Small-RNA analysis of pre-basic mother plants and conserved accessions of plant genetic resources for the presence of viruses

Pathogen-free stocks of vegetatively propagated plants are crucial in certified plant production. They require regular monitoring of the plant germplasm for pathogens, especially of the stocks maintained in the field. Here we tested pre-basic mother plants of Fragaria, Rubus and Ribes spp., and conserved accessions of the plant genetic resources of Rubus spp. maintained at research stations in Finland, for the presence of viruses using small interfering RNA (siRNA) -based diagnostics (VirusDetect). The advance of the method is that unrelated viruses can be detected simultaneously without resumptions of the viruses present. While no virus was detected in pre-basic mother plants of Fragaria and Ribes species, rubus yellow net virus (RYNV) was detected in pre-basic mother plants of Rubus. Raspberry bushy dwarf virus (RBDV), black raspberry necrosis virus (BRNV), raspberry vein chlorosis virus (RVCV) and RYNV were detected in the Rubus genetic resource collection. The L polymerase encoding sequence characterized from seven RVCV isolates showed considerable genetic variation. The data provide the first molecular biological evidence for the presence of RYNV in Finland. RYNV was not revealed in virus indexing by indicator plants, which suggests that it may be endogenously present in some raspberry cultivars. In addition, a putative new RYNV-like badnavirus was detected in Rubus spp. Blackcurrant reversion virus (BRV) and gooseberry vein banding associated virus (GVBaV) were detected in symptomatic Ribes plants grown in the field. Results were consistent with those obtained using PCR or reverse transcription PCR and suggest that the current virus indexing methods of pre-basic mother plants work as expected. Furthermore, many new viruses were identified in the collections of plant genetic resources not previously tested for viruses. In the future, siRNA-based diagnostics could be a useful supplement for the currently used virus detection methods in certified plant production and thus rationalize and simplify the current testing system.


Introduction
Virus-free and genetically true-to-type propagation materials are crucial for certified production of small fruits. Certified production of small fruits has been carried out in Finland since 1977 [1]. Pre-basic mother plants must be virus-free and are tested using recommended indexing procedures [1][2][3][4][5]. Basic methods consist of bioassays such as sap-inoculation and graftinoculation from the tested plant to recommended indicator plants that display symptoms when infected, and serological assays such as enzyme-linked immunosorbent assays (ELISA). In addition, nucleic acid analysis-based methods, such as polymerase chain reaction (PCR) and reverse transcription (RT)-PCR, have become commonplace. Pathogens not permissible in pre-basic mother plants of small fruits are regulated by a decree announced by the European Union (EU) in 2017. The new legislation will change the testing requirements and requires more frequent testing of pests in many plant species.
The national plant genetic resources program of agriculture, horticulture and forestry was established in Finland in 2003 with the aim to conserve genetic resources and promote their sustainable use. The horticultural plants are conserved in clonal field collections, in vitro cultures, and cryopreserved [6,7]. Because of the short growing season and cold winter, only a limited selection of plant species and cultivars can be maintained outdoors in Finland. On the other hand, in summer the long day length and daily variation in temperature are beneficial for accumulation of vitamins and aromatic substances in small fruits [8]. The most important plant species in 2017 were strawberry [Fragaria x ananassa (Weston) Royer] grown on 3800 ha, raspberry on 429 ha, and currants (Ribes nigrum L. and R. rubrum L.) on 1740 ha [9].
The maintenance of vegetatively propagated plants free of pathogens is challenging in the field, because plants such as small fruits tend to become infected with viruses [10]. The raspberry aphid (Aphis idaei van der Goot) and the European large raspberry aphid (Amphorophora idaei Börner) are the most common vectors that transmit viruses to red raspberry (Rubus idaeus L.) in Northern Europe [11][12], whereas various Cecidophyopsis mites transmit viruses to Ribes species [13]. Strawberry aphid (Chaetosiphon fragaefoli Cockerell) transmitting viruses to Fragaria species has not yet been detected in Finland but was recently detected in Sweden, a neighboring country [14].
