The motive forces in DNA-enabled nanomachinery

Summary

Building machines that can augment or replace human efforts to accomplish complex tasks is one of central topics for humanity. Especially, nanomachines made of discrete numbers of molecular components can perform intended mechanical movements in a predetermined manner. Utilizing free energies of Watson-Crick base pairing, different types of DNA nanomachines have been invented to operate intended stepwise or autonomous actions with external stimuli, and we here summarized the motive forces that drive DNA-based nanomachineries. DNA tweezers, DNA origami actuators, DNA walkers, and DNA machine-enabled bulk sensing are discussed including structural motif design, toehold creations for strands displacement reactions, and other input forces, as well as examples of biological motor-driven hybrid nanomachines. By addressing these prototypical artificial nanodevices, we envision future focuses should include developing various input energies, host cell-assisted structure self-replication, and nonequilibrium transportations.

Graphical abstract

Supramolecular chemistry; Biochemistry

Introduction

During a bike transportation from home to office, it follows the bike lane and safely carries the passenger to the directed place. As an example, cycling can be used to interpret the term machine which converts energy and does work.¹ The feet on pedals transmit the power of downstroke to forces that rotate wheels forward through roller chain and series of gears (Figure 1A). Such mechanical power transmission illustrates the basic endeavors that people have challenged to create machines and make jobs less difficult. Apart from macroscopic man-made machineries, life evolution has also developed tremendous amounts of molecular motors that conduct multi-level sophisticated tasks to drive vital activities. For instance, the small rotary motor ATP synthase catalyzes the formation of ATP—the energy currency—from ADP and orthophosphate. The transmembrane proton flow propels the c-ring subunit to spin and promote the ATP synthesis with the binding-change mechanism (Figure 1B).

Figure 1.

Bicycle and ATPase as macroscopic man-made machine and the smallest biomolecular motor, respectively

(A) Bicycle converts the inputted man power to the rotation of wheels. The forces (red arrow) that pushed down on the pedals rotate the crankset (yellow arrow), generating a torque which is transmitted to the wheel through the bicycle chain (green arrow) and rear cassette (blue arrow). The forces of the wheels against the ground cause the backward forces (F_rear-F_fwd) to wheels, propelling the bicycle forward.

(B) The electron-transport chain generates a proton gradient across the inner mitochondrial membrane. Protons then flow back into the matrix to equalize the distribution, creating a proton-motive force which drives the spinning movement of c-ring and the rotational catalytical phosphorylation of ADP to ATP.

To imitate the macroscopic and biological counterparts and expand the machinery design principle to nanoscale, it inspires to create synthetic nanomachines. Energies from small motions or coupled chemical reactions can be harnessed to supply powers for nanoscale actuations that can perform enormous roles in biomedical, mechanical, sensing, and monitoring applications. Resembling the real machines at the macroscopic level, synthetic nanomachines receive input energy and transform it into the work output, producing quasi-mechanical movement in artificial or biological microenvironment. Surrounded by various species, nanomachines encounter multi-level interactions and thermodynamic fluctuations and thus demand an accurate design with high energy-conversion efficiency. Moreover, the intuited creation of miniature-sized nanomachines is usually aimed to mimic the biological motors. Learning from coupled cascade reactions in biological systems, synthetic nanomachines arrange chemical stimulus one by one in orders so that it can operate according to the inputted information. Though far less complicated, nowadays molecular switches and shuttles from cis-trans isomerism to “rotaxanes” and “catenanes,” as well as cross-linked responsive chemical reactions chains at nanoscale, have been built up.

Here in this review, we focus on DNA-enabled synthetic nanomachines which either employ DNA structures as machine body or are driven by DNA inputted energies. Firstly, DNA nanotechnology has demonstrated the capacity of using DNA, omitting its biological functions but as programmable polymers, for numerous nanostructures fabrication. Through sequence design to achieve discrete hybridization events between certain segments of different DNA strands, many assembly strategies have been invented to implement the structural DNA nanotechnology. Owing to the highly specified DNA sequences, each DNA strand that weaved into nanostructures can be precisely identified at addressable locations. Hence the design and assembly of DNA nanostructures can be fine-tuned in a controllable manner to realize structures with defined shape, sizes, and curvatures. Though the reported sequence-averaged persistence length of single DNA duplex is merely ∼50 nm, multiple DNA duplex can be bundled up through crossover connections (e.g., in origami) to form larger rigid entities not subjected to this limitation. Moreover, the mature nucleic acid chemistry can introduce various functional groups on DNA strands such as amino and thiol, dyes, and click chemistry tags to enable DNA-templated assembly for many nanomaterials.

Secondly, in the classic Watson-Crick duplex model, the highly structured double-stranded DNA (dsDNA) is cooperatively stabilized by hydrogen bonding between A=T or G≡C and π-stacking of neighboring base pairs. DNA duplex is only a marginally stable system which depends on many factors such as temperatures, salt and buffer concentrations, GC content (the proportion of Guanine and Cytosine in a sequence), and sequence length. Even at the physiological environment, it is consistently undergoing thermally driven structural fluctuations (also known as “breathing”), the temporarily unwinding of a limited region of duplex to form a small single-stranded DNA (ssDNA) bubble. Such thermodynamic window allows for not only many fundamental biological reactions but also the sequence-specific hybridization and low-energy-favored strands displacement reactions (SDRs). In addition, DNA strands that are encoded with desired sequences possess several secondary structures, from duplex of A-form, B-form, and Z-form, to triplex and quadruplex which offer various means to realize regulatory dynamics. As organized in the current paper, we will first summarize state-of-the-art motive forces involved in DNA nanomachines. We termed intermolecular interactions as the engagement of different objects such as hybridization between two oligonucleotides and stacking interactions among structures, while intramolecular interactions are via formation of intramolecular structures such as hairpins. Non-DNA interactions as well as macroscopic manipulation techniques such as light and electric and magnetic field are later included. Then two unique research fields DNA walker and DNA machine-enabled sensing are discussed.

