A Covalently Dimerized Recombinant Human Bone Morphogenetic ...

REPRODUCTION-DEVELOPMENT

A Covalently Dimerized Recombinant Human Bone Morphogenetic Protein-15 Variant Identifies Bone Morphogenetic Protein Receptor Type 1B as a Key Cell Surface Receptor on Ovarian Granulosa Cells Minna M. Pulkki,* David G. Mottershead,* Arja H. Pasternack, Pranuthi Muggalla, Helen Ludlow, Maarten van Dinther, Samu Myllymaa, Katri Koli, Peter ten Dijke, Mika Laitinen, and Olli Ritvos Department of Bacteriology and Immunology (M.M.P., D.G.M, A.H.P., P.M., S.M., M.L., O.R.), Haartman Institute, University of Helsinki and HUSLAB, University Central Hospital of Helsinki, FIN-00029 Helsinki, Finland; Robinson Institute (D.G.M.), Research Centre for Reproductive Health, School of Paediatrics and Reproductive Health, Medical School, The University of Adelaide, Adelaide 5005, Australia; Centre for Proteins and Peptides (H.L.), School of Life Sciences, Oxford Brookes University, Headington, Oxford OX3 0BP, United Kingdom; Department of Molecular Cell Biology and Centre for Biomedical Genetics (M.v.D., P.t.D.), Leiden University Medical Center, 2333 ZA Leiden, The Netherlands; Molecular Cancer Biology Program and Haartman Institute (K.K.), University of Helsinki, Helsinki, FIN-00014 Helsinki, Finland; and Biotechvisions Ltd. (M.L.), FIN-00270 Helsinki, Finland

Genetic studies have identified bone morphogenetic protein-15 (BMP15) as an essential regulator of female fertility in humans and in sheep. Oocyte-derived BMP15 is a noncovalently linked dimeric growth factor mediating its effects to ovarian somatic cells in a paracrine manner. Although receptor ectodomains capable of binding BMP15 have previously been reported, no cell surface receptor complex involved in BMP15 signaling has previously been characterized. Here we have expressed and purified recombinant human BMP15 noncovalent and covalent dimer variants. The biological effects of these BMP15 variants were assessed in cultured human granulosa-luteal cells or COV434 granulosa cell tumor cells using BMP-responsive transcriptional reporter assays and an inhibin B ELISA. Biochemical characterization of ligand-receptor interactions was performed with affinity-labeling experiments using [125I]iodinated BMP15 variants. Both ligand variants were shown to form homodimers and to stimulate Smad1/5/8 signaling and inhibin B production in human granulosa cells in a similar manner. [125I]Iodination of both ligands was achieved, but only the covalent dimer variant retained receptor binding capacity. The [125I]BMP15S356C variant bound preferentially to endogenous BMP receptor 1B (BMPR1B) and BMPR2 receptors on COV434 cells. Binding experiments in COS cells with overexpression of these receptors confirmed that the [125I]BMP15S356C variant binds to BMPR1B and BMPR2 forming the BMP15 signaling complex. The results provide the first direct evidence in any species on the identification of specific cell surface receptors for a member of the GDF9/BMP15 subfamily of oocyte growth factors. The fact that BMP15 uses preferentially BMPR1B as its type I receptor suggests an important role for the BMPR1B receptor in human female fertility. The result is well in line with the demonstration of ovarian failure in a recently reported human subject with a homozygous BMPR1B loss-of-function mutant. (Endocrinology 153: 1509 –1518, 2012)

ISSN Print 0013-7227 ISSN Online 1945-7170 Printed in U.S.A. Copyright © 2012 by The Endocrine Society doi: 10.1210/en.2010-1390 Received December 2, 2010. Accepted December 28, 2011. First Published Online January 31, 2012

* M.M.P. and D.G.M. contributed equally to this study. Abbreviations: BMP15, Bone morphogenetic protein-15; BMPR2, BMP receptor 2; BS3, bis(sulfosuccinimidyl) suberate; DSS, disuccinimidyl suberate; DTT, dithiothreitol; ECD, ectodomain; GDF9B, growth differentiation factor-9B; hBMP15, human BMP15; hBMP15wt, hBMP15 wild type; hGL, human granulosa luteal; mAb-28, monoclonal antibody 28.

Endocrinology, March 2012, 153(3):1509 –1518

endo.endojournals.org

1509

1510

Pulkki et al.

Covalent BMP15 Dimer Bioactivity/Receptor Binding

one morphogenetic protein-15 (BMP15), also known as growth differentiation factor-9B (GDF9B) (1–3), and GDF9 (4) are two closely related members of the TGF-␤ superfamily. In recent years, a number of genetic mutations have been identified showing the physiological relevance of these two oocyte-derived growth factors (5– 12). BMP15-null mice display cumulus cell dysfunction, whereas GDF9-null mice display failed folliculogenesis (13, 14), and in humans, abnormal expression of GDF9 has been linked to polycystic ovarian syndrome (15). Although mice can reproduce without BMP15, it is an essential factor for both human (6 – 8, 11) and sheep (5, 9, 10, 12) fertility. In general, sheep heterozygous for mutations in GDF9 or BMP15 have a higher mean ovulation rate than their respective wild-type controls, whereas sheep homozygous for the BMP15 mutation are infertile due to follicular growth being impaired from the primary stage of ovarian follicle development (5, 9, 10, 12). Ligands of the TGF-␤ superfamily are synthesized as proproteins, which are proteolytically processed to a Cterminal biologically active mature region, and a larger N-terminal proregion (16). The mature regions of the TGF-␤ superfamily share six conserved cysteine residues giving rise to three disulfide bonds that stabilize the folding of the mature region monomer, forming the so-called cysteine knot structure characteristic of the superfamily (17). A seventh conserved cysteine is involved in intersubunit disulfide bonding, giving rise to a covalent dimer, characteristic of the bioactive protein. Members of the TGF-␤ superfamily act on target cells via binding to specific type I and type II receptors, which exhibit Ser/Thr kinase activity. The human genome encodes seven type I receptors (ACVRL1, ACVR1, BMPR1A, ACVR1B, TGFBR1, BMPR1B, and ACVR1C) and five type II receptors (TGFBR2, BMPR2, ACVR2A, ACVR2B, and AMHR2) that are paired in different combinations as receptor complexes for the various members of the TGF-␤ superfamily (16). Membrane-anchored proteoglycans, known as type III receptors (also known as betaglycan or endoglin), aid this process by capturing TGF-␤ for presentation to the signaling receptors I and II (16). After ligand binding, the activated type I receptor further activates downstream effector molecules known as Smads (16). Of the type I receptors, it has been shown that several BMP bind to type I activin A receptor, BMP type 1A receptor, or BMP type 1B receptor leading to phosphorylation of Smads 1, 5, and 8, whereas TGF-␤, activin, or nodal bind to type I activin B receptor, TGF-␤ type I receptor, or type I activin C receptor leading to phosphorylation of Smads 2 and 3 (18). BMP15 and GDF9 are unique among the TGF-␤ superfamily in that they are lacking the fourth conserved

