Nucleotide Sequence and Infectivity of a Full-Length Cdna Clone of Panicum Mosaic Virus

Home , Eremochloa ophiuroides, Panicovirus, Panicum mosaic virus

VIROLOGY 241, 141±155 (1998) ARTICLE NO. VY978939

Nucleotide Sequence and Infectivity of a Full-Length cDNA Clone of Panicum Mosaic Virus

Massimo Turina,1 Miko Maruoka,*,2 Judit Monis,*,3 A. O. Jackson,* and Karen-Beth G. Scholthof4

Department of Plant Pathology and Microbiology, Texas A&M University, College Station, Texas 77843; and *Department of Plant and Microbial Biology, University of California, Berkeley, California 94720 Received September 4, 1997; returned to author for revision October 2, 1997; accepted November 9, 1997

The sequence of an infectious cDNA clone of panicum mosaic virus (PMV) showed that the single-stranded RNA genome is 4326 nucleotides (nt) and a single highly abundant subgenomic (sg) RNA of 1475 nt was synthesized during PMV infection of pearl millet plants and protoplasts. Computer comparisons revealed strong similarities between the predicted amino acid sequences of the p48 and p112 open reading frames (ORFs) and replicase proteins of members of the Tombusviridae. The sgRNA has the potential to encode five proteins. Three small ORFs, p8, p8-FS, and/or p6.6 have similarity to ORFs of carmo-, necro-, and machlomoviruses thought to be involved in virus spread in plants. The sgRNA also has the potential to encode a 26-kDa capsid protein and a 15-kDa nested gene (p15) of unknown function. PMV transcripts also supported replication and movement of SPMV, the satellite virus. Genome organization, physicochemical properties, and biological features indicate that PMV is a member of the Tombusviridae family. However, PMV differs sufficiently from previously described members to warrant its placement in a new genus provisionally designated Panicovirus. © 1998 Academic Press

INTRODUCTION reported on centipedegrass (Eremochloa ophiuroides)in Florida, Louisiana, Arkansas, Mississippi, North Caro- Panicum mosaic virus (PMV) is a 30-nm icosahedral lina, and South Carolina (Haygood and Barnett, 1992). monopartite plant virus composed of a plus sense sin- The causal agent was characterized as a mixed infection gle-stranded RNA genome of approximately 4300 nucle- of a strain of PMV and a satellite virus (SPMV) that otides (nt) and a 26-kilodalton (kDa) capsid protein (CP) depends on PMV for movement and replication, but en- (Buzen et al., 1984; Masuta et al., 1987). The host range codes its own capsid protein (Buzen et al., 1984; Niblett of PMV is restricted to a few grass species (Niblett and et al., 1977). Depending on the host plant, PMV or mixed Paulsen, 1975; Niblett et al., 1977; Sill and Pickett, 1957), infections of PMV and SPMV can elicit a chlorotic mottle, where it typically causes a systemic mosaic. PMV was stunting, and seed reductions in forage grasses. During first reported on switchgrass (Panicum virgatum)inKan- the warm summer months, severe symptoms develop in sas in 1953 (Sill and Pickett, 1957) and was recently infected St. Augustinegrass (Dale and McDaniel, 1982). confirmed in breeding plots in Oklahoma (K.-B. G. Thus far, a biological vector has not been identified, but Scholthof, unpublished observations). Several serologi- infections can be spread by virus-contaminated lawn- cal variants of PMV have been described on St. Au- mower blades and from infected St. Augustinegrass gustinegrass (Stenotaphrum secundatum) (Holcomb et plugs or sod (Haygood and Barnett, 1992; Holcomb et al., al., 1972). In the mid-1960s, a virus disease named ``St. 1989). In the laboratory and under natural field condi- Augustine decline'' became so severe in the Lower Rio tions, PMV can support the replication of SPMV (Buzen et Grande Valley that it limited the production, mainte- al., 1984; Masuta et al., 1987) and at least two satellite nance, and establishment of St. Augustinegrass in Texas RNAs (satRNAs). The complete nucleotide sequence of (McCoy et al., 1969). A similar disease has since been two strains of SPMV have previously been determined (Berger et al., 1994; Masuta et al., 1987). The taxonomy of small isometric viruses infecting the The nucleotide sequence data reported in this paper have been Poaceae is complex, and their characterization has been submitted to the GenBank nucleotide sequence database and as- based primarily on serological relationships. In addition signed Accession No. U55002. 1 Present address: Istituto di Patologia Vegetale, Via Filippo Re 8, to PMV isolates in North America, a serological analysis 40126 Bologna, Italy. of 17 isolates of monopartite single-stranded RNA vi- 2 Present address: Genentech, 1 DNA Way, South San Francisco, CA ruses with icosahedral particles from various species of 94080. Poaceae in Europe revealed that PMV and molinia streak 3 Present address: Agritope, 8505 SW Creekside Place, Beaverton, virus (MSV) are related (Huth et al., 1974; Paul et al., OR 97005. 4 To whom correspondence and reprint requests should be ad- 1980). Furthermore, MSV is able to support the replica- dressed. Fax: (409) 845-6483. E-mail: [email protected]. tion of SPMV (Buzen et al., 1984). The sequence analysis

