RNA
RNA as a potential target for drug design
The field of RNA ligand design is much less advanced than that for DNA or proteins, but given the much more irregular nature of cavities and grooves formed by RNA, compared to the DNA minor groove’s essentially endless helical ramblings, the situation for RNA holds great promise. The nature of enzyme active-sites is highly disymmetric and the cavities beginning to be recognised for RNA ligand binding are also disymmetric. This probably means that the success of ligand design in enzymology will transfer to RNA, but such optimism must be tempered with recognition that RNA provides a very more flexible template and the problems of conformational change which plague ligand design de novo for proteins may also haunt the RNA world. In view of the accessibility of cellular RNAs to xenobiotics and the paucity of RNA repair mechanisms it is now recognised by that RNA may offer a therapeutically relevant target for drug design.
An understanding of the molecular features that lead to strong, specific binding to RNA is a prerequisite to designing RNA-directed ligands as drug leads. Recently in vitro selection of RNA sequences has been used to determine preferred RNA sequences and structures for antibiotics, eg for neomycin and tobramycin, but this approach does not produce new RNA binding structures. The complementary approach is to search for novel RNA ligand structures.
The Douglas and Bichenkova groups have studied various aspects of RNA systems, including the maturation of transfer RNA by the ribonuclease P ribozyme in collaboration with groups in Japan and the design of ligands for tRNA and ribonuclease P. They have also worked with collaborators in Russia on the design and synthesis of artificial ribonucleases to specifically cleave RNA strands.
Ligand design against Transfer RNA and the ribonuclease P ribozyme
To develop our understanding of ligand design and recognition principles for RNA sites, we have chosen transfer RNA as a model system with a well defined tertiary structure. The relatively few classes of reversible ligands so far reported for tRNA are: metal ions, intercalators, polyamines, soft electrophiles, porphyrins, distamycin and netropsin, benzimidazole heterocycles. There are also some reagents which cleave tRNA and obviously these will have a binding/ recognition component to their actions.
The Douglas and Bichenkova groups have shown that novel heterocyclic compounds provide an additional class of ligands for tRNA. Ligands of the Hoechst 33258 family bind selectively to yeast tRNAPhe with 1:1 stoichiometry and binding constants in the micromolar range [S. E. Ebrahimi, A. N. Wilton, K. T. Douglas, Chem. Commun., (1997), 385] led to our discovery of strong, specific, reversible binding ligands of other benzimidazole families for transfer RNA and their study by fluorescence and NMR spectroscopies [E.V.Bichenkova, S.E.Sadat-Ebrahimi, A.N.Wilton, N.O’Toole, D.S.Marks and K.T.Douglas, Nucleosides and Nucleotides (1998), 17, 1651].
From this they were able to develop inhibitors of the ribozyme ribonuclease P, which cleaves a precursor tRNA molecule to yield the mature functional tRNA [Y. Hori, E.V. Bichenkova, A.N. Wilton, M.N. El-Attug, S. Sadat-Ebrahimi, T. Tanaka, Y. Kikuchi, M. Araki, Y. Sugiura, and K.T. Douglas, Biochemistry (2001) 40, 603] as RNase P recognizes the three-dimensional structure of tRNA, which is maintained across both precursor (pre-tRNA) and mature tRNA. As the T-stem groove of E. coli tRNAPhe in the region of y-55 is a possible ligand-binding site for these benzimidazole ligands and this T-loop is also implicated in recognition properties of RNase P, they were tested in collaboration with Japanese colleagues in Toyohashi University (Professor Yo Kikuchi) and Kyoto University (Professor Yukio Sugiura) and found to be very strong inhibitors of RNase P, with I50 values in the low micromolar range. These provided the first examples of synthetic inhibitors of a ribozyme and were the strongest inhibitors at that time of ribonuclease P. However, the cationic porphyrin, meso-tetrakis(N-methyl-pyridyl)porphine also binds specifically to tRNA and some porphyrins and porphines inhibit the activity of E. coli RNase P even more strongly than the benzimidazoles [Y. Hori, E.V. Bichenkova, A.N. Wilton, T. Tanaka, K.T. Douglas and Y. Kikuchi Nucl. Acids Res. Suppl. 2, (2002) 111].
Artificial ribonucleases
Another project related to DNA/RNA-based research involves development of oligonucleotide-based non-metallated artificial ribonucleases (Professor Ken Douglas, Dr Elena Bichenkova). Design of artificial ribonucleases (AR) is normally based on mimicking the compositional and structural properties of the active sites of natural ribonucleases (eg RNase A or T1) by chemical conjugation of peptide-like fragments that normally constitute their hydrolytic domain. The binding site of an AR, which is composed of an oligonucleotide complementary to the target RNA, designed to deliver a hydrolytic construct(s) to the RNA target via specific Watson-Crick hydrogen-bonding. We can now correlate observed hydrolytic abilities against a tRNA target of a series of synthesised ARs with conformations obtained from molecular modelling. Our future projects in this area will study structural aspects of AR action by high-field NMR to determine the structural rules that govern binding and hydrolytic properties [M.M. Fabani, M.A. Zenkova, E. V. Bichenkova, N. N. Polushin, V.V. Silnikov, K.T.Douglas and V.V. Vlassov, Nucleic Acids Res. (2004) in press].