INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Bamboo mosaic virus (BaMV), a member of the potexvirus group of plant viruses, has a flexuous rod-shaped morphology (15). It causes a serious mosaic disease on bamboo, especially in Ma Chu (Dendrocalamus latiflorus) and Green Bamboo (Bambusa oldhamii) in Taiwan. The BaMV genome consists of a single-stranded positive-sense RNA molecule with a 5' cap structure and a 3' poly(A) tail. The entire nucleotide sequences of two isolates, O and V, comprising 6,366 nucleotides [excluding the 3' poly(A) tail], have been determined (17, 31). Five major open reading frames (ORFs 1 to 5) encode polypeptides of 155, 28, 13, 6, and 25 kDa, respectively (Fig. 1). Two major subgenomic RNAs of 2.0 and 1.0 kb in length are not encapsidated (16).

View larger version (10K): [in this window] [in a new window]   FIG. 1.   Diagram of the BaMV strain O infectious clone pBaMV-O, a pUC-derived clone containing a full-length insert of BaMV genomic cDNA. The BaMV insert, with its 5' and 3' termini and significant restriction sites as marked, is flanked at the 5' end by a bacteriophage T7 promoter and at the 3' end by a unique BamHI site. Five major ORFs encoded by BaMV RNA, ORF1 to ORF 5, are shown under the cDNA clone. When this linearized template is transcribed with T7 RNA polymerase in the presence of cap analogue, full-length capped transcripts beginning at the viral 5' terminus as indicated in the lower panel are synthesized.

Constructing a full-length infectious cDNA clone is fundamental for further understanding of the functions of the products encoded by each ORF (3, 5, 10), the cis-elements required for viral replication (28, 29, 31), and the relationships between the viral RNAs and host symptom development (12). Infectious RNA transcripts have been produced from cDNA clones of several RNA viruses isolated from bacteria, animals, and plants (reviewed in reference 4). The strategies employed to obtain infectious clones and to determine the parameters that affected the infectivities of those RNA transcripts have also been described in detail (4).

The sequences of the 3' untranslated region (UTR) of positive-sense RNA viruses have been reported to play an important role in RNA amplification, whether the end structure is a tRNA-like structure (25, 27, 28) or a poly(A) tail (8, 21, 26, 30). As with eukaryotic mRNAs, the poly(A) tail of potexviral RNA is expected to play a general role in RNA stabilization and in translation initiation, in addition to possibly providing recognition elements for the replicase complex. Experiments with white clover mosaic potexvirus (WClMV) RNA (10) showed some marginal infectivity in plants with a completely deleted poly(A) tail. However, in the cases of hepatitis A virus (14), cowpea mosaic virus (8), poliovirus (21), and encephalomyocarditis virus (6), deletion of the poly(A) tail led to a loss of infectivity.

Here we report the cloning and generation of infectious RNA transcripts from a full-length cDNA clone of BaMV and the use of this clone to characterize the role of the 3' poly(A) tail in the replication of BaMV RNA. We have used enzymatic and chemical structural mapping techniques to define a potential pseudoknot in the 3' UTR that involves some nucleotides of the poly(A) tail. Disrupting and compensating mutations in one stem of the predicted pseudoknot support a function for this structural element. The effects of mutations in the 3' terminus were consistent with a functional role for the putative pseudoknot in viral amplication.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

BaMV-O infectious cDNA and chimeric viral genomic cDNA construction. Overlapping BaMV cDNA clones used for sequencing (17) were used to construct the full-length cDNA clone. The 5' primer d(GCTCTAGATAATACGACTCACTATAGAAAACCACTCCAAACGAA), containing an XbaI site (italics) and T7 promoter (underlined), and the 5'-BaMV sequence, and downstream primer d(ATCTCCTCTTCTCCGGAA) priming at nucleotide position 421 (17) were used to generate a PCR fragment to provide the 5' region of the full-length clone of the BaMV-O genome.

In vitro transcription and inoculation with protoplasts. Plasmid DNAs were prepared from 50-ml bacterial cultures, and the mutated sequences were confirmed by sequencing. Capped genomic transcripts labeled with [-³²P]UTP (0.1 Ci/mmol) were generated with T7 RNA polymerase from the BamHI-linearized plasmid templates containing wild-type and chimeric viral genomic cDNAs and analyzed as described previously (29) prior to inoculation. Chenopodium quinoa, Nicotiana benthamiana, and N. tabacum were grown in a greenhouse under natural light or in a growth chamber under a 16-h day length at 28°C. Mesophyll protoplasts (4 × 10⁵) of C. quinoa, N. benthamiana, and N. tabacum were isolated, inoculated with 5 µg of transcript RNAs, and incubated at 25°C for 48 h under constant illumination as described previously (27, 28).

