key: cord-0004970-sb6bevb2 authors: Castaño, A.; Hernández, C. title: Complete nucleotide sequence and genome organization of Pelargonium line pattern virus and its relationship with the family Tombusviridae date: 2005-01-13 journal: Arch Virol DOI: 10.1007/s00705-004-0464-y sha: 47fc25cd32675a5936241cea9a6cd41fc27fa63b doc_id: 4970 cord_uid: sb6bevb2 The complete nucleotide sequence of Pelargonium line pattern virus (PLPV) has been determined. The PLPV genomic RNA comprises 3884 nt and contains six open reading frames (ORFs) potentially encoding proteins of 27 (p27), 13 (p13), 87 (p87), 7 (p7), 6 (p6), and 37 kDa (p37), respectively. The arrangement of these ORFs on the PLPV genome closely resembles that of members of the genus Carmovirus in the family Tombusviridae and, moreover, most of the putative PLPV gene products showed high identity with proteins of this viral group. However, several striking differences were noticed. Carmoviruses generate two subgenomic RNAs whereas PLPV produces a single one. In addition, only p7 showed similarity with movement proteins of carmoviruses whereas p6 (as p13) has no viral (or other) homologs. This protein might be expressed from a non-canonical start codon or, alternatively, through a −1 frameshift (FS) mechanism. Both, the production of one subgenomic RNA and the likely involvement of a −1 FS for expression of an internal ORF parallel the translation strategies reported for the unique species of the genus Panicovirus, belonging also to the family Tombusviridae. Overall, the results support the placement of PLPV in this family although its peculiar characteristics preclude its direct assignment to any of the current genera. Pelargonium line pattern virus (PLPV) has a broad experimental host range including more than 30 plant species from nine families, although reports on PLPV infections in nature are limited to geranium (Pelargonium ssp.). This virus may induce the appearance of yellow/green spots and line patterns on the leaves of the affected plants leading to significant quality deterioration [42, 43] . In Spain, a survey done on approximately one thousand geranium samples collected from different regions of the country revealed the presence of PLPV in more than 50% of the tested scions (M. Borja, personal communication). This situation is probably similar to that existing in other countries where this ornamental crop has been largely cultivated, as studies made in the nineties showed a high incidence of PLPV in the Netherlands and Israel [4, 11] . PLPV is transmitted by vegetative propagation and mechanical inoculation, and no vector has been identified. PLPV virions are icosahedral in shape, about 30 nm in diameter and contain a linear positive sense single stranded RNA of approximately 4 kb [35, 44] . The taxonomic position of PLPV is unclear. It was considered as a tentative member of the genus Carmovirus (family Tombusviridae) on the basis of its morphological and physico-chemical properties [32] , but the last report of the ICTV does not include PLPV within any established viral group or in the list of unassigned viruses [51] . This is most likely due to the poor characterization of PLPV and of other small isometric viruses infecting geranium which, as a consequence, has led to some controversies about their identity and to a confusing nomenclature [4, 18, 19, 27, 41, 43] . This situation underlines the need for additional information about these pathogens to facilitate their discrimination and to shed light on their phylogenetic relationships. A partial sequence corresponding to the PLPV coat protein gene has only recently become available at the sequence databases (Accession No. AY038067). In this work we report the entire primary structure of the PLPV genomic RNA and compare the genomic organization and the deduced amino acid sequences from its ORFs with those of related viruses. In addition, we have characterized the single PLPV subgenomic RNA detected in infected plants to gain an insight into the expression strategies of this virus. The results support the inclusion of PLPV in the family Tombusviridae but its assignment to genus level is still not unequivocal, and