Template-Independent Repair of the 3′ End of Cucumber Mosaic Virus Satellite RNA Controlled by RNAs 1 and 2 of Helper Virus

Abstract

RNA viruses which do not have a poly(A) tail or a tRNA-like structure for the protection of their vulnerable 3′ termini may have developed a different strategy to maintain their genome integrity. We provide evidence that deletions of up to 7 nucleotides from the 3′ terminus of cucumber mosaic cucumovirus (CMV) satellite RNA (satRNA) were repaired in planta in the presence of the helper virus (HV) CMV. Sequence comparison of 3′-end-repaired satRNA progenies, and of satRNA and HV RNA, suggested that the repair was not dependent on a viral template. The 3′ end of CMV satRNA lacking the last three cytosines was not repaired in planta in the presence of tomato aspermy cucumovirus (TAV), although TAV is an efficient helper for the replication of CMV satRNA. With use of pseudorecombinants constructed by the interchange of RNAs 1 and 2 of TAV and CMV, evidence was provided that the 3′-end repair was controlled by RNAs 1 and 2 of CMV, which encode subunits of the viral RNA replicase. These results, and the observation of short repeated sequences close to the 3′ terminus of repaired molecules, suggest that the HV replicase maintains the integrity of the satRNA genome, playing a role analogous to that of cellular telomerases.

The majority of plant viruses have messenger sense, single-stranded RNA genomes which are frequently exposed to the action of cellular nucleases during their life cycle. This is particularly true for the vulnerable 3′ terminus of replicating viral RNA, which has several essential functions for its replication (e.g., promotion of negative-strand synthesis, regulation of transcription, etc.) (7). Therefore, the existence of protective (active or passive) mechanisms which maintain the integrity of the 3′-terminal sequences would be very advantageous for viral RNAs. Basically, three major types of protective elements at the 3′ end of plant viral RNAs have been described elsewhere (16). (i) One type is heteropolymeric sequences ending in tRNA-like structures, which are characteristic for six groups of plant viruses (20, 24) and may represent the remnants of a primordial RNA world (23, 25). The tRNA-like structures at 3′ ends have important information for the initiation and regulation of RNA replication and for the aminoacylation of viral RNAs and may provide protection against the loss of 3′ sequences (39). Alterations in the 3′-terminal sequence of brome mosaic bromovirus RNAs, which have a tRNA-like structure, can be efficiently repaired in vivo in a telomere-like fashion, probably by a cellular nucleotidyltransferase (32). (ii) Another type is heteropolymeric sequences ending in poly(A) tails. A large number of plant RNA viruses have genomes terminating with a poly(A) tail that mimics the 3′ termini of cellular mRNAs and may also provide protection against 3′-end degradation. An important difference between viral RNAs and cellular mRNAs is that viral poly(A) tails are genetically derived, whereas mRNA tails are added posttranscriptionally through nontemplated RNA synthesis. (iii) The last type is heteropolymeric sequences with no particular structures or elements exhibiting intergroup homologies. In the last few years, a limited number of reports have been published about the 3′-end repair of some viruses having heteropolymeric sequences at the 3′ terminus. The first reports demonstrated that the truncated or altered 3′-terminal -CCC residues in genomic or satellite RNAs (satRNAs) of cymbidium ringspot tombusvirus (CymRSV) were restored in vivo (14, 15), and 3′-end repair was proposed to be catalyzed by either the virus-encoded polymerase or a host terminal transferase. Evidence has also been reported for 3′-end repair of truncated satRNAs of turnip crinkle carmovirus (TCV) (10, 11), and it was suggested that the 3′-end repair mechanism involved the production of 4- to 8-nucleotide (nt) oligoribonucleotides by abortive synthesis with the helper virus (HV) genome as a template (29). Furthermore, the deletion of up to 5 nt from the 3′ end of tobacco necrosis necrovirus (TNV) RNA was repaired in vivo (41). CymRSV, TCV, and TNV have 3′-end reparable RNA genomes and belong to different genera of the Tombusviridae, which suggests that a common repair mechanism may characterize this family of plant viruses.

