Chorionic gonadotropin (hCG) is a heterodimeric placental glycoprotein hormone essential for human reproduction. Twenty hCG β-subunit residues, termed the seatbelt, are wrapped around α-subunit loop 2 (α2) and their positions “latched” by a disulfide formed by cysteines at the end of the seatbelt (Cys 110) and in the β-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 β-subunit. If the seatbelt remained latched during the process of subunit combination, formation of the heterodimer would require α2 and its attached oligosaccharide to be threaded through a small β-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 β-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 β-subunit hole. We propose a mechanism for hCG assembly in vitro that depends on movements of α2 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.