is discussed. PLPV isolate PV-0193 was obtained from the German Collection of Microorganisms and Cell Cultures (DSMZ, Braunschweig, Germany) and maintained by mechanical sap inoculation onto the experimental hosts Chenopodium quinoa and Nicotiana clevelandii. Total RNA preparations of healthy and infected C. quinoa and/or N. clevelandii leaves were obtained by phenol extraction and lithium precipitation [52] . Double stranded (ds) RNAs were purified from infected tissue by total nucleic acid extraction and subsequent partitioning on nonionic cellulose (CF11, Whatman) with STE (50 mM Tris-HCl pH 7.2, 100 mM NaCl, 1 mM EDTA) containing 16% ethanol [33] . The dsRNAs were separated in non-denaturing 5% polyacrylamide gels and after ethidium bromide staining, the bands corresponding in size to the PLPV genomic and subgenomic RNAs were eluted and recovered by ethanol precipitation. For Northern blot analysis, 4 µg of total RNA from healthy and PLPV infected leaves were denatured by glyoxal-dimethyl sulfoxide treatment, electrophoresed in 1% agarose gels, blotted to nylon membranes (Scheicher & Schuell), and cross-linked by UV irradiation. The membranes were hybridized with 32 P-radioactive DNA probes at 42 • C for at least 12 h in the presence of 50% formamide, 5 × SSC and 5% sodium dodecyl sulfate (SDS). The probes, prepared with Ready to go DNA labelling beads (Amersham Biosciences), corresponded to selected PLPV genome regions: a 5 terminal probe which included from nt 1 to nt 181 of the PLPV genomic RNA, and a 3 terminal probe which comprised from nt 3095 to 3884. After hybridization, membranes were washed for 15 min, twice at room temperature in 2 × SSC plus 0.1% SDS and once at 60 • C in 0.1 × SSC plus 0.1% SDS. A partial PLPV cDNA clone (PL-1030) containing an uncharacterized PLPV cDNA ligated to the pUC9 vector, was provided by the DSMZ plant virus collection. Sequencing of the insert termini allowed the design of two PLPV specific oligonucleotides, CH17 (5 -GAAAATGGCCTTCTACGGGGAC-3 ), homologous to nt 2067-2088 of the PLPV genomic RNA, and CH18 (5 -GCCTCCCTGCTGGCATAAACTAATAC-3 ), complementary to nt 3578-3603. These primers were used for RT-PCR reactions with Titan One Tube RT-PCR system (Roche) employing total RNA preparations from PLPV infected leaves as templates. The amplified cDNA product of 1536 bp was inserted into the pGEM-T easy vector (Promega) and after determination of its sequence, new oligonucleotides were designed to fully characterize the primary structure of PLPV. Oligonucleotide CH32 (5 -CTGACCCCAAACGCCAAGTGGAATG-3 ), complementary to nt 2156-2180, was used together with a degenerate primer, able to anneal to carmovirus related polymerase sequences (31) , in RT-PCR reactions that yielded a cDNA of approximately 1000 bp. To obtain clones containing the 5 and 3 termini of PLPV genomic RNA, the corresponding dsRNA was denatured with 10 mM methylmercuric hydroxide and polyadenylated using yeast poly(A) polymerase (U.S. Biochemicals). After phenol-chloroform extraction and ethanol precipitation, the polyadenylated RNAs were subjected to reverse transcription (RT) with Superscript II-RT (Invitrogen) using primer GeneRacer Oligo dT (Invitrogen), which has a dT tail of 18 nt at its 3 end and contains the priming sites for the GeneRacer 3 and GeneRacer 3 Nested oligonucleotides (Invitrogen) at 5 end. RT products were PCR amplified with the Expand High Fidelity PCR System (Roche) using primers GeneRacer 3 and GeneRacer 3 Nested in combination with others derived from internal regions: CH43 (5 -TAGCTACTGGTTTCTGGAATTGGAG-3 ) or CH46 (5 -TGGCCAACCCATGGA CGCCAATCCG-3 ), complementary to nt 1378-1402 and nt 158-181, respectively, for the 5 terminus, and CH36 (5 -ATAGAGGCGGAGCAACCAGGTAGTGC-3 ) or CH53 (5 -CCTTAATGAAGAACGGGGGGGTAACC-3 ), homologous to nt 3548-3573 and nt 3840-3865, respectively, for the 3 terminus. Complete cDNA clones of the PLPV subgenomic RNA were obtained using the singleprimer method for cloning dsRNAs with unknown sequences [17, 54] with minor modifications. Briefly, the purified dsRNA was polyadenylated and reverse transcribed with primer GeneRacer Oligo dT. After treatment with RNase H (Roche), the cDNA complementary strands were annealed for 10 min at 80 • C, 16 h at 65 • C and 3 h at 30 • C and the resulting 952 A. Castaño and C. Hernández hybrid was filled in with T4 DNA polymerase (Roche) and PCR amplified with Expand High Fidelity PCR System (Roche) and primer GeneRacer Oligo dT. PCR products were separated by electrophoresis in 1% agarose gels or in non-denaturing 5% polyacrilamide gels, depending on the expected size. The DNA fragments were eluted and cloned into the pGEM-T easy vector (Promega) or the plasmid pTZ19R (Fermentas). At least two clones were selected for each DNA fragment and the nucleotide sequence of the inserts was determined with an ABI PRISM DNA sequencer 377 (Perkin-Elmer). The 5 terminal sequence of PLPV genomic RNA and the transcriptional start site of the subgenomic RNA were corroborated by primer extension assays. The substrates for the RT reactions were total RNA preparations from PLPV infected tissue in conjunction with primer CH46 (to map the 5 end of the genomic RNA) or primer CH50 (5 -GCACCGA CAAGATCGCGAGATGTAC-3 ), complementary to nt 2386-2410 (to map the 5 end of the subgenomic RNA). RT reactions were performed with Superscript-RT II in the presence of (α-32 )P(dCTP). The RT products were run in sequencing gels together with sequencing ladders obtained with the corresponding primer and proper recombinant plasmids with PLPV derived DNA inserts. Analysis of the coding capacity of the nucleotide sequences was carried out with the TRANSLATE program (available at the ExPASy proteomics server; http://www.expasy.org/ tools/dna.html). Database searches for nucleotide and protein similarities were done using the programs BLASTN, BLASTP and BLASTX (www.genome.ad.jp, www.ncbi.nlm.nih.gov, www.ebi.ac.uk). Pairwise comparisons of the PLPV nucleotide and amino acid sequences with those of related viruses were performed using the ALIGN program at the Genestream network (http://www2.igh.cnrs.fr/bin/align-guess.cgi) [34] . The RNA secondary structure of the PLPV 3 region was predicted at 37 • C using MFOLD version 3.1 [30, 56] . Multiple alignments of protein sequences were generated with the CLUSTAL X program [47] . These alignments were used to conduct phylogenetic analysis employing programs of the MEGA 2.1 package (24) . Sequences used for comparison were retrieved from the sequence databases and correspond to the following accession numbers: CarMV ( The complete nucleotide sequence of PLPV and its deduced amino acid sequences have been determined. The PLPV genomic RNA consists of 3884 nt and contains six open reading frames (ORFs) (Fig. 1 ). The first ORF potentially encodes a protein of 27 kDa (p27). This ORF(p27) begins at AUG (nt 7-9) in a favourable translation context (CAAAAUGGCAC), with an A at position −3 and a G at position +4 (23, 28) , and terminates at an UAG codon (nt 727-729). If this amber codon is read-through, the ORF extends to a downstream UAA (nt 2302-2304) giving rise to a protein of 87 kDa. Both p27 and p87 present homology with proteins of members of the family Tombusviridae (Table 1 ) which have been shown of PLPV and PMV, respectively, might be translated from non-canonical initiation codons. A −1 frameshift (FS) event could be required for the generation of the putative p7-FS and p8-FS proteins of PLPV and PMV, respectively. These potential frameshift products would have an N-terminus corresponding to p7 (for PLPV) or p8 (for PMV), and a C-terminus corresponding to almost the complete p6 (for PLPV) or the complete p6.6 (for PMV) to be involved in replication [1, 14, 55] and, moreover, p87 contains the motifs characteristic of the viral RNA dependent RNA polymerases (RdRp) including the highly conserved GDD box [22] . Another ORF, starting at AUG (nt 530-532, in the context CCGGAUGAAAA) and ending at UAA (nt 911-913), is nested between ORFs (p27) and (p87) and potentially encodes a protein of 13 kDa (p13) that is unique to this virus. A small centrally located ORF begins at AUG (nt 2274-2276, in the context UCGAAUGGAUA) and stops at UGA (nt 2463-2465), and may encode a protein of about 7 kDa (p7) which exhibits high identity with movement proteins from species