Plantago asiatica mosaic virus (PlAMV) (genus Potexvirus; family Alphaflexiviridae; order Tymovirales) infects lily plants (Lilium spp.) among other hosts (Komatsu et al. 2017). PlAMV was initially described from Plantago asiatica in Russia (Kostin and Volkov 1976, in: Solovyev et al. 1994). Currently, PlAMV is distributed in several countries in Asia, Europe and the Americas, including Chile (Vidal et al. 2016), China (Xu et al. 2017), Hungary (Pájtli et al. 2015), Italy (Parrella et al. 2015), Japan (Ozeki et al. 2006), Korea (Lim et al. 2016), and the USA (Hammond et al. 2015), and causing up to 80% losses in The Netherlands (Anonymous 2010). In addition, a virus infecting heavenly bamboo (Nandina domestica) in California, named Nandina mosaic virus (Zettler et al. 1980), is now recognized as a strain of PlAMV (Hughes et al. 2005).

Reports suggest that PlAMV is disseminated internationally by trading of infected bulbs (Hammond et al. 2015; Pájtli et al. 2015; Parrella et al. 2015). Moreover, PlAMV is mechanically transmitted (by sap) and no arthropod vector is known. The virus naturally infects hosts belonging to diverse plant families including Plantaginaceae (plantago), Liliaceae (lily), Primulaceae (primrose), Violaceae (violet), Scrophulariaceae (Chinese foxglove), and Berberidaceae (nandina) (Kostin and Volkov 1976, in: Solovyev et al. 1994; Komatsu et al. 2008, 2017; Kwon et al. 2017); and experimental hosts from additional plant families (Kostin and Volkov 1976; J. Hammond, unpublished).

During a survey to detect viruses belonging to the genus Orthotospovirus (family Tospoviridae; order Bunyavirales) (Adams et al. 2017) in Costa Rica, two Asiatic lily hybrid plants were collected at a greenhouse in Cartago Province on June 2013 (samples T188 and T189). The symptoms included brown necrotic streaking and irregular leaf chlorosis (Fig. 1a, b). RNA was extracted (RNeasy Plant Mini Kit, Qiagen, Germany) and cDNA was obtained by retrotranscription (Maxima H Minus First Strand cDNA Synthesis Kit, Thermo Scientific, Lithuania) primed with random hexamers. cDNAs were tested by PCR (DreamTaq PCR Master Mix, Thermo Scientific, Lithuania) with five different primer pairs for orthotospovirus detection (Eiras et al. 2001; Uga and Tsuda 2005). An amplicon under 500 bp was obtained for both plants, T188 and T189, using primer pair IYSV-837 / Tos-R15 (Uga and Tsuda 2005) and sequenced (Macrogen, Korea) in both directions. The resulting sequences (T188, KY807780; T189, KY807781) had top BLAST hits in GenBank (National Center for Biotechnology Information, USA) to isolates of PlAMV. The unexpected amplification of PlAMV sequence using primers designed to detect tospoviruses was investigated by aligning the sequences of primers IYSV-837 and Tos-R15 against the PlAMV genome. All alignments identified included multiple mismatches, particularly with primer Tos-R15, which has five two-fold degenerate positions near the 3′ end, and potential to prime on both strands of the PlAMV genome. The initial detection of PlAMV using ‘tospovirus-specific’ primers was verified by testing with primer pairs for general potexvirus detection (Potex 4 / Potex 5, Miglino et al. 2006) and for specific detection of PlAMV (PlAMV CP-F2 / CP-R3, Hammond et al. 2015). As negative controls, four symptomless lily plants (two pots, 2 plants per pot) were purchased at a supermarket in San José Province. RNA was extracted from each plant (L1 to L4) and processed as previously described. As a positive control for potexvirus detection, RNA was extracted from a Phaius tankervilliae plant infected with Cymbidium mosaic virus (CymMV) (genus Potexvirus) (Moreira et al. 1998). The two greenhouse lily plants (T188 and T189) and one of the supermarket (symptomless) plants (L3) yielded positive results with both generic potexvirus, and PlAMV-specific, primer pairs (Table 1).

Fig. 1
figure 1

Detection and pathogenicity of Plantago asiatica mosaic virus (PlAMV) in Costa Rica. Symptoms of PlAMV observed on a Lilium sp. plant (T188) (a, b). Transmission electron microscopy (TEM) image from sample T188 leaf tissue with several flexuous rods in the cytoplasm of a cell (c); scale bar represent 500 nm. Symptoms observed on Nicotiana benthamiana after mechanical inoculation with sample T189, 22 days after inoculation (dai) (d), with sample T188, 36 dai (e), and buffer-inoculated plant, 22 dai (f)

Full size image
Table 1 Plantago asiatica mosaic virus (PlAMV) detectiona in lily samples in Costa Rica
Full size table
Table 2 Pairwise nucleotide comparisonsa (estimates of evolutionary divergence between sequences) for 24 Plantago asiatica mosaic virus isolates and one isolate each of tulip virus x and Cymbidium mosaic virus
Full size table

Small leaf segments (1 × 2 mm) were harvested around chlorotic and necrotic lesions on leaves of one greenhouse lily sample (T188) and were processed for transmission electron microscopy (TEM) as previously described (Montero-Astúa et al. 2017). We detected individual flexuous-rod shaped particles scattered in the cytoplasm (Fig. 1c) similar to those associated with potexviruses (Purcifull et al. 1966). Particle average length was measured from micrographs of tissue sections using ImageJ software (Schneider et al. 2012), selecting only those particles that were clearly visible and not in aggregates. Average particle length was estimated at 516 nm with a 95% confidence interval 499–533 nm (n = 60), consistent with PlAMV particle size (490 to 530 nm, Kostin and Volkov 1976, in: Solovyev et al. 1994).