Intermolecular DNA interactions

The primary motive force for DNA nanomachines is the standard free energy ΔG° of hybridization between complementary DNA strands. First of all, DNA strands hybridization is low-energy-favored reaction and the free energy ΔG° associated with per base pair hybridization can be calculated with a nearest-neighbor (NN) model (around −4.0 kJ/mol at 20°C^2,3,4). The real hybridization events, however, will be influenced by various factors such as length, GC content, salt concentration, buffer composition, competitive inhibitors, as well as temperature and DNA oligonucleotides qualities. Based on hybridization-enabled FRET (fluorescence resonance energy transfer) measurement of two dye-labeled single strands, the calculated k_{hybridization} for both strands are in six orders of magnitude,⁵ while the k_dissociation for 10-mer and 20-mer DNA (at 25°C in a buffer of 1 M NaCl/10 mM NaH₂PO₄, pH 8) are 2.1x10⁻³ and 4.9x 10⁻¹² s⁻¹, respectively.⁶ This indicates the differences of half-life for dissociation of the 10-mer and 20-mer duplex are minutes to millenniums. For DNA nanostructures that assembled with a few tens of bases long oligomers, it would be sufficiently stable in commonly used buffers containing magnesium or sodium ions at room temperature. Moreover, there are numerous online tools (such as NUPACK,⁷ OligoCalc,⁸ and OligoAnalyzer at IDTDNA; www.idtdna.com) to facilitate the sequence design and minimize undesired secondary structures.

SDRs

Hence, aside from static DNA nanostructures, by labeling responsive molecules, extending selected DNA strands with functional sequences or for SDR,⁹ it introduces the mostly used driving forces on DNA substructures. Taking SDR as example, while DNA duplex can spontaneously dissociate, invader strand-enabled three-way branch migration can accelerate this process as the competitive binding of invader and incumbent strands with the substrate strand will result in different free energies (Figure 2A). The displacement rates will be much dependent on the toehold length; the longer the toehold, the higher the attempt frequencies for three-way branch migration. Experimentally measured rate constant for displacement reactions, however, will reach saturation when toehold length is about 5–6 bases. For more details about underlying mechanisms of SDRs, we refer to the comprehensive review paper on Chemical Reviews.¹⁰

Figure 2.

The principal strands displacement reactions (SDRs) and SDR-enabled nanomachines

(A) Schematic illustration of free-energy changes in a typical strands displacement reaction. The invader strand (red) releases the protector strand and hybridizes to the incumbent strand in an exergonic manner.

(B) 2-fold strands displacement reaction design examined the mismatch positions are critical to the displacement rates.¹⁴

(C) Instead of mismatches, a random sequence pool was introduced to serve as the first round for strands displacement kinetics control.¹⁶

(D) The DNA tweezer is the pioneer DNA machine driven by DNA strands displacement reactions.¹⁷ The switchable movement can be monitored by fluorescence signal of FRET (fluorescence resonance energy transfer) pair dyes TET (5‘-tetrachloro-fluorescein phosphoramidite) and TAMRA (carboxy-tetramethylrhodamine) that labeled on two arms.

(E) DNA origami box with lid that opened or closed by strands displacement reaction.¹⁸

(F) DNA origami-based nano-winch which can operate on cell surface via single- and double-stranded DNA linkages.¹⁹

(G) DNA origami-based nanoprinter using three independent DNA origami linear actuators can position the write head over a two-dimensional canvas driven by strands displacement reactions.²⁰ Figure 2B and 2F reproduced from Haley et al.¹⁴ and Mills et al.¹⁹ that are open-access articles distributed under the terms of the Creative Commons CC BY license; Figure 2C and 2G are adapted with permission from Mayer et al.¹⁶, copyright ©2023 American Chemical Society and Benson et al.²⁰, copyright ©2022 The American Association for the Advancement of Science, respectively.

In addition, parallel and sequential logic gate reactions and multiprobe detection usually need several SDR design which requires high hybridization specificity. The closely related DNA sequences, defective oligonucleotide synthesis, and side products usually induce the imposition of large thermodynamic penalties on designed hybridization events. To retain robust strand displacement kinetics and introduce additional control on SDR, attempts have been focused on adding a spacer to separate toehold and displacement domain,¹¹ a purposely designed mismatches at different locations to create frustrated complement molecules(Figure 2B),^12,13,14,15 or minimizing the crosstalk binding of toehold, invader, or incumbents to other background sequences by using only three bases in sequences design, or intramolecular hairpin design in invader(Figure 2C).¹⁶

The DNA tweezer is the pioneer work of creating SDR-driven DNA machine, which can be closed and opened for many cycles consuming “fuel” DNA strands (Figure 2D).¹⁷ The body structure is assembled with three SSDs. Leaving few bases in the middle as flexible hinge, the first strand function as the central strand and its two ends bind the second and third DNA to form duplex tweezer arm. The leftover overhangs from the second and third strands are free for further hybridization. FRET pair dyes TET (5‘-tetrachloro-fluorescein phosphoramidite) and TAMRA (carboxy-tetramethylrhodamine) were labeled at the end of the central strand. The addition of fuel strand F initiates the intramolecular hybridization with two overhangs to close the tweezer. The additional 8-base overhang section of F is a toehold that can be recognized by removal of strand $\overline{\mathrm{F}}$. By forming a double-stranded waste product F$\overline{\mathrm{F}}$, this removes F and returns the tweezers back to the open state. Along the sequential addition of F and $\overline{\mathrm{F}}$ strands, fluorescence quenching of TET by TAMRA and recovery indicate the closing and opening of the tweezer. Fluorescence signal damping along each cycle showed gradually decreased robustness of DNA tweezer which can be attributed to contaminated DNA sequences as well as the formation of higher-ordered structures. Needless to say, that DNA tweezer represents the earliest efforts to create switchable actuation which inspired many follow-up DNA nanomachines.

The dynamic regulation using SDR can be site-specifically placed on larger DNA machine bodies. In particular, many DNA origami actuators have taken this advantage to arrange docking strands or desired molecules and to execute complex tasks. Earlier efforts can be traced to a three-dimensional origami box with a controllable lid (Figure 2E) which can be opened or closed with SDR by labeling protector and incumbent strand on different substructures.¹⁸ The initial hybridization of protector and incumbent closed the lid. The “key” strand hybridizes with the incumbent strand to release protector and open the box. Recently, Mills et al. designed an SDR-driven nano-winch which can land on cell membrane to explore the mechanical force landscape of living cells (Figure 2F).¹⁹ Another recent paradigm for integrated DNA nanomachine is the nanoprinter for placing ink DNA strands on 2D canvas by moving the write head in two dimensions with an SDR-driven 2-fold linear actuator (Figure 2G).²⁰ Relying on unique sequences on tracks, the write head can position single molecules on specific site as confirmed by the super-resolution method DNA-PAINT. More examples of stepwise movement are concluded later in the DNA walkers section.

