Role of type I receptors for anti-Müllerian hormone in the SMAT-1 Sertoli cell line

Role of type I receptors for anti-Mu¨llerian hormone in the SMAT-1 Sertoli

cell line

Corinne Belville

_{, Soazik P Jamin}

_{, Jean-Yves Picard}

_{, Nathalie Josso}

_and

Nathalie di Clemente*

1_{Institut National de la Sante´ et de la Recherche Me´dicale, Unite´ 493 sur l’Endocrinologie du De´veloppement, Universite´ Paris XI, 32} rue des Carnets, 92140 Clamart, France

Anti-Mu¨llerian hormone (AMH) is a member of the transforming growth factor-b family responsible for regression of Mu¨llerian ducts during male sexual differ-entiation and for regulation of gonadal steroidogenesis. AMH is also a gonadal tumor suppressor which mediates its effects through a specific type II receptor andthe bone morphogenetic protein (BMP)-specific Smadproteins, suggesting that AMH andBMPs couldalso share type I receptors, namely activin-like kinases (ALKs)2, 3 or 6. However, attempts to identify a unique AMH type I receptor among them were unsuccessful. Here, using kinase-deficient type I receptors and small interfering RNA technology, we demonstrate that, in an AMH Sertoli target cell line, ALK3 mediates AMH effects on both Smad1 activation and P450 side-chain cleavage enzyme. In addition, transfecting a combination of normal andkinase-deficient receptors, we show that ALK2 can compensate for the absence of ALK3 andprobably acts in synergy with ALK3 at high concentrations of AMH to activate Smad1, whereas ALK6 has a competitive inhibitory effect. These results are a first step in under-standing how AMH transduces its effects in immature Sertoli cells.

Oncogene(2005) 24, 4984–4992. doi:10.1038/sj.onc.1208686; published online 16 May 2005

Keywords: anti-Mu¨ llerian hormone; Mu¨llerian inhibit-ing substance; transforminhibit-ing growth factor-b; receptors; activin-like kinases; Smad

Introduction

Anti-Mu¨llerian hormone (AMH), also called Mu¨llerian inhibiting substance (MIS), is a member of the transforming growth factor-b (TGF-b)family, produced exclusively by Sertoli and granulosa cells (reviewed in Teixeira et al., 2001). During male sexual differentiation,

AMH is responsible for the regression of Mu¨llerian ducts, the anlagen of uterus and tubes in females. AMH also regulates steroidogenesis in gonads of both sexes and can act as a gonadal tumor suppressor. Several AMH target genes have been identified in immature Sertoli cells, namely the P450 side-chain cleavage (P450scc)and aromatase enzymes, and the AMH type II receptor (Messika-Zeitoun et al., 2001). However, the mechanism that underlies the effect of AMH on these cells remains poorly understood.

TGF-b family members signal across the plasma membrane by inducing the formation of type I and type II serine/threonine kinase receptor complexes, which initiate intracellular signaling through phosphorylation of receptor-regulated Smads (R-Smad)proteins (Shi and Massague´, 2003). Once activated, R-Smads form heteromeric complexes with a common partner, Smad4, and accumulate in the nucleus, where they control gene expression. This signaling pathway can be regulated at different levels, in particular by a third group of Smad proteins, the inhibitory Smads (I-Smad). TGF-b family members can also activate Smad-independent signaling pathways.

Until recently, nothing was known of the signaling pathway downstream of the AMH type II receptor (AMHR-II). Since 2000, we and others have shown that AMH activates R-Smad1, 5 or 8 in all its target organs (Goue´dard et al., 2000; Clarke et al., 2001; Dutertre et al., 2001; Visser et al., 2001). These are the cytoplasmic effectors of the bone morphogenetic pro-teins (BMPs), suggesting that AMH and BMPs could also share type I receptors, namely activin-like kinases (ALKs)2, 3 or 6 (Lo et al., 1998). However, attempts to identify a unique type I receptor for AMH among these have been inconclusive (Josso and di Clemente, 2003). Here, we approach this issue by using a cell line, SMAT-1, derived from immature Sertoli cells by targeted oncogenesis (Dutertre et al., 1997). This cell line is particularly appropriate because, like its progenitor cells, it expresses both AMH and AMHR-II and is therefore a natural target of AMH. We investigate the effect of BMP type I receptors on Smad1 activation, P450scc and AMHR-II expression, and explore the possible existence of Smad-independent signaling path-ways in this cell line.

Received 13 April 2004; revised 2 March 2005; accepted 8 March 2005; published online 16 May 2005

*Correspondence: N di Clemente;

E-mail: nathalie.diclemente@inserm.ipsc.u-psud.fr

Results

Characteristics of activation/regulation of the Smad1 signaling pathway by AMH

We had shown that 1 h of treatment with 357 nM of AMH activated the Smad1, 5, 8 signaling pathway (also called Smad1 signaling pathway)in SMAT-1 cells (Goue´dard et al., 2000). Here, we determine the lowest concentration of AMH required for this activation, its kinetics and its regulation by the BMP-specific I-Smad, Smad6. The functional activation of Smad1 was assessed by a reporter system composed of a Gal4– Smad1 fusion protein and a Gal4-luc reporter construct (Clarke et al., 2001). We show that 3.6 nMof AMH can induce phosphorylation of proteins of the Smad1 group, their interaction with Smad4 and the expression of Smad6 (Figure 1a). This concentration of AMH also induces a 5677% stimulation of the Smad1 reporter system, which is not significantly modified at higher concentrations (Figure 1b). BMP2 stimulates this

reporter system by 5974% (result not shown).

