The Serine/Threonine Transmembrane Receptor ALK2 Mediates Müllerian Inhibiting Substance Signaling

Abstract

Müllerian inhibiting substance (MIS or anti-Müllerian hormone) is a member of the transforming growth factor-β family and plays a pivotal role in proper male sexual differentiation. Members of this family signal by the assembly of two related serine/threonine kinase receptors, referred to as type I or type II receptors, and downstream cytoplasmic Smad effector proteins. Although the MIS type II receptor (MISRII) has been identified, the identity of the type I receptor is unclear. Here we report that MIS activates a bone morphogenetic protein-like signaling pathway, which is solely dependent on the presence of the MISRII and bioactive MIS ligand. Among the multiple type I candidates tested, only ALK2 resulted in significant enhancement of the MIS signaling response. Furthermore, dominant-negative and antisense strategies showed that ALK2 is essential for MIS-induced signaling in two independent assays, the cellular Tlx-2 reporter gene assay and the Müllerian duct regression organ culture assay. In contrast, ALK6, the other candidate MIS type I receptor, was not required. Expression analyses revealed that ALK2 is present in all MIS target tissues including the mesenchyme surrounding the epithelial Müllerian duct. Collectively, we conclude that MIS employs a bone morphogenetic protein-like signaling pathway and uses ALK2 as its type I receptor. The use of this ubiquitously expressed type I receptor underscores the role of the MIS ligand and the MIS type II receptor in establishing the specificity of the MIS signaling cascade.

INTRODUCTION

Müllerian inhibiting substance (MIS), also known as anti-Müllerian hormone (AMH), is a member of the large transforming growth factor-β (TGFβ) superfamily that includes TGFβs, activins, and the bone morphogenetic proteins (BMPs) (1). While many TGFβ factors exert diverse effects in multiple tissues, the principle function of MIS is to induce regression of the Müllerian ducts during vertebrate male sexual differentiation (2, 3). Biologically active MIS is secreted from embryonic Sertoli cells of the testis and through unknown mechanisms is delivered to the paramesonephric region in the genital ridge. There, it binds the MIS type II receptor that is expressed in the mesenchyme surrounding the epithelial Müllerian duct. MIS signaling results in cell death of the Müllerian duct, thereby eliminating the female reproductive tract in males (4–6). In addition to this well defined role during male sexual differentiation, more subtle but important roles for MIS in somatic cell proliferation and gonadal germ cell maturation have been revealed by genetic experiments in mice (7, 8). For example, loss of function studies show that male MIS null mice develop Leydig cell hyperplasia (8). Careful analyses of adult MIS −/− ovaries also suggest that MIS-deficient ovaries exhibit enhanced primordial follicle recruitment eventually leading to premature ovarian failure (9). Finally, gain-of-function studies in mice in which MIS is overexpressed suggest a putative role for MIS in attenuating steroid production (7, 10). Presently, it is uncertain whether the same molecular machinery mediates all biological effects of MIS in both the embryo and adult. However, it is generally accepted that there is only one type II receptor for MIS in mediating Müllerian duct regression based on data obtained from both mouse and human mutants (11, 12).

All members of the TGFβ family presumably mediate their biological effects by two related transmembrane serine/threonine kinase receptors, type II and type I. Extensive studies with some members of the TGFβ family have established that upon ligand binding to type II receptors, such as the MISRII, type I receptors are recruited to form a heteromeric receptor complex. The type I receptors are then transphosphorylated by the type II receptor in the GS box, which is located close to the transmembrane region in the cytoplasm. Activation of the type I receptor leads to phosphorylation of receptor-specific Smad proteins. Phosphorylated Smads associate with the common Smad4 and are then translocated to the nucleus where they participate in regulating gene expression (13, 14). Selective recruitment of Smad proteins by ligand- receptor complexes suggests that two major signaling pathways exist for members of the TGFβ superfamily. In the first pathway, Smad2 or Smad3 are recruited by both TGFβ and activin through their respective type I receptors, ALK5 (TβRI) and ALK4 (ActRIB). In the second, Smad1, Smad5, or Smad8 are recruited by BMPs through ALK2, ALK3, and ALK6 or so-called BMP type I receptors (15, 16). Thus far, almost all TGFβ-like factors examined appear to use these two distinct intracellular signaling pathways.

