Characterisation of a novel Emaravirus identified in mosaic‐diseased Eurasian aspen (Populus tremula)

Humboldt-Universität zu Berlin, Albrecht Daniel Thaer-Institute for Crop and Animal Sciences, Division Phytomedicine, Berlin, Germany Landesamt für Ländliche Entwicklung, Landwirtschaft und Flurneuordnung, Zossen, Germany Division of Plant Health and Biotechnology, Norwegian Institute of Bioeconomy Research – NIBIO, Ås, Norway Independent Scholar, Marseille, France Natural Resources Institute Finland, Rovaniemi Research Unit, Rovaniemi, Finland


| INTRODUCTION
The objectives of this study were to (a) determine the genome of a presumably novel Emaravirus in Eurasian aspen (Populus tremula L.), (b) develop a reliable detection protocol, (c) provide evidence for association of the virus with the observed disease and (d) evaluate the incidence of the virus in P. tremula exhibiting virus-suspected symptoms in Fennoscandia.
We were observing mosaic, mottle, yellow blotching, variegation and chloroses along veins of leaves of Eurasian aspen since 1991 in Norway and since 2009 in Finland. Generally, whole branch parts or larger areas of the crown were affected and trees in the Rovaniemi area (Kivalo Research Forest) were increasingly showing signs of decline. Comparable symptoms on Eurasian aspen in Finland have first been reported by Bremer, Lehto, and Kurkela (1991). Erroneously, typical "mosaic-disease" symptoms have been considered a special genetic form of Eurasian aspen (Oskarsson & Nikkanen, 1998). This mosaicdisease of Eurasian aspen has also been documented by us since 2015 at additional sites in southern Norway and Sweden (Figure 1).
Few studies have addressed plant viruses affecting poplars (Populus spp.), which may also contribute to observed degeneration of trees (Büttner, von Bargen, Bandte, & Mühlbach, 2013;Nienhaus & Castello, 1989). The most widespread virus in Populus is the poplar mosaic virus (PopMV), which has been detected in different regions of Europe, Asia and the Americas and in various species or their hybrids including Eurasian aspen. The virus is transmissible by mechanical means and is also  (Biddle & Tinsley, 1971;Navratil, 1979). Further, populations of black poplars (Populus × euamericana) in the United Kingdom have been found to contain nematode-transmissible viruses such as arabis mosaic virus (ArMV) and tobacco rattle virus (TRV). From American aspen (P. tremuloides) the soil-borne tobacco necrosis virus (TNV) was isolated from declining trees also showing ring blotches and chlorosis of the leaf veins (Hibben, Reese, & Borzarth, 1979). However, with the exception of PopMV, the viruses could not be confirmed as causal agents of the observed diseases of the poplars studied (Navratil, 1979). Information on the distribution and the host range of these viruses is therefore very incomplete because of the lack of systematic studies on the occurrence and importance of plant viruses affecting poplars.
Some of the Eurasian aspen trees with virus-suspected leaf symptoms surveyed in Finland, Sweden and Norway were also infested by gall mites (Acari, Eriophyidae, Figure 1), which were reported as vectors for several members of the genus Emaravirus (Hassan et al., 2017;Mielke-Ehret & Mühlbach, 2012). Based on these findings, an initial RT-PCR for the genus-specific detection of Emaraviruses (Elbeaino, Whitfield, Sharma, & Digiaro, 2013) was carried out using total RNA from leaf material of Eurasian aspen with mosaic, mottle and chloroses along veins collected in 2015 in Sweden in comparison with material collected from a tree without any disease symptoms. Only from diseased Eurasian aspen a specific fragment was amplified with the generic primers. This suggested that a previously unknown putative Emaravirus is associated with the observed mosaic-disease symptoms in Eurasian aspen.
The genus Emaravirus (family Fimoviridae, order Bunyavirales) contains enveloped plant viruses with a monocistronic, segmented RNA genome of negative polarity, which are transmitted by gall mites (Elbeaino et al., 2018). Since the first description of the type species European mountain ash ringspot-associated Emaravirus (EMARaV) in 2007 (Mielke & Mühlbach, 2007), the genus has significantly increased in numbers of acknowledged species and putative members. Recently, several previously unknown Emaraviruses were identified to be prevalent in different important fruit or desirable ornamental tree species and shrubs, respectively. For instance, they could be associated with a disease in pistachio (Buzkan et al., 2019), blue palo verde tree (Ilyas, Avelar, Schuch, & Brown, 2018), blackberry (Hassan et al., 2017), kiwi (Zheng et al., 2017), jujube tree (Yang et al., 2019) and ti plant (Olmedo-Velarde et al., 2019).

