Conversion of human choriogonadotropin into a follitropin by protein engineering

Correlation in HeLa cells of anchorage-independent growth and synthesis of the glycoprotein hormone alpha-subunit

In addition to its normal synthesis in pituitary and placenta for production of the four glycoprotein hormones (GPH), the free alpha-subunit is produced by a variety of tumors and tumor-derived cell lines. It has long been used clinically as a tumor marker, but the question remains as to whether ectopic production of GPH alpha represents the random and chance activation of the hormone gene during cancer development or whether GPH alpha may contribute directly or indirectly to the biology of the tumor. One characteristic of tumorigenic cells in culture is their ability to proliferate in an anchorage-independent way. Data are presented in the following paper to show that the cloning efficiency of HeLa cervical carcinoma cells in soft agar is directly correlated with their production of the GPH alpha-subunit. HeLa variants that differ over 400-fold in their production of GPH alpha similarly exhibited a marked difference in their ability to form colonies in 0.3% noble agar. When HeLa SR3 cells (a variant that produces GPH alpha at high levels) was stably transfected with a vector producing the antisense strand of GPH alpha cDNA, the synthesis of the GPH alpha-subunit was reduced in these cells as was their cloning efficiency in soft agar. Similarly, when HeLa A5F cells (a variant producing little or no GPH alpha) were stably transfected with an alpha-subunit expression vector, the production of GPH alpha was increased significantly in concert with an increased ability to form colonies in 0.3% noble agar. Somatic cell hybrids between HeLa SR3 and HeLa A5F exhibited intermediate levels of GPH alpha gene expression and colony formation in soft agar compared to the parental strains. These data suggest that some parameters of the tumorigenic phenotype, such as anchorage-independent growth, are responsive to, or are dependent upon, the production of free alpha-subunit.

Effect of modification of the beta-hairpin and long loops simultaneously in both alpha- and beta-subunits on the function of human choriogonadotropin: part II

According to the X-ray diffraction, human choriogonadotropin has four beta-hairpin and two long loops, equally distributed in each of the alpha and beta subunits. Radical mutations such as the replacement of alpha 18Phe and alpha 74Phe with Thr in the alpha1 and alpha3 loops respectively and the replacement of alpha 45Lys with Asp in the alpha2 loop in the alpha-subunit were introduced while the loop sequences in the beta-subunit were replaced with the corresponding sequences in hFSH beta. Nine different double mutants with simultaneous mutations in both the alpha and beta loops including hCG alpha1 beta1, hCG alpha1 beta2, hCG alpha1 beta3, hCG alpha2 beta1, hCG alpha2 beta2, hCG alpha2 beta3, hCG alpha3 beta1, hCG alpha3 beta2 and hCG alpha3 beta3 were partially purified from insect High-Five cells. As previously reported (Shao et al., 1996, Mol. Cell. Endocrinol. 122, 173-182), the mutation in the alpha1 loop in the mutant, hCG alpha1 beta, the mutants hCG alpha1 beta1 and hCG alpha1 beta3 caused 200%, increase in the receptor binding, cAMP and progesterone stimulation. The mutant, hCG alpha1 beta2 and all other mutants behaved like the recombinant hCG (rehCG) in the receptor binding and post-receptor signaling activities. The molecular cause for this increase is probably due to a conformational change in the heterodimers caused by the mutation in the alpha1 loop. This conclusion is based on the results of the dissociation studies of the mutants heterodimers which indicated a decreased affinity between the subunits. The first order rate constants for the dissociation of the mutants hCG alpha1 beta1, hCG alpha1 beta2 and hCG alpha beta3 were 3.7 x 10(-2) min(-1), 1.4 x 10(-2) min(-1) and 4.6 x 10(-2) min(-1) respectively, as compared with 4.6 x 10(-3) min(-1) for the rehCG. It seems from the data that alpha 18Phe is located in, or in proximity to the receptor binding site and probably plays a critical role in maintaining either directly or indirectly its conformational integrity.

