|Home | About | Journals | Submit | Contact Us | Français|
Follicle-stimulating hormone beta subunit (Fshb) expression is regulated by transforming growth factor beta superfamily ligands. Recently, we demonstrated that bone morphogenetic proteins (BMPs) stimulate Fshb transcription alone and in synergy with activins. Also, transfection of the BMP type II receptor (BMPR2) and constitutively active forms of the type I receptors (activin A receptor type I [ACVR1] or BMP receptor type IA [BMPR1A]) in immortalized gonadotroph cells (LbetaT2) stimulated murine Fshb promoter-reporter activity. A third type I receptor (BMP receptor type IB [BMPR1B]) is also expressed in LbetaT2 cells, but we did not previously assess its functional role. A point mutation in BMPR1B (Q249R) is associated with increased ovulation rates and elevated FSH levels in Booroola (FecB) sheep. Herein, we assessed whether BMPR1B can regulate Fshb transcription in LbetaT2 cells and whether its ability to do so is altered by the Q249R mutation. As with ACVR1 and BMPR1A, coexpression of BMPR1B with BMPR2 increased Fshb promoter-reporter activity in BMP2-dependent and BMP2-independent fashions. Unexpectedly, the BMPR1B-Q249R mutant was equivalent to the wild type in its ability to stimulate SMAD1/5 phosphorylation and Fshb transcription. Pharmacological inhibition of ACVR1, BMPR1A, and BMPR1B confirmed that one or more of these receptors are required for BMP2-stimulated SMAD1/5 phosphorylation and Fshb reporter activity. Knockdown of endogenous BMPR1A, but not ACVR1 or BMPR1B, significantly impaired the synergism of BMP2 with activin A. Collectively, these data suggest that BMPR1A is the preferred BMP2 type I receptor in LbetaT2 cells and that neither ACVR1 nor BMPR1B compensates for its loss. The specific mechanism(s) through which the Booroola FecB mutation alters BMPR1B function remains to be determined.
The gonadotropins, follicle-stimulating hormone (FSH) and luteinizing hormone (LH), are essential reproductive hormones. Both are secreted from gonadotrophs of the anterior pituitary, but the two act to regulate different aspects of gonadal function. The gonadotropins are heterodimeric glycoproteins (α/β), with their β subunits determining rates of mature hormone synthesis and biological specificity. Both FSH and LH are regulated by gonadotropin-releasing hormone secreted from the hypothalamus and gonadal sex steroids; however, endocrine-paracrine transforming growth factor beta (TGFB) superfamily ligands such as activins and inhibins act to selectively regulate FSH synthesis. Activins signal via a combination of type II receptors (ACVR2 or ACVR2B), type I receptors (ACVR1B and ACVR1C, also known as activin receptor-like kinase [ALK] 4 and ALK7), and downstream signaling effectors (SMAD2 and SMAD3) to up-regulate FSH beta subunit (Fshb) transcription [1–5]. In contrast, inhibins suppress Fshb expression by blocking the actions of activins through a competitive binding mechanism .
Recently, other members of the TGFB superfamily, the bone morphogenetic proteins (BMPs), were shown to stimulate Fshb transcription alone and in synergy with activins [7–10]. The BMPs are expressed in LβT2 cells (an immortalized murine gonadotroph cell line) and in adult murine pituitary [7–9] and might regulate FSH synthesis in vivo. We previously reported  that BMP6 and BMP7, although endogenously expressed in LβT2 cells, only modestly regulate Fshb transcription. In contrast, BMP2 and BMP4 stimulate Fshb transcription more potently, but their expression in LβT2 cells is very low. However, BMP2 and BMP4 are highly expressed in the murine pituitary and may act as paracrine regulators of gonadotroph function. Relative to equimolar activins, BMP2 and BMP4 only weakly stimulate Fshb transcription, but they are nonetheless potent synergistic regulators when applied in combination with the activins. Physiologically, BMPs may be more important in terms of their cooperative rather than independent actions.
