Chemistry & Biology ArticleIRDs have been reported to act as multiple kinase inhibitors (Meijer et al., 2006). Based on the cocrystallization structure of indirubin E231 in the ATP-binding

Chemistry & Biology

Article

Indirubin Derivatives Modulate TGFb/BMP Signalingat Different Levels and Trigger Ubiquitin-MediatedDepletion of Nonactivated R-SmadsXinlai Cheng,1,2 Hamed Alborzinia,1 Karl-Heinz Merz,2 Herbert Steinbeisser,3 Ralf Mrowka,4 Catharina Scholl,1

Igor Kitanovic,1 Gerhard Eisenbrand,2 and Stefan Wolfl1,*1Institute of Pharmacy and Molecular Biotechnology, Heidelberg University, Im Neuenheimer Feld 364, D-69120 Heidelberg, Germany2Department of Chemistry, University of Kaiserslautern, Erwin-Schrodinger-Str. 52, D-67663 Kaiserslautern, Germany3Institute of Human Genetics, Heidelberg University, Im Neuenheimer Feld 366, D-69120 Heidelberg, Germany4Experimentelle Nephrologie. KIM III, Universitatsklinikum Jena, Am Nonnenplan 2-4, D-07734 Jena, Germany

*Correspondence: [email protected]

http://dx.doi.org/10.1016/j.chembiol.2012.09.008

SUMMARY

Regulatory Smads (R-Smads), Smad1/5/8 andSmad2/3, are the central mediators of TGFb andBMP signaling pathways. Here, we screened indiru-bin derivatives, known kinase inhibitors, and ob-served strong interference with BMP signaling. Wefound that indirubin derivative E738 inhibited bothTGFb and BMP pathways through ubiquitin-protea-some-mediated depletion of total R-Smad pools,although phospho-R-Smad levels were initiallystabilized by GSK3b and cyclin-dependent kinaseinhibition. E738 also enhanced p38 and JNK phos-phorylation, involved in Smad-independent TGFb/BMP signaling. Additionally, using a small siRNAscreen, we showed that depletion of ubiquitin prote-ases USP9x and USP34 significantly reduced totalR-Smad levels, mimicking E738 treatment. In fact,both USP9x and USP34 levels were significantlyreduced in E738-treated cells. Our findings not onlydescribe the complex activity profile of the indirubinderivative E738, but also reveal a mechanism forcontrolling TGFb/BMP signaling, the control ofR-Smad protein levels through deubiquitination.

INTRODUCTION

Transforming growth factor b (TGFb) and bone morphogenetic

protein (BMP) belong to the cytokine growth factor family and

play key roles in embryonic development and organ homeo-

stasis, by regulating cellular differentiation, extracellular matrix

remodeling and proliferation (Hong and Yu, 2009; De Robertis

and Kuroda, 2004; Eivers et al., 2009). Dysregulation of TGFb

and BMP signaling is associated with several diseases, including

cancer (Walsh et al., 2010; Wharton and Derynck, 2009; Pardali

et al., 2010). TGFb/BMP signaling is initiated by ligand binding

to receptor kinases on the cell surface. This induces formation

of heteromeric receptor complexes, which phosphorylate

regulatory Smads (R-Smads), Smad1/5/8 for BMP, or Smad2/3

Chemistry & Biology 19, 1423–143

for TGFb signaling at the C-terminal Mad homology 2 (MH2)

domain. Upon binding, the common partner Smad4 (Co-Smad)

R-Smads are translocated to the nucleus, where activated

Smad complexes bind DNA and regulate gene expression,

including IDs (Massague et al., 2005), CTGF (Abreu et al.,

2002), and SnoN (Pot and Bonni, 2008). Receptor-specific

inhibitors, the BMP-receptor inhibitor dorsomorphin (DM) (Yu

et al., 2008), and the TGFb-receptor inhibitor RepSox (Gellibert

et al., 2004; Ichida et al., 2009) efficiently suppress expression

of target genes by preventing phosphorylation of respective

R-Smads. R-Smads also comprise an N-terminal MH1 domain,

which is connected with the MH2 domain via an interdomain

linker region (Hariharan and Pillai, 2008). Phosphorylation of

this region enhances signaling and mediates turnover (Pera

et al., 2003). TGFb and BMP also activate MAP-kinases,

including ERK, p38, and JNK, independent of R-Smads (Derynck

and Zhang, 2003).

Indirubin is a major active ingredient of a Chinese herbal

prescription named Dangui Luhui Wan, used in traditional

Chinese medicine (TCM) for treatment of various diseases

and has been used clinically to treat chronic myelogenous

leukemia (CML) (Eisenbrand et al., 2004). Indirubin and indiru-

bin derivatives (IRDs) are potent protein kinase inhibitors.

Target kinases include serine/threonine kinases, such as

cyclin-dependent kinases (CDKs) (Hoessel et al., 1999) and

glycogen synthase kinase 3b (GSK3b) (Meijer et al., 2003),

and receptor and nonreceptor tyrosine kinases, like c-Src

(Nam et al., 2005) and FGFR1 (Zhen et al., 2007). The IRD

6-bromoindirubin-30-oxime (BIO) has been used to maintain

pluripotency of mouse and human embryonic stem cells

(ESCs) (Sato et al., 2004) and for generation of epiblast stem

cells (Chou et al., 2008). Because TGFb/BMP signaling is impli-

cated in these processes, we investigated whether IRDs target

this pathway.

In the present work, we establish IRDs as potent regulators

of TGFb/BMP signaling acting at different levels and demon-

strate that total R-Smad levels are regulated by specific ubiqui-

tination signals. Blocking deubiquitination, R-Smads were

rapidly depleted and TGFb/BMP/Smad signaling was sup-

pressed. With these observations, we add an important aspect

to the understanding of the intricate network of TGFb/BMP

signaling.

6, November 21, 2012 ª2012 Elsevier Ltd All rights reserved 1423

mailto:[email protected]


Chemistry & Biology

Modulation of Nonactivated R-Smad Levels

RESULTS

IRDs Interfere with BMP Signaling in HeLa Cellsand Human FibroblastsIn an exploratory structure-activity investigation, several IRDs

synthesized in our lab were tested at two concentrations (1 and

5 mM; Figure 1A) in HeLa cells treated with BMP2 (100 ng/ml)

for 24 hr. Although treatment with indirubin-30-oxime (E231) did

not significantly affect phospho- and total Smad1/5/8 levels, ana-

logs carrying either a propanediol group on 30-position (E804), or

a methoxy group at 5-position (E728) or both (E738) led to sig-

nificant depletion of phospho- and total Smad1/5/8. In contrast

5-methoxyindirubin (E723) and the 5-carboxamide (E748a)

showed only a decrease of phospho-Smad1/5/8 at 5 mM. For

E860a, an oxime ether with a large basic substituent in 30-position, no effect was observed at 1 mM, whereas 5 mM almost

completely depleted total and phospho-Smad1/5/8 levels.

The observed differences suggest that separate mechanisms

are required for the regulation of BMP/Smad signaling by IRDs.

Whereas E723 and E748a specifically blocked phosphorylation,

other IRDs also reduced total Smad1/5/8 levels.

In presence of 10% fetal bovine serum (FBS), the influence of

IRDs on BMP/Smad signaling was much weaker (Figure 1B) and

depletion of phospho- and total Smad1/5/8 was observed only

with 5 mM E738. The lack of activity of other IRDs in presence

of FBS indicates that FBS activates pathways, which attenuate

the effect of IRDs.

Experiments were repeated with human foreskin fibroblasts

(HFF) and comparable results were obtained (Figure 1C; Fig-

ure S1A available online). The influence of IRDs on phospho-

and total Smad1/5/8 levels correlated with their structures in all

experiments. E738, being the most active IRD in both cell lines,

was chosen for further experiments.

E738 Decreases Total Smad1/5/8 Levels

in Different Cell Lines

We next studied the effects of E738 on BMP/Smad signaling in

C2C12, HFF, NIH 3T3, HEK293, HaCat, and Jurkat cells. Probing

cell lysates with phospho- and total-Smad1/5/8-specific anti-

bodies a dose-dependent reduction of phospho- and total

Smad1/5/8 was observed in all six cell lines (Figures 1D and

S1B–S1D). Consistent with these results, immunostaining of

C2C12 cells treated 12 hr with 1 mM E738 showed nearly

complete depletion of the total Smad1/5/8 reservoir (Figure 1E).

