This study examines the effectiveness of a constitutively expressed group I intron that targets and trans-splices conserved sequences and induces apoptotic cell death upon infection as a means of suppressing DENV virus infection of transformed mosquito cells. Group I trans-splicing introns have established potential for targeting RNA virus genomes in infected cells
[18, 32, 33]. Previously we determined an optimal group I intron target sequence following an alignment of 98 instances of DENV that identified one highly conserved region positioned within the capsid coding sequence at nucleotides C131 to G151
[38, 62]. These nucleotides are a part of the 5’-3’ CS domain of the DENV genome
 that is essential for DENV replication
. We designed an anti-DENV group I trans-splicing intron, with firefly luciferase serving as the 3’ exon, and demonstrated its ability to effectively cleave at nucleotide position U143
. We also demonstrated its ability to effectively trans-splice an infecting DENV-2 NGC when constitutively expressed as RNA in transformed C6/36 cells
In this report we demonstrate the feasibility of using the U143 targeting group I intron, αDENV-U143, to catalyze trans-splicing of the 5’ CS region of DENV genomes to a 3’ ΔN Bax exon to induce apoptotic death of cells upon infection. A UAA stop codon inserted in the trans-splicing domain of the intron prevents premature expression of the ΔN Bax 3’ exon that would induce cell death in uninfected cells. Upon infection, αDENV-U143-ΔN Bax targeting and cleavage of DENV genomes at uracil 143 forms a chimeric mRNA that consists of the 5’ cap, 5′ UTR, 143 nucleotides of the DENV capsid (DCA) coding sequence, and the 3’ ΔN Bax exon. This chimeric RNA is capable of expressing a DCA-ΔN Bax fusion protein that induces apoptotic cell death precluding productive virus infection.
The strategy of targeting conserved sequences in the CS region of the genome cannot be considered immune to the evolution of escape mutations, but the extreme conservation of this region among all DENV, and even among Flaviviruses, suggests a markedly decreased potential for these mutations to develop. One obvious drawback to using these catalytic RNA molecules as simple genome degrading agents is that if the rate of virus replication exceeds the rate of group I intron catalytic suppression the evolution of escape mutants may be enhanced. An added level of insurance against the development of escape mutants should be achieved through the induction of cellular apoptotic pathways in response to DENV infection. Coupling the splicing activity of the group I intron to a death-upon-infection strategy insures that DENV replication rates do not exceed group I intron expression and catalytic rates, and should decrease the probability of generating escape mutants.
The use of a group I intron to induce cellular death upon infection has potential advantages over an RNAi suppression strategy since the length of conserved sequence necessary for group I intron targeting can be discontinuous as well as smaller than that required for RNAi-mediated responses. While successful RNAi responses in mosquitoes have been developed to directly target individual dengue serotype genomes
[15, 16, 63–66], the RNAi approach may have difficulties targeting all serotypes simultaneously, and there is the possibility that escape mutants may amplify without restriction in response to the RNAi suppression.
The targeting and cleavage capability of our intron constructs was demonstrated with transient transfection assays in C6/36 mosquito cells challenged with infectious DENV. Sequencing analysis confirmed that the correct splice product was obtained, indicating proper targeting and site-specific cleavage of the DENV genome by the transiently expressed αDENV-U143 introns. Addition of the IRES/mCherry reporter configuration immediately downstream of the 3’ ΔN Bax exon did not appear to alter the trans-splicing capabilities of the intron, or affect the ability of the DCA-ΔN Bax resulting from the splice product to initiate apoptosis in DENV infected cells.
Our 5’-RLM-RACE results demonstrate that there is a 56 nt 5’ extension of RNA sequence in our anti-DENV group I intron resulting from expression by the A5c promoter that does not prohibit targeting and trans-splicing of DENV genomes (Additional file
1: Figure S1). However, we cannot rule out the possibility that an enhancement in anti-DENV group I intron activity could be achieved if the 56nt 5’ extension could be eliminated. This does not appear to be possible at this time since all RNA pol II promoters add a 5’ extension of considerable length due to the placement of their respective TSS, and elimination of this sequence typically results in greatly diminished RNA pol II promoter activity
Expression and pro-apoptotic function of ΔN Bax is not inhibited by the 19 amino acids of the Dengue CA protein fused to its N-terminus. Expression and activity of ΔN Bax or DCA-ΔN Bax expressed in cells is not significantly different, and trans-splicing of the DENV RNA genome by αDENV-U143-ΔN Bax leads to the activation of cellular apoptosis as indicated by annexin V-FITC (Figure
8A), caspase 3 assays (Figure
8B), and DNA ladder analysis (Figure
9 and Additional file
3: Figure S3).
