Butyrylcholinesterase (BChE) in human serum consists predominantly of tetramers. Recombinant BChE, however, expressed in Chinese hamster ovary (CHO) cells, consists of approx. 55% dimers, 10-30% tetramers and 15-40% monomers. To determine the origin of the monomer species we added the FLAG epitope (epitope tag, amino acid sequence DYKDDDDK) to the C-terminus of the enzyme, and expressed BChE-FLAG in CHO cells. We found that secreted, active monomers had lost their FLAG epitope, suggesting that the monomers were made by proteolysis of dimers or tetramers at the C-terminus. To estimate the number of amino acids that could be deleted from the C-terminus without losing BChE activity, we expressed deletion mutants. We found that deletion of up to 50 amino acids from the C-terminus yielded active monomers, but that deletion of 51 amino acids destroyed BChE activity and caused the inactive protein to remain within the cell. Deletion of eight or more amino acids from the N-terminus also resulted in inactive protein that remained inside the cell. Monomeric BChE had wild-type Km and kcat values (8 microM and 24000 min-1 for butyrylthiocholine) and showed substrate activation. The Cys-571-->Ala mutant, though incapable of forming the interchain disulphide bond, had nearly the same amount of tetrameric BChE as recombinant wild-type BChE. These results support the conclusion that the tetramerization domain of BChE is at the C-terminus, within the terminal 50 amino acids, and that the interchain disulphide bond is not essential for tetramerization. Molecular modelling suggested that the tetramerization domain was a four-helix bundle, stabilized by interactions of seven conserved aromatic amino acids.