Biomimetic Autonomous Enzymatic Nanowalker of ... - ACS Publications

Jun 13, 2016 - rationally designed autonomous bipedal nanowalker made of DNA that achieves a fuel efficiency of less than two fuel molecules decompose...
2 downloads 0 Views 2MB Size
Download PDF
Biomimetic Autonomous Enzymatic Nanowalker of High Fuel Efficiency Meihan Liu,†,§ Juan Cheng,†,§ Shern Ren Tee,† Sarangapani Sreelatha,† Iong Ying Loh,† and Zhisong Wang*,†,‡ †

Department of Physics and ‡NUS Graduate School for Integrative Sciences and Engineering, National University of Singapore, Singapore 117542 S Supporting Information *

ABSTRACT: Replicating efficient chemical energy utilization of biological nanomotors is one ultimate goal of nanotechnology and energy technology. Here, we report a rationally designed autonomous bipedal nanowalker made of DNA that achieves a fuel efficiency of less than two fuel molecules decomposed per productive forward step, hence breaking a general threshold for chemically powered machines invented to date. As a genuine enzymatic nanomotor without changing itself nor the track, the walker demonstrates a sustained motion on an extended double-stranded track at a speed comparable to previous burn-bridge motors. Like its biological counterparts, this artificial nanowalker realizes multiple chemomechanical gatings, especially a biasgenerating product control unique to chemically powered nanomotors. This study yields rich insights into how pure physical effects facilitate harvest of chemical energy at the single-molecule level and provides a rarely available motor system for future development toward replicating the efficient, repeatable, automatic, and mechanistically sophisticated transportation seen in biomotor-based intracellular transport but beyond the capacity of the current burn-bridge motors. KEYWORDS: molecular machines, nanomotors, fuel efficiency, DNA, enzyme

N

the sole synthetic system qualified as a strict enzymatic nanomotor. This fact reflects some obstacles impeding this important research direction. One major obstacle is to find generally applicable physical mechanisms for harvesting chemical energy at the single-molecule level. These mechanisms should be scientifically advanced to break the two fuel per step threshold and facilitate sustainable autonomous motion. We report here a rationally designed enzymatic nanowalker meeting the requirements by implementing multiple chemomechanical gating mechanisms, which are characters of biowalkers22−25 known to be crucial for their efficient fuel use and sustainable motion.

anotechnology has long pursued the invention of rationally designed nanowalkers1−17 to replicate the chemically powered biomotors that autonomously transport cargos in the cell. The biomimetic systems also will provide an access to the science governing efficient chemical energy utilization at a single-molecule level, which is important to nanotechnology and energy technology but remains difficult to decipher from the complex biosystems. Indeed, the biomotors consume one fuel molecule at a time and may convert the chemical energy to work by ∼100% efficiency18 or drive directional motion up to one forward step per fuel molecule.19,20 Furthermore, the biomotors are enzymes in the strictest sense as they catalyze the fuel reaction without changing themselves and tracks, thus enabling repeatable transport seen in biology but yet to be done by synthetic motors. Chemically powered synthetic nanowalkers reported to date are predominately burn-bridge motors3−5,7,9−11,14 that consume the traversed track in an unrepeatable, domino-like chemical cascade. An enzymatic nanowalker beyond the burn-bridge design was reported in 2009 by Bath et al.,6 but each forward step costs no less than two fuel molecules, which turns out to be a general threshold21 for enzymatic nanomotors. Sustainable motion by any synthetic enzymatic motor has yet to be demonstrated because this walker relies on a soft, single-stranded DNA track that coils to halt the walker between nonadjacent binding sites. Six years after its publication, the motor of Bath et al. remains © 2016 American Chemical Society

RESULTS Motor Design and Fabrication. The walker is a DNA biped made of a 7 nm double-helix bridge connecting two identical single-stranded legs (Figure 1). The track is a double helix hosting identical binding sites, spaced ∼20 nm apart, with each site comprising two single-stranded segments ∼5 nm apart (D1* and D2*′, with the D2*′ to D1* pointing to a unique end of the track, henceforth termed “plus end”). Each leg of the motor can bind to a site by forming two helices (D1−D1* and D2−D2*′, ∼2 and 5 nm long, respectively). Should either helix Received: February 9, 2016 Accepted: June 13, 2016 Published: June 13, 2016 5882

DOI: 10.1021/acsnano.6b01035 ACS Nano 2016, 10, 5882−5890

Article

www.acsnano.org

Article

ACS Nano

Figure 2. Gel images of the fabricated motors and tracks from a native PAGE analysis. Lanes L are DNA ladders and lanes I−IV are the motors and tracks: Lane I for the motor (illustrated in Figure 1a) (I# for a motor mutant for binding bias measurement, I+ for an elongated motor variant, and I+# for the variant further mutated for binding bias measurement); lane II for the track (illustrated in Figure 1c); lanes III and IV, respectively, for truncated 3-site track and 2-site track; lane III# for a 3-site track mutant for binding bias measurement; and lane III* for the single-stranded template (180 nt) for the 3-site tracks.

leg (Figure S1). The bands for a preincubated motor−fuel mix were obtained without and with the enzyme. Without the enzyme, the motor−fuel binding complex and the fuel appear in separate bands, which is particularly true when the fuel− motor ratio is greater than one. After the enzyme is added, the motor−fuel complex band is reduced and the fuel band disappears. The two observations confirm the enzyme’s cutting of the fuel bound to the motor’s leg as well as the cyclic fuel turnover, namely, a leg-bound fuel is cut for spontaneous product dissociation that frees the leg for yet another fuel binding until exhaustion of the fuel supply. “Gated” Chemomechanical Coupling. The motor is designed to move directionally via three automatic gating mechanisms regulating leg-track interaction and fuel turnover. Following biomotor studies,23 “gating” means a physically coordinated stalling of one motor leg in a specific mechanochemical state until a certain “gatekeeper” process is completedoften at the other leg. Gating I. Fuel binding to a track-bound leg and the ensuing leg dissociation are gated by the other leg’s forward binding to the track. This gating is site-selective because the track-bound leg is protected in a single-leg motor-track binding state, but the short D1 segmenta recognition site for the fuelis exposed preferentially at the rear leg over the front leg in a two-leg state due to a mechanical asymmetry between the two legs. Due to its small size, the motor needs to be stretched to form a stable two-leg state, resulting in a nontrivial intramotor stress (see Figure 1d, state 1). The stress pulls the front leg’s D1−D1* and D2−D2*′ helices backward collinearly in a shearing mode, which is known to require a large force26 for duplex breaking, but the two helices are bent at the rear leg because its D2− D2*′ is pulled forward and the D1−D1* remains backward (state 4). This twists the rear leg into a looplike form, which requires minimally ∼4 nt27 in the single-stranded form due to DNA rigidity. Hence, the 6 bp long D1−D1* is readily broken for fuel binding at the rear leg but not the front leg. The rear leg twisting finds support in simulation using the oxDNA package28 (Figure S2), wherein a track-bound leg was simulated under forward or backward pulling forces of 5, 10, and 20 pN to mimic the intramotor stress on the rear or front leg. At all magnitudes of force, the simulation yields more D1−D1* breaking events for forward pull than backward pull (Figures S3 and S4). This mechanical asymmetry, which is between the two

