TRB3 interacts with SMAD3 promoting tumor cellmigration and invasion

Fang Hua*, Rong Mu*, Jinwen Liu, Jianfei Xue, Ziyan Wang, Heng Lin, Hongzhen Yang, Xiaoguang Chenand Zhuowei Hu`

Molecular Immunology and Pharmacology Laboratory, State Key Laboratory of Bioactive Substances and Functions of Natural Medicines, Institute ofMateria Medica, Chinese Academy of Medical Sciences and Peking Union Medical College, 1 Xian Nong Tan Street, Beijing, 100050, P R China

*These authors contributed equally to this work`Author for correspondence (huzhuowei@imm.ac.cn)

Accepted 27 May 2011Journal of Cell Science 124, 3235–3246� 2011. Published by The Company of Biologists Ltddoi: 10.1242/jcs.082875

SummaryTribbles homolog 3 (TRB3, also known as TRIB3, NIPK and SKIP3), a human homolog of Drosophila Tribbles, has been found to

interact with a variety of signaling molecules to regulate diverse cellular functions. Here, we report that TRB3 is a novel SMAD3-interacting protein. Expression of exogenous TRB3 enhanced the transcriptional activity of SMAD3, whereas knocking downendogenous TRB3 reduced the transcriptional activity of SMAD3. The kinase-like domain (KD) of TRB3 was responsible for the

interaction with SMAD3 and the regulation of SMAD3-mediated transcriptional activity. In addition, TGF-b1 stimulation oroverexpression of SMAD3 enhanced the TRB3 promoter activity and expression, suggesting that there is a positive feedback loopbetween TRB3 and TGF-b–SMAD3 signaling. Mechanistically, TRB3 was found to trigger the degradation of SMAD ubiquitin

regulatory factor 2 (Smurf2), which resulted in a decrease in the degradation of SMAD2 and phosphorylated SMAD3. Moreover, TRB3–SMAD3 interaction promoted the nuclear localization of SMAD3 because of the interaction of TRB3 with the MH2 domain of SMAD3.These effects of TRB3 were responsible for potentiating the SMAD3-mediated activity. Furthermore, knockdown of endogenous TRB3expression inhibited the migration and invasion of tumor cells in vitro, which were associated with an increase in the expression of E-

cadherin and a decrease in the expression of Twist-1 and Snail, two master regulators of epithelial-to-mesenchymal transition,suggesting a crucial role for TRB3 in maintaining the mesenchymal status of tumor cells. These results demonstrate that TRB3 acts as anovel SMAD3-interacting protein to participate in the positive regulation of TGF-b–SMAD-mediated cellular biological functions.

Key words: TRB3, SMAD3, Protein interaction

IntroductionThe transcription factor SMAD3 is a downstream modulator of TGF-

b signaling, and is involved in regulating a variety of physiological

and pathological processes, such as development, differentiation,

apoptosis, immunomodulation, fibrogenesis and carcinogenesis

(Kim et al., 2005). Following the activation and phosphorylation

by TGF-b type I receptor, SMAD3 forms a heterodimer with

SMAD4. The activated heteromeric SMAD complex is translocated

into the nucleus and regulates the transcription of target genes. The

specificity for target genes and biological effects of SMAD3-

mediated signaling is under the control of many SMAD-interacting

proteins including cytoplasmic partners (Conery et al., 2004; Xue et

al., 2010; Seong et al., 2007) and transcriptional cofactors (Ding et

al., 2009; Feng et al., 2000). The epithelial-to-mesenchymal

transition (EMT) is important in development, as well as in

fibrogenesis and tumor metastasis. SMAD3 is a key signaling

molecule for the induction of EMT by TGF-b and overexpression of

SMAD3 enhanced metastases of mice injected with MCF10CA1a

tumor cells but SMAD3DC, the dominant-negative mutant,

suppressed the metastases (Tian et al., 2003).

Tribbles was first identified in Drosophila as an inhibitor of

mitosis that regulates cell proliferation, migration and

morphogenesis during development (Grosshans and Wieschaus,

2000; Mata et al., 2000; Seher and Leptin, 2000). In mammals,

there are three Tribbles homologs – TRB1, TRB2 and TRB3. They

are considered to be members of the pseudokinase family, which

contain a Ser/Thr protein kinase-like domain but lack the ATP

binding pocket and catalytic residues. Despite lacking kinase

activity, tribbles has a scaffold-like function and participates in

protein complex assembly. TRB3 is the best-studied member of

the mammalian tribbles family. The interacting partners of TRB3

range from transcription factors, ubiquitin ligase, BMP type II

receptor to members of the MAPK and PI3K signaling pathways.

Through interacting with these proteins, it coordinates crucial

cellular processes, including glucose and lipid metabolism,

apoptosis, adipocyte differentiation, cell stress and regulation of

collagen expression (Bezy et al., 2007; Chan et al., 2007; Du et al.,

2003; Ohoka et al., 2005; Qi et al., 2006; Tang et al., 2008).

Recently, emerging evidence suggests that the three mammalian

tribbles homologs are crucial modulators of tumorogenesis. TRB1

and TRB2 are both involved in myeloid leukemogenesis (Jin et al.,

2007; Keeshan et al., 2006), whereas, TRB3 is highly expressed in

many human tumor cell lines and in several human tumor tissues

(Bowers et al., 2003; Xu et al., 2007). However, the function and

mechanism of interaction of TRB3 in cancer remains unknown.

In an effort to discover SMAD3-interacting proteins using a

yeast two-hybrid screening system, we identified a number of

molecules, including TRB3, that physically interact with SMAD3

Research Article 3235

Jour

nal o

f Cel

l Sci

ence

Fig. 1. TRB3 interacts with SMAD3. (A) SMAD3–Myc was

co-transfected with TRB3–HA into HEK293T cells. Whole cell

extracts were immunoprecipitated with an anti-Myc antibody

and blotted with an anti-HA. (B) Mapping of TRB3 domains

involved in SMAD3 binding. Top: deletion mutants of TRB3.

Below: HEK293T cells were transiently transfected with the

indicated constructs of TRB3. Whole cell extracts were

immunoprecipitated (IP) with anti-SMAD3 antibody. (C) TGF-

b1 promotes the endogenous interaction between SMAD3 and

TRB3. HepG2 cells were quiescent overnight and treated with

TGF-b1 (10 ng/ml) for 2 hours, or left untreated. Whole cell

extracts were immunoprecipitated with anti-SMAD3 antibody or

equal amount of rabbit IgG and blotted with an anti-TRB3

antibody. (D) Phosphorylation of SMAD3 increases the

SMAD3–TRB3 interaction. SMAD3–Myc or its two mutants

were co-transfected with TRB3–HA into HEK293T cells. Whole

cell extracts were immunoprecipitated with an anti-Myc

antibody and blotted with an anti-HA antibody. (E) In vitro

binding between TRB3 and SMAD3. Cell lysates from

HEK293T cells overexpressing TRB3–HA were incubated with

equal amounts of GST or GST–SMAD3 and analyzed by

western blotting using the anti-HA antibody. The presence of the

GST fusion proteins was confirmed by staining gels with

Coomassie Blue.

Fig. 2. Overexpression of TRB3 augments SMAD3-mediated

transcriptional activity and targeted gene expression. (A) HEK293T

cells were transiently transfected in 12-well plates with 500 ng of SBE4–

luc, 10 ng of TK-Renilla, with (+) or without (–) 1.0 mg of SMAD3–Myc

and 1.0 mg of HA-tagged TRB3 or its truncated mutants, as indicated.

