RNA and DNA as Catalysts

F. Eckstein
Max-Planck-Institut für experimentelle Medizin
Hermann-Rein-Str. 3, 37075 Göttingen ((ext. link)http://www.mpiem.gwdg.de/Abtlg/Arb-Gr/Eckstein/index.html)

DNA and RNA are usually seen as carriers of genetic information, with the occasional function of RNA as a scaffold, as for example in the ribosome. This picture has changed with the discovery of self-splicing of certain RNAs, most notably that of the Tetrahymena group I intron by Tom Cech, and the active role of RNA in the RNase P in the process of maturation of tRNAs by Sid Altman. These observations laid the foundation for the concept of catalytic RNA for which the Nobel prize of 1989 was awarded to these two colleagues.

Natural catalytic RNAs
These two first examples of catalytic RNA have been joined by others such as the group II intron, the hepatitis delta RNA and the plant viroid RNAs of the hammerhead and hairpin type (1). Although some of these reactions are facilitated by proteins in vivo, their reactions can be simulated in vitro without the aid of proteins, supporting the view that indeed the RNA by itself can do the job. With the exception of RNase P RNA, the RNAs undergo an intramolecular, or in cis, transesterification in these transformations. These reactions are sequence specific, and more importantly, occur with rates orders of magnitude faster than expected from the reactivity of RNA. Thus there must be a catalytic element involved, and therefore these active RNAs are called ribozymes, in analogy to the classical protein enzymes. To study the mechanism of these ribozyme-catalysed reactions and their structural requirements, they were redesigned for intermolecular, or in trans, reactions. Thus substrate and ribozyme are localised on different RNA strands. This arrangement facilitates the determination of kinetic constants such as kcat and Km values. An example for an in trans arrangement is shown for the hammerhead ribozyme (Fig. 1). A comparison of the efficiency of ribozyme- with enzyme-catalysed reactions shows that even though the reaction rate kcat is often smaller for the ribozymes, they compensate for it by a more favourable Km so that in the end efficiency is comparable for both catalysts (2).

map size 9K

Fig. 1. Generalised two dimensional structure of the hammerhead ribozyme for cleavage in trans. N, any nucleotide; N´, nucleotide complementary to N; H, any nucleotide except G; underlined, invariant core.

Mechanistic aspects
Most of these RNA-catalysed reactions depend on the presence of a divalent metal ion such as Mg²⁺ but some, such as the hairpin and the hammerhead ribozyme, can also be coaxed into action with high concentrations of monovalent ions such as Na⁺. Very interestingly, certain positively charged aminoglycosides can replace the metal ions for the hairpin ribozyme. Thus the question as to the role of the metal ions arises. Formation of a kinetically competent three-dimensional structure is quite certainly one of them. However, particularly in the presence of Mg²⁺ where most ribozymes work best, a more active role can be envisaged where it acts as a base and as a Lewis acid as originally proposed in the "two metal ion" mechanism for enzymes (3). Such mechanisms have been formulated particularly for the group I RNA and the hammerhead ribozymes. In these cases, one metal ion is thought to abstract a proton from the nucleophilic hydroxy group to attack the phosphate group, and the other coordinates to the departing hydroxy group (Fig. 2). Although there is experimental support for such a metal ion hydrate role, no metal ions which could take on this function have been detected in x-ray structures of these ribozymes. Thus, the final proof for the involvement of the metal ions as active participants in the chemical step of the reactions is still lacking. This of course becomes a concern in light of some of the ribozymes not requiring metal ions. Looking for alternative functional groups one has to consider whether the nucleobases could not play such a role. The argument against this idea in the past has been that the pKa values of these groups are outside the pH optimum of the reactions. NMR studies have however shown that the pKa of adenosine, for example, can be shifted upwards by 2.5 units to approximately 6.5 as a consequence of a change in the micro-environment in the three-dimensional structure of the ribozymes. Thus it might well be that such groups can function as acid-base catalysts at neutral pH and that we have to reconsider our prejudice against nucleotides as active participants in such catalytic reactions.

map size 6K

Fig. 2. Suggested hammerhead ribozyme cleavage mechanism with participation of two metal ions.

New ribozymes
The catalytic power of the naturally occuring ribozymes which all catalyse phosphoryl tranfers within RNA molecules, has stimulated the search for ribozymes with other activities. This has been made possible by the technique of in vitro selection. It involves the construction of double-stranded DNA containing a T7 promoter followed by a stretch of randomised nucleotides. Sequences necessary for binding of oligonucleotides as polymerase primers and for cloning are at both ends of such constructs. After transcription the RNA is incubated with the substrate and active RNA is separated from inactive RNA sequences. The success of the procedure depends crucially on the design of this selection step. Often substrates with a biotin moiety are used which is transferred to the active RNA by the catalysed reaction. This facilitates separation of the active from the inactive RNA on a Streptavidin column. The active species are amplified by PCR, transcribed and incubated with the substrate again. This cycle is repeated between 6 and a dozen times until enough active RNA has accumulated for characterisation.. The method is extremely powerful, enabling the selection of active molecules from a large pool of sequences. Thus, for example, a DNA with 22 randomised positions has 4²² sequences which is equivalent to 1.7x10¹³ sequences. This represents a rather small degree of randomisation in the generally performed selection procedures. However, there are certain constraints created by Avogadro´s number, i. e. the number of molecules in one mol which is 6x10²³. It so happens that 39 randomisations is equivalent to 3.7x10²³ sequences. Thus to cover the whole sequence space one would have to work with a 1 M DNA solution in one liter for the first selection. This becomes forbiddingly expensive in terms of enzyme required. Of course, larger degrees of randomisations can be used with the penalty that not the entire sequence space will be searched.

