RESEARCH ARTICLE Identification of virus-encoded microRNAs in divergent Papillomaviruses Rachel Chirayil1³, Rodney P. Kincaid1³, Christine Dahlke2³, Chad V. Kuny1³, Nicole DaÈlken2, Michael Spohn2, Becki Lawson3, Adam Grundhoff2*, Christopher S. Sullivan1* 1 Center for Systems and Synthetic Biology, Center for Infectious Disease and Dept. Molecular Biosciences, The University of Texas at Austin, Austin, TX, United States of America, 2 Heinrich Pette Institute, Leibniz Institute for Experimental Virology, Hamburg, Germany, 3 Institute of Zoology, Zoological Society of London, London, United Kingdom a1111111111 a1111111111 ³ These authors share first authorship on this work. a1111111111 * [email protected] (AG); [email protected] (CSS) a1111111111 a1111111111 Abstract MicroRNAs (miRNAs) are small RNAs that regulate diverse biological processes including multiple aspects of the host-pathogen interface. Consequently, miRNAs are commonly OPEN ACCESS encoded by viruses that undergo long-term persistent infection. Papillomaviruses (PVs) are Citation: Chirayil R, Kincaid RP, Dahlke C, Kuny capable of undergoing persistent infection, but as yet, no widely-accepted PV-encoded miR- CV, DaÈlken N, Spohn M, et al. (2018) Identification of virus-encoded microRNAs in divergent NAs have been described. The incomplete understanding of PV-encoded miRNAs is due in Papillomaviruses. PLoS Pathog 14(7): e1007156. part to lack of tractable laboratory models for most PV types. To overcome this, we have https://doi.org/10.1371/journal.ppat.1007156 developed miRNA Discovery by forced Genome Expression (miDGE), a new wet bench Editor: Paul Francis Lambert, University of approach to miRNA identification that screens numerous pathogen genomes in parallel. Wisconsin Madison School of Medicine and Public Using miDGE, we screened over 73 different PV genomes for the ability to code for miRNAs. Health, UNITED STATES Our results show that most PVs are unlikely to code for miRNAs and we conclusively dem- Received: April 4, 2018 onstrate a lack of PV miRNA expression in cancers associated with infections of several Accepted: June 15, 2018 high risk HPVs. However, we identified five different high-confidence or highly probable Published: July 26, 2018 miRNAs encoded by four different PVs (Human PVs 17, 37, 41 and a Fringilla coelebs PV (FcPV1)). Extensive in vitro assays confirm the validity of these miRNAs in cell culture and Copyright: © 2018 Chirayil et al. This is an open access article distributed under the terms of the two FcPV1 miRNAs are further confirmed to be expressed in vivo in a natural host. We Creative Commons Attribution License, which show that miRNAs from two PVs (HPV41 & FcPV1) are able to regulate viral transcripts cor- permits unrestricted use, distribution, and responding to the early region of the PV genome. Combined, these findings identify the first reproduction in any medium, provided the original canonical PV miRNAs and support that miRNAs of either host or viral origin are important author and source are credited. regulators of the PV life cycle. Data Availability Statement: The miDGE sequencing data are available via the European Nucleotide Archive (ENA, https://www.ebi.ac.uk/ ena) under accession number PRJEB25054. The sequencing data generated from chaffinch samples Author summary are available via the NCBI Sequence Read Archive (SRA, https://www.ncbi.nlm.nih.gov/sra) under the Papillomaviruses (PVs) are causative agents of cancer. Currently, there is an incomplete accession number SRP133175. understanding as to why only some infections lead to cancer. Developing a better compar- Funding: This work was supported by a Burroughs ative evolutionary understanding of PV gene products and their regulation is key to com- Wellcome Investigators in Pathogenesis Award and prehending the life cycle of these pathogens. An emerging concept of viral gene regulation a grant from the Cancer Prevention and Research PLOS Pathogens | https://doi.org/10.1371/journal.ppat.1007156 July 26, 2018 1 / 31 Large scale parallel identification of pathogen miRNAs Institute of Texas [RP140842] to CSS. AG is supported by the Free and Hanseatic City of is that many persistent viruses will utilize small regulatory RNAs called miRNAs to opti- Hamburg and the Federal Ministry of Health mize host and viral gene expression. Yet despite obvious interest, there have been no cred- (BMG). The funders had no role in study design, ible reports identifying canonical PV-encoded miRNAs. Here we develop new broadly data collection and analysis, decision to publish, or preparation of the manuscript. applicable methodology to identify miRNAs from organisms lacking a laboratory culture system. We identify the first examples of bona fide canonical PV miRNAs and provide Competing interests: The authors have declared evidence supporting both host and viral miRNA-mediated regulation as relevant to con- that no competing interests exist. trol of PV gene expression. These findings resolve the issue of PV miRNAs and further the notion of miRNA importance to persistent virus infection. Introduction Papillomaviruses (PVs) comprise a large family of circular double-stranded DNA viruses. Numerous PV genomes have been described including over 200 human PV (HPV) types. A minority of these are known as carcinogenic agents [1±3], however only a small fraction of hosts infected with these high risk types will go on to develop high grade lesions. It remains incompletely understood what factors dictate whether or not HPV infection will develop into malignant cancer [1,3]. Further, it is unclear why HPVs that share a high level of sequence sim- ilarity can have stark differences in tropism and infect different regions of the body. Develop- ing a better understanding of PV gene products and their regulation throughout diverse PV lineages provides an evolutionary foundation for deciphering the mechanisms resulting in dif- ferential outcomes of infection. MicroRNAs (miRNAs) are small regulatory RNAs that are an emerging class of viral gene products found in select virus families [4±7]. miRNAs are approximately 22 nucleotides long and function by docking to specific target mRNAs to repress translation [8,9]. miRNAs derive from primary transcripts containing a hairpin structure that is processed by a series of endo- nucleases (Drosha in the nucleus, Dicer in the cytosol) generating the final effector RNA [10± 15]. The miRNA then enters a multi-protein complex called the RNA Induced Silencing Com- plex (RISC) where it scans mRNAs and docks to regions of partial sequence complementarity [8]. The so-called seed region of miRNAs, approximately 6 or more nucleotides towards the 5' end, typically binds transcripts with perfect complementarity and this feature can be used to help identify mRNA targets [8,16,17]. miRNAs of a particular sequence typically function by subtly regulating numerous (10 or more) mRNA transcripts involved in a particular biological outcome. Although individual miRNA regulation of any single transcript can be subtle, the sum of regulation by multiple miRNA-bound-RISC complexes (miRISC) can add up to signifi- cant biological activity [8]. miRNAs of host or viral origin have been implicated in various pro- cesses relevant to virus infection including the control of the immune response, cell death, transformation and virus gene expression [18±29]. Over 300 viral encoded miRNAs have been described, all from viruses able to undergo long term persistent infection [4,5,7,30]. Most of these viruses have DNA genomes including the herpes, polyoma, and anello virus families [5]. However, some retroviruses including the delta retrovirus bovine leukemia virus (BLV) and foamy retroviruses also encode miRNAs [21,31± 33]. One likely role of viral miRNAs is to foster long-term interactions within the host [30]. In this regard, at least some members of the PV family would be expected to encode miRNAs. However, to date, no credible examples of PV canonical miRNAs have been described. Two studies examining fully infectious experimental systems of the high-risk HPV18 & 31 types report that these viruses do not encode miRNAs [34,35]. Further, at least two studies examin- ing transformed cell lines report that HPV16 does not encode miRNAs [36,37]. There have been reports in transformed cells of small RNAs from high risk PVs such as HPVs 16 & 18, PLOS Pathogens | https://doi.org/10.1371/journal.ppat.1007156 July 26, 2018 2 / 31 Large scale parallel identification of pathogen miRNAs however these studies did not demonstrate a connection of PV-derived small RNAs to the miRNA biogenesis (Dicer/Drosha) or effector (RISC) machinery, nor did they confirm a bio- logically meaningful abundance [38±40]. Consequently, these RNAs are not widely accepted as miRNAs and the signal detected likely represents degradation fragments derived from the turnover of longer transcripts. In contrast, it is well documented that PV infection and individ- ual PV gene products can alter the host miRNA repertoire, likely contributing to the biology of cancer [25,35,41±44]. Furthermore, at least one PV, HPV31, utilizes a host miRNA to directly regulate early viral gene expression [25]. Thus, what emerges is that although host miRNAs are involved in the PV life cycle and pathogenesis, of the few PV types that have been examined, no canonical PV miRNAs are yet established. One barrier to discovery of PV miRNAs is the dearth of facile fully-infectious laboratory systems. There are experimental systems established for a few PV types (approximately < 5) [45], but technological barriers have limited any comprehensive large-scale study of viral miR- NAs in the majority of PV types. Here we describe a new wet bench approach for the discovery of miRNAs that assays numerous viruses in parallel for the ability to express miRNAs. We identify bona fide PV miRNAs encoded by divergent PVs, and demonstrate these miRNAs depend on canonical miRNA biogenesis effector machinery. Our analysis also rules out canon- ical papillomaviral miRNAs in cancers associated with high-risk PV (HPVs 16, 18, 31, 45, and 58) infection. These results provide further evidence for the relevance of miRNA-mediated regulation of PV transcripts.