News and Publications

Catalytic Nucleic Acids for Treating the Molecular Basis of Disease

In recent years, antisense technology has emerged as a promising approach for the treatment of cancer and viral diseases. The antisense strategy uses short oligonucleotides that bind particular cellular RNAs, leading to inactivation of the RNAs. The discovery and subsequent development of catalytic nucleic acids have enhanced antisense technology by providing agents that both recognize and inactivate a target. We used in vitro evolution techniques to develop novel nucleic acid enzymes that cleave RNA or DNA targets in a sequence-specific manner. We are investigating the structure and mechanism of these enzymes and are applying them to the inactivation of disease-related genes in living cells.

A DNA Enzyme That Cleaves RNA

Four years ago, using in vitro selection, we developed the first example of a DNA enzyme. Since then, we have produced several additional examples, including a DNA enzyme that can be directed to cleave almost any targeted RNA under simulated cellular conditions. Compared with synthetic RNA enzymes, DNA enzymes are easier to prepare and are more stable in biological tissues. Our preferred DNA enzyme is the "10-23" motif, which contains a catalytic domain of 15 nucleotides flanked by substrate-recognition domains of 7--10 nucleotides each. The enzyme binds the RNA substrate through Watson-Crick pairing, enabling us to target different RNAs simply by altering the sequence of the recognition domains.

The 10-23 DNA enzyme appears to use a magnesium cofactor to assist in deprotonation of a particular 2´-hydroxyl group within the RNA substrate, resulting in cleavage of the adjacent phosphoester bond. The catalytic efficiency of the enzyme is the highest of any known nucleic acid enzyme. Catalytic efficiency is determined by the rate of enzyme-substrate association. After binding, the substrate is cleaved by the enzyme, and the 2 cleavage products are released at an even faster rate, allowing the enzyme to turn over rapidly. Substrate recognition involves 14--20 Watson-Crick base pairs, which is sufficient to specify a unique sequence in the human genome. Substrates that contain a single base mismatch relative to the enzyme are not cleaved efficiently because they are released from the enzyme faster than they are cleaved.

Application to Human Disease

The favorable kinetic properties of the 10-23 DNA enzyme make it an attractive candidate for development as a therapeutic agent. We have directed the enzyme to cleave a variety of disease-related target RNAs, including HIV type 1 gag-pol mRNA, human CCR5 mRNA (encodes a coreceptor for HIV type 1), human insulin-like growth factor 1 mRNA (associated with tumor cell proliferation), and human Her2/neu mRNA (associated with tumor metastasis). Chemical modifications have been introduced into the DNA enzyme to improve its stability and bioavailability.

Structure of a DNA Enzyme

Because DNA enzymes were discovered so recently, little is known about their 3-dimensional structure. In collaboration with W. Chazin, The Scripps Research Institute, we used nuclear magnetic resonance spectroscopy to study the DNA enzyme in complex with a cleavage-resistant analog of the RNA substrate. The data confirmed the existence of Watson-Crick pairing interactions between the enzyme and substrate, but the structure of the catalytic core of the enzyme could not be resolved. In collaboration with D. Stout, The Scripps Research Institute, we did x-ray crystallographic studies of the same molecules. We obtained crystals of the enzyme-substrate complex that diffracted to high resolution, as well as 6 heavy-atom derivatives that diffracted in a similar manner. Relying on data collected at the synchrotron at Stanford University, we solved the crystal structure of the complex at 3.0-Å resolution (Fig. 1).

We were disappointed to find that the enzyme and substrate had rearranged upon entering the crystal lattice to form an H-shaped 3-helix junction. The long arms of the H are composed of RNA-DNA duplexes, each with a complete strand of the RNA substrate and a substrate-recognition domain from each of 2 different DNA enzymes. The catalytic cores of these 2 DNA enzymes form the crossarm of the H, which is composed of a DNA-DNA duplex with loops of single-stranded DNA at each end. This structure is not the active structure of the DNA enzyme. It is nonetheless a remarkable 82-nucleotide structure that indicates the ability of nucleic acids to form tight helical junctions. We are continuing to pursue the crystal structure of the enzyme-substrate complex in its active conformation, which would guide us in formulating the DNA enzyme for therapeutic applications.

Publications

Joyce, G.F. Nucleic acid enzymes: Playing with a fuller deck. Proc. Natl. Acad. Sci. U.S.A. 95:5845, 1998.