BMP2 and BMP4 signaling is initiated by the interaction of the ligands with BMP type I receptors such as BMPR1A and BMPR1B (also known as ALK3 and ALK6). A type II receptor such as BMP type II receptor (BMPR2) is then recruited into the complex and phosphorylates the type I receptors [11, 12]. The activated type I receptors then phosphorylate intracellular signaling proteins, the most thoroughly characterized of which are the receptor-regulated SMADs (R-SMADs) SMAD1, SMAD5, and SMAD8. Once phosphorylated, R-SMADs form heteromeric complexes with the coregulatory SMAD (SMAD4), accumulate in the nucleus, and act as transcription factors, either activating or repressing gene expression [11, 13, 14]. Activins stimulate FSH synthesis by up-regulating Fshb subunit gene transcription at least in part through the SMAD2 and SMAD3 signaling proteins [2–4, 15]. The available data suggest that BMP2 might preferentially signal through SMAD8 to regulate the Fshb gene .
The BMP family members show some promiscuity in their binding to type I and type II receptors within the TGFB superfamily. For example, BMP2 and BMP4 preferentially signal through the type II receptor BMPR2 but can use ACVR2A in its absence . Similarly, BMPs can bind to several type I receptors, including ACVR1, BMPR1A, and BMPR1B . Each of these type I receptors is expressed in LβT2 cells [7, 9, 18]; however, our previous overexpression data suggested a preferred role for ACVR1 in mediating BMP2 responses . Nonetheless, a role for BMPR1B was not assessed, and the data with wild-type and constitutively active BMPR1A yielded conflicting results.
A potential role for BMPR1B in FSH regulation is particularly intriguing in light of the phenotype of so-called Booroola (FecB) sheep. These animals show increased ovulation rates, leading to multiple births [19–21], and FSH levels are elevated in some Booroola flocks . The FecB mutation was mapped to the Bmpr1b locus and a missense point mutation (CAG→CGG [Q249R]) discovered in the highly conserved intracellular serine-threonine kinase domain of the receptor [23–25]; however, the specific alteration in receptor function, at a mechanistic level, has not been determined. Some data suggest that the mutation leads to a partial loss of receptor function, particularly at the ovarian level [26, 27], but alterations at the pituitary level have not been ruled out definitively. In fact, recent data show differences in BMP signaling in pituitary cultures from Booroola and wild-type sheep . These effects may not be mediated directly at the gonadotroph level, as previous investigators failed to detect BMPR1B expression in ovine gonadotrophs by immunofluorescence . Nonetheless, one cannot rule out the possibility of low-level expression in these cells that evaded detection by this method. Indeed, Bmpr1b mRNA is expressed at low levels in LβT2 cells [7, 18]. Herein, we assessed the relative roles of endogenous ACVR1, BMPR1A, and BMPR1B in BMP2-regulated Fshb transcription in LβT2 cells and examined potential functional changes in the mutant BMPR1B receptor (Q249R) at the level of the gonadotroph.
Human recombinant activin A and BMP2 were purchased from R&D Systems (Minneapolis, MN). Wisent (St-Bruno, QC) was the supplier of gentamycin, 1× PBS, and Dulbecco modified Eagle medium (DMEM) with 4.5 g/l glucose, l-glutamine, and sodium pyruvate. We obtained 1× passive lysis buffer (PLB) from Promega (Madison, WI). Protease inhibitor cocktail tablets (CompleteMini) were purchased from Roche (Nutley, NJ). Aprotinin, leupeptin, pepstatin, PMSF, SB431542, mouse monoclonal β-actin (No. A5441), mouse monoclonal HA (No. H9658) and MYC (No. 9E10) antibodies, and rabbit monoclonal FLAG (No. F3165) antibody were from Sigma (St. Louis, MO). The pSMAD1/5/8 rabbit polyclonal antibody (No. 9511) was from Cell Signaling Technology, Inc. (Danvers, MA). Horseradish peroxidase-conjugated secondary antibodies were from Bio-Rad (Hercules, CA), and enhanced chemiluminescence Plus reagent was from GE Healthcare (Piscataway, NJ). Compound C (No. 171261) was purchased from Calbiochem (San Diego, CA). The following short-interfering (si) RNAs were purchased from Dharmacon, Inc. (Lafayette, CO): control (catalog No. D-001210–05), ACVR1 (catalog No. D-042047–01), BMPR1A (catalog No. D-040598–01), BMPR1B (catalog No. D-051071–01), ACVR2 (catalog No. D-040676–01), ACVR2B (catalog No. D-040629–02), and BMPR2 (catalog No. D-040599–01). Sodium bisulfite was purchased from Fisher Scientific (Fair Lawn, NJ) (catalog No. S654–500), and quinol hydroquinone was purchased from BDH AnalaR (Poole, England) (catalog No. 10312).
