74
Biochimica et Biophysica Acta, 1159 (1992) 74-80 © 1992 Elsevier Science Publishers B.V. All rights reserved 0167-4838/92/$05.00
BBAPRO 34288
Immunochemical study of equine chorionic gonadotropin (eCG/PMSG)" antigenic determinants on a- and/3-subunits Marie-Christine
Maurel
a, E l i s a b e t h
B a n a, J e a n - M i c h e l
and Yves Combarnous
Bidart b
a
" 1.N.R.A., Station de Physiologie de la Reproduction, Nouzilly (France) and h Service c~'lmmunologie Mol~culaire, lnstitut Gustave-Roussy, Vi#ejuif (France)
(Received 17 February 1992)
Key words: Chorionic gonadotropin; Monoclonal antibody; Antigenic determinant; Structure-function relationship In the present study we have established an immunochemical mapping of equine Chorionic Gonadotropin (eCG/PMSG) using three monoclonal antibodies (mAbs), namely the antibodies ECG01, El0 and D7, raised against the native hormone. These antibodies do not bind to reduced, alkylated hormone, suggesting that they recognize discontinuous rather than continuous epitopes. We have also assessed the reactivity of mAbs towards human CG, and ovine, porcine, equine and bovine LH and FSH. The antigenic determinant recognized by ECG01 is localized on the a-subunit of equine gonadotropins and of human CG and LH. The epitopes recognized by El0 and D7 mAbs appear to be very similar and are present on the/3-subunit of eCG and of LHs from all species tested, except hLH, as well as on porcine and equine FSHs. Attempts to specify the amino-acid residues involved in these epitopes suggest that ECG01 mAb might preferentially bind to residues around position 70 whereas the region around disulfide bridges Cys-88-Cys-90 might be involved in the epitopes recognized by D7 and El0 mAbs. Topographical relationships of epitopes show that ECG01 mAb never binds to eCG simultaneously with either D7 or El0 mAbs. Furthermore, simultaneous binding of D7 and El0 mAbs on eCG could not be achieved. Thus, these three epitopes appear to be closely located on the surface of eCG. Finally, ECG01 mAb inhibits eCG binding to LH and FSH receptors, suggesting that its antigenic site is closely related to hormone-receptor interaction site(s).
Introduction Equine chorionic gonadotropin (eCG), previously known as Pregnant M a r e Serum G o n a d o t r o p i n (PMSG), is secreted by trophoblastic cells during gestation. It belongs to the family of glycoprotein hormones which comprises the pituitary hormones Luteinizing H o r m o n e (LH), Follicle Stimulating H o r m o n e (FSH) and Thyroid Stimulating H o r m o n e (TSH). These hormones are heterodimers composed of two non-covalently associated a- and /3-subunits. Within a species the amino-acid sequence of a-subunits is identical for the different glycoprotein hormones whereas the /3subunits of each hormone possess unique sequences
Correspondence to: M.C. Maurel, I.N.R.A., Station de Physiologie de la Reproduction, F-37380 Nouzilly, France. Abbreviations: eCG, equine chorionic gonadotropin; PMSG, pregnant mare serum gonadotropin; hCG, human chorionic go_ nadotropin; LH, luteinizing hormone; FSH, follicle stimulating hormone; TSH, thyroid stimulating hormone; mAb, monoclonal antibody.
and confer the hormonal specificity [1]. These two subunits are encoded by separate genes, a unique copy for a [2] and one or multiple copies for the /3-subunit [3,4]. Chorionic gonadotropins have only been demonstrated in primates and in equidaes. Equine C G is detected in the serum of the pregnant mare between days 40 and 130 of pregnancy [5]. Its molecular weight has been estimated to be about 45 000, 15 000 for the a- and 30 000 for the /3-subunit, and it is highly glycosylated [5,6]. The amino-acid sequences of both eCG subunits have been determined [7,8] and they are exactly identical to those of e L H subunits [9,10]. Equine C G only exhibits LH-like activity in the horse [11], but it possesses both LH and FSH activities when administered to non-equine species [12,13]. As highly purified e C G presents this dual activity, the expression of both L H and FSH activities is likely to be due to the same molecule, rather than to distinct variants having FSH or LH activity [14]. Monoclonal antibodies (mAbs) constitute powerful topographical probes that can be used to examine the 3-dimensional structure of hormones and to define their functional sites. There have been many reports
75 relating the immunochemical mapping of human CG [15-19], TSH [20], LH [21] and FSH [22-24]. In contrast to the case for human gonadotropins, little information is available on the immunochemistry of eCG [25-27]. In the present study, we have investigated the immunochemical features of eCG using three mAbs, namely ECG01, El0 and D7, raised against eCG. To define the antigenic determinants on the eCG molecule, we have analysed the binding of mAbs to gonadotropins from various species. Binding inhibition experiments on LH and FSH receptors have been performed so as to localize the antigenic determinants in relation to the interaction site(s) of eCG with its receptors. Materials and Methods
Hormone standards. Highly purified eCG preparations were prepared in our laboratory (eCG-CY-940) or were a generous gift from Dr. D.N. Ward. The aand /3-subunits of eCG were isolated as previously described by Moore and Ward [28]. A crude preparation of eCG (600 IU/mg) was also used, representing material that had not been submitted to ion-exchange chromatography and gel filtration in the preparation of highly purified eCG. Purified hCG (hCG-CR-125) and hLH (hLH-I-3) were kind gifts from the Center for Population Research of the NICHHD, NIH (Bethesda, MD, USA). Purified equine LH (11.4 x NIH-LH-S1), equine FSH (1016 × NIH-FSH-S1), ovine LH (3.1 x NIH-LH-S1) and ovine FSH (22.2 x NIH-FSH-S1) were prepared in our laboratory as previously described [29-32]. Porcine LH (1.2 x NIH-LH-S1) and pFSH (81 x NIH-FSH-S1) as well as bovine FSH (40 x NIHFSH-B1) were kindly given by Dr. Hennen and Dr. Closset (Universit6 de Liege, Belgium). Preparation of reduced S-carboxymethylated eCG. Native eCG was reduced by mercaptoethanol and alkylated with iodoacetic acid as previously reported [33]. After extensive dialysis the reduced and carboxymethylated eCG (RC-eCG) was lyophilized and stored at 4°C. Production of monoclonal antibodies. Monoclonal antibodies were produced as previously reported [18,19]. Anti-eCG antibodies in hybridoma cell culture supernatants were detected by radioimmunoassay using 12SI-eCG. Hybridomas were cloned twice at limiting dilutions, and three monoclonal antibodies, namely ECG01, El0 and D7, were selected. Subtyping of monoclonal antibodies was carried out by ELISA using subclass-specific anti-mouse antibodies (Nordic Immunological Lab., Tilburg, Netherlands).