After 24 hours of transfection, cells were incubated in the presence (gray

bars) or absence (black bars) of TGF-b1 for another 24 hours and then the

luciferase activity was measured. (B) HepG2 cells were transfected with

500 ng of 3TP–lux, 10 ng of TK-Renilla, and with (+) or without (–)

1.0 mg of SMAD3–Myc and 1.0 mg of TRB3 for 24 hours. The cells were

incubated with or without TGF-b1 (10 ng/ml) for another 24 hours and

luciferase activity was measured. (C) The effect of TRB3 on TGF-b

signaling is SMAD3 dependent. HEK293T cells were transiently

transfected in 12-well plates with 500 ng of SBE4–luc, 10 ng of TK-

Renilla and with (+) or without (–) 2.0 mg of HA-tagged TRB3, and 50 nM

of universal scrambled negative control siRNA duplexes or siRNA

duplexes targeting SMAD3. After 24 hours of transfection, cells were

incubated in the presence (gray bars) or absence (black bars) of TGF-b1

for another 24 hours and then the luciferase activity was measured.

(D) HepG2 cells were transiently transfected with TRB3–HA plasmid.

After 24 hours of transfection, cells were incubated in the presence (+) or

absence (–) of TGF-b1 for another 24 hours. Whole cell extracts were

prepared and the expression of p21, PAI-1, SMAD2/3, and the

phosphorylation of Ser465 and Ser467 of SMAD2, and Ser423 and Ser425

of SMAD3 (pSmad2 and 3, respectively) were analyzed by western

blotting. Values are means ¡ s.e.m. of three independent assays.

*P,0.05, **P,0.01, ***P,0.001.

Journal of Cell Science 124 (19)3236

Jour

nal o

f Cel

l Sci

ence

(Xue et al., 2010). In this study, the functional significance and

regulatory mechanism of the TRB3–SMAD3 interaction wasfurther explored. We found that the TRB3–SMAD3 interactionplays an important role in the regulation of SMAD3-mediated

transcriptional activity and in the reciprocally positive regulationof TRB3 expression. Knockdown of endogenous TRB3expression inhibited the migration and invasion of HepG2 cellsin vitro, through inhibition of EMT. Our studies suggest a pivotal

role of TRB3 in tumor progression and metastasis.

ResultsIdentification of TRB3 as an interacting protein of SMAD3

In a previous study, we identified a number of proteins, includingTRB3, as partners of SMAD3 in a yeast two-hybrid screening

system (Xue et al., 2010). To confirm the TRB3–SMAD3interaction in mammalian cells, co-immunoprecipitation assaywas carried out using HEK293T cells co-transfected with Myc–

SMAD3 and HA–TRB3. TRB3 could be precipitated with

SMAD3 (Fig. 1A) and the kinase-like domain (KD) of TRB3

mediated the interaction (Fig. 1B). To determine whether TRB3

specifically interacted with SMAD3 or generally interacted with

other members of the SMAD family, we also examined the

possible interaction of HA–TRB3 with Myc–SMAD2 and Myc–

SMAD4. We found that SMAD2 and SMAD4 could also be co-

immunoprecipitated with HA-tagged TRB3, similar to SMAD3

(supplementary material Fig. S1A). SMAD3 is a downstream

molecule of TGF-b1 signaling, so we therefore examined whether

TGF-b1 regulated the endogenous TRB3–SMAD3 interaction. In

the absence of TGF-b1, endogenous TRB3 was weakly associated

with SMAD3. However, TGF-b1 stimulation enhanced this

interaction (Fig. 1C). Because phosphorylation of the C-terminal

serine residues in SMAD3 is a crucial step in TGF-b1 signaling

(Massague et al., 2005), we tested whether the phosphorylation

status of SMAD3 might regulate the SMAD3–TRB3 interaction.

Fig. 3. Silencing endogenous TRB3 inhibits SMAD3-

mediated transcription and the expression of TGF-b1

target genes. (A) Silencing TRB3 inhibits SBE4–luc

reporter activity. HepG2 cells stably expressing control

shRNA or TRB3 shRNA were transfected with the SBE4–

luc reporter construct for 24 hours. The cells were incubated

with (gray bars) or without (black bars) TGF-b1 (10 ng/ml)

for another 24 hours. The cells were harvested and

luciferase activity was measured. (B) Silencing of TRB3

inhibits 3TP–lux reporter activity. The assay was conducted

as described in A, except for transfecting with the 3TP–lux

reporter construct instead of SBE4–luc reporter construct.

(C) Silencing TRB3 inhibits the expression of TGF-b

signaling target genes. HepG2 cells stably expressing

control shRNA or TRB3 shRNA were incubated with (+) or

without (–) TGF-b1 (10 ng/ml) for 24 hours. Whole cell

extracts were prepared and the expression of p21, PAI-1,

SMAD2/3, and the phosphorylation of Ser465 and Ser467 of

SMAD2 and Ser423 and Ser425 of SMAD3 (p-Smad2 and

3, respectively) were analyzed by western blotting. Values

are means ¡ s.e.m. of three independent assays. **P,0.01.

Fig. 4. Activation of TGF-b1–SMAD3

increases the expression of TRB3.

(A) TGF-b1 stimulates TRB3 expression.

HepG2 cells were treated with TGF-b1

(10 ng/ml) for the indicated time. Whole

cell extracts were prepared and the

expression of TRB3, p21, PAI-1,

phosphorylated SMAD3 (pSmad3),

SMAD3 and actin were analyzed by

western blotting. (B) A diagram of the

TRB3 promoter reporter gene. Position +1

is the initiation site for TRB3 transcription.

(C) Overexpression of SMAD3 causes

TRB3 induction. HEK293T cells were

transiently transfected with the indicated

amounts of SMAD3-expressing plasmid

plus the TRB3 promoter luciferase

(pTRB3-luc) plasmid. After 24 hours of

transfection, the cells were incubated with

(grey bars) or without (black bars) TGF-b1

(10 ng/ml) for another 24 hours and

luciferase activity was measured. Values

are means ¡ s.e.m. of three independent

assays. ***P,0.001.

TRB3 interacts with SMAD3 3237

Jour

nal o

f Cel

l Sci

ence

We found that the SMAD3D mutant, which has three serine

residues replaced by aspartate to mimic phosphorylation, bound to

TRB3 as well as the wild-type SMAD3 did. By contrast, the 3A

mutant, lacking the phosphorylation sites, showed no interaction

(Fig. 1D). Using a GST pull-down assay, TRB3 was found to bind

to a GST–SMAD3 fusion protein in vitro but not GST protein

alone (Fig. 1E). These results not only provide evidence

demonstrating an interaction between SMAD3 and TRB3 but

also show that the KD of TRB3 is crucial for the interaction.

However, the phosphorylation status of SMAD3 is important, but

not absolutely required, for this interaction because the non-

phosphorylated GST–SMAD3 fusion protein expressed in

Escherichia coli can also interact with TRB3 (Fig. 1E).

TRB3 regulates SMAD3-mediated transcription and target

gene expression

To determine the effect of the TRB3–SMAD3 interaction on

SMAD3-mediated transcription, HEK293T cells were co-

transfected with Myc–SMAD3, HA-tagged TRB3 or its truncated

mutants, and the SBE4–luciferase (luc) reporter, which contains

four copies of a SMAD consensus-binding element. Together with

SMAD3, TRB3 overexpression induced an increase in SBE4–luc

reporter activity and the KD fragment but not the N- or C-terminus

of TRB3 that induced the SBE4–luc activity. Stimulation with TGF-

b1 further strengthened the TRB3-induced activity (Fig. 2A).

Additionally, we found that TRB3 induced SMAD2-mediated

transcriptional activity (supplementary material Fig. S1B).

Next, we assessed the effect of TRB3 on TGF-b1-responsive

gene transcription. Overexpression of TRB3 promoted an

increase in 3TP–lux reporter activity in the presence or absence

of TGF-b1 (Fig. 2B). This effect of TRB3 is SMAD3 dependent

because SMAD3 knockdown led to a significant reduction in the

SBE4–luc activity induced by TRB3 expression and TGF-b1

stimulation (Fig. 2C).

