Bimodal degradation of MLL by SCFSkp2
and APCCdc20assures cell cycle
execution: a critical regulatory circuit
lost in leukemogenic MLL fusions
Han Liu,1,2Emily H.-Y. Cheng,1,2,3and James J.-D. Hsieh1,2,4,5
1Molecular Oncology, Department of Medicine, Washington University School of Medicine, St. Louis, Missouri 63110, USA;
2Siteman Cancer Center, Washington University School of Medicine, St. Louis, Missouri 63110, USA;3Department
of Pathology and Immunology, Washington University School of Medicine, St. Louis, Missouri 63110, USA;4Department
of Molecular Biology and Pharmacology, Washington University School of Medicine, St. Louis, Missouri 63110, USA
Human chromosome 11q23 translocations disrupting MLL result in poor prognostic leukemias. It fuses
the common MLL N-terminal ∼1400 amino acids in-frame with >60 different partners without shared
characteristics. In addition to the well-characterized activity of MLL in maintaining Hox gene expression,
our recent studies established an MLL–E2F axis in orchestrating core cell cycle gene expression including
Cyclins. Here, we demonstrate a biphasic expression of MLL conferred by defined windows of degradation
mediated by specialized cell cycle E3 ligases. Specifically, SCFSkp2and APCCdc20mark MLL for degradation
at S phase and late M phase, respectively. Abolished peak expression of MLL incurs corresponding defects in
G1/S transition and M-phase progression. Conversely, overexpression of MLL blocks S-phase progression.
Remarkably, MLL degradation initiates at its N-terminal ∼1400 amino acids, and tested prevalent MLL
fusions are resistant to degradation. Thus, impaired degradation of MLL fusions likely constitutes the
universal mechanism underlying all MLL leukemias. Our data conclude an essential post-translational
regulation of MLL by the cell cycle ubiquitin/proteasome system (UPS) assures the temporal necessity
of MLL in coordinating cell cycle progression.
[Keywords: MLL; Taspase1; Skp2; Cdc20; cell cycle; leukemia]
Supplemental material is available at http://www.genesdev.org.
Received May 22, 2007; revised version accepted August 3, 2007.
The eukaryotic cell cycle consists of a tightly orches-
trated circular progression of a sequence of distinct
phases—namely, G1, S, G2, and M—designated for the
execution of inherent genetic programs (Nurse 2000).
Key regulatory components of the mammalian cell cycle
machinery include E2Fs, Rbs, Cyclins, Cyclin-dependent
kinases (CDKs), and CDK inhibitors (CDKIs), which
form complex positive and negative epistatic circuits en-
suring accurate cell cycle progression (Murray 2004). The
molecular blueprint of a normal cell cycle details the
central controls held by a series of CDKs that positively
influences proliferation through phosphorylating target
proteins such as Rbs (Morgan 1997). This releases the
repressive activities of Rbs on E2Fs, the major executors
of cell cycle gene expression programs (Dyson 1998;
Nevins 2001; Trimarchi and Lees 2002; Blais and Dyn-
lacht 2004; Bracken et al. 2004; Giacinti and Giordano
2006). Individual CDKs need to complex with special-
ized Cyclins to form catalytically active Cyclin/CDKs;
therefore, their activities depend on the availability of
involved Cyclins. Additional regulations of Cyclin/
CDKs come from negative factors, the CDKIs (Sherr and
Roberts 1999). Levels of Cyclins and CDKIs undulate
during cell division, which in part is due to the defined
windows of degradation by the ubiquitin–proteasome
system (UPS) (Reed 2003). UPS commences at substrate
recognition and the subsequent covalent conjugation of
ubiquitin by the responsible E3 ligases, followed by deg-
radation executed by the 26S proteasome (Hershko
2005). The two major E3 complexes involved in proteo-
lyzing core components of the cell cycle machinery are
SCF (Skp1–Cul1–F-box protein) and APC (anaphase-
promoting complex/cyclosome) complexes, the sub-
strate recognition of which is conducted by the variable
components—F-box proteins for SCF, and Cdc20 and
Cdh1 for APC (Cardozo and Pagano 2004; Nakayama
and Nakayama 2006; Peters 2006). Thus, two types of
post-translational modifications—phosphorylation and
E-MAIL email@example.com; FAX (314) 362-1589.
Article is online at http://www.genesdev.org/cgi/doi/10.1101/gad.1574507.
GENES & DEVELOPMENT 21:2385–2398 © 2007 by Cold Spring Harbor Laboratory Press ISSN 0890-9369/07; www.genesdev.org 2385
ubiquitination—constitute the two mainstream controls
of the cell cycle.
Polycomb group (PcG) and trithorax group (trxG) pro-
teins are chromatin modifiers required for the epigenetic
maintenance of repressive versus active states of cell fate
specification genes (Schumacher and Magnuson 1997;
Yu et al. 1998; Hanson et al. 1999; Ringrose and Paro
2004; Schuettengruber et al. 2007). The most-character-
ized targets are the Hox genes in vertebrates and homeo-
tic genes in invertebrates. Beyond perpetuating cellular
memory through multiple cell divisions, recent studies
on mammals and flies highlighted the active, dynamic
participation of PcG proteins in the transcriptional
control of cell cycle genes such as the p16Ink4a/ARF lo-
cus and Cyclin A (Jacobs et al. 1999a,b; Martinez and
Cavalli 2006; Martinez et al. 2006). For example, Bmi-1
can function as an oncogene through its positive influ-
ence on cell proliferation by suppressing p16Ink4a(Jacobs
et al. 1999b). On the other hand, genetic evidence also
supports certain PcG proteins, such as Mel-18, as tumor
suppressors in that they negatively regulate proliferation
(Guo et al. 2007). Despite ample elucidation on the inti-
mate relationship between PcG and the cell cycle, we are
only at the beginning in realizing a similar cell cycle
duty carried out by the trxG members.
MLL, the mammalian ortholog of Drosophila tritho-
rax, is the founding member of the trxG proteins (Gu et
al. 1992; Tkachuk et al. 1992; Djabali et al. 1993; Domer
et al. 1993; Thirman et al. 1993). Genetic studies con-
firmed the evolutionarily conserved role of trx/MLL in
maintaining homeotic/Hox gene expression, therefore
specifying segmental identity (Yu et al. 1995; Ringrose
and Paro 2004). Recurrent human chromosome band
11q23 translocations disrupting MLL cause leukemias of
poor prognosis. MLL leukemic cells carry pathogno-
monic MLL fusion proteins resulting from an in-frame
fusion between the N-terminal ∼1400 amino acids of
MLL with a wide spectrum of translocation partners,
ranging from nuclear transcription factors to cytoplas-
mic structural proteins (Rowley 1998; Ayton and Cleary
2001; Canaani et al. 2004; Daser and Rabbitts 2004;
Gilliland et al. 2004; Hess 2004; Eguchi et al. 2005). As
implicated by the broad diversity of translocation part-
ners, elegant murine models specifically examined their
possible dispensability and nonspecificity (Corral et al.
1996; Dobson et al. 2000; Wang et al. 2005). Remarkably,
mice engineered to carry the MLL-lacZ but not the MLL-
Myc tag allele developed leukemia, confirming the in-
dispensability and, yet, nonspecificity of these partners
(Dobson et al. 2000). Although certain shared properties
have been recognized among some fusion partners (Slany
et al. 1998; Martin et al. 2003; So et al. 2003), there is no
universally applicable mechanism identified.
MLL encodes a 3969-amino-acid nuclear protein con-
sisting of multiple conserved domains with distinct
biological properties. The most well-characterized bio-
chemical property of MLL is its ability to methylate his-
tone H3 at Lys 4 through the conserved SET domain at
the very C terminus (Milne et al. 2002; Nakamura et al.
2002). The complexity of MLL gene regulation was fur-
ther illustrated when we and others showed that the
500-kDa full-length MLL precursor (MLLFL) undergoes
evolutionarily conserved site-specific proteolysis to gen-
erate a mature MLLN320/C180heterodimer, consisting of
processed N-terminal 320-kDa and C-terminal 180-kDa
fragments (Nakamura et al. 2002; Yokoyama et al. 2002;
Hsieh et al. 2003b). Cleavage of MLL at QL(V)D/GXDD
sites is mediated by Taspase1 (threonine aspartase 1)
(Hsieh et al. 2003a), which is required for the full activity
of MLL (Takeda et al. 2006). In other words, unprocessed
MLL precursor functions as a hypomorphic allele (Ta-
keda et al. 2006).
The discovery of Taspase1 initiated a novel class of
endopeptidases that utilizes its N-terminal threonine of
the mature ? subunit to proteolyze substrates after an
aspartate (Hsieh et al. 2003a). Taspase1 is translated as a
proenzyme that undergoes autoproteolysis to generate a
mature ?/? heterodimer that displays an overall ?/?/?/?
structure (Khan et al. 2005). In addition to its antici-
pated roles in Hox gene regulation, our recent studies on
Taspase1−/−mice linked MLL proteolysis to the heart of
cell cycle gene programming (Takeda et al. 2006). We
showed that MLL directly participates in cell cycle pro-
gression through an MLL–E2F axis that controls expres-
sion of Cyclins E, A, and B. In the absence of Taspase1,
MLL remains as a precursor with reduced histone H3 K4
methyltransferase (HMT) activity (Takeda et al. 2006).
Here we further examine how MLL controls the cell
cycle and whether its activity is as tightly regulated as
most of the key cell cycle regulators.