B


cysteine residue involved in dimer formation (1) found in most TGF-␤ family members (19 –21). In the present study, we substituted a cysteine residue for the corresponding serine 356 (numbering from the amino terminus of the unprocessed protein) found in the human BMP15 (hBMP15) C-terminal mature region. Here we characterize the hBMP15 wild type (hBMP15wt) and S356C variant mature regions, purified without the addition of any epitope tags (22). We find that the resulting recombinant hBMP15S356C mature region forms a covalent dimer that exhibits identical bioactivity as the wild-type protein. This protein is amenable to iodination due to stabilization of the hBMP15 variant by an intersubunit disulfide bond. We were able to use this BMP15S356C variant to reveal that hBMP15 uses a complex of the BMP receptor 2 (BMPR2) and BMPR1B receptors for signaling in a human granulosa cell line.

Materials and Methods Expression vector construction The human BMP15 expression constructs were constructed as previously described (22). The proregion tags were inserted after the signal sequence as shown in Fig. 1A. The S356C mutation was inserted using QuikChange II Site-Directed Mutagenesis Kit (Stratagene, La Jolla, CA) by following the manufacturer’s instructions.

Protein expression and analysis Cell lines expressing epitope-tagged processed human BMP15 variants were developed as described previously (22). The epitope-tagged and untagged versions of BMP15 were detected in medium conditioned by these cell lines using monoclonal antibody 28 (mAb-28), which is specific for an epitope in the mature region of the human BMP15 protein (22). As such, this antibody detects the processed mature region of hBMP15, as well as any unprocessed precursor forms. The wild-type and the S356C variant hBMP15 mature regions were purified via a twostep procedure. First, a Ni2⫹-based immobilized metal ion-affinity resin step was carried out targeting the His6-tag at the amino terminus of the proregion (Fig. 1A). This was followed by a reverse-phase HPLC step to obtain the purified mature region, as used previously (22).

Human BMP15 wt

N

S FH

N

S FH

Pro

Mature

C

Human BMP15 S356C Pro

Mature

C

S356C FIG. 1. Schematic of the recombinant proteins produced in this study based on wild-type and S-C mutated hBMP15 (C, cysteine substitution in amino acid 356).


Reagents and growth factors Escherichia coli-produced BMP2 was a kind gift from Dr. Peter Mace (University of Otago, Dunedin, New Zealand). E. coli-produced activin A was a kind gift from Dr. Marko Hyvönen (University of Cambridge, Cambridge, UK). Fetal calf serum was purchased from Euroclone Ltd. (Devon, UK). DMEM and Ham’s F-12 were purchased from Invitrogen Life Technologies, Inc. (Gaithersburg, MD). Heparin (Fragmin) was purchased from Pharmacia & Upjohn (Stockholm, Sweden). BSA was purchased from Roche (Mannheim, Germany). Peroxidase-conjugated rabbit antimouse IgG was purchased from Jackson ImmunoResearch Laboratories, Inc. (West Grove, PA). Activin receptor 2A and -B ectodomain (ECD)/Fc chimera proteins were purchased from R&D Systems (Minneapolis, MN). BRII ECD/FC (BMPR2ecd-Fc) was produced in mammalian cells and purified in-house as described previously (23).

Reporter gene constructs The use of the pGL3BRE-luciferase reporter plasmid (24) and generation of the recombinant adenovirus BRE-Luc has been previously described (25).

Human granulosa luteal (hGL) cell cultures hGL cells were obtained with informed consent from women undergoing in vitro fertilization treatments. For each experiment, cells from one to six patients were pooled, and the cultures were prepared as described previously (22).

Adenovirus infections Adenovirus infections with adenovirus BRE-Luciferase were performed as described previously (22).

Transient transfections and luciferase assays Transient transfections with BRE-luciferase plasmid and luciferase assays were performed as described previously (22).

Inhibin B ELISA The inhibin B ELISA was performed as described previously (22).

Coimmunoprecipitation using receptor ECD/Fc chimeras The wild-type hBMP15 or hBMP15S356C mature regions (0.3 ␮g) were incubated with the indicated receptor ECD/Fc chimeric proteins (1.5 ␮g) in PBS containing 0.01% BSA for 45 min at room temperature. Protein A agarose beads (50 ␮l/reaction) were added for an additional 30 min and incubated for 30 min at 4 C in rotation. The agarose beads were collected by centrifugation and were washed (1 ml/reaction) three times with buffers containing subsequently increasing salt concentration. Bound proteins were eluted from the agarose beads in SDS-PAGE elution buffer containing 10 mM dithiothreitol (DTT) by boiling 3 min. Proteins eluted from the agarose beads were subjected to SDS-PAGE analysis and silver staining or Western blotting and immunostaining with mAb-28 (22).

Cross-linking Noncleavable cross-linkers, disuccinimidyl suberate (DSS) or bis(sulfosuccinimidyl) suberate (BS3), (Thermo Scientific Pierce,


1511

Rockford, IL) were added to 100 ng purified hBMP15 (diluted in 0.2% fetal calf serum DMEM containing 1:10 10 mM HCl) at a final concentration of 0.5 mM, incubated at room temperature for 30 min, before SDS-PAGE immunoblotting.

