Targeting Highly Structured RNA by Cooperative Action of siRNAs and Helper Antisense Oligomers in Living Cells

RNA target accessibility is one of the most important factors limiting the efficiency of RNA interference-mediated RNA degradation. However, targeting RNA viruses in their poorly accessible, highly structured regions can be advantageous because these regions are often conserved in sequence and thus less prone to viral escape. We developed an experimental strategy to attack highly structured RNA by means of pairs of specifically designed small interfering RNAs and helper antisense oligonucleotides using the 5’ untranslated region (5’UTR) of coxsackievirus B3 as a model target. In the first step, sites accessible to hybridization of complementary oligonucleotides were identified using two mapping methods with random libraries of short DNA oligomers. Subsequently, the accessibility of the mapped regions for hybridization of longer DNA 16-mers was confirmed by an RNase H assay. Using criteria for the design of efficient small interfering RNAs (siRNA) and a secondary structure model of the viral 5’UTR, several DNA 19-mers were designed against partly double-stranded RNA regions. Target sites for DNA 19-mers were located opposite the sites which had been confirmed as accessible for hybridization. Three pairs of DNA 19-mers and the helper 2’-O-methyl-16-mers were able to effectively induce RNase H cleavage in vitro. For cellular assays, the DNA 19-mers were replaced by siRNAs, and the corresponding three pairs of siRNA-helper oligomer tools were found to target 5’UTR efficiently in a reporter construct in HeLa cells. Addition of the helper oligomer improved silencing capacity of the respective siRNA. We assume that the described procedure will generally be useful for designing of nucleic acid-based tools to silence highly structured RNA targets.     

Terbium-mediated footprinting probes a catalytic conformational switch in the antigenomic hepatitis delta virus ribozyme

The two forms of the hepatitis delta virus ribozyme are derived from the genomic and antigenomic RNA strands of the human hepatitis delta virus (HDV), where they serve a crucial role in pathogen replication by catalyzing site-specific self-cleavage reactions. The HDV ribozyme requires divalent metal ions for formation of its tertiary structure, consisting of a tight double-nested pseudoknot, and for efficient self- (or cis-) cleavage. Comparison of recently solved crystal structures of the cleavage precursor and 3' product indicates that a significant conformational switch is required for catalysis by the genomic HDV ribozyme. Here, we have used the lanthanide metal ion terbium(III) to footprint the precursor and product solution structures of the cis-acting antigenomic HDV ribozyme. Inhibitory Tb(3+) binds with high affinity to similar sites on RNA as Mg(2+) and subsequently promotes slow backbone scission. We find subtle, yet significant differences in the terbium(III) footprinting pattern between the precursor and product forms of the antigenomic HDV ribozyme, consistent with differences in conformation as observed in the crystal structures of the genomic ribozyme. In addition, UV melting profiles provide evidence for a less tight tertiary structure in the precursor. In both the precursor and product we observe high-affinity terbium(III) binding sites in joining sequence J4/2 (Tb(1/2) approximately 4 microM) and loop L3, which are key structural components forming the catalytic core of the HDV ribozyme, as well as in several single-stranded regions such as J1/2 and the L4 tetraloop (Tb(1/2) approximately 50 microM). Sensitized luminescence spectroscopy confirms that there are at least two affinity classes of Tb(3+) binding sites. Our results thus demonstrate that a significant conformational change accompanies catalysis in the antigenomic HDV ribozyme in solution, similar to the catalytic conformational switch observed in crystals of the genomic form, and that structural and perhaps catalytic metal ions bind close to the catalytic core.     

The influence of imperfectly paired helices I and III on the catalytic activity of hammerhead ribozymes.