Analysis of viral products by Western and Northern blotting. The levels of coat protein in harvested protoplasts were analyzed in Western blots with anti-BaMV capsid protein serum as a primary antibody (15) and horseradish peroxidase-labeled secondary antibody and the chromogenic substrate 4-chloro-1-naphthol as described previously (29). Results were quantitated by scanning densitometry (Intelligent Quantifier; Bioimage). RNAs were extracted from protoplasts, glyoxalated, electrophoresed through 1% agarose, and transferred to nylon membranes as described previously (27). The hybridization probe was a ³²P-labelled RNA transcript complementary to 0.6 kb at the 3' end of BaMV RNA (17).

Prediction of the BaMV 3' UTR structure. To predict the secondary structure of the BaMV 3' UTR, we employed the STAR (structural analysis of RNA) computer program (1). STAR is able to predict not only secondary structures but also aspects of tertiary structures, particularly pseudoknots.

Preparation of end-labeled RNA transcripts for structural mapping. BaMV/6282 RNA (an RNA whose 5' end is at nucleotide position 6282 of the BaMV RNA) was transcribed from PCR-generated DNA amplified from linearized pBaMV-OM (original cDNA clone containing 32 3' adenylate residues and 23 nonviral nucleotides) with the upstream primer T7/6282 d(TAATACGACTCACTATAGGGTTTACACGGACT) located in nucleotide position 6282 of the BaMV genomic sequence (17) and T7 promoter (underlined) and the downstream primer d(CGGCAACGAAGGTACCATGG) located at the nonviral region downstream of the poly(A) sequences. Transcripts were separated on a 5% polyacrylamide gel and eluted by soaking the sliced bands in elution buffer (0.5 M ammonium acetate, 10 mM magnesium acetate, 1 mM EDTA, 0.1% sodium dodecyl sulfate with shaking at 37°C overnight, followed by phenol-chloroform extraction and ethanol precipitation.

Structural mapping with ribonucleases and chemicals. To localize the cleavage sites within the RNA structure, the size of labeled RNA cleavage fragments was determined by electrophoretic separation on denaturing (7 M urea) polyacrylamide gels. For each assay, a control without enzyme or chemical treatment was run in parallel. DNA sequencing reactions were used as markers to locate the cleavage sites. Primer (5'-GTTTACACGGACT-3'), which corresponds to the 5' ends of the labelled transcripts (except for the lack of two 5' guanine residues) was used in a dideoxy sequencing reaction to provide marker fragments.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Full-length cDNA cloning and mutant construction. The full-length cDNA clone of BaMV-O, pBL, was constructed by connecting five overlapping cDNA clones previously used to sequence the genome (17) into the BamHI site of the pUC119 vector. In order to obtain a unique restriction site at each end of the full-length cDNA, the BamHI site at the 5' end was replaced with an XbaI site. A primer annealing to nucleotides 404 to 421 of BaMV-O RNA (17) was used for reverse transcription, and the 5'-end primer containing a T7 promoter and an XbaI restriction site was used to synthesize the second strand. This newly synthesized 5' fragment was subcloned into pBL by using 5' XbaI and BspEI (nucleotide 406) restriction sites. The resulting plasmid, pBaMV-OM, contains a unique BamHI site at the 3' end of the viral sequence that can be used for linearization prior to runoff transcription or for subcloning 3'-end mutations (Fig. 1). Since this full-length cDNA clone is mainly constructed from previous clones used for identifying the genomic sequences, the transcripts derived from those clones inherited 32 adenine residues plus 23 nonviral nucleotides at the 3' end when the template was linearized with the BamHI site. To obtain transcripts with fewer nonviral nucleotides, the 3' end of BaMV-OM (32 adenines plus 23 nonviral nucleotides) was replaced with 25 adenines and 5 nonviral nucleotides derived from the restriction site BamHI. The transcript derived from the resulting plasmid, pBaMV-O, was then designated as wild type.

View larger version (37K): [in this window] [in a new window]   FIG. 2.   Replication of BaMV RNAs in N. benthamiana protoplasts. Representative experiments that have contributed to the quantitative data in Table 1 are shown. Protoplasts (4 × 105 cells) were inoculated with 1 µg of virion RNA or 5 µg of transcripts from pBaMV-O and its derivatives (indicated at the top of each lane of panels A, B, and C [Western immunoblots] and panel D [Northern blot]). Protoplasts (2 × 105 cells) were harvested 48 h after inoculation for either Western or Northern analysis with the indicated derivatives of BaMV RNA. (A, B, and C) Extracts were separated on a 14% polyacrylamide-sodium dodecyl sulfate gel, blotted, and probed with anti-BaMV coat protein serum. The blot was developed by using horseradish peroxidase-linked second antibodies and 5-chloro-1-naphthol color reagent. (D) Detection of viral genomic (G; 6.4 kb) and two subgenomic (SG; 2.0 and 1.0 kb) RNAs in by Northern blotting. RNAs were probed with a 32P-labelled RNA transcript complementary to 0.6 kb at the 3' end of the genomic RNA.