Figure 1c shows that proteins of the Smad1 group are phosphorylated 30 min after treatment of SMAT-1 cells by AMH and can interact with Smad4 after 1 h of treatment, and that Smad6 appears in cell lysates after 4 h.

AMH does not activate the Wnt nor the NF-kB signaling pathways in SMAT-1 cells

As AMH induces b-catenin accumulation in peri-Mu¨llerian mesenchymal cells (Allard et al., 2000)and DNA binding of p50 and p65 NF-kB subunits in breast and prostatic cells (Hoshiya et al., 2003), we then asked whether AMH could activate the Wnt or NF-kB transduction pathways in SMAT-1 cells.

No accumulation of b-catenin was detected by immunocytochemistry in the nucleus of SMAT-1 cells after a treatment period with AMH ranging from 30 min to 24 h (results not shown). To confirm this result, we then used the reporter construct TOPflash in which expression of the luciferase gene is controlled by a synthetic promoter that contains binding sites for nuclear factors LEF/TCF. Transfection of a constitu-tively active form of b-catenin (CA b-cat), which cannot be phosphorylated and interacts with LEF/TCF in the absence of ligand, stimulates the luciferase activity of TOPflash but not that of FOPflash, in which LEF/TCF-binding sites have been mutated. Recombinant mouse Wnt-3a, a ligand of the Wnt family, also activates this reporter gene in SMAT-1 cells. In contrast, 12 (not shown)or 24 h of treatment with BMP2 or AMH have no effect on TOPflash (Figure 2a), even after transfec-tion of TCF-1 (not shown).

The activation of the NF-kB transduction pathway in SMAT-1 cells was first assessed with a pNF-kB-luc construct, in which expression of the luciferase gene is controlled by a synthetic promoter that contains binding sites for NF-kB nuclear factors. Transfection of SMAT-1 cells with the pFC-MEKK-positive control induces a

stimulation of pNF-kB-luc luciferase activity, which is inhibited in the presence of a dominant-negative form of IkBa. TNF-a also activates pNF-kB-luc, whereas AMH has no effect (Figure 2b). Using a nonradioactive NF-kB p50/p65 transcription factor assay which combines the

T 3.6 7.1 35.7 71.4 178.5 nM AMH Smad1P _{WB anti-Smad1P} Smad4 IP anti-Smad1 + WB anti-Smad4 Ig G AMH (nM) % stimulation Gal4-luc 0 50 100 150 200 0 25 50 75 100 WB anti-Smad1 WB anti-αtubulin Smad1 α tubulin WB anti-Smad6 Smad6 T 0.5h 1h 2h 4h 6h 8h 24h Smad1P Smad4 Ig G WB anti-Smad1P IP anti-Smad1 + WB anti-Smad4 Smad6 WB anti-Smad6 Smad1 IP anti-Smad1 + WB anti-Smad1 Ig G Smad1 IP anti-Smad1 + WB anti-Smad1 Ig G WB anti-Smad1 Smad1 WB anti-αtubulin α tubulin AMH a b c

Figure 1 Characteristics of activation of Smad1 signaling pathway

by AMH. SMAT-1 cells were (a)treated with increasing concentrations of AMH for 24 h and cell lysates were: successively subjected to SDS–PAGE and immunoblotting with phosphory-lated Smad1 antibody (WB anti-Smad1P), Smad1 antibody (WB anti-Smad1), Smad6 antibody (WB anti-Smad6) and a-tubulin antibody (WB anti-a-tubulin), or immunoprecipitated using Smad1 antibody (IP anti-Smad1), before SDS–PAGE and Western blotting with Smad4 antibody (WB anti-Smad4)and Smad1 antibody (WB anti-Smad1). (b)At 24 h after plating, SMAT-1 cells were transiently transfected with Gal4-Smad1 (1 mg), Gal4-luc (1 mg)and pRLTK (50 ng)constructs and cultured for 24 h in the presence of increasing concentrations of AMH. Results are expressed as a percentage of stimulation of Firefly luciferase activity by AMH compared to cells cultured in control medium.

Data are means7s.e.m. of three experiments, each performed in

triplicate. (c)SMAT-1 cells were treated during various periods of

time with AMH (35.7 nM), and 20 mg of cell lysates was

immunoblotted with either Smad1P, Smad1, Smad6 or a-tubulin antibodies. The remaining cell lysates were immunoprecipitated with the Smad1 antibody, followed by Western blotting with Smad4 and Smad1 antibodies

Arbitr a ry units AMH BMP2 CA ß-cat FOPflash TOPflash -+ + + -+ TOPflash + -+ -+ -+ + + + + -Wnt-3a - - - + - - - -0 2 4 NS NS *** Arbitra ry units NS AMH IκB_α pFC-MEKK + -+ -+ + pNF-κB-luc TNF-_α - - - -** -+ -0 50 100 150 + + + + + + pNF-κB-luc *** 0 0.02 0.04 0.06 0.08 0.1 O D 450 n m p50 p65 negat ive cont rol posi tive cont rol _TN F-α _1h 24h control _{AMH ne}ga tive cont rol posi tive cont rol _TN F-α 1h ₂₄h cont rol AMH 0.12 a b c

