Raspberry bushy dwarf virus in Slovenia - geographic distribution, genetic diversity and population structure

Raspberry bushy dwarf virus (RBDV) is a long-known virus naturally infecting Rubus and grapevine. It is also one of the economically most important viruses of raspberries, but there are only a limited number of sequences covering a substantial part of the genome available in the databases. The aim of this study was: i) to study the geographic distribution of RBDV in Slovenia, and ii) to sequence RNA2 of several red raspberry and grapevine RBDV isolates and study their phylogeny and population structure. Geographic distribution studies were performed over a period of 13 years in three wine-growing regions of Slovenia (Primorska, Podravje and Posavje). The highest incidence of RBDV was found in Podravje (58.8%) and the lowest in Primorska (5.1%). Big differences were observed between Vipavska dolina (10.2%) and three other wine-growing districts of Primorska region (0.4–1.2%). Almost complete RNA2 sequences were obtained for four red raspberry isolates and seven grapevine isolates. Additionally, only coat protein sequences were obtained for three red raspberry isolates. Phylogenetic and population diversity analyses were performed on all available RBDV sequences. Phylogenetic analysis has shown clear differences in sequences from Rubus and grapevine that form two highly supported clades. In RNA2 analysis additional two sub-clades were found in grapevine clade. Two major subclades were identified also in the Rubus clade with further differentiation within these subclades. Purifying or stabilizing selection was found to be acting on both, CP and MP genes while few codons were found to be under positive selection.

The genome of RBDV is bipartite single-stranded RNA encapsidated in quasi-isometric particles. RNA1 contains one large open reading frame (ORF) encoding a 188 kDa protein with motifs of viral RNA helicases and polymerases (Ziegler et al. 1992). RNA2 has two ORFs, the one at the 5′-end encodes a 39 kDa movement protein (MP) and the one at the 3′-end a 30 kDa coat protein (CP) (Natsuaki et al. 1991). Parts of the genome were sequenced for RBDV isolates originating from different countries but there is still a very limited number of sequences available in the database.
Studies of RBDV in Slovenia started in 2002 when the virus was first detected in red raspberries from collection plantation of Agricultural Institute of Slovenia. In 2003, the RBDV infection was confirmed in grapevine grafts of cvs. Laški Rizling (Italian Riesling) a n d Š t a j e r s k a B e l i n a b y D A S -E L I S A a n d immunocapture RT-PCR and a few grapevine and red raspberry isolates were further characterized (Mavrič et al. 2003;Mavrič et al. 2004;Mavrič Pleško et al. 2009). RNA2 sequences used for molecular characterization of these isolates are some of the few RBDV RNA2 sequences available.
In this study the presence of RBDV was surveyed in grapevine in three wine-growing regions of Slovenia during the period of 13 years. Selected grapevine and red raspberry samples were used for molecular characterization and phylogenetic analysis of RBDV. Additionally, the population structure of RBDV was studied using available CP, MP and RNA2 sequences.
Over 30 grapevine cultivars collected in all winegrowing regions were analyzed. The largest number of samples was collected in Italian Riesling, Refošk, Merlot, Malvazija, Chardonnay and Pinot Gris. The highest infection rate was detected in Chardonnay, Rizvanec, Šipon, and Zweigelt (80% or higher) followed by Rhein Riesling (72.3%) and Italian Riesling (53.5%). Most of the RBDV positive samples from Primorska were collected in Vipavska dolina (Table 1) where the infection was confirmed in Italian Riesling (59%), Sauvignon (38.7%) and Malvazija (2.6%). In Posavje, the infected cultivars were Italian Riesling and Rhein Riesling (27.1% and 59.1% respectively) while in Podravje the majority of tested cultivars were found to be infected. The widespread infections with RBDV in Podravje indicate that the infection in Slovenia might originate in this region and had spread to other regions through infected planting material of cultivars grown in all regions. This would explain the high infection rate in Vipavska dolina in comparison with the rest of Primorska since Italian Riesling is a predominant cultivar of Slovenia but in the Primorska region it is almost exclusively grown in Vipavska dolina. In spite of importance of RBDV in raspberry production, there are only eight whole genome sequences and an additional seven sequences of complete or almost complete RNA2 available in the databases. Obtaining additional RNA2 sequences of RBDV from grapevine and red raspberry was the main aim of this study that allowed for better understanding of its diversity. Several grapevine and red raspberry samples, confirmed to be infected with RBDV by DAS-ELISA, were selected for diversity study. The selected grapevine samples were from different wine-growing regions while raspberry samples all originated from collection plantation of Agricultural Institute of Slovenia at Brdo pri Lukovici in central Slovenia. For the purpose of this study, all samples used for diversity study are referred to as isolates. The IC RT-PCR or total RNA extraction and RT-PCR were used to confirm RBDV infection of seven grapevine and seven red raspberry samples and to amplify two PCR products covering almost complete RNA2 of RBDV. The IC RT-PCR, RT-PCR, cloning and sequencing were performed as previously described (Mavrič Pleško et al. 2009). Sequences were analysed using Geneious version 8.1.8 and 11.1.4 (http://www. geneious.com, Kearse et al. 2012) and assembled into contigs. Almost complete RNA2 sequences (including complete ORFs for CP and MP) were obtained for seven grapevine and four red raspberry isolates and only CP sequences were obtained for three red raspberry isolates. Sequences were deposited in the GenBank under Accession Numbers KY417868 KY417881.
