Research in chemistry can be roughly divided into two categories: analysis-the measurement of existing objects and phenomena-and synthesis-the construction of those objects and phenomena from simpler pieces. Typically, synthesis lags behind analysis: one first determines the formula of a molecule (and often its structure) before attempting to make it. However, as chemistry advances, investigators are increasingly attempting to synthesize first, designing chemical systems that realize desired

more »

... nomena. A team led by André Estevez-Torres at CNRS in France and co-workers [1] has demonstrated an experimental toolkit that could be used for the rational engineering of nonlinear chemical effects in solution. Using DNA strands moving in a narrow channel and reacting under the action of enzymes, the authors are able to create chemical waves whose shapes and velocity can be finely controlled. Their setup could be programmed to yield a broad spectrum of other nonlinear phenomena in systems governed by a combination of chemical reactivity and molecular diffusion ("reaction-diffusion" systems). Nonlinear chemical dynamics [2] characterizes many natural and industrial processes and is a quintessential feature of living organisms: most of their chemistry occurs far from equilibrium, has a nonlinear dependence on parameters like molecular concentrations, and may exhibit temporal oscillations (many biological functions are, for instance, synchronized to a 24-hour cycle). Researchers have applied a number of algorithms to design systems featuring such nonlinear behavior, engineering chemical oscillators [3] (reactions in which the concentration of one or more components exhibits periodic changes), propagating reaction fronts [4] or "Turing patterns" [5]-spatial patterns of concentrations that, as Alan Turing proposed, might be related to biological morphogenesis (e.g., the formation of leopard spots or zebra stripes). But design algorithms are severely limited by the realities imposed by nature. One may write down a set of equations that generates a desired phenomenon, e.g., spatiotemporal chaos, for a set of reaction rates and diffusion coefficients. There is, however, no guarantee that an actual collection of molecules can be found that realizes the theorized behavior. For example, over the past several decades, researchers have successfully engineered, through systematic studies, new chemical oscillators with desired parameters for a variety of applications. These efforts have resulted in the discovery of several oscillators, but none of these oscillators "by design" has attained the importance of the Belousov-Zhabotinsky (BZ) reactions [6], a family of oscillating reactions (discovered serendipitously) that remains the most versatile and reliable chemical oscillator for most applications. Most efforts to design reaction-diffusion phenomena have utilized small inorganic molecules, largely because these substances are cheap, easy to work with, and produce visible color changes when they undergo reduction and oxidation (redox) reactions. Unfortunately, these reaction mixtures are typically not biocompatible, and they cannot be used in applications that place them in contact with components of living systems like proteins. The BZ reaction, for example, only works at acidity levels lethal to most biological cells. A solution may be offered by oligonucleotides (short single strands of DNA or RNA), which have been utilized as versatile building blocks for oscillators [7] , computational elements [8] , and structures with arbitrary shapes [9] . In their new work, Estevez-Torres et al. [1] have focused on dynamical aspects: They demonstrated that DNA strands can be used to realize an experimental model for reaction-diffusion systems whose spatiotemporal dynamics is fully controllable by programming three key elements of the system: the reaction rates, the diffusion coefficients, and the topology of the chemical reaction network (i.e., which reactions are linked to each other in ways that generate positive or negative feedback).