0013-7227/91/1283-1485$03.00/0 Endocrinology Copyright © 1991 by The Endocrine Society

Vol. 128, No. 3 Printed in U.S.A.

Epitope Mapping of Human Follicle Stimulating Hormone-a Using Monoclonal Antibody 3A Identifies a Potential Receptor Binding Sequence* RUSSELL S. WEINER, JAMES A. DIAS, AND THOMAS T. ANDERSEN Department of Biochemistry (R.S. W., J.A.D., T.T.A.), Albany Medical College, Albany, New York 12208; Wadsworth Center for Laboratories and Research (R.S. W., J.A.D.), New York State Department of Health, Empire State Plaza, Albany, New York 12201-0509; and School of Public Health (J.A.D.), University at Albany, State University of New York, Albany, New York 12201

hFSH/3. In contrast, 5F bound only [125I]hFSH. hFSH effectively competed with [125I]hFSH for 5F and 3A. In contrast, hFSHa competed with [125I]hFSH for 5F but not for 3A even though 3A could bind hFSHa in the ELISA and the RIA. These results suggest that 3A and 5F recognize different epitopes. The epitope recognized by 3A is unique in that its conformation appears to be dependent on association with hFSH/3. Since 3A was a more potent inhibitor of receptor binding than 5F, its epitope specificity was characterized further by epitope mapping. This was accomplished utilizing a peptide ELISA and by affinity chromatography. The results from epitope mapping demonstrated that 3A recognizes sequences 61-78 and 73-92 with binding to 73-92 being 4-fold greater than to 61-78. Thus, the epitope comprised of sequence 73-92 (and to a lesser extent 61-78) appears to be important for receptor binding. (Endocrinology 128:1485-1495,1991)

ABSTRACT. Five monoclonal antibodies (mabs) were generated (3A, 4B, 5F, 2E, IE) by immunizing BALB/c mice with human (h) FSH. The mabs were used to relate antigenic structures (epitopes) to function (receptor binding). All five mabs could immunoneutralize (inhibit binding to receptor) hFSH and could be placed into two groups based on potency (degree of neutralization). Group I mabs (5F, 2E, IE) were less potent than group II mabs (3A, 4B) even though group I mabs had a 2-fold higher average affinity constant than group II. Those data suggested that group II mabs recognize an epitope near or in the receptor binding site of hFSH. Immunoradiometric epitope cross-matching demonstrated that group I and group II mabs recognize different epitopes. Further characterization of 5F and 3A (representative of group I and group II, respectively) utilized an enzyme-linked immunosorbent assay (ELISA) and a RIA. In the ELISA, both mabs bound hFSH and hFSHa but not hFSH/3. In the RIA, 3A bound [125I]hFSH and [125I]hFSHa but not [125I]

T

HE HUMAN (h) FSH crystal structure has not been determined, and therefore structure-function analysis continues to be intensely investigated in order to gain information concerning its tertiary and quaternary structure. This holds true for hLH, hTSH, and hCG which together with hFSH constitute the homologous family of glycoprotein hormones. The hormones in this family share a common heterodimeric structure which consists of a common a-subunit in noncovalent association with a unique hormone-specific /3-subunit (1). In addition to the common structural motif the hormones share a great deal of interspecies sequence homology (2). Structure-function analysis of the a-subunit has been investigated thoroughly utilizing a vast number of differ-

September 24,1990. Address requests for reprints to: James A. Dias, Ph.D., Wadsworth Center for Laboratories and Research, New York State Department of Health, Empire State Plaza, P.O. Box 509, Albany, New York 122010509. * Supported By NIH Grants HD-18407 and GM-4344301 and National Science Foundation Grant DIR 8914757. James A. Dias is the recipient of NIH Grant 5K04-HD 00682 RCDA.

ent techniques. The results of such studies are discussed in several detailed reviews (1, 3, 4). Classically, this was effected through chemical modification and cross-linking of the hormone in an attempt to determine which residues were on the surface of the subunit and which of these surface-oriented residues are in the subunit interface (5-7). More recently, monoclonal antibodies (mabs), obtained by immunization with either heterodimer or free a-subunit, have been employed. These studies investigated conformational changes within the subunits upon formation of heterodimer (8), the orientation of the hormone while in contact with its receptor (9), and the number of epitopes present on the surface of the subunit (10, 11). Mabs, when used to screen hormone fragments, synthetic peptides, or chimeric hormones, have proven useful for epitope mapping experiments (12-14). These studies provide structural information as to which sequences comprise epitopes, receptor binding surfaces, or subunit association surfaces. This approach has been taken one step further, namely to relate the determined structure

1485

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 18 November 2015. at 18:47 For personal use only. No other uses without permission. . All rights reserved.