Phylogenetic relationships between the isolates were inferred from UPGMA clustering method. Trees were constructed from (1) near complete RNA2 alignment of 27 sequences with 2124 nucleotides (Fig. 2), and (2) CP gene alignment from 41 RBDV sequences with 822 nucleotides (Fig. 3); 14 obtained from this study and 27 from Genbank database (Table 2). Data were bootstrapped with 1000 re-samplings to test the robustness of the lineages in the trees. Analyses were performed using Tamura 3-parameter method implemented in MEGAX software (Kumar et al. 2018). Four sequences were used as an outgroup: DQ120126 (RBDV . Sequences from raspberry and grapevine formed two distinct and highly supported clades (> 92%) in RNA2-based and CP-based phylogenetic trees (Figs. 2 and 3). All grapevine sequences clustered together in one clade. Lower isolate diversity was observed within grapevine clade of CP gene. Overall, no correlation between white and red cultivar isolates was noticed. The clade of RNA2 sequences from raspberry showed higher internal clustering with two major sub-clades, one with Slovenian and Ecuador sequences (93%), and the other with Belarus and UK sequences (100%). Further differentiation was observed also within these subclades. Higher diversity was also observed in phylogenetic tree of raspberry CP gene.
It has already been demonstrated that grapevine and raspberry RBDV sequences differ genetically and phylogenetically from each other. The distinction was first observed after the discovery of RBDV in grapevine (Mavric Plesko et al. 2009) and also by Valasevich et al. (2011) when RBDV raspberry sequences from Belarus and Sweden were compared to all available RBDV sequences. Phylogenetic analysis of RNA2 sequences performed in the study of Valasevich et al. (2011) confirmed clear differentiation between grapevine and raspberry sequences. Further differentiation was observed within raspberry clade. Analysis performed in our study confirmed this differentiation with additional sequences in each of the two clades. No obvious geographical differentiation could be observed from phylogenetic analysis, however, considering the low coverage of countries or geographical regions, this was understandable.
The selection pressure working on CP and MP genes from the RBDV populations was studied based on the d N /d S ratio calculated from the average number of nonsynonymous substitutions per non-synonymous site (d N ), and the average number of synonymous substitutions per synonymous site (d S ) using algorithms implemented in DnaSP v6 (Rozas et al. 2017). Nucleotide diversity of CP and MP sequences was similar, but in both cases lower in grapevine (GR) ( Table 3). The ratio Cumulative behaviour plots (Fig. 4) for synonymous and non-synonymous substitutions along the coding regions were constructed using the 'xyplot' option of Fig. 3 Predicted relationships between RBDV sequences originating from Rubus spp. and grapevine (Vitis sp.), based on the nucleotide sequences of CP gene (822 nt). Horizontal lines are in proportion to the number of nucleotide differences between branch nodes. Numbers at branch nodes represent bootstrap support of 1000 replicates, but only values higher than 50% are shown. Sequences in bold were obtained during this study  Korber 2000). They indicated that nonsynonymous substitutions (red line) are distributed equally throughout the length of the MP (Fig. 4a) and CP (Fig. 4b) coding region. Generally, the number of synonymous substitutions over non-synonymous substitutions is greater in protein-coding regions. However, in the case of MP (Fig. 4a), the synonymous substitutions showed a biphasic distribution with its rate being lower than the rate of non-synonymous in the codon region of 1 to 106 codons. Both GR and RU sequences c o n t r i b u t e d t o o b s e r v e d b i p h a s i c distribution/phenomenon (data not shown). However, a few codons were found to be under positive selection.
