Plant pararetroviral sequences in wild Dahlia species in their natural habitats in Mexican mountain ranges




Selected wild Dahlia species in their natural habitats from west-central Mexico were tested for the presence of three caulimoviruses known to be associated with cultivated dahlia (Dahlia variabilis), viz. Dahlia mosaic virus (DMV), DMV-D10 and Dahlia common mosaic virus. Virus species-specific primers and PCR were used followed by cloning and sequencing of the amplicons. Results showed that the wild dahlia species in their natural habitat contained DMV-D10, which is an endogenous plant pararetrovirus. Viral sequences were found in 91% of the samples (n = 56) representing four different wild species. The gene coding for the movement protein of DMV-D10 from Dahlia coccinea and all other species was cloned and sequenced. Sequence comparisons showed divergence of this gene when compared to that of DMV-D10 from cultivated dahlias. The discovery of plant pararetroviruses in wild dahlia species in their natural habitats suggests a possible emergence, co-existence and co-evolution of pararetroviruses and their host plants.


Dahlia (Dahlia variabilis) is an economically important ornamental crop in the USA and several other parts of the world. The genus Dahlia, whose vernacular name is also ‘dahlia’, is in the family, Asteraceae. Currently, 35 ‘wild’ species are recognized (Saar & Sørensen, 2005), plus the cultivated forms, known as either D. pinnata or D. variabilis. There are over 2000 named cultivars of D. variabilis in the USA. The natural geographic range for the genus Dahlia includes the natural Sierra Madre Occidental region of Mexico (Fig. 1). Many of the Dahlia spp. have very limited ranges; several are known from only one or two populations (Saar, 1999). Exceptions are D. australis, which occurs at least as far south as southwestern Guatemala, and D. coccinea and D. imperialis, which have been reported throughout Central America into northern South America. Plant size varies from the small D. tenuis and D. scapigera, which only average 30–60 cm in height on slender stems, to tall arborescent species such as D. tenuicaulis, to D. macdougallii, which grows from under the mosses and ferns covering tree trunks and produces long shoots that sprawl across the canopy branches of tropical hardwoods.

Figure 1.

 Map of Mexico showing the occurrence of the four wild Dahlia spp. Samples were collected from Nayarit, Sinaloa, Chihuahua and Durango states. Modified from

More than a dozen viruses have been found to infect dahlia (Albouy, 1995). All but one are RNA viruses. Dahlia mosaic virus (DMV), a DNA virus, is widely prevalent in cultivated dahlias and causes an economically important disease referred to as dahlia mosaic (Brunt, 1971; Albouy, 1995). Based on molecular characterization, DMV is considered a distinct species in the genus Caulimovirus (Richins & Shepherd, 1983). The most characteristic symptoms of the dahlia mosaic disease include mosaic, vein clearing, vein-banding, leaf distortion, systemic chlorosis, flower breaking, and overall stunting of the plant. Symptoms vary depending on the cultivar, and cultivars with few or no symptoms are common.

In addition to DMV, it was recently shown that there are at least two more distinct caulimoviruses associated with dahlia. Tentatively referred to as Dahlia common mosaic virus (DCMV) and DMV-D10 (Pahalawatta et al., 2007; Pappu et al., 2008), these two new caulimoviruses showed significant sequence divergence compared to that of DMV (Pahalawatta et al., 2008b; Pappu et al., 2008). Both DCMV and DMV-D10 were found to be widespread based on surveys conducted in the USA and Europe (Pappu et al., 2005; Pahalawatta et al., 2007; Eid et al., 2009).

Of the three caulimoviruses found in dahlia, there is evidence that DMV-D10 is integrated into the dahlia genome and thus exists as an endogenous plant pararetrovirus (EPRV) (Pahalawatta et al., 2008a). Whilst much is known about plant pararetroviruses in cultivated plant species, their incidence in natural habitats of wild species not influenced by modern agricultural practices remains to be investigated. Using wild Dahlia collected in their native habitats as a model system to study plant–pararetrovirus ecology and evolution could provide insights into the evolutionary pathways, and relationships of pararetroviral sequences with their host plants. To this end, this study investigated the presence of caulimoviruses in Dahlia species in west-central Mexico and used molecular techniques to determine their phylogenetic relationships.

Materials and methods

Sampling site and procedure

Wild dahlia species were collected from the slopes of the Sierra Madre Occidental including the states of Nayarit, Sinaloa, Chihuahua and Durango in western Mexico during August and early September 2006 (Fig. 1). Plants were collected irrespective of the presence of any virus-induced symptoms. Fifty-six samples that included four wild dahlia species (D. coccinea, D. rupicola, D. tenuicaulis and D. sherffii) were tested for DMV, DCMV and DMV-D10 incidence. Each sample was from a separate population. The tissues were dried and stored at −80°C.

