Article

Loss of the Tumor Suppressor CYLD Enhances Wnt/β-Catenin Signaling through K63-Linked Ubiquitination of Dvl

Department of Cell Biology, University Medical Center Utrecht, Heidelberglaan 100, 3584CX Utrecht, The Netherlands.
Molecular cell (Impact Factor: 14.02). 03/2010; 37(5):607-19. DOI: 10.1016/j.molcel.2010.01.035
Source: PubMed
ABSTRACT
The mechanism by which Wnt receptors transduce signals to activate downstream beta-catenin-mediated target gene transcription remains incompletely understood but involves Frizzled (Fz) receptor-mediated plasma membrane recruitment and activation of the cytoplasmic effector Dishevelled (Dvl). Here, we identify the deubiquitinating enzyme CYLD, the familial cylindromatosis tumor suppressor gene, as a negative regulator of proximal events in Wnt/beta-catenin signaling. Depletion of CYLD from cultured cells markedly enhances Wnt-induced accumulation of beta-catenin and target gene activation. Moreover, we demonstrate hyperactive Wnt signaling in human cylindroma skin tumors that arise from mutations in CYLD. At the molecular level, CYLD interacts with and regulates K63-linked ubiquitination of Dvl. Enhanced ubiquitination of the polymerization-prone DIX domain in CYLD-deficient cells positively links to the signaling activity of Dvl. Together, our results argue that loss of CYLD instigates tumor growth in human cylindromatosis through a mechanism in which hyperubiquitination of polymerized Dvl drives enhancement of Wnt responses.
Molecular Cell
Article
Loss of the Tumor Suppressor CYLD
Enhances Wnt/b-Catenin Signaling
through K63-Linked Ubiquitination of Dvl
Daniele V.F. Tauriello,
1,5
Andrea Haegebarth,
3,5
Ineke Kuper,
1
Mariola J. Edelmann,
4
Marre Henraat,
1
Marijke R. Canninga-van Dijk,
2
Benedikt M. Kessler,
4
Hans Clevers,
3
and Madelon M. Maurice
1,
*
1
Department of Cell Biology
2
Department of Pathology
University Medical Center Utrecht, Heidelberglaan 100, 3584CX Utrecht, The Netherlands
3
Hubrecht Institute, Uppsalalaan 8, 3584CT Utrecht, The Netherlands
4
Henry Wellcome Building for Molecular Physiology, Nuffield Department of Clinical Medicine, University of Oxford, Roosevelt Drive,
Oxford OX3 7BN, UK
5
These authors contributed equally to this work
*Correspondence: m.m.maurice@umcutrecht.nl
DOI 10.1016/j.molcel.2010.01.035
SUMMARY
The mechanism by which Wnt receptors transduce
signals to activate downstream b-catenin-mediated
target gene transcription remains incompletely
understood but involves Frizzled (Fz) receptor-medi-
ated plasma membrane recruitment and activation
of the cytoplasmic effector Dishevelled (Dvl). Here,
we identify the deubiquitinating enzyme CYLD, the
familial cylindromatosis tumor suppressor gene, as a
negative regulator of proximal events in Wnt/b-cate-
nin signaling. Depletion of CYLD from cultured cells
markedly enhances Wnt-induced accumulation of
b-catenin and target gene activation. Moreover, we
demonstrate hyperactive Wnt signaling in human
cylindroma skin tumors that arise from mutations
in CYLD. At the molecular level, CYLD interacts
with and regulates K63-linked ubiquitination of Dvl.
Enhanced ubiquitination of the polymerization-prone
DIX domain in CYLD-deficient cells positively links
to the signaling activity of Dvl. Together, our results
argue that loss of CYLD instigates tumor growth in
human cylindromatosis through a mechanism in
which hyperubiquitinatio n of polymerized Dvl drives
enhancement of Wnt responses.
INTRODUCTION
The Wnt/b-catenin signaling pathway is critical in tissue
patterning in development and adult tissue homeostasis, and
is frequently misregulated in cancer (Clevers, 2006; Logan and
Nusse, 2004). Binding of Wnt to its seven-span transmembrane
receptor, Frizzled (Fz), and single-span coreceptor, Lrp5/6, at
the cell surface initiates a signaling cascade that results in
disruption of the default degradation of cytoplasmic b-catenin,
allowing for its nuclear translocation and promotion of target
gene transcription. Central to the earliest signaling events is
the recruitment of the cytosolic protein, Dishevelled (Dvl), to
the C terminus of Fz (Axelrod et al., 1998; Umbhauer et al.,
2000). The molecular events that couple Fz/Dvl complex forma-
tion to stabilization of b-catenin remain incompletely understood
but clearly involve Dvl-mediated recruitment and inhibition of the
negative regulator, Axin (Cliffe et al., 2003; Kishida et al., 1999;
Penton et al., 2002; Smalley et al., 1999). One of the key molec-
ular features of Dvl held responsible for its role in Wnt/b-catenin
signaling is its ability to self-associate and form a dynamic poly-
meric scaffold during signaling responses (Schwarz-Romond
et al., 2005, 2007a). Dvl-mediated coclustering of Fz, Lrp6,
Axin, and associated kinases at the plasma membrane triggers
phosphorylation of the Lrp6 cytoplasmic tail, thereby promoting
further Axin recruitment and b-catenin stabilization (Bilic et al.,
2007; Davidson et al., 2005; Tamai et al., 2004; Zeng et al.,
2005).
To better understand the molecular basis of Wnt signaling, we
investigated the regulatory role of ubiquitination in these
signaling events. Posttranslational modification of proteins with
ubiquitin is crucial in a vast array of cellular and biological
processes (Kerscher et al., 2006; Kirkin and Dikic, 2007). Ubiqui-
tin is selectively attached to substrate proteins by members of
the large family of E3 ubiquitin ligases that act in concert with
E2-conjugating enzymes. Proteins can be modified with a single
ubiquitin molecule or with elongated polyubiquitin chains that
arise through repetitive attachment of ubiquitin to one of seven
lysines in the preceding ubiquitin molecule. The selective use
of lysines in chain formation allows for formation of structurally
and functionally distinct polyubiquitin chains. Whereas attach-
ment of lysine 48 (K48)-linked ubiquitin chains destines proteins
for proteasomal degradation, K63-linked polyubiquitination
controls nonproteolytic events in DNA repair, endocytosis, and
kinase-mediated signaling responses ( Sun and Chen, 2004).
A number of E3 ubiquitin ligases were identified to control the
stability and/or activity of the core Wnt pathway proteins b-cat-
enin (Jiang and Struhl, 1998), APC (Tran et al., 2008), and Dvl
(Angers et al., 2006), thus implicating ubiquitin-mediated control
of Wnt signaling events at multiple levels.
Molecular Cell 37, 607–619, March 12, 2010 ª2010 Elsevier Inc. 607
Page 1
0
Relative fold activation
A
B
Fold activation
Luciferase activity
AXIN2/GAPDH
β-catenin/actin
D
E
G
F
C
L-CM
Wnt3a-CM
*
*
shCYLD-1 (h)
shCYLD-2 (h)
1
1.6
1.2
0.8
0.4
0
2 3 4
Wnt3a-CM
shCYLD-1
con
Wnt3a-CM
- + + -
β-catenin
actin
shCYLD-1
con
- + + -
0 24 48
0 48
dub1
dub3
dub5
dub8
dub11
dub13
dub15
dub17
dub20
dub22
dub24
dub26
dub28
dub30
dub32
dub34
dub36
dub38
dub40
dub42
dub45
dub47
dub49
800
600
400
200
0.020
0.016
0.012
0.008
0.004
0
TOPflash
FOPflash
2.5
1.5
0.5
2.0
1.0
0
1 2 3 4
1 2 3 4
0
2
4
6
8
Fold activation
Wnt3a-CM
293T
shRNA
con
CYLD
con
CYLD
con
CYLD
0
20
40
60
80
0
10
20
30
40
50
0
10
20
30
40
--++ - -++ - -++
HeLa
BT549
shCYLD
con
HA-CYLD
shUSP7
con
con
shCYLD-1
shCYLD-2
shCYLD-1
shCYLD-2
Molecular Cell
CYLD Regulates Dvl Ubiquitination in Wnt Signaling
608 Molecular Cell 37, 607–619, March 12, 2010 ª2010 Elsevier Inc.
Page 2
Ubiquitination of proteins is a reversible process. The removal
of ubiquitin is carried out by deubiquitinating enzymes (DUBs),
comprising five protease subclasses that can release ubiquitin
chains from proteins targeted for degradation, reverse regulatory
ubiquitination, and edit inappropriately ubiquitinated proteins.
Aberrant DUB expression and function has been implicated in
growth control and oncogenesis, neurodegenerative diseases,
and defective embryonic development (Nijman et al., 2005).
Despite the critical role of DUBs, our knowledge of their regula-
tion and substrate specificity remains limited.
Here, we have screened for involvement of DUBs in the Wnt/
b-catenin signaling pathway. We identify the DUB and tumor
suppressor, CYLD, as a negative regulator of proximal Wnt/
b-catenin signaling events. We confirm the clinical relevance of
this finding by demonstration of hyperactive Wnt signaling in
skin appendage tumors that arise in CYLD-mutant patients. At
the molecular level, we show that CYLD interacts with Dvl and
regulates K63-linked ubiquitination of the polymerization-prone
DIX domain of Dvl. These results uncover a mechanism by which
hyperubiquitination of polymeric Dvl drives transmission of the
Wnt signal in CYLD-depleted cells.