Effect of poly(A) tail length. To determine the role of the poly(A) tail in the infectivity of BaMV RNA, a series of mutants with changes at the 3' end were constructed. Transcripts lacking a poly(A) tail or with 4, 7, 10, 15, 18, and 22 3' adenines, were generated. Like the wild type containing 25 3' adenines, each RNA terminated with 5 nonviral nucleotides (GGAUC) at the end of the transcripts when the template was linearized with BamHI. To test whether another type of 3'-stabilizing sequence could replace the poly(A) tail, the infectivity of BaMV-O/TYtRNA RNA, in which the TYMV tRNA-like structure replaces the poly(A) tail, was constructed. The tRNA-like structure of TYMV can be aminoacylated at its 3' end with valine (7), and tRNA-like structures are thought to function as telomeres for stable retention of the genomic 3' end (20).

Putative pseudoknot involving the poly(A) tail. To determine the structure of the 3' region of the genomic RNA, the 3' end of BaMV RNA was transcribed from a PCR-generated fragment containing a T7 promoter. Three guanine residues including two nonviral guanines at the very 5' end of the transcripts are inherited from the T7 promoter for more efficient transcription. To optimize the reaction conditions for each ribonuclease and DEPC, we tried several buffer systems such as "standard" structure probing buffer (13), RNase digestion buffer (28), TMK buffer (11), and sodium cacodylate buffer (9). The banding patterns generated by the RNase cleavage were the same among all of the buffers tested. However, the condition modified from the RNase digestion buffer (28) could resolve the banding pattern better and give less background than those of the other buffers (data not shown). For each ribonuclease tested, we used a serial dilution of salt or enzyme concentrations to optimize the condition for each cleavage reaction. Selected reactions for each ribonuclease and for DEPC are presented in Fig. 3 and 4. Comparison of the banding patterns of the lanes labeled T2 (single-stranded specific RNase T₂) to those of the lanes labeled V1 (double-stranded specific RNase V1) showed clearly complementary banding. The deduced basepairing matches the prediction of the computer algorithm STAR (1), with a classical pseudoknot structure (nucleotides 23 to 12, the first A residue of the poly(A) tail connected to the 3' UTR is numbered 1) comprising at least 13 adenylate residues localized downstream of a major stem-loop (Fig. 5).

View larger version (77K): [in this window] [in a new window]   FIG. 3.   Chemical and enzymatic probing of the BaMV 3' UTR around the putative pseudoknot region. The RNA transcripts were 5' end labeled and treated with RNase T1 (lanes T1), RNase A (lanes A), RNase T2 (lanes T2), cobra venom nuclease V1 (lanes V1), and DEPC (lane DEPC). The amount of enzyme or chemical used in each reaction is indicated above each lane. The cleavage products were resolved on a 6% sequencing gel. Lane c corresponds to the control untreated RNA sample, whereas lane a corresponds to aniline-treated RNA used for sizing the cleavage fragments. Lanes G, A, T, and C correspond to the sequencing reaction as markers for identifying the cleavage sites.

View larger version (48K): [in this window] [in a new window]   FIG. 4.   Enzymatic probing of the BaMV 3' UTR corresponding to the pseudoknot region. The RNA transcripts were 5'-end labeled and treated with RNase T1 (lane T1), RNase A (lane A), and RNase T2 (lane T2). The concentration of enzyme used in each reaction is indicated above each lane. The cleavage products were resolved on an 8% sequencing gel. Lane c corresponds to the untreated RNA sample. Lanes G, A, T, and C contain DNA products from dideoxy sequencing reactions that are used as markers; note that the primer used for the dideoxy sequencing reaction was two residues shorter at the 5' end than the otherwise identical primer used for structure probing; there is thus a two-nucleotide offset between the markers and the products of structure probing experiments.

View larger version (17K): [in this window] [in a new window]   FIG. 5.   Proposed folding of the 3'-end 81 nucleotides of the UTR plus the poly(A) tail of BaMV RNA. Nucleotides are numbered from the 3'-end cytosine just upstream of the poly(A) tail. The structure has been deduced from chemical and enzymatic probing experiments of the predicted structure of BaMV 3' UTR assisted by computer predictions made with the STAR program (1). A summary of the cleavages or modifications induced by enzymatic or chemical probes is indicated by symbols explained in the figure.

