Intracellular ribozyme applications.

The exquisite target selectivity of trans -acting ribozymes has fostered their use as potential therapeutic agents and tools for down-regulating cellular transcripts. In living cells, free diffusion of RNAs is extremely limited, if it exists at all. Thus, getting ribozymes to base-pair with their cognate targets requires co-localizing the ribozyme transcript with the target RNA. In addition, not all sites along a given target RNA are equally accessible to ribozyme base pairing. Cellular proteins greatly influence the trafficking and structure of RNA, and therefore making ribozymes work effectively in cells a significant challenge. This article addresses the problems of getting engineered ribozymes to effectively pair with and cleave targets in cells. The work described here illuminates methods for target-site selection on native mRNAs, methods for ribozyme expression, and strategies for obtaining a discrete intracellular localization of ribozymes.

Detection of antisense and ribozyme accessible sites on native mRNAs: application to NCOA3 mRNA.

The efficacies of antisense oligonucleotides and ribozymes are greatly dependent on the accessibility of their mRNA targets. Target site accessibility is affected by both RNA structure and the proteins associated along the length of the RNA. To mimic the native state of mRNA for site identification, we have previously used endogenous mRNAs in cellular extracts as targets for defined sequence oligodeoxynucleotides (ODNs) designed to identify both antisense pairing and potential ribozyme cleavage sites. The rationale for this approach is that the specific pairing of an ODN with a mRNA forms a DNA:RNA hybrid that is cleaved by the endogenous RNaseH in the cell extract. To extend the usefulness of this basic approach, we report here the use of semi-random ODN libraries to identify hammerhead ribozyme cleavage sites. Thus, the most accessible sites for antisense and ribozyme base pairing are selected by this approach. A novel feature of the approach described here is the use of terminal transferase-dependent PCR (TDPCR) after reverse transcription to estimate the cleavage efficiency and to precisely determine the RNaseH and ribozyme cleavage sites on mRNAs in cell extracts following treatment with ODN or ribozyme libraries. As a model system, we have targeted the NCOA3 (also known as AIB-1) mRNA in cell extracts. The NCOA3 mRNA encodes a nuclear receptor co-activator that is amplified and over-expressed in a high proportion of breast and ovarian cancers. A highly accessible site on this mRNA was identified, and a ribozyme targeted to this site was demonstrated to effectively downregulate NCOA3 function in cells.

In vivo, high-resolution analysis of yeast and mammalian RNA-protein interactions, RNA structure, RNA splicing and ribozyme cleavage by use of terminal transferase-dependent PCR.

We have investigated the analysis of RNA by use of terminal transferase-dependent PCR (TDPCR), a procedure previously used for the analysis of DNA and chromatin [J. Komura and A.D.Riggs, Nucleic Acids Res.,26, 1807-1811 (1998)]. When preceded by reverse transcription (RT), TDPCR provides an extremely sensitive, versatile, quantitative and nucleotide-level assay for detecting RNA lesions or structures that block primer extension during the RT step. The procedure is: (i) RT using a gene-specific oligonucleotide; (ii) ribo-tailing of the single-stranded cDNA product by use of terminal deoxy-nucleotidyl transferase; (iii) ligation of a DNA linker to the tailed cDNA by use of T4 DNA ligase; and (iv) PCR using a nested, gene-specific primer and a linker-specific primer. This procedure combines the versatility of a primer extension assay with nucleotide-level resolution, the specificity of nested primers and the sensitivity of PCR. Band patterns obtained are reproducible and quantifiable. We successfully used the technique for the study of yeast RNA structure, splicing intermediates and ribozyme cleavage. Also, in vivo footprint experiments, using mammalian cells and RNase T1, revealed the binding of iron-responsive element binding protein to iron responsive elements in the mRNAs of transferrin receptor and ferritin H-chain.

Cooperative interaction of branch signals in the actin intron of Saccharomyces cerevisiae.

In pre-mRNA splicing, specific spliceosomal components recognize key intron sequences, but the mechanisms by which splice sites are selected arenot completely understood. In the Saccharomyces cerevisiae actin intron a silent branch point-like sequence (UACUAAG) is located 7 nt upstream of the canonical sequence. Mutation of the canonicalUACUAAC sequence to UAAUAAC reduces utilization of this signal and activates the cryptic UACUAAG. Splicing-dependent beta-galactosidase assays have shown that these two splice signals cooperate to enhance splicing. Analyses of several variants of this double branch point intron demonstrate that the upstream UACUAAG sequence significantly increases usage of the UAAUAAC as a site of lariat formation. This activation is sequence-specific and unidirectional. However the ability of the UACUAAG signal to activate the downstream branch point is dependent on the presence of a short non-conserved sequence located a few nucleotides upstream of the UACUAAG. Mutation of this sequence leads to the disappearance of the cooperative interactions between the two branch signals. Our results show that this non-conserved sequence and the UACUAAG signal must both be present to achieve activation of the downstream branch point and suggest that a specific structure may be necessary to allow efficient recognition of the UAAUAAC.