Analysis of the PMV RNA sequence The complete sequence of PMV reveals several unusual features of the 4326-nt genome. Among these is that PMV RNA has the potential to encode seven proteins (Figs. 3 and 5). Sequence analyses of PMV cDNAs indicate that if all seven ORFs are expressed in vivo, the virus uses a variety of translational control mechanisms for the expression of its genes, including translational readthrough, a ribosomal frameshift, leaky scanning, a noncanonical translational start site, and possibly one or more internal ribosomal entry sites. FIG. 1. Alignment of the cDNA clones derived from PMV genomic RNA. The plasmids pHindIII-PMV-5Ј, pM1, pPMV3, pPMV7, pPMV10, pPMV12, pPMV13, pPMV17, pPMV18, and pPMV21 are aligned along The 5Ј nontranslated leader sequence the 4326-nt PMV genome. Unique restriction enzymes sites on the PMV genome are indicated above the pPMV85 full-length infectious cDNA The 5Ј end of the genome was determined by primer clone which was constructed by ligation of pHindIII-PMV-5Ј, pPMV12, extension of the genomic RNA using a 32P end-labeled and pPMV17. Open rectangles on pPMV85 indicate significant open DNA primer (MT30) complementary to a portion of the 5Ј reading frames. The transcriptional start site of the sgRNA at nucleo- proximal end of the virus RNA (Figs. 4A and 4B). The tide 2851 is indicated by an arrowhead (). primer extension products produced in the absence of dideoxynucleotides lacked the terminal G doublet nor- of a full-length infectious clone of PMV described in this mally associated with a capped RNA (Goldberg et al., study provides further justification for the contention that 1991). The 5Ј end of PMV RNA (5Ј-GGGUAUU-3Ј) (Figs. 3 PMV forms a new genus of plant viruses. and 4) is similar to the 5Ј termini of the genomic RNA of carnation mottle carmovirus (CarMV; 5Ј-GGGUAAG-3Ј) RESULTS AND DISCUSSION (Carrington et al., 1989), maize chlorotic mottle machlo- Construction and infectivity of full-length cDNA movirus (MCMV; 5Ј-AGGUAAU-3Ј) (Nutter et al., 1989) clones and cowpea mottle carmovirus (CPMoV; 5Ј-NGGUAUC- 3Ј)(Youet al., 1995). Comparisons of the predicted 5Ј To determine the primary sequence of PMV, 10 cDNA untranslated leaders of PMV and the tombusviruses clones (Fig. 1) were selected from viral RNA that had within the Tombusviridae revealed no obvious similari- been either polyadenylated and primed with an oligo(dT) ties between these RNAs. primer (pPMV10, pPMV13, and pPMV17), primed with One difference, a C instead ofaGatposition 8, was random oligonucleotides (pPMV3, pPMV7, pPMV12, detected within the untranslated leader sequence de- pPMV18, and pPMV21), or primed with specific oligonu- rived from the pPMV85 clone compared to the sequence cleotides complementary to internal regions of PMV RNA obtained from RNA sequencing (compare Figs. 3 and 4B (pM1 and pHindIII-PMV-5Ј). Each cloned insert was se- with Fig. 4A). This variant was also observed in two of quenced on both strands and the sequence was con- three PCR-derived cDNA clones of pHindIII-PMV-5Ј and firmed on the full-length infectious cDNA clone after it one clone was identical to the wild-type RNA sequence. was constructed. Either PCR-associated error rates or the diversity of the Pearl millet plants inoculated with full-length tran- viral RNA population (Holland, 1993) could account for scripts derived from pPMV85 developed mild mosaic these differences. However, this difference did not result symptoms on the upper uninoculated leaves within 12 in any obvious change in replication, movement, or phe- days postinoculation. The mild symptoms were typical of notype on pearl millet plants compared to wild-type PMV those elicited by PMV lacking satellites. In contrast, co- infections (data not shown). inoculation of infectious PMV and SPMV transcripts resulted in a severe systemic mosaic within 2 weeks post- Relatedness of p48 and p112 to Tombusviridae inoculation which was indistinguishable from wild-type replicases mixed infections. RNA analysis of total RNA extracted from PMV- or PMVϩSPMV-infected plants or protoplasts, The first AUG on PMV genomic RNA at nt 29±31 is the using a 32P-labeled probe to the 3Ј end of the PMV start codon of a putative 48-kDa protein (p48) which genome, revealed the presence of the genomic RNA terminates at an UAG codon at positions 1304±1306 (gRNA) and one sgRNA (Figs. 2A and 2C). Immunoblot (Figs. 3 and 5). A translational readthrough could extend assays of plants indicated that the 26-kDa PMV CP was the ORF to a second stop codon (UAA) at nt 3005±3007 present in the PMV infections and that in the mixed to generate a 112-kDa readthrough protein (p112). A infections, both the PMV and the 17-kDa SPMV capsid second in-frame AUG codon is present at nt 77±79 but proteins were present (Fig. 2B). this putative initiation codon has a suboptimal transla- PANICUM MOSAIC VIRUS GENOME 143

FIG. 2. Infectivity of in vitro transcripts of pPMV85 and pSPMV1 in pearl millet plants and protoplasts. Leaves were extracted at 12 days postinoculation and protoplasts at 2 days posttransfection. (A and C) Northern blot analyses were performed on total RNA extracts following electrophoresis through 1% agarose gels and transfer to nylon membranes. (A) RNA from (left to right) uninoculated leaves (U) and plants inoculated with transcripts from pPMV85 (P) or pPMV85ϩpSPMV1 (S). The three panels (left to right) represent 32P-labeled DNA primers described under Materials and Methods, which were used to recognize the 3Ј terminus of PMV (3Ј; primer PMV3Ј), the sgRNA 5Ј end (sg; primer PMV2919R), or the 5Ј genome end of PMV (5Ј; primer PMV231R), respectively. An SPMV-specific primer probe (SPMV; primer SPMV150R) was similarly prepared, as shown in the far right panel. The PMV genomic (g) and subgenomic (sg) RNA and the SPMV RNA are indicated. (B) Detection of PMV and SPMV capsid proteins from extracts after 12% SDS±PAGE, transfer to nitrocellulose membranes, and probing with a PMV or SPMV capsid protein-specific polyclonal antibody. The lanes indicated represent upper uninoculated leaves of plants infected with PMV (P) or PMVϩSPMV (S) transcripts. Molecular weight markers are indicated in kilodaltons (kDa) and the polyclonal rabbit antibody (Ab) probe is indicated above each panel. (C) Northern blot analysis of total RNA from pearl millet protoplasts transfected with PMV transcripts (P) or the uninoculated control (U) using a probe to the full-length PMV cDNA clone. The positions of the gRNA and sgRNA are indicated. tional start context of UCCAUGC (Fig. 3), when compared thesis of p48 and the predicted p112 readthrough prod- to the context AAGAUGG surrounding the first potential uct at a much reduced level (Figs. 6A and 6B). This start site (Kozak, 1995, 1996; Lutcke et al., 1987). Eight readthrough strategy is used by other members of the nucleotide variants were detected on the putative repli- Tombusviridae to express replicase-associated proteins case ORFs p48 and p112. Within the p48 ORF, C or A (Scholthof et al., 1995; White et al., 1995). In an earlier polymorphisms at nt 248 encode Leu or Met, respec- report using RNA isolated from purified virions, Buzen et tively; U or C variants at nt 529 maintained the Thr al. (1984) also observed two PMV-related proteins during codons;aCoranAatnt848coded for Leu or Ile in vitro translation. These proteins were speculated to be respectively; andaTorCatnt1074encodes Ile or Thr, the capsid protein and an unknown protein of ca. 50 kDa. respectively. The region downstream of the predicted Our results with full-length infectious transcripts that are readthrough codon contains four nucleotide variants. free of contaminating sgRNA suggest that the larger Three of these, at nt 2041 withaUorC,atnt2470 with protein is p48. a C or A, and at nt 2578 with an A or G, result in no Sequence comparisons indicate that p48 and p112 are changes in amino acids. At nt 2175, U or C variants the viral replicase proteins (Fig. 5 and Table 1). The encode Leu or Ser, respectively. None of these nucleo- sequence context also implies that these two proteins tide differences resulted in reading frame changes or the are expressed from the genomic RNA by a translational introduction of stop codons. strategy similar to those used by tomato bushy stunt Translation of full-length transcripts from pPMV85 in a tombusvirus (TBSV) (Scholthof et al., 1995) and turnip wheat germ in vitro translation system resulted in syn- crinkle carmovirus (TCV) (White et al., 1995). The regions 144 TURINA ET AL.

FIG. 3. Nucleotide sequence of PMV RNA derived from cDNA clones and the predicted open reading frames (ORFs) corresponding to the viral polypeptides. The nucleotide position and the ORFs are indicated to the left of the sequence. Within the sequence an asterisk (*) designates a stop codon and the GDD motif is underlined. The asterisk indicating the stop codon of the p48 replicase is also the putative readthrough (UAG) codon for p112. The transcriptional start site for the sgRNA () is at nt 2851. The putative shifty heptanucleotide site for expression of p8-FS by a ribosomal frameshift is indicated with a double underline and the noncanonical (GUG) start of p6.6 with diamonds (ࡗࡗࡗ). Note that the 3Ј terminal nucleotide is shown in parentheses because we have been unable to determine whether the PMV genome terminates with the sequence CCC or CCCA. encompassing the first ORF of the replicases of the Methyltransferase and helicase domains are absent on representative members of the Tombusviridae are nec- PMV, as is characteristic of other members of the Tom- essary for replication (Andriessen et al., 1995; Scholthof busviridae (Buck, 1996; Koonin and Dolja, 1993). et al., 1995; White et al., 1995). Proteins derived from The high degree of similarity among replicase regions analogous ORFs are also implicated in the formation of of the positive-strand RNA viruses was previously used multivesicular bodies within the infected plant cell fol- for phylogenetic analysis (Koonin, 1991; Koonin and lowing infection with cymbidium ringspot and carnation Dolja, 1993). A ``signature region,'' DX3[FYWLCA]X0±1- Italian ringspot tombusviruses (Burgyan et al., 1996). DXn[STM]GX3TX3[NE]Xn[GS]DD, is characteristic of RNA PANICUM MOSAIC VIRUS GENOME 145

