help button home button Endocrine Society Endocrinology
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Endocrinology, doi:10.1210/en.2006-0131
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
147/6/2744    most recent
Author Manuscript (PDF)
Right arrow Purchase Article
Right arrow View Shopping Cart
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow Request Copyright Permission
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Harrison, C. A.
Right arrow Articles by Robertson, D. M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Harrison, C. A.
Right arrow Articles by Robertson, D. M.
Endocrinology Vol. 147, No. 6 2744-2753
Copyright © 2006 by The Endocrine Society

Activin-A Binds Follistatin and Type II Receptors through Overlapping Binding Sites: Generation of Mutants with Isolated Binding Activities

Craig A. Harrison, Karen L. Chan and David M. Robertson

Prince Henry’s Institute of Medical Research, Clayton, Victoria 3168, Australia

Address all correspondence and requests for reprints to: Craig A. Harrison, Prince Henry’s Institute of Medical Research, 246 Clayton Road, Clayton, Victoria 3168, Australia. E-mail: craig.harrison{at}princehenrys.org.


   Abstract
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 References
 
Follistatin is a potent extracellular antagonist of members of the TGFß superfamily that use activin type II receptors (ActRII/IIB) as part of their signaling complex. A recent crystallographic study indicates that follistatin contacts activin-A residues at both the type I (ALK4) and type II receptor binding interfaces. However, the relative contribution of these two sites on human activin-A to follistatin binding has not been determined. Residues at these sites were mutated to alanine and mutants were screened for their ability to bind follistatin and ActRII and induce FSH secretion from a gonadotrope cell line. Despite extensive mutagenesis across the type I receptor interface, activin-A affinity for follistatin was not significantly diminished. In contrast, mutagenesis of residues at the type II binding interface had pronounced effects on activin’s interaction with follistatin. In particular, residues Leu92, Tyr94, Ile100, and Lys102 were critical for high-affinity follistatin binding. Interestingly, mutation of another primary determinant of ActRII/IIB binding, Ser90, did not affect follistatin affinity, suggesting that the interaction surfaces for type II receptors and follistatin were overlapping but not identical. In support, mutation of Asp95, on the opposite edge of the common ActRII/follistatin interface, was disruptive for follistatin binding without affecting ActRII/IIB interactions. Activin-S90A was able to compete with wild-type activin for follistatin binding, whereas activin-D95A, due to its 8-fold lower affinity for follistatin, is a potent activin agonist. These reagents could be used to modulate follistatin antagonism of activin and related ligands in processes such as cancer, wound healing, and reproduction.


   Introduction
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 References
 
ACTIVINS, MEMBERS OF the TGFß superfamily, are homo- or heterodimeric cysteine knot proteins composed of related ß-subunits. Four ß-subunit genes (ßA, ßB, ßC, and ßE) have been described in humans, but only the ßA-ßA (activin-A), ßB-ßB (activin B), and ßA-ßB (activin-AB) dimers have been extensively characterized (1). Activins were initially recognized for their important roles in stimulating the synthesis and secretion of FSH by the anterior pituitary (2); however, it is now recognized that they act as pleiotropic hormones/growth factors with powerful actions on erythropoiesis (3), liver proliferation (4), immune function (5), bone formation (6), angiogenesis (7), neuronal survival (8), skin morphogenesis, and cutaneous wound repair (9).

Activins and other members of the TGFß superfamily exert their biological effects by interacting with two types of transmembrane receptors (types I and II), called receptor serine kinases (10). The relatively small number of receptor serine kinases (seven type I and five type II) necessitates that the receptors have multiple specificities. The activin type II receptors (ActRII) are particularly promiscuous, bridging subfamilies by binding numerous ligands, including activin-A, activin-B, bone morphogenetic protein (BMP)-2, BMP-4, BMP-7, myostatin, growth and differentiation factor-11, and nodal (11). Activin binding to either of its two type II receptors, ActRII or ActRIIB (12), is followed by recruitment, phosphorylation, and activation of the type I receptor, activin-like kinase-4, to initiate signaling via intracellular Smad proteins.

In the crystal structure of the ActRIIB extracellular domain bound to activin-A (13) as well as that of ActRII bound to BMP-7 (11), the ActRIIs make contact with the convex outer face of the finger region of each ligand. The activin-A/ActRIIB interface involves hydrophobic and ionic/polar residues in finger 1 (Phe17, Ile30, AL31, Pro32, His36) and finger 2 (Arg87, Pro88, Ser90, Leu92, Tyr94, Ile100, Lys102, Glu111) of activin-A (Fig. 1Go) (13). There is no crystal structure available for activin-A bound to activin-like kinase-4. However, using a mutagenesis approach, we recently identified activin residues Met91, Ile105, and Met108 on the concave surface of finger 2 and Ser60 and Ile63 in the {alpha}-helix that were important for binding to the type I receptor (14, 15).


Figure 1
View larger version (40K):
[in this window]
[in a new window]
  		FIG. 1. Schematic of the activin-A dimer. One chain only is emphasized. Residues mutated in this study (black outline) and those involved in interactions with type I (I) and type II (II) receptors are indicated. Figure adapted, with permission, from Ref.38 .

 
Activins are secreted in their processed, biologically active (25 kDa) form. The ability of activins to assemble their receptor complex, however, is regulated by a number of extracellular binding proteins, chief among them follistatin (16, 17). Follistatin is a glycoprotein that exists in two splice variants: follistatin-288 (FS288), which associates with the cell surface and is involved in activin internalization and degradation (18); and FS315, which is the predominant circulatory form (16). Both forms of follistatin bind activin-A with high affinity (19), approaching irreversibility due to slow dissociation rates, and thereby restrict activin bioavailability (20).

A recent study examining the activin/follistatin crystal structure indicated that follistatin contacts two contiguous surfaces on activin-A (21). One surface overlaps the type II receptor binding site and includes residues on the convex outer side of the activin ß-strands (Ile30, AL31, Pro32, Leu92, Tyr94, Ile100, and Lys102) and in the ß-strand fingertips (Asp95-Asn99). The second surface maps to the proposed type I receptor binding site and is formed by activin residues from the concave ß-strand of one subunit (Trp25, Trp28, Met91, Tyr93, and Ile105) and the helical wrist region of the second subunit (His47, Ile48, Gly50, Thr51, Ser52, Phe58, Thr61, and His65). It was predicted that in burying these residues, follistatin would inhibit both types I and II receptor binding (21). However, this study was unable to establish whether the function of these receptor binding sites was neutralized.

