Results and discussion

Flavivirus genomes possess inverted complementary sequences at the ends of the RNA, similar to that observed in the negative strand RNA viruses (bunya-, arena-and orthomyxoviruses) (Kohl et al 2004, Barr & Wertz 2004, Hsu et al 1987, Raju & Kolakofsky 1989, Mir & Panganiban 2005, Ghiringhelli et al 1991). These complementary sequences have been suggested to allow the ends of the genome to associate through base pairing, leading to circular conformations of the RNA (panhandle-like structures). Do flavivirus genomes acquire a circular conformation? Is the long-range RNA—RNA interaction required for dengue virus replication?

To investigate a possible association between the 5' and 3' ends of dengue virus genome, we analyzed the formation of RNA-RNA complexes using electrophoresis mobility shift assays (EMSA) with in vitro transcribed radiolabelled RNA molecules. We found that an RNA molecule corresponding to the first 160 nucleotides of dengue virus RNA (5'UTRC62 RNA) specifically binds to a second RNA molecule carrying the last 106 nucleotides of the viral genome (3 SL Probe, Fig. 2A). This RNA-RNA interaction shows high affinity (Kd about 8 nM) and is absolutely dependent on the presence of Mg2+. In dengue and other mosquito-borne flavivi-ruses it was proposed that the complementary sequences 5 -3 CS are a potential cyclization element. Folding predictions of the sequences representing the RNA-RNA complex formed by the ends of the dengue virus genome show two pairs of complementary regions (Fig. 2B). One of these regions is the 5'-3' CS, the second is located at the 5' end just upstream of the initiator AUG and at the 3' end within the stem of the 3' SL (named UAR, upstream AUG region, Fig. 2B). To investigate the RNA determinants for complex formation, we generated 5 UTRC62 RNA molecules with substitutions generating mismatches in either 5'-3' CS or 5'-3' UAR. We tested the binding ability of these mutated 5'RNAs in EMSA using a 3' SL wild-type probe. Mutations in 5' CS or 5' UAR greatly decreased the binding of the mutated 5 UTRC62 RNAs to the radiolabelled 3 SL, confirming that both complementary sequences were necessary for RNA-RNA complex formation (Fig. 2C).

To investigate whether the RNA-RNA contacts observed between two RNA molecules representing the ends of dengue virus genome also occur in a single RNA molecule as long-range interactions, we analysed the conformation of individual molecules by atomic force microscopy (AFM). The in vitro transcribed full-length dengue virus RNA was deposited on mica, dried and visualized by tapping mode AFM. The single-stranded RNA molecules acquire compact structures, precluding visualization of intramolecular contacts (Fig. 3A). To overcome this problem, we hybridized the viral RNA with an antisense molecule of 3.3 kb (complementary to a region encompassing NS4B-NS5), which yielded an extended double stranded region flanked by the single stranded ends of the viral RNA. These molecules were observed in both circular and linear conformations in the absence of proteins (Fig. 3B and C). Statistical analysis of this and other model RNA molecules carrying specific mutations confirmed that viral RNA circularizes through direct RNA-RNA contacts involving CS and UAR sequences (Alvarez et al 2005b).

Previous reports have suggested that base pairing between 5 and 3 CS of fla-vivirus genomes is necessary for viral replication. Using Kunjin and West Nile virus replicons it has been shown that substitution mutations in either 5 or 3 CS that disrupted base pairing were lethal for RNA replication (Khromykh et al 2001, Lo et al 2003). However, when both CS sequences were mutagenized with respect

RNA-RNA Complex

RNA-RNA Complex

UCAAUAUGCUG CAGCAUAUUGA

AGAGAGCAGAUCUCUG 5'UAR

CAGAGAUCCUGCUGUCU 3'UAR

200 100 20 5

200 100 20 5

200 100 20 5

200 100 20 5

200 100 20 5

200 100 20 5

RNA-RNA Complex

3'SL Probe

RNA-RNA Complex

3'SL Probe

FIG. 2. RNA-RNA complex formation between the end sequences of dengue virus RNA. (A) Mobility shift assays shows RNA-RNA complex formation. Uniformly labelled 3' SL RNA (1 nM, 30 000 cmp), corresponding to the last 106 nucleotides of dengue virus 2, was incubated with increasing concentrations of the 5'UTR-C62 RNA corresponding to the first 160 nucleotides of the viral genome, as indicated at the top of the gel. (B) Schematic representation of dengue virus genome showing the location and nucleotide sequence of 5' CS, 5' UAR, 3' CS, and 3' UAR. Folding prediction of the proposed circular conformation of the RNA is also shown. (C) RNA mobility shift analysis showing the effect of mutations in CS or UAR within the 5' UTR-C62 RNA on binding to the 3' SL RNA. Uniformly labelled 3' SL RNA was incubated with increasing concentrations of wild-type and mutated 5' UTR-C62 RNAs as indicated at the top of the gel.

