Fitness of a Turnip Crinkle Virus Satellite RNA Correlates with a Sequence-Nonspecific Hairpin and Flanking Sequences That Enhance Replication and Repress the Accumulation of Virions

Abstract

satC, a satellite RNA associated with Turnip crinkle virus (TCV), enhances the ability of the virus to colonize plants by interfering with stable virion accumulation (F. Zhang and A. E. Simon, unpublished data). Previous results suggested that the motif1-hairpin (M1H), a replication enhancer on minus strands, forms a plus-strand hairpin flanked by CA-rich sequence that may be involved in enhancing systemic infection (G. Zhang and A. E. Simon, J. Mol. Biol. 326:35-48, 2003). In this study, sequence and structural requirements of the M1H were further assayed by replacing the 28-base M1H with 10 random bases and then subjecting the pool of satellite RNA to functional selection in plants. Unlike previous results with 28-base replacement sequences (G. Zhang and A. E. Simon, J. Mol. Biol. 326:35-48, 2003), only a few of the 10-base SELEX (systematic evolution of ligands by exponential enrichment) assay winners contained short motifs in their minus-sense orientation that were similar to TCV replication elements. However, all second- and third-round winning replacement sequences folded into hairpins flanked by CA-rich sequence predicted to be more stable on plus strands than minus strands. Plus strands of several of the most fit satellite RNAs contained insertions of CA-rich sequence at the base of their hairpins whose presence correlated with enhanced replication and reduced detection of virions. Deletion of the M1H resulted in no detectable virions despite very low satellite accumulation. These results support the hypothesis that a sequence-nonspecific plus-strand hairpin brings together flanking CA-rich sequences in the M1H region that confers fitness to satC by reducing the accumulation of stable virions.

Most plant viruses possess plus-sense single-stranded RNA genomes. Replication of these genomes is mediated by a replicase complex that in general comprises a virus-encoded RNA-dependent RNA polymerase (RdRp), other virally encoded proteins, and possibly host factors (17). The replicase must recognize its cognate RNA through direct or indirect interaction with specific cis-acting elements in the RNA template (4, 9), which for Brome mosaic virus are functional at various locations relative to the minus-strand initiation site (36). Replicases can accurately identify the correct start site for transcription of the genomic RNA, even when the site is artificially placed distal to the 3′ end of the template (24, 40, 43). In addition, many viral RdRps synthesize 3′-coterminal subgenomic RNAs by using internal promoters located on minus-strand replication intermediates (18, 25, 48) that do not necessarily contain sequence or structural similarity to promoters that mediate full-length complementary-strand synthesis. Some plant viruses also provide the replication machinery for associated satellite RNAs (satRNAs), most of which share little sequence similarity with the viral genomic RNA (38). Thus, a single viral replicase must recognize a variety of seemingly different promoter elements.

In addition to core promoter elements, many RNA virus genomes contain additional cis-acting elements that enhance but are not required for basal level transcription (29, 37). These RNA replication enhancers, which can be found in both plus and minus strands of viral RNAs, are located at variable positions in relation to the initiation site for RNA transcription and are thought to facilitate the recruitment of RdRp to the template (17). Replication enhancers have been characterized for a number of RNA viruses such as bacteriophage Qβ (2), Alfalfa mosaic virus (46), Tomato bushy stunt virus (37), Brome mosaic virus (10, 35, 36), and Sindbis virus (11).

Since most satRNAs associated with plant viruses are fewer than 500 bases in size and do not encode proteins, they are ideal templates for examining the structure and function of cis replication elements. The Turnip crinkle virus (TCV) (family Tombusviridae, genus Carmovirus) isolate TCV-M, a single-stranded RNA virus of 4,054 bases, was originally found to be associated with three related satRNAs, including satD (194 bases) and satC (356 bases) (1, 22, 39). satD appears to have originated from numerous short noncontiguous stretches of TCV genomic RNA sequence (5), while satC is a chimeric molecule containing nearly full-length satD at its 5′ end and two discontinuous segments from TCV genomic RNA at its 3′ end (Fig. 1A) (39). satC is also an unusually virulent satRNA, strongly intensifying the symptoms of TCV on hosts that display discernible symptoms when infected with TCV alone (20).

FIG. 1.

Elements required for satC replication and/or fitness. (A) cis-acting elements involved in satC replication. satC is a chimeric molecule composed of satD sequence at the 5′ end (gray box) and two regions of TCV at the 3′ end (open box is derived from the CP ORF and hatched box is from the 3′ untranslated region). Plus-strand elements are shown above satC, and minus-strand elements (shown in their minus-sense, 3′-to-5′ orientation) are shown below the satRNA. Pr, the core promoter on plus strands required for synthesis of minus strands. Minus-strand elements CCS, 3′ PE, and 5′ PE are described in the text. The minus-strand M1H is a replication enhancer that spans all three satC segments and contains motifs found in TCV promoter-like elements. ∗, the sequence is also found in the 5′ PE; ∗∗, the sequence is identical to the TCV 3′ CCS. (B) Structure of the wt satC plus-strand M1H and surrounding sequence (left); structure of the M1H in satC_B, with the single site alteration required for the construction of satC with randomized sequence shown boxed and the residues that were replaced with 10 random bases underlined (middle); and plus-strand structure formed by the M1H replacement sequence (underlined) in clone UC, generated during the previous SELEX analysis (55) (right).

