0013-7227/91/1294-1987$03.00/0 Endocrinology Copyright © 1991 by The Endocrine Society
Vol. 129, No. 4 Printed in U.S.A.
Precursors of a-Inhibin Modulate Follicle-Stimulating Hormone Receptor Binding and Biological Activity* ALAN L. SCHNEYER, PATRICK M. SLUSS, RANDALL W. WHITCOMB, KATHRYN A. MARTIN, ROLF SPRENGEL, AND WILLIAM F. CROWLEY, JR. Reproductive Endocrine Unit, Department of Medicine, Massachusetts General Hospital, Boston, Massachusetts 02114; the Department of Urology, University of Rochester Medical Center (P.M.S.), Rochester, New York 14642; and the Laboratory of Molecular Neuroendocrinology, Center for Molecular Biology (R.S.), University of Heidelberg, Germany
ABSTRACT. Although several forms of monomeric a-inhibin have been isolated from follicular fluid, no biological function has yet been ascribed to these posttranslationally processed forms of the a-subunit precursor protein. Moreover, previous studies of a FSH receptor binding competitor (FRBC) isolated and characterized from porcine follicular fluid (pFF) suggested certain biochemical similarities between this protein and ainhibin precursors. We, therefore, investigated the hypothesis that a-inhibin and/or its precursors might represent autocrine and/or paracrine modulators of FSH action in the ovary, accounting for some of this FRBC activity and thereby exerting some degree of regulation over follicular maturation. Three separate sources of a-inhibin proteins were investigated for FRBC activity, including pFF, human FF (hFF), and a 293 cell line into which the full-length human a-inhibin cDNA had been stably transfected. Conditioned medium from these transfected cells contained several forms of a-inhibin precursors as well as mature a-inhibin, but no /?-subunit or intact inhibin. a-Inhibin proteins from all three sources, purified by a variety of methods, including immunoaffinity chromatography on an anti-a-inhibin column, inhibited FSH binding to both natural tissue FSH receptors as well as recombinant rat FSH receptors expressed in 293 cells. Furthermore, dimeric inhibin and activin, medium from untransfected 293 cells, and non-a-inhibin-con-
taining purification fractions were inactive in either assay. In addition, purified recombinant a-inhibin proteins were partial in vitro FSH antagonists in a bioassay in which cAMP generation from 293 cells expressing the recombinant FSH receptor is used as an index of FSH biological activity. These same fractions of hFF containing FRBC activity did not bind to LH receptors, thereby demonstrating receptor specificity for this activity. Using sodium dodecyl sulfate-polyacrylamide gel electrophoresis and Western blotting with a-inhibin or FRBC antisera, a 57,000 mol wt protein was identified in FRBC-active fractions from all three sources, suggesting that the active moiety was the fulllength a-inhibin precursor protein or a large mol wt fragment, but not mature a-inhibin. Lastly, all FRBC activity from all three sources was extracted by an a-inhibin immunoaffinity column and was recoverable upon elution. These results demonstrate that proteins derived from the ainhibin precursor modulate FSH binding to its receptor as well as its biological activity. Since a-inhibin precursors have been reported in FF at concentrations exceeding 2.5 Mg/ml> the potency of the FSH antagonism determined for a-inhibin is consistent with a potential physiological role for a-inhibin as an autocrine or paracrine FSH modulator. (Endocrinology 129: 1987-1999,1991)
I
NHIBIN, a heterodimeric glycoprotein comprised of a- and /3-subunits, was originally purified from follicular fluid (FF) (1). Recently, significant quantities of monomeric a-inhibin precursor proteins have also been isolated and/or characterized from bovine FF (bFF) (24) and serum (3), in addition to human FF (hFF) (5) and serum (6), for which no inhibin-like biological activity could be demonstrated. One major form of a-inhibin monomer, termed Pro-a-C (4) includes a 6,000 mol wt (Mr) peptide derived from the NH2-terminus of the full-
Received May 10,1991. Address requests for reprints to: Dr. Alan L. Schneyer, Reproductive Endocrinology Unit, Massachusetts General Hospital, BHX-5, Harvard University School of Medicine, Boston, Massachusetts 02114. * This work was supported in part by NIH Grants HD-25941 (to A.L.S.), HD-19302 (to P.M.S.), and HD-15080 and HD-15788 (to W.F.C.).
length pro-a-inhibin molecule linked through a disulfide bond to the processed mature a-inhibin protein, creating a molecule of 26,000-29,000 Mr (2-4). The intervening sequence, which is removed during posttranslational processing and termed pro-a-N, also has been purified from bFF (4). Although no biological activity has yet been ascribed to this pro-a-inhibin fragment, immunoneutralization against the pro-a-N sequence results in lower fecundity in sheep (7). In addition, expression of the a-inhibin gene appears to be regulated separately from that of the 0-subunit (8, 9), particularly in follicles at different maturational stages (10,11), and is expressed at considerably higher levels than the /?-subunit intragonadally (12). These observations raise the possibility that free a-inhibin, and particularly the forms of ainhibin comprised of various precursor portions of the
1987
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 17 November 2015. at 00:24 For personal use only. No other uses without permission. . All rights reserved.
1988
a-INHIBIN PRECURSORS MODULATE FSH ACTION
translated protein, such as pro-a-C, pro-a-N, or the entire precursor, may have paracrine or autocrine activities quite distinct from inhibin's systemic role (1) as a regulator of FSH secretion from the pituitary. FF has also been shown to contain numerous local modulators of FSH biological action, including insulinlike growth factor-I (IGF-I) (13-15), insulin (16, 17), IGF-I-binding protein (18), transforming growth factor|8 (TGFa) (19-21), inhibin, andactivin (22-25). However, all of these hormones and growth factors appear to modulate FSH-stimulated steroidogenesis at a point distal to the hormone-receptor binding event. Yet another distinct group of FSH modulators inhibits FSH binding to its receptor (26-28) and can have agonist or antagonist activity (29). To distinguish receptor binding inhibition from modulation of FSH biological action, these molecules can generally be classified as FSH receptor binding competitors (FRBCs). We have previously described a fraction of porcine FF (pFF) containing a FRBC identified by FSH receptor assay (RRA) and in vitro bioassay, but not containing FSH itself (29, 30). Available evidence suggested that the FRBC activity might be related to a 57,000 Mr monomeric protein that was present in all active fractions (3032). We have now further investigated the nature of this FRBC protein from both pFF and hFF, and have discovered a number of biochemical and immunological similarities to a-inhibin precursor proteins. While this FRBC has been purified substantially from the pooled FF starting material, final purification and sequencing to establish its identity have been extremely difficult to achieve due to losses during processing and its challenging chemistry. Therefore, we have also examined the relationship of the FRBC activity to a-inhibin by obtaining conditioned medium from cells into which the human ainhibin cDNA (but not the /?-subunit cDNA) was stably transfected. Therefore, these cells secrete a number of a-inhibin proteins, including both precursor and mature forms. The results of these studies suggest that a-inhibin precursor proteins bind to both natural and recombinant (33) FSH receptors, which supports the hypothesis that the large Mr FRBC identified in FF is related if not identical to a-inhibin precursor proteins. Materials and Methods pFF pFF was provided by Dr. Gabe Bialy (Contraceptive Development Branch, NICHHD). Initial processing of this material has been previously described in detail (29) and includes acidified acetone and diethyl ether extraction, followed by ultrafiltration and concentration to remove a low Mr FRBC (26, 29, 34). This fraction was then lyophilized, reconstituted in 0.05 M HEPES buffer (pH 7.6), and chromatographed on a HR 10/10 MONO-Q anion exchange (Pharmacia, Piscataway, NJ) col-
Endo • 1991 Voll29«No4