Comparison of E738 with Other Inhibitors

and Modulators of BMP/TGFb Signaling

The IRD BIO is a well-known GSK3b (IC50, 5 nM) and CDK

(CDK1/CycB IC50, 320 nM and CDK5/p25 IC50, 83 nM) inhibitor

(Meijer et al., 2003) and has been used in mouse somatic cell

reprogramming and maintaining of mouse iPSCs (induced

pluripotent stem cells), which involves modulation of BMP and

TGFb signaling (Ichida et al., 2009; Li et al., 2009; Yuan et al.,

2011; Xu et al., 2008). We therefore compared the effect of

E738 on BMP2 signaling in HeLa cells with BIO and BMP and

TGFb inhibitors, dorsomorphin (DM) and RepSox, (Zhu et al.,

2010; Ichida et al., 2009; Yuan et al., 2011). As shown in Fig-

ure 2A, treatment with E738 (1 mM) reduced the concentration

of total and phospho-Smad1/5/8, whereas BIO (5 mM) lowered

Smad1/5/8 phosphorylation, but only marginally decreased

total Smad1/5/8 levels. DM (1 mM) blocked Smad1/5/8 phos-

1424 Chemistry & Biology 19, 1423–1436, November 21, 2012 ª2012

phorylation, but did not alter total Smad1/5/8 levels, consistent

with inhibition of BMP-receptor signaling (Yu et al., 2008). Treat-

ment with 10 mM RepSox neither inhibited Smad1/5/8 phos-

phorylation nor changed total Smad1/5/8 levels. Interestingly,

Repsox augmented phosphorylation of Smad1/5/8.

Further analysis of TGFb/BMP/Smad signaling in HeLa cells

stimulated with TGFb or FBS (Figure 2B) confirmed that only

E738 strongly reduced both, phosphorylated and total Smad

levels. Using quantitative western blot analysis of total R-Smad

levels, we determined concentration/activity profiles for E738,

BIO, and GSK3b inhibitors, CHIR99021 (GSKi) and LiCl (Li

et al., 2011; Klein and Melton, 1996), in HeLa cells in presence

of BMP2, TGFb, or 10% FBS (Figure S2A). Whereas GSKi and

LiCl did not influence total R-Smad levels at all concentrations,

higher concentrations of BIO (5 and 10 mM) also mediated

depletion of R-Smads in presence of BMP2 and TGFb. In pres-

ence of FBS (10%) BIO up to 10 mM did not influence R-Smad

levels. IC50 values are summarized in Figure 2C. Depletion of

R-Smads by E738 occurred also in medium without FBS or

growth factors (Figure S2A).

To investigate the specificity of R-Smad depletion by E738, we

analyzed Smad4 and Smad7 protein levels. These inhibitory

Smads were not changed by E738 in HeLa cells in presence of

BMP2 or TGFb (Figure 2D). Thus, only R-Smad levels are sensi-

tive to E738 treatment.

The effect of E738 on BMP/TGFb target gene transcription

was evaluated by qRT-PCR (Figure 2E). Expression of ID1, a

well-established BMP target (Clement et al., 2000) regulating

cell differentiation by interaction with MyoD (Benezra et al.,

1990), was significantly repressed. Expression of ID2, another

BMP and TGFb signaling target (Clement et al., 2000), was

reduced to 50%. Expression of CTGF, a BMP antagonist and

TGFb agonist (Derynck and Zhang, 2003), and of SnoN, a

TGFb downstream gene (Deheuninck and Luo, 2009), was

decreased to �30% and 10%, respectively.

BMP2 target gene expression was also analyzed in compar-

ison with DM, RepSox, and BIO. The results obtained clearly

show that E738 inhibits both, BMP and TGFb-specific target

genes. DM only blocks expression of BMP target genes (Fig-

ure 2F). Treatment with RepSox enhanced expression of ID1

and ID2 and slightly decreased CTGF mRNA. The comple-

mentary activity of DM and RepSox fits with specific receptor

inhibition and opposing effects of BMP2 and TGFb signals.

Distinct from other inhibitors, BIO suppressed ID1 expression,

whereas ID2 and CTGF mRNA levels were clearly increased.

Taken together, the IRD E738 blocks TGFb/BMP/Smad

signaling through specific depletion of R-Smad pools without

affecting other Smad levels and thus uniquely modulates

TGFb/BMP/Smad signaling.

IRD E738 Is a Multiple Kinase InhibitorKinase Inhibitor Activity Is Indispensable

for R-Smad Depletion

IRDs have been reported to act as multiple kinase inhibitors

(Meijer et al., 2006). Based on the cocrystallization structure of

indirubin E231 in the ATP-binding pocket of CDK2, replacement

of C-atom at 70-position with N-atom should deprive IRDs of

protein kinase inhibitory activity (Hoessel et al., 1999; Eisenbrand

et al., 2008). The resulting 70-aza-indirubin showed no inhibitory

Elsevier Ltd All rights reserved

Figure 1. Impact of IRDs on the Activation of BMP-Signaling through Smad1/5/8

(A–C) Phospho-Smad1/5/8 ([(1(Ser463/465)/5(Ser463/465)/8(Ser426/428)) and total Smad1/5/8 levels (pan-Smad1/5/8) were detected by immunoblot

after 24 hr treatment with IRDs as indicated in HeLa cells with the stimuli of BMP2 (100 ng/ml) (A), with 10% FBS (B), or in HFF cells with 15% FBS (C).

(D) C2C12, HFF, and NIH 3T3 cells were treated with increasing concentrations of E738 in presence of BMP2 (100 ng/ml). DMSO was used as mock and b-Actin

as loading control. Forty micrograms of protein were loaded per lane.

(E) C2C12 cells were treated with 0.5 mMor 1 mME738 for 12 hr in the presence of BMP2 and analyzed by immunostaining with antibody against total Smad1/5/8

including DAPI staining of nuclei.

See also Figure S1.

Chemistry & Biology


Chemistry & Biology 19, 1423–1436, November 21, 2012 ª2012 Elsevier Ltd All rights reserved 1425

Figure 2. E738-Mediated Inactivation of BMP- and TGFb-Associated R-Smads

(A and B) Interference of E738 (1 mM) on TGFb/BMP/Smad signaling in comparison to BIO (5 mM), DM (1 mM), and RepSox (10 mM)with stimuli of BMP2 (100 ng/ml)

(A) or TGFb (20 ng/ml) (B), or FBS (10%) in HeLa cells for 24 hr treatment detected by immunoblot.

(C) IC50 values of compound concentrations depleting total R-Smad levels quantitatively analyzed by immunoblot from at least three independent experiments.

HeLa cells were treated with various concentration of E738, BIO, LiCl, or GSKi for 6 hr in presence of BMP (100 ng/ml), TGFb (20 ng/ml), 10% FBS, or without

growth factor.

(D) Smad4 and Smad7 protein levels were not influenced by E738 after 24 hr treatment.

Chemistry & Biology


1426 Chemistry & Biology 19, 1423–1436, November 21, 2012 ª2012 Elsevier Ltd All rights reserved

Chemistry & Biology


activity when screened against �200 protein kinases, although

70-aza-indirubin was clearly cytotoxic in tumor cells (Eisenbrand

et al., 2008). We used two 70-aza-indirubins E846 and E847

(structures in Figure S2B). These aza-IRDs neither inhibited

phosphorylation of Stat3, nor mediated depletion of R-Smads,

whereas counterpart IRDs (E728 and E804) did (Figure 2G).

Thus, inhibition of protein kinase activity could be indispensable

for R-Smad depletion.

E738-specific inhibition of protein kinases was analyzed

by in vitro inhibition of selected recombinant protein kinases

measuring residual activity after treatment with 1 mM E738 in

presence of 1 mM ATP (Cheng et al., 2010). E738 reduced the

residual activity of GSK3b to only 6% and significantly

decreased activity of CDKs (CDK1/CycB1, 16%; CDK2/CycA,

10%; CDK2/CycE, 8%; CDK4/CycD1, 23%; CDK5/p35, 16%).

E738 also inhibited kinases associated with Stat3 phosphory-

lation (JAK2, 44%; KIT, 44%; Src, 25%) and activity of b-Raf

was reduced to 45% (Figure S3A).

We next analyzed the effect of E738 on GSk3b/Wnt and

Stat3 signaling in the cellular context. Treatment of HeLa cells

with increasing concentrations of E738 for 1 hr (Figures 3A and

S3B) clearly showed a concentration-dependent decrease of

Stat3 phosphorylation and stabilization of nonphosphorylated

b-catenin (NP-b-catenin, Ser33/37/Thr41), which accumulates

upon GSK3b inhibition. E804 (5 mM) a known inhibitor of phos-

pho-Stat3 signaling was included as reference (Nam et al.,

2005).