Most satRNAs are considered to be parasitic agents, selfishly coopting the replication machinery of their helper virus without conferring any obvious advantage to the virus. satC, by interfering with the replication of TCV in protoplasts (16) and accumulation in plants (20), was thought to fit that general description. However, the ability of satC to enhance the symptoms of TCV only in plants that display visible symptoms when infected with TCV alone suggested that the satRNA enhances the ability of the virus to colonize the plant and interact with cellular factors (20, 54). Thus, the fitness of satC is likely a function of attributes that allow the satRNA to both replicate efficiently and interact with the helper virus in a way that enhances systemic infection. Zhang and Simon's recent finding of a surprising correlation between enhanced virulence of TCV and reduced levels of virions (54) suggests that satC fitness may reflect in part the ability of the satRNA to interfere with virion formation.

Analysis of satC replication in plant protoplasts and transcription of complementary strands in vitro using extracts from infected plants that contain partially purified RdRp led to the identification of a number of cis-acting elements required for efficient accumulation of the satRNA. These include a 3′-terminal hairpin on plus strands that is the core promoter for the synthesis of complementary minus strands (Fig. 1A) (6, 41, 42) and three linear elements, termed the 3′ carmovirus consensus sequence (CCS), 3′ proximal element (PE), and 5′ PE (Fig. 1A) (12-14), on minus strands. The CCS (C_2-3A/U_3-7) is found at the 3′ ends of all carmovirus genomic, subgenomic, and satRNA minus strands identified to date and is required for replication in vivo (12). The 3′ PE, which is required for transcription of minus strands in vitro in the absence of the 5′ PE, also contains a CCS (12). The 5′ PE is required for transcription of minus strands in vitro in the absence of the 3′ PE (13). Both the 3′ PE and 5′ PE can function as independent promoters of cRNA synthesis in vitro (14).

A 28-base internal hairpin (M1H) is also required for normal levels of satC accumulation in plants and protoplasts and thus functions as a replication enhancer (Fig. 1A) (26, 27, 29). In its minus-sense orientation, the M1H contains three short motifs identical to sequence in either the TCV 3′ CCS or the satC 5′ PE (55). The minus-sense M1H enhances transcription from the 3′ PE promoter by nearly 10-fold in vitro (29) and is also a hot spot for recombination between satC and satD in vivo (8, 28). Recently reported results obtained by using an in vivo functional SELEX (systematic evolution of ligands by exponential enrichment) assay, where 28 random bases replaced 28 bases of the M1H, resulted in the identification of winners that replicated to higher levels in protoplasts compared with satC containing nonselected 28-base sequences (55). As with the wild-type (wt) M1H, most winning M1H replacement sequences contained one to three short motifs in their minus-sense orientation that were identical, or nearly identical, to sequence from TCV or satC replication elements (55).

With one exception, the replacement sequences enhanced replication of satC in protoplasts to levels that correlated with fitness of the satRNAs to accumulate in plants. This exception, clone UC (Fig. 1B), contained only a 7-base replacement sequence, indicating that a deletion of 21 bases had occurred. Although competition assays determined that UC was the second-most-fit winner, it did not replicate more effectively in protoplasts than did random 28-base replacement sequences (55). The 7-base UC replacement sequence, together with downstream sequence, folded into a hairpin that was 16% more stable on plus strands than minus strands and was, like the wt M1H, flanked by CA-rich sequences (Fig. 1B). Surprisingly, the replacement sequences of the other second-round winners (and most of the first-round winners) also folded into plus-strand hairpins flanked by CA-rich sequences that were predicted to be more stable than minus-sense hairpins, suggesting that a sequence-nonspecific plus-strand hairpin might contribute to satRNA fitness. UC was substantially better at reducing TCV virion accumulation compared with other M1H replacement sequence winners that replicated to higher levels, suggesting that the ability to reduce virion levels contributed to the fitness of the satRNA. In addition, it seemed possible that the role of the hairpin was to bring the flanking CA-rich sequences into proximity, contributing to virion reduction.

For the present report, we tested this hypothesis and found that 10-base M1H replacement sequences also folded into hairpins predicted to be more stable on plus strands than minus strands. Several of the most fit satRNAs contained inserts of adenylates and cytidylates at the base of the hairpins whose presence correlated with enhanced replication and reduction of virion levels. These results confirm that a sequence-nonspecific plus-strand hairpin flanked by CA-rich sequence in the M1H region confers fitness to satC by reducing virion assembly.

MATERIALS AND METHODS

In vivo functional SELEX using 10-base M1H replacement sequences.

