Complete genomic sequence of a Rubus yellow net virus isolate and detection of genome-wide pararetrovirus-derived small RNAs

Rubus yellow net virus (RYNV) was cloned and sequenced from a red raspberry ( Rubus idaeus L.) plant exhibiting symptoms of mosaic and mottling in the leaves. Its genomic sequence indicates that it is a distinct member of the genus Badnavirus , with 7932 bp and seven ORFs, the ﬁrst three corresponding in size and location to the ORFs found in the type member Commelina yellow mottle virus . Bioinformatic analysis of the genomic sequence detected several features including nucleic acid binding motifs, multiple zinc ﬁnger-like sequences and domains associated with cellular signaling. Subsequent sequencing of the small RNAs (sRNAs) from RYNV-infected R. idaeus leaf tissue was used to determine any RYNV sequences targeted by RNA silencing and identiﬁed abundant virus-derived small RNAs (vsRNAs). The majority of the vsRNAs were 22-nt in length. We observed a highly uneven genome-wide distribution of vsRNAs with strong clustering to small deﬁned regions distributed over both strands of the RYNV genome. Together, our data show that sequences of the aphid-transmitted pararetrovirus RYNV are targeted in red raspberry by the interfering RNA pathway, a predominant antiviral defense mechanism in plants.


Introduction
Red raspberry (Rubus idaeus L.) is grown in temperate regions worldwide for fresh market, jam, juice, purée or individually quick frozen fruit. Raspberry is a host to many different viruses and mixed virus infections appear to be common (Quito-Avila and Martin, 2012). In North America and Europe, Rubus yellow net virus isolates cause raspberry mosaic disease (RMD) while in association with members of Black raspberry necrosis virus and Raspberry leaf mottle virus (McGavin and MacFarlane, 2010). Although complete genomic sequence is available for some of the viruses involved in RMD, the complete genomic sequence of a Rubus yellow net virus isolate is currently unavailable, limiting the development of immunological and nucleic acid diagnostics for the molecular characterization of this disease that reduces berry growth and yield. Rubus yellow net virus (RYNV) produces a bacillifom particle and partial sequence from a highly conserved portion of the virus genome, previously identified RYNV as a putative member of the family Caulimoviridae, genus Badnavirus (Jones et al., 2002).
Badnaviruses include some of the most destructive plant viruses, often causing devastating losses to infected crops (Harper et al., 2004;Huang and Hartung, 2001;Thresh and Owusu, 1986). Initially, it was reported that the badnaviruses were found infecting hosts growing exclusively in tropical and sub-tropical climates. However, badnaviruses have also been found infecting host plants occupying North American and other temperate climates. These include host shrub-like species including red raspberry, gooseberry and ornamental spiraea infected with strains of Rubus yellow net virus, Gooseberry vein banding virus and Spiraea yellow leaf spot virus, respectively.
Badnaviruses have a 120-150 × 30 nm non-enveloped bacilliform capsid containing a singular circular dsDNA genome that is 7.3-8.0 kb in size (Lockhart, 1990). Virions enter a host in a non-circulative semi-persistent manner by a mealybug or an aphid vector. In the case of raspberry, RYNV is vectored by Amphorophora agathonica Hottes also known as the large raspberry aphid. The typical genomic structure for badnaviruses consists of three open reading frames (ORFs), and all of the genes are encoded on the same discontinuous strand (Xu et al., 2011). The first two ORFs encode proteins that are 17 and 15 kDa, respectively and little is known about their function. Most information regarding the badnaviruses is derived from the characterization, organization and highly homologous nature of ORF 3 which encodes a 216 kDa polyprotein that is cleaved post-translationally by the viral-encoded aspartic protease to produce a movement protein, a coat protein and a replicase comprised of a reverse transcriptase and ribonuclease H (Laney et al., 2012;Sether et al., 2012).
High throughput small RNA sequencing has increasingly been used to characterize plant viruses and host-pathogen interactions (Hwang et al., 2013;Kreuze et al., 2009). RNA silencing is a universal system among most eukaryotes to developmentally and temporally regulate gene expression through the production of small RNAs (Huntzinger and Izaurralde, 2011). RNA silencing also plays an important role in antiviral defense whereby, small RNAs (sRNAs) accumulate to abundant levels in infected organisms (Ratcliff et al., 1997;Waterhouse et al., 2001). In plants, RNA silencing is orchestrated by a diverse family of endonucleases, known as Dicers-like proteins (DCLs). In Arabidopsis thaliana, four Dicers (DCL1-4) have been identified (Margis et al., 2006). The dicer-like proteins involved in antiviral defense include DCL2 and DCL4 that generate 22-24-nt small interfering RNAs, from double-stranded replicative intermediates of plant RNA viruses (Voinnet, 2001).
To examine the molecular structure and characteristics of RYNV, nucleic acid from the virus was cloned, sequenced and compared to other members of the genus Badnavirus, family Caulimoviridae. The RYNV genomic sequence most closely resembled that of GVBV, the only other member of the genus Badnavirus from North America with a fully sequenced genome. Small RNA sequence profiling of RYNV infected red raspberry leaf tissue yielded a highly uneven genome-wide distribution of virus-derived RNAs (vsRNAs) with strong clustering to small defined regions distributed over both strands of the RYNV genome. This shows that the raspberry interfering RNA pathway targets specific regions of a plant dsDNA virus sequence, possibly as an antiviral defense mechanism.