Study of the monoclonal antibodies binding specific# ties. A competitive ELISA test was developed to quantify the cross-reactivities of mAbs for various gonadotropins. Competition curves were generated by
co-incubating a constant amount of mAb with increasing concentrations of competitive hormone onto coated eCG. Coating was achieved by incubating a crude preparation of eCG (1 /xg/ml), diluted in carbonate buffer 0.1 M (pH 9.6), in Luxlon microtitre plates (C.M.L., Nemours, France). 100 izl was distributed per well and plates were incubated for 1 h at 37°C and for 18 h at 4°C, then washed five times with 10 mM potassium phosphate buffer/NaC1 0.15 M (pH 7.4) containing 0.1% Tween 20 (PBS-Tween). Plates were subsequently saturated with PBS-Tween containing 0.2% BSA (PBS-Tween-BSA) and incubated for 30 min at 37°C. Diluted supernatant from hybridoma cell culture (50 ~l/weli) was added and the plates were incubated for 1 h at 37°C together with increasing concentrations of competitive hormone (50 /zl/well), each diluted in PBS-Tween-BSA. After washing, mAbs bound to coated eCG were detected by using an antimouse IgG antibody coupled to /3-galactosidase (Amersham, UK), diluted 1:1000 in PBS-Tween-BSA (100 /zl/well). After incubation at 37°C for 1 h and washing the plate, fl-galactosidase activity was analysed using 4-methylumbelliferyl-/3-o-galactoside (200 #I/well) as the substrate (Sigma, St. Louis, USA) and tested after 2 h at 42°C. The released 4-methylumbelliferone was measured with a Dynatech Microfluor automatic microtitre plate reader and expressed in absolute fluorescent units at 365 nm for excitation and 450 nm for emission. Concentrations of eCG were calculated, assuming that 10000 IU represent 1 mg of pure hormone. The coating concentration and the supernatant dilution were optimized so as to get the highest sensitivity in competition experiments as described by Nieto et al. [34]. Results were expressed as B / B o (%), where B is the fluorescent signal obtained for each concentration of competitive hormones and Bo the signal obtained without competitive hormone. For quantitative purposes, 3 + 3 or 2 + 2 point parallel line assays were performed with 95% confidence limits [35].
Analysis of the topographical relationships between epitopes. In order to study the topographical relationship of epitopes, two-site 'sandwich' ELISAs were designed. Simultaneous binding of two mAbs onto different regions of eCG was tested by using a coated capture antibody and a second antibody, conjugated to horseradish peroxidase, as the labelled indicator. Briefly, each supernatant that had been diluted 1:4 with carbonate buffer 0.1 M (pH 9.6) was coated and washed five times as described previously. Next, increasing concentrations of eCG were added, the plate was incubated for 1 h at 37°C and then washed. Finally, ECG01 conjugated with peroxidase, diluted at 1 /xg/ml concentration, was incubated as described above. After washing, peroxidase activity was revealed by using orthophenylene-diamine (OPD), diluted in 0.1
76 M acetate buffer (pH 5.6), as the substrate (0.5 m g / m l ) in presence of H 2 0 2 (0.2%). Absorbances were measured with a Titertek Multiscan reader, model MK II (Flow laboratories, McLean, USA) at 492 nm. The preparation of ECG01-peroxidase conjugate was performed by the two-step glutaraldehyde method of Ternynck and Avrameas [36]. Radio-receptor binding inhibition assay. The ability of ECG01 mAb to inhibit the binding activity of eCG to porcine LH and FSH receptors was analysed using a radioligand receptor assay (RRA), performed as previously described [37]. Enriched receptor membrane fractions were prepared from piglet testis. I25I-oLH and ~25I-oFSH in 10 mM Tris-HC1 (pH 7.4) containing 0.1% BSA were used as the tracer hormones. For the binding assay increasing concentrations of eCG (0.5250 n g / 1 0 0 pA), 50/~1 of radiolabelled hormone (20 000 cpm) plus 50 txl of buffer containing 54 mM CaCI 2 were incubated with testicular membranes (400 ~zg protein in 50 pA/tube). After 18 h at 20°C, the testicular membranes were washed with 10 mM Tris-HCl (pH 7.4), centrifuged, and the pellets counted in a gamma counter. To evaluate the inhibitory effect of ECG01 mAb on eCG binding activity, 1 Izg of mAb (50 pA) was pre-incubated for 1 h at 37°C with both increasing concentrations of eCG and radiolabelled hormone prior to mixing with testicular membranes as described above. Results Subtyping of mAbs indicated that all selected antibodies were of the IgG1 subclass.
TABLE I
Characteristics of the monoclonal antibodies Relative binding activities (%) of the three mAbs to the different hormones and subunits, as determined by the ELISA techniques described in Materials and Methods, were calculated on a molar basis, using eCG as the reference (100%). Hormone eCG eCG-a eCG-/3 RC-eCG hCG eLH hLH oLH pLH eFSH oFSH bFSH pFSH
mAb ECG01 100 17 3 .0 11 113 8 0 0 96 0 ND 0
El0 100 1 133 0 1 47 0 22 19 21 0 0 53
D7 100 1 60 0 0 222 1 137 71 26 0 0 44
gins, one from NIH, oFSH-13-AFP-2846-C (15 × NIHFSH-S1) and two from our laboratory, oFSH-CY-1115 (22.2 × NIH-FSH-S1) and oFSH-CY-1756 (51 × NIHFSH-S1). In contrast to E l 0 and D7 mAbs, ECG01 mAb failed to cross-react with pLH, oLH or pFSH but recognized equine L H and FSH exclusively. In summary, ECG01 mAb appears to be specific for equine and, to a lower extent, human a-subunits. E l 0 and D7 mAbs display a specificity for LH /3-subunits excluding human LH, and an FSH specificity restricted to that from porcine and equine species.
Binding specificities of mAbs Table I summarizes the binding specificities of ECG01, E l 0 and D7 mAbs for different glycoprotein hormones and their subunits. The binding of mAbs to chorionic gonadotropins and subunits was analyzed first. Fig. 1 shows that ECG01 mAb recognized eCG-a better than either hCG or eCG-/3 (Panel 1A). In contrast, E 1 0 mAb cross-reacted strongly with eCG-/3 but not with eCG-a or hCG (Panel 2A). D7 mAb displayed a 60% cross-reactivity with eCG-/3 but no significant binding to either eCG-a or hCG (Panel 3A). Further, it appears that ECG01 mAb bound to e L H but did not cross-react with pLH nor o L H (Panel 1B). By contrast, E l 0 and D7 mAbs bound to eLH, p L H and o L H (Panels 2B and 3B). ECG01 mAb displayed a low binding affinity for hLH, while E l 0 and D7 mAbs did not bind to h L H at all. As expected from these results, ECG01 mAb bound to eFSH but not to pFSH or oFSH (Panel 1C). Surprisingly E l 0 mAb, as well as D7 mAb, were found to bind to both eFSH and pFSH but they cross-reacted with neither oFSH nor bFSH (Panels 2C and 3C). These results were confirmed by using three highly purified FSH preparations from different ori-
Recognition of reduced-alkylated eCG by mAbs We next studied the binding of mAbs to RC-eCG using the ELISA competition test. As shown in Table I, none of the mAbs bound to RC-eCG, suggesting that they recognize discontinuous rather than continuous epitopes.
Topographical relationships between epitopes Sandwich ELISAs were performed to assess the topographical relationships of the epitopes. First, we checked the binding of peroxidase-labelled ECG01 mAb to various concentrations of eCG captured either by coated D7 or E l 0 mAb. As shown in Fig. 2, ECG01 mAb did not bind to eCG that had been captured previously by D7 or E l 0 mAbs, regardless of the concentration of hormone used. These results indicate that both E l 0 and D7 mAbs bind to epitopes that are close to the epitope recognized by ECG01 mAb. Control experiments were performed in order to verify that ECG01 mAb is able to bind to eCG that has been captured previously by a polyclonal anti-eCG antibody. In this case, a positive signal, increasing with the con-
77 anti-mouse IgG antibody coupled to /3-galactosidase (data not shown). Whatever the sequence of incubation of the two mAbs, the fluorescent signals were not additive, indicating that the binding of the first mAb prevents the binding of the second. Thus, El0 and D7 mAbs seem to recognize closely related epitopes on eCG. Furthermore, although ECG01 mAb displays
centration of eCG, was obtained. This indicates that ECG01 mAb is able to bind to eCG captured at several different antigenic sites by the polyclonal antibody. The topographical relationship between El0 and D7 mAbs epitopes was further assessed by successively incubating the two mAbs on coated eCG after which the binding of the mAbs was determined by using an
El0
ECG01
percent of binding to coated
D7
mAbs
eCG
2A
IA Ioo
8o oo CGs 40 20
o
1
i
i
i
1
1
I
I
1B
2B
3B
lOO +
8o
oO
fills 40
2O
o
1
1
I
1
I
I
I
I
2C
lC
1
I
3C B
100
80 Q
B
60 FSHs 40 20 0 0
I
,o
,02
,0 3 lo 4
0
1
10
102
10 3 10 4
0
1
10
10 2
10 3 10 4
ng/ml
Fig. 1. Binding specificities of ECG01, El0 and D7 mAbs for: (A), Chorionic Gonadotropins, eCG (o o), hCG (D O) and the subunits eCG-a ( + - - + ) and eCG-fl (* *); (B), LHs from various species, eLH (* *), oLH (D D) and pLH (+ +); (C), FSHs from various species, eFSH (* *), oFSH (13 - - D ) and pFSH ( + +). Curves were generated by the presence of increasing concentrations of competitive hormones and with native eCG as the reference. Dilution of supernatants were 1:250 for El0 and D7 mAbs, and 1:500 for ECG01 mAb, corresponding to 20% of the binding of each antibody to coated eCG in the absence of competitive hormone.