umn attached to a FPLC (Pharmacia) apparatus. The activities of eluted fractions were determined by FSH RRA, as described below, and absorbance was measured at 280 nm. Active proteins were pooled, dialyzed, and chromatographed on a 2.5 X 40-cm Blue-Sepharose CL-6B (Pharmacia) column developed in 0.05 M HEPES, pH 7.5. Retained proteins were eluted with 1.5 M potassium chloride in 0.05 M HEPES buffer. All FSH RRA activity was unretained by this procedure, and this fraction was then circulated on an immunoaffinity cartridge (see below) prepared using protein-G (Pharmacia)-purified immunoglobulin G (IgG) from a rabbit immunized against a synthetic peptide comprising the first 25 amino acids of human a-inhibin (see KM-1 description below). Retained proteins were eluted with 0.2 M glycine buffer (pH 2.0), adjusted to pH 6.0, and electroconcentrated using an Elutrap apparatus (Schleicher and Schuell, Keene, NH), followed by dialysis against RRA buffer in an inverted microcentrifuge tube capped with an 8000 Mr cut-off dialysis membrane (Spectrapor, Houston, TX). Nonretained protein fractions and dialysis buffers were assayed as controls. hFF Pooled hFF collected at oocyte retrieval from in vitro fertilization (IVF) patients was used as starting material for this study. Before purification, protease inhibitors, including phenylmethylsulfonylfluoride (100 jig/ml; Sigma, St. Louis, MO), leupeptin (0.5 Mg/ml; Boehringer Mannheim, Indianapolis, IN), and aprotinin (1.0 fig/m\; Boehringer) were added to prevent proteolytic degradation of active proteins during processing. Purification method 1 This protocol was modified from purification schemes successfully applied to purification of inhibin and a-inhibin proteins from pFF (35). Briefly, 25 ml hFF were chromatographed on a 2.6 x 40-cm Red-Sepharose CL-6B (Pharmacia) column equilibrated in 0.05 M HEPES buffer containing 0.005 M 3-[cholamidopropyl)dimethylammonio]l-propanesulfonate (pH 7.5). Proteins were eluted by a gradient to 100% buffer B, which consisted of equilibration buffer plus 1.5 M NaCl and 6 M urea. Subsamples of eluted fractions were dialyzed against RRA buffer and analyzed for a-inhibin immunoreactivity as well as FSH RRA activity. Active fractions were pooled, dialyzed against 0.05 M HEPES buffer, and applied to the MONO-Q column, as described above for pFF. Eluted active fractions were pooled, dialyzed against 0.01 M phosphate buffer (pH 7.0), and applied to the same immunoaffinity cartridge used for pFF FRBC. Eluted proteins were assayed by RRA and analyzed for protein composition by sodium dodecyl sulfatepolyacrylamide gel electrophoresis (SDS-PAGE) and Western blot, using anti-FRBC antiserum ASP-401 (see below). Purification method 2 To increase the yield of active FRBC and determine whether any FRBCs exist in hFF that are not retained by Red-Sepharose or our a-inhibin immunoaffinity cartridge, a second purification scheme was employed. hFF (100 ml) was dialyzed against 25% acetic acid, centrifuged at 2000 X g to remove precipitate, and chromatographed on a 2.5 X 38-cm S-Sepharose fast flow
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 17 November 2015. at 00:24 For personal use only. No other uses without permission. . All rights reserved.
a-INHIBIN PRECURSORS MODULATE FSH ACTION (Pharmacia) column. The column was equilibrated with 0.01 M ammonium acetate (AmAc) buffer, pH 3.5, and eluted with 1.5 M AmAc buffer, pH 7.2. The retained protein peaks from two 50-ml runs were pooled and chromatographed on a 2.5 X 40cm Blue-Sepharose 6B column (Pharmacia) that was equilibrated with 0.05 M HEPES buffer, pH 7.5. Nonretained proteins containing all of the FRBC and a-inhibin activity were then chromatographed using a 1.6 X 10-cm hydrophobic interaction column (phenyl-Sepharose fast flow, Pharmacia) by FPLC technology (Pharmacia). This column was equilibrated with 0.05 M HEPES buffer, pH 7.5, containing 2.5 M ammonium sulfate and eluted with 0.05 M HEPES, pH 7.5, without additions. a-Inhibin precursor proteins typically elute in two peaks, the most active peak is weakly retained (~1.8 M ammonium sulfate), and the later peak is strongly retained (0-0.4 M ammonium sulfate). Inhibin and activin typically remain on the column under these elution conditions and can be eluted along with other strongly retained proteins with 8 M urea or 80% 2-propanol. Fractions containing peak a-inhibin activity were pooled, dialyzed, concentrated using Centricon-10 cartridges (Amicon, Lexington, MA) into RRA or RIA buffer, and assayed for activity. Active fractions were analyzed for protein composition using SDS-PAGE, as described below. Recombinant a-inhibin A full-length clone of the human a-inhibin subunit was transfected into mammalian 293 cells using a cytomeglovirusbased expression vector (36). Conditioned medium containing immunoactive a-inhibin proteins was concentrated by tangential flow (10,000 Mr cut-off), desalted into 25 mM 2-[AT-morpholino]ethanesulfonic acid (MES), pH 6.0, containing 6 M urea, and chromatographed on an S-Sepharose fast flow (Pharmacia) column (5 X 20 cm) equilibrated in the same buffer. Immunoactive a-inhibin proteins were eluted with tetraethyl ammonium chloride (0-1.0 M) in MES buffer, with 90% of the immunoactivity recovered. Eluted proteins were electroconcentrated and dialyzed in an Elutrap system (Schleicher and Schuell). These proteins were then dialyzed against RRA buffer (see below) for 48 h immediately before each receptor assay, since both RRA and a-inhibin RIA activity were lost within 35 days after dialysis, despite the addition of protease inhibitors (Boehringer Mannheim) to each batch of a-inhibin proteins and each stage of purification. This first batch of conditioned medium was used for initial characterization of a-inhibin proteins secreted by transfected cells and to test for the presence of a FRBC in crude extracts. Protein concentrations in this and subsequent preparations were determined using the BioRad (Richmond, CA) protein assay. A second batch of conditioned medium was used for further purification of the FRBC activity. After concentration of the conditioned medium, ammonium sulfate was added to 2.5 M, and the medium was loaded onto a phenyl-Sepharose Fast Flow (Pharmacia) hydrophobic interaction column (1 X 10 cm) equilibrated with 2.5 M ammonium sulfate in 50 mM HEPES, pH 6.5, and eluted by a gradient to 50 mM HEPES, pH 6.5. Fractions were analyzed for a-inhibin by RIA using antiserum KM-1 (see below), and immunoactive peaks were pooled, concentrated, and dialyzed into 50 mM HEPES, pH 6.8, in Centricon-10 cells (Amicon) for further analysis by RRA, RIA, and
1989
SDS-PAGE. Conditioned medium from untransfected 293 cells (control medium) was processed similarly to the a-inhibin-conditioned medium, dialyzed into RRA buffer, and tested in the FSH RRA at doses equivalent to those of the medium containing recombinant a-inhibin. This conditioned medium did not contain immunoprecipitable proteins using any inhibin antiserum, including antiserum 29A. Activin and inhibin Recombinant inhibin and activin were produced in mammalian cells and purified to homogeneity, as previously described (37, 38). Since these materials are in limited supply, they were tested in the RRA at levels up to 1 ;ug. Inhibin antisera Antiserum KM-1 was raised in a rabbit to a synthetic peptide comprising amino acids l-25-(Gly-Tyr) of mature human ainhibin conjugated to keyhole limpet hemocyanin (35, 39). This antiserum recognizes human inhibin and its a-subunit, as determined by RIA and Western blot, and was used to identify and characterize a-inhibin fractions during purification. For RIA, the antiserum was used at a 1:18,000 final dilution along with radioiodinated synthetic peptide (a-inhibin 1-25-GlyTyr), where the minimum detectable dose of peptide was 10 pg/tube. Antiserum 29A was raised in a rabbit to recombinant human inhibin and recognizes free a-subunit as well as dimeric inhibin. This antiserum was used for immunoprecipitation of radiolabeled proteins from culture medium conditioned by cells into which the a-inhibin cDNA was cloned. FRBC antiserum This antiserum (ASP-401) was raised in a rabbit to partially purified FRBC from pFF, as described previously (31). It recognizes proteins in pFF-, hFF-, human seminal plasma-, rat granulosa cell-, and Sertoli cell-conditioned medium and Follistatin Reference Preparation (FSRP 119-6-4, NHPP, NIDDK) (31) as well as recombinant a-inhibin proteins of 29,000 Mr and larger (data not shown). ASP-401 does not recognize FSH from any species tested (including human, ovine, and bovine), recombinant human 32,000 Mr dimeric inhibin, activin-A, or Muellerian inhibiting substance (data not shown), thereby demonstrating its specificity for the a-subunit of inhibin somewhere in the precursor region N-terminus to mature a-inhibin. Immunoaffinity chromatography Antiserum KM-1 was used to produce two immunoaffinity devices: an analytical sized cartridge used for pFF and hFF, and a larger immunoaffinity column used later for recombinant a-inhibin proteins. The immunoaffinity cartridge was prepared by purifying the IgG fraction from 10 ml KM-1 antiserum on a protein-G-Sepharose (Pharmacia) column (1.0 x 4 cm). After dialysis and concentration in Centriprep-30 cartridges (Amicon), the IgG fraction was immobilized onto activated support (Amino Zetaaffinity-60, CUNO, Inc., Meriden, CT) according to manufacturer's directions. Approximately 2 mg IgG were immobilized by this procedure. A larger capacity column was
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 17 November 2015. at 00:24 For personal use only. No other uses without permission. . All rights reserved.