In regards to the difference between Annexin V and Caspase-3 assays with reference to the C-11 cell clone, several labs have successfully demonstrated that the early stages of apoptosis can be reversed
[68–72]. The ability of U143-ΔN Bax mRNA to translocate out of the nucleus may be hampered, leading to a decrease in the number of DCA-ΔN Bax mRNA fusion molecules that are produced following DENV-U143-ΔN Bax targeting of DENV RNA in the cytoplasm of C-11 clonal cells. The diminished protein expression of DCA-ΔN Bax perturbs the progression of C-11 from the early stages of apoptosis (positive Annexin V staining) to the latter stages of apoptosis (negative Caspase 3 activity and DNA fragmentation). None of these assays indicate apoptotic cell death when the trans-splicing negative αDENV-ΔU143 or mCherry are expressed or when FL is used as the 3’ exon, confirming our results are a consequence of the presence of a trans-spliced RNA encoding the DCA-ΔN Bax. Other researchers have analyzed N-terminal epitope-tagged variants of tBax with little alteration in activity
[40, 73, 74]. This is likely due to the fact that the C-terminal residues of Bax possess the pore forming function of this pro-apoptotic protein.
TCID50-IFA results demonstrate suppression of infectious virus production from our transformed and hygromycin selected cell lines upon challenge with each of the four serotypes (Figure
6). While we observe as much as a 5 log decrease in viral titer with each of the four serotypes targeted, the effector gene is even more potent than these uncloned, hygromycin-selected transformed cells demonstrate because these cultures necessarily produce hygromycin-resistant, non-transformed susceptible cells. Support for this reasoning is provided by the observation that removal of hygromycin selection results in a rapid recovery of virus susceptibility for our transformed cultures.In contrast, a greater antiviral effect was observed with transformed clonal cell populations in which every cell is confirmed to express the αDENV-U143-ΔN Bax intron by detection of the DCV-IRES expression of the mCherry marker from the same transcript (Figure
7). The enhanced DENV suppression observed for αDENV-U143-ΔN Bax vs. αDENV-U143-FL clones reflects a lack of dependency upon complete cleavage of all DENV genomes within infected cells expressing αDENV-U143-ΔN Bax due to the potency of the proapoptotic DCA-ΔN Bax product generated. Our results validate the utility of this single antiviral effector gene as a means for producing transgenic mosquitoes that will be refractory for DENV transmission.
Recently, a DENV-5 serotype has been identified in non-human primates from Malaysia that is characterized by a different antibody profile than the four known DENV serotypes
. This discovery will significantly impact vaccine development efforts, and may further enhance the attractiveness of anti-DENV transgenic mosquito strategies that can affect all serotypes. Although no sequence data is available for DENV-5 at the time of this submission, there is a high likelihood that the 5’-3’ CS domain will be conserved in this isolate as well, making it susceptible to our anti-DENV group I intron, U143.
We now have an anti-DENV group I intron that allows us to target at least four, and likely all five, DENV serotypes simultaneously. Targeting all serotypes with a single catalytic ribozyme or siRNAs eliminates the necessity to construct and test separate catalytic RNAs or siRNA molecules. However, unlike siRNA molecules that have been designed to target conserved regions of DENV in mammalian cells, the induction of cellular apoptosis by our αDENV- U143-ΔN Bax construct following DENV trans-splicing eliminates escape mutants that may evolve in the infected cell and prevents virus replication from overriding the catalytic activity of the anti-DENV group I intron.
These results foreshadow the potential efficacy of our U143-ΔN Bax constructs against DENV infection of transgenic mosquitoes expressing these antiviral effectors. Based on natural infection rates of midgut cells and the regenerative capabilities of midgut epithelia we do not expect that the loss of cells upon ingestion of a blood meal will have a significant impact on the survivability of the transgenic mosquitoes. This is a potential advantage in the dissemination of this transgene within the native population. Finally, our demonstration that the appended 56 nucleotide 5’ extension resulting from transcription of the U143 intron does not inhibit targeting or trans-splicing of the DENV RNA genome suggests to us that an antiviral group I intron construct capable of targeting multiple viruses simultaneously should be possible. For us the most likely virus candidates for such a dual targeting construct would be DENV and chikungunya viruses since these co-endemic pathogens have been shown to simultaneously infect humans and vector mosquitoes