Introduction

In DNA nanotechnology, toehold-mediated DNA strand displacement reaction (SDR) is a powerful tool to construct many dynamic devices such as nanomachines1,2,3, sensors4, 5, logic gates6, 7, transformable structures8,9,10 and catalytic amplifiers11. In general, the key feature of SDR is DNA strand-exchange reaction in which a prehybridized strand from a DNA duplex is displaced by an invading DNA strand. The introduction of toehold significantly enhances the migration rate of invading strand toward original duplex to over 106-fold12, laying a foundation for dynamic DNA nanotechnology.

Originally, SDR is roughly manipulated by varying the toehold binding strength. Later, in order to enrich the activatable toolbox of SDR, different kinds of improved SDRs which are based on the toehold design, including toehold exchange13, remote toehold14, associative toehold15, and allosteric toehold16, have been successively developed. Moreover, considering the above strategies only concentrate on pure DNA systems, many efforts have also been made to realize SDR that is responsive to other stimuli instead of oligonucleotides. For example, aptamers have been integrated into toehold sequence, making the SDR responds to adenosine triphosphate (ATP)17, 18. Protein enzymes such as glutathione transferase and urease that can induce the pH change of solution have been used to trigger the activation of SDR19. Environmental factors such as light, pH, and metal ions have also been employed as stimulating elements for SDR through photocleavable oligonucleotides and DNA tetraplexe structures20, 21. For example, Ricci group developed a set of pH-responsive DNA devices in which the SDR can be finely triggered and controlled by using the parallel Hoogsteen (T, C)-motif in triplex DNA22. Altogether, the exploration of novel responsive elements for SDR has two fundamental significances: (1) it helps in improving the programmability of this reaction since in the presence of invading strand the process can’t be further controlled; (2) it expands potential utilization of this reaction in biomedical areas such as biosensor and targeted drug delivery.

Nucleic acid-related enzymes (NAEs), including nuclease, polymerase, and nicking enzymes, etc., have been widely used in molecular biology and many important techniques such as polymerase chain reaction (PCR) and gene recombination technology. Several groups have also made use of NAEs to dynamically control DNA nanodevices23, 24. Nevertheless, we note that there has been no report on controlling SDR through NAEs so far25. Considering SDR has been involved in the real biological environment (i.e., translation regulator)26, 27, the interaction between SDR’s elements and cellular NAEs is unavoidable. So, it is vital to study NAEs-controlled SDRs, which not only provides novel regulation approaches to the current strand displacement toolbox, but also helps in understanding the influence of NAEs on SDR and SDR-based DNA nanodevices. Motivated by the above arguments, we have conducted a series of NAEs-based studies for flexible SDR control and have further made use of these strategies to realize keypad lock systems and biosensing.

Results and Discussion

Nuclease-regulated SDR

Generally, there are two primary classifications of nuclease, namely exonucleases and endonucleases. However, some nucleases such as DNAase I and S1 nuclease indiscriminately cleave single-stranded DNA (ssDNA) or double-stranded DNA (dsDNA) without sequence-selectivity, making them impossible to regulate sequence-dependent SDR. So, we choose exonuclease III (Exo III), a classic exonuclease that selectively digests the 3′-end of strands in duplex DNAs, to realize SDR control. Figure 1 schematically outlines two principles of the Exo III-controlled SDR, which employ Exo III as an activator (strategy #1) or an inhibitor (strategy #2), respectively. In the first strategy, two ssDNA strands, F1 (fluorophore unit modified strand) and Q1 (quencher modified strand), both have domains complementary to each other, and the two strands also contain additional unpaired domains (4 nucleotides, green) in between. Strand F1 has an overhang at its 3′-end that can efficiently block the unfavorable cleavage of Exo III after hybridization of F1 and Q1, and the same overhang also appears in the 3′-end of strand I1 (invading strand). After the formation of F1Q1 duplex, the unpaired domain of Q1 forms a loop close to the toehold (red) that is hidden in the duplex, resulting in an extremely slow strand displacement between strand I1 and F1Q1 duplex. In the presence of Exo III, enzyme cleave happens at the 3′-end of Q1 until Exo III meets the bulge-loop structure without destroying the whole strand Q1. As a result, the hidden toehold of the F1Q1 duplex can be exposed and hybridize with its complementary strand I1 to initiate the SDR. In the second strategy, since the toehold has been exposed directly, we expect that the Exo III could rapidly cleave from the 3′-end of strand I2 once the SDR starts, leading to the failure of DNA displacement.