EPITOPE MAPPING OF HUMAN FSH«

1486

to a function, specifically receptor binding or subunit contact sequence determination (15). Limited proteolysis of the hormones has revealed specific sequences that are surface oriented in the heterodimer or masked by subunit association (16, 17). Recently, we reported the results of a novel approach to this problem which utilized highly specific antipeptide antisera obtained by immunizing rabbits with seven different synthetic hFSHa peptides coupled to hemocyanin (18). Using this approach we generated antipeptide antisera of known specificity. Each antipeptide antisera was then tested to determine which would bind to hFSH, hFSHa, or reduced and alkylated hFSHa. Antipeptide antisera to a-peptide 73-92 identified this region as a surface of the «-subunit. Sequences potentially involved in subunit contact included 1-15,11-27, and 33-58. The latter two sequences appear partially inaccessible in heterodimeric hFSH. Based on these data, we hypothesized that the receptor binding determinant for hFSHa might include one or all of these sequences. Therefore, the objectives of the present investigation were to identify monoclonal anti-hFSH mabs that could inhibit hormone binding to receptor. Once identified, it would be possible to determine the epitope specificity of mabs most potent as inhibitors of receptor binding, and thus determine a potential receptor binding sequence.

Materials and Methods Hormones hFSH/3 (NIADDK-NIH, AFP-1194B) and human pituitaries were kindly provided by the National Hormone and Pituitary Distribution Program (Baltimore, MD). Preparations hFSH1781-2 (4000 IU/mg) and 10065 (80 IU/mg) were from Dr. L. E. Reichert Jr. Otherwise, hFSH and hFSHa were purified from frozen human pituitaries as previously described (19), omitting zone electrophoresis and finishing with immunoaffinity purification. Purity and potency were verified as described previously (18). Peptide synthesis Seven peptides derived from the primary structure of the asubunit of hFSH were synthesized and purified as previously described (18) with the exception of sequence 22-39, which was synthesized on an Applied Biosystems solid phase peptide synthesizer (Forest City, CA; model 431A). The composition of each peptide was confirmed by amino acid analysis using a Beckman System Gold amino acid analyzer (Palo Alto, CA; model 126AAA) employing protocols supplied by Beckman. The following peptides corresponding to the primary structure of the a-subunit of hFSH were synthesized: iAPDVQDCPECTLQED^ nTLQEDPFFSQPGAPILQ27 22GAPILQCMGCCFSRAYPT39

Endo • 1991 Voll28»No3

33FSRAYPTPLRSKKTMLVQKNVTSEST58 51KNVTSESTCCVAKSY65 61VAKSYNRVTVMGGFKVEN78 73GFKVENHTACHCSTCYYHKS92

Preparation of mabs The procedure used for fusion of myeloma and spleen cells was that of Claflin and Williams (20), with some modifications. The myeloma cell line was recloned from P3X63-Ag8.653. The spleen cells were from female BALB/c mice. The immunogen was hFSH (80 IU/mg; National Pituitary Agency, Baltimore, MD). Spleen cells were isolated by mincing the spleen and expressing the cells through a stainless steel mesh screen using a 10 ml sterile syringe plunger. Lysis of red cells by ammonium chloride was not done. The screen was rinsed with Hanks buffered salt solution. The spleen cells were then triturated and passed through sterile nylon. Then 1 X 107 myeloma cells (cultured in log phase) and 2.07 x 108 spleen cells were combined in a round bottom 50 ml tube. The tube and contents were centrifuged for 5 min at 1200 rpm in a Beckman table top refrigerated centrifuge. The supernatant was aspirated, and cells were resuspended by swirling in 0.2 ml 50% polyethylene glycol (American Type Culture Collection, Rockville, MD). The suspension was centrifuged at 700 rpm (room temperature) for 3 min. After a total exposure time to polyethylene glycol of 5 min (37 C), 5 ml Dulbecco's modified Eagle's medium were added without resuspending the cells. After 1-2 min the cell pellet was resuspended by gentle intermittent swirling for 3-4 min. Then, the suspension was centrifuged for 5 min at 1200 rpm. Next, Dulbecco's modified Eagle's medium with 0.1 mM hypoxanthine and 16.0 mM thymidine (HT media) was added, and the entire fusion was plated in 96-well plates. Then 24 h later, 2x HAT media (HT media + 0.8 mM Aminopterin) was added. The fusion that yielded 4B, 3A, and 5F (10.4B6,10.3A6, and 10.5F1) used 100 ng hFSH sc in Freunds complete adjuvant for the first immunization. Then eight months later a booster of 100 tig hFSH in Freunds complete adjuvant was given. Four months later, 50 ng hFSH in saline were given ip, and the fusion was carried out 4 days later. For this fusion, the ratio of splenocytes to myelomas was 10:1. The resultant suspension was distributed into five 96 well plates. After 2 weeks, 136 wells had colonies (28%). Three of the colonies were found to have stable secretion of mabs which bound to [125I]hFSH in the double antibody RIA. The splenocytes that gave rise to IE and 2E (22.1E3 and 22.2E9.3) were from a mouse injected sc with 1 mg hFSH in Freunds complete adjuvant. This was followed 1 month later with 1 mg hFSH in incomplete Freunds adjuvant. One month later the mouse was boosted with 1 mg hFSH in saline, and the fusion was performed 3 days later with a ratio of 16:1 splenocytes. The entire fusion was distributed in two 96 well plates. Colonies were present in 64 wells (33%). Eight of the colonies were found to have stable secretion of antibody which bound to [125I]hFSH in the double antibody RIA.