The aim of this study was to test viruses in pre-basic mother plants and conserved accessions of plant genetic resources for the presence of viruses. The plants were tested using a newly developed method, deep sequencing of virus-derived small interfering RNAs (siRNAs) in plants, which relies on the ability of plants to recognize and inactivate double-stranded RNA by splicing it into fragments of certain sizes [15,[16][17][18][19]. The method is fast and costeffective, as it allows detection of unrelated viruses simultaneously. Furthermore, RNA samples can be pooled prior to analysis, enabling the detection of large numbers of viruses from different samples simultaneously.

Plant material
Pre-basic mother plants of economically important cultivars and other cultivars of Fragaria, Rubus and Ribes spp. suitable for Finnish climate were maintained at the Natural Resources Institute Finland (Luke) in Laukaa (Luke-Laukaa, 62˚33'N, 25˚99'E) (Fig 1). The plants had been tested for viruses at the intervals stipulated in the rules for maintenance using the recommended methods (ELISA, PCR, graft inoculation and sap inoculation to herbaceous hosts) [3][4][5] and were maintained in an insect-proof greenhouse ( Fig 1C). Also, a few samples of Rubus breeding lines (not yet approved as pre-basic mother plants) and indicator plants of Fragaria and Rubus maintained in vitro were tested (Fig 1D, S1 Table). Each propagation line originating from a single meristem was tested individually. For each cultivar/line, there were 1-7 propagation lines. In 2013, several leaves were taken from each pre-basic mother plant of Ribes and Rubus to obtain representative leaf samples of each propagation line. Samples were stored at −80˚C at Luke in Jokioinen (Luke-Jokioinen, 220 km southwest of Laukaa). In addition, similar representative samples were taken from the pre-basic mother plants of Fragaria and from the Rubus stock plants used as virus indicators and maintained in vitro.
The Finnish plant genetic resources collection of Rubus was developed in the 1980s. Accessions in the collection originate from different parts of Finland and consist of material selected based on interesting breeding traits, such as winter hardiness. After several rounds of selection, the Rubus collection currently consists mainly of Rubus cultivars bred in Finland. They are maintained in pots in the field during the summer and in cold storage during the winter at Luke in Piikkiö by the Baltic Sea (Luke-Piikkiö; 350 km southwest from Luke-Laukaa). Plants of the Rubus collection had not been previously tested for viruses. The plants were sampled at Luke-Piikkiö in June 2015 (S1 Table), except Rubus idaeus cv. Indian Summer, which was maintained and sampled at Luke-Mikkeli (330 km northeast from Luke-Piikkiö). Samples were collected from one to three plants and parallel samples were taken from each cultivar or plant stock. Several leaves, mostly those showing symptoms, were sampled from each plant. In addition, samples were collected from Ribes nigrum and Ribes rubrum that showed distinct symptoms of malformed leaves and conspicuous yellowing, respectively (Fig 1E-1H). These Ribes plants were from a field trial at Luke-Piikkiö and were not included in the germplasm collection (S1 Table). All samples were stored at −80˚C.

RNA isolation, siRNA sequencing and deposition of raw siRNA reads
Total RNA was extracted from leaves as described [20][21] with some modifications to diminish interference of secondary metabolites, phenolics, and polysaccharides, especially for the R. rubrum samples. RNA concentration and purity were determined with Gene Quant or Nanodrop 2000c UV-vis Spectrophotometer (Thermo Scientific, Wilmington, DE, USA). Equal amounts of total RNA (100 ng) were combined from each leaf sample to obtain six pools, each of which contained 21-43 samples (S1 Table). Four pools (GEN17-20) contained samples from the pre-basic mother plants of certified plant production, including Fragaria (GEN17), Fragaria and Rubus (GEN18), Rubus and Ribes (GEN19), and Ribes (GEN20) (S1 Table). Samples in two pools (HXR1 and HXR2) were from the plant genetic resource collection (S1 Table). Pool HXR1 contained only Rubus samples, whereas HXR2 also contained two samples from symptomatic Ribes plants grown in the field.