We provide evidence that MLL undergoes a specialized
bimodal degradation resulting in its biphasic expression
through the cell cycle. This unique expression is con-
ferred by SCFSkp2and APCCdc20E3 ligases of the cell cy-
cle UPS. Importantly, deregulation of this organized ex-
pression, either through overexpression or knockdown,
incurs corresponding cell cycle defects. Therefore, the
observed biphasic appearance of MLL is essential in as-
suring the correct execution of inherent genetic pro-
grams underlying every single cell division. Our model
predicts that tight control of MLL activity is essential in
cell proliferation, and deregulated expression of MLL im-
pedes normal cell proliferation. Thus, it explains pub-
lished seemingly contradictory positive or negative regu-
latory activity of MLL/MLL fusion in cell proliferation
(Muyrers-Chen et al. 2004; Milne et al. 2005; Xia et al.
2005; Takeda et al. 2006). Furthermore, MLL degradation
signals through its N-terminal ∼1400 amino acids, which
are universally retained in MLL leukemia fusions. Im-
portantly, tested prevalent MLL fusions—including
MLL-AF4, MLL-AF9, MLL-ENL, and MLL-ELL, and leu-
kemogenic MLL-lacZ—all exhibit diminished interac-
tion with Skp2 and Cdc20, resulting in resistance to deg-
radation mediated by the respective cell cycle UPS and a
constant presence through the cell cycle phase transi-
tion. Here, we uncover a functional commonality among
structurally diversified fusion partners and propose a
model in which fusion partners interfere with the recog-
nition of respective MLL fusions by the cell cycle E3
ligases, resulting in their nonoscillating expression. We
Liu et al.
2386GENES & DEVELOPMENT
hypothesize that loss of the biphasic expression of MLL
fusions represents the long-sought-after universal defect
underlying all MLL leukemias. Complementary to our
demonstrated MLL–E2F axis utilized by MLL in activat-
ing Cyclin expression, the current study recognizes a
critical post-translational regulation of MLL by the cell
cycle degradation machinery that is lost in MLL fusions.
Biphasic expression of MLL through the cell cycle
To determine whether the activity of MLL, like other
key cell cycle regulators, needs to be temporally regu-
lated, we examined MLL protein expression through pro-
gressive phases of the cell cycle. HeLa cells were first
synchronized at G1/S and early M phases with double
thymidine and nocodazole, respectively, before being
subjected to release for the indicated periods of time.
Cells at distinct phases of the cell cycle were confirmed
by fluorescence-activated cell sorter (FACS) analysis fol-
lowing propidium iodine (PI) staining for DNA contents
(Fig. 1A,B, left panels) or by fluorescent microscopy (Fig.
1C). Remarkably, we uncovered a unique biphasic ex-
pression pattern of MLL protein when cells progressed
through individual phases (Fig. 1A,B, right panels). Both
N-terminal (MLLN320) and C-terminal (MLLC180) MLL
fragments exhibited a coordinated expression pattern:
down-regulation at S and late M phases following respec-
tive peaks at G1/S and G2/M transitions (Fig. 1A,B). The
degradation of MLL in late M phase was further illus-
trated by fluorescent microscopy (Fig. 1C). The same
biphasic expression pattern was observed in human em-
bryonic kidney 293T cells and human osteosarcoma
U2OS cells, irrespective of the synchronization tech-
niques utilized, such as centrifugal elutriation and mi-
mosine block (Supplementary Fig. S1; data not shown).
These data unveil a previously unrecognized temporal
regulation of MLL activity during the cell cycle progres-
MLL is degraded by the UPS during the cell cycle
To investigate the mechanism(s) underlying the biphasic
expression of MLL, we examined the MLL mRNA level
in cells at different phases of the cell cycle by quantita-
tive RT–PCR (qRT–PCR). The amounts of MLL tran-
scripts did not correlate with its protein levels (Fig. 2A),
suggesting a dominating post-transcriptional control. As
protein degradation constitutes the major post-transla-
tional control of cell cycle regulators, we examined
whether cell cycle phase-specific degradation occurs.
Immunoprecipitated Flag-MLLN320/C180was incubated
with cellular extracts purified from the indicated phases
of the cell cycle in an attempt to recapitulate the in vivo
degradation of MLL. Degradation of immunoprecipitated
Flag-MLLN320/C180was then monitored by anti-Flag im-
munoblots. MLL was degraded by extracts derived from
S- and late M-phase cells but not from G1-, G2-, or early
M-phase cells (Fig. 2B). These data indicate the utiliza-
tion of specialized degradation machinery in controlling
MLL protein expression. Since a specialized UPS consti-
tutes the major protein degradation machinery of the cell
cycle, we tested whether the addition of MG132, a pro-
teasome inhibitor, will increase the protein half-life of
MLL in the presence of cyclohexamide, which blocks
translation. Indeed, a nearly eightfold extension of its
half-life was observed (Fig. 2C). The cell cycle UPS
marks proteins for degradation through polymerized
ubiquitin conjugation. To determine whether MLL also
undergoes UPS-mediated post-translational modifica-
tions, HA-Ubiquitin and Flag-MLL were coexpressed in
cells followed by anti-Flag immunoprecipitation. Flag-
MLL was covalently modified by HA-Ubiquitin and an
increased abundance of the ubiquitin-modified high-mo-
lecular-weight MLL was observed when MG132 was
added (Fig. 2D).
Specialized cell cycle E3 ligases mediate cognate
degradation of MLL
The two major E3 complexes of UPS involved in degra-
dation of core cell cycle regulators are SCF and APC
complexes, the substrate recognition of which is con-
ducted by the variable components—F-box proteins for
SCF, and Cdc20 and Cdh1 for APC (Nakayama and Na-
kayama 2006). The bimodal degradation of MLL at S and
late M phases suggests the involvement of both SCF and
APC for respective degradation. We first examined
whether overexpression of any of the major substrate
recognition F-box proteins of SCF—including Fbw7,
Skp2, and ?TrCP—affects the expression of endogenous
MLL. Among tested F-box proteins, only Skp2 enhanced
the degradation of MLL (Fig. 3A). Similarly, Cdc20 but
not Cdh1 degraded MLL in vivo (Fig. 3A). To determine
whether Skp2 and Cdc20 directly target MLL for degra-
dation, we probed the direct interactions between these
two substrate recognition modules and MLL by coimmu-
noprecipitation assays in the presence of MG132. Indeed,
direct intracellular interactions were identified between
endogenous MLL and Skp2 as well as MLL and Cdc20
(Fig. 3B). In addition to the observed in vivo interac-
tions, an in vitro degradation assay was employed to con-
firm the direct degradation of MLLN320/C180by SCFSkp2
and APCCdc20. Protein lysates prepared from synchro-
nized S- and late M-phase cells were incubated with
equal amounts of immunoprecipitated Flag-MLLN320/C180
for 1 h at 30°C followed by anti-Flag Western blot analy-
sis. Since protein modifications such as phosphorylation
are known important signals required for substrate rec-
ognition/degradation by the cell cycle UPS, the utiliza-
tion of immunoprecipitated Flag-MLLN320/C180as test
substrates abrogates the need for in vitro modifications
before being subjected to in vitro degradation assays.
Both S- and late M-phase lysates efficiently degraded
Flag-MLLN320/C180(Fig. 3C). Importantly, MLL degrada-
tion was impaired when SCFSkp2and APCCdc20were
immunodepleted, confirming the direct degradation of
MLLN320/C180at S and late M phases by SCFSkp2and
APCCdc20, respectively (Fig. 3C). These data demonstrate
SCFSkp2and APCCdc20keep MLL in check
GENES & DEVELOPMENT2387
direct degradation of MLL by SCFSkp2and APCCdc20, and
predict an increase of MLL protein when the expres-
sion of Skp2 or Cdc20 is suppressed. To test this hypoth-
esis, small interfering RNA (siRNA)-mediated knock-
down of Skp2 and/or Cdc20 was performed to determine
whether the MLL protein level is altered. Singular
knockdown of Skp2 or Cdc20 led to an increase of the
MLL level, and simultaneous knockdown of Skp2 and
Cdc20 resulted in a robust ∼3.5-fold increase of MLLC180
and a 1.7-fold increase of MLLN320. The observed lesser
increase of MLLN320is likely due to the known poor gel
transfer of MLLN320onto a membrane. Furthermore, it
may also reflect the more unstable nature of the N-ter-
minal MLL (Hsieh et al. 2003b). Taken together, our
studies detail the blueprint underlying MLL degradation
through the cell cycle, confirming SCFSkp2and APCCdc20
as two major E3 complexes in orchestrating MLL expres-
sion. Demonstrated regulation of MLL by the cell cycle
UPS raises the possibility that the temporal necessity
of MLL for gene activation is of quintessential impor-
tance in executing inherent programs for cell cycle pro-
strated two individual peaks of MLL expression. HeLa cells were synchronized at the G1/S transition by double thymidine block
followed by the release into thymidine-free media for the indicated time. (B) Nocodazole block and release of HeLa cells identified the
same two peak expressions of MLL as A. (A,B) The levels of MLLN320and MLLC180were determined with anti-NT and anti-CT
immunoblots, respectively. (A,B, left panels) The successful synchronization and subsequent release from indicated chemical inhibitor
was confirmed by FACS analyses of the DNA contents and by immunoblots using the anti-phospho-histone H3 Ser-10 antibody, which
marks mitosis. (A,B) The expression of ?-actin served as a loading control. (C) MLL protein is decreased at late M phase. Immuno-
fluorescence studies were performed on asynchronous HeLa cells using anti-MLL (red) and anti-phospho-histone H3 S10 (green)
antibodies. Individual phase assignments were based on nuclear morphology, with representative photos presented.