78 quite different immunoreactive properties its epitope L, likely to be overlapping the antigenic region recognized by E l 0 and D7 mAbs.
Absorbonce 0,2 _
0,15
0,1
0,05
i
I
1
I
10
t
102
I
103
104
eCG
ng/ml
Fig. 2. Topographical relationships of epitopes of the three mAbs as assessed by the sandwich ELISA test. The binding of ECG01 m A b conjugated with peroxidase to e C G captured by coated D7 (+ - + ) or E l 0 ( [] [] ) m A b was measured. Positive and negative controls were obtained using polyclonal anti-eCG antibody (* *) or ECG01 m A b itself (A - ix) for coating. Percentage of Specific binding
2,5
............. " ~ ' - ' ;
'lie ~
""
"
2O
15
10
5-
0
i
0
!
i i1|.1
|
1
i
I iiiiii
!
10
i
! i11.!
!
!
I ~lll~|
tO0
1000 concentr'ot ion ng/tube
eCG
Fig. 3. Binding inhibition experiments on porcine FSH testicular receptors. T h e displacement of 1251-oFSH bound to a preparation of porcine testicular m e m b r a n e s was assessed by increasing the concentration of e C G (* *) on FSH receptors. T h e displacement of bound 12SI-oFSH was abolished by previously incubating the e C G with ECG01 m A b ( I I).
Binding inhibition experiments on porcine FSH and LI-t testicular receptors The displacement of ~25I-oFSH bound to a preparation of porcine testicular membranes was first assessed by increasing the concentration of eCG on FSH receptors (Fig. 3). When eCG had previously been incubated with ECG01 mAb (1 tzg/tube) the displacement ot bound t25I-oFSH was abolished for up to a concentration of 125 ng of eCG per tube. Similar results were observed by using t25I-oLH as the tracer in the LH receptor assay. ECG01 mAb was able to inhibit the displacement induced by eCG. These results demonstrate that ECG01 mAb prevents the interaction of eCG on both FSH and LH receptors, suggesting that its antigenic site is probably located close to the binding site of eCG to its receptors. Unfortunately, we were unable to detect any influence of the anti-eCG/3 antibodies on the binding of eCG to its receptors, probably due to a low concentration of mAbs in the supernatants of hybridomas. Discussion
The aim of the present study was to obtain a preliminary immunochemical characterization of eCG. To our knowledge, such a study has not yet been performed on equine gonadotropins and in particular on eCG. Presently, only two anti-eCG mAbs have been elicited and described. One, named Neutra-PMSG, was produced and selected by Intervet International (Holland) for the purpose of shortening the action of eCG in vivo [25]. A second one, ECG01 mAb, has previously been shown to be specific for the eCG-a subunit [26]. An immunochemieal study of eCG has been carried out using polyclonal antisera [27] to examine the immunochemical relationships between eCG, equine pituitary gonadotropins and their subunits. In the present study, we have selected three mAbs El0, D7 and ECG01 raised against native eCG. Competition tests showed that not all the mAbs cross-reacted with reduced, alkylated eCG, suggesting that the epitopes recognized by these mAbs are discontinuous. Moreover, all three mAbs cross-reacted with eLH, as well as with eCG. As the amino-acid sequences, but not the oligosaccharide chains of both e L H and eCG are identical [9,10], these mAbs are likely to be directed against epitopes assembled from the polypeptidic moiety. In order to define the location of the epitopes of these three mAbs, we analysed their pattern of recognition of various gonadotropins from different species. In agreement with a previous report [26], ECG01 mAb
79 was found to be specific for eCG-a. Nevertheless, the affinity of the ECG01 mAb is more than 5-fold less for eCG-a than for native dimeric eCG, indicating that its epitope on the a-subunit has undergone probably a conformational change upon dissociation from /3-subunit. Moreover, our results showed that ECG01 mAb is highly specific for equine gonadotropins and to a lesser extent for human gonadotropins. Consequently, one or several residues in the equine and human a-subunits, that differ in porcine and ovine a-subunits, might be involved in the antigenic site of the ECG01 mAb. Inspection of the sequences of a-subunits [38] suggests that two contiguous residues, Arg-71 and Val-72 might be critical in the binding of ECG01 mAb. Moreover, all a-subunits possess a Thr in position 70, except equine (Ile) and human (Asn) a-subunits. Therefore, the sequence Ile-70-Arg-71-Val-72 of the equine c~-subunit might be involved in the epitope recognized by ECG01 mAb. The difference at position 70 (Ile/Asn) between e L H a and hCGa might explain why ECG01 mAb recognizes the equine subunit better than the human one. ECG01 mAb inhibits binding of eCG to LH and FSH receptors, indicating that its epitope might be partly localized on a region specifically involved in the interaction of eCG with its receptors. In previous reports, the C-terminal portion of the a-subunit, particularly regions a-71-85 and a-81-92 in hCG-a [39], has been shown to be critical in the binding of gonadotropins to their receptors [1]. Moreover, this region contains Tyr-87 and His-93 in the equine a-subunit [9], in contrast to His-87 and Tyr-93 in the other species. This interchange might contribute to the specificity of ECG01 mAb for equine a-subunit but, as position 87 is likely to be buried in native gonadotropins [40], only His-93 is expected to be accessible. In conclusion, our data suggest that in the equine a-subunit the short sequence Ile-70-Arg-71-Val-72 together with His-93 might be involved in the epitope recognized by ECG01 mAb. D7 and El0 mAbs have been found to bind to similar or overlapping epitopes located on the eCG /3-subunit, since they display very similar but not strickly identical patterns of recognition. Indeed, both mAbs cross-react with eLH, pLH and oLH, but they exhibit different binding activities values towards these three gonadotropins. They do not bind to either hLH or hCG. Interestingly, these mAbs also bind to equine and porcine FSH, exhibiting very similar values of binding activity, but do not recognize ovine or bovine FSH. In an attempt to localize the amino-acid residues potentially involved in the epitopes recognized by D7 and El0 mAbs, a comparison of the sequences of C G / L H /3-subunits from human, equine, porcine and ovine species was made [38]. It appears that 16 residues are identical in both eCG and eLH, pLH and oLH, but differ in the hLH and hCG/3-subunits. Since El0 and
D7 mAbs bind to both equine and porcine FSH but not to ovine or bovine FSH, a comparison of the sequences indicate that His-89 might be a critical residue in the overlapping epitopes recognized by D7 and El0 mAbs. It is noteworthy that this residue is located within two disulfide bridges, namely /3-34-88 and /3-9-90, connecting the highly conserved CAGYC region and the N-terminal region, respectively. Though speculative, our hypothesis is substantiated by recent results, indicating that /3-8-10 and /3-89-92 are involved in adjacent epitopes on hCG/3 [41]. Although the epitopes recognized on the a- (ECG01 mAb) and the /3-subunit (D7 and El0 mAbs) of eCG are distinct, they appear to be overlapping, as proven by the steric hindrance that prohibits the simultaneous binding of the different mAbs on the surface of eCG. Several studies, using chemical modifications [42], immunochemical [19] and mutagenetic approaches [43], have indicated that the C-terminal region of the a-subunit is at the surface of the molecule, but that it is partly shielded by the /3-subunit. Our results are in agreement with these observations, since: (i), the epitope of ECG01 mAb appears to be located mostly on the C-terminal region of the a-subunit; (ii), ECG01 mAb binding is inhibited by the binding of El0 or D7 mAbs to the /3-subunit and (iii), ECG01 mAb inhibits the binding of eCG to LH and FSH porcine testicular receptors. Altogether, the present study indicates that, as detected for the human hormone, distinct antigenic determinants are located on the equine choriogonadotropin. The mAbs selected constitute powerful tools for the study of the structure-activity relationship of this hormone that presents an unique pattern of bioactivity, i.e., C G / L H and FSH, among mammalian gonadotropins.
