Cooperative immunoassays: ultrasensitive assays with mixed monoclonal antibodies.

Mixtures of certain monoclonal antibodies appear to bind human chorionic gonadotropin in a "cooperative" fashion because they form circular complexes with the hormone. Experiments illustrate how this property might be exploited to develop very sensitive immunoassays for human chorionic gonadotropin or any other antigen. Since the assays are not based on competitive inhibition between radiolabeled and unlabeled antigen, they are much more sensitive than a traditional radioimmunoassay in which either one of the same antibodies is used alone.

Chimeric proteins can exceed the sum of their parts: implications for evolution and protein design.

Chimeric analogs derived from pairs of homologous proteins routinely exhibit activities found in one or both parents. We describe chimeras of two glycoprotein hormones, human chorionic gonadotropin (hCG) and human follitropin (hFSH), that exhibit activity unique to a third family member, human thyrotropin (hTSH). The results show that biological activity can be separated from hormone-specific amino acid residues. This is consistent with a model for the evolution of homologous ligand-receptor pairs involving gene duplication and the creation of inhibitory determinants that restrict binding. Disruption of these determinants can unmask activities characteristic of other members of a protein family. Combining portions of two ligands to create analogs with properties of a third family member can facilitate identifying key determinants of protein-protein interaction and may be a useful strategy for creating novel therapeutics. In the case of the glycoprotein hormones, this showed that two different hormone regions (i.e., the seat-belt and the intersubunit groove) appear to limit inappropriate contacts with receptors for other members of this family. These observations also have important caveats for chimera-based protein design because an unexpected gain of function may limit the therapeutic usefulness of some chimeras.

The surface of alpha-subunit loop 1 distant from the subunit interface is exposed in the hCG lutropin receptor complex.

Interactions of the placental glycoprotein hormone human choriogonadotropin (hCG) with lutropin receptors (LHR) are required for maintenance of early pregnancy. Knowledge of how hCG interacts with LHR is useful for understanding the mechanism of receptor function, an issue of considerable debate. A large surface of hCG remains exposed after the hormone binds the LHR and can be readily detected with monoclonal antibodies. Here we show that the surface of hCG alpha-subunit loop 1 furthest from the beta-subunit interface can also be recognized by a monoclonal antibody when hCG is bound to the LHR. This extends the area of hCG known to be exposed in the hormone receptor complex, an observation that further restricts models of hCG-LHR interaction.

Functional homodimeric glycoprotein hormones: implications for hormone action and evolution.

BACKGROUND: Human chorionic gonadotropin (hCG), lutropin, follitropin, and thyrotropin act as alpha beta heterodimers to control reproduction and thyroid function. The alpha and beta subunits of these proteins are divided into three loops (alpha 1,alpha 2,alpha 3; beta 1,beta 2,beta 3) by cysteine knots and the heterodimer is stabilized by 20 beta-subunit residues wrapped around alpha 2 like a seatbelt. Understanding how these hormones interact with their receptors, a matter of considerable dispute, would facilitate design of pro- and anti-fertility agents. RESULTS: By swapping alpha 2 for beta 2 and vice versa and, in some cases, adding an amino-terminal coiled-coil dimerization domain, we prepared homodimeric analogs that have the conformation found in each 'half' of hCG. Homodimers containing loops beta 1,alpha 2,beta 3 and none, part, or all of the seatbelt stimulated signal transduction to the same extent as hCG, albeit with lower potency. Those containing alpha 1,beta 2,alpha 3 were inactive. CONCLUSIONS: The activities of homodimers containing the beta 1,alpha 2,beta 3 groove exceed those of other minimized analogs more than 100-1000-fold, suggesting this portion of the hormone forms the major receptor contact. The discovery that glycoprotein hormone heterodimers can be converted to functional homodimers supports the proposal that this protein family evolved from an active homodimeric ancestor by gene duplication and acquisition of mutations to loop 2 that prevent homodimerization. This approach to protein minimization should be applicable to other proteins composed of architecturally related subunits, including those that might have arisen by gene duplication.

Quantitative explanation for increased affinity shown by mixtures of monoclonal antibodies: importance of a circular complex.

