MOLECULAR AND CELLULAR BIOLOGY, Aug. 2006, p. 6261–6271
Copyright © 2006, American Society for Microbiology. All Rights Reserved.
Vol. 26, No. 16
Identification of a Conserved Negative Regulatory Sequence That
Influences the Leukemogenic Activity of NOTCH1
Mark Y. Chiang,1† Mina L. Xu,2† Gavin Histen,2Olga Shestova,3Monideepa Roy,2Yunsun Nam,2
Stephen C. Blacklow,2David B. Sacks,2Warren S. Pear,3and Jon C. Aster2*
Department of Hematology/Oncology, Abramson Family Cancer Research Institute, University of Pennsylvania, Philadelphia,
Pennsylvania 191041; Department of Pathology, Brigham and Women’s Hospital and Harvard Medical School,
Boston, Massachusetts 021152; and Department of Pathology and Laboratory Medicine,
Abramson Family Cancer Research Institute, Institute for Medicine and Engineering,
University of Pennsylvania, Philadelphia, Pennsylvania 191043
Received 28 December 2005/Returned for modification 8 February 2006/Accepted 26 May 2006
NOTCH1 is a large type I transmembrane receptor that regulates normal T-cell development via a
signaling pathway that relies on regulated proteolysis. Ligand binding induces proteolytic cleavages in
NOTCH1 that release its intracellular domain (ICN1), which translocates to the nucleus and activates
target genes by forming a short-lived nuclear complex with two other proteins, the DNA-binding factor
CSL and a Mastermind-like (MAML) coactivator. Recent work has shown that human T-ALL is frequently
associated with C-terminal NOTCH1 truncations, which uniformly remove sequences lying between res-
idues 2524 and 2556. This region includes the highly conserved sequence WSSSSP (S4), which based on
its amino acid content appeared to be a likely site for regulatory serine phosphorylation events. We show
here that the mutation of the S4 sequence leads to hypophosphorylation of ICN1; increased NOTCH1
signaling; and the stabilization of complexes containing ICN1, CSL, and MAML1. Consistent with these
in vitro studies, mutation of the WSSSSP sequence converts nonleukemogenic weak gain-of-function
NOTCH1 alleles into alleles that cause aggressive T-ALLs in a murine bone marrow transplant model.
These studies indicate that S4 is an important negative regulatory sequence and that the deletion of S4
likely contributes to the development of human T-ALL.
NOTCH receptors and downstream mediators participate in
a signaling pathway that variously regulates the specification of
cell fate, proliferation, self-renewal, survival, and apoptosis in
a dose- and context-dependent fashion (2). Like other mem-
bers of the NOTCH receptor family, human NOTCH1 is a
large multimodular type I transmembrane glycoprotein (Fig.
1A). Newly synthesized NOTCH1 is cleaved by furin at a site
termed S1 just external to the transmembrane domain (21),
yielding two noncovalently associated extracellular (NEC) and
transmembrane (NTM) subunits (5, 21, 29). Binding of ligands
to NECtriggers two sequential proteolytic events within the
NTMsubunit at sites S2 and S3. S2 cleavage occurs just external
to the transmembrane domain and is catalyzed by ADAM-type
metalloproteases (6, 24). This creates a short-lived intermedi-
ate, NTM*, which is recognized by nicastrin (33), a component
of the protease complex called ?-secretase (8, 18, 32). Addi-
tional cleavages by ?-secretase free the intracellular domain of
NOTCH1 (ICN1), allowing it to translocate to the nucleus,
where it activates transcription through the formation of a
ternary complex with the DNA-binding factor CSL (16, 19, 36,
44) and coactivator proteins of the Mastermind-like (MAML)
family (27, 28, 43).
Nuclear ICN1 is short-lived. One mechanism that appears
to promote the rapid turnover of the CSL/ICN1/MAML
transcription complex involves the recruitment of mediator
complexes and CycC-CDK8 through the C-terminal tail of
MAML1 (13). Phosphorylation of ICN1 on multiple C-ter-
minal serine residues by CycC-CDK8 is hypothesized to
create recognition sites for E3 ligases such as FBW7/Sel10
(13), which has been implicated in the ubiquitylation and
subsequent degradation of ICN (38, 39). Some of the sites
targeted by CycC-CDK8 lie in the far C-terminal portion of
NOTCH1 (13), an unstructured region that is enriched for
the amino acids proline, glutamate, serine, and threonine
(PEST). PEST sequences regulate the degradation of a
number of proteins (30), sometimes by serving as substrates
for phosphorylation events that mark a protein for degra-
dation (22). In the case of CycC-CDK8, phosphorylation of
ICN is hypothesized to couple MAML-dependent transcrip-
tional activation to rapid ICN degradation (12, 13). How-
ever, there is evidence that NOTCH stability is also regu-
lated at other levels. For example, phosphorylation by
GSK? appears to promote the degradation of the intracel-
lular domain of NOTCH2 (9), and E3 ligases of the Itch
family have been implicated in the ubiquitylation and reg-
ulation of membrane-associated NOTCH receptors (23, 34).
Thus, inputs from multiple pathways regulate NOTCH at
the level of protein stability during different stages of re-
ceptor activation and trafficking.
Increased NOTCH1 signaling plays a central part in the
pathogenesis of T-cell acute lymphoblastic leukemia (T-
ALL), a tumor derived from T-cell progenitors. We ob-
served that human T-ALLs commonly harbor frameshift
and stop codon mutations that delete various numbers of
C-terminal residues from NOTCH1 (41), a finding that was
* Corresponding author. Mailing address: Department of Pathology,
Brigham and Women’s Hospital, Boston, MA 02115. Phone: (617)
278-0032. Fax: (617) 264-5169. E-mail: firstname.lastname@example.org.
† M.Y.C. and M.L.X. contributed equally to this study.
presaged by the detection of retroviral insertions in murine
T-ALLs that cause similar truncations (10, 14). Although
these mutations are scattered throughout the 3? end of exon
34, all of the deletions found to date eliminate at least
residues 2524 to 2556, suggesting that this minimal region
contains at least one important motif that negatively regu-
lates NOTCH1 signal strength. Here, we analyze the role of
a short conserved sequence, WSSSSP (referred to as S4),
found within the minimal deleted region that influences not
only the function and stability of activated NOTCH1 but
also its leukemogenic activity.
MATERIALS AND METHODS
Expression plasmids. A diagram depicting the various forms of NOTCH1
used in these studies is shown in Fig. 1. Expression constructs that encode
full-length human NOTCH1 (residues M1 to K2556); ?EGF, a form bearing a
deletion that removes the coding region for epidermal growth factor (EGF)-like
repeats 1 to 36 (residues R23 to I1446); ?EGF?LNR, a form bearing a deletion
that removes the coding region for EGF-like repeats 1 to 36 and the three
Lin12/NOTCH repeats (residues R23 to C1562); and ICN1 (residues 1762 to
2556) have been described (31). Expression constructs for forms of ICN1 bearing
a N-terminal FLAG-tag were created by PCR with primers containing a consen-
sus Kozak start codon followed by the coding sequence for the FLAG epitope.
A C-terminal deletion removing residues 2473 to 2556 (originally identified in
the cell line ALL-SIL) has been described (41). NOTCH1-GAL4 chimeric
cDNAs were creating by ligating a PCR product encoding the DNA-binding
domain of GAL4 in frame to ICN1 cDNA cut with the restriction enzymes
Bsu36I and NcoI, which removes sequences encoding the RAM and ANK do-
mains. In other constructs, premature stop codons and point mutations were
introduced by using the QuikChange kit (Stratagene). cDNAs were assembled
variously in pcDNA3 (Invitrogen); pcDNA5 (Invitrogen); or the retroviral vector
MSCV-GFP, which drives expression of NOTCH1 and green fluorescence pro-
tein (GFP) from a single bicistronic RNA containing an internal ribosomal entry
sequence (IRES). Expression plasmids for CSL-MYC (4), MAML1-GFP (43),
and dominant-negative MAML1-GFP (42) have all been described. A pCMV2
plasmid (Sigma) encoding a “kinase-dead” dominant-negative form of CDK8
was kindly provided by Andrew Rice, Baylor University.