a-INHIBIN PRECURSORS MODULATE FSH ACTION
1990
prepared by immobilizing affinity-purified (to synthetic peptide immobilized on Amino-link gel, Amicon) IgG from 100 ml KM1 antiserum (10 mg) to 5 ml amino link-activated gel according to manufacturer's directions. The affinity of the IgG appears to be greater when immobilized by this procedure, as elution with 0.1 M glycine (pH 1.7) containing 6 M guanidine HC1 or 1 M acetic acid is necessary to elute radioiodinated synthetic ainhibin peptide or radioiodinated recombinant inhibin. At least some of the activity of retained proteins is lost by this procedure, as determined by RIA of recombinant inhibin after immunoaffinity chromatography. This large capacity column was used for purification of recombinant a-inhibin proteins from partially purified conditioned medium. Tissue FSH receptor assay The tissue FSH receptor assay was performed as previously described (29, 40), using crude calf testes membranes as a receptor source and PAGE-purified radioiodinated NIH hFSH 1-3 (29) (kindly provided by the National Hormone and Pituitary Program, NIDDK). The FSH standard was NIH hFSH I3, which has 3.8 mlU/ng activity against the Second International Reference Preparation (78/549) in this receptor assay (41). Assay performance characteristics have been previously reported in detail (40-42). Recombinant FSH receptor assay The specificity of FRBC inhibition of FSH binding to membrane receptors was analyzed using recombinant rat testicular FSH receptors cloned and expressed in 293 cells (33). To prepare membranes, cells were grown in 45% F-12, 45% Dulbecco's Modified Eagle's Medium, and 10% fetal calf serum (containing 10 mM HEPES, 2 mM glutamine, and 5 mg/100 ml gentamicin) to confluence in T-175 flasks (~100 X 106 cells). The growth medium was removed, and the cells were washed with PBS. Cells were removed by rinsing for 1 min with 10 ml trypsin-EDTA solution and resuspended in 10 ml warm growth medium. They were then pelleted at 400 X g, washed in homogenization buffer (50 mM HEPES, 0.25 M sucrose, 5 mM MgCl2, 6 mM CaCl2, 2 mM EDTA, and 0.01% thimerosal), homogenized, and pelleted at 30,000 X g for 30 min. These membranes were resuspended to 20 mg wet weight/ml in membrane buffer [50 mM HEPES (pH 7.5), 0.1 M sucrose, 5 mM MgCl2, 6 mM CaCl2, 2 mM EDTA, 0.01% thimerosal, and 0.1% ovalbumin] and stored at —20 C for up to 6 months. All cell culture reagents and media were obtained from Gibco (Grand Island, NY), and chemicals were from Sigma. Membranes were diluted to 1 mg wet weight/ml (0.1 ml added/tube for the receptor assay), precipitated by centrifugation at 30,000 X g for 15 min, washed, and recentrifuged. Other assay procedures were exactly as described for the tissue receptor assay. This receptor exhibits specificity and affinity characteristics similar to those of tissue FSH receptors (33) (our unpublished observations). FSH bioassay Cells containing the recombinant FSH receptor were used to determine the in vitro biological activity of FRBC-containing fractions according to previously described methods (33), ex-
Endo • 1991 Vol 129 • No 4
cept that 1 x 106 cells/well were incubated in 24-well plates for 24 h, hormone or treatment (three wells for each dose) was applied for 30 min at 37 C, and the lysed cells were extracted into 0.1 N HC1. cAMP was quantitated as a measure of FSH receptor stimulation using a RIA kit (New England Nuclear, Boston, MA). Samples were diluted 1:10 and assayed at 25 /il sample/tube. Results are expressed as the mean (of three wells) ± SEM for each treatment and/or dose. Significance of differences between treatments was determined using Student's t test. Representative data from one of two experiments are presented. LH receptor assay Membranes containing functional LH receptors were prepared from MA-10 cells and used for this RRA, as previously described in detail (43). Briefly, cells were removed from flasks, washed, homogenized, and pelleted by centrifugation, as described above for FSH receptors. Membranes were diluted to 0.5 mg/100 (A and added to tubes containing 100 /x\ radiolabeled hCG (-250,000 cpm) and sample or standard in 200 /xl (0.4 ml total volume). After shaking for 24 h at 20 C, bound hormonereceptor complex was precipitated by centrifugation at 30,000 X g for 15 min, washed with 1 ml assay buffer, and recentrifuged at 30,000 X g for 15 min. SDS-PAGE and Western blot Electrophoresis and Western blotting procedures were performed as previously described (6, 30), except that antiserum KM-1 or ASP-401 was used with alkaline phosphatase-conjugated goat antirabbit second antibody (Bio-Rad) to detect immunoactive a-inhibin or FRBC, respectively, and SDSPAGE was accomplished using either the Mighty Small gel apparatus (Hoeffer, San Francisco, CA) or a Phast system (Pharmacia). In both cases, 10% acrylamide SDS-gels were used for protein separation. Data analysis All RIA and RRA data were statistically analyzed using the NIHRIA program (44) for slope and ED50. Differences in slope greater than 2 SE were considered significant (P < 0.05).
Results Three different sources of FRBC activity with increasing specificity and relevance to human reproductive processes were examined. pFF was first examined due to its availability and proven source of numerous bioactive gonadal peptides. FRBCs were next identified in pooled hFF from IVF patients, demonstrating the applicability of this molecule to human studies. Both of these sources yielded evidence that the molecule responsible for identified FRBC activity was related to precursors of «inhibin, so a recombinant source of a-inhibin was also examined in detail. FRBC from pFF As previously reported (29, 30), a large Mr inhibitor of FSH binding to receptor (FRBC) was identified in pFF
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 17 November 2015. at 00:24 For personal use only. No other uses without permission. . All rights reserved.
a-INHIBIN PRECURSORS MODULATE FSH ACTION
by FSH RRA. The initial purification steps of acidacetone extraction and anion exchange chromatography were amended to increase the recovery of active protein by FPLC technology for the anion exchange and BlueSepharose affinity steps. Pooled active fractions were then purified by immunoaffinity chromatography on an analytical scale, using a cartridge to which purified IgG fractions from KM-1 (antihuman a-inhibin) antiserum were immobilized. The activity in the FSH RRA of the eluted proteins is shown in Fig. 1. The slope of the inhibition curve for FRBC from pFF was parallel that of the hFSH standards (0.98 ± 0.09 us. 0.91 ± 0.05), while the ED50 was 480 ng compared to the ED50 of FSH, which was 2.6 ng. Thus, the activity of this purified FRBC relative to FSH is approximately 180-fold lower on a mass basis, or 100-fold lower on a molar basis (assuming the active moiety is the 57,000 Mr protein identified below). Since active FRBC was obtained using a-inhibin immunoaffinity chromatography, the protein(s) responsible for FRBC activity appears to be related to a-inhibin proteins. However, due to the low recovery of activity after immunoaffinity chromatography, the lability of the activity after storage for greater than 72 h, and loss of activity in organic buffers necessary for reverse phase chromatography, it was not possible to purify the active moiety to homogeneity for sequencing. This problem recurred for FRBC purified from all three sources. Further characterization by SDS-PAGE and Western blot of proteins in this RRA-active fraction of FRBC is shown in Fig. 2. A total of four proteins can be identified with total protein staining by AuroDye (lane 1) at 66,000, 57,000, and 44,000 Mr, and a faint, diffuse band is found at approximately 29,000 Mr. Using either anti-FRBC (lane 2) or anti-a-inhibin (lane 3) immunostaining, only the 57,000 Mr protein was recognized, suggesting that an a-inhibin precursor protein might be responsible for
PROTON CONTROL
-
hFF
hFF FRBC
100
.001
L
1000
PROTEIN
FIG. 1. Activity of FRBC in FSH receptor assay. FSH receptor assay showing activity in pooled hFF and FRBC purified from hFF and pFF using immunoaffinity chromatography on an a-inhibin immunoaffinity column. All inhibition curves are parallel. Unretained proteins from immunoaffinity chromatography of hFF (protein control) were innactive in the FSH RRA. FSH receptors for this assay were derived from calf testes.
