Deoxyribozymes performing logic operations and simple computations
| Joanne Macdonald and Darko Stefanovic (2012), Scholarpedia, 7(6):10021. | doi:10.4249/scholarpedia.10021 | revision #137054 [link to/cite this article] |
Ribozymes are nucleic acid enzymes, naturally occurring single-stranded catalytic RNA, with various biological functions. Their DNA counterparts, deoxyribozymes, are not known in nature. However, they have been identified and can be readily synthesised. In vitro systems that use deoxyribozymes to perform logic operations and simple computations have been reported starting in the early 2000s.
The catalytic activity effected by deoxyribozymes that has been used for computational purposes is the cleavage of nucleic acid strands through the breaking of a phosphodiester bond. To modulate the catalytic activity, deoxyribozyme molecules can be structurally modified; for example, through the addition of binding domains that specifically bind to yet other nucleic acid strands. These modifications are designed such that the presence of a ligand nucleic acid strand either strongly promotes, or strongly inhibits, the catalytic activity of the deoxyribozyme. The ligand nucleic acid strand is thus the controlling input that switches the activity of the modified deoxyribozyme on or off. When active, the deoxyribozyme produces a nucleic acid strand that can represent an output. The modified deoxyribozyme can then be thought of as a logic gate. In this fashion, logic operations and simple computations are carried out over signals represented as concentrations of oligonucleotides, short single-stranded DNA molecules.
Contents |
Catalysis in deoxyribozymes, its control, and its observation
Deoxyribozymes
The switch part of a deoxyribozyme logic gate is derived from a deoxyribozyme, a nucleic acid enzyme that catalyzes DNA reactions. In particular, the enzyme can be a phosphodiesterase, which cleaves an oligonucleotide substrate; commonly, the 8-17 deoxyribozyme (Santoro and Joyce, 1997) is used for this purpose. The cleavage results in two shorter oligonucleotide products. Either product, or both, can be used as the output signal of the logic gate.
Monitoring the output of deoxyribozyme logic gates
The output signal from a gate can be used as an input for a downstream gate to make a larger circuit. Ultimately, some gate outputs represent the end result of the computation in the circuit. To monitor those outputs, fluorescent readout can be used. Tsubstrates are labelled with fluorescent dyes, e.g., the “red”-channel TAMRA (T) dye, whose fluorescence is quenched by the Black-Hole 2 (BH2) quencher, which absorbs all of the TAMRA fluorescence. After cleavage, the TAMRA is separated from the BH2 and the fluorescence is no longer absorbed. This leads to an increase in fluorescence within the mixture, which is monitored via fluorescence spectroscopy. Other dye-quencher pairs can be used, and more than one pair can be used simultaneously, for real-time readout of multiple signals.
Controlling deoxyribozyme activity using stem-loop modules
To make a deoxyribozyme into a switch sensitive to an input DNA strand, a stem-loop module can be added to a branch of the deoxyribozyme. The stem component of the module blocks the access of the substrate to its recognition region in the deoxyribozyme. However, the loop component contains a single-stranded region available for binding by a complementary single-stranded DNA, the input oligonucleotide. When this input oligonucleotide is added, it binds to the loop, which destabilizes the stem, and the stem-loop module opens, allowing the substrate access to the deoxyribozyme. Thus, the conformational change in the stem-loop module dependent on the presence of an input oligonucleotide causes a change in the catalytic activity of the deoxyribozyme-stem-loop complex, i.e., the deoxyribozyme logic gate is switched on.
By using sufficiently long DNA inputs, and paying attention to any cross-reactivity, it is possible to realize circuits in which many different inputs selectively bind to specific deoxyribozyme logic gates so that all inputs and stem-loop regions function in the same mixture without undesirable gate activation from the wrong inputs. Because the stem-loop module is sufficiently separated from the catalytic core of the deoxyribozyme, the choice of input sequences is relatively unconstrained, allowing circuits with large numbers of inputs to be constructed in principle.
Basic deoxyribozyme logic gates
A number of deoxyribozyme-based logic gates, forming a set functionally complete for Boolean logic, have been developed (Stojanovic et al., 2002).
YES gate
The simplest deoxyribozyme logic gate is a repeater, also known as a YES gate. In this gate, the deoxyribozyme has been modified to include a single stem-loop module to regulate the binding of substrate to the enzyme. If the stem-loop is closed, the substrate cannot bind, the enzyme is inactivated, and no outputs are formed. However, when input DNA is added, it hybridizes to the stem-loop region and alters the conformation of the gate molecule. The conformational change causes the enzyme to become active, allowing cleavage of the substrate to produce the output DNA. That is, a YESx gate is active in the presence of a single input ix.