FIG. 2. RNA-RNA complex formation between the end sequences of dengue virus RNA. (A) Mobility shift assays shows RNA-RNA complex formation. Uniformly labelled 3' SL RNA (1 nM, 30 000 cmp), corresponding to the last 106 nucleotides of dengue virus 2, was incubated with increasing concentrations of the 5'UTR-C62 RNA corresponding to the first 160 nucleotides of the viral genome, as indicated at the top of the gel. (B) Schematic representation of dengue virus genome showing the location and nucleotide sequence of 5' CS, 5' UAR, 3' CS, and 3' UAR. Folding prediction of the proposed circular conformation of the RNA is also shown. (C) RNA mobility shift analysis showing the effect of mutations in CS or UAR within the 5' UTR-C62 RNA on binding to the 3' SL RNA. Uniformly labelled 3' SL RNA was incubated with increasing concentrations of wild-type and mutated 5' UTR-C62 RNAs as indicated at the top of the gel.

FIG. 3. Visualization of genome—length dengue virus RNA of 10.7 kb by tapping mode AFM. (A) Visualization of a representative single stranded dengue virus RNA molecule. (B, C) Images of circular and linear conformations of individual dengue virus RNA molecules, respectively. The 10.7 kb RNA molecule was hybridized with an antisense RNA of 3.3 kb resulting in a linear double stranded region with single stranded regions of 6970 and 451 nucleotides at the 5' and 3' ends, respectively.

FIG. 3. Visualization of genome—length dengue virus RNA of 10.7 kb by tapping mode AFM. (A) Visualization of a representative single stranded dengue virus RNA molecule. (B, C) Images of circular and linear conformations of individual dengue virus RNA molecules, respectively. The 10.7 kb RNA molecule was hybridized with an antisense RNA of 3.3 kb resulting in a linear double stranded region with single stranded regions of 6970 and 451 nucleotides at the 5' and 3' ends, respectively.

to the wild-type sequence such that their capacity to base pair was maintained, RNA replication was restored. We performed similar experiments using recombinant dengue viruses to address the importance of 5'—3' UAR complementarity during viral replication. Our data showed that specific substitutions within 5' or 3' UAR, in the context of the infectious dengue virus 2, yielded no viable viruses. Importantly, mutations at the 5' and 3' UAR that restored complementarity were sufficient to rescue viral replication. The replication of this recombinant virus displayed slow growth and small plaque phenotype when compared with the wildtype virus, suggesting that the nucleotide sequence/structure of 5' and/or 3' UAR is also important for efficient viral replication (Alvarez et al 2005b).

Our results indicate that long range RNA-RNA interactions result in circular conformation of the viral genome. In addition, the functional studies on dengue virus together with previous reports obtained with other flaviviruses strongly suggest that sequence complementarity is required for viral replication. However, many questions remain open: what is the role of the long-range 5'-3' end interactions during dengue virus replication? Is the cyclization of the viral genome necessary for efficient translation, similar to that observed for cellular mRNAs? Is the complex between the 5' and 3' ends of the genome required for NS5 polymerase binding during RNA synthesis? Does the structure of the RNA involving both ends of the genome play a role in coordinating translation and RNA synthesis? Does the long range RNA-RNA interaction constitute a signal for RNA encapsidation?

To dissect the role of the cyclization sequences during the viral processes, we constructed a dengue virus replicon that allows discrimination between translation of the input RNA and RNA synthesis (Alvarez et al 2005a). Similar replicons have been recently developed for West Nile and yellow fever viruses (Lo et al 2003, Jones et al 2005). In the context of dengue virus 2, we replaced the viral structural proteins by the firefly luciferase coding sequence (Luc). The trans membrane domain (TM) corresponding to the C-terminal 24 amino acids of E protein was retained in order to maintain the topology of the viral protein NS1 inside of the ER compartment (Fig. 4A). The Luc was fused in-frame to the first 102 nucleotides of the capsid protein (C), which contain the 5' CS sequence. To ensure proper release of the Luc from the viral polyprotein, we introduced the cis-acting FMDV 2A protease (Fig. 4A).

