Biochemical and structural characterization of the VPg uridylylation complex of poliovirus

The replication of poliovirus requires the formation of a variety of protein-RNA complexes during the course of its replicative cycle, an example of which is the VPg uridylylation complex. Uridylylated VPg is attached to the 5' end of all viral transcripts and serves as the primer for RNA synthesis. Uridylylation of VPg is catalyzed by the viral polymerase 3D pol , templated by a c is -acting r eplication e lement ( cre ) in the viral 2C coding region, and stimulated by up to 50-fold by the addition of the viral protease 3CD. Both the topology of the complex and the mechanism through which 3CD stimulates uridylylation is poorly understood: we investigated 3CD-stimulated VPg uridylylation through biochemical studies, structural studies and modeling. In vitro VPg uridylylation assays of mutant cres established that important characteristics of the cre stem-loop are terminal loop size (14-nt), position of the "AAACA" template sequence within the loop, and stem length. The crystal structure of 3CD revealed two protein-protein interfaces (D-C and D-D') between symmetry-related molecules: in vitro uridylylation assays of interface mutants confirmed that these interfaces are indeed relevant to VPg uridylylation. Structural modeling of the VPg uridylylation complex, using homologous structures and biochemical data, resulted in a complex consisting of a central polymerase molecule, which contains both VPg and the minimal required cre in its active site, and which is decorated by one copy of 3CD, plus two additional copies of a 3D domain bound at the previously characterized Interface I. Shape complementarity between the components is apparent. Sensitive in vitro uridylylation assays confirmed that very low levels of uridylylation activity result from mutating either Interface I, or the D-C interface, or Arg17 of VPg, or from truncating the C-terminal one-third of VPg. Remarkably, however, complexes containing any of these mutated components were incapable of being stimulated by 3CD. Since, according to our model, the three components required for stimulated uridylylation do not contact one another directly, and the bound 3C domain fails to contact the minimal cre, we favor an argument previously advanced in the literature that the catalytic activity of 3D pol is highly sensitive to allosteric control.