Itamar Willner and colleagues from The Hebrew University of Jerusalem, Israel, discuss the applications of DNA-based enzymes.

RNA and DNA molecules, like proteins, have complex three-dimensional structures that depend on the sequence of their building blocks - though whereas proteins have twenty amino acids, RNA or DNA molecules have only four types of nucleotide to play with. Still, this variety, together with single-stranded and double-stranded domains, can give complex structures that, much like enzymes, can selectively bind substrates and catalyse useful chemical reactions. Such nucleic acid based catalysts are called DNAzymes and ribozymes and hold great promise as chemical sensors; tools to construct nanostructures; and molecular machines and computing systems.

The idea of designing structures that selectively bind substrates (aptamers), or act as catalytic enzyme-like DNA strands, became practical in the 1990s, following the development of the Systematic Evolution of Ligands by Exponential Enrichment (SELEX) process. Here, nucleic acids with specific binding properties, or affinities towards a particular transition-state analogue, are fished out of a library of 10¹⁵ nucleic acids and amplified by the polymerase chain reaction (PCR). The aptamers and catalytic nucleic acids made this way are, in effect, man-made analogues of protein-based antibodies and enzymes, respectively. But the nucleic acid enzymes have advantages over their protein analogues: DNA is chemically very stable; the enzymes can be efficiently machine-synthesised by PCR; and one can even couple aptamers with DNAzymes, yielding hybrids which not only bind to specific substrates, but also have enzyme-like catalytic activity.

Some of the broadest applications of DNAzymes have been in the development of biosensors. For example, among the many DNAzymes and ribozymes prepared in recent years is a sequence of nucleic acids that mimics the action of the enzyme horseradish peroxidase. Together with hydrogen peroxide and an appropriate substrate, this DNAzyme generates a colour change, which can be used to detect nucleic acids or as a marker for cancer cells.

Similarly, metal ions such as lead or copper can be detected by nucleic acid strands that become catalytic when they bind around metals. Once catalytically active, these metal-dependent DNAzymes cleave a specific part of a DNA sequence, which acts as a fluorescent signal advertising the metal's presence.

DNAzyme-based systems have even been suggested as potential substitutes for the PCR protocol as a way to detect small amounts of DNA. Nucleic acids have been designed that, upon recognising a required piece of DNA, stimulate self-assembling syntheses of DNAzyme units. The accumulation of the DNAzyme provides a catalytic label, amplifying the original sensing event.

DNAzymes can also be used as tools for shaping and correcting nanostructures. Gold nanoparticles, for instance, can be forced into a blue-coloured crosslinked assembly when nucleic acid strands attached to each particle hybridise. A DNAzyme can cleave these nucleic acid crosslinks, turning the blue assembly into red-coloured individual nanoparticles: a sensitive read-out signal. In this way, DNAzymes have been used as proof-reading units that check through nanoparticle mixtures and remove any erroneous crosslinking.

Some ingenious molecular machines based on DNA are driven by DNAzymes. DNAzyme-containing nucleic acid structures have duplicated the mechanical functions of a scissor, while DNAzymes have also been used to cleave nucleic acid strands, allowing them to 'walk' along a DNA or RNA track. At first glance, these concepts seem only to satisfy scientific curiosity, but the emerging systems highlight some extremely valuable and promising applications of DNA-based machines.

One further application of DNAzymes relates to their use as 'smart' biomolecules that perform logic operations for computing systems. Nucleic acids of pre-designed sequences have been used as templates that activate, in the presence of appropriate nucleic acids as inputs, logic gate operations. These logic functions have been used, most famously, to make molecular calculators and a system that can play noughts and crosses (tic-tac-toe). Such DNA-based computing systems are not aimed to substitute man-made computers but to complement computer science and could eventually furnish new perspectives for drug design and nanomedicine in the future.

Read Willner et al's tutorial review 'DNAzymes for Sensing, Nanobiotechnology and Logic Gate Applications' in issue 6 of Chemical Society Reviews.