Inhibition of GSK3b was also analyzed in HEK293 cells con-

taining a TopFlash-luciferase reporter for recording of Wnt/

b-catenin signaling (Xu et al., 2004). Treatment with E738 at

1 mM induced luciferase quite similar to 20 mM LiCl (Figure 3B).

E738 Shows a Distinct Gene Activation Profile

in Xenopus Embryos

We also analyzed expression of Wnt and TGFb/nodal target

genes in Xenopus embryos treated with E738 in comparison

with the GSK3b inhibitor LiCl and SB505124 (SB5), a selective

ALK4, 5, and 7 inhibitor (Klein and Melton, 1996; Skirkanich

et al., 2011). Treatment with E738 resulted in a very distinct

expression profile of Wnt and TGFb/nodal target genes (Fig-

ure 3C). Expression of the Wnt target genes nodal related 3

(nr3) and siamois (sia) (McKendry et al., 1997; Carnac et al.,

1996) was strongly induced by LiCl and also increased after

treatment with E738. In contrast to the LiCl-specific response,

the TGFb target genes brachyury (bra) and goosecoid (gsc)

(Skirkanich et al., 2011) were strongly repressed by E738. In

addition, expression of fox1e, which is repressed by nodal-

signaling (Luxardi et al., 2010), was upregulated by E738. These

findings are in good agreement with previous expression pro-

files from embryos treated with SB5 (Skirkanich et al., 2011)

and prove that E738 inhibits both GSK3b and TGFb/nodal

signaling in early Xenopus embryos.

(E) Expression of ID1, ID2,CTGF, and SnoN in HeLa cells treated with 1 mME738

value of nontreated cells and show the mean ± SD of quadruplicates of three ind

(F) Expression of ID1, ID2, andCTGF in HeLa cells treatedwith 1 mME738, 1 mMDM

before and show the mean ± SD.

(G) HeLa Cells were treated with IRDs or Aza-IRDs (1 mM for each) for 24 hr with

(Smad1/5/8 or Smad2/3).

See also Figure S2.


The above results characterize E738 as a potent kinase inhib-

itor, affecting GSK3b as well as Stat3 signaling, and indicate

that kinase inhibition might be indispensable for depletion of

R-Smads. Nevertheless, inhibition of BMP/TGFb and nodal

signaling is clearly distinct from effects observed with other

GSK3b inhibitors.

E738-Mediated R-Smad Depletion Is Independentof R-Smad ActivationIt had been shown that inhibition of GSK3b and CDKs is involved

in controlling turnover of phosphorylated R-Smads and termina-

tion of signaling (Fuentealba et al., 2007; Alarcon et al., 2009;

Sapkota et al., 2007). Because E738 inhibited GSK3b and

CDKs in vitro and showed Wnt-like signaling in reporter cells,

we investigated the influence of E738 on the duration of

Smad1/5/8 phosphorylation and on degradation of activated

p-Smad1/5/8. In a first experimental design HeLa cells were

stimulated with a pulse of BMP2 for 1 hr prior to E738 treatment

(Figure 3D, upper). Without E738, the BMP2 induced p-Smad1/

5/8 signal peaked 0.5 hr after removal of BMP2 (1.5 hr after

adding BMP2) and decreased in the next 1 hr; no change of

total Smad1/5/8 was observed (Figure 3D). When E738 was

added at the time point of BMP2 removal, p-Smad1/5/8 levels

did not decrease for �1.5 hr after removal of BMP2 (2.5 hr after

adding BMP2) (Figure 3D). In presence of E738, a slight reduc-

tion of total Smad1/5/8 was already detectable after 2 hr,

whereas the level of p-Smad1/5/8 was still higher than in control

without E738.

In the second experimental design, we pretreated HeLa cells

with E738 for 1 hr before adding BMP2, removed E738 after

3 hr, and continued BMP2 treatment for 2 hr. In a parallel time

course experiment cells were treated with DM instead of E738

(Figure S3C). E738 pretreatment enhanced Smad1/5/8 phos-

phorylation even before BMP2 was added, whereas BMP2-

induced phosphorylation was blocked, as with DM. After

removal of compounds, Smad1/5/8 phosphorylation recovered

in DM treated cells but not after E738 treatment.

Influence of E738 on Phosphorylation of Linker

Region of Activated Smad1/5/8

Phosphorylation of Smad1/5/8 at the C termini and the linker

region are a prerequisite for ubiquitination by Smurf1 E3 ligase.

Linker phosphorylation at S206 is crucial for this specific recog-

nition (Fuentealba et al., 2007; Alarcon et al., 2009, Zhu et al.,

1999). We therefore performed time course studies of Smad1/

5/8 phosphorylation at S206 with E738 (1 mM). In absence of

E738 (Figure 3E, left) maximum intensity of BMP-induced phos-

phorylation at C termini and S206 simultaneously appeared

after incubation for 1 hr and decreased afterward. In presence

of E738 (Figure 3E, right) concurrent C-tail and linker phos-

phorylation of Smad1/5/8 occurred as well but with 1 hr delay.

Similar stabilization of linker phosphorylation was observed in

for 24 hr in presence of BMP2 quantified by qRT-PCR. Data were normalized to

ependent experiments.

, 5 mMBIO, or 10 mMRepSox relative to BMP2 alone. Data were normalized as

BMP2 and analyzed by immunoblot on the level of p-Stat3 and total R-Smads


Figure 3. Regulation of Protein Kinases Involved in Smad-Signaling by E738 and Modulation of Target Gene Expression

(A) HeLa cells were treated with or without IRDs as indicated for 1 hr in the presence of BMP2 and protein levels were analyzed by immunoblot as labeled on

the panel.

(B) Induction of Wnt-luciferase reporter assay (TopFlash) in HEK293 cells measured online in living cells after 1, 2, and 3 days with 1 mM E738 or 20 mM LiCl

in presence of 10%, 0.5%, or without FBS. Luminescence counts are presented as mean ± SD of 24 measurements.

Chemistry & Biology



Chemistry & Biology


the experiment adding E738 after a 1 hr BMP2 pulse (Figure 3D).

This delay in degradation of phosphorylated R-Smads is in good

agreement with the effect of other GSK3b inhibitors, which

prevent degradation of activated R-Smads (Fuentealba et al.,

2007). In contrast, looking at total Smad1/5/8 levels, we found

a decrease as early as 30 min upon addition of E738. After 6 hr

the amount was markedly reduced, and after 24 hr total

Smad1/5/8 was no longer detectable (Figure 3E).

E738 Treatment Leads to MAPK Phosphorylation

BMPs can activate MAPKs including ERK, p38, and JNK in

a Smad-independent fashion to regulate various biological

processes, like osteoblast apoptosis (Hay et al., 2001) and

vascular development (Yang et al., 2007). We analyzed regula-

tion of MAPKs in E738-treated HeLa cells. Treatment with

E738 enhanced phosphorylation of p38 and JNK, but not

ERK1/2 in the presence of BMP2 after 1 hr (Figure 3A). E738-

mediated JNK and p38 phosphorylation significantly increased

after 6 hr (Figure 3E). The activation of p38 by E738 was further

confirmed in a dose-dependent analysis after 6 hr treatment

(Figure 3F).

Taken together, the results presented so far demonstrate

complex E738-orchestrated TGFb/BMP signaling: (1) inhibition

of GSK3b/CDKs stabilizing phosphorylated R-Smads in the

short-term range, (2) degradation of total R-Smads decreasing

signaling activity in the long-term range, and (3) phosphoryla-

tion of p38/JNK MAPKs to activate Smad-independent TGFb/

BMP signaling.

Role of Ubiquitination in E738-Mediated Depletionof R-Smad PoolsR-Smad Depletion by E738 Is Independent

of MAP-Kinase Signaling

As already mentioned, turnover of activated R-Smads is medi-

ated by phosphorylation of the linker region through MAP-

kinases p38 and JNK, which facilitates binding of the ubiquitin

ligase Smurf1 (Sapkota et al., 2007). Because E738 induced

phosphorylation of p38 and JNK (Figures 3A, 3E, and 3F), we

analyzed whether p38 or JNK are required for R-Smad degrada-

tion by E738. Coincubation of HeLa cells with E738 and MAP-

kinase inhibitors SP600125 (JNKi), SB203580 (P38i), and

U0126 (ERKi) alone or in multiple combinations did not attenuate

the level of phospho-Smad1/5/8 and did not block the reduction

of total Smad1/5/8 levels after 6 and 24 hr of BMP2 treatment

(Figures 4A and 4B) or when stimulated with TGFb (Figure S4A).