Structural similarities between hammerhead ribozymes and the spliceosomal RNAs could be responsible for lack of ribozyme cleavage in yeast.

Hammerhead ribozymes have been proposed as potential therapeutic agents for the treatment of viral and other diseases. However, a clear understanding of the cleavage reaction in vivo is not available at present. In these studies, we chose the yeast Saccharomyces cerevisiae as a model system to study the effects of hammerhead cleavage on gene expression in vivo. Several reporter genes were employed to monitor the self-cleaving characteristics of three different ribozymes. We show that these ribozymes decrease expression of some reporter genes by interfering with splice site selection or translation initiation and not by in vivo cleavage of the RNA transcripts. In fact, it appears that although these ribozymes can efficiently self-cleave the RNA in vitro, they are not able to function in vivo. We have identified a yeast splicing protein that interacts in vivo with our cis-ribozyme by specifically recognizing the ribozyme structure (Lin and Rossi, 1996). This interaction does not occur if different secondary structures are used in place of the ribozyme. The binding of this protein to the ribozyme can account for the inability of ribozymes to efficiently cleave in yeast. Remarkably, when yeast extracts are added to in vitro trans-cleavage reactions, the cleavage ability of the ribozyme is hampered, whereas the addition of mammalian extracts yields an enhancement of the reactions. These results confirm the presence of factor(s) that can block ribozyme function in the yeast intracellular environment.

Unusual interactions between cleavage products of a cis-cleaving hammerhead ribozyme.

We have synthesized and tested a cis-cleaving ribozyme designed to have thermodynamically stable stem-loop structures. This cis-ribozyme cleaves very efficiently in vitro, with a cleavage rate of about 0.5/min. Surprisingly, during the course of in vitro transcription and cleavage of our ribozyme, a product of unusual mobility accumulates and coincides with a sharp decline in the rate of formation of cleavage products. Analyses of this electrophoretic variant demonstrated that it is formed by interactions of the cleavage products. Despite the fact that the products and ribozyme transcript are of identical sequence, the cleavage products interact only with one another and not with the uncleaved precursor. This suggests a significant structural difference between the cleaved and uncleaved ribozyme transcripts. Testing of this cis-ribozyme in both yeast and mammalian cells shows no significant cleavage activity in vivo. We conclude that the structure of the ribozyme flanking sequences is important for optimizing the rate of ribozyme cleavage, but this enhanced rate does not necessarily correlate with enhanced in vivo function.

The expression cassette determines the functional activity of ribozymes in mammalian cells by controlling their intracellular localization.

In order to better understand the influence of RNA transcript context on RNA localization and catalytic RNA efficacy in vivo, we have constructed and characterized several expression cassettes useful for transcribing short RNAs with well defined 5' and 3' appended flanking sequences. These cassettes contain promoter sequences from the human U1 snRNA, U6 snRNA, or tRNA Meti genes, fused to various processing/stabilizing sequences. The levels of expression and the sub-cellular localization of the resulting RNAs were determined and compared with those obtained from Pol II promoters normally linked to mRNA production, which include a cap and polyadenylation signal. The tRNA, Ul, and U6 transcripts were nuclear in localization and expressed at the highest levels, while the standard Pol II promoted transcripts were cytoplasmic and present at lower levels. The ability of these cassettes to confer ribozyme activity in vivo was tested with two assays. First, an SIV-growth hormone reporter gene was transiently transfected into human embryonic kidney cells expressing an anti-SIV ribozyme. Second, cultured T lymphocytes expressing an anti-HIV ribozyme were challenged with HIV. In both cases, we found that the ribozymes were effective only when expressed as capped, polyadenylated RNAs transcribed from Pol II cassettes that generate a cytoplasmically localized ribozyme that facilitates co-localization with its target. We also show that the inability of the other cassettes to support ribozyme-mediated inhibitory activity against their cytoplasmic target is very likely due to the resulting nuclear localization of these ribozymes. These studies demonstrate that the ribozyme expression cassette determines its intracellular localization and, hence, its corresponding functional activity.