Cell culture. U2OS and 293T cells (American Type Culture Collection) were
maintained in Dulbecco modified Eagle medium (DMEM; Invitrogen) supple-
mented with 10% fetal bovine serum (Invitrogen), 2 mM L-glutamine (Invitrogen),
100 U of penicillin (Invitrogen)/ml, and 100 ?g of streptomycin (Invitrogen)/ml. 293
TRex cells were obtained from Invitrogen. Cells were grown at 37°C under 5% CO2.
Transcriptional activation assays. NOTCH1 expression plasmids were intro-
duced into U2OS cells by transient transfection with Lipofectamine Plus
(Invitrogen) and assessed for their ability to activated a NOTCH-sensitive lucif-
erase reporter gene, as described previously (4). Briefly, cells in 24-well dishes
were cotransfected in triplicate with 10 ng of various pcDNA3-NOTCH1 expres-
sion constructs, a NOTCH-sensitive firefly luciferase reporter gene (15), and an
internal control Renilla luciferase plasmid (Promega). Experiments involving
NOTCH1-GAL4 fusion constructs used a GAL4-luciferase reporter gene (Clon-
tech). Total introduced DNA was kept constant by adding empty pcDNA3
plasmid. Normalized firefly luciferase activities were measured in whole-cell
extracts prepared 44 to 48 h after transfection using the Dual Luciferase kit
(Promega) and a specially configured luminometer (Turner Systems). In some
experiments, the cells were treated posttransfection with the ?-secretase inhibi-
tor compound E (kindly provided by Michael Wolfe) at 1 ?M or with carrier
alone (0.01% dimethyl sulfoxide [DMSO]).
ICN1 immunoprecipitation. 293T cells transfected with pcDNA plasmids en-
coding various NOTCH signaling components were lysed in 50 mM Tris (pH
8.0), containing 1% NP-40, 100 mM NaCl, 30 mM NaF, 20 mM Na pyrophos-
phate, 2 mM Na vanadate, 2 mM Na molybdate, and 5 mM Na EDTA (buffer A).
ICN1 polypeptides were immunoprecipitated with a rabbit polyclonal antibody
raised against the transcriptional activation domain of NOTCH1, as described
previously (3). In some experiments, the immunoprecipitation polypeptides were
treated with lambda phosphatase (New England Biolabs) according to the man-
ufacturer’s recommendations. In other experiments, complexes containing CSL-
MYC were immunoprecipitated with the mouse monoclonal antibody 9E10, as
described previously (3).
Phosphoamino acid analysis. 293 TRex cells (Invitrogen) were cotransfected
with pcDNA5-FLAG-ICN1 plasmids and pOGG44, which expresses Flp recom-
binase. Isogenic 293 recombinants were selected with hygromycin B and then
split into 100-mm dishes. After the induction of ICN1 expression by the addition
of tetracycline (1 ?g/ml) for 24 h, cells were incubated twice for 1 h in phosphate-
free DMEM (Invitrogen) containing 10% dialyzed fetal calf serum (depletion
medium) and then grown overnight in a 9:1 mixture of depletion medium and
complete medium containing 2.5 mCi of [32P]orthophosphate (New England
Nuclear). After three washes with ice-cold phosphate-buffered saline, the cells
were lysed in ice-cold buffer A for 15 min and centrifuged at 14,000 ? g for 15
min. Proteins in the resulting supernatants were immunoprecipitated by adding
FLAG-M2-antibody agarose beads (Sigma), followed by mixing for 2 h at 4°C.
The beads were washed four times with buffer A, and bound proteins were
released by adding sodium dodecyl sulfate-polyacrylamide gel electrophoresis
(SDS-PAGE) loading buffer and heating the mixture for 10 min to 100°C. ICN1
proteins were separated by SDS-PAGE in 8% gels, which were dried and ana-
lyzed by autoradiography.
After digestion of phosphorylated ICN1 with trypsin-TPCK (tolylsulfonyl phe-
nylalanyl chloromethyl ketone; Worthington), acid hydrolysis was carried out in
6 M HCl at 110°C for 2 h. Phosphoamino acids were separated by thin-layer
electrophoresis at pH 1.9 as described previously (17).
Phosphopeptide analysis.32P-labeled ICN1 polypeptides were prepared from
293 TRex cells by immunoprecipitation, followed by SDS-PAGE in 8% gels as
described above. Portions of the dried gels containing phosphorylated ICN1
polypeptides were excised, rehydrated, oxidized with performic acid, and di-
gested with trypsin (Worthington) as described previously (17). Peptides were
then resuspended in 5 ?l of formic acid (96%) and separated by thin-layer
electrophoresis on cellulose plates (Sigma) for 5 h at 400 V in formic acid-acetic
acid-water (10:31:359 [pH 1.9]). After drying the cellulose plate, ascending chro-
matography was performed in butanol-pyridine-acetic acid-water (50:33:10:40).
32P-labeled peptides were visualized by autoradiography.
Pulse-chase analysis. 293 cells in 60-mm dishes were transfected with
pcDNA3-FLAG-ICN1 plasmids (1 ?g) on day 1, split into six-well dishes on day
2, and then subjected to pulse-chase labeling on day 3 as follows. Cells were
incubated twice for 1 h in DMEM without L-methionine containing 10% dialyzed
fetal calf serum, followed by incubation in the same medium containing 2.5 mCi
washes with Hanks buffered saline, the cells were either harvested immedi-
ately or incubated for up to 6 additional hours in replete DMEM containing
10% fetal calf serum and 2 mM cold L-methionine.35S-labeled ICN1 polypep-
tides were immunoprecipitated from whole-cell detergent lysates on FLAG-
M2-antibody beads (Sigma). After elution from the beads by heating for 10
35S-labeled L-methionine (New England Nuclear) for 30 min. After two
FIG. 1. NOTCH1 expression constructs. A schematic representa-
tion of the mature full-length human NOTCH1 receptor and of ex-
pression constructs bearing N-terminal NOTCH1 deletions is shown.
Furin cleavage in the extracellular domain creates NEC(NOTCH1
extracellular) and NTM(NOTCH1 transmembrane) subunits that re-
main associated through noncovalent interactions between the N-ter-
minal and C-terminal portions of the heterodimerization domain
(HD). Other important functional domains include the LNR domain,
which comprises the three LIN12/Notch repeats; the transmembrane
segment (TM); the intracellular domain (ICN); the RAM domain; the
ankyrin repeat domain (ANK); the transactivation domain (TAD);
and the PEST domain. Some experiments used chimeric polypep-
tides, in which the RAM and ANK domains of NOTCH1 were
replaced with the DNA-binding domain of the transcription factor
6262 CHIANG ET AL.MOL. CELL. BIOL.
min at 100°C in SDS-PAGE loading buffer, the proteins were separated by
SDS-PAGE in 10% polyacrylamide gels, and detected within dried gels by
Western blot analysis. Whole-cell extracts and immunoprecipitated polypep-
tides were resolved by SDS-PAGE in 8% gels and transferred to polyvinylidene
difluoride membranes (Millipore) as described previously (3). Membranes were
stained with rabbit polyclonal antibodies against the intracellular domain of
NOTCH1 (3) or ICN1 (V1744 antibody, Cell Signaling) or with mouse mono-
clonal antibodies against GFP (Clontech), the MYC epitope (clone 9E10), or the
FLAG epitope (Sigma).