Figure 1. Motor design. Asterisk (*) marks complementary sequences, “nt” for nucleotides and “bp” for base pairs. See the main text for explanation.

break, a single-stranded fuel can perform a toehold-mediated strand displacement to dissociate the leg from the track. The leg-bound fuel is then recognized and cut by nicking enzyme N.BbvC IB,6 which is prevented from cutting the track via a point mutation6 in D2*′. To study the motor’s performance and working mechanisms, we have fabricated the motor and several variants plus multiple tracks of different length, which include an extensive, 6-site track as well as truncated 3- and 2-site tracks. The 6-site track follows the design shown in Figure 1c, while the 3- and 2-site tracks follow a similar design except for use of a single-stranded template to form the track’s entire helical backbone hosting the binding sites. The tracks were self-assembled from their constituent DNA strands through a single-pot annealing at a low cooling rate (∼5.4 °C/h from 95 to 20 °C). Figure 2 shows the typical gel images of the assembled tracks and motors from native polyacrylamide gel electrophoresis (native PAGE). In almost every case, the annealing resulted in a single prominent band that can be identified as the targeted product by molecular weight comparison. For example, the motor band is slightly lower in molecular weight than an elongated variant; the track bands lie far above and follow the right order according to their lengths, and the track template is lower than the corresponding track product. A PAGE experiment also confirms that the nicking enzyme (N.BbvC IB) cuts the fuel when it is bound with the motor’s 5883

DOI: 10.1021/acsnano.6b01035 ACS Nano 2016, 10, 5882−5890

Article

ACS Nano

Sustainable Directional Motion. The motor’s sustainable motion was tested by operating it on a 6-site track, which is labeled with three dyes at the first site located at the minus end, the fourth site, and the sixth site at the plus end (Figure 1c). The motor carries two quenchers so that the leg-track binding subjects a dye to a ∼100% effective contact quenching.34 Adding the fuel and enzyme to a long incubated equimolar motor-track mix started the motor’s operation, which was monitored by fluorescence detection. Each operation is compared to a control experiment in which the same procedure of fuel/enzyme administration and dye excitation was applied to an equal amount of tracks without any motor. The fluorescence of the operated motor-track mix divided by that of the bare tracks yields site occupation by the motor (see Materials and Methods). This control-calibrated signal and the extracted occupation probability are shown in Figure 3a,b for motor operation on the 6-site track. The occupation increases most at the plus end, increases moderately at the intermediate site, and decreases at the minus end. The patterns indicate a processive motor flow across these three sites toward the plus end. The same trend is observed upon a second addition of fuel, demonstrating repeatability of the motor’s operation. The average occupation over the three sites reaches a final value after operation, which is close to the initial value (Figure S5), indicating a negligible derailment of the motor throughout the operation. The on-track flow and low derailment support the motor’s processivity and thereby gatings I and III. Stepping Fidelity and Speed. The motor was also tested on a truncated 3-site track with each site labeled by a dye (see the track design in Figure S11c). The motor’s direction and a low derailment are confirmed again (Figure 3c−f and Figure S6). Notably, the control-calibrated signal for the middle-site (hence the site’s occupation) recovers to the initial value quickly, especially for a low initial fuel concentration (Figure 3c,d). This feature suggests that the fluorescence decrease and increase at the plus and minus ends are caused primarily by a full-step, minus- to plus-end swing of the motor. This offers a chance to estimate the motor’s fidelity and speed for minus to plus translocation. Following the control-calibrated signal in Figure 3c, the plusend drop yields the net increase of leg population, which is ∼0.1 at the end of operation. The initial motor population bound between the minus-end site and the middle site is the initial minus-end value of the control-calibrated signal subtracted from one (∼0.68). Since the motor’s actual fuel consumption is capped by the fuel supply (i.e., one fuel molecule per four motors), the fidelity of a motor making a fullstep translocation to the plus end, per fuel molecule consumed, is at least D = 0.1/(0.68 × 1/4) = 58.8%. The measurement of the fuel versus step ratio is made possible by taking advantage of the fully labeled 3-site track plus the controlled fuel supply in the ensemble fluorescence experiments. We note that a direct measurement of this ratio by single-molecule experiments remains a challenge for the technical difficulty to simultaneously detect a motor’s stepping and fuel consumption. In fact, the previous single-molecule studies19,20 reporting nearly one fuel per step by biomotor kinesin did not reach the conclusion by counting fuels versus steps but inferred it merely from the stepping data through stochastic modeling. The motor’s rate for overall leg dissociation immediately following fuel addition rises linearly with the initial fuel concentration (Figure 4a), consistent with fuel-induced leg dissociation. The minus to plus translocation probability (i.e., net increase of the plus-end population divided by the initial minus-end