The RNA pools were sent to Fasteris SA (Plan-les-Outes, Switzerland) for sequencing of the small RNAs. RNA samples were subjected to acrylamide gel electrophoresis in Fasteris, and small RNAs < 45 nt long were purified from the gel. Single-stranded 3' adapters and barcoded 5' adapters were ligated to the small-RNA oligonucleotides, followed by reverse-transcription and amplification by PCR to generate DNA colony template libraries. The PCR products were purified and diluted to 10 nM prior to high-throughput DNA sequencing by Illumina Genome Analyzer (HiSeq 2500). The raw siRNA reads obtained by Illumina were deposited to the European Nucleotide Archive (ENA) with accession numbers ERP108051 (gene pools GEN19-GEN20) and PRJEB30660 (gene pools HXR1 and HXR2).

DNA isolation and cDNA synthesis used for PCR analysis
For PCR analysis of DNA viruses, DNA was extracted from fresh or frozen leaves using DNeasy Plant Mini Kit (Qiagen, Hilden, Germany), the cetyl trimethyl ammonium bromide (CTAB) method, or a combination of organic extraction and the CTAB method [22], http:// cshprotocols.cshlp.org/content/2010/11/pdb.prot5515. For PCR analysis of RNA viruses, RNA was extracted from frozen leaves with the RNeasy Plant Mini Kit (Qiagen). Total RNA was treated with DNase I (Thermo Scientific) and cDNA was synthesized using the Revert Aid RT Kit (Thermo Scientific) and random hexamer primers. The quality of the cDNA was checked in control PCR reactions using the Phire Plant Direct PCR Kit (Thermo Scientific).

PCR analysis of samples
For PCR analysis, cDNA was diluted 10-fold to minimize the concentration of possible compounds that may inhibit PCR reactions. PCR was carried out using puReTaq Ready-To-Go PCR Beads (GE Healthcare UK Limited), or Phusion DNA polymerase or PCR Phire Hot Start II DNA polymerase (Thermo Scientific, Vilnius, Lithuania), and virus-specific primers (S2 Table). PCR products were purified using the E.Z.N.A. Gel Purification Kit (Omega BioTech Inc., Norcross, GA, USA) or QIAquick Gel Extraction Kit (Qiagen). They were sequenced by Macrogen (The Netherlands) or in the sequencing facility at Luke-Jokioinen. Most samples were sequenced directly without cloning, but a few PCR products were cloned into the vector pJET (Thermo Scientific) for sequencing using standard methods.

Data analysis
The reads of 21-24 nt from high-throughput DNA sequencing data were assembled into contigs using Velvet software [23]. The contigs were then used in database searches to find homologous sequences and identify viruses present in the sample pools. Mapping of siRNA reads to the virus sequences identified by BLAST (E-value threshold of 1.00E-10) was carried out with Bowtie [24] in order to examine the coverage and depth of siRNA reads in virus sequences. Analyses were carried out in parallel using the VirusDetect pipeline [25] freely available at http://bioinfo.bti.cornell.edu/tool/VirusDetect/.
Nucleotide and amino acid sequences were aligned using MultAlign [26]. Phylogenetic relationships were analyzed using the neighbor-joining method [27] implemented in MEGA7 [28] using the Kimura two-parameter model [29] for nucleotides and the Poisson model for amino acid sequences. Statistical significance of tree branching was tested by performing 1000 bootstrap replications.

Viruses detected by siRNA sequencing
Sequencing of the pooled RNA from the plant samples resulted in 10.0-15.9 million reads (21-24 nt) per pool. No virus was detected in sample pools GEN17 and GEN20, which contained samples from pre-basic mother plants of Fragaria and Ribes, respectively (Table 1, S1 Table).
Raspberry bushy dwarf virus (RBDV) and raspberry yellow net virus (RYNV) were detected in pool GEN18, whereas only RYNV was detected in pool GEN19 (Table 1). GEN18 contained samples from pre-basic mother plants of Fragaria, a few breeding lines of Rubus, and a few virus indicator plants of Fragaria and Rubus. GEN19 contained samples from pre-basic mother plants of Rubus and Ribes, a few breeding lines of Rubus, and virus indicator plants of Rubus (S1 Table).
