with CDK4
and increases concentrations of baculovirus encoding p21
Cip1
using cyclin D1 antibody. The resulting immunoprecipitates were assayed
for co-precipitated CDK4 and p21
Cip1
. Total immunoprecipitated cyclin D1 is confirmed by immunoblot using cyclin D1 antibody.
(b) Following co-infection of CDK4 either with D1, D1-T286A or D1-KE in Sf9 cells, cyclin D1 was immunoprecipitated using either
M2 or cyclin D1 antibody and assayed for the ability to support Rb kinase activity. Phosphorylated Rb was visualized by
autoradiography, following transfer onto a nitrocellulose membrane. The same membrane was probed for Rb phospho-serine 780,
cyclin D1 and co-precipitating CDK4. (c) Moderate concentrations of p21
Cip1
promote D1-K114E kinase activity. D1 and D1-K114E
in vitro kinase assays using recombinant glutathione-S-transferase-Rb and [g
32
-P]ATP in the absence and presence of increasing
concentrations of p21
Cip1
. Phosphorylated Rb was resolved by electrophoresis, transferred onto a nitrocellulose membrane and
visualized by autoradiography.
Constitutively nuclear cyclin D1 in human cancer
S Benzeno et al
6300
Oncogene
As a canonical priming phosphorylation event that is
characterized by the transfer of a phosphate group at
position (n þ 4) is not required to initiate GSK-3b-
mediated phosphorylation of cyclin D1 at Thr-286, the
classification of cyclin D1 as an ‘unprimed’ GSK-3b
substrate is appropriate at face value (Diehl et al., 1998).
Our data, on the other hand, suggest that recognition of
cyclin D1 by GSK-3b requires an acidic residue at the
(n þ 3) position (Asp-289); the negative charge provided
by Asp-289 may in fact substitute for a canonical
priming phosphorylation event at (n þ 4). Indeed,
aspartic acid 289 is conserved in both mouse and human
cyclin D1 and is spatially conserved in all three D-type
cyclins, highlighting its significance. Mutation Asp-289
to a non-charged residue or its deletion inhibits Thr-286
phosphorylation. Previous work has revealed that a
specific peptide inhibitor, GID, preferentially inhibits
GSK-3b phosphorylation of unprimed substrates such
as I-2 (Hedgepeth et al., 1999; Zhang et al., 2003).
Consistent with GSK-3b recognizing cyclin D1 as a
primed substrate (Hedgepeth et al., 1999; Zhang
et al., 2003), phosphorylation of cyclin D1 by GSK-3b
is insensitive to the GID peptide inhibitor. These
data strongly support the idea that D289 facilitates
recognition and phosphorylation of cyclin D1 at
Thr-286 and may do so via providing the negative
charge generally provided by phosphorylation at the
(n þ 4) position.
Nuclear accumulation of cyclin D1 during S phase
correlates with its oncogenic capacity and ascribes a
gain-of-function activity to cancer-derived, nuclear
export-defective cyclin D1 mutants. The precise nature
of cyclin D1’s nuclear S-phase function remains unclear
but may be related to the formation of active complexes
with CDK4. Alternatively, there is evidence for the
CDK-independent functions of cyclin D1 as well (Hirai
and Sherr, 1996; Neuman et al., 1997; Zwijsen et al.,
1997; Lamb et al., 2000; Petre et al., 2002; Benzeno
et al., 2004). With the advent of anticancer therapeutics
that target the active site of cyclin-dependent kinases, it
is critical to discern between CDK-dependent versus
CDK-independent functions (Fry et al., 2001, 2004).
To begin to address the contribution of cyclin D1 kinase
activity in cell transformation, we investigated the
potential of a K114E mutation in the context of
constitutively nuclear TA-KE and TR-KE, to affect
cellular transformation in vitro. The K114E mutation
had previously been reported as ‘kinase-dead owing to
its failure to bind CDK4 in vitro (Hinds et al., 1994).