Murine bone marrow transplantation assays. All experiments were performed
as described previously (1, 4), in accordance with National Institutes of Health
guidelines for the care and use of animals, and with an approved animal protocol
from the University of Pennsylvania Animal Care and Use Committee. Briefly,
cDNAs cloned into the MigRI vector were packaged into retroviruses by tran-
sient transfection of 293T cells. After the virus titers were determined on NIH
3T3 cells, GFP-normalized retroviral supernatants were used to “spinoculate”
5-fluorouracil-treated bone marrow cells from female 4- to 8-week-old C57BL/6
mice (Taconic Farms). Transduction was performed over 48 h in a cocktail
consisting of DMEM, 10% heat-inactivated fetal bovine serum (Gibco-BRL,
Gaithersburg, MD), 5% WEHI-conditioned medium, 6 U of recombinant mouse
interleukin-3 (Genzyme Corp., Cambridge, MA)/ml, 10,000 U of recombinant
mouse interleukin-6 (Genzyme)/ml, 5 U of recombinant mouse stem cell factor
(Genzyme)/ml, 1 ?g of Polybrene (Sigma Chemical Co., St. Louis, MO)/ml, 100
U of streptomycin (Gibco-BRL)/ml, 100 U of penicillin (Gibco-BRL)/ml, and 2
mM L-glutamine (Gibco-BRL). The retrovirally transduced bone marrow cells
were then injected into lethally irradiated (900 rads) 4- to 8-week-old female
Flow cytometry. Peripheral blood samples and tumor cell suspensions were
assessed for the presence of GFP?immature T cells by flow cytometric analysis
(FACSCalibur; Becton Dickinson). Cells were incubated with phycoerythrin-
labeled anti-CD8? (53-6.7), biotinylated anti-TCR? (H57-597), and allophyco-
cyanin-labeled anti-CD4 (RM4-5) antibodies (Pharmingen). Biotinylated anti-
bodies were revealed with streptavidin-PerCP. Dead cells, identified by forward
scatter and side scatter, were excluded from the analysis. fluorescence-activated
cell sorting results were analyzed by using Flowjo software.
Southern blot analysis. High-molecular-weight DNA was isolated from fresh or
restriction enzymes overnight, fractionated by electrophoresis on a 0.8% agarose gel,
and blotted overnight onto Nytran membrane (Schleicher & Schuell, Keene, NH)
via alkaline transfer. Blots were hybridized overnight with gel-purified32P-labeled
probes corresponding to the IRES or GFP fragments of MigR1.
Histology and immunohistochemistry. To assess histology, paraffin-embedded
sections of mouse tissues fixed in 10% phosphate-buffered formalin were stained
with hematoxylin and eosin. For immunohistochemistry, sections were deparaf-
finized in xylene and graded alcohols, subjected to antigen retrieval in citrate
FIG. 2. Functional effects of C-terminal NOTCH1 deletions and mutations. (A) Conservation of C-terminal NOTCH sequences. Residues in
boldface highlight the highly conserved S4 sequence; amino acids in boldface italics correspond to S residues that are phosphorylated by
CycC:CDK8; and the asterisk denotes S2524, which is the site of the most C-terminal mutation yet detected in human T-ALL. The arrows denote
the positions of nested deletions engineered to test the function of residues in this region. Key: hN1, human NOTCH1; mN1, mouse NOTCH1;
cN1; chicken NOTCH1; xN1, Xenopus NOTCH1; fN, Drosophila NOTCH; hN2, human NOTCH2; Con., consensus. (B to D) Effects of C-terminal
deletions and mutations on NOTCH1 signal strength. In each set of experiments, NOTCH1 signaling was assessed by cotransfection of U2OS cells
with 10 ng of pcDNA3-NOTCH1 plasmid, 250 ng of CSLx4-luciferase reporter plasmid (15), and 5 ng of pRL-TK-Renilla luciferase internal control
plasmid. Normalized luciferase activities were measured in triplicate and expressed relative to an empty plasmid control. Error bars represent
standard deviations. (B) Relative effects of deletions removing residues 2473 to 2556 (?2473), 2520 to 2556 (?2520), 2535 to 2556 (?2535), or 2545
to 2556 (?2545) and the A4 mutation on signals produced by NOTCH1 polypeptides bearing a point mutation in the heterodimerization domain,
L1601P, which causes modest activation of NOTCH1 signaling (41). (C) Relative effects of a deletion removing residues 2473 to 2556 (D2473) and
the A4 mutation on signals produced by ?EGF?LNR, a form of NOTCH1 bearing a deletion removing the extracellular EGF repeats and LNR
domain (31). (D) Relative effects of the indicated point mutations in the S4 sequence on signals produced by ?EGF?LNR.
VOL. 26, 2006NEGATIVE REGULATION OF NOTCH1 6263
buffer using a pressure cooker, and then stained with rabbit polyclonal antibodies
specific for the intracellular domain of NOTCH1 (3), CD3 (Dako), and terminal
deoxytransferase (Dako). Antibody staining was developed by using the Dako
Envision kit, per the manufacturer’s instructions, and the horseradish peroxidase
Identification of WSSSSP as a NOTCH1 negative regulatory
sequence. The commonly deleted region in T-ALL contains a
number of conserved residues (Fig. 2A). We commenced our
FIG. 3. S4 affects ICN1 phosphorylation. (A) Mutation of S4 affects the electrophoretic mobility of ICN1. 293T cells transfected with the plasmids
pcDNA3-ICN1-S4 or pcDNA3-ICN1-A4 were lysed 2 days posttransfection. Immunoprecipitates prepared from these lysates with an antibody specific
for ICN1 were treated with lambda phosphatase (Sigma). Untreated whole-cell extracts (WCE) and phosphatase-treated immunoprecipitates (IP ? ?)
were analyzed on a Western blot stained with anti-NOTCH1 antibody. (B) Metabolic labeling of ICN1-S4 and ICN1-A4 with [32P]orthophosphate.
Isogenic 293 Flp-in cells engineered to express FLAG-ICN1-S4 and FLAG-ICN1-A4 were metabolically labeled with [32P]orthophosphate.32P-labeled
ICN1-S4 and ICN1-A4 were immunoprecipitated on FLAG-M2-antibody beads (Sigma) and electrophoresed in an 8% SDS-PAGE gel, which was dried
and exposed to X-ray film. The resultant autoradiogram is shown. (C) Phosphoamino acid analysis. Immunoprecipitated32P-labeled ICN1-S4 and
ICN1-A4 were excised from SDS-PAGE gels and converted to amino acids by proteolysis, followed by acid hydrolysis.32P-labeled amino acids were
spotted onto thin-layer cellulose plates (O, origin) along with phosphoserine (pS), phosphothreonine (pT), and phosphotyrosine (pY) amino acid
standards and separated by electrophoresis. After staining with ninhydrin to detect the positions of the amino acid standards, the32P-labeled amino acids
in the test samples were detected by autoradiography. (D) Two-dimensional phosphopeptide analysis. Immunoprecipitated32P-labeled ICN1-S4 and
ICN1-A4 were excised from SDS-PAGE gels, digested exhaustively with trypsin-TPCK, and spotted onto thin-layer cellulose plates (O, origin). The
32P-labeled peptides were subjected to thin-layer electrophoresis (TLE), followed by thin-layer chromatography (TLC) in the second dimension. After
drying, the positions of32P-labeled peptides were determined by autoradiography. The left pair of autoradiograms shows the complex, but similar,
patterns of phosphopeptides that are produced from ICN1-S4 and ICN1-A4. The right pair of autoradiograms focuses in on the most pronounced
differences in the observed patterns of spots, which lie in the area outlined by the boxes in the left pair of autoradiograms. Individual spots are identified
arbitrarily as A to I. The data are representative of two independent experiments. Dotted circles denote the position of spots that are absent from the
map shown and present in the corresponding map prepared from the other form of ICN1.
