These genes were made for walkin'...
The above (embellished) depiction of our walker is in flattened form.
Motion is a necessary component of life. While this does include the motion of animals and people as they move about, there is also crucial motion which takes place on a cellular level. Cells could not function without the proteins kinesin and myosin , which carry cargo to and from our organelles. The movements employed by these proteins are very similar to the way people walk in the fact that they have two “feet” which step forward, one at a time . Because of the crucial role these proteins play in supporting life, we began to wonder if we could create our own molecular walker that could perform similar tasks.
In order for this walker to be comparable to (or even better than) kinesin, it must be autonomous and run on a reusable track. In other words, it should not require an external driver and should not damage the track when taking steps. Other walkers have been created in the past, but none has yet been able to accomplish both of these tasks. There are some that are at least capable of either walking autonomously  or of reusing their track . Our walker is able to accomplish both tasks by means of two alterations. The use of two discrete fuel strands which can only interact with the walker system in a specific order allow us to add an excess of fuel while still keeping the walker on the track. We also make use of a nicking enzyme which restores the track to its original form without harming it, allowing it to be reused.
However, our design does sacrifice some functionality by using this system of fuels. Because all molecules are free-floating in solution and have no preference to location, we expect there to be leakage of the walkers to some extent. This is due to the method by which the F1:F2 complex, see below for details, can displace the lagging leg while the leading leg is still in transition to its new foothold. We expect the probability of this occurring to be very low due to the functional breakdown of F1:F2 by the nicking enzyme when the fuels successfully bind to the track.
Design and walking mechanism
Our walker design consists of four main components: a walker, specifically a DNA molecule with two "legs" and a body capable of storing cargo; a track, a long platform supporting footholds to which the legs of the walker can attach; fuel molecules that supply energy to the walker system; and a nicking enzyme which reverts the footholds along the track to a usable state after the walker has walked across them. The sections below illustrate how we designed each of these components of the walker system.
The components of the walker can be defined terms of short common DNA sequences, called domains, numbered 1-6 and a-f. An asterisk (*) after a domain name indicates the reverse complement sequence.
The "legs" of the walker are the two toeholds (5'- 2*1* -3' and 5'- 6*4* -3') of W1 and W2. By careful construction of the track, the walker will bind to corresponding toeholds on the first two hairpins (aA, bB) of the track. This is the starting point of the system. When the walker binds to the track, the hairpins will open, exposing a domain that was inaccessible. The fuel F1 binds to this toehold and displaces the lagging leg. We designed the sequence lengths such that when a fuel displaces a leg from the track, there is a smaller chance of the leg retaking its position on the track. This means that the forward intermediate is a more favorable reaction than the reverse. This is what gives the "push" our walker needs to move in one particular direction. This leg will then bind to the next hairpin in line, cA, as it did with the first, now becoming the leading leg by hopping over the other.
The duplex formed by aA and F1 expose a binding domain specific to the second hairpin F2. F2 then displaces F1 from aA, forming a duplex F1:F2. This duplex is free-floating in solution. Because of this, it can bind to bB and displace the now lagging leg, allowing it to bind to dB. Now the walker has taken a full two steps.
Finally a nicking enzyme, Nb.BsmI, specific to a domain formed only when F1:F2 is bound to a hairpin with a B loop will cut the formation. This cut causes the total number of bonds in the duplex to decrease to the point at which they can freely dissociate from the track , becoming waste. This will allow the hairpin bB to return to its initial state, thus regenerating the track.
This process can continue as the track is lengthened and as long as there is fuel available for consumption. However our design incorporates elements that limit the length of the track. Almost all things observed in nature tend to follow exponential curves. This is no different in the reaction times for DNA strand displacement. We know from observation that not all walkers will complete their first steps at the same time, and some may not even complete one. This is due to the subtle differences in free energies of the displacements that our design employs. Because the drive is not very strong, we expect the walkers to decrease in number as they continue moving forward due to energy dips and mismatched base pairs.
Fueling the walker
Directional walking cannot be achieved without a source of energy, such as a fuel molecule . The fuel strand system behind our walker utilizes catalyzed hairpin assembly  to move the walker down the track. Simplified, the mechanism of the system can be illustrated with three DNA strands. Two of these are fuel strands. These strands are hairpins that are kinetically trapped and cannot form a duplex with each other. The third strand is the initiator strand. When this strand is present, the two fuels can interact with each other and form a new duplex.
Why and how does this reaction take place? The initiator strand has bases that are complementary to one of the fuel strands. This gives it a toehold that allows it to attach to the fuel strand. After this attachment, by branch migration, the initiator opens up the fuel hairpin. This opening in turn exposes bases that are complementary to the other fuel strand. Toehold binding occurs between the two fuel strands, and by branch migration, they kick off the initiator strand. The final products of this reaction are a stable duplex between the two fuel strands and the initiator strand. This regeneration of the fuel strand gives it a catalytic property.
The CHA reaction has been studied and its kinetics are understood . Using kinetic models we can simulate this reaction.
2-D representation of DNA forms
It is sometimes easier to view DNA molecules in a "flattened" representation. In this representation, a duplex is shown as a pair of parallel lines, with filled circles in between to denote base pairing. Additionally, in flattened form, the 3' end of every strand is embellished with a half-arrow. An example of this notation, applied to the fuel molecule F1, is shown below.
We are sponsored by Integrated DNA Technologies (Coralville, IA, USA). We received travel support from the College of Natural Sciences, the University of Texas at Austin.
Background photo taken by Linhao Zhang in Fort Davis, Texas. Used with permission.