The expression constructs for rat ACVR1-HA, FLAG-ACVR2, and FLAG-ACVR2B and for human FLAG-SMAD1 were provided by Dr. Teresa Woodruff (Northwestern University, Evanston, IL). Human BMPR1A-HA (Q233D) and murine BMPR1B-HA (Q203D) were provided by Dr. Mitsuyasu Kato (University of Tsukuba, Tsukuba, Japan). The following variants were constructed by site-directed mutagenesis using the QuikChange protocol (Stratagene, La Jolla, CA) and the primers listed in Supplemental Table S1 (available at www.biolreprod.org): constitutively active and siRNA-sensitive ACVR1-HA (Q207D); wild-type and siRNA-sensitive BMPR1A-Q233D-HA; wild-type BMPR1B-HA, BMPR1B-Q249R-HA, BMPR1B-Q249R/Q203D-HA, and BMPR1B-Q203D/D265A-MYC; methylated BMPR1B-Q249R-HA; and siRNA-resistant BMPR1B-Q203D-HA. In the case of methylated BMPR1B-Q249R, primers containing methylated cytosines (Supplemental Table S1) were used, and the resulting PCR products were purified by ethanol precipitation following DpnI digestion of the parental plasmid and utilized directly in transfection experiments. Methylation was confirmed by bisulfite sequencing [30, 31]. All BMPR1A and BMPR1B constructs were subcloned into pcDNA4 (Invitrogen, San Diego, CA). This removed the HA tag and replaced it with a C-terminal MYC-HIS tag. Human FLAG-SMAD5 was provided by Dr. Tetsuro Watabe (Tokyo University, Tokyo, Japan). The human BMPR2 expression construct  and BREX4-luc  were provided by Dr. Joan Massague (Memorial Sloan-Kettering Cancer Center, New York, NY). The murine Fshb promoter-reporter constructs were described previously .
Immortalized murine gonadotroph LβT2 cells were provided by Dr. Pamela Mellon (University of California, San Diego, CA) and were cultured in 10% fetal bovine serum (FBS)/DMEM and 4 μg/ml gentamycin as described previously . For luciferase assays, cells were plated in 24-well plates (2.5 × 105 cells/well) approximately 36 h before transfection. Cells were transfected with Lipofectamine 2000 according to the manufacturer's instructions (Invitrogen). Twenty-four hours after transfection, cells were washed in 1× PBS and then treated with 1 nM (25 ng/ml) activin A and/or BMP2 in DMEM or with DMEM alone (control) for the indicated times. In overexpression experiments, 450 ng of the reporter and 100 ng of each receptor and/or effector were used per well. Cells were placed into serum-free media 24 h after transfection. In some experiments, 10 μM SB431542, an ACVR1B/ACVR1C/TGFBRI inhibitor , was included to block the effects of endogenous activin B (or other ligands signaling through these receptors). In RNA interference (RNAi) experiments, siRNAs were transfected at 5 nM. Resulting data were calibrated to cells transfected with the 1× siRNA buffer only (20 mM KCl, 6 mM HEPES [pH 7.5], and 0.2 mM MgCl2) or to cells transfected with the control siRNA. Lysates were collected 24 h after transfer to serum-free medium. CHO cells were obtained from Dr. Patricia Morris (Population Council, New York, NY) and were cultured in F-12/DMEM containing 10% FBS and 4 μg/ml gentamycin. Except for the BMPR1B/D265A experiment (where 4 μg FLAG-SMAD1 and 4 μg of receptor were transfected in CHO cells seeded in 10-cm plates), CHO cells in 6-well plates were transfected when 70%–80% confluent using Lipofectamine reagent, 300 ng of the indicated receptor expression vectors, and 1 μg FLAG-SMAD1 or FLAG-SMAD5 for 6 h and were then placed in growth media. The repeat of this experiment in LβT2 cells in 6-well plates was performed in a similar fashion, except that Plus reagent and 1200 ng of the indicated receptor expression vectors were included. Cell lysates were then harvested the following day. HepG2 cells (No. HB-8065) were purchased from ATCC (Manassas, VA) and were cultured in 10% FBS/Eagle minimum essential medium (modified by ATCC) and 4 μg/ml gentamycin. Transfection protocols were identical to those used for the LβT2 cells.