6264 CHIANG ET AL.MOL. CELL. BIOL.
functional analysis by creating a series of nested deletions in
full-length NOTCH1 cDNAs bearing the leukemia-associated
heterodimerization domain mutation L1601P, which causes a
modest activation of NOTCH1 signaling (Fig. 2B). We ob-
served that the deletion of residues 2545 to 2556 had little
effect on NOTCH1 signaling, whereas larger deletions span-
ning residues 2535 to 2556, 2520 to 2556, and 2473 to 2556
increased NOTCH1 signal strength in a stepwise fashion. Prior
work from Jones’ group showed that serine residues at posi-
tions 2514, 2517, and 2538 can be phosphorylated by CycC:
CDK8, an event that is hypothesized to mark ICN1 for tran-
scription-dependent degradation (13). However, the largest
stepwise increase in activity was associated with the deletion of
residues 2520 to 2534, a region where none of the serines have
been reported to be phosphorylated by CycC-CDK8. This re-
gion encompasses the position of the most C-terminal deletion
we have yet identified in human T-ALL (a stop codon muta-
tion at residue 2524 in the cell line MOLT-15 ; Fig. 2B) and
includes a short sequence, WSSSSP (designated S4), that is
restricted in the protein sequence database to a subset of
NOTCH receptors. Specifically, WSSSSP is 100% conserved
within fly NOTCH and vertebrate NOTCH1 and NOTCH2
receptors (Fig. 2A) but has diverged in NOTCH3 (WSDSTP)
and is completely absent from NOTCH4.
To test the idea that the S4 sequence might regulate
NOTCH1 function, we mutated S4 to AAAA (A4). In the
context of either full-length NOTCH1 L1601P (Fig. 2B) or a
second relatively weak gain-of-function form of NOTCH1,
?EGF?LNR (Fig. 2C), the A4 mutation stimulated signaling
to an extent comparable to the ?2473-2556 deletion. We also
investigated the effects of mutating individual S4 residues.
Each single S-to-A substitution produced a modest stimulation
in NOTCH1 activity in the context of the ?EGF?LNR
polypeptide, but no single mutation was as strong as the A4
substitution (Fig. 2D). Thus, multiple S residues within the S4
sequence appear to contribute to negative regulation of
S4 affects ICN1 phosphorylation. The S4 sequence is distinct
from previously identified sites of NOTCH1 phosphorylation.
To determine whether the S4 sequence influences ICN1 phos-
phorylation, we first compared the electrophoretic mobilities
of ICN1-S4 and ICN1-A4. The mean electrophoretic mobility
of ICN1-A4 was consistently greater than that of ICN1-S4 (Fig.
3A), a finding that could be explained by ICN-A4 being un-
derphosphorylated relative to ICN-S4. In support of this inter-
pretation, the broad bands corresponding to ICN1-S4 and
ICN1-A4 on Western blots were resolved into tight bands of
faster, but identical, electrophoretic mobility by treatment with
lambda phosphatase (Fig. 3A), suggesting that the S4 sequence
influences the phosphorylation of ICN1 by one or more pro-
To further characterize the effect of the S4 sequence on
phosphorylation, we compared the phosphoamino acid content
and the phosphopeptide maps of ICN1-S4 and ICN1-A4 pre-
pared from metabolically labeled isogenic 293 TRex cells.32P-
labeled ICN1-S4 and ICN-A4 again demonstrated a difference
FIG. 4. S4 influences ICN1 stability. (A) U2OS cells in 6-well dishes were transfected in duplicate with 1 ?g of pcDNA3 plasmids encoding
inactive ?EGF (lanes 1 and 2), constitutively active ?EGF?LNR (lanes 3 and 4), or ?EGF?LNR bearing the deletion of residues 2473 to 2556
(lanes 5 and 6) or the A4 mutation (lanes 7 and 8). Immediately posttransfection, cells were refed medium containing 0.1% DMSO or 0.1% DMSO
plus 1 ?M compound E, a ?-secretase inhibitor. At two days posttransfection, whole-cell extracts were prepared and analyzed on Western blots
stained with a polyclonal antibody specific for an epitope in the NOTCH1 transcriptional activation domain (TAD) or anti-V1744 (Cell Signaling),
a polyclonal antibody that is specific for a neo-epitope created at the N-terminal end of ICN1 by ?-secretase cleavage. (B) ICN1-A4 is more stable
than ICN1-S4. 293 cells in 60-mm dishes were transfected with 1 ?g of the pcDNA3 plasmids FLAG-ICN1-S4 or FLAG-ICN1-A4. The cells were
split 24 h posttransfection into six-well dishes. Two days posttransfection, the cells were depleted twice for 1 h in medium lacking methionine,
“pulsed” for 30 min with medium containing 2.5 mCi of [35S]methionine, and then “chased” for the indicated time periods with complete medium
supplemented with 2 mM unlabeled L-methionine.35S-labeled ICN1 polypeptides were immunoprecipitated from whole-cell lysates on FLAG-
M2-antibody beads, separated by SDS-PAGE, and detected by autoradiography. NS, nonspecific band.
VOL. 26, 2006 NEGATIVE REGULATION OF NOTCH16265
in electrophoretic mobility consistent with underphosphoryla-
tion of the A4 mutant (Fig. 3B). Phosphoserine was the major
phosphoamino acid in both ICN1-S4 and ICN1-A4 (Fig. 3C), a
finding consistent with previous analysis (7). Two-dimensional
tryptic maps prepared from ICN1-S4 and ICN1-A4 revealed a
complex, but reproducible, set of
3D). Of these phosphopeptides, one major peptide (desig-
nated peptide A) and two minor peptides (peptides B and C)
were absent from digests prepared from ICN1-A4. We also
noted that the labeling of a number of other peptides (such as
32P-labeled peptides (Fig.
peptides D to H in Fig. 3D) was reduced in ICN1-A4 com-
pared to ICN1-S4. This tendency toward a global decrease in
labeling of ICN1-A4 was consistent with the results of counting
of the excised ICN1-S4 and ICN1-A4 bands prior to tryptic
digestion, which typically revealed that ICN1-A4 incorporated
40 to 50% less32P (data not shown). On the other hand, while
the overall effect of the A4 mutation was to decrease ICN1
phosphorylation, one major phosphopeptide (designated I)
and several minor phosphopeptides (designated J and K) re-
producibly showed relatively increased labeling in ICN1-A4
FIG. 5. CDK8 is unlikely to be the S4 kinase. (A) Dominant-negative CDK8 fails to affect the difference in ICN1-S4 and ICN1-A4 phosphor-
ylation. 293 cells in 60-mm dishes were transfected with the indicated combinations of pcDNA3-ICN1-S4 or pcDNA3-ICN1-A4 (100 ng),
pCMV2-FLAG-dominant-negative CDK8 (DN-CDK8) (1 ?g), and pcDNA3-CSL-MYC (1 ?g) plasmids. At 2 days posttransfection, whole-cell
detergent extracts were prepared from each dish of transfected cells. Extracts prepared from cells cotransfected with CSL-MYC (lanes 5 to 8) were
further subjected to immunoprecipitation with anti-MYC 9E10 antibody. The upper panel shows a Western blot containing whole-cell extracts that
was stained with the FLAG M2 antibody, which recognizes FLAG-DN-CDK8. The lower panel shows a blot stained with anti-NOTCH1. In this
blot, lanes 1 to 4 contain whole-cell extracts, while lanes 5 to 8 contain NOTCH1 polypeptides that were coprecipitated with CSL-MYC.