In vivo genetic selection was carried out as previously described (6, 12, 13, 55). To produce full-length satC RNAs containing 10 random bases replacing 28 bases of the M1H, two fragments were generated by separate PCRs using pT7C+ (40), a full-length cDNA clone of wt satC. The 5′ fragment was produced using primers T7C5′ (5′-GTAATACGACTCACTATAGGGATAACTAAGGG-3′), which contains a T7 polymerase promoter at the 5′ end, and SEL5′ (5′-GACTGGATCCTTTTGAGTGGGAAACAG-3′). The 3′ fragment containing 10 random bases was generated using primers SEL10BAS (5′-AAAAGGATCCNNNNNNNNNNACCAAAAACGGCGGCAGCAC-3′) and oligo7 (5′-GGGCAGGCCCCCCGTCCGA-3′). In these two PCRs, a new BamHI site was generated in satC cDNA by changing a U to C at position 176 for convenient ligation of the purified PCR fragments. satC with this new BamHI site is referred to as satC_B. Both 5′ and 3′ PCR fragments were treated with BamHI, purified through 1.2% agarose gels, and then ligated to produce full-length cDNA. satC transcripts were directly synthesized from the ligated product by using T7 RNA polymerase. Five micrograms of satC transcripts and 2 μg of TCV transcripts were coinoculated onto each of 29 turnip seedlings. Total RNA was extracted from uninoculated leaves of each plant at 21 days postinoculation (dpi), amplified by reverse transcription-PCR using primers T7C5′ and oligo7, and then cloned into the SmaI site of pUC19 and sequenced. For the second-round SELEX, equal amounts of total RNA extracted from all first-round plants were pooled and approximately 5 μg of the pooled RNA was used to inoculate each of six new turnip seedlings. Total RNA was extracted from uninoculated leaves at 21 dpi, and satC was cloned and sequenced as described above. The third-round SELEX was accomplished by combining in vitro transcripts of all recovered second-round clones and then inoculating 5 μg of the pooled transcripts along with 2 μg of TCV transcripts onto six new turnip seedlings. satC accumulating in uninoculated leaves was cloned and sequenced as described above.

Secondary-structure analysis of SELEX winners.

The 10-base M1H replacement sequences (along with 11 upstream and 8 downstream bases) of all second-round SELEX winners were analyzed for secondary structures by using mFOLD (23). Each sequence was also subjected to three randomizations by using the Shuffle program from Arizona Research Labs, and the stability of the folded sequences was determined and the values were averaged.

Fitness comparison of the third-round SELEX winners in plants.

To compare the fitness of the third-round SELEX winners for accumulation in plants, equal amounts of transcripts (0.2 μg/plant) were combined and used to inoculate a single leaf of three turnip seedlings along with TCV genomic RNA transcripts (2 μg/plant). Total RNA was extracted at 21 dpi, and satC RNAs were cloned and sequenced as described above.

Construction of satC variants.

Constructs ΔA₅C, 2xA₅C, and R1/A₅C were generated by ligation of two PCR fragments. The 5′ fragment was obtained by PCR using pT7C+ and primers T7C5′ and SEL5′, while the 3′ fragments were amplified by using pT7C+ and primers oligo7 and either 5′-GACTGGATCCTTTTACGGGAACCAAAAACGGCGGCAGCAC-3′, 5′-GACTGGATCCTTTTACGGGAACCAAAAACAAAAACAAAAACGGCGGCAGCAC-3′, or 5′-GACTGGATCCATCCGGACCAACCAAAAACAAAAACGGCGGCAGCAC-3′. For MOV/A₅C, the 5′ fragment was obtained by PCR using pT7C+ and primers T7C5′ and 5′-GACTGGATCCGTTTTTTTTTGAGTGGGAAACAGCC-3′ and the 3′ fragment was identical to that for construction of ΔA₅C. The 5′ and 3′ PCR fragments were treated with BamHI, gel purified, ligated together, and inserted into the SmaI site of pUC19. All clones were confirmed by sequencing.

Protoplast preparation, inoculation, and RNA gel blots.

Transcripts for protoplast inoculations were synthesized in vitro by using T7 polymerase. Protoplasts (5 × 10⁶) prepared from callus cultures of Arabidopsis thaliana ecotype Col-0 were inoculated with 20 μg of TCV genomic RNA transcripts and 2 μg of satC transcripts as described previously (15). Equal amounts of total RNA extracted from protoplasts at 36 to 40 h postinoculation (hpi) were subjected to RNA gel blot analysis (48). An oligonucleotide complementary to both positions 3950 to 3970 of TCV genomic RNA and 250 to 269 of satC was labeled with [γ-³²P]ATP by using T4 polynucleotide kinase and was used as a probe for simultaneous detection of TCV genomic RNA and satC.

Protein extraction and analysis.

Total protein was extracted from protoplasts at 40 hpi with TCV alone or TCV and various satC transcripts as described previously (49). Proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) on 12% polyacrylamide gels and then stained by using Coomassie brilliant blue.

Virion isolation and Western blots.

Virion isolation and Western blots were performed as previously described (15, 33). Virions isolated from protoplasts (2.5 × 10⁶) at 40 hpi with TCV alone or TCV and various satC transcripts were analyzed by Western blotting using the Western Lighting chemiluminescence reagent kit (Perkin Elmer Life Sciences) and anti-TCV coat protein (CP) polyclonal antibody.

RESULTS

In vivo functional SELEX of satC containing 10 random bases replacing the M1H.