RESULTS
Identification of CYLD as a Suppressor
of Wnt/b-Catenin Signaling
To identify functional DUBs in Wnt/b-catenin signaling, we used
an shRNA-based library that targets 50 different DUBs of the
USP family (Brummelkamp et al., 2003). In this library, each
DUB is targeted by a pool of four shRNA vectors. As a model
system, we used human embryonic kidney (HEK) 293T cells
that carry no known mutations in Wnt signaling components,
allowing us to screen for DUB involvement in both proximal
and distal signaling events. We investigated the effects of DUB
knockdown on Wnt-induced b-catenin-mediated transcriptional
activation with the TOPFlash luciferase reporter, a highly specific
and quantitative readout of Wnt pathway activity (Korinek et al.,
1997). Knockdown of one DUB (no. 36) significantly enhanced
TOPFlash reporter activity induced by Wnt1 transfection
(Figure 1A). The targeted DUB in this library sample is the cylin-
dromatosis tumor suppressor, CYLD (Bignell et al., 2000). Similar
results were obtained when targeting CYLD with the single active
shRNA (shCYLD-1) isolated from the pooled sample (Brummel-
kamp et al., 2003) while stimulating cells with Wnt3a-conditioned
medium (Figures 1B and 1C). CYLD knockdown did not enhance
basal levels of TOPFlash reporter activity in HEK293T cells, but
strongly increased the response to Wnt3a (Figure 1C). To reduce
the chances of off-target effects as an explanation of our results,
we generated a second, nonoverlapping shRNA with compa-
rable ability to reduce endogenous CYLD levels (shCYLD-2;
Figure 1B). Knockdown of CYLD with this second shRNA equally
enhanced Wnt3a-induced TOPFlash reporter activity (Figure 1C).
Knockdown of another DUB, USP7, did not influence Wnt
pathway activation (Figure 1D). Furthermore, overexpression of
CYLD reduced TOPFlash reporter activity (Figure 1D). We
conclude that Wnt/b-catenin signaling is controlled by CYLD
function. These findings are corroborated by enhanced Wnt3a-
induced stabilization of endogenous b-catenin in cells treated
with both shCYLD-1 and shCYLD-2 shRNAs (Figure 1E; see
also Figure S1A available online) and enhanced transcription of
AXIN2, a well-defined target gene of Wnt signaling (Behrens
et al., 1998), in CYLD-depleted cells (Figure 1F and Figure S1B).
The enhancement of Wnt-induced signaling by removal of CYLD
was reproduced in two other cell lines of cervical and breast
epithelial cell origin (Figure 1G), suggesting mechanistic conser-
vation between tissues of CYLD-mediated regulation of Wnt
signaling. In
HeLa cells, we consistently observed enhanced
basal levels of Wnt signaling upon deletion of CYLD, suggesting
that, in some cases, loss of CYLD may be sufficient for activation
of the pathway.
CYLD-Mediated Regulation of Wnt Signaling
Is Independent of NF-kB
In previous studies, CYLD was identified as a regulator of NF- kB
signaling (Brummelkamp et al., 2003; Kovalenko et al., 2003;
Trompouki et al., 2003). We investigated whether indirect,
NF-kB-mediated effects could be responsible for the modulation
of Wnt signaling by CYLD. Stimulation of CYLD knockdown cells
with TNFa did not induce TOPFlash reporter activity (Figure 2A),
whereas NF-kB reporter activity was strongly induced, as
expected (Figure 2B). Wnt3a-induced Wnt signaling activity
remained unaffected by costimulation with TNFa (Figure 2C).
These results argue that CYLD-mediated control of Wnt
signaling is independent of its effects on the NF-kB pathway.
Figure 1. CYLD Is a Negative Regulator of Wnt Signaling
(A) Screen for functional DUBs in Wnt signaling. HEK293T cells were transfected with either TCF optimal (TOPFlash) or mutant negative control (FOPFlash)
luciferase reporter, together with TK-Renilla reporter, individual samples of the DUB knockdown library, and Wnt1. Relative TOPFlash ov er FOPFlash luciferase
ratios SD) of duplicate experiments, normalized against the average value, are shown.
(B) Western blot analysis of endogenous CYLD knockdown (arrow) by two nonoverlapping shRNA vectors in HEK293T cells. Asterisks indicate a background
band.
(C) HEK293T cells were transfected with shCYLD-1, shCYLD-2, or shGFP control (con), together with TOPFlash or FOPFlash, and TK-Renilla. Bars represent
TOPFlash over FOPFlash luciferase ratios (mean ± SD) of cells treated with Wnt3a conditioned medium (Wnt3a-CM) or control L-cell medium (L-CM). Results
are representative of at least three experiments done in duplicate.
(D) Comparison of shCYLD, shUSP7, HA-CYLD, and mock plasmid expression in HEK293T cells on TOPFlash and FOPFlash luciferase activity.
(E) (Top) Western blot analysis of Wnt-induced accumulation of endogenous cytosolic b-catenin in contro l and shCYLD-1 knockdown cells. Actin loading control
is shown. (Bottom) quantification of the results from three independent experiments (mean ± SD).
(F) Quantified results of Northern blot analysis of endogenous Axin2 and control GAPDH mRNA ratios in HEK293T cells, transfected with either shCYLD-1 or
shGFP vector and treated overnight with L-CM or Wnt3a-CM. Results (mean ± SD) of three independent experiments are shown.
(G) CYLD regulation of Wnt signaling is conserved between tissues. HEK293T, HeLa, and BT549 cells were treated as in (C). Results are representative of at least
three experiments done in duplicate (mean ± SD).
Molecular Cell
CYLD Regulates Dvl Ubiquitination in Wnt Signaling
Molecular Cell 37, 607–619, March 12, 2010 ª2010 Elsevier Inc. 609
Page 3
Human Cylindroma Tumors Arising from CYLD
Mutations Display Hyperactive Wnt Signaling
Mutations in the CYLD gene are known to cause human familial
cylindromatosis (Bignell et al., 2000). Patients develop cylindro-
mas, spiradenomas, and trichoepitheliomas, benign skin
appendage tumors that arise from sweat gland and hair follicle
tissue, respectively. Regarding the involvement of Wnt signals
in skin appendage morphogenesis (Fuchs, 2007), we hypothe-
size that dysregulation of Wnt signals due to CYLD mutations
may be responsible for development of cylindromatosis. To
investigate whether Wnt signaling is active in these tumors, we
collected tumor samples of 15 cylindromatosis patients (Table
S1), 6 of which were previously determined to carry a mutation
in the CYLD gene (E868X) ( Bignell et al., 2000). All cylindroma
tumors showed accumulation of nuclear b-catenin, a hallmark
of hyperactive Wnt signaling (Figures 3A and 3B and Table S1).
In contrast, in adjacent normal epidermis, b-catenin predomi-
nantly localized to the plasma membrane, where it binds E-cad-
herin to serve a role in maintaining adherens junctions (Figure 3B)
(Ozawa et al., 1989). In addition, by in situ hybridization we
observed enhanced levels of mRNA expression of the Wnt target
gene, AXIN2, in distinct groups of cells in the majority of tumors
(Figure 3C and Table S1). These results reveal a previously
unknown link between cylindromatosis and hyperactive Wnt
signaling.
As the increase in Wnt signaling in CYLD knockdown cells is
ligand dependent (Figure 1C) in two out of three cell lines that
were analyzed, we presumed that Wnt signaling in CYLD-defi-
cient cylindroma tumors may be sustained by a local source of
Wnts. We observed expression of WNT3 mRNA (Figure 3D and
Table S1), but not WNT10B or WNT5A mRNA (data not shown),
within cylindroma tumor samples, in a patchy staining pattern
similar to AXIN2. Wnt3 is a known activator of Wnt/b-catenin
signaling. These results indicate that cylindroma tumors provide
their own supply of Wnt, thereby reinforcing a role of Wnt
pathway hyperactivation in cylindromatosis.
CYLD Regulates K63-Linked Ubiquitination of Dvl
How does CYLD regulate Wnt signaling at the molecular level?
We first positioned the effect of CYLD in the Wnt cascade by
epistasis experiments. TOPFlash reporter activity initiated by
enhanced expression of the upstream components, Fz5 or
Dvl1, was substantially increased by CYLD knockdown, whereas
signaling induced by ectopically enhanced expression of b-cat-
enin or by a dominant-negative variant of the b-catenin ubiquitin
ligase, b-TrCP (DF-TrCP) (Hart et al., 1999), remained unaffected
(Figure 4A). These results argue that CYLD acts upstream of
b-catenin activation.
Remarkably, TOPFlash reporter activity
induced by a constitutively active form of the Wnt receptor,
Lrp6, lacking the extracellular domain (DN-Lrp6), was insensitive
to CYLD knockdown. The mechanism by which DN-Lrp6 initiates
Wnt signaling remains unclear, but does not involve Dvl function
(Cong et al., 2004), suggesting that Dvl-dependent signaling
events in the Wnt cascade are regulated by CYLD.
