Date Published: May 5, 2017
Publisher: Public Library of Science
Author(s): Elizabeth Jaworski, Andrew Routh, Adam S. Lauring.
Defective-Interfering RNAs (DI-RNAs) have long been known to play an important role in virus replication and transmission. DI-RNAs emerge during virus passaging in both cell-culture and their hosts as a result of non-homologous RNA recombination. However, the principles of DI-RNA emergence and their subsequent evolution have remained elusive. Using a combination of long- and short-read Next-Generation Sequencing, we have characterized the formation of DI-RNAs during serial passaging of Flock House virus (FHV) in cell-culture over a period of 30 days in order to elucidate the pathways and potential mechanisms of DI-RNA emergence and evolution. For short-read RNAseq, we employed ‘ClickSeq’ due to its ability to sensitively and confidently detect RNA recombination events with nucleotide resolution. In parallel, we used the Oxford Nanopore Technologies’s (ONT) MinION to resolve full-length defective and wild-type viral genomes. Together, these accurately resolve both rare and common RNA recombination events, determine the correlation between recombination events, and quantifies the relative abundance of different DI-RNAs throughout passaging. We observe the formation of a diverse pool of defective RNAs at each stage of viral passaging. However, many of these ‘intermediate’ species, while present in early stages of passaging, do not accumulate. After approximately 9 days of passaging we observe the rapid accumulation of DI-RNAs with a correlated reduction in specific infectivity and with the Nanopore data find that DI-RNAs are characterized by multiple RNA recombination events. This suggests that intermediate DI-RNA species are not competitive and that multiple recombination events interact epistatically to confer ‘mature’ DI-RNAs with their selective advantage allowing for their rapid accumulation. Alternatively, it is possible that mature DI-RNA species are generated in a single event involving multiple RNA rearrangements. These insights have important consequences for our understanding of the mechanisms, determinants and limitations in the emergence and evolution of DI-RNAs.
RNA viruses are extremely diverse and rapidly evolving. Their RNA-dependent RNA polymerases (RdRps) readily generate single-nucleotide variants whilst lacking proof-reading capabilities. RdRps are also highly prone to RNA recombination; either through template-switching or through non-replicative end-joining. RNA recombination has been demonstrated to be responsible for the emergence of new strains or species of viruses such as rhinoviruses and dengue virus, and the formation of vaccine-derived poliovirus. Non-homologous RNA recombination is also responsible for the generation of defective RNAs[8, 9]. These are versions of the parental viral genome that can arise naturally during the course of viral passaging but have been truncated and rearranged by RNA recombination. While not encoding for functional viruses themselves, they can be amplified and co-passaged with the help of the wild-type ‘helper’ virus that provides the necessary machinery for replication, encapsidation and transmission. A defective RNA that accumulates to such an extent as to compete with or otherwise attenuate the replication of the parental virus is known as a Defective-Interfering RNA (DI-RNA).
In this manuscript, we sought to provide a thorough and comprehensive analysis of the frequency and identity of recombination events present during the serial passaging of Flock House virus in cell culture in order to elucidate the pathways and mechanism of DI-RNA emergence and evolution. We began with a homogenous inoculum derived from plasmid cDNAs of each of the FHV genomes. In the inoculum and in the early passages, we find a wide range of low-frequency recombination events corresponding to deletions and duplications that are dispersed through-out the viral genomic RNAs. We can be confident that these species do not constitute sequencing artifacts as we made our RNAseq libraries using ‘ClickSeq’ that has previously been demonstrated to reduce artifactual recombination in RNAseq data to fewer than 3 events per million reads. Further confidence in the low rates of artifactual recombination in our study is provided internally by inspecting the numbers of inter-RNA recombination events (RNA1 to RNA2 and vice versa), which are always low. Furthermore, the majority of the detected inter-RNA recombination events correspond to genomic RNA hetero- and homo-dimers, which have previously been characterized as replication intermediates .