Biphasic expression of MLL during cell cycle progression. (A) Double thymidine block and release of HeLa cells demon-
Liu et al.
2388 GENES & DEVELOPMENT
Concerted expression of MLL is necessary for proper
cell cycle progression
In addition to our prior report, which indicates direct
participation of MLL in the cell cycle gene expres-
sion program through the MLL–E2F axis (Takeda et al.
regulation of the MLL level conferred by the cell cycle
UPS. The next important questions are why the ac-
tivity of MLL needs to be temporally controlled and
whether deregulation of such an exquisite expres-
sion causes any cell cycle aberrations. To tackle these
questions, RNA interference (RNAi)-mediated knock-
down and overexpression of MLL were performed to
interrogate possible cell cycle defects. Retrovirus-
mediated stable knockdown of MLL in HeLa cells was
performed, and the consequences of failed MLL induc-
tion at G1/S and G2/M transitions were examined. Two
independent MLL-short hairpin RNA (shRNA) con-
structs stably reduced the expression of MLL as com-
pared with the scramble shRNA (scr-shRNA) control
(Fig. 4A; Supplementary Fig. S2). MLL knockdown
cells were impaired in proliferation—a nearly twofold
reduction of total cell number was observed after 6 d in
culture (Fig. 4B; Supplementary Fig. S2). To specifically
pinpoint the defect at G1/S transition, cells were syn-
chronized by double thymidine treatment before being
released into standard culture medium for the indi-
cated periods of time. The S-phase entrance rate was
evaluated with a 30-min pulse of BrdU, which specifi-
cally labels active DNA synthesis. Strikingly, a 1-h delay
in S-phase entry was observed in MLL-deficient cells
(Fig. 4C), indicating that the G1/S peak of MLL is re-
quired for proper G1/S transition. We then examined the
defect in M phase following the loss of the second MLL
peak. MLL-deficient HeLa cells were synchronized with
nocodazole in prometaphase before being released into
standard culture medium for the indicated periods of
time. Cellular DNA contents were stained with PI and
analyzed by FACS. In MLL knockdown cells, the im-
paired G2/M peak expression resulted in a marked delay
in M-phase progression (Fig. 4D,E). Thus, the up-regula-
tion of MLL in G1/S and G2/M transitions is apparently
required for S-phase entry and M-phase progression, re-
We next investigated why MLL needs to be degraded.
In other words: What are the consequences if MLL
continues to express at high levels? To answer this
question, GFP-MLL was transiently overexpressed in
293T cells, the cell cycle profile of which was monitored
by costaining with 7-AAD for DNA contents and BrdU
for active DNA synthesis. An increased appearance
(22.9%) of cells with an intermediate incorporation of
BrdU was identified in GFP-MLL-expressing cells, sug-
gestive of deterred replication (Fig. 4F). Taken together,
these data emphasize the importance of the biphasic ex-
pression of MLL in coordinating the cell cycle progres-
UPS. (A) The levels of MLL mRNA do not correlate
with the biphasic MLL protein expression through spe-
cific phases of the cell cycle. qRT–PCR was performed
on HeLa cells that were synchronized by double thymi-
dine treatment and released for the indicated periods of
time. MLL expression was normalized against GAPDH.
The average MLL expression in asynchronous HeLa
cells was arbitrarily assigned as 1.0. Values shown are
mean ± 1 SD obtained from three independent experi-
ments. (B) MLL is degraded at specific phases of the cell
cycle. On-bead in vitro degradation assays were carried
out by incubating purified Flag-MLLN320/C180with cel-
lular lysates prepared from cells synchronized at the
indicated phases of the cell cycle for 30 min at 30°C.
The abundance of Flag-MLL was analyzed by anti-Flag
immunoblots. (C) The degradation of MLL can be pre-
vented by inhibitors of the 26S proteasome. HeLa cells
were treated with 10 µg/mL cycloheximide for the in-
dicated periods of time in the presence of either DMSO
as a control or 10 µM MG132 to inactivate 26S protea-
some. (Left panel) The expression of MLL was evaluated
by anti-MLL immunoblots. The expression of ?-actin
served as a loading control. (Right panel) The half-life of
MLLN320/C180was determined by quantifying the levels
of normalized MLLC180. (D) MLL undergoes polyubiq-
uitination. 293T cells transfected with indicated ex-
pression vectors for 48 h were treated with MG132 or
MLL undergoes bimodal degradation by the
DMSO vehicle for an additional 24 h. Cells were lysed in RIPA buffer containing 20 mM N-ethylmaleimide, which inhibits deubiq-
uitination. Cellular extracts were subjected to anti-Flag immunoprecipitation, SDS-PAGE, and immunoblots with the indicated
SCFSkp2and APCCdc20keep MLL in check
GENES & DEVELOPMENT2389
The MLL N-terminal ∼1400 amino acids target
To further characterize how MLL degradation is regu-
lated, the region responsible for the observed degradation
was mapped. Flag-tagged constructs representing differ-
ent regions of MLL were cotransfected with either Skp2
or Cdc20, followed by examinations of their expressions
by anti-Flag immunoblots (Fig. 5A). The first ∼1400
amino acids of MLL were found to recapitulate the deg-
radation of MLLN320/C180(Fig. 5A). In contrast, MLLC180
alone was insensitive to degradation by coexpressed
Skp2 or Cdc20 (Fig. 5A). As endogenous MLLC180derived
from proteolytic processing of MLLFLprecursor het-
erodimerizes with MLLN320(Hsieh et al. 2003b) and is
susceptible to degradation, it is possible that MLLC180is
degraded through its interaction with MLLN320; i.e.,
MLLN320brings heterodimerized MLLC180to the 26S
proteasome for degradation. To test this hypothesis,
MLLN320(amino acids 1–2664) and MLLC180(amino ac-
ids 2772–3969) were cotransfected with or without spe-
cific E3 substrate recognition modules, Skp2 and Cdc20.
In accordance with our data on the degradation of endog-
enous MLLN320/C180(Fig. 3A), MLLC180was degraded
along with the cotransfected MLLN320in the presence of
either Skp2 or Cdc20 (Fig. 5B). If the MLL C terminus
was degraded while complexed with MLLN320, a MLL
C-terminal mutant that no longer interacts with
MLLN320would be resistant to such a degradation. We
employed a mutant MLL C terminus (MLLC180?F/SET)
that lacks FYRC and SET domains—a region that medi-
ates its heterodimerization with MLLN320(Hsieh et al.
2003b). When Flag-MLLC180?F/SET(amino acids 2722–
3620) was coexpressed with Flag-MLLN320(amino acids
1–2664) plus either Myc-Skp2 or Myc-Cdc20, its level
remained unchanged (Fig. 5B). This was in stark contrast
to the nearly complete disappearance of Flag-MLLC180
(Fig. 5B). Thus far, our data establish a model in which
the very N-terminal ∼1400 amino acids of MLL are re-
sponsible for the coordinated biphasic control of
MLLN320/C180. Since all of the MLL leukemia fusions
retain the degradation-prone region of MLL (Rowley
1998; Ayton and Cleary 2001; Canaani et al. 2004; Daser
and Rabbitts 2004; Gilliland et al. 2004; Hess 2004), it is
possible that the fusion partners confer resistance to deg-
radation to initiate the first universal insult underlying
all MLL leukemias.
MLL fusions are refractory to the cell cycle UPS,
resulting in stable expression through the cell cycle
The findings that the common MLL N-terminal ∼1400
amino acids are responsible for the degradation of
MLLN320/C180are consistent with our prior observation
that MLLN320is less stable (Hsieh et al. 2003b). These
data rekindle a prior proposal that fusion partners may
contribute to leukemogenesis in part by stabilizing the
N-terminal MLL (Dobson et al. 2000; Hsieh et al. 2003b).
We next examined whether leukemogenic MLL fusions
are refractory to degradation. Indeed, tested preva-
lent MLL fusions, including MLL-AF4, MLL-AF9, MLL-
ENL and MLL-ELL, were all resistant to SCFSkp2- and
APCCdc20-mediated degradation (Fig. 6A). Remarkably,
the same resistance to degradation was also observed
with the leukemogenic MLL-lacZ but not the nonleuke-
mogenic MLL-Myc tag fusion (Fig. 6A). Subsequent
experiments were performed to determine whether en-
dogenous MLL fusions undergo cell cycle-coordinated
SCFSkp2and APCCdc20. (A) Specialized substrate recognition
modules (Skp2 and Cdc20) of the cell cycle E3 ligases are re-
quired for the degradation of MLL. Specific degradation of the
endogenous MLL protein occurred in cells with overexpression
of Skp2 or Cdc20. 293T cells were transfected with the indi-
cated Myc-tagged substrate recognition module of cell cycle E3
ligases, and the levels of endogenous MLL were detected with
the indicated anti-MLL antibodies. The level of ?-actin served
as a loading control. (B) MLL interacts with Skp2 and Cdc20 in
vivo. Cellular lysates of 293T cells treated with MG132 were
subjected to immunoprecipitation (IP) with anti-MLL antibody.
Coprecipitated Skp2 and Cdc20 were detected by immunoblots.