[125I]hBMP15S356C binding assay Iodination of the purified hBMP15S356C mature region was performed according to the chloramine T method, and cells were subsequently affinity labeled with the radioactive ligand as described before (26). In brief, cells were incubated on ice for 3 h with the radioactive ligand. After incubation, cells were washed and cross-linking was performed using 54 mM DSS and 3 mM BS3 (Pierce) for 15 min. Cells were washed, scraped, and lysed. Lysates were incubated with the respective antisera overnight, and immune complexes were precipitated by adding protein A Sepharose (Amersham Biosciences AB, Uppsala, Sweden). Samples were washed, boiled in SDS sample buffer, and subjected to SDSPAGE. Gels were dried and scanned with the STORM imaging system (Amersham).

Statistical analysis The statistical analysis of the data from this study focuses on a comparison of the hBMP15wt with the S356C variant. Hence, the results in Figs. 2C– 4A have been assessed by two-way ANOVA, data being log transformed when the normality test and/or equal variance test failed. There was no statistical main effect of type of BMP15 (P ⬎ 0.05) for all the data, except in Fig. 3A, in which only the highest concentration of BMP15 showed a significantly different effect (P ⬍ 0.001) as assessed by post hoc analysis using the Holm-Sidak multiple-comparison procedure.

Results Production of unpurified forms of recombinant BMP15 produced by HEK-293T cells The recombinant wild-type and the S356C variant hBMP15 proteins derived from the expression constructs shown in Fig. 1 were produced in stable HEK-293T cell lines. The various forms of recombinant hBMP15 can be detected in medium conditioned by these cell lines using mAb-28, which is specific for an epitope in the human BMP15 protein (22). The reduced hBMP15wt and S356C variant mature regions appear as two distinct bands corresponding to approximately 16 and 17 kDa as previously described for the wild-type (22) and tagged (27, 28) proteins. The 17-kDa band represents an O-linked glycosylated form of the human BMP15 mature region (28). Characterization of the purified hBMP15wt and S356C variant mature region The wild-type and the S356C variant hBMP15 mature regions were purified via a two-step procedure, first using the His6-tag at the amino terminus of the proregion, followed by a reverse-phase HPLC step to obtain the purified mature region, as used previously (22). We characterized

1512

Pulkki et al.


the purified hBMP15wt and S356C variant mature regions via SDS-PAGE under reducing (with 10 mM DTT) and nonreducing (without DTT) conditions and detected the proteins by silver staining and Western blotting using mAb-28 (Fig. 2). The reduced hBMP15wt and S356C variant mature regions appear as two distinct bands corresponding to approximately 16 and 17 kDa. The two distinct bands of the hBMP15wt protein can be separated without reduction as well. However, the two bands could not be separated without reduction as clearly as with reduction, presumably because the cysteine knot (17) would be still intact under these conditions. When the hBMP15wt protein was subjected to chemical cross-linking using DSS or BS3, the approximately 34-kDa homodimer of hBMP15 was observed (Fig. 2A). The hBMP15S356C variant protein, with the substituted cysteine residue known to be involved in forming the intermolecular disulfide bond between monomers (19 –21), forms a covalent homodimer (Fig. 2B). The nonreduced hBMP15S356C migrates as three bands corresponding to 30- to 34-kDa forms (Fig. 2B). The bioactivities of the various purified hBMP15 forms were tested on an immortalized human granulosa cell line COV434, which we and others have previously used to monitor BMP15 bioactivity (22, 29). The COV434 cells were transiently transfected with Smad1/5/8-responsive transcriptional reporter construct, BRE-luciferase (24), and the cells were stimulated with various doses of hBMP15wt or hBMP15S356C. The purified hBMP15S356C mature region was equally potent in activating the Smad1/5/8 pathway as the respective


wild-type protein (Fig. 2C), with two-way ANOVA indicating that there was no main effect of the type of BMP15 (P ⬎ 0.05). The purified BMP15wt and S356C variant mature regions stimulate inhibin B production in a doseand time-dependent manner We have shown previously that GDF9 stimulates the production of inhibin B by hGL cells (30 –32). In our previous study focusing on BMP15, we found that purified recombinant hBMP15wt protein also stimulates inhibin B production in hGL cells in a time- and dose-dependent manner (22). In this study, we find that purified hBMP15wt and S356C variant mature regions stimulate inhibin B production in hGL cells in a broadly equivalent dose-dependent manner (Fig. 3A), with two-way ANOVA indicating that there was no main effect of type of BMP15 (P ⬎ 0.05) for all doses, except for the highest concentration of BMP15 (P ⬍ 0.001). Furthermore, a time-course effect on inhibin B production in hGL cells by both variants was seen (Fig. 3B), which was equivalent (two-way ANOVA, P ⬎ 0.05). The BMPR2 ECD/Fc fusion molecule (BMPR2ecd-Fc) antagonizes the hBMP15 action Previously, we have shown that the BMPR2ecd-Fc protein inhibits the effects of the hBMP15wt on rat primary granulosa cell DNA synthesis (22). Here we have determined whether the BMPR2ecd-Fc can inhibit the biological effects of the hBMP15wt or S356C proteins. The

FIG. 2. Characterization of the purified hBMP15wt and hBMP15S356C mature region. A, When hBMP15wt protein was subjected to chemical cross-linking using DSS or BS3, the approximately 34-kDa hBMP15wt homodimer was observed. B, The reverse-phase HPLC-purified hBMP15wt and hBMP15S356C mature region were SDS-PAGE analyzed in nonreducing (without DTT) and reducing conditions (with 10 mM DTT) detected by silver staining (upper panel) and Western blotting (lower panel) by using the BMP15-specific mAb-28. The reduced mature region of hBMP15wt and hBMP15S356C migrates as two distinct bands corresponding to 16 and 17 kDa. The two distinct bands of the hBMP15wt protein can be separated without reduction as well. The nonreduced hBMP15S356C migrates as three bands corresponding to 30- to 34-kDa forms. C, COV434 cells transfected with the Smad1/5/8 (BRE-luc) reporter were incubated for 24 h in the absence (an adjusted value of 1.0 for the mean of the control wells) or presence of various concentrations of HPLC-purified hBMP15wt or hBMP15S356C protein. The purified hBMP15wt and S356C mature regions are equally potent in activating the Smad1/5/8 pathway.