pH-responsive triplex DNA

Triplex DNA, also known as H-DNA, comprises three oligonucleotides that wind around each other. In addition to the B-form DNA double helix, the third oligonucleotide formed Hoogsteen base pairs through hydrogen bonding in the major groove (Figure 3A). In the parallel triplex consisting of two pyrimidine and one purine (CG∗C or TA∗T), the CG∗C triad is stable in lower pH less than 6.5 when N3 of cytosine is protonated, but TA∗T triad can only decompose when pH is above 10. Thus it is possible to engineer reversible triplex by sequence design that is responsible for an exact threshold pH value.²¹ Chen incorporated reversible triplex in a tweezer system.²² The tweezer consisting of three short duplexes and a single-stranded region is in an open state at pH 8.0. When pH decreased to 5.0, the random coiled single strand binds to the nearby duplex and closed the tweezer. The oscillated pH between 5.0 and 8.0 triggered the continuous cycles between open and close states. It is challenging to distinguish parallel and antiparallel triplex in solution but possible to characterize the different triplex with nanoalignment in an origami framework. Yamagata labeled the triplex component in a DNA origami frame and investigated the real-time triplex formation with fast atomic force microscopy.²³ Ijäs employed reversible triplex to assemble origami-based nanocapsules that can be switched between open and close state according to altered pH (Figure 3B).²⁴ Julin labeled triplex subcomponent on a cross-shaped DNA origami and build a pH-sensitive reconfigurable DNA origami (Figure 3C). Such origami unit can be further assembled into one-dimensional and two-dimensional lattices which still possess pH responsiveness.²⁵ Moreover, several groups recently^26,27 have explored using triplex formation to fold duplex scaffold into defined nanostructures. In comparison to standard Watson-Crick-mediated duplex origami, triplex origami is based on Hoogsteen binding to align neighboring duplex. Owing to the mature production of double-stranded scaffold-like plasmids, triplex origami is promising to further lower the material cost but also robust to assembly DNA structure that is responsive to pH but also resistant to DNase I degradation (Figure 3D).²⁷

Figure 3.

Triplex formation and base pair stacking enabled nanomachines

(A) Two pyrimidine and one purine triad TA∗T and CG∗C formed through the combination of Watson-Crick and Hoogsteen interactions.

(B) Reversibly opened and closed origami nanocapsule using pH-responsive triplex DNA motif.²⁴

(C) Triplex DNA-enabled configurable DNA origami lattices that are responsive to pH.²⁵

(D) Hoogsteen triplex origami structures (up: the design principle; down: left is formed via parallel triplex design and right is via nonparallel triplex design).²⁷

(E) Schematic illustration of the free-energy changes with/without one extra base pair stacking interaction (left) and the corresponding super-resolution data (right).²⁸ The cyan and magenta docking strands are for identifying the origami location and for analyzing base-stacking interactions, respectively.

(F) Base pair stacking-enabled origami assembly in one-dimensional array.²⁹

(G) Switchable shape-complementary DNA objects in open and close conformations that are responsive to temperature.³⁰ Figure 3B (Ijäs et al.²⁴, copyright © 2019 American Chemical Society), 3C (Julin et al.²⁵, copyright © 2023 The Authors; Published by American Chemical Society), and 3E (Banerjee et al.²⁸) are reproduced from articles that licensed under CC-BY. Figure 3D and 3G are adapted with permission from Ng eta al.²⁷, copyright ©John Wiley and Sons and Gerling et al.³⁰ copyright ©The American Association for the Advancement of Science, repectively.

Base pair stacking

As mentioned earlier, π-stacking of neighboring base pairs is a significant stability force during duplex formation. Using DNA-based point accumulation in nanoscale topography (DNA-PAINT), Banerjee conducted super-resolution analysis with imager strands with and without dinucleotide stacking and have found that one extra base pair stacking interaction can enhance 250-fold stabilization (Figure 3E).²⁸ During DNA origami design and assembly, however, blunt-ends interactions should be prevented; otherwise it is almost not possible to obtain monodispersed origami nanostructures due to the blunt ends-induced aggregation. However, it is also possible to employ base pair stacking to build reversible assemblies. Woo explored the first example using geometric complementarity of blunt-end stacking to assemble discrete numbered origami patterns (Figure 3F).²⁹ Gerling further developed three-dimensional shape-complementary DNA objects that can assemble in response to cation concentration, temperature, as well as the triggering oligonucleotide (Figure 3G).³⁰ Later on, large capsid assembly solely with shape complementarity was also created,³¹ demonstrating that base-pairing stacking can play as a universal molecular glue for higher-order functional structures.

Intramolecular DNA interactions

The efficiency of SDR-driven reversible opening and closing, however, decays after each cycle due to higher-ordered side products, uneven stoichiometric DNA fuel strands that titrated in, or the accumulated fuel waste in solution. To increase the robustness of DNA nanomachines, intramolecular DNA tethers can be introduced, creating an intramolecular contraction force and facilitating the reconfigurable actuations. Typically, tether DNA is encoded with sequences that form intramolecular secondary structures. Hence the tether DNA sequence design and quality control are the key tasks to avoid sequence errors and maintain the required metastable state. For instance, even a single extra PAGE purification can significantly improve the hairpin strands quality.³² In this section, we summarize the role of intramolecular DNA interactions in building DNA machines.

B-Z transitions

Before the real tether DNA, conformational transition from right-handed B-DNA to left-handed Z-DNA at a considerable cation concentration is one specific motif. When DNA is encoded with alternating purine and pyrimidine nucleotides, in particular with guanine and cytosine, a high-salt solution reduces the electrostatic repulsion between phosphate backbones and triggers the transition from B-DNA to Z-DNA. Mao connected two rigid DNA “double-crossover” (DX) motif with a short B-Z transforming duplex.³³ Upon high salt concentration, the transition from B- to Z-form leads to cis- to trans-isomerization of DX motif and the corresponding efficiency changes in fluorescence resonance energy transfer. Rajendran labeled the B-Z transition motif in a DNA origami frame and investigated the salt-dependence of B-Z transition and the real-time dynamic with fast atomic force microscopy (Figure 4A).³⁴ Despite the immobilized origami frame on mica surface, they found the B-Z transition takes place at the MgCl₂ concentration range of 10–25 mM. The real-time dynamics showed the B-Z transition takes minutes to complete while the non-Z-DNA coded region has non-rotary motions.

Figure 4.