The virus was mechanically inoculated to Nicotiana benthamiana plants and symptoms were observed. Frozen leaf tissue (stored at −35 °C) from each greenhouse lily sample (T188 and T189) was ground in ice-cold 100 mM Na2SO3 using chilled mortars and pestles. The extracts were rub-inoculated with carborundum and cotton swabs on two-week old N. benthamiana plants (four plants per sample). No symptoms were visible at 15 days after inoculation (dai). A mild mottle was observed at 22 dai (Fig. 1d) and young leaf corrugation and severe chlorotic patterns were observed 36 dai (Fig. 1e) on three and four plants inoculated with T188 and T189 extracts, respectively. One mock inoculated plant (buffer) presented no symptoms (Fig. 1f).

Leaf tissue from the two greenhouse lily samples (T188 and T189) and from the corresponding mechanically inoculated N. benthamiana plants tested positive by DAS-ELISA with commercial antibodies against the Nandina mosaic strain of PlAMV (Agdia Inc., Indiana, USA). Moreover, RT-PCR and sequencing (as described previously) from inoculated N. benthamiana tissue yielded positive results and sequences corresponding to PlAMV (Table 1). Virus partial sequences obtained from inoculated N. benthamiana plants (GenBank accession numbers: KY807783 and KY807785) had the same SNPs as the corresponding source virus sequences (KY807782 and KY807784). Therefore, we confirmed the identity of the virus and its transmission and pathogenicity to N. benthamiana. Additionally, we obtained for the first time sequence confirmation (KY807788) for CymMV from Costa Rica (potexvirus positive control in P. tankervilliae, Moreira et al. 1998). The nucleotide sequences obtained for PlAMV and CymMV are available in GenBank under accession numbers KY807776-KY807788.

The partial sequences obtained with primer pairs Potex 4 / Potex 5 (Miglino et al. 2006) and CP-F2 / CP-R3 (Hammond et al. 2015) for samples T188, T189 and L3 were aligned with corresponding sequences from available PlAMV genomes from GenBank (Table 2) using MUSCLE (Edgar 2004) algorithm in MEGA7 (Kumar et al. 2016). The species Tulip virus X (Yamaji et al. 2001) and Cymbidium mosaic virus (Wong et al. 1997) were included as outgroups. The two sequence alignments: (i) a fragment of the polymerase ORF (3350–3616, nucleotide positions from Russian reference genome sequence NC003849) and (ii) a fragment of the coat protein ORF (5664–6059) were concatenated with FaBox (Villesen 2007). Phylogenetic analysis was done using a maximum likelihood method with a Hasegawa-Kishino-Yano model (HKY) and a gamma-shaped rate variation across sites with a proportion of invariable sites (G + I) and 5000 permutations. A Japanese isolate of tulip virus X (TVX) grouped closely with isolates of PlAMV as previously reported (Yamaji et al. 2001). Three putative PlAMV groups were indicated by the dendrogram (Fig. 2). Two clades are supported by a 96% bootstrap value: one group includes the three lily isolates from Costa Rica, plus isolates from China, Hungary, Korea and The Netherlands; a second clade includes Japanese isolates from lily and one from primrose (Pr, Primula sieboldii). A third loose group includes isolates from N. domestica (USA and Japan) and from P. asiatica (Russia and Korea). Two Japanese isolates, one from N. domestica (NJ) and another from Viola grypoceras (Vi) were separated from other Japanese isolates in accordance with a previous report (Komatsu et al. 2017). Komatsu et al. (2008) indicated that the species Plantago asiatica mosaic virus shows high biological and genomic variability. Pairwise nucleotide comparisons (calculated with MEGA7, Table 2) for the concatenate sequence dataset of 24 PlAMV isolates used herein has a mode of 0.141. The highest distance (0.221 ± 0.017) was found between two Japanese PlAMV isolates; one from viola (LC155795) and another from nandina (LC155795). Costa Rican isolates were very similar to one another (0.003 ± 0.002, or 0.007 ± 0.003), and showed the highest distance (0.204 ± 0.016, or 0.206 ± 0.016) to the Japanese PlAMV isolate from viola (LC155795), with the lowest distance values (0.008 ± 0.004) to three Korean lily isolates (KU159089, KU159090, and KU159091).

Fig. 2
figure 2

Phylogenetic tree constructed with Plantago asiatica mosaic virus (PlAMV) sequences, each corresponding to two concatenated fragments of the virus genome, including a partial sequence of the RdRp ORF and a partial sequence of the coat protein ORF adding to a final alignment of 689 nucleotide positions with gaps. Each sequence code indicates GenBank accession number, virus species, (isolate, when available) and geographical origin (three letter country code). Sequences were aligned with MUSCLE algorithm in MEGA7 and phylogeny inferred with the maximum likelihood method using a Hasegawa-Kishino-Yano model (HKY) with G + I variation rate across sites and 5000 bootstrap replicates. Scale bar represents 0.05 substitutions per site

Full size image

To our knowledge, this is the first report of PlAMV in Costa Rica. We hypothesize the virus was introduced with imported lily bulbs as has been indicated for other countries (Hammond et al. 2015; Pájtli et al. 2015; Parrella et al. 2015; Vidal et al. 2016; Xu et al. 2017). Symptomless infection of lily was noted in Chile (Vidal et al. 2016) and China (Xu et al. 2017). We also detected a symptomless infection in a commercialized lily plant increasing the risk for unobserved movement and transmission of the disease. These results suggested the need for enhanced virus testing schemes for bulbs and plants to avoid virus dissemination worldwide and emergence of novel diseases due to new virus-host encounters.