(B) Altered MAML1 function fails to abrogate the difference in ICN1-S4 and ICN1-A4 phosphorylation. 293T cells in six-well format were
cotransfected with pcDNA3-CSL-MYC (1 ?g), pEGFP-MAML1-GFP or dominant-negative MAML1-GFP (0.5 ?g), and pcDNA3-ICN1 or
ICN1-A4 (0.1 ?g). At 2 days posttransfection, whole-cell detergent extracts and CSL-MYC immunoprecipitates were prepared as described
previously (3). Lanes 1 to 6 contain 1% of total protein inputted into the immunoprecipitates; lanes 7 to 12 show 20% of the proteins in the
corresponding immunoprecipitates. The blot was sequentially stained with anti-MYC (9E10), anti-GFP (Clontech), and anti-NOTCH1, using an
antibody against the intracellular transcriptional activation domain (3). HC and LC, immunoglobulin heavy chain and light chains, respectively.
(C) The A4 mutation increases the function of NOTCH1-GAL4 fusion proteins lacking the ankyrin repeats. U2OS cells in 24-well format were
cotransfected with 10 ng of empty pcDNA3 plasmid or pcDNA3 plasmids encoding GAL4 fusion proteins in which the RAM and ankyrin repeat
domains of NOTCH1 were replaced with the DNA-binding domain of GAL4. N1 corresponds to a full-length NOTCH1-GAL4 construct, whereas
?EGF?LNR corresponds to a construct bearing a deletion removing the EGF and LNR repeat coding regions of NOTCH1. ?EGF?LNR-A4 is
a GAL4 fusion construct in which the S4 site has been mutated to A4. These plasmids were cotransfected with 250 ng of a GAL4-luciferase reporter
plasmid bearing four GAL4 binding sites and 5 ng of the pRL-TK-Renilla luciferase internal control plasmid. Normalized luciferase activities were
measured in triplicate and are expressed relative to an empty plasmid control. Error bars represent standard deviations.
6266 CHIANG ET AL.MOL. CELL. BIOL.
compared to ICN1-S4 (Fig. 3D). Thus, these data indicate that
the A4 mutation has complex effects on ICN1 phosphorylation
that are generally, but not uniformly, inhibitory.
S4 influences ICN1 stability. The simplest way for the A4
mutation to stimulate ICN1 activity is by increasing ICN1 pro-
tein levels. We first addressed this by comparing the levels of
ICN1 that are present in cells expressing various forms of
?EGF?LNR, a NOTCH1 polypeptide that is susceptible to
ligand-independent S2 and S3 cleavages due to the absence of
the protective LNR domain (31). Relative to intact ?EGF?LNR,
?EGF?LNR polypeptides bearing the deletion ?2473-2556 or
the A4 mutation generated substantially higher steady-state levels
of ICN1 in a fashion that was sensitive to a ?-secretase inhibitor,
compound E (Fig. 4A).
To determine directly whether the A4 mutation stabilized
ICN1, pulse-chase experiments were performed with FLAG-
tagged ICN1-S4 and ICN1-A4 (Fig. 4B). This revealed that the
A4 mutation has a modest, but appreciable, stabilizing effect
on ICN1. Together with experiments, such as those in Fig. 4A,
showing that the A4 mutation allows ICN1 to accumulate to
higher levels in cells, these results provide a likely explanation
for the stimulatory effect of the A4 mutation on ICN1 function.
S4 is unlikely to be a target sequence for CDK8. Three sites
in the C terminus of ICN1 near the S4 site (S2514, S2517, and
S2538) can be phosphorylated by CDK8, which appears to be
recruited to ICN1 through the C terminus of MAML cofactors
(13). We thus performed a series of experiments to explore the
relationship of the S4 site to CDK8. Most directly, we first
FIG. 6. S4 influences the development of T-ALL. (A) ?EGF?LNR?P(?2473-2556) and ?EGF?LNR-A4 cause the development of leukemia,
whereas ?EGF?LNR causes only a transient lymphocytosis of CD4?CD8?T cells. Representative flow cytometric analyses of peripheral blood
samples drawn from mice at 6 and 13 weeks posttransplant with bone marrow cells transduced with the indicated MigRI retroviruses are shown.
(B) Kaplan-Meier curve showing that leukemia develops in mice reconstituted with bone marrow cells expressing ?EGF?LNR?P(?2473-2556),
?EGF?LNR-A4, and ICN1 NOTCH1 polypeptides but not in ?EGF?LNR or MigRI animals. Each group in this experiment contained at least
VOL. 26, 2006 NEGATIVE REGULATION OF NOTCH16267
investigated whether a “kinase-dead” dominant-negative form
of CDK8 altered the difference in phosphorylation between
ICN1-S4 and ICN1-A4, as judged by electrophoretic mobility.
We observed that ICN1-A4 remained underphosphorylated
relative to ICN1-S4 in whole-cell extracts and in CSL com-
plexes even in the presence of a 10-fold excess of dominant-
negative CDK8 plasmid (Fig. 5A). Additional experiments
were performed in cells overexpressing CSL and full-length
MAML1 (which should enhance CDK8 recruitment), or CSL
and a dominant-negative form of MAML1 (DN-MAML1).
DN-MAML consists of a 62-amino-acid kinked ?-helix that
forms a stable ternary complex through contacts on both CSL
and the ankyrin repeats of NOTCH1 (25) but which lacks the
C-terminal portions of MAML1 that are responsible for re-
cruitment of p300 and CycC:CDK8 (12, 13, 40). Thus, if S4-
targeted phosphorylation is dependent on CycC:CDK8, DN-
MAML should render ICN1-S4 equivalent to ICN1-A4, both
in terms of stability and phosphorylation, by preventing the
recruitment of CycC:CDK8. In these experiments, we adjusted
the input of the ICN1-A4 and ICN1-S4 plasmids, relative to
the CSLmyc and MAML1 plasmids, to make ICN1 the limiting
factor for complex assembly (Fig. 5B). Whether judged by
Western blots of whole-cell extracts (lanes 1 to 6) or CSL
immunoprecipitates (lanes 7 to 12), we observed differences in
electrophoretic mobility consistent with underphosphorylation
of ICN1-A4 in the presence of endogenous MAMLs (compare
lanes 1 and 2 and lanes 7 and 8), overexpressed full-length
MAML1 (compare lanes 3 and 4 and lanes 9 and 10), and
DN-MAML1 (compare lanes 5 and 6 and lanes 11 and 12). We
also observed a stabilizing effect of the A4 mutation on full-
length MAML1 and CSL when these proteins were coex-
pressed (compare the recovery of MAML1 in CSL complexes
immunoprecipitated in lanes 9 and 10). The ability of overex-
pressed MAML1 to promote degradation of CSL and ICN1 is
consistent with prior data implicating assembly of the CSL/
ICN/MAML ternary complex in ICN1 turnover (12, 13, 40).
We also studied whether the stabilizing effect of the A4
mutation could be transferred to other types of transcriptional
activation complexes. Experiments were performed with chi-
meric ?EGF?LNR polypeptides in which the RAM and
ankyrin repeat domains of ICN1 were replaced with the DNA-
binding domain of GAL4 (an approach used first by Struhl and
Adachi ). This form of NOTCH1 activates GAL4 reporter
genes in a ?-secretase-dependent fashion but fails to assemble
a ternary complex and cannot stimulate transcription from
CSL-dependent promoters (data not shown). As seen in Fig.