B-Z transition, entropic forces, and intramolecular secondary structure enabled nanomachines

(A) B-Z transition investigated by fast atomic force microscopy by labeling B-Z transition flag inside an origami frame.³⁴

(B) Force clamp based on entropic elasticity of single-stranded DNA (ssDNA) spring. On a given DNA origami rigid supporter, the longer the ssDNA, the higher entropy and the lower entropic force.³⁶

(C) Assembling macroscopic machine design, DNA origami technique can assemble angular, linear motion, and crank-slider mechanism at nanoscale.³⁷

(D) Different ssDNA spring can bend DNA origami compliant nanostructures to different angles. The shorter connection, the larger bending angles.³⁸

(E) dsDNA-to-ssDNA transition created a strong contractive force which was employed to bend a DNA origami switch.⁴⁰

(F) Two silver nanoparticles labeled at the ends of DNA arms. The closed state of DNA tweezer induced the proximity of sliver nanoparticles which further triggers surface-enhanced Raman scattering (SERS) biosensing applications.⁴⁴

(G) Induced size changes of one-layer DNA origami upon the actuation of i-motif.⁵⁴

(H) DNA tweezer controlled distances of cascade enzymes glucose oxidase (GOx) and horseradish peroxidase (HRP), regulating cascade reactions efficiencies.⁵⁶

(I) An i-motif interconnected DNA tweezer on cell surfaces for real-time imaging of cell surface pH changes.⁵⁷

Figure 4A, 4D, 4F, 4G, and 4I are reproduced with permission from previous studies^{34,38,44,54,57} copyright © 2012, 2013, 2020, 2017, and 2018 American Chemical Society, respectively; Figure 4H is adapted with permission from Xin et al.⁵⁶, copyright © 2013, John Wiley and Sons; Figure 4B and 4E are reproduced from Kramm et al.³⁶ and Gür et al.⁴⁰ that are open-access articles distributed under the terms of the Creative Commons CC BY license, repectively.

Entropic elasticity

Natural and synthetic polymers chains are often highly flexible as bond angles are usually not 180° and rotatable. Entropically, polymers tend to contract to maximize the possible chain conformations. This contraction force in the opposite direction of external stretching can be exploited to drive nanoactuators. For ssDNA that encoded sequence possessing no intramolecular structures, when both ends were fixed on a rigid supporter, the tendency of maximizing its entropy generates entropic force to reduce the end-to-end distance. Given a certain spanning distance, the shorter the DNA spring, the higher the entropic force is. Nickels anchored the ssDNA spring on a stiff origami structure. Employing the dynamics of the Holliday junction to calibrate the force clamp, as an example, the binding dynamics of TATA binding protein is investigated with high sensitivity.³⁵ Kramm et al. further applied the DNA origami-based force clamp to monitor the binding dynamics of several other general transcription factors, demonstrating the universality of DNA origami force clamp for biophysical studies (Figure 4B).³⁶

In addition to static origami supporters, ssDNA spring can be also anchored on dynamic origami substructures. Castro group demonstrated the capacity of DNA origami to make machinery elements such ssDNA spring-facilitated origami slider in linear motions³⁷ and built a series of tunable geometries³⁸ as well as demonstrated localized transient hybridization events³⁹ (Figures 4C and 4D). Gür reported an open-close dual states origami switch relying on the entropic force generated from dsDNA-to-ssDNA transition, realizing reversible plasmonic coupling events of assembled gold nanoparticles (Figure 4E).⁴⁰ The bending up of several duplex bundle confirmed that the contractive force from ssDNA spring is considerable which could have further applications in building complex nanoactuators. Another example regarding ssDNA-dsDNA transition is a molecular “domino array.”⁴¹ Song designed a square-shaped antijunction unit containing four short DNA duplex regions of equal length and four dynamic nicking points. These four nicking points hence enable four base pair stacking interactions which switch the antijunction unit between two conformations. If an extra ssDNA loop was included at one nicking point, interestingly, it frustrates one base pair interaction and stabilizes the antijunction unit at one state. When adding the trigger strand, it hybridizes the loop region and drives the antijunction to another state. This transformation can be programmed to implement long-range information relay and complex dynamic control.

Hairpins and quadruplex

Single-stranded DNA encoded with random sequences mainly possesses entropic forces, while ssDNA encoded self-complementary regions can form intramolecular hairpin or quadruplex. Comparing to contractive forces from entropic elasticity, hairpins and quadruplex have a clearer secondary structure that can be used to drive nanomachines. In particular for the tweezer system, intramolecular hairpin can be incorporated to hold two arms within a certain radius and increase the opening and closing efficiency. For instance, a tweezer integrated plasmonic nanoparticles to introduce alternative output signals such as plasmonic quenched fluorescence^42,43 and approximating two silver nanoparticles to enable SERS (surface-enhanced Raman scattering) (Figure 4F).⁴⁴

For quadruplex tether we here focus on the specific intercalated tetraplex (i-motif), a cytosine-rich DNA strand that can fold four-stranded DNA structure in acidic conditions.⁴⁵ The folding and unfolding of i-motif have been investigated with methodologies such as circular dichroism, gel electrophoresis, ultraviolet (UV) absorbance, trapped ion mobility spectrometry-mass spectrometry (TIMS-MS),⁴⁶ and X-ray crystallography.⁴⁷ To realize protonated cytosine-cytosine (C:C⁺) base pair formation, pH oscillations generated by various means have been studied.^48,49 In particular, kinetics study with stopped-flow circular dichroism (SFCD) technique showed that the DNA i-motif can fold and unfold on a timescale of 100 ms in a solution of pH ∼5 and pH ∼8, respectively.⁵⁰ At single-molecular level, the rupture forces to unfold an i-motif is in a range of <30 pN.^51,52 In addition, the formation of duplex DNA from i-motif and its complementary strand is more competitive than i-motif quadruplex. Thus, when incorporating i-motif in DNA substructures, several mismatches are required to add in the complementary strands so that i-motif DNA can reversibly hybridize and dehybridize to form quadruplex in a pH-dependent manner.⁵³ Anyhow, the formation of intramolecular quadruplex shortened the end-to-end distance. Majikes implemented multiple i-motif strands to seam two origami substructures (Figure 4G).⁵⁴ The molecular-level contraction generates nanoscale deformation. Instead of parallel array, aligning hairpin or quadruplex in series may generate multifold shape changes. Eventually these nanoscale actuators offer various means to accelerate diffusion-controlled reactions (Figure 4H)^55,56 or to monitor signals in bulk solution, such as DNA tweezers immobilized by Zeng on cell surface to probe real-time cell surface pH changes, demonstrating the great potential for investigating local biochemical processes (Figure 4I).⁵⁷