The specific phosphorylation of MAPKs was clearly reduced by

the corresponding inhibitors. Thus, MAPKs should not play a role

in E738-mediated degradation of R-Smads.

We also applied several other inhibitors in combination with

E738 (Figures 4C, 4D, S4B, and S4C). Cells treated with the

translation inhibitor cycloheximide (CHX) showed a weaker

signal for total Smad1/5/8, most likely due to reduced Smad

synthesis (Figures 4C and S4B), whereas DM even enhanced

E738-mediated total R-Smad reduction (Figures 4D and S4C).

(C) Expression of Wnt and nodal/TGFb target genes during gastrula in Xenopus la

normalized to ODC, in comparison to untreated DMSO control and show the me

(D–F) HeLa cells were treated with 1 mM E738 after a 1 hr pulse of BMP2 (D), or in

lysates were analyzed by immunoblot using specific antibodies as indicated.

See also Figure S3.


Effect of E738 Treatment on R-Smad Expression

The effect of CHX promoted us to investigate if E738 inter-

feres with Smad expression on the mRNA level. Quantitative

RT-PCR revealed that 24 hr E738 treatment repressed

Smad1 and Smad2 expression to 60% and 30%, respectively,

whereas Smad4 mRNA levels remained unchanged (Figure 4E).

Expression of Smad-associated E3 ligases, Smurf1 and

Smurf2, was also reduced to 50% (Smurf1) and 30% (Smurf2).

This effect of E738 on R-Smad transcription is not sufficient

to explain the strong reduction observed on the protein level.

Thus, major degradation must occur directly on the protein

level.

Proteasome Inhibition Prevents E738-Mediated

R-Smad Degradation

The ubiquitin-proteasome pathway is the major route of pro-

tein degradation. Treatment with the proteasome inhibitor

MG132 efficiently rescued E738-mediated R-Smad degrada-

tion in HeLa cells stimulated by BMP2 (Figures 4F, 4G, and

S4D), or TGFb (Figure S4E), and in HFF cells (Figure S4F).

Because MG132 counteracted E738 activity on R-Smad degra-

dation, we postulated that MG132 might also promote cell

survival in presence of E738. Cotreatment with MG132 improved

cell survival in presence of FBS, BMP2, or TGFb (Figure 5A).

Cleavage of PARP, a marker of apoptosis, was significantly

prevented by cotreatment compared to either E738 or MG132

alone (Figure 5B).

Looking at ubiquitination, we found that E738 changed the

size distribution of ubiquitinated proteins and ubiquitination

precursors also in the presence of MG132, showing a significant

reduction of lower molecular weight fragments (Figure 5C).

DM and LiCl did not influence the ubiquitination pattern (Fig-

ure 5C). Moreover, accumulation of p-Smad1/5/8 was observed

in presence of MG132 alone or in combination with E738, but

not with DM (Figures 4G and 5C).

Smad1/5/8-specific ubiquitination was analyzed by immuno-

precipitation of Smad1/5/8 following incubation with MG132

and E738 in HeLa cells. Cotreatment with E738 clearly enhanced

Smad1/5/8 ubiquitination, but did not change Smurf1 coimmu-

noprecipitation (Figure 5D).

These results illustrate that E738 does not inhibit BMP

receptor-mediated phosphorylation of Smad1/5/8 and that

ubiquitin-mediated degradation is required to remove both,

nonactivated and activated Smad1/5/8.

USP9x and USP34 Are Involved in Maintenance

of R-Smad Pools

Our results demonstrated that E738 is able to promote ubiquiti-

nation of R-Smads (Figure 5D) and thus interferes with TGFb/

BMP signaling. We next performed a small siRNA screen target-

ing 16 proteins, which are crucial in TGFb/BMP-, WNT-, Stat3-

signaling, or play a role in ubiquitination or apoptosis. For each

gene, two published siRNAs were used (Table S1).

Smurf1 is the ubiquitin ligase required for degradation of

activated R-Smads (Fuentealba et al., 2007). Because E738

evis analyzed by qRT-PCR. Fold changes were calculated from mRNA levels,

an ± SD of triplicate measurements.

a time course study (E), or with various concentrations of E738 for 6 hr (F). Cell


Figure 4. Cotreatment with the Proteasome Inhibitor MG132 Blocks E738-Induced R-Smad Degradation

(A and B) HeLa cells were incubated with E738 (5 mM) or in combination with SP600125 (JNKi, 10 mM), SB203580 (P38i, 10 mM), and U0126 (ERKi, 10 mM), or with

inhibitor combinations (JNKi+P38i, 2i) or (JNKi+P38i+ERKi, 3i) for 6 hr (A) or 24 hr (B).

(C) HeLa cells were stimulated with BMP2 for 24 hr in the presence or absence of E738 (1 mM), cycloheximide (CHX, 10 mM), or both.

(D) E738 at increasing concentrations as indicated in combination with DM (1 mM).

(E) The expression levels of Smad1, Smad2, Smad4, Smurf1, and Smurf2 were analyzed by qRT-PCR. Data were normalized to the value of nontreated cells,

show the mean ± SD of quadruplicates, and are representative of three independent experiments.

(F and G) BMP2-stimulated HeLa cells were treated with DM (1 mM), E738, MG132 (MG), or in combination as indicated for 24 hr (F) or 6 hr (G).

See also Figure S4.

Chemistry & Biology


clearly reduced total Smad1/5/8 levels in presence of DM (Fig-

ure 4D) and in absence of growth factors (Figures 2C and

S2A), and reduced expression of Smurf1 and Smurf2 (Figure 4E),

we assumed that Smurfs might not be involved in degradation

of nonactivated Smad1/5/8. Indeed, in Smurf1-depleted HeLa

cells the reduction of R-Smad degradation was insignificant

(Figure 5E).


TAB1 and TAB2 are key elements of Smad-independent

TGFb/BMP-signaling (Derynck and Zhang, 2003). Depletion of

TAB1 and TAB2 by siRNA significantly reduced phosphorylation

of p38 in HeLa cells, but did not influence total R-Smad levels

(Figure 5F). Thus, E738-mediated activation of MAPKs is

TAB1/2-dependent, whereas degradation through ubiquitination

of R-Smads is TAB1/2-independent.


Figure 5. Effects of E738 and MG132 on Cell Proliferation and Ubiquitination

(A) SRB assays were performed to analyze the inhibitory effect of E738 (1 mM) with or without MG (5 mM) on the proliferation of HeLa cells after 24 hr treatment in

the presence of FBS (10%), BMP2, or TGFb for 24 hr. Data were normalized to value of nontreated cells. Mean ± SD of quadruplicates representing three

independent experiments.

(B) Effect of E738 and MG132 on PARP cleavage. Cells were treated with BMP2 or TGFb in presence of E738 (1 mM) with or without MG (5 mM) for 24 hr.

(C) BMP2-stimulated HeLa cells were treated with DM (1 mM), E738 (1 mM), BIO (5 mM), MG (5 mM), LiCl (30 mM), or E738 (1 mM) in combination with MG as

indicated for 24 hr.

(D) Immunoprecipitation of 4 hr compound-treated HeLa cells with Smad1/5/8 antibody, analyzed by immunoblot for ubiquitin and Smurf1.

(E and F) HeLa cells were transfected with Smurf1 siRNA or random siRNA control and treated with E738 (1 mM) in presence of BMP2 for 24 hr (E), or with TAB1/2

siRNAs and then treated with E738 (5 mM) for 6 hr (F).

Chemistry & Biology


Ubiquitin proteases USP9x and USP34 were reported to con-

tribute to accumulation of b-catenin (Murray et al., 2004; Lui et al.,

2011). Moreover, USP9x had been shown to influence ubiq-


uitination of Smad4 to enhance TGFb/Smad signaling (Dupont

et al., 2009). RNA interference of USP34 in HeLa cells reduced

the level of total Smad1/5/8, but not Smad2/3 (Figures 6A and


Figure 6. Role of USP9x and USP34 in the Control of R-Smad Levels

(A and B) HeLa cells were transfected with increasing concentrations of USP34 siRNA (A) or USP9x siRNA (B). Protein levels were analyzed after 6 hr

BMP2-stimulation in presence or without with E738 (5 mM).

(C and D) Cell extracts of BMP2-stimuated HeLa cells with or without E738 were prepared at indicated time points and used for immunoblot, or for immuno-

precipitation with antibodies specific for USP9x (C) or Smad1/5/8 (D) antibody and subsequent immunoblot analysis.

(E) Schematic model of BMP signaling orchestrated by E738.