We then examined the regulatory effect of TRB3 on TGF-b1–

SMAD3 signaling and target gene expression. We found that

overexpression of TRB3 enhanced the expression of PAI-1 and p21

even in the absence of TGF-b1. Interestingly, TRB3 overexpression

promoted the phosphorylation of SMAD3 and upregulated the

expression of SMAD2 but had little effect on the phosphorylation

of SMAD2 (Fig. 2D), whereas knock down of TRB3 with sequence

specific shRNA (Du et al., 2003) reduced the SBE4–luc and 3TP–

lux reporter activity to 50% of the control level even in the presence

of TGF-b1 (Fig. 3A,B). In addition, knock down of TRB3 reduced

Fig. 5. TRB3 promotes cell migration and invasion.

(A) Cell migration was measured using transwell assays for

parental, control-shRNA and TRB3-shRNA HepG2 cells.

The cells were plated in the upper chamber of the filters

that had been coated with fibronectin on the underside, and

stimulated with TGF-b1 (10 ng/ml) for 24 hours or left

unstimulated. Cells migrating to the underside of the

transwell insert were measured. (B) A transwell invasion

assay was performed as described in A, except that the

chambers were coated with basement membrane Matrigel

(30 mg/well) and cells migrating to the underside of the

insert were measured 48 hours after the start of the

experiment. (C) Cell invasion was measured using

transwell assays for parental, control-shRNA and TRB3-

shRNA HCT-8 cells. 24 hours after plating, the cells

migrating to the underside of the transwell insert were

measured. The expression levels of TRB3 in these cells

were detected by western blotting. (D) Transwell migration

assays were performed using NIH3T3 cells transfected

with or without TRB3–HA for 24 hours. The cells

migrating to the underside of the transwell insert were

measured. The expression levels of TRB3 in these cells

were detected by western blot. The value from parental

cells was arbitrarily set at 100%. Values are means ¡

s.e.m. of three independent assays. **P,0.01. The original

magnification of all images was 2006.

Journal of Cell Science 124 (19)3238

Jour

nal o

f Cel

l Sci

ence

Fig. 6. TRB3 positively regulates EMT process. (A) Parental HepG2 cells and cells expressing either control shRNA or TRB3 shRNA were stimulated with

TGF-b1 (10 ng/ml) for 24 hours or left unstimulated, and then immunofluorescence staining of E-cadherin and a-SMA was examined. Scale bars: 20 mm. (B)

Parental HepG2 cells and cells expressing either control shRNA or TRB3 shRNA were treated as described in A; whole cell extracts were prepared and the

expression of E-cadherin and a-SMA was examined by immunoblotting. (C) NIH3T3 cells were transiently transfected with TRB3–HA for 24 hours. Whole cell

extracts were prepared, and the expression of E-cadherin, a-SMA, TRB3 (anti-HA antibody) and actin were analyzed by western blotting. (D) Semi-quantitative

RT-PCR analysis of the E-cadherin repressors. Cells were grown to confluence in six-well plates. Total RNA was extracted from parental, control-shRNA and

TRB3-shRNA HepG2 cells. The expression of mRNA encoding for Twist-1, Twist-2, Zeb-2, Snail, Slug, TRB3 and b-actin was determined by RT-PCR using

specific primer sets. The value from parental cells was arbitrarily set at 1.0. (E) Protein expression level of the E-cadherin repressors. Cells were grown to

confluency in six-well plates. Whole cell extracts were prepared from control-shRNA and TRB3-shRNA HepG2 cells. The expression of Twist-1, Snail, Slug,

TRB3 and actin were analyzed by western blotting. The value from control-shRNA cells was arbitrarily set at 1.0. Values are means ¡ s.e.m. of three independent

assays. **P,0.01.

TRB3 interacts with SMAD3 3239

Jour

nal o

f Cel

l Sci

ence

the expression of PAI-1, p21, SMAD2 and phosphorylated SMAD3

(Fig. 3C), which was consistent with the result of Fig. 2D. Taken

together, these results indicated that TRB3 is a positive modulator

of TGF-b–SMAD3 signaling.

TGF-b1–SMAD3 stimulates TRB3 expression and activates

TRB3 promoter activity

Because TRB3 promoted SMAD3-mediated transcriptional

activity, we wondered if the expression of TRB3 could be

regulated by TGF-b1–SMAD3 signaling. We found that TGF-b1

induced a time-dependent expression of TRB3 as well as the

SMAD3-targeted proteins, such as p21 and PAI-1 in HepG2 cells

(Fig. 4A). The SMAD3-targeted proteins were induced much

more quickly than TRB3. The expression of p21 and PAI-1 was

induced 12 hours after TGF-b1 stimulation, whereas the

phosphorylation of SMAD3 was observed 1 hour after

stimulation and a high level of phosphorylation occurred at

6 hours after stimulation. TGF-b1 stimulation also induced a time-

Fig. 7. See next page for legend.

Journal of Cell Science 124 (19)3240

Jour

nal o

f Cel

l Sci

ence

dependent expression of these proteins in HCT-8 cells, a human

colorectal carcinoma cell line (supplementary material Fig. S2).

Using the promoter region of human TRB3 (–1038 , +847) as a

luciferase reporter system (TRB3–luc; Fig. 4B), we found that

overexpression of SMAD3 induced a concentration-dependent

transactivation of TRB3 and that stimulation with TGF-b1 further

induced its activation (Fig. 4C). Taken together with the data in

Figs 2 and 3, our study suggests that TRB3 can be regulated by

TGF-b1–SMAD3 signaling, and there exists a positive feedback

loop between TRB3 and TGF-b1–SMAD3 signaling.

TRB3 is essential for the migration and invasion of

tumor cells

TRB3 mRNA is highly expressed in several human tumor tissues

and multiple human tumor cell lines (Bowers et al., 2003; Xu et

al., 2007), indicating a potential role of TRB3 in tumorigenesis

and tumor progression. A large number of studies have provided

strong evidence for the role of TGF-b1–SMAD3 in tumor

invasion and metastasis (Micalizzi et al., 2009; Wesolowska et

al., 2007). Therefore, we examined the regulatory effect of TRB3

on migratory and invasive ability in HepG2 cells using transwell

migration and invasion assays. We found that TGF-b1 promoted

the migration and invasion of HepG2 cells, whereas knocking

down of TRB3 blocked cell migration and invasion in the

absence or presence of TGF-b1 (Fig. 5A,B). HCT-8 is a human

colorectal carcinoma cell line, with a high level of expression of

TRB3. Knockdown of TRB3 inhibited cell invasion as observed

in HepG2 cells (Fig. 5C). By contrast, overexpression of TRB3in NIH3T3 cells, which express low levels of endogenous TRB3,

promoted migration of the cells (Fig. 5D). Taken together, thesedata suggest that TRB3 participates in the regulation of tumorcells migration and invasion.

TRB3 is required for maintaining the mesenchymal statusof HepG2 cells

Epithelial-mesenchymal transition (EMT) plays an important rolein the conversion of early stage tumors into invasive malignancies(Kang and Massague, 2004). Epithelial cells undergoing EMT

typically lose the expression of epithelial markers such as E-cadherin and acquire the expression of mesenchymal markersincluding a-SMA and vimentin. Because we found that silencing

endogenous TRB3 blocked the migration and invasion of HepG2cells, we reasoned that TRB3 might have a role in the regulationof EMT. We examined EMT markers in HepG2 cells. As a

hepatoma cell line, parental and control short hairpin (shRNA)-expressing HepG2 cells showed modest a-SMA staining and lostthe staining of E-cadherin (Fig. 6A). It is generally accepted thatTGF-b can promote an invasive phenotype of tumor cells in later

stages of carcinoma by inducing EMT. We found that TGF-b1stimulation enhanced the expression of a-SMA but reduced theexpression of E-cadherin in HepG2 cells. Silencing TRB3

significantly reduced the expression of a-SMA but markedlyenhanced the expression of epithelial marker E-cadherin, even inthe presence of TGF-b1 (Fig. 6A). Western blot analysis of

cellular extracts from TRB3-silenced or TRB3-overexpressingcells confirmed the expression changes of a-SMA and E-cadherin(Fig. 6B,C).