In the current study, we sought to verify and extend the crystal structure data with particular reference to the importance of the type I and type II interfaces for follistatin binding. Residues at these sites and on surrounding surfaces were mutated to alanine. Mutants were screened for their ability to bind follistatin and ActRII and induce FSH secretion from a pituitary gonadotrope cell line. Using these approaches we confirmed that there is a partial overlap of the follistatin and ActRII binding sites on finger 2 of activin-A. In addition, we generated several mutants that dissociate these binding properties of activin.


   Materials and Methods
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 References
 
Mutagenesis and protein expression
An overlapping PCR strategy was used to incorporate mutations in the mature region of the full-length human activin-ßA cDNA. First, a unique NheI site was introduced just 5' of the mature region in the activin-ßA construct, allowing us to make and subclone mutant PCR products (~600 bp) spanning only the mature region of activin. Primers were constructed to incorporate a 5' NheI site and a 3' XhoI site for subcloning back into the full-length construct. Gel-purified PCR products were digested with NheI and XhoI and then subcloned into NheI/XhoI-digested activin-A vector. For each construct, the mutated amino terminal mature region was confirmed by DNA sequencing.

High-level production of mutant activins was achieved using polyethyleneimine (PEI) to transiently transfect HEK293T cells (14). HEK293T cells were plated at 2 x 107 cells per 15-cm plate coated with poly-D-lysine. After 16 h, cells were transfected using the PEI transfection reagent. Briefly, for 15-cm plates, 24 µg activin mutant DNA was diluted to 1.2 ml with serum-free DMEM. PEI (10 mg/ml stock) was diluted to 1 mg/ml, and 36 µl were added to the DNA solution, vortexed, and incubated at room temperature for 10 min. During the incubation, cells were washed with DMEM, and 12 ml serum-free DMEM were added to each plate. After 10 min the DNA/PEI complexes were added directly to the plates and incubated for 72 h at 37 C in 5% CO2.

Activin A immunofluorometric assay (IFMA)
To measure the concentration of mutant activins in the conditioned media of HEK293T cells, an activin IFMA was used. Briefly, a 96-well plate (C12 MaxiSorb, Nunc, Glostrup, Denmark) was coated with a monoclonal antibody (E4, Oxford BioInnovation, Oxfordsire, UK) directed against the activin ßA-subunit (750 ng/well) and blocked with 50 mM Tris-HCl (pH 7.4) containing 1% BSA. Activin A standard or activin mutant-conditioned media (100 µl) was added to the wells in duplicate and incubated overnight in the presence of 6% H2O2 (50 µl) and 0.05 M Tris/HCl (pH 7.5) containing 0.154 M NaCl, 0.01% Tween 40, 0.2% sodium dodecyl sulfate, 0.5% BSA, 0.1% sodium azide, 20 µM diethylenetriaminepentaacetic acid, and Amaranth dye (100 µl). The following day, the plates were washed and incubated for 2 h with biotinylated E4 antibody (100 µl, Oxford BioInnovation). Plates were then washed and incubated for 1 h with Europium-labeled streptavidin (100 µl, 13 ng; Wallac, Turku, Finland). The plates were washed and enhancement solution (200 µl, Wallac) added and the plates read on a time-resolved fluorometer (Victor; Wallac). The data were analyzed using Multicalc software (Wallac). The working range of the assay is 0.03–3 ng/well. The sensitivity of the assay is 0.03 ng/well. Parallelism was observed between response curves in the IFMA derived from serial dilutions of activin A standard and activin-M108A- (14) conditioned medium diluted in culture medium.

The epitope recognized by the E4 monoclonal antibody (residues 88–93) overlaps the regions of interest on the ßA-subunit. To estimate the concentration of the activin mutants R87A-P88A, S90A, M91A, L92A, and Y93A, which fall within this epitopic region, conditioned media from transfected 293T cells were partially purified by heparin Sepharose affinity chromatography as previously described (22). Eluted proteins were separated by SDS-PAGE and silver stained. The amount of activin mutant present was determined by densitometric comparison with activin A standards loaded on the same gel. As a control, the concentration of activin mutant M108A, which is recognized by the E4 antibody, was also determined densitometrically. The densitometric analysis consistently gave results that were ± 10% of the concentrations obtained in the activin IFMA.

Follistatin binding assay
The follistatin binding assay used a polyclonal antiserum (no. 204C) raised in a rabbit against 35 kDa bovine follistatin that binds the protein with high affinity but does not interfere with its binding to activin (Phillips, D., unpublished observations). The recombinant human FS288 used as competitor was a heterogeneous preparation of nonglycosylated, monoglycosylated and diglycosylated forms of 31, 35, and 40 kDa recombinant human follistatin (Biotech Australia, Sydney, Australia).

Recombinant human FS288 (8 ng) was incubated at room temperature overnight with a constant amount of 125I-activin-A [10,000 cpm, iodinated by the chloramine T procedure (23)] and varying concentrations of mutant or wild-type activin-A (0.8–20 ng) in the presence of the antifollistatin antibody (1:6000 no. 204C final dilution) in a final volume of 500 µl. After this time, goat antirabbit IgG (100 µl) was added and incubated for 30 min. Polyethylene glycol (6% final concentration in 0.9% NaCl at 4 C) was added, and the precipitate was collected by centrifugation after 30 min incubation at 4 C. The supernatant was aspirated and the precipitated counts measured in a {gamma}-counter.

ActRII binding assay
HEK293T cells were plated at 2 x 105 cells/well in 24-well plates coated with poly-D-lysine. The following day, the cells were transfected with 100 ng ActRII cDNA using the Lipofectamine transfection reagent (10 µl transfection reagent per 500 µl media, Invitrogen, Carlsbad, CA) and incubated for 48 h at 37 C. Cells were washed in binding buffer (DMEM + 0.1% BSA). Binding buffer (100 µl), unlabeled competitor (50 µl, activin-A or activin-A mutant at various dilutions in binding buffer as indicated; see Fig. 3Go), and 125I-activin-A (50 µl, 40,000 cpm/well) were then added. Plates were incubated for 3 h at room temperature, the cells washed in PBS, and cells solubilized in 1% Triton X-100. Radioactivity was measured using a {gamma}-counter. Binding data were analyzed using the Prism program (version 2.0; GraphPad Software, San Diego, CA).