See also Figure S5.

Chemistry & Biology



Chemistry & Biology


S5A). In USP9x-depleted HeLa cells, both R-Smads were

reduced (Figures 6B and S5B). In direct comparison with E738

treatment, the effect of USP9x and USP34 depletion on R-

Smad degradation was less pronounced, which may be ex-

plained by limited efficiency of siRNA-mediated depletion of

USP9x and USP34. Interestingly, E738 treatment led to a sig-

nificant decrease of USP34 and USP9x (Figures 6A and 6B).

Further analysis of the effect of E738 onUSP9x andUSP34 levels

revealed that both were reduced as early as 2 hr after treatment

with E738 in presence of BMP2 (Figure 6C) and TGFb (Fig-

ure S5C). Coimmunoprecipitation with an USP9x-specific anti-

body and subsequent immunoblot for Smad1/5/8 showed a clear

interaction of Smad1/5/8 with USP9x (Figures 6C and S5C). The

reverse experiment, coimmunoprecipitation with a Smad1/5/8

antibody, confirmed USP9x and Smad1/5/8 interaction and re-

vealed interaction of USP34 with Smad1/5/8 (Figure 6D).

These experiments clearly demonstrate a direct control of

nonactivated R-Smad levels through ubiquitination, which

involves ubiquitin proteases, USP9x and USP34, to control

ubiquitination and prevent depletion of nonactivated R-Smads.

This mechanism is efficiently blocked by E738 and can explain

the initially observed inhibition of BMP and TGFb signaling

through this IRD.

DISCUSSION

TGFb signaling, including TGFb- and BMP-mediated pathways,

is a central component in the control of tissue formation, embry-

onic development, and maintenance of stem cells. IRDs have

been shown to prevent migration and invasion of tumor endothe-

lial cells in vitro and in vivo (Williams et al., 2011) and to generate

ESC-like cells (Chou et al., 2008). Because TGFb/BMP signaling

plays an important role in these events (Xu et al., 2008), we postu-

lated that IRDs influence TGFb/BMP signaling. Screening of our

own indirubin derivatives collection resulted in identification of

IRDs, in particular E738, which inhibited R-Smad signaling by

promoting R-Smad degradation in all of seven cell lines tested.

In the TGFb/BMP signaling cascade, Smad2/3 is phosphory-

lated by TGFb-receptor type 1, ALK4/5/7, whereas Smad1/5/8

is phosphorylated by ALK1/2/3/6 after activation by BMP (Mas-

sague et al., 2005). Cross-activation between these pathways is

possible (Valdimarsdottir et al., 2006; Goumans et al., 2002) and

even TGFb/BMP independent phosphorylation of R-Smads was

reported (Zhu et al., 2007). Treatment with receptor inhibitors,

like DM and RepSox, directly inhibited R-Smad phosphorylation.

In comparison, E738 showed a long-term and robust block of R-

Smad activation, including noncanonically activated R-Smads.

Like other IRDs, E738 is a potent protein kinase inhibitor that

clearly inhibits GSK3b, CDKs, and other kinases. The strong

inhibition of GSK3b, an activity shared by other IRDs including

BIO (Eisenbrand et al., 2004), is sufficient to explain stabiliza-

tion of phospho-Smad1/5/8 in the short term and fits well with

results by Fuentealba et al. (2007). Reduced expression of

Smurf1, required for ubiquitination and degradation of activated

Smad1/5/8, should further contribute to stabilization of phos-

phorylated R-Smads by E738. These mechanisms, however,

cannot explain the strong inhibition of TGFb/BMP signaling

observed with E738. Other GSK3b inhibitors, except for BIO in

some conditions, did not directly inhibit TGFb/BMP signaling.


BIO, however, is also an IRD and thus may share inhibitory activ-

ities with E738.

Inhibition of TGFb/BMP signaling through R-Smad depletion

was completely blocked by MG132, an inhibitor of the ubiqui-

tin-proteasome degradation pathway. Ubiquitin-mediated

degradation is the known mechanism for turnover of activated

R-Smads with Smurf1 as main E3 ligase (Fuentealba et al.,

2007; Zhu et al., 1999).

With our experiments, we clearly show that ubiquitination

also controls nonactivated R-Smad levels, which is independent

of Smurf1. Consistent with this, knock-down of Smurf1 only

rescued a trace amount of total Smad1/5/8, because activated

Smad1/5/8 were no longer degraded.

MAPKs p38 and JNK are known to mediate turnover of

R-Smads after activation by TGFb, BMP, and other growth

factors phosphorylating R-Smads in the linker region (Pera

et al., 2003; Sapkota et al., 2007). E738 also activated p38 and

JNK. Using small molecule MAP-kinase inhibitors and siRNA-

mediated knock-down of TAB1/2 we could exclude this pathway

from participation in the observed R-Smad degradation.

The proteasome inhibitor MG132 not only rescued R-Smads

from degradation, but also prevented E738-induced PARP

cleavage and apoptosis, linking E738 activity directly with the

ubiquitin-proteasome pathway. Indeed, two ubiquitin-specific

proteases, USP9x and USP34, were identified to play a role in

stabilization and degradation of R-Smads. Both were dramati-

cally repressed by E738 in HeLa cells as early as 2 hr after

treatment.

USP9x and USP34 had been reported to stabilize b-catenin

and enhance Wnt signaling (Murray et al., 2004; Lui et al.,

2011). Our results add an additional regulatory activity, control

of R-Smad levels, and reveal a direct interaction with R-Smads.

This provides, to our knowledge, a novel interaction between

Wnt and TGFb/BMP signaling. In fact, treatment of Xenopus

embryos with E738 induced Wnt/Gsk3b target genes and simul-

taneously blocked TGFb (activin/nodal) activity.

Taken together, our findings clearly show that E738 regulates

TGFb/BMP signaling at four regulatory levels: (1) stabilization

of receptor-activated R-Smads through inhibition of CDKs

and GSK3b, (2) TAB1/2-dependent activation of p38 and JNK

MAP-kinases, and (3) reduction of R-Smad levels by down-

regulation of R-Smad expression, and (4) increased ubiquitina-

tion of nonphosphorylated R-Smads by blocking USP9x and

USP34 activity (Figure 6E).

The intricate regulation of TGFb/BMP signaling by IRDs may

explain the versatile applications of IRDs, including a potential

benefit of IRDs in stem cell reprogramming, which had been

shown for BIO (Sato et al., 2004; Chou et al., 2008). Furthermore,

our results add an important part to the control of TGFb/BMP

signaling, namely the control of R-Smad abundance by ubiquiti-

nation and deubiquitination. Given that IRDs influence these

pathways in several distinct layers, our results suggest, that

E738 and other IRDs may lead to new therapies in regenerative

medicine, fibrosis, and chronic diseases.

SIGNIFICANCE

TGFb and BMP signaling pathways are crucial in the con-

trol of cell differentiation, embryonic development, and


Chemistry & Biology


proliferation. Overactivation of these signaling pathways is

associated with a number of diseases. It is well known that

TGFb/BMP receptor signaling is regulated at several levels.

In particular, duration of R-Smad activation is controlled by

ubiquitination of activated R-Smads, which leads to degra-

dation of transcriptionally active phosphorylated R-Smads.

So far, total R-Smad levels had not been considered as

an important regulatory element in TGFb/BMP signaling.

Here, we show drastic depletion of total R-Smad levels of

bothmajor TGFb-family signaling pathways. In recent years,

it had been shown that protein stability is not only controlled

at the level of ubiquitination but also by deubiquitination

through ubiquitin-specific proteases. Using the new indiru-

bin derivative, E738, we not only observed impaired TGFb/

BMP signaling but also a rapid decline of total R-Smad

levels. Analyzing the effects of E738 in detail, we found inter-

ference with TGFb/BMP signaling at clearly separated

levels: short term stabilization of R-Smad activation through

GSK3b inhibition, activation of noncanonical signaling

through MAP-kinases p38 and JNK, and strong depletion

of total R-Smads, through protein degradation and reduced

transcription. Using RNA interference and protein interac-

tion studies, we could clearly show that total R-Smad level

control could be linked to the deubiquitinating proteases

USP9x and USP34. Depletion of these enzymes by siRNA

led to a concomitant reduction of total R-Smad levels of

both mayor pathways, Smad1/5/8 for BMP-signaling, as

well as Smad2/3 for TGFb downstream signals. Our work

clearly demonstrates that TGFb/BMP-signaling is efficiently

regulated by the relative abundance of total R-Smad

pools. Furthermore, this mechanism is accessible for small

molecule intervention, which will lead to new options for

intervention with prosurvival and cell differentiation TGFb/

BMP signals.