Acknowledgements We are grateful to the National Institute of Child Health and Human Development and to Drs. D.N. Ward, C. Cahoreau, G. Hermen and J. Closset for providing purified hormones and subunits.
References 1 Pierce, G.J. and Parsons, T.F. (1981) Annu. Rev. Biochem. 50, 465-495. 2 Boothby, M., Ruddon, R.W., Anderson, C., McWilliams, D. and Boime, I. (1981) J. Biol. Chem. 256, 5121-5127. 3 Fiddes, J.C. and Talmage, K. (1984) Recent Prog. Horm. Res. 40, 43-75. 4 Policastro, P.F., Daniels-McQueen, S., Carle, G. and Boime, I. (1986) J. Biol. Chem. 261, 5907-5916. 5 Murphy, B.D. and Martinuk, S.D. (1991) Endocr. Rev. 12, 27-44. 6 Combarnous, Y., Salesse, R. and Garnier, J. (1981) Biochim. Biophys. Acta 667, 267-276.
80 7 Moore, W.T., Ward, D.N. and Burleigh, B.D. (1979) Fed. Proc. 38, 462. 8 Sugino, H., Bousfield, G.R., Moore, W.T. and Ward, D.N. (1987) J. Biol. Chem. 262, 8603-8609. 9 Ward, D.N., Moore, W.T. and Burleigh, B.D. (1982) J. Protein Chem. 1,263-280. 10 Bousfield, G.R., Liu, W.K., Sugino, H. and Ward, D.N. (1987) J. Biol. Chem. 262, 8610-8620. 11 Stewart, F. and Allen, W.R. (1979) J. Reprod. Fert. Suppl. 27, 431-440. 12 Moore, W.T. and Ward, D.N. (1980) J. Biol. Chem. 255, 69306936. 13 Combarnous, Y., Guillou, F., Martinat, N. and Cahoreau, C. (1984) Ann. Endocr. 45, 261-268. 14 Moyle, W.R., Erickson, G., Bahl, O.P., Christakos, S. and Gutowski, J. (1978) Am. J. Physiol. 235, E218-E230. 15 Schwarz, S., Berger, P. and Wick, G. (1986) Endocrinology 118, 189-196. 16 Bidart, J.M., Troalen, F., Salesse, R., Bousfield, G.R., Bohuon, C.J. and Bellet, D.H. (1987) J. Biol. Chem. 262, 8551-8556. 17 Bidart, J.M., Troalen, F., Bohuon, C.J., Hennen, G. and Bellet, D.H. (1987) J. Biol. Chem. 262, 15483-15489. 18 Bidart, J.M., Troalen, F., Bousfield, G.R., Birken, S. and Bellet, D.H. (1988) J. Biol. Chem. 263, 10364-10369. 19 Troalen, F., Bellet, D.H., Ghillani, P., Puisieux, A., Bohuon, C.J. and Bidart, J.M. (1988) J. Biol. Chem. 263, 10370-10376. 20 Benkirane, M.M., Bon, D., Costagliola, S., Paolucci, F., Darbouret, B., Prince, P. and Carayon, P. (1987) Endocrinology 121, 1171-1177. 21 Alonso-Whipple, C., Couet, M.L., Doss, R., Koziarz, J., Ogunro, E.A. and Crowley, W.F. (1988) Endocrinology 123, 1854-1860. 22 Krystek, S.R., Dias, J.A., Reichert, L.E. and Andersen, T.T. (1985) Endocrinology 117, 1125-1131. 23 Hojo, H. and Ryan, R.J. (1985) Endocrinology 117, 2428-2434. 24 Berger, P., Panmoung, W., Khaschabi, D., Mayregger, B. and Wick, G. (1988) Endocrinology 123, 2351-2359. 25 Dieleman, S.J. and Bevers, M.M. (1987) J. Reprod. Fert. 81, 533-542.
26 Bidart, J.M., Troalen, F., Bousfield, G.R., Bohuon, C. and Bellet, D. (1989) Endocrinology 124, 923-929. 27 Farmer, S.W. and Papkoff, H. (1979) Biol. Reprod. 21,425-431. 28 Moore, W.T. and Ward, D.N. (1980) J. Biol. Chem. 255, 69236929. 29 Combarnous, Y. and Henge, M.H. (1981) J. Biol. Chem. 256, 9567-9572. 30 Guillou, F. and Combarnous, Y. (1983) Biochim. Biophys. Acta 755, 229-236. 31 Papkoff, H., Gospodarowicz, D. and Li, C.H. (1967) Arch. Biochem. Biophys. 120, 434-440. 32 Papkoff, H. and Gan, J. (1970) Arch. Biochem. Biophys. 136, 522-528. 33 Combarnous, Y. (1976) Etude de l'hormone lut6inisante: purification, structure covalente et conformation, Th?~se de Doctorat bs Sciences. 34 Nieto, A., Gaya, A., Jansa, M., Moreno, C. and Vives, J. (1984) Mol. Immunol. 21,537-543. 35 Colquhoum, D. (1971) Lectures on biostatistics, Clarenton press, Oxford. 36 Ternynck, Th. and Avrameas, S. (1987) Techniques immunoenzymatiques, pp. 34-35, INSERM, Paris. 37 Combarnous, Y., Hennen, G. and Ketelslegers, J.M. (1978) FEBS Lett. 90, 65-68. 38 Ward, D.N., Bousfield, G.R. and Mar, A.O. (1989) in Structurefunction relationship of gonadotropins (Bellet, D. and Bidart, J.M., eds.), pp. 1-19, Raven Press, New York. 39 Charlesworth, C., McCormick, D.J., Benjamin, M. and Ryan, R. (1987) J. Biol. Chem. 262, 13409-13416. 40 Weiner, R.S., Andersen, T.T. and Dias, J.A. (1990) Endocrinology 127, 573-579. 41 Moyle, W.R., Matzuk, M.M., Campbell, R.K., Cogliani, E., Dean-Emig, D.M., Krichevsky, A., Barnett, R.W. and Boime, I. (1990) J. Biol. Chem. 265, 8511-8518. 42 Combarnous, Y. and Maghuin-Rogister, G. (1974) Eur. J. Biochem. 42, 13-19. 43 Yoo, J., Ji, 1. and Ji, T.H. (1991) J. Biol. Chem. 266, 17741-17743.