Mixtures of some but not all monoclonal antibodies which bind to separate epitopes on human chorionic gonadotropin (hCG) show an increased affinity for the hormone. To find an explanation for the increase in affinity, we developed a mathematical model which predicts the quantities of intermediates formed when pairs of IgG1 mouse monoclonal antibodies having affinities of approximately 10(8) M-1 for hCG are mixed with the hormone. At low antibody concentrations (i.e. less than 1 nM or 0.15 micrograms/ml) analysis of possible antibody-hormone combinations, including linear and circular chains composed of less than 12 molecules of antibody and 12 molecules of hCG, suggests the increase in affinity is due to formation of a circular complex containing two molecules of antibody and two of hCG. Further, the model predicts that the circular complex will be the major species formed at antibody-antigen equivalence. This prediction is supported by experimental observations on the molecular weight of a new complex formed in the presence of hCG and the mixture of the monoclonal antibodies. In addition, based on experimental values of binding constants for individual antibodies to hCG, the model correctly quantifies the loss in complex observed in the presence of excess hCG antigen. At high antibody concentrations (i.e. greater than 10 nM or 1.5 micrograms/ml) the formation of linear chains of antibody hCG pairs becomes appreciable and contributes to the increase in apparent affinity of the mixture for hCG. These results suggest that the observed affinity of complex mixtures of antibody for antigens containing multiple epitopes calculated from Scatchard plots may not be related to the affinity or avidity of any of the antibody species for a given epitope.

Threading of a glycosylated protein loop through a protein hole: implications for combination of human chorionic gonadotropin subunits.

Chorionic gonadotropin (hCG) is a heterodimeric placental glycoprotein hormone essential for human reproduction. Twenty hCG beta-subunit residues, termed the seatbelt, are wrapped around alpha-subunit loop 2 (alpha 2) and their positions "latched" by a disulfide formed by cysteines at the end of the seatbelt (Cys 110) and in the beta-subunit core (Cys 26). This unique arrangement explains the stability of the heterodimer but raises questions as to how the two subunits combine. The seatbelt is latched in the free beta-subunit. If the seatbelt remained latched during the process of subunit combination, formation of the heterodimer would require alpha 2 and its attached oligosaccharide to be threaded through a small beta-subunit hole. The subunits are known to combine during oxidizing conditions in vitro, and studies described here tested the idea that this requires transient disruption of the latch disulfide, possibly as a consequence of the thioredoxin activity reported in hCG. We observed that alkylating agents did not modify either cysteine in the latch disulfide (Cys 26 or Cys 110) during heterodimer formation in several oxidizing conditions and had minimal influence on these cysteines during combination in the presence of mild reductants (1--3 mM beta-mercaptoethanol). Reducing agents appeared to accelerate subunit combination by disrupting a disulfide (Cys 93--Cys 100) that forms a loop within the seatbelt, thereby increasing the size of the beta-subunit hole. We propose a mechanism for hCG assembly in vitro that depends on movements of alpha 2 and the seatbelt and suggest that the process of glycoprotein hormone subunit combination may be useful for studying the movements of loops during protein folding.

Acute inhibition of corticosteroidogenesis by inhibitors of calmodulin action.

To identify the possible role of calmodulin in ACTH function, we tested the ability of chlorpromazine (CP) and other calmodulin antagonists to inhibit steroidogenesis of isolated adrenocortical cells of the rat. CP reversibly inhibited maximal ACTH-induced corticosterone (B) production. The presence of the drug did not alter the ED50 of ACTH stimulation (3.2 X 10(3) pg/ml), suggesting that it inhibited ACTH-induced steroidogenesis in a noncompetitive manner. The CP concentration required for half-maximal inhibition was 8.2 microM, a value close to the dissociation constant of the CP-calmodulin complex (5.3 microM). Concentrations greater than 40 microM resulted in complete inhibition. Similar concentrations of CP inhibited ACTH-induced cAMP accumulation in a dose-dependent manner, indicating an effect of the drug on early events in ACTH action. In addition, CP also apparently acted at a site distal to the point of cAMP formation, as shown by the finding that it inhibited cAMP-induced B production. CP inhibition of ACTH-induced B production was independent of the Ca2+ concentration, suggesting that the drug did not compete with Ca2+ directly. Concentrations of CP greater than 20 microM inhibited protein synthesis as measured by leucine incorporation into cellular proteins. Thus, although the inhibitory effect of high concentrations of CP on steroidogenesis might be explained by an effect on protein synthesis, the inhibition seen at 10 microM appeared to be independent of protein synthesis. Other antagonists of calmodulin action inhibited maximal ACTH-induced B production with the following relative potencies: trifluoperazine greater than CP greater than haloperidol greater than chlordiazepoxide. This order is similar to that reported for inhibition of calmodulin-activated phosphodiesterase and for binding to calmodulin. These findings suggest that calmodulin may modulate the effect of ACTH on steroidogenesis at multiple sites.