We next examined the effect of TRB3 on the expression of the E-cadherin repressors known to regulate EMT. RT-PCR analysisshowed that Zeb-2 and slug (snail2) were constitutively expressed

in parental HepG2 cells and in the control-shRNA and TRB3-shRNA transfectants. However, expression of Twist-2 mRNA wasundetectable in all the three cell types. The expression of Twist-1

and Snail, two transcription factors contributing to EMT and

metastasis, was detectable in parental HepG2 cells and the control-shRNA transfectants but was significantly decreased in TRB3-shRNA transfectants (Fig. 6D). Western blotting confirmed the

expression change of Twist-1 and Snail (Fig. 6E). These dataindicate that TRB3 is crucial in maintaining the mesenchymalstatus of HepG2 cells and in the regulation of TGF-b1-induced

EMT.

TRB3–SMAD3 interaction maintains the nuclearlocalization of SMAD3

Previous work indicated that SMAD3 is distributed predominantlyin the cytoplasm (Xu et al., 2002) but TRB3 localizes in the nucleus

(Xu et al., 2007) in quiescent cells. When SMAD3 and TRB3 werecoexpressed, SMAD3 was concentrated and colocalized withTRB3 in the nucleus (Fig. 7A). To confirm those observations in

intact cells, nuclear and cytosolic fractions were extracted fromHepG2 cells transfected with TRB3–HA-expressing plasmid andthe level of SMAD3 in each fraction was determined. The cytosolic

and nuclear fractions were immunoblotted for histone H3 as anuclear marker and GAPDH as a cytosolic marker. We found thatno histone H3 was detected in the cytosolic fractions and no

GAPDH was detected in the nuclear fractions, verifying an efficientisolation of these fractions (Fig. 7B). Consistent withimmunofluorescence results, SMAD3 localized predominantly in

Fig. 7. TRB3 interacts with the MH2 domain of SMAD3 to maintain the

nuclear localization of SMAD3. (A) HepG2 cells were transiently

transfected with plasmid expressing SMAD3–Myc, TRB3–HA or co-

transfected with the both plasmids for 24 hours. The cells were

immunostained with an anti-Myc or anti-HA antibody, followed by FITC-

conjugated anti-mouse secondary antibody (green, for TRB3) or Alexa Fluor

594 anti-rabbit secondary antibody (red, for SMAD3). Scale bar: 25 mm. The

number of cells showing a predominant nuclear distribution of SMAD3 was

divided by the total number of cells to obtain the percentage of cells with

nuclear localization. Values are means ¡ s.e.m. of three independent

experiments. (B) HepG2 cells were transfected with TRB3–HA-expressing

plasmid or left untransfected. After 24 hours of transfection, the nuclear and

cytosolic fractions were extracted and assessed by western blot analysis. To

verify separation of nuclear and cytosolic fractions, histone H3 (nuclear

marker) and GAPDH (cytosolic marker) were immunoblotted. C.E. and N.E.,

cytosolic extract and nuclear extract, respectively. Values are means ¡ s.e.m.

of three independent assays. The value from untransfected HepG2 cells was

arbitrarily set at 1.0. (C) HepG2 cells expressing either control shRNA or TRB3

shRNA were stimulated with TGF-b1 (10 ng/ml) for 2 hours or left

untransfected. The cells were immunostained with an anti-SMAD3 antibody,

followed by FITC-conjugated anti-rabbit secondary antibody. DAPI was used

to stain the nuclei. Scale bars: 30 mm. The number of cells showing a

predominant nuclear distribution of SMAD3 was divided by the total number of

cells to obtain the percentage of cells with nuclear localization. Values are

means ¡ s.e.m. of three independent experiments. (D) HepG2 cells expressing

either control shRNA or TRB3 shRNA were stimulated with TGF-b1 (10 ng/

ml) for 2 hours or left untransfected. The nuclear and cytosolic fractions were

extracted and assessed as described in B. Values are means ¡ s.e.m. of three

independent assays. Cells expressing control shRNA without TGF-b1

stimulation were used as a control and the value from these cells was arbitrarily

set at 1.0. (E) Cell lysates from HEK293T overexpressing TRB3–HA were

incubated with equal amounts of GST–MH1, GST–MH2 or GST–linker and

analyzed by western blotting using the anti-HA antibody. The presence of the

GST fusion proteins (arrows) was confirmed by staining gels with Coomassie

Blue. Data is representative of two assays with identical results.

TRB3 interacts with SMAD3 3241

Jour

nal o

f Cel

l Sci

ence

the cytoplasm under basal condition. Transient transfection ofTRB3–HA resulted in accumulation of SMAD3 in the nucleus

(Fig. 7B). The effect of endogenous TRB3 on SMAD3 nuclear

localization was also examined. As shown in Fig. 7C,D, knockingdown TRB3 decreased the nuclear accumulation of SMAD3 and

inhibited TGF-b1-induced nuclear translocation of SMAD3 to acertain extent. Because SMAD proteins are continuously shuttling

between the nucleus and the cytoplasm regardless of the presenceof a TGF-b signal (Hill, 2009), we further examined how TRB3

regulates the localization of SMAD3. SMAD3 consists of twoglobular domains, namely the N-terminal MH1 domain and the C-

terminal MH2 domain, coupled by a linker region (Fig. 7E, upperpanel). The MH2 domain of SMAD3 is responsible for nuclear

export of SMAD3 through interaction with exportin 4 (Kurisaki etal., 2006). Using a GST pull-down assay we detected which domain

of SMAD3 interacted with TRB3. As illustrated in the lower panelof Fig. 7E, there was clear interaction between TRB3 and GST–

MH2, but not TRB3 and GST–linker. A weak interaction betweenGST–MH1 and TRB3 was also observed. These results indicatethat TRB3 binds directly to the MH2 domain of SMAD3, which

interferes with the nuclear export of SMAD3 (Kurisaki et al., 2006).

TRB3 interacts with Smurf2 and promotes its degradation

Overexpression of TRB3 increased the phosphorylation of SMAD3,whereas silencing of TRB3 caused the opposite effect. Because

TRB3 is a pseudokinase and lacks the kinase activity, it mightregulate the phosphorylation of SMAD3 by changing the stability ofphosphorylated SMAD3. Wu et al. reported recently that

phosphorylated SMAD3 is a target for SMAD ubiquitin regulatoryfactor 2 (Smurf2) ubiquitylation (Wu et al., 2008). TRB3 has been

Fig. 8. TRB3 interacts with Smurf2 and promotes its ubiquitylation. (A) Semi-quantitative RT-PCR analysis of Smurf2. Cells were grown to confluence in six-

well plate. Total RNA was extracted from parental, control-shRNA and TRB3-shRNA HepG2 cells. The expression of mRNA encoding for Smurf2, TBR3 and b-

actin was determined by RT-PCR using specific primer sets. The value from parental cells was arbitrarily set at 1.0. Values are means ¡ s.e.m. of three independent

assays. (B) The whole cell extracts of parental HepG2 cells and HepG2 cells stably expressing control shRNA or TRB3 shRNA were prepared, and the expression of

Smurf2, TRB3 and actin were analyzed by western blotting. The value from parental cells was arbitrarily set at 1.0. Values are means ¡ s.e.m. of three independent

assays. **P,0.01. (C) HEK293T cells were transiently transfected with increasing amounts of TRB3–HA-expressing plasmid together with Smurf2-DDK plasmid

for 24 hours. The cells were also transfected with an identical amount of pEGFP-N1 plasmid to monitor the transfection efficiency. The expression of Smurf2,

TRB3 or GFP was examined by western blotting with an anti-DDK antibody (for Smurf2), an anti-HA antibody (for TRB3), or an anti-GFP antibody, respectively.