5C, the A4 mutation strongly stimulated the activity of
?EGF?LNR-GAL4 on a GAL4-reporter gene. Taken to-
gether, these data suggest that the S4 site is phosphorylated by
a kinase or kinases other than CycC-CDK8 and that this phos-
phorylation event does not depend on the entry of ICN1 into
the ternary complex.
S4 influences leukemogenesis. A striking feature of the
NOTCH1 tumor-associated mutations is that extracellular HD
mutations frequently occur in cis with deletions of the intra-
cellular PEST region (41). The data described above suggested
that loss of S4-associated regulation of ICN1 might enhance
the leukemogenic activity of weak gain-of-function forms of
activated NOTCH1, which are not themselves leukemogenic,
and might thus mimic the effect of tumor-associated mutations
found in cis. To assess this possibility, we compared the leu-
kemogenic activity of the weak gain-of-function ?EGF?LNR
form of NOTCH1 with ?EGF?LNR-A4 in a murine bone
marrow transplantation assay. The selection of ?EGF?LNR
for these experiments was based on the observations showing
that although this form of NOTCH1 generated signals of suf-
ficient strength to drive T-cell development (31), it did not
induce T-ALL (Fig. 6A). Typically, mice reconstituted with
bone marrow cells expressing ?EGF?LNR developed a GFP?
CD4?CD8?immature “double-positive” T-cell population by
6 weeks after transplantation that disappeared by 13 weeks
posttransplantation. None of these animals have developed
leukemia at times greater than 1 year posttransplantation (Fig.
6B and data not shown).
In contrast, the double-positive T-cell count in the periph-
eral blood of mice reconstituted with bone marrow cells ex-
pressing the ?EGF?LNR-A4 continued to rise (Fig. 6A), and
all of these animals eventually became moribund and suc-
cumbed to disseminated leukemia (Fig. 6B). The disease la-
tency in ?EGF?LNR-A4 mice was similar to that seen in
animals reconstituted with bone marrow cells expressing
?EGF?LNR?PEST (which bears a deletion removing the 73
C-terminal amino acids of NOTCH1) and ICN1, a strong gain-
of-function form of NOTCH1 (Fig. 6B). At necropsy, leukemic
blasts replaced the bone marrow and heavily infiltrated the
spleen, liver, lymph nodes, and viscera such as the kidneys
FIG. 7. Pathology of ?EGF?LNR-A4-induced T-ALL. Sections
stained with hematoxylin and eosin (H?E) show lymphoblasts infil-
trating the liver and the kidney. In the bottom four panels, tumor cells
are shown in sections of liver stained with H?E or antibodies specific
for CD3, terminal deoxytransferase (TdT), and NOTCH1. Immuno-
staining was developed by method that produces a brown color (he-
6268CHIANG ET AL.MOL. CELL. BIOL.
(Fig. 7). Immunohistochemical stains confirmed that these
blasts were immature T cells expressing CD3, terminal de-
oxytransferase (TdT), and NOTCH1 (Fig. 7). Western blots
prepared from heavily infiltrated spleens confirmed the pres-
ence of NOTCH1 polypeptides of the expected size of the
precursor and furin-processed forms of ?EGF?LNR-A4 (Fig.
8). Further workup of ?EGF?LNR-A4 tumors included flow
cytometry, which revealed that the tumors expressed surface
CD3 and variable levels of CD4 and CD8 (data not shown). As
anticipated given the results of Western blotting, the tumors
contained intact proviruses and were monoclonal or oligo-
clonal based on the presence of one or several dominant pro-
viral insertions (Fig. 9 and data not shown).
Our studies indicate that the sequence WSSSSP (S4), which
spans residues 2521 to 2526 of human NOTCH1, exerts a
functionally important restraint on NOTCH1 signal strength,
and influences the development of T-ALL in a murine model.
It follows that the C-terminal NOTCH1 deletions that are
often seen in human T-ALL contribute to leukemogenesis, at
least in part, by removing this sequence. Consistent with this
possibility, the most C-terminal deletion that we have yet ob-
served in T-ALL starts within this sequence at residue 2524
(41). The gains in ICN1 function that occur when the S4
sequence is mutated are correlated with changes in ICN1 phos-
phorylation, making it probable that this sequence acts by
altering the recognition of ICN1 by one or more kinases. Be-
cause it is necessary to mutate multiple S residues within S4 to
maximally activate NOTCH1, it is possible that one or more
kinases phosphorylate more than one site within this sequence.
More broadly, mutation of the S4 site appears to affect the
phosphorylation of multiple ICN1 tryptic peptides, suggesting
that the phosphorylation status of S4 influences the recognition
of ICN1 by other kinases. NOTCH signaling is tightly coordi-
nated with other signaling pathways, and it is possible that S4
phosphorylation may serve to prime ICN1 for recognition by
other kinases. The existence of multiple kinases that target
ICN1 for degradation through S4 and adjacent residues such as
those recognized by CDK8 provides a reasonable explanation
for why deletions involving the C terminus of NOTCH1 are
common in T-ALL, whereas point mutations are rare (41).
In addition to previously recognized retroviral insertions
(10, 14), several recent reports describe acquired frameshift or
stop codon mutations in diverse murine T-ALL models that
FIG. 8. Detection of NOTCH1 polypeptides in ?EGF?LNR?P(?2473-2556) and ?EGF?LNR-A4-induced T-ALLs. Detergent extracts from
spleens heavily involved by T-ALLs arising in ?EGF?LNR?P(?2473-2556) and ?EGF?LNR-A4 mice were analyzed on Western blots stained
with a polyclonal antibody specific for the intracellular domain of NOTCH1. (A) Cross-reactive NOTCH1 polypeptides are observed in both types
of tumors of the expected size for pro-?EGF?LNR?, pro-?EGF?LNRA4, and the furin-processed mature forms of each of these polypeptides
(NTM?PEST and NTM-A4, respectively). (B) The size of the Notch1 polypeptides in panel A is compared to NOTCH1 polypeptides expressed in
the cell line ALL-SIL, which has a normal NOTCH1 allele that gives rise to NTM, and a mutated NOTCH1 allele containing the same PEST
deletion (residues 2473 to 2556) that has been engineered into the ?EGF?LNR? expression plasmid (41). The deleted allele gives rise to NTM?P
after furin processing.
VOL. 26, 2006 NEGATIVE REGULATION OF NOTCH16269
produce C-terminal truncations of NOTCH1 (20, 26). The
most C-terminal deletion noted in murine tumors to date falls
at amino acid residue 2490 in murine NOTCH1 (20), which is
the equivalent of residue 2516 in human NOTCH1; thus, the
mutations that occur in murine T-ALL also consistently delete
the S4 sequence. While these correlations (and the studies
directed at the S4 sequence described here) do not preclude
additional important contributions of other sequences in the
C-terminal tail to the negative regulation of normal and patho-
physiologic NOTCH1 signaling, they are consistent with a ma-
jor role for the S4 sequence in T-ALL.
Several questions arise from this work, the most immediate
of which concerns the identity of the kinase(s) that targets
ICN1 through S4. After MAML-dependent recruitment to the
NOTCH1 transcription complex on DNA, CycC-CDK8 can
phosphorylate at least three serine residues in the far C-ter-
minal region of NOTCH1, including one within the minimal
deleted region (41). However, ICN1-A4 is still underphosphor-
ylated relative to ICN1-S4 even under a variety of conditions in
which ternary complex formation and recruitment of CDK8
are defective. These findings suggest that the S4 kinase is likely
to be different than CycC-CDK8. Other candidates include the
GSK3? (9, 11) and MEK/ERK kinases, based on the well-
recognized, but complex, functional interactions between the
RAS and NOTCH signaling pathways (37). However, experi-
ments conducted with pharmacologic inhibitors of MEK and
GSK3?, RNAi directed against GSK3?, and constitutively ac-
tive forms of MEK1, have failed to detect evidence of epistasis
between these kinases and the S4 sequence (J. C. Aster, data
not shown). It is likely that unbiased functional screens will be
necessary to identify the kinase(s) that is responsible for S4-
dependent ICN1 phosphorylation.