1513

A 1200

Inhibin B (pg/mll)

1000

800

600

400

200

0 -

ng/ml LIGAND

100

2% FCS BMP2

3

10

30

100

300

3

10

30

100

300

BMP15 (S356C)

BMP15 wt

B 1600 1400

Inhibin B (pg/m ml)

1200

1000

800

600

400

200

0 24

48

72

96

Time (hours) 2% FCS

BMP15 wt

BMP15 (S356C)

FIG. 3. hBMP15 stimulates inhibin B production in cultured hGL cells. Three different cell pools were treated with 3–300 ng/ml hBMP15wt or hBMP15S356C protein for 48 h or with 150 ng/ml hBMP15wt or hBMP15S356C protein for 24, 48, 72, or 92 h, and the spent media were harvested for the measurement of inhibin B concentrations with an inhibin B ELISA. Data were similar in three different hGL cell pools (each pool contains two to six women), and representative experiments are shown. A, hBMP15wt and hBMP15S356C proteins stimulate hGL cell inhibin B production in a dose-dependent manner with the same potency. B, Time course of hBMP15wt or hBMP15S356C action on inhibin B production in hGL cells.

BMPR2ecd-Fc was a potent inhibitor of both the wildtype and the S356C hBMP15 variant as monitored via activation of the Smad1/5/8 (BRE-luc) signaling pathway (Fig. 4A). Indeed, the BMPR2ecd-Fc inhibitor did not discriminate between the two types of BMP15 (two-way ANOVA, P ⬎ 0.05). BMP2 acted as a specificity control because this ligand is active in the BRE-luc assay on COV434 cells but has a low affinity for the BMPR2 receptor alone; hence the BMPR2ecd-Fc fusion protein had no effect on BMP2 bioactivity. We also characterized the biochemical binding ability of hBMP15 to specific receptor ECD/Fc chimeras. Pull-down experiments revealed that both hBMP15wt and S356C variant bind only to BMPR2 and not to Activin receptor 2A or 2B (Fig. 4B).

hBMP15 forms a complex with BMPR2 and BMPR1B at the cell surface Previous in vitro studies using a C-terminally Flagtagged form of hBMP15 and soluble receptor ectodomain IgG Fc chimera have indirectly pointed out that BMPR2 and BMPR1B receptors could possibly be involved in BMP15 action (29). To study directly the binding of hBMP15 to the cell surface receptors, we analyzed the ability of hBMP15 to bind to the endogenous type I and II receptors on human COV434 ovarian granulosa tumor cells, a cell line that we have found by quantitative RTPCR to express all the potentially relevant signaling receptors (data not shown). The cells were incubated with [125I]hBMP15S356C, and cross-linked receptor-ligand

1514

Pulkki et al.


A


COV434 BRE-luc bioassay

15 14

Luciferase activity / Fold L F Change

13 12

BMPR2 ecd-Fc µg/ml:

11 10 9

0 0.03 0.1 0.3 1 3

8 7 6 5 4 3 2 1 0 Cont

BMP2 100

BMP15 wt 100

BMP15 (S356C) 100

( / l) (ng/ml)

ACVR2B

BMP15 (S356C) BMPR2 ACVR2A

ACVR2B

ACVR2A

BMPR2

ActA

ACVR2B

BMP15 wt BMPR2

ACVR2B

ACVR2A

BMP2

ACVR2A

B

17

17

mAb-28

FIG. 4. hBMP15wt and the hBMP15S356C variant bind the BMPR2 ECD/Fc fusion protein. A, The purified hBMP15wt mature region was preincubated with the BMPR2ecd-Fc protein before addition of the ligand-receptor complexes to COV434 cells. B, Coimmunoprecipitation of specific receptor ECD/Fc chimeras and hBMP15wt, hBMP15S356C, activin A, or BMP2 and subjected to SDS-PAGE silver staining (top) and immunoblotting (bottom).

complexes were immunoprecipitated with antisera against endoglin, ACVRL1, ACVR1, BMPR1A, ACVR1B, TGFBR1, BMPR1B, ACVR1C, TGFBR2, BMPR2, and ACVR2A and -B (26), or betaglycan (33). We found that human BMP15 strongly bound to BMPR1B and BMPR2 (Fig. 5A). We further confirmed this result by transfecting COS-7 cells with cDNA for BMPR1B or BMPR2 or with their combination. The cells were first affinity labeled with [125I]hBMP15S356C, and then ligand-receptor complexes were chemically crosslinked and subsequently immunoprecipitated with specific type I or II receptor antisera. Resolution of immunoprecipitated complexes by SDS-PAGE and detection of the radioactive signal on a phosphoimager screen showed that hBMP15 binds to BMPR1B and BMPR2, and the binding to BMPR1B is greatly facilitated when coexpressed with BMPR2 (Fig. 5B). To verify that hBMP15 binding to BMPR1B and BMPR2 is specific, we affinity labeled a hBMP15 nonresponsive human granulosa tumor cell line, KGN, with [125I]hBMP15S356C and found no binding of the ligand (data not shown).

Discussion In this study, we are providing the first direct evidence of specific cell surface receptors for a member of the GDF9/ BMP15 subfamily of oocyte growth factors. Our study confirmed that BMP15 preferentially binds to BMPR1B and BMPR2 in human granulosa cells. Here, we produced mammalian cell expressed and purified wild-type noncovalent dimer hBMP15 form in parallel to an engineered covalent dimer hBMP15 form. The engineered hBMP15S356C resembles all the other BMP harboring the endogenous intermolecular disulfide bridge commonly found in nearly all the other TGF-␤ family ligands. Our results demonstrate that the covalent hBMP15 dimer behaves biologically in a similar manner as the wild-type form on hGL cells and on COV434 granulosa tumor cells. Importantly, the availability of the hBMP15S356C variant enabled us to label the ligand with [125I]iodine without changing its structure, and this subsequently made it possible for the first time to detect hBMP15 interacting with its endogenous receptors. Previously, we and others have shown biological effects for mouse and ovine GDF9 and BMP15. These studies



