Novel method for selection of tRNA-driven ribozymes with enhanced stability in mammalian cells

Intracellular stability is a critical determinant of the activity of a ribozyme in vivo. In previous studies, we succeeded in constructing an effective system for the expression of ribozymes using the promoter of a human gene for tRNA(Val). The resultant tRNA(Val)-driven ribozymes (tRNA-ribozymes) had a half-life of approximately 100 minutes. In the present study, we established a novel system for the selection of tRNA-ribozymes that were more stable than a previously generated optimally designed tRNA-ribozyme, and we confirmed that the newly selected tRNA-ribozymes worked well. Selective pressure was applied by treating cells that expressed tRNA-ribozymes with actinomycin D, and the system yielded tRNA-ribozymes with enhanced stability. The sequences isolated after selection exhibited some similarities. Furthermore, some selected tRNA-ribozymes had almost the same activity as or higher activity than that of the optimally designed tRNA-ribozyme despite the fact that the selective pressure was not aimed at enhancing the cleavage activity. Our approach might be very useful for selection not only of ribozymes with enhanced stability but also of other functional nucleic acids in vivo.     

Construction of a tRNA-embedded-ribozyme trimming plasmid.

We have combined Cotten and Birnstiel's tRNA-embedded ribozymes and our 5'- and 3'-trimming system. Although the activity of the tRNA-embedded ribozyme was ca. 30% lower than those of naked ribozymes, since the stability of the former in bovine serum was higher than those of naked ribozymes, the tRNA-embedded ribozymes appear useful especially when the 5'- and 3'-trimming units are concatenated in tandem.     

Design, characterization and testing of tRNA3Lys-based hammerhead ribozymes

A hammerhead ribozyme targeted against the HIV-1 env coding region was expressed as part of the anticodon loop of human tRNA3Lys without sacrificing tRNA stability or ribozyme catalytic activity. These tRNA-ribozymes were isolated from a library which was designed to contain linkers (sequences connecting the ribozyme to the anticodon loop) of random sequence and variable length. The ribozyme target site was provided in cis during selection and in trans during subsequent characterization. tRNA-ribozymes that possessed ideal combinations of linkers were expected to recognize the cis target site more freely and undergo cleavage. The cleaved molecules were isolated, cloned and characterized. Active tRNA-ribozymes were identified and the structural features conducive to cleavage were defined. The selected tRNA-ribozymes were stable, possessed cleavage rates lower or similar to the linear hammerhead ribozyme, and could be transcribed by an extract containing RNA polymerase III. Retroviral vectors expressing tRNA-ribozymes were tested in a human CD4+ T cell line and were shown to inhibit HIV-1 replication. These tRNA3Lys-based hammerhead ribozymes should therefore prove to be valuable for both basic and applied research. Special application is sought in HIV-1 or HIV-2 gene therapy.     

Selection of intracellularly active ribozymes in mammalian cells.

Ribozymes are expected to be useful as antiviral agents and powerful tools of functional analysis of unknown gene products in vivo. For use of ribozymes in vivo, they must be fully functional in the intracellular environment. Not all ribozymes selected in vitro would be expected to work in vivo, whereas ribozymes selected in the intracellular environment should retain their function in vivo. With the eventual aim of using ribozymes as antiviral agents or biological tools in mammalian cells, we then devised a novel selection system in mammalian cells of active ribozymes by targeting at a gene for the cyclin dependent kinase inhibitor (CDKI), p16INK4a. In this system, we found that p16INK4a-knockdown cells became malignant and they formed foci. In the mammalian system, we confirmed that the selected cells harbored the active ribozyme, indicating that our positive selection systems in vivo were operational.     

Suppression of BCR-ABL mRNA by various ribozymes in HeLa cells.

Ribozymes are RNA molecules with enzymatic activity that can cleave target RNA molecules in a sequence specific manner. To date, various types of ribozyme have been constructed to cleave other RNAs and such trans-acting ribozymes include hammerhead, hairpin and HDV ribozymes. External guide sequence (EGS) can also induce the suppression of a gene-expression by taking advantage of cellular RNase P. Here we compared the activities of various functional RNA cleavers both in vitro and in vivo. The first purpose of this comparison was intended to determine the best ribozyme motif with the highest activity in cells. The second purpose is to know the correlation between the activities of ribozymes in vitro and in vivo. Our results indicated that the intrinsic cleavage activity of ribozymes is not the sole determinant that is responsible for the activity of a ribozyme in cultured cells.     

Expression of a chimeric ribozyme gene results in endonucleolytic cleavage of target mRNA and a concomitant reduction of gene expression in vivo.

