Divergent microRNA Targetomes of Closely - related Circulating 4 Strains of a Polyomavirus 5 6 7

23 24 Hundreds of virus-encoded microRNAs (miRNAs) have been uncovered, but an 25 in depth functional understanding is lacking for most. A major challenge for 26 the field is separating those miRNA targets that are biologically relevant from 27 those that are not advantageous to the virus. Here we show that miRNAs from 28 related variants of SV40 polyomavirus have differing host target repertoires 29 (targetomes) while completely preserving their direct autoregulatory activity 30 on virus-encoded early gene products. These results underscore the 31 importance of miRNA-mediated viral gene autoregulation in some 32 polyomavirus lifecycles. More broadly, these findings imply that some host 33 targets of viral-encoded miRNAs are likely to be of little selective advantage to 34 the virus and our approach provides a strategy for prioritizing relevant 35 targets. 36 37 38 39 40 41 42 43 44 45 on A uust 0, 2017 by gest http/jvi.asm .rg/ D ow nladed fom INTRODUCTION 46 47 miRNAs are a class of eukaryotic small RNA molecules that play a regulatory role in 48 several biological processes relevant to virus infection including the immune 49 response, apoptosis and tumorigenesis (1). Virus-encoded miRNAs identified from 50 several different families, including the herpes, retro, and polyoma viruses, have 51 generated much interest as potential effectors of pathogenesis. Over 300 viral 52 miRNAs have been identified, yet only a small fraction have well-understood 53 functions (2–5). Unlike host miRNAs, most viral miRNAs are not well conserved and 54 only ~25% or less are likely to serve as mimics or “analogs” of host miRNAs (1, 2). 55 As such, identifying the most relevant targets of viral miRNAs is not straightforward. 56 A valuable approach towards understanding the functions of viral miRNAs relies on 57 high-throughput target transcript identification (6–13). However, it is unlikely that 58 all of these identified targets are relevant to the virus infectious cycle, thus limiting 59 the utility of such approaches as stand-alone platforms for determining viral miRNA 60 function. Here, we take advantage of natural variations in miRNA gene products 61 from closely related virus strains, with the assumption that important miRNA target 62 transcripts will be preserved throughout evolution. 63 64 miRNAs are ~22 nucleotides long (reviewed in 14), and are derived from primary 65 transcripts (pri-miRNA) containing hairpin precursor molecules (pre-miRNA) (15, 66 16). The pri-miRNA is cleaved by the double-stranded RNA-specific endonuclease 67 Drosha to liberate the pre-miRNA (17–19) that is then exported to the cytoplasm 68 on A uust 0, 2017 by gest http/jvi.asm .rg/ D ow nladed fom (20–22). There, Dicer further cleaves the pre-miRNA (23, 24), and typically a single69 stranded 22mer, enriched from one arm of the hairpin, is more abundantly retained 70 in the RNA induced silencing complex (RISC). The other less abundant strand is 71 sometimes referred to as the “star” strand or “passenger” strand (14). The 5’ end of 72 the 22mer, referred to as the “seed” region (nucleotides 2-8), is especially important 73 for mRNA target binding and typically binds with perfect complementarity to the 3’ 74 untranslated (UTR) region of the target transcript (25). Most miRNA-targeted 75 transcripts display impaired translation followed by subsequent increased turnover 76 (26), which can manifest as an overall decreased steady-state level of the targets 77 (25, 27). In addition, although rare for most animal miRNAs, some plant and viral 78 miRNAs can bind with perfect complementarity (all ~22 nucleotides) to their 79 targets and direct “siRNA-like” cleavage resulting in robust decreases in the steady 80 state levels of the targeted transcripts. 