Figure 1
figure 1

General scheme of the Exo III-controlled, toehold-based DNA strand displacement strategies.

Full size image

Figure 2A depicts the results of strategy #1 using Exo III as an activator. When the F1Q1 duplex forms, the fluorophore HEX (λem = 536 nm and λex = 556 nm) and the black-hole quencher (BHQ1) are in proximity to each other, so the fluorescence is quenched. As expected, the addition of strand I1 cannot recover fluorescence signal, demonstrating that no displacement occurs. Similarly, Exo III alone also fails to enhance any signal response. However, once in the presence of Exo III and strand I1, SDR can be activated in an enzyme-concentration-dependent manner, consistent with the formation of the energetically stabilized duplex F1I1. Upon treatment of enzyme (10 units) within 4 min, almost 92% strand Q1 has been displaced by strand I1, suggesting a high activation level. These results demonstrate that Exo III indeed acts as an activator and can regulate SDR by adjusting the concentration of Exo III molecules.

Figure 2
figure 2

(A) Time-dependent fluorescence spectra changes upon subjecting the Exo III activation module to only invading strand I1 (a) and I1 and variable concentrations of Exo III: (b) 0, (c) 1, (d) 2, (e) 3, (f) 5, (g) 10 units, and (h) only strand F1, [F1] = [Q1] = [I1] = 50 nM. (B) Real-time fluorescence changes upon treatment of the Exo III inhibition module with I2 and variable concentrations of Exo III: (a) 0, (b) 5, and (c) 10 units, [F1] = [Q2] = [I2] = 50 nM.

Full size image

Figure 2B depicts the results of strategy #2 using Exo III as an inhibitor. Since the toehold (8 nucleotides) has already been exposed, the addition of strand I2 can significantly increase fluorescence signal. However, when Exo III has been introduced into the system, the fluorescence signal doesn’t reduce obviously even at the highest Exo III concentration (10 units), indicating that Exo III has weak ability to stop SDR. It is worthy to note that the source of signal recovery is not due to the Exo III-induced dissociation of F1 from the formed F1/I2 duplex, considering that the completely cleavage of I2 in the F1/I2 duplex requires ~8 min (Figure S1). Actually, SDR can be divided into two processes, which are toehold binding and subsequently branch migration. If Exo III has enough hydrolysis speed, the toehold of strand I2 will be timely destroyed before the branch migration begins. Nevertheless, these results suggest that SDR is so fast that Exo III doesn’t has enough time to completely cleave toehold to stop SDR, which may be attributed to the relatively low processivity of the enzyme. Although Exo III fails to act as an inhibitor, this phenomenon may be helpful to understand the DNA displacement behavior when the system appears Exo III that has been widely used in nucleic acid-related areas.

Nicking enzyme-controlled SDR

Nicking enzyme (NEase) is a common endonuclease that recognizes specific nucleotide sequences in double-stranded DNA and cleaves only one of the two strands. Different from Exo III, NEases don’t cleave nucleotides one by one, while they can create a nick in the DNA duplex, which may be superior candidates for SDR regulation. Figure 3A schematically outlines two principles of the NEase-controlled SDR, which use NEase as an inhibitor. The recognition sequence of (blue) Nt.BbvCl is integrated into toehold, and we expect that Nt.BbvCl could rapidly cleave and shorten the toehold nucleotides in invading strand from seven to two, resulting in an oppressive SDR.