NOT gate
Another type of molecular logic gate is the NOT gate. In it, the catalytic core of the enzyme is modified to include a stem-loop module that regulates enzyme activity. If the stem-loop is closed, the enzyme is active. However, when an input oligonucleotide is added, it hybridizes to the loop component and alters the conformation of the gate molecule, deforming the catalytic core. The conformational change causes the enzyme to become inactive, preventing cleavage of the substrate even if it is bound to the enzyme. Thus, a NOTz gate is active unless a single input iz is added, which inactivates the gate.
AND gate
By using combinations of the modular structures described above for YES and NOT loop structures, further Boolean logic gate structures can be made. The AND gate is made by combining two activating stem-loop structures:
ANDANDNOT gates
The AND gate and the NOT gate are a complete set, in the sense that any logic function can be expressed as a composition of some number of AND and NOT gates, under the assumption that they can be arbitrarily connected. Thus, in principle, no other gates are necessary. But the stem-loop module design allows one more kind of logic gate to be implemented using a single enzyme, and this may be preferable to a cascaded two-enzyme circuit. The ANDANDNOT gate is created by combining two activating stem-loop modules and one inhibitory stem-loop module. For input signals x, y, and z, it computes x AND y AND NOT z.
Applications
In deoxyribozyme logic gates the inputs and outputs are of the same kind (namely oligonucleotides). This allows the cascading of gates without any external interfaces. The inputs are compatible with sensor molecules (aptamers) that could detect cellular disease markers, and the outputs can be tied to the release of small molecules, such as drugs. Thus it may eventually be possible to make therapeutic decisions cell-by-cell according to a complex decision function based on many attributes of the cell. This is sometimes referred to as "intelligent drug delivery". Clinically informed therapeutic decision procedures are likely to be simple computations.
Simple computations
Arithmetic primitives
Half-adder
A molecular half-adder has been constructed (Stojanovic and Stefanovic, 2003a) that is able to calculate a sum and carry digit by adding two DNA "bits" of information, using three molecular logic gates and a two-color fluorogenic output system in a single tube.
Full Adder
A molecular full-adder has been constructed (Lederman et al. 2006) that is able to calculate a sum and carry digit by adding three bits together, using 7 logic gates and a two-color fluorogenic output system in a single tube.
Automata for games of strategy
Automata are autonomous machines that are able to analyze a series of stimuli and respond to them. Deoxyribozyme logic gates have been used to build automata based on a logical abstraction to play simple games of strategy.
MAYA-I
MAYA-I (Stojanovic and Stefanovic 2003b) is an automaton that plays a symmetry-pruned restricted game of tic-tac-toe (noughts and crosses). The automaton always goes first in the middle well of a nine-well (3 by 3) game board representation, and the human player can respond by moving in only two wells.
MAYA-II
MAYA-II (Macdonald et al. 2006) plays a full, non-symmetry-pruned complete game of tic-tac-toe. The automaton still always goes first in the middle well, but its human opponent may move in any remaining square. The automaton displays both human and automaton moves using a two-colour fluorogenic output system.
References
- Lederman, H; Macdonald, J; Stefanovic, D and Stojanovic, M N (2006). Deoxyribozyme-Based Three-Input Logic Gates and Construction of a Molecular Full Adder Biochemistry 45(4): 1194-1199. doi:10.1021/bi051871u.
- Macdonald, J et al. (2006). Medium Scale Integration of Molecular Logic Gates in an Automaton Nano Letters 6(11): 2598-2603. doi:10.1021/nl0620684.
- Pei, R; Matamoros, E; Liu, M; Stefanovic, D and Stojanovic, M N (2010). Training a molecular automaton to play a game Nature Nanotechnology 5: 773-777. doi:10.1038/nnano.2010.194.
- Santoro(1997). A general purpose RNA-cleaving DNA enzyme Proc. Natl. Acad. Sci. USA 94: 4262–4266. doi:10.1073/pnas.94.9.4262.
- Stojanovic, M N; Mitchell, T E and Stefanovic, D (2002). Deoxyribozyme-Based Logic Gates Journal of the American Chemical Society 124(14): 3555-3561. doi:10.1021/ja016756v.
- Stojanovic(2003). Deoxyribozyme-based Half-Adder Journal of the American Chemical Society 125(22): 6673-6676. doi:10.1021/ja0296632.
- Stojanovic(2003). A Deoxyribozyme-Based Molecular Automaton Nature Biotechnology 21: 1069-1074. doi:10.1038/nbt862.
Further reading
- Elbaz, J et al. (2010). DNA computing circuits using libraries of DNAzyme subunits Nature Nanotechnology 5: 417-422. doi:10.1038/nnano.2010.88.
- Stojanovic, M N and Stefanovic, D (2005). Computing with Nucleic Acids. In: Bioelectronics: From Theory to Applications I. Willner and E. Katz (editors). Wiley-VCH, Weinheim. Chapter 14.
See also
Implementations of computational state transitions with biomolecules