We have also failed to see evidence of epistasis between
dominant-negative forms of FBW7 (a homolog of Sel-10) and
S4 (J. C. Aster, data not shown). In this regard, it is relevant
that the stability of mammalian NOTCH4 is regulated by
FBW7/Sel-10 (38, 39) despite the absence of the S4 sequence
from this Notch receptor. We thus favor the idea that S4-
dependent modulation of ICN1 levels depends on factors other
than FBW7/Sel-10. It will be of interest to determine whether
mutations in the kinases or the destruction machinery respon-
sible for S4-dependent ICN1 degradation will be identified in
human T-ALLs lacking C-terminal NOTCH1 deletions.
The C-terminal NOTCH1 deletions that are found in human
T-ALL often occur in concert with mutations involving the
heterodimerization domain of NOTCH1 that lead to increased
ICN1 production (41). We have also further characterized here
a form of NOTCH1 with a mutation affecting the ectodomain,
?EGF?LNR, which drives abnormal double positive T-cell
development (31) without causing T-ALL. This demonstrates
for the first time that the threshold dose of NOTCH1 signals
that is required for efficient induction of T-ALL development
is higher than that which is required to induce T-cell develop-
ment. This raises the question of whether certain heterodimer-
ization domain mutations will suffice to create signals that are
strong enough to induce T-ALL, or whether they will require
additional events in cis or in trans. The “hypoleukemic” phe-
notype induced in vivo by weak gain-of-function NOTCH1
alleles such as ?EGF?LNR should be useful in identifying not
only cis-acting elements such as the S4 sequence but also ele-
ments that act in trans. Such trans-acting factors could either
elevate NOTCH1 signaling tone directly or, by complementing
NOTCH1 functions that require strong signals, lower the dose
of NOTCH1 that is required for efficient induction of T-ALL.
We thank Hong Sai and Zhigang Li for excellent technical assis-
tance. We are also grateful to the John Morgan and Stemmler ASU
(University of Pennsylvania), the Abramson Cancer Center Flow Cy-
tometry Core (University of Pennsylvania), the AFCRI Core (Univer-
sity of Pennsylvania), and the Hematopathology Core of the Dana
Farber/Harvard Cancer Center.
This study was supported by grants from the National Institutes of
Health to W.S.P. (CA93615 and AI47833), J.C.A. (CA82308), S.C.B.
(CA92433), and D.B.S. (CA75205) and from the Leukemia and Lym-
phoma Society SCOR Program. M.L.X. was the recipient of an Amer-
ican Society of Hematology Medical Student Trainee award. M.C. was
1. Allman, D., F. G. Karnell, J. A. Punt, S. Bakkour, L. Xu, P. Myung, G. A.
Koretzky, J. C. Pui, J. C. Aster, and W. S. Pear. 2001. Separation of Notch1
promoted lineage commitment and expansion/transformation in developing
T cells. J. Exp. Med. 194:99–106.
2. Artavanis-Tsakonas, S., M. D. Rand, and R. J. Lake. 1999. Notch signaling:
cell fate control and signal integration in development. Science 284:770–776.
3. Aster, J. C., E. S. Robertson, R. P. Hasserjian, J. R. Turner, E. Kieff, and J.
Sklar. 1997. Oncogenic forms of NOTCH1 lacking either the primary bind-
ing site for RBP-J? or nuclear localization sequences retain the ability to
FIG. 9. Analysis of proviral integrity and integration in A4-induced
T-ALLs. (A) Southern blot of EcoRV-digested genomic DNAs iso-
lated from four spleens involved by A4-induced T-ALLs. EcoRV,
which cleaves twice within the provirus at sites flanking the
?EGF?LNR-A4 cDNA, is expected to release a proviral fragment of
5.9 kb from the intact ?EGF?LNR-A4 provirus when hybridized to a
probe specific for the IRES sequence. The correctly sized fragment is
shown in the right-hand lane, which contains MigRI-?EGF?LNR-A4
and MigRI-NOTCH1 plasmid DNAs digested with Eco RV. Each of
the left-hand four lanes contain genomic DNA cut with EcoRV obtained
from spleens heavily infiltrated by different MigRI-?EGF?LNR-A4 T-
ALLs. (B) Genomic DNAs from the same four spleens shown in panel A
were analyzed by digestion with EcoRI, which cleaves once within the
DNA fragment. The Southern blot hybridized to a 592-bp ECMV IRES
probe. Single proviruses are readily seen in tumors 1, 4, and 5, whereas a
single faint proviral band is seen in tumor 3 (arrowheads). NS, nonspecific
band. In both panels A and B, the numbers correspond to sizes in kilo-
6270 CHIANG ET AL.MOL. CELL. BIOL.
associate with RBP-J? and activate transcription. J. Biol. Chem. 272:11336– Download full-text
4. Aster, J. C., L. Xu, F. G. Karnell, V. Patriub, J. C. Pui, and W. S. Pear. 2000.
Essential roles for ankyrin repeat and transactivation domains in induction
of T-cell leukemia by notch1. Mol. Cell. Biol. 20:7505–7515.
5. Blaumueller, C. M., H. Qi, P. Zagouras, and S. Artavanis-Tsakonas. 1997.
Intracellular cleavage of Notch leads to a heterodimeric receptor on the
plasma membrane. Cell 90:281–291.
6. Brou, C., F. Logeat, N. Gupta, C. Bessia, O. LeBail, J. R. Doedens, A.
Cumano, P. Roux, R. A. Black, and A. Israel. 2000. A novel proteolytic
cleavage involved in Notch signaling: the role of the disintegrin-metallopro-
tease TACE. Mol. Cell 5:207–216.
7. Cagan, R. L., and D. F. Ready. 1989. Notch is required for successive cell
decisions in the developing Drosophila retina. Genes Dev. 3:1099–1112.
8. De Strooper, B., W. Annaert, P. Cupers, P. Saftig, K. Craessaerts, J. S.
Mumm, E. H. Schroeter, V. Schrijvers, M. S. Wolfe, W. J. Ray, A. Goate, and
R. Kopan. 1999. A presenilin-1-dependent ?-secretase-like protease medi-
ates release of Notch intracellular domain. Nature 398:518–522.
9. Espinosa, L., J. Ingles-Esteve, C. Aguilera, and A. Bigas. 2003. Phosphory-
lation by glycogen synthase kinase-3? down-regulates Notch activity, a link
for Notch and Wnt pathways. J. Biol. Chem. 278:32227–32235.
10. Feldman, B. J., T. Hampton, and M. L. Cleary. 2000. A carboxy-terminal
deletion mutant of Notch1 accelerates lymphoid oncogenesis in E2A-PBX1
transgenic mice. Blood 96:1906–1913.
11. Foltz, D. R., M. C. Santiago, B. E. Berechid, and J. S. Nye. 2002. Glycogen
synthase kinase-3? modulates notch signaling and stability. Curr. Biol. 12:
12. Fryer, C. J., E. Lamar, I. Turbachova, C. Kintner, and K. A. Jones. 2002.
Mastermind mediates chromatin-specific transcription and turnover of the
Notch enhancer complex. Genes Dev. 16:1397–1411.
13. Fryer, C. J., J. B. White, and K. A. Jones. 2004. Mastermind recruits CycC:
CDK8 to phosphorylate the Notch ICD and coordinate activation with
turnover. Mol. Cell 16:509–520.