Protein-A beads served as a negative control. (C) In vitro degra-
dation assays confirm the direct recognition and subsequent
destruction of MLL by SCFSkp2and APCCdc20in S and late M
phases, respectively. Immunoprecipitated Flag-MLL was incu-
bated with the indicated cellular lysates obtained from synchro-
nized HeLa cells at 30°C for 30 min before being subjected to
Western blot analyses. Anti-Skp2 or anti-Cdc20 immunopre-
cipitation was performed to deplete SCFSkp2or APCCdc20from
the indicated lysates. Mouse IgG was used as mock depletion.
(D) Knockdown of Skp2 and/or Cdc20 leads to the accumulation
of MLL protein. HeLa cells were transfected with the indicated
siRNA specific for Skp2 or Cdc20. Protein levels of endogenous
Skp2, Cdc20, and MLL were analyzed by Western blots at 72 h
after transfection and quantified using ImageGauge software
(FujiFilm). The level of ?-actin served as a loading control, and
the expression of MLL in control knockdown (siScr, si scramble)
was arbitrarily assigned as 1.0.
The bimodal degradation of MLL is mediated by
Liu et al.
2390GENES & DEVELOPMENT
degradation in various MLL leukemia cell lines. Inter-
estingly, unlike commonly utilized tissue culture cell
lines such as HeLa, 293T, and U2OS cells, we were un-
able to use mimosine, double thymidine, or nocodazole
to synchronize available MLL leukemia cells, including
MLL-AF4-bearing SEM and RS(4;11) and MLL-AF9-bear-
ing THP-1, NOMO-1, and MOLM13 cell lines. This sur-
prising observation suggests inherent cell cycle defects
of these MLL leukemia cells. One possible scenario is
that these cells have cycle checkpoint defects that confer
resistance to chemical-induced synchronization. Never-
theless, individual populations representing distinct
phases of the cell cycle were separated using centrifugal
elutriation based on their physical properties. The purity
of each population was assessed by FACS analysis of
DNA contents (Fig. 6B). Remarkably, we witnessed a
loss of the unique biphasic expression of MLL-AF4 in
SEM human MLL leukemia cells (Fig. 6B). In contrast to
the biphasic expression of wild-type MLL, MLL-AF4 pro-
tein did not fluctuate through the cell cycle while its
level was not markedly increased (Fig. 6B). As over-
expression of MLL fusion is likely to block cell cycle
progression and induce apoptosis (Muyrers-Chen et al.
2004; Xia et al. 2005), additional compensatory mecha-
nism(s) must have evolved to maintain the level of MLL
fusions within a tolerable range. Since all of the MLL
fusions contain the common N-terminal ∼1400 amino
acids that signal degradation, it is possible that indi-
vidual fusion partners directly interfere with the recog-
nition of the MLL N terminus by the cell cycle E3 li-
gases. Coimmunoprecipitation assays were performed in
cells transfected with indicated Flag-MLL fusions to ex-
amine this hypothesis. Indeed, diminished interactions
between tested MLL fusions and endogenous Skp2 or
Cdc20 were observed (Fig. 6C). To further demonstrate
the refractoriness of MLL fusions to the cell cycle UPS,
in vitro degradation assays were performed on immuno-
precipitated MLL-AF4. In contrast to the nearly com-
plete degradation of MLL by S- or late M-phase extracts,
no significant degradation of MLL-AF4 was observed
(Fig. 6D). In conclusion, we discovered a common defect
associated with all of the tested prevalent MLL fusions
that were not previously appreciated. The resistance of
MLL fusions to preprogrammed degradation likely con-
corresponding cell cycle progression defects. (A) Stable
knockdown of MLL in HeLa cells was achieved with
shRNA against MLL. HeLa cells were infected with
retrovirus-expressing shRNA targeting MLL (MLL-
shRNA) or control scramble shRNA (Scr-shRNA). The
successful suppression of MLL expression was con-
firmed by anti-MLL Western blots. (B) Depletion of
MLL impairs cell proliferation. The proliferation curves
of HeLa cells with or without MLL knockdown are pro-
vided. Replicate cultures of 10,000 cells were plated and
trypsinized at the indicated time points to obtain cell
numbers. Values shown are mean ± 1 SD obtained from
three independent experiments. (C) A marked impair-
ment in S-phase entry is observed in MLL knockdown
cells. HeLa cells with or without MLL knockdown were
synchronized at the G1/S transition by double thymi-
dine block, followed by a release for the indicated time
points. Cells were chased with BrdU for 30 min before
collections, and S-phase entry was assessed by the posi-
tive BrdU staining determined by FACS. (D,E) MLL is
required for proper M-phase progression. HeLa cells
with MLL or control knockdown were synchronized at
the prometaphase by nocodazole block plus shake-off
procedure before being released into regular culture me-
dium. Cells were stained with PI for DNA content be-
fore being subjected to FACS analyses to monitor phase
progression of the cell cycle. Quantification of the M-
phase progression from serial graphs presented in E is
presented in D. (F) Overexpression of MLL induces an
S-phase block. 293T cells were transfected with GFP-
MLL or GFP-expressing plasmids for 48 h before being
subjected to a 30-min BrdU pulse immediately before
analyses. GFP-positive cells were gated for analyses of
BrdU incorporation and DNA content. Additional con-
trols using untransfected cells were included. (C–F) Re-
sults presented are representative data of three indepen-
Deregulation of MLL protein level incurs
SCFSkp2and APCCdc20keep MLL in check
GENES & DEVELOPMENT2391
stitutes the first universal insult committed by all MLL
fusions, leading to the ultimate development of frank
MLL and the cell cycle machinery
MLL, the mammalian homolog of Drosophila trithorax,
is best known for its positive regulation of Hox gene
expression. Homozygous disruption of MLL in mice re-
sults in early embryonic lethality, and heterozygous de-
ficiency results in homeotic transformation due to im-
paired maintenance of Hox genes (Yu et al. 1995). The
early lethality of MLL−/−mouse embryos precludes de-
tailed investigation of its involvement in other signaling
pathways. Our demonstration that MLL is regulated by
Taspase1-mediated proteolytic cleavage broadens the av-
enue in studying MLL in that noncleaved precursor MLL
functions as a hypomorphic allele. Although the under-
lying mechanisms were not further investigated, initial
characterizations on MLL−/−mice suggest a role of MLL
tion of MLLN320/C180. (A) The N-terminal 1400 amino acids of
MLL are targeted by SCFSkp2and APCCdc20for degradation.
Various Flag-MLL fragments were coexpressed in 293T cells
with either Myc-Skp2 or Myc-Cdc20. The protein levels of
various MLL fragments were determined by anti-Flag Western
blots. The diagram on the right depicts the domain composi-
tions of transfected MLL fragments. Flag-MLL(1–3969) served as
a positive control. (B) MLLN320brings MLLC180to degradation.
Flag-MLLN320plus Flag-MLLC180or Flag-MLLN320plus Flag-
MLLC180?F/SETwere coexpressed in 293T cells with either Myc-
Skp2 or Myc-Cdc20 before being subjected to indicated Western
The MLL N terminus is responsible for the degrada-
mediated degradation. (A) Prevalent MLL fusion proteins, in-
cluding MLL-AF4, MLL-AF9, MLL-ENL, and MLL-ELL, and leu-
kemogenic MLL-lacZ, are resistant to Skp2- or Cdc20-mediated
degradation. Various Flag-tagged MLL fusions were coexpressed
in 293T cells with Myc-Skp2 or Myc-Cdc20. The levels of MLL
fusions were analyzed by anti-Flag immunoblots. (B) Human
MLL leukemia cell line SEM t(4;11) exhibits constant expres-
sion of endogenous MLL-AF4 protein through the cell cycle
progression. (Bottom panel) To examine the expression of MLL-
AF4, we synchronized the cells with centrifugal elutriation and
confirmed individual phases of the cell cycle with PI staining of
the DNA content followed by FACS analyses. The same
amounts of lysates purified from indicated cellular fractions
were subjected to immunoprecipitation with anti-AF4 anti-
body, which recognizes the C terminus of AF4. Immunoprecipi-
tates were resolved by SDS-PAGE and analyzed with an anti-
MLL antibody (Bethyl Laboratories) that recognizes amino acids
720–780 of MLL. (C) Fusion partners interfere with the recog-
nition of MLL N terminus by Skp2 and Cdc20. Flag-MLL fu-
sions and Flag-MLL(1–1400) were transfected in 293T cells for 2
d, followed by treatment with MG132 for 4 h. Cellular extracts
were subjected to anti-Flag immunoprecipitation assays. Immu-
noprecipitates were analyzed by Western blots using anti-Flag,
anti-Skp2, and anti-Cdc20 antibodies. The abundance of copre-
cipitated endogenous Skp2 and Cdc20 was determined with
anti-Skp2 and anti-Cdc20 Western blots, respectively. Immuno-
precipitation with extracts prepared from untransfected cells or
Flag-MLL(1–1400) transfected cells served as negative and posi-
tive controls, respectively. (D) MLL-AF4 is resistant to degrada-
tion by S- or late M-phase lysates. Flag-MLL-AF4 was tran-
siently expressed in 293T cells and purified with anti-Flag an-
tibody. Immunoprecipitated Flag-MLL-AF4 was aliquoted and
subjected to in vitro degradation experiments using the indi-
cated cellular lysates. The abundance of Flag-MLL-AF4 was
then determined by anti-Flag Western blots.
MLL fusions are resistant to SCFSkp2- or APCCdc20-
Liu et al.