Enzyme assistance, DNA aptamer, and non-DNA interactions

Enzyme and DNA aptamers

The intrinsic DNA-binding proteins such as restriction enzyme can create DNA oligonucleotides of various lengths and generate discriminable SDRs to trigger DNA machines. While DNA aptamers are a short ssDNA which can recognize specific targets with high affinity. The targets of DNA aptamer can compete with the complementary strand of aptamer DNA and enable SDR-like reversible interactions. Xu employed the classic DNA tweezer design but using fuel strands generated from telomerase reactions (Figure 5A).⁵⁸ In the presence of telomerase, the telomerase substrate strands can be properly elongated to efficiently close and open the tweezer. A logic-gated nanorobot for target delivery designed by Shawn Douglas described a robust aptamer lock mechanism (Figure 5B).⁵⁹ The cargos in robot are exposed in response to binding target antigen keys, and combinational aptamers are designed to proof the aptamer-encoded logic gating. Alternatively, metal ions binding DNAzymes fold ssDNA sequences into complex tertiary structures which can cleave phosphodiester backbone at a ribonucleotide or deoxyribonucleotide site.^60,61 Many metal ions and corresponding DNA sequences have been well coordinated for establishing metal-sensing methods, isolating new DNA aptamer sequences, as well as for building DNA machine systems for various purposes.^62,63

Figure 5.

Enzyme, DNA aptamer, and non-DNA interactions enabled DNA machines and superstructures

(A) DNA strands produced from telomerase reactions as fuel to close the classic DNA tweezer for monitoring human telomerase activities.⁵⁸

(B) DNA aptamer enabled logic-gated nanorobot that in response to target antigen keys.⁵⁹

(C) Cholesterol molecules placed on one side of single layered DNA origami introduced hydrophobic interactions to fold up the DNA origami.⁶⁴ The addition of surfactant Tween 80 neutralizes the hydrophobic effect and opened the origami bilayer.

(D) Thermo-responsive poly(N-iso-propylacrylamide) (PNIPAM) possesses hydrophilic-to-hydrophobic phases transition and enabled switchable close-open of DNA origami tweezer.⁶⁶

(E) Employing asymmetric sequence design for binding gold nanoparticles, higher temperature can release gold nanoparticles from the lower affinity bottom arm, hence realizing thermal actuation design.⁶⁷

(F) DNA origami polymerization based on adamantane/b-cyclodextrin host/guest interactions.⁶⁸

(G) Dimeric origami assembly based on formation of peptide coiled-coil heterodimers.⁶⁹ Figure 5A and 5E are reproduced with permission from Xu et al.⁵⁸ and Johnson et al.⁶⁷ copyright © 2018 and 2019 American Chemical Society, respectively; Figure 5B is reproduced with permission from Douglas et al.⁵⁹ Copyright © 2012 The American Association for the Advancement of Science; Figure 5C, 5D are reproduced with permission from List et al.⁶⁴ and Turek et al.⁶⁶ copyright WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim; Figure 5F is from Loescher et al.⁶⁸ © 2019 The Authors, published by Wiley-VCH Verlag GmbH & Co. KGaA; Figure 5G is reproduced with permission from Jin et al.⁶⁹ Copyright © 2019 American Chemical Society and is licensed under CC-BY.

Non-DNA interactions: Hydrophobic effect, host-guest recognition, and peptide-peptide interactions

Along with the mature chemical synthesis of nucleic acid, many functional groups can be conjugated to DNA oligonucleotides to enrich reversible interactions for building DNA nanomachines. Intrinsic hydrophobic molecules such as cholesterol⁶⁴ (Figure 5C) and poly(arylether) dendron⁶⁵ can introduce hydrophobic interactions between DNA origami objects for regulatory actuation. Polymer poly(N-iso-propylacrylamide) (PNIPAM) is negatively thermosensitive and possesses a lower critical solution temperature (LCST). In aqueous solution, PNIPAM contracts upon temperature increasing above its LCST and experiences coil-to-globule transition by forming more intramolecular hydrogen bonds. Hence, PNIPAM introduced a temperature-controlled hydrophilic-to-hydrophobic phase transition to reversibly open and close the DNA flexor (Figure 5D).⁶⁶ In addition, Johnson achieved the angular distributions of an origami hinge by placing nanoparticle of different sizes at different distance from the hinge vertex and realized rapid actuation dynamics following bulk solution temperature changes (Figure 5E).⁶⁷ Loescher introduced adamantane/b-cyclodextrin host/guest recognition on DNA origami objects to achieve one-dimensional polymerization (Figure 5F),⁶⁸ and Jin introduced coiled-coil interactions to build dimeric DNA nanostructures (Figure 5G).⁶⁹ Though switchable parameters and dynamic assemblies require more systematic investigations, non-DNA interactions are deemed to play a significant role for higher-order assembly and dynamic control.

Macroscopic manipulation techniques: Light, electricity, and magnetic field

For photic driving nanomachines, the role of light radiation is to generate conformational transitions, photothermal effects, or solution condition changes via specific light-induced reactions. In many cases, light-assisted DNA nanomachines require well-characterized conjugation of light photosensitizer and nucleic acid. Liang synthesized azobenzene-modified DNA as lock strand to close the tweezer (Figure 6A).⁷⁰ While planer trans-azobenzene can intercalate between duplex base pairs, nonplanar cis-azobenzene generates steric hindrance which can break down the hydrogen bonding and stacking between paired DNA bases and destabilize the duplex. The trick is that the trans- vs. cis-configurations of azobenzene are switched by irradiating with UV and visible light. So a photoresponsive DNA tweezer is built consuming no fuel DNA strands but only requiring alternate irradiation with visible and UV light. Owing to the fast photoregulation, the synthesized azobenzene-modified DNA can efficiently switch the ON and OFF states of the DNA tweezer which later inspired vast light-enabled DNA machines.⁷¹

Figure 6.

