DNA nanomachines and their functional evolution

DNA nanomachines and their functional evolution

(Parte 1 de 4)

DNA nanomachines and their functional evolution

Huajie Liuw and Dongsheng Liu*

Received (in Cambridge. UK) 17th December 2008, Accepted 10th March 2009 First published as an Advance Article on the web 9th April 2009 DOI: 10.1039/b822719e

Since the establishment of the Watson–Crick model more than five decades ago, the understandings of DNA structures are well sufficient to enable applications of DNA in designing and assembling two-dimensional (2D) and three-dimensional (3D) structures at the nanoscale. Furthermore, the conformational switchability of DNA also enables the fabrication of nanoscale molecular machines, which can perform movements upon stimuli. In this article, we will summarize the present efforts on constructions of DNA nanomachines based on different driven mechanisms, and further discuss their evolutional processes, in order to find applications and future development directions.


DNA has been seen to play an extraordinary important role in life science more than five decades since the establishment of the Watson–Crick model. Tracing back to its chemical essence, DNA has also received attention in material sciences, especially in nanoscience.1 Based on specific base-pair formation and programmable sequence, DNA nanostructure assembly, pioneered by Seeman et al.2–6 in the 1980s, has now reached the stage of facile fabrication of complicated 2D7–12 and even 3D nanostructures via designed hybridization processes.13–16 However, compared to these static nanostructures, a more challenging aspect in this field is fabricating nanomachines which can perform nanoscale movements in response to external stimuli.17–20 Although protein is the material chosen by Nature to facilitate nanomachines in living beings,21,2 the clearer structures, established synthesis and modification methods and clearer driven mechanisms of DNA nanomachines, have been demonstrated to be of interest for material research as well as theoretical studies. From the first effort to control DNA motion,23 the last decade has witnessed an explosion of interest and effort in this field.24–29

In this review, we have sorted out several basic DNA nanomachine types by driven mechanism. The consequent evolution on the power input method to improve the kinetics of each type is summarized to give a perspective on their development trends. We then highlight the efforts on measuring the mechanical outputs of DNA nanomachines as well as employing these outputs to achieve new functional devices and materials. Through these analyses, we will try to provide some perspectives on the development of DNA nanomachines in the near future.

National Centre for NanoScience & Technology, No. 1 Beiyitiao, Zhongguancun, Beijing, 100190, China. E-mail: liuds@nanoctr.cn; Fax: +86-10-62656765; Tel: +86-10-82545589

Huajie Liu

Dr Huajie Liu received his BS in applied chemistry and MS in inorganic chemistry at Tongji University. From 2005 to 2008, he was a PhD student under the supervision of Prof. Dongsheng Liu at the National Center for Nanoscience and Technology. During his PhD period, he worked on developing new control-modes and applications of DNA nanomotors. He is currently working as a postdoctoral fellow with Prof. Kurt V. Gothelf at the University of

Aarhus, where his research concentrates on DNA directed assembly of nanomaterials.

Dongsheng Liu

Professor Dongsheng Liu graduated from the University of Science and Technology of China with a BS degree in 1993. After working in the Institute of Chemistry, Chinese Academy of Sciences for six years, he went to the Hong Kong Polytechnic University and finished his PhD there under the supervision of Professor A. S. C. Chan in 2002. He moved to the UK afterwards and worked as a postdoc research associate in NanoIRC and Department of

Chemistry, Cambridge University. In 2005, he took the position as a principal investigator in the National Centre for Nano- Science and Technology, China. His researches are mainly focused on using biomolecules to fabricate nanostructures and nanodevices.

w Current address: Centre for DNA Nanotechnology at Department of Chemistry and iNANO, University of Aarhus, Denmark.