EXPERIMENTAL PROCEDURES

Materials

IRDs were synthesized. Structures and purities were determined by 13C- and1H-nuclear magnetic resonance (NMR) spectroscopy and elemental analyses

(Eisenbrand et al., 2004; Hoessel et al., 1999; Cheng et al., 2010). Dorsomor-

phin (DM), LiCl, and RepSox were purchased from Calbiochem (Germany).

P38i (SB203580), JNKi (SP600125), ERKi (U0126), SB505124 (SB5), and

MG132 were obtained from Sigma-Aldrich (Germany). The Xenopus experi-

ments were performed according to the German animal protection laws and

approved by the government of the state of Baden-Wurttemberg.

Cell Culture

HeLa, HEK293, HaCat, and C2C12 cells were cultured in DMEM (GIBCO,

Germany) containing 10% FBS (PAA, Germany) and 1% Pen/Strep (GIBCO,

Germany). Jurkat cells were cultured in RPMI with 10% FBS and 1%

Pen/Strep. HFF cells were cultured in Dulbecco’s modified Eagle medium

(DMEM) containing 15% FBS and 1% Pen/strep. NIH 3T3 cells were cultured

in DMEM with 10% NCS (GIBCO, Germany) and 1% Pen/Strep. Cells were

maintained under 5% CO2 at 37�C in humidified atmosphere. After 24 hr pre-

incubation in serum-free medium, cells were treated with compounds and

proteins as detailed in the text. Oligonucleotides for siRNA interference were

synthesized using previously reported sequences (Table S1) and riboxx

siRNA-design (Riboxx, Radebeul, Germany). Transfections were performed

using the riboxxFECT reagent according to themanufacturer’s instructions fol-

lowed by 30 hr incubation in DMEM with 5% FBS. Before adding substances,

transfected cells were maintained in medium without FBS for 18 hr. Then cells


were treated with TGFb (20 ng/ml) or BMP2 (100 ng/ml) and compounds as

indicated in the text.

Immunoblot

Cell pellets were homogenized in lysis buffer containing 1 mM EDTA, 0.5%

Triton X-100, 5 mM NaF, 6 M Urea, 1 mM Na3VO4, 10 mg/ml Pepstatin,

100 mMPMSF, and 3 mg/ml Aprotinin in PBS. Forty micrograms of total protein

per sample were resolved on 10% SDS-PAGE and blotted to membrane

(Immobilon, Millipore, Germany) by semi-dry transfer. Antibodies against the

following proteins were used for immunoblotting: p-Smad1/5/8 (C-terminal),

p-Smad1 (S206), p-JNK, JNK, p-p38-MAPK, p38-MAPK, p-Erk1/2, Erk1/2,

p-Smad3, pan-Smad2/3, Smad4, p-Stat3, Stat3, non-p-b-catenin 41/37/33,

ubiquitin, USP9x, TAB1, TAB2 (all from Cell Signaling, NEB, Germany), b-actin

(Sigma Aldrich), Smad1/5/8 (Santa Cruz, Germany), Smad7 (Epitomics,

Biomol, Germany), and USP34 (Bethyl, Biomol, Germany). Corresponding

goat anti-mouse and goat ant-rabbit IgG/HRP conjugates were from

KPL (Germany). All antibodies were applied according to manufacturers’

recommendations. Immunoblots were developed with ECL solution (Pierce,

Germany) and images were recorded with a digital imaging system (LAS

3000, Fuji).

Immunostaining

Serum-deprived C2C12 cells were coincubated with BMP2 (100 ng/ml) and

E738 at 0.5 mM or 1 mM for 12 hr. Cells were fixed in 4% paraformaldehyde,

permeabilized at room temperature for 10 min using PBS containing 0.1%

Triton X-100, and afterward blocked 1 hr at room temperature in PBS con-

taining 5% goat serum and 0.1% Tween 20. Cells were incubated overnight

at 4�C with Smad1/5/8 antibody (Santa Cruz) and corresponding signals

were detected using FITC-conjugated secondary antibody (BD Bioscience,

Germany). Nuclear staining was performed with DAPI (1 mg/ml) for 5 min.

Immunoprecipitation

Immunoprecipitation was performed according to the manufacturers protocol

(Cell Signaling Technologies). Briefly, cells were lysed in lysis buffer (100 mM

TRIS/HCl pH 7.5, 150 mM NaCl, 1 mM MgCl2, 0.25% NP-40 substitution,

5 mM NaF, 1 mM Na3VO4, 10 mg/ml pepstatin, 100 mM PMSF, and 3 mg/ml

aprotinin) and incubated with rabbit serum (DAKO) for 1 hr on ice and

protein-A Sepharose beads (Upstate, Millipore) for 30 min. Supernatant was

incubated with anti-Smad1/5/8 (1:100, Santa Cruz) or anti-USP9x (1:100,

Cell Signaling Technologies,) antibodies overnight at 4�C. Immune complexes

were precipitated with protein-A Sepharose bead for 1 hr at 4�C and analyzed

by immunoblotting. Clean-Blot IP Detection Reagent (Thermo, Germany) was

used for background reduction.

Quantitative RT-PCR

Quantitative real-time (qRT) reverse-transcription-PCR was performed ac-

cording to manufacturer’s instructions (Light Cycler, Roche, Germany). Briefly,

total RNAwas isolated fromcells using TRIzol reagent. cDNAwasgeneratedby

reverse-transcription with AMV reverse transcriptase (Promega, Germany) of

equivalent quantities of RNA, and qRT-PCRwas performed using SYBRGreen

PCR master mix on Light Cycler 480 (Roche, Germany). The following primer

(Eurofins, Germany) pairs were used for the amplification of ID1, ID2, CTGF,

SnoN, Smad1, Smad2, Smad4, Smurf1, and Smurf2, respectively: (ID1) 5 s:

CATGAACGGCTGTTACTCAC; 3as: GTTCCAACTTCGGATTCCGAG; (ID2)

5 s:GACTGCTACTCCAAGCTCAAG; 3as: GTGATGCAGGCTGACAATAGTG;

(CTGF) 5 s: CCAGACCCACTATGATTAGAGC; 3as:GAGGCGTCATTGGTAAC;

(SnoN) 5 s: GCGAACCTGTACTTCTGTTCCTG; 3as:GAAATGCTGCTGGTC

AC; (Smad1) 5 s: GTGCTGGTTCCAAGACACAG; 3as ATTGGGAGAGTGAG

GAAACG; (Smad2) 5 s: AGAAGTCAGCTGGTGGGTCT; 3as: TGAGTGGTGA

GGCTTTCTC; (Smad4) 5 s: TGCCATAGACAAGGTGGAG; 3as: AGCCTCCC

ATCCAATGTTC; (Smurf1) 5 s: GATACCGGATACCAGCGTTTG; 3as:GGTTC

CTATTCTGTCTCGTG and (Smurf2) 5 s: ACAGAGGACAACGCAACAAG; 3as:

ATTACGGATCTCCCATCAG. b-Actin 5 s: CTGACTACCTCATGAAGATCCTC;

3as: CATTGCCAATGGTGATGACCTG was used as control.

Kinase Assay and Cell Proliferation Assay

A radiometric protein kinase assay (33PanQinase activity assay) was used for

measuring activity of protein kinases performed by ProQinase as previously


Chemistry & Biology


described (Cheng et al., 2010). Sulforhodamine B binding (SRB) assay was

used to evaluate cell proliferation (Cheng et al., 2010).

Analysis of Gene Expression Patterns in Xenopus laevis Embryos

Embryos were obtained by in vitro fertilization and cultured with inhibitors

at 14�C from stage 2 to stage 10.5 (early gastrula), RNA was isolated from

embryonic tissue, and cDNA synthesis was performed using random primers.

For each experiment, individual samples were analyzed in triplicate. Quantita-

tive PCR was performed in a q-TOWER (Jena Analytics) under conditions

described in Houston et al. (2002) and Skirkanich et al. (2011). Samples

were normalized to ornithine decarboxylase (ODC) and compared to appro-

priate DMSO/medium control. Primer sequences for qRT-PCR were: ODC:

Forward: 50-TGCACATGTCAAGC-30, Reverse: 50-GCCCATCACACGTT-30;fox1e: Forward: 50-CATGGAGCCCCAGATAAAAG-30, Reverse: 50-TTGGGTC

CAAGGTCCAATAA-30; bra: Forward: 50-TTCTGAAGGTGAGCATGTCG-30,Reverse: 50-GTTTGACTTTGCTAAAAGAGACAGG-30; goosecoid: Forward: 50-TTACCAATGAACAACTGGA-30, Reverse: 50-TTCCACTTTTGGGCATTTTC-30;nr3: Forward: 50-CCAAAGCTTCATCGCTAAAAG-30, Reverse: 50-AAAAGAAG

GGAGGCAAATACG-30; sia: Forward: 50-TCTGGTAGAACTTTACTCTGTTTT

GG-30, Reverse: 50-AACTTCATGGTTTTGCTGACC-30.