Source of infected plant material
To obtain RYNV genomic sequence, leaf tissue was collected from one red raspberry (R. idaeus L.) plant located near Lethbridge, Alberta, Canada expressing symptoms characteristic of RYNV infection (Kalischuk et al., 2008). Tissue was ground in liquid nitrogen and total DNA extracted using hexadecyltrimethylammonium bromide (CTAB) (Doyle and Doyle, 1990). Total RNA for the small RNA sequencing was extracted using the miRNeasy spin column kit (Qiagen, Toronto, Canada) from three RYNV infected plants showing the symptoms similar to those described above and located at the same location. The quantity and quality of the RNA were checked using spectrophotometry and agarose gel electrophoresis.

Polyclonal antibodies to the coat protein of RYNV
A DNA fragment containing a portion of the putative coat protein was prepared by PCR with the oligonucleotides 5 CTGGAATATCTACAAAACGTCTGCAAG3 and 5 GATGCTACAGATTTCATCGTCATCTGA3 (Kalischuk et al., 2008). The fragment was cloned and expressed in the glutathione S-transferase gene fusion system pGEX (GE Healthcare, USA) following the manufacturer's protocol. Amplified products inserted into the recombinant clones were confirmed by DNA sequencing. Protein from inclusion bodies was purified using glutathione sepharose 4B (GE Healthcare, USA). The purified fusion protein was used to immunize rabbits for the production of polyclonal antibodies.

Development of a full length RYNV infectious clone
A full length clone was constructed by ligating two PCR products to produce a greater than unit length fragment of the RYNV genome based on preliminary sequence. NheI was identified as a unique endonuclease restriction site in initial RYNV genomic sequence. Oligonucleotides 5 ATATAGGAGCTAGCCGCAGCTGTGGA3 and 5 AGCGAATTCGCCTGATTCCGAGCTGCTTGTTG3 were used to PCR amplify 4501 bp from the NheI area of the genome (bold nucleotides represent the NheI site and the underlined nucleotides represent an added EcoRI site). A second PCR fragment was amplified 5871 bp from NheI in the opposite direction of the RYNV genome using oligonucleotides 5 AGCGTCGACCTGGTGAACAAGGAGGTACAGAGC3 and 5 TCCACAGCTGCGGCTAGCTCCTATAT3 (bold nucleotides represent the NheI site and underlined nucleotides represent a SalI site). Phusion high-fidelity DNA polymerase (New England Biolabs, Canada) was used for PCR and procedures followed the manufacturer's recommendations. PCR products were examined using agrose gel electrophoresis and DNA was excised, purified using a Qiaex II gel extraction kit (Qiagen, Toronto) and ligated using T4-DNA ligase into the binary vector pBINPLUS. Ligation reactions were directly transformed into Agrobacterium tumefaciens EHA105 by electroporation. Cells were incubated for 3 h and then 20 l of the inoculum was injected using a Hamilton syringe into the leaf veins (5 spots per leaf) of young vegetatively propagated woodland strawberry (Fragaria vesca L.) or R. idaeus. Another construct containing empty pBINPLUS was prepared and tested in a similar manner. Transformed and healthy plants were evaluated for RYNV infection with an immunocapture and subsequent PCR amplification of a 451 bp fragment that was negative with blank and preserum controls (Kalischuk et al., 2008).