Figure 2 Absence of activation by AMH of the Wnt and NB-kB signaling pathways in SMAT-1 cells. Cells were transiently

transfected with pRLTK (50 ng)and (a)TOPflash (1 mg)or FOPflash (1 mg)and when indicated constitutively active catenin (CA b-cat)(1 mg), (b)pNF-kB-luc (500 ng)and when indicated pFC-MEKK positive control (200 ng)with or without dominant-negative IkBa

(200 ng). Cells were treated with AMH (35.7 nM), Wnt-3a (1.3 nM) , BMP2 (2 nM)or TNF-a (5.9 mM)for 24 h. Firefly luciferase activity

was measured in cell lysates and results were normalized to Renilla luciferase activity. Three independent experiments were performed.

Data are expressed as means7s.e.m. of a single representative experiment performed in triplicate. NS: not significant compared to cells

cultured in the absence of AMH, **Po0.01, ***Po0.001. (c)Cells were treated either with TNF-a (5.9 mM)during 1 h or with AMH

(350 nM)during 1 and 24 h. After nuclear extraction, NF-kB p50 and p65 subunits were assayed using an oligonucleotide containing

the DNA-binding sequence of NF-kB in an ELISA-based colorimetric reaction. Results are expressed as OD at 450 nm. The negative control was incubated with a nonspecific oligonucleotide and, for the positive control, TNF-a-treated HeLa whole extracts were used instead of SMAT-1 nuclear extracts

principle of electrophoretic mobility shift assay with that of enzyme-linked immunosorbent assay, we show that 1 h of treatment with TNF-a induces the binding of p50 and p65 subunits contained in SMAT-1 cell nuclear extracts, with an oligonucleotide specific of the con-sensus DNA-binding sequence for NF-kB, whereas AMH treatment has no effect after 1 and 24 h (Figure 2c).

The type I receptors ALK2 and ALK3 are predominant in SMAT-1 cells

We then studied the relative content of the mRNAs for the three type I receptors, ALK2, ALK3 and ALK6, in SMAT-1 cells by real-time PCR (Figure 3), using primers specific for their extracellular domain (Table 1). ALK2 and ALK3 are expressed predominantly, at a level comparable to that of AMHR-II; in contrast, the amount of ALK6 mRNAs is, respectively, four- and sixfold lower.

ALK3 mediates AMH activation of Smad1 and AMH repression of P450scc mRNAs

To identify which type I receptors mediate AMH activation of Smad1, we specifically blocked ALK2, ALK3 or ALK6 by transfection of their kinase-deficient version (ALK-KD). These receptors, whose ATP-bind-ing site is mutated, have been shown to exert a dominant-negative effect, by forming nonfunctional complexes with the type II receptor (Wieser et al., 1993). Figure 4a shows that activation of Smad1 by AMH, measured with the Gal4-luc reporter, is inhibited only by the kinase-deficient version of ALK3 and not by that of ALK2. Addition of ALK6-KD increases the stimulatory effect of AMH on Gal4-luc.

To confirm these results, we disrupted endogenous ALK3 mRNAs in SMAT-1 cells by transfecting a small interfering RNA (siRNA)specific of the mouse ALK3 (mALK3 siRNA). Figure 4b shows that mALK3 siRNAs disrupt 42.172.8% of ALK3 mRNAs, but have no effect on ALK2 and ALK6 mRNAs. An siRNA specific of the human ALK3 (hALK3 siRNA)has no significant effect on endogenous ALK3 mRNAs. The mALK3 siRNAs can reduce AMH stimulation of Gal4-luc by 31.4712.8% (Figure 4c).

Next, we tested whether ALK3 also mediated AMH effects on its target genes in SMAT-1 cells: P450scc enzyme and AMHR-II (Messika-Zeitoun et al., 2001). Real-time PCR experiments (Table 1)show that both genes are repressed by 54.3711.1 and 27.6712.2%,

respectively, after 24 h of treatment with AMH

(Figure 4d). Transfection of mALK3 siRNAs reduces AMH repression of P450scc mRNAs to 22.576.6%, but does not significantly inhibit AMH effect on AMHR-II mRNAs.

ALK2 can mediate AMH activation of Smad1, whereas ALK6 has an antagonistic effect

To test whether these results were due to the content of type I receptors in SMAT-1 cells, we co-transfected the normal versions of the three ALKs together with the

ALK2 ALK3 ALK6 AMHR-II

Relative expression

0.0 0.5 1.0 1.5

Figure 3 ALKs and AMHR-II expression in SMAT-1 cells. Total

RNAs were isolated from SMAT-1 cells after 24 h of culture. ALK2, ALK3, ALK6 and AMHR-II transcripts were quantified by real-time PCR and results are expressed as their relative expression. The murine housekeeping gene GAPDH was used for the normalization

Table 1 Primers and PCR conditions used in the study

Gene Sequence primers 50-30 Length (bp) Annealing

temperature (1C)

Polymerisation time (s)

mALK2 For ATG GCT TCC ACG TCT ACC AG 363 59 15

Rev TCC GCT AGA GTG CTG TCT CC

mALK3 For ACA TCA GAT TAC TGG GAG CC 360 54 15

Rev TCC GAC AAC ATT CTA TTG TC

mALK6 For ACT CAG TCA ACA ATA TCT GCA G 423 61 17

Rev TGC TCG ATC AAG TCT CTC AGG

mAMHR-II For TCC AGC TGG CAT CCT TTT GC 366 58 15

Rev AGT CAG TGC CAC AGG AAC AC

mP450scc For ACA CAG ACG CAT CAA GCA GC 256 59 11

Rev CTG CAT GGT CCT TCC AGG TC

mGAPDH For TGG ATC TGA CAT GCC GCC TG 264 59 11

Gal4-Smad1/Gal4-luc reporter system. Figure 5a shows that none increases AMH stimulation of luciferase activity. In contrast, ALK6 decreases this effect.