| Determination of complete viral genome segments
Different approaches were applied to determine complete genomic segments of the novel virus from mosaic-diseased Eurasian aspen trees.
Initially, Illumina RNAseq was applied to total RNA extracted from symptomatic leaf samples taken from two diseased trees originating from different sites in Sweden in 2016 (Säffle, E55056) and in 2017 Fossum (E56750) as described in Mielke and Mühlbach (2007), but omitting the phenol/chloroform extraction. The first sample from Säffle was prepared for large scale sequencing (high-throughput sequencing, HTS) including data analyses carried out by the company BaseClear (Leiden, the Netherlands), as described in von Bargen et al. (2019) using the generic PDAP213 primer (Di Bello, Ho, & Tzanetakis, 2015) for preparation of double stranded cDNA. A standard-paired end library (125 bp) was prepared by the sequencing-company BaseClear (the Netherlands) and sequenced on a HiSeq2500 machine (Illumina). For the second HTS sample E56750 RNA was extracted from fresh leaf material collected in 2017. The only other difference was that a random hexamer-primed RNAseq approach was used, in order to allow identification of additional viruses which may contribute to the observed disease. Both datasets only yielded small genome fragments of the assumed novel Emaravirus.
Therefore, full-length RT-PCR for amplification, cloning and Sangersequencing of missing emaraviral genome segments as outlined in von Bargen et al. (2019) was carried out with total RNA leaf-extracts from the same diseased Eurasian aspen in Fossum sampled in 2016 (E55089).
It was not possible to amplify the complete RNA1 of the novel virus by full-length RT-PCR from either of the samples taken in 2016 (E55089) or 2017 (E56750. As leaf material from 2016 was not available anymore, missing sequence information was generated by Sanger-sequencing of PCR amplicons generated from random hexamer primed cDNA from sample E56750 with newly developed primers (Table S1) Raw data obtained by RNAseq were quality trimmed and assembled de novo into contigs and scaffolded with the in-house standard procedure of the sequencing company BaseClear. Delivered scaffolds with a minimum length of 300 nucleotides were compared via blastX (National Center for Biotechnology Information, NCBI) against the non-redundant protein database (NRPROT). By filtering against the taxid 675,845 "Emaravirus" and taxid 10,239 "viruses" respectively, all scaffolds relating to plant viruses were identified and aligned with reference genomes in BioEdit, version 7.0.5.3 (Hall, 1999) using the integrated ClustalW option. Analysis of Sanger-sequencing generated RT-PCR products was also performed in BioEdit as well as alignments of complete genome segments of Emaraviruses. Sequence identity matrices were calculated with the according tool in BioEdit from aligned nucleotide and aa sequences.

| Analyses and comparison of viral genome segments and encoded proteins
ORF finder at NCBI (https://www.ncbi.nlm.nih.gov/orffinder/) was used to find open reading frames (ORFs) and the proteins encoded by the identified viral genome segments. All ORFs with 300 or more nucleotides were considered.
Phylogenetic analyses were conducted using MEGA version X (Kumar, Stecher, Li, Knyaz & Tamura, 2018). Trees were calculated applying the Maximum Likelihood method based on the JTT matrixbased model (Jones, Taylor, & Thornton, 1992) and boot-strapping with n = 1,000 replicates. Branches with values below 50% were collapsed. Initial tree(s) for the heuristic search were obtained automatically by applying Neighbour-Join and BioNJ algorithms to a matrix of pairwise distances estimated using a JTT model, and then selecting the topology with superior log likelihood value. All positions containing gaps and missing data were eliminated.
HHpred was run either against the databases PDB70 or PFAM, or in pairwise mode comparison (option "Align two sequences or MSAs").
Additionally, double-stranded RNA (dsRNA) was isolated according to the protocol developed for EMARaV (Benthack, Mielke, Büttner, & Mühlbach, 2005)  (2019). The integrity of the RNA and absence of PCR inhibitors were evaluated using an internal control according to Menzel, Jelkmann, and Maiss (2002). For specific detection of all genome segments of the novel virus identified in Eurasian aspen, primer pairs were developed within the coding region for each of the five RNAs, according to the scaffolds assembled from the diseased sample tree in Säffle (E55056). PCR conditions for primer pairs for each genome segment of the novel virus were optimised in order to generate specific fragments of the expected size. Amplified PCR products were bidirectionally sequenced and aligned in order to confirm that they originated from the same virus species (data not shown). The same samples were tested by RT-PCR for PopMV infection applying a primer pair established by Werner, Mühlbach, and Büttner (1997).

