The hepatitis B virus binds avidly to albumin polymers which in turn may mediate the initial binding of viral particles to the liver cell. However, the interaction of albumin polymers with the liver remains poorly characterized, and the possibility that hepatic binding reflects an artifact of polymerization with glutaraldehyde has not been excluded. We therefore characterized the binding of '9-labeled natural and synthetic albumin polymers to suspensions of rat hepatocytes.
Saturable binding was demonstrated for all preparations of monomeric and polymeric albumin studied. Glutaraldehyde-polymerized albumin (mean polymerization number = 15) bound much more avidly than naturally occurring albumin polymers (mostly dimers and trimers) or monomeric albumin. Competition between monomer and synthetic polymer was not observed. Reduction of free aldehyde groups on the synthetic polymer decreased nonsaturable binding without affecting saturable binding. Autoradiography confirmed binding of polyalbumin to hepatic parenchymal cells. Glutaraldehyde-polymerized ovalbumin, a protein unrelated to serum albumin, also bound hepatocytes saturably.
We conclude that hepatic binding of synthetic albumin polymers is not due to residual aldehyde groups on the polymer and is much more avid than for natural polymer. This difference may reflect the higher degree of polymerization or chemical modification of the synthetic polymer. The hepatic binding sites for synthetic polymer appear distinct from those previously deecribed for monomeric albumin and may not be specific for albumin.
In 1979, Imai et al. (1) showed binding of hepatitis B viral particles to glutaraldehyde-polymerized human albumin. Subsequent studies demonstrated binding of polymerized albumin to human liver biopsies (2) and isolated human hepahytes (3). On this basis, it has been proposed that naturally occurring host albumin polymers may play an important role in hepatitis B viral infection by acting as a bridge between the virus and the liver cell. This hypothesis is supported by the finding that the host range of the virus closely parallels the species specificity of viral albumin binding (1). In contrast, binding of