These findings imply that the substrate of CYLD is located
upstream of b-catenin activation, rendering the receptors for
Wnt at the cell surface and their immediate downstream effec-
tors likely candidates for (de)ubiquitination. To test this, we
coexpressed tagged constructs of Fz5, Lrp6, and Dvl1 with
His-ubiquitin in control and CYLD knockdown cells, and precip-
itated ubiquitinated proteins under denaturing conditions, in
which noncovalent protein-protein interactions are disrupted.
A selective increase in ubiquitination of Dvl1, but not Fz5 or
Lrp6, was apparent by immunoblot analysis in CYLD knockdown
conditions (Figure 4B, lanes 3 and 4). Ubiquitination of Dvl1 was
diminished by overexpression of wild-type CYLD and enhanced
by overexpression of a catalytically inactive CYLD (CYLD C601S)
(Figure 4C, lanes 4–6, and Figure 4D). These results reveal that
deubiquitination of Dvl depends on the catalytic activity of
CYLD. Next, we analyzed ubiquitination levels of endogenous
DVL3, a Dvl isoform expressed in HEK293T cells, in control
and CYLD-depleted cells and after treatment with Wnt3a.
Precipitation of His-ubiquitin modified proteins unexpectedly
Luciferase activity
Luciferase activity
Luciferase activity
0.30
0.20
0.10
0
0
5
10
15
20
25
30
CYLD
con
con con con con
con con
con con
CYLD
CYLD CYLD
FOPflash
TOPflash
stimulus:
shRNA:
shRNA:
Wnt3a Wnt3a TNFαTNFα
stimulus:
shRNA:
TNFα Wnt3aWnt3a
0.25
0.15
0.05
0.30
0.20
0.10
0
CYLD
stimulus 1
Wnt3a
Wnt3aWnt3a
stimulus 2 - - - TNFαTNFα
0.25
0.15
0.05
NFκB-luc
FOPflash
TOPflash
A
C
B
conconcon
conconcon
Figure 2. The Regulation of Wnt Signaling by CYLD Is Independent of Its Role in NF-kB Activation
(A) TNFa does not induce TOPFlash reporter activity. HEK293T cells were transfected with shCYLD, control shGFP, TOPFlash or FOPFlash, and TK-Renilla, as
indicated. Cells were subsequently stimulated with either Wnt3a or TNFa for 16 hr.
(B) Wnt3a does not induce NF-kB reporter activity. HEK293T cells were transfected with shCYLD or control shGFP, NF-kB luciferase reporter construct
(63kB-luc), and TK-Renilla. Cells were subsequently stimulated with TNFa or Wnt3a for 16 hr.
(C) Combined TNFa and Wnt3a stimulation does not further enhance Wnt3a-induced TOPFlash reporter activity. HEK293T cells were transfected as in (A) and
stimulated with either Wnt3a or TNFa, or a combination of both, for 16 hr. Black bars represent FOPFlash, and gray bars represent TOPFlash luciferase activity.
Molecular Cell
CYLD Regulates Dvl Ubiquitination in Wnt Signaling
610 Molecular Cell 37, 607–619, March 12, 2010 ª2010 Elsevier Inc.
Page 4
revealed the presence of a prominent, unmodified DVL3 protein
band in all samples. Sequence alignment of Dvl family members
showed the presence of a His-rich region at the C terminus of
the DVL3 isoform that may explain intrinsic binding propensity
of DVL3 for Ni-NTA beads (Figure 4E). We used the unmodified
DVL3 band to demonstrate that overall levels of DVL3 do not
change significantly upon modulation of CYLD. In contrast, we
observed a striking increase in high-molecular weight (MW)
ubiquitin DVL3 conjugates in CYLD knockdown (Figure 4E, lanes
2, 4, and 6) as compared with control knockdown (Figure 4E,
lanes 1, 3, and 5) cells. The abundance of the slowest migrating
forms of ubiquitinated DVL3 increased further upon treatment
with Wnt3a in CYLD-depleted cells (Figure 4E), indicating that
combined CYLD knockdown and Wnt treatment favors DVL3
ubiquitination. Moreover, both endogenous as well as overex-
pressed CYLD coprecipitated with endogenous DVL3 (Figure 4F
and Figure S2), but not with Fz5 or Lrp6 (data not shown), indi-
cating that CYLD and Dvl reside in similar protein complexes.
Next, we set out to determine the type of polyubiquitin linkage
with which Dvl is modified in CYLD-deficient conditions by
immunoprecipitation of flag-Dvl1 from cells that coexpress
K48R or K63R mutant ubiquitin and either wild-type or the domi-
nant-negative CYLD C601S variant (Figure 5A, lanes 3–8).
Expression of K48R ubiquitin still allowed for 70% of relative
levels of Dvl1 ubiquitination, whereas K63R ubiquitin usage
decreased ubiquitination to 30% of wild-type conditions (Fig-
ure 5B), suggesting preferential use of K63-linked ubiquitin
chains. To obtain conclusive evidence on the type of ubiquitin
linkage conjugated to Dvl1, we subjected purified ubiquitinated
Dvl1 from cells expressing CYLD C/S to MS/MS analysis. Of
note, two separate regions of ubiquitinated Dvl1 with lower and
higher MW could be distinguished on SDS-PAGE gel (Figure 5C).
As revealed by MS/MS analysis, the most abundant lower MW
fraction of ubiquitinated Dvl was selectively modified with K63-
linked ubiquitin chains (Figure 5E), whereas, in the higher MW
fraction, the majority of ubiquitin linkages were formed through
K48 (Figure 5D). These results demonstrate that, besides the
earlier reported modification of Dvl with K48-linked ubiquitin
that links to Dvl turnover (Angers et al., 2006), a major fraction
of Dvl1 is coupled to K63-linked ubiquitin chains in CYLD
C/S-expressing cells. These findings are in accordance with
the previously reported specificity of CYLD for disassembly of
HEIHC: β-CAT
IHC: β-CAT
ISH: AXIN2
ISH: WNT3
EA
C
D
B
* * *
*
Figure 3. CYLD-Mutant Cylindroma Tumors Display Hyperactive Wnt Signaling Activity
(A) b-Catenin (b-CAT) staining of CYLD-mutant cylindroma tumor and adjacent unaffected skin by immunohistochemistry (IHC). Magnification, 203; asterisk
indicates the epidermis.
(B) Tumor area stained for b-catenin (magnification, 403). Examples of tumor cells displaying nuclear accumulation of b-catenin are indicated by arrows.
E-cadherin-bound b-catenin in adherens junctions is shown in normal epidermis (asterisk). Results shown in (A) and (B) are representative of cylindroma tumors
of 15 patients.
(C) In situ hybridization (ISH) of AXIN2 mRNA in CYLD-mutant cylindroma tumor. Patches of AXIN2-positive cells are found throughout the tumor (arrows),
whereas adjacent unaffected epidermal skin (asterisk) is negative. Tumors of nine patients were analyzed for AXIN2. Results are representative of eight tumors;
one tumor was negative. Magnification, 203.
(D) In situ hybridization of WNT3 mRNA in cylindroma tumor. Similar patches of WNT3-expressing cells are found throughout the tumor (arrows). Tumors of six
patients were analyzed for WNT3. Results are representative of five tumors; one was negative. Magnification, 403.
(E) Hematoxylin-eosin (HE) stain of cylindromatosis tumor tissue, showing cylindroma tumor islets underneath the dermis and epidermis (asterisk). Magnification, 203.
Molecular Cell
CYLD Regulates Dvl Ubiquitination in Wnt Signaling
Molecular Cell 37, 607–619, March 12, 2010 ª2010 Elsevier Inc. 611
Page 5
A
E
CD
B
F
250
250
150
150
100
75
con shCYLD
Relative fold activation
Relative Dvl1
ubiquitination
shRNA
V5-Fz5
flag-Dvl1
250
150
100
75
flag-Dvl1
Input
Pulldown
myc-LRP6
con
CYLD
con
CYLD
250
150
100
75
50
1 2 3 4
6 54 3 2 1
2
1
3
4
0
Fz5 Dvl1
N-Lrp6F-βTrCP β-cat
immature
mature
Dvl1
Dvl1-Ub
n
dimer
Input (2%)
Pulldown (20%)
con
shCYLD
Wnt3a-CM
min
+
++
+
++
- - -
- - -
0 30 60
7 8 9 10 11 12
1 2 3 4 5 6
actin
DVL3
DVL3
250
150
100
100
Dvl1
Dvl1-Ub
DVL3
DVL3-Ub
n
Input
Pulldown
- -
-
-
- -
-
-
+
++
+
-
-
-
-
++
CYLD wt
Mock
CYLD /
C
S
CYLD /
CYLD wt
Mock
C
S
DVL3
CYLD
IP:
IgG
DVL3
Input
0
0.4
0.8
1.2
1.6
Figure 4. Loss of CYLD Enhances Ubiquitination of the Upstream Core Wnt Signaling Component Dvl
(A) Positioning the effect of CYLD in the Wnt pathway by epistatic analysis. Wnt signaling was induced by transfe ction of HEK293T cells with Fz5, Dvl1, DN-Lrp6,
DF-b-TrCP (DF-box domain), or b-catenin. TOPFlash reporter and shCYLD or control knockdown constructs were cotransfected as indicated. Relative TOPFlash
over FOPFlash luciferase ratios SD) of duplicate experiments normalized against control knockdown values are shown.