The subclass of catalytic RNAs termed ribozymes cleave specific target RNA sequences in vitro. Only circumstantial evidence supports the idea that ribozymes may also act in vivo. In this study, ribozymes with a hammerhead motif directed against a target sequence within the mRNA of the neomycin phosphotransferase gene (npt) were embedded into a functional chimeric gene. Two genes, one containing the ribozyme and the other producing the target, were cotransfected into plant protoplasts. Following in vivo expression, a predefined cleavage product of the target mRNA was detected by ribonuclease protection. Expression of both the ribozyme gene and the target gene was driven by the CaMV 35S promoter. Concomitant with the endonucleolytic cleavage of the target mRNA, a complete reduction of NPT activity was observed. An A to G substitution within the ribozyme domain completely inactivates ribozyme-mediated hydrolysis but still shows a reduction in NPT activity, albeit less pronounced. Therefore, the reduction of NPT activity produced by the active ribozyme is best explained by both hydrolytic cleavage and an antisense effect. However, the mutant ribozyme--target complex was more stable than the wildtype ribozyme--target complex. This may result in an overestimation of the antisense effect contributing to the overall reduction of gene expression.     

Design and specificity of hammerhead ribozymes against calretinin mRNA.

We obtained a partial sequence of mouse calretinin mRNA from cDNA clones, and designed hammerhead ribozymes to cleave positions within it. With a view to optimising hammerhead ribozymes for eliminating the mRNA in vivo, we varied the length and sequence of the three duplex 'arms' and measured the cleavage of long RNA substrates in vitro at 37 degrees C (as well as 50 degrees C). Precise cleavage occurred, but it could only go to completion with a large excess of ribozyme. The evidence suggests that the rate-limiting step with a large target is not the cleavage, but the formation of the active ribozyme: substrate complex. The efficiency varied unpredictably according to the target site, the length of the substrate RNA, and the length of the ribozyme; secondary structure in vitro may be responsible. We particularly investigated the degree of sequence-specificity. Some mismatches could be tolerated, but shortening of the total basepairing with the substrate to less than 14 bp drastically reduced activity, implying that interaction with weakly-matched RNAs is unlikely to be a serious problem in vivo. These results suggest that specific and complete cleavage of a mRNA in vivo should be possible, given high-level expression of a ribozyme against a favourable target site.     

A strategy for developing a hammerhead ribozyme for selective RNA cleavage depending on substitutional RNA editing

Substitutional RNA editing plays a crucial role in the regulation of biological processes. Cleavage of target RNA that depends on the specific site of substitutional RNA editing is a useful tool for analyzing and regulating intracellular processes related to RNA editing. Hammerhead ribozymes have been utilized as small catalytic RNAs for cleaving target RNA at a specific site and may be used for RNA-editing-specific RNA cleavage. Here we reveal a design strategy for a hammerhead ribozyme that specifically recognizes adenosine to inosine (A-to-I) and cytosine to uracil (C-to-U) substitutional RNA-editing sites and cleaves target RNA. Because the hammerhead ribozyme cleaves one base upstream of the target-editing site, the base that pairs with the target-editing site was utilized for recognition. RNA-editing-specific ribozymes were designed such that the recognition base paired only with the edited base. These ribozymes showed A-to-I and C-to-U editing-specific cleavage activity against synthetic serotonin receptor 2C and apolipoprotein B mRNA fragments in vitro, respectively. Additionally, the ribozyme designed for recognizing A-to-I RNA editing at the Q/R site on filamin A (FLNA) showed editing-specific cleavage activity against physiologically edited FLNA mRNA extracted from cells. We demonstrated that our strategy is effective for cleaving target RNA in an editing-dependent manner. The data in this study provided an experimental basis for the RNA-editing-dependent degradation of specific target RNA in vivo.     

Specific inhibition of macrophage TNF-alpha expression by in vivo ribozyme treatment.

The overproduction of the cytokine TNF-alpha is associated with inflammatory and autoimmune diseases. We have developed a means to block TNF-alpha production with ribozymes directed against TNF-alpha mRNA to selectively inhibit its production in vitro and in vivo. Following cationic lipid-mediated delivery to peritoneal murine macrophages in culture, anti-TNF-alpha ribozymes were more effective inhibitors of TNF-alpha secretion than catalytically inactive ribozyme controls. Inhibition of TNF-alpha secretion was proportional to the concentration of ribozyme administered, with an IC50 of approximately 10 nM. After i.p. injection of cationic lipid/ribozyme complexes, elicited macrophages accumulated approximately 6% of the administered ribozyme. The catalytically active ribozyme suppressed LPS-stimulated TNF-alpha secretion by approximately 50% relative to an inactive ribozyme control without inhibiting secretion of another proinflammatory cytokine produced by macrophages, IL-1alpha. Ribozyme-specific TNF-alpha mRNA degradation products were found among the mRNA extracted from macrophages following in vivo ribozyme treatment and ex vivo stimulation. Thus, catalytic ribozymes can accumulate in appropriate target cells in vivo; once in the target cell, ribozymes can be potent inhibitors of specific gene expression.     

Inhibition of HIV-1 replication by ribozymes that show poor activity in vitro.