The 3′ end of the satRNA of cucumber mosaic cucumovirus (CMV) ends with -CCC similarly to the above-mentioned viral RNAs, which could suggest a common repair mechanism for these RNAs. In this work, we show that in fact the altered 3′ end of CMV satRNA is repaired in planta. In addition, evidence is presented that the 3′-end repair of CMV satRNA is dependent on RNAs 1 and 2 of the HV, which encode subunits of the viral RNA replicase (31). The differences in primary sequence and structure between the 3′ termini of CMV genomic RNAs and satRNAs (2, 31, 34) do not support a template-dependent repair, as might also be the case for CymRSV satRNAs and genomic RNAs and for TNV RNA. These results and others presented in this study suggest that this 3′-end repair mechanism could be a widespread phenomenon among plant RNA viruses.

MATERIALS AND METHODS

Viruses.

The strains of CMV (Fny-CMV, in subgroup I of CMV strains [30], and Kin-CMV, in subgroup II), of tomato aspermy cucumovirus (TAV, 1-TAV, and V-TAV), and of CMV satRNA (Ix-satRNA) used in this work have been described elsewhere (6, 28, 33). Fny-CMV and Kin-CMV were derived from biologically active, full-length cDNA clones representing the genomic RNAs 1, 2, and 3, which have also been described elsewhere (6, 33). Plasmids pF109, pF209, and pF309 were the gift of Peter Palukaitis (Scottish Crop Research Institute, Dundee, United Kingdom), and pK1, pK2, and pK3 were the gift of David Baulcombe (The Sainsbury Laboratory, Norwich, United Kingdom). Biologically active full-length cDNA clones of RNAs 1 and 2 of V-TAV and of RNA 3 of 1-TAV have also been described elsewhere (27). Pseudorecombinant viruses with RNAs 1 and 2 of Fny-CMV or Kin-CMV and RNA 3 of 1-TAV (F₁F₂T₃ and K₁K₂T₃) or with RNAs 1 and 2 from V-TAV and RNA 3 from Kin-CMV (V₁V₂K₃) were also derived from these full-length cDNA clones. Ix-satRNA was derived from the full-length cDNA clone pIx5 (1).

Construction of 3′ deletion mutants of Ix-satRNA.

The biologically active full-length cDNA clone of Ix-satRNA, pIx5, was mutagenized by PCR with the following oligonucleotides as first primer: 5′ GATATCCTGCGGAGGAATGAT 3′, to obtain pIxΔ3C; 5′ GCCGGCGGAGGAATAATAGAC 3′, to obtain pIxΔ7; and 5′ CCCGGGAATAATAGACATT 3′, to obtain pIxΔ12. The mutagenizing oligonucleotides contained the complementary sequence of the modified 3′ terminus of Ix-satRNA. The underlined nucleotides indicate the introduced restriction sites used for linearization, EcoRV in pIxΔ3C, NaeI in pIxΔ7, and SmaI in pIxΔ12. In each PCR, the oligonucleotide 5′ GAATTCTAATACGACTCACTATAGTTTTGTTTGATGGAGA 3′, containing the first 17 bases (italics) of Ix-sat RNA, 17 bases of the bacteriophage T7 RNA polymerase promoter consensus sequence (boldface), and 6 nt contributing to the formation of an EcoRI I restriction site (underlined) at the 5′ end, was used as the second primer. PCR was carried out as described by Burgyán et al. (8), and its product was cloned into SmaI-digested and dephosphorylated pUC18. The resulting clones were checked by sequencing.

In vitro transcription and plant inoculation.

Biologically active full-length clones of CMV RNAs, TAV RNAs, or Ix-satRNA and its mutants were linearized with the appropriate restriction endonucleases for in vitro transcription (33). For viral RNAs, but not for satRNAs, transcription was in the presence of m⁷GpppG (New England Biolabs). Nicotiana tabacum L. cv. Xanthi-nc plants were inoculated with HV transcripts or a satRNA-free virus preparation (200 μg/ml) with or without satRNA transcripts (20 μg/ml) in 0.1 M Na₂HPO₄.

Analysis of wild-type (wt) and mutant Ix RNA progenies.