(B) Analysis of ubiquitination of Fz5, Lrp6, and Dvl1 upon loss of CYLD in vivo. HEK293T cells were cotransfected with V5-tagged Fz5 (top), myc-tagged Lrp6
(middle), or flag-tagged Dvl1 (bottom), together with control plasmid or His-tagged ubiquitin and shCYLD or control shRNA constructs. Ubiquitinated proteins
were precipitated under denaturing conditions and immunoblotted for V5 (Fz5), myc (Lrp6), or flag (Dvl1), respectively. Input and His pull-down (His-PD) results
are shown as indicated.
(C) Analysis of Dvl1 ubiquitination in cells transfected with wild-type (wt), catalytically inactive CYLD C601S (C/S), or control plasmid. Results represent at least
two independent experiments.
(D) Quantification of the results in (C). Bars represent the rati o of ubiquitinated Dvl over total Dvl levels.
(E) Ubiquitination of endogenous DVL3 is induced by loss of CYLD. HEK293T cells were transfected with His-tagged ubiquitin and shCYLD or control shRNA
constructs. After treatment with Wnt3a-CM for 0, 30, or 60 min, ubiquitinated proteins were precipitated under denaturing conditions and immunoblotted for
DVL3 (arrow). Input (lanes 1–6) and His pull-down (lanes 7–12) results are shown as indicated.
(F) CYLD interacts with Dvl in vivo in HEK293T cells. Endogenous DVL3 was immunoprecipitated, and samples were analyzed by immunoblot detection for DVL3
(bottom) and coprecipitated CYLD (top).
Molecular Cell
CYLD Regulates Dvl Ubiquitination in Wnt Signaling
612 Molecular Cell 37, 607–619, March 12, 2010 ª2010 Elsevier Inc.
Page 6
input IP: Flag
flag-Dvl1
Relative Dvl1
ubiquitination
HA-Ub
Ub wt Ub K48R Ub K63R
100
150
100
75
250
150
-
-
+
+
-
-
+
+
-
-
+
+
-
-
+
+
wildtype K48R K63R
HA-Ub
1 2 3 4 5 6 7 8
CYLD wt
CYLD /
C
S
Dvl
Dvl-Ub
n
0
0.2
0.4
0.6
0.8
1.0
BA
C
250
150
100
CYLD C/S
MOCK
75
D
E
387.1
500.3
637.4
750.5
851.6
938.6
1016.2
1067.7
1309.8
1437.9
1550.9
1665.1
1744.8
1858.0
1958.0
0
250
500
750
1000
1250
Intensity [a.u.]
200 400 600 800 1000 1200 1400 1600 1800 2000
m/z
T L S D Y N I Q K E S T L H L V L R
y3
y3
y4
y4
y5
y5
y6
y6
y7
y7
y8
y8
y9
y9
G
G
y10
y11y12
y14++
914.5
y14
y11
y12
y13
y13y16
y16++
b14
b15
b16
b14 b15 b16
*y10
*
K63
232.0
227.2
347.0
374.1
445.3
476.1
589.3
618.1
717.3
744.4
872.5
959.5
1016.6
1087.6
1114.6
1229.7
0
1000
2000
3000
4000
5000
6000
200 300 400 500 600 700 800 900 1000 1100 1200
Intensity [a.u.]
m/z
L I F A G K Q L E D G R
y2
y2
y3
y3
y4
y4
y5
y5
y6
y6
y7
y8
G
G
*
*
*
*y7
y10
y8
y9
y9
1234.7
y10
b2
b2
b3
b3
b4
b4
b6
b6
b7
b7
985.5
b8
y10++
b8
b9
b9
b10
b10
K48
II
I
Figure 5. Loss of CYLD Enhances K63-Linked Ubiquitination of Dvl1
(A) HA-tagged wild-type, K48R, or K63R mutant ubiquitin were cotransfected with wild-type (wt) CYLD or CYLD C/S constructs and flag-tagged Dvl1 in HEK293T
cells. Immunoprecipitation of flag-Dvl1 was performed on denatured lysates, and the ubiquitinated fraction of Dvl1 was analyzed by immunoblot detection of
HA-ubiquitin (top). Immunoprecipitated levels of flag-tagged Dvl1 are shown (bottom). Results represent at least three independent experiments.
(B) Quantification of the results in (A). Bars represent the ratio of ubiquitinated Dvl over total Dvl levels in CYLD C/S-expressing cells.
(C) HEK293T cells, transfected with Flag-Dvl1, His-ubiquitin, and either empty vector or CYLD C601S were lysed under denaturing conditions and subjected to
successive His and flag pull-down. Samples of purified ubiquitinated Dvl1 were resolved on SDS-PAGE and stained with Coomassie G250. Indicated regions
(brackets) of the gel were cut and analyzed separately.
(D and E) Tandem mass spectrometry (LC-MS/MS) analysis. The MS/MS spectra show the b and y fragment ions that were identified. (D) Ubiquitinated Dvl material
isolated from region II revealed the ubiquitin-derived tryptic peptide fragment 43–54 ([M+2H]
2+
, 732.4 Da; MW, 1459.8 Da; diamond) containing gly-gly-modified K48.
(E) From ubiquitinated Dvl in region I, we identified the trypticpeptide fragment55–72([M+2H]
2+
, 1123.9 Da; MW, 2243.2 Da; diamond) containing gly-glymodified K63.
Molecular Cell
CYLD Regulates Dvl Ubiquitination in Wnt Signaling
Molecular Cell 37, 607–619, March 12, 2010 ª2010 Elsevier Inc. 613
Page 7
DIX
FL
DIX-5KR
DIX-6KR
DIX-7KR
1-502
1-394
1-345
1-250
1-217
1
86
217
250
345
394
502
695
A
B
C
D
E
F
G
M1
M2
PDZ DEPDIX
250
150
100
75
FL
DIX
M1
M2
M4
FL
DIX
M1
M2
M4
input His-Pulldown
0
500
1000
1500
2000
Luciferase activity
Intensity [au]
Luciferase activity
1 2 3 4 5
1 2 3 4 5
6 7 8 9 10
PDZ-3KR
250
150
100
WT
DIX-5KR
DIX-6KR
DIX-7KR
PDZ-3KR
WT
DIX-5KR
DIX-6KR
DIX-7KR
PDZ-3KR
His-Pulldown
100
75
input
0
1000
2000
3000
250
150
100
75
50
37
FL
DIX
1-502
1-394
1-345
1-250
WT
M1
M2
M4
DIX
1-217
FL
DIX
1-502
1-394
1-345
1-250
1-217
1-502
1-394
1-345
1-250
1-217
NT
His-Pulldowninput
flag-Dvl1
flag-Dvl1
flag-Dvl1
1 2 3 4
DIX PDZ DEP
5 6 7 8 9 10 11 12 13 14
332.0
y3
447.1
y4
562.1
y5
709.2
y6
732.7
y13
781.3
b7
805.9
y14
880.5
b8
919.3
y16
1080.6
y9
1208.7
y10
1307.8
y11
1463.9
y13
1611.0
y14
1740.9
y27
0
200
400
600
800
1000
1600
N V L S N R P V H A Y K F F F K S M D Q D F G V V K E E I F D D N A K
822.4
y7
951.5
y8
y3y4y5y6y7y8y14 y13 y9y10y11
y27
b8
326.7
b3
b7
b3
G
Ox
G
G
G
1400
1200
1000800
600
400
m/z
b16
b16*++
K46 K50
Molecular Cell
CYLD Regulates Dvl Ubiquitination in Wnt Signaling
614 Molecular Cell 37, 607–619, March 12, 2010 ª2010 Elsevier Inc.
Page 8
K63-linked ubiquitin chains (Brummelkamp et al., 2003;
Komander et al., 2008; Kovalenko et al., 2003; Trompouki et al.,
2003). In addition, general reduction of K63-linked ubiquitin
chain formation in cells by shRNA-mediated knockdown of the
E2 enzyme, Ubc13, significantly impaired Wnt-induced reporter
activity (Figure S3), corroborating a positive role of K63-linked
chain formation in transmission of the Wnt signal.
We conclude that the core Wnt signaling component Dvl is a
substrate for K63-linked ubiquitination. CYLD interacts with Dvl
and edits its ubiquitin chains, supporting a model in which the
effect of CYLD on Wnt signaling is mediated by CYLD-depen-
dent deubiquitination of Dvl.
The DIX Domain of Dvl Is Hyperubiquitinated
in CYLD-Deficient Cells
How does Dvl function depend on ubiquitination? Dvl contains
three functionally important conserved domains: an N-terminal
DIX domain, a central PDZ, and a C-terminal DEP domain (Wall-
ingford and Habas, 2005). Remarkably, conserved lysine
residues that serve as potential acceptor sites for ubiquitin
attachment are mostly confined to these functional domains in
Dvl1 (Figure 6A, filled circles).
To identify target sites for ubiquitination in Dvl1, we applied
MS/MS and database searching on both K63- and K48-linked
pools of purified, ubiquitinated Dvl (figure 5C). We identified
DIX domain lysines 46 and 50 with high confidence (Figure 6B),
and lysines 20 and 34 with lower confidence (data not shown),
as principal sites of K63-linked ubiquitination in CYLD-deficient
cells. In addition, PDZ domain lysine 285 was identified as a
target of K48-linked ubiquitination (data not shown).