FIG. 3ÐContinued replicases, where ``X'' is any amino acid residue with tity that are considered ``in the twilight zone'' (Doolittle, alternative residues shown in brackets (Buck, 1996; Koo- 1987) be subjected to further statistical analyses. Signif- nin, 1991; Koonin and Dolja, 1993). This signature occurs icant relationships between putative ORFs can be deter- on the PMV genome upstream of the GDD motif within mined and grouped as marginally significant (scores of the p112 ORF (Fig. 3). Sequence motifs IV and VI of RNA 3±5), probably significant (a score of 5±10), or of certain polymerases are located upstream of the predicted relatedness (scores of 10 or more) (Doolittle, 1987). readthrough codon of the PMV p112 ORF, and this region Amino acid sequence comparisons within the signature is similar to those of other members of the Tombusviri- region confirm a highly statistically significant degree of dae (Habili and Symons, 1989). similarity among viruses belonging to the Tombusviridae With an increasing number of viruses within the Tom- (Fig. 5 and Table 1). From these analyses, the putative busviridae available for comparisons, it is important that replicase proteins of PMV appear to be more closely amino acid sequences within the range of 15±25% iden- related to the carmoviruses, including CPMoV (You et al., 146 TURINA ET AL.

FIG. 3ÐContinued

1995), than to the tombusviruses, necroviruses, or oat infected with PMV. The primer extension analyses sug- chlorotic stunt virus (OCSV) (Table 1). In the same way, gested that the 5Ј terminus of the sgRNA corresponds to the putative replicases of PMV and MCMV can be nt 2851 and that the sgRNA lacks a 5Ј terminal cap grouped together, since they have significantly higher structure (Figs. 4C and 4D). Interestingly, the first seven relatedness than pairwise comparisons between PMV nucleotides of this 1475 nt sgRNA exactly match the and other viruses within the Tombusviridae (Table 1). sequence of the 5Ј end of the gRNA (5Ј-GGGUAUU-3Ј) (Figs. 3 and 4). Northern blot analyses using pPMV85 or The sgRNA transcriptional start and predicted ORFs oligonucleotide-specific hybridization probes revealed To determine the transcription start site of a ca. that only one sgRNA accumulated to detectable amounts 1500-nt sgRNA, primer extension and RNA sequencing in protoplasts and plants (Fig. 2). were performed with sucrose density gradient-purified The sgRNA has the potential to encode five proteins. sgRNA recovered from total RNA of pearl millet plants Three ORFs are present within the 5Ј half of the sgRNA, PANICUM MOSAIC VIRUS GENOME 147

FIG. 3ÐContinued and the putative coat protein (p26) ORF, which contains a corresponding to positions 2875 to 2877, extends to a p15 nested ORF, resides within the 3' proximal portion of stop codon (UAG) at nt 3094±3096 (Figs. 1 and 3) to the sgRNA. Following a 24-nt untranslated region on the encode a putative 8-kDa protein (p8). The p8 start codon sgRNA, a start codon in a poor context (CGAAUGU), overlaps the p112 replicase ORF by ϩ2 nt (Fig. 3). A 148 TURINA ET AL.

FIG. 4. Determination of the 5Ј terminal nucleotides of the PMV gRNA and sgRNA. Synthetic oligonucleotide primers, complementary to the viral gRNA were labeled at their 5Ј termini with 32P and annealed to RNA isolated from pearl millet plants systemically infected with wild-type PMV. RNA (A) or cDNA (B) sequencing reactions primed with the same primer (5Ј-ATGGCATGGAAAGCGAAGATGTAGC-3Ј; nt 83 to 59) following separation on polyacrylamide gels are shown adjacent to the primer extension products (PE) to reveal the 5Ј end of the gRNA. The PE products of the sgRNA using a primer (5Ј-AACAGTAGACATTCGGG-3Ј; nt 2887±2870) downstream of the predicted transcriptional start site with comparisons based on total RNA (C) or sgRNA (D) isolated from PMV-infected plants. The nucleotide sequences are represented in the virus complementary sense. second ORF, that could be expressed as a Ϫ1 frameshift the coat protein (data not shown), even under the most (FS) immediately upstream of the stop codon of p8, could restrictive conditions for multicistronic expression of pro- result in a protein of 14.6 kDa (p8-FS) that terminates at teins from internal initiation sites; i.e., a wheat germ a UAG at positions 3273±3275. The canonical frameshift- extract translation system with a capped sgRNA tem- ing sequences (XXXYYYN) include a putative shifty plate versus rabbit reticulocyte lysate and uncapped codon of UUC representing the YYN nucleotides (Brier- sgRNA (Futterer and Hohn, 1996). These results suggest ley, 1995; Brierley et al., 1992) that is similar to the that efficient leaky scanning or internal ribosome entry 5Ј-CAAUUUC-3Ј heptad observed for PMV at nucleotides are possible strategies for coat protein expression from 3087±3093 (Fig. 3). Immediately downstream of the shifty the sgRNA. site, a predicted strong stem-loop appears, and this structure may facilitate frameshifting. In addition, or al- The relatedness of p8, p8-FS, and p6.6 to viral ternatively, a p6.6 protein could be expressed in the movement proteins same reading frame as p8-FS (diamonds on Fig. 3) from a noncanonical start site (CUAGUGG) at nucleotides The carmo-, necro-, and machlomoviruses have the 3096 to 3098 by a leaky scanning mechanism (Futterer potential to encode two or three small proteins from the and Hohn, 1996). Interestingly, no AUG codons occur in central region of the genome. Recently, tobacco necrosis any reading frame from the poor translational start con- necrovirus strain D (TNV-D) p7a protein was shown to be text of p8 until the CP ORF (nt 3272±3274). a nucleic acid binding protein that localizes in cell walls In vitro translation from wheat germ extracts suggest and membrane fractions (Drouzas et al., 1996; Offei et al., that the sgRNA template can express p8 and the putative 1995), suggesting that this protein may be associated p26 CP abundantly (Fig. 6C). A smaller amount of label with virus spread in infected plants. Similarly, disruption generally appeared as a doublet near the position of the of the p8 ORF of TCV abrogated virus spread (Hacker et 14.3-kDa molecular weight marker. These bands could al., 1992). Comparisons of the predicted amino acid se- represent the predicted p15 and p8-FS proteins. A p6.6 quences at the carboxy terminus of PMV p8 with three product was not clearly evident in these experiments, carmoviruses and a machlomovirus reveals conserva- possibly due to the inefficiency of initiation from the tion of a FNF motif within these putative movement pro- putative noncanonical (GUG) start codon. Mutational teins (Fig. 7). There are also similarities to the putative analysis or immunoprecipitation will be necessary for movement proteins of CPMoV, which ends in FNI (You et more precise determination of the ORFs that are required al., 1995), and that of TNV-D, which ends in FNH (Coutts to sustain PMV infections in planta. It is interesting to et al., 1991). note that the most abundant product translated in vitro is These small RNA viruses also encode a downstream PANICUM MOSAIC VIRUS GENOME 149