Previous results suggested that the M1H participates in two aspects of satC fitness: by serving as a sequence-specific enhancer of satC replication in its minus-sense orientation and possibly by interfering with stable assembly of TCV virions mediated by the hairpin structure in its plus-strand orientation (55). To test this hypothesis, in vivo functional SELEX was conducted by replacing 28 bases of the M1H with 10 random bases by using the PCR strategy outlined in Materials and Methods. Function-based in vivo SELEX identifies sequences that impart fitness to the molecule regardless of the mechanism of action. This differs from the more traditional in vitro SELEX, which selects for molecules with a specific function, such as binding to a particular polypeptide or ability to catalyze a specific reaction (50). The replacement of 28 bases with 10 bases was chosen for two reasons. (i) A shorter sequence should lead to the recovery of satC that is more similar to UC with its 7-base replacement sequence and thus might contain hairpins that could be compared with UC. (ii) The promoter-like motifs found in the minus-sense 28-base M1H replacement sequences were generally between 6 and 10 bases, suggesting that similar motifs that enhanced replication might be found when using a 10-base replacement sequence.

satC transcripts containing 10 random bases replacing the M1H were coinoculated onto 29 turnip seedlings along with transcripts of TCV genomic RNA. At 21 dpi, total RNA was isolated from uninoculated leaves. Examination of RNA following gel electrophoresis and ethidium bromide staining revealed that all plants contained detectable satRNA (data not shown) (wt satC normally accumulates to levels similar to that of 5S rRNA and thus is readily detectable by this method). Sixty-four satC species were cloned from 19 of the 29 inoculated plants. Sequencing of the clones revealed 43 different first-round winning sequences (data not shown).

These and other first-round winners were subjected to further competition by combining equal portions of total RNA extracted from all 29 infected plants and then inoculating six new seedlings. At 21 dpi, total RNA was extracted from uninoculated leaves and satC molecules were cloned. Sequencing of 8 to 11 clones from each plant revealed 25 different sequences (the second-round winners [Table 1 ]). The sequences are presented in their minus-sense 3′-to-5′ orientations since promoter-like motifs are present in the minus-strand M1H sequence (Fig. 1A) and were also found in the minus-sense 28-base replacement sequences from the previous study (55). Six of the second-round winning sequences were also found in the first round (II-6, II-10, II-11, II-15, II-17, and II-20). Sequences ΙΙ-1a and ΙΙ-1b differed by only 1 base and sequences ΙΙ-3a and ΙΙ-3b differed by only 2 bases; thus, these were possibly derived from the same original transcript. Since the ΙΙ-1a and ΙΙ-1b and the ΙΙ-3a and ΙΙ-3b replacement sequences folded into hairpins with different stabilities (see below), they are listed separately in Table 1. One of the second-round winners, II-16, was unusual in having an 18-base replacement sequence instead of the original 10 random bases, indicating that additional modification of replacement sequences is occurring in planta. The replacement sequence in II-16 (ACUAGCCCGUUAGCCCGU) contains a 10-base sequence (underlined) followed by a repeat of the terminal 8 bases (double underline). In addition, one second-round winner (II-21) contained only an 8-base replacement sequence.

TABLE 1.

Second-round SELEX winners

Name	M1H replacement sequencea	No. recovered from six plantsd	ΔG (kcal/mol)		Randomized ΔGe (kcal/mol)
			Plus strand	Minus strand	Plus strand	Minus strand
II-1a	CUUCCCCAAG	5	−9.2	−0.7	−5.0	−2.7
II-1b	CUCCCCCAAG	1	−6.7	−0.7	−5.2	−5.0
II-2	UACUCAAGGC	8	−2.1	−1.0	−2.1	−2.3
II-3a	UUAAACCAGG	1	−5.7	−3.9	−0.5	−1.6
II-3b	UUAUACUAGG	1	−6.3	−6.4	−0.7	−1.6
II-4	CAAAUCCAGG	4	−5.9	−2.5	−2.8	−3.7
II-5	ACUCUUUCCU	5	−7.5	−5.2	−2.5	−2.5
II-6	AUAAGUCCCU	2	−6.6	−3.8	−0.7	−2.0
II-7	GGUACCCCUA	2	−11.0	−9.5	−2.2	−2.3
II-8	AAUGCUAGAU	1	−3.5	−2.7	−0.2	−0.6
II-9	AUAACCCCGG	3	−9.2	−4.2	−5.5	−4.4
II-10	CGUUCCGCCG	1	−3.7	−2.4	−3.1	−2.5
II-11	AUAGUUGUCC	1	−6.1	−8.1	−0.5	−1.5
II-12	CGAAGCCCGC	1	−4.3	−4.4	−3.4	−4.1
II-13	UUUGCCUGGC	1	−5.0	−5.8	−1.8	−3.8
II-14	CAUUAGCCCA	2	−8.1	−4.9	−1.9	−2.3
II-15	CUUACCAGGU	2	−5.5	−3.1	−3.8	−4.0
II-16	ACUAGCCCGUUAGCCCGU	1	−7.8	−6.2	−7.6	−6.6
II-17	AUAUAUGUCCb	3	−6.1	−8.1	−1.7	−3.1
II-18	AAAAUGCCCUc	1	−6.2	−4.0	−1.9	−2.5
II-19	UACCGGACCU	1	−6.4	−5.4	−4.4	−4.3
II-20	UAUUCAGACC	3	−6.1	−6.8	−1.6	−3.3
II-21	UUGAGUCC	7	−5.2	−5.5	−1.0	−2.3
II-22	AUAACCUCGG	1	−6.0	−3.7	−2.8	−5.0
II-23	CUAUUUGAGC	1	−1.8	−3.2	−0.5	−0.4
Mean ± SD			−6.1 ± 2.1	−4.5 ± 2.3	−2.5 ± 1.9	−3.0 ± 1.4

^aSequences are shown in their minus-sense, 3′-to-5′ orientation.