Then, we studied the rescue of AMH activation of Smad1 blocked by ALK3-KD. Figure 5b shows that ALK3 and ALK2 compensate more or less equally for this effect, whereas ALK6 increases the blocking effect of ALK3-KD. Next, we co-transfected the normal version of ALKs in cells blocked with both ALK2-KD and ALK3-KD. Figure 5c shows that, again, ALK2 and ALK3 can abolish the blocking effect of the kinase-deficient receptors, but, in this case, ALK3 is more effective than ALK2. Addition of ALK6 amplifies the blocking effect of the kinase-deficient receptors.

Finally, we tested the compensating effect of type I receptors when 350 nM instead of 7.1 nM of AMH is used. Figure 5d shows that, under these conditions, ALK3-KD does not significantly block AMH

stimula-tion of Gal4-luc. ALK2 or ALK3 increase and ALK6 reduces this stimulatory effect of AMH.

Discussion

Members of the TGF-b family acting through the Smad cytoplasmic effectors are known for their bifunctional tumor suppressor/oncogenic role depending on the state of tumorigenesis. Similarly, AMH can act as a tumor suppressor gene (Mishina et al., 1996; Hoshiya et al., 2003)although it is highly expressed in the cancer of granulosa cells (Long et al., 2000). AMH induces activation of the Smad1, 5, 8 signaling pathway and the inhibitory Smad, Smad6 (reviewed in Josso and di Clemente, 2003). Here we report that, in the SMAT-1 Sertoli cell line, the concentration of AMH required for

Mock ALK2-KD ALK3-KD ALK6-KD

0 50 100 150 % stimulation by AMH _NS *** *** WB anti-HA

-

+

-

+

-

+

-

+

ALK3 mRNAs ALK6 mRNAs P450scc mRNAs AMHR-II mRNAs

Gal4-luc Gal4-luc Control siR NAhALK 3 siR NA mA LK3 siR NA _Control siRN A hALK3 siRN A mA LK3 siRN A a b c d

Figure 4 The type I receptor ALK3 mediates AMH effects in SMAT-1 cells. (a)SMAT-1 cells were co-transfected with Gal4-Smad1

(1 mg), Gal4-luc (1 mg), pRLTK (50 ng) constructs and the kinase-deficient version of type I receptors (ALK-KD) (1 mg). Luciferase

activity was measured after 24 h of culture in the presence of AMH (7.1 nM). Results are expressed as a percentage of stimulation of

luciferase activity by AMH compared to cells cultured in control medium. (b)SMAT-1 cells were transfected with control, mALK3 or

hALK3 siRNAs at 10 nMfor 24 h. Total RNAs were harvested, ALK2, ALK3 and ALK6 transcripts were quantified by real-time

PCR. Results are expressed as the relative expression of ALK2, ALK3 and ALK6 mRNAs. (c)Cells were co-transfected with

Gal4-Smad1 (1 mg), Gal4-luc (1 mg), pRLTK (50 ng) constructs and control or mALK3 siRNAs (10 nM), and luciferase activity was measured

after 24 h of treatment with AMH. Results are expressed as a percentage of stimulation of luciferase activity by AMH compared to

cells cultured in control medium. (d)Cells were transfected with 10 nMof control or mALK3 siRNAs and treated with AMH (7.1 nM)

for 24 h. Total RNAs were extracted and analysed by real-time PCR for (left)the P450scc and (right)the AMHR-II mRNAs. Results are expressed as a percentage of inhibition of P450scc and AMHR-II mRNAs in the presence of AMH compared to control cells. Data

are the mean7s.e.m. of three independent experiments. NS: not significant compared to mock-transfected cells or to cells transfected

with control siRNAs, *Po0.05, ***Po0.001 4988

activation of the Smad1 signaling pathway, its kinetics and its regulation by Smad6 are similar to those of other members of the TGF-b family (Lagna et al., 1996).