Total RNA from leaf tissue was isolated essentially according to the method described by Dalmay et al. (14). The presence of HV- and Ix-satRNA-related RNAs in total nucleic acid (TNA) extracts was assessed by Northern blot analysis. TNA samples (usually 500 ng) were denatured with formaldehyde and formamide, electrophoresed in formaldehyde-permeated 1.5% agarose gels, transferred to nylon membranes (Amersham), and probed with ³²P-labeled probes prepared by nick translation (36) of clone pF309, p13, or pIx5. Alternatively, ³²P-labeled cRNA obtained from pIx5 was used.

Sequence analysis of progeny molecules was performed on cDNA clones prepared as follows. TNA extracts were electrophoresed in 1.2% low-melting-point agarose (FMC Bioproducts), and the RNA species migrating in the position of wt Ix-satRNA was excised and purified as described by Burgyán et al. (9). The low-melting-point agarose-purified progeny RNA was polyadenylated with poly(A) polymerase (Amersham) and used as template for oligo(dT)-primed cDNA synthesis with the cDNA System Plus (Amersham) according to the manufacturer’s protocol. The double-stranded cDNA was cloned into SmaI-digested and dephosphorylated pUC18. The obtained clones were analyzed by sequencing them with modified T7 DNA polymerase (Sequenase; U.S. Biochemical).

Nucleotide sequence accession number.

The nucleotide sequence of the genomic RNA of TNV-D^H, the Hungarian isolate of strain D of TNV, was deposited in the EMBL and GenBank databases under accession no. U62546.

RESULTS

Replication of Ix-sat RNA mutants missing 3′-terminal residues.

To determine if the 3′ end of CMV satRNA (-AGGACCC) could be repaired in planta, a mutant derived from Ix-satRNA, with the 3′-most -CCC deleted, was prepared (designated IxΔ3C). The 3′ end of the in vitro-transcribed IxΔ3C RNA was -AGGAt, in which the last “t” was derived from the EcoRV site used to linearize the clone pIxΔ3C. Groups of 4 to 15 plants were inoculated with in vitro transcripts of IxΔ3C in the presence of the capped synthetic transcripts corresponding to RNAs 1, 2, and 3 of Fny-CMV. Two weeks after inoculation, total RNA was extracted from systemically infected leaves and subjected to Northern blot analysis. Control plants inoculated with Fny-CMV RNAs alone contained only the genomic and subgenomic RNAs of the HV (Fig. 1A, lanes 5 to 8) without any satRNA-like accumulation (Fig. 1B, lanes 5 to 8). However, each of the plants inoculated with the in vitro transcripts of IxΔ3C in the presence of the HV contained high levels of satRNA-like molecules (Fig. 1B, lanes 1 to 4). These satRNA-like molecules were gel purified, cloned, and sequenced. The sequence analysis of the obtained cDNA clones indicated that they were derived from the IxΔ3C transcripts, but the truncated 3′ end had been repaired. Repair of the 3′ end resulted in the restoration of the perfect (i.e., wt) sequence in 58% (11 of 19) of the sequenced clones. In 42% (8 of 19) of the clones, the 3′ end was not perfectly restored (Table 1). In some cases, only one or two cytosines were added to the truncated end (2 of 19 clones of each). Three different clones had the same -CCttt end (lowercase letters indicate nucleotides which were added in planta but are not present at the 3′ end of Ix- satRNA), but they contained deletions of 1, 2, and 3 nt, respectively, just upstream of the last two cytosines. Finally, one clone had the wt 3′-end sequence, with the -ggag sequence added to it. These results clearly show that a deletion of the -CCC end of Ix-satRNA can be efficiently restored in vivo. The 3′-end sequences of wt satRNA progenies were also determined. As was expected, most of the clones (72%) contained the wt sequence, but one clone lacked the last two cytosines and one lacked the last cytosine with an added “t” (Table 1). These data demonstrate that the 3′ end of the wt satRNA is quite stable but that a limited sequence alteration of the 3′ end could also happen.

FIG. 1.

