We have confirmed the effectiveness of expressed hRzs as suppressive agents of DENV in transduced mosquito cells. These ribozymes have the ability to cleave their target RNA at an NUH triplet site, and may recycle themselves provided there are short sequence homologies between the ribozyme arms and the corresponding target sequence. This could provide an advantage over standard antisense RNAs that act stoichiometrically and do not destroy the function of the targeted RNAs .
In this study, the ribozymes were expressed under the control of the Ae. aegypti tRNAval promoter. Human tRNAval has been used successfully to enable suppression of target genes by driving the expression of hRzs [29, 30]. The tRNAval utilizes RNA pol III which is involved in the transcription of short RNAs  and yields transcription levels 2–3 orders of magnitude greater than that of pol II systems . Linking hRzs to a tRNAval also enhances their activity by increasing resistance to RNases, since intracellular stability is one of the most important determinants of ribozyme efficacy , and facilitates the cytoplasmic transport of the attached ribozyme  where it can be co-localized with replicating DENV RNA.
Since intramolecular base pairing within the complex structure of the long DENV genomic RNA could preclude the association of ribozymes with their target sequences, we designed our hRz to include a 3' terminal 60 adenylic acids. This poly(A) sequence acts to recruit the unwinding activity of endogenous RNA helicase . Hybrid ribozymes with this poly(A) tail are able to cleave target RNAs possessing complex secondary structure that hRz without accessory helicases cannot act upon [26, 35]. In addition, the effect of helicase-attached hRzs is significantly greater than those of the conventional parental ribozymes .
Retroviral vectors offer a number of advantages over other gene transfer methods for the transduction of somatic cells, and have been widely used as gene delivery systems. They have the ability to stably integrate the genes they vector into host cell chromosomes with high transduction efficiencies [36–38]. Pseudotyped retroviruses displaying the vesicular stomatitis virus G glycoprotein (VSV-G) are able to transduce any dividing cells without the requirement for cell receptors . This pantropic retroviral system has been successfully employed for transduction of mosquito cells . In our experiments, we used self-inactivating retrovirus vectors having a defective 3' U3 that is duplicated as part of the 5' LTR during reverse transcription, allowing the hRz transgene to be expressed by the internal tRNAval promoter and decreasing the possibility of promoter interference [36, 41]. Transduced hRz cells were selected for hygromycin resistance for longer than 2 months. Though it is likely that less than 100% of these resistant cells were transformed, the majority of them appear to bear functional hRz vectors within the genome as indicated by the RT-PCR results of expressed hRz RNA. In addition, because these cells were transduced using lentivitus vectors, there is no real possibility that expression of the hRz came from the non-integrated vector.
Transduction of C6/36 cells by means of pantropic retrovirus vectors resulted in stable genomic integration and expression of the hRzs, rendering them persistently resistant to DENV. Since pools of transduced cells were used for the DENV challenge assays, the observed inhibition effects are not likely due to clonal variation or differing copy number of the constructs within the transduced cell population. However, this approach does allow the possibility that individual cells in the pools may be more susceptible to DENV and thus contribute to the majority of virus detected. In addition, the lower antiviral activity seen for some hRzs might also reflect a distribution effect leading to a wide range of hRz expression levels in the transduced cell cultures [17, 42].
We designed 14 hRzs that target different regions along the DENV-2 genome. Quantitative real-time PCR and immunofluorescence assays demonstrate that C6/36 cells expressing certain hRzs were able to suppress DENV-2 NGC viral replication by at least 25%, with hRz-C6/36 cell lines # 2, 5, 7 and 11 having a more pronounced effect, of nearly 2 logs (100 fold) reduction in viral titer compared to the untransduced C6/36 cells. These IFA results correspond well with data obtained from the Northern hybridization analyses for those hRzs examined.
We were unable to detect the cleaved RNA products by Northern analyses due to two factors. First, there is likely rapid degradation of the DENV RNA within the cells after they are cleaved. Second, the hRz expressing cells are "armed" with many hRz molecules, and upon infection by the virus there is little chance for initiation of a productive infection during which DENV target RNA might build up to levels that cleavage products might become apparent. We interpret these results to demonstrate that suppression of virus, when it occurs, is so effective that little new viral RNA target is actually produced, and few cleavage products would be expected to be evident.
With the exception of hRz # 11, the three most active hRzs target the GUC triplet site which is cleaved most efficiently compared to other triplets . GUC is frequently chosen as the target in many hRz studies because of its wide occurrence in natural hRz motifs . The No-hRz transduced cells, which lack the hRz insert, exhibited no significant suppression of DENV replication, verifying that decreased levels of viral titers requires the presence of the hRz  and is not simply due to the presence of the transducing retrovirus or hygromycin selection.