of the family Tombusviridae, especially with those of the genus Carmovirus (Table 1 ). In this PLPV putative product, a domain adopting a characteristic α-helix, similar to that able to bind RNA in the p7 of CarMV (29), can be predicted (data not shown). An additional small ORF, overlapping ORF(p7), potentially encodes a protein of 6 kDa (p6) which does not show clear similarity with any viral (or other) protein. This ORF might be expressed from a noncanonical start codon, AUU (nt 2441-2443, in the context UGGAAUUUGUG), and/or by a −1 frameshift (FS) mechanism which would lead to the production of a 12 kDa fusion protein (p12 or p7-FS). Supporting the latter possibility, an heptanucleotide is found immediately upstream of the stop codon of ORF(p7) whose sequence (Fig. 2) fits the canonical frameshifting motifs (5, 6) . Moreover, a stem-loop is predicted downstream of the putative shifty site (data not shown), a type of structure which seems to favour frameshifting (reviewed in [12] ). Finally, the 3 proximal ORF of PLPV, starting at AUG (nt 2621-2623, in the context GGAAAUGGCGG) and ending at UGA (nt 3635-3637), is predicted to encode a 37 kDa coat protein (p37 or CP), as indicates the high amino acid identity found between this product and proteins involved in encapsidation in the genus Carmovirus and in other genera of the family Tombusviridae (Table 1 ). The three different structural domains described in this type of proteins (39) can be also distinguished in the PLPV CP: (i) R, the N-terminal internal domain which Fig. 2 . Comparison of the nucleotide sequences of equivalent genome regions of PLPV and PMV. The heptanucleotide which fits the canonical frameshifting motifs [5] is shown in bold underlined letters. The stop codons of the PLPV ORF(p7) and PMV ORF(p8) are boxed contains many positively charged residues and interacts with RNA, (ii) S, the shell domain which forms a barrel structure made up of β strands and constitutes the capsid backbone and, (iii) P, the protruding C-terminal domain. When overlapping cDNA clones were compared, eighteen nucleotide substitutions were detected within coding regions. Only eight of these point mutations resulted in amino acid changes, five affecting the p27 and/or the p87 (two of them non-conservative, substitutions P → S and R → W at positions 40 and 336, respectively, of p87), two the p13 (both non-conservative, substitutions S → W and A → D at positions 64 and 86, respectively), and one the CP. Comparison of nucleotide sequence of the CP gene of the PLPV isolate characterized in this work with that available at the databases showed an overall sequence identity of 93.7%. The corresponding protein sequences differed at nine amino acid positions including two non-conservative changes affecting the S and P domains (substitutions E → G and L → Q at positions 140 and 296, respectively). The 5 and 3 terminal sequences of the PLPV genomic RNA were determined by sequencing of cDNA clones derived from these regions as indicated in the Methods section. This approach was complemented by primer extension assays (Fig. 3A) which confirmed that the 5 untranslated region (UTR) of PLPV RNA comprises only 6 nt, to our knowledge, the shortest 5 UTR described so far for the genomic RNA of a plant virus. The closest examples in this respect correspond to subgenomic RNAs encoding viral coat proteins as that of Brome mosaic virus (BMV) which has a 5 UTR composed of only 9 nt [3] . It is also worth mentioning that the primer extension assays suggested that the PLPV genomic RNA lacks a 5 terminal cap structure since the cDNA doublet normally associated with a capped RNA was absent [2, 13] . The 3 UTR of PLPV RNA contains 247 nt. As observed in most viral species of the family Tombusviridae, the 3 terminal sequence may fold into a stem-loop which is particularly stable in the case of PLPV because of the presence of six contiguous G-C pairs in the stem (Fig. 4) . Studies with several representatives of the family Tombusviridae have shown that this type of structure is required for the efficient in vivo replication of the viral RNAs [7, 9, 10, 46, 50] . In addition, a midrange base-pairing interaction