AURO KD DYE
1991
ANTI- ANTI-IN FRBC a(1-25)
97 66 43 31 21 14
1 FIG. 2. Characterization of purified porcine FRBC fraction after SDSPAGE under reducing conditions. Lane 1 shows total protein (AuroDue) staining of the purified FRBC fraction that was active in the FSH RRA (see Fig. 1). Four protein bands were detected in this fraction. Lane 2 shows this same fraction after Western blotting with anti-FRBC antiserum ASP-401, while lane 3 shows Western blotting with anti-a-inhibin antiserum KM-1. A 57,000 Mr protein is recognized by both antibodies, suggesting a relationship between the FRBC activity and a-inhibin precursor proteins. The positions of Mr markers are shown to the left of lane 1.
FRBC activity in this fraction of pFF that is also recognized by the anti-FRBC antibody. FRBC in hFF As shown in Fig. 1, pooled hFF recovered from individual follicles during IVF was active in the FSH RRA at relatively high doses. This activity was purified by an adaptation of the pFF protocol based on previously described inhibin purification protocols (35) and was terminated with immunoaffinity purification on the identical a-inhibin column as that used for pFF FRBC. Thus, as observed for pFF, all FRBC activity is removed by this analytical a-inhibin immunoaffinity column. As demonstrated in Table 1, eluted proteins from the immunoaffinity column were approximately 3000-fold more active in the FSH RRA than pooled hFF, with an inhibition slope parallel to that of hFSH standards and an ED50 of approximately 5 /xg (Fig. 1). Although pooled hFF has activity in both FSH and LH RRAs, partially purified FRBC fractions that were active in the FSH RRA were completely inactive in the LH RRA (data not shown), demonstrating the hormonal specificity for FRBC activity.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 17 November 2015. at 00:24 For personal use only. No other uses without permission. . All rights reserved.
1992
a-INHIBIN PRECURSORS MODULATE FSH ACTION
TABLE 1. Purification of FRBC from hFF Protein (mg) SA° Purification (fold)
Preparation hFF (25 ml) Matrix gel red Ion exchange (MONO-Q) Immunoaffinity
625 48 7.5 0.205
0.2 1.15 24 600
1 5.77 120 3000
Results at major steps in the purification are shown from two separate batches. Specific activity was determined using the calf testicular FSH receptor assay at each step. ° Specific activity is expressed as nanogram equivalents of mg protein, as quantitatedc in the FSH receptor assay. The mass of sample required to give 50% inhibition of radioligand binding (or maximal inhibition) was divided into nanograms of FSH 1-3 that resulted in the same degree of radioligand inhibition.
ANTI FRBC
66
43
31
21 14
Endo • 1991 Voll29«No4
minor bands observed at 44,000 and 29,000 Mr, similar to but less complex than the staining observed for pFF. For comparison, the immunoaffinity purified FRBC fraction from hFF was also characterized by SDS-PAGE and Western blot with this anti-FRBC antiserum, revealing proteins of 66,000, 57,000, and 42,000 Mr (Fig. 3, lane 3). As observed with FRBC purified from pFF, the activity in hFF was also labile after storage and was lost after reverse phase chromatography, making purification of active FRBC to homogeneity unrealistic using these methods. In an attempt to obtain a more potent and abundant FRBC fraction from hFF, a second pool of hFF was purified using a modified procedure. Matrix gel red and immunoaffinity chromatography were eliminated, and purification was initiated with dialysis of pooled (100 ml) hFF against 25% acetic acid, followed by cation exchange, Blue-Sepharose, and hydrophobic interaction chromatography (HIC). FRBC activity eluted in weakly retained fractions from the HIC column, which were pooled, concentrated, and dialyzed. As shown in Fig. 4, FRBC purified in this fashion had an inhibition slope parallel to that of the FSH standard, but the potency was increased 12-fold over that of the immunoaffinitypurified material and was approximately 100-fold lower than that of the FSH standard on a mass basis or 50fold lower than FSH on a molar basis (assuming the 57,000 immunoactive protein is responsible for the FRBC activity). Since the quantity of recovered protein from this process was extremely low, the protein concentration in eluted fractions was determined by compositional analysis, but there was insufficient material for microsequence determination. FRBC in recombinant a-inhibin supernatant Since the FRBC activity was retained by an a-inhibin immunoaffinity column, we next tested for the presence 100
KD FIG. 3. Characterization of pooled hFF and purified FRBC from hFF. Pooled pFF (lane 1) and hFF (lane 2) were electrophoresed under reducing conditions and Western blotted with anti-FRBC antiserum ASP-401. The staining pattern for hFF was similar to but less complex than that of pFF, but included a large 57,000 Mr band, a 44,000 Mr band, and a 29,000 Mr band. As shown in lane 3, immunoaffinity (on an a-inhibin antibody column)-purified FRBC from hFF (activity shown in Fig. 1) contains three bands, including the 57,000 and 44,000 Mr proteins, suggesting that the FRBC activity was due to large Mr ainhibin precursor proteins.
After SDS-PAGE under reducing conditions and Western blotting, proteins recognized by anti-FRBC antiserum (ASP-401) in hFF and pFF before purification were compared (Fig. 3, lanes 1 and 2). The major protein in hFF was at approximately 55,000-60,000 Mr, with
1.0
10.0
100.0
1000.0
ng FSH OR FRBC
FIG. 4. Recombinant FSH receptor assay of FRBC purified from hFF. FRBC was purified from hFF by method 2 (see Materials and Methods) and assayed in a FSH RRA using membranes from 293 cells expressing the recombinant rat FSH receptor. The inhibition curve for FRBC is parallel to that of hFSH and is approximately 50-fold lower in potency.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 17 November 2015. at 00:24 For personal use only. No other uses without permission. . All rights reserved.
a-INHIBIN PRECURSORS MODULATE FSH ACTION
of FRBC activity in culture medium conditioned by 293 cells that had been stably transfected with the human ainhibin cDNA. After concentration, dialysis, and cation exchange chromatography, these partially purified ainhibin proteins were analyzed for activity using the tissue FSH receptor assay. Binding of radiolabeled FSH to its receptor was inhibited by recombinant a-inhibin at doses between 10-90 ;ug, compared to doses of 0.8-25 ng of the highly purified FSH standard (Fig. 5), and at a significantly greater slope (2.66 ± 0.2 vs. 0.89 ± 0.04, respectively; P < 0.05). This activity is most likely due to secreted a-inhibin proteins from the transfected cells, since conditioned medium from untransfected 293 cells treated similarly and at similar doses was not active in this assay. In addition, dimeric inhibin and activin did not inhibit FSH binding to its receptor at doses up to 100-fold greater than that of FSH. Although the doses of dimeric inhibin and activin are less than that of recombinant a-inhibin proteins, they are also significantly more pure (37, 38). Additional controls, such as chromatography fractions that did not contain any immunoactive a-inhibin, and electroconcentration buffers had no activity in the FSH receptor assay. This receptorbinding activity was observed in multiple experiments using three separate batches of recombinant a-inhibin proteins. The partially purified recombinant a-inhibin-conditioned medium contains a number of proteins, including two major protein bands at an approximate Mr of 66,000 that were observed by total protein staining of transblotted gels (Fig. 6A). At least one of these could be removed by chromatography with Blue-Sepharose and may, therefore, be related to albumin, but was inactive by RIA and Western blot using anti-a-inhibin antiserum 100
o
80 60 •
— •
hfW (MH-FSH-l-3)
O — O ALPHA INHIHN SUBUtflT
40
• — • RECOMBUMNTACTMN-A • — • RECOMBMANT MH1BM DMER A — A UEDIUU FROM MOCK TIWNSFECnaN
20
.01
10
100
1.000
ng PURIFIED hFSH, ACTIVIN, INHIBIN OR fig ALPHA INHIBIN, CONTROL MEDIUM
FIG. 5. Activity of recombinant a-inhibin proteins in tissue FSH RRA. Standard inhibition curves are shown for highly purified hFSH standard and a-inhibin. Controls include medium from mock-transfected cell cultures in addition to pure recombinant inhibin and activin, for which no activity was observed in the RRA. Note that FSH is plotted in nanograms, while a-inhibin doses are presented graphically in micrograms.
