Background: Understanding how an organism replicates and assembles a multi-segmented genome with fidelity previously measured at 100% presents a model system for exploring questions involving genome assortment and RNA/protein interactions in general. The virus family Reoviridae, containing nine genera and more than 200 members, are unique in that they possess a segmented double-stranded (ds) RNA genome. Using reovirus as a model member of this family, we have developed the only functional reverse genetics system for a member of this family with ten or more genome segments. Using this system, we have previously identified the flanking 5' sequences required by an engineered s2 ssRNA for efficient incorporation into the genome of reovirus. The minimum 5' sequence retains 96 nucleotides and contains a predicted sequence/structure element. Within these 96 nucleotides, we have identified three nucleotides A-U-U at positions 79-81 that are essential for the incorporation of in vitro generated ssRNAs into new reovirus progeny viral particles. The work presented here builds on these findings and presents the results of an analysis of the required 3' flanking sequences of the s2 ssRNA.

Results: The minimum 3' sequence we localized retains 98 nucleotides of the wild type s2 ssRNA. These sequences do not interact with the 5' sequences and modifications of the 5' sequences does not result in a change in the sequences required at the 3' end of the engineered s2 ssRNA. Within the 3' sequence we discovered three regions that when mutated prevent the ssRNA from being replicated to dsRNA and subsequently incorporated into progeny virions. Using a series of substitutions we were able to obtain additional information about the sequences in these regions. We demonstrate that the individual nucleotides from, 98 to 84, 68 to 59, and 28 to 1, are required in addition to the total length of 98 nucleotides to direct an engineered reovirus ssRNA to be replicated to dsRNA and incorporated into a progeny virion. Extensive analysis using a number of RNA structure-predication software programs revealed three possible structures predicted to occur in all 10 reovirus ssRNAs but not predicted to contain conserved individual nucleotides that we could probe further by using individual nucleotide substitutions. The presence of a conserved structure would permit all ten ssRNAs to be identified and selected as a set, while unique nucleotides within the structure would direct the set to contain 10 unique members.

Conclusion: This study completes the characterization and mapping of the 5' and 3' sequences required for an engineered reovirus s2 ssRNA to be incorporated into an infectious progeny virus and establishes a firm foundation for additional investigations into the assortment and encapsidation mechanism of all 10 ssRNAs into the dsRNA genome of reovirus. As researchers build on this work and apply this system to additional reovirus genes and additional dsRNA viruses, a complete model for genome assortment and replication for these viruses will emerge.