^bClones containing this sequence also had an insertion of A₁₀C (Fig. 3C).

^cClones containing this sequence also had an insertion of A₅C (Fig. 3C).

^dTotal number recovered = 59.

^eMinus-strand sequences were shuffled and then folded three times, with the averages presented as described in Materials and Methods. The complements of the shuffled minus-strand sequences were also folded to give the averaged plus-strand randomized values.

The third round of competition was conducted by combining equal portions of full-length transcripts synthesized for each of the 25 second-round winners and then inoculating the mixture with TCV genomic RNA transcripts onto six seedlings. Of the 46 clones recovered from five of these plants at 21 dpi, 29 were clone II-18 and 10 were either II-3a (seven clones) or II-3b (three clones). Two additional clones, II-2 and II-17, were isolated from only one plant each. All of these clones were designated as third-round winners. To determine whether the number of clones recovered in the third round reflected the fitness of the individual clones to accumulate in plants, equal amounts of transcripts of pairs of third-round winners were combined and inoculated with TCV genomic RNA onto three turnip seedlings. At 21 dpi, total RNA was extracted from an uninoculated leaf and approximately equal numbers of clones were sequenced from each of the three plants. In direct competition between II-18 and II-3a, 70% of the clones recovered (21 of 30) were found to be II-18. When plants were inoculated with II-18 and II-2, 92% (24 of 26) of the recovered clones were found to be II-18. The second most prevalent winner, II-3a, out competed II-2 (22 of 34 clones), while II-2 and II-17 were of similar fitness when assayed together (19 of 35 clones were II-2). These results indicate that the number of clones accumulating in the third-round SELEX plants reflect the fitness of the clones under the infection conditions used.

Sequence and structural composition of the second-round winning sequences.

Analysis of the minus-strand M1H replacement sequences in the second-round 10-base SELEX winners indicated a disproportionate number of cytidylates (33% of the total bases) with a strong preference for multiple consecutive cytidylate residues (5.2-fold greater than expected for random sequences). Cytidylates also comprised 35% of the total number of residues found for the 28-base replacement sequence winners (55). Three of the 10-base SELEX second-round winners contained motifs found in TCV promoter elements (class I and class II [Fig. 2 ]), and two second-round winners contained sequence similar to a motif found in numerous 28-base SELEX winners (AACCCCU) but not present in any known TCV promoter-like element (class III [Fig. 2]). Seven winners from the first round and six winners from the second round contained a new motif consisting of an AU-rich sequence followed by C_1-4, A_1-2, G_1-2, and C/U (class IV [Fig. 2]). This motif was similar to the 7-base sequence in the 28-base SELEX winner UC (UCAGGAA). Several of the winners also contained base changes or inserted sequence outside of the replacement sequence region. Two of the third-round winners, ΙΙ-17 and ΙΙ-18, had the sequences UUUUUUUUUUG and UUUUUG” (minus-sense orientation), respectively, inserted downstream of their replacement sequences, while second-round winner II-9 had an upstream A-to-U alteration (Fig. 3B and C).

FIG. 2.

Motifs found in 10-base SELEX M1H replacement sequences. All sequences are shown in their minus-sense, 3′-to-5′ orientation. TCV and satC minus-strand promoter elements (5′ PE and 3′ CCS) are shown in Fig. 1A and described in the text. Common motifs (classes I through V) are underlined. Round 1, winners from the first round of the SELEX. An asterisk indicates that the clone was also a third-round winner.

FIG. 3.

Computer-predicted plus-strand structures encompassing the M1H replacement sequences (underlined). (A) Structures formed by the complement of class IV (II-1a, II-1b, and II-4) and class V (II-14) replacement sequences in second-round winners (Fig. 2). ΙΙ-1a and ΙΙ-1b differ by either an adenylate or guanylate in the 4-base loop. (B) Two second-round winners, ΙΙ-16 and ΙΙ-9, have additional alterations (a repeat of 8 bases [asterisks] or an adenylate-to-uridylate transversion [boxed]) that increase the stability of the plus-strand hairpins. ΙΙ-9 clones contained either a uridylate or a guanylate in the 4-base loop, as shown. (C) Structures of the five second-round winners that were recovered in the third round. ΙΙ-3b contains two base differences (U to A and G to A) compared with ΙΙ-3a. For II-18 and II-17, inserted bases outside the replacement sequences that are not found in wt satC are denoted by asterisks. (D) Hairpin H4, located in the 3′ untranslated region of the TCV genome, is in the same position relative to the 3′ end as the M1H of satC. The boxed sequence (and downstream sequence) is found in satC (hatched box region in Fig. 1A) and comprises the 3′ base of the plus-strand M1H. As with the M1H, H4 is also flanked on the 5′ side with CA-rich sequences.

To determine whether the 10-base M1H replacement sequences in the second-round winners formed hairpins in conjunction with nearby sequences, mFold structural predictions (23) were generated for all 25 winning sequences along with 11 upstream and 8 downstream bases. Sixty-eight percent of the winning sequences formed local hairpins that were more stable in their plus-sense orientation than in their minus-sense orientation, with the plus-strand structures being, on average, 36% more stable than the minus-strand structures (Table 1). All hairpins were present in the most stable structures predicted for full-length plus strands of third-round winners, while only two of five third-round winners had hairpins that were present in the most stable full-length satC minus-strand structures (data not shown).