Activation of the Smad1 signaling pathway implied that AMH exerts its effects through a BMP type I receptor (Lo et al., 1998). Indeed, biochemical, biolo-gical and genetic approaches have shown that ALK2, ALK3 and ALK6 can act as AMH type I receptors. However, none of these type I receptors fulfils all the characteristics of an AMH type I receptor (Josso and di Clemente, 2003). In the present study, we demonstrate that, in the SMAT-1 Sertoli cell line, AMH activation of Smad1 and repression of P450scc are mediated through ALK3. As in Mu¨llerian ducts (Jamin et al., 2002), endogenous ALK2 or ALK6 cannot compensate for the absence of ALK3 (Figure 4). They are likely engaged by BMP type II receptors (ten Dijke et al., 2003)expressed by SMAT-1 cells (Goue´dard et al., 2000)and by unknown ligands, probably BMPs, present in the serum added to the culture medium, and they phosphorylate part of the cellular pool of Smad1. In the presence of AMH, there is no competition between ALK3 and ALK2 for Smad1, as shown by the lack of effect of ALK2 (Figure 5a)or ALK2-KD (Figure 4a)on AMH activation of Smad1, and by the similar capacities of ALK2 and ALK3 to compensate for the blocking effect of ALK3-KD (Figure 5b). In contrast, there is a competition between ALK3 and ALK6 for Smad1; blocking of ALK6 increases AMH activation of Smad1 (Figure 4a)and addition of ALK6 decreases it under all conditions (Figure 5a–d).

These results suggest that endogenous ALK2 and ALK6 do not normally mediate AMH activation of Smad1 in SMAT-1 cells. However, it is possible that some Smad1-dependent effects are mediated exclusively by ALK3, whereas others could require the engagement of ALK2. Indeed, transfection of ALK2 restores AMH activation of Smad1 when ALK3 is no longer available (Figure 5b). Furthermore, ALK3-KD no longer blocks AMH activation of Smad1 in the presence of high concentrations of AMH, suggesting that another BMP type I receptor is engaged by AMH (Figure 5d). As, under these conditions, ALK2 amplifies the stimulatory effect of AMH on luciferase activity, whereas ALK6 has the opposite effect, this type I receptor is probably ALK2. In keeping with these results, Jamin et al. (2003) have recently shown that Mu¨llerian duct regress in mice in which ALK3 has been disrupted when AMH is overexpressed, suggesting that type I receptors other than ALK3 can transduce AMH signals. A synergistic effect of ALK2 with ALK3 or ALK6 has been shown for alkaline phosphatase induction in osteoblasts (Aoki et al., 2001). Is this hypothesis biologically relevant? Arango et al. (1999)have shown that Mu¨llerian duct regression requires less AMH than does degeneration of fetal ovaries. Furthermore, AMH concentration varies locally and during development. For example, during folliculogenesis, AMH concentration is high in the fluid from preantral and small antral follicles, and low in larger ones (Vigier et al., 1984).

Gal4-luc [AMH]=7.1nM

+ ALK2-KD + ALK3-KD +ALK2 +ALK3 +ALK6

% stimulation b y AMH 0 50 100 150 200 Mock *** *** *** ***

-Mock +ALK2 +ALK3 +ALK6

% stimulation b y AMH Gal4-luc [AMH]=7.1nM 0 50 100 150 200 NS _NS * Gal4-luc [AMH]=350nM + ALK3-KD

+ALK2 +ALK3 +ALK6 Mock % stimulation b y AMH 0 50 100 150 200 NS _* *** *

+ALK2 +ALK3 +ALK6

b

c

d

Figure 5 ALK2 can mediate AMH effect on Smad1 activation,

whereas ALK6 has an antagonistic effect. SMAT-1 cells

were transfected with Gal4-Smad1 (1 mg), Gal4-luc (1 mg), pRLTK (50 ng)vectors and 1 mg of the following constructs: (a) ALK2, ALK3 or ALK6 type I receptors, (b)ALK3-KD and either ALK2, ALK3 or ALK6, (c)ALK2-KD and ALK3-KD and either ALK2, ALK3 or ALK6, and treated for 24 h in the presence of

7.1 nM of AMH. (d)Cells were transfected with ALK3-KD and

either ALK2, ALK3 or ALK6, and treated for 24 h in the presence

of 350 nM of AMH. Results are expressed as a percentage of

stimulation of luciferase activity by AMH compared to cells

cultured in control medium. Data are means7s.e.m. of four

experiments, each performed in triplicate. NS: not significant, *Po0.05, **Po0.01, ***Po0.001

In SMAT-1 cells, ALK6 inhibits AMH activation of Smad1. However, ALK6 is the only type I receptor able to interact with AMHR-II in a CHO cell line permanently expressing high amounts of AMHR-II (Goue´dard et al., 2000). This might not be the case in SMAT-1 cells. It is quite possible that, depending on conditions, AMH can have access to different signaling receptors and that the presence or absence of accessory molecules (Massague´ and Wotton, 2000)determines the interaction of a given type I receptor in a given cell type. An attractive hypothesis is that, in SMAT-1 cells, AMH binds to AMHR-II and ALK3 (or ALK2)to activate Smad1-dependent effects, and perhaps to AMHR-II and ALK6 to induce putative Smad1-independent effects. Interestingly, ALK6 is required for activation of the Smad-independent PKC signaling pathway by BMP-2 in osteoblast cells (Hay et al., 2004).

Two other AMH transduction pathways have been identified, involving b-catenin/LEF-1 (Allard et al., 2000)or p50 and p65 NF-kB subunits (Hoshiya et al., 2003). None of these Smad-independent signaling path-ways seem to be activated by AMH in SMAT-1 cells (Figure 2), probably because they lack a component of the Wnt or the NF-kB signaling pathways. We have previously shown that SMAT-1 cells lack a factor, present in P19 cells, which is essential for activation of the Smad1-specific XVent2-luc reporter by AMH and BMP-2 (Goue´dard et al., 2000).