In contrast, we found that in cells this mutant cyclin D1
associates with both CDK4 and p21
Cip1
but is character-
ized by an attenuated and thus, hypomorphic capacity
to activate CDK4 catalytic activity. Although not
‘kinase-dead’, we used these compound mutants as a
tool to evaluate the effect of attenuated cyclin D1
activity on transforming potential. Our results reveal
that the attenuated activity of constitutively nuclear
T286A-KE and T286R-KE correlates with a significant
reduction in transformation relative to wild-type coun-
terparts, T286A and T286R, as assessed by anchorage-
independent growth. Our data provide strong evidence
that transformation by cyclin D1 is driven by its
constitutive nuclear retention. Our results along with
recently published work from the Sicinski and Hinds
laboratories (Landis et al., 2006; Yu et al., 2006) also
suggest that neoplastic growth is the direct result of
unrestricted catalytic activity.
Although cyclin D1 is frequently overexpressed in
human cancer, it is only weakly oncogenic in vitro or in
mouse model systems. In contrast, we have demon-
strated that cyclin D1 mutants that are refractory to
nuclear export exhibit an enhanced capacity to trigger a
cell transformation in vitro. We have recently demon-
strated that transgenic expression of the constitutively
nuclear mutant cyclin D1-T286A in murine lymphocytes
triggers an aggressive B-cell lymphoma (Gladden et al.,
2005). This is again in contrast to wild-type cyclin D1
expression of which is not sufficient to trigger B-cell
malignancies in mice (Bodrug et al., 1994; Lovec et al.,
1994a). Interpreted in sum, we suggest that overexpres-
sion of wild-type cyclin D1, although providing a
growth advantage and contributing to the proliferative
potential of tumor cells, is not in fact a transforming
event. In contrast, in cases where cyclin D1 contributes
to cancer initiation, accumulating data suggests that
cells will exhibit defects in cyclin D1 nuclear export.
These defects could result from mutations that impact
the GSK-3b phosphoacceptor site (data herein), induce
alternative splicing of cyclin D1 (Lu et al., 2003) or via
upstream mutations that impact the signal-transduction
pathway that regulates Thr-286 phosphorylation
(Rimerman et al., 2000).
In conclusion, our results emphasize the need for
further investigation into the downstream mechanism(s)
by which oncogenic mutants of cyclin D1 drive
neoplastic growth. The results provided by these
experiments will increase our understanding of the
contribution of cyclin D1/CDK activity to cyclin D1
transformation in human neoplasia and expand our
knowledge of downstream regulators including novel
substrates that potentiate the oncogenicity of a con-
stitutively active cyclin D1 mutant protein.
Materials and methods
Cell culture conditions and transfections
All mammalian cell lines were maintained in Dulbecco’s
modified Eagle’s medium containing glutamine supplemented
with antibiotics (Cellgro, Mediatech Inc., Herndon, VA, USA)
and 10% fetal calf serum (FCS) (BioWhitaker Europe,
Belgium). Insect Sf9 cells were grown in Grace’s medium
supplemented with 10% heat-inactivated FCS. Procedures for
manipulation of baculoviruses were described previously
(Summers, 1987). All cyclin D1 point mutants corresponding
to human somatic cyclin D1 mutations were engineered using
pFlex-murine cyclin D1 vector as a template for oligonucleo-
tide-directed mutagenesis with the Quick ChangeTM in vitro
mutagenesis system (Stratagene, La Jolla, CA, USA). All flag-
tagged cyclin D1 mutants were confirmed by sequencing.
Derivation of NIH-3T3 cells engineered to overexpress Flag-
tagged cyclin D1 and Flag-tagged cyclin D1 mutants were as
described previously (Diehl et al., 1998).