Biochimica et Biophysica Acta, 1159 (1992) 74-80 © 1992 Elsevier Science Publishers B.V. All rights reserved 0167-4838/92/$05.00
BBAPRO 34288
Immunochemical study of equine chorionic gonadotropin (eCG/PMSG)" antigenic determinants on a- and/3-subunits Marie-Christine
Maurel
a, E l i s a b e t h
B a n a, J e a n - M i c h e l
and Yves Combarnous
Bidart b
a
" 1.N.R.A., Station de Physiologie de la Reproduction, Nouzilly (France) and h Service c~'lmmunologie Mol~culaire, lnstitut Gustave-Roussy, Vi#ejuif (France)
(Received 17 February 1992)
Key words: Chorionic gonadotropin; Monoclonal antibody; Antigenic determinant; Structure-function relationship In the present study we have established an immunochemical mapping of equine Chorionic Gonadotropin (eCG/PMSG) using three monoclonal antibodies (mAbs), namely the antibodies ECG01, El0 and D7, raised against the native hormone. These antibodies do not bind to reduced, alkylated hormone, suggesting that they recognize discontinuous rather than continuous epitopes. We have also assessed the reactivity of mAbs towards human CG, and ovine, porcine, equine and bovine LH and FSH. The antigenic determinant recognized by ECG01 is localized on the a-subunit of equine gonadotropins and of human CG and LH. The epitopes recognized by El0 and D7 mAbs appear to be very similar and are present on the/3-subunit of eCG and of LHs from all species tested, except hLH, as well as on porcine and equine FSHs. Attempts to specify the amino-acid residues involved in these epitopes suggest that ECG01 mAb might preferentially bind to residues around position 70 whereas the region around disulfide bridges Cys-88-Cys-90 might be involved in the epitopes recognized by D7 and El0 mAbs. Topographical relationships of epitopes show that ECG01 mAb never binds to eCG simultaneously with either D7 or El0 mAbs. Furthermore, simultaneous binding of D7 and El0 mAbs on eCG could not be achieved. Thus, these three epitopes appear to be closely located on the surface of eCG. Finally, ECG01 mAb inhibits eCG binding to LH and FSH receptors, suggesting that its antigenic site is closely related to hormone-receptor interaction site(s).
Introduction Equine chorionic gonadotropin (eCG), previously known as Pregnant M a r e Serum G o n a d o t r o p i n (PMSG), is secreted by trophoblastic cells during gestation. It belongs to the family of glycoprotein hormones which comprises the pituitary hormones Luteinizing H o r m o n e (LH), Follicle Stimulating H o r m o n e (FSH) and Thyroid Stimulating H o r m o n e (TSH). These hormones are heterodimers composed of two non-covalently associated a- and /3-subunits. Within a species the amino-acid sequence of a-subunits is identical for the different glycoprotein hormones whereas the /3subunits of each hormone possess unique sequences
Correspondence to: M.C. Maurel, I.N.R.A., Station de Physiologie de la Reproduction, F-37380 Nouzilly, France. Abbreviations: eCG, equine chorionic gonadotropin; PMSG, pregnant mare serum gonadotropin; hCG, human chorionic go_ nadotropin; LH, luteinizing hormone; FSH, follicle stimulating hormone; TSH, thyroid stimulating hormone; mAb, monoclonal antibody.
and confer the hormonal specificity [1]. These two subunits are encoded by separate genes, a unique copy for a [2] and one or multiple copies for the /3-subunit [3,4]. Chorionic gonadotropins have only been demonstrated in primates and in equidaes. Equine C G is detected in the serum of the pregnant mare between days 40 and 130 of pregnancy [5]. Its molecular weight has been estimated to be about 45 000, 15 000 for the a- and 30 000 for the /3-subunit, and it is highly glycosylated [5,6]. The amino-acid sequences of both eCG subunits have been determined [7,8] and they are exactly identical to those of e L H subunits [9,10]. Equine C G only exhibits LH-like activity in the horse [11], but it possesses both LH and FSH activities when administered to non-equine species [12,13]. As highly purified e C G presents this dual activity, the expression of both L H and FSH activities is likely to be due to the same molecule, rather than to distinct variants having FSH or LH activity [14]. Monoclonal antibodies (mAbs) constitute powerful topographical probes that can be used to examine the 3-dimensional structure of hormones and to define their functional sites. There have been many reports
75 relating the immunochemical mapping of human CG [15-19], TSH [20], LH [21] and FSH [22-24]. In contrast to the case for human gonadotropins, little information is available on the immunochemistry of eCG [25-27]. In the present study, we have investigated the immunochemical features of eCG using three mAbs, namely ECG01, El0 and D7, raised against eCG. To define the antigenic determinants on the eCG molecule, we have analysed the binding of mAbs to gonadotropins from various species. Binding inhibition experiments on LH and FSH receptors have been performed so as to localize the antigenic determinants in relation to the interaction site(s) of eCG with its receptors. Materials and Methods
Hormone standards. Highly purified eCG preparations were prepared in our laboratory (eCG-CY-940) or were a generous gift from Dr. D.N. Ward. The aand /3-subunits of eCG were isolated as previously described by Moore and Ward [28]. A crude preparation of eCG (600 IU/mg) was also used, representing material that had not been submitted to ion-exchange chromatography and gel filtration in the preparation of highly purified eCG. Purified hCG (hCG-CR-125) and hLH (hLH-I-3) were kind gifts from the Center for Population Research of the NICHHD, NIH (Bethesda, MD, USA). Purified equine LH (11.4 x NIH-LH-S1), equine FSH (1016 × NIH-FSH-S1), ovine LH (3.1 x NIH-LH-S1) and ovine FSH (22.2 x NIH-FSH-S1) were prepared in our laboratory as previously described [29-32]. Porcine LH (1.2 x NIH-LH-S1) and pFSH (81 x NIH-FSH-S1) as well as bovine FSH (40 x NIHFSH-B1) were kindly given by Dr. Hennen and Dr. Closset (Universit6 de Liege, Belgium). Preparation of reduced S-carboxymethylated eCG. Native eCG was reduced by mercaptoethanol and alkylated with iodoacetic acid as previously reported [33]. After extensive dialysis the reduced and carboxymethylated eCG (RC-eCG) was lyophilized and stored at 4°C. Production of monoclonal antibodies. Monoclonal antibodies were produced as previously reported [18,19]. Anti-eCG antibodies in hybridoma cell culture supernatants were detected by radioimmunoassay using 12SI-eCG. Hybridomas were cloned twice at limiting dilutions, and three monoclonal antibodies, namely ECG01, El0 and D7, were selected. Subtyping of monoclonal antibodies was carried out by ELISA using subclass-specific anti-mouse antibodies (Nordic Immunological Lab., Tilburg, Netherlands).
