Introduction

Rose (Rosa spp.) is one of the most popular ornamental plants around the world and has been valued and cultivated for millennia [1]. Rose production has a global market value of about 24 billion euros [2]. Rose pathogens, including viruses, affect the economic and aesthetic value of the crop and are considered important threats to the industry. Viruses belonging to numerous taxa have been reported in rose (reviewed in reference 3), and many of these infectious agents have been associated with well-characterized or emerging diseases. During a preliminary study, Lockhart et al. [4] found that rose necrotic mosaic symptoms could be associated with filamentous virus particles 750-800 nm in length, but no further characterization was performed. Viruses in the genus Carlavirus (family Betaflexiviridae) have a positive-stranded RNA genome with six open reading frames (ORFs), including a triple gene block (TGB) involved in virus movement [5]. Members of the genus Carlavirus, whose type species is Carnation latent virus, are mainly transmitted by insects in a non-persistent manner [6, 7]. In this paper, the complete genome sequence of new carlavirus, rose virus A (RVA), and its phylogenetic relationships to other characterized virus members are described. Additionally, two detection assays were developed to screen the University of California-Davis Rose Virus Collection (RVC) for RVA.

In 2018, a memorial rose (R. wichuraiana Crep.) exhibiting leaf deformation and mosaic symptoms was collected for high-throughput sequencing (HTS) analysis. This plant originated from the former USDA-ARS rose breeding program in Beltsville, MD, and was later conserved in a private rosarium in Maryland, USA. Total nucleic acid (TNA) was extracted from symptomatic leaves using a MagMax Plant RNA Isolation Kit (Thermo Fisher Scientific). Subsequently, TNA extracts were used as template for cDNA library construction with a TruSeq Stranded Total RNA with Ribo-Zero Plant Kit (Illumina) as per the manufacturer’s protocol. The cDNA library was sequenced using the Illumina NextSeq 500 platform, yielding approximately 29 million raw reads, which were filtered and trimmed using Illumina bcl2fastq software. De novo contigs were generated from the cleaned Illumina reads using SPAdes v3.13 [8] and compared to viral sequences in the GenBank database using tBLASTx, which identified a total of 28 virus-like contigs. Subsequent analysis revealed that the memorial rose plant was infected with blackberry chlorotic ringspot virus (BCRV) and rose yellow vein virus (RYVV). Furthermore, 12 out of the 28 contigs, ranging between 477 to 8,809 nucleotides (nt), showed a distant relationship (71 to 78% nt sequence identity) to carlaviruses. The longest contig (8,809 nt), with an average coverage of 57X per nt position, was extended in both directions using a SMARTer RACE 5′/3′ Kit (Takara), following the provided instructions.

The full RVA genome was determined to be 8,849 nt long (GenBank no. MN053272), excluding a poly(A) tail. BLASTn comparisons revealed the closest homologous sequence to be poplar mosaic virus (PopMV; genus Carlavirus), with 69% nt sequence identity (26% query coverage). Coding regions present in the RVA genome were identified using ORF Finder [9] and were analyzed using SmartBLAST and the Conserved Domain search tool [10]. The molecular weights of predicted proteins were calculated using the ProtParam software [11]. The RVA genome (Fig. 1) contains six ORFs, which are organized as in other carlaviruses, with 5′ and 3′ untranslated regions of 79 and 78 nt, respectively. In addition, two intergenic regions of 71 and 102 nt were identified. ORF 1 encodes a protein of 1,984 amino acids (aa) (225.1 kDa), which is similar (51% aa sequence identity, 97% query coverage) to the replicase of elm carlavirus (ElmCV) and comprises the following domains: a viral methyltransferase domain (F43-V352), a 2OG-Fe(II) oxygenase superfamily domain (R723-R855), a carlavirus endopeptidase domain (K1004-Y1091), a viral (superfamily 1) RNA helicase domain (G1182-S1298), and an RNA-dependent RNA polymerase domain (P1595-S1973). Downstream of the replicase, ORFs 2 to 4 overlap each other and were predicted to form the TGB. ORF 2 is 235 codons in length, and its translational product (26.1 kDa) is most closely related to the TGB1 of ElmCV (44% aa sequence identity, 99% query coverage) and contains a second viral (superfamily 1) RNA helicase domain (V25-T226). The 110-aa-long product of ORF 3 shares 56% aa sequence identity (97% query coverage) with the carrot virus S TGB2, a predicted viral movement protein (A5-C106) of 12 kDa. The protein encoded by ORF 4 (70 aa, 7 kDa) has the highest similarity to the elderberry virus B TGB3 protein (41% aa sequence identity, 100% query coverage). A coat protein (CP) with a molecular weight of 36 kDa is encoded by ORF 5 (328 aa), based on 42% aa sequence identity (84% query coverage) to the corresponding product of helleborus net necrosis virus; additionally, a flexi CP N superfamily motif (R82-L133) and a flexi CP superfamily motif (N143-E281) were also identified. Lastly, the 122-aa-long protein encoded by ORF 6 is related to a carlavirus putative nucleic-acid-binding protein (13.8 kDa) and shares 49% aa sequence identity (84% query coverage) with the orthologous protein of birch carlavirus.