(D) The interaction between TRB3 and Smurf2 was examined by immunoprecipitation assay. Smurf2-DDK was co-transfected with TRB3–HA into HEK293T

cells. Whole cell extracts were immunoprecipitated with an anti-DDK antibody and blotted with the anti-HA antibody. (E) Overexpression of TRB3 mediates the

ubiquitylation of Smurf2. HEK293T cells were transfected with ubiquitin, Smurf2–Myc, and TRB3–HA expression constructs. Whole cell extracts were prepared

and immunoprecipitated with the anti-Myc antibody. The precipitates were then blotted with an anti-ubiquitin antibody. (F) Overexpression of Smurf2 reverses the

regulatory effect of TRB3 on SMAD3-mediated transcriptional activity. HEK293T cells or HepG2 cells were transfected with different plasmids as indicated for

24 hours and luciferase activity was measured. Values are means ¡ s.e.m. of three independent assays. **P,0.01.

Journal of Cell Science 124 (19)3242

Jour

nal o

f Cel

l Sci

ence

reported to interact with Smurf1, which shows a high homology withSmurf2 (Lin et al., 2000), and to promote its degradation through the

ubiquitin–proteasome pathway (Chan et al., 2007). We thusexamined the regulatory effect of TRB3 on the expression ofSmurf2 in HepG2 cells. We found that silencing TRB3 increased theprotein but not mRNA expression of Smurf2 (Fig. 8A,B), whereas

increasing TRB3–HA expression decreased the levels of Smurf2(Fig. 8C). Moreover, the immunoprecipitation assay revealed astrong interaction between TRB3 and Smurf2 (Fig. 8D). We further

determined whether the ubiquitylation of Smruf2 was TRB3dependent. Compared with the control cells, many moreubiquitylated proteins (molecular mass .86 kDa) were

immunoprecipitated with anti-Myc (Smurf2–Myc) from TRB3-transfected cells (Fig. 8E), suggesting that TRB3 is a negativeregulator of Smurf2 by promoting Smurf2 ubiquitylation. Indeed,knockdown of SMURF2 using SMURF2-targeting siRNA duplexes

caused the accumulation of SMAD2 and phosphorylated SMAD3 inthe cells (supplementary material Fig. S3). Similar results had beenreported previously (Lin et al., 2000; Zhang et al., 2001; Wu et al.,

2008). Therefore we investigated whether TRB3 regulation ofSMAD3-mediated transcriptional activity was Smurf2 dependent.We found that expression of TRB3 enhanced SBE4–luc reporter

activity but this effect was significantly attenuated in the cellscoexpressing TRB3 and Smurf2 (Fig. 8F). Taken together, ourstudies indicate that TRB3 augments TGF-b1–SMAD3 signaling by

promoting ubiquitylation and degradation of Smurf2, whichattenuates the degradation of SMAD2 and phosphorylatedSMAD3 (Fig. 9).

DiscussionThe major finding of this study is that TRB3, an intracellularadaptor molecule, augments TGF-b1–SMAD3-mediated

transcriptional activity and cellular functions by physicallyinteracting with SMAD2/3. Knockdown of TRB3 expression intumor cells significantly inhibits the invasive and metastatic

ability of the cells by promoting mesenchymal-epithelialtransition. Moreover, TRB3 interacts with SMAD2 andSMAD4, two members of TGF-b–SMAD signaling family. Thehighly conserved MH2 domain of the SMAD proteins, as

indicated previously (Massague et al., 2005) might mediatetheir interaction with TRB3. However, the highly conservedtribbles kinase-like domain in TRB3 is also necessary for TRB3–

SMAD3 interaction and SMAD3-mediated transcription activity,suggesting an important role for the biological function of thisregion. Although only receptors and SMAD proteins are essential

components of the canonical TGF-b1–SMAD pathway, thispathway can be modulated by, and have a crosstalk with, othersignaling molecules through interaction with SMAD2/3 (Chen et

al., 2007). Thus, our studies indicate that TRB3 is a novel partnerand modulator of TGF-b–SMAD3 signaling, which might play animportant role in tumor progression and metastasis.

There is evidence that the activation and nuclear translocation

of SMAD3 results from TGF-b-induced phosphorylation at theC-terminal serine residues of SMAD3 (Massague, 1998;Moustakas et al., 2001). The observations that the positive

regulation of the endogenous TRB3–SMAD3 interaction byTGF-b1 and that the SMAD3D mutant can interact with TRB3confirm that the phosphorylation of SMAD3 is important for the

formation of the TRB3–SMAD3 complex. However, we foundthat GST–SMAD3 can also interact with TRB3 in pull-downassay while GST–SMAD3 fusion protein cannot be

phosphorylated when it is expressed in E. coli. These results

suggest that the subcellular localization rather than the

phosphorylation status is essential for the interaction between

TRB3 and SMAD3. In particular, the enhancement of the TRB3–

SMAD3 interaction when SMAD3 is phosphorylated might occur

because phosphorylated SMAD3 moves into the nucleus more

readily (Massague, 1998; Moustakas et al., 2001) and is therefore

more accessible for interaction. Moreover, previous work

indicates that SMAD proteins are continuously shuttling

between the nucleus and the cytoplasm and the MH2 domain

of SMAD3 is responsible for nuclear export of SMAD3 (Kurisaki

et al., 2006). We found in the current study that TRB3 interacts

with the MH2 domain of SMAD3, and overexpression of TRB3

increases the nuclear localization of SMAD3. On the basis of

these observations, we propose that the interaction between

TRB3 and the MH2 domain of SMAD3 might form a steric

hindrance to the interaction of SMAD3 with exportin 4, which

blocks the nuclear export of SMAD3. Such an effect mimics a

continuous activation of TGF-b–SMAD3 to exert its

transcriptional activity and other cellular actions (Fig. 9).

Indeed, Chan et al. recently report that TRB3 interacts with

BMPRII and positively regulates BMP and TGF-b signaling.

Because TRB3 is predominantly localized in the nucleus, these

authors surmised that TRB3 might play a direct role in the

transcriptional regulation of BMP and TGF-b target genes in

concert with SMAD proteins in the nucleus (Chan et al., 2007).

SMAD proteins are crucial intracellular mediators of TGF-bsignaling pathways. An alteration in SMAD protein levels can

profoundly affect signaling transduction and cellular functions.

Smurf2 is a HECT class ubiquitin E3 ligase, which can interact

with SMAD1, SMAD2 and SMAD3 (Lin et al., 2000). Several

studies have shown that Smurf2 ubiquitylates and degrades

SMAD1 and SMAD2, but not SMAD3 at steady state (Lin et al.,

2000; Zhang et al., 2001). Recently, Wu et al. reported

phosphorylated SMAD3 is a target for Smurf2 ubiquitylation

(Wu et al., 2008). We found that TRB3 interacts with Smurf2 to

promote degradation of Smurf2. Overexpression of TRB3

increases, whereas silencing of TRB3 decreases the levels of

both SMAD2 and phosphorylated SMAD3. SMAD2 and SMAD3

Fig. 9. Schematic diagram of the TRB3–SMAD3 interaction in the

regulation of TGF-b–SMAD3 signaling. (1) TRB3 keeps SMAD3 in the

nucleus by physically interacting with it; (2) TRB3 forms a complex with

Smurf2 that leads to the degradation of Smurf2, which in turn stabilizes the

targets of Smurf2, such as SMAD2 and phosphorylated SMAD3 to enhance

reciprocally the TGF-b–SMAD3 signaling activity.