In conclusion, we have studied the respective role of BMP type I receptors in AMH activation of Smad1 in an AMH target cell, the SMAT-1 Sertoli cell line. We show that they have differential roles: (1)ALK3 is the main AMH type I receptor, and probably the natural one, (2)ALK2 can also mediate AMH effect on Smad1, probably in synergy with ALK3 at high AMH concentrations and (3)ALK6 has a competitive inhi-bitory role. These results are a first step in understand-ing how AMH transduces its effects in Sertoli cells.

Materials andmethods

Reagents and antibodies

Recombinant human AMH was produced in a Chinese hamster ovary cell line permanently transfected with the human AMH gene (Imbeaud et al., 1995). BMP2 was a kind gift of Biogen Inc. (Cambridge, MA, USA). Recombinant mouse Wnt-3a and human TNF-a were, respectively, obtained from R&D Systems (Minneapolis, MN, USA)and Upstate Biotechnology (Waltham, MA, USA). Rabbit polyclonal antibodies against Smad1 and phosphorylated Smad1 were from Upstate Biotechnology (Waltham, MA, USA), mouse monoclonal anti-Smad4 and anti-Smad1 and goat polyclonal anti-Smad6 were from Santa Cruz Biotechnology (Santa Cruz, CA, USA), mouse monoclonal anti-hemagglutinin (HA) (clone 12CA5)was from Roche Diagnostics (Indianapolis, IN, USA), mouse monoclonal anti-a tubulin (clone B-5-1-2)was from Sigma-Aldrich (St Louis, MO, USA)and peroxidase-labeled anti-mouse, anti-goat and anti-rabbit antibodies were from Jackson ImmunoResearch Laboratories (West Grove, PA, USA). All Smad1 antibodies recognize Smad1, Smad5 and Smad8 proteins.

Cell line and transfection

The mouse immature Sertoli cell line SMAT-1 was cultured in Dulbecco’s modified Eagle’s medium (Life Technologies, Rockville, MD, USA)as described (Dutertre et al., 1997), and easily transfected with the LipofectAMINE Plus reagent (Invitrogen, Carlsbad, CA, USA).

DNA constructs

The influenza virus HA-tagged ALKs and kinase-deficient mutant human ALK6-K231R (ALK6-KD) were provided by Dr J Massague´ (New York). Human ALK2 and ALK3 cDNAs from pCMV5 were cloned into BamHI/HindIII digested pALTER vector and used as templates for site-directed mutagenesis to generate kinase-deficient receptors, ALK2-K235R (ALK2-KD)and ALK3-K261R (ALK3-KD). Site-directed mutagenesis was performed according to the Altered Sites Mutagenesis System kit (Promega, Madison, WI, USA)using mutagenesis reverse primers 50_{-GGGAGGAGAAGATCCTCACGGCAACAT} TC-30_{for ALK2-KD and 5}0_{-TGGTAAAGAATACTCTCAC} CGCCACT-30_{for ALK3-KD. Mutated cDNAs were digested} by BamHI/HindIII and introduced in pCMV5 upstream of HA epitope tag. Constructs were checked by enzymatic digestion and DNA sequencing. Gal4-Smad1 and Gal4-luc plasmids were a gift from Dr A Atfi (Paris). Constitutively active b-catenin, TOPflash, FOPflash and TCF-1 constructs were provided by Dr G Rawadi (Paris). IkBa construct was a gift from Dr S Maheswaran (Boston). pNF-kB-luc and pFC-MEKK plasmids were from PathDetect in vivo Signal Transduction Pathway cis-Reporting System (Stratagene, La Jolla, CA, USA).

SiRNA synthesis

siRNAs were generated by in vitro transcription from DNA primers using the Silencer siRNA Construction Kit (Ambion, Austin, TX, USA)according to the manufacturer’s protocol. All primer sequences were subjected to BLAST searches to ensure that there were no matches with the known sequences of other genes. The 21 base primers used to generate the mouse ALK3 (GenBank Accession Number AY365062)specific siRNAs (mALK3-siRNA)were 50-AACTTTCGGT GAATCCTTGCA-30 (sense)and 50-AATGCAAGGATTCA CCGAAAG-30 (antisense), and 50-AATTTTCTGGTCG GAGTCTGA-30 (sense)and 50-AATCAGACTCCGACCA GAAAA-30 (antisense)to generate the human ALK3 (Gen-Bank accession number NM_004329)specific siRNAs (hAlk3-siRNA). They correspond to the positions 326–346 and 107– 127 relative to the start codon of mouse ALK3 and human ALK3, respectively. Control siRNAs were generated using nonspecific primers: 50 -AAACGTGACACGTTCGGAGAA-30 _{(sense)and 5}0_{-AATTCTCCGAACGTGTCACGT-3}0 (anti-sense). SMAT-1 cells were transfected with 10 nMsiRNAs and treated 24 h with 7.1 nMAMH.