Use of antisera to follicle-stimulating hormone (FSH) to detect non-FSH factors in human serum which modulate rat granulosa cell steroidogenesis.

We measured the ability of serum from women to stimulate steroidogenesis in cultured granulosa cells. Serum promoted estradiol and progesterone synthesis in proportion to its FHS content measured by RIA [i.e. serum from postmenopausal women (PM) greater than serum from the midcycle at the time of the gonadotropin surge (MC) greater than serum from the first day of the menstrual cycle (D1) greater than serum from a hypophysectomized woman (AP)]. The FSH activity of these sera was reduced but not eliminated when we included excess antisera to ovine or human FSH in the culture medium (i.e. PM greater than MC greater than D1 greater than AP). These antisera completely neutralized the actions of ovine FSH, human FSH, and menopausal gonadotropin (Pergonal) added to serum. In contrast to the stimulation seen with 5% or lower concentrations of serum in the culture medium, we observed that 10-20% serum inhibited FSH-induced androgen aromatization and progesterone accumulation. The degree of stimulation or inhibition of steroidogenesis depended on the number of granulosa cells added to each culture. High initial cell concentrations inhibited the ability of the cells to respond to either serum or PMSG. In addition to factors which stimulate or inhibit FSH-induced steroidogenesis, human serum contains factors distinct from FSH which cause the cells to flatten and adhere more tightly to the culture dishes. Although progesterone synthesis was increased in cells which had flattened on the surface of the culture dishes, this phenomenon was not a prerequisite for serum-induced steroidogenesis. We conclude that serum contains factors immunologically distinct from FSH, possibly of pituitary origin, which induce granulosa cell steroidogenesis. In addition, serum contains inhibitory substances which block hormone-induced steroidogenesis and which tend to obscure the stimulatory effects of FSH. Detection of both factors depends in part on the number of granulosa cells used to innoculate the cell cultures.

Nonreceptor binding of human chorionic gonadotropin (hCG): detection of hCG or a related molecule bound to endometrial tissue during pregnancy using labeled monoclonal antibodies that bind to exposed epitopes on the hormone.

When hCG adsorbs to surfaces, including membranes from tissues that lack specific hCG receptors, it adsorbs with a particular orientation. Some sites on the alpha- and beta-subunits project away from the surface and can be detected with radiolabeled monoclonal antibodies. Other epitopes, which are located on a region on the hormone that presumably contacts the surface, lose their ability to bind antibody. Using antibodies specific for epitopes on hCG which remain exposed and can be detected when the hormone is adsorbed to rat brain homogenates, we found hCG or closely related substances bound to progestational decidual tissues. Immunologically reactive material adsorbed to the decidual tissue increased and decreased in parallel with the serum levels of hCG throughout pregnancy. Binding of labeled monoclonal antibody to substances similar or identical to hCG in other tissues, including placenta and fetal lung, but not red cells, also was identified. Unlike material adsorbed to decidual tissues, receptor-bound hCG was not recognized by any of our alpha-subunit-specific antibodies. This finding suggests either that the adsorbed hormone has a different orientation than receptor-bound hormone or that the adsorbed hormone has dissociated into subunits. These studies represent the first detection of nonreceptor binding of hCG or related molecules to tissues lacking receptors or presumed not to synthesize the hormone. The role of nonreceptor-bound hCG, if any, is unknown. Other than its effects on stimulation of luteal steroidogenesis during early pregnancy, the role of hCG during most of pregnancy has not been determined. Conceivably, the nonreceptor binding we identified is related to a role for hCG in pregnancy that is not associated with an action on the ovarian LH receptor.

Rat cholesterol side-chain cleavage enzyme (P-450scc): use of a cDNA probe to study the hormonal regulation of P-450scc mRNA levels in ovarian granulosa cells.