Ribozymes expressed within the loop of a natural antisense RNA form functional transcription terminators.

As an initial step towards developing a widely applicable system for expressing small ribozymes (Rz) in various cell types using T7 RNA polymerase, we have replaced the loop domain of a natural prokaryotic antisense RNA (RNAout from Tn10), with hammerhead (Hh) Rz. RNAout was chosen, because the stem of its secondary structure gives it an unusually long half life in Escherichia coli which should also confer in vivo stability to small RNA sequences expressed within its loop domain. In order to define the 3' end of the Rz-RNAout chimeric RNAs, a poly(U) tract was inserted just 3' of the RNAout stem. Molecular analysis indicates that these RNAs function both as transcription terminators and Rz. In addition, the RNAs are stable in E. coli and can be expressed in mammalian cells. These results show that certain characteristics of a naturally evolved RNA, RNAout in this case, can be used to provide additional functions to short RNAs containing Hh Rz without disrupting the enzymatic activity of the Rz.

Isolation and functional analysis of a Kluyveromyces lactis RAP1 homologue.

Saccharomyces cerevisiae RAP1 (Sc RAP1) is an essential protein which interacts with diverse genetic loci within the cell. RAP1 binds site-specifically to the consensus sequence, 5'-AYCYRTRCAYYW (UASRPG, where R = A or G, W = A or T, Y = C or T). In Kluyveromyces lactis (Kl) ribosomal protein-encoding genes (rp) retain functional RAP1-binding elements, suggesting the presence of a RAP1-like factor. Kl extracts display an activity capable of specifically binding to rp fragments bearing UASRPG. We subsequently isolated the Kl RAP1-encoding gene by homology to a subfragment which encodes the N terminus of the DNA-binding domain of Sc RAP1. The predicted amino acid (aa) sequence of Kl RAP1 indicates it is smaller than Sc RAP1 (666 vs. 827 aa) with the N terminus being truncated. The DNA-binding domain is virtually identical between the two RAP1 proteins, while the RIF1 domain is moderately conserved. The region between these two domains and the N-termini are highly divergent. Two potential UASRPG were identified in the 5' flanking region, suggesting an autoregulatory role for RAP1. Despite the similarities between the two proteins, KI RAP1 is unable to complement Sc rap1ts mutants, suggesting that domains essential for function in Sc are absent from the Kl protein.

Small sequence insertions within the branch point region dictate alternative sites of lariat formation in a yeast intron.

The problem of intron recognition in S. cerevisiae appears to be in part solved by the strong conservation of intron encoded splicing signals, in particular the 5' GUAUGU and the branch point UACUAAC which interact via base pairing with the RNA components of U1 and U2 snRNPs respectively. Nevertheless, the mere presence of such signals is insufficient for splicing to occur. In the S. cerevisiae ACT1 intron, a silent UACUAAC-like sequence (UACUAAG) is located 7 nucleotides upstream of the canonical branch point signal. In order to investigate whether other factors, in addition to the U2-UACUAAC base-pair interactions, affect branch point selection in yeast, we created a cis-competition assay by converting the UACUAAG to a strong branch point signal (UACUAAC). If simply having a canonical UACUAAC sequence were sufficient for lariat formation, a 1:1 ratio in usage of the two signals should have been observed. In this double branch point intron, however, the downstream UACUAAC is utilized preferentially (4:1). Results obtained from the analyses of numerous sequence variants flanking the two UACUAAC sequences, demonstrate that non-conserved sequences in the branch point region are able to define lariat formation. Consequently, we conclude that U2 base-pairing is not the only requirement determining branch point selection in yeast, and local structure in the vicinity of the branch point could play a critical role in its recognition.

Biological and functional aspects of catalytic RNAs.

Several naturally occurring ribozymes have now been well characterized with respect to their in vivo and in vitro activities. Through detailed biochemical and genetic analyses, it has become possible to alter the substrate specificity of each ribozyme using simple Watson-Crick base pairing. Several laboratories, therefore, have designed ribozymes to cleave viral or other cellular transcripts in vitro with the hope of developing these molecules as antiviral or therapeutic agents. In addition to Watson-Crick base pairing, however, other factors such as protein or RNA tertiary interactions are involved in the ribozyme cleavage activity. Although several engineered ribozymes have been used successfully to reduce gene expression in vivo, it is difficult to determine whether gene expression has been reduced by the cleaving activity of the ribozyme or by its inherent antisense activity. In order to discriminate between these two activities and optimize potentially therapeutic ribozymes, it is imperative to develop in vivo assays in which the antisense activity of ribozymes is negligible.