Phylogenetic analysis of Tomato mosaic virus from Hemerocallis sp. and Impatiens hawkeri

The culture and commercialization of ornamental plants have considerably increased in the last years. To supply the commercial demand, several Hemerocallis and Impatiens varieties have been bred for appreciated qualities such as flowers with a diversity of shapes and colors. With the aim of characterizing the tobamovirus isolated from Hemerocallis sp. (tobamo-H) and Impatiens hawkeri (tobamo-I) from the USA and Sao Paulo, respectively, as well as to establish phylogenetic relationships between them and other Tobamovirus species, the viruses were submitted to RNA extraction, RT-PCR amplification, coat-protein gene sequencing and phylogenetic analyses. Comparison of tobamovirus homologous sequences yielded values superior to 98.5% of identity with Tomato mosaic virus (ToMV) isolates at the nucleotide level. In relation to tobamo-H, 100% of identity with ToMV from tomatoes from Australia and Peru was found. Based on maximum likelihood (ML) analysis it was suggested that tobamo-H and tobamo-I share a common ancestor with ToMV, Tobacco mosaic virus, Odontoglossum ringspot virus and Pepper mild mottle virus. The tree topology reconstructed under ML methodology shows a monophyletic group, supported by 100% of bootstrap, consisting of various ToMV isolates from different hosts, including some ornamentals, from different geographical locations. The results indicate that Hemerocallis sp. and I. hawkeri are infected by ToMV. This is the first report of the occurrence of this virus in ornamental species in Brazil.

Although the states of Santa Catarina, Rio Grande do Sul, Minas Gerais, Rio de Janeiro, Pernambuco and Ceará produce ornamentals for sale, São Paulo State is the main producer with US$ 9.0 million in exports of live plants and floricultural products in 2001 (9). Other ornamentals which deserve mention are Impatiens hawkeri and I. walleriana (Balsaminaceae) which are among the most marketable species, due to the introduction of a wide range of new flower colors and forms (19).
Since Impatiens and Hemerocallis species are vegetatively propagated and constantly introduced, viruses tend to appear and spread quickly in the crop.
Several viruses are described as infecting Impatiens species, including Impatiens necrotic spot virus, Tomato spotted wilt virus, Tobacco streak virus, Tobacco ringspot virus, Helenium virus S and Cucumber mosaic virus (CMV) (19). In Brazil, flexuous particles (6), CMV (7), Tomato mosaic virus (ToMV) (8) and an unidentified potyvirus (5) have been reported, although there have been no previous reports of Hemerocallis virus in Brazil.
The present study identified a tobamovirus isolated from Hemerocallis sp. from the USA and compared its sequence with another tobamovirus previously reported (8) infecting I. hawkeri from São Paulo State, Brazil, with emphasis on phylogenetic analysis.

Virus source, RNA extraction and Reverse Transcriptase-Polymerase Chain Reaction (RT-PCR)
The tobamovirus was isolated from Hemerocallis sp. (tobamo-H) showing necrotic spots, and from I. hawkeri (tobamo-I) exhibiting mosaic and leaf deformation. Viral RNA was extracted from infected leaves by using the Trizol LS reagent according to the manufacturer ‚ s instructions (Gibco BRL).

Sequencing
Amplified products of tobamo-H and tobamo-I were purified using the CONCERT Rapid Gel Extraction Kit according to the manufacturer ‚ s instructions (Gibco-BRL) and sequenced using the ABI Prism Big Dye Terminator System (PE Applied Biosystems) and the primers PS1 and PAS2.

Sequence analyses
A fragment of 480bp corresponding to the complete coat protein (CP) gene of tobamo-H and tobamo-I was analyzed. The sequences were submitted to the Basic Local Alignment Search Tool (Blast 2.0) (http://www.ncbi.nlm.nih.gov/BLAST/) and aligned with Tobamovirus species, ToMV isolates and Tobacco rattle virus (TRV) using the program CLUSTAL X or by eye with the Se-Al program, Version 1.0 alpha 1 (24). The degree of identity between them was determined by using the PAUP* v. 4.0b10 program (26). Moreover, the tobamo-H and tobamo-I nucleotide sequences of the CP gene was aligned with Tobacco mosaic virus isolated from Petunia sp. (TMV-p) and 9 ToMV isolates.

