Restoration of the 3' End of Potyvirus RNA Derived from Poly(A)

Virology 265, 147–152 (1999) Article ID viro.1999.0027, available online at http://www.idealibrary.com on

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector Restoration of the 3Ј End of Potyvirus RNA Derived from Poly(A)-Deficient Infectious cDNA Clones

Yoko Tacahashi1 and Ichiro Uyeda2

Pathogen-Plant Interactions Group, Graduate School of Agriculture, Hokkaido University, Sapporo 060-8589, Japan Received July 1, 1999; returned to author for revision July 14, 1999; accepted September 17, 1999

Poly(A)-deficient full-length cDNA clones of clover yellow vein virus (ClYVV), a member of the genus Potyvirus, were found to be infectious when expressed from the CaMV 35S promoter. The poly(A) tail was replaced with different short sequences and the infectivities of the cDNA constructs were examined. Although the infectivity of the plasmid varied depending on the sequences introduced, all the constructs were infectious. In all cases, progeny viral RNAs from the cDNA clones had an authentic viral sequence at their 3Ј regions with poly(A) tails and the downstream nonviral sequences were completely lost. However, two minor mutations, a two-nucleotide deletion at the 3Ј end and a single-nucleotide addition at the second nucleotide position downstream of the poly(A) site, were also observed. The clones of the viral (Ϫ) strand RNAs had poly(U) tracts at their 5Ј ends, suggesting that their synthesis is primed by the poly(U) sequence. It furthermore suggests that the mutations were introduced during or after primary transcription from the cDNA and were maintained during authentic viral replication. Although the mechanism involved is not known, recovery of the poly(A) tail is an essential step for maintaining the infectivity of the viral cDNAs. © 1999 Academic Press Key Words: infectious cDNA; Potyvirus; clover yellow vein virus; poly(A) tail.

INTRODUCTION to viral RNA lacking the poly(A) tail. The (Ϫ) strand of Coronavirus requires a poly(A) tail to initiate its synthesis The 3Ј terminal region plays an important role in ge- (Lin et al., 1994). nome replication and gene expression for viruses having Although a poly(A) tail is essential for viral replication, a(ϩ) ssRNA as a genome. The poly(A) tail at the 3Ј end some viruses were still infectious when viral RNAs with- of the genome of some viruses is assumed to have a out a tail were introduced into host cells (Jupin et al., protective function against exoribonuclease attack and is 1990; Guilford et al., 1991; Hill et al., 1997). In all cases, involved in translational regulation, similar to the func- tions associated with the tails of cellular mRNAs (Blee- the viral progeny had restored the long poly(A) tail. man and Parker, 1995). Little is known about the effect of poly(A) tail length on Besides cellular mRNA mimicry, the poly(A) tail also the infectivity of the viruses in the genus Potyvirus.Po- ϩ ϳ plays a crucial role in (Ϫ) strand viral RNA synthesis. The tyviruses have ( ) ssRNA 10 kb long as a genome, with Ј (Ϫ) strand RNA of many poly(A)-tailed viruses has poly(U) a protein (VPg) linked to its 5 end and a poly(A) tail at the Ј at its 5Ј end (Spector and Baltimore, 1975; Sawicki and 3 end. Infectious cDNA clones of potyviruses fused to Gomatos, 1976; Dolja et al., 1987; Hofmann and Brian, the cauliflower mosaic virus (CaMV) 35S promoter con- Ͼ 1991), suggesting that the poly(A) tail is essential for tained relatively long poly(A) tails ( 50 nucleotides in poly(U)-primed (Ϫ) strand synthesis. Studies involving length) (Maiss et al., 1992; Gal-On et al., 1995; Johansen, viral RNAs transcribed in vitro from cDNAs introduced 1996; Jakab et al., 1997). However, a clover yellow vein into host cells to determine the effect of the length of the virus (ClYVV) cDNA clone that contained only 10 A resi- poly(A) tail showed that a shortened poly(A) tail de- dues under the control of the 35S promoter was also creased or diminished infectivity (Spector and Baltimore, highly infectious (Takahashi et al., 1997). The effect of 1975; Eggen et al., 1989; Sarnow, 1989; Guilford et al., poly(A) tail length on the infectivity of potyvirus infectious 1991). Cui et al, (1993) showed that Encephalomyocardi- cDNAs or transcripts has not been systematically inves- tis virus RNA polymerase binds in vitro specifically to a 3Ј tigated. This prompted us to investigate the threshold noncoding region of the poly(A)-tailed viral RNA, but not number of A residues necessary for maintaining the infectivity of a cDNA clone. In this study, the effect of the poly(A) tail on the infectivity of ClYVV was examined using infectious cDNAs. We introduced mutations in the 1 Present address: Department of Molecular Biology and Microbiol- ogy, Tufts University School of Medicine, Boston, MA 02111. poly(A) region and tested their effects on infectivity; the 2 To whom reprint requests should be addressed. Fax: ϩ81–11-706– poly(A) regions of the genomes of the recovered viral 2483. E-mail: [email protected]. progeny were sequenced.

