A major challenge in nanotechnology is to precisely transport a nanoscale object from one location on a nanostructure to another location along a designated path. The successful construction of self-assembled DNA nanostructures provides a solid structural foundation to meet this challenge. DNA, with its immense information-encoding capacity and welldefined Waston-Crick complementarity, has been explored as an excellent building material for nanoconstruction. [1,2] In particular, recent years have seen remarkable success in the construction of both self-assembled nanostructures and individual nanomechanical devices. For example, one-and two-dimensional DNA lattices have been constructed from a rich set of branched DNA molecules. [3][4][5][6][7] These DNA lattices could provide a platform for embedded DNA nanomechanical devices to perform the desired transportation. A diverse group of DNA nanomechanical devices have also been demonstrated. These include DNA nanodevices that execute cycles of motions, such as opening/closing, [8][9][10][11] extension/ contraction, [12][13][14] and reversible rotation. [15,16] Such DNAbased nanodevices can be cycled between well-defined states by means of external intervention, for example, by the sequential addition of DNA "fuel strands" [8-10, 12-14, 16] or by changing the ionic composition of the solution. [11,15] However, these devices are unsuitable for the above challenge for two reasons: First, they demonstrate only local conformation changes, rather than progressive motion. Second, they do not move autonomously. Various schemes of autonomous DNA walker devices based on DNA cleavage and ligation have been explored theoretically but not experimentally; [17] these schemes have been limited to random bidirectional movement. The use of DNA hybridization as an energy source for autonomous molecular motors has also been proposed.