Mapping of the virus-derived siRNA reads to the RBDV RNA1 and RNA2 showed high coverage of 72.6% and 80.3%, respectively, and high depth of coverage of 27.9 and 19.6, respectively (Table 1). High coverage with siRNA reads (81.8%) of the RYNV genome was observed in pool GEN19, whereas only low coverage (10.4%) was observed in pool GEN18 (Table 1).  3 Average number of times the nucleotides in the reference genome were covered by the siRNA reads of the sample (identical nucleotides, no mismatches allowed). 4 RBDV, raspberry bushy dwarf virus. 5 RYNV, rubus yellow net virus.
Pools HXR1 and HXR2 contained samples from the collection of Rubus plant genetic resources (S1 Table). Four known raspberry viruses-RBDV, RYNV, black raspberry necrosis virus (BRNV), and raspberry vein chlorosis virus (RVCV)-were detected in pool HXR1 (Table 2). Five viruses-RBDV, RYNV, BRNV, blackcurrant reversion virus (BRV), and gooseberry vein banding associated virus (GVBaV)-that had infected raspberries or Ribes plants were identified in the sample pool HXR2 (Table 2), which contained also one symptomatic plant each of R. nigrum and R. rubrum grown in the field. Mapping of the virus-derived siRNAs to viral genomes, including to the partial sequence of RVCV, showed high coverage (60.7-89.9%) of all identified viruses, except BRNV, which had a low coverage of 11.6-24.2% ( Table 2).
The samples included in the four pools tested by VirusDetect (and usingVelvet and BLAST search) that were found to contain viruses were also tested by PCR or RT-PCR to confirm the results. As expected, RYNV and RBDV were detected in pool GEN18 (S1 Table). RYNV was detected in Rubus breeding line Z-22, whereas RBDV was identified in Rubus breeding line Z-13 (maintained as a positive control for RBDV). RYNV was detected by PCR in 10 samples from pool GEN19 (S1 Table). The RYNV-positive samples were pre-basic mother plants of  Tables 3 and 4). BRV was detected in a sample of R. nigrum, and GVBaV was detected in a sample of R. rubrum (Table 4). RVCV, raspberry vein chlorosis virus; RYNV, rubus yellow net virus. 3 Coverage of identical 21-to 24-nt siRNAs relative to the full-length viral reference sequence; in case of RVCV, coverage of the RNA-dependent RNA polymerase encoding region (partial L gene). 4 Average number of times the nucleotides in the reference genome were covered by the siRNA reads of the sample (identical nucleotides, no mismatches allowed). https://doi.org/10.1371/journal.pone.0220621.t002

Sequence variability of RYNV
The sequence variability of RYNV was further characterized because only a partial sequence of the open reading frame 3 (ORF3) (AF468454, [29]) and two full-length sequences of RYNV, namely RYNV-Ca (accession number KF241951, [31]) and RYNV-BS (KM078034, [32]), are known and available in GenBank. In contrast, RBDV and BRNV sequences were not further characterized here, as their sequence variability has been relatively well studied [21,[33][34][35]. A small genomic region (559-844 nt) of ORF3 of RYNV was amplified and sequenced from 22 RYNV-positive Rubus samples from pre-basic mother plants and from plants in the collection of genetic resources (Tables 3 and 4). The sequences of RYNV-Ca and RYNV-BS, as well as the variant of the latter (RYNV-BSa; AF468454) were included for comparison (S1 Fig). RYNV-Ca originates in a symptomatic red raspberry in Canada, whereas RYNV-BS has been described from red raspberry cv. "Baumforth's Seedling A", which originates in the United Kingdom and has been maintained in Canada. All Finnish RYNV isolates sequenced in this study showed higher sequence identity to RYNV-Ca than RYNV-BS. RYNV isolates from Finnish raspberry cultivars Maurin Makea and Takalan Herkku from pre-basic mother plants of Rubus (isolates MM-78, MM-79, MM-80, and TH-88), the Rubus collection of plant genetic resources (isolates MM-3dg and MM-4), and also the RYNV isolate PK-16 from raspberry cultivar Pisan keltainen (Rubus idaeus f. chlorocarpus Krause) were identical ( Table 5). The two sequences from raspberry cultivar Muskoka (Mu-81 and Mu-83) and Rubus breeding line Z23 (Z23-97 and Z23-98) were identical and also were very similar to each other, with substitutions Table 3. Viruses detected in pool HXR1 consisting of raspberry samples from the Rubus collection of plant genetic resources grown in the field in Piikkiö.