The biologic action of single-chain choriogonadotropin is not dependent on the individual disulfide bonds of the beta subunit

Disrupting disulfide loops in the human chorionic gonadotropin beta subunit (CGbeta) inhibits combination with the alpha subunit. Because the bioactivity requires a heterodimer, studies on the role of disulfide bonds on receptor binding/signal transduction have previously been precluded. To address this problem, we bypassed the assembly step and genetically fused CGbeta subunits bearing paired cysteine mutations to a wild-type alpha (WTalpha) subunit. The changes altered secretion of the single-chain mutants which parallel that seen for the CGbeta monomeric subunit. Despite conformational changes in CG disulfide bond mutants (assayed by gel electrophoresis and conformationally sensitive monoclonal antibodies), the variants bind to the lutropin/CG receptor and activated adenylate cyclase in vitro. The data show that the structural requirements for secretion and bioactivity are not the same. The results also suggest that the extensive native subunit interactions determined by the cystine bonds are not required for signal transduction. Moreover, these studies demonstrate that the single-chain model is an effective approach to structure-activity relationships of residues and structural domains associated with assembly of multisubunit ligands.

Influence of subunit interactions on lutropin specificity. Implications for studies of glycoprotein hormone function

Bovine lutropin (bLH) and human chorionic gonadotropin (hCG) are heterodimeric glycoprotein hormones required for reproduction. Both bind rat LH receptors (rLHRs), but hCG binds human LH receptors (hLHRs) 1000-10,000 fold better than bLH. We tested the premise that this difference in affinity could be used to identify lutropin receptor contacts. Heterodimers containing hCG/bLH alpha- or beta-subunit chimeras that bound hLHR like hCG (or bLH) were expected to have hCG (or bLH) residues at the receptor contact sites. Analogs containing one subunit derived from hCG bound hLHR much more like hCG than bLH, indicating that each bLH subunit contains all the residues sufficient for high affinity hLHR binding. Indeed, the presence of bovine alpha-subunit residues increased the activities of some hCG analogs. The low hLHR activity of bLH was due primarily to an interaction between its alpha-subunit and beta-subunit residue Leu95. Leu95 does not appear to contact the hLHR since it did not influence the hLHR activity of heterodimers containing human alpha-subunit. These observations show that interactions within and between the subunits can significantly influence the activities of lutropins, thereby confounding efforts to identify ligand residues that contact these receptors.

A defined epitope on the human choriogonadotropin alpha-subunit interacts with the second extracellular loop of the transmembrane domain of the lutropin/choriogonadotropin receptor

The monoclonal antibody, HT13 recognizes human choriogonadotropin (CG) bound to the extracellular domain of its receptor, but not to the full-length receptor. The HT13 epitope is located in the regions of residues 15-17 and 73-75 of the human CG alpha-subunit. Only one synthetic peptide, lutropin (LH)/CG-receptor-(481-497)-peptide (EL2 peptide), which spans the second putative extracellular loop of the LH/CG-receptor endodomain, prevents recognition of human CG by HT13 mAb. EL2 peptide decreases hormone-induced cAMP production, but not high-affinity binding. An anti-EL2 serum also displays the capacity to inhibit human CG-stimulated cAMP production. These results suggest that the second extracellular loop of the receptor is in contact with the HT13 epitope of human CG alpha-subunit and is involved in signal transduction. A relative orientation of the hormone versus the endodomain is proposed.

Engineering human glycoprotein hormone superactive analogues

We report the generation of superactive analogues of human glycoprotein hormones, with potential applications in thyroid and reproductive disorders. Current biological and structural data were used to rationalize mutagenesis. The 11-20 region in the alpha-subunit with a cluster of lysine residues forms a previously unrecognized domain critical for receptor binding and signal transduction, as well as an important motif in the evolution of glycoprotein hormone activities. The gradual elimination of basic residues in the alpha-subunit coincided with the evolutionary divergence of the hominids from the Old World monkeys. By selective reconstitution of certain critical residues present in homologous nonhuman hormones we have developed human thyroid stimulating hormone and chorionic gonadotropin analogues with substantial increases in receptor binding affinity and bioactivity, thus providing a paradigm for the design of novel therapeutic protein analogues.