Genomic sequencing of RYNV
The complete genomic sequence of RYNV was determined from the RYNV full length infectious clone. The 4501 bp NheI/EcoRI restricted product was subcloned into the SpeI/EcoRI site of pBluescript II SK(−) (Stratagene, Santa Clara, CA, USA). However, due to instability in Escherichia coli, direct cloning of the 5871 bp fragment was unsuccessful. Instead, the 5871 bp fragment was digested with BamHI and each of the two DNA fragments subcloned separately. The SalI/BamHI fragment was cloned into the corresponding sites of pBluescript II SK and the NheI/BamHI fragment was cloned into the corresponding sites of pLitmus 38i (New England Biolabs, Ipswich, MA, USA). Standard procedures were used for the remaining cloning and Sanger sequencing reactions. Primer walking was used to obtain sequence from the entire length of each insert. Sequencing was performed on both DNA strands to produce at least three copies of overlapping sequence for the entire genome (primers are listed in Supplemental Material).

Phylogenetic analysis
Phylogenetic relationships among the badnaviruses were estimated using CLUSTAL X v2 multiple alignment with ORF 3 amino acid sequence. The neighbor-joining clustering algorithm was used to estimate the phylogeny and a rooted tree constructed using CLC Main Workbench software (CLC Biosciences, Cambridge, MA, USA). Confidence estimates were based on bootstrap sets of re-sampling alignments with 1000 replicates and these values were added to the nodes of the trees. GenBank accessions used in the analysis are listed in Supplemental Material.

Small RNA sequencing and mapping of RYNV infected leaf tissue
The RNA was sequenced on an Illumina Genome Analyzer IIx platform (Illumina, Inc., Madison, WI). Briefly, the processing by Illumina consisted of the following successive steps: purification of 20-30 nt RNAs by filtering through a polyacrylamide gel, ligation of 3 and 5 adapters to the 20-30 nt RNAs, c-DNA synthesis by reverse transcription followed by acrylamide gel purification and a final steps of bridge amplification and paired-end sequencing.
RYNV-derived sRNA sequences were identified by mapping the reads to the RYNV genome. Raw reads were filtered for quality and adaptor sequences trimmed using FastX Test Kit (http://hannonlab.cshl.edu/fastx toolkit). A mappable read had a Phred score greater than 30, a 3 ADT and greater than 15 bases after the 3 ADT cut. Reads that were 18-25 nt and that met the filtering criteria were mapped to the RYNV genome using Bowtie (http://bowtie-bio.sourceforge.net/index.shtml). The seed length was set to 16 nt and one mismatch was allowed in the seed region during mapping. The command line for Bowtie on a Mac Unix system was as follows: bowtie -q -l 16 -n 1 -e 40 -k 1 raspberry virus Rubida.fq. Output data were exported to a Microsoft Excel spreadsheet and summarized using standard graphics software (http://www.corel.com).

Results
The complete genomic sequence of RYNV was determined and consisted of 7932 bp (Genbank Accession: KF241951).

Non-coding regions of RYNV genomic DNA
The intergenic region (IR) of RYNV was 969 bp and contained many of the conserved nucleic acid sequences previously described for plant dsDNA viruses (Benfey and Chua, 1990;Medberry and Olszewski, 1993). A plant tRNA Met sequence was predicted with 67 percent nucleotide identity to the complementary sequence of 3 ACCAUAGUCUCGGUCCAA5 , and this region has been previously described as one of the priming site for reverse transcription (Medberry et al., 1990). The 5 end of the tRNA Met was designated as the first nucleotide for numbering the RYNV genome as done with previously characterized badnaviruses. Nucleic acid motifs previously identified and located in the intergenic region of other members of the family Caulimoviridae and identified along the RYNV intergenic region included a TATA box (nt 7523-7536), a cap signal associated with the TATA box (nt 7561-7564), a hexamer motif (nt 7633-7638), an as-1-like sequence (nt 7363-7379) and a GATA box motif (nt 7486-7490) (see Supplemental Material for details on conserved sequences).

