CRYSTAL STRUCTURE OF COXSACKIEVIRUS A9. John Tate*, Michael Smyth, Timo Hyypiä** and David Stuart, Laboratory of Molecular Biophysics, Oxford, U. K. and ^**Department of Virology, University of Turku, Turku, Finland, *Present address: The Scripps Research Institute, 10666 North Torrey Pines Road, La Jolla, California, 92037, U.S.A

Coxsackievirus A9 (CAV9) has been purified and crystallized in the presence of an antiviral compound and its structure determined.

Crystals belong to space group C2 with cell dimensions a=488.5Å, b=358.5Å, c=306.5Å and ß=128.1*. There is half a virion in the asymmetric unit giving rise to 30-fold non-crystallographic symmetry. Oscillation images have been collected and processed to 4.2Å, yielding approximately 20% of possible reflections to this resolution.

The structure solution used molecular replacement with a model composed from human rhinovirus 14 and poliovirus. An electron density map was calculated using all available data, and cyclic 30-fold averaging performed. Following 50 cycles the R factor was 15.5%. An atomic model has been fitted into this electron density map using the published sequence.

The major proteins VP1-3 show the same eight-stranded ß-barrel motif as the other picornaviruses and spherical plant viruses. A striking canyon is the dominant feature of the surface, with adjacent loops similar to poliovirus. A hydrophobic pocket lies within VP1 below the canyon floor and contains electron density commensurate with the antiviral compound (WIN51711) which was necessary for capsid stabilization during the crystallization step. A discrete node of density at the exterior of the 5-fold annulus may indicate a calcium binding site, while density on the interior of the capsid surrounding the 5-fold may be due to the myristate of VP4.

It is known from the CAV9 sequence that the carboxy-terminal of VP1 is extended by approximately 15 amino acids when compared to the closely related coxsackie B viruses, and this region contains the RGD motif implicated in receptor binding. In our structure to date, the 19 terminal residues have not been modelled due to ambiguities in the electron density.

We await data to higher resolution which will elucidate some of these ambiguities.