Follitropin conformational stability mediated by loop 2 beta effects follitropin-receptor interaction

Follicle-stimulating hormone (FSH) is in the family of pituitary/placental glycoprotein hormones which also includes luteinizing hormone (LH), chorionic gonadotropin (hCG), and thyroid-stimulating hormone. These hormones are heterodimers composed of common alpha- and similar but unique beta-subunits. The 21 amino acid loop between Y33 and F53 of the FSH beta-subunit (L2 beta) can be switched into L2 beta of hCG beta without a loss of receptor binding, yet mutation of hFSH beta 37LVY39 to 37AAA39 was antecendent to a 20-fold reduction in receptor binding (based on ID50). A mutation in the LH beta gene, which causes Q54 to be R, causes hypogonadism. This residue is conserved in the glycoprotein hormones and corresponds to Q48 in hFSH beta. Mutation of hFSH beta 48QKTCT52 to 48AAACA52 resulted in a failure of heterodimer formation. In the current study single mutations were made to pinpoint which of the seven hFSH beta residues in the 37LVY39 to 37AAA39 and the 48QKTCT52 to 48AAACA52 mutants were responsible for the observed phenotypes. A single mutation of T52 to alanine was sufficient to cause a reduction in expression of heterodimeric hormone. Single mutants Q48A, T50A, V38A, Y39A, and, to a lesser extent, T52A formed heterodimer. However, these hFSH mutants were markedly unstable at pH 2.0. Thus, acid dissociation can be used to reveal metastable forms of this protein. Mutant hFSH beta Q48A was also 8-fold less active than wild-type hFSH when assayed for binding to hFSH receptors. hFSH beta V38A and Y39A mutants affected receptor binding; however, neither mutation alone caused greater than a 2-fold decrease in receptor binding activity. In summary, these results identify single important residues in the long loop (between Y33 and F53) of the hFSH beta-subunit which are required for proper subunit interactions that provide conformational stability which in turn is necessary for FSH-receptor interaction.

Structural predictions for the ligand-binding region of glycoprotein hormone receptors and the nature of hormone-receptor interactions

BACKGROUND

Glycoprotein hormones influence the development and function of the ovary, testis and thyroid by binding to specific high-affinity receptors. The extracellular domains of these receptors are members of the leucine-rich repeat (LRR) protein superfamily and are responsible for the high-affinity binding. The crystal structure of a glycoprotein hormone, namely human choriogonadotropin (hCG), is known, but neither the receptor structure, mode of hormone binding, nor mechanism for activation, have been established.

RESULTS

Despite very low sequence similarity between exon-demarcated LRRs in the receptors and the LRRs of porcine ribonuclease inhibitor (RI), the secondary structures for the two repeat sets are found to be alike Constraints on curvature and beta-barrel geometry from the sequence pattern for repeated beta alpha units suggest that the receptors contain three-dimensional structures similar to that of RI. With the RI crystal structure as a template, models were constructed for exons 2-8 of the receptors. The model for this portion of the choriogonadotropin receptor is complementary in shape and electrostatic characteristics to the surface of hCG at an identified focus of hormone-receptor interaction.

CONCLUSIONS

The predicted models for the structures and mode of hormone binding of the glycoprotein hormone receptors are to a large extent consistent with currently available biochemical and mutational data. Repeated sequences in beta-barrel proteins are shown to have general implications for constraints on structure. Averaging techniques used here to recognize the structural motif in these receptors should also apply to other proteins with repeated sequences.

Model of human chorionic gonadotropin and lutropin receptor interaction that explains signal transduction of the glycoprotein hormones