RNA extraction and reverse transcription

Total RNAs were extracted as described previously (di Clemente et al., 1994), from cells cultured in 10 cm dishes. Reverse transcription was performed in a total of 20 ml, with the First Strand cDNA Synthesis Kit for RT–PCR (Roche Diagnostics, Indianapolis, IN, USA)using 1 mg RNA, AMV reverse transcriptase, and random primers p(dN)6 as

recom-mended by the manufacturer. 4990

Quantitative real-time PCR

Evaluation of gene expression levels was achieved by real-time quantitative PCR kinetics using the SYBR Green I chemistry. Real-time PCR was performed with 5 ml of appropriate diluted cDNA, 500 nM of forward and reverse specific primers for murine ALK2, ALK3, ALK6, AMHR-II, P450scc (Table 1), 3 mMMgCl2, 1 LightCycler FastStart Reaction Mix SYBR

Green I (Roche Diagnostics, Indianapolis, IN, USA)in a LightCycler (Roche Diagnostics, Indianapolis, IN, USA). The PCR protocol used an initial denaturating step at 951C for 10 min, followed by 45 cycles of 951C for 10 s, annealing temperature (Table 1)for 10 s, 721C for 11–17 s (Table 1), with a transition rate of 201C/s. Crossing point (CP)values were acquired by using the second derivative maximum method of the LightCycler software 3.3 (Roche Diagnostics, Indiana-polis, IN, USA). The specificity of the desired product was documented with the analysis of the melting curve. The melting curve was achieved by first cooling samples to 601C at a transition rate of 201C/s after 30 s of incubation, and a slow-heating step at a rate of 0.11C/s until a maximum temperature of 951C. The mix was next cooled at 401C for 1 min, at a transition rate of 201C/s. Quantification of gene expression is based on a standard curve for each target gene with known amounts of testicular cDNAs (diluted from 20 to 0.5 ng/ml), and is included in each LightCycler real-time PCR experiment. Relative gene expression was calculated as a ratio of each target gene concentration to housekeeping gene GAPDH concentration. Data are means7s.e.m. of at least three experiments.

Immunoprecipitation and Western blot analysis

Cells were washed and solubilized in 500 ml lysis buffer (20 mMTris, 150 mMNaCl, 1% (v/v)Triton, 1 mM phenylmethyl-sulfonyl fluoride, 1 proteinase inhibitor mixture (Sigma-Aldrich, St Louis, MO, USA). Protein concentration was determined using the BCA assay (Pierce Chemical Co., Rockford, IL, USA). Total immunoprecipitates or 20 mg of cell lysates were subjected to Western blot analysis after 7.5% SDS–PAGE (Bio-Rad Laboratories, Hercules, CA, USA) using the indicated primary antibodies (1 mg/ml)and perox-idase-labeled secondary antibodies at 1 : 5000 as described previously (Faure et al., 1996). Proteins were visualized with the ECL Plus Kit Detection System (Amersham Pharmacia Biotech, Piscataway, NJ, USA).

Luciferase assays

SMAT-1 cells were co-transfected with luciferase reporter and expression constructs, and 50 ng of pRLTK as a control for transfection efficiency. After 24 h of treatment in the presence of AMH (7.1 or 350 nM), BMP2 (2 nM), Wnt-3a (1.3 nM)or TNF-a (5.9 mM), cells were washed twice with PBS, and lysed for 20 min under rocking in 500 ml of passive lysis buffer (Promega, Madison, WI, USA). In all, 20 ml was analysed for Firefly and Renilla luciferase activity according to the manufacturer (Dual Luciferase kit, Promega, Madison, WI,

USA), using a Lumat LB 9507 luminometer (Perkin-Elmer, Norwalk, CT, USA). Except when indicated, results were expressed as a percentage of stimulation of Firefly luciferase activity (after normalization to Renilla luciferase activity)in the presence of AMH compared to cells cultured in control medium. Data are means7s.e.m. of at least three experiments, each performed in triplicate.

Transcription factor assay

SMAT-1 cells were grown to 70% confluence, then serum deprived during 1 h before treatment with TNF-a (5.9 mM) , or AMH (350 nM)during 1 h and 24 h. Cells were harvested after trypsination, washed with cold PBS 1 and nuclear extracts were obtained by the protocol recommended by the NF-kB transcription factor assay manufacturer (Chemicon interna-tional, Temecula, CA, USA). Briefly, cells were resuspended in a hypotonic lysis buffer (10 mM Hepes (pH 7.9), 1.5 mM MgCl2, 10 mM KCl, 0.5 mM DTT, 0.1% Triton X-100,

protease inhibitor cocktail). After centrifugation and collection of the cytoplasmic fractions, the nuclei were lysed and nuclear proteins were solubilized in the extraction buffer (20 mMHepes (pH 7.9), 1.5 mM MgCl2, 10 mM KCl, 0.42M NaCl, 0.2 mM

EDTA, 0.5 mM DTT, 1% Nonidet P-40, 25% glycerol, protease inhibitor cocktail). Protein content in nuclear extracts was determined by the BCA assay (Pierce Chemical Co.). A NF-kB p50/p65 transcription factor assay ELISA (Chemicon international, Temecula, CA, USA)was used to determine the presence of the p65 et p50 subunits in the nuclei. Briefly, 40 mg of nuclear proteins was incubated during 1 h with a double-stranded biotinylated oligonucleotide containing the NF-kB consensus-binding site (50_{-GGGACTTTCC-3}0_{)in the provided} transcription factor assay buffer. Then the mixture was loaded during 1 h in a streptavidin-coated plate at room temperature. The wells were rinsed with a washing buffer, and nuclear proteins were hybridized with either a NF-kB p65 or p50 antibody at a dilution of 1/1000 during 1 h. After rinsing, a final incubation with horseradish peroxidase (HRP)-conju-gated secondary antibody diluted at 1/500 resulted in a colorimetric reaction, which was stopped by 0.5MHCL and quantified at 450 nm with a reference wavelength of 650 nm. The specificity of the reaction was controlled by the use of nonspecific oligonucleotide.