identical legs bound to two identical sites, can be traced back to the local asymmetry inside each site made of two different overhangs (D2*′, D1*). The intramotor stress pulls the rear leg forward but the front leg back so that the two legs experience differing mechanical response from the binding sites. Such a track-motor design plus the DNA rigidity at short length amplifies the intrasite asymmetry into interleg binding asymmetry for gating I. Gating II. Leg-track binding is gated by controlled product releaseagain with a site-selectivity that ensures preferential leg binding to the front site over the back site. The fuel’s cutting by the nicking enzyme is so arranged that a 5 nt long product (P1 in Figure 1d) and a 7 nt long product (P2) cover the leg’s D2 segment, and a third, 8 nt long product (P3) covers the D1 segment plus the two adjacent nucleotides of D2. The hybridization free energy for the products and the leg is ∼8.6, 15.2, 12.7kBT for P1, P2, and P3, respectively, as predicted from the DNA nearest-neighbor thermodynamics29 for the experimental temperature (T = 25 °C, kB is the Boltzmann constant). The free-energy gaps between the duplexes suggest that the shortest P1 is released from the leg by thermal fluctuations before P2 and P3, on average. The diffusing leg free of P1 but still retaining the longer products can hybridize with the D2*′ overhang of the front site but not the equally distant D1* at the back site, while the back site’s D2*′ is ∼5 nm further than the front site’s D2*′ from the track-bound leg (counted from the leg’s D1−D2 junction; see state 3). Hence the diffusing leg binds preferentially forward. Gating III. Full product release is gated by consolidated forward leg binding. The two longer products (P2, P3) are actively displaced as the D2*′ and D1* overhangs complete their hybridization with the incoming leg. This suppresses premature fuel turnover at the diffusing leg independent of the motor’s motion, thereby prohibiting uncontrolled decay of the chemical energy into heat. The three gatings ensure a coordinated chemomechanical cycle that facilitates autonomous and sustainable walking of the motor toward the plus end. Gating I causes an asymmetric state (state 1) in which the fuel preferentially dissociates the rear leg instead of the protected front leg (by D1* or fuel products). In the ensuing single-leg state (states 2 and 3), the diffusing leg undergoes fuel cutting and binds preferentially to the front site after spontaneous P1 release (gating II). The forward binding creates a new two-leg state in which the previous front leg becomes a twisted rear leg (state 4). The complete D2−D2*′ hybridization at the front leg raises the intermotor stress to expose the rear leg for fuel recognition, bringing the motor back to the initial state 1 after P2 and P3 dissociation. The full chemomechanical cycle i → ii → iii → iv → i consumes one fuel molecule and translocates the motor a step to the plus end. The motor’s processive stepping is ensured by gating I that suppresses fuel-induced derailment from a single-leg state. Gatings I and III combine to delay the enzymatic cycle of the front leg relative to the rear lega feature called “alternating catalysis” for biomotors30 and proven crucial for processivity of bipedal enzymatic motors. The motor’s sustainable walking is further supported by the double-stranded track because the motor’s step size (∼20 nm) is well within the track’s persistence length (∼50 nm, as the track retains a virtually perfect helical structure despite the presence of nicks, as found by previous studies31−33). Hence the track, being flexible over a larger scale though, behaves like a rigid rod at a “zoom-in” scale that is really relevant to stepping of the small motor. 5884

DOI: 10.1021/acsnano.6b01035 ACS Nano 2016, 10, 5882−5890

Article

ACS Nano

Figure 3. Motor operation on a 6-site track (a,b) and a truncated 3-site track (c−f). The control-calibrated fluorescence in the top panels (a), (c), and (e) are the fluorescence from an operation experiment with a motor-track mix divided by the fluorescence from a track-only control experiment (with the same amount of track and the same procedure of fuel/enzyme addition). The occupation in the bottom panels (b), (d), and (f) is extracted from the control-calibrated signal in the top ones. The data shown are obtained for 1:1 fuel−motor ratio per fuel addition, 40 nM motor/track, 560 nM enzyme for 6-site operation and 80 nM motor/track, 560 nM enzyme for 3-site operation. The fuel addition is done twice for the 6-site operation under the consideration that the average number of forward steps per motor is only two from a rough estimation assuming equal probability for the motor’s different initial binding positions along the 6-site track.

release (and leg binding) might possibly occur before full dissociation of the enzyme from a leg because a single enzyme likely catalyzes both cuttings by a sliding mechanism36 due to the close proximity of the two nicking sites. Two Directional Biases and Their Dependence on Fuel and Enzyme. The preferential fuel binding to the rear leg and the resultant leg dissociation bias (i.e., gating I) are verified by the faster rise of the control-calibrated fluorescence at the minus end compared to that at the plus end shortly after the fuel addition, which is a common pattern of the 6-site operation and the 3-site operation (Figure 3). The same pattern was also observed in a fuel-only dissociation experiment using a truncated track with two binding sites (see the track design in Figure S11). In this experiment, the fuel was added to a preincubated equimolar motor-track mix, and the fluorescence from the dyes labeled at the track’s two sites was recorded. Unlike the 3-site and 6-site operation driven by the fuel plus enzyme, the fuel-only dissociation experiment shows a monotonic fluorescence increase signifying the fuel-induced leg dissociation in the absence of the enzyme. The fluorescence from the dissociation experiment divided by that from the pure track yields the occupation at the plus and minus ends, which in turn yields the dissociation rate ratio between the motor’s rear leg and front leg (see Materials and Methods). The minus to plus dissociation rate ratios deduced from the 2-site dissociation experiment and from the 3-site and 6-site operation experiments are invariably above one, consistent with gating I. The highest ratio is above 100 (Figure 5). The rate ratios scale linearly with the site-fuel ratio for all three types of experiments (Figure 5d), which explains the motor’s speed scaling. The scaling is a feature of the ensemble experiments in which two-leg motor-track states occur over the entire track to share the fuel supply, hence reducing available fuels at individual sites. To study the leg binding bias, we introduced a motor mutant in which one D2 segment is mutated to exclusively bind the

Figure 4. Motor translocation and speed extracted from operation on a fully labeled 3-site track. The data are for 80 nM motor/track, 560 nM enzyme unless stated otherwise. Rates in (a) are obtained by summing slopes of three dyes in Figure 3c,e from the last data point before the fuel addition to the first data point after. Translocation in (b) is for 1:1 (greenish blue) and 1:4 fuel−motor (orange). The lines in (a,d) are linear fits.

population) follows a Michaelis−Menten-like saturation with increasing enzyme concentration (Figure 4b). The speed of the minus to plus translocation drops quickly with fuel consumption (Figure 4c). The peak speed scales linearly with the number ratio of binding sites over supplied fuel molecules (Figure 4d). The highest detected speed is ∼3 nm per minute, which is the same magnitude as that previously achieved by burn-bridge nanowalkers.11,14 The present motor is limited by the enzyme turnover. The reported turnover rate35 of ∼0.0037/s caps the motor’s speed to ∼4.5 nm per minute at best, which is 50% higher than the achieved speed. The low turnover rate is due to the enzyme’s slow dissociation,35 but the fuel product 5885