| Grafting experiment
Total of 276 Eurasian aspen scions were whip grafted or side grafteddepending on the diameter of the grafted scion-to Populus tremula seedlings (nursery Bunk, Germany) in March 2017 in order to confirm the graft-transmissibility of the observed disease. Twenty-eight Eurasian aspen seedlings were grafted with healthy scions from the seedling population, serving as negative controls. Grafted seedlings were cultivated for 2 months in a fogged tunnel, then transferred to the experimental garden and cultivated in 2 L pots for 2 years in peat substrate. The scions used for graft-inoculation originated from a mosaicdiseased Eurasian aspen tree growing in the Kivalo Research Forest (Finland) that tested positive for the novel virus by RT-PCR applying genus specific primers according to Elbeaino et al. (2013) and speciesspecific primers targeting RNA3 and RNA4 of the novel virus. For two consecutive vegetations periods the grafted seedlings were evaluated for virus-suspected symptoms such as mottle, mosaic, development of yellow blotches and/or vein yellowing. Symptomatic leaf material from 53 grafted scions and three graft-inoculated rootstocks exhibiting leaf symptoms in May and June 2018 were tested by RT-PCR targeting the RNA3-RNA5 of the novel identified virus, to confirm its presence in the scions and the rootstocks.

| Geographical distribution of the novel virus in Eurasian aspen
Primer sets targeting the RNA3 and RNA4 that produced a 317 bp long fragment of RNA 3 and a 288 bp long amplicon of the RNA4 pro- The sample E55089 was further analysed to determine the complete sequences of the five genome segments of the novel virus identified in mosaic-diseased Eurasian aspen (Figure 2).
The complete RNA1 was assembled by Sanger sequencing of eight overlapping RT-PCR amplified fragments applying primer sets (Table S1) developed according to the available sequence information from the RNAseq data sets. Additionally, both ends of RNA1 were confirmed by RACE PCR with cDNA-synthesis-primers specific for the 5 0 and the 3 0 terminus respectively (Table S1) (Table 1). The 3 0 end of the genomic RNA2 could be confirmed by RACE PCR as described above.
Cloning and sequencing of the 1.6 kb fragment amplified by fulllength RT-PCR revealed that it contained two different viral genome To verify this, we compared the alignment of P28 and its close homologues identified above with the alignment of the close homologues of EMARaV P4 identified above, by using HHpred in pairwise comparison mode. HHpred reported that their central region is similar, with an E value of 3.6 × 10-8, below the stringent significance cutoff of 10-5, thereby confirming homology. In contrast, we could  The novel virus from Eurasian aspen clustered consistently with PiVB, RRV and PPSMV-2 supporting a closer relation with these viruses.
Furthermore, high bootstrap values clearly demonstrate that AsMaV represents a separate novel species within the genus.
As a result of the observed symptoms in Eurasian aspen, the determined genome structure and phylogenetic relations of the novel virus, we propose the name aspen mosaic-associated virus (AsMaV).
F I G U R E 3 Gel electrophoresis of amplification products after fulllength PCR applying the PDAP213-primer targeting the conserved terminal sequence regions of Emaraviruses according to Di Bello et al. (2015). Positions of emaraviral genome segments amplified from PDAP213-primed cDNA of the AsMaV-infected sample E55089 from Fossum, Sweden as confirmed by Sanger-sequencing of at least three individual clones are indicated at the right side of the gel. At the left side of the gel, unspecific amplification products from an Eurasian aspen sample without leaf symptoms are displayed. M = 1 kb ladder (ThermoScientific) 3.2 | Identification of AsMaV as causal agent of the mosaic-disease affecting P. tremula All five RNAs could be confirmed in leaf material from the mosaicdiseased P. tremula sampled in consecutive years (2016, 2017 and 2018) in Säffle and Fossum by RT-PCR using specific primer pairs targeting each genome segment. Primers were derived from the RNAseq-generated sequences of the virus variant in the Eurasian aspen tree E55056 from Säffle (Table S1).
RNA3-RNA5 were also consistently detected by this RT-PCR in 59 leaf samples of 62 mosaic-diseased Eurasian aspen trees collected from different locations in Norway, Sweden and Finland in the years 2016 and 2017 (Table 2). However, RNA1 and RNA2 specific fragments could only be amplified in 53 leaf samples showing mosaic, mottle, vein chloroses and/or chlorotic ringspots with the respective primer pairs, but RNA1 was also detectable with the generic primer set (Elbeaino et al., 2013) in these samples. Sequencing of selected PCR products confirmed that the mosaic-diseased trees were infected by AsMaV (data not shown). In none of the 22 Eurasian aspen trees with no typical virus-suspected leaf symptoms sampled adjacently to the diseased P. tremula could AsMaV be detected by RT-PCR using RNA1-RNA5 specific primers. These results demonstrate a clear F I G U R E 4 Sequence alignment of the homologues of the P28 protein encoded by RNA5 of AsMaV. Strictly conserved or semi-conserved positions are boxed, and the corresponding amino acid is indicated above the alignment F I G U R E 5 Molecular phylogenetic analysis by Maximum Likelihood method of RdRp protein encoded by RNA1 (a) and N protein encoded by RNA3 (b) of aspen mosaic-associated virus (AsMaV). Groups a to c of Emaraviruses described in Elbeaino et al. (2018) are indicated at the right side of the trees. The evolutionary analyses were conducted in MEGA X (Kumar, Stecher, Li, Knyaz, & Tamura, 2018) and were inferred by using the Maximum Likelihood method based on the JTT matrix-based model (Jones et al., 1992). The percentage of trees in which the associated taxa clustered together is shown next to the branches. Initial tree(s) for the heuristic search were obtained automatically by applying Neighbour-Join and BioNJ algorithms to a matrix of pairwise distances estimated using a JTT model, and then selecting the topology with superior log likelihood value. All positions containing gaps and missing data were eliminated.  (data not shown).
In scions received from an AsMaV-infected P. tremula from the Kivalo Research Forest, Finland, grafted in 2017 onto healthy P.
tremula rootstocks mottle and mosaic symptoms first appeared 2 months after grafting at the end of May in leaves of 17 seedlings. In July of the first vegetation period 70 scions showed mosaic-disease symptoms and AsMaV was confirmed by RT-PCR in 10 symptomatic scions applying RNA3 and RNA5 specific primers (Table S1)