17
Uusikirkko at only three nucleotide positions. Fourteen of the 22 RYNV isolates sequenced in this study differed at the nucleotide level (S1 Fig). Sequencing of a few isolates resulted in double peaks (degenerate nucleotides) that were detected by direct sequencing of the PCR products (S1 Fig).
The internal nucleotide sequence variability was identical or very similar in samples that were derived from the same raspberry cultivars, especially in the samples from pre-basic mother plants of Rubus. They included MM-78, MM-79, and MM-80 (raspberry cultivar Maurin Makea); TH-87 and TH-88 (raspberry cultivar Takalan Herkku); and Jatsi-109 (raspberry cultivar Jatsi) in pre-basic mother plants of raspberry and also in raspberry isolates MM-3dg, MM-4 (cultivar Maurin Makea), and PK-16 (cultivar Pisan keltainen) in the Rubus collection of genetic resources (S1 Table, Table, S1 Fig). In general, RYNV sequences amplified from the Rubus collection of plant genetic resources grown in the field showed more variability than the isolates from pre-basic mother plants.
Comparison of the deduced amino acid sequences indicated that all Finnish RYNV isolates were very similar (97.3-100% aa identity) and nearly identical to RYNV-Ca (96.8-98.9% aa Ribes rubrum (Valkoinen Suomalainen) ----+ identity) (Fig 2) over the part of the viral genome analyzed. In addition, the degenerate nucleotides detected in the sequences were synonymous at amino acid level, and only 9 RYNV isolates out of 22 differed at the amino acid level. Sequence variability of RYNV was further characterized by cloning and sequencing PCR products of RYNV isolates from raspberry cultivar Maurin Makea. Analysis of 10 PCR clones The previously characterized isolates RYNV-Ca and RYNV-BS differ in their genomic organization [32]. To characterize the genomic organization of the Finnish RYNV isolates, a larger genomic region (2348 nt) of isolate MM-3 was amplified and sequenced (see nucleotide alignment in S4 Fig). Genomic organization of MM-3 was similar to that of RYNV-Ca, and the nucleotide sequence of MM-3 was 98.7% identical to that of RYNV-Ca. MM-3 expresses ORF5 from the sense strand and ORF7 from the antisense strand, which is similar to RYNV-Ca. These ORFs were missing from RYNV-BS [32]. ORF5 and ORF7 differed only at four nucleotide positions and one nucleotide position, respectively, which was predicted to cause no change or changes in two amino acids, respectively, between MM-3 and RYNV-Ca. ORF4 was identical among the isolates (S4 Fig). Phylogenetically accessed, all Finnish RYNV isolates grouped to the same large cluster together with RYNV-Ca, whereas RYNV-BS was placed on a separate branch. Finnish isolates were further divided into two subclusters containing the isolates Z23-97, Z23-98, Mu-81, and Mu-83 and the isolates IS-39 and IS-40 ( Fig  3).
In addition to isolates of RYNV, a few sequences resembling RYNV were detected in three samples from raspberry cultivars RU168 Krusenbergs (isolates Krus-11 and Krus-12) and HY 6230 (isolate HY-20) (Tables 3 and 5). These three identical isolates were only 71.6% and 83.4% identical to RYNV-Ca and RYNV-BS, respectively, at the nucleotide level, and 75.6% and 85.6% identical, respectively, at the deduced amino acid level (S5 Fig). In phylogenetic analysis, the three identical RYNV-like isolates formed a clearly distinct clade supported by a high (99%) bootstrap value (Fig 3). Hence, they may represent a new badnavirus species or, alternatively, be remnants of an endogenous virus sequence.