FIG. 5. Comparison of the RNA genome organization predicted for selected members of the Tombusviridae. The viruses include panicum mosaic virus (PMV), maize chlorotic mottle machlomovirus (MCMV), cowpea mottle carmovirus (CPMoV), melon necrotic spot carmovirus (MNSV), turnip crinkle carmovirus (TCV), carnation mottle carmovirus (CarMV), and tobacco necrosis necrovirus, strain D (TNV-D). The genomes are aligned to the putative replicase readthrough codon (UAG) as suggested by Riviere and Rochon (1990). The similarly shaded boxes indicate ORFs with significant amino acid sequence similarity, excluding PMV p15, CpMoV p28/38, and MCMV p31.6 and p32.7. The carboxy end of PMV p8-FS and the entire ORF of PMV p6.6 are identical, but expression of both proteins would require two translational strategies. A Ϫ1 frameshift (FS) would be required for p8-FS and noncanonical initiation would be necessary for translation of p6.6. The approximate sizes of the ORFs are indicated in kilodaltons and the capsid protein (CP) ORFs are shown.

ORF which may be translated by different strategies. We p8-FS (equivalent to p6.6) with the analogous regions of suggest that the putative 14.6-kDa ORF (p8-FS) may re- melon necrotic spot carmovirus (MNSV), CarMV, TCV, sult from a -1 frameshift from the PMV p8 ORF. The and MCMV (Table 1; Riviere and Rochon, 1990). The hydrophobicity profiles (Devereux et al., 1984; Kyte and results imply that the gene products are conserved, but Doolittle, 1982) of the amino acid sequences of TCV p9, that more than one translational strategy may exist for TNV-D p7b, and the carboxy terminus of PMV p8-FS (or their expression. For example, the TCV p9 ORF is pre- p6.6) are remarkably similar (data not shown), suggest- sumably expressed by leaky scanning of ribosomes ing that these proteins may have common functions or (Riviere and Rochon, 1990) that bypass the start codon sites of cellular localization during the infection process. for p8 (5Ј-UAGAUGG-3Ј) to initiate 139 nt downstream at We have also aligned the carboxy-terminal half of PMV the stronger p9 context (5Ј-GAAAUGA-3Ј) (Fig. 5). MNSV 150 TURINA ET AL.

FIG. 6. In vitro translation of transcripts from the full-length infectious cDNA clone (pPMV85) and the PCR-amplified sgRNA template. RNA transcripts were translated in a wheat germ system optimized with 110 mM potassium acetate. The [14C]leucine-labeled proteins were analyzed on (A and C) 16.5% or (B) 12.5% SDS±PAGE. (A and B) pPMV85 templates with 1-day (A) or 10-day (B) exposures with p48 and the p112 readthrough product indicated by an arrow and an asterisk, respectively. (C) Translation products from the in vitro synthesized sgRNA with predicted ORFs of p8, p8-FS (14.6 kDa), p15, and p26 (capsid protein). Designations above the panels are: M, molecular weight markers; P(g), the full-length PMV transcript; P(sg), the sgRNA transcript; and C, the optimized wheat germ extract lacking template RNA. The molecular weight markers in kDa are indicated at the left of each panel. has two possible coding strategies (Riviere and Rochon, putative p15 nested gene product. A single nucleotide 1990), which may be mediated either by leaky scanning variant was detected on this region of the PMV genome. to express p7B or by an amber codon readthrough to At nt 3625, eitheraCoranAwasobserved; this variant result in p7A and p14 (Fig. 5). If leaky scanning occurs, maintained the Leu codon for the 26-kDa ORF, which is the ribosome would have to scan past the p7A start predicted to encode the CP, but caused a change from codon (5Ј-CUCAUGG-3Ј) to reach the p7B start codon in Ser to Tyr within the nested p15 ORF. The CP ORF initates a similar context (5Ј-UGUAUGG-3Ј) 201 nt downstream. at nucleotides 3272±3274 (AUG) and extends to a termi- Alternatively, a readthrough could result in expression of nation codon (UAA) at nucleotides 3998±4000. The p15 p7A and p14 (instead of p7B), but the nucleotide context ORF begins 11 nt downstream of the CP start site (nt surrounding the putative MNSV p14 readthrough codon 3282±3284) and is terminated by a UGA stop codon at is less conserved than the context surrounding the nucleotides 3678±3680 (Fig. 3). The translation start con- readthrough region of the MNSV replicase (Riviere and text of p15 (GCAAUGG) is somewhat more favorable than Rochon, 1990). Both the TNV-D p7a and p7b proteins are that of the CP (GAAAUGA), and would be a candidate for similar in sequence to the corresponding proteins leaky scanning. present in the Carmovirus genus (Table 1). Internal initi- The 3Ј terminal location of the p15 and CP ORFs on the ation or ribosomal frameshifting are possible mecha- sgRNA suggests that PMV employs an unusual transla- nisms for translation of these three TNV-D proteins (p7a, tional strategy when compared with the coat protein p7b, and p7a-7b) from a single sgRNA (Offei et al., 1995). expression strategies of other small RNA viruses. An A similar mechanism may exist for TCV, MNSV, and abundant sgRNA whose sole translation product is the TNV-A (Offei et al., 1995; Riviere and Rochon, 1990). CP is present in the carmo-, necro-, and tombusviruses. MCMV may have the potential to express the small ORF Thus far, Northern blot analyses from RNA extracted from downstream of p9 by a ϩ1 frameshift (Riviere and plants or protoplasts at various times postinoculation Rochon, 1990). In addition, a translational readthrough in with PMV indicate that only a single abundant sgRNA is the same region probably results in p32.7 (Fig. 5) (Lom- transcribed during infection (Figs. 2A and 2C). However, mel et al., 1991; Nutter et al., 1989). Inspection of the PMV transient transcription of a low abundance mRNA cannot genome as well as our in vitro translation data suggest yet be excluded. Nevertheless, it is difficult to explain the that internal initiation and a frameshift are plausible high level of in vivo accumulation of the coat protein from mechanisms for expression of p8, p8-FS, and/or p6.6 a low abundance RNA during infection (Fig. 2). In addi- from the sgRNA (Figs. 3 and 6). tion, our in vitro translation data show that the CP is the most abundant product of the synthetic sgRNA tran- The 3Ј proximal CP and p15 ORFs and their possible scripts (Fig. 6C). Thus, these data are compatible with a translational strategies model in which the sgRNA is multicistronic, with the As indicated above, the 3Ј half of the 1475-nt sgRNA of capability of expressing multiple proteins (p8, p8-FS, PMV contains two ORFs which encode the CP and a p6.6, CP, and p15) in vivo. However, the exact mechanism PANICUM MOSAIC VIRUS GENOME 151