1515

FIG. 5. Human BMP15 binds BMPR2 and BMPR1B individually and more efficiently when both receptors are present. A, Human BMP15S356C binds to endogenous receptors in COV434 cells. The COV434 cells were affinity labeled with [125I]hBMP15S356C, and cross-linked ligand-receptor complexes were immunoprecipitated with specific antisera as indicated. B, COS-7 cells were transiently transfected with cDNA for BMPR1B, BMPR2, or both together and affinity labeled with [125I]hBMP15S356C. Cross-linked complexes were immunoprecipitated with specific antisera against the receptors and subjected to SDS-PAGE and autoradiography. Human BMP15S356C bound BMPR2 and BMPR1B individually and more efficiently when both receptors were present.

were performed using unpurified conditioned media (34 – 36) or partially purified (37) mouse or ovine GDF9 and BMP15 proteins. Both the unpurified and the partially (hydrophobic interaction chromatography) purified preparations contain the pro- and the mature regions of the growth factors. Studies with these preparations have shown that GDF9 and BMP15 have different roles between different species, and the effects of GDF9 and BMP15, when added together, are different from those observed for the growth factors alone (34 –37). Also purified preparations have been used throughout the published literature (reviewed in Ref. 38). The purified forms, however, contain epitope tags in different locations of the pro- or mature region of the encoded protein. This may well be the prime cause for the differing results that have been reported concerning the bioactivity of GDF9 (25) and BMP15 (22). The most recent studies of ours show that placing the epitope tag at the C terminus of GDF9 (25) or BMP15 (22) has a detrimental effect on the bioactivity of the protein. Our (22, 25) and the commercially available (R&D Systems) untagged GDF9 and BMP15 as well as an N-terminally epitope tagged BMP15 (27) (our unpublished results) are showing consistent bioactivity and stability. However, C-terminally epitope-tagged forms (29) (reviewed in Ref. 39) need comparative studies in parallel with the untagged GDF9 and BMP15 forms.

We have shown in our previous studies that GDF9 stimulates inhibin B production in human granulosa cells (30 – 32). We have also observed previously that activation of both the activin- and BMP-activated Smad pathways results in significantly stimulated inhibin B production in human granulosa cells (40). The most recent study of ours shows that wild-type BMP15 stimulates inhibin B production in human granulosa cells in a dose-responsive manner (22). Here we show that the purified hBMP15wt and S356C variant mature regions stimulate inhibin B production in hGL cells in a time- and dose-responsive manner with a broadly similar potency. Furthermore, the purified hBMP15wt and S356C variant mature regions were tested in transcriptional reporter assays specific for Smad1/5/8 activation in human ovarian granulosa cells. The purified hBMP15wt and S356C variant proteins were equally potent activators of the Smad1/5/8 (BRE-luc)-responsive transcriptional reporter in these assays. A typical feature of most TGF-␤ family members is the seven conserved cysteines in their mature region. Six of these cysteines form a knot-like structure, whereas the remaining conserved cysteine is responsible for linking the two ligand monomers to each other, thus forming a dimer via disulfide bond (19, 21). GDF9 and BMP15 lack this conserved cysteine that is involved in peptide dimerization. However, it is presumed that hydrophobic interactions between the two

1516

Pulkki et al.


monomers are sufficient for binding the proteins together. Previous studies using C-terminally tagged GDF9 and BMP15 (41) have shown that these growth factors can form homodimers. Here, we are showing that when the wild-type hBMP15 protein is subjected to chemical crosslinking an approximately 34-kDa homodimer is observed. Also, the hBMP15S356C variant protein that contains a cysteine residue in the position where most other TGF-␤ family members have a dimerizing disulfide bridge, forms a covalent approximately 34-kDa homodimer. Therefore, this approximately 34-kDa homodimer is likely to be the signaling form of BMP15 at the receptor level. We have also unpublished results showing that wild-type GDF9 forms homodimers when it is subjected to chemical crosslinking. We conclude that because the hBMP15wt and S356C mature region proteins are equally potent in biological assays, and they resemble each other also biochemically, hBMP15S356C can be used as a tool in studying the function of hBMP15. So far, no previous study has been able to directly address the question of which type I or type II serine threonine kinase receptors GDF9 or BMP15 are binding on the cell surface. The published literature provides indirect evidence indicating that GDF9 would use TGFBR1 as a type I receptor (31, 42) and BMPR2 as a type II receptor (43). It has also been suggested that BMP15 would use the BMPR1B type I receptor (29, 44) and the BMPR2 type II receptor (22, 29, 44) for signaling. Furthermore, the bioactivity of mouse GDF9 (23) and human BMP15 (22) and the cooperative effect of mouse (35) and ovine (45) GDF9 and BMP15 have been clearly inhibited using the BMPR2ecd-Fc protein. Interestingly, our results showing that the BMPR2ecd-Fc fusion protein has the ability to pull down BMP15 (Fig. 4B) are in contrast with the published literature on this matter (29, 44). These two previous studies have been carried out using a C-terminally Flag-tagged BMP15, and in our recent studies, we have shown that a C-terminal modification has an effect on the structural integrity of both GDF9 (25) and BMP15 (22). The C-terminal modification may well be the cause for these discrepancies. Here, in the current study, characterization of the biochemical binding abilities of the hBMP15wt and hBMP15S356C mature regions revealed clearly that both proteins bind specifically to the ectodomain of the BMPR2 receptor in pull-down assays. We also find that the BMPR2ecd-Fc antagonized the BRE-luc bioactivity of hBMP15wt and S356C in a dose-dependent manner in COV434 cells. Furthermore, we have been able to radioactively label both BMP15wt and the S356C variant, but only the BMP15S356C mutant retained receptor binding capacity. Here we have used the [125I]hBMP15S356C protein to study the ability of hBMP15 to bind to the endogenous