This journal is c The Royal Society of Chemistry 2009 Chem. Commun., 2009, 2625–2636 | 2625

FEATURE ARTICLE w.rsc.org/chemcomm | ChemComm

Prototype DNA nanomachines

A machine could be defined as any device that uses energy to perform some activity, e.g. mechanical movement. We also use word ‘‘motor’’ to describe a device that converts various forms of energy into kinetic energy for mechanical work. As outlined in an editorial18 by Stoddart: ‘‘organization, the power source, and work of a repetitive nature’’ are the most important characteristics of a molecular machine. From this point of view, the DNA nanomachine could be regarded as a kind of molecular machine that is made up of assembled DNA structures integrated with an external stimuli responding mechanism. Thus all established DNA nanomachines could be sorted into several catalogues by their original driven mechanisms. In the following, we will summarize the variation of each prototype to map their development trends, respectively.

DNA nanomachines controlled by ‘‘fuel-strands’’

It is well known that a short DNA strand can be replaced by a longer strand to form a more stable duplex, which is called ‘‘chain-exchange reaction’’ or ‘‘strand-exchange reaction’’. This reaction has been employed to induce motions to DNA-based nanostructures. In 2000, Yurke et al. reported the first hybridization energy-driven DNA nanomachine which resembles a pair of tweezers (Fig. 1).30 Their device is assembled by three single-strands which can form two rigid duplex arms connected by a hinge section and two dangling ends linked to arms. At the ‘‘open’’ state, two ends of the arms are thermodynamically separated. To ‘‘close’’ the tweezers, an additional fourth strand F hybridizes with both dangling ends and pulls the two arms together. The device could be reopened by F’s fully complementary strand F0 because duplex FF0 has a lower free energy (note an overhang section on F is the starting point for FF0 hybridization). The alternating addition of F and F0 will cycle the machine and produce duplex FF0 as waste.

Since the machine is powered by competitive hybridization, the authors called the stimuli, DNA F and F0, as ‘‘fuel’’ and ‘‘anti-fuel’’. Overall, one switching is able to generate a force of about 15 pN, with a maximum separation of the arm ends of about 6 nm or 501; FRET and recently sp-FRET31 techniques have been used to monitor the motions. The modifications of this design in following years have led to some variants: an actuator32 and a three-state nanomachine33 that looks like a combination of tweezers and actuator.

This fuel-strands strategy has also been employed to drive different DNA assemblies to move: Yan and Seeman proposed a robust DNA nanomachine34 whose motions are between two topological motifs: four-stranded PX and JX2 complexes. The addition of DNA fuels will induce a four-step rotation.

A noticeable experimental detail is the use of biotinylated fuel-strands to remove duplex wastes. By covalently linking

PX–JX2 machines linearly, the rotations of DNA machines in these arrays could be controlled synchronously and be visualized by atomic force microscopy (AFM). RNA has also been exploited to control this device.35 Extension of this two-state machine to a three-state one has recently been done by the same group.36

In addition to duplexes, DNA can form unusual hybridized structures such as triplexes and quadruplexes. It has been proven by the groups of Tan37 and Mergny38 that the quadruplex–duplex transition could generate mechanical force. Their designs of G-quadruplex-based DNA nanomachines are simpler than that of duplex-based models. In their cases, only one G-rich strand is used to construct the main body of the machine, that is, an intramolecular G-quadruplex. The G-quadruplex state could be switched to the duplex form by adding DNA fuel. Through this transition, the distance between two ends of G-rich strand could be controlled. Hence, we may regard the G-quadruplex and duplex forms as closed and open states, respectively.

DNA nanomachines controlled by non-DNA stimuli

In principle, the above fuel-strands strategy could be applied to all strand-exchange reaction-powered DNA nanomachines, since, as we have mentioned, hybridization is the common feature of DNA. However, the main disadvantage is these reactions will result in cumulated duplex wastes. These useless duplexes may compete with surrounding nanomachines. And from the point of entropy flow, the accumulation of waste DNA will increase the entropy of the system and will eventually destroy the machine.