Ng et al. reported 12% reduction of the krr1 gene expression in Giardia canis in cells without a hRz motif and attribute the observation to an antisense effect of the hRz forming a complex with the mRNA, inhibiting its translation and promoting its degradation . While we cannot rule out an antisense effect for our hRzs, the in vitro analyses of ribozyme cleavage activity demonstrated that they were catalytically active. Hence, the cleavage activity of ribozymes within cellular environments would be expected to occur. The reduction in functional full-length DENV RNA was consistent with the cleavage effects of the ribozymes since the effectiveness of the hRzs we examined correlates very well with the realative effectiveness of the core cleavage site in that three of the four most effective hRzs target a GUC triplet, supporting the catalytic nature of the hRzs effectiveness and arguing against a simple antisense effect.
For the most part, the variations in relative effectiveness that we observed among the 14 hRz we analyzed can be explained by less efficient cleavage activity resulting in incomplete inhibition of DENV replication. However, there are several alternative possibilities unrelated to hRz catalysis that could influence their effectiveness. These include (i) the presence of mismatched bases in the hRz arm, especially with shorter effectors , due to viral escape mutants, (ii) integration of the hRz construct into an inactive gene expression region of the host cell genome, and (iii) obscuring of the target triplet site by RNA-binding proteins [28, 47]. Since our experiments were done over a short time frame and with multiple replicates we can rule out the selection of escape mutations as the cause of weak suppression. Our RT-PCR analyses of the expressed hRz RNA in transformed cells did not show an appreciable difference in levels of expression, allowing us to rule out position effects on expression levels as a factor. The only possibility we cannot rule out in our experiments is the interference by RNA binding proteins at the target site.
hRz # 6 exhibited a somewhat different result from all other hRzs when comparing the relative levels of DENV-specific RNA between cellular and extracellular RNAs by qRT-PCR. The results for hRz # 6 suggest cellular DENV-specific RNA levels were not reduced to comparable levels as extracellular DENV-specific RNA levels. This implies that the reductions in virus titer observed with hRz # 6 are not related to reductions in viral RNA synthesis in infected cells, but may result from affects on packaging of the viral genomes and/or assembly and release of virions. Because hRz # 6 was not among the most effective hRzs tested, we did not pursue examination of this phenomenon.
While we were able to target several sequences that are common to all DENV serotype 2 strains, we were unable to identify suitable cleavage sites that were conserved among all DENV serotypes. Several highly conserved sequences are present among all DENV serotypes, but the requirement for an NUH triplet cleavage site makes these conserved sequences useless as targets for the hRz strategy. Other ribozymes having less stringent cleavage site requirements may be more appropriate for these conserved sequences.
Our results are comparable with results obtained from sense/antisense and siRNA strategies for suppression of DENV. dsSIN may be used in transient expression systems to validate these strategies in either cell cultures or mosquitoes [6, 7] but retroviral vectors and transduced mosquito cells are more appropriate for examining the effectiveness of transgene suppression strategies in somatic cells because they allow a closer approximation of the actual endpoint of these strategies, transgenic mosquito tissues. Our results suggest the best application of this effector molecule is to constitutively express the hRz in target cells since DENV undergoes rapid RNA synthesis at 3–6 hours post infection and release of infectious virions after 12 hours. Transgenic cells pre-primed with expressed hRzs are able to act instantly in inhibiting viral replication upon entry.
Neither our hRz approach nor the siRNA strategy is wholey effective against viral escape mutants. However, one potential obstacle in using RNAi is the evidence for viral RNAi suppressors which have been reported in plants, poliovirus, and HIV-1 variants [48–50]. In the case of DENV, the presence of viral RNAi suppressors may contribute to persistent infections in which DENV is able to replicate in cells without causing cytopathic effect. hRzs are not subject to suppression by viral gene products, which could be an advantage when applied as an anti-DENV effector gene.
A number of RNA viruses exhibit significant genetic variation due to error-prone replication machinery. The simultaneous use of multiple ribozymes attacking two or more sites on the same target RNAs may be advantageous since it helps minimize loss of ribozyme activity due to base-pairing mismatches from mutation of one or the other targeted sites. However, simultaneous expression of two different ribozymes within the same transduced cell pool did not enhance the reduction of DENV titers (data not shown). Our results indicate the same levels of inhibition as obtained with the single hRz approach. This lack of synergy when two independent hRzs area employed might be caused by structural changes within the target RNA following the cleavage by one hRz that alter the cleavage domain for the second hRz, interfering with its ability to anneal with the substrate. Similar results have been reported by Xu et al., 1999, in which the expression of human insulin-like growth factor II is not completely inhibited by double hRzs but yields the same efficacy as single hRzs. As an alternative approach, we are currently exploring combining two hRzs as a so-called Maxizyme [25, 29, 30, 51, 52] that targets two regions with low tolerance for sequence variation, and may ultimately provide even better suppression of DENV replication in transduced mosquito cells and tissues.
Our results demonstrate that the hRz approach has significant potential as a strategy for suppressing DENV in transgenic mosquitoes, either alone or in combination with alternative genetic suppression strategies. While complete suppression of virus infectivity in infected mosquito tissues may not be possible, significant reduction of virus titers within mosquito tissues could result in a decreased efficiency of transmission for the virus, thereby reducing the prevalence of the disease among associated human populations.