Total nucleic acid extraction

Total nucleic acids were extracted from dahlia leaf material using the modified Dellaporta procedure (Presting et al., 1995; Pappu et al., 2005). To avoid contamination, each sample was treated with care. The sample tube was opened and approximately 20–40 mg leaf tissue placed in a new tube. To avoid possible contamination, weighing and extraction procedures were conducted in a room where no DMV work had been done previously.

Polymerase chain reaction amplification

Polymerase chain reaction (PCR) was performed by two different persons at two different periods of time to determine the relative incidence of the three caulimoviruses. Primer pairs specific to each of the virus species were used (Fig. 2; Table 1). The quality of extracted DNA was controlled by PCR using a host-specific primer set designed on the internal transcribed spacer (ITS) region of dahlia.

Figure 2.

 DMV-D10 genome organization graph showing the relative positions (forward A; reverse B) of each ORF-specific primer pair used in this study.

Table 1.   Primers used to detect three distinct caulimoviruses in dahlia
PrimerSequence (5′–3′)Annealing temperature (°C)Expected size (bp)
  1. aDMV, Dahlia mosaic virus.

  2. bDCMV, Dahlia common mosaic virus.

  3. cInternal transcribed spacer region of dahlias.

Dahlia ITSc

A 1:20 dilution of the total nucleic acid extract was used in PCR. Each 20-μL PCR volume contained 2 μL total nucleic acid extract, 1× PCR buffer (20 mm Tris, pH 8·4 and 500 mm KCl), 150 μm dNTP mix, 2 mm MgCl2, 0·6 pmol each of forward and reverse primers, 12·7 μL sterile H2O and 0·2 μL GoTaq® Flexi DNA Polymerase (Promega). The amplification was performed in a DNA thermal cycler (BioRad) with the following parameters: an initial incubation of 94°C for 4 min; followed by 50 cycles at 94°C for 30 s, the required annealing temperature (Table 1) based on primer pair used for 20 s, and a 72°C extension step determined based on the size of the amplicon to be synthesized at the rate of 1000 bp min−1; and terminated by a final incubation at 72°C for 7 min. PCR products (7·5 μL) were analysed by agarose gel electrophoresis (1·2%) in 0·5× TAE (Tris–Acetate–EDTA) buffer. The identity of amplicons was verified by cloning and sequencing. Amplicons were cloned into pGEM-T (Promega). Nucleotide sequences were determined using the ABI Prism Sequencing System at the Molecular Biology Core Laboratory of Washington State University and compared with available sequences in GenBank.

Sequence analysis

DNA fragments were assembled into contigs and the contigs were assembled to give complete sequences using contigexpress (Vector NTI Suite 9.0.0, Informax Inc.). Phylogenetic analysis was done using clustalw version 1.83 (Thompson et al., 1994) and mega4 (Tamura et al., 2007).

Results and discussion

Leaf samples from four wild dahlia species were tested for three caulimoviruses using virus-specific primers in a PCR assay conducted by two different persons, as described above, and the outcome was similar. Results showed that the wild dahlia species contained DMV-D10. The four wild species that were tested for DMV, DCMV and DMV-D10 represented each of the three major clades within the genus Dahlia (Saar et al., 2003). There was no evidence of infection by DMV and DCMV. DMV-D10 was found in 91% of the wild dahlia samples (n = 56) tested, and the rest were found to be virus-free (Table 2). The virus-free populations of D. sherffii were collected from Chihuahua, whilst those found infected with DMV-D10 were from both Chihuahua and Durango (Fig. 1).

Table 2.   Incidence of caulimoviruses in wild Dahlia species
Dahlia sp.Samples with DMV-D10Samples free of viral sequences
  1. Data show number of plants infected or free or virus out of the total number of samples tested.

D. coccinea33/341/34
D. rupicola1/10/1
D. sherffii16/204/20
D. tenuicaulis1/10/1

Primer pairs specific to various viral genes (Fig. 2) were used to determine the relative distribution of the viral genes of DMV-D10 in the wild Dahlia spp. ORF VI coding for the inclusion body protein was found at the highest frequency (100%, detected in 51 out of 51 infected samples), followed by the movement protein gene, ORF I (80%, detected in 41 out of 51 infected samples), and the reverse transcriptase gene, ORF IV-V (45%, detected in 23 out of 51 infected samples) (Table 3). To further investigate the potential sequence divergence of this caulimoviral genome in wild dahlia species from that of cultivated dahlias, the gene coding for the movement protein (ORF I) was cloned and sequenced. Pairwise comparisons at nucleotide and amino acid levels with the corresponding gene of DMV-D10 from cultivated dahlias (Pahalawatta et al., 2008a) showed 93% and 89% identity at nucleotide and amino acid levels, respectively. Phylogenetic analysis showed that the nucleotide and deduced amino acid sequences of the ORF I (encoding the movement protein) from wild and cultivated dahlias formed one cluster within the genus Caulimovirus (Fig. 3a,b).