To avoid duplex wastes, non-DNA stimuli should also be choices for controlling motions. In fact, this approach has already been proposed in the construction of the first DNA-based nanomechanical device,23 in which case ethidium ions are used as intercalators to induce branch point migration in a tetramobile branched junction structure.

Simpler ions than ethidium have also been explored. Mao and Seeman have demonstrated a DNA machine based on a

B–Z transition (Fig. 2).39 In the absence of Co(NH3)63+ ion, sequence (CG)10 forms normal right-handed B-DNA. This B-DNA can be transformed to left-handed Z-DNA upon the addition of a high concentration of Co(NH3)63+ ion. The

Fig. 1 DNA tweezers controlled by ‘‘fuel’’ and ‘‘anti-fuel’’ strands. Fuel strand F hybridizes with the dangling ends of the open state machine (shown in blue and green) to pull the tweezers closed. Hybridization with the overhang section of F (red) allows anti-fuel strand F0 to remove F from the tweezers, forming a double-stranded waste product FF0 and allowing the tweezers to open (reprinted with permission from ref. 30; copyright 2000, Nature Publishing Group).

2626 | Chem. Commun., 2009, 2625–2636 This journal is c The Royal Society of Chemistry 2009

whole transition will generate a rotary motion with about 2 nm displacement in space.

Based on the G-quadruplex structure, Fahlman et al. have constructed a nanopinching device which is sensitive to Sr2+ ion.40 This device’s working mechanism is due to sequenced guanisines’ ability of forming G-quadruplexes in the presence of metal ions. The addition of chelant EDTA to the solution will enable the removal of Sr2+ ion from G-quadruplex and break the pinched structure. Another ion-sensitive G-quadruplex nanomachine was developed by Sugimoto’s group.41 Their machine is assembled from modified DNA strands, in which a coordination unit for divalent metal ions, 2,20-bipyridine, is used to link two G fragments. In the absence of M2+ ions, the modified sequence will self-assemble into an antiparallel G-quadruplex. Adding Ni2+ ion to the system will induce the rotation of the 2,20-bipyridine unit and switch the structure to a parallel G-wire. Reverse reaction could be started by EDTA. Strictly speaking, this machine could not be regarded as an absolute DNA machine since its driving force is derived from the ion-sensitivity of the chemical unit.

For almost every chemical or biological system, the pH value is a very important factor. For some DNA structures, this is also true. For example, acidic pH favours the formation of the four-stranded C-rich i-motif structure and C+GC triplet because C base protonation (C+) will lead to extra hydrogen bonds. This concept has first been applied to DNA nanomachine construction by Liu and Balasubramanian.42 The system comprises a 21mer strand X containing four C stretches and its partial complementary strand Y (Fig. 3). At slightly acidic pH, half protonated C stretches will form intercalated C+C base pairs and induce X to fold into a compact i-motif structure, corresponding to the machine’s closed state. Raising pH to slightly basic value will unfold X and form an extended duplex structure XY (open state). Reversible switches between compact and extended states can be produced by changing the pH and both processes are completed in less than 5 s. In total, the operations of this machine will result in a 5 nm linear movement and opening and closing forces both larger than 10 pN; the wastes are only salt and water from the neutralization reaction. The advantages of this proton-driven nanomachine are obvious: it is clear, quick, reliable and efficient. Similarly, the pH-sensitive DNA triplex–duplex transitions have also been utilized to build DNA nanomachines by the groups of Mao43 and Zuccheri.4

Another approach to stimulate DNA structural transformation is changing the temperature. Although all nucleic acid structures are sensitive to temperature change, Sugiyama’s work demonstrated that DNA and RNA may have inverse responses to thermal stimuli and therefore produce opposite rotations.45 On the other hand, Isambert et al. concentrated on controlling the folding kinetics of a DNA nanoswitch by modulating its annealing cooling rate.46 By fast cooling, a metastable conformation is favoured kinetically.