To confirm these findings and gain insight into how DIX
domain ubiquitin (de)conjugation connects to Dvl function, we
generated truncation and functional Dvl1 mutants (Figure 6A),
and analyzed their signaling ability and levels of ubiquitination
in CYLD-deficient cells. Deletion of the DIX domain, involved in
Dvl polymerization and formation of dynamic cytoplasmic
protein assemblies (Schwarz-Romond et al., 2007a), strongly
diminished ubiquitination (Figure 6D, lane 9, and Figure 6E,
lane 10). This construct was unable to induce TOPFlash reporter
activity (Figure 6C). In addition, three previously reported point
mutants that interfere with DIX domain function, M1 (F33A), M2
(V59A/K60A), and M4 (Y17D) (Schwarz-Romond et al., 2007a),
showed a strongly correlated, stepwise decrease in Dvl1 poly-
merization properties, signaling ability, and ubiquitination
(Figures 6C and 6E, lanes 7–9, and Figure S4). To interfere with
DIX domain ubiquitination, we substituted DIX domain lysines
(K5, 20, 34, 46, 50, 60, and 69) by arginines. In accordance
with MS/MS results that indicate ubiquitination of multiple DIX
domain lysines, single lysine mutants did not significantly alter
overall Dvl ubiquitination levels (data not shown). We therefore
replaced all DIX domain lysines by arginines (DIX-7KR). All Dvl1
DIX-KR mutants (DIX-5KR to -7KR) formed polymeric protein
assemblies in cells, and bound the Fz receptor at the plasma
membrane equally well as wild-type Dvl1, showing that the
introduction of multiple lysine-to-arginine mutations did not
generally perturb DIX-domain folding and function (Figure S4).
Noticeably, levels of K63-linked ubiquitination, as well as signal-
ing ability, were progressively decreased for Dvl DIX-5KR, -6KR,
and -7KR mutants (Figure 6F, lanes 2–4, and Figure 6G). We
conclude that K63-linked ubiquitination of the Dvl DIX domain
is tightly and positively linked to Dvl signaling activity. Moreover,
interference with Dvl polymerizing propensity, without perturbing
DIX domain folding (mutant M4) (Schwarz-Romond et al.,
2007a), strongly reduces ubiquitination (Figures 6C and 6E,
lane 9). Conversely, impairing DIX domain ubiquitination leaves
Dvl polymerization unaffected, yet hampers signaling ability
(Figure 6F and Figure S4B). Together, these data strongly argue
that the polymeric form of Dvl is ubiquitinated in CYLD-deficient
cells.
Lysine 285 of the Dvl1 PDZ domain, implicated in Fz binding,
was identified as a substrate for K48-linked ubiquitination. To
address if and how ubiquitination of the PDZ domain links to
Dvl function, we replaced the three conserved lysines in this
domain (K225, K285, and K337) by arginines (PDZ-3KR). PDZ-
3KR demonstrated neither altered signaling activity nor changes
in ubiquitination levels in cells with impaired CYLD activity
(Figure 6F, lane 5, and Figure 6G). We therefore conclude that
ubiquitination of the PDZ domain is not instrumental to CYLD-
mediated regulation
of Dvl activity.
Finally, we tested whether the DEP domain, implicated in
membrane targeting of Dvl, is linked to Dvl ubiquitination (Roth-
bacher et al., 2000; Simons et al., 2009 ). Truncation of the C
terminus up to the DEP domain showed only a moderate effect
on ubiquitination and Dvl-mediated TOPFlash activity (Figures
6C and 6D, lane 10). Further truncation to remove the DEP
domain entirely abolished TOPFlash activity as well as Dvl ubiq-
uitination (Schwarz-Romond et al., 2007b)(Figures 6C and 6D,
lanes 11–14), whereas Dvl polymerization properties of this
Figure 6. DEP Domain-Dependent Hyperubiquitination of the N-Terminal Dvl DIX Domain in CYLD-Deficient Cells
(A) Schematic overview of the relative location of the conserved domains and lysines in Dvl1. Lysines conserved between Drosophila, human, and mouse
(filled circles), conserved between human and mouse Dvl1, -2, and -3 (half-open circles), and those specific to the Dvl1 isoform (open circles) are indicated.
Truncation and deletion mutants are indicated. All constructs bear an N-terminal flag tag.
(B) Identification of ubiquitination sites in the Dvl DIX domain. LC-MS/MS analysis of ubiquitinated Dvl identified tryptic peptide fragment 35–69 ([M+4H]
4+
,
1095.1 Da; MW, 4378.1 Da). Lysine residues K46 and K50 were both modified with a gly-gly tag (+114.1 Da). The MS/MS spectrum shows the identified
b and y fragment ions that were detected.
(C) TOPFlash reporter activity induced by wild-type and Dvl1 deletion constructs. HEK293T cells were cotransfected wit h indicated Dvl1 constructs and
TOPFlash or FOPFlash reporter. Relative TOPFlash over FOPFlash luciferase ratios SD) of duplicate experiments are shown. Results represent at least
four independent experiments.
(D–F) Ubiquitination analysis of flag-Dvl1 wild-type and mutant proteins. HEK293T cells were transfected with His-tagged ubiquitin, HA-CYLD C/S, and Dvl1
variants, as indicated. His-ubiqui tin-conjugated Dvl1 was precipitated and analyzed by Western blot using anti-flag antibodies.
(G) TOPFlash reporter activity induced by wild-type and lysine mutant Dvl1 constructs. Relative TOPFlash over FOPFlash luciferase ratios SD) of triplicate
experiments are shown.
Molecular Cell
CYLD Regulates Dvl Ubiquitination in Wnt Signaling
Molecular Cell 37, 607–619, March 12, 2010 ª2010 Elsevier Inc. 615
Page 9
mutant remained intact (data not shown). We conclude that ubiq-
uitination of the Dvl DIX domain in CYLD-depleted cells depends
on an intact DEP domain. We propose that DEP domain-medi-
ated membrane targeting and/or protein binding is required for
the modification of polymerized Dvl with K63-linked ubiquitin
chains.
DISCUSSION
Wnt/b-catenin signaling plays key roles in metazoan develop-
ment, adult tissue homeostasis, and human malignancies. To
gain a deeper understanding of the molecular mechanisms by
which Wnt signals are transmitted, we examined the regulatory
role of DUBs in the Wnt/b-catenin pathway. We identify the
DUB and tumor suppressor, CYLD, as a negative regulator of
proximal events in Wnt/b-catenin signaling. Knockdown of
CYLD in cultured cells strongly enhances Wnt-induced b-catenin
stabilization and target gene transcription. Moreover, human
skin appendage tumors that arise from mutations in CYLD
display hyperactive Wnt signaling. Based on the established
role of Wnt signaling in self-renewal of the skin, associated cell
lineage decisions, and skin appendage formation (Fuchs,
2007), we propose that increased or prolonged activation of
Wnt signaling in CYLD-mutant cells may drive uncontrolled
proliferation of early sweat gland and hair follicle cell lineages
to form cylindromas and trichoepitheliomas, respectively.
At the molecular level, CYLD reverses the ubiquitination of the
upstream Wnt signaling component, Dvl. Collectively, our work
implicates a positive role of K63-linked ubiquitination at multiple
lysines of the Dvl N-terminal DIX domain in the transmission of
the Wnt signal, thereby uncovering a layer of regulation in the
Wnt pathway. Ubiquitination of the Dvl DIX domain in CYLD-
deficient cells is tightly and positively linked to Dvl polymerization
and signaling ability.
Remarkably, a functional C-terminal DEP domain, implicated
in membrane targeting of Dvl in both b-catenin-dependent
and -independent Wnt pathways (Boutros et al., 1998; Li et al.,
1999; Pan et al., 2004; Rothbacher et al., 2000; Simons et al.,
2009; Yanagawa et al., 1995), is also essential for ubiquitination
to occur. Possibly, the DEP domain provides a binding site for
the yet unidentified E3 ubiquitin ligase. The inferred E3 ligase
that ubiquitinates Dvl in CYLD-deficient cells is likely different
from the previously reported KLHL12 enzyme that targets Dvl
for K48-linked ubiquitin-mediated degradation, and binds to
the region C terminal to the DEP domain of Dvl (Angers et al.,
2006). This region is dispensable for ubiquitination of Dvl in the
absence of CYLD activity (Figure 6D, lane 10). Another ubiquitin
ligase, Smurf2, was recently reported to bind to the DEP domain
of phosphorylated Dvl to target Prickle1, an essential protein
involved in the b-catenin-independent Wnt/planar cell polarity
pathway, for ubiquitin-mediated degradation (Narimatsu et al.,
2009). Smurf E3 ligases generally target their substrates for
K48-linked ubiquitination and subsequent proteasomal degra-
dation, rendering them unlikely candidates for the K63-linked
modification of Dvl as described in this work.
