The human glycoprotein hormones, hCG, TSH, LH, and FSH, are composed of a common α-subunit assembled to a hormone-specific β-subunit. The subunits combine noncovalently early in the secretory pathway and exist as heterodimers but not as multimers. LH/FSH are synthesized in the pituitary gonadotrophs, and several of the α-subunit sequences required for association with either the LHβ or FSHβ subunits are different. Thus, it is intriguing that no ternary complexes are observed for LH and FSH in vivo (e.g. two different β-assembled to a single α-subunit). To examine whether the α-subunit can interact with more than one β-subunit, and to study the conformational relationships between the ligand and the receptor, we constructed a vector encoding two tandemly arranged β-subunits fused to a single α-subunit gene (FSHβ-CGβ-α). This approach permitted structure-function analyses of α/β domain complexes without the possibility of subunit dissociation. We reported previously that the CGβ or FSHβ subunit gene can be genetically fused to the α-gene and the resulting single chains (CGβα and FSHβα, respectively) were biologically active. Here we demonstrate that a triple-domain single chain bearing the configuration FSHβ-CGβ-α is efficiently secreted from transfected Chinese hamster ovary (CHO) cells and exhibits high-affinity receptor binding to both FSH and LH/hCG receptors, comparable to the native heterodimers. These results indicate that the α-subunit can interact with each β-subunit in the same complex and that an α-domain fused to a β-domain can still interact with an additional β-subunit. The data also demonstrate the remarkable flexibility of the receptor to accommodate the increased bulkiness of the triple-domain ligand. In addition, the formation of intrachain FSH- and CG-like complexes observed in a triple-domain single chain suggests that the α-subunit can resonate, i.e. shuttle between α-β heterodimeric intermediates during the early stages of synthesis and accumulation in the endoplasmic reticulum. Such model compounds could be useful as substrates to generate a new class of analogs in which the ratio of the LH/FSH activity is varied. This could aid in the design of analogs that could be used to mimic the in vivo hormonal profiles.