Mechanism of stimulation of plus-strand synthesis by an RNA replication enhancer in a tombusvirus

Abstract

Replication of RNA viruses is regulated by cis-acting RNA elements, including promoters, replication silencers, and replication enhancers (REN). To dissect the function of an REN element involved in plus-strand RNA synthesis, we developed an in vitro trans-replication assay for tombusviruses, which are small plus-strand RNA viruses. In this assay, two RNA strands were tethered together via short complementary regions with the REN present in the nontemplate RNA, whereas the promoter was located in the template RNA. We found that the template activity of the tombusvirus replicase preparation was stimulated in trans by the REN, suggesting that the REN is a functional enhancer when located in the vicinity of the promoter. In addition, this study revealed that the REN has dual function during RNA synthesis. (i) It binds to the viral replicase. (ii) It interacts with the core plus-strand initiation promoter via a long-distance RNA-RNA interaction, which leads to stimulation of initiation of plus-strand RNA synthesis by the replicase in vitro. We also observed that this RNA-RNA interaction increased the in vivo accumulation and competitiveness of defective interfering RNA, a model template. We propose that REN is important for asymmetrical viral RNA replication that leads to more abundant plus-strand RNA progeny than the minus-strand intermediate, a hallmark of replication of plus-strand RNA viruses.

FIG. 1.

(A) A schematic presentation of a typical tombusvirus genome (TBSV; the minus strand is shown in a 3′-to-5′ orientation) and a prototypical DI RNA (DI-72). The four noncontiguous regions that are present in the DI-72 RNA are indicated with gray boxes. RIII(−) bears the replication enhancer function, whereas the site of plus-strand initiation is indicated by a solid arrow. The minimal plus-strand initiation promoter (termed cPR, indicated by a triangle), which is located at the 3′ end of RI(−), contains the base sequence (boxed). The predicted long-distance RNA-RNA base pairing (indicated by a two-headed arrow) takes place between the bridge and base sequences, which is predicted to disrupt the hairpin formed in cPR (shown at the bottom). The stem-loop structures SL1-III(−) and SL2-III(−) in RIII(−) are also shown schematically. (B) Time course of RNA synthesis by CNV replicase in vitro, based on denaturing gel analysis. The RNA templates were used in equal molar amounts in the in vitro assays. The ³²P-labeled replicase products were normalized, based on the number of templated urilydates. The templates used contain the cPR promoter and the similar-sized RIII(−) REN, an artificial GC sequence, and RIII(+), respectively. The length of the incubation period during the in vitro assay is shown below the gel. Graphic representation of the quantified and normalized replicase products obtained from the experiments shown in panel B is also shown.

FIG.2.

Stimulation of RNA synthesis by REN in trans. (A, top) Schematic representation of the in vitro trans-replication system. Each of the shown templates consists of two RNA strands. The template strand contains the cPR promoter, which is fused to RII(+) of DI-72 (Fig. 1A). The nontemplate strand carrying RII(−) can base pair with the template RNA, as indicated by the label “clamp” (shown schematically by a ladder). Further details can be found in Fig. 1A. Note that the template strand is the same in these experiments and that each nontemplate strand can form identical clamp structures with the template strand. Only the boxed sequences are different among the RNA constructs. The bridge sequence, which can base pair with cPR, is boxed. A representative denaturing gel of radiolabeled replicase products synthesized by in vitro transcription with CNV replicase is shown at the bottom. The gel-isolated, annealed RNAs were used in equal molar amounts. The template-sized tombusvirus replicase products are marked by an arrow. (B) The effect of REN location on trans-replication activity. Note that the presence of REN sequences (lanes 3 and 4) reduced template activity by ∼50%, likely due to competition between the distant cPR and REN for replicase binding (Fig. 3A). See further details in panel A.

FIG. 3.