Cells were washed with 1× PBS and lysed in 1× PLB. Luciferase assays were performed on an Orion II microplate luminometer (Berthold Detection Systems, Oak Ridge, TN) using standard reagents. All treatments were performed in duplicate or triplicate as described in the text or figure legends. Data presented are from at least 2–3 independent experiments.
Cells were washed with 1× PBS, and whole-cell protein extracts (WCEs) were prepared with 1× RIPA (1% NP-40, 1% sodium deoxycholate, 0.1% SDS, 0.15 M NaCl, 0.01 M sodium phosphate [pH 6.8], 2 mM edetic acid, 50 mM sodium fluoride, and CompleteMini tablets) and centrifuged at 13000 × g for 0.5 h at 4°C to remove cellular debris. The WCEs were subjected to immunoblot analyses as previously described . Briefly, equivalent amounts of protein were separated by SDS-PAGE and were transferred to Protran nitrocellulose filters (Schleicher and Schuell, Keene, NH). Blots were probed with the indicated antibodies using standard techniques.
Data from three replicate experiments were highly similar, and their means were pooled for statistical analyses. Data are presented as fold change from the control condition in each experiment. Differences between means were compared using one-way, two-way, or three-way ANOVAs, followed by post hoc pairwise comparison with Bonferroni or Tukey adjustment where appropriate (SYSTAT 10.2; Systat Software, Inc., Richmond, CA), as indicated in the figure legends. Significance was assessed relative to P < 0.05.
Results of previous transfection investigations in our laboratory suggested that ACVR1 might be the preferred type I receptor mediating the regulation of Fshb transcription by BMP2 . Although we and others  observed Bmpr1b mRNA expression in the murine pituitary and LβT2 cells, we did not previously assess its role in BMP2 signaling. In addition, we discovered that the wild-type BMPR1A expression vector we had used previously  harbored an unwanted frameshift mutation that truncated the receptor within the kinase domain. This potentially invalidated the interpretation of our previous results  using this reagent. Therefore, we transfected LβT2 cells with the −846/+1 mouse Fshb-luc reporter and validated wild-type BMPR1A or BMPR1B receptor expression vectors alone or together with the type II receptor BMPR2. As observed previously with ACVR1 , either BMPR1A or BMPR1B with BMPR2 conferred heightened BMP2-independent and BMP2-dependent Fshb promoter activity (Fig. 1A). These effects were only observed when BMPR1A or BMPR1B was expressed in conjunction with BMPR2 and not when either was expressed alone. Similarly, constitutively active forms of BMPR1A (Q233D) and BMPR1B (Q203D) when expressed together with BMPR2, but not alone, stimulated Fshb promoter activity (Fig. 2A). Collectively, these results suggest that overexpressed ACVR1 (as shown previously ), BMPR1A, and BMPR1B can all regulate Fshb transcription in conjunction with BMPR2.