Several catalytic antisense RNAs directed against different regions of the genomic or antigenomic RNA of Sendai virus were constructed. All RNAs contained the same catalytic domain based on hammerhead ribozymes but some had deletions or mutations resulting in imperfect helices I and III. Pre-annealed substrate/ribozyme complexes were used to determine the rates of the cleavage process for the different ribozymes under single-turnover conditions. It was found that the sequence context surrounding the cleavable motif influenced the cleavage efficiencies. Deletions or mutations of nucleotides 2.1 or 15.1 and 15.2 according to the numbering system for hammerhead ribozymes of Hertel et al. destroyed catalytic activity. Deletions of nucleotide 2.2 or additional nucleotides in the helix I-forming region of the ribozyme did not destruct, but only reduced the cleavage efficiencies. Similar results were observed for a deletion of nucleotide 15.3. Simultaneous deletions within helices I and III resulted in alternative cleavage sites. The potential consequences for the specificity of the ribozyme reaction are discussed.     

Antigenomic delta ribozyme variants with mutations in the catalytic core obtained by the in vitro selection method

We have used the in vitro selection method to search for catalytically active variants of the antigenomic delta ribozyme with mutations in the regions that constitute the ribozyme active site: L3, J1/4 and J4/2. In the initial combinatorial library 16 nt positions were randomized and the library contained a full representation of all possible sequences. Following ten cycles of selection-amplification several catalytically active ribozyme variants were identified. It turned out that one-third of the variants contained only single mutation G80U and their activity was similar to that of the wild-type ribozyme. Unexpectedly, in the next one-third of the variants the C76 residue, which was proposed to play a crucial role in the ribozyme cleavage mechanism, was mutated. In these variants, however, a cytosine residue was present in a neighboring position to the polynucleotide chain. It shows that the ribozyme catalytic core possesses substantial ‘structural plasticity’ and the capacity of functional adaptation. Four selected ribozyme variants were subjected to more detailed analysis. It turned out that the variants differed in their relative preferences towards Mg²⁺, Ca²⁺ and Mn²⁺ ions. Thus, the functional properties of the variants were dependent on both the structure of their catalytic sites and divalent metal ions performing catalysis.     

Oligonucleotides comprised of alternating 2'-deoxy-2'-fluoro-beta-D-arabinonucleosides and D-2'-deoxyribonucleosides (2'F-ANA/DNA 'altimers') induce efficient RNA cleavage mediated by RNase H

Chimeric oligonucleotides comprised of alternating residues of 2'-deoxy-2'-fluoro-D-arabinonucleic acid (2'F-ANA) and DNA were synthesized and evaluated for an important antisense property-the ability to elicit ribonuclease H (RNase H) degradation of complementary RNA. Experiments used both human RNase HII and Escherichia coli RNase HI. Mixed backbone oligomers comprising alternating three-nucleotide segments of 2'F-ANA and three-nucleotide segments of DNA were the most efficient at eliciting RNase H degradation of target RNA, and were significantly better than oligonucleotides entirely composed of DNA, suggesting that these mixed backbone oligonucleotides may be potent antisense agents.     

RNase H is responsible for the non-specific inhibition of in vitro translation by 2'-O-alkyl chimeric oligonucleotides: high affinity or selectivity, a dilemma to design antisense oligomers.

Ribonuclease H (RNase H) which recognizes and cleaves the RNA strand of mismatched RNA-DNA heteroduplexes can induce non-specific effects of antisense oligonucleotides. In a previous paper [Larrouy et al. (1992), Gene, 121, 189-194], we demonstrated that ODN1, a phosphodiester 15mer targeted to the AUG initiation region of alpha-globin mRNA, inhibited non-specifically beta-globin synthesis in wheat germ extract due to RNase H-mediated cleavage of beta-globin mRNA. Specificity was restored by using MP-ODN2, a methylphosphonate-phosphodiester sandwich analogue of ODN1, which limited RNase H activity on non-perfect hybrids. We report here that 2'-O-alkyl RNA-phosphodiester DNA sandwich analogues of ODN1, with the same phosphodiester window as MP-ODN2, are non-specific inhibitors of globin synthesis in wheat germ extract, whatever the substituent (methyl, allyl or butyl) on the 2'-OH. These sandwich oligomers induced the cleavage of non-target beta-globin RNA sites, similarly to the unmodified parent oligomer ODN1. This is likely due to the increased affinity of 2'-O-alkyl-ODN2 chimeric oligomers for both fully and partly complementary RNA, compared to MP-ODN2. In contrast, the fully modified 2'-O-methyl analogue of ODN1 was a very effective and highly specific antisense sequence. This was ascribed to its inability (i) to induce RNA cleavage by RNase H and (ii) to physically prevent the elongation of the polypeptide chain.     