DOI: 10.1021/acsnano.6b01035 ACS Nano 2016, 10, 5882−5890

Article

ACS Nano

was first mixed with the fuel at a saturating concentration to cover the native leg; the mutated track was then added to bind the motor’s mutated leg at the middle site. Adding enzyme triggers the native leg’s binding to either the plus or the minus end. This enzyme-induced leg binding experiment versus a corresponding track-only control yields the control-calibrated fluorescence for the plus and minus ends and thereby the forward to backward binding rate ratio of the enzyme-liberated leg (see Materials and Methods). The control-calibrated fluorescence decays faster at the plus end than at the minus end (Figure 6b), verifying the forward-biased binding and thereby gating II. The forward to backward binding rate ratio is as high as 50 shortly after the enzyme addition and decreases quickly with decreasing single-leg population (Figure 6c). The binding bias is also reflected16 in the different fluorescence decay at the plus and minus ends in the normal 3-site operation; the extracted rate ratio has similar magnitude (Figure 6e). The binding rate ratio versus enzyme concentration follows the same Michaelis− Menten-like pattern as for the translocation probability (Figure 6f). More Experimental Evidence of Asymmetric Leg Binding. The motor’s rear leg is more exposed to the fuel than the front leg, as has become clear from the observed preferential rear leg dissociation in the 3-site/6-site operation experiments as well as in the fuel-only dissociation experiment on a 2-site track. To further elucidate whether the exposed leg segment is D1 or D2, we delete the D1* segment from the fuel for a control experiment. The incomplete fuel (now with the D2* segment only) does not affect the motor-track binding in the case of D1 exposure but should cause a fluorescence increase if the D2 segment is exposed instead because the fuel

Figure 5. Bias for rear leg dissociation. (a−c) Rate ratios extracted from 6-site operation (1:4 fuel−motor), 3-site operation (1:8) in Figure 3, and from a specially designed 2-site dissociation experiment (1:4 fuel−motor, 100 nM motor/track). Unlike the operation experiments involving both fuel and enzyme, the dissociation experiment is a fuel-only experiment in which adding fuel triggers the motor’s leg dissociation from a 2-site track in the absence of the enzyme. (d) Average ratio over the first 2 min after fuel addition for the three types of experiments. The left vertical axis is for the 2-site and 3-site data and the right axis for the 6-site data. The lines are linear fits to each type of data.

middle site (mutated too accordingly) of a 3-site track (see a schematic illustration in Figure 6a). This heteropedal motor

Figure 6. Bias for forward leg binding. (a) Schematic illustration of a specially designed leg binding experiment starting from a single-leg state in the middle of a 3-site track (by mutating the track’s middle site and the motor’s leg-bound leg). Adding enzyme liberates the fuel-protected leg for its binding forward or backward to the exposed sites. (b,c) Fluorescence and rate ratio from the leg binding experiment (80 nM motor/ track, 560 nM enzyme). (d) Rate ratio from the leg binding experiments for an elongated motor (80 nM motor/track, 560 nM enzyme) and an elongated motor with further fuel/leg mutations to lose product control (83 nM motor/track, 325 nM enzyme). Compared to the native motor illustrated in Figure 1a, both of the motor variants have a 30 bp duplex between its two legs. The first variant has two native legs for normal operation (as for (f)), but for the leg binding experiment, one leg is replaced by the mutated leg designed for the experiment. The second motor variant, designed just for the leg binding experiment, has its second leg further mutated to be protected by a new mutated fuel for the purpose of making P3 the shortest product (instead of P1 as shown in Figure 1d). The two exposed sites in (a) are also mutated accordingly for leg binding after the new fuel is cut by the enzyme. (e,f) Rate ratio from 3-site operation as in Figure 3 but for 3:4 fuel−motor and averaged ratio over the first 2 min after the rate ratio stabilizes to positive values. Squares are for 3-site operation (1:1 fuel−motor); diamond for the leg binding experiment in (c). 5886

DOI: 10.1021/acsnano.6b01035 ACS Nano 2016, 10, 5882−5890

Article

ACS Nano

Leg Orientation and Binding Bias. The oxDNA simulation suggests a ∼23° tilting of the track-bound leg toward the minus end from the track’s normal (Figure S10). The back-tilting is expected because the adjoining D1−D1* and D2−D2*′ helices tend to stack onto each other, resulting in a backward displacement of the top end of the D2−D2*′ helix. However, the simulation also found that a forward force of merely 2 pN applied at the helix’s top end reverses this back-tilting to a ∼9° forward tilting. The 2 pN force results in a ∼2.6 nm forward displacement of the D2−D2*′ top, on average, indicating a low barrier of less than 1.3kBT for the transition from the backward to forward tilting. These simulation results suggest that the leg’s back-tilting is readily reversed by thermal fluctuations, in line with the experimental data of the third motor variant. The back-tilting of the trackbound leg implies the presence of an intrinsic bias for the other leg to bind backward, which is consistent with the slightly minus-end-directed bias observed in the variant of reversed product control, but the influence of the back-tilting on the leg binding is overcome by the product control, as proven by the observed forward binding bias in the native motor as well as in the two motor variants retaining the control. Under the product control, the thermally susceptible backtilting does not change the diffusing leg’s preferential binding to the front D2*′ over the back one because the motor, due to its small size, hardly reaches either D2*′ at the tilting angle but must fall down to the track in an almost parallel configuration to form the first 5 bp duplex with a D2*′ over its P1* segment. This first contact requires the remaining 18 nt long segment of the D2*′−S overhang to extend ∼4.6 or ∼9.7 nm, respectively, for forward or backward hybridization, resulting in a higher barrier for the latter than for the former. The barrier difference is ∼10−16kBT from a rough estimation using a worm-like chain formula, that is, eq 2 of ref 37, with the previously reported contour length per nucleotide of ∼0.63−0.70 nm38,39 and persistence length of ∼1 nm38,40 for single-stranded DNA.