A missense mutation, Q249R, was mapped to the kinase domain of BMPR1B in Booroola (FecB) sheep [23–25]. Given that Bmpr1b is expressed in the pituitary and may mediate BMP2 effects on Fshb (Figs. 1A and and2A),2A), we investigated the effects of the BMPR1B-Q249R mutation on BMP2 signaling in gonadotroph cells. We introduced the mutation in the context of a murine BMPR1B expression vector. As already observed, expression of type I or type II receptors alone in LβT2 cells had no effect, whereas BMPR1B with BMPR2 up-regulated Fshb transcription, and this effect was further potentiated in the presence of BMP2 (Fig. 1B). Unexpectedly, the BMPR1B-Q249R mutant produced results equivalent to those of the wild-type BMPR1B receptor. Next, we examined potential functional differences between constitutively active forms of BMPR1B and BMPR1B-Q249R. The advantage of this approach is that it allowed us to examine functional changes in BMPR1B-Q249R that were independent of the particular ligand used in our experiments. As shown in Figure 2A, BMPR1A-QD, BMPR1B-QD, and BMPR1B-QD/Q249R all stimulated Fshb reporter activity when cotransfected with BMPR2 and did so to comparable extents.
To determine whether the results in LβT2 cells were cell specific, we assessed functionality of the constitutively active BMPR1A and BMPR1B receptors in HepG2 cells. Because Fshb reporters are inactive in nongonadotrophs, we used a validated BMP-responsive reporter, BREX4-luc . We previously observed that constitutively active ACVR1 and BMPR1A regulated this reporter in these cells without the need for BMPR2 coexpression (data not shown). BMPR1A-QD, BMPR1B-QD, and BMPR1B-QD/Q249R all stimulated BREX4-luc activity in HepG2 cells and did so equivalently (Fig. 2B). Therefore, no obvious functional impairment in BMPR1B-Q249R was noted in two distinct cellular contexts.
In LβT2 and HepG2 cells, we failed to detect functional changes in BMPR1B-Q249R. One study  used molecular modeling to predict the effects of the Q249R mutation on receptor function, and the results suggested that the mutated receptor might more stably interact with the inhibitory protein FKBP12. This would be predicted to impair signaling by the receptor to its downstream effectors, including SMAD1 and SMAD5 [11, 13, 14]. Therefore, we examined the relative abilities of BMPR1A-QD, BMPR1B-QD, and BMPR1B-QD/Q249R to stimulate SMAD1 and SMAD5 phosphorylation. Use of constitutively active forms of the receptors obviated the need for exogenous ligand treatment. CHO cells were transfected with combinations of the indicated receptors and FLAG-SMAD1 or FLAG-SMAD5. Western blots using a phospho-SMAD1/5/8 antibody showed that all three receptors were equivalent in their abilities to stimulate SMAD1 and SMAD5 phosphorylation (Fig. 3A, top panel [compare lanes 5, 6, 8, 9, 11, and 12 vs. lanes 2 and 3]). Reprobing of the blots with FLAG (second panel) and MYC (third panel) antibodies confirmed equivalent expression of the SMADs and receptors, respectively. Similar results were observed in LβT2 cells (Fig. 3B). Thus, the BMPR1B-Q249R receptor seemed capable of stimulating SMAD1/5 phosphorylation to the same extent as the wild-type BMPR1B.
To confirm that point mutations can, in fact, impair BMPR1B function in these assays, we generated a novel mutation in BMPR1B, D265A. The aspartic acid at position 265 is only 16 amino acids C-terminal to Q249R and is located within the L45 loop of the receptor. This receptor subdomain has been implicated in SMAD activation by type I receptors [35–37]. The analogous mutation in TGFBR1 (also known as ALK5), D266A, has been reported to impair the ability of the receptor to stimulate SMAD2 phosphorylation . Whereas BMPR1B-QD and BMPR1B-QD/Q249R stimulated SMAD1 phosphorylation, BMPR1B-QD/D265A was incapable of doing so (Fig. 3C). All three receptors were expressed at equivalent levels. Thus, our assays are able to detect impairments in receptor function.