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 18 November 2015. at 18:47 For personal use only. No other uses without permission. . All rights reserved.

EPITOPE MAPPING OF HUMAN FSHa Iodination of hormones The hormones were iodinated as previously described (21, 22) with modifications. To 10 pg hormone (1 mg/ml in 0.01 M sodium phosphate, pH 7.0), 25 pi 0.5 M sodium phosphate pH 7.0,10 pi lactoperoxidase (1 U/ml in 0.01 M sodium phosphate, pH 7.0) (Sigma, St. Louis, MO; no. L-8257) and 1 mCi carrier free Na125I (New England Nuclear, Boston, MA; no. NEZO33H) were added. Next, 5 pi H2O2 (1:30,000) were sequentially added at times 0, 2, 4, 6, 8 min with gentle mixing after each addition. Eleven minutes after the first addition the reaction was terminated by the addition of 100 pi transfer solution (16% sucrose, 1% KI, and 0.01% bromphenol blue), and the solution was applied to a Sephadex G50-150 column. Next, 100 pi transfer solution were added to the vial, and this too was applied to the column. The column was developed in PBS, and 30 1 ml fractions were collected into tubes containing 0.5 ml PBS made 1% with BSA. FSH radioligand receptor assay The receptor assay was performed as previously described (22). Briefly, 200 pi of a 1:10 dilution (wt/vol) of 30,000 x g calf testes membranes were incubated with 1.2 ng [125I]hFSH (25 pCi/pg; preparation 1781-2, 4,000 IU/mg; 30% bindable) previously incubated for 2 h with increments (0.0025-2.5 pi) of each mab. Each tube contained 2.0 mg bovine 7-globulin to obviate the possibility of nonspecific association of mabs with the membranes. The final volume was 500 pi. Nonspecific binding was determined in the presence of 200 pg hFSH (10065, 80 IU/mg). After 16 h at room temperature, the contents of each tube were diluted with 1 ml ice-cold 0.05 M Tris buffer, pH 7.5, and centrifuged at 2,500 x g for 1 h to separate membrane-bound from free hormone. Analysis of mab affinity and concentration In order to determine the affinity constant of each mab and the number of moles of binding sites for FSH in each antibody preparation, a double antibody RIA was used. Briefly, 100 pi normal mouse serum (1:400), 100 pi of each mab, 100 pi [I25I] hFSH (1.2 ng, 1781-2), and 200 pi rabbit antimouse 7-globulin (1:80) were coincubated with graded doses of unlabeled hFSH in a final volume of 1 ml. The reaction was allowed to proceed for 16 h at room temperature, at which time 1 ml ice-cold PBS, pH 7.5, was added, and bound antibody was separated from free by centrifugation at 2,500 X g for 1 h. The resultant data were analyzed with the LIGAND (23) program. Enzyme-linked immunosorbent assay (ELISA) for determining the hFSH subunit specificity of mabs Briefly, hFSH, hFSHa, and hFSH/3 (60 pmol/ml) were diluted in coating buffer (0.05 M Tris buffer, pH 9.5) and 100 pi were added to the wells of a 96 well Immulon I microtiter plate. The plates were incubated for a minimum of 16 h at 4 C at which point the wells were emptied, and 200 pi 1% BSA [immunoglobulin G (IgG)-free] in coating buffer were added and incubated for 2 h at room temperature. The wells were then washed with a PBS wash (PBS containing 0.01% NaN3 and 0.025% Tween 20) using a plate washer. Next, 100 pi mab diluted in ELISA binding buffer (PBS containing 0.25% IgG-