Sequence variability of RVCV
The only published data on RVCV sequences is a 3030-nt-long region of the L polymerase gene [36]. Therefore, part (827 nt) of the L polymerase-encoding sequence was characterized from seven Finnish RVCV isolates ( Table 5). The sequences showed considerable variability at  the nucleotide level and differed also from the previously characterized RVCV isolate. Direct sequencing of PCR products revealed double peaks, indicating the presence of different RVCV isolates in individual raspberry plants in the Rubus collection of genetic resources in the field. For example, two or more RVCV variants were found in raspberry cultivars RU53 and RU54. Therefore, some of the PCR products were cloned and sequenced, and isolates designated as RVCV-8 (cultivar RU53), RVCV-9-3, RVCV-9-4 and RVCV-10 (Table 3). Identical RVCV isolates, as judged by the sequenced 827-nt long region of the viral genome, were detected in cultivar RU54 and RU55 (Fig 4, Table 5). Isolate RVCV-21 (cultivar RU25 Norna) was most similar-99.2% at the nucleotide level (S6 Fig) and 100% at the amino acid level (Fig 4A)-to the RVCV isolate from raspberry in Scotland (accession number FN812699). These two isolates formed a subcluster in the phylogenetic tree (Fig 4B). Isolates RVCV-13 and RVCV-14 from raspberry cultivar RU55 and isolate RVCV-9-4 from RU54 were identical and showed 93.1% and 99.3% similarity at the nucleotide and amino acid levels, as compared with the previously identified RVCV (Fig 4A,  S6 Fig). They formed a small subcluster in the cluster that contained the previously characterized RVCV and RVCV-21 ( Fig 4B). Isolates RVCV-8 from cultivar RU53 and RVCV-9-3 and RVCV-10 from cultivar RU54 were also identical to each other. They also differed most from the Scottish RVCV (84.7% nt and 97.8% aa identity to FN812699) and thus formed a separate small cluster (Fig 4, S6 Fig).

Discussion
In this study, we tested the presence of viruses in pre-basic mother plants and conserved accessions of plant genetic resourses using the newly developed virus detection method (deep sequencing of virus-derived siRNAs). Four viruses (RBDV, BRNV, RYNV, and RVCV) were detected in the Rubus collection of plant genetic resources. These plants have not been tested for the presence of viruses before. In addition, two viruses (BRV and GVBaV) were identified in symptomatic Ribes plants grown in a field trial.
Only one virus (RYNV) was detected in pre-basic mother plants of Rubus by siRNA-based diagnostics, and no virus was detected in pre-basic mother plants of Fragaria and Ribes species. The results confirm that thermotherapy combined with meristem tip culture and regular virus indexing generally produce virus-free plants [37]. RYNV and other viruses belonging to the raspberry mosaic virus complex have been tested previously by grafting using Rubus idaeus 'Malling Landmark', R. idaeus 'Malling Delight' and Rubus occidentalis 'Cumberland' as indicator plants. However, the results of this study indicate that graft inoculation to indicator plants may be insufficient for indexing method to detect RYNV. RYNV (genus Badnavirus, family Caulimoviridae) is one of the viruses associated with raspberry mosaic disease complex, together with BRNV (unassigned genus, family Secoviridae) and raspberry leaf mottle virus (RLMV, genus Closterovirus, family Closteroviridae) [10]. Both RYNV and BRNV were detected in this study, whereas no RLMV was found. BRNV is an aphid-transmitted bipartite RNA virus that is present in very low titers in raspberries [38]. Sequences of BRNV isolates from the United Kingdom, the United States, and Finland show large genetic variation [21,33,34]. Mixed infection of different Rubus viruses (RYNV, BRNV, RBDV, and RVCV) appeared common among plants from the Rubus collection of genetic resources grown in the open field in Finland. Of 40 analyzed plant samples, 17 were found to contain two or three viruses. RBDV and BRNV have been previously identified in both cultivated and wild raspberries in using the neighbor-joining algorithm [26]. GVBaV (HQ852248) was included as a root. Numbers at branches represent bootstrap values of 1000 replicates.  Finland [7,21,39,40]. In addition, RBDV has been detected in arctic brambles (Rubus arcticus L.) in Finland [39,41]. RBDV is a pollen-transmitted virus with a bipartite RNA genome [10]. It has a wide host range and worldwide distribution, and its genomic sequence has been well characterized [35,[42][43][44][45].