type I and type II receptors in the human ovarian granulosa tumor cell line COV434. We find that human BMP15 strongly bound to BMPR1B and BMPR2 on the COV434 cell surface, an association that was further verified in COS-7 cells by overexpression of these receptors. Mutations causing modulatory effects on fertility of mono- and polyovulatory mammals have been identified in BMP15 (5, 9, 10, 12, 14) and BMPR1B (46 –50). These mutations show the importance of these individual proteins in fertility regulation. A study by Demirhan et al. (46) describes a homozygous mutation in the BMPR1B gene in a 16-yr-old girl resulting in a severe condition with a skeletal and genital phenotype, leading to ovarian dysfunction. Similarly to humans with a homozygous mutation in BMPR1B, sheep with homozygous mutations in BMP15 (5, 9, 10, 12) suffer from nonfunctional streak ovaries and infertility. Furthermore, Booroola sheep, a strain with a natural point mutation in the highly conserved intracellular serine threonine kinase signaling domain of the BMPR1B receptor (47– 49), show a significant increase in ovulation rate. The basis for this phenotype has recently been reported to be lower BMP15 mRNA levels together with an earlier onset of LH responsiveness in granulosa cells in sheep homozygous for the Booroola mutation (51). Although the functional nature of the Booroola mutant is not completely understood, experimental evidence suggests that this mutant has somewhat suppressed signaling activity (47). Our present study suggesting that BMP15 is using BMPR1B as a signaling receptor is consistent with the idea that the Booroola BMPR1B mutant would primarily mimic a slightly compromised BMP15 signaling situation. A striking observation from the identified homozygous mutations of the BMPR1B (46) and BMP15 (5, 9, 10, 12) genes is their similar ovarian phenotype of streak ovaries leading to ovarian dysfunction. These genetic findings are consistent with the suggestion that BMPR1B is the main type I receptor for BMP15. It should be pointed out that there is a marked species difference in the phenotype of BMPR1B mutants when comparing the situation in the mouse (50) with that of humans (46), in that BMPR1B knockout mice are infertile due to a lack of cumulus expansion; however, ovaries are present and contain follicles. This situation is consistent with the mild phenotype in mice for the BMP15 knockout, in that the mice are subfertile with a defect in cumulus expansion (14). The more severe phenotype in the human for an BMPR1B mutant is consistent with the greater importance of BMP15 in the human for normal levels of fertility. The present studies have started to unveil which receptor subtypes are involved in the signaling of the BMP15 homodimer. However, for better understanding of the physiology of these intriguing oocyte-derived growth fac-


tors, the heterodimer and co-operational signaling and receptor binding activities of GDF9 and BMP15 will have to be characterized in detail in future studies. The current availability of purified bioactive hGDF9 (25) and hBMP15 (22) will certainly facilitate these follow-up studies.


8.

9.

Acknowledgments The skillful technical assistance of Mr. Jarmo Koponen, Mr. Juha Sund, and Ms. Marjo Rissanen is kindly acknowledged. Drs. Peter Mace and Marko Hyvönen are acknowledged for donating BMP2 and activin A proteins. Dr. Robert Gilchrist is thanked for his assistance in all matters statistical. Address all correspondence and requests for reprints to: Minna Pulkki, Vactech Oy, Biokatu 8, FIN-33520 Tampere, Finland. E-mail: [email protected]. These studies were supported by grants from European Molecular Biology Organization, the Academy of Finland, the Finnish Funding Agency for Technology and Innovation, the Juselius Foundation, the Jalmari and Rauha Ahokas Foundation, the Novo Nordisk Foundation, the Orion-Farmos research foundation, the Helsinki University research funds, the Paulo Foundation, the Jenny and Antti Wihuri Foundation, New Zealand Foundation for Research Science and Technology, the Scientific Research Grant 918.66.606 from The Netherlands Organization and Centre for Biomedical Genetics. Disclosure Summary: The authors have nothing to disclose.

10.

11.

12.

13.

14.

15.

References 1. Laitinen M, Vuojolainen K, Jaatinen R, Ketola I, Aaltonen J, Lehtonen E, Heikinheimo M, Ritvos O 1998 A novel growth differentiation factor-9 (GDF-9) related factor is co-expressed with GDF-9 in mouse oocytes during folliculogenesis. Mech Dev 78:135– 140 2. Dube JL, Wang P, Elvin J, Lyons KM, Celeste AJ, Matzuk MM 1998 The bone morphogenetic protein 15 gene is X-linked and expressed in oocytes. Mol Endocrinol 12:1809 –1817 3. Aaltonen J, Laitinen MP, Vuojolainen K, Jaatinen R, Horelli-Kuitunen N, Seppä L, Louhio H, Tuuri T, Sjöberg J, Bützow R, Hovata O, Dale L, Ritvos O 1999 Human growth differentiation factor 9 (GDF-9) and its novel homolog GDF-9B are expressed in oocytes during early folliculogenesis. J Clin Endocrinol Metab 84:2744 – 2750 4. McPherron AC, Lee SJ 1993 GDF-3 and GDF-9: two new members of the transforming growth factor-␤ superfamily containing a novel pattern of cysteines. J Biol Chem 268:3444 –3449 5. Bodin L, Di Pasquale E, Fabre S, Bontoux M, Monget P, Persani L, Mulsant P 2007 A novel mutation in the BMP15 gene causing defective protein secretion is associated with both increased ovulation rate and sterility in Lacaune sheep. Endocrinology 148:393– 400 6. Di Pasquale E, Beck-Peccoz P, Persani L 2004 Hypergonadotropic ovarian failure associated with an inherited mutation of human bone morphogenetic protein-15 (BMP15) gene. Am J Hum Genet 75: 106 –111 7. Di Pasquale E, Rossetti R, Marozzi A, Bodega B, Borgato S, Cavallo

16. 17. 18.

19.

20.

21.

22.

23.