Structural features of target RNA molecules greatly modulate the cleavage efficiency of trans-acting delta ribozymes

The aim of this work was to shed some more light on factors influencing the effectiveness of delta ribozyme cleavage of structured RNA molecules. An oligoribonucleotide that corresponds to the 3'-terminal region X of HCV RNA and yeast tRNAPhe were used as representative RNA targets. Only a few sites susceptible to ribozyme cleavage were identified in these targets using a combinatorial library of ribozyme variants, in which the region responsible for ribozyme-target interaction was randomized. On the other hand, the targets were fairly accessible for binding of complementary oligonucleotides, as was shown by 6-mer DNA libraries and RNase H approach. Moreover, the specifically acting ribozymes cleaved the targets precisely but with unexpectedly modest efficacy. To explain these observations, six model RNA molecules were designed, in which the same seven nucleotide long sequence recognized by the delta ribozyme was always single stranded but was embedded into different RNA structural context. These molecules were cleaved with differentiated rates, and the corresponding k2 values were in the range of 0.91-0.021 min-1; thus they differed almost 50-fold. This clearly shows that cleavage of structured RNAs might be much slower than cleavage of a short unstructured oligoribonucleotide, despite full accessibility of the targeted regions for hybridization. Restricted possibilities of conformational transitions, which are necessary to occur on the cleavage reaction trajectory, seem to be responsible for these differences. Their magnitude, which was evaluated in this work, should be taken into account while considering the use of delta ribozymes for practical applications.     

A pH-sensitive RNA tertiary interaction affects self-cleavage activity of the HDV ribozymes in the absence of added divalent metal ion

Self-cleavage of the genomic and antigenomic ribozymes from hepatitis delta virus (HDV) requires divalent cation for optimal activity. Recently, the HDV genomic ribozyme has been shown to be active in NaCl in the absence of added divalent metal ion at low pH (apparent pKa 5.7). However, we find that the antigenomic ribozyme is 100 to 1000-fold less active under similar conditions. With deletion of a three-nucleotide sequence (C41-A42-A43) unique to the genomic ribozyme, the rate constant for cleavage decreased substantially, while activity of the antigenomic ribozyme was enhanced by introducing a CAA sequence. From the crystal structure, it has been proposed that C41 in this sequence is protonated. To investigate a possible connection between activity at low pH and protonation of C41, mutations were made that were predicted to either eliminate protonation or alter the nature of the tertiary interaction upon protonation. In the absence of added Mg2+, these mutations reduced activity and eliminated the observed pH-rate dependence. Thermal denaturation studies revealed a pH-sensitive structural feature in the genomic ribozyme, while unfolding of the mutant ribozymes was pH-independent. We propose that, in the absence of added Mg2+, protonation of C41 contributes to enhanced activity of the HDV genomic ribozyme at low pH.     

Selective binding of trisamine-modified phosphorothioate antisense DNA to target mRNA improves antisense activity and reduces toxicity

Antisense activity in living cells has been thought to occur via a mechanism involving both DNA-mediated hybridization arrest of target mRNA and RNase H-mediated mRNA digestion. Therefore an ideal antisense agent should be permeable to the cell and possess capacities (1) to form a thermally stable duplex in vivo with its target, (2) to discriminate between mRNAs with different degrees of complementarity, and (3) to form antisense/RNA complexes that are susceptible to RNase H hydrolysis. A trisamine-modified deoxyuridine derivative of a novel phosphorothioate DNA 15-mer that meets all these criteria is described here. Compared with the unmodified phosphorothioate oligomer, the phosphorothioate derivative exhibits a higher antisense activity as well as reduced cytotoxicity in cells infected with HIV-1. Our data suggest that the melting temperature (T(m)) between antisense DNA and the target mRNA is not only one of the factors contributing to this derivative's improved antisense activity. Also important are an enhanced ability to discriminate between sequences and an increased susceptibility of the DNA/mRNA complex to RNase H hydrolysis. These results will be useful in designing more active, clinically useful antisense drugs.     