RYNV was not revealed in virus indexing by indicator plants in pre-basic mother plants of Rubus, which suggests that it may actually be an endogenous virus that is present in some raspberry cultivars, as suggested [46,47]. In plants of Rubus collection of plant genetic resources, RYNV sequences were variable. All isolates, except the one from Maurin Makea, were unique for their sequence. Some samples also contained sequences that resembled RYNV but were, however, notably different from them. The nucleotide sequence identity of the sequenced ORF3 region of these RYNV-like isolates was <84% as compared with previously characterized RYNV isolates. Thus, they may form a new virus species in the genus Badnavirus. Alternatively, they may represent sequences of endogenous badnaviruses that have integrated into the plant genome.
The new RYNV-like virus and RYNV detected in this study were the most common partners in mixed virus infections, followed by RBDV (genus Idaeovirus). RBDV in combination with RYNV or other aphid-transmitted viruses, such as RLMV and raspberry latent virus (RpLV), cause severe crumbly fruit symptoms in raspberries [10]. Symptoms of raspberry yellows and yellow mosaic disease and of raspberry vein chlorosis have been described in Finnish raspberry cultivars [48] that could correspond to RYNV and RVCV, respectively. So far, however, no definitive reports or sequence information of these viruses have been available from Finland before this study.
The host range of RYNV is restricted to Rubus species. This virus is transmitted by aphid species Amphorophora agathonica Hottes in North America and A. idaei Börner in Europe [29]. Infection by RYNV alone results in the development of a netlike chlorosis along the leaf veins in some raspberry cultivars, but in most raspberry cultivars it is asymptomatic [49]. Jones et al. [29] identified RYNV as a distinct badnavirus. Thus far, two complete sequences of RYNV have been characterized, each of which shows altered genomic organization [31,32]. The DNA genome of the first sequenced isolate, RYNV-Ca, originated from a red raspberry plant in Canada [50] and consists of 7932 nt and seven ORFs. Five of the ORFs are in a sense orientation and two are on the antisense strand [31]. The isolate RYNV-BS from red raspberry cultivar Baumforth's Seedling A, originally from England, is 7836 nt long and shows 82% nucleotide identity to isolate Ca. It encodes, however, only five ORFs, of which four are on the sense strand and one is on the antisense strand [32].
This study shows the first molecular biological evidence for the presence of RYNV in Finland. We analyzed a 559-nt region of the conserved ORF3 of RYNV from 22 Rubus samples. The samples represented different red raspberry cultivars and breeding lines grown in Finland. All identified Finnish RYNV isolates showed higher sequence similarity to RYNV-Ca, including identical genomic organization, as compared with RYNV-BS. Thus, the data do not support the previously suggested assumption that RYNV-Ca represents a North American lineage and RYNV-BS a European lineage of RYNV [32]. The sequence variation among Finnish RYNV isolates was also relatively large and these isolates were further divided into three subclusters, as the RYNV isolates from cultivars Z23 and Muskoka and those isolates from the cultivar Indian Summer formed two smaller subclusters.