Macroscopic manipulation technique light, electricity, and magnetic field powered nanomachine

(A) Nanotweezers photoswitched by light irradiation. The hybridization of azobenzene-labeled lock strands to tweezer arms is controlled via UV- or visible light-enabled trans-cis isomerization.⁷⁰

(B) With one end anchored on electrode surface, applying different bias voltage can absorb or repel the negatively charged origamis from the surface.⁷²

(C) Electric field-controlled robotic arm on a DNA origami platform.⁷³

(D) DNA origami base of the nanorobotic arm is fixed on surface through biotin-NeutrAvidin binding, and the electrical actuation force on the nanorobotic arm is from an interplay between Coulomb force (F_C), electrophoresis(F_ep), and electroosmotic flow(F_eo).⁷⁴

(E) Built by attaching twisted DNA tile-tubes to magnetic beads via biotin-streptavidin coupling, free magnetic swimmer travels along with the directions of applied magnetic field.⁷⁹

(F) Surface-anchored origami-based magnetic nanodevices and the circular movement under in-plane but different strengths magnetic fields.⁸⁰ Figure 6A is adapted with permission from Liang et al.⁷⁰ Copyright © 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim. Figure 6B and 6D are adapted with permission from Kroener et al.⁷² and Vogt et al.⁷⁴ Copyright © 2017 & 2023 American Chemical Society, respectively; Figure 6C is adapted with permission from Kopperger et al.⁷³, copyright © 2018 The American Association for the Advancement of Science; Figure 6E and 6F are reproduced with permission from Maier et al.⁷⁹ copyright ©2016 American Chemical Society and 80 that are licensed under CC-BY, repectively.

Electric-powered DNA actuators are correlated to the physicochemical properties of DNA. The primary property of hydrosoluble DNA molecule is their low pKa of phosphate backbone. In commonly used PBS (phosphate-buffered saline) or TAE (Tris-acetate-EDTA, EDTA: ethylenediaminetetraacetate) buffer, deprotonation renders DNA negatively charged backbones, which is associated to many biological functions. For instance, nucleosome—the structure unit of chromosome—is composed of two turns of DNA around positively charged histone core based on charge-charge interactions. In the scope of DNA assembly, deprotonized phosphate backbones are also crucial factors in DNA structure folding, cationic strength dependent stability, and electrophoresis-based purification. When immobilizing DNA nanostructure on surface, their orientations can be operated by externally applied electrical field. At negative potentials, negatively charged DNA structures are repelled from the surface and stand up, while at positive potentials they are attracted and compacted on the surface (Figure 6B).⁷² Kopperger designed a robotic arm on a DNA-based molecular platform and demonstrated computer-controlled arm switching (Figure 6C).⁷³ With arm length extended to more than 400 nm, it can transport objects of interest under the electrically controlled movement. In SDR-based DNA machines, the hybridization rate constants k_{hybridization} and k_dissociation determined the reaction speed and often sequential reactions are required which slow down the overall transformation speed. In comparison, electrical control is contact free and can be at least five orders of magnitude faster. And multiple parameters such as ionic strength, solution viscosity, and robotic arm geometries can be tuned to control the electrical “actuatability” (Figure 6D).⁷⁴ While mechanical energy can be stored in the linker region of robotic arm,⁷⁵ Pumm et al. positioned a rotor arm on triangle-shaped pedestal through three unpaired nucleotides as a pivot point to explore Brownian ratcheting.⁷⁶ To achieve that, three rectangular origami plates protruding with an inclination as physical obstacles are designed on the edges of the triangular platform. As revealed by total internal reflection fluorescence (TIRF) microscopy, the rotor arm showed unbiased random rotary movements preferentially at six discrete positions when the external electrical field was off. Also, the robotic arm can be controlled by the orientation of the alternating current field to perform ATPase-like rotary movement. In addition, Shi et al.^77,78 embedded DNA origami turbines in a solid-state nanopore and observed unidirectional rotation driven by a salt gradient or applied voltage. The chemical compatibility of silicon nitride nanopore and DNA-based rotor may further enhance the rotary apparatuses.

The third macroscopic manipulation method is employing magnetic field. In this case, DNA machine bodies have to be connected with magnetic beads. Maier attached chiral-shaped tile tube to streptavidin-coated magnetic nanoparticles and obtained free DNA-flagellated magnetic bead hybrids. Under external applied homogeneous magnetic field, the free swimmer was propelled hydrodynamically by DNA flagellar bundles (Figure 6E).⁷⁹ Lauback assembled a mechanically stiff micron-scale origami arm with one end fixed on surface and the other end attached to a digoxigenin-coated magnetic bead (Figure 6F). Applying in-plane magnetic field with different strengths, rotation motions of the lever arm showed magnetic actuations allowing a direct control with high spatial position and sub-second response times.⁸⁰

Hybrid system with motor proteins

An alternative strategy is to build hybrid system with DNA-based walker body and biological motors such as myosins. Relying on DNA-enabled molecular engineering, it offers a controlled motor organization to analyze the effects of motor density and spacing on gliding speed, as well as to reconstitute muscle motilities in an artificial system.⁸¹ Another example demonstrated polarized radial arrays of microtubules with the help of kinesin dimer-bounded DNA duplex.⁸² Through accurate positioning on DNA duplex, two attached kinesins are oriented approximately antiparallel which align parallel microtubule together to mimic natural asters for directional transport. A recent example by Ibusuki (Figure 7)⁸³ re-engineered protein motor dynein by replacing microtubules-binding domain with a DNA-binding protein at the end of the dynein stalk. The ATP-driven conformational changes of AAA+ modules move the stalk and hence the DNA-binding domain along the track. By attaching modified dynein on surface, they performed the gliding assays of DNA nanotubes and found the velocity of DNA nanotubes dependents on the type of hybrid motors. Moreover, the modified dyneins can be placed on synthesized DNA tracks to gain a robust walker system.

Figure 7.

DNA machines driven by biological motors

Gliding assays with dyneins immobilized on surface to move DNA nanotubes.⁸³ Human cytoplasmic dyneins were re-engineered to attach different DNA-binding domains, which can recognize specific sequences that were incorporated in the DNA tube design. Figure reproduced with permission from Ibusuki et al. ⁸³, copyright © 2022 The American Association for the Advancement of Science.

DNA walkers

Individual DNA origami actuators realize localized conformational changes, but DNA walkers can transport cargo molecules along the track to predefined destinations, even in an autonomous manner. The design of DNA walker usually consists of the body and the track. Due to the consistent thermal fluctuations, the walker body has to spare at least one foot to be in contact with the track. A loss of contact thus will free the walker to the solution of no preferential direction of flow. Like other DNA-based nanomachines, DNA walkers require driving forces to operate, mostly relying on strands hybridization-based SDR. To exchange walker legs, either new invader strands are added or new toehold overhangs have to be created by restriction enzymes for driving the energetically favorable sequence hybridizations. Apart from developing DNA-catalyzed loop reactions,⁸⁴ early efforts have pioneered and established one hybridization-driven autonomous molecular motor with the help of ligation and restriction enzymes.⁸⁵ Three anchorages are labeled on a DNA duplex track, and a single six-nucleotide DNA “walker” (not a duplex segment but a staggered 2x three-base overhang on the anchorage) moves from one anchorage to another. Through toehold-mediated hybridization, the walker is linked to the next anchorage. With the help of PflM I and BstAP I for alternate enzymatic ligation and cleavage and owing to the unique recognition sites and restriction patterns, the walker motion is defined to be unidirectional.

