Direct RNA sequencing of respiratory syncytial virus infected human cells generates a detailed overview of RSV polycistronic mRNA and transcript abundance

Abstract

To characterize species of viral mRNA transcripts generated during respiratory syncytial virus (RSV) infection, human fibroblast-like MRC-5 lung cells were infected with subgroup A RSV for 6, 16 and 24 hours. In addition, we characterised the viral transcriptome in infected Calu-3 lung epithelial cells at 48 hours post infection. Total RNA was harvested and polyadenylated mRNA was enriched and sequenced by direct RNA sequencing using an Oxford nanopore device. This platform yielded over 450,000 direct mRNA transcript reads which were mapped to the viral genome and analysed to determine the relative mRNA levels of viral genes using our in-house ORF-centric pipeline. We examined the frequency of polycistronic readthrough mRNAs were generated and assessed the length of the polyadenylated tails for each group of transcripts. We show a general but non-linear decline in gene transcript abundance across the viral genome, as predicted by the model of RSV gene transcription. However, the decline in transcript abundance is not uniform. The polyadenylate tails generated by the viral polymerase are similar in length to those generated by the host polyadenylation machinery and broadly declined in length for most transcripts as the infection progressed. Finally, we observed that the steady state abundance of transcripts with very short polyadenylate tails less than 20 nucleotides is less for N, SH and G transcripts in both cell lines compared to NS1, NS2, P, M, F and M2 which may reflect differences in mRNA stability and/or translation rates within and between the cell lines.