1993
B
92 ^ ^
66
•
43 ^H 31 flB
(Ok
21 &P
Vol. 129, No. 4 Printed in U.S.A.
Precursors of a-Inhibin Modulate Follicle-Stimulating Hormone Receptor Binding and Biological Activity* ALAN L. SCHNEYER, PATRICK M. SLUSS, RANDALL W. WHITCOMB, KATHRYN A. MARTIN, ROLF SPRENGEL, AND WILLIAM F. CROWLEY, JR. Reproductive Endocrine Unit, Department of Medicine, Massachusetts General Hospital, Boston, Massachusetts 02114; the Department of Urology, University of Rochester Medical Center (P.M.S.), Rochester, New York 14642; and the Laboratory of Molecular Neuroendocrinology, Center for Molecular Biology (R.S.), University of Heidelberg, Germany
ABSTRACT. Although several forms of monomeric a-inhibin have been isolated from follicular fluid, no biological function has yet been ascribed to these posttranslationally processed forms of the a-subunit precursor protein. Moreover, previous studies of a FSH receptor binding competitor (FRBC) isolated and characterized from porcine follicular fluid (pFF) suggested certain biochemical similarities between this protein and ainhibin precursors. We, therefore, investigated the hypothesis that a-inhibin and/or its precursors might represent autocrine and/or paracrine modulators of FSH action in the ovary, accounting for some of this FRBC activity and thereby exerting some degree of regulation over follicular maturation. Three separate sources of a-inhibin proteins were investigated for FRBC activity, including pFF, human FF (hFF), and a 293 cell line into which the full-length human a-inhibin cDNA had been stably transfected. Conditioned medium from these transfected cells contained several forms of a-inhibin precursors as well as mature a-inhibin, but no /?-subunit or intact inhibin. a-Inhibin proteins from all three sources, purified by a variety of methods, including immunoaffinity chromatography on an anti-a-inhibin column, inhibited FSH binding to both natural tissue FSH receptors as well as recombinant rat FSH receptors expressed in 293 cells. Furthermore, dimeric inhibin and activin, medium from untransfected 293 cells, and non-a-inhibin-con-
taining purification fractions were inactive in either assay. In addition, purified recombinant a-inhibin proteins were partial in vitro FSH antagonists in a bioassay in which cAMP generation from 293 cells expressing the recombinant FSH receptor is used as an index of FSH biological activity. These same fractions of hFF containing FRBC activity did not bind to LH receptors, thereby demonstrating receptor specificity for this activity. Using sodium dodecyl sulfate-polyacrylamide gel electrophoresis and Western blotting with a-inhibin or FRBC antisera, a 57,000 mol wt protein was identified in FRBC-active fractions from all three sources, suggesting that the active moiety was the fulllength a-inhibin precursor protein or a large mol wt fragment, but not mature a-inhibin. Lastly, all FRBC activity from all three sources was extracted by an a-inhibin immunoaffinity column and was recoverable upon elution. These results demonstrate that proteins derived from the ainhibin precursor modulate FSH binding to its receptor as well as its biological activity. Since a-inhibin precursors have been reported in FF at concentrations exceeding 2.5 Mg/ml> the potency of the FSH antagonism determined for a-inhibin is consistent with a potential physiological role for a-inhibin as an autocrine or paracrine FSH modulator. (Endocrinology 129: 1987-1999,1991)
I
NHIBIN, a heterodimeric glycoprotein comprised of a- and /3-subunits, was originally purified from follicular fluid (FF) (1). Recently, significant quantities of monomeric a-inhibin precursor proteins have also been isolated and/or characterized from bovine FF (bFF) (24) and serum (3), in addition to human FF (hFF) (5) and serum (6), for which no inhibin-like biological activity could be demonstrated. One major form of a-inhibin monomer, termed Pro-a-C (4) includes a 6,000 mol wt (Mr) peptide derived from the NH2-terminus of the full-
Received May 10,1991. Address requests for reprints to: Dr. Alan L. Schneyer, Reproductive Endocrinology Unit, Massachusetts General Hospital, BHX-5, Harvard University School of Medicine, Boston, Massachusetts 02114. * This work was supported in part by NIH Grants HD-25941 (to A.L.S.), HD-19302 (to P.M.S.), and HD-15080 and HD-15788 (to W.F.C.).
length pro-a-inhibin molecule linked through a disulfide bond to the processed mature a-inhibin protein, creating a molecule of 26,000-29,000 Mr (2-4). The intervening sequence, which is removed during posttranslational processing and termed pro-a-N, also has been purified from bFF (4). Although no biological activity has yet been ascribed to this pro-a-inhibin fragment, immunoneutralization against the pro-a-N sequence results in lower fecundity in sheep (7). In addition, expression of the a-inhibin gene appears to be regulated separately from that of the 0-subunit (8, 9), particularly in follicles at different maturational stages (10,11), and is expressed at considerably higher levels than the /?-subunit intragonadally (12). These observations raise the possibility that free a-inhibin, and particularly the forms of ainhibin comprised of various precursor portions of the
1987
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 17 November 2015. at 00:24 For personal use only. No other uses without permission. . All rights reserved.
1988
a-INHIBIN PRECURSORS MODULATE FSH ACTION
translated protein, such as pro-a-C, pro-a-N, or the entire precursor, may have paracrine or autocrine activities quite distinct from inhibin's systemic role (1) as a regulator of FSH secretion from the pituitary. FF has also been shown to contain numerous local modulators of FSH biological action, including insulinlike growth factor-I (IGF-I) (13-15), insulin (16, 17), IGF-I-binding protein (18), transforming growth factor|8 (TGFa) (19-21), inhibin, andactivin (22-25). However, all of these hormones and growth factors appear to modulate FSH-stimulated steroidogenesis at a point distal to the hormone-receptor binding event. Yet another distinct group of FSH modulators inhibits FSH binding to its receptor (26-28) and can have agonist or antagonist activity (29). To distinguish receptor binding inhibition from modulation of FSH biological action, these molecules can generally be classified as FSH receptor binding competitors (FRBCs). We have previously described a fraction of porcine FF (pFF) containing a FRBC identified by FSH receptor assay (RRA) and in vitro bioassay, but not containing FSH itself (29, 30). Available evidence suggested that the FRBC activity might be related to a 57,000 Mr monomeric protein that was present in all active fractions (3032). We have now further investigated the nature of this FRBC protein from both pFF and hFF, and have discovered a number of biochemical and immunological similarities to a-inhibin precursor proteins. While this FRBC has been purified substantially from the pooled FF starting material, final purification and sequencing to establish its identity have been extremely difficult to achieve due to losses during processing and its challenging chemistry. Therefore, we have also examined the relationship of the FRBC activity to a-inhibin by obtaining conditioned medium from cells into which the human ainhibin cDNA (but not the /?-subunit cDNA) was stably transfected. Therefore, these cells secrete a number of a-inhibin proteins, including both precursor and mature forms. The results of these studies suggest that a-inhibin precursor proteins bind to both natural and recombinant (33) FSH receptors, which supports the hypothesis that the large Mr FRBC identified in FF is related if not identical to a-inhibin precursor proteins. Materials and Methods pFF pFF was provided by Dr. Gabe Bialy (Contraceptive Development Branch, NICHHD). Initial processing of this material has been previously described in detail (29) and includes acidified acetone and diethyl ether extraction, followed by ultrafiltration and concentration to remove a low Mr FRBC (26, 29, 34). This fraction was then lyophilized, reconstituted in 0.05 M HEPES buffer (pH 7.6), and chromatographed on a HR 10/10 MONO-Q anion exchange (Pharmacia, Piscataway, NJ) col-
Endo • 1991 Voll29«No4
umn attached to a FPLC (Pharmacia) apparatus. The activities of eluted fractions were determined by FSH RRA, as described below, and absorbance was measured at 280 nm. Active proteins were pooled, dialyzed, and chromatographed on a 2.5 X 40-cm Blue-Sepharose CL-6B (Pharmacia) column developed in 0.05 M HEPES, pH 7.5. Retained proteins were eluted with 1.5 M potassium chloride in 0.05 M HEPES buffer. All FSH RRA activity was unretained by this procedure, and this fraction was then circulated on an immunoaffinity cartridge (see below) prepared using protein-G (Pharmacia)-purified immunoglobulin G (IgG) from a rabbit immunized against a synthetic peptide comprising the first 25 amino acids of human a-inhibin (see KM-1 description below). Retained proteins were eluted with 0.2 M glycine buffer (pH 2.0), adjusted to pH 6.0, and electroconcentrated using an Elutrap apparatus (Schleicher and Schuell, Keene, NH), followed by dialysis against RRA buffer in an inverted microcentrifuge tube capped with an 8000 Mr cut-off dialysis membrane (Spectrapor, Houston, TX). Nonretained protein fractions and dialysis buffers were assayed as controls. hFF Pooled hFF collected at oocyte retrieval from in vitro fertilization (IVF) patients was used as starting material for this study. Before purification, protease inhibitors, including phenylmethylsulfonylfluoride (100 jig/ml; Sigma, St. Louis, MO), leupeptin (0.5 Mg/ml; Boehringer Mannheim, Indianapolis, IN), and aprotinin (1.0 fig/m\; Boehringer) were added to prevent proteolytic degradation of active proteins during processing. Purification method 1 This protocol was modified from purification schemes successfully applied to purification of inhibin and a-inhibin proteins from pFF (35). Briefly, 25 ml hFF were chromatographed on a 2.6 x 40-cm Red-Sepharose CL-6B (Pharmacia) column equilibrated in 0.05 M HEPES buffer containing 0.005 M 3-[cholamidopropyl)dimethylammonio]l-propanesulfonate (pH 7.5). Proteins were eluted by a gradient to 100% buffer B, which consisted of equilibration buffer plus 1.5 M NaCl and 6 M urea. Subsamples of eluted fractions were dialyzed against RRA buffer and analyzed for a-inhibin immunoreactivity as well as FSH RRA activity. Active fractions were pooled, dialyzed against 0.05 M HEPES buffer, and applied to the MONO-Q column, as described above for pFF. Eluted active fractions were pooled, dialyzed against 0.01 M phosphate buffer (pH 7.0), and applied to the same immunoaffinity cartridge used for pFF FRBC. Eluted proteins were assayed by RRA and analyzed for protein composition by sodium dodecyl sulfatepolyacrylamide gel electrophoresis (SDS-PAGE) and Western blot, using anti-FRBC antiserum ASP-401 (see below). Purification method 2 To increase the yield of active FRBC and determine whether any FRBCs exist in hFF that are not retained by Red-Sepharose or our a-inhibin immunoaffinity cartridge, a second purification scheme was employed. hFF (100 ml) was dialyzed against 25% acetic acid, centrifuged at 2000 X g to remove precipitate, and chromatographed on a 2.5 X 38-cm S-Sepharose fast flow
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 17 November 2015. at 00:24 For personal use only. No other uses without permission. . All rights reserved.
