Hepatitis B surface antigen assembles in a post-ER, pre-Golgi compartment

Previous studies have provided much information on the biogenesis and the secretion of HBsAg. Eble et al. (1986) demonstrated that in a cell-free translation system HBsAg is initially synthesized as a transmembrane polypeptide that spans the lipid bilayer at least twice. Cultured cells which are transfected with the coding region for HBsAg secrete 22-nm particles that are morphologically and antigenically indistinguishable from those in serum (Crowley et al., 1983; Dubois et al., 1980; Liu et al., 1982; Moriarty et al., 1982; Patzer et al., 1984 Patzer et al., , 1986 Simon et al., 1988) . Using such a system, Simon et al. (1988) showed that HBsAg becomes carbonate extractable with a very similar half-time to that of particle secretion. This indicated that the rate-limiting step in the export process is the budding of the lipoprotein particles from the membrane. All of the oligosaccharides on the secreted HBsAg lipoprotein particles are of the N-linked complex type. However Patzer et al. (1984) showed that all of the glycans on cellular HBsAg were of the high mannose type. Hence, the rate limiting step in secretion must lie prior to the medial Golgi. EM of HBsAg in infected or in stably transfected cells reveals particles within the lumen of smooth intracellular membranes (Gerber et al., 1974; Shibayama et al., 1984; Patzer et al., 1986) . These regions have been interpreted as elements of the ER and contributed to a consensus view that particle assembly in the membrane and budding which is the rate limiting step in lipoprotein secretion, occurs in the ER (reviewed in Ganem and Varmus, 1987) .