Constitutively nuclear cyclin D1 in human cancer
S Benzeno et al
6301
Oncogene
Immunoblotting and kinase assays
For Western analysis, cells were lysed in Tween-20 buffer
(50 m
M N-2-hydroxyethylpiperazine-N-2-ethanesulfonic acid
at pH 7.5, 150 m
M NaCl, 10 mM MgCl
2
,1mM ethylenediamine
tetraacetate, 0.1% Tween-20, 1 m
M phenylmethylsulfonyl
fluoride, 20 U/ml aprotinin, 5 mg/ml leupeptin, 0.4 m
M NaVO
4
,
0.4 m
M NaF). Proteins were resolved on denaturing poly-
acrylamide gels, electrophoretically transferred to nitrocellu-
lose membranes (MSI, Westborough, MA, USA), and blotted
with the indicated antibodies. For the detection of cyclin D1-
dependent kinase activity, cells were harvested in Tween 20
buffer. Following precipitation with the M2 or cyclin D1
antibody, protein kinase assays using 1 mg of recombinant
GST-Rb were performed as described previously (Diehl and
Sherr, 1997; Rimerman et al., 2000). Detection of GSK-3b was
as described previously (Diehl et al., 1998); GSK-3b kinase
assays and GID inhibition assays were performed as described
previously (Zhang et al., 2003).
Immunofluorescence
NIH-3T3 cells seeded on glass coverslips were fixed using
methanol–acetone (1:1). Coverslips were stained with a mouse-
specific cyclin D1 monoclonal antibody (D1-17-13G) in
phosphate-buffered saline (PBS)/10% FCS. Secondary fluor-
escein isothiocyanate-conjugated anti-mouse (Amersham
Pharmacia Biotech, Uppsala, Sweden) antibody staining was
performed for 30 min. DNA was visualized using Hoechst
33258 dye at a 1:500 dilution. Coverslips were mounted on
glass slides with Vectashield medium (Vector Laboratories
Inc., Burlingame, CA, USA).
Protein turnover analysis
NIH-3T3 cells overexpressing either Flag-tagged wild-type or
mutant cyclin D1 were seeded in 10 cm dishes. The following
day, cells were treated with the protein synthesis inhibitor
cycloheximide (100 mg/ml; Sigma, St Louis, MO, USA) for the
indicated intervals, and harvested in sodium dodecylsulfate
(SDS)-sample buffer.
Tumor samples and DNA sequence analysis
Cyclin D1 exon 5 polymerase chain reaction products from
normal and tumor-derived genomic DNA samples were generated
using human cyclin D1 forward: 5
0
-CTCAGGTCCAGAG
GAGGCAG-3
0
and human cyclin D1 reverse: 5
0
-GAGATG
GAAGGGGGAAAGAG-3
0
primers that generated overlapping
amplicons. All mutations were confirmed via bidirectional
sequencing in two independent polymerase chain reaction
reactions. Tumor-derived cell lines utilized were available in our
laboratory. Mutations in TE3/7 were confirmed upon acquisition
of independent samples of these from the Rustgi laboratory.
Cell transformation assays
NIH-3T3 cells and derivatives overexpressing the indicated
cyclin D1 isoforms were plated at 1.5 10
5
cells/well of a
six-well dish. Cells were cultured in media containing 5% FCS.
Foci were visualized after 21–28 days with Wright–Giemsa
stain (Sigma). Anchorage-independent growth was determined
by analysing cellular growth in semisolid medium. Cells
(5 10
3
) were seeded in Iscove’s media containing 0.65%
noble agar/10% FCS. Cells were grown for 3 weeks in 8%
CO
2
. For visualization, colonies were incubated with 1 ml/well
of 0.01% neutral red (Sigma N7005) in PBS for 1 h at 371C.
Quantification of results represents the average number of
colonies scored with standard error calculated from three
independent experiments.
Acknowledgements
We thank J Woodgett for providing GSK-3b cDNAs. This
work was supported by a grant from the National Institutes of
Health (CA93237, CA111360) and the WW Smith Charitable
Trust (JAD).
References
Adelaide J, Monges G, Derderian C, Seitz JF, Birnbaum D.
(1995). Br J Cancer 71: 64–68.
Aktas H, Cai H, Cooper GM. (1997). Mol Cell Biol 17:
3850–3857.