81 82 The polyomaviruses are a family of small, circular, double-stranded DNA circular 83 genome viruses. Most polyomaviruses are thought to take up lifelong infections of 84 their hosts, albeit the mechanisms for how this occurs are poorly understood. In 85 addition, polyomaviruses can undergo robust lytic infection. There are currently 12 86 known human polyomaviruses, of which at least four: Merkel Cell Polyomavirus 87 (MCPyV), Trichodysplasia Spinulosa Polyomavirus (TSPyV), BK Virus (BKPyV), and 88 JC Virus (JCPyV) are associated with serious disease in immunosuppressed humans 89 (28–32). Simian Vacuolating Virus 40 (SV40), a prototypic polyomavirus, undergoes 90 lytic infection in cultured African green monkey cells and as such has been a 91 on A uust 0, 2017 by gest http/jvi.asm .rg/ D ow nladed fom valuable laboratory model for polyomavirus infection (33). We have previously 92 demonstrated that several members of the Polyomaviridae (SV40, BKV, JCV, Simian 93 Agent 12 (SA12), murine polyomavirus (muPyV) and Merkel Cell Polyomavirus 94 (MCPyV)) express miRNAs that lie antisense to the early transcripts and possess the 95 ability to cleave these transcripts via an siRNA-like mechanism (34–38). The 96 conserved nature of this mode of autoregulation amongst divergent polyomaviruses 97 implies importance. However, at least three observations could suggest otherwise. 98 First, the degree of regulation imparted by the miRNA is partial. That is, at least in 99 the laboratory models of lytic infection(35, 37), a high fraction (~50%) of intact 100 early transcripts remains uncleaved by the viral miRNA-RISC (35, 37). Second, a 101 host target has been reported for the JCV star strand miRNA (39). Third, our 102 unpublished data suggest that at least some strains of polyomavirus likely do not 103 encode miRNAs (Cox and Sullivan, unpublished). Therefore, it remains to be 104 determined if autoregulation of the antisense early transcripts is truly important in 105 the polyomaviral lifecycle. 106 107 Here, we address the question of whether polyomaviral miRNA-mediated 108 autoregulation of the early transcripts is a relevant activity, or rather results as an 109 off consequence of the genomic location of the polyoma miRNAs (antisense and 110 therefore necessarily perfectly complementary to the early transcripts). We 111 screened all 63 deposited fully-sequenced isolates of SV40 for possible variations in 112 their pre-miRNAs and derivative miRNAs. We uncovered 17 different classes of pre113 miRNA primary sequence variants, some of which result in different miRNA 114 on A uust 0, 2017 by gest http/jvi.asm .rg/ D ow nladed fom products. We identified a naturally-circulating variant virus (RI257) that generates 115 derivative miRNAs– all possessing different seeds than the miRNA derivatives from 116 the majority of SV40 isolates. We show that, as would be predicted from the altered 117 seed repertoires, the reference strain 776 miRNAs target a different repertoire of 118 host transcripts than RI257. However, strikingly, the RI257 miRNAs efficiently 119 autoregulate early transcript levels to a similar degree as strain 776. These results 120 underscore the likely importance of SV40 miRNA-mediated autoregulation of viral 121 gene expression. Furthermore, this work demonstrates that highly similar viruses 122 can tolerate substantial variability in their miRNA targetomes. 123 124 125 126 127 on A uust 0, 2017 by gest http/jvi.asm .rg/ D ow nladed fom MATERIALS AND METHODS 128 129 SV40 sequence analysis and alignment. 63 unique SV40 complete genome 130 sequences were aligned based on a ~120bp region encompassing the pre-miRNA, 131 using the Geneious software (Biomatters, New Zealand). 132 133 Cell culture and RNA isolation. Human embryonic kidney (HEK) cells 293 134 and 293T, African green monkey kidney epithelial cells BSC-40, and African green 135 monkey kidney fibroblast cells COS-7, were obtained from the American Type 136 Culture Collection (Manassa, VA). All cells were maintained in Dulbecco’s modified 137 Eagle’s medium supplemented with 10% fetal bovine serum (Life Technologies, 138 New York). Total RNA was harvested using an in-house PIG-B solution as described 139 previously (36, 40–42). 140 141 Vector construction, transfection, and High-resolution northern blot 142 analysis. All DNA vector constructs were confirmed by sequence analysis through 143 the Institute of Cellular and Molecular Biology Sequencing Facility at the University 144 of Texas at Austin. The primers used in the construction of the 17 representative 145 SV40 microRNA expression vectors are listed in Table S1. Briefly, the primers are 146 annealed and filled-in using Phusion High-Fidelity DNA Polymerase (New England 147 BioLabs, Massachusetts) according to the manufacturer’s protocol. The PCR 148 products are cloned into the KpnI/XhoI sites of the pcDNA3.1neo expression vector. 149 on A uust 0, 2017 by gest http/jvi.asm .rg/ D ow nladed fom 293T cells were plated in 6-well plates and transfected using the Lipofectamine 15