Figure 3
View larger version (21K):
[in this window]
[in a new window]
  		FIG. 3. ActRII binding assay. 293T cells were transfected with ActRII and 48 h later subjected to competition binding as described in Materials and Methods. Displacement curves for activin-A analogs with mutations at the type II receptor interface (A), surrounding the type II receptor interface (B), and at the proposed type I receptor interface (C) are shown. The amount of bound 125I-activin-A was determined in triplicate for each experiment, and the values are the mean ± SD. The experiment was repeated three times. The displacement curve generated in the presence of unlabeled activin-A is shown for comparison (closed squares; A–C).

 
Activin in vitro bioassay using a mouse pituitary gonadotrope cell line (LßT2)
The in vitro biological activity of the mutant activins was assessed by their ability to promote the release of FSH from a mouse pituitary gonadotrope cell line (LßT2). LßT2 cells were plated in 48-well plates at a density of 250,000 cells/well. The cells were allowed to recover for 24 h in DMEM supplemented with 10% fetal calf serum. The cells were then washed with DMEM + 0.2% fetal calf serum and treated with increasing doses of activin or activin mutants for 24 h in the same media.

Rat FSH IFMA
The secretion of FSH was quantified using the supersensitive rat IFMA (24) using reagents kindly provided by Arijan Grootenhuis (N.V. Organon, Oss, The Netherlands). Briefly, 96-well plates (Nunc) were coated with a monoclonal antibody (MCA {alpha}FSH 56A) directed against human FSH. Rat FSH reference preparation (2–2000 pg/well, RP-2, National Institute of Diabetes and Digestive and Kidney Diseases) or conditioned media from treated LßT2 cells were added to the plates in triplicate and incubated overnight. The following day the plates were washed and incubated for 3 h with a biotinylated anti-FSH antibody (biot-PCA29-7515) before incubation with Europium-labeled streptavidin for 1 h. The fluorescence levels were determined using a fluorometer (Victor; Wallac). Mean activin-induced FSH release in the presence or absence of follistatin and activin-S90A was compared using a paired t test with a confidence interval of 95% and P < 0.05 considered statistically significant.


   Results
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 References
 
Selection of amino acid residues on activin-A for mutagenesis
Follistatin is a potent antagonist of members of the TGFß superfamily that use ActRIIs (ActRII/IIB) as part of their signaling complex. The affinity of follistatin for TGFß ligands reflects their affinity for ActRII/IIB, suggesting that the binding sites overlap. Thompson et al. (13) identified a discrete epitope for high-affinity ActRIIB binding on the convex outer surface of the fingers of activin-A (Fig. 1Go), and residues within this epitope were the initial targets for mutagenesis in this study. Residues were substituted by alanine, although future studies may look to incorporate charged residues because these would be predicted to be more disruptive. Residues Phe17, Ser90, Leu92, Tyr94, Ile100, and Lys102 were mutated to alanine, and the effects of the mutations on follistatin and ActRII binding were assessed. Arg87, Pro88, and Asp104 mutations were incorporated in cassettes (K85A-M89A and K103A-N107A) and because these variants retained the ability to bind follistatin (data not shown), the individual point mutants were not generated. Secondary mutations targeted residues in the broader region implicated by the activin-A/ActRIIB complex structure focusing, in particular, on residues in finger 1 (Ser19, Lys21, Ser33, and Tyr35) and the fingertips of finger 2 (Asp95, Asp96, Gly97, Gln98, and Asn99). Finally, because the proposed type I receptor binding interface of activin-A was envisaged to be well removed from the follistatin binding site, a series of {alpha}-helical mutants were screened for their ability to bind follistatin (Fig. 1Go). The recently solved activin/follistatin crystal structure implicated most of the residues mutated in this study in follistatin binding (21). Indeed, the unexpected demonstration that the follistatin N-terminal domain occupies the type I receptor binding site meant that even the {alpha}-helical mutants could potentially disrupt follistatin binding. In all, of the 23 mutants examined, 17 incorporated residues shown to be involved in follistatin binding (21).

Synthesis and secretion of wild-type and mutant activin-A by HEK293T cells
Activin-A was synthesized in HEK293T cells using a transient transfection approach. Conditioned medium from 293T cells transfected with the activin ßA cDNA contained mature, dimeric activin-A protein as previously described (14). The concentration of activin-A in the conditioned media of transiently transfected 293T cells after 72 h was generally 5–10 nM based on an activin-A IFMA. All the activin mutants described in this study were expressed at levels comparable with wild-type activin-A (data not shown). In addition to the mutants described, several attempts were made to generate mutants W25A, N26A-D27A, W28A, and I30A-P32A. Although expressed, the levels of these mutants (~100 pM) were too low for testing in subsequent biological assays.

Follistatin binding of activin-A mutants
The activin mutants generated in this study were separated into three groups and assessed for their ability to displace 125I-activin-A from follistatin in a radioligand binding assay. Group A residues were those at the type II receptor binding interface, and, in general, mutation of these residues had the most profound effects on follistatin binding (Fig. 2AGo and Table 1Go). Of this group, only activin-S90A retained an affinity for follistatin (90 pM) comparable with wild-type activin-A (120 pM). Mutants F17A (300 pM), L92A (600 pM), K102E (600 pM), Y94A (2600 pM), and I100A (2600 pM) displayed progressively lower affinity for follistatin (Fig. 2AGo and Table 1Go). These results support the activin/follistatin crystal structure data that indicated there was a partial overlap of the ActRII and follistatin binding sites (13). Group B residues were either peripherally implicated in receptor binding (11, 13) or surrounded the type II binding interface. This group included residues in finger 1 (Ser19, Lys21, Ser33, and Tyr35) and the fingertips of finger 2 (Asp95, Asp96, Gly97, Gln98, and Asn99). Of the resultant mutants, only activin-D95A had a reduced affinity for follistatin (800 pM) (Fig. 2BGo and Table 1Go). This was surprising, given that six of these residues are apparently buried by follistatin in the crystal structure (21). The final group of mutants, group C, targeted residues at the proposed type I binding interface (Fig. 2CGo). Despite most of these mutants (P45A-I48A, S52A-S54A, S55A-L56A, S60A-T61A, V62A-I63A, H65A/M68A, R69A-H71A) having multiple residues substituted to alanine, there was very little effect on follistatin binding (Fig. 2CGo and Table 1Go). Similarly, activin-M108A, which had previously been shown to completely disrupt type I receptor binding (14), retained high affinity for follistatin (85 pM). Again these results were difficult to reconcile with the crystal structure data, which indicated that the majority of residues at the proposed type I receptor binding interface were covered by the N-terminal domain of follistatin (21). Of the group C mutants, only M91A (3200 pM) and Y93A (950 pM) had reduced affinity for follistatin (Fig. 2CGo and Table 1Go). Together these results suggest that follistatin primarily contacts activin residues on the convex outer surface of the fingers.