between the 5 nt in the 3 terminus and a sequence located about 50 nt upstream, similar to that described recently as a replication silencer in Tomato bushy stunt virus and related viruses [36] , can also be predicted for PLPV (Fig. 4) . As stated above, analysis of the coding capacity of PLPV predicts the synthesis of at least three proteins from downstream ORFs. To determine if subgenomic RNAs are generated during PLPV infection which could serve as mRNAs for the production of these proteins, total RNA from C. quinoa leaves infected by PLPV were analyzed by Northern blot hybridization with cDNA probes derived from different regions of the genomic RNA. When a 3 terminal probe was used, a strongly hybridizing RNA species of about 1.6 kb was detected in infected tissue in addition to the genomic size RNA (Fig. 5A) . Such 1.6 kb RNA corresponded very likely to a subgenomic RNA since a 5 specific probe did not hybridize with it (Fig. 5B) . Analysis of dsRNA extracts from PLPV infected tissue revealed the presence of two dsRNA species of about 4.0 kbp and 1.6 kbp, in agreement with the results of the Northern blot (Fig. 5C) . To determine the precise transcription start site of the 1.6 kb subgenomic RNA, the corresponding dsRNA was eluted from the gel, reverse-transcribed and cloned using the single-primer approach described in the Methods section. Sequencing of some of the clones showed that the 5 terminus of the PLPV subgenomic RNA maps 23 nt upstream (position 2251 of the genomic RNA) from the initiation codon of ORF(p7). Primer extension analysis confirmed this result and suggested that such RNA lacks a 5 terminal cap structure as seems to be the case for the genomic RNA (Fig. 3B) . The molecular structure of the PLPV subgenomic RNA suggests that is generated for the translation of the small internal ORFs and the CP ORF having, therefore, the potential to encode at least three proteins. Multiple alignments with the sequences of different proteins of PLPV and those from members of the family Tombusviridae were generated and used for phylogenetic analysis. Comparisons based on the p27 and p7 homologs yielded tentative phylogenetic trees in which the distinction by genera was not clearly discernible since the viruses showed a tendency to form a continuum rather than to cluster (data not shown) in agreement with recent reports [8, 45] . When the p87 homologs were used to perform the analysis, the resulting tree exhibited a better correlation between grouping and genera (with the same exceptions reported in 45) and revealed a close association of PLPV with carmoviruses (Fig. 6A) . In the analysis based on the CP sequences, those corresponding to three additional isometric viruses infecting Pelargonium spp., PelRSV, PCRPV and ELV, were included. The CP tree showed PLPV grouped with most carmoviruses but, remarkably, appeared tightly associated to PelRSV, PCRPV and ELV, three viral entities whose taxonomic position remains to be defined (see below). We have determined the complete 3884 nt sequence of the genomic RNA of PLPV. The arrangement of ORFs in this virus resembles that of members of the genus Carmovirus and, moreover, phylogenetic analyses have shown that PLPV is closely related to this viral group. However, PLPV presents some distinctive features that prevent its direct assignment to this genus. The central region of the genomic RNA of carmoviruses contains two small out-of-frame ORFs whose translation gives rise to two proteins, p7 (or p8) and p9, involved in cell-tocell movement [14, 25] . PLPV genome is not predicted to code for a product homologue of the second movement protein of carmoviruses and presents, instead, a small ORF which potentially encodes a polypeptide, p6, with no clear similarity to any viral protein. In addition, only one subgenomic RNA seems to be produced for translation of downstream PLPV ORFs, whereas carmoviruses generate two subgenomic RNAs for expression of the movement proteins and the CP, respectively [39] . The ORF(p13) is also unique to PLPV although its location on the viral genome is similar to that of the novel ORF(p23) of Hibiscus chlorotic ringspot carmovirus which encodes a