Coding regions of RYNV genomic DNA
The nucleic acid sequence of RYNV ORF 1 potentially encodes a 210 amino acid protein of 24 kDa (Table 1) with two nuclear export signals (NES), two coiled-coil regions and a proline rich C-terminus. Homology was detected between sequences of RYNV ORF 1 and both P24 of Rice tungro bacilliform virus and ORF 1 of Gooseberry vein banding virus isolates.
Nucleotide sequence of RYNV ORF 3 was predicted to produce a 1971 amino acid protein with a molecular mass of 210 kDa (Table 1). RYNV ORF 3 resembled that of other badnaviruses encoding a polyprotein containing the movement protein (MP), coat protein (CP), aspartic protease (PR), reverse transcriptase (RT) and ribonuclease H (RNaseH) (Laco et al., 1995;Marmey et al., 1999;Medberry et al., 1990;Tzafrir et al., 1997). The active site HXGX 9 HRX 3 GX 8 D. . .EXDX 3 G (where X represents any amino acid) for the MP was located at RYNV ORF 3 amino acid position 137-182. In addition to similarities to the badnaviruses (Bouhida et al., 1993;Tzafrir et al., 1997), the active site for the MP region of the RYNV genome also showed homology to the 30 K superfamily movement protein from members of the families Caulimoviridae and Flexiviridae and the genus Tobamovirus (Melcher, 2000). A second motif GX 2 SXRFXNYX 7 P (where X represents any amino acid) recognized to be involved in systemic infection was also found within the RYNV amino acids 299-315 (Tzafrir et al., 1997). A total of five proline, glutamic acid, serine and threonine (PEST) sequences found in proteins with short cell half-lives, were located within the putative movement protein region of RYNV. PEST sequences have been described previously (Kawchuk et al., 2001).
The badnavirus coat proteins were previously identified in ORF 3 as containing two cysteine-histidine rich motifs (Cys-His motif), with sequences CXCX 2 CX 4 HX 4 C and downstream CX 2 CX 5 HX 5 CX 2 CX 4 CX 2 C (where X represents any amino acid) (Medberry et al., 1990). These two Cys-His motifs within ORF 3 of RYNV were located at amino acids 864-879 and 998-1024, respectively. At amino acids 1209-1216, 3 to the RYNV putative CP, was an aspartic protease (PR) active site motif AX 2 DXGXT (where X represents any amino acid). The consensus sequences for RT and RNaseH were identified at the 3 end of ORF 3 at amino acids 1553-1556 and 1553-1823, respectively.
Open reading frame 4 (136 aa) encodes a predicted protein of 15 kDa (Table 1). RYNV ORF 4 overlapped the 3 region of ORF 3 but in a different reading frame. Homology was absent between RYNV ORF 4 and other known proteins. Comparative genomics between ORF 4 and a partial sequence for another RYNV isolate (Jones et al., 2002) showed 76% amino acid similarity.
The nucleotide sequence of RYNV ORF 5 was located 305 bp downstream of ORF 3, in the same reading frame as ORF 3 and potentially encoded a 17 kDa protein of 152 amino acids (Table 1). This ORF 5 sequence overlapped the putative consensus sequence complementary to the plant tRNA Met and other predicted promoter elements. Features predicted for RYNV ORF 5 included a signal peptide for localization within the secretory pathway near the N-terminus (1-21 aa), a transmembrane helix near the central portion of the sequence (66-85 aa) and an endocytosis signal (YxxF) towards the C-terminus, with all of these features being previously described (Kawchuk et al., 2001).

Open reading frames along the antisense strand
There were two ORFs predicted at greater than 10 kDa along the antisense strand of the RYNV genome and they were designated as ORFs 6 and 7 (Fig. 1). ORF 6 would produce a 145 amino acid 16.2 kDa protein. Similarly, ORF 7 would encode a 143 amino acids 16.7 kDa protein that contained a zinc finger-like motif with 86% amino acid similarity to the second Cys-His motif of the RYNV putative coat protein.

Infectivity of the full length clone
To confirm that the RYNV genomic DNA sequence was infectious, a full length clone was developed and agro-inoculated into raspberry and strawberry plants. At 14 days post-inoculation, a PCR product corresponding to the RYNV genomic sequence was detected in the immunocapture assay of newly emerging leaves of all plants inoculated with A. tumefaciens carrying the full length clone. The PCR product was not detected in any plants injected with the RYNV full length clone nucleic acid or agro-inoculated with the empty pBINPLUS vector. Leaves agro-inoculated with the RYNV full length clone produced necrotic lesions in F. vesca and R. idaeus (data not shown). Transmission electron microscopy (TEM) of RYNV agro-inoculated plants detected 140 × 30 nm bacilliform particles (data not shown).