A protein-driven DNA nanomechanical device has been proposed by Seeman’s group.47 In the report, E. coli integration host factor (IHF) has the ability to distort the device by recognizing and binding specific DNA sequences in the device. This device was suggested to be effective in measuring the interaction between protein and DNA.

Unidirectional motions

The machines already described above have fixed links between their primary elements. The motions of those DNA machines are all fundamentally intramolecular. A challenge arises as to mimic biological molecular motors21,2 such as myosins, kinesins and dyneins that can perform unidirectional motions on external tracks. One of the most attractive features of this kind of motion is its ability to transport substances at the nano- or micro-scale.

The simplest kind of unidirectional motion is linear walking.

This has now been achieved by DNA-based devices in a few instances. In 2004, Seeman’s group reported a biped DNA

Fig. 2 An ion-triggered DNA rotary machine based on a B–Z conformational transition. The motor consists of two DNA double crossover (DX) molecules and at the centre of the connecting helix is a 20-nucleotide region of proto-Z DNA shown in yellow. Fluorescent dyes fluorescein (green) and Cy3 (magenta) are attached to the free hairpins near the middle of the molecule. When the transition occurs, the two DX molecules change their relative positions and the separation of the dyes (reprinted with permission from ref. 39; copyright 1999, Nature Publishing Group).

Fig. 3 An i-motif DNA-based nanomotor driven by pH change. Strand X contains four C stretches and can form a four-stranded i-motif structure in slightly acidic environment. With the increase of pH, the i-motif structure will be destroyed and X0s partial complementary strand Y will hybridize with X to form a duplex (reprinted with permission from ref. 42; copyright 2003, Wiley Interscience).

This journal is c The Royal Society of Chemistry 2009 Chem. Commun., 2009, 2625–2636 | 2627 walker machine48 whose working mechanism is shown in Fig. 4. Their device consists of two parts: a three-step track and a biped walker. The track itself is in fact a triple crossover (TX) molecule49 with three single-stranded overhangs which serve as ‘‘footholds’’. The biped region has two duplex ‘‘legs’’ and on the end of each leg there is a singlestranded ‘‘foot’’. Three additional single-stranded linkers between the two legs make the biped walker flexible. A foot attaches to a foothold when a ‘‘set’’ strand complementary to both is added. The biped region will walk on the track in a certain direction if specific set and unset strands are added sequentially.

Shin and Pierce have proposed another simple approach for a DNA walker.50 Their device has also two main components: a walker and a track which are both in duplex forms. The walker is a partially complementary duplex with two singlestranded overhangs as legs. The track consists of six oligonucleotides and has four protruding single-stranded branches. With the similar walking principle to Seeman’s device, the walker can move on the track by set and unset strands.

The above two DNA walkers, although their structures are different, have similar working mechanisms. One of their common features is that the walking steps in a certain direction is limited to the number of footholds or branches on the track. We could imagine if the track is a circle rather than a line, the walker may always move unidirectionally. This kind of motion has been realized in a DNA nanogear model by Mao’s group.51 The gear system has two DNA duplex circles. Each circle contains four oligonucleotides and three singlestranded overhangs as ‘‘teeth’’. These two circles may roll against each other driven by the ‘‘fuel-strands’’ mechanism which was also applied in the above two DNA walkers.

Evolution of kinetics and driven modes

One of the critical issues involved in all kinds of nanomachines is how to control their motions precisely. In natural systems, the sophisticated motions of protein-based nanomotors are the results of complicated biological cooperations, and natural evolution makes these machines work at their best. For most DNA-based nanomachines, motions are generated from DNA hybridizations. So the question now should be how to evolve the kinetics and efficiency of the hybridizations in the DNA nanomachines. Basically, many factors such as the machine’s structure, environment, fuel and control mode will have certain impacts.

Evolution of hybridization kinetics of ‘‘fuel-strands’’

(Parte 1 de 4)