RESULTS Plant infection We previously reported that pClYVV had high infectivity and that it was infectious at a concentration of 0.5 ␮g/ml, although the nos terminator was absent and its poly(A) tail contained only 10 A residues (Takahashi et al., 1997). We then decided to test the effects of a shorter poly(A) tail and various oligonucleotide sequences downstream from the poly(A) site on the infectivities of cDNA constructs. Figure 1a shows the construction of the plasmid containing the full-length ClYVV cDNA clone (positions 1–9584). The full-length cDNA clone was placed under the control of the CaMV 35S promoter and upstream of the BamHI, HindIII, and EcoRI restriction sites and the pUC8 sequence. The sequences downstream from the poly(A) site of each plasmid DNA are shown in Fig. 1b. The infectivities of the nine constructs were tested in broad bean plants by both mechanical inoculation and particle bombardment (Table 1). Almost all the pClYVV-inoculated plants were infected, and the upper leaves developed mosaic and necrosis symptoms 4–7 days postinoculation. The infectivity of pClYVV-PA5 was slightly lower than that of pClYVV, although they were the same as those seen with pClYVV. pClYVV⌬PA was still infectious, but the infectivity was reduced to about one-sixth of that of pClYVV. The symptoms appeared later than those in plants inoculated with pClYVV. Viral progeny cultures recovered from plants inoculated with pClYVV⌬PA were capable of inducing wild-type symptoms with no observed delays in symptom appearance or development. The location of T residues in the poly(A) tail affected the number of infected plants. The pClYVV PAmut series FIG. 1. Construction of pClYVV series. (a) Schematic representation consisted of plasmids with T residues in the poly(A) tract. of pClYVV. The full-length ClYVV cDNA is fused to the CaMV 35S pClYVV PAmut 2 and pClYVV PAmut 3 had infectivities promoter (35S). The open reading frame is shaded. P1, protein 1; similar to that of pClYVV. The inserted T residues had no HC-pro, helper component-protease; P3, protein 3; CI, cylindrical inclu- effect on the infectivities of these plasmids. On the other sion protein; VPg, genome-linked protein; NIa-pro, nuclear inclusion hand, the infectivities of pClYVV PAmut 1 and pClYVV protein a-protease; NIb, nuclear inclusion protein b; CP, capsid protein. (b) Sequences surrounding the BamHI site. The viral sequence (3Ј PAmut 4 were lower than the infectivity of pClYVV but ⌬ noncoding region) and nonviral sequence are represented by capital higher than that of pClYVV PA. Symptom development letters and lower-case letters, respectively. The poly(A) sequence and was the same as that observed with pClYVV. the mutations in the poly(A) region are shown in bold letters. In pClYVV 3Јdup 1, the last five nucleotides, CGAGA, were duplicated. The infectivity of pClYVV 3Јdup 1 was Analysis of progeny RNA higher than that of pClYVV⌬PA, but lower than that of pClYVV. Symptoms developed in a manner similar to We next examined the nucleotide sequences of the those seen with pClYVV. In pClYVV 3Јdup 4, the last 10 3Ј-terminal regions of the viral progeny RNAs derived nucleotides, TAGAGCGAGA, were duplicated. The infec- from the various cDNA constructs. The 3Ј-terminal re- tivity of pClYVV 3Јdup 4 was higher than that of pClYVV gions of the progeny RNAs obtained from the infected 3Јdup 1. The symptoms again developed in the same plants were cloned. Analysis of the cDNA clones showed manner as pClYVV symptoms. It seems that the length of that each progeny RNA had a longer poly(A) tail than the the duplicated sequence enhances the infectivity of the parental plasmid. The lengths of the poly(A) chains of the plasmid. These results showed that the poly(A) tract was cDNA clones ranged from 20 to 60 nucleotides (Fig. 2). nonessential and that the context of sequences down- The poly(A) tails of progeny RNAs may be longer than stream of the poly(A) site affected plasmid infectivity. those of the cDNA clones because the efficiency of RESTORATION OF THE 3Ј-END SEQUENCE OF ClYVV RNA 149