Albanese C, Johnson J, Watanabe G, Eklund N, Vu D, Arnold
A et al. (1995). J Biol Chem 270: 23589–23597.
Alt JR, Cleveland JL, Hannink M, Diehl JA. (2000). Genes
Dev 14: 3102–3114.
Alt JR, Gladden AB, Diehl JA. (2002). J Biol Chem 277:
8517–8523.
Arber N, Lightdale C, Rotterdam H, Han KH, Sgambato A,
Yap E et al. (1996). Cancer Epidemiol Biomarkers Prev 5:
457–459.
Bani-Hani K, Martin IG, Hardie LJ, Mapstone N, Briggs JA,
Forman D et al. (2000). J Natl Cancer Inst 92: 1316–1321.
Bax B, Carter PS, Lewis C, Guy AR, Bridges A, Tanner R
et al. (2001). Structure (Cambridge) 9: 1143–1152.
Beals CR, Sheridan CM, Turck CW, Gardner P, Crabtree GR.
(1997). Science 275: 1930–1934.
Benzeno S, Diehl JA. (2004). J Biol Chem 279: 56061–56066.
Benzeno S, Narla G, Allina J, Cheng GZ, Reeves HL, Banck
MS et al. (2004). Cancer Res 64: 3885–3891.
Bodrug SE, Warner BJ, Bath ML, Lindeman GJ, Harris AW,
Adams JM. (1994). EMBO J 13: 2124–2130.
Calbo J, Parreno M, Sotillo E, Yong T, Mazo A, Garriga J
et al. (2002). J Biol Chem 277: 50263–50274.
Cheng M, Olivier P, Diehl JA, Fero M, Roussel MF, Roberts
JM et al. (1999). EMBO J 18: 1571–1583.
Cheng M, Sexl V, Sherr CJ, Roussel MF. (1998). Proc Natl
Acad Sci USA 95: 1091–1096.
Collecchi P, Santoni T, Gnesi E, Giuseppe Naccarato A,
Passoni A, Rocchetta M et al. (2000). Cytometry 42:
254–260.
Coqueret O. (2002). Gene 299: 35–55.
Dajani R, Fraser E, Roe SM, Young N, Good V, Dale TC
et al. (2001). Cell 105: 721–732.
Davies H, Bignell GR, Cox C, Stephens P, Edkins S, Clegg S
et al. (2002). Nature 417: 949–954.
Diehl JA, Sherr CJ. (1997). Mol Cell Biol 17: 7362–7374.
Diehl JA, Cheng M, Roussel MF, Sherr CJ. (1998). Genes Dev
12: 3499–3511.
Diehl JA, Zindy F, Sherr CJ. (1997). Genes Dev 11: 957–972.
Diehl JA. (2002). Cancer Biol Ther 1: 226–231.
Doble BW, Woodgett JR. (2003). J Cell Sci 116: 1175–1186.
Farkas T, Hansen K, Holm K, Lukas J, Bartek J. (2002).
J Biol Chem 277: 26741–26752.
Fiol CJ, Haseman JH, Wang YH, Roach PJ, Roeske RW,
Kowalczuk M et al. (1988). Arch Biochem Biophys 267:
797–802.
Fiol CJ, Wang A, Roeske RW, Roach PJ. (1990). J Biol Chem
265: 6061–6065.
Frame S, Cohen P, Biondi RM. (2001). Mol Cell 7: 1321–1327.
Constitutively nuclear cyclin D1 in human cancer
S Benzeno et al
6302
Oncogene
Fraser E, Young N, Dajani R, Franca-Koh J, Ryves J,
Williams RS et al. (2002). J Biol Chem 277: 2176–2185.
Fry DW, Bedford DC, Harvey PH, Fritsch A, Keller PR, Wu
Z et al. (2001). J Biol Chem 276: 16617–16623.
Fry DW, Harvey PJ, Keller PR, Elliott WL, Meade M,
Trachet E et al. (2004). Mol Cancer Ther 3: 1427–1438.