Phylogenetic analyses
Sequences from 23 Tobamovirus species or isolates including those from tobamo-H and tobamo-I as well as from 9 isolates from ToMV were subjected to phylogenetic analyses. Swofford´s (26) PAUP* v. 4.0b10 with heuristic search and equal weighting in the maximum parsimony (MP) analysis was employed. Bootstrap values were obtained using the branch-and-bound method and computed after 1000 resamplings followed by an MP reconstruction. The likelihood ratio test (LRT) (12) for the nucleotide substitution models comparison was performed using Modeltest v. 3.06 (23). The tree reconstruction was also performed under maximum likelihood (ML) criterion using PAUP*.
The primers PS1 (3) and PAS2, designed based on the partial sequence of tobamovirus isolated from I. hawkeri (8), were able to amplify a fragment with 480 nucleotides corresponding to the complete coat protein (CP) gene of tobamoviruses isolated from Hemerocallis sp. and I. hawkeri here named tobamo-H and tobamo-I, respectively. The sequences from tobamo-H and tobamo-I were submitted to GenBank, and the accession numbers are DQ230836 and AY063743, respectively.
The BLAST search revealed significant alignment of the tobamo-H and tobamo-I with the CP of ToMV, TMV and Pepper mild mottle virus (PMMoV).
Analyses involving ToMV sequences from several hosts were made because a high identity with tobamo-H and tobamo-I sequences was observed (Table 1). Thus, values superior to 98.5% were verified for ToMV isolates, including those from the ornamental plants Camellia sp. (AJ 417701), Hibiscus rosa-sinensis (11) and Eustoma russellianum (13). It is worth mentioning that an identity of about 86% with the Brazilian ToMV sequence was observed (Table 1), supporting the suggestion that the virus might be considered an intermediary strain of ToMV (22).
In relation to the tobamo-H sequence, 100% of identity with ToMV isolated from tomatoes from Australia and Peru was found. It is important to note that the Hemerocallis was imported from the USA and hence the plant was already introduced infected.
In the MP analysis of the nucleotide sequences of tobamo-H, tobamo-I and other tobamoviruses, using Tobacco rattle virus (TRV) as the out group, only one tree was found with 2266 steps. When the analysis was carried out with ToMV isolates, using the sequence from the TMV-p as the out group two most-parsimonious trees with 567 steps were found.
Based on LRT with tobamovirus sequences, the model HKY presented the best likelihood value of -lnL = 8488.0928. The distribution of rates at variable sites yielded a gamma with the shape parameter alpha = 0.8800. The statistical tests yielded the following nucleotide frequencies among the sequences: A (0.2594), C (0.1889), G (0.2323) and T (0.3194). The tree reconstructed under the LRT criterion is shown in Figure 1A.
Tobamo-H and tobamo-I revealed a monophyletic group with ToMV-Br supported by a bootstrap value of 94% (data not shown). In the tree reconstructed after ML methodology the studied viruses share a common ancestor with ToMV-Br, TMV-p, ORSV-Br and PMMoV (Fig. 1A).
ML analysis also revealed two groups sharing a common ancestor:     (17) and the monophyletic group formed by Brassicaceae, Curcubitaceae and Malvaceae viruses (Fig. 1A). Three main tobamovirus groups have been proposed based on genomic location of their origin of virion assembly, structure of the junction between movement protein and CP open reading frames, phylogenetic position, and botanical family of the hosts (17,2). The hosts of subgroup I are mainly solanaceous, those of subgroup II are cucurbits and legumes, and those of subgroup III are cruciferous.
The tree topology obtained from the data suggested that the lineage from Brassicaceae hosts (Subgroup III) had a divergence more basal than those from cucurbitaceous. On the other hand, the tree topology presented by Adkins (1) shows that cucurbit subgroup lineage seems to be the most basal divergence, probably between or within the Cucumber green mottle mosaic virus (CGMMV) and Sunn-hemp mosaic virus (SHMV) (10).  Table 1, TRV (AJ272199) Based on our phylogenetic analyses, HLSV and HLFPV clustered with 100% bootstrap and made a monophyletic group with cucurbit viruses, Frangipani mosaic virus (FrMV) and SHMV. This analysis corroborated with results obtained by Adkins (1) who proposed that HLSV and HLFPV could be new viruses belonging to the malvaceous subgroup.
In relation to ToMV sequences, the nucleotide substitution model obtained was also HKY, with likelihood value of -lnL = 1344.4966, gamma with the shape parameter alpha = 0.4270 and nucleotide frequencies among the sequences A (0.2958), C (0.2034), G (0.2086) and T (0.2922). The tree reconstructed under LRT criterion is shown in Figure 1B.
With the aim of comparing tobamo-H and tobamo-I with isolates from different hosts including ornamental and from different regions of the world, phylogenetic analyses based on MP (data not shown) and ML methodology were made. The tree topology proposed allowed us to point out that ToMV-Br lineage divergence was most basal in relation to the other taxa. Based on ML analysis using TMV-p as the out group, the monophyletic group formed by ToMV isolated from lisianthus and other isolates, including those from Hemerocallis (tobamo-H) and Impatiens (tobamo-I) are strongly supported by a bootstrap value of 100% (Fig. 1B). Tobamo-H share a common ancestor with the monophyletic group formed by tobamo-I, and the ToMV isolated from tomatoes from Australia, China, Germany and Peru (Fig. 1B).
Based on the results, the present report describes the occurrence of an infection by ToMV (subgroup I) in two non-solanaceous plants, Hemerocallis sp. (Hemerocallidaceae) and I. hawkeri (Balsaminaceae). It is important mentioning that this is the first report of ToMV in ornamental species in Brazil.
The ability of ToMV to infect Camellia, E. russellianum (lisianthus), Hemerocallis, Hibiscus and Impatiens can be explained by the virus jumping from a solanaceous host to another family, indicating a relaxed constraint involving the host plant in the subgroup I proposed by Lartey (17). Gibbs (10) suggested that the ability of Ribgrass mosaic virus to prosper in both brassicas and plantains, like the gene sequences encoding their movement-and/or coat protein junction, may be a "derived" rather than an "ancestral" state.
The jumping of virus lineage has also been used to explain the evolution of the tobamovirus subgroup III (17). In the present case, the jumping is observed inside the subgroup I, which is significant for the management of economically important plants especially those vegetatively propagated. Since ToMV is a very infective virus, rigid phytosanitary measures must be used to prevent its transmission. Thus, the incidence of virus infections in stocks will decrease dramatically and will improve the quality standards for impatiens species.