Collection and motif-based prediction of phosphorylation sites in human viruses

Although various databases have been established that are designed to compile an ever-growing list of protein phosphorylation sites in plants and animals, no such repository exists for viruses. Here, we developed the viral posttranslational modification (virPTM) database, which contains a comprehensive list of 329 accurately localized phosphorylation sites in proteins from 52 human viruses published between 1986 and the present. Additionally, to aid in the detection of new viral phosphorylation sites, we used the scan-x tool to make thousands of high-specificity serine, threonine, and tyrosine phosphorylation predictions in 229 viruses that replicate in human cells. By cross-validating our prediction results with the literature-based entries in the virPTM database, we highlight the effectiveness of the scan-x tool with viral data and extrapolate the existence of at least 4000 as yet unidentified phosphorylation sites on hundreds of viral proteins. Together, these results imply a substantial role for human kinases in mediating viral protein functions and suggest, more generally, that viral primary structure may provide important clues to aid in the rational design of therapeutic agents.