The precise delineation of the MIS signaling pathway has been hampered because bona fide MIS target genes and a partner type I receptor for MIS have not been identified. One previous report proposed that ALK6 could serve as the MIS type I receptor based on in vitro data (17). However, mice deficient in ALK6 survive into adulthood and appear to show normal Müllerian duct regression (18, 19). Unfortunately, mouse knockouts of other type I receptors are uninformative because of embryonic lethality; the early developmental delays observed in ALK2, ALK3, and ALK4 knockouts (12, 20–22) preclude examination of MIS-triggered events, such as Müllerian duct regression. To address the molecular nature of the MIS signaling pathway, we surveyed different cell types and potential reporters that might be useful in dissecting the components of the MIS signaling pathway. These studies suggested that the mouse embryonic carcinoma cell line P19 and the mouse Tlx-2 homeobox gene promoter fused to a reporter gene would be useful in identifying type I receptor candidates. Here we report that nanomolar concentrations of bioactive MIS ligand induce a BMP-like signal through Smad1 and Smad5. Furthermore, our collective data obtained from cell studies, expression analyses, and antisense studies done in both cells and organ culture all strongly support the hypothesis that ALK2 is the type I receptor for MIS signaling in Müllerian duct regression and possibly in adult gonads.

RESULTS

A BMP Response Is Induced by Bioactive MIS Ligand and the MISRII

Attempts to identify a MIS type I receptor have been hampered by the lack of known MIS target genes. Moreover, MIS does not activate commonly used TGFβ-responsive reporter genes, such as 3TP-Lux or PAI-Luc, suggesting that MIS does not signal via the classic TGFβ pathway. To overcome these limitations, we investigated whether MIS induced a response in P19 mouse embryonic carcinoma cells, which have been shown to contain both BMP and TGFβ/activin signaling pathways. Two reporters were tested; the BMP responsive mouse Tlx-2 promoter reporter (TLX2-Lux) and the activin responsive Xenopus Mix.2 gene (A3-Lux) (23, 24). We observed significant induction (6-fold) of the Tlx-2 promoter after transiently transfecting P19 cells with the MISRII and MIS stimulation; this MIS-induced response was completely dependent on transfection of the MISRII (Fig. 1A). By contrast, MIS did not induce the activin-responsive A3-Lux reporter (Fig. 1A), whereas activin did (data not shown). MIS induction of the Tlx-2 promoter required the MISRII, whereas introduction of other type II receptors, such as BMPRII and ActRIIB or TβRII, failed to elicit MIS-dependent reporter activity (Fig. 1A). Transfection of a kinase-inactive MISRII (MISRII-KR), either alone or in the presence of the wild-type MISRII, completely inhibited MIS-induced Tlx-2 activation (Fig. 1B). Thus, when coexpressed with wild-type MISRII, the MISRII-KR mutant served as a dominant-negative mutant receptor. Our findings are consistent with a previous report showing a 2-fold activation of another BMP-type promoter (Xvent-Luc) in P19 cells upon MISRII transfection and MIS stimulation (17).

Using the TLX2-Lux reporter assay in P19 cells, we then determined the effective concentration of bioactive MIS ligand needed to activate this reporter. A saturable and dose-dependent profile of MIS-induced Tlx-2 reporter activity was observed with significant activity observed at subnanomolar levels. The estimated ED₅₀ (effective dose) for recombinant bioactive MIS (MIS-RARR) was determined to be approximately 18 nm MIS (Fig. 2); this value corresponds with the necessary concentration required for full Müllerian duct regression in organ explant culture assays of 15–20 nm (25). Taken together, these results demonstrate that MIS is capable of generating a BMP-like signal in P19 cells that does not require addition of type I receptors, but is solely dependent on the presence of the MISRII.