Given that our analyses failed to show impairments in BMPR1B-Q249R function, we next examined whether the mutation affects receptor expression. Our initial analyses revealed equivalent expression of wild-type and Q249R forms of BMPR1B (Fig. 3, A–C, and data not shown). We noted that the mutation itself (CAG→CGG) introduces a novel CpG dinucleotide (underlined) that may be a substrate for DNA methylation. Although gene silencing is usually associated with methylation of cytosines in CpGs within promoter or enhancer regions, CpGs within coding regions might also be methylated and therefore have an effect on gene expression through their abilities to bind methyl DNA-binding proteins . Although there was no apparent effect of the mutation on expression in transfected cells (Fig. 3, A–C), the DNA used was propagated in Escherichia coli and would not be methylated at this or other CpGs. Therefore, we introduced methylated cytosines on both strands of the BMPR1B-Q249R construct by site-directed mutagenesis using primers methylated specifically at the sites of interest. The same procedure was followed using identical primers that lacked methylcytosines. The resulting PCR products were then purified and transfected directly into CHO cells, and their relative expression was measured by Western blot. Methylation of the amplified DNA was confirmed by bisulfite sequencing (data not shown). The methylated and unmethylated BMPR1B-Q249R constructs were expressed to equivalent extents (Fig. 3D); therefore, methylation at this site alone did not seem to affect receptor expression.
Although the data presented herein and previous findings  indicated that Acvr1, Bmpr1a, and Bmpr1b are expressed in LβT2 cells and can augment BMP2 actions when overexpressed in this cell line, the data did not definitely show whether BMP2 preferentially signals through one or more of these receptors. To confirm that ACVR1, BMPR1A, and/or BMPR1B is required for BMP2 signaling, we treated cells with compound C (also known as dorsomorphin), a small-molecule inhibitor of these three receptors . We treated LβT2 cells with 1 μM or 10 μM compound C 30 min before treatment with 25 ng/ml BMP2 or activin A for 1 h. At 1 μM and 10 μM, the inhibitor significantly impaired BMP2-stimulated SMAD1/5 phosphorylation but did not affect activin A-stimulated SMAD2 phosphorylation (Fig. 4A). Increasing the concentration to 20 μM antagonized the BMP2 effect more significantly but also had a small inhibitory effect on activin A (data not shown). Therefore, in subsequent analyses, we used 10 μM compound C. We next transfected cells with a murine Fshb reporter and treated them with BMP2 with or without activin A in the presence or absence of compound C. The inhibitor significantly impaired the independent and synergistic actions of BMP2 on Fshb transcription but did not significantly alter the activin A response or basal reporter activity (Fig. 4B). These data suggested a role for endogenous ACVR1, BMPR1A, and/or BMPR1B in BMP2 signaling in LβT2 cells.
We next knocked down expression of ACVR1, BMPR1A, and/or BMPR1B by RNAi to determine which might be the preferred receptor in this system. LβT2 cells were transfected with −846/+1 mouse Fshb-luc and siRNAs for ACVR1, BMPR1A, or BMPR1B, and they were then treated with 25 ng/ml BMP2 with or without 25 ng/ml activin A. We observed the synergistic actions of BMP2 and activin A under control conditions and in the presence of the ACVR1 or BMPR1B siRNAs (Fig. 5). In contrast, the BMPR1A siRNA significantly inhibited the synergistic actions of BMP2 and activin A on Fshb reporter activity but did not impair the independent activin A response. The BMPR1A siRNA did not significantly diminish the independent BMP2 effect in the context of this analysis, although the trend was in this direction. These data suggested that BMPR1A is the preferred BMP2 type I receptor in LβT2 cells.
We confirmed the functionality and specificity of the siRNAs used in these experiments. LβT2 cells were transfected with epitope-tagged expression vectors for ACVR1, BMPR1A, or BMPR1B that were predicted to be sensitive or resistant to their respective siRNAs based on sequence match or mismatch. That is, we introduced mutations that rendered the expression constructs perfect matches (in rat ACVR1 and human BMPR1A) or created mismatches (in murine BMPR1B) relative to the murine siRNAs used in the experiment shown in Figure 5. In all cases, mutations altered the nucleotide but not the amino acid sequences. As shown in Supplemental Figure S1, the siRNAs specifically impaired expression of their sequence-matched (“sensitive”) targets. The siRNAs directed against one receptor did not inhibit expression of the other receptors, and sequence-mismatched targets were resistant to their corresponding siRNAs. These data confirmed that the siRNA effects on receptor expression were sequence specific and did not reflect nonspecific or off-target effects.