Thus far, the precise mechanism by which CYLD edits ubiqui-
tination of selected substrates remains unclear. Recent in vitro
studies demonstrate that CYLD efficiently disassembles unan-
chored K63-linked ubiquitin chains, but not ubiquitin chains
conjugated to two of its established substrates, TRAF6 and
NEMO (Xia et al., 2009). Likewise, in vitro deubiquitination of
Dvl1 with purified CYLD was not successful in our hands (data
not shown). Possibly, CYLD requires unknown cofactors to
regulate substrate ubiquitination in vivo. Alternatively, CYLD
may interfere with the activity of selected E2/E3 ubiquitin ligases
that transfer K63-linked ubiquitin chains onto specific sub-
strates. Identification of the responsible E2/E3 ubiquitin ligase
involved in K63-linked ubiquitination of Dvl will be of help in
solving these issues.
To this end, we show that Ubc13, a ubiquitin-conjugating E2
enzyme broadly involved in the generation of K63-linked ubiqui-
tin chains (Deng et al., 2000; Hofmann and Pickart, 1999), is
required for optimal Wnt/b-catenin signaling. However, depen-
dence of Wnt/b-catenin signaling on Ubc13 may differ between
cell types, as shown by a recent study in which a remaining
population of surviving Ubc13-deficient hematopoietic cells dis-
played hyperactive Wnt signaling (Wu et al., 2009). Of note, in
addition to its positive regulation of Dvl activity, K63-linked
ubiquitination was shown to exert negative effects on down-
stream signaling events in the Wnt pathway (
Tran et al., 2008).
How
Ubc13
may regulate these opposite roles of K63-linked
ubiquitin in Wnt signaling in different tissues awaits further
elucidation.
On the basis of our findings, we propose a model in which the
polymeric form of Dvl is a target for DEP domain-dependent
K63-linked ubiquitination (Figure 7). Conceivably, enhanced or
prolonged K63-linked ubiquitination of the Dvl DIX domain in the
absence of a functional CYLD molecule may contribute to the
amplification of the Wnt signal by selective recruitment of effector
molecules, by controlling Dvl-mediated regulation of receptor
endocytosis (Yu et al., 2007), or by enhancing the formation of
Dvl-based polymerized signaling complexes ( Schwarz-Romond
et al., 2007a). In line with the latter possibility, polyubiquitination
was recently demonstrated to directly promote molecular aggre-
gation and activation of receptor-bound caspase-8 in apoptosis
signaling (Jin et al., 2009 ).
Our findings may have implications for the development and
progression of other human disorders, such as multiple myeloma
and colon and hepatocellular carcinomas linked to loss of CYLD
(Hashimoto et al., 2004; Hellerbrand et al., 2007; Jenner et al.,
2007), and depend on dysregulated Wnt signaling (Clevers,
2006; Derksen et al., 2004; Lee et al., 2006 ). The vital question
of how CYLD interferes with cell lineage specification and growth
in vivo in the skin remains unanswered thus far because of the
current lack of animal models that phenocopy CYLD-mediated
human disease. Clearly, improved animal models are necessary
to address the vital question of how CYLD controls two major
signaling pathways, NF-kB and Wnt/b-catenin, in development
and cancer.
EXPERIMENTAL PROCEDURES
Please see the Supplemental Experimental Procedures for the following
sections: ‘Cell Culture and Transfection’’; ‘Constructs and Antibodies’’;
‘Reporter Assays’’; ‘RNA Preparation and Northern Blot Analysis’’; ‘Immuno-
histochemistry and In Situ Hybridization’’; and ‘Immunoprecipitations.’
Molecular Cell
CYLD Regulates Dvl Ubiquitination in Wnt Signaling
616 Molecular Cell 37, 607–619, March 12, 2010 ª2010 Elsevier Inc.
Page 10
b-Catenin Stabilization Assay
HEK293T cells were seed ed in six well plates, transfected with pSUPER
shRNA vectors, as indicated, for 48 hr prior to 5.5 hr treatment with Wnt3a-
CM. Cells were washed and harvested with ice-cold PBS and lysed in 0.5 ml
of hypotonic lysis buffer containing 10 mM Tris-HCl (pH 7.4), 2 mM EDTA,
10 mM KCl, and protease inhibitors (2 mg/ml aprotinin, 10 mg/ml leupeptin,
1 mg/ml pepstatin, and 1 mM PMSF) by resuspension on ice. The cytoplasmic
fraction was isolated by centrifugation at 16,000 3 g for 30 min at 4
!
C. The
levels of b -catenin versus actin were analyzed and quantified by Odyssey
digital immunoblot detection (LiCor Biosciences).
Ubiquitination Assays
HEK293T cells were transfected with His-ubiquitin, shRNA pSUPER vectors,
as indicated, and either empty vector or flag-Dvl1. Cells were washed with
and harvested in ice-cold PBS with 0.92 mM CaCl
2
, 0.50 mM MgCl
2
, and
10 mM N-ethylmaleimide (NEM; Sigma), and a 10% sample was lysed in
sample buffer. The remainder was lysed in a chaotropic lysis buffer containing
6 M guanidinium-HCl (Fluka), 100 mM P-buffer, 10 mM Tris-HCl (pH 8.0),
20 mM imidazole, and 10 mM b-mercaptoethanol. After sonification, lysates
were cleared of their residual 16,000 3 g centrifugation pellets and incubated
with Ni
2+
-NTA-agarose beads (Qiagen) for 3 hr in a head-over-head tumbler at
room temperature. Beads were washed with lysis buffer without imidazole and
inhibitors, but with 0.2% Triton X-100, then with 8 M urea washing buffers A, B,
and C (A, 8 M urea, 100 mM P-buffer, 10 mM Tris-HCl [pH 8.0], 10 mM b-ME,
0.2% Triton X-100; B, as buffer A, but pH 6.3; C, as B, but with 0.1% Triton
X-100). His-tagged proteins were eluted with two bead volumes of elution
buffer (6.2 M urea, 10 mM Tris-HCl [pH 7.0], 100 mM P buffer, and 200 mM
imidazole), and two volumes of 33 sample buffer.
Purification and MS Analysis of Ubiquitinated Dvl
HEK293T cells were grown in five 15 cm dishes and transfected with flag-Dvl1,
His-ubiquitin, and either HA-CYLD C601S (CYLD-C/S) or empty vector. At 48 hr
after transfection, cells were washed, harvested, lysed, and incubated with
Ni-NTA beads, as described above for the ubiquitination assays. The beads
were washed subsequently with guadinium-HCl washing buffer (3 M Gua-HCl,
T c f
wildtype
kdCYLD
β
β
β
β
β
β
β
β
β
β
β
β
β
β
β
T c f
GSK3
CK1
APC
Axin
GSK3
CK1
APC
Axin
K63-linked
polyubiquitin
β-catenin
β
Lrp5/6
Fz
CYLD
Wnt
Wnt
Wnt
Wnt
Figure 7. Model of CYLD-Mediated Regula-
tion of Wnt/ b-Catenin Signaling
(Left) In wild-typecells, Wnt binding of the receptors,
Fz and Lrp6, leads to recruitment of Dvl and induces
clustering of the receptor-Dvl complex. CYLD medi-
ates disassembly of K63-linked ubiquitin chains on
the DIX domain of polymerized Dvl, thereby sup-
pressing downstream activation of b-catenin.
(Right) In CYLD-deficient cells, K63-linked ubiquiti-
nation of polymeric Dvl facilitates transmission of
the Wnt signal, leading to enhanced levels of tran-
scriptionally active b-catenin.
100 mM P buffer [pH 8.0], 10 mM Tris [pH 8.0],
0.2% Triton X-100, 10 mM b-ME), with Tris washing
buffer (1% Triton X-100, 150 mM NaCl, 25 mM
Tris-HCl [pH 7.4], 0.2% SDS) with 1 M Gua-HCl,
and with Tris washing buffer without Gua-HCl.
Ubiquitinated proteins were eluted with hot PBS
lysis buffer (see Supplemental Information) and
treated as described above. After immunoprecipi-
tation, the samples were washed in Tris washing
buffer and eluted by boiling in SDS-sample buffer.
For details about sample preparation and MS
analysis, see the Supplemental Information.
SUPPLEMENTAL INFORMATION
Supplemental Information includes Supplemental
Experimental Procedures, four figures, and one
table and can be found with this article online at doi:10.1016/j.molcel.2010.
01.035.
ACKNOWLEDGMENTS
This work was supported by the Dutch Cancer Society (UU 2006-3508), the
European Research Council (ERC-StG no. 242958 to M.M.M.), the University
of Utrecht (High Potential grant), and the RUBICON EU network of excellence.
We thank H.C. Korswagen and C. Rabouille for critical reading of the manu-
script and discussions, and gratefully acknowledge J.J. van der Smagt for
selecting genotyped human cylindroma materials and H. Begthel for help
with immunohistochemistry.
Received: July 24, 2009
Revised: November 19, 2009
Accepted: January 14, 2010
Published: March 11, 2010
REFERENCES
Angers, S., Thorpe, C.J., Biechele, T.L., Goldenberg, S.J., Zheng, N.,
MacCoss, M.J., and Moon, R.T. (2006). The KLHL12-Cullin-3 ubiquitin ligase
negatively regulates the Wnt-beta-catenin pathway by targeting Dishevelled
for degradation. Nat. Cell Biol. 8, 348–357.
Axelrod, J.D., Miller, J.R., Shulman, J.M., Moon, R.T., and Perrimon, N. (1998).