Study of the monoclonal antibodies binding specific# ties. A competitive ELISA test was developed to quantify the cross-reactivities of mAbs for various gonadotropins. Competition curves were generated by
co-incubating a constant amount of mAb with increasing concentrations of competitive hormone onto coated eCG. Coating was achieved by incubating a crude preparation of eCG (1 /xg/ml), diluted in carbonate buffer 0.1 M (pH 9.6), in Luxlon microtitre plates (C.M.L., Nemours, France). 100 izl was distributed per well and plates were incubated for 1 h at 37°C and for 18 h at 4°C, then washed five times with 10 mM potassium phosphate buffer/NaC1 0.15 M (pH 7.4) containing 0.1% Tween 20 (PBS-Tween). Plates were subsequently saturated with PBS-Tween containing 0.2% BSA (PBS-Tween-BSA) and incubated for 30 min at 37°C. Diluted supernatant from hybridoma cell culture (50 ~l/weli) was added and the plates were incubated for 1 h at 37°C together with increasing concentrations of competitive hormone (50 /zl/well), each diluted in PBS-Tween-BSA. After washing, mAbs bound to coated eCG were detected by using an antimouse IgG antibody coupled to /3-galactosidase (Amersham, UK), diluted 1:1000 in PBS-Tween-BSA (100 /zl/well). After incubation at 37°C for 1 h and washing the plate, fl-galactosidase activity was analysed using 4-methylumbelliferyl-/3-o-galactoside (200 #I/well) as the substrate (Sigma, St. Louis, USA) and tested after 2 h at 42°C. The released 4-methylumbelliferone was measured with a Dynatech Microfluor automatic microtitre plate reader and expressed in absolute fluorescent units at 365 nm for excitation and 450 nm for emission. Concentrations of eCG were calculated, assuming that 10000 IU represent 1 mg of pure hormone. The coating concentration and the supernatant dilution were optimized so as to get the highest sensitivity in competition experiments as described by Nieto et al. [34]. Results were expressed as B / B o (%), where B is the fluorescent signal obtained for each concentration of competitive hormones and Bo the signal obtained without competitive hormone. For quantitative purposes, 3 + 3 or 2 + 2 point parallel line assays were performed with 95% confidence limits [35].
Analysis of the topographical relationships between epitopes. In order to study the topographical relationship of epitopes, two-site 'sandwich' ELISAs were designed. Simultaneous binding of two mAbs onto different regions of eCG was tested by using a coated capture antibody and a second antibody, conjugated to horseradish peroxidase, as the labelled indicator. Briefly, each supernatant that had been diluted 1:4 with carbonate buffer 0.1 M (pH 9.6) was coated and washed five times as described previously. Next, increasing concentrations of eCG were added, the plate was incubated for 1 h at 37°C and then washed. Finally, ECG01 conjugated with peroxidase, diluted at 1 /xg/ml concentration, was incubated as described above. After washing, peroxidase activity was revealed by using orthophenylene-diamine (OPD), diluted in 0.1
76 M acetate buffer (pH 5.6), as the substrate (0.5 m g / m l ) in presence of H 2 0 2 (0.2%). Absorbances were measured with a Titertek Multiscan reader, model MK II (Flow laboratories, McLean, USA) at 492 nm. The preparation of ECG01-peroxidase conjugate was performed by the two-step glutaraldehyde method of Ternynck and Avrameas [36]. Radio-receptor binding inhibition assay. The ability of ECG01 mAb to inhibit the binding activity of eCG to porcine LH and FSH receptors was analysed using a radioligand receptor assay (RRA), performed as previously described [37]. Enriched receptor membrane fractions were prepared from piglet testis. I25I-oLH and ~25I-oFSH in 10 mM Tris-HC1 (pH 7.4) containing 0.1% BSA were used as the tracer hormones. For the binding assay increasing concentrations of eCG (0.5250 n g / 1 0 0 pA), 50/~1 of radiolabelled hormone (20 000 cpm) plus 50 txl of buffer containing 54 mM CaCI 2 were incubated with testicular membranes (400 ~zg protein in 50 pA/tube). After 18 h at 20°C, the testicular membranes were washed with 10 mM Tris-HCl (pH 7.4), centrifuged, and the pellets counted in a gamma counter. To evaluate the inhibitory effect of ECG01 mAb on eCG binding activity, 1 Izg of mAb (50 pA) was pre-incubated for 1 h at 37°C with both increasing concentrations of eCG and radiolabelled hormone prior to mixing with testicular membranes as described above. Results Subtyping of mAbs indicated that all selected antibodies were of the IgG1 subclass.
TABLE I
Characteristics of the monoclonal antibodies Relative binding activities (%) of the three mAbs to the different hormones and subunits, as determined by the ELISA techniques described in Materials and Methods, were calculated on a molar basis, using eCG as the reference (100%). Hormone eCG eCG-a eCG-/3 RC-eCG hCG eLH hLH oLH pLH eFSH oFSH bFSH pFSH
mAb ECG01 100 17 3 .0 11 113 8 0 0 96 0 ND 0
El0 100 1 133 0 1 47 0 22 19 21 0 0 53
D7 100 1 60 0 0 222 1 137 71 26 0 0 44
gins, one from NIH, oFSH-13-AFP-2846-C (15 × NIHFSH-S1) and two from our laboratory, oFSH-CY-1115 (22.2 × NIH-FSH-S1) and oFSH-CY-1756 (51 × NIHFSH-S1). In contrast to E l 0 and D7 mAbs, ECG01 mAb failed to cross-react with pLH, oLH or pFSH but recognized equine L H and FSH exclusively. In summary, ECG01 mAb appears to be specific for equine and, to a lower extent, human a-subunits. E l 0 and D7 mAbs display a specificity for LH /3-subunits excluding human LH, and an FSH specificity restricted to that from porcine and equine species.
Binding specificities of mAbs Table I summarizes the binding specificities of ECG01, E l 0 and D7 mAbs for different glycoprotein hormones and their subunits. The binding of mAbs to chorionic gonadotropins and subunits was analyzed first. Fig. 1 shows that ECG01 mAb recognized eCG-a better than either hCG or eCG-/3 (Panel 1A). In contrast, E 1 0 mAb cross-reacted strongly with eCG-/3 but not with eCG-a or hCG (Panel 2A). D7 mAb displayed a 60% cross-reactivity with eCG-/3 but no significant binding to either eCG-a or hCG (Panel 3A). Further, it appears that ECG01 mAb bound to e L H but did not cross-react with pLH nor o L H (Panel 1B). By contrast, E l 0 and D7 mAbs bound to eLH, p L H and o L H (Panels 2B and 3B). ECG01 mAb displayed a low binding affinity for hLH, while E l 0 and D7 mAbs did not bind to h L H at all. As expected from these results, ECG01 mAb bound to eFSH but not to pFSH or oFSH (Panel 1C). Surprisingly E l 0 mAb, as well as D7 mAb, were found to bind to both eFSH and pFSH but they cross-reacted with neither oFSH nor bFSH (Panels 2C and 3C). These results were confirmed by using three highly purified FSH preparations from different ori-
Recognition of reduced-alkylated eCG by mAbs We next studied the binding of mAbs to RC-eCG using the ELISA competition test. As shown in Table I, none of the mAbs bound to RC-eCG, suggesting that they recognize discontinuous rather than continuous epitopes.
Topographical relationships between epitopes Sandwich ELISAs were performed to assess the topographical relationships of the epitopes. First, we checked the binding of peroxidase-labelled ECG01 mAb to various concentrations of eCG captured either by coated D7 or E l 0 mAb. As shown in Fig. 2, ECG01 mAb did not bind to eCG that had been captured previously by D7 or E l 0 mAbs, regardless of the concentration of hormone used. These results indicate that both E l 0 and D7 mAbs bind to epitopes that are close to the epitope recognized by ECG01 mAb. Control experiments were performed in order to verify that ECG01 mAb is able to bind to eCG that has been captured previously by a polyclonal anti-eCG antibody. In this case, a positive signal, increasing with the con-
77 anti-mouse IgG antibody coupled to /3-galactosidase (data not shown). Whatever the sequence of incubation of the two mAbs, the fluorescent signals were not additive, indicating that the binding of the first mAb prevents the binding of the second. Thus, El0 and D7 mAbs seem to recognize closely related epitopes on eCG. Furthermore, although ECG01 mAb displays
centration of eCG, was obtained. This indicates that ECG01 mAb is able to bind to eCG captured at several different antigenic sites by the polyclonal antibody. The topographical relationship between El0 and D7 mAbs epitopes was further assessed by successively incubating the two mAbs on coated eCG after which the binding of the mAbs was determined by using an
El0
ECG01
percent of binding to coated
D7
mAbs
eCG
2A
IA Ioo
8o oo CGs 40 20
o
1
i
i
i
1
1
I
I
1B
2B
3B
lOO +
8o
oO
fills 40
2O
o
1
1
I
1
I
I
I
I
2C
lC
1
I
3C B
100
80 Q
B
60 FSHs 40 20 0 0
I
,o
,02
,0 3 lo 4
0
1
10
102
10 3 10 4
0
1
10
10 2
10 3 10 4
ng/ml
Fig. 1. Binding specificities of ECG01, El0 and D7 mAbs for: (A), Chorionic Gonadotropins, eCG (o o), hCG (D O) and the subunits eCG-a ( + - - + ) and eCG-fl (* *); (B), LHs from various species, eLH (* *), oLH (D D) and pLH (+ +); (C), FSHs from various species, eFSH (* *), oFSH (13 - - D ) and pFSH ( + +). Curves were generated by the presence of increasing concentrations of competitive hormones and with native eCG as the reference. Dilution of supernatants were 1:250 for El0 and D7 mAbs, and 1:500 for ECG01 mAb, corresponding to 20% of the binding of each antibody to coated eCG in the absence of competitive hormone.