A rat ovarian cDNA library was constructed and screened by differential colony hybridization to detect cDNA clones specific for mRNA induced by follicle-stimulating hormone (FSH). The cDNA clone which demonstrated the greatest degree of induction contained a 766-bp insert which was characterized and sequenced. We conclude that this cDNA is specific for the rat gene coding for cholesterol side-chain cleavage enzyme (P-450scc) by virtue of nucleotide sequence homology to the bovine and human P-450scc cDNA sequences. Southern blotting of rat genomic DNA suggests the presence of a single P-450scc gene. Northern blot analysis indicates that P-450scc mRNA is present in steroidogenic tissues (ovary, adrenal, testis), but not in brain, kidney, liver, lung, or heart. The rat P-450scc mRNA is induced by FSH or pregnant mare's serum gonadotropin in ovaries of estrogen-treated immature rats in vivo. In cultured granulosa cells, estradiol treatment alone did not increase P-450scc mRNA levels, but in combination with FSH or 8-Br-cAMP resulted in three- to four-fold increase in this mRNA.

Use of the average antibody-antigen bond concept and probability theory to simplify modeling of linear and circular antibody-antigen complex formation.

We illustrate use of a simple approach to describe the equilibrium interactions of mixtures of monoclonal antibodies and antigens. This procedure is based on elementary concepts in probability theory and is readily suited to describing interactions of antibodies and antigens which form circular as well as linear complexes. The method is also suited to describing the inhibitory effects of antibodies which compete for overlapping epitopes and an example is provided to show how the procedure can be used to describe the interactions of antibodies which inhibit circular complex formation. We also outline simple strategies for preparing computer programs to simulate binding of antigens to defined antibody mixtures. The methods described should facilitate design of immunoassay procedures based on the use of defined mixtures of monoclonal antibodies.

Desensitization and resensitization of lutropin receptors expressed in transfected Y-1 adrenal cells.

Stimulation of gonadal cells by lutropins such as human chorionic gonadotropin (hCG) is often transient and followed by down-regulation and/or desensitization of lutropin receptors (LHR). Here we describe desensitization/resensitization of LHR in Y-1 adrenal cell lines (termed Y-1L) expressing a rat cDNA lacking most 5' and 3' LHR untranslated regions under the control of a metallothionein promoter. Using a simple morphological assay in which stimulated cells are round and unstimulated cells are flat, we identified clones that rounded and remained round and others that became insensitive to lutropin stimulation and reverted to their flat appearance within 2-4 h. Flattened cells were insensitive to further hormonal stimulation but rounded after treatments with cholera toxin, forskolin, or cyclic AMP, showing that loss of responsiveness was associated with an early step in signal transduction, not loss of rounding potential. Removing the lutropin stimulus for at least 90-120 min reversed hormone insensitivity, even in the presence of the protein synthesis inhibitor puromycin. The number of surface bound receptors did not change during a cycle of rounding/flattening and hCG bound to rounded or flattened cells was replaced equally by radioiodinated hCG during incubations at 4 degrees C. Thus, desensitization/resensitization of LHR in Y-1L cells occurred in the absence of new receptor synthesis, receptor degradation, or receptor recycling. These observations suggest that LHR desensitization/resensitization in Y-1L cells was closely coupled to receptor occupancy and that this cell line may be useful for identifying factors that modulate the activities of occupied receptors.

Bifunctional hCG analogs adopt different conformations in LH and FSH receptor complexes.

Human reproduction requires specific interactions between follitropin (hFSH) and its receptor (FSHR) and between lutropin (hLH) or choriogonadotropin (hCG) and the lutropin receptor (LHR). Substitution of hFSH residues between hCG beta-subunit cysteines 11-12 creates a bifunctional analog that binds both receptors. To understand the basis of this observation, we used antibody probes to compare the conformations of bifunctional analogs before and after they were complexed with each receptor. Introduction of hFSH residues between cysteines 11-12 changed a distant conformation-sensitive region created by the juxtaposition of the subunit aminotermini. This region, found not to contact either receptor, was altered further when bifunctional ligands bound FSHR. All other surfaces, detected in LHR complexes, were also recognized in FSHR complexes, an indication that bifunctional ligands bind both receptors in similar orientations. These observations suggest that unlike hCG or hFSH, bifunctional gonadotropins can acquire "lutropin" and "follitropin" conformations, a phenomenon accentuated by receptor contacts.

hCG beta residues 94-96 alter LH activity without appearing to make key receptor contacts.