Employing enzymes for conditional DNA cleavage later had been used to design more efficient free-running walkers. For instance, under a “burnt bridges” mechanism, Bath demonstrated an ssDNA cargo moving directionally on a one-dimensional track.⁸⁶ The ssDNA firstly hybridized to one stator on track; then nicking enzyme N.BbvC IB recognizes and cuts the stator in the duplex of stator-cargo. With a short stator fragment released, a single-stranded overhand is created on the cargo strand which serves as a toehold for the adjacent stator to hybridize. Driving by SDR, the ssDNA cargo thus steps onto stator_i+1 and one cycle is accomplished. The backwards step is inhibited because the stator was cut in the previous step and cargo DNA would keep moving until the endpoint stator containing mismatches is resistant to cutting. Such strategy also allowed the DNA cargo to transport along a two-dimensional track on DNA origami (Figure 8A),⁸⁷ and by selectively removing blocking strands on junction stators it can open specific path to process cargo (Figure 8B).⁸⁸ Alternatively, Lund et al. designed a DNA “spider” walker consisting of a streptavidin body, a capturing DNA leg, and three DNA enzyme legs. The DNA enzyme binds and cleaves docking strand on the track into shorter products creating toehold overhand for the next docking strand to capture the spider body.⁸⁹ While the tensegrity triangle walker driven by SDR can pick up several objects and carry them to designed locations,⁹⁰ Thubagere et al. designed a cargo-sorting system using a robot. By introducing dual-toehold-enabled reversible strand displacement reaction, the single-stranded robot can bind to different track DNA and travel around on a two-dimensional surface. Once the single-stranded cargo is picked up, the robot DNA hence carries the cargo until handing the cargo to the goal DNA. Owing to the robust DNA sequence design and hybridization, multi-cargos at unordered locations can be picked up and carried to specific destinations (Figure 8C).⁹¹

Figure 8.

DNA walkers

(A) A single-stranded DNA walker moves autonomously on predefined tracks. The nicking enzyme cutting reveals the toehold of the walker body which enables branch migration and the transfer of the motor to the neighboring stator.⁸⁷

(B) By selectively removing block strands at unique joints, DNA walker can navigate on tracks in a controllable manner.⁸⁸

(C) The single-stranded DNA robot bearing dual toehold of same length can reach equilibrium reactions with track strands. The extended robot arm thus can carry cargo and wonder around until placing the cargo strands at goal position.⁹¹ Figure reproduced with permission from Thubagere et al.⁹¹, copyright © 2017 The American Association for the Advancement of Science.

(D) Catalytic hairpin chain reaction on microparticle surfaces demonstrated that catalyst strand (red) triggered the hybridization of H2 (mars green hairpins) to the surface-immobilized H1 (gray hairpins).⁹⁴

DNA fuel-driven DNA walker is a strategy for a controlled stepwise movement. Such as a bipedal DNA walker alternately binds and releases legs according the added DNA fuels.⁹² Alternately adding DNA fuels to bind and another to release the legs takes time and causes DNA wastes. HCR, the hairpin hybridization chain reaction,⁹³ triggers the opening of a pool of intramolecular hairpins to form a nicked double helix principally with only one ssDNA initiator. When using hairpin molecules as anchorages and fuels, DNA walker can move autonomously too. Jung labeled hairpin 1 (H1) on microbeads surface and left hairpin 2 (H2) in bulk solution. The initiator DNA opens one H1 and displaces with H2 and keeps wandering around the microbead surface before H1 is depleted (Figure 8D).^94,95 Chao et al. designed HCR-enabled pathfinding platform in a molecular maze on two-dimensional origami surface. At the presence of hairpin DNA fuel, the initiator DNA triggered a localized chain reaction and formed a molecular pattern. With an additional set of hairpin DNA fuels and corner hairpin in the hidden pattern, the DNA initiator can navigate and reveal a specific graph.⁹⁶

Moreover, reactants that attached to DNA adapters can be brought in close proximity along with DNA assembly. DNA-templated molecular conjugation thus is achieved with controlled orientation and stoichiometry. Through DNA sequence design, a large number of DNA adapters can carry respective reactant for the parallel synthesis of diverse products in a one-pot reaction. Secondly, autonomous DNA hybridization cascade reactions will create oligomers in a one-by-one manner, resembling mRNA-templated peptide synthesis during ribosomal translation. In either cases, reactants that terminated at one end of DNA strands should be able to conduct peculiar cleavage and coupling reactions—cutting off from one DNA adapter and conjugating to the next by building block molecule, such as Wittig reaction developed by McKee,⁹⁷ nucleophilic attack of the incoming ylide on the aldehyde, and handing the growing chain to the incoming monomer through an olefin linkage. Stepwise addition of template and removal strands controlled the monomer that can be coupled to oligomer. They later explored an autonomous, programmable covalent synthesis employing a specially designed HCR containing cargo-carrying initiator duplex, cargo-carrying chemistry hairpin group, and instruction hairpin group.⁹⁸ The split toehold from initiator duplex opens one instruction hairpin and transfers the cargo-carrying strands to the other end of instruction hairpin. The newly exposed loop region from this instruction hairpin then can bind to one chemistry hairpin allowing the group transfer via Wittig reaction. The added chemistry hairpin loop then can bind to other instruction hairpin for a new cycle to begin. Taken aforementioned enzyme-drive autonomous walker,^86,89 He incorporated N-hydroxysuccinimidyl (NHS) ester and amine group, respectively, on walker and track DNA molecules and demonstrated a series of DNA-templated amine acylation reactions.⁹⁹ Additional types of DNA-templated synthesis have been reported, and we refer readers to documents specialized on this topic.^100,101

DNA machine-enabled sensing

The utilization of DNA machines is closely connected to macroscopic signal output from collective effects enabled by DNA actuators. Despite a large number of studies on DNA sensing, we here summarize several sensing examples with DNA-colloidal systems, surface-anchored DNA, and DNA-enabled bulk materials. Firstly, we refer DNA-colloidal systems to solution-based monodispersed DNA-nanoobjects. Many reported sensing examples belong to this category with optical readout such as intensity changes of light absorption or emission. For example, target DNA molecules hybridize with DNA on gold nanoparticles and aggregated gold nanoparticles induce plasmonic optical effects reflected on absorption peak shifts. When half i-motif oligos were labeled on gold nanoparticles and i-motif quadruplex formation aggregates gold nanoparticles, which functioned as a plasmonic sensing method to monitor pH changes.¹⁰² One particular type plasmonic chiral sensing arises from switchable enantiomers of plasmonic nanoparticles on DNA actuators. It allows to detect trace amounts of target molecules in visible wavelength (Figure 9A) or pick up chiral signals by plasmonic hotspot from UV range down to the near-infrared.^{103,104,105,106}

Figure 9.