a-INHIBIN PRECURSORS MODULATE FSH ACTION (Pharmacia) column. The column was equilibrated with 0.01 M ammonium acetate (AmAc) buffer, pH 3.5, and eluted with 1.5 M AmAc buffer, pH 7.2. The retained protein peaks from two 50-ml runs were pooled and chromatographed on a 2.5 X 40cm Blue-Sepharose 6B column (Pharmacia) that was equilibrated with 0.05 M HEPES buffer, pH 7.5. Nonretained proteins containing all of the FRBC and a-inhibin activity were then chromatographed using a 1.6 X 10-cm hydrophobic interaction column (phenyl-Sepharose fast flow, Pharmacia) by FPLC technology (Pharmacia). This column was equilibrated with 0.05 M HEPES buffer, pH 7.5, containing 2.5 M ammonium sulfate and eluted with 0.05 M HEPES, pH 7.5, without additions. a-Inhibin precursor proteins typically elute in two peaks, the most active peak is weakly retained (~1.8 M ammonium sulfate), and the later peak is strongly retained (0-0.4 M ammonium sulfate). Inhibin and activin typically remain on the column under these elution conditions and can be eluted along with other strongly retained proteins with 8 M urea or 80% 2-propanol. Fractions containing peak a-inhibin activity were pooled, dialyzed, concentrated using Centricon-10 cartridges (Amicon, Lexington, MA) into RRA or RIA buffer, and assayed for activity. Active fractions were analyzed for protein composition using SDS-PAGE, as described below. Recombinant a-inhibin A full-length clone of the human a-inhibin subunit was transfected into mammalian 293 cells using a cytomeglovirusbased expression vector (36). Conditioned medium containing immunoactive a-inhibin proteins was concentrated by tangential flow (10,000 Mr cut-off), desalted into 25 mM 2-[AT-morpholino]ethanesulfonic acid (MES), pH 6.0, containing 6 M urea, and chromatographed on an S-Sepharose fast flow (Pharmacia) column (5 X 20 cm) equilibrated in the same buffer. Immunoactive a-inhibin proteins were eluted with tetraethyl ammonium chloride (0-1.0 M) in MES buffer, with 90% of the immunoactivity recovered. Eluted proteins were electroconcentrated and dialyzed in an Elutrap system (Schleicher and Schuell). These proteins were then dialyzed against RRA buffer (see below) for 48 h immediately before each receptor assay, since both RRA and a-inhibin RIA activity were lost within 35 days after dialysis, despite the addition of protease inhibitors (Boehringer Mannheim) to each batch of a-inhibin proteins and each stage of purification. This first batch of conditioned medium was used for initial characterization of a-inhibin proteins secreted by transfected cells and to test for the presence of a FRBC in crude extracts. Protein concentrations in this and subsequent preparations were determined using the BioRad (Richmond, CA) protein assay. A second batch of conditioned medium was used for further purification of the FRBC activity. After concentration of the conditioned medium, ammonium sulfate was added to 2.5 M, and the medium was loaded onto a phenyl-Sepharose Fast Flow (Pharmacia) hydrophobic interaction column (1 X 10 cm) equilibrated with 2.5 M ammonium sulfate in 50 mM HEPES, pH 6.5, and eluted by a gradient to 50 mM HEPES, pH 6.5. Fractions were analyzed for a-inhibin by RIA using antiserum KM-1 (see below), and immunoactive peaks were pooled, concentrated, and dialyzed into 50 mM HEPES, pH 6.8, in Centricon-10 cells (Amicon) for further analysis by RRA, RIA, and
1989
SDS-PAGE. Conditioned medium from untransfected 293 cells (control medium) was processed similarly to the a-inhibin-conditioned medium, dialyzed into RRA buffer, and tested in the FSH RRA at doses equivalent to those of the medium containing recombinant a-inhibin. This conditioned medium did not contain immunoprecipitable proteins using any inhibin antiserum, including antiserum 29A. Activin and inhibin Recombinant inhibin and activin were produced in mammalian cells and purified to homogeneity, as previously described (37, 38). Since these materials are in limited supply, they were tested in the RRA at levels up to 1 ;ug. Inhibin antisera Antiserum KM-1 was raised in a rabbit to a synthetic peptide comprising amino acids l-25-(Gly-Tyr) of mature human ainhibin conjugated to keyhole limpet hemocyanin (35, 39). This antiserum recognizes human inhibin and its a-subunit, as determined by RIA and Western blot, and was used to identify and characterize a-inhibin fractions during purification. For RIA, the antiserum was used at a 1:18,000 final dilution along with radioiodinated synthetic peptide (a-inhibin 1-25-GlyTyr), where the minimum detectable dose of peptide was 10 pg/tube. Antiserum 29A was raised in a rabbit to recombinant human inhibin and recognizes free a-subunit as well as dimeric inhibin. This antiserum was used for immunoprecipitation of radiolabeled proteins from culture medium conditioned by cells into which the a-inhibin cDNA was cloned. FRBC antiserum This antiserum (ASP-401) was raised in a rabbit to partially purified FRBC from pFF, as described previously (31). It recognizes proteins in pFF-, hFF-, human seminal plasma-, rat granulosa cell-, and Sertoli cell-conditioned medium and Follistatin Reference Preparation (FSRP 119-6-4, NHPP, NIDDK) (31) as well as recombinant a-inhibin proteins of 29,000 Mr and larger (data not shown). ASP-401 does not recognize FSH from any species tested (including human, ovine, and bovine), recombinant human 32,000 Mr dimeric inhibin, activin-A, or Muellerian inhibiting substance (data not shown), thereby demonstrating its specificity for the a-subunit of inhibin somewhere in the precursor region N-terminus to mature a-inhibin. Immunoaffinity chromatography Antiserum KM-1 was used to produce two immunoaffinity devices: an analytical sized cartridge used for pFF and hFF, and a larger immunoaffinity column used later for recombinant a-inhibin proteins. The immunoaffinity cartridge was prepared by purifying the IgG fraction from 10 ml KM-1 antiserum on a protein-G-Sepharose (Pharmacia) column (1.0 x 4 cm). After dialysis and concentration in Centriprep-30 cartridges (Amicon), the IgG fraction was immobilized onto activated support (Amino Zetaaffinity-60, CUNO, Inc., Meriden, CT) according to manufacturer's directions. Approximately 2 mg IgG were immobilized by this procedure. A larger capacity column was
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 17 November 2015. at 00:24 For personal use only. No other uses without permission. . All rights reserved.