then can invade and open the D2−D2*′ duplex to remove the contact quenching between the dye at the tip of the D2*′ overhang and the quencher at the motor’s duplex bridge (see panel 1 of Figure 1d for the D2−D2*′ duplex in leg-track binding and the dye/quencher labeling in Figure 1a,c). The loss of contact quenching increases the fluorescence, especially at the rear leg even if it, by any chance, is retained on the track by the D1−D1* duplex. This is because the rear leg’s release from the D2−D2*′ duplex allows the intramotor tension to pull the motor bridge forward away from the D2*′ overhang to separate the quencher from the dye. Moreover, a leg with its D2 segment covered by the incomplete fuel instead of the D2*′ overhang readily dissociates from the track to further increase fluorescence because the remaining 6 bp D1−D1* duplex can spontaneously break by thermal fluctuation. In the control experiment, the incomplete fuel is added to a preincubated mix of the motor and 3-site track. Unlike the 3-site operation in Figure 3, the incomplete fuel fails to induce a fluorescence increase (Figure S7). This observation indicates that the exposed leg segment is not D2 but D1. The conclusion is further supported by a second control in which the motor’s binding to a 2-site track is monitored by recording the fluorescence before and after the motor-track mixing (Figure S8). More than 90% quenching is observed for the dyes at both sites, suggesting that the motor’s two legs both have their D2 segments fully protected within the D2−D2*′ duplex. Again, what the motor exposes at the rear leg must be the D1 segment. Hence, the two control experiments confirm the preferential D1 exposure at the rear leg over the front leg predicted from the oxDNA simulation. Motor Variants Testing Size Dependence and Product Control. Three motor variants further elucidate the motor’s working mechanisms. In the first variant, the duplex bridge is elongated to 30 bp. The second variant has the elongated bridge plus mutations in the fuel and related motor-track segments to reverse the product control: P3 becomes the shortest (6 nt) versus a 7 nt long P1 now and the same P2 as for the native motor. The hybridization free energy is ∼7.6, 15.2, and 13.7kBT for P3, P2, and P1 from the nearest-neighbor thermodynamics,29 suggesting a preferential dissociation of P3 instead of P1 on average and hence the possibility of a backward leg binding to the D1* overhang. The first variant has a lower forward binding bias than the native motor (Figure 6f), as the longer bridge allows more access to the rear D2*′ overhang by the diffusing leg. For the variant with a reversed product control, the forward bias is further deteriorated and slightly reverses direction (Figure 6d). The bias difference between the two variants is experimental evidence for the effectiveness of the product control and its critical role in generating the forward binding bias. The third motor variant introduces more flexibility into the two bridge-leg junctions by inserting four extra nucleotides (TTTT) at either junction. This variant shows no improvement of performance (Figure S9), suggesting that the mechanical flexibility necessary for the diffusing leg’s binding to the track is not from within the motor but from the track-bound leg. The oxDNA simulation indeed finds a broad orientation distribution of the D2−D2*′ duplex of the track-bound leg under zero pulling force at the operation temperature (Figure S10). The flexibility is due to the soft 9 nt long S segment as well as spontaneous breaking and re-formation of the short D1−D1* duplex as found in the simulation.

DISCUSSION Fuel Efficiency. The measured dissociation rate ratios are essentially the probability ratio for the motor’s rear leg dissociation over the front leg dissociation (α). Likewise, the binding rate ratios are the probability ratio for the dissociated leg to bind the front site over the back site (β). The measured stepping fidelity is the previously proposed concept of directional fidelity,21,41 which is defined as the probability for a motor’s forward step minus that for a backward step divided by the total probability for forward, backward, and futile steps. The directional fidelity D can be counted from α and β ratios by a simple stepping statistics: the probabilities for forward, backward, and futile steps are [α/(1 + α)] × [β/(1 + β)], [1/(1 + α)] × [1/(1 + β)], and [α/(1 + α)] × [1/(1 + β)] + [1/(1 + α)] × [β/(1 + β)] (two terms for futile forward or backward step returning the motor to its previous location). Hence, D = (αβ − 1)/[(α + 1)(β + 1)], indicating symmetric and complementary roles of the dissociation and binding biases in producing a motor’s direction. With either bias missing (i.e., α = 1 or β = 1), a motor is capped to D ≤ 50%, and on average, two fuel molecules or more must be decomposed per forward step. This important limit, found general for enzymatic nanomotors by previous studies,21 is now broken by our motor: the measured biases are up to α ∼ 100 and β ∼ 50, potentially affording a D well above the experimentally deduced lower limit of ∼60%. 5887

DOI: 10.1021/acsnano.6b01035 ACS Nano 2016, 10, 5882−5890

Article

ACS Nano Previous studies42,43 found that the second law of thermodynamics requires a least energy price for directional fidelity of isothermal nanomotors. A biomotor with ∼100% efficiency evidently42,43 exhausts the chemical energy to produce high directional fidelity at the second law-decreed least price, with chemical energy to work conversion resulting from loadreduced direction. Hence, a vital capability of efficient nanomotors is the effective channeling of fuel energy into directional fidelity, which can be achieved mechanistically via complementary dissociation and binding biases. The present motor powers both biases by a single fuel molecule through three gating mechanisms at major energy-releasing stages of the fuel turnover cycle, resulting in a controlled release of the chemical energy and its effective use for the motor’s directional motion before decaying into random heat. The motor is likely among the most fuel-efficient machines ever invented to operate on chemical energy from fuel decomposition and differs scientifically from macroscopic machines that burn a massive quantity of fuel molecules into heat and then produce motion or work from the heat. Product Control Enabling Directional Bias as a Unique Mechanism for Chemically Powered Nanomotors. The control over which product to be first released from the fuel-consuming leg modulates the rate for the leg to form the first contact with the front or back sites to differing extents, thereby resulting in the leg’s preferential forward binding. This is a key to break the two fuel per step threshold as the product control reverses an intrinsic bias that is otherwise opposite to the leg dissociation bias. Similar product controls exist in bipedal biomotors consuming ATP (adenosine triphosphate) as the fuel: the earlier release of the small product phosphate than the sizable ADP (adenosine diphosphate) serves as a “nucleotide gating”23 or “checkpoint”44 conducive to the directional motion of kinesin and myosin V; a phosphatedependent switch in cytoplasmic dynein reverses45 this motor’s direction entirely. The bias-generating product control is a mechanism unique to chemically powered nanomotors but unseen in motors driven by light or electric/magnetic fields. The product control-enabled bias between identical binding sites demonstrated in this study is further beyond chemical motors of a burn-bridge design but is pertinent to genuine enzymatic nanomotors that draw chemical energy from fuels outside the motor-track system. We note that a product control similar to the present one was previously suggested46 to improve the autonomous motor of Bath et al.6

knowledge). For example, this study may be extended to collective transport by a tandem of motors along a long track, which is demonstrated routinely by biomotors but is impossible for burn-bridge motors. Single-molecule characterization of the motorpreferably on a two-dimensional DNA origami platform11,17 hosting a stiffer longer track and better control of the motor’s initial positionis a worthy target for future study, too, which may yield more clues beyond the present ensemble measurement to further improve the motor’s performance in speed and fuel efficiency.