TABLE 1 Pairwise Amino Acid Sequence Comparisons of Proteins Predicted from the PMV Genome and Analogous Proteins from Representative Machlomoviruses (MCMV), Carmoviruses (CPMoV, TCV, MNSV, and CarMV), Necroviruses (TNV-A and TNV-D), Tombusviruses (TBSV), Sobemo- viruses (SBMV and RYMV), Luteoviruses (BYDV-PAV), and OCSV

p48 p112 p8 p8-FS CP p15

MCMV 17.5 64.5 3.1 6.7 11.3 Ͻ1 (X14736) (26.1) (58.7) (23.5) (31.6) (26.8) (24.0) CPMoV 11.4 40.2 2.3 1.1 4.5 1.2 (U20976) (23.6) (43.2) (14.7) (20.3) (23.3) (22.1) TCV 12 37.4 5.7 2.6 3.2 Ð (M22445) (23.4) (44.6) (25.3) (30.6) (23.4) Ð MNSV 12.4 38.9 5.0 5.6 4.1 Ð (M29671) (23.1) (46.7) (26.7) (15.5) (21.2) Ð CarMV 14.6 42.1 7.0 2.2 4.7 Ð (X02986) (28.5) (43.7) (39.0) (23.7) (24.7) Ð TNV-A 9.6 51.3 2.8 3.6 6.8 Ð (M33002) (26.8) (46.0) (21.2) (35.4) (22.1) Ð TNV-D 1.4 38.6 3.7 2.1 8.0 Ð (D00942) (17.9) (39.6) (26.3) (18.9) (23.9) Ð OCSV 1.9 31.2 Ð Ð 2.1 Ͻ1 (X83964) (27.6) (36.4) Ð Ð (18.6) (18.2) TBSV 1.2 33.6 Ð Ð 7.1 Ð (M31019) (18.9) (39.6) Ð Ð (22.0) Ð BYDV-PAV Ð Ð Ð Ð 2.3 Ͻ1 (X07653) Ð Ð Ð Ð (18.0) (22.5) SBMV Ð Ð Ð Ð 5.5 Ð (L34672) Ð Ð Ð Ð (22.3) Ð RYMV Ð Ð Ð Ð 3.7 Ð (L20893) Ð Ð Ð Ð (19.8) Ð

Note. The GenBank Accession numbers for each virus are indicated. Pairwise comparison was carried out through the BESTFIT program with the algorithm of Gribskov and Burgess (1986). For each pairwise comparison, the statistical significance was calculated as the adjusted alignment score (ASϭSobs-Sran/␴), where Sobs is the alignment score obtained upon comparison of the given pair of sequences, Sran is the average score obtained upon comparison of 25 pairs of random sequences generated by jumbling the real sequences, and ␴ is the standard deviation. Scores of 3 to 5 are marginally significant, scores of 5±10 are probably significant, and scores of Ͼ10 indicate certain relatedness between the putative ORFs (Doolittle, 1987). The percentages of identity for the pairwise comparisons are shown in parentheses, directly below each score. The dashes (Ð) indicate comparisons that were not performed or in which no analogous ORF was present. The p112 comparisons correspond to the portion of the ORF that is downstream of the predicted replicase readthrough codon (Figs. 3 and 5). Comparison of p8-FS includes the carboxy-terminal region that is equivalent to the p6.6 ORF (Figs. 3 and 5). of expression will require extensive additional study. but which are hallmarks of many animal virus genomes Expression of the CP and p15 may be due to internal as well as some plant virus genomes (Futterer and Hohn, ribosomal entry sites which have not yet been identified, 1996; Ivanov et al., 1997).

Sequence analyses of the capsid protein and the nested p15 ORF Pairwise comparisons of the capsid protein gene of PMV with the capsid proteins of the viruses shown in Table 1 reveal that the amino acid alignment scores range from 2.1 (OCSV) to 11.3 (MCMV) and amino acid sequence identity ranges from 18% with barley yellow dwarf luteovirus strain PAV (BYDV-PAV) to 26.8% with MCMV. Phylogenetic distances of multiple alignments of FIG. 7. Comparisons of the predicted amino acid sequences at the the shell (S) domain of the capsid protein are in agree- carboxy termini of PMV p8 and similar regions on MNSV, TCV, MCMV, ment with results previously derived by Dolja and Koonin and CarMV. Identical nucleotides are indicated by a colon and the (1991), and our analyses indicate a very close related- nucleotides that are conserved for all the viruses are indicated by asterisks. The stop codons of the putative movement proteins are ness to MCMV (Table 1). underlined. The conserved amino acid sequence (FNF) is shown under Another interesting feature found within the CP ORF is the corresponding nucleotide codons. the presence of a nested 15-kDa ORF (p15) beginning 11 152 TURINA ET AL. nucleotides downstream of the CP start codon. Com- that PMV represents a new entity within the Tombusviri- puter comparisons have failed to reveal any obvious dae. These analyses have prompted us to provisionally match or similarity of p15 with proteins of related viruses, designate PMV as the type member of a new genus, and pairwise comparisons with the ORFs nested within Panicovirus. the CP genes of MCMV, CPMoV, BYDV-PAV, and OCSV The genome organization of PMV is distinct from also showed no similarity to p15 (Table 1). Only the that of viruses within the Carmovirus genus, because BYDV-PAV p17 ORF, which is nested within the CP gene, none of the recognized members have a gene nested has previously been analyzed experimentally, and al- in the 3Ј ORFs from one abundant sgRNA. However, though this protein is dispensable for virus replication, it pairwise comparisons revealed significant scores of appears to be required for movement (Mohan et al., 17.5 for the replicase and 11.3 for the PMV-MCMV coat 1995). protein alignments versus the PMV, carmo-, and necrovirus alignments (Table 1). Both PMV and MCMV The 3Ј trailer sequence have larger amino terminal regions (48 or 50 kDa) of Following the PMV CP stop codon is an untranslated the replicase proteins than the carmoviruses (27 to 36 trailer sequence extending from positions 4001 to 4326 kDa). Nevertheless, we propose splitting MCMV and (Fig. 3). An ORF with the potential to encode a 34-amino- PMV for several reasons. First, MCMV has a 5Ј prox- acid polypeptide occurs in the trailer region, but no imal 32-kDa gene overlapping the 5Ј replicase ORF, evidence exists for the presence of a small sgRNA or while PMV does not (Lommel et al., 1991). Second, translational strategies for its expression. We have not wild-type MCMV RNA has a 7-methylguanosine cap yet determined if this small ORF is required for replica- structure at the 5Ј end of the genomic RNA (Lommel et tion in the same manner as the tombusvirus pX se- al., 1991), and PMV does not. A third argument for PMV quence (Boyko and Karasev, 1992; Dalmay et al., 1993; placement in a new viral genus is the puzzling trans- Scholthof and Jackson, 1997). Within the untranslated lational strategy for the CP (and subsequently p15). trailer of PMV, at nt 4020, a variant U or C was detected. The type members of the carmo-, necro-, and tombus- Despite numerous attempts using various techniques virus genera produce two abundant subgenomic (Masuta et al., 1987; Shirako and Wilson, 1993; Zuidema RNAs, but from our results, PMV appears to use only et al., 1986) we were unable to determine whether the 3Ј one multicistronic RNA to facilitate internal expression terminal nucleotide of PMV ends with an A at position of the CP and p15, by an unknown mechanism. 4027 oraCatposition 4026. The A nucleotide may be An additional important distinguishing feature contrib- derived from in vitro polyadenylation of the genomic RNA uting to phylogeny may be the nested gene (p15) in the followed by oligo(dT) priming for the production of the CP region on the PMV genome. Both MCMV and CPMoV first-strand cDNA. However, SPMV, a satellite virus of have genes overlapping the CP. However, our results fail PMV,hasaCastheterminal nucleotide (Masuta et al., to reveal obvious amino acid sequence relatedness 1987), suggesting that the terminal nucleotide of PMV among the putative PMV, MCMV, and CPMoV translation may also be a C. Our inability to resolve the 3Ј end of products, despite the fact that the CP similarity is con- PMV also mirrors problems associated with determining served (Table 1). Identity up to 25% is questionable the termini of other members of the Tombusviridae in- (Doolittle, 1987), and only following a statistical analysis cluding CarMV (Carrington et al., 1989), MCMV (Nutter et can any association be determined when the signifi- al., 1989), TCV (Heaton et al., 1989), cardamine chlorotic cance of relatedness has scores of Ͼ3. A novel gene fleck carmovirus (CCFV) (Skotnicki et al., 1993), TBSV that originates by overprinting of a coding sequence (Hearne et al., 1990), and MNSV (Riviere and Rochon, (such as the CP or replicase) has been proposed to be a 1990). These viruses all share a consensus of 5Ј- step in generation of a new virus genus (Keese and GCCC(A)-3Ј, but have been refractory to direct identifica- Gibbs, 1992). This argument can be applied to PMV in the tion of the 3Ј terminal nucleotide. Thus far, pothos latent sense that p15 was likely to be generated by overprinting tombusvirus is the only member of the Tombusviridae of the coat protein ORF to generate a new gene for a which has clearly been determined to have a 3Ј terminal particular novel or specialized function. This suggests C residue (Rubino et al., 1995). that the CP is the original gene and that the occurrence of p15 is an evolutionarily important intermediate in gen- Taxonomic placement of PMV erating a new virus genus, which we provisionally des- The restriction of PMV primarily to grasses belonging ignate Panicovirus. Following further characterization, to the Paniceae tribe (Powell, 1994) of the Poaceae has molinia streak virus and other monopartite isometric previously been noted (Niblett and Paulsen, 1975; Sill RNA viruses, which are known to infect the Poaceae and and Pickett, 1957). The genome organization of PMV, which share physicochemical and serological properties including the number of ORFs, their position in the ge- with PMV (Catherall and Chamberlain, 1977; Huth et al., nome, the putative translational strategy of the proteins, 1974; Paul et al., 1980), may be placed in the Panicovirus and their relatedness, provides a persuasive argument genus. PANICUM MOSAIC VIRUS GENOME 153