1517

L, Einaudi S, Radetti G, Russo G, Sacco M, Wasniewska M, Cole T, Beck-Peccoz P, Nelson LM, Persani L 2006 Identification of new variants of human BMP15 gene in a large cohort of women with premature ovarian failure. J Clin Endocrinol Metab 91:1976 –1979 Dixit H, Rao LK, Padmalatha VV, Kanakavalli M, Deenadayal M, Gupta N, Chakrabarty B, Singh L 2006 Missense mutations in the BMP15 gene are associated with ovarian failure. Hum Genet 119: 408 – 415 Galloway SM, McNatty KP, Cambridge LM, Laitinen MP, Juengel JL, Jokiranta TS, McLaren RJ, Luiro K, Dodds KG, Montgomery GW, Beattie AE, Davis GH, Ritvos O 2000 Mutations in an oocytederived growth factor gene (BMP15) cause increased ovulation rate and infertility in a dosage-sensitive manner. Nat Genet 25:279 –283 Hanrahan JP, Gregan SM, Mulsant P, Mullen M, Davis GH, Powell R, Galloway SM 2004 Mutations in the genes for oocyte-derived growth factors GDF9 and BMP15 are associated with both increased ovulation rate and sterility in Cambridge and Belclare sheep (Ovis aries). Biol Reprod 70:900 –909 Laissue P, Christin-Maitre S, Touraine P, Kuttenn F, Ritvos O, Aittomaki K, Bourcigaux N, Jacquesson L, Bouchard P, Frydman R, Dewailly D, Reyss AC, Jeffery L, Bachelot A, Massin N, Fellous M, Veitia RA 2006 Mutations and sequence variants in GDF9 and BMP15 in patients with premature ovarian failure. Eur J Endocrinol 154:739 –744 Martinez-Royo A, Jurado JJ, Smulders JP, Martí JI, Alabart JL, Roche A, Fantova E, Bodin L, Mulsant P, Serrano M, Folch J, Calvo JH 2008 A deletion in the bone morphogenetic protein 15 gene causes sterility and increased prolificacy in Rasa Aragonesa sheep. Anim Genet 39:294 –297 Dong J, Albertini DF, Nishimori K, Kumar TR, Lu N, Matzuk MM 1996 Growth differentiation factor-9 is required during early ovarian folliculogenesis. Nature 383:531–535 Yan C, Wang P, DeMayo J, DeMayo FJ, Elvin JA, Carino C, Prasad SV, Skinner SS, Dunbar BS, Dube JL, Celeste AJ, Matzuk MM 2001 Synergistic roles of bone morphogenetic protein 15 and growth differentiation factor 9 in ovarian function. Mol Endocrinol 15:854 – 866 Teixeira Filho FL, Baracat EC, Lee TH, Suh CS, Matsui M, Chang RJ, Shimasaki S, Erickson GF 2002 Aberrant expression of growth differentiation factor-9 in oocytes of women with polycystic ovary syndrome. J Clin Endocrinol Metab 87:1337–1344 Massagué J 1998 TGF-␤ signal transduction. Annu Rev Biochem 67:753–791 Sun PD, Davies DR 1995 The cystine-knot growth-factor superfamily. Annu Rev Biophys Biomol Struct 24:269 –291 Heldin CH, Miyazono K, ten Dijke P 1997 TGF-beta signalling from cell membrane to nucleus through SMAD proteins. Nature 390: 465– 471 Daopin S, Piez KA, Ogawa Y, Davies DR 1992 Crystal structure of transforming growth factor-␤2: an unusual fold for the superfamily. Science 257:369 –373 Griffith DL, Keck PC, Sampath TK, Rueger DC, Carlson WD 1996 Three-dimensional structure of recombinant human osteogenic protein 1: structural paradigm for the transforming growth factor ␤ superfamily. Proc Natl Acad Sci USA 93:878 – 883 Schlunegger MP, Cerletti N, Cox DA, McMaster GK, Schmitz A, Grütter MG 1992 Crystallization and preliminary x-ray analysis of recombinant human transforming growth factor ␤2. FEBS Lett 303: 91–93 Pulkki MM, Myllymaa S, Pasternack A, Lun S, Ludlow H, AlQahtani A, Korchynskyi O, Groome N, Juengel JL, Kalkkinen N, Laitinen M, Ritvos O, Mottershead DG 2011 The bioactivity of human bone morphogenetic protein-15 is sensitive to C-terminal modification: characterization of the purified untagged processed mature region. Mol Cell Endocrinol 332:106 –115 Myllymaa S, Pasternack A, Mottershead DG, Poutanen M, Pulkki MM, Pelliniemi LJ, Ritvos O, Laitinen MP 2010 Inhibition of oocyte

1518

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

Pulkki et al.


growth factors in vivo modulates ovarian folliculogenesis in neonatal and immature mice. Reproduction 139:587–598 Korchynskyi O, ten Dijke P 2002 Identification and functional characterization of distinct critically important bone morphogenetic protein-specific response elements in the Id1 promoter. J Biol Chem 277:4883– 4891 Mottershead DG, Pulkki MM, Muggalla P, Pasternack A, Tolonen M, Myllymaa S, Korchynskyi O, Nishi Y, Yanase T, Lun S, Juengel JL, Laitinen M, Ritvos O 2008 Characterization of recombinant human growth differentiation factor-9 signaling in ovarian granulosa cells. Mol Cell Endocrinol 283:58 – 67 Scharpfenecker M, van Dinther M, Liu Z, van Bezooijen RL, Zhao Q, Pukac L, Löwik CW, ten Dijke P 2007 BMP-9 signals via ALK1 and inhibits bFGF-induced endothelial cell proliferation and VEGFstimulated angiogenesis. J Cell Sci 120:964 –972 Li Q, Rajanahally S, Edson MA, Matzuk MM 2009 Stable expression and characterization of N-terminal tagged recombinant human bone morphogenetic protein 15. Mol Hum Reprod 15:779 –788 Saito S, Yano K, Sharma S, McMahon HE, Shimasaki S 2008 Characterization of the post-translational modification of recombinant human BMP-15 mature protein. Protein Sci 17:362–370 Moore RK, Otsuka F, Shimasaki S 2003 Molecular basis of bone morphogenetic protein-15 signaling in granulosa cells. J Biol Chem 278:304 –310 Kaivo-Oja N, Bondestam J, Kämäräinen M, Koskimies J, Vitt U, Cranfield M, Vuojolainen K, Kallio JP, Olkkonen VM, Hayashi M, Moustakas A, Groome NP, ten Dijke P, Hsueh AJ, Ritvos O 2003 Growth differentiation factor-9 induces Smad2 activation and inhibin B production in cultured human granulosa-luteal cells. J Clin Endocrinol Metab 88:755–762 Kaivo-Oja N, Mottershead DG, Mazerbourg S, Myllymaa S, Duprat S, Gilchrist RB, Groome NP, Hsueh AJ, Ritvos O 2005 Adenoviral gene transfer allows Smad-responsive gene promoter analyses and delineation of type I receptor usage of transforming growth factor-␤ family ligands in cultured human granulosa luteal cells. J Clin Endocrinol Metab 90:271–278 Roh JS, Bondestam J, Mazerbourg S, Kaivo-Oja N, Groome N, Ritvos O, Hsueh AJ 2003 Growth differentiation factor-9 stimulates inhibin production and activates Smad2 in cultured rat granulosa cells. Endocrinology 144:172–178 Piek E, Franzén P, Heldin CH, ten Dijke P 1997 Characterization of a 60-kDa cell surface-associated transforming growth factor-␤ binding protein that can interfere with transforming growth factor-␤ receptor binding. J Cell Physiol 173:447– 459 Gilchrist RB, Ritter LJ, Cranfield M, Jeffery LA, Amato F, Scott SJ, Myllymaa S, Kaivo-Oja N, Lankinen H, Mottershead DG, Groome NP, Ritvos O 2004 Immunoneutralization of growth differentiation factor 9 reveals it partially accounts for mouse oocyte mitogenic activity. Biol Reprod 71:732–739 McIntosh CJ, Lun S, Lawrence S, Western AH, McNatty KP, Juengel JL 2008 The proregion of mouse BMP15 regulates the cooperative interactions of BMP15 and GDF9. Biol Reprod 79:889 – 896 McNatty KP, Juengel JL, Reader KL, Lun S, Myllymaa S, Lawrence SB, Western A, Meerasahib MF, Mottershead DG, Groome NP, Ritvos O, Laitinen MP 2005 Bone morphogenetic protein 15 and growth differentiation factor 9 co-operate to regulate granulosa cell function. Reproduction 129:473– 480 Hussein TS, Thompson JG, Gilchrist RB 2006 Oocyte-secreted fac-