Northern blot analysis of RNA extracted from single plants inoculated with in vitro transcripts of IxΔ3C in the presence of HV. Capped synthetic transcripts corresponding to RNAs 1, 2, and 3 of Fny-CMV were used to inoculate plants with (lanes 1 to 4) or without (lanes 5 to 8) IxΔ3C transcripts. The positions of genomic and subgenomic RNAs of the HV and of satRNA are indicated at the right side. Northern hybridization was performed with ³²P-labeled nick-translated probes of Fny RNA 3 (A) or pIx5 (B).

TABLE 1.

3′-end sequences of progeny satRNA generated in vivo from wt and 3′-end-altered transcripts in the presence of Fny-CMV

Construct	No. of plants containing satRNA/ no. inoculated	3′-end sequencea
pIx5 wt	5/5	[TCCTCCGCAGGACCC]
		TCCTCCGCAGGACCC (5)
		TCCTCCGCAGGACCt (1)
		TCCTCCGCAGGAC (1)
IxΔ3C	13/15	[TCCTCCGCAGGAt]
		TCCTCCGCAGGACCC (11)
		TCCTCCGCAGGACCt (2)
		TCCTCCGCAGGAC (2)
		*TCCTCCGCAGGCCttt (1)
		*TCCTCCGCAGCCttt (1)
		*TCCTCCGCACCttt (1)
		TCCTCCGCAGGACCCggag (1)
Ixm4	10/10	*[TCCTCCGCAGGCCttt(A)16]
		TCCTCCGCAGGACCC (3)
		TCCTCCGCAGGAC (1)
		-39 (1)
Ixm5	9/10	*[TCCTCCGCAGCCttt(A)22]
		TCCTCCGCAGGACCC (3)
		-27 (1)
Ixm6	7/10	*[TCCTCCGCACCttt(A)19]
		TCCTCCGCAGGACCC (4)
		TCCTCCGCAGGAC (1)
		TCCTCCGCAGGCCtt (1)
		TCCTCCGCAGGACC (1)
IxΔ7	3/10	[TCCTCCGCc]
		TCCTCCGCAGGACCC (3)
		TCCTCCGCAGGACC (3)
		TCCTCCGCAt (1)
		TCCTCC (1)
		TCCTC (2)
		-27 (1)
		-40 (1)
		-52 (2)
IxΔ12	0/15	[TCC]

^aPlants were inoculated with transcripts of Ix-satRNA (either wt or containing the alterations shown). Transcript sequences used for inoculation are shown in brackets. Lowercase letters are non-satRNA-derived sequence. All progeny clones terminated in poly(A), which was used in the cloning procedure. Numbers in parentheses reflect the number of cDNA clones with the sequence indicated. Asterisks indicate those 3′-end sequences which were inserted as substitutes into the wt pIx5 clone, resulting in mutants Ixm4, Ixm5, and Ixm6. 

To determine the extent of 3′-end alteration which is reparable, three cDNA clones of the IxΔ3C progeny RNA carrying 2, 3, and 4-nt deletions upstream of the last -CC (indicated by asterisks in Table 1) were tested. The 3′ ends of these mutants were inserted as substitutes into the wt pIx5 clone, generating the mutants named Ixm4, Ixm5, and Ixm6, respectively. These plasmids were linearized downstream of the poly(A)_16-22 tail with BamHI, which resulted in the addition of 8 plasmid-derived nt to the added poly(A) tail. The in vitro transcripts of these constructs were again infectious when coinoculated with Fny-CMV, and wt-size satRNA accumulated in most of the inoculated plants (Table 1). The nucleotide sequence of the progeny RNAs indicated that 62% (10 of 16) of the satRNA molecules had regained the wt 3′ sequence of Ix-satRNA. Four clones had imperfectly restored 3′ ends, with only one or two cytosines of the -CCC, or one upstream deletion. Two clones had large 3′-end deletions of 27 and 39 nt (Table 1).