TABLE 1 Detection and sequence analysis of (Ϫ) strand RNA Infectivity of pCIYVV Series Assayed on Broad Bean Plants Because the viral progeny RNAs were found to have a

b poly(A) tail longer than that of the cDNA constructs and Infected/inoculated Days required ⌬ for symptom even pClYVV PA-derived progenies had a poly(A) tail, a Plasmida Ac BCD E development poly(A) addition mechanism must operate before normal replication of ClYVV. Furthermore a mechanism must pCIYVV 4/5 6/6 4–7 exist to eliminate any sequences downstream of the pCIYVV-PA5 3/7 5/6 4/9 5/6 5–10 ⌬ poly(A) site in the cDNA so that the progeny RNAs are pCIYVV PA 1/7 1/6 1/9 1/6 2/14 7–10 Ј pCIYVV PAmut 1 6/15 7 able to recover the true 3 end with a poly(A) tail. As the pCIYVV PAmut 2 15/15 7–8 first step to understand the mechanisms, we next ana- pCIYVV PAmut 3 11/13 7–8 lyzed the 3Ј end structure of the (Ϫ) strand RNA. pCIYVV PAmut 4 5/14 8–9 The 5Ј end of the (Ϫ) strand RNA was first amplified Ј pCIYVV 3 dup 1 4/15 6–10 using RT–PCR with sequence-specific primers. The pCIYVV 3Јdup 4 10/14 7–10 246-bp fragments from a pClYVV-infected plant were a Inoculation with circular formed plasmids. amplified, but not those from a healthy plant. This result b Number of plants showing symptoms (infected)/number of plants indicates that the (Ϫ) strand RNA has a sequence com- inoculated. Ј c plementary to the 3 noncoding region of the positive- A–E are independent experiments. A, B, C, and E are the mechan- strand RNA. The fragments were amplified from both the ical inoculation tests, and D is the particle bombardment test. pellet and the supernatant fractions from 2 M lithium chloride precipitation. It is not known whether the ma- jority of the in vivo (Ϫ) strand RNA population exists in Ј double-stranded or single-stranded form. dsRNA might amplification primed far from the 3 noncoding region be an RF artifact of the RNA extraction. was lower in the PCR step. All progeny derived from the The 5Ј end of the (Ϫ) strand RNA was cloned by cDNA constructs had poly(A) tails, suggesting that 5ЈRACE and the sequence was analyzed. As shown in poly(A) addition proceeds in a template-independent Fig. 3, the 5Ј-end sequence begins at Tn in all 12 clones manner. sequenced. This result indicates that the (Ϫ) strand RNA The sequences around the poly(A) sites were analyzed was polyuridylated at its 5Ј end. (Table 2). With few exceptions, the processing of the 3Ј end of the progeny RNA was precise, and the additional nonviral sequence present downstream in the plasmid was completely removed. Three clones of the 3Ј end of pClYVV-derived progeny RNA from one plant were sequenced; two of these had the authentic viral sequence and one had a mutant sequence. Table 2 shows this mutant sequence, in which either two nucleotides, AG, are deleted, or an A replaces the last G. This result suggests that this pClYVV-derived progeny RNA was heterogeneous and that the new mutation did not affect RNA replication. The 3Ј end of pClYVV-PA5-derived progeny RNA had the authentic sequence. The 3Ј end of pClYVV⌬PA-derived progeny RNAs had two types of the sequence. Three clones from one infected plant had the authentic sequence and all 10 clones from the other infected plant had a mutation in which a C residue was inserted at the second position in the poly(A) tail. There was no heterogeneity in the viral RNA population of either plant. cDNA clones of the progeny from two infected plants were obtained for both the pClYVV PAmut and pClYVV 3Јdup series. All the clones sequenced had the authentic viral sequence and the insertions or duplications introduced into the parental plasmids FIG. 2. Presence of the poly(A) tail in pClYVV⌬PA-derived progeny downstream of the poly(A) site were completely lost RNA. The cDNA clone to 3Ј-end region of pClYVV⌬PA-derived progeny (Table 2). RNA are sequenced. The number of A residues is indicated on the left. 150 TACAHASHI AND UYEDA