MATERIALS AND METHODS Motor−Track Fabrication. The nucleotide sequences for the double-stranded backbone of tracks were taken from λDNA; the sequences for the motor and other parts of the tracks were generated using Canada software.47 The secondary structures were checked using Mfold48 (for single strands) and NUPACK49 (for motor/track formation), with unwanted structures removed by manually adjusting the sequences. The final sequences of the strands are given in online Supporting Information. The motors and tracks were fabricated separately using constituent DNA strands (from Integrated DNA Technologies, Inc.). The strands were mixed stoichiometrically at 5 μM in TE buffer containing 10 mM Tris-HCl, 1 mM EDTA, and 200 mM NaCl (or 400 mM NaCl for 6-site tracks). The mixed samples were incubated in a heating block and gradually cooled from 95 to 20 °C over a period of 14 h using polymerase chain reaction (C1000 Touch Thermal Cycler, Bio-Rad). The final products were analyzed using 10% native polyacrylamide gel electrophoresis (PAGE) against a low molecular weight DNA ladder (New England BioLabs, Inc.). Motor Operation and Fluorescence Detection. For motor operation, an equimolar mix of motor/track was prepared at submicromolar concentrations in a buffer containing 40 mM Tris, 20 mM acetic acid, 2 mM EDTA, and 12.5 mM magnesium acetate (pH 8.0). The mixed sample was incubated for at least 15 h to ensure thermodynamic equilibration before operation. A fuel−enzyme mixture was separately prepared from 100 μM stock solution for the fuel (TE buffer) and 3.25 μM N.BbvC IB enzyme with a specific activity of 48 000 units/mg (New England BioLabs, Inc.). The motor operation was started by quickly mixing the fuel−enzyme with the incubated motor-track sample and was monitored by detecting fluorescence from track-tethered dyes with a Cary eclipse spectrophotometer (Varian, Inc.; kinetic mode, at excitation/emission wavelength of 495 nm/520 nm (FAM), 549 nm/563 nm (TYE), and 648 nm/668 nm (CY5)). The sample incubation and later operation were all done at 25 °C. Adding the fuel−enzyme and their associated buffer to the motor-track sample affects the optical properties of the dyes. This effect is removed by dividing the fluorescence from an operation experiment by that from a control experiment on the track sample only under the same procedure of fuel−enzyme addition and fluorescence detection. Such a control-calibrated signal also removes16 the influence of photobleaching (a negligible effect in this study as confirmed by a control experiment; see Figures S7 and S8). Extracting Occupation Probability, Translocation Probability, Speed, and Rate Ratios from the Control-Calibrated Fluorescence Data. Following previous studies,15,16 the probability for a site to be occupied by a motor is related to the fluorescence of the dye tethered to the site as P(t) = [1 − IM(t)]/γ. Here, IM(t) = I MT (t)/I T(t) is control-calibrated fluorescence, IMT(t) is the fluorescence collected from an operated motor-track sample at a time t, IT(t) is the fluorescence of an equal amount of bare tracks from the accompanying control experiment, and γ is the quenching efficiency of the dye by the motor-carried quencher; γ ≈ 1 holds for all the three dyes due to their ∼100% effective contact quenching.34 The change of occupation probability from time t1 to a later time t2 is ΔP = P(t2) − P(t1) = [IM(t1) − IM(t2)]/γ. The rate for leg dissociation off a site from t1 to t2 is rd = −ΔP/(t2 − t1) = [IM(t2) − IM(t1)]/γ(t2 − t1). For motor operation on a fully labeled 3-site track, the rate for overall leg dissociation off track (shown in Figure 4a) is rd,tot = rd+ + rd− + rdm, in

CONCLUSION In summary, this study achieves an efficient artificial enzymatic nanowalker and yields rich mechanistic insights into how pure physical effects enable effective harvest of chemical energy at the single-molecule level. Like biomotors, this rationally designed autonomous motor realizes multiple chemomechanical gating mechanisms but in a conceptually transparent way, representing an important step toward replicating key mechanistic characteristics of efficient biomotors. A long-term goal of nanotechnology is to achieve in designed systems the efficient, repeatable, automatic, and long-range transport seen in biomotor-based intracellular transport but beyond the capacity of current burn-bridge motors. This motor, with its demonstrated capability of autonomous and sustainable motion along a double-stranded track, provides a rarely available candidate nanowalker for future development in this direction (the only other available candidate is an early walker from ref 6 to our 5888

DOI: 10.1021/acsnano.6b01035 ACS Nano 2016, 10, 5882−5890

Article

ACS Nano which the subscripts + , −, and m mark the plus-end site, minus-end site, and the middle site. The probability for a motor initially at the minus end to translocate to the plus end (shown in Figure 4b) is Ptrans = ΔP+/P(t1) = [IM+(t1) − IM+(t2)]/[1 − IM−(t1)]. The average speed of the motor’s fuel-driven translocation is v = Ptrans × d/(t2 − t1) nf, in which d is the motor’s step size and nf is the number of fuel molecules consumed per motor. The v data shown in Figure 4c are a lower limit obtained using the supplied fuel molecules per motor, which is an upper limit of nf. The rate ratio for leg dissociation from the minus-end site over the plus-end site is α = rd−/rd+ = (γ+/γ−) × [IM−(t2) − IM−(t1)]/[IM+(t2) − IM+(t1)] (Figure 5). Similarly, the rate for leg binding from t1 to t2 is rb± = [P±(t2) − P±(t1)]/(t2 − t1) = [IM±(t1) − IM±(t2)]/γ(t2 − t1), and the rate ratio for leg binding to the plus-end site over the minus-end site is β = rb+/rb− = (γ−/γ+) × [IM+(t1) − IM+(t2)]/[IM−(t1) − IM−(t2)] (Figure 6). Computer Simulation of Motor-Track Binding. The oxDNA coarse-grained simulation package28 was used to simulate a leg (i.e., the 20 nt long D1−D2 segment) bound with the middle site of a 3-site track at the experimental temperature (25 °C). To exert a pulling force to the leg, we applied two harmonic traps ∼12 nm apart to immobilize a 35 bp segment of the track that sandwiches the 15 bp binding site in the middle. Following ref 28, a Monte Carlo simulation was done to sample base-pair breaking events at the leg−site complex under a forward or backward force, which is parallel to the track segment and placed at the other end of the D2−D2*′ duplex to model the intramotor stress in a two-leg state. Monte Carlo simulations of ∼108 steps were performed for each value of force.