Fu M, Wang C, Li Z, Sakamaki T, Pestell RG. (2004).
Endocrinology 145: 5439–5447.
Garnett MJ, Marais R. (2004). Cancer Cell 6: 313–319.
Gladden AB, Woolery R, Aggarwal P, Wasik MA, Diehl JA.
(2005). Oncogene 25: 1620–1628.
Hagen T, Di Daniel E, Culbert AA, Reith AD. (2002). J Biol
Chem 277: 23330–23335.
Hall M, Peters G. (1996). Adv Cancer Res 68: 67–108.
Harbour JW, Luo RX, Dei Santi A, Postigo AA, Dean DC.
(1999). Cell 98: 859–869.
Hatakeyama M, Brill JA, Fink GR, Weinberg RA. (1994).
Genes Dev 8: 1759–1771.
Hedgepeth CM, Deardorff MA, Rankin K, Klein PS. (1999).
Mol Cell Biol 19: 7147–7157.
Hinds PW, Dowdy SF, Eaton EN, Arnold A, Weinberg RA.
(1994). Proc Natl Acad Sci USA 91: 709–713.
Hirai H, Sherr CJ. (1996). Mol Cell Biol 16: 6457–6467.
Kauffmann-Zeh A, Rodriguez-Viciana P, Ulrich E, Gilbert C,
Coffer P, Downward J et al. (1997). Nature 385: 544–548.
Kerkhoff E, Rapp UR. (1997). Mol Cell Biol 17: 2576–2586.
Kitagawa M, Higashi H, Jung HK, Suzuki-Takahashi I, Ikeda
M, Tamai K et al. (1996). EMBO J 15: 7060–7069.
LaBaer J, Garrett MD, Stevenson LF, Slingerland JM,
Sandhu C, Chou HS et al. (1997). Genes Dev 11: 847–862.
Lamb J, Ladha MH, McMahon C, Sutherland RL, Ewen ME.
(2000). Mol Cell Biol 20: 8667–8675.
Landis MWPB, Li T, Sicinski P, Hinds PW. (2006). Cancer
Cell 9: 13–22.
Lavoie JN, L’Allemain G, Brunet A, Muller R, Pouyssegur J.
(1996). J Biol Chem 271: 20608–20616.
Leng X, Noble M, Adams PD, Qin J, Harper JW. (2002). Mol
Cell Biol 22: 2242–2254.
Lin J, Beerm DG. (2004). Semin Oncol 31: 476–486.
Lovec H, Grzeschiczek A, Kowalski MB, Moroy T. (1994a).
EMBO J 13: 3487–3495.
Lovec H, Sewing A, Lucibello FC, Muller R, Moroy T.
(1994b). Oncogene 9: 323–326.
Lu F, Gladden AB, Diehl JA. (2003). Cancer Res 63:
7056–7061.
Lukas J, Jadayel D, Bartkova J, Nacheva E, Dyer MJ, Strauss
M et al. (1994). Oncogene 9: 2159–2167.
Moreno-Bueno G, Rodriguez-Perales S, Sanchez-Estevez C,
Hardisson D, Sarrio D, Prat J et al. (2003). Oncogene 22:
6115–6118.
Muraoka RS, Lenferink AE, Law B, Hamilton E, Brantley
DM, Roebuck LR et al. (2002). Mol Cell Biol 22: 2204–2219.
Muraoka RS, Lenferink AE, Simpson J, Brantley DM,
Roebuck LR, Yakes FM et al. (2001). J Cell Biol 153:
917–932.
Neuman E, Ladha MH, Lin N, Upton TM, Miller SJ,
DiRenzo J et al. (1997). Mol Cell Biol 17: 5338–5347.
Park IK, Roach P, Bondor J, Fox SP, DePaoli-Roach AA.
(1994). J Biol Chem 269: 944–954.
Petre CE, Wetherill YB, Danielsen M, Knudsen KE. (2002).
J Biol Chem 277: 2207–2215.