Although the BMPR1A siRNA specifically impaired murine BMPR1A expression in LβT2 cells, we performed an additional control to show that decreases in Fshb reporter activity associated with the BMPR1A siRNA were attributable to receptor knockdown and not to some other off-target effect. We cotransfected LβT2 cells with −846/+1 mouse Fshb-luc and combinations of BMPR2 and siRNA-sensitive BMPR1A-QD or siRNA-resistant BMPR1A-QD along with control, BMPR1A, or BMPR1B siRNAs. The two forms of BMPR1A-QD equivalently stimulated reporter activity with BMPR2 (Supplemental Fig. S2). The BMPR1A, but not BMPR1B, siRNA inhibited the stimulatory effect of the sensitive, but not resistant, BMPR1A-QD expression vector, confirming that the BMPR1A siRNA effect was sequence specific.
Finally, having established BMPR1A as the relevant endogenous type I receptor in LβT2 cells, we examined with which endogenous type II it cooperates to mediate BMP2 activity. BMP2 can bind BMPR2, ACVR2, and ACVR2B [41–43], and we showed previously that all three of these receptors are expressed in LβT2 cells and in adult murine pituitary . We coexpressed BMPR1A-QD along with BMPR2, ACVR2, or ACVR2B expression vectors. None of the type II receptors had effects on their own, but all synergized with BMPR1A-QD to stimulate Fshb promoter activity (Fig. 6A). BMPR2 and ACVR2B had more pronounced effects than AVCR2. Next, we knocked down expression of the endogenous type II receptors using siRNAs. We cotransfected cells with the Fshb reporter and the indicated siRNAs, and we then treated them with 25 ng/ml BMP2 in the presence of the activin type I receptor inhibitor SB431542. Because we showed previously  that exogenous BMPs can synergize with endogenous activins in these cells, we needed to remove the potential confounding effects of activin signaling through ACVR2 or ACVR2B. Knockdown of BMPR2 or ACVR2 inhibited basal activity and the small (although not statistically significant) induction of Fshb transcription by BMP2 (Fig. 6B). The ACVR2B siRNA had no effect.
We reported previously that activin A and BMP2 synergistically regulate murine Fshb transcription . We postulated that BMP2 might signal preferentially through the type I receptor ACVR1 to mediate its effects. This was based on the observation that transfection of wild-type ACVR1, but not BMPR1A, with the type II receptor BMPR2 stimulated promoter-reporter activity alone and in the presence of BMP2. In contrast, constitutively active forms of ACVR1 and BMPR1A both synergized with BMPR2 to stimulate Fshb transcription. We subsequently discovered that our presumptive wild-type BMPR1A expression vector possessed a frameshift mutation, which prematurely truncated the kinase domain of the receptor. When we repeated the analysis using a validated full-length receptor, we observed that BMPR1A functioned similarly to ACVR1 (Fig. 1A). A third BMP type I receptor, BMPR1B, is also expressed in LβT2 cells [7, 18] and can similarly act in synergy with BMPR2 to regulate Fshb transcription. These observations suggest that one or more type I receptors may mediate BMP signaling in gonadotroph cells. Indeed, inhibition of ACVR1, BMPR1A, and BMPR1B with compound C confirmed a role for at least one of these receptors in BMP2-regulated SMAD1/5 phosphorylation and Fshb reporter activity (Fig. 4).