The goal of these studies was to devise a model that explains how human chorionic gonadotropin (hCG) interacts with lutropin (LH) receptors to elicit a hormone signal. Here we show that alpha-subunit residues near the N terminus, the exposed surface of the cysteine knot, and portions of the first and third loops most distant from the beta-subunit interface were recognized by antibodies that bound to hCG-receptor complexes. These observations were combined with similar data obtained for the beta-subunit (Cosowsky, L., Rao, S.N.V., Macdonald, G.J., Papkoff, H., Campbell, R.K., and Moyle, W.R. (1995) J. Biol. Chem. 270, 20011-20019), information on residues of hCG that can be changed without disrupting hormone function, the crystal structure of deglycosylated hCG, and the crystal structure of a leucine-repeat protein to devise a model of hCG-receptor interaction. This model suggest that the extracellular domain of the LH receptor is "U-" or "J"-shaped and makes several contacts with the transmembrane domain. High affinity hormone binding results from interactions between residues in the curved portion of the extracellular domain of the receptor and the groove in the hormone formed by the apposition of the second alpha-subunit loop and the first and third beta-subunit loops. Most of the remainder of the hormone is found in the large space between the arms of the extracellular domain and makes few, if any, additional specific contacts with the receptor needed for high affinity binding. Signal transduction is caused by steric or other influences of the hormone on the distance between the arms of the extracellular domain, an effect augmented by the oligosaccharides. Because the extracellular domain is coupled at multiple sites to the transmembrane domain, the change in conformation of the extracellular domain is relayed to the transmembrane domain and subsequently to the cytoplasmic surface of the plasma membrane. While the model does not require the hormone to contact the transmembrane domain to initiate signal transduction, small portions of both subunits may be near the transmembrane domain and assist in initiating the hormonal signal. This is the first model that is consistent with all known information on the activity of the gonadotropins including the amounts of the hormone that are exposed in the hormone-receptor complex, the apparent lack of specific contacts between much of the hormone and the receptor, and the roles of the oligosaccharides in signal transduction.(ABSTRACT TRUNCATED AT 400 WORDS)

The groove between the alpha- and beta-subunits of hormones with lutropin (LH) activity appears to contact the LH receptor, and its conformation is changed during hormone binding

Gonadotropins are heterodimeric glycoprotein hormones that control vertebrate fertility through their actions on gonadal lutropin (luteinizing hormone, LH) and follitropin (follicle-stimulating hormone, FSH) receptors. The beta-subunits of these hormones control receptor binding specificity; however, the region of the beta-subunit that contacts the receptor has not been identified. By a process of elimination we show this contact to be the portions of beta-subunit loops one and three found in a hormone groove created by the juxtaposition of the alpha- and beta-subunits. Most other regions of the beta-subunit can be recognized by antibodies that bind to human chorionic hormone (hCG)-receptor complexes or replaced without disrupting hormone function. Using a series of bovine LH/hCG and human FSH/hCG beta-subunit chimeras we identified key hCG beta-subunit residues in the epitopes of two antibodies that bind to hCG-receptor complexes. These epitopes include the surfaces of beta-subunit loops one and three near residue 74 on the outside of the hormone groove and parts of the C-terminal end of the "seat belt" that holds the two subunits together. The antibody that recognized residue 74 bound to receptor complexes containing most mammalian lutropins better than to the free hormones, an indication that the outside surface of the beta-subunit groove is altered during hormone binding. This region of the beta-subunit is furthest from the alpha-subunit and is recognized equally well in the free beta-subunit and in the heterodimer. Thus, the receptor associated increase in antibody binding appears due to an interaction of this portion of the beta-subunit with the receptor and not to an effect of the receptor on the relative positions of the alpha- and beta-subunits. Unlike most previous studies designed to identify portions of the beta-subunit likely to contact the LH receptor, this indirect approach provides data that are more easily interpreted because it does not rely on the use of mutations that disrupt hormone function. The approach described here should be valuable for studying the receptor interactions of other complex ligands.

On the role of the invariant glutamine at position 54 in the human choriogonadotropin beta subunit

The twelve Cys and eight of the non-Cys residues are invariant in the glycoprotein hormone beta subunits from a variety of mammalian species. beta-Gin-54 of human lutropin (hLH) and choriogonadotropin (hCG) is one of these invariant amino acid residues. A single A-->G mutation in the LH beta gene of a patient presenting with hypogonadism resulted in the replacement of Gin-54 with Arg [1]. The authors also reported that an expressed mutant of hLH beta, with Arg replacing Gin-54, associated with the alpha subunit, but there was no demonstrable binding of the mutant hormone to receptor. We have replaced Gin-54 in hCG beta with Glu and with Lys using site-directed mutagenesis. The expression plasmids pRSV-hCG beta (wild-type and mutants) were transiently transfected into CHO cells containing a stably integrated gene for bovine alpha, and the media were analyzed for holoproteins, which were characterized in vitro using competitive binding and steroidogenic assays with MA-10 cells. hCG beta(Glu-54) bound to alpha almost as well as hCG beta wild-type, and the resulting heterodimer competed with [125l]hCG binding to the LH/CG receptor and stimulated progesterone production to the same extent as the wild-type control. However, the apparent potencies, as judged by ED50s, were less than those of the wild-type control, the effect being more pronounced in binding than in steroidogenesis. In contrast, hCG beta(Lys-54) associated very poorly with alpha. Our results suggest that while Gin-54 in hCG beta participates in receptor binding, its major function appears to involve alpha binding. Such dual functionality leads to interesting models for holoprotein formation and receptor binding.