a-INHIBIN PRECURSORS MODULATE FSH ACTION
1990
prepared by immobilizing affinity-purified (to synthetic peptide immobilized on Amino-link gel, Amicon) IgG from 100 ml KM1 antiserum (10 mg) to 5 ml amino link-activated gel according to manufacturer's directions. The affinity of the IgG appears to be greater when immobilized by this procedure, as elution with 0.1 M glycine (pH 1.7) containing 6 M guanidine HC1 or 1 M acetic acid is necessary to elute radioiodinated synthetic ainhibin peptide or radioiodinated recombinant inhibin. At least some of the activity of retained proteins is lost by this procedure, as determined by RIA of recombinant inhibin after immunoaffinity chromatography. This large capacity column was used for purification of recombinant a-inhibin proteins from partially purified conditioned medium. Tissue FSH receptor assay The tissue FSH receptor assay was performed as previously described (29, 40), using crude calf testes membranes as a receptor source and PAGE-purified radioiodinated NIH hFSH 1-3 (29) (kindly provided by the National Hormone and Pituitary Program, NIDDK). The FSH standard was NIH hFSH I3, which has 3.8 mlU/ng activity against the Second International Reference Preparation (78/549) in this receptor assay (41). Assay performance characteristics have been previously reported in detail (40-42). Recombinant FSH receptor assay The specificity of FRBC inhibition of FSH binding to membrane receptors was analyzed using recombinant rat testicular FSH receptors cloned and expressed in 293 cells (33). To prepare membranes, cells were grown in 45% F-12, 45% Dulbecco's Modified Eagle's Medium, and 10% fetal calf serum (containing 10 mM HEPES, 2 mM glutamine, and 5 mg/100 ml gentamicin) to confluence in T-175 flasks (~100 X 106 cells). The growth medium was removed, and the cells were washed with PBS. Cells were removed by rinsing for 1 min with 10 ml trypsin-EDTA solution and resuspended in 10 ml warm growth medium. They were then pelleted at 400 X g, washed in homogenization buffer (50 mM HEPES, 0.25 M sucrose, 5 mM MgCl2, 6 mM CaCl2, 2 mM EDTA, and 0.01% thimerosal), homogenized, and pelleted at 30,000 X g for 30 min. These membranes were resuspended to 20 mg wet weight/ml in membrane buffer [50 mM HEPES (pH 7.5), 0.1 M sucrose, 5 mM MgCl2, 6 mM CaCl2, 2 mM EDTA, 0.01% thimerosal, and 0.1% ovalbumin] and stored at —20 C for up to 6 months. All cell culture reagents and media were obtained from Gibco (Grand Island, NY), and chemicals were from Sigma. Membranes were diluted to 1 mg wet weight/ml (0.1 ml added/tube for the receptor assay), precipitated by centrifugation at 30,000 X g for 15 min, washed, and recentrifuged. Other assay procedures were exactly as described for the tissue receptor assay. This receptor exhibits specificity and affinity characteristics similar to those of tissue FSH receptors (33) (our unpublished observations). FSH bioassay Cells containing the recombinant FSH receptor were used to determine the in vitro biological activity of FRBC-containing fractions according to previously described methods (33), ex-
Endo • 1991 Vol 129 • No 4
cept that 1 x 106 cells/well were incubated in 24-well plates for 24 h, hormone or treatment (three wells for each dose) was applied for 30 min at 37 C, and the lysed cells were extracted into 0.1 N HC1. cAMP was quantitated as a measure of FSH receptor stimulation using a RIA kit (New England Nuclear, Boston, MA). Samples were diluted 1:10 and assayed at 25 /il sample/tube. Results are expressed as the mean (of three wells) ± SEM for each treatment and/or dose. Significance of differences between treatments was determined using Student's t test. Representative data from one of two experiments are presented. LH receptor assay Membranes containing functional LH receptors were prepared from MA-10 cells and used for this RRA, as previously described in detail (43). Briefly, cells were removed from flasks, washed, homogenized, and pelleted by centrifugation, as described above for FSH receptors. Membranes were diluted to 0.5 mg/100 (A and added to tubes containing 100 /x\ radiolabeled hCG (-250,000 cpm) and sample or standard in 200 /xl (0.4 ml total volume). After shaking for 24 h at 20 C, bound hormonereceptor complex was precipitated by centrifugation at 30,000 X g for 15 min, washed with 1 ml assay buffer, and recentrifuged at 30,000 X g for 15 min. SDS-PAGE and Western blot Electrophoresis and Western blotting procedures were performed as previously described (6, 30), except that antiserum KM-1 or ASP-401 was used with alkaline phosphatase-conjugated goat antirabbit second antibody (Bio-Rad) to detect immunoactive a-inhibin or FRBC, respectively, and SDSPAGE was accomplished using either the Mighty Small gel apparatus (Hoeffer, San Francisco, CA) or a Phast system (Pharmacia). In both cases, 10% acrylamide SDS-gels were used for protein separation. Data analysis All RIA and RRA data were statistically analyzed using the NIHRIA program (44) for slope and ED50. Differences in slope greater than 2 SE were considered significant (P < 0.05).
Results Three different sources of FRBC activity with increasing specificity and relevance to human reproductive processes were examined. pFF was first examined due to its availability and proven source of numerous bioactive gonadal peptides. FRBCs were next identified in pooled hFF from IVF patients, demonstrating the applicability of this molecule to human studies. Both of these sources yielded evidence that the molecule responsible for identified FRBC activity was related to precursors of «inhibin, so a recombinant source of a-inhibin was also examined in detail. FRBC from pFF As previously reported (29, 30), a large Mr inhibitor of FSH binding to receptor (FRBC) was identified in pFF
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 17 November 2015. at 00:24 For personal use only. No other uses without permission. . All rights reserved.
a-INHIBIN PRECURSORS MODULATE FSH ACTION
by FSH RRA. The initial purification steps of acidacetone extraction and anion exchange chromatography were amended to increase the recovery of active protein by FPLC technology for the anion exchange and BlueSepharose affinity steps. Pooled active fractions were then purified by immunoaffinity chromatography on an analytical scale, using a cartridge to which purified IgG fractions from KM-1 (antihuman a-inhibin) antiserum were immobilized. The activity in the FSH RRA of the eluted proteins is shown in Fig. 1. The slope of the inhibition curve for FRBC from pFF was parallel that of the hFSH standards (0.98 ± 0.09 us. 0.91 ± 0.05), while the ED50 was 480 ng compared to the ED50 of FSH, which was 2.6 ng. Thus, the activity of this purified FRBC relative to FSH is approximately 180-fold lower on a mass basis, or 100-fold lower on a molar basis (assuming the active moiety is the 57,000 Mr protein identified below). Since active FRBC was obtained using a-inhibin immunoaffinity chromatography, the protein(s) responsible for FRBC activity appears to be related to a-inhibin proteins. However, due to the low recovery of activity after immunoaffinity chromatography, the lability of the activity after storage for greater than 72 h, and loss of activity in organic buffers necessary for reverse phase chromatography, it was not possible to purify the active moiety to homogeneity for sequencing. This problem recurred for FRBC purified from all three sources. Further characterization by SDS-PAGE and Western blot of proteins in this RRA-active fraction of FRBC is shown in Fig. 2. A total of four proteins can be identified with total protein staining by AuroDye (lane 1) at 66,000, 57,000, and 44,000 Mr, and a faint, diffuse band is found at approximately 29,000 Mr. Using either anti-FRBC (lane 2) or anti-a-inhibin (lane 3) immunostaining, only the 57,000 Mr protein was recognized, suggesting that an a-inhibin precursor protein might be responsible for
PROTON CONTROL
-
hFF
hFF FRBC
100
.001
L
1000
PROTEIN
FIG. 1. Activity of FRBC in FSH receptor assay. FSH receptor assay showing activity in pooled hFF and FRBC purified from hFF and pFF using immunoaffinity chromatography on an a-inhibin immunoaffinity column. All inhibition curves are parallel. Unretained proteins from immunoaffinity chromatography of hFF (protein control) were innactive in the FSH RRA. FSH receptors for this assay were derived from calf testes.