Figure 2
View larger version (24K):
[in this window]
[in a new window]
  		FIG. 2. Follistatin binding assay. Recombinant human FS-288 (8 ng) was incubated with a constant amount of 125I-activin-A (10,000 cpm) and varying concentrations of mutant or wild-type activin-A as described in Materials and Methods. Activin mutants were separated into those at the type II receptor interface (A), those peripherally implicated in type II receptor binding (B), and those at the proposed type I receptor interface (C). Follistatin-activin complexes were immunoprecipitated to separate bound from free ligand. Radioactivity in the precipitate was determined using a {gamma}-counter. Each point of the displacement curves is the mean ± SD of three measurements. The experiment was repeated three times. The displacement curve generated in the presence of unlabeled activin-A is shown for comparison (closed squares; A–C).

 

View this table:
[in this window]
[in a new window]
  		TABLE 1. Follistatin and ActRII binding properties of activin-A mutants

 
Receptor binding of activin-A mutants
To determine whether the activin-A mutants retained the ability to bind ActRII competition binding assays were performed in which the mutants were assessed for their ability to displace 125I-activin-A from 293T cells transfected with ActRII. The IC50 of activin for ActRII in these studies was determined to be approximately 200 pM (Table 1Go and Fig. 3Go). Mutation of residues at the type II binding interface reduced activin’s affinity for ActRII. Mutants F17A (600 pM) and Y94A (2600 pM) had 3- and 13-fold lower affinity for ActRII, respectively, whereas the ability of mutants S90A, L92A, I100A, and K102E to displace 125I-activin-A from 293T cells transfected with ActRII was severely compromised (Fig. 3AGo and Table 1Go). Mutation of residues surrounding the type II binding interface (group B residues) had little effect on activin binding to ActRII (Fig. 3BGo). Only mutants S19A/K21A (490 pM) and S33A/Y35A (900 pM) had reduced affinity, compared with wild-type activin (200 pM) (Fig. 3BGo and Table 1Go). Significantly, activin-D95A retained wild-type-like affinity for ActRII (260 pM) despite having a reduced affinity for follistatin. As expected, mutations of most of the residues at the proposed type I binding interface (group C residues) had little impact on the ability of activin to bind ActRII (Fig. 3CGo). However, mutants M91A and Y93A did display significantly reduced affinity (>10 nM) for ActRII (Fig. 3CGo and Table 1Go). These mutants, particularly activin-M91A, also had low affinity for follistatin (Fig. 2CGo). Together these results suggest that, despite their location on the concave inner surface of finger 2, mutation of Met91 or Tyr93 actually disrupts the common binding surface used by ActRII and follistatin on the convex outer surface of finger 2.

Separation of the follistatin and ActRII binding properties of activin-A
A major aim of this study was to separate the ActRII and follistatin binding properties of activin-A and, in the process, develop reagents that could modulate follistatin antagonism of TGFß ligands. The activin-A/ActRIIB binding interface is fairly discrete, involving 13 residues on the convex surface of the fingers of activin-A (13). In contrast, the activin-A/follistatin binding interface is more extensive with two follistatin molecules wrapped around the activin dimer contacting 39 of 116 residues in each subunit (21). As such, it was envisaged that it would be relatively easy to disrupt follistatin binding but retaining high-affinity ActRII binding. However, for most of the activin-A mutants examined, there was an excellent correlation between affinity for follistatin and affinity for ActRII (Fig. 4Go). The only mutants that deviated significantly from this relationship were activin-D95A and activin-S90A (Fig. 4Go). Both these residues are on the convex outer surface of finger 2, highlighting the importance of this region for binding protein interactions. Asp95 is located in the distal portion of finger 2 and has not been implicated in type II receptor binding (11, 13). However, mutation of this residue to alanine leads to an 8-fold decrease in activin’s affinity for follistatin. In contrast, mutation of Ser90, which is at the center of the type II binding interface, is disruptive for receptor binding but has no effect on follistatin binding. Thus, we made use of the fact that the interaction surfaces for ActRII and follistatin are overlapping but not identical to dissociate these binding properties of activin-A.


Figure 4
View larger version (13K):
[in this window]
[in a new window]
  		FIG. 4. IC50 values of activin and activin mutants for follistatin and ActRII binding sites. The IC50 values for each activin-A mutant binding to both follistatin (x-axis) and ActRII (y-axis) were determined using Prism GraphPad software. The labels for several mutants are abbreviated, including S19A/K21A (S19A), S33A/Y35A (S33A), P45A/S46A/H47A/I48A (P45A), S52A/G53A/S54A (S52A), S55A/L56A (S55A), S60A/T61A (S60A), V62A/I63A (V62A), H65A/M68A (H65A), and R69A/G70A/H71A (R69A).

 
Biological activity of activin-D95A
The mutant that appeared to have the most favorable agonist properties based on the initial competition binding assays was the D95A analog. This was due to its low affinity for follistatin (~15% that of wild-type activin) and its high affinity for ActRII (~260 pM). To determine whether a decreased affinity for follistatin would correspond to an increased biological activity, we assessed the ability of activin-D95A to promote the release of FSH from LßT2 gonadotrope cells, in both the absence and presence of follistatin. In the absence of exogenous follistatin, the ability of activin-D95A to stimulate the secretion of FSH was identical with that of wild-type activin-A (Fig. 5AGo). In contrast, activin-S90A and activin-I100A, which display low affinity for ActRII/IIB, had reduced activity in this assay (Fig. 5AGo). Treatment of LßT2 cells with a submaximal dose of activin or activin-D95A (100 pM) in the presence of increasing concentrations of follistatin led to a dose-specific decrease in FSH secretion (Fig. 5BGo). However, exogenous follistatin had a reduced capacity to antagonize the biological activity of activin-D95A (IC50 360 ± 49 pM), compared with the wild-type protein (IC50 178 ± 19 pM). In addition, maximal suppression of activin-A-induced FSH release was observed with a follistatin to activin molar ratio of 4:1, whereas a follistatin to D95A molar ratio of 10:1 was required to fully inhibit the D95A-induced response (Fig. 5BGo). These results suggest that activin-D95A, due to decreased binding and clearance by follistatin, would be at least 2-fold more potent than wild-type activin in vivo.