Testing for selective pressures operating on MP and CP genes was performed using an online tool Datamonkey Adaptive Evolution Server (Delport et al. 2010). The analyses showed that 0.16-0.21% of codons were under a negative (purifying) selection. The FEL analysis (Fixed Effects Likelihood) revealed a very small proportion of codons under positive selection in MP gene (0.006%) and none in CP gene (Table 4). In addition, we used MEME (Mixed Effects Model of Evolution) as it was shown to have greater resolving  protein sequences of RBDV isolates, as determined by SNAP program, is plotted along the coding region. The analysed coding region consisted of 358 codons for MP and 274 codons for CP gene sequences. Outgroup sequence DQ120126 was not included in the analysis power than FEL (Murrell et al. 2012). Test revealed that five codons for MP and two for CP gene could be under a positive selection, which represents 0.014-0.007% of codons in total. The majority of codons are therefore under neutral selection, which agrees with the calculated d N /d S ratio (Table 3). The sites of episodic diversifying selection that could be mapped to specific lineages are summarized in Table 4. The positively selected sites were found to be present in RBDV lineages originating from Rubus, and less from grapevine. A rapid screening for recombination events was performed using GARD (Kosakovsky Pond et al. 2006), a recombination detection method implemented in Datamonkey server (Delport et al. 2010) and RDP4 program using default parameters (Martin et al. 2015). We found no evidence for recombination events. Sequence divergence, genetic differentiation and gene flow were estimated by algorithms implemented in DnaSP v6 (Rozas et al. 2017). To estimate genetic differentiation, four subpopulation groups were created: (1) GR group with sequences from grapevine, (2) RU group with sequences from Rubus, (3) RU-SLO with Slovenian Rubus sequences only, and (4) RU-OTH group with Rubus sequences not originating from Slovenia. Genetic differentiation between groups was estimated with statistics K st *, Z*, H st , and S nn subjected to permutation tests (Hudson et al. 1992;Hudson 2000). The DnaSP tool was used also for estimating gene flow (F ST ) between populations (Hudson et al. 1992). The comparison of GR and RU groups revealed strong genetic differences in both CP and MP genes, as evidenced by statistically significant values of all statistics for detecting genetic differentiation (Table 5). Hudson's S nn indicated evidence for differentiation between GR and RU, and also between Slovenian Rubus and other Rubus isolates. Hudson's F ST also indicated the presence of inter-population diversity between abovementioned groups. The F ST value of 0.371 for CP group RU-SLO vs. RU-OTH is close to the threshold value of 0.33 (Rozas et al. 2003). This might indicate more frequent gene flow between population groups RU-SLO and RU-OTH as opposed to infrequent gene flow among all other groups. However, similar situation was not detected for MP gene.
In conclusion, our survey confirmed the presence of RBDV in grapevine in all wine-growing regions of  a Pi(tot): total estimate of nucleotide diversity between subgroups; b dN/dS: indicator of selective pressure operating on protein-coding genes (1neutral; >1positive; <1purifying); c K st *, Z*, H st , S nnstatistics with permutation tests for detecting genetic differentiation between subpopulations (*, P < 0.05; **, P < 0.01; ***, P < 0.001); d F ST -Wright's fixation index for quantification of the genetic differentiation (F ST > 0.33 suggests infrequent gene flow) Slovenia with the highest incidence in Podravje and the lowest in Primorska. Several economically important cultivars were found to be infected. The knowledge about RBDV infection in grapevine is still very limited. It has been reported from Slovenia, Hungary and Serbia (Mavrič Pleško et al. 2009;Mavrič Pleško et al. 2012;Czotter et al. 2018;Jevremović and Paunović 2011). However, we would expect that its distribution is wider than currently known. The present study is an important contribution to the knowledge about RBDV variability, especially for grapevine isolates. The study considerably raised the number of available sequences in the databases which may help to improve the available molecular tests for RBDV detection in research and diagnostics. The phylogenetic analysis of RBDV sequences confirmed the host-based differentiation of RBDV isolates. Further differentiation into subclades was observed within grapevine and within Rubus isolates. The data indicate the possible greater variability of RBDV than previously thought. No recombination events were detected within CP and MP genes.