TRB3 interacts with SMAD3 3243

Jour

nal o

f Cel

l Sci

ence

are both activated in response to TGF-b1 and act, in cooperation

with SMAD4, as effectors of the TGF-b response (Massague et

al., 2005). Thus, augmentation of TGF-b signaling by TRB3 may

not only rely on the decrease in the expression of Smurf1 (Chan

et al., 2007), but also depend on the TRB3 downregulation of

Smurf2 (Fig. 8) to maintain the stability of both phosphorylated

SMAD3 and SMAD2.

TGF-b1 plays an important role in tumor progression. An

increased level of TGF-b1 has been observed in many human

tumors and enhanced TGF-b1 can promote tumor invasion or

metastasis in the established tumor (Bierie and Moses, 2006). In the

present study, we found that silencing TRB3 inhibits cell migration

and invasion, even in the presence of TGF-b1. Overexpression of

TRB3 augments SMAD3-mediated transcription activity. TGF-b–

SMAD3 signaling in turn upregulates the expression of TRB3, thus

establishing a positive feedback loop that amplifies TGF-b–

SMAD3 signaling. Given that TRB3 is highly expressed in

human tumors (Bowers et al., 2003) and it positively promotes

TGF-b–SMAD3 signaling, as demonstrated by this study, the

expression level of TRB3 may, at least partly, be responsible for the

different roles of TGF-b–SMAD3 signaling in tumor inhibition or

tumor progression at the different stages of tumor development

(Bierie and Moses, 2006).

EMT is crucial in the conversion of early stage tumors into

invasive ones during tumor progression, because it allows tumor

cells to infiltrate the adjacent tissue, gain invasive properties and

ultimately form metastases at distal sites (Lee et al., 2006). TGF-b–

SMAD3 signaling regulates EMT through SMAD3-dependent or -

independent mechanisms (Xu et al., 2009). We found that silencing

TRB3 increases the expression of E-cadherin and decreases the

expression of mesenchymal marker a-SMA, suggesting that TRB3

is involved in maintaining the mesenchymal status of tumor cells.

The ability of TRB3 to maintain the mesenchymal status of cancer

cells could result from TRB3 enhancing SMAD3-mediated

transcription activity. Several developmentally important genes,

including those encoding for Twist-1, Twist-2, Snail, Slug and

Zeb-2 have been found to induce EMT through repression of the

gene encoding E-cadherin, a hallmark of EMT (Peinado et al., 2007).

Indeed, in our study, silencing TRB3 attenuated the expressions of

Twist-1 and Snail that have been identified as promoters of invasion

and metastasis in various types of human cancer (Batlle et al., 2000;

Yang et al., 2004; Yang et al., 2008). Moreover, Snail can form a

Snail–SMAD3 or –SMAD4 transcriptional repressor complex to

promote TGF-b-mediated EMT (Vincent et al., 2009). Thus, TRB3

induces EMT, cell migration and invasion through interacting with

SMAD3 to promote SMAD3-induced transcriptional activity and the

expression of EMT-related genes such as those encoding the

transcription factors Twist-1 and Snail.

In summary, our studies demonstrate that TRB3 is a novel

partner of SMAD3. The interaction of TRB3 and SMAD3 has a

crucial role in the regulation of SMAD3-mediated transcriptional

activity and in the reciprocally positive regulation of TRB3

expression. TRB3 augments TGF-b1–SMAD3 signaling through

two mechanisms: (1) keeping SMAD3 in the nucleus by physical

interaction; and (2) preventing the degradation of SMAD2 and

phosphorylated SMAD3 by promoting the degradation of Smurf2

(Fig. 9). Our studies indicate that TRB3 is important for

maintaining the mesenchymal status of tumor cells and may be

a potential therapeutic target for the treatment of human tumor

metastasis.

Materials and MethodsPlasmids constructionHA-tagged TRB3 and its truncations: DC (amino acids 1–179), DN (amino acids180–358) and KD (amino acids 71–315) were constructed in a pcDNA3.1-HAvector by standard subcloning. The TRB3-shRNA-targeting sequence has beenreported previously (Du et al., 2003). The sequence was subcloned intopSilencerTM3.1-H1 hygro-expressing vector (Ambion, Texas, USA). A controlshRNA oligonucleotide, which does not match any known human coding cDNA,was used as control. pTRB3-luc was generated by ligating the human TRB3promoter region (–1038 to +847) with pGL3-basic. Myc-tagged SMAD3A andSMAD3D mutants were generated by PCR as described previously (Liu et al.,1997). Full-length SMAD3 was subcloned in frame to the pGEX4T-1 to makeGST fusion proteins. Smurf2-DDK/Myc expression plasmid was purchased fromOrigene (Rockville, MD, USA). The ubiquitin gene was synthesized according tothe sequence published on NCBI web site (accession number: M17524) and clonedinto pFLAG-CMVTM-2 expression plasmid. SBE4-luc reporter plasmid wasobtained from Dr Bert Vogelstein of Johns Hopkins University MedicalInstitutions, Baltimore, MD. The 3TP–lux reporter plasmid was obtained fromDr Joan Massague of Sloan-Kettering Institute, New York, NY.

Cells lines and transfectionAll cell lines were obtained from the ATCC. Transient transfection was carried outusing LipofectAMINE2000 (Invitrogen, Carlsbad, CA, USA) according to themanufacturer’s instructions. To generate HepG2 cell populations stably expressingTRB3 shRNA, the control-shRNA and the TRB3-shRNA plasmids were bothtransfected into HepG2 cells with siPORTTM XP-1 tansfection reagent (Ambion)according to the manufacturer’s instructions. After 24 hours of transfection, stabletransfectants were selected in medium containing 200 mg/ml hygromycin(Calbiochem, San Diego, CA, USA) for 7 days. After two or three passages in thepresence of hygromycin, the cultures were used for experiments without cloning.

Immunoprecipitation and western blottingCells were washed three times with phosphate-buffered saline, harvested, and lysed inco-immunoprecipitation (co-IP) buffer that has been described before (Du et al., 2003).Total cell lysate (5 mg protein) was subjected to immunoprecipitation with appropriateantibodies, as indicated, overnight at 4 C with gentle agitation, followed by incubationwith protein A/G Plus–agarose for 2,4 hours at 4 C. The immunocomplex waswashed three times and then mixed with 23SDS sample buffer and boiled for5 minutes. For western blotting, co-precipitates or whole cell extracts were resolved bySDS-PAGE and blotted on PVDF membranes (Millipore, Bedford, MA, USA). Themembranes were immunoblotted with the indicated antibodies and developed using anECL detection system (Amersham Bioscience, Piscataway, NJ, USA).

GST pull-down assayGST fusion constructs of SMAD3 were expressed in Escherichia coli and purifiedusing glutathione–Sepharose 4B beads (Amersham Bioscience). Equal amounts ofGST or GST fusion proteins bound to glutathione–Sepharose beads were incubatedwith lysates from HEK293T cells transiently transfected with TRB3–HA. Beadswere washed three times and interacting proteins were detected by immunoblotting.Expression of GST fusion proteins was confirmed by Coomassie Blue staining.

Luciferase reporter assayHEK293T or HepG2 cells were seeded on 12-well plates at a density of 26105/well, and transfected with the indicated plasmids. TK-Renilla expression plasmidwas used as an internal control. The total amount of plasmid per well wasnormalized by the addition of pcDNA3 empty vector. After 24 hours oftransfection, cells were incubated in the presence (+) or absence (–) of TGF-b1for another 24 hours. Cells were harvested with lysis buffer, and the luciferaseactivity was then measured using the dual luciferase assay kit according to themanufacturer’s instruction (Promega, Madison, Wisconsin, USA).