Statistics

Results were analysed using Student’s t-test to determine the statistical significance between different conditions. NS: not significant, *Po0.5, **Po0.01, ***Po0.001.

Acknowledgements

We thank Drs J Massague´, A Atfi, G Rawadi and S Maheswaran for kindly providing DNA constructs. We are grateful to Nathalie Lede´e and Sylvie Dubanchet for advice on real-time PCR. We are grateful to the Association pour la Recherche sur le Cancer for its generous support to Nathalie di Clemente (grant no. 4253).

References

Allard S, Adin P, Goue´dard L, di Clemente N, Josso N, Orgebin-Crist MC, Picard JY and Xavier F. (2000). Development, 127, 3349–3360.

Aoki H, Fujii M, Imamura T, Yagi K, Takehara K, Kato M and Miyazono K. (2001). J. Cell Sci., 114, 1483–1489.

Arango NA, Lovell-Badge R and Behringer RR. (1999). Cell, 99, 409–419.

Clarke TR, Hoshiya Y, Yi SE, Liu XH, Lyons KM and Donahoe PK. (2001). Mol. Endocrinol., 15, 946–959.

di Clemente N, Wilson CA, Faure E, Boussin L, Carmillo P, Tizard R, Picard JY, Vigier B, Josso N and Cate RL. (1994). Mol. Endocrinol., 8, 1006–1020.

Dutertre M, Goue´dard L, Xavier F, Long WQ, di Clemente N, Picard JY and Rey R. (2001). Endocrinology, 142, 4040–4046.

Dutertre M, Rey R, Porteu A, Josso N and Picard JY. (1997). Mol. Cell. Endocrinol., 136, 57–65.

Faure E, Goue´dard L, Imbeaud S, Cate RL, Picard JY, Josso N and di Clemente N. (1996). J. Biol. Chem., 271, 30571–30575.

Goue´dard L, Chen YG, Thevenet L, Racine C, Borie S, Lamarre I, Josso N, Massague´ J and di Clemente N. (2000). J. Biol. Chem., 275, 27973–27978.

Hay E, Lemonnier J, Fromigue O, Guenou H and Marie PJ. (2004). J. Biol. Chem., 279, 1650–1658.

Hoshiya Y, Gupta V, Segev DL, Hoshiya M, Carey JL, Sasur LM, Tran TT, Ha TU and Maheswaran S. (2003). Mol. Cell. Endocrinol., 211, 43–49.

Imbeaud S, Faure E, Lamarre I, Mattei MG, di Clemente N, Tizard R, Carre´-Euse`be D, Belville C, Tragethon L, Tonkin C, Nelson J, McAuliffe M, Bidart JM, Lababidi A, Josso N, Cate RL and Picard JY. (1995). Nat. Genet., 11, 382–388.

Jamin SP, Arango NA, Mishina Y, Hanks MC and Behringer RR. (2002). Nat. Genet., 32, 408–410.

Jamin SP, Arango NA, Mishina Y, Hanks MC and Behringer RR. (2003). Mol. Cell. Endocrinol., 211, 15–19.

Josso N and di Clemente N. (2003). Trends Endocrinol. Metab., 14, 91–97.

Lagna G, Hata A, Hemmat-Brivanlou A and Massague´ J. (1996). Nature, 383, 832–836.

Lo RS, Chen YG, Shi YG, Pavletich NP and Massague´ J. (1998). EMBO J., 17, 996–1005.

Long WQ, Ranchin V, Pautier P, Belville C, Denizot P, Cailla H, Lhomme´ C, Picard JY, Bidart JM and Rey R. (2000). J. Clin. Endocrinol. Metab., 85, 540–544.

Massague´ J and Wotton D. (2000). EMBO J., 19, 1745–1754. Messika-Zeitoun L, Goue´dard L, Belville C, Dutertre M, Lins L, Imbeaud S, Hughes IA, Picard JY, Josso N and di Clemente N. (2001). J. Clin. Endocrinol. Metab., 86, 4390–4397.

Mishina Y, Rey R, Finegold MJ, Matzuk MM, Josso N, Cate RL and Behringer RR. (1996). Genes Dev., 10, 2577–2587. Shi YG and Massague´ J. (2003). Cell, 113, 685–700.

Teixeira J, Maheswaran S and Donahoe PK. (2001). Endocr. Rev., 22, 657–674.

ten Dijke P, Korchynskyi O, Valdimarsdottir G and Goumans MJ. (2003). Mol. Cell. Endocrinol., 211, 105–113.

Vigier B, Picard JY, Tran D, Legeai L and Josso N. (1984). Endocrinology, 114, 1315–1320.

Visser JA, Olaso R, Verhoef-Post M, Kramer P, Themmen APN and Ingraham HA. (2001). Mol. Endocrinol., 15, 936–945.

Wieser R, Attisano L, Wrana JL and Massague´ J. (1993). Mol. Cell. Biol., 13, 7239–7247.