In vitro binding of CNV replicase proteins to REN. (A) An RIII(−)-containing RNA is an efficient competitor for the CNV replicase. The template used was MDV(−)/cPR (20), which contains, in addition to the 3′-terminal cPR sequence (Fig. 1A), a 221-nt heterologous sequence derived from the minus-stranded satellite RNA of Qβ bacteriophage. The competitor RNAs were R3(+)/cPR and R3(−)/cPR (Fig. 1B). Lane −, sample lacking a competitor RNA in the replicase reaction. Filled triangles show the increasing amounts of competitor RNAs (from 0.1 μg to 1.0 μg). For graphic representations of data, the template activity of MDV(−)/cPR without competitor is given as 100%. The 50% inhibitory concentration for each competitor is shown by a dotted line. (B) Purified recombinant p92 binds to REN in a gel mobility shift assay. Each probe (as shown at the top) was labeled with [³²P]UTP and was used in four lanes. The left lane represents the free probe (no p92 was added), while the other three lanes have a combination of the free probe and the CNV p92 expressed and purified from E. coli. The nonspecific RNA competitor was used in increasing amounts (500, 1,000, and 2,000-fold excess over the probe). The migration of the free probe is indicated. Protein-RNA complexes are indicated by a bracket on the left. The RNA probes are boxed in Fig. 2A. (C) Efficient binding of recombinant p33 to REN in vitro. The same amounts of labeled probes (as shown on the top) were used in the presence of increasing amounts of p33 (note that p33C, a functional truncated form of the p33 protein, was used). In this native polyacrylamide gel, some of the RNA probes ran as multiple bands, likely due to the formation of alternative secondary structures. The relative amounts of the shifted probes were quantified and shown as percentages of the probe without p33 added (labeled as M). The sequence of R2(−)5′ derived from RII(−) of DI-72 (Fig. 1A) is shown on the right. Note that the presence of more than one shifted bands is likely due to the binding of different number of p33 molecules to the same RNA molecule.

FIG. 4.

Effect of mutations within the RIII bridge sequence on DI RNA accumulation in N. benthamiana. (A and B) A schematic representation of the predicted base pairing between the base and bridge sequences in wt (Fig. 1A) and mutated DI-72 RNAs. Deleted nucleotides are shown by Δ, while the mutated nucleotides are in boldface type. See further details in Fig. 1A. Northern blot analysis shows DI-72 accumulation over time in N. benthamiana protoplasts. The membrane was probed to detect either positive strands (top) or negative strands (middle). The ethidium bromide-stained gel is shown as a loading control at the bottom. The levels of plus-stranded RNA accumulation obtained with the mutated DI RNAs are compared to wt DI-72 RNA after 48 h incubation (100%) (A) or to Base-m1 (B), based on three independent experiments. Note that Base-m1 accumulates only to 10% of wt DI-72 RNA (not shown), suggesting that these mutations affected the function of either cPR or that of the 5′ end of the plus-strand RNA. (C) Accumulation of wt and mutated DI-72 RNA in yeast. Mutations within the bridge sequence are shown with small letters. The level of DI RNA accumulation was estimated based on Northern blots. (D) In vivo competition studies between the wt DI-72 RNA and the bridge deletion mutant (A). RT-PCR analysis of total RNA extracts obtained from N. benthamiana plants or from protoplasts (lanes 9 to 10) inoculated with an equal mixture of WT DI-72 and Δbridge RNAs (see panel A), along with the transcripts of CNV helper by using high-resolution polyacrylamide gels. The RT-PCR was performed either from the inoculation mixture (mixed transcripts) (lanes 1, 8), from inoculated leaf 7 days post inoculation (dpi) (lanes 2 and 5), from noninoculated leaves 7 and 10 dpi (lanes 2, 3 and 6, 7), or from protoplasts (first and second inoculations) (lanes 9 and 10).

FIG. 5.

Presence of conserved base and bridge sequences in tombusviruses. The complementary nucleotides are shown for the minus strands in a 3′-to-5′ orientation. TBSV, tomato bushy stunt virus; CNV, cucumber necrosis virus; AMCV, artichoke mottled crinkle virus; CymRSV, cymbidium ringspot virus; CIRV, carnation Italian ringspot virus. Note that CNV has an additional ACCUCU sequence at the 5′ end of RII(−).

FIG. 6.

A model explaining the stimulation of RNA synthesis from the cPR plus-strand initiation promoter by the RIII(−) REN. The long-range RNA-RNA interaction between the bridge and the base sequences, possibly enhanced by the folding of the full-length RNA, is proposed to alter the structure of cPR. In addition, the formation of base-bridge interaction might also participate in bringing the REN and cPR into proximity. The tombusvirus replicase, containing p33, p92, and host factors, is proposed to bind to one of the two stem-loops, which then could result in positioning the active site of the replicase over the initiation site. This is predicted to lead to efficient initiation of plus-strand synthesis from cPR.