Figure 3
figure 3

(A) General scheme of the Nt.BbvCl-inhibited, toehold-based DNA strand displacement strategies. (B) Real-time fluorescence changes upon treatment of the Nt.BbvCl-inhibition module i with variable concentrations of Nt.BbvCl: (a) 0, (b) 1, (c) 2, (d) 5, and (e) 10 units, [F2] = [Q3] = [I3] = 50 nM. (C) Real-time fluorescence changes upon treatment of the Nt.BbvCl-inhibition module ii with variable concentrations of Nt.BbvCl: (a) 0, (b) 1, (c) 5, and (d) 10 units, [F3] = [Q4] = [I4] = 50 nM.

Full size image

Figure 3B depicts the outcomes of strategy #3 using Nt.BbvCl as an inhibitor. Considering the property of NEase, the location of recognition sequence has two potential sites. One is entirely located in the toehold and the other is located between toehold and branch migration sequence (strategy #3i and ii). We have studied the displacement processes before and after injection of Nt.BbvCl through recording the real-time fluorescence change. For both strategies, Nt.BbvCl alone cannot induce any signal increase, while strand displacement proceeds with a fast kinetic upon invading strand I3/I4 addition (Fig. 3B and C, curve a). However, once in the presence of Nt.BbvCl, the strand displacement rates can be decelerated in a Nt.BbvCl concentration-dependent way. When Nt.BbvCl concentrations are increased from 1 to 10 units (Fig. 3B curve b–e, Fig. 3Cb–d), the rate constants can be regulated in the range from 8.69 × 105 to 7.22 × 104 M−1 s−1 (strategy #3i) and from 1.21 × 105 to 9.94 × 104 M−1 s−1 (strategy #3ii), respectively. These results are consistent with the Nt.BbvCl-stimulated SDR inhibition mechanism, and the strand displacement kinetics can be finely adjusted by varying the concentrations of Nt.BbvCl.

Obviously, Nt.BbvCl in strategy #3i has a much better inhibition effect compared with strategy #3ii. Considering the most difference between strategy #3i and ii is the location of enzyme’s recognition site, it indicates that toehold binding is the rate-determining step, so Nt.BbvCl has enough time to nick toehold; however, once the whole toehold completes hybridization and branch migration process proceeds, the displacement reaction rate greatly promotes, making Nt.BbvCl inefficiently prevents SDR. So, we conclude that the reaction rate order is rbranch migration > rNt.BbvCl > rtoehold binding > r Exo III.

Polymerase enzyme-controlled SDR

Polymerase can add nucleotides from 3′-end of template strand with extremely high efficiency (1000 nucleotides/min). Thus, it is easily to achieve polymerase-inhibited SDR. As shown in Fig. 4A, strand F4 and Q5 are partially complementary to each other, and the extension of strand Q5’s 3′-end by polymerase can efficiently block toehold and suppress SDR. Figure 4B shows the results obtained from addition of different concentrations of polymerase to the reaction system. Of note, without any optimization, the SDR is responsive to polymerase as low as 1 × 10−5 units, which is much more sensitive than nuclease-controlled SDR. This phenomenon indicates that polymerase-mediated interference should also receive enough attention in the construction of dynamic DNA nanodevices which work in complex biological media, considering that previous reports only focused on the influence of nuclease on DNA structure.

Figure 4
figure 4

(A) General scheme of the polymerase-inhibited, toehold-based DNA strand displacement strategy. (B) Fluorescence spectra changes observed upon treatment of the sensing module with variable concentrations of polymerase: (a) 0, (b) 1 × 10−5, (c) 1 × 10−4, (d) 1 × 10−3, (e) 0.01, and (f) 0.1 units, [F4] = [Q5] = [I6] = 50 nM. (C) General scheme of the polymerase-based keypad lock system. (D) Fluorescence intensities of the system at 520 nm with respect to 24 kinds of input sequences.