Immunofluorescence microscopyHepG2 cells were plated on sterile coverlips and transfected with the expressionplasmids, as indicated, for 24 hours. Cells were fixed with 4% paraformaldehydefor 10 minutes at room temperature, permeabilized with 0.5% Triton X-100 for10 minutes, and blocked with 3% BSA for 30 minutes at 37 C. The coverslipswere incubated with primary antibodies at 4 C overnight. After washing threetimes with PBS, cells were stained with Alexa Fluor 594 goat anti-mouse IgG andFITC-conjugated goat anti-rabbit antibodies for 30 minutes at 37 C then washedthree times with PBS and mounted on glass slides. Images were acquired using aconfocal microscope (Leica Microsystems Heidelberg GmbH, TCS SP2).

Cellular fractionationTo prepare the nuclear and cytoplasmic fractions, the cultured cells were lysedusing a Nuclear and Cytoplasmic Protein Extraction Kit (Beyotime Institute ofBiotechnology, Suzhou, China) following the manufacturer’s instructions. Briefly,cells was washed with iced PBS and added to buffer A for 15 minutes and then

Journal of Cell Science 124 (19)3244

Jour

nal o

f Cel

l Sci

ence

centrifuged at 12,000 g for 15 minutes at 4 C. The supernatants were collected asthe cytoplasmic fraction. The pellet was washed with PBS and resuspended in anuclear extraction buffer. The mixture was votexed vigorously for 15 seconds andput on ice for 10 minutes. The process was repeated 4 or 5 times, and thencentrifuged again (12,000 g for 15 minutes, at 4 C). The obtained supernatant wascollected as the nuclear fraction. The separated cytoplasmic and nuclear fractionswere subjected to SDS-PAGE and immunoblotting.

Cell migration and invasion assays

Transwell migration assays were performed using Transwell chambers with filtermembranes of 8 mm pore size (Millipore). Chambers were precoated with 10 mg/mlfibronectin on the lower surface as previously described (Katoh et al., 2006). Cellswere starved overnight in assay medium [Dulbecco’s modified Eagle’s medium(DMEM) containing 0.4% FBS] and then seeded into the upper chamber (56104

cells per well in 0.4% FBS in DMEM). After 24 hours, non-migrating cells on theupper side of the filter were removed with a cotton swab. Cells migrated were stainedand counted. Transwell invasion assays were done under the same conditions as thetranswell migration assays, but before invasion assays, the polycarbonate filter wascoated with Matrigel (30 mg/well; BD Bioscience, Bedford, MA, USA).

Semi-quantitative RT-PCR

Total RNA was extracted using TRIzol (Invitrogen) following the manufacturer’sinstructions. Reverse transcription of the total cellular RNA was carried out usingoligo(dT) primers and M-MLV reverse transcriptase (Promega). PCR wasperformed using an Mycycler thermal cycler and analyzed using agarose gels.

Statistical analysis

Data are expressed as mean ¡ standard error of the mean (s.e.m.). Student’s t-testwas used for two-group comparisons. Comparisons between three or more groupswere analyzed by one-way ANOVA followed by Duncan’s test in SPSS 13.0(SPSS Inc.). P,0.05 was considered statistically significant.

AcknowledgementsWe thank Dr Bert Vogelstein for the SBE4–luc reporter plasmid, andDr Joan Massague for the 3TP–lux reporter plasmid. This work wassupported by grants from the National Major Basic ResearchProgram of China (973: #2006CB503808), the National NaturalScience Foundation of China (30472025; 30973557) and Creation ofMajor New Drugs (2009ZX09301-003-13; 2009ZX09301-003-9-1).Dr Zhuowei Hu is also supported by the Cheung-Kong ScholarsProgramme of Ministry of Education and by a Senior OverseasChinese Scholar Fund from the Ministry of Personnel of PRC. DrFang Hua is supported by grants from the Basic Research Program ofthe Institute of Materia Medica (2006QN33; 2010CHX12).

Supplementary material available online at

http://jcs.biologists.org/lookup/suppl/doi:10.1242/jcs.082875/-/DC1

ReferencesBatlle, E., Sancho, E., Franci, C., Dominguez, D., Monfar, M., Baulida, J. and

Garcia De Herreros, A. (2000). The transcription factor snail is a repressor of E-cadherin gene expression in epithelial tumour cells. Nat. Cell Biol. 2, 84-89.

Bezy, O., Vernochet, C., Gesta, S., Farmer, S. R. and Kahn, C. R. (2007). TRB3blocks adipocyte differentiation through the inhibition of C/EBPb transcriptionalactivity. Mol. Cell. Biol. 27, 6818-6831.

Bierie, B. and Moses, H. L. (2006). Tumour microenvironment: TGF-b: the molecularJekyll and Hyde of cancer. Nat. Rev. Cancer 6, 506-520.

Bowers, A. J., Scully, S. and Boylan, J. F. (2003). SKIP3, a novel Drosophila tribblesortholog, is overexpressed in human tumors and is regulated by hypoxia. Oncogene

22, 2823-2835.

Chan, M. C., Nguyen, P. H., Davis, B. N., Ohoka, N., Hayashi, H., Du, K., Lagna, G.

and Hata, A. (2007). A novel regulatory mechanism of the bone morphogeneticprotein (BMP) signaling pathway involving the carboxyl-terminal tail domain ofBMP type II receptor. Mol. Cell. Biol. 27, 5776-5789.

Chen, Y. G., Wang, Z., Ma, J., Zhang, L. and Lu, Z. (2007). Endofin, a FYVE domainprotein, interacts with Smad4 and facilitates transforming growth factor-betasignaling. J. Biol. Chem. 282, 9688-9695.

Conery, A. R., Cao, Y., Thompson, E. A., Townsend, C. M., Jr, Ko, T. C. and Luo,

K. (2004). Akt interacts directly with Smad3 to regulate the sensitivity to TGF-binduced apoptosis. Nat. Cell Biol. 6, 366-372.

Ding, L., Wang, Z., Yan, J., Yang, X., Liu, A., Qiu, W., Zhu, J., Han, J., Zhang, H.,

Lin, J. et al. (2009). Human four-and-a-half LIM family members suppress tumorcell growth through a TGF-b-like signaling pathway. J. Clin. Invest. 119, 349-361.

Du, K., Herzig, S., Kulkarni, R. N. and Montminy, M. (2003). TRB3: a tribbleshomolog that inhibits Akt/PKB activation by insulin in liver. Science 300, 1574-1577.

Feng, X. H., Lin, X., Derynck, R. (2000). Smad2, Smad3 and Smad4 cooperate with

Sp1 to induce p15(Ink4B) transcription in response to TGF-b. EMBO J. 19, 5178-

5193.

Grosshans, J. and Wieschaus, E. (2000). A genetic link between morphogenesis and

cell division during formation of the ventral furrow in Drosophila. Cell 101, 523-531.

Hill, C. S. (2009). Nucleocytoplasmic shuttling of Smad proteins. Cell Res. 19, 36-46.

Jin, G., Yamazaki, Y., Takuwa, M., Takahara, T., Kaneko, K., Kuwata, T., Miyata,

S. and Nakamura, T. (2007). Trib1 and Evi1 cooperate with Hoxa and Meis1 in

myeloid leukemogenesis. Blood 109, 3998-4005.

Kang, Y. and Massague, J. (2004). Epithelial-mesenchymal transitions: twist in

development and metastasis. Cell 118, 277-279.

Katoh, H., Hiramoto, K. and Negishi, M. (2006). Activation of Rac1 by RhoG

regulates cell migration. J. Cell Sci. 119, 56-65.

Keeshan, K., He, Y., Wouters, B. J., Shestova, O., Xu, L., Sai, H., Rodriguez, C. G.,

Maillard, I., Tobias, J. W., Valk, P. et al. (2006). Tribbles homolog 2 inactivates

C/EBPa and causes acute myelogenous leukemia. Cancer Cell 10, 401-411.