AURO KD DYE
1991
ANTI- ANTI-IN FRBC a(1-25)
97 66 43 31 21 14
1 FIG. 2. Characterization of purified porcine FRBC fraction after SDSPAGE under reducing conditions. Lane 1 shows total protein (AuroDue) staining of the purified FRBC fraction that was active in the FSH RRA (see Fig. 1). Four protein bands were detected in this fraction. Lane 2 shows this same fraction after Western blotting with anti-FRBC antiserum ASP-401, while lane 3 shows Western blotting with anti-a-inhibin antiserum KM-1. A 57,000 Mr protein is recognized by both antibodies, suggesting a relationship between the FRBC activity and a-inhibin precursor proteins. The positions of Mr markers are shown to the left of lane 1.
FRBC activity in this fraction of pFF that is also recognized by the anti-FRBC antibody. FRBC in hFF As shown in Fig. 1, pooled hFF recovered from individual follicles during IVF was active in the FSH RRA at relatively high doses. This activity was purified by an adaptation of the pFF protocol based on previously described inhibin purification protocols (35) and was terminated with immunoaffinity purification on the identical a-inhibin column as that used for pFF FRBC. Thus, as observed for pFF, all FRBC activity is removed by this analytical a-inhibin immunoaffinity column. As demonstrated in Table 1, eluted proteins from the immunoaffinity column were approximately 3000-fold more active in the FSH RRA than pooled hFF, with an inhibition slope parallel to that of hFSH standards and an ED50 of approximately 5 /xg (Fig. 1). Although pooled hFF has activity in both FSH and LH RRAs, partially purified FRBC fractions that were active in the FSH RRA were completely inactive in the LH RRA (data not shown), demonstrating the hormonal specificity for FRBC activity.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 17 November 2015. at 00:24 For personal use only. No other uses without permission. . All rights reserved.
1992
a-INHIBIN PRECURSORS MODULATE FSH ACTION
TABLE 1. Purification of FRBC from hFF Protein (mg) SA° Purification (fold)
Preparation hFF (25 ml) Matrix gel red Ion exchange (MONO-Q) Immunoaffinity
625 48 7.5 0.205
0.2 1.15 24 600
1 5.77 120 3000
Results at major steps in the purification are shown from two separate batches. Specific activity was determined using the calf testicular FSH receptor assay at each step. ° Specific activity is expressed as nanogram equivalents of mg protein, as quantitatedc in the FSH receptor assay. The mass of sample required to give 50% inhibition of radioligand binding (or maximal inhibition) was divided into nanograms of FSH 1-3 that resulted in the same degree of radioligand inhibition.
ANTI FRBC
66
43
31
21 14
Endo • 1991 Voll29«No4
minor bands observed at 44,000 and 29,000 Mr, similar to but less complex than the staining observed for pFF. For comparison, the immunoaffinity purified FRBC fraction from hFF was also characterized by SDS-PAGE and Western blot with this anti-FRBC antiserum, revealing proteins of 66,000, 57,000, and 42,000 Mr (Fig. 3, lane 3). As observed with FRBC purified from pFF, the activity in hFF was also labile after storage and was lost after reverse phase chromatography, making purification of active FRBC to homogeneity unrealistic using these methods. In an attempt to obtain a more potent and abundant FRBC fraction from hFF, a second pool of hFF was purified using a modified procedure. Matrix gel red and immunoaffinity chromatography were eliminated, and purification was initiated with dialysis of pooled (100 ml) hFF against 25% acetic acid, followed by cation exchange, Blue-Sepharose, and hydrophobic interaction chromatography (HIC). FRBC activity eluted in weakly retained fractions from the HIC column, which were pooled, concentrated, and dialyzed. As shown in Fig. 4, FRBC purified in this fashion had an inhibition slope parallel to that of the FSH standard, but the potency was increased 12-fold over that of the immunoaffinitypurified material and was approximately 100-fold lower than that of the FSH standard on a mass basis or 50fold lower than FSH on a molar basis (assuming the 57,000 immunoactive protein is responsible for the FRBC activity). Since the quantity of recovered protein from this process was extremely low, the protein concentration in eluted fractions was determined by compositional analysis, but there was insufficient material for microsequence determination. FRBC in recombinant a-inhibin supernatant Since the FRBC activity was retained by an a-inhibin immunoaffinity column, we next tested for the presence 100
KD FIG. 3. Characterization of pooled hFF and purified FRBC from hFF. Pooled pFF (lane 1) and hFF (lane 2) were electrophoresed under reducing conditions and Western blotted with anti-FRBC antiserum ASP-401. The staining pattern for hFF was similar to but less complex than that of pFF, but included a large 57,000 Mr band, a 44,000 Mr band, and a 29,000 Mr band. As shown in lane 3, immunoaffinity (on an a-inhibin antibody column)-purified FRBC from hFF (activity shown in Fig. 1) contains three bands, including the 57,000 and 44,000 Mr proteins, suggesting that the FRBC activity was due to large Mr ainhibin precursor proteins.
After SDS-PAGE under reducing conditions and Western blotting, proteins recognized by anti-FRBC antiserum (ASP-401) in hFF and pFF before purification were compared (Fig. 3, lanes 1 and 2). The major protein in hFF was at approximately 55,000-60,000 Mr, with
1.0
10.0
100.0
1000.0
ng FSH OR FRBC
FIG. 4. Recombinant FSH receptor assay of FRBC purified from hFF. FRBC was purified from hFF by method 2 (see Materials and Methods) and assayed in a FSH RRA using membranes from 293 cells expressing the recombinant rat FSH receptor. The inhibition curve for FRBC is parallel to that of hFSH and is approximately 50-fold lower in potency.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 17 November 2015. at 00:24 For personal use only. No other uses without permission. . All rights reserved.
a-INHIBIN PRECURSORS MODULATE FSH ACTION
of FRBC activity in culture medium conditioned by 293 cells that had been stably transfected with the human ainhibin cDNA. After concentration, dialysis, and cation exchange chromatography, these partially purified ainhibin proteins were analyzed for activity using the tissue FSH receptor assay. Binding of radiolabeled FSH to its receptor was inhibited by recombinant a-inhibin at doses between 10-90 ;ug, compared to doses of 0.8-25 ng of the highly purified FSH standard (Fig. 5), and at a significantly greater slope (2.66 ± 0.2 vs. 0.89 ± 0.04, respectively; P < 0.05). This activity is most likely due to secreted a-inhibin proteins from the transfected cells, since conditioned medium from untransfected 293 cells treated similarly and at similar doses was not active in this assay. In addition, dimeric inhibin and activin did not inhibit FSH binding to its receptor at doses up to 100-fold greater than that of FSH. Although the doses of dimeric inhibin and activin are less than that of recombinant a-inhibin proteins, they are also significantly more pure (37, 38). Additional controls, such as chromatography fractions that did not contain any immunoactive a-inhibin, and electroconcentration buffers had no activity in the FSH receptor assay. This receptorbinding activity was observed in multiple experiments using three separate batches of recombinant a-inhibin proteins. The partially purified recombinant a-inhibin-conditioned medium contains a number of proteins, including two major protein bands at an approximate Mr of 66,000 that were observed by total protein staining of transblotted gels (Fig. 6A). At least one of these could be removed by chromatography with Blue-Sepharose and may, therefore, be related to albumin, but was inactive by RIA and Western blot using anti-a-inhibin antiserum 100
o
80 60 •
— •
hfW (MH-FSH-l-3)
O — O ALPHA INHIHN SUBUtflT
40
• — • RECOMBUMNTACTMN-A • — • RECOMBMANT MH1BM DMER A — A UEDIUU FROM MOCK TIWNSFECnaN
20
.01
10
100
1.000
ng PURIFIED hFSH, ACTIVIN, INHIBIN OR fig ALPHA INHIBIN, CONTROL MEDIUM
FIG. 5. Activity of recombinant a-inhibin proteins in tissue FSH RRA. Standard inhibition curves are shown for highly purified hFSH standard and a-inhibin. Controls include medium from mock-transfected cell cultures in addition to pure recombinant inhibin and activin, for which no activity was observed in the RRA. Note that FSH is plotted in nanograms, while a-inhibin doses are presented graphically in micrograms.
1993
B
92 ^ ^
66
•
43 ^H 31 flB
(Ok
21 &P
