Because of its critical role in HDL formation, significant efforts have been devoted to studying apolipoprotein A-I (APOA1) structural transitions in response to lipid binding. To assess the requirements for the conformational freedom of its termini during HDL particle formation, we generated three dimeric APOA1 molecules with their termini covalently joined in different combinations. The dimeric (d)-APOA1C-N mutant coupled the C-terminus of one APOA1 molecule to the N-terminus of a second with a short alanine linker, whereas the d-APOA1C-C and d-APOA1N-N mutants coupled the C-termini and the N-termini of two APOA1 molecules, respectively, using introduced cysteine residues to form disulfide linkages. We then tested the ability of these constructs to generate reconstituted HDL by detergent-assisted and spontaneous phospholipid microsolubilization methods. Using cholate dialysis, we demonstrate WT and all APOA1 mutants generated reconstituted HDL particles of similar sizes, morphologies, compositions, and abilities to activate lecithin:cholesterol acyltransferase. UnlikeWT, however, the mutants were incapable of spontaneously solubilizing short chain phospholipids into discoidal particles. We found lipid-free d- APOA1C-N and d-APOA1N-N retained most of WT APOA1's ability to promote cholesterol efflux via the ATP binding cassette transporter A1, whereas d- APOA1C-C exhibited impaired cholesterol efflux. Our data support the double belt model for a lipidbound APOA1 structure in nascent HDL particles and refute other postulated arrangements like the "double super helix." Furthermore, we conclude the conformational freedom of both the N- and Ctermini of APOA1 is important in spontaneous microsolubilization of bulk phospholipid but is not critical for ABCA1-mediated cholesterol efflux.