To more definitely establish which receptor(s) might be most critical, we used siRNAs to deplete endogenous expression of ACVR1, BMPR1A, or BMPR1B. Although all of the siRNAs were effective in depleting expression of their targets in sequence-specific fashion (Supplemental Figs. S1 and S2), only BMPR1A knockdown blocked the synergistic actions of BMP2 and activin A on Fshb transcription (Fig. 5). The BMPR1A siRNA did not hinder activin A signaling by itself. These observations suggest that the effect of the BMPR1A siRNA is principally through antagonism of BMP2 signaling. BMP2 can signal through multiple type I and type II receptors [16, 44], and there is evidence for functional redundancy of the different receptors. For example, in the absence of BMPR2, BMP2 and BMP4 can signal through ACVR2 . Herein, BMPR2 and ACVR2, but not ACVR2B, seemed to mediate the BMP2 response. Therefore, it is possible that ACVR1 and/or BMPR1B might compensate for the loss of BMPR1A, especially in light of the ability of these receptors to modulate Fshb transcription in overexpression experiments. However, the almost complete abrogation of BMP2-activin A synergism in the presence of the BMPR1A siRNA (Fig. 5) and the efficacy of ACVR1 and BMPR1B siRNAs in depleting their targets (Supplemental Fig. S1) suggest that neither ACVR1 nor BMPR1B compensates for the loss of BMPR1A in LβT2 cells, at least in these transient transfection assays. In light of these data and those with the type I receptor inhibitor (Fig. 4), we conclude that BMPR1A is the endogenous signal-propagating BMP2 receptor in these cells. Moreover, because overexpression of BMPR1B can potentiate the BMP2 response but knockdown of the endogenous receptor has no effect, we postulate that BMPR1B may be expressed at insufficient levels to propagate BMP2 signals in these cells.
Some Booroola (FecB) sheep that harbor a missense mutation (Q249R) in BMPR1B have increased FSH levels [22, 45, 46] in association with increased ovulation rates. Therefore, we hypothesized a priori that altered BMPR1B function might contribute to these phenotypes. The data presented herein failed to confirm this hypothesis on multiple levels. First (as already described), although it is expressed in gonadotroph cells, endogenous BMPR1B does not mediate BMP2 signaling. Second, the BMPR1B-Q249R receptor was functionally equivalent to the wild type in multiple assays. That is, the wild-type and mutant receptors stimulated two different reporters (Fshb-luc and BREx4-luc) in two different cell lines (LβT2 and HepG2) to equivalent extents (Figs. 1 and and2).2). Moreover, the receptors similarly stimulated SMAD1 and SMAD5 phosphorylation in CHO and LβT2 cells and were expressed at equivalent levels (Fig. 3, A–C). Most important, mutation of a nearby residue, D265A, completely abrogated BMPR1B-regulated SMAD1 phosphorylation (Fig. 3C), demonstrating the sensitivity of our experimental approach.
We also examined whether the Q249R mutation might affect receptor expression, perhaps through DNA methylation (Fig. 3D). However, the methylated and unmethylated Q249R receptors were expressed at equivalent levels, which is consistent with a previous study  showing equivalent Bmpr1b mRNA levels in wild-type and Booroola sheep ovaries.
In conclusion, the data presented herein show that BMP2 regulates murine Fshb subunit transcription independently and synergistically with activin A by signaling through the type I receptor BMPR1A and the type II receptors BMPR2 and ACVR2. Although ACVR1 and BMPR1B are expressed in LβT2 cells and in murine pituitary and both can act with BMPR2 to regulate Fshb promoter activity in overexpression analyses, neither seems necessary for BMP2 action, nor does either compensate for the loss of BMPR1A. We further show that the Q249R mutation observed in BMPR1B of Booroola sheep does not alter the ability of the receptor to stimulate SMAD1/5 phosphorylation or to activate target gene transcription in different cellular contexts. Future investigations will be required to confirm a role for BMPR1A in FSH regulation in vivo and to determine the nature of altered BMPR1B function in Booroola (FecB) sheep.
The authors acknowledge Drs. Mitsuyasu Kato, Joan Massague, Pamela Mellon, Patricia Morris, Tetsuro Watabe, and Teresa Woodruff for sharing reagents and cell lines. Dr. Paolete Soto and Beata Bak provided valuable feedback on a draft of the manuscript. Dr. Kathy Lee provided critical preliminary data.
1Supported by CIHR 86626 and NICHD HD047794 operating grants (to D.J.B.). C.C.H. was supported by a McGill University Faculty of Medicine Internal Studentship. D.J.B. is a Chercheur-boursier of the Fonds de la recherche en santé du Québec.