The plus-strand hairpins predicted to form for nine of the second-round winners, including the four third-round winners, are shown in Fig. 3A to C. The plus-strand structures of the related clones II-1a and b were not altered by the base differences. For II-3a and II-3b, the sequence differences did not affect the structure of the plus-sense hairpins (U:G or U:A pairing in the stem) but would affect the structure of the minus-strand hairpins. The additional 8-base repeat in ΙΙ-16, which contained the unusual 18-base replacement sequence, increased the stability of the plus-sense hairpin by 70%. The second site alteration in ΙΙ-9 increased the stability of the predicted plus-sense hairpin, but not that of the minus-strand hairpin, by 64%. Most of the class IV motifs found in first- and second-round winners allowed for the formation of hairpins with AU-rich loops and perfectly base-paired stems (for example, see II-4 in Fig. 3A and II-3a/b in Fig. 3C), with the replacement sequence extending down one side to the base of the stem. Class V motifs formed structures similar to that shown for II-14, where unaltered satC sequence formed the base on both sides of the stem, with the replacement sequence forming the loop and upper portion of the stem. All of these structures were similar to the wt M1H and the UC hairpin (Fig. 1C) in that they consisted of hairpins flanked by CA-rich sequences. The possible importance to satC fitness of maintaining CA-rich sequences at the base of a hairpin was also suggested by two of the five third-round winners (ΙΙ-17 and ΙΙ-18), which contained additional CA-rich sequence in their plus strands that was not found in the parental transcripts (Fig. 3C).

Effect of 10-base winning sequences on satC replication in protoplasts.

As described above, the fitness of satC to accumulate in plants depends on several factors, including the ability of the satRNA to replicate and to positively affect the systemic infection of the helper virus. Replication of satC can be directly assayed by using host Arabidopsis thaliana protoplasts, since alterations to the M1H, including deletion of the hairpin, have no effect on stability of the RNA (55). To determine how well the selected 10-base sequences contribute to replacing the replication enhancer function of the M1H, replication levels in protoplasts were determined for II-18 and II-3a, the two most fit third-round winners and were compared with those of wt satC and satC_B. satC_B, the parental satRNA of the SELEX constructs, contains a single base alteration near the base of the M1H resulting in the incorporation of several bases of flanking sequence into the hairpin according to computer models (Fig. 1B). As shown in Fig. 4A, satC_B replicates to 82% of wt satC levels at 36 hpi. Deletion of the 28-base M1H (ΔM1H) resulted in a 97% decrease in satC levels. satC with a randomly selected 10-base sequence replacing the M1H (Rd10) replicated at 10% of wt satC levels. This increase above levels obtained with ΔM1H may be due to size effects, since previous results have shown that satC and other TCV subviral RNAs can replicate more poorly when reduced in size (7, 21, 53). Third-round winners II-18 and II-3a replicated to levels that were 2.7- and 1.8-fold higher, respectively, than Rd10, with II-18 reaching levels that were 33% more than the less fit II-3a. However, II-18 accumulated 2.5-fold less than did parental satC_B, suggesting that the replacement sequence in II-18 is unable to enhance replication of satC comparable to the natural M1H enhancer. The 18-base size difference between II-18 and satC_B might also contribute to the difference in the levels of the satRNAs. II-18 and II-3a accumulated to levels similar to that of UC, the 7-base replacement sequence winner from the previous M1H SELEX (55). These results suggest that the 10-base M1H replacement sequences weakly enhance the replication of satC in the absence of the natural enhancer. The contribution to replication of the A₅C insert three bases downstream of the replacement sequence in II-18 is discussed below.

FIG. 4.

Accumulation of viral RNAs and virions in protoplasts. Arabidopsis protoplasts were coinoculated with TCV genomic RNA transcripts and transcripts of various satRNAs. (A) Total RNA was extracted at 36 hpi and subjected to RNA gel blot analysis using an oligonucleotide probe complementary to both TCV and satC. Values shown were calculated from at least three independent experiments. (B) Effect of different M1H replacement sequences on virion accumulation in protoplasts at 40 hpi. CP and virions were subjected to electrophoresis on SDS-PAGE gels and detected by Coomassie brilliant blue staining and chemiluminescence following treatment with anti-TCV CP antibody, respectively. satC, wt satC; satC_B, the parental satRNA used to generate satC with randomized sequences; II-18 and II-3a, third-round SELEX winners; Rd10, satC with a randomly chosen 10-base sequence replacing the 28 bases of the M1H; ΔM1H, satC with a deletion of the 28 bases of the M1H

Two most-fit third-round winners interfere with TCV virion accumulation in protoplasts.

Previous results suggested that the 7-base replacement sequence in UC might contribute to the fitness of the satRNA by interfering with TCV virion accumulation. To determine whether the replacement sequences of the third-round winners also repress virion accumulation, protoplasts were inoculated with TCV and various satRNA species and the levels of viral RNA, total CP, and virions were determined at 40 hpi. As shown in Fig. 4B (and data not shown), coinoculation of TCV with satRNAs did not reproducibly affect the total amount of CP. The wt satC and satC_B replicated to higher levels than did II-18 and II-3a, which replicated better than Rd10 and ΔM1H, as previously shown when protoplasts were assayed at 36 hpi (Fig. 4A). The wt satC, satC_B, II-18, and II-3a all reduced virion accumulation by similar amounts, although II-18 and II-3a replicated to levels that were 3- to 4.5-fold less than that of wt satC. II-18 and II-3a were more efficient at reducing virion levels than was Rd10, with a sevenfold greater reduction in virion levels for II-18 compared with Rd10, while satRNA levels varied by less than threefold. Surprisingly, the most efficient satRNA at reducing virion levels was ΔM1H, with no virions detected despite very low satRNA accumulation. This result suggests that a sequence-nonspecific hairpin in this plus-strand location may function to bring into proximity the CA-rich sequences that flank the hairpin, which would also result from a deletion of the hairpin sequence.

