Enhancement of VIPERdb, a virus structure knowledgebase

Viruses are common pathogens that cause disease, suffering and death in all ?kingdoms of life?. Three- dimensional (3D) structures of viruses provide accurate details on the organization and interactions of coat proteins (CPs) as well as the locations and environments of receptor binding sites and antigenic epitopes. This information can be used to counter viral infections and pathogenesis by developing vaccines and antiviral drugs employing the rational design approaches. We created a curated repository of non-enveloped virus (NEV) structures, namely VIPERdb (Virus Particle Explorer database; http://viperdb.scripps.edu), where all the virus capsids are oriented in a single (standard) icosahedral convention as opposed to a myriad of orientations and conventions found in the PDB. Such an organization not only enables easy generation of proper capsids, importantly facilitates rapid computational analysis of virus capsids en masse in terms of protein-protein interactions at the unique subunit-subunit interfaces as well as identification of the surface exposed regions (epitopes). The structure-derived properties and annotations (e.g., surface/interface/core residues and residue pairs at the subunit interfaces) are stored in a MySQL relational database along with the information related to virus taxonomy, genome type, capsid architecture, and natural host. Furthermore, a number of web-based utilities have been created and made available at VIPERdb that can be used to analyze virus structures and compare structure-derived properties across a group of capsids that belong to particular virus family or capsid architecture (T-number). Moreover, in addition to being an excellent research resource for virologists, VIPERdb is also being used an educational resource on virus structures. The tertiary and quaternary structures are highly conserved within each virus family (e.g., Picornaviridae or Papillomaviridae). On average, experimental structures are available for ~10% of sequenced and biochemically characterized viruses. With the advent of robust homology modeling methods/programs and greater structural conservation seen within each virus family, one can generate reasonably accurate homology models for the remaining viruses. Such models provide new structural knowledge and will be useful for molecular virologists in designing their experiments and to generate hypothesis Here, we propose to enhance the structural content and utility of VIPERdb as a knowledgebase by 1) including accurate and validated homology models, generated by restraining the interface residues in the presence of neighboring subunits, of structurally uncharacterized capsids and 2) Including the structures of helical and non-icosahedral viruses and analysis of the extent and nature of interactions at the CP-CP interfaces between different types of (icosahedral vs. helical) virus assemblies.