Dominant Negative Smad5, but not Smad2, Disrupts MIS Signaling

MIS-induced activation of a BMP-like signal suggests that MIS signaling is mediated through BMP receptor-specific Smads, such as Smad5, Smad1, or Smad8. Therefore, we investigated the effects of overexpressing representative Smads of either the BMP or TGFβ/activin signaling pathways. Increased reporter activity was observed after cotransfection of Smad5 or Smad1 and the MISRII (Fig. 3A and data not shown). In contrast, Smad2 failed to enhance MIS signaling above the baseline observed with transfection of only MISRII (Fig. 3A). To examine this issue further, we made use of a dominant-negative strategy to block endogenous Smad recruitment in P19 cells. Mutations of the carboxy-terminal phosphorylation sites of Smad proteins prevent dissociation from the activated receptor complex, thereby inhibiting activation (phosphorylation) of endogenous Smad proteins (26–28). We found that elevation of mutant Smad5 (Smad5–2SA), but not mutant Smad2 (Smad2–2SA), attenuated MIS-induced activation of TLX2-Lux reporter (Fig. 3B). Collectively, these data are consistent with the hypothesis that MIS signaling uses a BMP-like pathway.

ALK2 Is Involved in MIS Signaling

The results described above suggest that P19 cells contain an endogenous type I receptor capable of transducing MIS signaling when the MISRII is present. To test whether one of the previously identified type I receptors functions as a MIS type I receptor, multiple type I receptors (ALK1–6) were cotransfected with TLX2-Lux in P19 cells using lower concentrations of MISRII (1 ng/well) to observe additional activation. Under these conditions the strongest activation of the TLX2-Lux reporter was observed with a MISRII/ALK2 combination. Surprisingly, other BMP type I receptors known to signal through Smad5, such as ALK1, ALK3, and ALK6, failed to enhance the MIS response significantly, and ALK1 and ALK3 actually diminished MIS activation of the reporter (Fig. 4A). Western blot analysis revealed that each type I receptor was expressed at nearly equivalent levels (data not shown). That ALK2 participates in MIS signaling was further supported by using a dominant-negative or kinase-inactive ALK2 mutant receptor (ALK2-KR). Indeed, increasing amounts of ALK2-KR severely attenuated MIS-induced activation of TLX2-Lux (Fig. 4B). Although the dominant-negative ALK6-KR was previously reported to inhibit Xvent- reporter activity, in our hands we failed to observe inhibition of MIS-induced signaling with transfection of this mutant type I receptor (Fig. 4B and Ref. 17). Finally, increased activation of the MIS signaling by ALK2 was dependent on addition of biologically active MIS ligand and was not observed with either conditioned medium or inactive noncleavable MIS (MIS-RAGA, Fig. 4C).

ALK2 Is Expressed in All the MIS Target Tissues

Although our cell transfection studies suggested that ALK2 functions as a MIS type I receptor, supportive genetic data have been uninformative due to the early embryonic lethality observed in ALK2 null mice (21, 22). Therefore, we determined whether the expression profile of ALK2 coincides with the established expression patterns of MISRII. Based on previous studies regarding expression of the MISRII, we would expect ALK2 to be present in periductal mesenchymal cells surrounding the Müllerian duct as well as in both fetal and adult gonads. Indeed, we noted prominent expression of ALK2 in isolated urogenital ridges and gonads at several stages of embryonic development (Fig. 5, A and B). Interestingly, both ALK2 and MISRII levels diminished during the window of Müllerian duct regression, which commences at about E13 in male mice. Furthermore, similar to MISRII expression, ALK2 was easily detected in fetal gonads of both sexes, albeit at lower levels in fetal ovaries. By contrast, the other candidate MIS type I receptor, ALK6, is expressed at very low levels in both male and female urogenital ridges (Fig. 5A) and is barely detectable in embryonic gonads of both sexes (Fig. 5B). In situ hybridization analysis of E13 mouse lower body sections showed more precisely that ALK2 is expressed in mesenchymal cells adjacent to the concentric Müllerian duct (Fig. 5C). This result agrees with the expression pattern of the rat ALK2 (R1) in the genital ridge reported by Donahoe and co-workers (29). We also noted prominent ALK2 expression in the embryonic adrenal and liver (Fig. 5C). Given that the MISRII is also expressed in postnatal gonads (30–32), we determined the transcript levels of both ALK2 and ALK6 in these tissues. ALK6 mRNA expression was restricted to the ovary with virtually no signal observed in the testes; whereas ALK2 mRNA is expressed in both female and male gonads, albeit at a lower level in the testes (Fig. 5D). Taken together, our expression studies of ALK2 in both adult and embryonic reproductive tissues strongly support ALK2’s role as a MIS type I receptor.