38.

39. 40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.


tors enhance oocyte developmental competence. Dev Biol 296:514 – 521 Mottershead DG, Watson AJ 2009 Oocyte peptides as paracrine tools for ovarian stimulation and oocyte maturation. Mol Hum Reprod 15:789 –794 Moore RK, Shimasaki S 2005 Molecular biology and physiological role of the oocyte factor, BMP-15. Mol Cell Endocrinol 234:67–73 Bondestam J, Kaivo-oja N, Kallio J, Groome N, Hydén-Granskog C, Fujii M, Moustakas A, Jalanko A, ten Dijke P, Ritvos O 2002 Engagement of activin and bone morphogenetic protein signaling pathway Smad proteins in the induction of inhibin B production in ovarian granulosa cells. Mol Cell Endocrinol 195:79 – 88 Liao WX, Moore RK, Otsuka F, Shimasaki S 2003 Effect of intracellular interactions on the processing and secretion of bone morphogenetic protein-15 (BMP-15) and growth and differentiation factor-9. Implication of the aberrant ovarian phenotype of BMP-15 mutant sheep. J Biol Chem 278:3713–3719 Mazerbourg S, Klein C, Roh J, Kaivo-Oja N, Mottershead DG, Korchynskyi O, Ritvos O, Hsueh AJ 2004 Growth differentiation factor-9 signaling is mediated by the type I receptor, activin receptorlike kinase 5. Mol Endocrinol 18:653– 665 Vitt UA, Mazerbourg S, Klein C, Hsueh AJ 2002 Bone morphogenetic protein receptor type II is a receptor for growth differentiation factor-9. Biol Reprod 67:473– 480 McMahon HE, Sharma S, Shimasaki S 2008 Phosphorylation of bone morphogenetic protein-15 and growth and differentiation factor-9 plays a critical role in determining agonistic or antagonistic functions. Endocrinology 149:812– 817 Edwards SJ, Reader KL, Lun S, Western A, Lawrence S, McNatty KP, Juengel JL 2008 The cooperative effect of growth and differentiation factor-9 and bone morphogenetic protein (BMP)-15 on granulosa cell function is modulated primarily through BMP receptor II. Endocrinology 149:1026 –1030 Demirhan O, Türkmen S, Schwabe GC, Soyupak S, Akgül E, Tastemir D, Karahan D, Mundlos S, Lehmann K 2005 A homozygous BMPR1B mutation causes a new subtype of acromesomelic chondrodysplasia with genital anomalies. J Med Genet 42:314 –317 Mulsant P, Lecerf F, Fabre S, Schibler L, Monget P, Lanneluc I, Pisselet C, Riquet J, Monniaux D, Callebaut I, Cribiu E, Thimonier J, Teyssier J, Bodin L, Cognié Y, Chitour N, Elsen JM 2001 Mutation in bone morphogenetic protein receptor-IB is associated with increased ovulation rate in Booroola Merino ewes. Proc Natl Acad Sci USA 98:5104 –5109 Souza CJ, MacDougall C, MacDougall C, Campbell BK, McNeilly AS, Baird DT 2001 The Booroola (FecB) phenotype is associated with a mutation in the bone morphogenetic receptor type 1 B (BMPR1B) gene. J Endocrinol 169:R1–R6 Wilson T, Wu XY, Juengel JL, Ross IK, Lumsden JM, Lord EA, Dodds KG, Walling GA, McEwan JC, O’Connell AR, McNatty KP, Montgomery GW 2001 Highly prolific Booroola sheep have a mutation in the intracellular kinase domain of bone morphogenetic protein IB receptor (ALK-6) that is expressed in both oocytes and granulosa cells. Biol Reprod 64:1225–1235 Yi SE, LaPolt PS, Yoon BS, Chen JY, Lu JK, Lyons KM 2001 The type I BMP receptor BmprIB is essential for female reproductive function. Proc Natl Acad Sci USA 98:7994 –7999 Crawford JL, Heath DA, Reader KL, Quirke LD, Hudson NL, Juengel JL, McNatty KP 2011 Oocytes in sheep homozygous for a mutation in bone morphogenetic protein receptor 1B express lower mRNA levels of bone morphogenetic protein 15 but not growth differentiation factor 9. Reproduction 142:53– 61