Figure 5
View larger version (15K):
[in this window]
[in a new window]
  		FIG. 5. In vitro bioactivity of wild-type activin-A and the activin-D95A mutant in the absence and presence of follistatin. LßT2 cells were incubated with increasing concentrations of wild-type activin-A (open circles), activin-D95A (closed circles), activin-S90A (open squares), or activin-I100A (closed squares) for 24 h (A) or were treated with a submaximal dose (100 pM) of wild-type activin-A (open circles) or activin-D95A (closed circles) and increasing concentrations of follistatin as indicated (B). In each experiment medium was collected and FSH was measured by IFMA as described in Materials and Methods. The values are the mean ± SEM from a representative experiment.

 
Activin-S90A competes with wild type activin for follistatin binding
Mechanisms to liberate activin-A from follistatin inhibition would increase the effective concentration of activin available for receptor binding and signal transduction. Therefore, the activin-S90A mutant, with its high affinity for follistatin (90 pM) but low affinity for ActRII (>10 nM), was assessed for its ability to compete with wild-type activin-A for binding to follistatin (Fig. 6Go). LßT2 cells secreted high levels of FSH (2.7 ng/well) in response to activin-A treatment (Fig. 6Go, black bar), and this secretion was blocked in the presence of an inhibitory dose of follistatin (Fig. 6Go, white bar). However, when activin-S90A was added to LßT2 cells treated with a constant amount of activin/follistatin complex, it stimulated FSH release in a concentration-dependent manner (Fig. 6Go, gray bars). The bioactivity of activin-S90A under these conditions was much greater than that attributable to the S90A mutant alone (Fig. 6Go, hatched bar). This indicated that activin-S90A could compete with wild-type activin for binding to follistatin and thus increase the effective activin concentration. As a control, activin-I100A, which exhibits disrupted follistatin (IC50 2.6 nM) and receptor (IC50 > 10 nM) binding, was tested. This mutant was unable to compete with wild-type activin-A for follistatin binding and therefore did not increase FSH release in this assay (data not shown).


Figure 6
View larger version (14K):
[in this window]
[in a new window]
  		FIG. 6. FSH release from LßT2 cells treated with activin-S90A in the presence and absence of wild-type activin-A and follistatin. LßT2 cells were treated with activin-A alone (100 pM, black bar), activin-A in the presence of an inhibitory dose of follistatin (500 pM, white bar), or the activin·follistatin complex in the presence of increasing concentrations of activin-S90A (gray bars). Medium was collected and FSH was measured by IFMA as described in Materials and Methods. The effect on FSH release of a high dose of activin-S90A (800 pM) is shown for comparison (hatched bar). **, P < 0.01. n.s., Not significant.

 

   Discussion
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
References
 
Follistatin, a secreted glycoprotein, modulates the active local concentration and consequently the extent of action of members of the TGFß superfamily including activin-A (25), activin-B (26), myostatin (27), growth and differentiation factor-11 (28), and numerous BMP ligands (29). Follistatin knockout mice are small and present with skeletal and muscle defects. They also display improper tooth and whisker development, and their skin is taut and shiny (30). The mice fail to breathe and die soon after birth. These phenotypic effects are consistent with follistatin regulating multiple TGFß ligands in vivo. To provide a structural basis for understanding the specificity and function of this extracellular antagonist, activin-A mutants were generated and assessed for their ability to bind follistatin.

Mutagenesis of residues at the type II receptor binding interface, or knuckle region, of activin-A (Fig. 1Go) had pronounced effects on follistatin binding. In particular, residues Tyr94 and Ile100 and, to a lesser extent, Leu92 and Lys102 are critical for high-affinity follistatin binding. These residues form a contiguous surface toward the distal portion of finger 2 of activin-A. It is likely that Ile30 and Ala31 from the convex portion of finger 1 are also important for this interaction; however, mutagenesis of these residues effected protein expression. Interestingly, mutation of another primary determinant of ActRII/IIB binding, Ser90, did not affect follistatin binding, indicating that the interaction surfaces for type II receptors and follistatin are overlapping but not identical.

This suggested that residues specific for follistatin binding may reside outside the common type II receptor/follistatin binding interface. Consequently, residues through the ß-strand fingertips (Asp95-Asn99), the helical regions (Pro45-Ile48; Ser52-Ser56; Ser60-His65; and Met68-His71) and the concave surface of the fingers (Met91, Tyr93 and Met108) of activin-A were mutated to alanine. Within the fingertip, mutation of Asp95 resulted in an appreciably lower affinity for follistatin (800 pM) than the wild-type protein (120 pM). In the activin/ActRIIB crystal structure (13), Asp95 is in close proximity to Tyr94 and Ile100, both critical residues for follistatin binding. Interestingly, mutation of surrounding residues (Asp96, Gly97, Gln98, or Asn99) had no effect on follistatin binding. These results suggest that follistatin wraps around the fingertips of activin-A but that direct contacts in this region are minimal.

Similarly, the majority of the wrist region of activin-A, formed by residues from the concave ß-strand of one subunit (Trp25, Trp28, Met91, Tyr93, and Ile105) and the helical regions of the second subunit (His47-His71), appears to be unnecessary for high-affinity follistatin binding. Mutations through the {alpha}-helical region, despite incorporating between two and four amino acid substitutions, had little effect on follistatin binding (Fig. 2CGo). Indeed, the greatest decrease in affinity for follistatin among these mutants was only 2-fold (observed for mutants P45A/S46A/H47A/I48A, S52A/G53A/S54A, and V62A/I63A). These results indicate that residues within the helical region of activin are not primary determinants for binding to follistatin. It is also difficult to ascertain what role, if any, residues on the concave surface of the fingers play in follistatin binding. Mutation to alanine of residues Met91 and Tyr93 disrupts the juxtaposed binding surface used by ActRII and follistatin on the convex surface of finger 2, thereby decreasing activin’s affinity for both these binding proteins. The ability of these residues to directly contact follistatin has not been determined. However, activin-M108A, which we have previously shown to completely disrupt type I receptor binding (14), has no effect on follistatin binding. Together, the mutagenesis results indicate that high-affinity follistatin binding is determined by residues on the convex outer surface of the fingers of activin-A, with minimal contributions from surrounding surfaces.