Differential recruitment of Dishevelled provides signaling specificity in the
planar cell polarity and Wingless signaling pathway s. Genes Dev. 12, 2610–
2622.
Behrens, J., Jerchow, B.A., Wurtele, M., Grimm, J., Asbrand, C., Wirtz, R.,
Kuhl, M., Wedlich, D., and Birchmeier, W. (1998). Functional interaction of an
Axin homolog, conductin, with beta-catenin, APC, and GSK3beta. Science
280, 596–599.
Bignell, G.R., Warren, W., Seal, S., Takahashi, M., Rapley, E., Barfoot, R.,
Green, H., Brown, C., Biggs, P.J., Lakhani, S.R., et al. (2000). Identification
Molecular Cell
CYLD Regulates Dvl Ubiquitination in Wnt Signaling
Molecular Cell 37, 607–619, March 12, 2010 ª2010 Elsevier Inc. 617
Page 11
of the familial cylindromatosis tumour-suppressor gene. Nat. Genet. 25, 160–
165.
Bilic, J., Huang, Y.L., Davidson, G., Zimmermann, T., Cruciat, C.M., Bienz, M.,
and Niehrs, C. (2007). Wnt induces LRP6 signalosomes and promotes Dishev-
elled-dependent LRP6 phosphorylation. Science 316, 1619–1622.
Boutros, M., Paricio, N., Strutt, D.I., and Mlodzik, M. (1998). Dishevelled
activates JNK and discriminates between JNK pathways in planar polarity
and wingless signaling. Cell 94, 109–118.
Brummelkamp, T.R., Nijman, S.M., Dirac, A.M., and Bernards, R. (2003). Loss
of the cylindromatosis tumour suppressor inhibits apoptosis by activating
NF-kappaB. Nature 424, 797–801.
Clevers, H. (2006). Wnt/beta-catenin signaling in development and disease.
Cell 127, 469–480.
Cliffe, A., Hamada, F., and Bienz, M. (2003). A role of Dishevelled in relocating
Axin to the plasma membrane during wingless signaling. Curr. Biol. 13,
960–966.
Cong, F., Schweizer, L., and Varmus, H. (2004). Wnt signals across the plasma
membrane to activate the beta-catenin pathway by forming oligomers contain-
ing its receptors, Frizzled and LRP. Development 131, 5103–5115.
Davidson, G., Wu, W., Shen, J., Bilic, J., Fenger, U., Stannek, P., Glinka, A.,
and Niehrs, C. (2005). Casein kinase 1 gamma couples Wnt receptor activation
to cytoplasmic signal transduction. Nature 438, 867–872.
Deng, L., Wang, C., Spencer, E., Yang, L., Braun, A., You, J., Slaughter, C.,
Pickart, C., and Chen, Z.J. (2000). Activation of the IkappaB kinase complex
by TRAF6 requires a dimeric ubiquitin-conjugating enzyme complex and
a unique polyubiquitin chain. Cell 103, 351–361.
Derksen, P.W., Tjin, E., Meijer, H.P., Klok, M.D., MacGillavry, H.D., van Oers,
M.H., Lokhorst, H.M., Bloem, A.C., Clevers, H., Nusse, R., et al. (2004).
Illegitimate WNT signaling promotes proliferation of multiple myeloma cells.
Proc. Natl. Acad. Sci. USA 101, 6122–6127.
Fuchs, E. (2007). Scratching the surface of skin development. Nature 445,
834–842.
Hart, M., Concordet, J.P., Lassot, I., Albert, I., del los Santos, R., Durand, H.,
Perret, C., Rubinfeld, B., Margottin, F., Benarous, R., and Polakis, P. (1999).
The F-box protein beta-TrCP associates with phosphorylated beta-catenin
and regulates its activity in the cell. Curr. Biol. 9, 207–210.
Hashimoto, K., Mori, N., Tamesa, T., Okada, T., Kawauchi, S., Oga, A., Furuya,
T., Tangoku, A., Oka, M., and Sasaki, K. (2004). Analysis of DNA copy number
aberrations in hepatitis C virus-associated hepatocellular carcinomas by
conventional CGH and array CGH. Mod. Pathol. 17, 617–622.
Hellerbrand, C., Bumes, E., Bataille, F., Dietmaier, W., Massoumi, R., and
Bosserhoff, A.K. (2007). Reduced expression of CYLD in human colon and
hepatocellular carcinomas. Carcinogenesis 28, 21–27.
Hofmann, R.M., and Pickart, C.M. (1999). Noncanonical MMS2-encoded
ubiquitin-conjugating enzyme functions in assembly of novel polyubiquitin
chains for DNA repair. Cell 96, 645–653.
Jenner, M.W., Leone, P.E., Walker, B.A., Ross, F.M., Johnson, D.C., Gonzalez,
D., Chiecchio, L., Dachs Cabanas, E., Dagrada, G.P., Nightingale, M., et al.
(2007). Gene mapping and expression analysis of 16q loss of heterozygosity
identifies WWOX and CYLD as being important in determining clinical outcome
in multiple myeloma. Blood 110, 3291–3300.
Jiang, J., and Struhl, G. (1998). Regulation of the Hedgehog and Wingless
signalling pathways by the F-box/WD40-repeat protein Slimb. Nature 391,
493–496.
Jin, Z., Li, Y., Pitti, R., Lawrence, D., Pham, V.C., Lill, J.R., and Ashkenazi, A.
(2009). Cullin3-based polyubiquitination and p62-dep endent aggregation of
caspase-8 mediate extrinsic apoptosis signaling. Cell 137, 721–735.
Kerscher, O., Felberbaum, R., and Hochstrasser, M. (2006). Modification of
proteins by ubiquitin and ubiquitin-like proteins. Annu. Rev. Cell Dev. Biol.
22, 159–180.
Kirkin, V., and Dikic, I. (2007). Role of ubiquitin- and Ubl-binding proteins in cell
signaling. Curr. Opin. Cell Biol. 19, 199–205.
Kishida, S., Yamamoto, H., Hino, S., Ikeda, S., Kishida, M., and Kikuchi, A.
(1999). DIX domains of Dvl and Axin are necessary for protein interactions
and their ability to regulate beta-catenin stability. Mol. Cell. Biol. 19, 4414–
4422.
Komander, D., Lord, C.J., Scheel, H., Swift, S., Hofmann, K., Ashworth, A., and
Barford, D. (2008). The structure of the CYLD USP domain explains its speci-
ficity for Lys63-linked polyubiquitin and reveals a B box module. Mol. Cell 29,
451–464.
Korinek, V., Barker, N., Morin, P.J., van Wichen, D., de Weger, R., Kinzler,
K.W., Vogelste in, B., and Clevers, H. (1997). Constitutiv e transcriptional
activation by a beta-catenin-Tcf complex in APC
"/"
colon carcinoma. Science
275, 1784–1787.
Kovalenko, A., Chable-Bessia, C., Cantarella, G., Israel, A., Wallach, D., and
Courtois, G. (2003). The tumour suppressor CYLD negatively regulates
NF-kappaB signalling by deubiquitination. Nature 424, 801–805.
Lee, H.C., Kim, M., and Wands, J.R. (2006). Wnt/Frizzled signaling in hepato-
cellular carcinoma. Front. Biosci. 11, 1901–191 5.
Li, L., Yuan, H., Xie, W., Mao, J., Caruso, A.M., McMahon, A., Sussman, D.J.,
and Wu, D. (1999). Dishevelled proteins lead to two signaling pathways.
Regulation of LEF-1 and c-Jun N-terminal kinase in mammalian cells. J. Biol.
Chem. 274, 129–134.
Logan, C.Y., and Nusse, R. (2004). The Wnt signaling pathway in development
and disease. Annu. Rev. Cell Dev. Biol. 20, 781–810.
Narimatsu, M., Bose, R., Pye, M., Zhang, L., Miller, B., Ching, P., Sakuma, R.,
Luga, V., Roncari, L., Attisano, L., and Wrana, J.L. (2009). Regulation of planar
cell polarity by Smurf ubiquitin ligases. Cell 137, 295–307.
Nijman, S.M., Luna-Vargas, M.P., Velds, A., Brummelkamp, T.R., Dirac, A.M.,
Sixma, T.K., and Bernards, R. (2005). A genomic and functional inventory of
deubiquitinating enzymes. Cell 123, 773–786.
Ozawa, M., Baribault, H., and Kemler, R. (1989). The cytoplasmic domain of
the cell adhesion molecule uvomorulin associates with three independent
proteins structurally related in different species. EMBO J. 8, 1711–1717.
Pan, W.J., Pang, S.Z., Huang, T., Guo, H.Y., Wu, D., and Li, L. (2004).
Characterization of function of three domains in Dishevelled-1: DEP domain
is responsible for membrane translocation of Dishevelled-1. Cell Res. 14,
324–330.
Penton, A., Wodarz, A., and Nusse, R. (2002). A mutational analysis of Dishev-
elled in Drosophila defines novel domains in the Dishevelled protein as well as
novel suppressing alleles of Axin. Genetics 161, 747–762.
Rothbacher, U., Laurent, M.N., Deardorff, M.A., Klein, P.S., Cho, K.W., and
Fraser, S.E. (2000). Dishevelled phosphorylation, subcellular localization and
multimerization regulate its role in early embryogenesis. EMBO J. 19, 1010–
1022.
