Comparative analysis of flavivirus sfRNA dynamics and secondary structure

The accumulation of subgenomic flavivirus RNAs (sfRNAs) modulates viral fitness and pathogenicity in culture and in vivo. These noncoding RNAs are produced by incomplete digestion of the flavivirus genome by the cellular 5-3 exoribonuclease (XRN1). Diverse flaviviruses have conserved RNA structural elements (RSEs) that map to their 3-untranslated region (3-UTR): Xrn-resistant RNA structures, dumbbell structures, and a 3-stem loop (3SL). Despite the importance of the 3-UTR RSEs for flavivirus replication, the structural dynamics of sfRNA during flavivirus infection are understudied. Here, we use digital droplet PCR to quantify sfRNA levels during infection for a panel of mosquito-borne flaviviruses (MbFV) including dengue virus serotypes 1 (DENV1), 2 (DENV2), and 4 (DENV4), and Zika virus (ZIKV). We then used SHAPE-MaP on XRN1-digested, in vitro-transcribed sfRNAs from each virus to determine their secondary structures compared to the corresponding sfRNAs obtained from flavivirus-infected A549 cells. Results seen in-cell and in vitro were largely similar; however, motifs within the dumbbell, the small hairpin (sHP) directly upstream of the 3-SL, and 3-SL regions showed significant differences in the extent of nucleotide reactivity. These differences were consistent among the four flaviviruses examined and may indicate regions of sfRNA that are shielded by interaction with proteins or other nucleic acids during infection. However, strong protection indicative of sustained interaction was not apparent. Our findings suggest that sfRNA interactions with viral and host factors within the cell are few, occur via base-paired regions, or are highly transient. ImportanceFlaviviruses are highly prevalent human pathogens. The flavivirus genome contains RNA structural elements (RSEs), including those encoded in the 3-UTR, that are necessary for viral replication. Subgenomic flavivirus RNAs (sfRNAs) are produced by incomplete digestion of flavivirus genomic RNA due to the cellular exoribonuclease XRN1 encountering 3-UTR RSEs that promote its stalling and disassociation. Viruses unable to produce sfRNAs are highly attenuated, underlining their biological importance. sfRNA secondary structure has been investigated previously but little information is available on sfRNA secondary structure dynamics in infected cells. By comparing SHAPE-MaP reactivities in vitro and in cells, we determined that previously inferred structures are likely maintained within infected cells. We also identified differences in the extent of SHAPE reactivity between in vitro and in-cell environments that were common to multiple mosquito-borne flaviviruses. These differences suggest that sfRNAs may engage in transient interactions within the cell that may be important for their function.