(7) Omabegho, T.; Sha, R.; Seeman, N. C. A Bipedal DNA Brownian Motor with Coordinated Legs. Science 2009, 324, 67−71. (8) von Delius, M.; Geertsema, E. M.; Leigh, D. A. A Synthetic Small Molecule That Can Walk down a Track. Nat. Chem. 2010, 2, 96−101. (9) Lund, K.; Manzo, A. J.; Dabby, N.; Michelotti, N.; Johnson-Buck, A.; Nangreave, J.; Taylor, S.; Pei, R.; Stojanovic, M. N.; Walter, N. G.; Winfree, E.; Yan, H. Molecular Robots Guided by Prescriptive Landscapes. Nature 2010, 465, 206−210. (10) He, Y.; Liu, D. R. Autonomous Multistep Organic Synthesis in a Single Isothermal Solution Mediated by a DNA Walker. Nat. Nanotechnol. 2010, 5, 778−782. (11) Wickham, S. F. J.; Endo, M.; Katsuda, Y.; Hidaka, K.; Bath, J.; Sugiyama, H.; Turberfield, A. J. Direct Observation of Stepwise Movement of a Synthetic Molecular Transporter. Nat. Nanotechnol. 2011, 6, 166−169. (12) You, M.; Chen, Y.; Zhang, X.; Liu, H.; Wang, R.; Wang, K.; Williams, K. R.; Tan, W. An Autonomous and Controllable LightDriven DNA Walking Device. Angew. Chem., Int. Ed. 2012, 51, 2457− 2460. (13) Cheng, J.; Sreelatha, S.; Hou, R.; Efremov, A.; Liu, R.; van der Maarel, J. R. C.; Wang, Z. Bipedal Nanowalker by Pure Physical Mechanisms. Phys. Rev. Lett. 2012, 109, 238104. (14) Cha, T.-G.; Pan, J.; Chen, H.; Salgado, J.; Li, X.; Mao, C.; Choi, J. H. A Synthetic DNA Motor That Transports Nanoparticles along Carbon Nanotubes. Nat. Nanotechnol. 2014, 9, 39−43. (15) Liu, M.; Hou, R.; Cheng, J.; Loh, I. Y.; Sreelatha, S.; Tey, J. N.; Wei, J.; Wang, Z. Autonomous Synergic Control of Nanomotors. ACS Nano 2014, 8, 1792−1803. (16) Loh, I. Y.; Cheng, J.; Tee, S. R.; Efremov, A.; Wang, Z. From Bistate Molecular Switches to Self-Directed Track-Walking Nanomotors. ACS Nano 2014, 8, 10293−10304. (17) Liber, M.; Tomov, T. E.; Tsukanov, R.; Berger, Y.; Nir, E. A Bipedal DNA Motor That Travels Back and Forth between Two DNA Origami Tiles. Small 2015, 11, 568−575. (18) Toyabe, S.; Watanabe-Nakayama, T.; Okamoto, T.; Kudo, S.; Muneyuki, E. Thermodynamic Efficiency and Mechanochemical Coupling of F1-ATPase. Proc. Natl. Acad. Sci. U. S. A. 2011, 108, 17951−17956. (19) Schnitzer, M. J.; Block, S. M. Kinesin Hydrolyses One ATP per 8-Nm Step. Nature 1997, 388, 386−390. (20) Hua, W.; Young, E. C.; Fleming, M. L.; Gelles, J. Coupling of Kinesin Steps to ATP Hydrolysis. Nature 1997, 388, 390−393. (21) Efremov, A.; Wang, Z. Maximum Directionality and Systematic Classification of Molecular Motors. Phys. Chem. Chem. Phys. 2011, 13, 5159−5170. (22) Vale, R. D. Myosin V Motor Proteins: Marching Stepwise towards a Mechanism. J. Cell Biol. 2003, 163, 445−450. (23) Gennerich, A.; Vale, R. D. Walking the Walk: How Kinesin and Dynein Coordinate Their Steps. Curr. Opin. Cell Biol. 2009, 21, 59− 67. (24) Wang, Z.; Feng, M.; Zheng, W.; Fan, D. Kinesin Is an Evolutionarily Fine-Tuned Molecular Ratchet-and-Pawl Device of Decisively Locked Direction. Biophys. J. 2007, 93, 3363−3372. (25) Xu, Y.; Wang, Z. Comprehensive Physical Mechanism of TwoHeaded Biomotor Myosin V. J. Chem. Phys. 2009, 131, 245104. (26) Kufer, S. K.; Puchner, E. M.; Gumpp, H.; Liedl, T.; Gaub, H. E. Single-Molecule Cut-and-Paste Surface Assembly. Science 2008, 319, 594−596. (27) Varani, G. Exceptionally Stable Nucleic Acid Hairpins. Annu. Rev. Biophys. Biomol. Struct. 1995, 24, 379−404. (28) Šulc, P.; Romano, F.; Ouldridge, T. E.; Rovigatti, L.; Doye, J. P. K.; Louis, A. A. Sequence-Dependent Thermodynamics of a CoarseGrained DNA Model. J. Chem. Phys. 2012, 137, 135101. (29) SantaLucia, J. A Unified View of Polymer, Dumbbell, and Oligonucleotide DNA Nearest-Neighbor Thermodynamics. Proc. Natl. Acad. Sci. U. S. A. 1998, 95, 1460−1465. (30) Hackney, D. D. Evidence for Alternating Head Catalysis by Kinesin during Microtubule-Stimulated ATP Hydrolysis. Proc. Natl. Acad. Sci. U. S. A. 1994, 91, 6865−6869.

ASSOCIATED CONTENT S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsnano.6b01035. Extra data, DNA sequences, and supplementary Figures S1−S11 (PDF)

AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected]. Author Contributions §

M.L. and J.C. contributed equally.