The functional analysis of the importance of individual amino acids to follistatin binding coincided with the resolution of the activin/follistatin crystal structure (21). In this complex, two FS-288 molecules wrap around the activin dimer, burying 3000 Å of the ligand surface and contacting 39 of 116 residues in each subunit. One of the two primary contact sites is formed by hydrophobic residues on the convex surface of activin-A (Ile30, AL31, Pro32, Leu92, Tyr94, Ile100, and Lys102) packing against hydrophobic follistatin residues in follistatin domain (FSD)1 and FSD2 (21). This site corresponds to the high-affinity binding site identified by mutagenesis.

Previously it had been assumed that follistatin acts by inhibiting type II, but not type I, receptor binding to ligands (22, 31). However, the crystal structure revealed that the N-terminal domain of follistatin, which is required for biological activity (20), mimics the BMP type I receptor and occupies the proposed type I binding surface of activin-A (21). This surface comprises activin residues from the concave ß strand fingers of one subunit (Trp25, Trp28, Met91, Tyr93, and Ile105) and the helical wrist region of the second subunit (His47, Ile48, Gly50, Thr51, Ser52, Phe58, Thr61, and His65). Residues present in activin ligands, but absent from BMP ligands, map to this region and have been proposed to contribute to the specificity and affinity of the follistatin interaction. Our inability to generate mutants within the wrist region of activin that directly disrupt follistatin binding; however, questions the importance of this interaction site. Follistatin may well loop around the activin dimer with the N-terminal domain occupying the type I receptor interface and FSD2 burying residues in the fingertips (21); however, our mutagenesis strategy has indicated that these regions of activin-A are unnecessary for high-affinity follistatin binding. In support of this conclusion, follistatin binds myostatin with an affinity (580 pM) approaching that for activin-A (50–280 pM) and equivalent to that for activin-B (26, 27, 29), yet none of the activin-A helical residues implicated in follistatin binding are conserved in myostatin. Furthermore, we recently generated an activin-A/activin-C chimera in which residues 46–78 of activin-A were replaced with the corresponding residues from activin-C (32). Despite low conservation through this region (11 of 33 residues), the resultant chimera had only a 3-fold reduction in its affinity for follistatin (32). Finally, BMP ligands, which use a different signaling mechanism to activin by binding initially to high affinity type I receptors, can simultaneously bind follistatin suggesting that follistatin does not block the type I receptor binding site (31, 33).

A major aim of this study was to separate the ActRII and follistatin binding properties of activin-A and, in the process, to develop reagents that could modulate follistatin antagonism of TGFß ligands. The activin mutants S90A and D95A fulfill this aim and show promise as first-generation activin agonists. Activin-S90A maintains high-affinity follistatin binding despite a marked decrease in its affinity for ActRII (Table 1Go). In the LßT2 FSH assay, activin-S90A effectively competed for follistatin binding thus releasing wild-type activin from endogenous or exogenous complexes (Fig. 6Go). This activity is reminiscent of that previously attributed to activin-K102E (22). However, the wild-type-like affinity of activin-S90A for follistatin (90 pM) is 7-fold higher than that observed for activin-K102E (600 pM), suggesting that the S90A mutant would be considerably more effective at releasing bound activin-A from the follistatin complex. The attractiveness of activin-S90A is that it should be capable of differentially releasing multiple ligands from follistatin blockade. Indeed, whereas activin-S90A increases the effective activin-A concentration, it should be even more potent in freeing those ligands (e.g. activin-B, myostatin, and BMPs) with lower affinity for follistatin.

Of the 23 mutants examined in this study, only activin-D95A displayed an appreciable reduction in follistatin affinity (8-fold) but maintaining high-affinity binding to the ActRII (Figs. 2AGo and 3AGo). In the LßT2 assay, follistatin had a reduced capacity to antagonize the biological activity of activin-D95A (IC50 360 pM), compared with wild-type activin-A (IC50 180 pM). One curious aspect of these experiments was that the ability of follistatin to antagonize activin-D95A was not further diminished. Based solely on binding data (Fig. 2Go), we expected an 8-fold reduction in follistatin activity. These results replicate a previous study that compared follistatin’s interaction with activin-A and activin-B (26). In this report, activin-B was approximately 10-fold less potent than activin-A for binding to follistatin, and yet it took only 3-fold more follistatin to neutralize activin-B, compared with activin-A (26). Whatever the reason for these differences between binding and biological activity, a 2- to 3-fold reduction in follistatin’s ability to antagonize activin-D95A could translate to important biological consequences if this agonist is used in vivo.

There is currently considerable interest in developing clinical applications for members of the TGFß superfamily. For example, recombinant human BMP-2 has been successfully tested for the treatment of open tibial fractures (34), whereas topical application of TGFß3 has proven beneficial in the treatment of pressure ulcers (35). Clinical trials using activin have yet to commence; however, its efficacy in animal disease models (36, 37) suggests it may also be beneficial for the treatment of related human conditions. To be effective, mechanisms must be developed to maintain these ligands in their active state for as long as possible. Our development of reagents to sequester or reduce the activity of the endogenous TGFß superfamily antagonist follistatin will aid in this endeavor.


   Acknowledgments
 
The authors thank Yogeshwar Makanji and Sara Gruskin for technical assistance. We also thank Dr. David Phillips and Anne O’Connor for the provision of reagents for the follistatin binding assay.


   Footnotes
 
This work was supported by C. J. Martin Fellowship, National Health and Medical Research Council of Australia (registration key 169013) (to C.A.H.) and Program Grant, National Health and Medical Research Council of Australia (registration key 241000) (to D.M.R.).

Disclosure of potential conflicts of interest: C.A.H., K.L.C., and D.M.R. have nothing to declare.