Antisense ALK2 Blocks MIS Signaling in Cells and in Organ Cultures

To determine whether endogenous ALK2 in P19 cells is capable of conferring MIS-induced activity in the presence of the MISRII, an antisense approach was undertaken to disrupt ALK2 protein expression. P19 cells were treated with morpholino antisense oligomers 24 h before cotransfection with TLX2-Lux and MISRII. The efficiency of morpholino oligomers resides in their prolonged stability and the efficient delivery system [ethoxylated polyethylenimine solution (EPEI)] in which almost 100% of the cells transfected take up these morpholino oligomers (Ref. 33 and Fig. 6A). Treatment of cells with an antisense ALK2 oligomer greatly diminished MIS-induced activation of TLX2-Lux to near basal levels (Fig. 6B). Incubation of cells with the delivery solution without antisense oligomers or with a control antisense ALK2 oligomer showed no inhibition of MIS-induced activity. The ALK2 control oligomer was identical to ALK2 except for four nucleotide mismatches. In contrast to the marked inhibition observed with the antisense ALK2 oligo, no inhibition of TLX2-Lux was observed using an ALK6 antisense oligomer (Fig. 6B). However, treatment of P19 cells with the ALK6 antisense oligomer decreased BMP4 signaling by approximately 50% (data not shown). Collectively, our findings suggest that MIS signaling is mediated through an endogenous ALK2 type I receptor present in P19 cells.

To assess whether ALK2 mediates MIS signaling in vivo, a similar antisense approach was used in the Müllerian duct regression assay. In the absence of MIS, female urogenital ridges exhibit both the Müllerian (MD) and Wolffian (WD) ducts. In the presence of MIS, regression of the Müllerian duct occurs leaving only a visible Wolffian duct. Treatment of urogenital ridges with the antisense ALK2 oligomer partially or fully blocked MIS-induced regression (Fig. 6, left panel, as-ALK2). In contrast, treatment of the urogenital ridges with the control antisense ALK2 oligomer or the antisense ALK6 oligomer did not block MIS- induced regression (Fig. 6, m4-ALK2, as-ALK6). Furthermore, MIS-induced regression was unaffected by the special delivery reagent as judged by the normal loss of the Müllerian duct after incubation with MIS and the EPEI solution (Fig. 6, EPEI). In all cases, addition of morpholino oligomers alone had no effect on duct morphology (Fig. 6, right panel, no MIS, as-ALK2). These results strongly suggest that ALK2 mediates MIS-induced signaling in Müllerian duct regression.

DISCUSSION

The molecular signaling mechanisms of some TGFβ superfamily members have been studied in great detail, and it is now established that at least two distinct signaling pathways exist. The first of these is a Smad2 and 3 signaling pathway shared by TGFβ and activin, and the second is a Smad1 and 5 signaling pathway shared by the BMP family (14). Here, we report that MIS activates a BMP-like pathway using Smad1 or Smad5 and requires bioactive MIS and the MIS type II receptor. Moreover, our cellular studies and antisense experiments indicate that the type I receptor that bridges the MISRII and Smad intracellular signaling is ALK2.

The best known function of MIS is induction of Müllerian duct regression in male fetuses (2, 3) where MIS induces apoptosis in the epithelial layer of the Müllerian duct by activating a signaling cascade in the mesenchymal layer (4). Although Müllerian ducts normally differentiate in female mice because MIS synthesis is restricted to embryonic testes, female Müllerian ducts are competent to respond to MIS in vitro and in vivo (8, 25). Indeed, several groups have established expression of MISRII in the mesenchyme surrounding the Müllerian duct of both sexes (4–6, 32). Here we report that the expression of ALK2 correlates well with the known expression pattern of the MISRII and is found in the urogenital ridge of male and female fetuses. In situ hybridization analysis revealed expression of ALK2 in the mesenchymal cells surrounding the Müllerian ducts. Interestingly, Smad5 is also expressed at a relatively high level in the mesenchymal cells surrounding the Müllerian duct (34). Most important, we believe that ALK2 plays an important role in Müllerian duct regression based on the blockage of duct regression after ALK2 antisense treatment in organ cultures. However, it is clear from our study and previous reports that ALK2 is not restricted to reproductive tissues. Despite the role of ALK2 in MIS signaling, the specificity and propagation of this signaling pathway are conferred by the MIS ligand and the MIS type II receptor and their subsequent association.