Quelle DE, Ashmun RA, Shurtleff SA, Kato JY, Bar-Sagi D,
Roussel MF et al. (1993). Genes Dev 7: 1559–1571.
Rimerman RA, Gellert-Randleman A, Diehl JA. (2000). J Biol
Chem 275: 14736–14742.
Rodriguez-Viciana P, Warne PH, Dhand R, Vanhaesebroeck
B, Gout I, Fry MJ et al. (1994). Nature 370: 527–532.
Sears R, Nuckolls F, Haura E, Taya Y, Tamai K, Nevins JR.
(2000). Genes Dev 14: 2501–2514.
Sherr CJ, Roberts JM. (1999). Genes Dev 13: 1501–1512.
Solomon DA, Wang Y, Fox SR, Lambeck TC, Giesting S,
Lan Z et al. (2003). J Biol Chem 278: 30339–30347.
Summers MD, Smith GE. (1987). Tex Agric Exp St Bull 1555:
1–56.
ter Haar E, Coll JT, Austen DA, Hsiao HM, Swenson L,
Jain J. (2001). Nat Struct Biol 8: 593–596.
Welcker M, Orian A, Jin J, Grim JA, Harper JW, Eisenman
RN et al. (2004). Proc Natl Acad Sci USA 101: 9085–9090.
Welcker M, Singer J, Loeb KR, Grim J, Bloecher A, Gurien-
West M et al. (2003). Mol Cell 12: 381–392.
Winston JT, Coats SR, Wang YZ, Pledger WJ. (1996).
Oncogene 12: 127–134.
Yost C, Farr III GH, Pierce SB, Ferkey DM, Chen MM,
Kimelman D. (1998). Cell 93: 1031–1041.
Yu QSE, Geng Y, Ahnstrom M, Zagozdzon A, Kong Y,
Gardner H et al. (2006). Cancer Cell 9: 23–32.
Zhang F, Phiel CJ, Spece L, Gurvich N, Klein PS. (2003).
J Biol Chem 278: 33067–33077.
Zwijsen RM, Wientjens E, Klompmaker R, van der Sman J,
Bernards R, Michalides RJ. (1997). Cell 88: 405–415.
Constitutively nuclear cyclin D1 in human cancer
S Benzeno et al
6303
Oncogene
    • "Second binding site contain sixteen residues (Ser3, Glu4, Lys5, Thr6, Phe7, Arg10, Glu36, Arg37, Tyr38, Lys39, Gly40, Glu41, Lys42, Gly85, His86, Val112), while third binding site contain ten residues (Thr 6, Phe 7, Lys 8, Ile34, Ile35, Glu 36, Gly85, Tyr 110, Met111, Val 112). COACH server predicted 88 residues being potential sites involved in binding including 6, 7, 8, 10, 11, 14, 18, 19, 20, 22, 23, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 57, 58, 59, 60, 62, 63, 64, 66, 67, 69, 70, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 89, 99, 104, 108, 103, 105, 107, 109, 110, 111, 112, 113, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125). Further analyses revealed that fifteen high risk missense SNPs were located in these binding sites (11, 32, 33, 37, 40, 68, 70, 78, 79, 104, 113, 117, 120) (see Figure 7B). "
    [Show abstract] [Hide abstract] ABSTRACT: Autophagy, an evolutionary conserved multifaceted lysosome-mediated bulk degradation system, plays a vital role in liver pathologies including hepatocellular carcinoma (HCC). Post-translational modifications (PTMs) and genetic variations in autophagy components have emerged as significant determinants of autophagy related proteins. Identification of a comprehensive spectrum of genetic variations and PTMs of autophagy related proteins and their impact at molecular level will greatly expand our understanding of autophagy based regulation. In this study, we attempted to identify high risk missense mutations that are highly damaging to the structure as well as function of autophagy related proteins including LC3A, LC3B, BECN1 and SCD1. Number of putative structural and functional residues, including several sites that undergo PTMs were also identified. In total, 16 high-risk SNPs in LC3A, 18 in LC3B, 40 in BECN1 and 43 in SCD1 were prioritized. Out of these, 2 in LC3A (K49A, K51A), 1 in LC3B (S92C), 6 in BECN1 (S113R, R292C, R292H, Y338C, S346Y, Y352H) and 6 in SCD1 (Y41C, Y55D, R131W, R135Q, R135W, Y151C) coincide with potential PTM sites. Our integrated analysis found LC3B Y113C, BECN1 I403T, SCD1 R126S and SCD1 Y218C as highly deleterious HCC-associated mutations. This study is the first extensive in silico mutational analysis of the LC3A, LC3B, BECN1 and SCD1 proteins. We hope that the observed results will be a valuable resource for in-depth mechanistic insight into future investigations of pathological missense SNPs using an integrated computational platform.