DNA machine-enabled sensing

(A) Chiral plasmonic metamolecule (CPM) with pH-responsive DNA locks. Light illumination induced merocyanine-based photoacid to release protons which promotes the triplex formation and hence the closed right-handed state of CPM. Reaction in dark reversed the reaction and multiple cycles can be realized.¹⁰⁶

(B) i-motif controlled surface hydrophobicity. With one end of i-motif labeled with hydrophobic groups, the other end immobilized on gold substrate, the formed i-motif and extended duplex can bury and expose hydrophobic groups hence change surface wettability.¹⁰⁷

(C) Hairpin chain reaction (HCR) is implemented in a DNA-cross-linked polyacrylamide hydrogel and the in situ HCR expands the volume to a well-defined final size¹¹⁵ Figure 9A is adapted from Ryssy et al.¹⁰⁶ that is distributed under the terms of the Creative Commons CC BY license; Figure 9B and 9C are reproduced with permission from Wang et al.¹⁰⁷, copyright © 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim and 115, copyright © 2017, The American Association for the Advancement of Science, respectively.

Secondly, surface-immobilized DNA molecules can introduce electrochemical surface sensing and desired surface properties. Electrochemical surface sensing is a big topic which is skipped in this paper as it has been covered in many other places. Wang labeled a thiol group to the five prime and a hydrophobic group to the three prime of i-motif strand and anchored them to gold surface via gold-thiol bond. The i-motif structure on the surface in basic buffer (pH 8.5) is in a state of stretched single-stranded structure to expose the hydrophobic molecule, and, during the proton-driven quadruplex folding, the hydrophobic end is hidden inside. Contact-angle measurement showed this realized proton-driven approach to manipulate macroscopic surface wettability (Figure 9B).¹⁰⁷ When a dye molecule is labeled at the three prime, dye molecules are brought to gold surface and quenched in the quadruplex state, demonstrating a pH-sensing surface.¹⁰⁸ Moreover, the transition of single-stranded i-motif to larger-diameter quadruplex generates significant steric and electrostatic repulsion which can bend a micrometer cantilever,¹⁰⁹ encapsulate drug molecules in the cavity close to the surface,¹¹⁰ and tune the pore sizes.¹¹¹

Thirdly, DNA-assembled smart hydrogels are the DNA sensors in bulk phase.^112,113,114 DNA as building blocks for hydrogel inherited the advantages of DNA such as biocompatibility, sequence designability, recognizability, self-healing, rigidity, as well as in situ polymerization to achieve volumetric expansion (Figure 9C).¹¹⁵ Importantly, the stiffness of DNA hydrogels can be tuned through not only DNA concentrations but also incorporating mismatched sequence design to change permeability. A recent work by Zhou et al. introduced different amounts of mismatch bases to regulate the mechanical strength of hydrogels. As all mismatch bases are located in the linear linker but not the Y-shape scaffold, the network topology and chemical composition remained unchanged. They found that storage modulus of such hydrogel ranging from ∼1 to 3 kPa does not change the differentiation of neural progenitor cells, but the degradation of the hydrogel can enhance cell-cell interactions.¹¹⁶ An earlier report from Liu group went one step further applying DNA hydrogels to carry homologous neural stem cell (NSC) and repair a 2 mm long spinal cord cavity.¹¹⁷ Within 8 weeks, basic hindlimb function was rebuilt after the formation of a continuous renascent neural network. Different from other flexible polymer-based hydrogels, DNA supramolecular hydrogel possesses fully rigid molecular substructure due to the segmental rigid DNA duplex. This not only avoided the formation of topological pores caused by polymer entropic recoiling forces but obtained an ultra-high permeable material that is similar to extracellular matrix. This work exploited a novel and efficient approach using robust DNA hydrogel to treat spinal cord as well as a wide range of other injuries.

Conclusion

Humans always want to build a human-like machine to do all kinds of work instead of humans. Man-made uncomplicated energy-conversion systems such as water mill and windmill can transform the kinetic energy of wind or water into mechanical energy, assisting humans to accomplish hard jobs. Other than energy converters, robots are decidedly more complex and can replace humans to perform repetitive and dangerous tasks with high accuracy and reliability. Guided by external controls or preset programs, robots operate autonomously and exhibit predetermined behaviors in response to their environment. In the field of synthetic biology, it shares a similar objective to build smart entities that are comparable to beings than to machines. As summarized here, the particular example of DNA nanomachineries has exploited tremendous efforts to assemble nanostructures conducting a complex series of predesigned actions.

Despite many obstacles remaining to be surmounted, DNA-based smart and functional motor systems operated by desired inputs (Table 1) have been established. To further advance DNA nanomachines, self-replication to reproduce itself is another aim.¹¹⁸ In modern factory automation, robotic machines such as drilling, shaper, welding, and painting can work coordinately to manufacture designed machines. There have been several profound studies on bioproduction of ssDNA oligonucleotides.¹¹⁹ Combining isothermal assembly assisted by enzymes¹²⁰ and kinetics control, it may be feasible to harvest designed DNA nanomachines with microbial production in a prescribed manner. With the fast developed computer-aided sequence design with cross-linked reactions and a sufficient supply of DNA materials, we believe DNA-enabled synthetic nanomachines with spatial and temporal controls can accomplish specified assignments in real-life biomedical and clinical applications.

Table 1.

Different motive forces that drive DNA nanomachine

Interaction type	Specific method
Intermolecular DNA interactions	Strands displacement reactions pH-responsive triplex DNA Base pair stacking
Intramolecular DNA interactions	B-Z transitions Entropic elasticity Hairpins and quadruplex
Hybrid driving forces	Enzyme assistance DNA aptamer Motor proteins
Non-DNA interactions	Hydrophobic effect Host-guest recognition Peptide-peptide interactions
Macroscopic manipulation	Light, electric and magnetic field