78 quite different immunoreactive properties its epitope L, likely to be overlapping the antigenic region recognized by E l 0 and D7 mAbs.
Absorbonce 0,2 _
0,15
0,1
0,05
i
I
1
I
10
t
102
I
103
104
eCG
ng/ml
Fig. 2. Topographical relationships of epitopes of the three mAbs as assessed by the sandwich ELISA test. The binding of ECG01 m A b conjugated with peroxidase to e C G captured by coated D7 (+ - + ) or E l 0 ( [] [] ) m A b was measured. Positive and negative controls were obtained using polyclonal anti-eCG antibody (* *) or ECG01 m A b itself (A - ix) for coating. Percentage of Specific binding
2,5
............. " ~ ' - ' ;
'lie ~
""
"
2O
15
10
5-
0
i
0
!
i i1|.1
|
1
i
I iiiiii
!
10
i
! i11.!
!
!
I ~lll~|
tO0
1000 concentr'ot ion ng/tube
eCG
Fig. 3. Binding inhibition experiments on porcine FSH testicular receptors. T h e displacement of 1251-oFSH bound to a preparation of porcine testicular m e m b r a n e s was assessed by increasing the concentration of e C G (* *) on FSH receptors. T h e displacement of bound 12SI-oFSH was abolished by previously incubating the e C G with ECG01 m A b ( I I).
Binding inhibition experiments on porcine FSH and LI-t testicular receptors The displacement of ~25I-oFSH bound to a preparation of porcine testicular membranes was first assessed by increasing the concentration of eCG on FSH receptors (Fig. 3). When eCG had previously been incubated with ECG01 mAb (1 tzg/tube) the displacement ot bound t25I-oFSH was abolished for up to a concentration of 125 ng of eCG per tube. Similar results were observed by using t25I-oLH as the tracer in the LH receptor assay. ECG01 mAb was able to inhibit the displacement induced by eCG. These results demonstrate that ECG01 mAb prevents the interaction of eCG on both FSH and LH receptors, suggesting that its antigenic site is probably located close to the binding site of eCG to its receptors. Unfortunately, we were unable to detect any influence of the anti-eCG/3 antibodies on the binding of eCG to its receptors, probably due to a low concentration of mAbs in the supernatants of hybridomas. Discussion
The aim of the present study was to obtain a preliminary immunochemical characterization of eCG. To our knowledge, such a study has not yet been performed on equine gonadotropins and in particular on eCG. Presently, only two anti-eCG mAbs have been elicited and described. One, named Neutra-PMSG, was produced and selected by Intervet International (Holland) for the purpose of shortening the action of eCG in vivo [25]. A second one, ECG01 mAb, has previously been shown to be specific for the eCG-a subunit [26]. An immunochemieal study of eCG has been carried out using polyclonal antisera [27] to examine the immunochemical relationships between eCG, equine pituitary gonadotropins and their subunits. In the present study, we have selected three mAbs El0, D7 and ECG01 raised against native eCG. Competition tests showed that not all the mAbs cross-reacted with reduced, alkylated eCG, suggesting that the epitopes recognized by these mAbs are discontinuous. Moreover, all three mAbs cross-reacted with eLH, as well as with eCG. As the amino-acid sequences, but not the oligosaccharide chains of both e L H and eCG are identical [9,10], these mAbs are likely to be directed against epitopes assembled from the polypeptidic moiety. In order to define the location of the epitopes of these three mAbs, we analysed their pattern of recognition of various gonadotropins from different species. In agreement with a previous report [26], ECG01 mAb
79 was found to be specific for eCG-a. Nevertheless, the affinity of the ECG01 mAb is more than 5-fold less for eCG-a than for native dimeric eCG, indicating that its epitope on the a-subunit has undergone probably a conformational change upon dissociation from /3-subunit. Moreover, our results showed that ECG01 mAb is highly specific for equine gonadotropins and to a lesser extent for human gonadotropins. Consequently, one or several residues in the equine and human a-subunits, that differ in porcine and ovine a-subunits, might be involved in the antigenic site of the ECG01 mAb. Inspection of the sequences of a-subunits [38] suggests that two contiguous residues, Arg-71 and Val-72 might be critical in the binding of ECG01 mAb. Moreover, all a-subunits possess a Thr in position 70, except equine (Ile) and human (Asn) a-subunits. Therefore, the sequence Ile-70-Arg-71-Val-72 of the equine c~-subunit might be involved in the epitope recognized by ECG01 mAb. The difference at position 70 (Ile/Asn) between e L H a and hCGa might explain why ECG01 mAb recognizes the equine subunit better than the human one. ECG01 mAb inhibits binding of eCG to LH and FSH receptors, indicating that its epitope might be partly localized on a region specifically involved in the interaction of eCG with its receptors. In previous reports, the C-terminal portion of the a-subunit, particularly regions a-71-85 and a-81-92 in hCG-a [39], has been shown to be critical in the binding of gonadotropins to their receptors [1]. Moreover, this region contains Tyr-87 and His-93 in the equine a-subunit [9], in contrast to His-87 and Tyr-93 in the other species. This interchange might contribute to the specificity of ECG01 mAb for equine a-subunit but, as position 87 is likely to be buried in native gonadotropins [40], only His-93 is expected to be accessible. In conclusion, our data suggest that in the equine a-subunit the short sequence Ile-70-Arg-71-Val-72 together with His-93 might be involved in the epitope recognized by ECG01 mAb. D7 and El0 mAbs have been found to bind to similar or overlapping epitopes located on the eCG /3-subunit, since they display very similar but not strickly identical patterns of recognition. Indeed, both mAbs cross-react with eLH, pLH and oLH, but they exhibit different binding activities values towards these three gonadotropins. They do not bind to either hLH or hCG. Interestingly, these mAbs also bind to equine and porcine FSH, exhibiting very similar values of binding activity, but do not recognize ovine or bovine FSH. In an attempt to localize the amino-acid residues potentially involved in the epitopes recognized by D7 and El0 mAbs, a comparison of the sequences of C G / L H /3-subunits from human, equine, porcine and ovine species was made [38]. It appears that 16 residues are identical in both eCG and eLH, pLH and oLH, but differ in the hLH and hCG/3-subunits. Since El0 and
D7 mAbs bind to both equine and porcine FSH but not to ovine or bovine FSH, a comparison of the sequences indicate that His-89 might be a critical residue in the overlapping epitopes recognized by D7 and El0 mAbs. It is noteworthy that this residue is located within two disulfide bridges, namely /3-34-88 and /3-9-90, connecting the highly conserved CAGYC region and the N-terminal region, respectively. Though speculative, our hypothesis is substantiated by recent results, indicating that /3-8-10 and /3-89-92 are involved in adjacent epitopes on hCG/3 [41]. Although the epitopes recognized on the a- (ECG01 mAb) and the /3-subunit (D7 and El0 mAbs) of eCG are distinct, they appear to be overlapping, as proven by the steric hindrance that prohibits the simultaneous binding of the different mAbs on the surface of eCG. Several studies, using chemical modifications [42], immunochemical [19] and mutagenetic approaches [43], have indicated that the C-terminal region of the a-subunit is at the surface of the molecule, but that it is partly shielded by the /3-subunit. Our results are in agreement with these observations, since: (i), the epitope of ECG01 mAb appears to be located mostly on the C-terminal region of the a-subunit; (ii), ECG01 mAb binding is inhibited by the binding of El0 or D7 mAbs to the /3-subunit and (iii), ECG01 mAb inhibits the binding of eCG to LH and FSH porcine testicular receptors. Altogether, the present study indicates that, as detected for the human hormone, distinct antigenic determinants are located on the equine choriogonadotropin. The mAbs selected constitute powerful tools for the study of the structure-activity relationship of this hormone that presents an unique pattern of bioactivity, i.e., C G / L H and FSH, among mammalian gonadotropins.