1487

free BSA, 0.01% NaN3, and 1 mM EDTA) were added and incubated for 2 h at room temperature. The wells were then washed and incubated for 2 h at room temperature with 100 pi goat antimouse Ig-alkaline phosphatase conjugate (Tago, Burlingame, CA; no. 6543) diluted to 1:1000 in ELISA binding buffer. The wells were washed, 100 pi of substrate (Bio-Rad, Richmond, CA; no. 172-1063; p-nitrophenyl-phosphate in diethanolamine buffer) were added, and absorbance was determined using a Dynatech microplate reader. Analysis of mab hFSH subunit specificity using an RIA Briefly, 100 pi protein-A purified mab (10 pg/ml) were incubated with 100 pi of either [125I]hFSH, [125I]hFSHa, or [125I] hFSH/3 (40,000 cpm), 100 pi of normal mouse serum (diluted 1:20), and 100 pi of rabbit antimouse Ig (diluted 1:200) in a final volume of 500 pi for a minimum of 16 h at 4 C. The reaction was terminated by the addition of 1 ml ice-cold PBS and bound was separated from free as described above. For the determination of ID50 values for the hFSH subunits, this procedure was modified to use a dilution of mab which resulted in 30% binding of [125I]hFSH in the absence of unlabeled subunit. Displacement curves were then generated by the addition of increasing concentrations of unlabeled hormone. The resultant data were analyzed with the NIHRIA program (24). Immunoradiometric epitope cross-matching The two-site immunoradiometric assay for epitope crossmatching was performed as previously described (15). Each mab was prepared as a solid phase (15). Then each solid phase mab was saturated with hFSH (80 IU/mg). After washing, radiolabeled mabs were added to aliquots of each saturated solid phase mab and incubated followed by separation of bound from free radiolabeled mab (15). The anti-hFSH mabs (areactive) that were cross-matched were 2E, IE, 5F, 3A, and 4B. The positive control mab (/3-reactive) was 3G3. The positive control solid phase mab gives the maximal amount of binding possible for each radioiodinated mab, thus assuring that iodination did not appreciably damage each mab. All immunoradiometric epitope cross-matching data are expressed as percent of the maximal amount of radioiodinated mab bound to the positive control solid phase mab. Nonspecific binding of radiolabeled tracer was determined by incubation of an excess of unlabeled hFSH with the solid phase and the radiolabeled antibody. We reasoned that there may be two variables that could contribute to nonspecific binding. One variable was the solid phase support since each support was used at slightly different dilutions. Nonspecific binding variation due to solid support was not a significant variable. For example, when using radiolabeled 3G3 antibody the average nonspecific binding was 3868 ± 243 cpm per tube (n = 7 different solid phase antibodies). The average specific binding of radiolabeled mab 3G3 to the five monoclonal solid phase antibodies discussed in the manuscript was 45,557 ± 13,450 cpm. We found that most of the variability in nonspecific binding was due to the radiolabeled mab. For example nonspecific binding ranged from 7 to 27% (14.4 ± 4.0%) of the total counts specifically bound when each radiolabeled mab was tested with solid phase 3G3. In this

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 18 November 2015. at 18:47 For personal use only. No other uses without permission. . All rights reserved.

EPITOPE MAPPING OF HUMAN FSHa

1488

case the specific binding ranged from 34,000-100,000 cpm, and the nonspecific binding ranged from 4,300-27,000 cpm. Preparation of peptide affinity columns Briefly, 1 g CNBr-activated Sepharose-4B was washed on a sintered glass filter (G 3) and resuspended in 200 ml 1 mM HC1. Next, 5 mg synthetic peptide were dissolved in coupling buffer (0.1 M NaHCO3, 0.5 M NaCl, pH 8.3). The gel was then washed with 5 ml of coupling buffer, added immediately to peptide solution (a final gel to buffer ratio of 1:2) and mixed in an endover-end mixer for 2 h at room temperature. Next, the gel was transferred to a blocking solution (0.2 M glycine, pH 8.0) and incubated for 2 h at room temperature. Finally, the gel was washed with coupling buffer followed by acetate buffer (0.1 M acetate, 0.5 M NaCl, pH 4.0) and lastly by coupling buffer. The peptide-Sepharose conjugate was then stored in affinity column storage buffer (0.1 M potassium phosphate, 0.3 M NaCl, 0.02% NaN3) pH 7.0) at 4 C. Production and purification of rabbit polyclonal antipeptide antisera on peptide affinity columns Synthetic peptides were coupled to hemocyanin using the water-soluble heterobifunctional coupling reagents l-ethyl-3(3-methylaminopropyl) carbodiimide HC1 (for peptides 11-27, 22-39, 33-58, 51-65, 61-78, and 73-92) and sulfo-succinimidyl 4-(iV-maleimidomethyl) cyclohexane-1-carboxylate (for peptide 1-15) as previously described (18). New York State (Flemish Giant) rabbits (obtained from the Griffin Laboratories, NYSDH, protocol 88-199) were immunized as previously described (18). The serum was isolated after the blood was stored 1 h at room temperature, and proteins in serum were precipitated with a final concentration of 50% ammonium sulfate. The precipitated material was then dialyzed against PBS and stored at —20 C before subsequent purification. Each rabbit antipeptide antisera was purified on its respective peptide affinity column using the method of Karlsen et al. (25) with modifications. Briefly, the affinity column was first washed with 50 ml affinity column binding buffer (ABB) (PBS containing 0.01% NaN3 and 1 mM EDTA). Next, 1 ml antiserum was diluted with 1 ml ABB and applied to the column. The flow was stopped after the sample was completely included. Next, 2 ml ABB were layered on top of the column to prevent drying, and the column was incubated for 2 h at 37 C. After incubation, the column was washed with ABB, and 2 ml fractions were collected. The bound mabs were eluted with 1 M acetic acid collecting 2 ml fractions which were immediately neutralized with 1.1 ml 2 M Tris base. The column was then washed with subsequent 50 ml additions of 1 M acetic acid, affinity column wash buffer (0.1 M potassium phosphate, 0.3 M NaCl, pH 7.0) and affinity column storage buffer. The column fractions were analyzed for both protein concentration (by absorbance at 280 nm) and immunological reactivity to the corresponding peptide using the peptide ELISA as previously described. Epitope mapping of mab 3A using synthetic peptide antigens and affinity chromatography Briefly, 96-well Immulon I microtiter plates (Dynatech Laboratories, Inc. No. 011-010-3350) were coated with 0.19 ± 0.02

