Characterization of a carboxyl-terminal deletion mutant of the human choriogonadotropin beta subunit

The $\beta$ subunit of human choriogonadotropin contains an extension of 24 amino acids, glycosylated at 4 serines, that is not present in the structurally and functionally similar $\beta$ subunit of human lutropin. In addition to this carboxyl-terminal extension in human choriogonadotropin $\beta$, the two gonadotropin $\beta$ subunits differ in the number of N-asparagine carbohydrate moieties and in 20% of the amino acid residues. These subunits associate with a common $\alpha$ subunit, and the resulting hormones, acting through the same receptor, exhibit different potencies in certain cell types. In order to ascertain whether the O-linked carboxyl-terminal extension of human choriogonadotropin $\beta$ contributes to this difference in activity, a premature termination codon was introduced at position 122 to produce a mutant $\beta$ subunit lacking the carboxyl-terminal extension. The cDNA was transiently expressed under the constitutive transcriptional control of an RSV promoter in a Chinese hamster ovarian cell line with a stably integrated and expressed bovine $\alpha$ gene, and the resulting heterologous bovine $\alpha$-des(122-145) human $\beta$ gonadotropin was compared with a heterologous bovine $\alpha$-human $\beta$ gonadotropin prepared using the cDNA from the human $\beta$ gene. Results demonstrate that the carboxyl-terminal region of human choriogonadotropin $\beta$ is not required for proper folding of the $\beta$ subunit, for association with the complementary $\alpha$ subunit, or for secretion of heterodimer. It also makes no discernable contribution to acute steroidogenesis in rat or mouse Leydig cells.Transformed Leydig cells (MA-10), however, respond similarly to both standard gonadotropins as well as to both expressed gonadotropins in their steroidogenic potencies. The magnitude of the cyclic 3$\sp\prime$,5$\sp\prime$-adenosine monophosphate response in MA-10 cells is greater for human lutropin than it is for human choriogonadotropin, but is equivalent for both of the expressed gonadotropins. Thus, it appears the potency differences that can be demonstrated with the two human gonadotropins in rodent Leydig cells arise from the 20% difference in amino acid sequence of the core region, or from the glycosylation differences. It is likely that the carboxyl-terminal region of human choriogonadotropin $\beta$ may contribute to the greater circulatory half-life of human choriogonadotropin compared to human lutropin. (Abstract shortened with permission of author.