First Published Online March 9, 2006

Abbreviations: ActRII, Activin type II receptor; BMP, bone morphogenetic protein; FS288, follistatin-288; FSD, follistatin domain; IFMA, immunofluorometric assay; PEI, polyethyleneimine.

Received February 1, 2006.

Accepted for publication March 2, 2006.


   References
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 References
 

  1. Phillips DJ 2000 Regulation of activin’s access to the cell: why is mother nature such a control freak? Bioessays 22:689–696[CrossRef][Medline]
  2. Vale W, Rivier C, Hsueh A, Campen C, Meunier H, Bicsak T, Vaughan J, Corrigan A, Bardin W, Sawchenko P, Petraglia F, Yu J, Plotsky P, Spiess J, Rivier J 1988 Chemical and biological characterization of the inhibin family of protein hormones. Recent Prog Horm Res 44:1–34[Medline]
  3. Eto Y, Tsuji T, Takezawa M, Takano S, Yokogawa Y, Shibai H 1987 Purification and characterization of erythroid differentiation factor (EDF) isolated from human leukemia cell line THP-1. Biochem Biophys Res Commun 142:1095–1103[CrossRef][Medline]
  4. Matzuk MM, Finegold MJ, Su JGJ, Hsueh AJW, Bradley A 1992 {alpha}-Inhibin is a tumor-suppressor gene with gonadal specificity in mice. Nature 366:313–319
  5. Jones KL, de Kretser DM, Patella S, Phillips DJ 2004 Activin A and follistatin in systemic inflammation. Mol Cell Endocrinol 225:119–125[CrossRef][Medline]
  6. Ogawa Y, Schmidt DK, Nathan RM, Armstrong RM, Miller KL, Sawamura SJ, Ziman JM, Erickson KL, Deleon ER, Rosen DM, Seyedin S, Glaser C, Chang R, Corrigan A, Vale W 1992 Bovine bone activin enhances bone morphogenetic protein-induced ectopic bone formation. J Biol Chem 267:14233–14237[Abstract/Free Full Text]
  7. McCarthy SA, Bicknell R 1993 Inhibition of vascular endothelial cell growth by activin-A. J Biol Chem 268:23066–23071[Abstract/Free Full Text]
  8. Schubert D, Kimura H, LaCorbiere M, Vaughan J, Karr D, Fischer WH 1990 Activin is a nerve cell survival molecule. Nature 344:868–870[CrossRef][Medline]
  9. Beer HD, Gassmann MG, Munz B, Steiling H, Engelhardt F, Bleuel K, Werner S 2000 Expression and function of keratinocyte growth factor and activin in skin morphogenesis and cutaneous wound repair. J Investig Dermatol Symp Proc 5:34–39[Medline]
  10. Massague J 2000 How cells read TGFß signals. Nat Rev Mol Cell Biol 1:169–178[CrossRef][Medline]
  11. Greenwald J, Groppe J, Gray P, Wiater E, Kwiatkowski W, Vale W, Choe S 2003 The BMP7/ActRII extracellular domain complex provides new insights into the cooperative nature of receptor assembly. Mol Cell 11:605–617[CrossRef][Medline]
  12. Mathews LS 1994 Activin receptors and cellular signaling by the receptor serine kinase family. Endocr Rev 15:310–325[CrossRef][Medline]
  13. Thompson TB, Woodruff TK, Jardetzky TS 2003 Structures of an ActRIIB:activin A complex reveal a novel binding mode for TGFß ligand:receptor interactions. EMBO J 22:1555–1566[CrossRef][Medline]
  14. Harrison CA, Gray PC, Fischer WH, Donaldson C, Choe S, Vale W 2004 An activin mutant with disrupted ALK4 binding blocks signaling via type II receptors. J Biol Chem 279:28036–28044[Abstract/Free Full Text]
  15. Kirsch T, Sebald W, Dreyer MK 2000 Crystal structure of the BMP-2-BRIA ectodomain complex. Nat Struct Biol 7:492–496[CrossRef][Medline]
  16. Phillips DJ, de Kretser DM 1998 Follistatin: a multifunctional regulatory protein. Front Neuroendocrinol 19:287–322[CrossRef][Medline]
  17. Sidis Y, Tortoriello DV, Holmes WE, Pan Y, Keutmann HT, Schneyer AL 2002 Follistatin-related protein and follistatin differentially neutralize endogenous vs. exogenous activin. Endocrinology 143:1613–1624[Abstract/Free Full Text]
  18. Hashimoto O, Nakamura T, Shoji H, Shimasaki S, Hayashi Y, Sugino H 1997 A novel role of follistatin, an activin-binding protein, in the inhibition of activin action in rat pituitary cells. Endocytotic degradation of activin and its acceleration by follistatin associated with cell-surface heparan sulfate. J Biol Chem 272:13835–13842[Abstract/Free Full Text]
  19. Wang Q, Keutmann HT, Schneyer AL, Sluss PM 2000 Analysis of human follistatin structure: identification of two discontinuous N-terminal sequences coding for activin A binding and structural consequences of activin binding to native proteins. Endocrinology 141:3183–3193[Abstract/Free Full Text]
  20. Keutmann HT, Schneyer AL, Sidis Y 2004 The role of follistatin domains in follistatin biological action. Mol Endocrinol 18:228–240[Abstract/Free Full Text]
  21. Thompson TB, Lerch TF, Cook RW, Woodruff TK, Jardetzky TS 2005 The structure of the follistatin:activin complex reveals antagonism of both type I and type II receptor binding. Dev Cell 9:535–543[CrossRef][Medline]
  22. Fischer WH, Park M, Donaldson C, Wiater E, Vaughan J, Bilezikjian LM, Vale, W 2003 Residues in the C-terminal region of activin A determine specificity for follistatin and type II receptor binding. J Endocrinol 176:61–68[Abstract]
  23. Bittencourt JC, Presse F, Arias C, Peto C, Vaughan J, Nahon JL, Vale W, Sawchenko PE 1992 The melanin-concentrating hormone system of the rat brain: an immuno- and hybridization histochemical characterization. J Comp Neurol 319:218–245[CrossRef][Medline]
  24. van Casteren JI, Schoonen WG, Kloosterboer HJ 2000 Development of time-resolved immunofluorometric assays for rat follicle-stimulating hormone and luteinizing hormone and application on sera of cycling rats. Biol Reprod 62:886–894[Abstract/Free Full Text]
  25. Tsuchida K 2004 Activins, myostatin and related TGFß family members as novel therapeutic targets for endocrine, metabolic and immune disorders. Curr Drug Targets Immune Endocr Metab Disord 4:157–166
  26. Schneyer A, Schoen A, Quigg A, Sidis Y 2003 Differential binding and neutralization of activins A and B by follistatin and follistatin like-3 (FSTL-3/FSRP/FLRG). Endocrinology 144:1671–1674[Abstract/Free Full Text]
  27. Amthor H, Nicholas G, McKinnell I, Kemp CF, Sharma M, Kambadur R, Patel, K 2004 Follistatin complexes Myostatin and antagonises Myostatin-mediated inhibition of myogenesis. Dev Biol 270:19–30[CrossRef][Medline]
  28. Gamer LW, Cox KA, Small C, Rosen V 2001 Gdf11 is a negative regulator of chondrogenesis and myogenesis in the developing chick limb. Dev Biol 229:407–420[CrossRef][Medline]
  29. Glister C, Kemp CF, Knight PG 2004 Bone morphogenetic protein (BMP) ligands and receptors in bovine ovarian follicle cells: actions of BMP-4, -6 and -7 on granulosa cells and differential modulation of Smad-1 phosphorylation by follistatin. Reproduction 127:239–254[Abstract/Free Full Text]
  30. Matzuk MM, Lu N, Vogel H, Sellheyer K, Roop DR, Bradley A 1995 Multiple defects and perinatal death in mice deficient in follistatin. Nature 374:360–363[CrossRef][Medline]
  31. Iemura S, Yamamoto TS, Takagi C, Uchiyama H, Natsume T, Shimasaki S, Sugino H, Ueno, N 1998 Direct binding of follistatin to a complex of bone-morphogenetic protein and its receptor inhibits ventral and epidermal cell fates in early Xenopus embryo. Proc Natl Acad Sci USA 95:9337–9342[Abstract/Free Full Text]
  32. Muenster U, Harrison CA, Donaldson C, Vale W, Fischer WH 2005 An activin-A/C chimera exhibits activin and myostatin antagonistic properties. J Biol Chem 280:36626–36632[Abstract/Free Full Text]
  33. Canalis E, Economides AN, Gazzerro E 2003 Bone morphogenetic proteins, their antagonists, and the skeleton. Endocr Rev 24:218–235[Abstract/Free Full Text]
  34. Govender S, Csimma C, Genant HK, Valentin-Opran A, Amit Y, Arbel R, Aro H, Atar D, Bishay M, Borner MG, Chiron P, Choong P, Cinats J, Courtenay B, Feibel R, Geulette B, Gravel C, Haas N, Raschke M, Hammacher E, van der Velde D, Hardy P, Holt M, Josten C, Ketterl RL, Lindeque B, Lob G, Mathevon H, McCoy G, Marsh D, Miller R, Munting E, Oevre S, Nordsletten L, Patel A, Pohl A, Rennie W, Reynders P, Rommens PM, Rondia J, Rossouw WC, Daneel PJ, Ruff S, Ruter A, Santavirta S, Schildhauer TA, Gekle C, Schnettler R, Segal D, Seiler H, Snowdowne RB, Stapert J, Taglang G, Verdonk R, Vogels L, Weckbach A, Wentzensen A, Wisniewski T 2002 Recombinant human bone morphogenetic protein-2 for treatment of open tibial fractures: a prospective, controlled, randomized study of four hundred and fifty patients. J Bone Joint Surg Am 84-A:2123–2134
  35. Hirshberg J, Coleman J, Marchant B, Rees RS 2001 TGFß3 in the treatment of pressure ulcers: a preliminary report. Adv Skin Wound Care 14:91–95[Medline]
  36. Munz B, Smola H, Engelhardt F, Bleuel K, Brauchle M, Lein I, Evans LW, Huylebroeck D, Balling R, Werner S 1999 Overexpression of activin A in the skin of transgenic mice reveals new activities of activin in epidermal morphogenesis, dermal fibrosis and wound repair. EMBO J 18:5205–5215[CrossRef][Medline]
  37. Wankell M, Munz B, Hubner G, Hans W, Wolf E, Goppelt A, Werner S 2001 Impaired wound healing in transgenic mice overexpressing the activin antagonist follistatin in the epidermis. EMBO J 20:5361–5372[CrossRef][Medline]
  38. Harrison CA, Gray PC, Vale WW, Robertson DM 2005 Antagonists of activin signaling: mechanisms and potential biological applications. Trends Endocrinol Metab 16:73–78[CrossRef][Medline]