Table 3.   Frequency of occurrence of DMV-D10 ORFs in wild Dahlia species
Dahlia sp.DMV-D10 primers corresponding to ORF I, ORF IV and ORFVI
  1. Data are numbers of plants showing amplification out of total numbers of samples tested.

D. coccinea17/3414/340/42/34
D. rupicola0/10/11/10/1
D. sherffii5/204/200/207/20
D. tenuicaulis0/11/10/10/1
Figure 3.

 Phylogenetic analysis of selected caulimoviruses including DMV-D10 from cultivated dahlias and DMV-D10 Wild spp. from wild dahlias, from the mountain ranges of Mexico. Cluster dendrograms were based on (a) nucleotide and (b) amino acid sequences of the movement protein gene (ORF I) of selected members of the family Caulimoviridae. BRRV, Blueberry red ringspot virus; CaMV, Cauliflower mosaic virus; CERV, Carnation etched ring virus; CYLCV, Cestrum yellow leaf curling virus; ComYMV, Commelina yellow mottle virus; DMV, Dahlia mosaic virus; FMV, Figwort mosaic virus; HLV, Horseradish latent virus; MMV, Mirabilis mosaic virus; PCSV, Peanut chlorotic streak virus; SbCMV, Soybean chlorotic streak virus; SVBV, Strawberry vein banding virus. Bootstrap values are indicated at branching points in the phylogram as a percentage of 1000 iterations.

The nature, extent and timing of the spread of DMV-D10 from wild species to cultivated plants or vice versa is not clear since the ORF II (aphid transmission factor) was not found in the cultivated plants infected with DMV-D10 (Pahalawatta et al., 2008a). DMV-D10 was previously shown to exist as an EPRV in cultivated dahlia (D. variabilis, Pahalawatta et al., 2008a). It remains to be seen if the wild species that were found to contain DMV-D10 sequences also have these sequences as EPRVs. Southern hybridization of total plant genomic DNA from wild species should provide this information. If DMV-D10 were present as an EPRV, it could have spread vertically from these ancestors to present-day cultivated dahlias. Moreover, whether the observed viral ORFs in wild species are functional or not remains to be investigated.

It is well known that plant viruses co-evolved with wild plants in their centres of origin, before they were domesticated, which make these locations the main source, of virus diversity (Lovisolo et al., 2003). Investigating the occurrence of viruses in natural settings has gained significant attention in recent years and raised the possibility of future emergence of such viruses as crop pathogens (Melcher et al., 2006). MacClement & Richards (1956) found that total annual infection was about 10% of herbaceous annual and perennial plants, and these plants were infected by one or more viruses as indicated by mechanical inoculation to test plants. Gibbs et al. (2000) reported Diuris virus Y (genus Potyvirus) and Pterostylis blotch virus (genus Tospovirus) in plants of Diuris orientis and Pterostylis spp., respectively, in their natural habitats in Australia. Barley yellow dwarf viruses (BYDVs) are prevalent in natural grasslands (Garrett et al., 2004) and were also detected in herbarium specimens (Malmstrom et al., 2007). Whilst these findings mostly dealt with RNA viruses, the present study confirmed the existence of reverse transcribing, DNA viral elements in wild plant species for the first time.

Molecular characterization and phylogenetic studies of viral genomes from wild and cultivated plant species may provide important clues about the relation between ‘source’ and ‘sink’, as was the case with Cotton leaf crumple virus (Brown, 2002). Malmstrom et al. (2007) used the historical virus sequences of BYDVs to determine rough time estimates of relevant phylogenetic events. Similarly, Bousalem et al. (2003) analysed the diversification of viral sequences to reveal the course of virus invasion in Dioscorea spp. Studies such as these on plant–virus interactions in natural and managed ecosystems provide insight into plant virus ecology and the human influence on plant virus diversification and spread within natural ecosystems (Malmstrom et al., 2007; Malmstrom & Melcher, 2008).

This study’s findings of viral sequences in plant species in their centres of diversity show the possible co-evolution of reverse transcribing elements and plant pararetroviruses with their plant hosts. Further studies correlating the incidence of viral sequences with the species and population distribution of wild dahlias would provide new avenues of research into the evolutionary pathways of plant-associated pararetroviruses.


Research was supported in part by the Samuel and Patricia Smith Endowment for Dahlia Virus Research, created by the American Dahlia Society, and by funding from the USDA Northwest Nursery Crop Research Center. SE is supported, in part, by a graduate student research assistantship from the WSU-Agricultural Research Center, PPNS no. 0526, Department of Plant Pathology, College of Agricultural, Human and Natural Resource Sciences, Agricultural Research Center, Project # WNPO 0545, Washington State University, Pullman, WA 99164-6430, USA. Field work in Mexico was supported, in part, by KY NSF EPSCoR grant no. 3046884400-06-403 to DES. DES expresses her appreciation for field assistance from Jeffrey R. Bacon, Universidad Juárez del Estado de Durango.