Role of CA-rich sequence flanking the M1H in satC replication and virion accumulation.

Previous results indicated that deletion of the CA-rich single-stranded sequences flanking the M1H reduced satC levels in protoplasts (30). Thus, these sequences may serve a dual function by inhibiting virion levels and enhancing satC replication. Since two of the five third-round winners had insertions of additional CA-rich sequence just downstream of the M1H, we tested whether such inserts conferred additional fitness to the satRNA. Mutations were introduced into II-18 to generate a series of derivatives in which the A₅C insert was deleted, duplicated, or moved upstream of the hairpin (Fig. 5A). In addition, the 10-base M1H replacement sequence in II-18 was exchanged with the replacement sequence from a randomly selected first-round winner (R1), which was not isolated in further rounds and thus judged less fit than third-round winning sequences.

FIG. 5.

Effect of the A₅C insert in third-round SELEX winner II-18 on replication and virion accumulation. (A) Predicted plus-strand structures encompassing the M1H replacement sequence and flanking regions. The M1H replacement sequences are underlined, and inserted sequences are marked by asterisks. II-18, most fit third-round winner; ΔA₅C, II-18 with a deletion of the A₅C insertion; 2xA₅C, II-18 with a duplication of the A₅C insert; MOV/A₅C, II-18 with the A₅C insert deleted and moved upstream of the predicted hairpin; R1, randomly selected first-round winner; R1A₅C, R1 with an insert of A₅C. (B) Accumulation of satRNAs in Arabidopsis. Total RNA was extracted at 36 hpi and subjected to RNA gel blot analysis using an oligonucleotide probe complementary to both TCV and satC. Values shown were calculated from at least three independent experiments. (C) Accumulation of viral RNAs, CP, and virions in protoplasts inoculated with TCV transcripts alone or together with transcripts of satRNAs. TCV genomic RNA and satRNA were detected by RNA gel blot analysis using an oligonucleotide probe complementary to both TCV and satC. CP and virions were subjected to electrophoresis on SDS-PAGE gels and detected by Coomassie brilliant blue staining and chemiluminescence following treatment with anti-TCV CP antibody, respectively.

Transcripts of the different II-18 derivatives were coinoculated with TCV genomic RNA into protoplasts, and total RNA was extracted at 36 hpi and analyzed by Northern hybridization (Fig. 5B). Deletion of the A₅C insert from II-18 reduced RNA accumulation of the resultant satRNA (ΔA₅C) by nearly 50%, suggesting a role for the additional sequence in enhancing replication of II-18. Relocating the A₅C insert upstream of the hairpin increased replication by 30% over ΔA₅C, suggesting that the CA-rich sequence on either side of the hairpin can contribute to satRNA levels in protoplasts. In contrast, replication of first-round winner R1 was not enhanced by the addition of A₅C downstream of the hairpin. However, the R1 replacement sequence terminates with ACCA, which may already be contributing towards enhancement of replication not impacted by additional CA-rich sequence. This possibility was supported by the finding that tandem duplication of the A₅C sequence in II-18 (2xA₅C) also did not further enhance replication.

To examine the effect of the A₅C insert on virion accumulation, TCV genomic RNA transcripts along with transcripts of II-18 and its derivatives were inoculated onto protoplasts and total RNA, protein, and virions were extracted at 40 hpi. As shown in Fig. 4B, association of TCV with a satRNA did not affect the total amount of CP accumulating in infected protoplasts (Fig. 5C). Movement of the A₅C insert upstream of the hairpin did not appreciably affect virion levels compared with II-18. However, deletion of the A₅C insert from II-18 increased virion accumulation by nearly 2.5-fold, suggesting that an insert of A₅C upstream or downstream of the hairpin enhanced the satRNA's ability to repress the accumulation of virions. Additional CA-rich sequence (2xA₅C) did not further contribute to virion repression compared with II-18. The first-round winner R1 was less effective at reducing virion accumulation compared with II-18, and addition of A₅C did not enhance the ability of R1 to reduce virion levels. However, R1 was almost twice as efficient at reducing virions compared with ΔA₅C (i.e., the II-18 replacement sequence alone). Since R1 also accumulated in protoplasts to 38% higher levels than ΔA₅C, the enhanced fitness of II-18 is primarily due to the A₅C insert and not its M1H replacement sequence. These results strongly suggest that a sequence-nonspecific plus-strand hairpin flanked by CA-rich sequences can contribute to satRNA fitness by repressing the accumulation of virions.