protein indispensable for host specific replication [26] . Database searches have not demonstrated similarities between the putative PLPV p13 and any known protein and, moreover, computer-aided sequence analysis has not revealed any motif in p13 which could provide an insight about its function. Hence, the biological significance of p13, if produced in infected plants, is yet to be determined. Caution is required when the possible translation events on RNA are deduced only from sequence data but should all proteins potentially encoded by PLPV be produced in vivo, the virus must use a variety of mechanisms for expression of its genes. Translation of PLPV ORFs (p27) and (p87) will probably occur from the genomic RNA involving a read-through mechanism in the case of ORF(p87), as demonstrated for related ORFs of members of the family Tombusviridae [16, 40, 48, 55] . Indeed, the sequence surrounding the amber stop codon of ORF(p27) is AAA-UAG-GGA, which is consistent with the consensus sequence required for efficient read-through AA(A/G)-UAG-G(G/U)(G/A) [16] . In the case of ORF(p13), an internal ribosomal entry is likely for its translation because the optimal sequence context of the initiation codon of ORF(p27) makes its expression via a leaky scanning mechanism unlikely [12] . The remaining downstream ORFs may be translated from the abundant 1.6 kb subgenomic RNA detected in PLPV infected plants. ORF(p7) would be expressed by conventional ribosomal scanning whereas ORF(p6) might be translated from a non-canonical start codon by leaky scanning and/or by a −1 FS mechanism giving rise to a p12 fusion protein (p7-FS). This also distinguishes PLPV from carmoviruses, whose small internal ORFs present canonical initiation codons and are presumably translated by a leaky scanning mechanism, with the exception of Melon necrotic spot virus for which a readthrough event has been predicted [38] . While we cannot exclude the production of a subgenomic RNA accumulating to very low amounts for the CP gene's expression, over exposure of the Northern blots did not reveal such potential RNA species. Thus, a leaky scanning mechanism or an internal ribosomal entry may account for its translation. Interestingly, the arrangement and predicted expression strategies of the ORFs on the PLPV subgenomic RNA parallel those proposed for the unique subgenomic RNA of Panicum mosaic virus (PMV), the sole member of the genus Panicovirus in the family Tombusviridae, with the exception of an additional ORF overlapping the CP gene within PMV (Fig. 1) [48] . Both subgenomic RNAs have a region upstream of the CP gene devoid of AUG codons except for the one opening the 5 proximal ORF which, moreover, has not an optimal translation context. Furthermore, the proposed "shifty" heptanucleotide of PLPV is identical in sequence and in an equivalent position to one found immediately upstream of the stop codon of the PMV ORF(p8) (Fig. 2) , thus reinforcing the hypothesis of the production of a frameshift fusion protein. Attempts to detect the putative PMV p8-FS in vivo have failed however [49] . In contrast, in vitro translation assays have demonstrated that the non-canonical start codon of PMV ORF(p6.6) is active despite its suboptimal initiation context (CUAGUGG). Subsequent mutagenesis experiments have shown this ORF is required for efficient infection in plants [49] . Despite not share clear sequence similarity, the PMV p6.6 and the potential PLPV p6, also presumably translated from a non-canonical codon in a suboptimal context, do have similar hydrophobicity profiles and, moreover, resemble those of the p9 movement proteins of carmoviruses (data not shown). Only one membranespanning domain however is predicted in the PMV p6.6 and the PLPV p6 whereas the carmovirus proteins contain two transmembrane segments [53] . Bioassays with PMV mutants have suggested that the p6.6 plays a role in virus spread [49] and, considering the analogies found between the two proteins, an equivalent function can be anticipated for the PLPV p6. A series of small isometric viruses naturally restricted to geranium has been described including PFBV, Pelargonium leaf curl virus (PLCV), PNSV, PelRSV, PCRPV, and