TABLE 2 3Ј-End Sequence of Progeny RNA from Each Plasmid

Progeny sequence Total No. of No. of the clones analyzed No. of infected Plasmid . . .AUAGAGCGAGAn (wt) sequenced from a single infected plant plants analysed

... pCIYVV AUAGAGCGAGAn 2 31 ...AUAGAGCGAn 1 pCIYVV-PA5 ...AUAGAGCGAGAn 6 6 1 pCIYVV⌬PA ...AUAGAGCGAGAn 3 3 1 ...AUAGAGCGAGACAn 10 10 1 pCIYVV PAmut 1 ...AUAGAGCGAGAn 12 6 2 pCIYVV PAmut 2 ...AUAGAGCGAGAn 12 6 2 pCIYVV PAmut 3 ...AUAGAGCGAGAn 12 6 2 pCIYVV PAmut 4 ...AUAGAGCGAGAn 10 5 2 pCIYVV 3Јdup 1 ...AUAGAGCGAGAn 10 5 2 pCIYVV 3Јdup 4 ...AUAGAGCGAGAn 11 5 or 6 2

DISCUSSION vitro transcribed RNA with altered poly(A) tails (Guilford et al., 1991; Hill et al., 1997). In vitro synthesized Sindbis This is the first report of cDNA clones of the potyvirus genome that lack the poly(A) tail and yet are still infec- viral RNA lacking a poly(A) tail but having five nucleo- Ј tious. The poly(A)-deficient, full-length cDNA clones gen- tides of nonviral sequence at its 3 end initiated infection; erated viral progeny RNAs with authentic poly(A) tail the poly(A) tail was restored and nonviral sequences sequences at their 3Ј-terminal regions, the downstream were eliminated (Hill et al., 1997). The specific infectivity nonviral sequences of the primary transcripts were com- was also 15-fold lower than the wild-type virus and CPE pletely lost. Although infectious cDNAs were used in this development was delayed. A reduction in infectivity re- study, the results are similar to those seen after infection sulting from a shortened poly(A) tail has also been re- with Sindbis and white clover mosaic virus (WClMV) in in ported for white clover mosaic virus (WClMV), a member of the genus Potexvirus (Guilford et al., 1991). A WClMV transcript terminating at the 3Ј end without a poly(A) tail or nonviral sequences produced 50-fold fewer lesions than a transcript with 74 A residues, whereas a transcript with 10 A residues produced only twofold fewer lesions. The poly(A) tail was also restored in the progeny. The presence of the poly(A) tail is required for the binding of viral RNA polymerase to the 3Ј end of picor- navirus RNA (Cui et al., 1993) and is essential for initia- tion of (Ϫ) strand synthesis of the Coronavirus RNA (Lin et al., 1994). Furthermore the (Ϫ) strand of many viruses with a poly(A) tail, including ClYVV, has a poly(U) tail at the 5Ј end (Spector and Baltimore, 1975; Sawicki and Gomatos, 1976; Dolja et al., 1987; Hofmann and Brian, 1991), and the poly(U) sequence primes the (Ϫ) strand synthesis of picornaviruses in vitro (Andrews and Balti- more, 1986; Sanker and Porter, 1991). If this is also the case for potyviruses, restoration of the poly(A) tail after inoculation with the pClYVV must occur by an as yet unknown mechanism before the primary transcripts be- come a template for (Ϫ) strand synthesis. And it is unlikely that the infective progeny were revertants gen- erated from degraded primary transcripts. If that was the FIG. 3. Demonstration of the presence of poly(U) at the 5Ј end of (Ϫ) case, the infectivities of the various constructs made by strand RNA. The sequence of the cDNA clone corresponding to the 5Ј introducing different oligosequences downstream of the end of the ([minus) strand RNA is shown. The cDNA is derived from the Ј wild-type viral RNA. The sequence is given from the 5Ј (top) to 3Ј 3 end would have been the same. However, the possi- (bottom) ends. bility that oligo(U) primes (Ϫ) strand synthesis at the RESTORATION OF THE 3Ј-END SEQUENCE OF ClYVV RNA 151 specific internal site of the primary transcript without upstream from the CP gene; an A was substituted for the base-pairing to