5

7

Effects of mutations in the stem of the putative pseudoknot. Results from structural probing indicated that the pseudoknot might exist in the buffer condition tested. However, we cannot rule out the existence of an alternative structure. To further investigate the pseudoknotted structure involved in BaMV RNA replication, mutations were introduced into the pseudoknot to destabilize stem S1 or to restore it with compensatory changes. Mutants BaMV-O/G¹, -O/C¹⁹, -O/C¹⁸, and -O/C¹⁸C¹⁹ were expected to disrupt stem S1 of the putative pseudoknot. These mutants accumulated to levels 12 to 17% of the wild-type levels, as measured by coat protein accumulation (Table 2). Combinations of the BaMV-O/G¹C¹⁹ or BaMV-O/G²C¹⁸ mutations, expected to restore stem S1 and pseudoknot formation, resulted in amplification to 75 and 56%, respectively. However, the compensatory mutants could not accumulate the coat protein level to that of the wild type, implying that the primary sequence is involved in the BaMV RNA replication as well as the secondary structure. These data provide strong evidence that maintaining the stem formation of the pseudoknot is important for BaMV RNA replication.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

An infectious cDNA clone is an important tool for studying the RNA viruses at a molecular level. Although the original construct of the full-length cDNA of BaMV, pBaMV-OM, has 23 nonviral extra nucleotides after 32 adenine residues at the very 3' end, it still has substantial replication efficiency in N. benthamiana protoplasts (only fivefold less than the virion RNAs) (Fig. 2A). It has been reported that 28 (32) and even up to 198 (2) nonviral residues at the very 3' end of the transcripts did not significantly affect infectivity. The addition of up to 2,434 nonviral nucleotides at the 3' end of transcripts derived from the papaya mosaic virus infectious cDNA clone decreased but did not abolish infectivity completely (24). However, to prevent any possible artifact of these long nonviral nucleotides at the very 3' end of the genome, the end was replaced with 25 adenylate residues and 5 nonviral nucleotides (GGAUC) derived from the BamHI restriction site. The resulting transcript, BaMV-O, was then designated as wild type.

It has been reported that the transcripts of WClMV with shortened poly(A) tails were less infectious than wild type; however, the transcripts without any 3'-terminal (A) residue still produced three lesions while the wild type (74 A residues) or mutants with 27 or 10 A residues at the 3' end showed 125, 105, or 70 local lesions, respectively, on the inoculated leaves of cowpea plants (10). In contrast to these observations with WClMV, we could not detect any amplification of mutants with no or short poly(A) tails despite higher sensitivities than in the WClMV studies (Fig. 2B). These results suggested that the poly(A) tail may be involved in BaMV amplification, perhaps affecting stability as for mRNA in eukaryotic cells, or else playing a structural role in the recognition by the replicase complex, as in the case of hepatitis A virus, for which the poly(A) tail is part of a recognition element involved in viral replication (14). If the poly(A) tail of the BaMV RNA plays only a stabilizing role, then a tRNA-like structure at the 3' end might support a similar function. However, no signals could be detected in the protoplasts when inoculated with the mutant pBaMV-O/TYtRNA (Fig. 2B and Table 1).

Structural analysis of the 3' noncoding region of BaMV RNA showed a potential classic pseudoknot involving a number of adenylate residues of the poly(A) tail. Besides, a similar structure was found among the most stable structures predicted by MFOLD (program in the Genetics Computer Group sequence analysis software package [33]). If pseudoknot formation is required for BaMV RNA replication, the number of adenylate residues might be a critical factor. Transcripts with only 4 or 7 A residues are unable to form the proposed pseudoknot, but probably form a 6-bp stem with a large loop of 15 nucleotides. Although transcripts with 10 A residues can form a pseudoknot with a shorter S2 stem with the help of two wobble GU pairs, where the two guanine residues next to the 10 adenine residues were derived from BamHI cleavage site (GGAUC), no amplification could be detected in protoplasts. Transcripts with more than 15 A residues can form a stable pseudoknot but also have extra adenylate residues at the very 3' end that might be important for RNA stability in host cells. The transcripts of pBaMV-O/15A, which possessed a pseudoknot and seven unstructured adenylate residues at the end replicated to ca. 26% of the wild-type levels. Replication of transcripts derived from pBaMV-O/18A or -O/22A, which possessed a pseudoknot and 10 or 14 additional nucleotides amplified to 69 or 98%, respectively, the level of the wild type.

Each mutant with disruptions in S1 of the pseudoknot failed to accumulate viral products efficiently, whereas the compensatory mutants with the base-pairing restored could only amplify up to 75% of the level of the wild type. It is likely that the functional properties of this pseudoknot are conferred by both the secondary and primary structure. Based on the structural mapping, the sufficient length of adenylate residues required to maintain the downstream stem formation, and the compensatory mutational analysis of the upstream stem formation of the predicted pseudoknot structure, we conclude that the formation and maintenance of a stable pseudoknot at the 3' UTR of BaMV is important for viral RNA replication.