Endo«1991 Voll28«No3

nmol (mean ± SE) of each synthetic peptide diluted in 100 ^1 coating buffer (15 mM sodium carbonate, 35 mM sodium bicarbonate, 3 mM sodium azide, pH 9.6) and incubated for a minimum of 16 h at 4 C. The wells were then washed with successive 200 n\ aliquots of wash buffer [1.5 mM potassium phosphate, 20 mM sodium phosphate, 2.7 mM potassium chloride, 0.14 M sodium chloride, pH 7.4, containing 0.05% Tween 20 (vol/vol)], and incubated with 200 n\ 5% Carnation nonfat dry milk in coating buffer for 1 h at room temperature. The wells were washed and then incubated with 100 (A normal mouse Ig (500 ng), 3A (500 ng), or the respective rabbit antipeptide antisera (1:100) diluted in ELISA binding buffer for 16 h at 4 C. The wells were washed and incubated for 2 h at room temperature with 100 n\ goat antimouse or goat antirabbit (Fisher, Rochester, NY; no. OB1294-52) Ig-alkaline phosphatase conjugate diluted 1:1000 in TBST [10 mM Tris-HCl, 150 mM NaCl, pH 8.0, containing 0.05% Tween 20 (vol/vol)]. The wells were washed, 100 ^1 substrate were added, and the absorbance at 410 nm was determined as above.

Results Immunoneutralization of hFSH Each of the a-specific mabs (5F, IE, 2E, 3A, and 4B) as well as negative control mabs (4E, 4F, and 1H) were tested to determine whether they could neutralize (inhibit binding to receptor) hFSH. These results are illustrated in Fig. 1. Whereas the control mabs 4E, 4F, and 1H did not inhibit the binding of [125I]hFSH to receptor, all the hFSHa-specific mabs neutralized hFSH. However, the five hFSHa specific mabs appeared to be divided into two groups with different relative potencies. The mabs IE, 2E, and 5F appeared less potent than r^~^#==_j, =% ^^-^m ——~-

100 T3 C

o CD

• 4E v 4F

80 60

\

^4

O 5F

40

20 \ \

s

*

A 3A

0 .001

.01

.1

1

10

Microliters of Antibody Added FIG. 1. Inhibition of hFSH receptor binding by a-specific mabs 5F, IE, 2E, 3A, and 4B and negative control mabs 4E, 4F, and 1H. Each tube contained [125I]hFSH (1.2 ng) and various dilutions of mab containing ascites, preincubated at room temperature for 2 h. Then 200 n\ of a 1:10 dilution (wt/vol) of 30,000 x g calf testis membranes were added followed by further incubation for 16 h at room temperature. The reaction was terminated by the addition of 1 ml ice-cold 0.05 M Tris buffer, pH 7.5, and centrifuged at 2,500 x g for 1 h. All ascites were dialyzed before use in the assay.

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 18 November 2015. at 18:47 For personal use only. No other uses without permission. . All rights reserved.