The Finnish RYNV isolates from raspberry cultivars Maurin Makea, Takalan Herkku, and Pisan Keltainen in the largest subcluster contained several variable degenerate nucleotides but, nevertheless, were identical at the amino acid level. Two of these isolates are from the prebasic mother plants of Rubus and represent two Finnish cultivars (Maurin Makea and Takalan Herkku). The cultivars are progeny from crossings by free pollination between wild and cultivated Rubus species. The original mother plants show no viral symptoms themselves, and virus-indexing using indicator plants has not revealed any virus. There is increasing evidence that endogenous forms of RYNV exist in some raspberry genomes [46,47]. Therefore, these isolates may represent an endogenous form of RYNV instead of actively replicating exogenous viruses, as suggested by the data above. The integration of endogenous badnaviruses is assumed to have taken place by illegitimate recombination into host genomes, and their presence is not necessarily associated with infection [51][52][53].
RVCV belongs to family Rhabdoviridae, in which the viruses have a negative-sense RNA genome of 12000-14500 nt, and is common in Europe [10]. Based on symptoms typical to RVCV, it has been reported that RVCV may be common in wild raspberries and the raspberry cultivars Asker, Malling Promise, and Preussen in Finland [48]. Our study provides the first molecular evidence of the presence of RVCV in cultivated raspberries in Finland. RVCV is transmitted by raspberry aphid (Aphis idaei) and may easily spread between plants. Symptoms of RVCV include chlorosis of minor veins and reductions in plant vigor and raspberry yields, especially in mixed infections [10]. There is only one published partial RVCV sequence available. It is from the L polymerase-encoding region of the virus [36]. Therefore, no information about genetic variability of RVCV is available. In this study, we show that there is substantial sequence variation among Finnish RVCV isolates, and many plants are simultaneously infected by two or more different RVCV isolates. The characterized Finnish RVCV isolates were detected in a representative collection of cultivated raspberries obtained from different parts of Finland and maintained in Luke-Piikkiö in the field. The nucleotide sequence identity of individually characterized Finnish RVCV isolates, as compared with the previously characterized RVCV isolate, was 93.1-99.2%. However, actual variability is probably higher, as indicated by the numerous degenerate nucleotides in plants infected with a mixture of RVCV isolates.
Symptomatic Ribes plants in a field trial in Piikkiö were shown to contain BRV (genus Nepovirus) in R. nigrum and GVBaV (genus Badnavirus) in R. rubrum. BRV is a bi-partite RNA virus that is the causative agent of the reversion disease of blackcurrants. Eriophyid gall mite of blackcurrant (Cecidophyopsis ribis Westwood) transmits BRV and also causes substantial damage, including distortion of leaves, galling, and sterility of buds [54]. BRV was originally isolated from reverted blackcurrants and is present in Finland [55,56]. GVBaV is a DNA virus related to RYNV and is transmitted by aphids. It causes gooseberry vein banding disease in Ribes plants and is present in Europe and North America [57].
This study also confirms that siRNA-based diagnostics can be used to detect different viruses in plant samples without previous knowledge about the infecting virus. Recent studies show that VirusDetect reaches a similarly high sensitivity relative to RT-PCR in detection of plant viruses [19]. However, sensitivity might vary depending on the method used for sampling and library preparation, downstream manipulation of RNA and the infection cycles of viruses. That no prior knowledge about viruses that may be present in plant samples is required is an advantage of this method. The method is also relatively fast. This study did raise some questions, such as how to proceed in situations where new, unexpected viruses are detected in the samples. The rules for situations mentioned above are being discussed by the plant inspection authorities [58] who decide the methods used in phytosanitary testing.
In summary, siRNA-based diagnostics revealed only one virus (RYNV) that is putatively endogenous in pre-basic mother plants suggesting that the current virus indexing methods are working as expected. On the other hand, several viruses were detected in accessions of plant genetic resources previously not tested for viruses. This is not unexpected as these plants are maintained in the field. In general, siRNA-based diagnostics proved useful for testing plants for viruses and could be a valuable supplement or replacement of some of the existing methods used for virus detection in pre-basic mother plants for certified plant production, and in the analysis of collections of plant genetic resources.  Table. Plant samples from pre-basic mother plants, breeding lines, and indicator plants from Luke-Laukaa (GEN17 to GEN20), samples from field-grown Rubus plants of the genetic resource collection at Luke-Piikkiö and two samples of field grown Ribes plants (HXR1 and HXR2).