MATERIALS AND METHODS Assembly and infectivity of full-length cDNA clones

Virus propagation and purification and RNA extraction A cDNA clone representing a 5Ј-terminal portion of PMV RNA was generated by RT±PCR using reverse PMV 109S, a Kansas strain obtained from Chuck Nib- primer MT33 (5Ј-CCCAACTAATGGATTGGTCAC-3Ј;nt lett, was maintained and purified as previously described 1234 to 1214) and forward primer (5Ј-ATGCAAGCTTTA- (Buzen et al., 1984; Masuta et al., 1987). RNA was ex- ATACGACTCACTATAGGGTATTGGCTGCAACC-3Ј) com- tracted from purified virions according to the methods posed of the first 17 nucleotides at the 5Ј end of PMV described previously (Jackson and Brakke, 1973), with RNA (double underline), a T7 RNA polymerase promoter modifications (Hunter et al., 1986). The extracted RNA sequence (bold), and a HindIII site (underlined). This was further purified on 7.5±30% sucrose density gradi- clone, pHindIII-PMV-5Ј, in combination with pPMV12 and ents (Goldberg and Brakke, 1987). pPMV17, was used to generate a full-length cDNA clone of PMV (Fig. 1). The full-length clone, designated cDNA cloning and sequence analyses pPMV85, was subsequently used for sequence analyses and infectivity experiments. Cesium chloride-purified Standard molecular biology protocols (Sambrook et plasmid DNA of pPMV85 was digested with EcoICR I al., 1989) or modifications suggested by the commercial (Promega Corp., Madison, WI) to generate a template suppliers of the reagents were followed. PMV RNA was which extends the PMV cDNA by 30 bp downstream from denatured with 10 mM methyl mercuric hydroxide fol- the putative 3Ј terminus of the genome. SPMV RNA lowed by a polyadenylation reaction using protocols de- (Masuta et al., 1987) was prepared from pSPMV1 (Monis scribed by Sambrook et al. (1989). A library of cDNA et al., 1992) by digestion with BglII to linearize the plas- clones was generated using oligo(dT) and random or mid resulting in a single extra nucleotide following the 3Ј specific oligonucleotide primers, and the resulting DNA ϩ terminus of the cDNA. The in vitro transcription reactions fragments were ligated into plasmid vectors pBSKS and inoculations of pearl millet plants were performed as (Stratagene, La Jolla, CA) or pUC119. A series of clones described (Scholthof et al., 1993). Preparation and trans- designated pM1 and pPMV 3, 7, 10, 12, 13, 17, 18, and 21 fections of protoplasts derived from 2-week-old pearl were selected for sequence analyses (Fig. 1) following millet plants were essentially as described previously restriction enzyme mapping and Southern blot hybridiza- (Jones et al., 1990; Scholthof et al., 1993), except that the tion. Primer-based sequencing of both strands of the leaves were sliced into 1-mm strips and at each step the cDNA clones was performed with Sequenase 2.0 (Am- protoplasts were centrifuged at 1100 rpm due to their ersham, Arlington Heights, IL) using custom primers syn- small size. thesized by Genosys Biotechnologies, Inc. (The Wood- Protein and immunoblot analyses of leaves and pro- lands, TX) or Life Technologies (Gaithersburg, MD). The toplasts were performed as previously described initial database search was performed through the (Scholthof et al., 1995). Rabbit anti-PMV antibody was BLAST E-mail server (Altschul et al., 1990), followed by used at a dilution of 1:4000 and rabbit anti-SPMV anti- evaluation with the University of Wisconsin Genetics body at 1:2000 (Masuta et al., 1987). RNA isolation from Computer Group (GCG) software package (Devereux et leaves and protoplasts and hybridizations with [32P]- al., 1984). dCTP-labeled pPMV85 were performed as described previously (Jones et al., 1990; Scholthof et al., 1993). Determination of the 5Ј and 3Ј terminal nucleotides of Oligonucleotide-specific hybridization was performed PMV genomic and subgenomic RNA with 1 ␮gof32P-5Ј labeled primer at 42°C. The oligonucleotides used for PMV were: PMV231R (5Ј-GGT- The 5Ј terminal RNA sequence of PMV genomic RNA GAGAGATGGGAATAGC-3 ; nt 231 to 213); PMV2919R (5 - and the transcriptional start site of the sgRNA were deter- Ј Ј TTTGGTTGCAAGGGTATCTTGGG-3 ; nt 2919 to 2897); mined directly by a dideoxynucleotide termination method Ј and PMV 3 (5 -GGGCCTGGTGAGGGGATT-3 ; nt 4326 to with reverse transcriptase (Gibson et al., 1990) and primer Ј Ј Ј 4309). Primer SPMV150R (5 -GATGTGGCAGCAGC- extension (Zhou and Jackson, 1996). The substrates for Ј CCGG-3 ; nt 150 to 133) was used for the detection of these reactions were either virion RNA or total RNA from Ј SPMV RNA. PMV-infected plants fractionated on sucrose density gradi- ents (Goldberg and Brakke, 1987) in conjunction with In vitro translation analyses of the genomic primer MT30 (5Ј-ATGGCATGGAAAGCGAAGATGTAGC-3Ј; and sgRNA nt 83 to 59), primer 2887R (5Ј-AACAGTAGACATTCGGG-3Ј; nt 2887±2870), or primer MT36 (5Ј-ACGTCCTCTGG- The PMV cDNA representing the subgenomic RNA TAGGGG-3Ј; nt 3307±3291). Determination of the 3Ј termi- was amplified by PCR with a 3Ј primer (PMV 3Ј:5Ј- nal nucleotide was attempted using previously described GGGCCTGGTGAGGGGATT-3Ј; nt 4326 to 4309) and a 5Ј methods (Masuta et al., 1987; Shirako and Wilson, 1993; primer (HindIII-T7-Sg1: 5Ј-GGAAGCTTTAATACGACT- Zuidema et al., 1986). CACTATAGGGTATTCCTGCAAG-3Ј). This primer consists 154 TURINA ET AL. of a HindIII site (underlined), a T7 polymerase promoter Functional analysis of cymbidium ringspot virus genome. Virology sequence (bold), and a PMV sgRNA sequence from po- 194, 697±704. sitions 2851 to 2865 (double underline) representing the Devereux, J., Haeberli, P., and Smithies, O. (1984). A comprehensive set of sequence analysis programs for the VAX. Nucleic Acids Res. 12, first 15 nt of the sgRNA. Substrate RNAs were produced 387±395. from linearized pPMV85 or the PCR-amplified sgRNA Doolittle, R. F. (1987). ``Of URFs and ORFs.'' Univ. Science Books, Mill templates followed by in vitro translation from wheat Valley. germ extract (Promega Corp.) optimized with a final con- Drouzas, A. D., Offei, S. K., and Coutts, R. H. A. (1996). Subcellular centration of 110 mM potassium acetate. The [14C]- location and expression of tobacco necrosis Necrovirus p7a protein. J. Phytopathol. 144, 297±299. leucine-labeled in vitro translation products were sepa- Futterer, J., and Hohn, T. (1996). Translation in plantsÐrules and ex- rated by electrophoresis through 12.5% (Goldberg et al., ceptions. Plant Mol. Biol. 32, 159±189. 