Linear actuators made from DNA

In this article we show two different approaches to produce linear actuators made from DNA

In all levels of engineering, complex machinery is based on the concerted activity of many different subunits of heterogeneous nature. One of such subunits are the linear actuators we refer to in the title. Basically, a linear actuator is a device or construct capable of producing a motion confined in one axis; this type of devices, as they are currently found in mechanisms and machines, can be as large as the ones found in the hydraulic arm of an excavator or as precise as the piezoelectric actuators capable of movements in the nanometer range.

To reproduce this useful subunit in the nanoscale and try to apply it to more ambitious manipulations in this scale, we followed two different strategies based on DNA origami to create linear actuators based on the architecture of the rotaxane (a ring threaded through a rod and topologically locked by the presence of stoppers at both ends of the rod). For some schematics see Figure 1.

Figure 1. Construction of slider rail devices from one or two scaffold strands. Top: overview of one-pot and two-pot assembly scheme, with top and side renders of slider and rail, based on cadnano designs. Bottom: TEM images of assembled structures, scale bars 100 nm.

Once we had the properly formed structures, the next steps would be studying its dynamic characteristics, starting by seeing if the threaded ring (slider) is able to freely diffuse the whole length of the rod (rail). We did so in two different ways, with transmission electron microscopy (TEM) and with DNA-PAINT. The former allows us to see with detail structures so small that are under the diffraction limit of light and are not possibly seen directly, the later allows us to obtain a more “dynamic” image of our structures by the clever use of transiently binding fluorophores. (Figure 2)

Figure 2. Assembly of the linear actuators and diffusive motion of slider on rail. A) rendering and B) TEM images of the one-pot and two-pot devices with the slider freely diffusing, scale bars 100 nm. C) TEM images of one-pot device with freely diffusing slider; images are sorted by position of slider and cropped in 100 × 300 nm boxes. D) Distribution of distance from the slider center to the closest end of the rail for one-pot (left) and two-pot (right) devices. Dotted lines indicate distances to the mid-point of the rail. E) DNA–PAINT super-resolution micrographs of the freely moving slider (blue) and the stoppers (red) of the one-pot device, cropped in 500 nm boxes.

From this collection of images, we can first see the properly assembled rotaxanes and quantify the homogeneous distribution of the slider along the rail. Secondly, with DNA-PAINT, we can confirm that the threaded slider is not stuck in place but rather diffusing freely. This we can tell by the fact that the blue labelled slider shows as a blur between the stoppers, demonstrating that it can occupy any position along the rail.

Having achieved the rotaxane structure and having demonstrated the free diffusion of the slider along the rail, it follows that we tried to achieve the controlled positioning of the slider on specific points of the rail. We did so by the introduction of two different styles of attachment, non-interacting and interacting addresses, that consist in the reversible connections between single-stranded linker oligonucleotides protruding from the inside of the slider and others forming an addressable track along the rail. Both using the concept of strand displacement to be able to undo the connections and allowing the slider to freely diffuse to a new attachment point of our choice. By means of TEM we validated the precise positioning of the slider along the different sliders using the different addresses systems (Figure 3).

Figure 3. Controlling the position of the linear actuators. A) Schematic of the two systems for reversibly binding the slider to a specific position on the rail. The number of linker duplexes used to bind the slider at each position is two with non-interacting addresses (one-pot device) and four with interacting addresses (two-pot device). B) The non-interacting and interacting addresses were used to drive the slider of the one-pot device to five pre-defined positions and the two-pot device to three positions. C) TEM images of the actuators locked in position, cropped in 100 × 300 nm boxes. D) Distribution of slider positions (normalized to rail length) for devices locked in different positions; dotted line indicates designed position. E) Distribution of distance from slider to closest end for devices released to freely diffusing state after being locked. Dotted lines indicate distances to the mid-point of the rail.

There are several factors that could affect the precision of the positioning, one of them that we studied was the thermal fluctuations of the system. In order to see this effect we produced simulations using a program called OxDNA. This program is designed to do coarse-grained molecular dynamics simulations of DNA and it allowed us to study the fluctuations that the slider underwent when attached to a specific position of the rail. From these simulations we saw that the fluctuations were of greater magnitude in the structure with the larger slider and, therefore, looser fit around the rail; the flexibility of both the slider structures and the connections to the rail affect greatly the precision of the position (Figure 4).

Figure 4. Study of actuator precision for one- and two-pot devices, with 18- and 31-helix sliders respectively, by oxDNA. The origami structures were prepared locked into their most central binding site and simulated for 108 time steps. The images show the final configurations. Fluctuations in slider position were quantified by calculating the distance from the slider to a point on the rail directly inside the slider (proximal) and to a point at the end of the track near a stopper (distal). Deviations from the mean distances are shown.

To wrap up, we have developed two different methods of producing rotaxanes. We have proven the capabilities of the systems to perform reversible positioning of the sliders in predetermined locations along the rail, with good precision. These results are encouraging for the further development of modular linear actuators to incorporate in nanoscale machinery, the main strength of such systems laying in the great levels of simultaneous actuation that we achieved even in such an early stage of the design. Down the line, this could be used to improve on the traditional fabrication methods by producing programmable assemblies in industry at a reduced cost and increased scale.

Written by Rafael Carrascosa, PhD student in Turberfield grou p at Oxford University

Reference: Benson, E., Carrascosa, R., Bath, J., Turberfield, A. J., Strategies for Constructing and Operating DNA Origami Linear Actuators. Small 2021, 17, 2007704. https://doi.org/10.1002/smll.202007704