    Full-text · Article · Jan 2017
    • "Recently, large-scale studies have shown that damaged PTMs caused by numerous inherited and somatic amino acid substitutions [10] have profound impact on both gene and protein function [11], and they are associated with human cancer [12]. One instance is that mutation S215R occurring on the PTMs of TP53 could result in breast cancer [13]; another is mutation of T286 in cyclin D1 (CCND1) causing the loss of phosphorylation of T286 is involved in nuclear accumulation of cyclin D1 in esophageal cancer [14]. However, some of these previous studies concluded the relationship between damaged PTMs and human health based on predications; some focused only on cancers and many focused on only unique type of PTM. "
    [Show abstract] [Hide abstract] ABSTRACT: Protein posttranslational modifications (PTMs) play key roles in a variety of protein activities and cellular processes. Different PTMs show distinct impacts on protein functions, and normal protein activities are consequences of all kinds of PTMs working together. With the development of high throughput technologies such as tandem mass spectrometry (MS/MS) and next generation sequencing, more and more nonsynonymous single-nucleotide variations (nsSNVs) that cause variation of amino acids have been identified, some of which result in the damage of PTMs. The damaged PTMs could be the reason of the development of some human diseases. In this study, we elucidated the proteome wide relationship of eight damaged PTMs to human inherited diseases and cancers. Some human inherited diseases or cancers may be the consequences of the interactions of damaged PTMs, rather than the result of single damaged PTM site.
    Full-text · Article · Oct 2015
    • "Since cyclin D1 proteasomal degradation has been regarded as one of important anti-cancer mechanisms (Spinella et al., 1999; Mukhopadhyay et al., 2002; Huang et al., 2005 ), cyclin D1 proteasomal degradation may be one of the molecular targets for the anti-cancer activity of NAR. Threonine-286 phosphorylation is associated with cyclin D1 phosphorylation and its mutation is observed in human cancer cells (Benzeno et al., 2006 ). Thus, failure of phosphorylationdependent degradation of cyclin D1 may contribute to the development of cancer. "
    [Show abstract] [Hide abstract] ABSTRACT: Naringenin (NAR) as one of the flavonoids observed in grapefruit has been reported to exhibit an anti-cancer activity. However, more detailed mechanism by which NAR exerts anti-cancer properties still remains unanswered. Thus, in this study, we have shown that NAR down-regulates the level of cyclin D1 in human colorectal cancer cell lines, HCT116 and SW480. NAR inhibited the cell proliferation in HCT116 and SW480 cells and decreased the level of cyclin D1 protein. Inhibition of proteasomal degradation by MG132 blocked NAR-mediated cyclin D1 downregulation and the half-life of cyclin D1 was decreased in the cells treated with NAR. In addition, NAR increased the phosphorylation of cyclin D1 at threonine-286 and a point mutation of threonine-286 to alanine blocked cyclin D1 downregulation by NAR. p38 inactivation attenuated cyclin D1 downregulation by NAR. From these results, we suggest that NAR-mediated cyclin D1 downregulation may result from proteasomal degradation through p38 activation. The current study provides new mechanistic link between NAR, cyclin D1 downregulation and cell growth in human colorectal cancer cells.
    Full-text · Article · Jul 2015
Show more