home - index - site map   -   if you don't see a molecular model:
http://www.biologie.uni-hamburg.de/
lehre/bza/virus/1dgi/1dgim.htm

Poliovirus - receptor interaction

Picornaviruses face two problems regarding their surface. On one hand, to infect the proper type of target cell they have to interact reliably with some receptor found specificially on that type of cells. This requires a highly specific surface pattern undisturbed by genetic fluctuation. On the other hand viruses like to change their surface topology to escape adaptive immune responses of their hosts, requiring a constant genetic drift. The tropism of different viruses goes with the constant receptor specificity, so a rhinovirus will keep to your nose and wouldn't trouble you with a neurodegenerative disease like polio, though both viruses behave rather alike once they enter a cell. The differentiation between the required docking surface to the receptor and the rather unwanted interaction surface for antibodies is done by placing the receptor interface in a groove not easyly accessible to bulky antibodies.

The polio receptor (CD155) is a membrane anchored protein with three extracellular globular domains. According to the folding pattern these domains belong to the superfamily of immunoglobulins. To visualize the interaction of polio (or other picorna) viruses with respective receptors the extracellular domains were (expressed from engineered genes) mixed with virions and subjected to cryo-electron microscopy. Due to the rather global resolution of these data (22 Å) both the receptor and viral proteins could only be modelled with their C_alpha atoms, giving no clue to the actual orientation of the amino acid's side chains. Therefore, substituting the complete atomic coordinate data gained from uncomplexed virus is to be taken with due caution.

To the left the position of the receptor within the canyon of a poliovirus protomer is shown. The receptor binds with its aminoterminal globular domain. Binding is only possible when the fatty acid component beneath the canyon is removed  . This indicates a) the fatty acid component is not very tightly bound so it can diffuse out of it's binding pocket in VP1, and b) there must subsequently be some structural rearrangement in the shape of the canyon to trigger the binding of the receptor. For b) there are no structural details known at atomic resolution, so the model shown here lacks precision. The fatty acid looks at you through an escape hole at the bottom of the canyon from which it may emerge.

Interactions between amino acids of receptor and virus are (tentatively because of poor resolution) derived from genetic experiments and model building. Mutants with amino acid replacements affecting binding constants (or the whole infection process) demonstrate either direct interactions between changed residues or conformational changes affecting the interface topology. Mutations found in the receptor map to both the interface and the adverse side on the binding domain  . Viral mutants map in the canyon to the next protomer too  .

By model building it is possible to place the receptor into the canyon by matching as many as possible amino acids according to their properties in relation to binding interactions. So the walls of the canyon are lined with hydrophobic (van der Waals forces), hydrophilic (hydrogen bridges), or charged (acidic or basic) amino acids in contacting distance to the receptor (use the mouse to have a thorough look at the canyon!). The counterparts in the receptor may form the possible bonds then.