Schwarz-Romond, T., Merrifield, C., Nichols, B.J., and Bienz, M. (2005). The
Wnt signalling effector Dishev elled forms dynamic protein assemblies rather
than stable associations with cytoplasmic vesicles. J. Cell Sci. 118, 5269–
5277.
Schwarz-Romond, T., Fiedler, M., Shibata, N., Butler, P.J., Kikuchi, A.,
Higuchi, Y., and Bienz, M. (2007a). The DIX domain of Dishevelled confers
Wnt signaling by dynamic polymerization. Nat. Struct. Mol. Biol. 14, 484–492.
Schwarz-Romond, T., Metcalfe, C., and Bienz, M. (2007b). Dynamic recruit-
ment of Axin by Dishevelled protein assemblies. J. Cell Sci. 120, 2402–2412.
Simons, M., Gault, W.J., Gotthardt, D., Rohatgi, R., Klein, T.J., Shao, Y., Lee,
H.J., Wu, A.L., Fang, Y., Satlin, L.M., et al. (2009). Electrochemical cues
regulate assembly of the Frizzled/Dishevelled complex at the plasma
membrane during planar epithelial polarization. Nat. Cell Biol. 11, 286–294.
Smalley, M.J., Sara, E., Paterson, H., Naylor, S., Cook, D., Jayatilake, H., Fryer,
L.G., Hutchinson, L., Fry, M.J., and Dale, T.C. (1999). Interaction of Axin and
Dvl-2 proteins regulates Dvl-2-stimulated TCF-dependent transcription.
EMBO J. 18, 2823–2835.
Sun, L., and Chen, Z.J. (2004). The novel functions of ubiquitination in
signaling. Curr. Opin. Cell Biol. 16, 119–126.
Molecular Cell
CYLD Regulates Dvl Ubiquitination in Wnt Signaling
618 Molecular Cell 37, 607–619, March 12, 2010 ª2010 Elsevier Inc.
Page 12
Tamai, K., Zeng, X., Liu, C., Zhang, X., Harada, Y., Chang, Z., and He, X. (2004).
A mechanism for Wnt coreceptor activation. Mol. Cell 13 , 149–156.
Tran, H., Hamada, F., Schwarz-Romond, T., and Bienz, M. (2008). Trabid,
a new positive regul ator of Wnt-induced transcription with preference for
binding and cleaving K63-linked ubiquitin chains. Genes Dev. 22, 528–542.
Trompouki, E., Hatzivassiliou, E., Tsichritzis, T., Farmer, H., Ashworth, A., and
Mosialos, G. (2003). CYLD is a deubiquitinating enzyme that negatively regu-
lates NF-kappaB activation by TNFR family members. Nature 424, 793–796.
Umbhauer, M., Djiane, A., Goisset, C., Penzo-Mendez, A., Riou, J.F., Boucaut,
J.C., and Shi, D.L. (2000). The C-terminal cytopl asmic Lys-thr-X-X-X-Trp motif
in Frizzled receptors mediates Wnt/beta-catenin signalling. EMBO J. 19,
4944–4954.
Wallingford, J.B., and Habas, R. (2005). The developmenta l biology of Dishev-
elled: an enigmatic protein governing cell fate and cell polarity. Development
132, 4421–4436.
Wu, X., Yamamoto, M., Akira, S., and Sun, S.C. (2009). Regulation of hemato-
poiesis by the K63-specific ubiquitin-conjugating enzyme Ubc13. Proc. Natl.
Acad. Sci. U.S.A. 106, 20836–20841.
Xia, Z.P., Sun, L., Chen, X., Pineda, G., Jiang, X., Adhikari, A., Zeng, W., and
Chen, Z.J. (2009). Direct activation of protein kinases by unanchored polyubi-
quitin chains. Nature 461, 114–119.
Yanagawa, S., van Leeuwen, F., Wodarz, A., Klingensmith, J., and Nusse, R.
(1995). The Dishevelled protein is modified by wingless signaling in Drosophila.
Genes Dev. 9, 1087–1097.
Yu, A., Rual, J.F., Tamai, K., Harada, Y., Vidal, M., He, X., and Kirchhausen, T.
(2007). Association of Dishevelled with the clathrin AP-2 adaptor is required for
Frizzled endocytosis and planar cell polarity signaling. Dev. Cell 12, 129–141.
Zeng, X., Tamai, K., Doble, B., Li, S., Huang, H., Habas, R., Okamura, H.,
Woodgett, J., and He, X. (2005). A dual-kinase mechanism for Wnt co-receptor
phosphorylation and activation. Nature 438, 873–877.
Molecular Cell
CYLD Regulates Dvl Ubiquitination in Wnt Signaling
Molecular Cell 37, 607–619, March 12, 2010 ª2010 Elsevier Inc. 619
Page 13
  • Source
    • "By contrast, another K63-specific DUB called CYLD attenuates its signalling activity [24], which implies that ubiquitylation can also have a positive impact on Dvl. Examining the ubiquitylation of Dvl2 by mass spectroscopy, we found K11-Ub in the absence of Wnt, and additionally K63- Ub and K48-Ub after Wnt stimulation, consistent with previous reports [20,24]. Because each of these ubiquitin linkages can be generated by NEDD4 family E3 ligases [22], they could be the enzymes imparting these ubiquitylations. "
    [Show abstract] [Hide abstract] ABSTRACT: Dishevelled is a pivot in Wnt signal transduction, controlling both β-catenin-dependent transcription to specify proliferative cell fates, and cell polarity and other non-nuclear events in post-mitotic cells. In response to Wnt signals, or when present at high levels, Dishevelled forms signalosomes by dynamic polymerization. Its levels are controlled by ubiquitylation, mediated by various ubiquitin ligases, including NEDD4 family members that bind to a conserved PPxY motif in Dishevelled (mammalian Dvl1-3). Here, we show that Dvl2 binds to the ubiquitin ligase WWP2 and unlocks its ligase activity from autoinhibition. This disinhibition of WWP2 depends on several features of Dvl2 including its PPxY motif and to a lesser extent its DEP domain, but crucially on the ability of Dvl2 to polymerize, indicating that WWP2 is activated in Wnt signalosomes. We show that Notch intracellular domains are substrates for Dvl-activated WWP2 and their transcriptional activity is consequently reduced, providing a molecular mechanism for cross-talk between Wnt and Notch signalling. These regulatory interactions are conserved in Drosophila whose WWP2 orthologue, Suppressor-of-deltex, downregulates Notch signalling upon activation by Dishevelled in developing wing tissue. Attentuation of Notch signalling by Dishevelled signalosomes could be important during the transition of cells from the proliferative to the post-mitotic state.
    Preview · Article · Dec 2015 · Open Biology
  • Source
    • "To date, there are at least eight DUBs involved in this pathway, including CYLD, USP4, USP7, USP14, USP15, USP34, TRABID and OTULIN. In fact, CYLD serves as a negative regulator of Wnt signaling and β-catenin activation by deubiquitylating the cytoplasmic effector Dishevelled (Dvl) [104]. USP4 negatively regulates Wnt signaling by interacting with Nemolike kinase [105] and USP15 promotes β-catenin degradation through the stabilization of adenomatous polyposis coli (APC), a negative regulator of Wntmediated transcription [106]. "
    [Show abstract] [Hide abstract] ABSTRACT: Deubiquitinases are deubiquitinating enzymes (DUBs), which remove ubiquitin from proteins, thus regulating their proteasomal degradation, localization and activity. Here, we discuss DUBs as anti-cancer drug targets.
    Full-text · Article · May 2015 · Oncotarget
  • Source
    • "As a member of USPs subfamily, CYLD can antagonize Lysine-63 polyubiquitin chain conjugation (Kovalenko et al., 2003; Trompouki et al., 2003b). As mentioned previously, CYLD is involved in NF-κB, Wnt/β-catenin and JNK signaling pathway (Reiley et al., 2004; Tauriello et al., 2010; Trompouki et al., 2003b). By using CYLD knock-out mice, a recent study shows that in TGF-β-treated T cells, CYLD deficiency causes enhanced TAK1 and p38 mitogen-activated protein kinase activities (Zhao et al., 2011). "
    [Show abstract] [Hide abstract] ABSTRACT: Transforming growth factor-β (TGF-β) members are key cytokines that control embryogenesis and tissue homeostasis via transmembrane TGF-β type II (TβR II) and type I (TβRI) and serine/threonine kinases receptors. Aberrant activation of TGF-β signaling leads to diseases, including cancer. In advanced cancer, the TGF-β/SMAD pathway can act as an oncogenic factor driving tumor cell invasion and metastasis, and thus is considered to be a therapeutic target. The activity of TGF-β/SMAD pathway is known to be regulated by ubiquitination at multiple levels. As ubiquitination is reversible, emerging studies have uncovered key roles for ubiquitin-removals on TGF-β signaling components by deubiquitinating enzymes (DUBs). In this paper, we summarize the latest findings on the DUBs that control the activity of the TGF-β signaling pathway. The regulatory roles of these DUBs as a driving force for cancer progression as well as their underlying working mechanisms are also discussed.
    Full-text · Article · Apr 2014 · Protein & Cell
Show more