Self-cleaving RNAs (ribozymes) can be engineered to cleave target RNAs of choice in a sequence-specific manner (1). Consequently, they could be used to inhibit virus replication or to analyse host gene function in vivo. However, ribozymes that are catalytic in vitro are generally disappointing when analysed in cells unless expressed at high levels relative to their target RNAs (2, 3). Here we provide evidence that this can be overcome by optimizing ribozyme structure using cellular rather than cell-free assays. We show that ribozymes of relatively long flanking complementary regions (FCRs), while poor catalysts in vitro, can produce profound inhibition of HIV replication in cells. By examining a series of ribozymes in which the FCRs vary from 9 to 564 nucleotides, we establish that the optimum length for activity in the cell is > or = 33 nucleotides.     

Ribozyme inhibition of alphavirus replication

A model system to examine the expression and antiviral activity of trans-acting ribozymes in mammalian cells has been developed and evaluated. Hairpin ribozymes were engineered to cleave a specific site, identified by a combinatorial activity-based selection method, within genomic and subgenomic RNA species of Sindbis virus. Transiently transfected cells expressed moderate levels of ribozyme (approximately 50,000 molecules/cell) with predominant nuclear localization and a short half-life (23 min). Stable cell lines expressed ribozymes at modest levels (approximately 2,000 molecules/cell). Ribozyme-mediated RNA cleavage activity was detected in cell extracts. Clonal cell lines were challenged with recombinant Sindbis virus, and viral replication was examined using plaque formation and green fluorescent protein assays. Significant inhibition of viral replication was observed in cells expressing the active antiviral ribozyme, and lower levels of inhibition in control cells expressing inactive or irrelevant ribozymes. These findings are consistent with a model in which inhibition of viral replication occurs via ribozyme cleavage of viral RNAs, suggesting that ribozymes may represent useful antiviral agents.     

Can ribozymes be used to regulate procaryote gene expression?

The in vivo activity of ribozymes designed against mRNA coding for E. coli beta-galactosidase was tested both in intramolecular and in intermolecular conditions. When recombinant M13 phage DNA carrying on the same molecule the information for both the ribozyme and the target was transfected into bacterial cells, ribozyme activity was observed. Conversely, a ribozyme coded by a recombinant M13 vector, but targeted against an mRNA transcribed from the F episome including the remaining part of the beta-galactosidase gene, was inefficient.     

The effect of structure in a long target RNA on ribozyme cleavage efficiency.

Inhibition of gene expression by catalytic RNA (ribozymes) requires that ribozymes efficiently cleave specific sites within large target RNAs. However, the cleavage of long target RNAs by ribozymes is much less efficient than cleavage of short oligonucleotide substrates because of higher order structure in the long target RNA. To further study the effects of long target RNA structure on ribozyme cleavage efficiency, we determined the accessibility of seven hammerhead ribozyme cleavage sites in a target RNA that contained human immunodeficiency virus type 1 (HIV-1) vif - vpr . The base pairing-availability of individual nucleotides at each cleavage site was then assessed by chemical modification mapping. The ability of hammerhead ribozymes to cleave the long target RNA was most strongly correlated with the availability of nucleotides near the cleavage site for base pairing with the ribozyme. Moreover, the accessibility of the seven hammerhead ribozyme cleavage sites in the long target RNA varied by up to 400-fold but was directly determined by the availability of cleavage sites for base pairing with the ribozyme. It is therefore unlikely that steric interference affected hammerhead ribozyme cleavage. Chemical modification mapping of cleavage site structure may therefore provide a means to identify efficient hammerhead ribozyme cleavage sites in long target RNAs.     

CTAB-mediated enrichment for active forms of novel dimeric maxizymes

We demonstrated previously that shortened forms of (stem II-deleted) hammerhead ribozymes with low intrinsic activity form very active dimers with a common stem II (very active short ribozymes capable of forming dimers were designated maxizymes). As a result of such a dimeric structure, heterodimeric maxizymes are potentially capable of cleaving a substrate at two different sites simultaneously. In this case, active heterodimers are in equilibrium with inactive homodimers. Longer forms of common stem II can lead to enrichment of the active heterodimers in vitro. In this study, we investigated whether the cationic detergent CTAB, which is known to enhance strand displacement of nucleic acids, might inhibit the dimerization of maxizymes. Significantly, under all conditions examined, CTAB instead enhanced the activity of a variety of maxizymes, with the extent of enhancement depending on the conditions. The activity of our least stable, least active maxizyme was enhanced 100-fold by CTAB. The strand displacement activity of CTAB thus appears to enhance the conversion of alternative conformations of inactive maxizymes, with intra- and inter-molecular hydrogen bonds, to active forms. Thus, our smallest maxizyme can also be considered a potential candidate for a gene-inactivating agent in vivo, in view of the fact that various facilitators of strand displacement reactions are known to exist in vivo (indeed, a separate experiment in cell culture supported the conclusion that our smallest maxizyme is a good gene-inactivating agent). Although activities of ribozymes in vitro do not necessarily reflect their activities in vivo, our findings suggest that the activity of ribozymes in vivo can be better estimated by running ribozyme kinetics in the presence of CTAB in vitro.