The ability of human chorionic gonadotropin (hCG) to distinguish lutropin (LHR) and follitropin (FSHR) receptors is controlled principally by beta-subunit residues 94-117. To learn how residues 94-96 (Arg-Arg-Ser) influence LHR binding, we studied the effects of replacing them on the LH and FSH activities of a bifunctional hCG analog in which residues 101-109 were derived from FSH. Analogs containing 1-3 arginines and no aspartates at residues 94-96 bound LHR with 25-400% the potency of hCG. When residues 94-96 were neutral or contained 1-3 aspartates, LHR binding was reduced 6-100 fold but remained at least ten-fold greater than the negative control analog containing residues 94-117 derived from FSH. Residues 94-96 had little influence on FSHR binding. These observations support a model [Moyle et al. (1995) J. Biol. Chem. 270:20,020] in which residues 94-96 influence LHR binding specificity primarily through an effect on hormone conformation rather than by direct participation in essential high affinity receptor contacts.

The lutropin beta-subunit N-terminus facilitates subunit combination by offsetting the inhibitory effects of residues needed for LH activity.

Human chorionic gonadotropin (hCG) contains a beta-subunit N-terminal amino acid extension that contacts the alpha-subunit and is needed for efficient alpha and hCG beta-subunit combination. Here we report that an hCG beta-subunit analog, lacking residues 2-8, combined with the alpha-subunit more efficiently when positively charged residues between beta-subunit cysteines 10 and 11 were replaced with negatively charged residues found in the corresponding portion of follitropin. Residues 2-8 had no influence on binding of hCG to lutropin receptors. Positive charges between cysteines 10 and 11 are essential for high affinity binding of lutropins to their receptors. Therefore, the N-terminal extension found in all lutropin beta-subunits appears to have evolved to offset the inhibition of subunit combination by beta-subunit residues that are essential for lutropin activity. This beta-subunit extension is not found in follitropins or thyrotropins, hormones that have negatively charged residues between cysteines 10 and 11.

Assembly and expression of a synthetic gene encoding the bovine glycoprotein hormone alpha-subunit.

The glycoprotein hormones are a family of alpha beta heterodimeric proteins which are responsible for gonadal and thyroid function. In previous studies we employed chimeric glycoprotein hormone beta-subunits to identify amino acid residues critical for binding to receptors and antibodies. To facilitate similar studies of the alpha-subunit of these hormones, we assembled a 406 bp synthetic gene which encodes the human alpha-subunit leader sequence and the secreted portion of the bovine alpha-subunit. It contains unique restriction sites that can be used for cassette mutagenesis or for making human/bovine alpha-subunit chimeras. The gene was assembled from eight long oligodeoxynucleotides in a single ligation and its structure verified by DNA sequencing. Co-transfection of COS-7 cells with the synthetic gene and the cDNA for human chorionic gonadotropin (hCG) beta-subunit resulted in the secretion of a functional alpha beta heterodimer which bound to luteinizing hormone receptors. The protein was recognized by several monoclonal antibodies including B109, an antibody to a conformational epitope which binds hCG but not the free bovine alpha-, human alpha-, or hCG beta-subunits. This suggests that the binding site for B109 may be formed by residues located primarily within the hCG beta-subunit and that formation of this epitope requires a change in conformation of the beta-subunit when it combines with the alpha-subunit.

Cloning of rat lutropin (LH) receptor analogs lacking the soybean lectin domain.

cDNAs coding for rat ovarian luteinizing hormone receptor analogs lacking three of the leucine repeats were detected in a library which had been prepared from rat luteal tissue undergoing human chorionic gonadotropin-induced luteinization. These leucine repeats correspond to amino acids 206-267 and contain the portion of the receptor that is homologous to the soybean lectin. The cDNA library also contained a receptor analog lacking amino acids 321-700 which code for the transmembrane domain. S-1 mapping suggests that this latter form constitutes approximately half of all receptor-related mRNA.

Addition of an N-terminal dimerization domain promotes assembly of hCG analogs: implications for subunit combination and structure-function analysis.