Notes

The authors declare no competing financial interest.

ACKNOWLEDGMENTS This work is supported by grants to Z.W. from the Ministry of Education of Singapore under R-144-000-290-112, R-144-000325-112, and R-144-000-320-112. REFERENCES (1) Sherman, W. B.; Seeman, N. C. A Precisely Controlled DNA Biped Walking Device. Nano Lett. 2004, 4, 1203−1207. (2) Shin, J.-S.; Pierce, N. A. A Synthetic DNA Walker for Molecular Transport. J. Am. Chem. Soc. 2004, 126, 10834−10835. (3) Yin, P.; Yan, H.; Daniell, X. G.; Turberfield, A. J.; Reif, J. H. A Unidirectional DNA Walker That Moves Autonomously along a Track. Angew. Chem., Int. Ed. 2004, 43, 4906−4911. (4) Bath, J.; Green, S. J.; Turberfield, A. J. A Free-Running DNA Motor Powered by a Nicking Enzyme. Angew. Chem., Int. Ed. 2005, 44, 4358−4361. (5) Tian, Y.; He, Y.; Chen, Y.; Yin, P.; Mao, C. A DNAzyme That Walks Processively and Autonomously along a One-Dimensional Track. Angew. Chem., Int. Ed. 2005, 44, 4355−4358. (6) Bath, J.; Green, S. J.; Allen, K. E.; Turberfield, A. J. Mechanism for a Directional, Processive, and Reversible DNA Motor. Small 2009, 5, 1513−1516. 5889

DOI: 10.1021/acsnano.6b01035 ACS Nano 2016, 10, 5882−5890

Article

ACS Nano (31) Hagerman, P. J. Flexibility of DNA. Annu. Rev. Biophys. Biophys. Chem. 1988, 17, 265−286. (32) Aymami, J.; Coll, M.; van der Marel, G. A.; van Boom, J. H.; Wang, A. H.; Rich, A. Molecular Structure of Nicked DNA: A Substrate for DNA Repair Enzymes. Proc. Natl. Acad. Sci. U. S. A. 1990, 87, 2526−2530. (33) Zhang, Y.; Crothers, D. M. High-Throughput Approach for Detection of DNA Bending and Flexibility Based on Cyclization. Proc. Natl. Acad. Sci. U. S. A. 2003, 100, 3161−3166. (34) Marras, S. A. E.; Kramer, F. R.; Tyagi, S. Efficiencies of Fluorescence Resonance Energy Transfer and Contact-Mediated Quenching in Oligonucleotide Probes. Nucleic Acids Res. 2002, 30, e122. (35) Bellamy, S. R. W.; Milsom, S. E.; Scott, D. J.; Daniels, L. E.; Wilson, G. G.; Halford, S. E. Cleavage of Individual DNA Strands by the Different Subunits of the Heterodimeric Restriction Endonuclease BbvCI. J. Mol. Biol. 2005, 348, 641−653. (36) Gowers, D. M.; Wilson, G. G.; Halford, S. E. Measurement of the Contributions of 1D and 3D Pathways to the Translocation of a Protein along DNA. Proc. Natl. Acad. Sci. U. S. A. 2005, 102, 15883− 15888. (37) Wang, Z. Synergic Mechanism and Fabrication Target for Bipedal Nanomotors. Proc. Natl. Acad. Sci. U. S. A. 2007, 104, 17921− 17926. (38) Smith, S. B.; Cui, Y.; Bustamante, C. Overstretching B-DNA: The Elastic Response of Individual Double-Stranded and SingleStranded DNA Molecules. Science 1996, 271, 795−799. (39) Murphy, M. C.; Rasnik, I.; Cheng, W.; Lohman, T. M.; Ha, T. Probing Single-Stranded DNA Conformational Flexibility Using Fluorescence Spectroscopy. Biophys. J. 2004, 86, 2530−2537. (40) Saleh, O. A.; McIntosh, D. B.; Pincus, P.; Ribeck, N. Nonlinear Low-Force Elasticity of Single-Stranded DNA Molecules. Phys. Rev. Lett. 2009, 102, 068301. (41) Efremov, A.; Wang, Z. Universal Optimal Working Cycles of Molecular Motors. Phys. Chem. Chem. Phys. 2011, 13, 6223−6233. (42) Wang, Z.; Hou, R.; Efremov, A. Directional Fidelity of Nanoscale Motors and Particles Is Limited by the 2nd Law of thermodynamicsvia a Universal Equality. J. Chem. Phys. 2013, 139, 035105. (43) Hou, R.; Wang, Z. Role of Directional Fidelity in Multiple Aspects of Extreme Performance of the F1-ATPase Motor. Phys. Rev. E 2013, 88, 022703. (44) Kad, N. M.; Trybus, K. M.; Warshaw, D. M. Load and Pi Control Flux through the Branched Kinetic Cycle of Myosin V. J. Biol. Chem. 2008, 283, 17477−17484. (45) Walter, W. J.; Koonce, M. P.; Brenner, B.; Steffen, W. Two Independent Switches Regulate Cytoplasmic Dynein’s Processivity and Directionality. Proc. Natl. Acad. Sci. U. S. A. 2012, 109, 5289−5293. (46) Ouldridge, T. E.; Hoare, R. L.; Louis, A. A.; Doye, J. P. K.; Bath, J.; Turberfield, A. J. Optimizing DNA Nanotechnology through Coarse-Grained Modeling: A Two-Footed DNA Walker. ACS Nano 2013, 7, 2479−2490. (47) Feldkamp, U. CANADA: Designing Nucleic Acid Sequences for Nanobiotechnology Applications. J. Comput. Chem. 2010, 31, 660− 663. (48) Zadeh, J. N.; Steenberg, C. D.; Bois, J. S.; Wolfe, B. R.; Pierce, M. B.; Khan, A. R.; Dirks, R. M.; Pierce, N. A. NUPACK: Analysis and Design of Nucleic Acid Systems. J. Comput. Chem. 2011, 32, 170−173. (49) Zuker, M. Mfold Web Server for Nucleic Acid Folding and Hybridization Prediction. Nucleic Acids Res. 2003, 31, 3406−3415.

5890

DOI: 10.1021/acsnano.6b01035 ACS Nano 2016, 10, 5882−5890