Rubus yellow net virus (RYNV) is a member of the genus Badnavirus (family Caulimoviridae) [1,2] and has been detected in several European countries and the Americas (USA, Canada and Chile) [3,4]. RYNV host range is, to date, restricted to Rubus spp. [5]. The most common symptoms elicited by RYNV on susceptible genotypes are net-like chlorosis along the veins, giving the plant a pale green appearance with some leaves being slightly cupped downwards. While there may be distortion or stunting in some, most genotypes remain symptomless [6]. Synergistic interactions with black raspberry necrosis virus and raspberry leaf mottle virus cause raspberry mosaic disease (RMD) [7], possibly the most devastating virus disease of raspberry [4]. In the field, RYNV spreads by the European large raspberry aphid, Amphorophora idaei Börner in Europe and the large raspberry aphid, A. agathonica Hottes, in North America, most likely in a semi-persistent manner [6]. Moreover, these vectors are also able to transmit the other two viruses in the RMD complex [8].
As other members of the genus Badnavirus, RYNV may be found in two forms: endogenous - integrated in the host genome or episomal - forming virions [9]. The endogenous pararetroviruses (EPRVs) are inactive and do not cause symptoms, yet they cause major issues in international plant movement and trade given that plants test positives in PCR assays [9–11]. However, there is evidence that, at least in the case of banana streak virus, integrated forms are able to reactivate and form virions, initiating replication and cause disease [12].
There are only two full length RYNV sequences available in GenBank (as of January 2022), both from raspberry. RYNV-Ca from Canada [2] and RYNV-BS, an isolate that likely originated in Europe but was characterized in the United States [13]. Here we describe the first RYNV episomal genome from blackberry (Rubus fruticosus L.; cultivar Jumbo) collected in Bosnia and Herzegovina (BiH).
During the 2018 growing season, asymptomatic blackberry samples were collected from orchards in Banja Luka, Bugojno and Teslic. Total nucleic acids (TNA) from 28 samples were extracted and reverse transcribed using random primers and the Maxima® reverse transcriptase essentially as described in Poudel et al. [14] protocol. The reverse transcription step was performed to increase detection by also targeting the RNA intermediate form and mRNA transcripts of the virus. cDNA quality was evaluated using the NADH dehydrogenase ND-2 subunit (ndhB gene) transcript as an internal control [15]. Samples were screened with primers RYNV6F/RYNV6R [16] and DDF/DDR (this study) (Supplementary Table 1) that amplify a 463 bp and 203bp of the genome respectively. Two samples, K20 and K26 tested positive and were further assayed to determine whether these were infected by the episomal or endogenous forms of the virus. To remove host genome and any integrated viral forms, samples were digested by TurboTM DNase (ThermoFisher Scientific, USA) for 30 min at 37°. Successful digestion was assessed using the ndhB control. The primers can amplify the genomic copy that incorporates an intron and give a band of ~1300bp [15] that easily differentiates from the transcripts of ~ 720 bp. Only a 720 bp band was amplified after digestion (data not shown). Both samples were retested with RT-PCR (Fig. 1). K26 was chosen for further analysis, including rolling circle amplification (RCA), which preferentially amplifies circular DNA using the Illustra TempliPhi 500 Amplification Kit (GE Healthcare, Buckinghamshire, UK) [17]. Back-to-back primers RYNVB2BF/RYNVB2BR were designed within the DDF/DDR screening region were used on the RCA product; yielding a product of approximately 8000 bp (Fig. 1). Using RYNV-BS as a reference, primers were designed across ~1kb overlapping regions to amplify the full genome (Fig. 1; Supplementary Table 1). Each section was amplified and sequenced at least three times, and the consensus genome was deposited in Genbank under accession MZ358192.
The blackberry isolate was compared against the only two other, fully sequenced isolates, both from raspberry. Sequences were aligned using MUSCLE in MEGA 5.2 [18]. ORFinder was used to identify ORFs (http://www.ncbi.nlm.nih.gov/projects/gorf) and the Sequence Demarcation Tool Version 1.2 (SDTv1.2) was applied to assess pairwise identity [19]. Each ORF was analyzed individually to determine the selection pressure using FUBAR [20] within Selection Map version1 [21].
The RYNV-BiH is 7,816 nucleotides (nt) long and shares 82% and 97% full genome identity with the RYNV-Ca and RYNV-BS respectively (Table 1). For this analysis we follow the RYNV-Ca ORF designation of as was the first published. A single nucleotide insertion in RYNV-BiH disrupts the presumed beginning of ORF1, resulting in a premature stop codon. Badnaviruses may utilize alternative start codons [22–24], therefore, RYNV-BiH was reanalyzed with ORFinder allowing for alternative start codons and a UUG start codon was identified. Evidence for the alternative start codon is the fact that the presumed ORF shares 92-98% pairwise amino acid (aa) identity to its raspberry counterparts. The function of ORF 1 and downstream ORF 2 have not been characterized in depth and their function is unknown. All isolates contain a large polyprotein (ORF 3) [2,13] with all essential replication-associated motifs including ribonuclease H and reverse transcriptase, pepsin-like aspartate protease and a zinc binding motif. Both ORF 4 and -6 code for putative proteins with nuclear export signals [13]. RYNV-BiH and RYNV-BS have five ORFs whereas RYNV-CA encodes seven (additional ORFs 5 and 7) (Table 1).
RYNV–BiH and RYNV-BS share higher % pairwise identities when compared with RYNV-Ca. RYNV-Ca ORF 4 and ORF 6 have low identity of 74-79% nt and 57-67% aa with the other isolates. Interestingly, RYNV-BiH and RYNV-BS also show lower aa (90%) than nt identity (96%) in ORF 6, with nearly all mutations resulting in aa changes, even though there are codon degeneracies, suggesting positive selection for those positions. ORF 2 also shows lower aa identities compared to nt between RYNV-BiH and RYNV-BS as RYNV-BS is missing two nt near the end of the ORF which changes all the preceding aa.
FUBAR was used to identify evidence of selection for each ORF. Only ORF 6 has a site identified as evolving under positive selection. ORF 6 also has a single site which shows evidence of negative selection; overall ORF 6 is evolving under positive selection with an average dN/dS = 1.248 where dN are the non-synonymous substitutions per nonsynonymous site whereas dS as the synonymous substitutions per synonymous site. The other ORFs are all evolving under negative selection, with a number of sites showing evidence of negative selection; ORF 1 (10 sites) with an average dN/dS of 0.28, ORF 2 (9 sites) and dN/dS of 0.35, ORF 4 (6 sites) and dN/dS of 0.40 and ORF 3 which has a large number of sites spread across the whole ORF, with average dN/dS of 0.10 (Supplementary Fig. 1).
Herein we present the first RYNV episomal genome from blackberry whereas this is first report of the virus in Bosnia and Herzegovina. This isolate is most similar to RYNV-BS, an isolate of possible European origin, whereas it is quite divergent when compared to the RYNV-Ca. This may be because of the geographic isolation of the isolates or because as indicated by Ho et al. [11]. RYNV-Ca may be an endogenous form which integrated into the raspberry genome more than 150 years ago. While sequence availability is scarce and more genome data is needed to fully explore this virus, we hypothesize that the host (raspberry vs blackberry) does not significantly affects virus fitness. This is particularly interesting if we take into account that the RYNV-BS isolate is likely at a genetic equilibrium, as it is infecting the plant assayed positive for the virus for more than 20 years, whereas the blackberries sampled in BiH have been in the ground for 5 years at the time of sampling. This work further expands our understanding of RYNV diversity and host range.