Acknowledgements We are grateful to the National Institute of Child Health and Human Development and to Drs. D.N. Ward, C. Cahoreau, G. Hermen and J. Closset for providing purified hormones and subunits.
References 1 Pierce, G.J. and Parsons, T.F. (1981) Annu. Rev. Biochem. 50, 465-495. 2 Boothby, M., Ruddon, R.W., Anderson, C., McWilliams, D. and Boime, I. (1981) J. Biol. Chem. 256, 5121-5127. 3 Fiddes, J.C. and Talmage, K. (1984) Recent Prog. Horm. Res. 40, 43-75. 4 Policastro, P.F., Daniels-McQueen, S., Carle, G. and Boime, I. (1986) J. Biol. Chem. 261, 5907-5916. 5 Murphy, B.D. and Martinuk, S.D. (1991) Endocr. Rev. 12, 27-44. 6 Combarnous, Y., Salesse, R. and Garnier, J. (1981) Biochim. Biophys. Acta 667, 267-276.
80 7 Moore, W.T., Ward, D.N. and Burleigh, B.D. (1979) Fed. Proc. 38, 462. 8 Sugino, H., Bousfield, G.R., Moore, W.T. and Ward, D.N. (1987) J. Biol. Chem. 262, 8603-8609. 9 Ward, D.N., Moore, W.T. and Burleigh, B.D. (1982) J. Protein Chem. 1,263-280. 10 Bousfield, G.R., Liu, W.K., Sugino, H. and Ward, D.N. (1987) J. Biol. Chem. 262, 8610-8620. 11 Stewart, F. and Allen, W.R. (1979) J. Reprod. Fert. Suppl. 27, 431-440. 12 Moore, W.T. and Ward, D.N. (1980) J. Biol. Chem. 255, 69306936. 13 Combarnous, Y., Guillou, F., Martinat, N. and Cahoreau, C. (1984) Ann. Endocr. 45, 261-268. 14 Moyle, W.R., Erickson, G., Bahl, O.P., Christakos, S. and Gutowski, J. (1978) Am. J. Physiol. 235, E218-E230. 15 Schwarz, S., Berger, P. and Wick, G. (1986) Endocrinology 118, 189-196. 16 Bidart, J.M., Troalen, F., Salesse, R., Bousfield, G.R., Bohuon, C.J. and Bellet, D.H. (1987) J. Biol. Chem. 262, 8551-8556. 17 Bidart, J.M., Troalen, F., Bohuon, C.J., Hennen, G. and Bellet, D.H. (1987) J. Biol. Chem. 262, 15483-15489. 18 Bidart, J.M., Troalen, F., Bousfield, G.R., Birken, S. and Bellet, D.H. (1988) J. Biol. Chem. 263, 10364-10369. 19 Troalen, F., Bellet, D.H., Ghillani, P., Puisieux, A., Bohuon, C.J. and Bidart, J.M. (1988) J. Biol. Chem. 263, 10370-10376. 20 Benkirane, M.M., Bon, D., Costagliola, S., Paolucci, F., Darbouret, B., Prince, P. and Carayon, P. (1987) Endocrinology 121, 1171-1177. 21 Alonso-Whipple, C., Couet, M.L., Doss, R., Koziarz, J., Ogunro, E.A. and Crowley, W.F. (1988) Endocrinology 123, 1854-1860. 22 Krystek, S.R., Dias, J.A., Reichert, L.E. and Andersen, T.T. (1985) Endocrinology 117, 1125-1131. 23 Hojo, H. and Ryan, R.J. (1985) Endocrinology 117, 2428-2434. 24 Berger, P., Panmoung, W., Khaschabi, D., Mayregger, B. and Wick, G. (1988) Endocrinology 123, 2351-2359. 25 Dieleman, S.J. and Bevers, M.M. (1987) J. Reprod. Fert. 81, 533-542.
26 Bidart, J.M., Troalen, F., Bousfield, G.R., Bohuon, C. and Bellet, D. (1989) Endocrinology 124, 923-929. 27 Farmer, S.W. and Papkoff, H. (1979) Biol. Reprod. 21,425-431. 28 Moore, W.T. and Ward, D.N. (1980) J. Biol. Chem. 255, 69236929. 29 Combarnous, Y. and Henge, M.H. (1981) J. Biol. Chem. 256, 9567-9572. 30 Guillou, F. and Combarnous, Y. (1983) Biochim. Biophys. Acta 755, 229-236. 31 Papkoff, H., Gospodarowicz, D. and Li, C.H. (1967) Arch. Biochem. Biophys. 120, 434-440. 32 Papkoff, H. and Gan, J. (1970) Arch. Biochem. Biophys. 136, 522-528. 33 Combarnous, Y. (1976) Etude de l'hormone lut6inisante: purification, structure covalente et conformation, Th?~se de Doctorat bs Sciences. 34 Nieto, A., Gaya, A., Jansa, M., Moreno, C. and Vives, J. (1984) Mol. Immunol. 21,537-543. 35 Colquhoum, D. (1971) Lectures on biostatistics, Clarenton press, Oxford. 36 Ternynck, Th. and Avrameas, S. (1987) Techniques immunoenzymatiques, pp. 34-35, INSERM, Paris. 37 Combarnous, Y., Hennen, G. and Ketelslegers, J.M. (1978) FEBS Lett. 90, 65-68. 38 Ward, D.N., Bousfield, G.R. and Mar, A.O. (1989) in Structurefunction relationship of gonadotropins (Bellet, D. and Bidart, J.M., eds.), pp. 1-19, Raven Press, New York. 39 Charlesworth, C., McCormick, D.J., Benjamin, M. and Ryan, R. (1987) J. Biol. Chem. 262, 13409-13416. 40 Weiner, R.S., Andersen, T.T. and Dias, J.A. (1990) Endocrinology 127, 573-579. 41 Moyle, W.R., Matzuk, M.M., Campbell, R.K., Cogliani, E., Dean-Emig, D.M., Krichevsky, A., Barnett, R.W. and Boime, I. (1990) J. Biol. Chem. 265, 8511-8518. 42 Combarnous, Y. and Maghuin-Rogister, G. (1974) Eur. J. Biochem. 42, 13-19. 43 Yoo, J., Ji, 1. and Ji, T.H. (1991) J. Biol. Chem. 266, 17741-17743.