PLPV. The characterization of these viruses has been mainly based on their biological/physicochemical properties and serological relationships, raising questions as to their taxonomy and status as distinct virus species [4, 18, 19, 27, 41, 43] . This has been further complicated by designating the same virus with different names, for example PLPV, which was formerly known as Pelargonium ring pattern virus [4] . Information on the genomic organization of these pathogens will undoubtedly shed light on this issue. The primary structures of PFBV, a carmovirus, and of PNSV, a proposed new species of the genus Tombusvirus, have been recently reported [15, 37] . In addition, some authors have determined the complete nucleotide sequence of other pelargonium viruses such as PCRPV, PelRSV and also PLPV [20] but have not made the data public. Recently, the genomic organization of PCRPV which potentially encodes five proteins very similar to those involved in replication (p27 and p87), movement (p7 and p9) and encapsidation (p37) in the genus Carmovirus has been described [21] . Interestingly, PCRPV resembles PLPV in producing a single subgenomic RNA for expression of downstream ORFs and in the likely involvement of a −1 FS mechanism for translation of the second small internal ORF. It has been suggested that such traits are also shared by PelRSV and ELV [21] and, based on these common features and sequence similarities, it has been proposed to place the four viruses in a new genus within the family Tombusviridae for which the name Pelarspovirus has been advanced [21] . Only the CP sequences of PCRPV, PelRSV and ELV are available and show high identity with that of PLPV (44, 48 and 50%, respectively). Consistently, the phylogenetic tree based on CP sequences grouped the four pathogens together (Fig. 6) . However, the limited sequence information has not allowed us to further analyze similarities and differences between these viruses and, therefore, commenting on the taxonomic proposal done. To conclude, the results of this study support the classification of PLPV as a member of the family Tombusviridae. Although the PLPV genomic organization and the characteristics of its putative proteins indicate a close association with members of the genus Carmovirus, the uniqueness in the arrangement of some ORFs and proposed expression strategies hinder the direct assignment of the virus to this genus. The detailed description of the genomes of PCRPV, PelRSV and ELV, which resemble PLPV, will allow the establishment of more accurate relationships and the performance of phylogenetic analysis based on whole genome sequences as described recently [45] . We are currently attempting to elucidate the PLPV translation mechanisms and to ascertain the role of the novel ORFs found through the viral genome. Expression of Tobacco necrosis virus open reading frames 1 and 2 is sufficient for the replication of satellite Tobacco necrosis virus cDNA cloning and in vitro transcription of the complete Brome mosaic virus genome Complete nucleotide sequence of Brome mosaic virus RNA 3 Pelargonium flower-break and Pelargonium line pattern viruses in the Netherlands; purification, antiserum preparation, serological identification, and detection in pelargonium by ELISA Ribosomal frameshifting on viral RNAs Mutational analysis of the "slippery sequence" component of a coronavirus ribosomal frameshifting signal Mutation analysis of cis-elements in the 3 -and 5 -untranslated regions of satellite Tobacco necrosis virus strain C RNA Molecular variability of twenty-one geographically distinct isolates of Carnation mottle virus (CarMV) and phylogenetic relationships within the Tombusviridae family Analysis of sequences and predicted structures required for viral satellite RNA accumulation by in vivo genetic selection 3 -Terminal RNA secondary structures are important for accumulation of Tomato bushy stunt DI RNAs Virus and virus-like diseases of pelargonium in Israel Translation in plants-rules and exceptions Structure of the glycoprotein gene of Sonchus yellow net virus, a plant rhabdovirus Turnip crinkle virus genes required for RNA replication and virus