Kim, I. Y., Kim, M. M. and Kim, S. J. (2005). Transforming growth factor-beta:

biology and clinical relevance. J. Biochem. Mol. Biol. 38, 1-8.

Kurisaki, A., Kurisaki, K., Kowanetz, M., Sugino, H., Yoneda, Y., Heldin, C. H. and

Moustakas, A. (2006). The mechanism of nuclear export of Smad3 involves exportin

4 and Ran. Mol. Cell. Biol. 26, 1318-1332.

Lee, J. M., Dedhar, S., Kalluri, R. and Thompson, E. W. (2006). The epithelial-

mesenchymal transition: new insights in signaling, development, and disease. J. Cell

Biol. 172, 973-981.

Lin, X., Liang, M. and Feng, X. H. (2000). Smurf2 is a ubiquitin E3 ligase mediating

proteasome-dependent degradation of Smad2 in transforming growth factor-bsignaling. J. Biol. Chem. 275, 36818-36822.

Liu, X., Sun, Y., Constantinescu, S. N., Karam, E., Weinberg, R. A. and Lodish,

H. F. (1997). Transforming growth factor b-induced phosphorylation of Smad3 is

required for growth inhibition and transcriptional induction in epithelial cells. Proc.

Natl. Acad. Sci. USA 94, 10669-10674.

Massague, J. (1998). TGF-b signal transduction. Annu. Rev. Biochem. 67, 753-791.

Massague, J., Seoane, J. and Wotton, D. (2005). Smad transcription factors. Genes

Dev. 19, 2783-2810.

Mata, J., Curado, S., Ephrussi, A. and Rørth, P. (2000). Tribbles coordinates mitosis

and morphogenesis in Drosophila by regulating String/CDC25 proteolysis. Cell 101,

511-522.

Micalizzi, D. S., Christensen, K. L., Jedlicka, P., Coletta, R. D., Baron, A. E.,

Harrell, J. C., Horwitz, K. B., Billheimer, D., Heichman, K. A., Welm, A. L. et al.

(2009). The Six1 homeoprotein induces human mammary carcinoma cells to undergo

epithelial-mesenchymal transition and metastasis in mice through increasing TGF-bsignaling. J. Clin. Invest. 119, 2678-2690.

Moustakas, A., Souchelnytskyi, S. and Heldin, C. H. (2001). Smad regulation in TGF-

b signal transduction. J. Cell Sci. 114, 4359-4369.

Ohoka, N., Yoshii, S., Hattori, T., Onozaki, K. and Hayashi, H. (2005). TRB3, a

novel ER stress-inducible gene, is induced via ATF4-CHOP pathway and is involved

in cell death. EMBO J. 24, 1243-1255.

Peinado, H., Olmeda, D. and Cano, A. (2007). Snail, Zeb and bHLH factors in tumour

progression: an alliance against the epithelial phenotype? Nat. Rev. Cancer 7, 415-

428.

Qi, L., Heredia, J. E., Altarejos, J. Y., Screaton, R., Goebel, N., Niessen, S.,

Macleod, I. X., Liew, C. W., Kulkarni, R. N., Bain, J. et al. (2006). TRB3 links the

E3 ubiquitin ligase COP1 to lipid metabolism. Science 312, 1763-1766.

Seher, T. C. and Leptin, M. (2000). Tribbles, a cell-cycle brake that coordinates

proliferation and morphogenesis during Drosophila gastrulation. Curr. Biol. 10, 623-

629.

Seong, H.-A., Jung, H., Kim, K.-T. and Ha, H. (2007). 3-Phosphoinositide-dependent

PDK1 negatively regulates transforming growth factor-beta-induced signaling in a

kinase-dependent manner through physical interaction with Smad proteins. J. Biol.

Chem. 282, 12272-12289.

Tang, M., Zhong, M., Shang, Y., Lin, H., Deng, J., Jiang, H., Lu, H., Zhang, Y. and

Zhang, W. (2008). Differential regulation of collagen types I and III expression in

cardiac fibroblasts by AGEs through TRB3/MAPK signaling pathway. Cell. Mol. Life

Sci. 65, 2924-2932.

Tian, F., DaCosta Byfield, S., Parks, W. T., Yoo, S., Felici, A., Tang, B., Piek, E.,

Wakefield, L. M. and Roberts, A. B. (2003). Reduction in Smad2/3 signaling

enhances tumorigenesis but suppresses metastasis of breast cancer cell lines. Cancer

Res. 63, 8284-8292.

Vincent, T., Neve, E. P., Johnson, J. R., Kukalev, A., Rojo, F., Albanell, J., Pietras,

K., Virtanen, I., Philipson, L., Leopold, P. L. et al. (2009). A SNAIL1-SMAD3/4

transcriptional repressor complex promotes TGF-beta mediated epithelial-mesench-

ymal transition. Nat. Cell Biol. 11, 943-950.

Wesolowska, A., Kwiatkowska, A., Slomnicki, L., Dembinski, M., Master, A., Sliwa,

M., Franciszkiewicz, K., Chouaib, S. and Kaminska, B. (2007). Microglia-derived

TGF-b as an important regulator of glioblastoma invasion-an inhibition of TGF-b -

dependent effects by shRNA against human TGF-b type II receptor. Oncogene 27,

918-930.

Wu, Q., Kim, K. O., Sampson, E. R., Chen, D., Awad, H., O’Brien, T., Puzas, J. E.,

Drissi, H., Schwarz, E. M., O’Keefe, R. J. et al. (2008). Induction of an

osteoarthritis-like phenotype and degradation of phosphorylated Smad3 by Smurf2 in

transgenic mice. Arthritis Rheum. 58, 3132-3144.

TRB3 interacts with SMAD3 3245

Jour

nal o

f Cel

l Sci

ence

Xu, J., Lv, S., Qin, Y., Shu, F., Xu, Y., Chen, J., Xu, B. E., Sun, X. and Wu, J. (2007).TRB3 interacts with CtIP and is overexpressed in certain cancers. Biochim. Biophys.

Acta 1770, 273-278.Xu, J., Lamouille, S. and Derynck, R. (2009). TGF-b-induced epithelial to

mesenchymal transition. Cell Res. 19, 156-172.Xu, L., Kang, Y., Col, S. and Massague, J. (2002). Smad2 nucleocytoplasmic shuttling

by nucleoporins CAN/Nup214 and Nup153 feeds TGF-b signaling complexes in thecytoplasm and nucleus. Mol. Cell 10, 271-282.

Xue, J. F., Hua, F., Lv, Q., Lin, H., Wang, Z. Y., Yan, J., Liu, J. W., Lv, X. X., Yang,

H. Z. and Hu, Z. W. (2010). DEDD negatively regulates transforming growth factor-b1 signaling by interacting with Smad3. FEBS Lett. 584, 3028-3034.

Yang, J., Mani, S. A., Donaher, J. L., Ramaswamy, S., Itzykson, R. A., Come, C.,

Savagner, P., Gitelman, I., Richardson, A. and Weinberg, R. A. (2004). Twist, a

master regulator of morphogenesis, plays an essential role in tumor metastasis. Cell

117, 927-939.

Yang, M. H., Wu, M. Z., Chiou, S. H., Chen, P. M., Chang, S. Y., Liu, C. J., Teng,

S. C. and Wu, K. J. (2008). Direct regulation of TWIST by HIF-1alpha promotes

metastasis. Nat. Cell Biol. 10, 295-305.

Zhang, Y., Chang, C., Gehling, D. J., Hemmati-Brivanlou, A. and Derynck, R.

(2001). Regulation of Smad degradation and activity by Smurf2, an E3 ubiquitin

ligase. Proc. Natl. Acad. Sci. USA 98, 974-979.

Journal of Cell Science 124 (19)3246

Jour

nal o

f Cel

l Sci

ence