Delta ribozyme has the ability to cleave in transan mRNA.

We report here the first demonstration of the cleavage of an mRNA in trans by delta ribozyme derived from the antigenomic version of the human hepatitis delta virus (HDV). We characterized potential delta ribozyme cleavage sites within HDV mRNA sequence (i.e. C/UGN6), using oligonucleotide binding shift assays and ribonuclease H hydrolysis. Ribozymes were synthesized based on the structural data and then tested for their ability to cleave the mRNA. Of the nine ribozymes examined, three specifically cleaved a derivative HDV mRNA. All three active ribozymes gave consistent indications that they cleaved single-stranded regions. Kinetic characterization of the ability of ribozymes to cleave both the full-length mRNA and either wild-type or mutant small model substrate suggests: (i) delta ribozyme has turnovers, that is to say, several mRNA molecules can be successively cleaved by one ribozyme molecule; and (ii) the substrate specificity of delta ribozyme cleavage is not restricted to C/UGN6. Specifically, substrates with a higher guanosine residue content upstream of the cleavage site (i.e. positions -4 to -2) were always cleaved more efficiently than wild-type substrate. This work shows that delta ribozyme constitutes a potential catalytic RNA for further gene-inactivation therapy.     

Fast and accurate determination of sites along the FUT2 in vitro transcript that are accessible to antisense oligonucleotides by application of secondary structure predictions and RNase H in combination with MALDI-TOF mass spectrometry

Alteration of gene expression by use of antisense oligonucleotides has considerable potential for therapeutic purposes and scientific studies. Although applied for almost 25 years, this technique is still associated with difficulties in finding antisense-effective regions along the target mRNA. This is mainly due to strong secondary structures preventing binding of antisense oligonucleotides and RNase H, playing a major role in antisense-mediated degradation of the mRNA. These difficulties make empirical testing of a large number of sequences complementary to various sites in the target mRNA a very lengthy and troublesome procedure. To overcome this problem, more recent strategies to find efficient antisense sites are based on secondary structure prediction and RNase H-dependent mechanisms. We were the first who directly combined these two strategies; antisense oligonucleotides complementary to predicted unpaired target mRNA regions were designed and hybridized to the corresponding RNAs. Incubation with RNase H led to cleavage of the RNA at the respective hybridization sites. Analysis of the RNA fragments by matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometry, which has not been used in this context before, allowed exact determination of the cleavage site. Thus the technique described here is very promising when searching for effective antisense sites.     

Substrate shape specificity of E coli RNase P ribozyme is dependent on the concentration of magnesium ion

The bacterial RNase P ribozyme can accept a hairpin RNA with CCA-3' tag sequence as well as a cloverleaf pre-tRNA as substrate in vitro, but the details are not known. By switching tRNA structure using an antisense guide DNA technique, we examined the Escherichia coli RNase P ribozyme specificity for substrate RNA of a given shape. Analysis of the RNase P reaction with various concentrations of magnesium ion revealed that the ribozyme cleaved only the cloverleaf RNA at below 10 mM magnesium ion. At 10 mM magnesium ion or more, the ribozyme also cleaved a hairpin RNA with a CCA-3' tag sequence. At above 20 mM magnesium ion, cleavage site wobbling by the enzyme in tRNA-derived hairpin occurred, and the substrate specificity of the enzyme became broader. Additional studies using another hairpin substrate demonstrated the same tendency. Our data strongly suggest that raising the concentration of metal ion induces a conformational change in the RNA enzyme.     