TopFlash-Luciferase Reporter Assay

Gsk3b inhibition in living cells was analyzed in Super TOP-Flash (STF) cells

(kindly provided by Jeremy Nathans, Johns Hopkins School of Medicine,

Baltimore, MD). STF cells are HEK293 cells stably transfected with a luciferase

gene containing seven LEF/TCF-binding sites in the promoter (Xu et al., 2004).

For on line measurement, cells were cultured in DMEM without phenol red

containing 10 mM HEPES buffer, 10% FCS, 4.5 g/l glucose and luciferin

(250 mM) in 96-well plates. Measurements were performed in a customized

TopCount plate reader (Perkin-Elmer) modified in a way that plates and

measurement chamber are kept constantly at 35�C ensuring cell viability

during measurement periods of several days. Substances were added at

beginning of measurements as indicated in the text.

Statistical Analysis

The statistical significance of compared measurements was assessed using

Student’s one-tailed or two-tailed t test (Microsoft Excel).

SUPPLEMENTAL INFORMATION

Supplemental Information includes one table and five figures and can be found

with this article online at http://dx.doi.org/10.1016/j.chembiol.2012.09.008.

ACKNOWLEDGMENTS

We thank Elke Lederer, Marie-Sofie Karipidis, and Holger Treiber for technical

assistance; Anna Ahomaa, Kristina Buchner, and Kristina Herold for help per-

forming experiments; Pavlo Holenya for suggestions on experiments and

methods; Peter Hortschansky for recombinant BMP2; Joachim Nickel for

recombinant TGFb; Steven Dooley for providing inhibitors and reagents; and

Jeremy Nathans for making available the Super-TopFlash cells. We thank

Christiane Peuckert for critical reading and corrections. We also thank Aris

Moutsakas and Carl-Hendrik Heldin for critical discussions and comments

on our manuscripts. This work was supported by BMBF MedSys, Drugs-iPS

(FKZ0315298[A,B]) coordinated by Ralf Mrowka.

Received: December 20, 2011

Revised: September 1, 2012

Accepted: September 6, 2012

Published: November 21, 2012

REFERENCES

Abreu, J.G., Ketpura, N.I., Reversade, B., and De Robertis, E.M. (2002).

Connective-tissue growth factor (CTGF) modulates cell signalling by BMP

and TGF-b. Nat. Cell Biol. 4, 599–604.


Alarcon, C., Zaromytidou, A.I., Xi, Q., Gao, S., Yu, J., Fujisawa, S., Barlas, A.,

Miller, A.N., Manova-Todorova, K., Macias, M.J., et al. (2009). Nuclear CDKs

drive Smad transcriptional activation and turnover in BMP and TGF-b path-

ways. Cell 139, 757–769.

Benezra, R., Davis, R.L., Lockshon, D., Turner, D.L., and Weintraub, H. (1990).

The protein Id: a negative regulator of helix-loop-helix DNA binding proteins.

Cell 61, 49–59.

Carnac, G., Kodjabachian, L., Gurdon, J.B., and Lemaire, P. (1996). The

homeobox gene Siamois is a target of the Wnt dorsalisation pathway and

triggers organiser activity in the absence of mesoderm. Development 122,

3055–3065.

Cheng, X., Rasque, P., Vatter, S., Merz, K.-H., and Eisenbrand, G. (2010).

Synthesis and cytotoxicity of novel indirubin-5-carboxamides. Bioorg. Med.

Chem. 18, 4509–4515.

Chou, Y.-F., Chen, H.-H., Eijpe, M., Yabuuchi, A., Chenoweth, J.G., Tesar, P.,

Lu, J., McKay, R.D.G., and Geijsen, N. (2008). The growth factor environment

defines distinct pluripotent ground states in novel blastocyst-derived stem

cells. Cell 135, 449–461.

Clement, J.H., Marr, N., Meissner, A., Schwalbe, M., Sebald, W., Kliche, K.-O.,

Hoffken, K., and Wolfl, S. (2000). Bone morphogenetic protein 2 (BMP-2)

induces sequential changes of Id gene expression in the breast cancer cell

line MCF-7. J. Cancer Res. Clin. Oncol. 126, 271–279.

De Robertis, E.M., and Kuroda, H. (2004). Dorsal-ventral patterning and neural

induction in Xenopus embryos. Annu. Rev. Cell Dev. Biol. 20, 285–308.

Deheuninck, J., and Luo, K. (2009). Ski and SnoN, potent negative regulators

of TGF-b signaling. Cell Res. 19, 47–57.

Derynck, R., and Zhang, Y.E. (2003). Smad-dependent and Smad-indepen-

dent pathways in TGF-b family signalling. Nature 425, 577–584.

Dupont, S., Mamidi, A., Cordenonsi, M., Montagner, M., Zacchigna, L.,

Adorno, M., Martello, G., Stinchfield, M.J., Soligo, S., Morsut, L., et al.

(2009). FAM/USP9x, a deubiquitinating enzyme essential for TGFbeta sig-

naling, controls Smad4 monoubiquitination. Cell 136, 123–135.

Eisenbrand, G., Hippe, F., Jakobs, S., and Muehlbeyer, S. (2004). Molecular

mechanisms of indirubin and its derivatives: novel anticancer molecules with

their origin in traditional Chinese phytomedicine. J. Cancer Res. Clin. Oncol.

130, 627–635.

Eisenbrand,G.,Merz, K.-H., andCheng, X. (2008). 7-Azaindirubins, 7’-azaindir-

ubins, 7,7’-diazaindirubin and the corresponding �30-oxime ether derivatives:

production thereof and use as a medicament thereof. Patentanmeldung

EP 08 022 288.8-2101.

Eivers, E., Demagny, H., and De Robertis, E.M. (2009). Integration of BMP and

Wnt signaling via vertebrate Smad1/5/8 and Drosophila Mad. Cytokine Growth

Factor Rev. 20, 357–365.

Fuentealba, L.C., Eivers, E., Ikeda, A., Hurtado, C., Kuroda, H., Pera, E.M., and

De Robertis, E.M. (2007). Integrating patterning signals: Wnt/GSK3 regulates

the duration of the BMP/Smad1 signal. Cell 131, 980–993.

Gellibert, F., Woolven, J., Fouchet, M.-H., Mathews, N., Goodland, H.,

Lovegrove, V., Laroze, A., Nguyen, V.-L., Sautet, S., Wang, R., et al. (2004).

Identification of 1,5-naphthyridine derivatives as a novel series of potent and

selective TGF-b type I receptor inhibitors. J. Med. Chem. 47, 4494–4506.

Goumans, M.-J., Valdimarsdottir, G., Itoh, S., Rosendahl, A., Sideras, P., and

ten Dijke, P. (2002). Balancing the activation state of the endothelium via two

distinct TGF-b type I receptors. EMBO J. 21, 1743–1753.

Hariharan, R., and Pillai, M.R. (2008). Structure-function relationship of inhibi-

tory Smads: Structural flexibility contributes to functional divergence. Proteins

71, 1853–1862.

Hay, E., Lemonnier, J., Fromigue, O., and Marie, P.J. (2001). Bone morphoge-

netic protein-2 promotes osteoblast apoptosis through a Smad-independent,

protein kinase C-dependent signaling pathway. J. Biol. Chem. 276, 29028–

29036.

Hoessel, R., Leclerc, S., Endicott, J.A., Nobel, M.E.M., Lawrie, A., Tunnah, P.,

Leost, M., Damiens, E., Marie, D., Marko, D., et al. (1999). Indirubin, the active

constituent of a Chinese antileukaemia medicine, inhibits cyclin-dependent

kinases. Nat. Cell Biol. 1, 60–67.



Chemistry & Biology


Hong, C.C., and Yu, P.B. (2009). Applications of small molecule BMP inhibitors

in physiology and disease. Cytokine Growth Factor Rev. 20, 409–418.

Houston, D.W., Kofron, M., Resnik, E., Langland, R., Destree, O., Wylie, C.,

and Heasman, J. (2002). Repression of organizer genes in dorsal and ventral

Xenopus cells mediated by maternal XTcf3. Development 129, 4015–4025.