DISCUSSION

The previously reported correlation between fitness and repression of virions was based on only a single SELEX winner, clone UC with a 7-base M1H replacement sequence. We have now repeated the in vivo functional SELEX by using a 10-base replacement sequence. While several winning replacement sequences in the 28-base SELEX were able to enhance satC replication to near satC_B levels (55), the 10-base replacement sequences in the most fit satRNA of the present study were only weakly effective at enhancing replication. II-18, the most fit 10-base SELEX winner, accumulated to levels that were 2.7-fold greater than a randomly selected sequence (Rd10) but were only 40% of parental satC_B levels (Fig. 4A). In addition, while 48% of first- and second-round winners of the 28-base SELEX winners contained one or more minus-sense motifs found in a variety of replication-like elements, only 3 of 83 first- and second-round winners of the present 10-base SELEX had similar replication-like motifs and none were found in third-round winners. However, all 25 of the 10-base SELEX second-round winners contained replacement sequences that were predicted to fold into similar hairpins flanked by unaltered single-stranded CA-rich sequences. The principal motifs (classes IV and V) found in a number of 10-base SELEX winners differed primarily between whether the replacement sequences extended midway down or to the base of the hairpin (compare II-a/b and II-4 with II-14 [Fig. 3A]). This result suggests that a 10-base region is insufficient space to specify both replication-like motifs and the sequence required to form plus-strand hairpins. Since all of the second-round winning sequences formed similar hairpins, this suggests that the major selective pressure on the 10-base M1H replacement sequences was for plus-strand hairpin formation and not replication enhancement. We cannot rule out, however, that the high percentage of selected sequences with multiple C residues in the minus-sense orientation in both the 28-base and 10-base SELEX may also weakly enhance replication, since recent reports have indicated that CCA or CCCA repeats are capable of serving as independent promoters of transcription in vitro by using RdRp from several viruses including TCV (45, 51, 52).

The replacement sequences of two of the third-round winners, II-17 and II-18, formed hairpins such that a downstream, originally single-stranded ACC formed the 3′ base of the hairpin. These two winners contained unusual downstream inserts of A₁₀C and A₅C, respectively, which extended the single-stranded CA-rich sequences at the 3′ base of the hairpin. For II-18, A₅C inserted either upstream or downstream of the hairpin contributed to both the replication of the satRNA and the ability of the satRNA to reduce virions. While it is currently unknown whether the plus-sense CA-rich sequence or its complementary minus-strand sequence is responsible for enhancement of replication, the strong preference for plus-strand hairpins in the second- and third-round winners suggests that the A₅C insert in its plus-sense orientation enhanced virion reduction and was responsible for the selection of II-18 as the most fit of the 10-base SELEX winners.

ΔM1H replicated very poorly in protoplasts yet was highly effective at reducing virion accumulation. This result suggested that a hairpin in this location primarily functions to bring together the flanking CA-rich sequences, which would also be achieved by deletion of the region between the CA-rich sequences. The mechanism by which the CA-rich sequences repress virion accumulation is not known. Qu and Morris (33) determined that a bulged hairpin located near the 3′ end of the CP open reading frame (ORF) was a specific packaging signal for TCV genomic RNA based on its ability to independently promote the packaging of a heterologous RNA by the TCV CP in protoplasts. Interestingly, this hairpin in the TCV-M isolate used in our study (which differs at a single position from the TCV-B isolate used by Qu and Morris) contains an A-rich (five of nine residues) loop. While the sites within the hairpin that specifically interact with the CP are not known, hairpin loops are frequently targets of RNA-binding proteins (31). It is therefore possible that the multiple consecutive adenylates within the CA-rich satC sequences compete for CP binding, thus disrupting virion assembly.

A second possibility is that virion assembly requires sequences in addition to the hairpin in the CP ORF. The lack of a similar hairpin in packaged subviral RNAs satC, satD, and DI-G, a defective interfering RNA containing no CP ORF sequence, indicates that other packaging signal sequences have yet to be identified. The 3′ untranslated region of TCV genomic RNA contains a plus-strand hairpin (H4 [Fig. 3D]), which is similar to the plus-strand M1H in structure, location relative to the 3′ end of the RNA, and flanking CA-rich sequences. Since the second recombination event producing satC occurred near the 3′ base of H4, satC and TCV share related downstream sequences, including the 3′-side CA-rich sequence at the base of the M1H (boxed sequence in Fig. 3D). However, the majority of TCV H4 and 5′ CA-rich sequence is unrelated to sequence in satC. We are currently testing whether the U-rich loop of H4 and flanking CA-rich sequences, elements also found in DI-G, are involved in virion assembly, which might be inhibited by similar CA-rich sequences in satC.

Recent reports demonstrating the multifunctional nature of the TCV CP have provided a possible explanation for how inhibition of TCV virions contributes to the fitness of satC. Most plant viruses encode repressors of virus-induced gene silencing (VIGS), an antiviral protective system in plants (3, 19, 47) which might serve a similar function in other organisms (32). The TCV CP was recently shown to be a very strong suppressor of VIGS when assayed independently of the virus (34, 44). However, when expressed from the virus genome, the CP is a weak suppressor, possibly due to suppression of an early step in gene silencing, while the CP is expressed mainly later in infection (34). Furthermore, the N terminus of the CP, required for suppressor activity, is also the RNA-binding domain and is unavailable for suppressor function when sequestered within the virus capsid (44). By interfering with virion formation, satC could be enhancing the abundance of free CP, leading to more efficient suppression of VIGS and thus facilitating the systemic infection of TCV (54).