EPITOPE MAPPING OF HUMAN FSHa mabs 4B and 3A. In order to determine whether the lower potency was due to a lower affinity or a lower concentration of mab, or both, ligand binding isotherms with each mab were performed, and the results are described below. Analysis of mab affinity and concentration Ligand binding analysis of the five mabs revealed that, as expected for a mab, a single site high affinity model best fit the data. The results of the analysis are presented in Table 1. Mab 5F had the highest affinity constant (K a ) for hFSH, and 2E had the lowest affinity for hFSH. Mab 5F had approximately a 5-fold higher affinity than 4B or 3A. Also listed is the dose of antibody required to inhibit hFSH from binding to receptor by 50% (ID 50 ). Table 1 illustrates that the greater relative potency of the group II mabs (ID50) is not likely due to a greater affinity (K a ) of those mabs for hFSH, nor is it likely due to a greater number of hFSH binding sites (molarity) in the ascites tested. Although the concentration of binding sites in each ascites preparation varied, it is evident that the group II mabs require less mab molecules than the group I mabs to cause receptor inhibition. The estimated collective receptor binding inhibitory dose estimated to give a 50% displacement (ID50) of the group I mabs was 15.9 ± 5.3 X 10"10 mol mab. In contrast, the group II mabs had an ID 50 of approximately 0.024 ± 0.0048 X 10~10 mol mab. In order to determine whether the group I and II antibody epitopes differed, immunoradiometric cross-matching was performed. Immunoradiometric mabs

epitope cross-matching of a-reactive

The results of the immunoradiometric epitope crossmatching studies are listed in Table 2. The data demonstrate that group II mabs 3A and 4B recognize identical epitopes. However, their epitope recognition is different from the other three a-specific group I mabs. TABLE I. Determination of affinity constants and concentration of mabs Group

Antibody

KB (NT1 X 109)

I

5F 2E IE

1.7 ± 0.28 0.099 ± 0.01 0.22 ± 0.035

II

4B 3A

0.332 ± 0.03 0.37 ± 0.06

ID50 (mol x 10~13) 4.23 34.2 12.37 0.0095 0.014

Molarity (X 10"10) 3.32 ± 0.33 26.0 ± 0.21 9.9 ± 1.1 15.8 ± 1.1 24.0 ± 2.4

Results of the analysis of the data from ligand binding isotherms of each monoclonal anti-hFSH« antibody using the LIGAND (23) program. The hFSH used as radioligand and for displacement was LER1781-2 (4,000 IU/mg). Values (Ka, molarity) are the final parameter estimates and the approximate SEs. The ID50 values are the moles of each antibody required to prevent 50% of radiolabeled hFSH from binding to receptor.

1489

TABLE 2. Immunoradiometric cross-matching of antihFSH a-subunit reactive mabs. Solid phase mabs 4B 3A 5F IE 2E

Radiolabeled mabs 4B 4.2 ± 0.9 ± 16.3 ± 67.8 ± 62.0 ±

3A

5F

IE

2E

0.5 2.4 ± 0.1 27.6 ± 1.7 16.9 ± 1.6 ND 0.5 2.3 ± 0.6 34.7 ± 1.7 25.3 ± 1.7 1.0 ± 0.4 0.5 15.9 ± 0.8 1.3 ± 2.5 3.9 ± 1.6 16.7 ± 0.5 1.1 58.3 ± 1.6 8.3 ± 2.7 4.1 ± 0.8 67.6 ± 1.3 0.3 52.2 ± 1.0 ND 0.8 ± 0.2 61.1 ± 0.6

Each of the solid phase antibodies listed, as well as mab 3G3 (15) were saturated with hFSH (80 IU/mg). After washing, the saturated solid phase mab were further incubated with each of the radiolabeled mabs listed. The data are expressed as percent ± SE of the maximal amount of radiolabeled mab which bound to the /^-specific solid phase mab 3G3. The specific counts bound to the positive mab 3G3 (100%) in each case was not less than 35,000 cpm (not shown here). Nonspecific binding for each solid phase mab was determined by coincubating radiolabeled mab with excess hFSH as described in Materials and Methods. ND, Not detectable.

These data complement and confirm the immunoneutralization results. It is unclear to us why 2E exhibited nonreciprocal cross-matching. Indeed, 2E did not exhibit identity with itself. This finding does not impact on this study. It is possible that this is a chimeric mab. Characterization of mab binding specificities Mabs 5F (group I) and 3A (group II) were selected for further characterization due to their different receptor neutralization properties and nonidentity in immunoradiometric cross-matching. Two different techniques, including ELISA and RIA, were used to characterize accurately the binding specificities of these two mabs. In the case of the RIA we first determined whether each radiolabeled subunit could bind to each mab. Then we confirmed those results by displacing radiolabeled hFSH with graded doses of heterdimeric FSH and each subunit, obviating the possibility of radioiodination damage obscuring the results. First, an ELISA was conducted which determined whether mabs 3A and 5F bound to solid phase hFSH, hFSHa, and hFSH|8. The results presented in Table 3 illustrate that both mabs bound hFSH as well as hFSHa with the binding to hFSH being 2-fold higher than to hFSHa. There was no detectable binding of either mab to hFSH/3, and thus these mabs were classified as aspecific. Positive control mab 3H (46.3H6.B7, specific for hFSH/3) bound to hFSH and hFSH,8. The mabs were then characterized using a double antibody RIA which tested whether 3A and 5F bound to [125I]hFSH, [125I]hFSH«, and [125I]hFSH/3. As illustrated in Table 3, mab 3A bound [125I]hFSH and [125I]hFSH«. In contrast, mab 5F only bound to [125I]hFSH. Neither mab had detectable binding to [125I]hFSH|3. Positive