Such selections have yielded a most impressive number of ribozymes, catalysing the most diverse reactions. Examples are phosphoryl transfer like a kinase reaction, nucleoside triphosphate polymerisation, acyl transfer, N-glycosidic bond formation to obtain a nucleotide, and formation of a carbon-carbon bond in a Diels-Alder reaction. An example for peptide bond formation is shown in fig. 3 (4).

map size 10K

Fig. 3. Example of an in vitro selected ribozyme for peptide bond formation. Bold line, ribozyme. Taken from ref. 4.

Surprisingly, the activity of ribozymes can be allosterically controlled. An example is the flavin-dependent activity of a hammerhead ribozyme to which a flavin-binding aptamer has been fused. In contrast to this positive effect, an ATP aptamer-ribozyme construct is inhibited in the presence of ATP. These examples obviously make ribozymes more akin to proteozymes. Besides the interest in mechanistic questions, such ribozymes might be very useful as therapeutic agents, as discussed below, as their activity can be modulated by small molecules.

DNAzymes
It was generally assumed that RNA was better suited as a catalyst than DNA because of its flexibility and the presence of the 2´-hydroxy group. It thus came as a big surprise when Joyce selected a DNAzyme capable of cleaving RNA, and this with somewhat better efficiency than a ribozyme (Fig. 4) (2). Thus our preconceived ideas have been proven wrong. DNA as a single-stranded molecule can apparently fold into three dimensional structures similarly to RNA, and the 2´-hydroxy group is dispensable for catalytic action. As ribozymes, DNAzymes can also be made, by selection, to depend on a cofactor. This has been demonstrated for a histidine-dependent DNAzyme for RNA hydrolysis. This narrows the gap to proteozymes and offers the possibility of extending the range of catalysed reactions.

So far, selections have been performed with the aim in mind of mimicking reactions which are also catalysed by proteozymes. Besides the interesting mechanistic aspects of the RNA- or DNA-catalysed reactions, the ribozyme-catalysed reactions particularly are often considered as support for the prebiotic RNA world where RNA might have played the part of our present day enzymes. As the number of various types of reactions catalysed by ribozymes increases, this hypothesis gains in strength. Another aspect for the future is the development, by in vitro selection, of nucleic acid enzymes to catalyse reactions for which no natural enzymes exist. This could be of industrial importance.

map size 3K

Fig. 4.COLOR="#009999"> The first RNA-cleaving DNAzyme. R, either adenosine or guanosine. Taken from ref. 2.

Application
A very active area of research for ribozymes is their application for the inhibition of gene expression with the aim to develop them as therapeutic agents (5). Although this area is still in its infancy it will probably gain momentum as the advantages of such an approach for this purpose are significant. In this methodology the ribozyme is used for the sequence specific destruction of a mRNA, thus preventing the expression of a gene. The inherent catalytic power of ribozymes to cleave the mRNA is considered an advantage over the presently often used antisense-oligodeoxynucleotide approach which relies on the presence of RNase H for this cleavage reaction. Obviously if successful, They can be introduced for this purpose into cells either exogenously, where the ribozyme is synthesised first and then applied to cells or organisms, very much like a classical drug, or endogenously where the ribozyme gene is transfected as part of a vector into cells where it is then transcribed. This latter method resembles that of gene therapy and, as there, can lead to stable expression of the ribozyme. Clinical trials using both methods are under way.

Summary
Nucleic acid enzymes represent a new, powerful and interdisciplinary, field of interest. It holds great promise not only to gain better insight into catalysis as such but also to be developed for various applications, in particular in medicine.

References
1. Carola, C. and Eckstein, F. (1999) Nucleic acid enzymes. Curr. Opin. Chem. Biol. 3: 274-283.
2. Santoro, S. W. and Joyce, G. F. (1997) A general purpose RNA-cleaving DNA enzyme. Proc. Natl. Acad. Sci. USA 94: 4262-4266.
3. Steitz, T. A. and Steitz, J. A. (1993) A general two-metal-ion mechanism for catalytic RNA. Proc. Natl. Acad. Sci. USA 90: 6498-6502.
4. Zhang, B. and Cech, T. R. (1998) Peptidyl-transferase ribozymes: trans reactions, structural characterisation and ribosomal RNA-like structures. Chem. Biol. 5: 539-553.
5. Bramlage, B., Luzi, E. and Eckstein, F. (1998) Designing ribozymes for the inhibition of gene expression. TIBTECH 16: 434-438.

Impressum