poly(A) can not be ruled out. T at nucleotide 8518 to create an XbaI site), and primers The addition of a poly(A) tract to viral RNA was dem- containing each mutation (complementary to the 3Ј end onstrated for white clover mosaic virus (WClMV). WClMV of ClYVV followed by a BamHI site). The PCR products RNAs transcribed from cDNAs lacking a poly(A) tail were were digested with XbaI and BamHI and cloned into infectious and had the poly(A) tail restored. They favored pBluescript II SK- or pHSG298 (Takara Shuzo). These poly(A) addition by cytoplasmic poly(A) polymerase, and cDNA clones were sequenced using a Thermo Seque- the AAUAAA motif in the 3Ј terminal region appeared to nase fluorescent labeled primer cycle sequencing kit act as a polyadenylation signal. A search of the 3Ј non- with 7-deaza-dGTP (Amersham) to verify that the muta- coding regions of several potyviruses failed to detect a tions were introduced; they were then inserted into the consensus motif for the poly(A) signal. This, however, SpeI–BamHI-digested plasmid pClYVV⌬CP. pClYVV⌬CP does not rule out the possibility that potyviral genomes, was the plasmid in which both the CP and the 3Ј non- including the ClYVV genome, have a poly(A) signal rec- coding regions were deleted from pClYVV, and T re- ognized by host nuclear factors, because the canonical placed A at nucleotide 8523 to create a SpeI site up- AAUAAA sequence is not present in Ͼ50% of plant nu- stream from the CP gene. The resulting plasmids had the clear mRNAs (Joshi, 1997). same sequence as the parental pClYVV, except for the A poly(A) tract was also added to transcribed beet sequence of the poly(A) tail. necrotic yellow vein virus RNA 3 lacking a poly(A) tail (Jupin et al., 1990). However, a nonviral U-rich sequence Plant inoculation Ј was introduced between the 3 end and the poly(A) tail in The plasmids were prepared by alkaline miniprep and the progeny RNA 3. Although it was proposed that the resuspended in TE buffer. Twenty microliters (8–10 ␮g) of restoration of the poly(A) tail was due to the action of each of the circular-form plasmids was inoculated onto either a viral or host poly(A) polymerase, a different Carborundum-dusted leaves of 10-day-old broad bean mechanism may operate in viruses that have a seg- seedlings. Successful plant infection was determined by Ј mented genome. The possibility that the 3 end is re- the development of symptoms in the upper leaves. paired by the poly(A)–RNA polymerase complex gener- Particle bombardment was performed according to the ated from RNA 1 or 2, which has recently been proposed method of Gal-On et al. (1997). The plasmids prepared by Ј for the repair mechanism of the 3 ends of turnip crinkle alkaline miniprep were mixed with gold particles and virus sat-RNA (Nagy et al., 1997), cannot be ruled out. calcium nitrate, pH 10.5, and the plasmids bound to the This study shows that the minimum number of A res- gold. Between 0.06 and 0.1 ␮g of plasmid was used for idues required for maintaining high infectivity is between each broad bean plant, shot onto the first fully expanded 5 and 10. The infectivities of mutants containing substi- leaves. The target leaves were 2 cm from the discharge tutions of T residues for A’s showed that those having nozzle. more than five continuous A residues (pClYVV PA5, -PAmut 2, and -PAmut 3) are more infectious than those Sequence analysis of the 3Ј ends of progeny RNAs having zero to two A residues (pClYVV⌬PA, -PAmut 1, and -PAmut 4). If cleavage and poly(A) addition occur with all Progeny virus particles derived from each plasmid of the constructs, then the downstream sequence also were partially purified from