Structure of human chorionic gonadotropin at 2.6 A resolution from MAD analysis of the selenomethionyl protein

BACKGROUND

Human chorionic gonadotropin (hCG) is a placental hormone that stimulates secretion of the pregnancy-sustaining steroid progesterone. It is a member of a family of glycoprotein hormones that are disulfide-rich heterodimers, with a common alpha-chain and distinctive beta-chains specific to their particular G-protein linked receptors.

RESULTS

We have produced recombinant hCG in mammalian cells as the selenomethionyl protein, and have determined its structure (after partial deglycosylation) at 2.6 A resolution from multiwavelength anomalous diffraction (MAD) measurements. Despite only limited sequence similarity (10% identity), the alpha- and beta-subunits of hCG have similar tertiary folds. Each subunit has a cystine-knot motif at its core of extended hairpin loops. There is a very extensive subunit interface featuring two inter-chain beta-sheets and a unique, disulfide-tethered 'arm' from the beta-subunit which 'embraces' the alpha-subunit. The carboxy-terminal peptide of the beta-subunit, which is rich in O-linked sugars, is disordered.

CONCLUSIONS

Structural and sequence comparisons indicate an evolutionary homology, albeit remote, between the glycoprotein hormone chains and other cystine-knot proteins, notably platelet-derived growth factor. Segments of the alpha- and beta-chains that have been convincingly implicated in receptor binding by hCG are juxtaposed on one side of the molecule. A glycosylation site implicated in signal transduction but not in binding is also close to the presumed binding site suggesting a possible coupling between ligand binding and signaling. This study with selenomethionyl protein produced in mammalian cells extends the realm of MAD phasing.

Co-evolution of ligand-receptor pairs

Specific receptors for lutropin (luteinizing hormone; LH) and follitropin (follicle-stimulating hormone; FSH) mediate the actions of human chorionic gonadotropin (hCG) and FSH5 on the gonads. Here we report that short independent sequences of the beta-subunit enable hCG to distinguish between the receptors for FSH and LH. Residues between the 11th and 12th cysteines restrict FSH receptor binding; residues between the 10th and 11th cysteines and, to a much lesser extent, residues carboxy-terminal to the 12th cysteine also affect LH receptor binding. CF101-109, an hCG analogue containing hFSH beta residues between the 11th and 12th cysteines, had high affinity for both LH and FSH receptors. Modifications to CF101-109 that reduce binding to either LH or FSH receptors yield gonadotropin analogues having differing ratios of LH:FSH activity. Ligand-binding specificity of the LH receptor is determined by residues encoded by parts of exons 2-4 and 7-9 which prevent hFSH binding but have little effect on hCG binding. FSH receptor specificity is controlled primarily by residues encoded by exons 5 and 6 that prevent hCG binding but have little effect on hFSH binding. These determinants can be interchanged to create receptor analogues that bind hCG and hFSH. Our observations support a model in which distinct negative determinants restrict ligand-receptor interaction. This explains coevolution of binding specificity in families of homologous ligands and their receptors. Natural or designed manipulation of these determinants leads to the 'evolution' of new, specific protein-protein interactions.

Primary amenorrhoea and infertility due to a mutation in the beta-subunit of follicle-stimulating hormone

We report a woman with primary amenorrhoea and infertility associated with an isolated deficiency of pituitary follicle-stimulating hormone (FSH), but normal luteinizing hormone (LH) secretion. Ovulation was induced by administration of exogenous FSH and resulted in a successful pregnancy. Sequence analysis of the FSH beta-subunit gene indicated that she is homozygous for a two nucleotide frameshift deletion in the coding sequence. Her mother and son are heterozygous for this mutation. This deletion results in an alteration of amino acid codons 61-86 followed by a premature termination codon. The predicted truncated beta-subunit peptide lacks regions which are important for association with the alpha subunit and for binding to and activation of the FSH receptor. Abnormalities of FSH structure or function might be an under recognised but treatable cause of infertility.