Dengue virus replicon was efficiently translated and amplified in transfected BHK and mosquito cells. After RNA transfection with lipofectamine, Luc activity increases as a function of time reaching the highest levels between 8 and 10 h, reflecting the translation of the input RNA. After 24 h of transfection, Luc activity increases exponentially as a result of replicon RNA amplification (Fig. 4A). A replicon with a mutation in the GDD active site of the RNA dependent RNA polymerase NS5 (MutNS5) showed the translation peak, but after 24 h the levels of Luc were indistinguishable from the background, showing the lack of RNA amplification. These results indicate that replicon RNA amplification by the viral replicase machinery can be monitored through the expression of Luc as a function of time in transfected cells.

We used the replicon system to generate specific substitutions disrupting 5'-3' CS or 5'-3' UAR complementarity. To this end, we incorporated substitutions within the 3' CS sequence, generating 4 mismatches (the wild-type 3' CS CAG-CAUAUUGA was replaced by UAUCAUUUGGA, Mut.3'CS replicon). In addition, a second replicon was designed carrying point mutations at the 5' CS that restored sequence complementarity with the mutated 3'CS (Rec.5'-3'CS replicon). RNAs corresponding to the wild-type, Mut.NS5, Mut.3'CS, and the double mutant Rec.5'-3'CS were in vitro transcribed and equal amounts of the RNA were trans-

DENGUE RNA A

FMDV2A

FMDV2A

C

Luc | NS1

NS2AB

NS3

NS4AB

NS5

<D 1000000

10000

1000

10000

1000

Translation

RNA Replication

DVRep

FIG. 4. Dengue virus replicon allows discrimination between translation and RNA replication in transfected cells. (A) At the top, schematic representation of dengue virus replicon. Boxes denoting coding sequences of capsid (C), Luciferase (Luc), and non-structural proteins (NS) are shown. Also, the position of FMDV2A protease is indicated by an arrow. At the bottom, replication of dengue virus replicon in BHK cells is shown. Time-course of luciferase activity was detected in cytoplasmic extracts prepared from BHK cells transfected with wildtype dengue virus replicon (WT) or replication-incompetent MutNS5 RNAs. (B) Translation and RNA replication of WT, 3' CS mutant (3'CSMut), double mutant at the 3' and 5' CS (Rec.5'-3'CS), and replication-incompetent MutNS5 were analysed in transfected BHK cells. Normalized Luc levels are shown in logarithmic scale at 10 h after transfection to estimate translation of input RNA and at 3 days after transfection to evaluate RNA replication.

fected into BHK cells (Fig. 4B). Renilla Luc mRNA was cotransfected with all the replicons and used to standardize the transfection efficiency in each time point. After 10 h of transfection no significant differences in Luc activity were observed between cells transfected with the wild-type or mutated replicons, suggesting that translation of the input RNA was not dependent on 5'—3' CS interactions. In contrast, RNA synthesis of the Mut 3'CS replicon was undetectable. Reconstitution of 5'—3' base pairing in the Rec.5'—3'CS replicon also restored RNA synthesis (Fig. 4B). The level of RNA replication detected with this replicon was lower than the one observed with the wild-type RNA, suggesting that the highly conserved WT sequences within CS provide an advantage during RNA amplification.

To study the importance of 5'—3' UAR complementarity during viral translation and RNA synthesis we introduced specific mutations in these regions in the replicon system. UAR sequences are located within very conserved stem loops of the viral 5' and 3' UTRs, therefore, it is difficult to introduce mutations in 5' or 3' UAR sequences that disrupt complementarity without altering secondary structures. We designed point substitutions in both sides of the stem or loop of SLB (5' UAR) and/or in their complementary sequences within the stem of the 3'SL (3' UAR) (Fig. 5A). We generated three groups of mutations (Mut 1, Mut 2 and Mut 3), each group was composed of three different replicons: (a) with mutations in 5' UAR, (b) with mutations in 3' UAR, and (c) with mutations a + b in the same replicon, restoring 5'—3' UAR complementarity with sequences that differed from the wildtype sequences (Fig. 5A). The RNAs corresponding to the wild-type and the nine mutated replicons described above were transfected into BHK cells and Luc activity was measured as a function of time. The translation efficiency was determined measuring Luc activity after 10 h of transfection. The results showed that all mutant replicons were translated with similar efficiencies to that observed with the wildtype replicon (data not shown). To evaluate RNA synthesis, we compared the RNA amplification of each mutated replicon with the wild-type levels and expressed the results as a percentage of the wild-type (Fig. 5B). From this analysis we observed that: (i) single mutations disrupting 5'—3' UAR base pairing decreased or abolished RNA synthesis; (ii) in the three groups of mutants the reconstitution of the 5'—3' complementarity (5'3'Mut 1, 5'3'Mut 2 and 5'3'Mut 3) also increased the levels of RNA synthesis; (iii) not all base pairings within 5'—3' UAR are equally important for RNA synthesis (compare the replication levels of 3'Mut-2 with 3'Mut-3, both with only one mismatch); and (iv) in some cases reconstitution of 5'—3' complementarity is not sufficient to rescue RNA synthesis to wild-type levels, suggesting that specific nucleotides within the 5' or 3' UAR sequences can be critical. In this regard, it is important to mention that the sequence of 5' and 3' UAR are absolutely conserved in all dengue virus serotypes, suggesting that even though co-evolution of the two complementary sequences could occur there must be a growth advantage to preserve the wild-type nucleotide sequences.