1991) or 16.5% (Schagger and von Jagow, 1987) polyacryl- Gibson, C. A., Dunster, L. M., Bouloy, M., Minor, P. D., Sanders, P. G., and amide gels containing sodium dodecyl sulfate (SDS± Barrett, A. D. T. (1990). Primer extension dideoxy chain termination PAGE). nucleotide sequencing of partially purified RNA virus genomes: A technique for investigating low titre viruses with extensive genome secondary structure. J. Virol. Methods 29, 167±176. ACKNOWLEDGMENTS Goldberg, K.-B., and Brakke, M. K. (1987). Concentration of maize chlorotic mottle virus increased in mixed infections with maize dwarf Research support was obtained through funding from Texas Agricul- mosaic virus, strain B. Phytopathology 77, 162±167. tural Experiment Station Grant H-8388, USDA Competitive Grant 96- Goldberg, K.-B., Modrell, B., Hillman, B. I., Heaton, L. A., Choi, T.-J., and 35303-3714, and NIH Postdoctoral Fellowship AI-08710 to K.-B. G. S.; Jackson, A. O. (1991). Structure of the glycoprotein gene of sonchus AI-08541 to J. M.; and USDA Grant 87-CRCR-2556 to A. O. J. We yellow net virus, a plant rhabdovirus. Virology 185, 32±38. appreciate the helpful comments of Herman Scholthof, Diane Law- Gribskov, M., and Burgess, R. R. (1986). Sigma factors from E. coli, B. rence, and Jeff Batten. We also thank Joan Kuecker for technical subtilis, phage SP01, and phage T4 are homologous proteins. Nu- assistance and maintenance of the greenhouse facilities. cleic Acids Res. 14, 6745±6763. Habili, N., and Symons, R. H. (1989). Evolutionary relationship between luteoviruses and other RNA plant viruses based on sequence motifs REFERENCES in their putative RNA polymerases and nucleic acid helicases. Nu- cleic Acids Res. 17, 9543±9555. Altschul, S. F., Gish, W., Miller, W., Myers, E. W., and Lipman, D. J. (1990). Hacker, D. L., Petty, I. T. D., Wei, N., and Morris, T. J. (1992). Turnip Basic local alignment search tool. J. Mol. Biol. 215, 403±410. crinkle virus genes required for RNA replication and virus movement. Andriessen, M., Meulewaeter, F., and Cornelissen, M. (1995). Expres- Virology 186, 1±8. sion of tobacco necrosis virus open reading frames 1 and 2 is Haygood, R. A., and Barnett, O. W. (1992). Widespread occurrence of sufficient for the replication of satellite tobacco necrosis virus. Virol- centipedegrass mosaic in South Carolina. Plant Dis. 76, 46±49. ogy 212, 222±224. Hearne, P. Q., Knorr, D. A., Hillman, B. I., and Morris, T. J. (1990). The Berger, P. H., Shiel, P. J., and Gunasinghe, U. (1994). The nucleotide complete genome structure and synthesis of infectious RNA from sequence of satellite St. Augustine decline virus. Mol. Plant-Microbe clones of tomato bushy stunt virus. Virology 177, 141±151. Interact. 7, 313±316. Heaton, L. A., Carrington, J. C., and Morris, T. J. (1989). Turnip crinkle Boyko, V. P., and Karasev, A. V. (1992). Tombusvirus genome may virus infection from RNA synthesized in vitro. Virology 170, 214±218. encode the sixth small protein near its 3Ј terminus. Virus Genes 6, Holcomb, G. E., Derrick, K. S., Carver, R. B., and Toler, R. W. (1972). St. 143±148. Augustine decline virus found in Louisiana. Plant Dis. Rep. 56, 69±70. Brierley, I. (1995). Ribosomal frameshifting on viral RNAs. J. Gen. Virol. Holcomb, G. E., Liu, T. Z., and Derrick, K. S. (1989). Comparison of 76, 1885±1892. isolates of panicum mosaic virus from St. Augustinegrass and cen- Brierley, I., Jenner, A. J., and Inglis, S. C. (1992). Mutational analysis of tipedegrass. Plant Dis. 73, 355±358. the `slippery sequence' component of a coronavirus ribosomal frame- Holland, J. (1993). Replication error, quasispecies populations, and shifting signal. J. Mol. Biol. 227, 463±479. extreme evolution rates of RNA viruses. In ``Emerging Viruses'' (S. S. Buck, K. W. (1996). Comparison of the replication of positive-stranded Morse, Ed.), pp. 317. Oxford Univ. Press, New York. RNA viruses of plants and animals. Adv. Virus Res. 47, 159±251. Hunter, B. G., Heaton, L. A., Bracker, C. E., and Jackson, A. O. (1986). Burgyan, J., Rubino, L., and Russo, M. (1996). The 5Ј-terminal region of Structural comparison of poa semilatent and barley stripe mosaic a tombusvirus genome determines the origin of multivesicular bod- virus. Phytopathology 76, 322±326. ies. J. Gen. Virol. 77, 1967±1974. Huth, W., Paul, H.-L., Querfurth, G., and Lesemann, D. E. (1974). Molinia Buzen, F. G., Niblett, C. L., Hooper, G. R., Hubbard, J., and Newman, streak virus: A virus with isometric particles from Molinia coerulea. M. A. (1984). Further characterization of panicum mosaic virus and its Intervirology 2, 345±351. associated satellite virus. Phytopathology 74, 313±318. Ivanov, P. A., Karpova, O. V., Skulachev, M. V., Tomashevskaya, O. L., Carrington, J. C., Heaton, L. A., Zuidema, D., Hillman, B. I., and Morris, Rodionova, N. P., Dorokhov, Yu, L., and Atabekov, J. G. (1997). A T. J. (1989). The genome structure of turnip crinkle virus. Virology 170, tobamovirus genome that contains an internal ribosome entry site 219±226. functional in vitro. Virology 232, 32±43. Catherall, P. L., and Chamberlain, J. A. (1977). Relationships, host- Jackson, A. O., and Brakke, M. K. (1973). Multicomponent properties of ranges and symptoms of some isolates of phleum mottle virus. Ann. barley stripe mosaic virus ribonucleic acid. Virology 55, 483±494. Applied Biol. 87, 147±157. Jones, R. W., Jackson, A. O., and Morris, T. J. (1990). Defective-interfering Coutts, R. H. A., Rigden, J. E., Slabas, A. R., Lomonossoff, G. P., and RNAs and elevated temperatures inhibit replication of tomato bushy Wise, P. J. (1991). The complete nucleotide sequence of tobacco stunt virus in inoculated protoplasts. Virology 176, 539±545. necrosis virus strain D. J. Gen. Virol. 72, 1521±1529. Keese, P. K., and Gibbs, A. (1992). Origins of genes: ``Big bang'' or Dale, J. L., and McDaniel, M. C. (1982). St. Augustinegrass decline in continuous creation? Proc. Natl. Acad. Sci. USA 89, 9489±9493. Arkansas. Plant Dis. 66, 259±260. Koonin, E. V. (1991). The phylogeny of RNA-dependent RNA poly- Dalmay, T., Rubino, L., Burgyan, J., Kollar, A., and Russo, M. (1993). merases of positive-strand RNA viruses. J. Gen. Virol. 72, 2197±2206. PANICUM MOSAIC VIRUS GENOME 155

