Researchers are developing custom DNA molecules to mimic the logical operations carried out in silicon-based computers as a way to improve biomedical technologies or aid in assembling nano-size building blocks into new materials. In one approach, a series of wells is filled with hairpin-shaped strands of DNA, called gates, which respond to inputs in ways chosen by the designers. A group had already shown that such a system, called a molecular array of YES and AND gates, or MAYA, can play a game of tic-tac-toe restricted to certain moves.
In the November Nano Letters the team will report that their second version, MAYA-II, allows a player to make a move in any well after the computer's first move, which is in the center square. The game consists of a three-by-three array of test tubes, each filled with its own set of DNA gates. Each gate is constructed so that when it is exposed to a specific combination of shorter input strands of DNA, its hairpin structure unravels and activates a fluorescent substrate. To make a move, a player adds to all the wells an input strand that represents a move in the chosen square during that round of play. For instance, there is a DNA strand that corresponds specifically to "middle left square, first round." Although a player adds this strand to all the wells, only one well (the one the player wants to mark) has the gate that is activated by that input strand. That well then fluoresces at one of two frequencies of light. The computer's responses, which cause wells to fluoresce at the other frequency, are encoded in the gates in the other wells. For example, the human player's middle left input would also be designed to trip a gate in the upper right that fluoresces at the computer's frequency.
MAYA-II required 128 different DNA gates spread among the wells and 32 input molecules, the group reports. "We're excited about the potential given we've got so many gates working together," says first author Joanne Macdonald, a virologist at Columbia University. The goal, Macdonald says, is to simplify detection schemes for viruses or cancer, so that a sample only elicits a signal if it contains the right combinations of DNA sequences. Others agree about the potential of the technique. "This one has much greater complexity and they've demonstrated that even at that level it works," says Nadrian Seeman of New York University. "It opens the door to more complex computations.