2392 GENES & DEVELOPMENT
in proliferation in addition to differentiation. For ex-
ample, MLL−/−fetal liver or yolk sac hematopoietic cells
grow more slowly and form smaller colonies in methyl
cellulose assays (Hess et al. 1997). Our subsequent stud-
ies on cells deficient for Taspase1 or bearing noncleav-
able alleles of MLL (MLLNC/NC) recognized a participa-
tion of MLL through E2Fs in regulating cell proliferation
(Takeda et al. 2006). The active involvement of MLL/trx
in the cell cycle is conserved through evolution, in that
genetic studies in Drosophila also highlighted a signifi-
cant role of trx in cell proliferation (Muyrers-Chen et al.
2004). Recent studies, including ours, began to elucidate
the downstream targets for MLL in the cell cycle regu-
lation that include Cyclins and CDKIs (Muyrers-Chen et
al. 2004; Milne et al. 2005; Xia et al. 2005). We demon-
strated a MLL–E2F axis in regulating Cyclin E/A/B
expression for progressive cell cycle phase transition
(Takeda et al. 2006). Here, we further investigated the
participation of MLL in the cell cycle control and dis-
covered a tightly controlled biphasic expression of MLL.
This unique expression is conferred by defined windows
of degradation mediated by specialized cell cycle E3
ligases: SCFSkp2and APCCdc20(Fig. 7A). Importantly, in-
dividual peak expressions of MLL precede the induction
of Cyclin E/A and Cyclin B to ensure proper G1/S tran-
sition and M-phase progression, respectively (Fig. 7A).
Deregulation of this unique expression of MLL by
shRNA-mediated knockdown causes corresponding de-
fects in G1/S entry and M-phase progression (Fig. 7B).
Furthermore, overexpression of MLL incurs specific S-
phase defects, indicating the importance of down-regu-
lating its activity in S phase (Fig. 7C). However, whether
this resulted from sustained expression of Cyclin E/A
and/or invoked not-yet-identified insults remains to be
determined. Our data highlight the significance of this
biphasic expression of MLL in regulating cell prolifera-
tion and uncover a novel mechanism in regulating MLL
through protein degradation—another post-translational
regulatory scheme in addition to Taspase1-mediated
site-specific proteolysis. As MLL directly activates the
transcription of Cyclins that exhibit periodic expression
during cell proliferation, it is necessary for a cell to in-
corporate MLL expression into the intricately assembled
cell cycle circuitry to ensure correct transition of pro-
gressive phases. Current data not only consolidate the
role of MLL in activating cell cycle but also discover a
built-in program elegantly designed to turn off MLL with
an impeccable temporal sequence.
The observed undesirable cell cycle consequences from
perturbations of MLL levels led us to postulate that MLL
fusions may also interfere with the cell cycle—a mecha-
nism that is not fully appreciated in MLL leukemogen-
esis. Human chromosome 11q23 aberrations disrupt
MLL, leading to infant and chemotherapy-related leuke-
mias. These balanced translocations fuse the N-terminal
∼1400 amino acids of MLL in-frame with a wide spec-
trum of fusion partners to generate leukemogenic MLL
fusions. Mysteriously, among the >60 diverse fusion
partners identified so far, there are no commonly shared
characteristics identified based on sequence or struc-
ture homology (Rowley 1998; Ayton and Cleary 2001;
Canaani et al. 2004; Daser and Rabbitts 2004; Gilliland
et al. 2004; Hess 2004). Genetic evidence provided by
studying mice carrying individual MLL fusions reveals
several fundamental aspects of MLL leukemias. First,
the fusion partner is indispensable (Corral et al. 1996;
Dobson et al. 2000). Second, the fusion partner can be as
nonspecific as bacterial galactosidase (lac Z) in that mice
bearing MLL-lacZ developed myeloid leukemia after a
prolonged latency (Dobson et al. 2000). Third, the fusion
partner determines the phenotypes of the resulting leu-
kemia (Corral et al. 1996; Wang et al. 2005; Chen et al.
2006). For example, mice carrying MLL-AF4 or MLL-AF9
develop lymphoid versus myeloid malignancies (Corral
et al. 1996; Chen et al. 2006; Metzler et al. 2006), mim-
icking human counterparts. Detailed analyses of indi-
vidual MLL fusions using retrovirus-mediated gene
through cell cycle progression, the consequences of its deregu-
lation, and the resistance of MLL fusions to degradation.
(A) MLL protein oscillates throughout the cell cycle with two
distinct peaks at the G1/S and G2/M transitions. This unique
expression was conferred by a bimodal degradation executed
by SCFSkp2and APCCdc20at S and late M phases, respectively.
(B) Cells with constitutively low expression of MLL have de-
fects in S-phase entry and M-phase progression. (C) Constitu-
tively high expression of MLL incurs an S-phase block. (D) In
human MLL leukemia cells, MLL fusions remain constant
through the cell cycle due to their impaired interactions with
Skp2 and Cdc20. This deregulated expression may represent the
first initial biological insult contributing to the ultimate devel-
opment of MLL leukemias.
Models depict a critical biphasic expression of MLL
SCFSkp2and APCCdc20keep MLL in check
GENES & DEVELOPMENT2393
transduction of the hematopoietic stem cells provide in-
sightful mechanistic explanations regarding MLL leuke-
mias. These studies established important models, such
as transactivation and dimerization (Slany et al. 1998;
Martin et al. 2003; So et al. 2003). However, these two
models can only explain subsets of MLL leukemias since
not all fusion partners contain transactivation or dimer-
ization domains. As mice bearing MLL-lacZ developed
leukemia, it had long been postulated that lacZ induces
leukemia through either oligomerizing or stabilizing the
N terminus of MLL—two non-mutually exclusive
mechanisms. Although stabilization may contribute to
MLL leukemogenesis, it has not been further examined.
Our initial studies on the site-specific proteolysis of
MLL indicated that proteolytic cleavage of MLL is not
only required for its full activation but also regulates its
stability (Hsieh et al. 2003a,b; Takeda et al. 2006). Inter-
estingly, Taspase1-mediated site-specific proteolysis
also controls the protein levels of another recently iden-
tified Taspase1 substrate, TFIIA (transcription factor II
A) (Hoiby et al. 2004; Zhou et al. 2006). As our prior
studies implicated that MLL N terminus is unstable, we
envisioned that fusion partners must stabilize the com-
mon MLL N terminus for downstream gene regulation
(Hsieh et al. 2003b). Based on our current observations
that levels of MLL need to be tightly monitored during
cell proliferation and the degradation of MLLN320/C180
signals through its N-terminal ∼1400 amino acids (the
common denominator present in all MLL fusions), we
hypothesize that deregulated expression of MLL fusions
through cell division may constitute the long-awaited
universal insult underlying all MLL leukemias. To inter-
rogate this model, we examined whether overexpression
of SCFSkp2and APCCdc20affects the levels of leukemo-
genic MLL fusions, including MLL-AF4, MLL-AF9, MLL-
ENL, MLL-ELL, and MLL-lacZ. Remarkably, our in vivo
and in vitro degradation assays demonstrated that these
MLL fusions, unlike MLLN320/C180, have acquired resis-
tance to specialized cell cycle UPS. Furthermore, when
the human t(4;11) MLL leukemia cell line, SEM, was
examined, we witnessed a constant expression of MLL-
AF4 throughout progressive phases of the cell cycle. Al-
though the consequences of stable expression of MLL
fusions through the cell cycle progression remain to be
determined, several studies focused on individual MLL
fusions such as MLL-AF4 and MLL-AF9 did report asso-
ciated cell cycle defects (Pession et al. 2003; Caslini et al.
2004; Xia et al. 2005). Most interestingly, we noticed
that all of the five tested human MLL leukemia cell lines
failed to be synchronized by drugs like mimosine, thy-
midine, and nocodazole, implicating global defects in
cell cycle checkpoints. As shown in several knock-in
murine MLL leukemia models—including MLL-AF4
(Chen et al. 2006; Metzler et al. 2006), MLL-AF9 (Corral
et al. 1996), MLL-lacZ (Dobson et al. 2000), MLL-ELL-
(Luo et al. 2002), and MLL-CBP (Wang et al. 2005)—en-
gineered mice only developed leukemia after a long la-
tency or the challenge with carcinogens such as ENU
(Luo et al. 2002; Wang et al. 2005), suggesting that MLL
fusion alone is necessary but insufficient in MLL leuke-
mogenesis. Therefore, additional mutations are likely to
contribute to the full-blown MLL leukemia phenotypes
(Felix et al. 1998; Mahgoub et al. 1998; Armstrong et al.
2003; Oguchi et al. 2003). The loss of the biphasic ex-
pression of MLL fusions may initiate the first universal
mechanistic insult toward the ultimate development of
MLL leukemias by disrupting cell cycle checkpoints (Fig.
7D). Future studies focusing on dissecting the mecha-
nisms by which MLL fusions compromise cell cycle
checkpoints and evaluating the contribution of such de-
fects in MLL leukemogenesis will certainly provide
therapeutic vantage points in treating this deadly illness.
The intertwined relationship between PcG and trxG
members in cell fate and cell cycle
Deregulation of essential developmental pathways com-
monly leads to dire outcomes, including cancer; the PcG
and trxG proteins are such examples (van Lohuizen
1999). Initial genetic evidence obtained from flies along
with subsequent studies in mammals established the
fundamental roles of PcG and trxG proteins in maintain-
ing Hox gene expression, thus ensuring the correct in-
stallment of complex body plans in higher organisms. In
addition to their essential roles in embryonic develop-
ment, deregulated PcG and trxG proteins also contribute
to various oncogenic processes. For example, Bmi-1, a
PcG protein, is well recognized for its ability to acceler-
ate Myc-induced lymphoma in mice (van Lohuizen et al.