Our studies demonstrate that MIS signaling in P19 cells is mediated through an endogenous type I receptor. Interestingly, P19 cells express multiple endogenous BMP type I receptors, such as ALK2, ALK3, and ALK6, as well as endogenous activin type I receptor (ALK4) (35). In these cells, the identity of this endogenous BMP type I receptor involved in MIS signaling is likely to be ALK2. Consistent with this hypothesis, MIS signaling increased after cotransfection of ALK2 and was decreased in either the presence of a dominant-negative ALK2 or after treatment with antisense ALK2 oligomers. Although we found that other type I receptors were unable to significantly enhance MIS responsiveness in P19 cells, we noted that ALK2, ALK5, and ALK6 are all capable of forming a receptor complex with MISRII in vitro (data not shown). Interestingly, formation of these receptor complexes was ligand-independent, and addition of ligand showed no further increases in type I receptor recruitment by the MISRII. The lack of ligand-induced ALK2 recruitment could suggest that very few MISRII/ALK2 receptor complexes are needed for signal transduction and thus these low levels would fall below the limits of detection. Alternatively, the MIS ligand may induce a conformational change in the MISRII/ALK2 receptor complex rather than simply increasing their association. Whether all of these MISRII-type I receptor complexes are biologically relevant or merely reflect overexpression in a heterologous system remains uncertain. Nonetheless, our functional data suggest strongly that a MISRII/ALK2 receptor complex is capable of transducing the MIS signal.

It is plausible that additional type I receptors or other plasma membrane proteins contribute to MIS signal transduction by their association with the MISRII/ALK2 receptor complex. For instance, ALK6 may indeed be part of the MIS signal transduction complex, as previously suggested, but at least in our system, ALK2 appears to be the essential signaling type I receptor. In contrast to a previous study, which suggested that MIS signals by recruiting ALK6 (17), we failed to observe any attenuation of MIS-induced signaling after overexpression of a dominant-negative ALK6-KR or treatment with antisense ALK6 oligomers. Potential reasons for this discrepancy may rest with the reporter used or the amounts of mutant type I receptor used in each study. In our hands, significant inhibition of the Tlx-2 reporter was noted at a ratio of 1:10, MISRII:ALK2-KR. Collectively our data suggest that a role for ALK6 in the MIS signaling cascade will be less straightforward than previously proposed.

While a clearly defined role for MIS in adult reproduction remains to be established, the apparent modulation of gonadal cell function by MIS infers the presence of an active signaling pathway(s) in these tissues. In addition to a developmental role for MIS in Leydig cell proliferation, several groups also report that MIS dampens steroidogenesis in the ovary and testis by repressing transcription of key steroidogenic enzymes (7, 10, 36, 37). Whether steroidogenic enzymes are directly downstream of MIS signaling is still uncertain, given that responses in steroidogenic cell lines are not robust and require MIS treatment for a minimum of 2 days (Ref. 36 and our unpublished results). Nevertheless, gonadal expression of both MISRII and ALK2 in the embryo and adult suggests that, similar to the urogenital ridge, a competent MISRII/ALK2-receptor complex is able to function in these reproductive tissues. Likewise, the strong ALK6 expression in the adult ovary (Fig. 5D) raises the possibility that a MISRII/ALK6 complex could function in the adult female. Clearly, identification of MIS target genes may elucidate the full physiological relevance of MIS in gonadal function and may also reveal its potential interplay with additional TGFβ signaling pathways in reproductive function.

Our findings imply that MIS signaling activates a BMP-like pathway through ALK2 and Smad5. Presently, our studies are unable to exclude the possibility that novel type I receptors play a role in MIS signaling. The ability to conditionally ablate ALK2 in mice in the genital ridge would circumvent the embryonic lethality at E9.5 during the early stages of gastrulation (21, 22) and should provide definitive data as to its role in mediating MIS signaling during Müllerian duct regression. Obviously, an ideal promoter to execute this in vivo genetic strategy is the MISRII promoter. It will also be of interest to determine whether similar factors known to modulate ligand availability and ligand binding (i.e. follistatin, inhibin binding proteins) also function in the MIS signaling pathway. Continued investigation of this signaling cascade should provide new insights into the molecular mechanisms of developmental and adult reproductive physiology.

