Probing the sequence constraints for the stable incorporation of chimeric NSP1 segments into infectious rotaviruses

Rotaviruses are a global cause of acute paediatric gastroenteritis. They present a genome of 11 dsRNA segments that have been successfully incorporated into a plasmid-based reverse-genetics system that enables the generation of reassortant and chimeric viruses as a possible foundation for multi-virus vaccines. Studies have shown that these chimeric segments are pruned across successive passages, highlighting a need to better define the constraints that govern segment recognition and stable incorporation into mature viruses. This work leveraged the reverse-genetics system to investigate these constraints by introducing a series of modifications to a chimeric NSP1 sequence with an internally cloned GFP insert (NSP1gfp). Across ten serial passages, a pruned NSP1gfp segment emerged and eventually outcompeted the original construct, resulting in a loss of GFP expression. The removal of the entire NSP1 ORF sequence 3' of the insert resulted in prolonged GFP expression, but closer examination of rotavirus RNA suggested that these modifications were merely delaying replacement from pruned segments, rather than avoiding their emergence. A series of systematic truncations to the remaining NSP1 ORF sequence 5' of the GFP insert showed that as little as 27 nt of the original coding sequence is needed for rescue of infectious chimeric rotavirus. Despite a near-total replacement of NSP1 with GFP, pruned segments emerged and were preferentially incorporated into later passages, which was correlated with increased infectivity at the expense of GFP expression. These results suggest that factors beyond excess segment size drive pruning of chimeric sequences. Closer sequence analysis of pruned segments revealed that pruning occurred at high GC regions of the insert, suggesting an inherent incompatibility with the genome composition of rotavirus segments. These findings shed light on the requirements for rotavirus segment recognition and stability, as well as provide important considerations for the usage of this model as a potential vaccine platform.