Affinity chromatographic separation of gonadotropic hormone agonist and antagonist antibodies. Implications for structure, immunological and biological properties of glycoproteins

Gonadotropic hormones which have lost peripheral sugar residues in their oligosaccharide chains display antagonistic properties and produce antibodies that fail to recognize the fully glycosylated hormone (agonist form). These polyclonal antibodies were separated by successive affinity chromatography on divinylsulfonyl-Sepharose coupled agonist and antagonist columns. The immunoglobulin G fraction from the agonist affinity columns recognizes both free agonist and antagonist in solution radioimmunoassays and also when these hormone forms are bound to receptors on gonadal cells. However, antagonist-specific antibodies recognize only the free antagonist in solution but not when it is receptor bound, implying that the conformation of the receptor-bound antagonist is different from that of the agonist. Affinity-purified antibodies against the different forms are useful in analyzing immunological and biological properties of the hormones. The studies with these glycosylated hormones serve as a useful model for other glycoproteins of pharmaceutical value.

Design of a long-acting follitropin agonist by fusing the C-terminal sequence of the chorionic gonadotropin beta subunit to the follitropin beta subunit

Follitropin (FSH) is a pituitary glycoprotein hormone that is essential for the development of ovarian follicles and testicular seminiferous tubules. FSH is used clinically to stimulate follicular maturation for in vitro fertilization and treatment of anovulatory women. One issue regarding the clinical use of FSH is its short half-life in the circulation. To address this point, we constructed chimeric genes containing the sequence encoding the C-terminal peptide of the chorionic gonadotropin beta subunit (CG beta) fused to the translated sequence of the human FSH beta subunit (FSH beta). This region of CG beta is important for maintaining the prolonged plasma half-life of human CG dimer. The presence of the C-terminal peptide sequence did not significantly affect assembly of FSH beta with the alpha subunit or secretion of the dimer. In vitro receptor binding and steroidogenic activity of dimer bearing the FSH beta-C-terminal peptide chimera were the same as wild-type FSH. However, both the in vivo potency and half-life in circulation of the dimer bearing either one or two C-terminal peptide units were enhanced. Dimers containing FSH beta-CG beta chimeras could serve as potent FSH agonists for clinical use, and the present strategy may have wide applications for enhancing the in vivo half-life of diverse proteins.

Contributions of arginines-43 and -94 of human choriogonadotropin beta to receptor binding and activation as determined by oligonucleotide-based mutagenesis

Members of the glycoprotein hormone family contain a common alpha subunit and a hormone-specific beta subunit. Human choriogonadotropin (hCG) beta is a 145 amino acid residue protein glycosylated at 6 positions (2 N-linked and 4 O-linked oligosaccharides). In an effort to elucidate receptor determinants on hCG beta, we have used site-directed mutagenesis to prepare and express several mutant cDNAs with replacements at arginines-43 and -94. Arg-43 is invariant in all known mammalian CG/lutropin beta amino acid sequences, and Arg-94 is conserved in 10 of the 12 sequences. Moreover, various studies involving synthetic peptides and enzymatic digestions of intact beta chains suggest that these residues may be important in hCG receptor binding. Point mutants were made in which these two arginines were replaced with the corresponding residues in human follitropin beta, Leu-43 and Asp-94. The wild-type and mutant beta chains were expressed in CHO cells containing a stably integrated gene for bovine alpha, and heterodimer formation occurred. These heterologous gonadotropins were active in assays using transformed Leydig cells, competitive binding with standard 125I-hCG, and cAMP and progesterone production, but the potency was considerably less than that associated with the hCG beta wild-type-containing gonadotropin. The double-mutant protein Arg-43 to Leu/Arg-94 to Asp also associated with bovine alpha, but the resultant heterodimer exhibited only low activity. Replacement of each arginine with lysine yielded heterodimers that were at least as potent as bovine alpha-hCG beta wild type, but the Lys-43-containing beta chain appeared to exhibit a low degree of subunit association or reduced stability relative to the expressed hCG beta wild type. These results demonstrate that arginines-43 and -94 contribute to receptor binding through a positive charge.