MATERIALS AND METHODS

Plasmids

The full-length rat MISRII cDNA was isolated from E14 genital ridges by RT-PCR using the following primers: sense, 5′-GACGAATTCCTTTAGTAGGATGCTG-3′; antisense, 5′-CC G GTCGACGGACTTAGAGCCAGAGCC-3′. The sequence of this MISRII cDNA was verified by comparison to the published sequence (6) and subcloned into EcoRI/SalI sites of pRK5 containing a carboxy-terminal Flag epitope tag. Type I receptor expression vectors were provided by Dr. J. Massagué (Sloan Kettering, New York, NY; human ALK1, mouse ALK3, and mouse ALK6), Dr. R. Derynck (University of California, San Francisco, CA; mouse ALK2 and rat ALK5), and Dr. L. Mathews (University of Michigan, Ann Arbor, MI; human ALK4) (38–40). Kinase inactive mutants, MISRII-K228R, ALK2-K235R, and ALK6-K230R, were generated by PCR-based mutagenesis strategy. TLX2-Lux and A3-Lux reporter genes and human Smad1 and human Smad2 expression vectors were provided by Dr. J. Wrana (Samuel Lunenfeld Research Institute, Toronto, Ontario, Canada) (23, 24, 26, 41). Mouse Smad5 expression vector was provided by Dr. X. Wang (Duke University, Durham, NC) (42), and human Smad2-SA and mouse Smad5–2SA constructs were provided by Dr. P. ten Dijke (The Netherlands Cancer Institute, Amsterdam, The Netherlands) (28, 43). The human BMPRII, mouse ActRIIB, and human TβRII were provided by Dr. M. Kawabata (Cancer Institute, Tokyo, Japan) (44), Dr. L. Attisano (University of Toronto, Ontario, Canada) (45), and Dr. R. Derynck (46), respectively.

Cell Culture, Transfections, and Luciferase Assays

P19 cells were cultured in α-MEM containing 7.5% calf serum and 2.5% FBS. For MIS-induced luciferase assays, P19 cells were seeded at 20% confluency in 12-well plates and transfected with TLX2-Lux or A3-Lux reporter plasmid (100 ng/well) and the indicated receptor expression vectors (25 ng/well) using Fugene 6 transfection reagent (Roche Molecular Biochemicals, Indianapolis, IN). Twenty-four hours after transfection, cells were cultured for 2 h in medium containing 0.2% serum followed by 16 h treatment with the MIS ligand (15 nm). Luciferase was measured using the luciferase assay system (PharMingen, San Diego, CA) in a Monolight 2010 luminometer (Analytical Luminescence Laboratory, San Diego, CA). In all transfections,β -galactosidase expression plasmid (pCMV5-βGal) served as internal control to normalize for transfection efficiency.

MIS Preparation and Enzyme-Linked Immunosorbent Assay (ELISA)

Recombinant bioactive (MIS-RARR) and inactive (MIS-RAGA) conditioned medium were obtained from stably transfected HEK-293S cells expressing these rat MIS cDNAs as described previously (25). Conditioned medium of wild-type HEK-293S cells served as the control. Collected media were concentrated approximately 30-fold using a Centriprep system (Millipore Corp., Bedford, MA). The integrity of recombinant MIS protein was determined by Western blot analysis using purified antirat MIS antibody (25), and the amount of MIS was measured by an in-house ELISA assay. The standard curve was generated using purified MIS at concentrations ranging from 0.025–1μ g/ml. Purified MIS or MIS- conditioned medium were serially diluted in coating buffer (0.015 m Na₂CO₃, 0.035 m NaHCO₃, pH 9.6) overnight at 4 C in a 96-well Immulon plate (Nunc, Rochester, NY), followed by incubation with blocking buffer (3% BSA, 0.02% NaN₃ in PBS) for 2 h at room temperature. Plates were incubated overnight at 4 C with a polyclonal MIS rabbit antibody (1:250 dilution) (25). The following day, wells were washed with PBST (0.1% Tween in PBS) and incubated with biotinylated conjugated donkey antirabbit IgG (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA) (1:500 dilution) for 1 h at room temperature. After extensive washing with PBST, Streptavidin-peroxidase conjugate (Caltag Laboratories, Inc. Burlingame, CA) was added (1:500 dilution) for 30 min at room temperature. Wells were washed with PBST and incubated with BM blue (Roche Molecular Biochemicals) to visualize the complex. The reaction was stopped by adding 100 μl 1 m H₂SO₄, and the optical density was measured at 490 nm on a reader plate (BioTek, Winooski, VT). The molar concentrations of our recombinant HIS-tagged MIS were calculated based on an apparent molecular mass of 155 kDa as judged by SDS-PAGE in the absence of β-mercaptoethanol.

