Variable and conserved features of copy-back viral genome populations generated de novo during Sendai virus infection

Copy-back viral genomes (cbVGs) are generated during the replication of negative-sense RNA viruses when the polymerase drops off from the genome and reattaches to the nascent strand. cbVGs have strong immunostimulatory properties and impact infection outcomes. Despite their importance, the composition and mechanisms of de novo cbVG generation and accumulation remain unclear due to challenges in obtaining cbVG-free virus stocks (clean stocks). Here, we obtained several clean stocks by independently rescuing recombinant Sendai virus (SeV) six times and verified their cleanliness through PCR, RNA sequencing, and absence of immunostimulatory activity. High multiplicity-of-infection (MOI) passaging of clean stocks produced six high-MOI passaged stocks, each with distinct cbVG populations. Among them, polymerase drop-off (break) positions occurred throughout the genome, while polymerase reattachment (rejoin) positions preferentially occurred near the trailer end. Few common breaks were observed between stocks, while there was a hot rejoin region near the trailer end. In each stock, a few cbVG species dominated and remained stable across passages, all conforming to the "rule of six," regardless of length. Low-abundance cbVGs were variable across passages, indicating the continuous generation of new cbVGs, despite the stabilization of a subset of species. Intriguingly, cbVG species that originated from polymerase drop-off at or close to nucleotide 1 were present in all stocks, suggesting that cbVG species originating at the 3' end of the genome are conserved products of SeV replication.IMPORTANCECopy-back viral genomes (cbVGs) are generated during infection when the polymerase drops off from the template and reattaches to the nascent strand, and they are major drivers of antiviral immunity. However, natural isolates contain pre-existing cbVGs, limiting our ability to understand how cbVGs are generated and accumulated. Here, we used cbVG clean stocks obtained from cDNA to address these questions. Comparative analysis of six parallel cbVG-high stocks showed that polymerase drop-off sites are broadly distributed across the genome, with a recurrent origin near nucleotide 1, while polymerase reattaches near the trailer end. Longitudinal analysis revealed that dominant cbVG species remain stable across passages of the same stock, whereas some cbVGs are dynamic. cbVG accumulation was independent of cbVG length but strictly followed the rule of six. These findings reveal conserved and variable features of cbVG generation from clean stocks and shed light on how cbVGs accumulate during infection.