Potent gene-specific inhibitory properties of mixed-backbone antisense oligonucleotides comprised of 2'-deoxy-2'-fluoro-D-arabinose and 2'-deoxyribose nucleotides

Phosphorothioate deoxyribonucleotides (PS-DNA) are among the most widely used antisense inhibitors. PS-DNA exhibits desirable properties such as enhanced nuclease resistance, improved bioavailability, and the ability to induce RNase H mediated degradation of target RNA. Unfortunately, PS-DNA possesses a relatively low binding affinity for target RNA that impacts on its potency in antisense applications. We recently showed that phosphodiester-linked oligonucleotides comprised of 2'-deoxy-2'-fluoro-D-arabinonucleic acid (FANA) exhibit both high binding affinity for target RNA and the ability to elicit RNase H degradation of target RNA [Damha et al. (1998) J. Am. Chem. Soc. 120, 12976]. In the present study, we evaluated the antisense activity of phosphorothioate-linked FANA oligonucleotides (PS-FANA). Oligonucleotides comprised entirely of PS-FANA were somewhat less efficient in directing RNase H cleavage of target RNA as compared to their phosphorothioate-linked DNA counterparts, and showed only weak antisense inhibition of cellular target expression. However, mixed-backbone oligomers comprised of PS-FANA flanking a central core of PS-DNA were found to possess potent antisense activity, inhibiting specific cellular gene expression with EC(50) values of less than 5 nM. This inhibition was a true antisense effect, as indicated by the dose-dependent decrease in both target protein and target mRNA. Furthermore, the appearance of mRNA fragments was consistent with RNase H mediated cleavage of the mRNA target. We also compared a series of PS-[FANA-DNA-FANA] mixed-backbone oligomers of varying PS-DNA core sizes with the corresponding 2'-O-methyl oligonucleotide chimeras, i.e., PS-[2'meRNA-DNA-2'meRNA]. Both types of oligomers showed very similar binding affinities toward target RNA. However, the antisense potency of the 2'-O-methyl chimeric compounds was dramatically attenuated with decreasing DNA core size, whereas that of the 2'-fluoroarabino compounds was essentially unaffected. Indeed, a PS-FANA oligomer containing a single deoxyribonucleotide residue core retained significant antisense activity. These findings correlated exactly with the ability of the various chimeric antisense molecules to elicit RNase H degradation of the target RNA in vitro, and suggest that this mode of inhibition is likely the most important determinant for potent antisense activity.     

Efficient reduction of target RNAs by small interfering RNA and RNase H-dependent antisense agents. A comparative analysis

RNA interference can be considered as an antisense mechanism of action that utilizes a double-stranded RNase to promote hydrolysis of the target RNA. We have performed a comparative study of optimized antisense oligonucleotides designed to work by an RNA interference mechanism to oligonucleotides designed to work by an RNase H-dependent mechanism in human cells. The potency, maximal effectiveness, duration of action, and sequence specificity of optimized RNase H-dependent oligonucleotides and small interfering RNA (siRNA) oligonucleotide duplexes were evaluated and found to be comparable. Effects of base mismatches on activity were determined to be position-dependent for both siRNA oligonucleotides and RNase H-dependent oligonucleotides. In addition, we determined that the activity of both siRNA oligonucleotides and RNase H-dependent oligonucleotides is affected by the secondary structure of the target mRNA. To determine whether positions on target RNA identified as being susceptible for RNase H-mediated degradation would be coincident with siRNA target sites, we evaluated the effectiveness of siRNAs designed to bind the same position on the target mRNA as RNase H-dependent oligonucleotides. Examination of 80 siRNA oligonucleotide duplexes designed to bind to RNA from four distinct human genes revealed that, in general, activity correlated with the activity to RNase H-dependent oligonucleotides designed to the same site, although some exceptions were noted. The one major difference between the two strategies is that RNase H-dependent oligonucleotides were determined to be active when directed against targets in the pre-mRNA, whereas siRNAs were not. These results demonstrate that siRNA oligonucleotide- and RNase H-dependent antisense strategies are both valid strategies for evaluating function of genes in cell-based assays.     