Full size image

Inspired by the property of polymerase, we have made use of polymerase-inhibited SDR to establish a keypad lock system (Fig. 4C), whose output signals not only depend on the proper combination but also the correct sequence of input signals, namely password. Since the polymerase only adds nucleotide from 3′-end of DNA strand, we can alternatively inhibit SDR through the selective addition of substrate (i.e., dATP, dTTP, dCTP, and dGTP), leading to 24 kinds of input sequences. The designed sequence of toehold is 5′-CCCCCCGAT-3′, so it requires polymerase to add nucleotide from 3′-end of Q in an order of ATCG, resulting in a complete suppression of SDR. The output signal is defined as “ON” when the SDR is totally inhibited and as “OFF” when SDR is inadequately inhibited. Stimulated by the nucleotide-encoded inputs, the system exhibits different fluorescent responses as shown in Fig. 4D. Obviously, only one correct order of the input signals (ATCG) results in the “ON” state, while all others produce the “OFF” state. Thus, a four-input molecular keypad lock has been successfully mimicked that can stop SDR only when the correct “password” is entered. The system keeps “OFF” at all other combinations to deny the illegal invasion.

Also, inspired by the sensitivity of polymerase, we have made use of this inhibition strategy to detect telomerase. Telomerase is a ribonucleoprotein that contains a RNA template in the protein backbone28. It is over-expressed in cancer cells, and catalyzes the addition of the telomeric repeats (TTAGGG)n onto the 3′-end of the human chromosomes29. As shown in Fig. 5A, the sequence of strand Q7 is used as telomerase primer, and the toehold in the strand F6 is complementary to the production of telomerase (only one copy). Similar results with polymerase, the elongation of template results in complete blocking of toehold and invalid SDR. Figure 5B depicts the fluorescence spectra generated by the telomerase-treated F6Q7 that are interacted with invading strand I7, in the presence of the dNTP mixture, for a fixed time-interval of 40 min. As the telomerase content increases, the resulting fluorescence is decreased, consistent with the higher level of toehold blocking. Figure 5B, inset, reveals the calibration curve corresponding to the fluorescence intensities generated upon detecting various concentrations of the telomerase. The detection limit for analyzing telomerase corresponded to 8 × 10−10 IU. Actually, the potential interference of polymerase can’t be ignored (Figure S2), so we have updated the sequence design by adding a mismatch base (G) in the toehold and employed a novel dNTP mixture without addition of dCTP. As a result, the interference of polymerase can be completely eliminated (Figure S3). All in all, these results demonstrate that SDR can be developed to detect telomerase conveniently and sensitively.

Figure 5
figure 5

(A) General scheme of the telomerase detection based on DNA displacement reaction. (B) Fluorescence spectra changes observed upon treatment of the sensing module with variable concentrations of polymerase: (a) 0, (b) 8 × 10−10, (c) 8 × 10−9, (d) 8 × 10−8, (e) 8 × 10−7, (f) 8 × 10−6, (g) 8 × 10−5 IU, [F6] = [Q7] = [I7] = 50 nM.

Full size image

Conclusion

In summary, we report a series of NAE-based strategies for SDR regulation and offers additional information on how NAEs affect SDR. We have found that (1) Exo III cannot timely stop SDR once the reaction starts, indicating a slower cleavage rate of enzyme compared with DNA displacement; (2) NEase allows the fine regulation of SDR kinetics through adjusting the concentration and recognition site of the enzyme; (3) polymerase should be a unneglected factor that affects nucleic acid system other than widely studied nuclease and can be used to construct biocomputing security system; (4) SDR can be developed to detect nucleic acid-related enzyme. We anticipate that our strategy can also be applied to construct diverse NAE-fueled DNA nanodevices, which may provide a means for biosensing in biological system.

