Traditional robots 1 rely on computing to coordinate sensing and actuating components and to store internal representations of their goals and environment. Any implementation of single-molecule based robotics must overcome the limited ability of individual molecules to store complex programs and, for example, use architectures that obtain complex behaviors from the interaction of simple robots with their environment [2][3][4] . Previous research in DNA walkers 5 focused on transitioning from non-autonomous systems 6, 7 to directed but brief motion on one-dimensional tracks8 -11. Herein, we obtain elementary robotic behaviors from the interaction between a random walker incorporating deoxyribozymes 12 and a precisely defined environment. Using singlemolecule microscopies we demonstrate that such walkers achieve directionality by sensing and modifying their environment, following trails of recognition elements ("bread crumbs") laid out on a two-dimensional DNA origami landscape 13 . These molecular robots autonomously carry out sequences of actions such as "start", "follow", "turn", and "stop", thus laying the foundation for the synthesis of more complex robotic behaviors at the molecular level by incorporating additional layers of control mechanisms. For example, interactions between multiple molecular robots could lead to collective behavior14 , 15, while the ability to read and transform secondary cues on the landscape could provide a mechanism for Turing-universal algorithmic behavior2 ,16,17 .Author Contributions: AFM experiments were performed by K.L. (majority), J.N., and N. D.; analysis was performed by N. D., K.L., J.N., S.T., and supervised by E.W., and H.Y. Fluorescence microscopy and particle tracking analysis were performed by A.J.M., N.M., and A.J.B, supervised by N. G. W. Spiders were synthesized, purified, and their integrity confirmed and monitored by S.T. SPR experiments were performed by R. P. Research coordination by M.N.S., material transfer coordination by S.T., J.N., and K.L. Experimental design and manuscript was done with input from all authors. Author Information: Reprints and permissions information is available at npg.nature.com/reprintsandpermissions. The authors declare no competing financial interests. Correspondence and requests for materials should be addressed to: mns18@columbia.edu, winfree@caltech.edu, nwalter@umich.edu, hao.yan@asu.edu Supplementary Information is linked to the online version of the paper at www.nature.com/nature. NIH Public Access Author ManuscriptNature. Author manuscript; available in PMC 2010 November 1. Restated in a biochemically more intuitive manner: A deoxyribozyme on a site that was previously converted to a product will dissociate faster, whereas it will stick longer on the substrates and eventually cleave them. Because spiders have multiple legs, a single dissociated leg will quickly reattach to nearby product or substrate. It follows that the body of a spider positioned at the interface between products and substrates will move toward the s...

We describe polycatalytic assemblies, comprising one or two streptavidin molecules and two to six attached nucleic acid catalysts (deoxyribozymes), with phosphodiesterase activity. When exposed to a matrix covered at high densities with oligonucleotide substrates, these molecules diffuse through the matrix continuously cleaving the substrate at rates comparable to those of individual catalysts in solution. Rates of diffusion (movement), processivity, and resident times of assemblies can be controlled through the number of catalytic units and the length of substrate/product recognition regions. The assemblies were characterized at the ensemble level using surface plasmon resonance.

We report a straightforward evolutionary procedure to build an optimal sensor array from a pool of DNA sequences oriented toward three-way junctions. The individual sensors were mined from this pool under separate selection pressures to interact with four steroids, while allowing cross-reactivity, in a manner designed to achieve perfect classification of individual steroids. The resulting sensor array had three sensors and displayed discriminatory capacity between steroid classes over full ranges of concentrations. We propose that similar protocols can be used whenever we have two or more classes of samples, with individual classes being defined through gross differences in ratios of dominant families of responsive components.