| DISCUSSION
The overall genome structure of the novel virus identified in mosaicdiseased P. tremula is in accordance with the genus Emaravirus (Elbeaino et al., 2018). (a) RNA1-RNA4 encode the viral replicase, structural proteins (GPP and N) of the virus particle and the MP essential for cell-to-cell transport of nucleoprotein complexes. RNA5 encodes a small protein (P28), which is distantly related to other emaraviral proteins (P4-P6, transmission of AsMaV is in general possible, but seems not very effective. This is in accordance with a report from Bremer et al. (1991) stating that their trials to transmit the mosaic and veinbanding symptoms in Eurasian aspen by grafting were not successful at all. However, as natural propagation of Eurasian aspen most often occurs vegetatively via suckers grown from root runners (Caudullo & de Rigo, 2016;MacKenzie, 2010), this mechanism has to be considered an important mode of AsMaV transmission. Effective virus dispersal via vegetative propagation may also explain our observations that many Eurasian aspen trees growing as clonal populations next to each other in different investigated locations and sampled sites (Figure 1) show the characteristic mosaic-disease symptoms and tested positive for the virus. In parallel our investigation of the geographic distribution of AsMaV in Eurasian aspen trees needs to be continued in the future and extended to other regions of the wide natural range of the host plant in Europe and Asia.
Our results also change the knowledge of the assumed prevalent distribution and impact of PopMV in poplar species (Biddle & Tinsley, 1971;Navratil, 1979). We could demonstrate that symptoms such as mosaic, mottle, variegation, yellow leaf blotches and veinal chloroses in leaves of Eurasian aspen are rather caused by an infection with the novel Emaravirus AsMaV and are not associated with PopMV. We could show that the mosaic-disease is widely distributed in Eurasian aspen in Scandinavia and Finland. AsMaV could be detected in many of the affected trees, but not in other investigated diseased poplar species from different European locations. Leaf symptoms especially vein chloroses and yellow blotches, which are similar to the observed mosaic-disease in Eurasian aspen that could be associated with AsMaV in this study, have also been described to occur in North American aspen (P. tremuloides) and the hybrid P. x euamericana (Navratil, 1979). However, no virus could be identified as the causal agent to date. European aspen readily hybridises with P. tremuloides (MacKenzie, 2010) and other aspen species (Caudullo & de Rigo, 2016 Financial support from the European Cooperation in Science and Technology (COST action FA1407 "DIVAS") enabled the scientific collaboration and is also kindly acknowledged.