Koonin, E. V., and Dolja, V. V. (1993). Evolution and taxonomy of positive- Riviere, C. J., and Rochon, D. M. (1990). Nucleotide sequence and strand RNA viruses: Implications of comparative analysis of amino genomic organization of melon necrotic spot virus. J. Gen. Virol. 71, acid sequences. Crit. Rev. Biochem. Mol. Biol. 28, 375±430. 1887±1896. Kozak, M. (1995). Adherence to the first-AUG rule when a second AUG Rubino, L., Russo, M., and Martelli, G. P. (1995). Sequence analysis of codon follows closely upon the first. Proc. Natl. Acad. Sci. USA 92, pothos latent virus genomic RNA. J. Gen. Virol. 76, 2835±2839. 2662±2666. Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989). ``Molecular Cloning: Kozak, M. (1996). Interpreting cDNA sequences: Some insights from A Laboratory Manual.'' Cold Spring Harbor Laboratory Press, Cold studies on translation. Mamm. Genome 7, 563±574. Spring Harbor, NY. Kyte, J., and Doolittle, R. F. (1982). A simple method for displaying the Schagger, H., and von Jagow, G. (1987). Tricine-sodium dodecyl sulfate± hydropathic character of a protein. J. Mol. Biol. 157, 105±132. polyacrylamide gel electrophoresis for the separation of proteins in Lommel, S. A., Kendall, T. L., Siu, N. F., and Nutter, R. C. (1991). the range from 1 to 100 kDa. Anal. Biochem. 166, 368±379. Characterization of maize chlorotic mottle virus. Phytopathology 81, Scholthof, H. B., and Jackson, A. O. (1997). The enigma of pX: A host- 819±823. dependent cis-acting element with variable effects on tombusvirus Lutcke, H. A., Chow, K. C., Mickel, F. S., Moss, K. A., Kern, H. F., and Scheele, G. A. (1987). Selection of AUG initiation codons differs in RNA accumulation. Virology 237, 56±65. plants and animals. EMBO J. 6, 43±48. Scholthof, H. B., Morris, T. J., and Jackson, A. O. (1993). The capsid Masuta, C., Zuidema, D., Hunter, B. G., Heaton, L. A., Sopher, D. S., and protein gene of tomato bushy stunt virus is dispensable for systemic Jackson, A. O. (1987). Analysis of the genome of satellite panicum movement and can be replaced for localized expression of foreign mosaic virus. Virology 159, 329±338. genes. Mol. Plant-Microbe Interact. 6, 309±322. McCoy, N. L., Toler, R. W., and Amador, J. (1969). St. Augustine decline Scholthof, K.-B. G., Scholthof, H. B., and Jackson, A. O. (1995). The (SAD)ÐA virus disease of St. Augustine grass. Plant Dis. Rep. 53, tomato bushy stunt virus replicase proteins are coordinately ex- 955±958. pressed and membrane associated. Virology 208, 365±369. Mohan, B. R., Dinesh-Kumar, S. P., and Miller, W. A. (1995). Genes and Shirako, Y., and Wilson, T. M. A. (1993). Complete nucleotide sequence cis-acting sequences involved in replication of barley yellow dwarf and organization of the bipartite RNA genome of soil-borne wheat virus-PAV RNA. Virology 212, 186±195. mosaic virus. Virology 195, 16±32. Monis, J., Sopher, D. S., and Jackson, A. O. (1992). Biologically active Sill, W. H., and Pickett, R. C. (1957). A new disease of switchgrass, cDNA clones of panicum mosaic virus satellites. Phytopathology 82, Panicum virgatum L. Plant Dis. Rep. 41, 241±249. 1175. Skotnicki, M. L., Mackenzie, A. M., Torronen, M., and Gibbs, A. J. (1993). Niblett, C. L., and Paulsen, A. Q. (1975). Purification and further char- The genomic sequence of cardamine chlorotic fleck carmovirus. acterization of panicum mosaic virus. Phytopathology 65, 1157±1160. J. Gen. Virol. 74, 1933±1937. Niblett, C. L., Paulsen, A. Q., and Toler, R. W. (1977). ``Panicum Mosaic White, K. A., Skuzeski, J. M., Li, W., Wei, N., and Morris, T. J. (1995). Virus.'' Descriptions of Plant Viruses, No. 177, Commonwealth Myco- Immunodetection, expression strategy and complementation of tur- logical Institute and Association of Applied Biologists, Kew, England. nip crinkle virus p28 and p88 replication components. Virology 211, Nutter, R. C., Scheets, K., Panganiban, L. C., and Lommel, S. A. (1989). 525±534. The complete nucleotide sequence of the maize chlorotic mottle virus genome. Nucleic Acids Res. 17, 3163±3177. You, X. J., Kim, J. W., Stuart, G. W., and Bozarth, R. F. (1995). The Offei, S. K., Coffin, R. S., and Coutts, R. H. A. (1995). The tobacco nucleotide sequence of cowpea mottle virus and its assignment to necrosis virus p7a protein is a nucleic acid-binding protein. J. Gen. the genus Carmovirus. J. Gen. Virol. 76, 2841±2845. Virol. 76, 1493±1496. Zhou, H., and Jackson, A. O. (1996). Expression of the barley stripe Paul, H. L., Querfurth, G., and Huth, W. (1980). Serological studies on the mosaic virus RNA ␤ ``triple gene block.'' Virology 216, 367±379. relationships of some isometric viruses of Gramineae. J. Gen. Virol. Zuidema, D., Heaton, L. A., Hanau, R., and Jackson, A. O. (1986). 47, 67±77. Detection and sequence of plus-strand leader RNA of sonchus Powell, A. M. (1994). ``Grasses of the Trans-Pecos and Adjacent Areas,'' yellow net virus, a plant rhabdovirus. Proc. Natl. Acad. Sci. USA 83, 1st ed. Univ. of Texas Press, Austin. 5019±5023.