Two additional deletion mutants, IxΔ7 (-TCCTCCGCc) and IxΔ12 (-TCC), with 7 and 12 bases deleted, respectively, were prepared to test longer deletions at the 3′ terminus (Table 1). RNA was transcribed in vitro from both clones and used to inoculate plants in the presence of the Fny-CMV RNAs. No satRNA was detected in 15 plants inoculated with IxΔ12 transcript either by Northern blotting or by reverse transcriptase PCR analysis. In contrast, 3 of 10 plants inoculated with IxΔ7 RNA contained high levels of satRNA (Table 1). The sequence analysis of the cDNA clones prepared from these progeny RNAs showed a less efficient repair compared with that for the mutants with shorter deletions: only 3 of 14 (21%) clones contained the wt terminal sequence. Three clones had only two instead of three cytosines at the 3′ end, one had a deletion of 6 nt, and seven clones had longer deletions of up to 52 nt (Table 1). These data clearly demonstrate that there is a repair mechanism which maintains the 3′ end of Ix-satRNA and that the extension of this repair is limited to deletions between 7 and 12 nt from the 3′ end.

The effect of HV on the 3′-end repair.

It is well known from previous studies that TAV can be an HV for CMV-satRNA, although no satRNA has ever been found associated with TAV in natural field conditions (reference 28 and references therein). To test if the restoration of the 3′ end of CMV-satRNA also occurred when TAV was the HV, transcripts from pIx5 and pIxΔ3C were each coinoculated with 1-TAV or Fny-CMV RNAs. wt Ix-satRNA transcripts were efficiently replicated and accumulated in the presence of either of the two HVs (Fig. 2B, lanes 1 to 5, and Tables 1 and 2), as was expected from previous reports (28). No satRNA was detectable in IxΔ3C–1-TAV-infected plants (Fig. 2B, lanes 6 to 10, and Table 2), in contrast with the high amounts of satRNA that accumulated in the plants inoculated with IxΔ3C-Fny RNAs (Fig. 1B, lanes 1 to 4, and Table 2). These results show that 1-TAV is not able to support the 3′-end repair of the mutant CMV satRNA, in spite of its ability to support its replication efficiently. The ability to support the 3′-end repair of IxΔ3C by a CMV strain belonging to subgroup II of CMV (31) and by viruses that are pseudorecombinants of TAV and CMV was also tested. Plants were coinoculated with IxΔ3C RNA with the different viruses and pseudorecombinants listed in Table 2. Two weeks after the inoculation, total RNA was extracted and tested by Northern blot analysis. satRNA accumulation was detected only when IxΔ3C was coinoculated with those helpers (Fny-CMV; F₁F₂T₃; Kin-CMV; K₁K₂T₃) that had RNAs 1 and 2 derived from CMV, regardless of their belonging to subgroup I (Fny-CMV) or to subgroup II (Kin-CMV) (Table 2). No satRNA accumulation was detected when IxΔ3C was coinoculated with HVs that had RNAs 1 and 2 from TAV (V₁V₂K₃ or 1-TAV) (Table 2). These results show that the repair of the 3′ end of IxΔ3C depends on the nature of RNAs 1 and 2 of the HV: it occurs when these RNAs are from CMV, and it does not occur when they are from TAV. The origin of RNA 3 has no effect on 3′-end repair. Thus, the restoration of 3′-truncated satRNA is controlled by RNA 1 and/or RNA 2 of the HV, which encodes a subunit of the viral RNA replicase (31).

FIG. 2.

Repair and accumulation of wt (Ix wt) and mutant (IxΔ3C) satRNA in the presence of different HVs. (A and B) Northern blot analysis of RNA extracted from plants inoculated with the in vitro transcripts of Ix5 (wt) (lanes 1 to 5) and IxΔ3C (lanes 6 to 10) in the presence of 1-TAV viral RNAs as helpers. (C) Northern blot analysis of RNA extracted from plants inoculated with IxΔ3C and V₁V₂K₃ (lanes 1 to 5) or F₁F₂T₃ (lanes 6 to 10). The hybridization was performed with ³²P-labeled nick-translated probe of 1-TAV RNA 3 (A) or pIx5 (B and C). RNAs 3B and 5 are a novel class of subgenomic RNAs derived from TAV RNA 3 (38).

TABLE 2.

