Mutagenesis of the ‘determinant loop’ region of human choriogonadotropin β

Abstract

The hormone-specific β subunits of the four human glycoprotein hormones are homologous, and mapping studies are underway in many laboratories to delineate the amino acid residues responsible for receptor binding and activation. Results on the human choriogonadotropin β (hCGβ) subunit, obtained using synthetic peptides, chemically modified derivatives, and mutant forms prepared via site-directed mutagenesis, have suggested that amino acid residues enclosed by the purported disulfide loop between Cys-93 and Cys-100 may contribute to receptor binding and perhaps specificity. Indeed, the 93–100 amino acid sequence is referred to as a determinant loop. We have used site-directed mutagenesis to prepare single amino acid residue replacements at positions not previously investigated in full length β subunits; these include Arg-95, Ser-96, Thr-97, and Thr-98. In addition, Leu-92 was studied in an effort to determine whether changes immediately adjacent to the determinant loop alter receptor binding. The wild-type and mutant cDNAs for hCGβ were subcloned into a P rsv expression vector and transiently transfected into Chinese hamster ovary cells containing a stably integrated gene for bovine α. The concentrations of total expressed hCGβ in heterodimer form with the bovine α subunit were determined by radioimmunoassays. The mutant gonadotropins were assayed in vitro using a competitive binding assay with [ 125I]hCG and progesterone production, both in the transformed murine Leydig cell line, MA-10. Mutant β subunits containing the replacements Lys-92, Ser-95, Asp-96, and Tyr-97 exhibited normal α subunit binding. The resulting heterodimers with β replacements at positions 92 and 97 gave comparable potency to control gonadotropin containing hCGβ wild-type, and mutant gonadotropins containing replacements at positions 95 and 96 yielded reduced potency in the binding assay. Interestingly, the potencies of these latter two mutant gonadotropins were comparable to the wild-type gonadotropin in the steroidogenic assay, although the maximal Steroidogenesis was less in the hormone with the Arg-95 → Ser replacement. Substitution of Thr-98 with Asp and with Arg resulted in mutant β subunits with reduced subunit assembly; however, both heterodimers were active in the in vitro bioassays. Accurate potencies could not be obtained with the gonadotropin containing hCGβ(Arg-98), but the activity associated with hCGβ(Asp-98) appeared less than with wild-type gonadotropin. A comparison of these results with other data obtained on the determinant loop region leads to the following conclusions. Arg-94, Arg-95, Ser-96, Thr-98, and Asp-99 contribute to CG/lutropin receptor binding, although neither is absolutely critical since certain substitutions can be tolerated without a total loss of activity. The determinant loop region also appears to participate in some manner to α subunit binding, perhaps involving Thr-98. It remains to be determined if this region serves as a crucial determinant for β chain folding or represents a direct contact site for the α subunit. The question of whether the determinant loop region is important in receptor specificity is still open, but available data argue strongly that several residues in this disulfide loop contribute to CG/lutropin receptor binding.