Phylogenetic analysis
The phylogeny derived from the badnavirus ORF 3 amino acid sequence (Fig. 2 & Table 2) showed that RYNV was most closely related to GVBV but they remain as two distinct species as they have differences in host range and differences in polymerase nucleotide sequences greater than 20%.

Virus derived small RNA profiling
A total of 14 million sRNA reads between 15 and 30 nt in length were obtained from high throughput sequencing of RYNV-infected raspberry. After filtering the data and using a greater than 99% accuracy cut-off value for base calls, 6.7 million reads were mapped against the RYNV genomic sequence. Of the mappable reads, a total of 0.4% showed sequence homology with RYNV, with a greater number of sRNAs represented by the sense strand than by the antisense strand. The genomic coverage of RYNV by sRNAs was 84% and the size classes of the RNAs were mainly 21-24 nt with 22nt being the most abundant size class (Fig. 3). Mapping the viral small RNAs (vsRNA) to the RYNV genome indicated several prominent areas targeted for RNA silencing. ORF 2 exhibited concentrated high levels of RYNV vsRNAs and these regions corresponded to predicted secondary structures (Fig. 3). There were forty-four regions dispersed across the RYNV genome that were devoid of vsRNAs. Although 82% of these regions were less than 100 nt in length, one identified sRNA desert comprised 6.5% of the RYNV genome or 514 nt (Fig. 3). This 514 nt region corresponded to the 5 portion of the large polyprotein sequence and contained the sequence for the active site for the MP (Fig. 3).