Human chorionic gonadotropin (hCG) is a heterodimeric placental glycoprotein hormone that acts through ovarian lutropin receptors (LHR) to maintain early pregnancy. Its ability to distinguish LHR and follitropin receptors (FSHR) is controlled by 20 beta-subunit 'seatbelt' residues that surround alpha-subunit loop 2. Positively charged amino acids between residues 93-100, a small loop within the seatbelt, have been postulated to make essential LH receptor contacts. Previous studies showed that analogs containing negatively charged amino acids in this small loop had 5-10% the activity of hCG and 1-10% the lutropin activities of hCG/hFSH chimeric analogs capable of binding LHR and FSHR. These effects might be due to the influence of these residues on receptor contacts or on hormone conformation. During efforts to distinguish these possibilities, we increased and decreased the number of residues in this loop, mutations we anticipated would distort its conformation. Consistent with this supposition, these changes inhibited dimer formation, precluding assessment of these mutations on hormone activity. Addition of Fos and Jun dimerization domains to the N-termini of hCGalpha- and hCG/hFSHbeta-subunit chimeras overcame the effects of the seatbelt mutations on subunit combination and enabled preparation of heterodimers containing six, seven, or nine residues in their seatbelt loops. These had 0.1-10% the lutropin and 3-60% the follitropin activities of bifunctional chimeras containing 8 residues derived from hCG in the seatbelt loop. The abilities of N-terminal dimerization domains to promote subunit combination may permit structure/function analysis of other residues that influence heterodimer formation.

Deglycosylation of a bifunctional lutropin-follitropin agonist reduced its follitropin activity more than its lutropin activity.

OBJECTIVE: To design a drug that blocks the gonadal actions of lutropins and follitropins. DESIGN: Controlled in vitro study. SETTING: Academic laboratory. PATIENT(S): None. INTERVENTION(S): We removed three glycosylation signals from an hCG-hFSH chimera known to have high affinity for LH and FSH receptors, expecting this would create a bifunctional antagonist (dgCFC). To offset the inhibition of subunit combination caused by deglycosylation of alpha-subunit loop 2, we prepared dgCFC as a single-chain fusion protein containing the alpha-subunit downstream of the chimeric beta-subunit. MAIN OUTCOME MEASURE(S): Receptor binding, cyclic adenosine monophosphate accumulation. RESULT(S): dgCFC bound LH or FSH receptors similar to hCG or hFSH. It was a partial agonist and had one tenth the efficacy of hFSH and two thirds the efficacy of hCG. CONCLUSION(S): The surprising high residual lutropin activity of dgCFC indicated that its FSH residues offset the effects of deglycosylation, suggesting this approach to preparing a bifunctional antagonist is unlikely to lead to a useful drug. The increased lutropin efficacy of dgCFC relative to deglycosylated hCG supports the idea that oligosaccharides modulate glycoprotein hormone efficacy through an influence on hormone conformation.

Partial restoration of lutropin activity by an intersubunit disulfide bond: implications for structure/function studies.

Gonadal function is controlled by lutropins and follitropins, heterodimeric cystine knot proteins that have nearly identical alpha-subunits. These heterodimeric proteins are stabilized by a portion of the hormone-specific beta-subunit termed the "seatbelt" that is wrapped around alpha-subunit loop 2 (alpha 2). Here we show that replacing human chorionic gonadotropin (hCG) alpha 2 residue Lys51 with cysteine or alanine nearly abolished its lutropin activity, an observation that implies that alpha Lys51 has a key role in hormone activity. The activity of the heterodimer containing alpha K51C, but not that containing alpha K51A, was increased substantially when beta-subunit seatbelt residue beta Asp99 was converted to cysteine. As had been reported by others, heterodimers containing alpha K51C and beta D99C were crosslinked by a disulfide. The finding that an intersubunit disulfide restored some of the activity lost by replacing alpha Lys51 suggests that this residue is not crucial for receptor binding or signaling and also that hCG and related hormones may be particularly sensitive to mutations that alter interactions between their subunits. We propose the unique structures of hCG and related family members may permit some subunit movement in the heterodimer, making it difficult to deduce key residues involved in receptor contacts simply by correlating the activities of hormone analogs with their amino acid sequences.