systemically infected leaves Ј affects the efficiency of the poly(A) addition process and and the RNA genomes were extracted. The 3 -terminal Ј leads to different infectivities. An alternative explanation region of each genome was amplified by 3 RACE. The is that the primary transcripts with a short poly(A) tail first-strand cDNA was synthesized using AMV RTase Ј (Ͼ5) can initiate (Ϫ) strand synthesis without cleavage (Seikagaku Kogyo) with primer 2, 5 GGCCACGCGTC- Ј and poly(A) addition. In this case, elongation of the GACTAGTACTTTTTTTTTTTTTTTTT3 . The first-strand poly(A) tail is not necessary, which is in agreement with cDNA was amplified using PCR with adapter primer 3, Ј Ј the fact that the poly(U) tract at the 5Ј end of Coronavirus 5 GGCCACGCGTCGACTAGTAC3 , and primer 1. The (Ϫ) strand RNA is only 8–20 nucleotides long (Hoffman PCR product was cloned into pBluescript II SK- and and Brian, 1991). sequenced using a Li-cor DNA sequencer, model 4000L. Detection of (Ϫ) strand RNA MATERIALS AND METHODS Broad bean tissue systemically infected with pClYVV cDNA modifications and plasmid constructions was homogenized in RNA extraction buffer (4.2 M gua- All mutations were created by PCR using TaKaRa Ex nidine thiocyanate, 0.5% sarcosyl, 25 mM sodium citrate, Taq (Takara Shuzo). The CP and 3Ј noncoding regions and 0.1 M 2-mercaptoethanol). The homogenate was were amplified from the template pClYVV (Takahashi et acidified and treated with H2O-saturated phenol. Total al., 1997) with primer 1, 5ЈAGTTGTGATCTAGAAGTGTA- RNA was precipitated with 2-propanol, and dsRNA was CAA3Ј (corresponding to genome positions 8510–8532, separated from ssRNA in 2 M lithium chloride. The pellet 152 TACAHASHI AND UYEDA was resuspended in TE buffer and the RNA was precip- encoding constructs by particle bombardment. J. Virol. Methods 64, itated with ethanol; both ssRNA and dsRNA were sub- 103–110, 1997. jected to RT–PCR. Gal-On, A., Meiri, E., Huet, H., Hua, W. J., Raccah, B., and Gaba, V. (1995). Particle bombardment drastically increases the infectivity of cloned After heat denaturation of the ssRNA and dsRNA, the DNA of zucchini yellow mosaic potyvirus. J. Gen. Virol. 76, 3223–3227. first-strand cDNA was synthesized with primer 4, 5ЈGAC- Guilford, P. J., Beck, D. L., and Forster, R. L. S. (1991). Influence of the GAGGAGAACACAGAA3Ј (corresponding to the positive- poly(A) tail and putative polyadenylation signal on the infectivity of strand RNA at positions 9332–9349). The ssDNA was white clover mosaic potexvirus. Virology 182, 61–67. then amplified by PCR using primer 5, 5ЈCCGGAAT- Hill, K. R., Hajjou, M., Hu, J. Y., and Raju, R. (1997). RNA-RNA recombi- Ј TCAGCAAATGATGTTAA CAGG3Ј (corresponding to the nation in sindbis virus: Roles of the 3 conserved motif, poly(A) tail, and nonviral sequences of template RNAs in polymerase recognition positive-strand RNA at positions 9358–9376 with an and template switching. J. Virol. 71, 2693–2704. EcoRI site at the 5Ј end), and primer 6, 5ЈCGCGGATC- Hofmann, M. A., and Brian, D. A. (1991). The 5Ј end of coronavirus CTCGCTCTATAAAGA TCAGA3Ј (complementary to the minus-strand RNAs contains a short poly(U) tract. J. Virol. 65, 6331– positive-strand RNA at positions 9565–9584 with a 6333. BamHI site at the 5Ј end). The PCR products were ana- Jakab, G., Droz, E., Brigneti, G., Baulcombe, D., and Malnoe, P. (1997). lyzed by agarose gel electrophoresis. Infectious in vivo and in vitro transcripts from a full-length cDNA clone of PVY-N605, a Swiss necrotic isolate of potato virus Y. J. Gen. Ј Ϫ Virol. 