Selective inhibition of cell-free translation by oligonucleotides targeted to a mRNA hairpin structure.

Using an in vitro selection approach we have previously isolated oligodeoxy aptamers that can bind to a DNA hairpin structure without disrupting the double-stranded stem. We report here that these oligomers can bind to the RNA version of this hairpin, mostly through pairing with a designed 6 nt anchor. The part of the aptamer selected against the DNA hairpin did not increase stability of the RNA-aptamer complex. However, it contributed to the binding site for Escherichia coli RNase H, leading to very efficient cleavage of the target RNA. In addition, a 2'- O -methyloligoribonucleotide analogue of one selected sequence selectively blocked in vitro translation of luciferase in wheat germ extract by binding to the hairpin region inserted upstream of the initiation codon of the reporter gene. Therefore, non-complementary oligomers can exhibit antisense properties following hybridization with the target RNA. Our study also suggests that in vitro selection might provide a means to extend the repertoire of sequences that can be targetted by antisense oligonucleotides to structured RNA motifs of biological importance.     

Mechanism of DNA chain growth. XI. Structure of RNA-linked DNA fragments of Escherichia coli

The size of RNA attached to nascent DNA fragments of Escherichia coli with a chain length of 400 to 2000 nucleotides is estimated to be about 50 to 100 nucleotides from: (a) the density of the molecules of known sizes; (b) the decrease of the molecular size produced by hydrolysis with RNases or alkali; and (c) the size of RNA released by DNase treatment. Only a small decrease in molecular size is produced by RNase or alkali treatment, excluding the possibility that the RNA is located in the middle of the fragment or that ribonucleotide sequences are scattered in the molecule. The RNA is not located at the 3′ end of the molecule either, since the DNA is degraded by 3′ → 5′ exonuclease action of bacteriophage T4 DNA polymerase which has neither RNase nor DNA endonuclease activity. Positive evidence for the covalent attachment of the RNA to the 5′ end of the DNA is provided by the finding that one 5′-OH terminus of DNA is created from each RNA-linked DNA fragment by alkaline hydrolysis. The quantitative production of the 5′-OH group at the 5′ end of DNA is also found upon hydrolysis with pancreatic RNase, indicating that the 3′-terminal base of the RNA segment of the fragments is a pyrimidine. On the other hand, when the RNA-linked DNA fragments hydrolysed with alkali or pancreatic RNase are incubated with [γ-³²P]ATP and polynucleotide kinase and the DNA thus labelled is degraded to constituent 5′-mononucleotides, the ³²P is found only in dCMP. Therefore, C is the specific 5′-terminal base of the DNA segment of the RNA-linked DNA fragments, and the RNA-DNA junction has the structure … p(rPy)p(dC)p …     

A nested double pseudoknot is required for self-cleavage activity of both the genomic and antigenomic hepatitis delta virus ribozymes.

The crystal structure of a genomic hepatitis delta virus (HDV) ribozyme 3' cleavage product predicts the existence of a 2 bp duplex, P1.1, that had not been previously identified in the HDV ribozymes. P1.1 consists of two canonical C-G base pairs stacked beneath the G.U wobble pair at the cleavage site and would appear to pull together critical structural elements of the ribozyme. P1.1 is the second stem of a second pseudoknot in the ribozyme, making the overall fold of the ribozyme a nested double pseudoknot. Sequence comparison suggests the potential for P1.1 and a similar fold in the antigenomic ribozyme. In this study, the base pairing requirements of P1.1 for cleavage activity were tested in both the genomic and antigenomic HDV ribozymes by mutagenesis. In both sequences, cleavage activity was severely reduced when mismatches were introduced into P1.1, but restored when alternative base pairing combinations were incorporated. Thus, P1.1 is an essential structural element required for cleavage of both the genomic and antigenomic HDV ribozymes and the model for the antigenomic ribozyme secondary structure should also be modified to include P1.1.