1991). Subsequent studies on Bmi-1-associated oncogen-
esis primarily emphasized the ability of Bmi-1 to sup-
press the expression of tumor suppressors p16ink4aand
ARF, thus promoting caner cell proliferation (Jacobs et
al. 1999b). Despite the well-established role in Hox gene
regulation, the significance of Bmi-1-induced Hox gene
alterations in Bmi-1-mediated oncogenesis remains un-
determined. On the contrary, studies on MLL, the
founder of trxG proteins, and its associated leukemias
recognized the potential mechanistic importance of
overexpression of Hox genes including HoxA7, HoxA9,
and HoxA11 in leukemogenesis (Armstrong et al. 2002;
Ayton and Cleary 2003). However, whether aberrant cell
cycle gene expression contributes to MLL leukemogen-
esis has not been largely questioned. Based on our find-
ings that link MLL/MLL fusions to cell cycle regulation,
we propose that a simultaneous deregulation of differen-
tiation and proliferation through Hox and cell cycle
genes by Bmi-1 and MLL is of paramount importance in
Materials and methods
Plasmid constructions, siRNA, and transfections
cDNAs encoding full-length MLL, various MLL fragments, and
MLL fusions were fused in-frame after an N-terminal Flag tag
(pCMV-3xFlag; Sigma) to generate individual eukaryotic expres-
sion constructs for transient transfection assays. Full-length
Fbw7, Skp2, ?TrCP1, Cdc20, and Cdh1 were cloned from a 293T
cDNA library and inserted into an N-terminal Myc tag ex-
Liu et al.
2394 GENES & DEVELOPMENT
pression vector (CMV-Myc; Clontech). The pMT123 plasmid
encoding HA-Ubiquitin was kindly provided by Dr. Dirk Boh-
mann (Treier et al. 1994). Lipofectamine and Oligofectamine
(Invitrogen) were used to transfect plasmid and siRNA oligos
(Dharmacon), respectively, according to the manufactures’ pro-
tocols. siRNA oligos against Skp2 (M-003324-03) and Cdc20
(M-003225-03) were purchased from Dharmacon.
Antibodies and Western blots
Anti-CT (C terminus) antibody that recognizes the transactiva-
tion domain of MLLC180has been described previously (Hsieh et
al. 2003b). A rabbit polyclonal anti-NT (N terminus) antibody
(MO-353) was raised against amino acids 1600–1985 of MLL.
Western blots using commercially available antibodies against
Skp2 (Santa Cruz Biotechnology), Cdc20 (Santa Cruz Biotech-
nology), Myc (Santa Cruz Biotechnolgoy), phospho-Histone H3
(Ser-10) (Upstate Biotechnology), MLL1 (recognizing amino ac-
ids 720–780 of MLL; Bethyl Laboratories), ?-Actin (Sigma), Flag
(Sigma), and HA (12CA5) were performed according to the
manufacturers’ recommendations. Antibodies were detected
using the enhanced chemiluminescence method (Western
Lightning, PerkinElmer). Western blot signals were acquired
with the LAS-3000 Imaging system (FujiFilm) and were ana-
lyzed by ImageGauge software (FujiFilm) as described previ-
ously (Kim et al. 2006).
HeLa cells were plated on LabTek II chamber slides (Nunc) over-
night, washed with PBS twice, fixed with 4% paraformaldehyde
for 15 min, permeabilized with 0.2% Triton X-100 for 10 min,
and blocked with 3% BSA in PBS for 1 h. These cells were
subsequently incubated with the indicated primary antibodies
for 3 h, followed by two PBS washes, and were then incubated
with Alexa488-conjugated anti-mouse or Alexa568-conjugated
anti-rabbit antibodies (Molecular Probes) for an additional 30
min. The nuclei were counterstained with Hoechst 33342 (Mo-
lecular Probes). Fluorescent images were captured by a CCD
camera (Diagnostic Instruments) attached to an Olympus IX51
microscope and were analyzed with the SPOT advanced soft-
Cell culture and synchronization
HeLa and 293T cells were obtained from American Type Cul-
ture Collection and grown in DMEM (Invitrogen) supplemented
with 10% fetal bovine serum, nonessential amino acids, L-glu-
tamine, and penicillin/streptomycin. The human leukemia cell
line SEM, carrying the t(4;11)(q21;q23) chromosomal transloca-
tion, was maintained in RPMI 1640 medium (Invitrogen) con-
taining the same supplements. For double thymidine block and
release experiments, HeLa cells were treated with 2 mM thy-
midine for 12 h and released for 12 h, followed by the second
12-h treatment before being released into thymidine-free me-
dium for the indicated periods of time. As to the nocodazole
block and release experiments, HeLa cells were blocked with
200 ng/mL nocodazole for 18 h. Mitotic cells were enriched by
shake-off and washed twice with PBS before being released into
nocodazole-free medium for the indicated time. Centrifugal
elutriation was carried out using a Beckman JE6B elutriation
rotor (Beckman). Small aliquots of cells collected at indicated
time points or fractions were subjected to FACS analysis of the
DNA contents for cell cycle phase assignment before subse-
Cells were harvested and lysed with Trizol (Invitrogen) for RNA
purification using RNeasy (Qiagen). Reverse transcriptions
were performed with oligo-dT plus random octamer primers
(Ambion) using SuperScript II (Invitrogen). qPCR was performed
with Power SYBR Green master mix (Applied Biosystems) in
triplicate using the following MLL primer set: MLL-forward,
ACACATTCCAGACCAAGAAACGAC, and MLL-reverse, AG
GCATCTTCAATACTTTCTGCACAG. Data were acquired
using an ABI PRISM 7000 system (Applied Biosystems) and ana-
lyzed as described previously (Takeda et al. 2006). GAPDH ex-
pression served as a control.
Cell cycle analyses and BrdU incorporation assays
Cells were trypsinized, washed with PBS, treated with 20 µg/mL
RNase A, and stained with 25 µg/mL PI for 1 h before being
subjected to cell cycle analyses. Flow-cytometric analyses were
performed using a FACSCalibur flow cytometer (Becton-Dick-
inson) to measure DNA contents. Ten-thousand events were
collected per each experiment, and the data were analyzed with
FlowJo software (Tree Star) or FCS Express (De Novo System).
Cells were pulsed with BrdU for 30 min and analyzed by
FACSCalibur according to the manufacturer’s manual (BD
Pharmingen; FITC/APC BrdU Flow Kit).
shRNA-mediated knockdown of MLL in HeLa cells
Target sequences (GTGCCAAGCACTGTCGAAA [Fig. 4] and
against human MLL and a control scrambled sequence (GCGC
GCTTTGTAGGATTCG) that has no significant homology
with the human genome were inserted into the pSUPER.
retro.puro vector, according to the manufacturer’s protocol
(Oligoengine). Generated retrovirus carrying indicated shRNA
was used to infect HeLa cells for 2 d before being subjected to
puromycin selection at 2 µg/mL.
Protein extracts and in vitro degradation assays
To obtain G1-phase lysates, HeLa cells were synchronized by
mimosine for 16 h. For S-phase lysates, HeLa cells were syn-
chronized by double thymidine block, followed by a 4-h release.
To synchronize cells at G2 phase, HeLa cells were released from
a double thymidine block for 5 h, followed by a 5-h treatment
with 0.5 µM etoposide and extensive washes to remove mitotic
cells. As to early M-phase lysates, HeLa cells were synchronized
at prometaphase by nocodazole treatment for 18 h. Cells re-
leased from the nocodazole block for 1 h were at late M phase.
Synchronized cells were resuspended in buffer containing 20
mM Tris-HCl (pH 7.2), 2 mM dithiothreitol (DTT), 0.25 mM
EDTA, and Complete Protease Inhibitor Cocktail (Roche) on ice
before being transferred to a nitrogen disruption bomb (Parr).
The pressure was subsequently brought up to 1000 psi, followed
by incubation for 30 min on ice. The pressure was then slowly
released to gently disrupt the cells. The homogenized materials
were spun down at 10,000g for 10 min. The resulting cellular
extracts were aliquoted and stored at −80°C. For immunodeple-
tion experiments, 10 µg of anti-Skp2 or anti-Cdc20 antibodies
were first adsorbed to 15 µL of protein-A beads (GE Healthcare)
by rocking for 90 min at 4°C. The antibody-bound beads were
washed and incubated with 40 µL of HeLa extracts (∼400 µg of
protein) for 2 h at 4°C. Bead-bound immunocomplexes were
subsequently removed by centrifugation. Immunodepleted ex-
tracts were aliquoted and used for the indicated in vitro degra-
SCFSkp2and APCCdc20keep MLL in check
GENES & DEVELOPMENT 2395
dation assays. 293T cells were transfected with the indicated
Flag-MLL-expressing constructs for 2 d before being subjected to
lysis with RIPA buffer in the presence of Complete Protease
Inhibitor Cocktail (Roche). Expressed Flag-MLL was immuno-
precipitated using anti-Flag M2 beads (Sigma). Bead-bound Flag-
MLL was aliquoted for the indicated experiments. The degrada-
tion reaction (10 µL) contained 40 mM Tris-HCl (pH 7.6), 5 mM
MgCl2, 1 mM DTT, 10% glycerol, 1 µM ubiquitin aldehyde, 10
mM phosphocreatine, 100 µg/mL creatine phosphokinase,
0.5 mM ATP, 10 µg of the indicated HeLa cellular extracts, 3 µL
of Flag-MLL immunoprecipitates, and Complete Protease In-
hibitor Cocktail (Roche). Following incubation for 1 h at 30°C,
reactions were stopped by the addition of SDS sample buffer
before being subjected to SDS-PAGE analyses.