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 18 November 2015. at 18:47 For personal use only. No other uses without permission. . All rights reserved.

EPITOPE MAPPING OF HUMAN FSH«

1490

TABLE 3. Binding specificities of mabs 5F and 3A



100

Mab

Hormone

ELISA0

RIA6

5F

hFSH hFSHa hFSH/3

0.753 ± 0.056 0.375 ± 0.041 0.001 ± 0.001

24,419 ± 219 359 ± 188 441 ± 65

3A

hFSH hFSHa hFSH p

0.673 ± 0.167 0.353 ± 0.078 0.002 ± 0.002

35,316 ± 742 8,534 ± 79 305 ± 30

3H

hFSH hFSHa hFSH/3

0.448 ± 0.033 0.074 ± 0.015 0.777 ± 0.072

38,225 ± 158 52 ± 147 12,297 ± 4

Endo • 1991 Vol 128 • No 3

i

^

^fc^i

control mab 3H for hFSH/3 bound to [125I]hFSH and [125I]hFSH/3. Since the discrepancy between the ELISA and the RIA could have been due to iodination damage of hFSHa!, the mabs were characterized using a RIA in which radiolabeled hFSH was displaced with graded doses of hFSH and each subunit. This assay has the advantage of the sensitivity of an RIA and also eliminates the use of [125IJ hFSHa which did not bind to 5F, thus allowing for the use of noniodinated native hormones for displacement. The results are illustrated in Fig. 2, and they proved quite interesting. hFSH effectively competed with [125I] hFSH with an ID50 = 0.53 x 10"13 mol and an ID50 = 1.05 X 10~13 mol for 5F and 3A respectively. In contrast, there was a large difference in the case of hFSHa. Specifically, hFSHa displaced [125I]hFSH with an ID50 = 0.96 X 1(T13 mol and an ID50 = 1484 x 1(T13 mol for 5F and 3A, respectively. This was unexpected since 3A binds [125]hFSHa, and 5F does not. In addition, hFSH/?, which bound neither 5F nor 3A in the ELISA, can at high dose levels, also displace [125I]hFSH with an ID50 = 90.6 x 10"13 mol and an ID50 > 556 x 10"13 mol, respectively. Affinity purification of rabbit antipeptide antisera for use as positive controls In order to demonstrate that the synthetic peptides remain adsorbed onto the ELISA plate during the course of the epitope mapping experiments (see below) rabbit antipeptide antisera were used as positive controls. The rabbit antipeptide antisera were produced by immunizing rabbits with the seven synthetic peptides coupled to hemocyanin as described in Materials and Methods. The antisera were purified first by ammonium sulfate precipitation and secondly by affinity chromatography on their

• hFSH/?

80 60

\

\

\

40

A

20

The assays were performed as described in Materials and Methods using 6 pmol hormone/100 ^1-well for the ELISA and 40,000 cpm of iodinated hormone/tube for the RIA. The concentration of protein-A purified mab used was 1 ng. Nonspecific binding values, as determined by substituting 1 ng protein-A purified normal mouse Ig for mab, have been subtracted. The values are expressed as the mean ± SE. 0 Absorbance at 410 nm after a 1.5 h incubation. 6 Counts per minute after incubation for 16 h at 4 C.

• hFSH A hFSHa

^

\

3A

O QD

• hFSH A hFSHa • hFSH/J

100 Q.

80 \ 60

\

40

\

\

\

20 5F n

0.1

\

\ 10

100

1,000

10.000

ng/tube FIG. 2. Determination of ID50 values for hFSH, hFSHa, and hFSH/3 using mabs 3A and 5F. Mab (100 MD, [125I]hFSH (40,000 cpm/100 /il), and unlabeled hormone (100 n\) were incubated with rabbit antimouse Ig antiserum (1:20, 100 nl) and normal mouse serum (1:200, 100 /xl) in a final volume of 500 fi\for16 h at 4 C. The reaction was terminated by the addition of 1 ml ice-cold PBS, and the separation of bound from free was accomplished by centrifugation at 2,500 x g for 45 min. Nonspecific binding values were determined by substituting normal mouse Ig for mab.