This article has been cited by other articles:


Home page
 J. Biol. Chem.Home page
Y. Makanji, K. L. Walton, M. C. Wilce, K. L. Chan, D. M. Robertson, and C. A. Harrison
Suppression of Inhibin A Biological Activity by Alterations in the Binding Site for Betaglycan
J. Biol. Chem., June 13, 2008; 283(24): 16743 - 16751.
[Abstract] [Full Text] [PDF]

Home page
 Hum ReprodHome page
C. J. Stoikos, C. A. Harrison, L. A. Salamonsen, and E. Dimitriadis
A distinct cohort of the TGF{beta} superfamily members expressed in human endometrium regulate decidualization
Hum. Reprod., June 1, 2008; 23(6): 1447 - 1456.
[Abstract] [Full Text] [PDF]

Home page
 J. Biol. Chem.Home page
J.-L. Zhang, Y. Huang, L.-Y. Qiu, J. Nickel, and W. Sebald
von Willebrand Factor Type C Domain-containing Proteins Regulate Bone Morphogenetic Protein Signaling through Different Recognition Mechanisms
J. Biol. Chem., July 6, 2007; 282(27): 20002 - 20014.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
147/6/2744    most recent
Author Manuscript (PDF)
Right arrow Purchase Article
Right arrow View Shopping Cart
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow Request Copyright Permission
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Harrison, C. A.
Right arrow Articles by Robertson, D. M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Harrison, C. A.
Right arrow Articles by Robertson, D. M.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Endocrinology Endocrine Reviews J. Clin. End. & Metab.
Molecular Endocrinology Recent Prog. Horm. Res. All Endocrine Journals