We thank Drs. Scott Armstrong and Jorg Faber for providing
human MLL leukemia cell lines. MLL-ENL and MLL-ELL ex-
pression constructs were kindly provided by Dr. Cleary and
Dr. Mitani, respectively. H.L. is supported by NIH CA119008 to
J.J.-D.H. E.H.-Y.C. is supported by the Searle Scholars Program,
Mallinckrodt Jr. Foundation, NCI Howard Temin Award, and
NIH CA125562. This work is supported by Mallinckrodt Jr.
Foundation, ASH Scholar Award, NCI Howard Temin Award,
and NIH CA119008 to J.J.-D.H.
Armstrong, S.A., Staunton, J.E., Silverman, L.B., Pieters, R., den
Boer, M.L., Minden, M.D., Sallan, S.E., Lander, E.S., Golub,
T.R., and Korsmeyer, S.J. 2002. MLL translocations specify a
distinct gene expression profile that distinguishes a unique
leukemia. Nat. Genet. 30: 41–47.
Armstrong, S.A., Kung, A.L., Mabon, M.E., Silverman, L.B.,
Stam, R.W., Den Boer, M.L., Pieters, R., Kersey, J.H., Sallan,
S.E., Fletcher, J.A., et al. 2003. Inhibition of FLT3 in MLL.
Validation of a therapeutic target identified by gene expres-
sion based classification. Cancer Cell 3: 173–183.
Ayton, P.M. and Cleary, M.L. 2001. Molecular mechanisms of
leukemogenesis mediated by MLL fusion proteins. Onco-
gene 20: 5695–5707.
Ayton, P.M. and Cleary, M.L. 2003. Transformation of myeloid
progenitors by MLL oncoproteins is dependent on Hoxa7 and
Hoxa9. Genes & Dev. 17: 2298–2307.
Blais, A. and Dynlacht, B.D. 2004. Hitting their targets: An
emerging picture of E2F and cell cycle control. Curr. Opin.
Genet. Dev. 14: 527–532.
Bracken, A.P., Ciro, M., Cocito, A., and Helin, K. 2004. E2F
target genes: Unraveling the biology. Trends Biochem. Sci.
Canaani, E., Nakamura, T., Rozovskaia, T., Smith, S.T., Mori,
T., Croce, C.M., and Mazo, A. 2004. ALL-1/MLL1, a homo-
logue of Drosophila TRITHORAX, modifies chromatin and
is directly involved in infant acute leukaemia. Br. J. Cancer
Cardozo, T. and Pagano, M. 2004. The SCF ubiquitin ligase:
Insights into a molecular machine. Nat. Rev. Mol. Cell Biol.
Caslini, C., Serna, A., Rossi, V., Introna, M., and Biondi, A.
2004. Modulation of cell cycle by graded expression of MLL-
AF4 fusion oncoprotein. Leukemia 18: 1064–1071.
Chen, W., Li, Q., Hudson, W.A., Kumar, A., Kirchhof, N., and
Kersey, J.H. 2006. A murine Mll-AF4 knock-in model results
in lymphoid and myeloid deregulation and hematologic ma-
lignancy. Blood 108: 669–677.
Corral, J., Lavenir, I., Impey, H., Warren, A.J., Forster, A., Lar-
son, T.A., Bell, S., McKenzie, A.N., King, G., and Rabbitts,
T.H. 1996. An Mll-AF9 fusion gene made by homologous
recombination causes acute leukemia in chimeric mice:
A method to create fusion oncogenes. Cell 85: 853–861.
Daser, A. and Rabbitts, T.H. 2004. Extending the repertoire of
the mixed-lineage leukemia gene MLL in leukemogenesis.
Genes & Dev. 18: 965–974.
Djabali, M., Selleri, L., Parry, P., Bower, M., Young, B., and
Evans, G.A. 1993. A trithorax-like gene is interrupted by
chromosome 11q23 translocations in acute leukaemias. Nat.
Genet. 4: 431.
Dobson, C.L., Warren, A.J., Pannell, R., Forster, A., and Rab-
bitts, T.H. 2000. Tumorigenesis in mice with a fusion of the
leukaemia oncogene Mll and the bacterial lacZ gene. EMBO
J. 19: 843–851.
Domer, P.H., Fakharzadeh, S.S., Chen, C.S., Jockel, J., Johansen,
L., Silverman, G.A., Kersey, J.H., and Korsmeyer, S.J. 1993.
Acute mixed-lineage leukemia t(4;11)(q21;q23) generates an
MLL-AF4 fusion product. Proc. Natl. Acad. Sci. 90: 7884–
Dyson, N. 1998. The regulation of E2F by pRB-family proteins.
Genes & Dev. 12: 2245–2262.
Eguchi, M., Eguchi-Ishimae, M., and Greaves, M. 2005. Molecu-
lar pathogenesis of MLL-associated leukemias. Int. J. Hema-
tol. 82: 9–20.
Felix, C.A., Megonigal, M.D., Chervinsky, D.S., Leonard, D.G.,
Tsuchida, N., Kakati, S., Block, A.M., Fisher, J., Grossi, M.,
Salhany, K.I., et al. 1998. Association of germline p53 mu-
tation with MLL segmental jumping translocation in treat-
ment-related leukemia. Blood 91: 4451–4456.
Giacinti, C. and Giordano, A. 2006. RB and cell cycle progres-
sion. Oncogene 25: 5220–5227.
Gilliland, D.G., Jordan, C.T., and Felix, C.A. 2004. The molecu-
lar basis of leukemia. Hematology 2004: 80–97.
Gu, Y., Nakamura, T., Alder, H., Prasad, R., Canaani, O.,
Cimino, G., Croce, C.M., and Canaani, E. 1992. The t(4;11)
chromosome translocation of human acute leukemias fuses
the ALL-1 gene, related to Drosophila trithorax, to the AF-4
gene. Cell 71: 701–708.
Guo, W.J., Datta, S., Band, V., and Dimri, G.P. 2007. Mel-18, a
polycomb group protein, regulates cell proliferation and se-
nescence via transcriptional repression of Bmi-1 and c-Myc
oncoproteins. Mol. Biol. Cell 18: 536–546.
Hanson, R.D., Hess, J.L., Yu, B.D., Ernst, P., van Lohuizen, M.,
Berns, A., van der Lugt, N.M., Shashikant, C.S., Ruddle,
F.H., Seto, M., et al. 1999. Mammalian Trithorax and poly-
comb-group homologues are antagonistic regulators of
homeotic development. Proc. Natl. Acad. Sci. 96: 14372–
Hershko, A. 2005. The ubiquitin system for protein degradation
and some of its roles in the control of the cell division cycle.
Cell Death Differ. 12: 1191–1197.
Hess, J.L. 2004. Mechanisms of transformation by MLL. Crit.
Rev. Eukaryot. Gene Expr. 14: 235–254.
Hess, J.L., Yu, B.D., Li, B., Hanson, R., and Korsmeyer, S.J. 1997.
Defects in yolk sac hematopoiesis in Mll-null embryos.
Blood 90: 1799–1806.
Hoiby, T., Mitsiou, D.J., Zhou, H., Erdjument-Bromage, H.,
Tempst, P., and Stunnenberg, H.G. 2004. Cleavage and pro-
teasome-mediated degradation of the basal transcription fac-
tor TFIIA. EMBO J. 23: 3083–3091.
Hsieh, J.J., Cheng, E.H., and Korsmeyer, S.J. 2003a. Taspase1:
A threonine aspartase required for cleavage of MLL and
Liu et al.
2396GENES & DEVELOPMENT
proper HOX gene expression. Cell 115: 293–303.
Hsieh, J.J., Ernst, P., Erdjument-Bromage, H., Tempst, P., and
Korsmeyer, S.J. 2003b. Proteolytic cleavage of MLL gener-
ates a complex of N- and C-terminal fragments that confers
protein stability and subnuclear localization. Mol. Cell. Biol.
Jacobs, J.J., Kieboom, K., Marino, S., DePinho, R.A., and van
Lohuizen, M. 1999a. The oncogene and Polycomb-group
gene bmi-1 regulates cell proliferation and senescence
through the ink4a locus. Nature 397: 164–168.
Jacobs, J.J., Scheijen, B., Voncken, J.W., Kieboom, K., Berns, A.,
and van Lohuizen, M. 1999b. Bmi-1 collaborates with c-Myc
in tumorigenesis by inhibiting c-Myc-induced apoptosis via
INK4a/ARF. Genes & Dev. 13: 2678–2690.
Khan, J.A., Dunn, B.M., and Tong, L. 2005. Crystal structure of
human Taspase1, a crucial protease regulating the function
of MLL. Structure 13: 1443–1452.
Kim, H., Rafiuddin-Shah, M., Tu, H.C., Jeffers, J.R., Zambetti,
G.P., Hsieh, J.J., and Cheng, E.H. 2006. Hierarchical regula-
tion of mitochondrion-dependent apoptosis by BCL-2 sub-
families. Nat. Cell Biol. 8: 1348–1358.
Luo, R.T., Kebriaei, P., Kaberlein, J.J., and Thirman, M.J. 2002.
Cooperating mutations are necessary for the development of
AML in mll-ELL knock-in mice. Blood 100: 136a.