Accumulation of satRNA in plants inoculated with IxΔ3C or wt Ix-satRNA in the presence of different HVs

Helpera	satRNA	No. of plants containing satRNA/no. inoculated
Kin-CMV	wt	4/5
Kin-CMV	IxΔ3C	5/5
K1K2T3	IxΔ3C	3/5
V1V2K3	IxΔ3C	0/5
F1F2T3	IxΔ3C	5/5
1-TAV	IxΔ3C	0/15
1-TAV	wt	14/15
Fny-CMV	IxΔ3C	13/15

^aK₁K₂T₃, RNAs 1 and 2 of Kin-CMV and RNA 3 of 1-TAV; V₁V₂K₃, RNAs 1 and 2 from V-TAV and RNA 3 from Kin-CMV; F₁F₂T₃, RNAs 1 and 2 of Fny-CMV and RNA 3 of 1-TAV. 

DISCUSSION

We present here evidence that mutants of CMV satRNA with deletions at the 3′ end are able to replicate in plants when coinoculated with the HV CMV and that restoration of the 3′ end does occur. satRNA accumulation and 3′-end repair were observed when 3 to 7 nt were deleted from the 3′ end, but not when the deletion involved 12 nt. Sequence analysis of the progeny satRNA showed that a high percentage of molecules had a wt 3′ end. The efficiency of 3′-end repair, however, depended on the length of the deletion: when the -CCC sequence at the very end of satRNA or 2, 3, and 4 nt just upstream of the last two cytosines (Ixm4, -m5, and -m6) were deleted, satRNA accumulation was detected in 87% of the inoculated plants, and a wt satRNA sequence was found in about 60% of the progeny molecules. When the last 7 nt were deleted, satRNA accumulation was found in only 30% of the inoculated plants and the restoration of a wt 3′ end was found in only 21% of the progeny molecules. In addition, the frequency of clones having longer deletions of 27 to 52 nt was also higher. The origin of these progeny molecules is not clear: they may be remnants of ill-repaired or unrepaired satRNAs which could be replicated at only a low rate, but in any case they should derive from replicating molecules since they were found in the upper noninoculated leaves, and most probably they are not viable, although this point was not tested. The progeny of wt satRNA transcripts also showed some heterogeneity, which indicates that some degradation of satRNA or aborted transcription could also occur during its replication. 3′-end heterogeneity of wt satellite RNA of CymRSV was also reported elsewhere (15). Therefore, a mechanism which repairs the altered 3′ ends of these molecules is clearly advantageous for them.

The sequence analysis of the repaired satRNAs shows that, even in mutants that were efficiently repaired, a high percentage (49%) of the progeny molecules were only partially repaired or ill repaired. These molecules showed a variety of satellite-derived 3′-end sequences, as well as the presence of nucleotides not found in the wt satRNA. The observed alteration of sequence restoration of the wt 3′ ends may indicate a nontemplate repair mechanism. In further support of this hypothesis, we must point out that there is no sequence similarity between the 3′ ends of CMV RNAs and those of satRNAs (Table 3). In addition, CMV genomic RNAs have tRNA-like structures which can be aminoacylated (tyrosylated) (22), while no tRNA-like structure is found at the 3′ end of different CMV satRNA variants, including Ix-satRNA (2, 18, 34). Thus, it is very unlikely that the 3′ end of CMV genomic RNA might be used as a template for the 3′-end repair of CMV satRNA.

TABLE 3.

3′-end sequences of Ix-satRNA and its HVs

RNA	Sequence
Ix-satRNA	TCCTCCGCAGGACCC
Fny RNAs 1, 2, and 3	TCTAAAAGGAGACCA
TAV RNAs 1, 2, and 3	CCCCTAGGGGTCCCA