Methods

Materials and Reagents

All HPLC-purified DNA oligonucleotides were purchased from Sangon Inc. (Shanghai, China). The DNA sequences and modifications are listed in Table S1. Exonuclease III (Exo III), Nt.BbvCl, Taq polymerase and their corresponding reaction buffers were purchased from NEB (Ipswich, USA). Pure telomerase was obtained from the ELISA kit that was purchased from Yibang biotech. Inc. (Suzhou, China). Proteins were diluted with NHS-magnetic bead (diameter = 500 nm) was purchased from Beaver biotech. Inc. (Suzhou, China). Tris(hydroxymethyl)metyl aminomethane (Tris), ethylene diamine tetraacetic acid (EDTA), deoxyadenosine triphosphate (dATP), deoxythymine triphosphate (dTTP), deoxycytosine triphosphate (dCTP), and deoxyguanosine triphosphate (dGTP) and all other chemicals were purchased from Sigma-Aldrich (St. Louis, MO, USA).

The buffers employed in this work were as follows. DNA storage buffer: 20 mM Tris-HCl, 1 mM EDTA (pH 7.4). Exo III reaction buffer: 1 × NEBuffer1. Nt.BbvCl reaction buffer: 1 × CutSmart. Taq polymerase reaction buffer: OneTaq standard reaction buffer. Telomerase reaction buffer: 10 mM phosphate buffered saline (PBS, pH 8.0) with 140 mM KCl, 1.5 mM MgCl2, 0.005% tween-20, 200 μM dNTP, and BSA 0.1 mg/ mL. Buffers for both electrochemistry and electrode washing are 10 mM PBS solution with 0.1 M NaCl and 0.05% Tween-20 (pH 7.4). All solutions were prepared with NANOpure H2O (>18.0 MΩ) from a Millipore system.

Fluorescence Measurements

For preparation of FQ duplex, strand F was mixed with strand Q labeled by the 3′- fluorophore and the 5′-quencher at 1:1 ratio in reaction buffer with a final concentration of 50 nM for each strand. The resulting solutions were annealed by heating at 95 °C for 5 min and then slowly cooled down to room temperature over 2 hours. Fluorescence measurements were performed by using an F-7000 fluorescence spectrometry (Hitachi, Japan) at 25 °C. Sample solutions were excited at 494 nm (FAM) and 536 nm (HEX), and the emission signals were recorded with wavelength of 518 nm and 556 nm, respectively. For all time-dependent fluorescence tests, appropriate volumes of DNA stock solutions were added to reaction buffer to achieve 50 nM final concentration with a total volume of 198 μL in the cuvettes. Afterwards, corresponding concentration of enzyme and invading strand I, with 1 μL and 1 μL of volume, respectively, were added respectively and mixed quickly within 20 s. In all graphs, time t = 0 indicates the time of strand I being added to the solutions.

Keypad lock system

For achieve keypad lock system, NHS-magnetic beads (d = 500 nm) were used as solid support for the DNA strands immobilization. The amino group modified strand F 5 was pre-mixed with equivalent strand Q 6 (5 μM) and the duplex was tethered onto the surface of NHS-magnetic beads according to the manufacturer instructions. Four deoxynucleotide triphosphates were used as the input signals (A, T, C, G). The beads were exposed to 1 mL of solution composed of 5 U Taq polymerase and one of the four deoxynucleotide triphosphates (0.5 mM) for 5 min intervals before being reacted with the next solution. After the reaction with four solutions, the beads were exposed to strand I 6 (1 μM) for 5 min intervals and the resulting supernatants were transferred to a cuvette for fluorescent measurements.

Telomerase detection

For telomerase detection, strand F 6 was mixed with strand Q 7 at 1:1 ratio in reaction buffer with a final concentration of 50 nM for each strand. The resulting solutions were annealed by heating at 95 °C for 5 min and then slowly cooled down to room temperature over 2 hours. Different concentrations of telomerase were incubated with the duplex for 40 min at 37 °C. Afterward, the resulting solution was introduced into strand I7 for 5 min of incubation and was subjected to fluorescent measurements.