respective peptide-Sepharose columns. The result of a representative purification is presented in Fig. 3. Clearly, the majority of the protein, 98%, in the sample remained in the breakthrough with 2% eluted with 1 M acetic acid as determined by absorbance at 280 nm. More importantly, the peptide immune reactivity was eluted by 1 M acetic acid as determined by screening the column fractions in a peptide ELISA. Epitope mapping of 3A using an ELISA and affinity chromatography Since 3A was clearly the most potent inhibitor of receptor binding (Table 1 and Fig. 1), its binding specificity was characterized further by epitope mapping. This was accomplished through the use of a peptide ELISA in which equimolar quantities of the seven synthetic peptides were coated onto ELISA plates followed by incubation with either protein-A purified normal mouse Ig or protein-A purified 3A as described in Materials and Methods. As positive controls, to ensure that each peptide remained adsorbed onto the ELISA plate, we also

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 18 November 2015. at 18:47 For personal use only. No other uses without permission. . All rights reserved.

EPITOPE MAPPING OF HUMAN FSHa

0

5

10 15 Fraction (number)

FIG. 3. Affinity purification of hFSHa peptide-specific antiserum (rabbit anti-73-92). hFSHa peptide (a73-92) was covalently coupled to CNBr-Sepharose. One milliliter of ammonium sulfate-precipitated and dialyzed rabbit anti-73-92 antiserum was diluted with 1 ml affinity column binding buffer and incubated for 3 h at 37 C on the affinity column. The column was washed, and the bound mabs were eluted in 1 M acetic acid as described in Materials and Methods. This figure illustrates the immune reactivity to a73-92 in an ELISA using a 1:5 dilution of each column fraction as well as the protein profile as determined by absorption at 280 nm. 20

* a. 15QL

10-

I o

5

1491

purified on a33-58-Sepharose (negative control) and a73-92-Sepharose (positive control). The fractions obtained from the affinity purification of 3A were screened for hFSH binding activity by an ELISA, and the results of such an experiment are illustrated in Fig. 5. When 3A was applied to the «33-58-Sepharose column, all of the hFSH binding activity remained in the breakthrough fractions. In contrast, when 3A was purified on the «7392-Sepharose column some hFSH binding activity was detectable in the breakthrough; however, a large peak of hFSH binding activity was eluted with 1 M acetic acid. Thus, both methods of epitope mapping are in agreement.

Discussion Structure-function analysis of the gonadotropin hormones has always relied on the use of a variety of indirect methods. A direct method would require attaining suitable crystals for x-ray diffraction analysis. This is a formidable task since there is microheterogeneity in the carbohydrate chains, and crystal trials alone would consume tens of milligrams of pure hormone. Several glycoprotein hormones have been reported to be crystallized. These include bovine TSHa, bovine LHa, deglycosylated hCG, and desialated hCG (26-28). However, the much anticipated three-dimensional structures have not yet been reported. Mabs, owing to their specificity, have for many years made excellent tools with which to probe the surface 2.0 TJ

Q

O

A •

'< T

1-15 11-27 22-39 33-58 51-65 61-78 73-92

o 4

1.5

-

•«

1.0

-

i'i 1

0.5

-

'h

hFSHa Synthetic Peptides FlG. 4. Epitope mapping of mab 3A using an ELISA. The assay was performed as described in Materials and Methods using 0.19 ± 0.02 nmol (mean ± SE) of each synthetic peptide per well. The wells were blocked by the addition of 5% nonfat dry milk followed by incubation with protein-A purified mab 3A (500 ng) for 16 h at 4 C. Color was developed for 1.5 h. Nonspecific binding values, as determined by substituting protein-A purified normal mouse Ig (500 ng) for 3A have been subtracted. The values represent the mean ± SE.

0.0 -0

included the seven affinity purified rabbit antipeptide antisera against their respective peptides. Figure 4 illustrates the results obtained by epitope mapping. Mab 3A clearly demonstrated binding to peptide sequence 73-92 and 4-fold less to peptide 61-78. There was no demonstrable binding activity between 3A and peptides 1-15, 11-27, 22-39, 33-58, or 51-65. The seven rabbit antipeptide antisera used as positive controls had absorbance values greater than or equal to 2.00 after incubation with substrate for 1.5 h. To further support these findings 3A was affinity

A

Epitope mapping of human follicle stimulating hormone-alpha using monoclonal antibody 3A identifies a potential receptor binding sequence.

Five monoclonal antibodies (mabs) were generated (3A, 4B, 5F, 2E, 1E) by immunizing BALB/c mice with human (h) FSH. The mabs were used to relate antig...
2MB Sizes 0 Downloads 0 Views