Search

The mechanism of intracellular retention for the large surface protein (L) of duck hepatitis B virus (DHBV) was analyzed by examination of the transmembrane topologies and secretory properties of a collection of DHBV L mutants and compared with that of human hepatitis B virus (HBV) L. Our results demonstrate that, in contrast to its HBV counterpart, intracellular retention of DHBV L does not depend on the cytosolic disposition of its preS domain. L mutants with either cytosolic or lumenal preS were mostly retained in the absence of the small surface protein (S), whereas coexpression with S resulted in efficient secretion of both topological forms. Coexpression of the wild-type DHBV L with S resulted in efficient incorporation of L into secreted S + L particles, whereas HBV L was partially excluded from secreted particles under the same conditions. We propose that HBV provides L retention even in the presence of an excess of S, by exclusion of molecules with cytosolic preS domains from secreted particles at the stage of their assembly. DHBV lacks such a retention mechanism due to the absence of topological selection in particulate assembly.

Analysis of deletion and/or site-specific mutants of the hepatitis B virus (HBV) env gene, expressed in human cells, provided clues about the mechanism that retains the L protein, the largest gene product, in a pre-Golgi compartment. Differences in secretability of the analyzed variants suggest that the N-terminal myristic acid and an internal sequence within the PreS1 region function as independent retention signals. N-terminal myristic acid alone neither prevented PreS1 + 2 N-linked glycosylation, which signals cotranslational translocation of the domain, nor strongly inhibited lumenal budding. Thus, myristic acid by itself acts by arresting secretion of lumenal, soluble Env particles. By contrast, the internal retention determinant, mapping in the C-terminal portion of PreS1, also prevented budding. In addition, the presence of this PreS1 segment correlated with the depression of PreS1 + 2 glycosylation. This suggests a connection between L retention and the recently described inhibition of PreS1 + 2 cotranslational translocation. A model can be proposed, according to which HBV surface proteins need to cotranslationally translocate their N-terminal moieties in order to assume a transmembrane topology suitable for particulate assembly and secretion. L protein, whose PreS1 + 2 domain undergoes translocation only posttranslationally, would fail to complete the secretion process. To support this model, we show that forced cotranslational translocation of the PreS1 + 2 domain (by attachment of an N-terminal processed signal sequence) results in secretion of L protein.