Fig. 1
figure 1

Genome organization of rose virus A (RVA) showing six predicted open reading frames and their corresponding products. CP, coat protein; NB, nucleic acid binding protein; TGB, triple gene block

Full size image

Phylogenetic analysis was performed using MEGA X [12] by the maximum-likelihood statistical method. Phylograms were generated from an amino acid sequence alignment of the predicted replicase and CP proteins of selected members of the family Betaflexiviridae and RVA. The maximum-likelihood trees were confirmed by bootstrap analysis using 1,000 replicates. Phylogenetic analysis of the replicase (Fig. 2a) placed RVA together with elderberry virus A, PopMV, and other representative members of the genus Carlavirus. A similar grouping was observed in a phylogenetic analysis based on the CP (Fig. 2b).

Fig. 2
figure 2

Phylogenetic relationships of rose virus A (RVA) to recognized members of the family Betaflexiviridae. The viruses used in the analyses and their abbreviations and GenBank accession numbers are as follows: poplar mosaic virus (PopMV), AY505475; elderberry virus A (ElVA), KJ572560; red clover vein mosaic virus (RCVMV), FJ685618; hop mosaic virus (HpMV), EU527979; potato virus M (PVM), NC001361; apple stem pitting virus (ASPV), NC003462; cherry necrotic rusty mottle virus (CNRMV), NC002468; apple stem grooving virus (ASGV), JX080201; carrot Ch virus 1 (CtChV-1), NC025469; citrus leaf blotch virus (CLBV), NC003877; diuris virus A (DiVA), JX173276; apricot vein clearing associated virus (AVCaV), KY132099; potato virus T (PVT), NC011062; apple chlorotic leaf spot virus (ACLSV), M58152; grapevine virus A (GVA), NC003604; watermelon virus A (WVA), KY363796; and tobacco mosaic virus (TMV), NC001367, family Virgaviridae as an outgroup. Maximum-likelihood trees based on the amino acid sequence of the replicase (a) and coat protein (b) are shown. Branch lengths are measured in the number of substitutions per site. Bootstrap values less than 50% are not shown

Full size image

To confirm RVA infection in the source plant, two PCR primer sets were designed. Primers RVA-RDRP-F (5′-GGCGCGACGATTGAAAATCA-3′) and RVA-RDRP-R (5′-TTCGAAGGCCATCACGAACT-3′) target the replicase and generate a 357-bp product. RVA-CP-F (5′-GCCTATGGACCATCTTGTGC-3′) and RVA-CP-R (5′-CGAAAGCAATGGTCTCCTCAT-3′) amplify a 521-bp region in the CP. Both sets of primers were used in one-step reverse transcription-PCR (RT-PCR). The RT-PCR reaction was performed with SuperScript III Reverse Transcriptase (Invitrogen) and GoTaq G2 Flexi DNA Polymerase (Promega), and the thermocycling conditions consisted of 30 min at 52 °C, 35 cycles of 30 s at 94 °C, 45 s at 56 °C, and 1 min at 72 °C, and a final step of 7 min at 72 °C. The memorial rose tested positive for RVA using both the replicase and CP primers. Amplicons were sequenced in both directions by the Sanger method and were 100% identical to the sequences obtained by HTS. To further investigate the distribution of RVA, the University of California-Davis RVC was screened using the RT-PCR assays. RVC has more than 650 rose plants infected by different viruses, such as apple mosaic virus, prunus necrotic ringspot virus, rose spring dwarf-associated virus, and RYVV. However, none of the RVC plants tested positive for RVA using either primer set.

Based on phylogenetics and genomic analysis, we propose that RVA should be classified as a member of the genus Carlavirus. Furthermore, using the current species demarcation criteria for the genus Carlavirus, which include 72% nt sequence identity or 80% aa sequence identity between the CP or replicase genes [13], RVA can be considered a member of a new species. To date, recognized carlaviruses are classified into 53 species (https://talk.ictvonline.org/); however, none of these viruses have been detected in roses. Thus, to our knowledge, RVA represents the first carlavirus identified in rose. Several carlaviruses, however, have been reported in other ornamental plants. Chrysanthemum viruses B and R are pathogens associated with leaf mottling and stunting in chrysanthemum [14, 15]. Lily symptomless virus and verbena latent virus are commonly occurring viruses in lily and verbena, respectively [16, 17].

The symptomatic memorial rose plant that was the original source of RVA was also infected with BCRV and RYVV. Biological studies are needed to identify mechanisms of transmission and assess effects of single and/or mixed RVA infection on rose plants. Limited testing suggests that RVA may not be widespread, because it was not detected in the University of California-Davis RVC, although testing other collections is warranted.