Morpholino Antisense Oligo Treatment and Regression Assays

Special delivery morpholino antisense oligos (33, 47) were designed for ALK2 (as-ALK2: 5′-CTCCATCGACCATTGTATAACC-3′) and ALK6 (as-ALK6: 5′-ATTTTCCAGAGC-TTCGTAAGAGCAT-3′) with guidance by Dr. Paul Marcos (Gene Tools LLC, Corvallis, OR). To create a control oligo, 4 mispairs were introduced into the antisense ALK2 oligo (4 m-ALK2: 5′-CTgCATgGACCATTGaATAtCC-3′). Antisense morpholino oligos were mixed with a delivery reagent, EPEI, according to the manufacturer’s directions (Gene Tools) for 20 min incubation at room temperature. A complete delivery solution containing 1.4 μm of oligo was added to P19 cells for 3 h and medium was replaced with fresh medium. Transfection and ligand treatment were performed (as described above) 24 h after antisense oligo treatment. Female rat urogenital ridges were dissected from E14.5 embryos and incubated for 3 h with the delivery solution containing 1.4 μm of morpholino antisense oligos (as described above). Treated organs were cultured on MilliCell-CM biopore membranes (Millipore Corp.) floated in 0.2 ml medium as previously described (4, 25), in the presence or absence of MIS (32 nm) for 3 days.

RNAse Protection Assay and in Situ Hybridization

To obtain embryos for RNA protection assays and in situ hybridization studies, pregnant FVB mice were killed by cervical dislocation on embryonic (E) day 13, E14, E15, or E18; vaginal plug detection was considered to be E0. Fetal tissues were isolated and snap frozen in liquid nitrogen and stored at −80 C. Fetal sex was determined by PCR using placental genomic DNA, and total RNA was isolated as previously described (48). An EcoRI-HindIII fragment containing 1–472 bp of the mouse ALK2 cDNA, and a HpaI-ApaI fragment containing 79–551 bp of the mouse ALK6, both encoding the extracellular domain, were subcloned in pBKS and used to generate[ ³²P]-UTP-labeled antisense RNA probes. Mouse MISRII and control glyceraldehyde 3-phosphate dehydrogenase (GAPD) antisense RNA probes were generated (6) and used in RNase protection assays using 10 μg of total RNA as described previously (48). In situ hybridization analyses was performed as described previously (6) on tissue sections fixed overnight in Bouin’s fixative, embedded in paraffin, and sectioned transversally at 8 μm. The same EcoRI-HindIII fragment of mouse ALK2 cDNA was used to generate sense and antisense[ ³⁵S]-UTP-labeled transcripts. Mice and rats were housed in accordance with NIH guidelines.

Note Added in Proof

Similar findings showing the use of ALK2 in the MIS signaling pathway are reported in the following article (pp. 946–959) by Clarke et al. (19A ).

Acknowledgments

We are especially grateful to all investigators for providing plasmids as noted in Materials and Methods. We also wish to thank Dr. Carol Basbaum for technical advice on aspects of this project.

This investigation was supported by Californian Division-American Cancer Society Fellowship 2–1-00 (to J.A.V.), The French ARC Foundation (Postdoctoral Fellowship to R.O.), and NIH-NICHD (Grant RO1, RCDA to H.A.I.).

References