movement Pelargonium necrotic spot virus: a new member of the genus Tombusvirus Complete nucleotide sequence and genome organization of Hibiscus chlorotic ringspot virus, a new member of the genus Carmovirus: evidence for the presence and expression of two novel open reading frames An efficient cloning strategy for viral double-stranded RNAs with unknown sequences Elderberry latent virus: its relationship to Pelargonium ringspot virus and its identification as a distinct member of the genus Carmovirus, family Tombusviridae Partial characterization of Pelargonium line pattern and Pelargonium ringspot viruses Genome organization of Pelargonium ringspot and Elderberry latent viruses Genome organization of Pelargonium chlorotic ring pattern virus: further implications for Tombusviridae taxonomy Evolution and taxonomy of positive-strand RNA viruses: implications of comparative analysis of amino acid sequences Point mutations define a sequence flanking the AUG initiator codon that modulates translation by eukaryotic ribosomes MEGA 2: molecular evolutionary genetics analysis software Cell-to-cell movement of Turnip crinkle virus is controlled by two small open reading frames that function in trans The p23 protein of Hibiscus chlorotic ringspot virus is indispensable for host-specific replication Viruses of pelargonium in Italy Selection of AUG initiation codons differs in plants and animals In vivo detection, RNAbinding properties and characterization of the RNA-binding domain of the p7 putative movement protein from Carnation mottle carmovirus (CarMV) Expanded sequence dependence of thermodynamic parameters improves prediction of RNA secondary structure Use of highly conserved motifs in plant virus polymerases as the tags for specific detection of carmovirus-related RNA-dependent RNA-polymerase genes Carnation mottle virus and viruses with similar properties Isolation and analysis of double-stranded RNA from virusinfected plant and fungal tissue Comparison of DNA sequences with protein sequences Some properties of a distinctive isometric virus from pelargonium A replication silencer element in a plus-strand RNA virus Complete nucleotide sequence and genome organization of Pelargonium flower break virus Nucleotide sequence and genomic organization of Melon necrotic spot virus Molecular Biology of Tombusviridae The Tomato bushy stunt virus replicase proteins are coordinately expressed and membrane associated Nine viruses isolated from pelargonium in the United Kingdom A whole genome perspective on the phylogeny of the plant virus family Tombusviridae Analysis in vivo of Turnip crinkle virus satellite RNA C variants with mutations in the 3 -terminal minus-strand promoter The CLUSTAL X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools Nucleotide sequence and infectivity of a full-length cDNA clone of Panicum mosaic virus A gene cluster encoded by Panicum mosaic virus is associated with virus movement Mutational analysis of cis-acting sequences in the 3 -and 5 -untranslated regions of RNA 2 of Red clover necrotic mosaic virus Virus Taxonomy. Seventh Report of the International Committee on Taxonomy of Viruses A small-scale procedure for the rapid isolation of plant RNAs Insertion and topology of a plant viral movement protein in the endoplasmic reticulum membrane Sequence-independent amplification and cloning of large dsRNA virus genome segments by poly(dA)-oligonucleotide ligation Inmunodetection, expression strategy and complementation of Turnip crinkle virus p28 and p88 replication components Mfold web server for nucleic acid folding and hybridization prediction Author's address: Carmen Hernández, Instituto de Biología Molecular y Celular de Plantas (UPV-CSIC) We thank Drs. R. Flores and V. Pallás for critical reading of the manuscript and suggestions, and Dr. M. Borja for valuable comments in the course of this work. This research was supported by grants AGL2000-0942 and AGL2003-04249 from the Ministerio de Ciencia y Tecnología de España and by grant Grupos 03/064 from the Generalitat Valenciana (to C. H.). A. C. was the recipient of predoctoral fellowships from the Generalitat Valenciana and from CSIC-Fundación Bancaja.