Our results also show that a second HV of CMV satRNA, TAV, is not able to support the 3′-end repair and accumulation of CMV satRNA mutants with a deleted 3′ end. Pseudorecombinant viruses can be obtained by the exchange of RNAs 1 and 2, or RNA 3, between CMV and TAV (31). With three such pseudorecombinant viruses, it was found that the repair of the 3′ end of CMV satRNA was associated with CMV RNAs 1 and 2, which encode proteins (1a and 2a) which are known to be subunits of CMV RNA-dependent RNA polymerase (RdRp). These results strongly suggest that CMV RdRp is responsible for the repair of the CMV satRNA 3′ end. If so, the replicase function of the RdRp must be separated from the 3′-end repair function, because 1-TAV is an efficient helper for the replication and accumulation of wt Ix-satRNA but not for the repair of the mutant IxΔ3C. Although the RdRps of TAV and CMV must be similar enough to support the replication of heterologous RNA 3 in CMV-TAV pseudorecombinants, there is also evidence of functional differences between them: the TAV RNA 3′-end structure, although tRNA-like, differs from that of CMV RNA and cannot be tyrosylated (21), and recombination leading to the restoration of a homologous 3′ end has been reported elsewhere to favor RNA 3 accumulation in TAV-CMV pseudorecombinants (17). An alternative hypothesis would be that the inoculum of the 3′-end deletion mutants would contain sequence variants due to nucleotide addition by T7 RNA polymerase during transcription (26). Suboptimal variants could be amplified for the more efficient HV, CMV, but not by TAV, to permit error-prone repair. However, this hypothesis is not supported by the fact that pseudorecombinant viruses that have RNAs 1 and 2 from CMV are able to multiply the 3′-end mutants, while they are not more efficient HVs than TAV (Fig. 2). In spite of the fact that TAV can act as an HV for CMV satRNA, no satRNA has ever been found associated with TAV under natural field conditions (31, 35), which is an unexplained phenomenon. We could speculate that the inability of TAV RdRp to efficiently maintain a 3′ end functional for satRNA replication might be a reason that no satRNAs are found in TAV field isolates.

A novel 3′-end repair mechanism was recently reported for the satRNAs of TCV (10, 11), involving the production of 4- to 8-nt oligoribonucleotides by the viral RdRp with the helper viral genome as a template (29). The restoration of 3′-truncated Ix-satRNA does not seem to occur by a similar mechanism; the HV RdRp also seems to be involved but, as discussed above, the HV RNAs cannot be templates for satRNA 3′-end repair. In fact, we do not know of any possible template for CMV satRNA repair, but we cannot exclude the possibility that an unknown cellular RNA could play this template role analogously to telomerases, which add repeated blocks of sequences to the ends of cellular chromosomes synthesized on the telomerase RNA moiety as a template (5, 19). Alternatively, the CMV RdRp itself would be responsible for the repair of the satRNA 3′ terminus by the addition of nontemplated nucleotides. Our data support this mechanism, as it would generate a random distribution of 3′-end sequence variants on which selection would subsequently operate according to replication efficiency, to yield the sequence distribution shown in Table 1, where the wt sequence (i.e., the most fit) is prevalent. In earlier reports, it was shown that double-stranded forms of the satRNAs of CMV and of peanut stunt cucumovirus contained an unpaired guanosine (12, 13). In fact, the ability to add terminal untemplated nucleotides has been described for different RdRps such as that of TCV (40), of Qβ bacteriophage (3, 4), and of uninfected tomato leaves (37).

The repair of the 3′ end of CMV satRNA shows several similarities with 3′-end repair of genomic RNA and satRNA of CymRSV (14, 15) and of TNV (41). There are no known templates with the sequences required for restoration in CymRSV, TNV, or CMV satRNA. All these RNAs have as a common feature a -CCC 3′ end. The 3′-terminal sequences of these RNAs are characterized by the presence of short repeating units (Table 4) which may be considered an indication of analogy with eukaryotic chromosomal telomeres (5).

TABLE 4.

Viral RNAs containing reparable 3′ terminus and CCC end

Viral RNA (reference)	3′-end sequencea	Repetitive sequence
CMV Ix-satRNA (1)	TCCGTGAATGTCTATCATTCCTCCGCAGGACCC	TCC
CymRSV (19a)	CGGACAACCGGAACATTGCAGCAATGCAGCCC	GCA
CymRSV satRNA (35a)	TTCAAACAATCTAATTGTTGAAAACAACAACCC	CAA
TNV-DH	CCTCTTTATTTACCTAGGATTTCCTAGGAACCC	TTT/CCT
TCV satRNA (38a)	GATAGCCTCCCTCCTCGGACGGGGGGCCTGCCC	CCT

^aRepeated sequences are underlined. 

Our results and those reported elsewhere (14, 15, 41) suggest that in vivo 3′-end repair is a common phenomenon among plant RNA viruses with a -CCC 3′ terminus.