78, 3141–3145. Sequence analysis of the 5 end of ( ) strand RNA Johansen, I. E. (1996). Intron insertion facilitates amplification of cloned The 5Ј-terminal region of the (Ϫ) strand RNA was virus cDNA in Escherichia coli while biological activity is reestab- cloned using the 5ЈRACE method. The first-strand cDNA lished after transcription in vivo. Proc. Natl. Acad. Sci. USA 93, 12400–12405. was synthesized from the soluble fraction in 2 M lithium Joshi, C. P. (1997). Putative polyadenylation signals in nuclear genes of chloride using primer 4. A poly(C) tail was added to the higher plants: A compilation and analysis. Nucleic Acids Res. 15, ssDNA using terminal deoxynucleotidyl transferase 9627–9640. (Gibco BRL). The tailed first-strand cDNA was amplified Jupin, I., Bouzoubaa, S., Richards, K., Jonard, G., and Guiley, H. (1990). by PCR with primer 5 and an oligo d(G) primer with a Multiplication of beet necrotic yellow vein virus RNA 3 lacking a 3Ј SpeI site at its 5Ј end. The PCR product was cloned into poly(A) tail is accompanied by reappearance of the poly(A) tail and novel short U-rich tract preceding it. Virology 178, 281–284. pBluescript II SK- and sequenced. Lin, Y.-J., Liao, C.-L., and Lai, M. M. C. (1994). Identification of the cis-acting signal for minus-strand RNA synthesis of a murine coro- ACKNOWLEDGMENT navirus: Implications for the role of minus-strand RNA in RNA replication and transcription. J. Virol. 68, 8131–8140. We thank Dr. Chikara Masuta for helpful suggestions. Maiss, E., Timpe, U., Brisske-Rode, A., Lesemann, D.-E., and Casper, R. (1992). Infectious in vivo transcripts of a plum pox potyvirus full- REFERENCES length cDNA clone containing the cauliflower mosaic virus 35S promoter. J. Gen. Virol. 73, 709–713. Andrews, N. C., and Baltimore, D. (1986). Purification of a terminal Nagy, P. D., Carpenter, C. D., and Simon, A. E. (1997). A novel 3Ј-end uridylyltransferase that acts as host factor in the in vitro poliovirus repair mechanism in an RNA virus. Proc. Natl. Acad. Sci. USA 94, replicase reaction. Proc. Natl. Acad. Sci. USA 83, 221–225. 1113–1118. Beelman, D. A., and Parker, R. (1995). Degradation of mRNA in eu- karyotes. Cell 81, 179–183. Sanker, S., and Proter, A. G. (1991). Expression, purification, and prop- Cui, T., Sankar, S., and Portar, A. G. (1993). Binding of encephalomyo- erties of recombinant encephalomyocarditis virus RNA-dependent carditis virus RNA polymerase to the 3Ј-noncoding region of the viral RNA polymerase. J. Virol. 65, 2993–3000. Ј RNA is specific and requires the 3Ј-poly(A) tail. J. Biol. Chem. 286, Sarnow, P. (1989). Role of 3 -end sequences in infectivity of poliovirus 26093–26098. transcripts made in vitro. J. Virol. 63, 467–470. Dolja, V. V., Grama, D. P., Morozov, S. Y., and Atabekov, J. G. (1987). Sawicki, D. L., and Gomatos, P. J. (1976). Replication of semliki forest Potato virus X-related single- and double-stranded RNAs; character- virus: Polyadenylate in plus-strand RNA and polyuridylate in minus- ization and identification of terminal structures. FEBS Lett. 214, 308– strand RNA. J. Virol. 20, 446–464. 312, 1987. Spector, D. H., and Baltimore, D. (1975). Polyadenylic acid on poliovirus Eggen, R., Verver, J., Wellink, J., Pleij, K., van Kammen, A., and Goldback, RNA. IV. Poly(U) in replicative intermediate and double-stranded R. (1989). Analysis of sequences involved in cowpea mosaic virus RNA. Virology 67, 498–505. RNA replication using site-specific mutants. Virology 173, 456–464. Takahashi, Y., Takahashi, T., and Uyeda, I. (1997). A cDNA clone to Gal-On, A., Meiri, E., Elman, C., Gray, D. J., and Gaba, V. (1997). Simple clover yellow vein potyvirus genome is highly infectious. Virus Genes hand-held devices for the efficient infection of plants with viral- 14, 235–243.