Mahgoub, N., Parker, R.I., Hosler, M.R., Close, P., Winick, N.J.,
Masterson, M., Shannon, K.M., and Felix, C.A. 1998. RAS
mutations in pediatric leukemias with MLL gene rearrange-
ments. Genes Chromosomes Cancer 21: 270–275.
Martin, M.E., Milne, T.A., Bloyer, S., Galoian, K., Shen, W.,
Gibbs, D., Brock, H.W., Slany, R., and Hess, J.L. 2003.
Dimerization of MLL fusion proteins immortalizes hemato-
poietic cells. Cancer Cell 4: 197–207.
Martinez, A.M. and Cavalli, G. 2006. The role of polycomb
group proteins in cell cycle regulation during development.
Cell Cycle 5: 1189–1197.
Martinez, A.M., Colomb, S., Dejardin, J., Bantignies, F., and
Cavalli, G. 2006. Polycomb group-dependent Cyclin A re-
pression in Drosophila. Genes & Dev. 20: 501–513.
Metzler, M., Forster, A., Pannell, R., Arends, M.J., Daser, A.,
Lobato, M.N., and Rabbitts, T.H. 2006. A conditional model
of MLL-AF4 B-cell tumourigenesis using invertor technol-
ogy. Oncogene 25: 3093–3103.
Milne, T.A., Briggs, S.D., Brock, H.W., Martin, M.E., Gibbs, D.,
Allis, C.D., and Hess, J.L. 2002. MLL targets SET domain
methyltransferase activity to Hox gene promoters. Mol. Cell
Milne, T.A., Hughes, C.M., Lloyd, R., Yang, Z., Rozenblatt-
Rosen, O., Dou, Y., Schnepp, R.W., Krankel, C., Livolsi,
V.A., Gibbs, D., et al. 2005. Menin and MLL cooperatively
regulate expression of cyclin-dependent kinase inhibitors.
Proc. Natl. Acad. Sci. 102: 749–754.
Morgan, D.O. 1997. Cyclin-dependent kinases: Engines, clocks,
and microprocessors. Annu. Rev. Cell Dev. Biol. 13: 261–
Murray, A.W. 2004. Recycling the cell cycle: Cyclins revisited.
Cell 116: 221–234.
Muyrers-Chen, I., Rozovskaia, T., Lee, N., Kersey, J.H., Naka-
mura, T., Canaani, E., and Paro, R. 2004. Expression of leu-
kemic MLL fusion proteins in Drosophila affects cell cycle
control and chromosome morphology. Oncogene 23: 8639–
Nakamura, T., Mori, T., Tada, S., Krajewski, W., Rozovskaia, T.,
Wassell, R., Dubois, G., Mazo, A., Croce, C.M., and Canaani,
E. 2002. ALL-1 is a histone methyltransferase that assembles
a supercomplex of proteins involved in transcriptional regu-
lation. Mol. Cell 10: 1119–1128.
Nakayama, K.I. and Nakayama, K. 2006. Ubiquitin ligases: Cell-
cycle control and cancer. Nat. Rev. Cancer 6: 369–381.
Nevins, J.R. 2001. The Rb/E2F pathway and cancer. Hum. Mol.
Genet. 10: 699–703.
Nurse, P. 2000. A long twentieth century of the cell cycle and
beyond. Cell 100: 71–78.
Oguchi, K., Takagi, M., Tsuchida, R., Taya, Y., Ito, E., Isoyama,
K., Ishii, E., Zannini, L., Delia, D., and Mizutani, S. 2003.
Missense mutation and defective function of ATM in a
childhood acute leukemia patient with MLL gene rearrange-
ment. Blood 101: 3622–3627.
Pession, A., Martino, V., Tonelli, R., Beltramini, C., Locatelli,
F., Biserni, G., Franzoni, M., Freccero, F., Montemurro, L.,
Pattacini, L., et al. 2003. MLL-AF9 oncogene expression af-
fects cell growth but not terminal differentiation and is
downregulated during monocyte–macrophage maturation in
AML-M5 THP-1 cells. Oncogene 22: 8671–8676.
Peters, J.M. 2006. The anaphase promoting complex/cyclosome:
A machine designed to destroy. Nat. Rev. Mol. Cell Biol. 7:
Reed, S.I. 2003. Ratchets and clocks: The cell cycle, ubiquityla-
tion and protein turnover. Nat. Rev. Mol. Cell Biol. 4: 855–
Ringrose, L. and Paro, R. 2004. Epigenetic regulation of cellular
memory by the Polycomb and Trithorax group proteins.
Annu. Rev. Genet. 38: 413–443.
Rowley, J.D. 1998. The critical role of chromosome transloca-
tions in human leukemias. Annu. Rev. Genet. 32: 495–519.
Schuettengruber, B., Chourrout, D., Vervoort, M., Leblanc, B.,
and Cavalli, G. 2007. Genome regulation by polycomb and
trithorax proteins. Cell 128: 735–745.
Schumacher, A. and Magnuson, T. 1997. Murine Polycomb- and
trithorax-group genes regulate homeotic pathways and be-
yond. Trends Genet. 13: 167–170.
Sherr, C.J. and Roberts, J.M. 1999. CDK inhibitors: Positive and
negative regulators of G1-phase progression. Genes & Dev.
Slany, R.K., Lavau, C., and Cleary, M.L. 1998. The oncogenic
capacity of HRX-ENL requires the transcriptional transacti-
vation activity of ENL and the DNA binding motifs of HRX.
Mol. Cell. Biol. 18: 122–129.
So, C.W., Lin, M., Ayton, P.M., Chen, E.H., and Cleary, M.L.
2003. Dimerization contributes to oncogenic activation of
MLL chimeras in acute leukemias. Cancer Cell 4: 99–110.
Takeda, S., Chen, D.Y., Westergard, T.D., Fisher, J.K., Rubens,
J.A., Sasagawa, S., Kan, J.T., Korsmeyer, S.J., Cheng, E.H.,
and Hsieh, J.J. 2006. Proteolysis of MLL family proteins is
essential for taspase1-orchestrated cell cycle progression.
Genes & Dev. 20: 2397–2409.
Thirman, M.J., Gill, H.J., Burnett, R.C., Mbangkollo, D., Mc-
Cabe, N.R., Kobayashi, H., Ziemin-van der Poel, S., Kaneko,
Y., Morgan, R., Sandberg, A.A., et al. 1993. Rearrangement of
the MLL gene in acute lymphoblastic and acute myeloid
leukemias with 11q23 chromosomal translocations. N. Engl.
J. Med. 329: 909–914.
Tkachuk, D.C., Kohler, S., and Cleary, M.L. 1992. Involvement
of a homolog of Drosophila trithorax by 11q23 chromosomal
translocations in acute leukemias. Cell 71: 691–700.
Treier, M., Staszewski, L.M., and Bohmann, D. 1994. Ubiquitin-
dependent c-Jun degradation in vivo is mediated by the ?
domain. Cell 78: 787–798.
Trimarchi, J.M. and Lees, J.A. 2002. Sibling rivalry in the E2F
family. Nat. Rev. Mol. Cell Biol. 3: 11–20.
van Lohuizen, M. 1999. The trithorax-group and polycomb-
group chromatin modifiers: Implications for disease. Curr.
Opin. Genet. Dev. 9: 355–361.
SCFSkp2and APCCdc20keep MLL in check
GENES & DEVELOPMENT 2397
van Lohuizen, M., Verbeek, S., Scheijen, B., Wientjens, E., van
der Gulden, H., and Berns, A. 1991. Identification of cooper-
ating oncogenes in E µ-myc transgenic mice by provirus tag-
ging. Cell 65: 737–752.
Wang, J., Iwasaki, H., Krivtsov, A., Febbo, P.G., Thorner, A.R.,
Ernst, P., Anastasiadou, E., Kutok, J.L., Kogan, S.C., Zinkel,
S.S., et al. 2005. Conditional MLL-CBP targets GMP and
models therapy-related myeloproliferative disease. EMBO J.
Xia, Z.B., Popovic, R., Chen, J., Theisler, C., Stuart, T., San-
tillan, D.A., Erfurth, F., Diaz, M.O., and Zeleznik-Le, N.J.
2005. The MLL fusion gene, MLL-AF4, regulates cyclin-de-
pendent kinase inhibitor CDKN1B (p27kip1) expression.
Proc. Natl. Acad. Sci. 102: 14028–14033.
Yokoyama, A., Kitabayashi, I., Ayton, P.M., Cleary, M.L., and
Ohki, M. 2002. Leukemia proto-oncoprotein MLL is proteo-
lytically processed into 2 fragments with opposite transcrip-
tional properties. Blood 100: 3710–3718.
Yu, B.D., Hess, J.L., Horning, S.E., Brown, G.A., and Korsmeyer,
S.J. 1995. Altered Hox expression and segmental identity in
Mll-mutant mice. Nature 378: 505–508.
Yu, B.D., Hanson, R.D., Hess, J.L., Horning, S.E., and Kors-
meyer, S.J. 1998. MLL, a mammalian trithorax-group gene,
functions as a transcriptional maintenance factor in morpho-
genesis. Proc. Natl. Acad. Sci. 95: 10632–10636.
Zhou, H., Spicuglia, S., Hsieh, J.J., Mitsiou, D.J., Hoiby, T.,
Veenstra, G.J., Korsmeyer, S.J., and Stunnenberg, H.G. 2006.
Uncleaved TFIIA is a substrate for taspase 1 and active in
transcription. Mol. Cell. Biol. 26: 2728–2735.
Liu et al.
2398 GENES & DEVELOPMENT