Ichida, J.K., Blanchard, J., Lam, K., Son, E.Y., Chung, J.E., Egli, D., Loh, K.M.,

Carter, A.C., Di Giorgio, F.P., Koszka, K., et al. (2009). A small-molecule inhib-

itor of TGF-b signaling replaces sox2 in reprogramming by inducing nanog.

Cell Stem Cell 5, 491–503.

Klein, P.S., and Melton, D.A. (1996). A molecular mechanism for the effect of

lithium on development. Proc. Natl. Acad. Sci. USA 93, 8455–8459.

Li, J., Khavandgar, Z., Lin, S.-H., and Murshed, M. (2011). Lithium chloride

attenuates BMP-2 signaling and inhibits osteogenic differentiation through

a novel WNT/GSK3-independent mechanism. Bone 48, 321–331.

Li, W., Wei, W., Zhu, S., Zhu, J., Shi, Y., Lin, T., Hao, E., Alberto, H., Deng, H.,

and Ding, S. (2009). Generation of rat and human induced pluripotent stem

cells by combining genetic reprogramming and chemical inhibitors. Cell

Stem Cell 4, 16–19.

Lui, T.T., Lacroix, C., Ahmed, S.M., Goldenberg, S.J., Leach, C.A., Daulat,

A.M., and Angers, S. (2011). The ubiquitin-specific protease USP34 regulates

axin stability and Wnt/b-catenin signaling. Mol. Cell. Biol. 31, 2053–2065.

Luxardi, G., Marchal, L., Thome, V., and Kodjabachian, L. (2010). Distinct

Xenopus Nodal ligands sequentially induce mesendoderm and control gastru-

lation movements in parallel to the Wnt/PCP pathway. Development 137,

417–426.

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

Genes Dev. 19, 2783–2810.

McKendry, R., Hsu, S.C., Harland, R.M., and Grosschedl, R. (1997). LEF-1/

TCF proteins mediate Wnt-inducible transcription from the Xenopus nodal-

related 3 promoter. Dev. Biol. 192, 420–431.

Meijer, L., Skaltsounis, A.-L., Magiatis, P., Polychronopoulos, P., Knockaert,

M., Leost, M., Ryan, X.P., Vonica, C.A., Brivanlou, A., Dajani, R., et al.

(2003). GSK-3-selective inhibitors derived from Tyrian purple indirubins.

Chem. Biol. 10, 1255–1266.

Meijer, L., Guyard, N., Skaltsounis, A.L., and Eisenbrand, G., eds. (2006).

Indirubin, the Red Shade of Indigo (Roscoff, France: Life in Progress),

pp. 235–246.

Murray, R.Z., Jolly, L.A., and Wood, S.A. (2004). The FAM deubiquitylating

enzyme localizes to multiple points of protein trafficking in epithelia, where it

associates with E-cadherin and b-catenin. Mol. Biol. Cell 15, 1591–1599.

Nam, S., Buettner, R., Turkson, J., Kim, D., Cheng, J.Q., Muehlbeyer, S.,

Hippe, F., Vatter, S., Merz, K.-H., Eisenbrand, G., and Jove, R. (2005).

Indirubin derivatives inhibit Stat3 signaling and induce apoptosis in human

cancer cells. Proc. Natl. Acad. Sci. USA 102, 5998–6003.

Pardali, E., Goumans, M.J., and ten Dijke, P. (2010). Signaling by members of

the TGF-beta family in vascular morphogenesis and disease. Trends Cell Biol.

20, 556–567.

Pera, E.M., Ikeda, A., Eivers, E., and De Robertis, E.M. (2003). Integration of

IGF, FGF, and anti-BMP signals via Smad1 phosphorylation in neural induc-

tion. Genes Dev. 17, 3023–3028.

Pot, I., and Bonni, S. (2008). SnoN in TGF-b signaling and cancer biology. Curr.

Mol. Med. 8, 319–328.


Sapkota, G., Alarcon, C., Spagnoli, F.M., Brivanlou, A.H., and Massague, J.

(2007). Balancing BMP signaling through integrated inputs into the Smad1

linker. Mol. Cell 25, 441–454.

Sato, N., Meijer, L., Skaltsounis, L., Greengard, P., and Brivanlou, A.H. (2004).

Maintenance of pluripotency in human and mouse embryonic stem cells

through activation of Wnt signaling by a pharmacological GSK-3-specific

inhibitor. Nat. Med. 10, 55–63.

Skirkanich, J., Luxardi, G., Yang, J., Kodjabachian, L., and Klein, P.S. (2011).

An essential role for transcription before the MBT in Xenopus laevis. Dev. Biol.

357, 478–491.

Valdimarsdottir, G., Goumans, M.-J., Itoh, F., Itoh, S., Heldin, C.-H., and ten

Dijke, P. (2006). Smad7 and protein phosphatase 1a are critical determinants

in the duration of TGF-b/ALK1 signaling in endothelial cells. BMC Cell Biol. 7,

16–26.

Walsh, D.W., Godson, C., Brazil, D.P., and Marrtin, F. (2010). Extracellular

BMP-antagonist regulation in development and disease: tied up in knots.

Trend. Cell Biol. 20, 244–256.

Wharton, K., and Derynck, R. (2009). TGFbeta family signaling: novel insights

in development and disease. Development 136, 3691–3697.

Williams, S.P., Nowicki, M.O., Liu, F., Press, R., Godlewski, J., Abdel-Rasoul,

M., Kaur, B., Fernandez, S.A., Chiocca, E.A., and Lawler, S.E. (2011).

Indirubins decrease glioma invasion by blocking migratory phenotypes in

both the tumor and stromal endothelial cell compartments. Cancer Res. 71,

5374–5380.

Xu, Q., Wang, Y., Dabdoub, A., Smallwood, P.M., Williams, J., Woods, C.,

Kelley, M.W., Jiang, L., Tasman, W., Zhang, K., and Nathans, J. (2004).

Vascular development in the retina and inner ear: control by Norrin and

Frizzled-4, a high-affinity ligand-receptor pair. Cell 116, 883–895.

Xu, R.-H., Sampsell-Barron, T.L., Gu, F., Root, S., Peck, R.M., Pan, G., Yu, J.,

Antosiewicz-Bourget, J., Tian, S., Stewart, R., and Thomson, J.A. (2008).

NANOG is a direct target of TGFbeta/activin-mediated SMAD signaling in

human ESCs. Cell Stem Cell 3, 196–206.

Yang, X., Lee, P.J., Long, L., Trembath, R.C., and Morrell, N.W. (2007). BMP4

induces HO-1 via a Smad-independent, p38MAPK-dependent pathway in

pulmonary artery myocytes. Am. J. Respir. Cell Mol. Biol. 37, 598–605.

Yu, P.B., Hong, C.C., Sachidanandan, C., Babitt, J.L., Deng, D.Y., Hoyng, S.A.,

Lin, H.Y., Bloch, K.D., and Peterson, R.T. (2008). Dorsomorphin inhibits BMP

signals required for embryogenesis and iron metabolism. Nat. Chem. Biol. 4,

33–41.

Yuan, X., Wan, H., Zhao, X., Zhu, S., Zhou, Q., and Ding, S. (2011). Combined

chemical treatment enables Oct4-induced reprogramming from mouse

embryonic fibroblasts. Stem Cells 29, 549–553.

Zhen, Y., Sørensen, V., Jin, Y., Suo, Z., and Wied1ocha, A. (2007). Indirubin-30-monoxime inhibits autophosphorylation of FGFR1 and stimulates ERK1/2

activity via p38 MAPK. Oncogene 26, 6372–6385.

Zhu, H., Kavsak, P., Abdollah, S., Wrana, J.L., and Thomsen, G.H. (1999). A

SMADubiquitin ligase targets the BMPpathway and affects embryonic pattern

formation. Nature 400, 687–693.

Zhu, S., Li, W., Zhou, H., Wei, W., Ambasudhan, R., Lin, T., Kim, J., Zhang, K.,

and Ding, S. (2010). Reprogramming of human primary somatic cells by

OCT4 and chemical compounds. Cell Stem Cell 7, 651–655.

Zhu, S., Wang, W., Clarke, D.C., and Liu, X. (2007). Activation of Mps1

promotes transforming growth factor-b-independent Smad signaling. J. Biol.

Chem. 282, 18327–18338.


Documents

Chemistry & Biology ArticleIRDs have been reported to act as multiple kinase inhibitors (Meijer et al., 2006). Based on the cocrystallization structure of indirubin E231 in the ATP-binding