Influenza virus polymerases: determination of the cap binding site and the crucial role of CA endonuclease cleavage site in the cap snatching mechanism for the initiation of viral messenger RNA synthesis

It has been established that influenza virus synthesizes its viral mRNAs through a "cap-snatching" mechanism; specifically, the viral mRNA synthesis is initiated with capped-RNA primers generated by the endonuclease intrinsic to the viral polymerase from host cellular mRNAs. To exert the cap-binding and endonuclease activity, the polymerase must be activated by the conserved terminal sequences in the viral genomic RNAs (vRNAs). By UV cross-linking and protein microsequencing, a short peptide (544-556 in the PB2 protein) is found to directly bind to the capped-RNAs. To identify the specific amino acids involved in direct cap recognition, five aromatic amino acids around this region were mutated individually to alanine. Four of the five mutated PB2 proteins are unable to form functional polymerase complexes with other subunits. The fifth mutant, W552A, which forms functional polymerase, has attenuated cap-binding and endonuclease activities. Therefore, some of the other four aromatic amino acids may be involved in direct cap recognition. Using different capped RNA as substrates for endonuclease, I compared the effect of the dinucleotides at the cleavage site on the endonuclease activity and transcription initiation in vitro. I demonstrate that only the substrate containing a 3' CA terminus is effectively used as a primer for viral mRNA synthesis. This explains the previous observation that the vast majority of cellular capped primers used during virus infection contain CA at their 3' termini. Using this system, I found, in contrast to previous reports, the 5' terminal sequence of viral RNA (vRNA) alone is sufficient for the activation of the endonuclease, and the 3' terminal sequence of vRNA functions solely as template for transcription initiation. I further identified the sequence/structure in the 5' and 3' termini of vRNA that are both necessary and sufficient for their function as endonuclease activator and transcription template, respectively. The 3'vRNA binds the polymerase by both RNA-polymerase and RNA-5’vRNA interactions. My results also indicate that the viral mRNA elongation may be bi-phasic. In early elongation, within 18 nucleotides from the transcription initiation site, the viral polymerase and the conserved vRNA terminal sequences are sufficient. In the second phase of elongation, additional factors/conditions are required.