Discussion
The genome of RYNV, a pararetrovirus that infects red raspberry and is transmitted by an aphid vector, was sequenced and found to be 7932 base pairs in size. All badnaviruses, including RYNV, share ORFs 1-3 which have approximately the same size and location within the genome (Lockhart, 1990;Medberry et al., 1990). Nucleotide sequence of ORF 3 encodes a large polycistronic transcript critical to the dsDNA viral lifecycle, as it produces the movement protein, coat protein and reverse transcriptase. Phylogenetic analysis of the amino acid sequence for ORF 3 suggests that RYNV is a member of genus Badnavirus and is most related to GVBV, another member belonging to genus Badnavirus found infecting a temperate climate host.
Analysis of ORF 3 indicated that RYNV closely resembled other members of the Badnavirus genus regardless of the many biological and molecular differences. Unlike ORF 3, ORF 1 and 2 had low homology with the badnaviruses and other known proteins. RYNV ORF 1 and 2 contained a proline rich C-terminus indicating possible non-sequence specific nucleic acid binding potential. Nonspecific nucleic acid binding was demonstrated at the proline rich Neighbor-joining dendrogram of sequence relationships determined using the amino acid sequence alignment of the reverse transcriptase among species within genus Badnavirus. The tree was rooted to rice tungro bacilliform virus (RTBV). Details of the accessions used in the analysis are in the Supplemental Material. Alignments were produced by a CLUSTAL algorithm and the dendrogram was produced by CDC Main Workbench software. Horizontal distances were proportional to sequence distances and vertical distances were arbitrary. The dendrograms was bootstrapped 1000 times (shown at nodes).
terminus region for ORF 2 of cacao swollen shoot virus (CSSV) and RTBV (Jacquot et al., 1996(Jacquot et al., , 1997. RYNV ORF 1 and 2 also contained coiled-coil regions that are associated with protein-protein interactions. Coiled-coil domains located in these ORFs are widespread across plant pararetroviruses (Leclerc et al., 1998) and protein-protein interactions were detected through tetramerization (Stavolone et al., 2001). The amino acid sequence of RYNV ORF 2 was especially interesting because in addition to possible protein-protein interactions and nucleic acid binding potential, the DXG motif necessary for aphid transmission of CaMV (Schmidt et al., 1994) was present at the C-terminus of the sequence and may function in aphid transmission. Evidence supporting a Needleman-Wunsch algorithm was used for alignments. Identity and similarity were calculated as # matches divided by longest total sequence length of either query or subject multiplied by 100. b Based on same position along the genome, RYNV ORF 4 was compared to CSSV ORF Y, CYMV ORF 6, and DMV ORF 6. c RYNV ORF 5 was compared to DMV ORF 7 and RTBV P46. additional function includes the prediction of nuclear export signals (NES) contained in ORFs 1 and 2, suggesting a nucleocytoplasmic function.
Unique to the RYNV genome were two small ORFs located on the sense strand. CSSV, Citrus mosaic virus (CMBV) and Dracaena mottle virus (DMV) also have an ORF located in relatively the same position as RYNV ORF 4 (Su et al., 2007;Huang and Hartung, 2001). Although homology was not detected between RYNV ORF 5 and any other proteins, DMV, CaMV and RTBV have an ORF of a similar size near the same vicinity. Another feature unique to the RYNV genome was that ORF 6 and ORF 7 occur in the antisense sequence and may encode several zinc finger-like motifs. ORF 7 encoded a zinc finger-like motif with 86% amino acid similarity to the second Cys-His motif identified in the RYNV putative coat protein. Zinc finger motifs are often involved in forming fingerlike protrusions involved in binding DNA, RNA, protein and/or lipid substrates. There are many occurrences of zinc finger signatures in viral genomes and functional studies have suggested diverse roles such as in structure and regulation (Tanchou et al., 1998). Since RYNV ORFs 6 and 7 are located on the antisense strand and are not preceded by sequences associated with promoter activity, the transcription of RYNV is likely asymmetric like that of the other members of genera Badnavirus and Caulimovirus (Medberry et al., 1990).
An increasing number of DNA viruses are being reported that are targeted by the interfering RNA pathway and produce RNA silencing suppressors (Blevins et al., 2006;Moissiard and Voinnet, 2006;Bronkhorst et al., 2012). In our study, approximately 0.4% of total sRNAs generated by raspberry were RYNV-derived vsRNAs which spanned 84% of the RYNV genome. Involvement of all known DCLs in the production of a diverse pool of 21-24 nt vsRNAs has been demonstrated as a general host plant response to pararetrovirus infection (Blevins et al., 2006;Moissiard and Voinnet, 2006). Our sequencing of sRNAs from RYNV-infected raspberry leaf tissue revealed a diverse pool of 21-24-nt vsRNAs, supporting the aforementioned interaction between pararetrovirus and host. The 22-nt vsRNAs were the predominant size class of vsRNAs spanning the RYNV coding regions and concentrated in areas such as ORF 2 that could form considerable secondary structure (Stavolone et al., 2001). Another study that partially examined size and distribution of vsRNAs in badnavirus-infected tissue showed that the majority of vsRNAs were also 22-nt in size and distributed along the genome (Kreuze et al., 2009).
Mapping the vsRNAs to the RYNV intergenic region revealed that they were predominately 21 and 22-nts in size (Fig. 3). Previously, it was demonstrated that in Arabidopsis thaliana and Brassica rapa exposed to CaMV, the 24-nt size class of vsRNAs was predominant and that most of the vsRNAs clustered within the long pararetrovirus leader sequence (Blevins et al., 2011;Moissiard and Voinnet, 2006), rather than being dispersed throughout the genome, as with RYNV. Mapping of the RYNV vsRNAs against the entire viral genome revealed a 514 nt region that appears to have escaped RNA silencing. This sequence lacking sRNAs, spanned the region of the RYNV genome associated with the binding domain in the badnavirus movement protein. Deep sequencing to identify novel viruses in sweet potato reported partial vsRNA profiles for two badnaviruses that also lacked vsRNA in a similar location (Kreuze et al., 2009).
Genomes of RNA viruses produce a double-stranded RNA intermediate during replication (Weber et al., 2006). These dsRNA products are recognized by the RNAi mechanism and cleaved into virus-derived siRNAs that target homologous ssRNA for further degradation. Although DNA viruses do not require dsRNA intermediates, they are known to produce dsRNA during infection (Weber et al., 2006). Recently it was shown that infection of Drosophila melanogaster with invertebrate iridescent virus 6 (IIV-6) elicited the RNAi pathway involving Dcr-2 and Argonaute-2 (Bronkhorst et al., 2012). The highly uneven genome-wide distribution with clustering of vsRNAs to defined regions designated hotspots in both strands of IIV-6 were also observed with RYNV. Together, our results show that RYNV induces the siRNA antiviral defense pathway in red raspberry, producing an uneven genome-wide distribution of vsRNAs with strong clustering to small defined regions distributed over both strands of the genome. These results increase the complexity of the pararetroviruses and provide evidence of intriguing host-badnavirus interactions in economically important plants.