14. Hoemann, C. D., N. Beaulieu, L. Girard, N. Rebai, and P. Jolicoeur. 2000.
Two distinct Notch1 mutant alleles are involved in the induction of T-cell
leukemia in c-myc transgenic mice. Mol. Cell. Biol. 20:3831–3842.
15. Hsieh, J. J., T. Henkel, P. Salmon, E. Robey, M. G. Peterson, and S. D.
Hayward. 1996. Truncated mammalian Notch1 activates CBF1/RBPJk-re-
pressed genes by a mechanism resembling that of Epstein-Barr virus
EBNA2. Mol. Cell. Biol. 16:952–959.
16. Jarriault, S., C. Brou, F. Logeat, E. H. Schroeter, R. Kopan, and A. Israel.
1995. Signalling downstream of activated mammalian Notch. Nature 377:
17. Joyal, J. L., and D. B. Sacks. 1994. Insulin-dependent phosphorylation of
calmodulin in rat hepatocytes. J. Biol. Chem. 269:30039–30048.
18. Kimberly, W. T., W. P. Esler, W. Ye, B. L. Ostaszewski, J. Gao, T. Diehl, D. J.
Selkoe, and M. S. Wolfe. 2003. Notch and the amyloid precursor protein are
cleaved by similar ?-secretase(s). Biochemistry 42:137–144.
19. Kopan, R., E. H. Schroeter, H. Weintraub, and J. S. Nye. 1996. Signal
transduction by activated mNotch: importance of proteolytic processing and
its regulation by the extracellular domain. Proc. Natl. Acad. Sci. USA 93:
20. Lin, Y. W., R. A. Nichols, J. J. Letterio, and P. D. Aplan. 2006. Notch1
mutations are important for leukemic transformation in murine models of
precursor-T leukemia/lymphoma. Blood 107:2540–2543.
21. Logeat, F., C. Bessia, C. Brou, O. LeBail, S. Jarriault, N. G. Seidah, and A.
Israel. 1998. The Notch1 receptor is cleaved constitutively by a furin-like
convertase. Proc. Natl. Acad. Sci. USA 95:8108–8112.
22. MacKichan, M. L., F. Logeat, and A. Israel. 1996. Phosphorylation of p105
PEST sequence via a redox-insensitive pathway up-regulates processing of
p50 NF-?B. J. Biol. Chem. 271:6084–6091.
23. McGill, M. A., and C. J. McGlade. 2003. Mammalian numb proteins pro-
mote Notch1 receptor ubiquitination and degradation of the Notch1 intra-
cellular domain. J. Biol. Chem. 278:23196–23203.
24. Mumm, J. S., E. H. Schroeter, M. T. Saxena, A. Griesemer, X. Tian, D. J.
Pan, W. J. Ray, and R. Kopan. 2000. A ligand-induced extracellular cleavage
regulates ?-secretase-like proteolytic activation of Notch1. Mol. Cell 5:197–
25. Nam, Y., P. Sliz, L. Song, J. C. Aster, and S. Blacklow. 2006. Structural basis
for cooperativity in recruitment of the mastermind coactivator to Notch
transcription complexes. Cell 124:973–983.
26. O’Neil, J., J. Calvo, K. McKenna, V. Krishnamoorthy, J. C. Aster, C. H.
Bassing, F. W. Alt, M. Kelliher, and A. T. Look. 2006. Activating Notch1
mutations in mouse models of T-ALL. Blood 107:781–785.
27. Petcherski, A. G., and J. Kimble. 2000. LAG-3 is a putative transcriptional
activator in the Caenorhabditis elegans Notch pathway. Nature 405:364–368.
28. Petcherski, A. G., and J. Kimble. 2000. Mastermind is a putative activator for
Notch. Curr. Biol. 10:R471–R473.
29. Rand, M. D., L. M. Grimm, S. Artavanis-Tsakonas, V. Patriub, S. C. Blacklow,
J. Sklar, and J. C. Aster. 2000. Calcium depletion dissociates and activates
heterodimeric notch receptors. Mol. Cell. Biol. 20:1825–1835.
30. Rechsteiner, M. 1989. PEST regions, proteolysis, and cell cycle progression.
Revis. Biol. Celular. 20:235–253.
31. Sanchez-Irizarry, C., A. C. Carpenter, A. P. Weng, W. S. Pear, J. C. Aster,
and S. C. Blacklow. 2004. Notch subunit heterodimerization and prevention
of ligand-independent proteolytic activation depend, respectively, on a novel
domain and the LNR repeats. Mol. Cell. Biol. 24:9265–9273.
32. Schroeter, E. H., J. A. Kisslinger, and R. Kopan. 1998. Notch-1 signalling
requires ligand-induced proteolytic release of intracellular domain. Nature
33. Shah, S., S. F. Lee, K. Tabuchi, Y. H. Hao, C. Yu, Q. LaPlant, H. Ball, C. E.
Dann III, T. Sudhof, and G. Yu. 2005. Nicastrin functions as a ?-secretase-
substrate receptor. Cell 122:435–447.
34. Shaye, D. D., and I. Greenwald. 2005. LIN-12/Notch trafficking and regula-
tion of DSL ligand activity during vulval induction in Caenorhabditis elegans.
35. Struhl, G., and A. Adachi. 1998. Nuclear access and action of notch in vivo.
36. Struhl, G., and I. Grenwald. 1999. Presenilin is required for activity and
nuclear access of Notch in Drosophila. Nature 398:522–525.
37. Sundaram, M. V. 2005. The love-hate relationship between Ras and Notch.
Genes Dev. 19:1825–1839.
38. Tetzlaff, M. T., W. Yu, M. Li, P. Zhang, M. Finegold, K. Mahon, J. W.
Harper, R. J. Schwartz, and S. J. Elledge. 2004. Defective cardiovascular
development and elevated cyclin E and Notch proteins in mice lacking the
Fbw7 F-box protein. Proc. Natl. Acad. Sci. USA 101:3338–3345.
39. Tsunematsu, R., K. Nakayama, Y. Oike, M. Nishiyama, N. Ishida, S.
Hatakeyama, Y. Bessho, R. Kageyama, T. Suda, and K. I. Nakayama. 2004.
Mouse Fbw7/Sel-10/Cdc4 is required for notch degradation during vascular
development. J. Biol. Chem. 279:9417–9423.
40. Wallberg, A. E., K. Pedersen, U. Lendahl, and R. G. Roeder. 2002. p300 and
PCAF act cooperatively to mediate transcriptional activation from chromatin
templates by notch intracellular domains in vitro. Mol. Cell. Biol. 22:7812–7819.
41. Weng, A. P., A. A. Ferrando, W. Lee, J. P. T. Morris, L. B. Silverman, C.
Sanchez-Irizarry, S. C. Blacklow, A. T. Look, and J. C. Aster. 2004. Activat-
ing mutations of NOTCH1 in human T-cell acute lymphoblastic leukemia.
42. Weng, A. P., Y. Nam, M. S. Wolfe, W. S. Pear, J. D. Griffin, S. C. Blacklow,
and J. C. Aster. 2003. Growth suppression of pre-T acute lymphoblastic
leukemia cells by inhibition of notch signaling. Mol. Cell. Biol. 23:655–664.
43. Wu, L., J. C. Aster, S. C. Blacklow, R. Lake, S. Artavanis-Tsakonas, and J. D.
Griffin. 2000. MAML1, a human homologue of Drosophila mastermind, is a
transcriptional coactivator for NOTCH receptors. Nat. Genet. 26:484–489.
44. Ye, Y., N. Lukinova, and M. E. Fortini. 1999. Neurogenic phenotypes and
altered Notch processing in Drosophila Presenilin mutants. Nature 398:525–
VOL. 26, 2006NEGATIVE REGULATION OF NOTCH1 6271