The human δ opioid receptor (hδOR) is one of the three opioid receptors that mediate signals of endogenous opioid peptides as well as exogenous opiate compounds. The exogenous drugs are highly efficient analgesics but they also induce tolerance and addiction. Our previous studies on hδOR biogenesis have shown that the receptor matures slowly and inefficiently, so that only 40% of synthesized receptors are able to gain stable native conformation and reach the cell surface, while the rest are targeted for degradation (1, 2). Importantly, the receptor cDNA used in these studies represents the less common polymorphic variant of the receptor, carrying 80T→G substitution that leads to exchange of Phe to Cys at position 27. The frequency of the 80G allele in Caucasians is about 9% (3). To compare the biochemical characteristics of the two receptor variants from early folding steps to the functional receptor, the two receptor constructs were expressed as Myc/Flag epitope-tagged fusion proteins in stably transfected HEK293 cells under tetracycline induction. Pharmacological characterization revealed no significant difference between the two variants in either expression levels or binding affinities for various opioid ligands. Similarly, both variants induced agonist-mediated downstream signaling with similar efficiency in GTP binding assays. In contrast, clear differences in the relative abundance of receptor mature and precursor forms were detected by Western blot analysis of immunoprecipitated receptors. Namely, the receptor mature to precursor ratio was clearly higher for the Phe27 variant, suggesting that the Cys27 variant accumulates in the endoplasmic reticulum (ER). Metabolic pulse-chase experiments showed that both receptors were converted to the mature form with similar kinetics and efficiency, but increased turn-over of the Phe27 precursor became clearly evident. This indicates effective clearance of the folding incompetent Phe27 variant from the ER. Prolonged ER retention of the Cys27 variant was not calnexin dependent, as the turn-over of calnexin-bound precursors was significantly faster than that of the total precursor pool in co-immunoprecipitation experiments. On the other hand, one of the two N-glycosylation sites (Asn33) of the Cys27 variant was found to be inefficiently occupied. Given the close proximity of Cys27 to Asn33, it can be hypothesized that the cysteine residue mediates interaction with oxidoreductases within the protein disulfide isomerase family (4). This might prevent efficient N-glycosylation of Asn33 of the hδOR Cys27 variant and also hinder proper folding and/or efficient retrotranslocation of incompletely folded receptors, subsequently causing accumulation of receptor precursors in the ER.