5' and 3'UAR Mutations

aA"

FIG. 5. Mutations within 5' or 3' UAR alter replicon RNA amplification (A) Schematic representation of dengue virus genome showing the predicted secondary structure of the RNA elements containing 5' and 3' UAR (shown in boxes). Nucleotide sequences of wild-type and the substitutions introduced at the 5' UAR (5'Mutl, 5'Mut2, and 5'Mut3); and at the 3' UAR (3'Mutl, 3'Mut2, and 3'Mut3) are shown. (B) RNA replication of mutant replicons disrupting and restoring 5'—3' UAR complementarity in transfected BHK cells. Normalized luciferase levels determined 3 days after transfection are shown as a percentage of the wild-type replicon for the three groups of mutants within UAR sequences (Mutl, Mut2 and Mut3). Underneath of each plot, base pairing between sequences corresponding to 5'—3'UAR for the mutants are compared with the wild-type sequences.

FIG. 5. Mutations within 5' or 3' UAR alter replicon RNA amplification (A) Schematic representation of dengue virus genome showing the predicted secondary structure of the RNA elements containing 5' and 3' UAR (shown in boxes). Nucleotide sequences of wild-type and the substitutions introduced at the 5' UAR (5'Mutl, 5'Mut2, and 5'Mut3); and at the 3' UAR (3'Mutl, 3'Mut2, and 3'Mut3) are shown. (B) RNA replication of mutant replicons disrupting and restoring 5'—3' UAR complementarity in transfected BHK cells. Normalized luciferase levels determined 3 days after transfection are shown as a percentage of the wild-type replicon for the three groups of mutants within UAR sequences (Mutl, Mut2 and Mut3). Underneath of each plot, base pairing between sequences corresponding to 5'—3'UAR for the mutants are compared with the wild-type sequences.

Taken together, the results discussed here indicate that the complementary sequences 5'—3' CS and 5'—3' UAR are not necessary for efficient translation of the input RNA but they are essential during RNA synthesis. To better define the role of sequence complementarity during the process of RNA synthesis, we are currently investigating NS5 polymerase association to circular and linear conformations of the RNA and the impact of RNA conformation on enzyme activity. Disruption of 5'—3' interactions did not alter significantly translation initiation of the input RNA, however, we cannot rule out a possible regulatory role of RNA-RNA interactions on translation after several rounds of protein synthesis had taken place in the infected cell. It is possible that dynamic RNA-RNA and RNA-protein interactions involving cellular and viral factors, induce conformational changes of the viral RNA rendering the viral genome more competent for translation, RNA synthesis, or encapsidation at different stages of viral infection. Understanding the role of secondary and tertiary structures of the viral RNA during the viral life cycle will help to clarify molecular details of dengue virus replication. At present neither specific antiviral therapy nor licensed vaccine exists to control dengue virus infections. Therefore, it is of crucial interest to investigate the biology of this virus at the molecular level as an essential step on designing novel antiviral strategies.

Was this article helpful?

0 0
Swine Influenza

Swine Influenza

SWINE INFLUENZA frightening you? CONCERNED about the health implications? Coughs and Sneezes Spread Diseases! Stop The Swine Flu from Spreading. Follow the advice to keep your family and friends safe from this virus and not become another victim. These simple cost free guidelines will help you to protect yourself from the swine flu.

Get My Free Ebook


Post a comment