Programming crystallization kinetics of self-assembled DNA crystals with 5-methylcytosine modification

Significance

Self-assembled DNA crystals provide a precise yet controllable platform for DNA nanotechnology due to their programmable three-dimensional structural feature. However, complex designs are often required to control the crystal structures. Here, we have developed an atomic-level 5-methylcytosine (5mC) modification strategy to program the crystallization kinetics of DNA crystals. We reveal that 5mC modification enhances the reaction rate of DNA hybridization using combined DNA-PAINT and FRET-labeled DNA strand displacement experiments. We thus modulate the crystallization kinetics and then the morphology of DNA crystals by tuning the site and number of 5mC modifications. This atomic-level modification strategy expands the design space of DNA crystals and offers a promising route toward the development of crystalline DNA materials with tunable properties.

Abstract

Self-assembled DNA crystals offer a precise chemical platform at the ångström-scale for DNA nanotechnology, holding enormous potential in material separation, catalysis, and DNA data storage. However, accurately controlling the crystallization kinetics of such DNA crystals remains challenging. Herein, we found that atomic-level 5-methylcytosine (5mC) modification can regulate the crystallization kinetics of DNA crystal by tuning the hybridization rates of DNA motifs. We discovered that by manipulating the axial and combination of 5mC modification on the sticky ends of DNA tensegrity triangle motifs, we can obtain a series of DNA crystals with controllable morphological features. Through DNA-PAINT and FRET-labeled DNA strand displacement experiments, we elucidate that atomic-level 5mC modification enhances the affinity constant of DNA hybridization at both the single-molecule and macroscopic scales. This enhancement can be harnessed for kinetic-driven control of the preferential growth direction of DNA crystals. The 5mC modification strategy can overcome the limitations of DNA sequence design imposed by limited nucleobase numbers in various DNA hybridization reactions. This strategy provides a new avenue for the manipulation of DNA crystal structure, valuable for the advancement of DNA and biomacromolecular crystallography.

Self-assembled DNA crystals are unique crystal structures that arise from self-assembly of DNA molecules with atomic-level spatial and information precision (1–3). Such DNA crystals were first demonstrated by Seeman and co-workers. They employed DNA tensegrity triangle motif bearing 2-nt DNA sticky ends along three axial directions as building blocks and grew DNA crystals with rhombohedral symmetry (4). DNA crystals provide a precise chemical platform for organizing nanoparticles or enzymes in three dimensions, achieving cascade and selective catalysis (5, 6), or modulating optical or acoustic properties (7–9). Their high porosity and density also offer opportunities for material separation and DNA data storage (10–12). Leveraging the programmability and manipulability of DNA molecules, accurate chemical modification and assembly of which can be realized, resulting in DNA crystals with specific lattice parameters and symmetry, as well as unique physical and chemical properties (13–19). For instance, Yan et al. achieved threefold or sixfold screw axes-structured DNA crystals through interlayer hybridization of square motifs (13, 14). Mao et al. constructed three or fourfold screw axes DNA crystals using two short strands as motifs (15). Di Michele et al. assembled 3D crystals with lattice parameters over 20 nm using amphiphilic DNA motifs (16, 17). Furthermore, Seeman and Mao et al. exploited DNA crystals to implement light modulation (9, 20), logic gate operation (21), and visualization of DNA molecular reactions (22, 23). The foundation for these applications is the single nucleobase designability of DNA crystal structure. Hence, how to precisely regulate the structure and growth of DNA crystal is crucial in this field.

Several prior studies have revealed that the design of sticky ends of DNA motif is essential for DNA crystal growth (24–27). For instance, Seeman and Mao et al. discovered that the sequences of sticky-end nucleobases in DNA motifs determine whether a DNA crystal forms, as well as how its symmetry varies (24, 25). They further observed that the preservation of 5′-phosphate groups could enhance the hydrogen bond interactions among DNA strands, thereby influencing the crystallization kinetics and modulating the morphology of DNA crystals (26). Yan et al. also reported that immobile sequences of Holliday Junction impact the crystallinity and symmetry of DNA crystals (27). However, the number of base pairs or the GC content in the sticky ends of DNA motifs cannot be arbitrarily increased, as this could lead to undesired crystal structures or even hinder the crystallization process (25, 28). How to program crystallization kinetics of DNA crystal and modulate their morphology without altering DNA sequence remains challenging.

Chemical modifications alter the intrinsic properties of DNA nucleobase molecules at the atomic level, such as charge, rigidity, and steric hindrance (29, 30). They play a key role in modulating DNA–protein interactions, chromosome replication, DNA mismatch repair, gene transcription, and epigenetic phenotype formation (31, 32). They also affect the diffusion, folding, and collision of DNA in solution while maintaining precise complementarity of DNA nucleobases, thus affecting DNA hybridization based on nucleobase recognition and pairing (33, 34). 5-Methylcytosine (5mC) is the most prevalent nucleobase chemical modification on DNA, accounting for ~1% of all nucleobases in mammalian DNA (35, 36). It can not only induce X chromosome inactivation, inhibit transposons, and act as a biomarker and therapeutic target for cancer but also affect the rigidity, electronegativity and diffusion coefficient of DNA (30, 33, 37–39). Interestingly, we found that each sticky-end of diverse tensegrity triangle motifs contains at least one G-C base pair, offering intrinsic sites for 5mC modification. Hence, exploring the influence of specific 5mC modifications on the DNA crystallization process holds significance for programming DNA crystal growth within the strict constraints of sequence limitations.

In this work, we demonstrated that atomic-level nucleobase 5mC modification kinetically controls DNA crystal growth and modulates DNA crystal morphology. First, we observed that 5mC modification on the DNA sticky-ends exerted a remarkable effect on the macroscopic growth habit and crystallization kinetics of DNA crystals. To elucidate the mechanism underlying this phenomenon, we conducted systematic investigations at the single-molecule and bulk solution level, using DNA-PAINT (40–42) and FRET-labeled DNA strand displacement experiments (43), respectively. Further supported by direct transmission electron microscopy (TEM) pieces of evidence for the acceleration of DNA crystallization kinetics through 5mC modification, we concluded that 5mC modification enhances DNA double-stranded hybridization affinity, expediting the reaction rate of sticky-end hybridization, facilitating binding between DNA crystal motifs, and promoting nucleation and crystallization rates.

Results and Discussion

Anisotropic Growth of DNA Crystals Regulated by 5mC Modification.

We utilized rhombohedral DNA crystal as a model system (26). It has periodic repeating units that can accumulate and amplify subtle interactions between the tensegrity triangle motifs. Such accumulated weak interactions will ultimately affect the bulk structure of DNA crystal. As shown in Fig. 1A, we constructed a DNA tensegrity triangle motif (Δ4T) with four helical turns per edge using six strands in three vector directions (1, 2, and 3) and a central DNA strand L: L+M1+M2+M3+S1+S2+S3. The strands were hybridized by 2-nt sticky-end cohesions (Fig. 1A, sequences shown in SI Appendix, Fig. S1A and Table S1), allowing periodic extension and DNA crystal formation. We applied the tricolor rule to label the 5mC modifications on the sticky ends for different directions (1: red; 2: blue; 3: green, Fig. 1 A, Right panel). The modified directions were indicated by colored rods, and the unmodified ones by gray rods. Previous studies have shown that 5mC modification could enhance the hydrogen bond interaction of the Guanine–Cytosine (G-C) base pair (Fig. 1B), thus increasing the thermal stability of the double helix (44–46).

Fig. 1.

DNA tensegrity triangle crystals (Δ4T) as an experimental model for studying 5mC regulation of DNA crystallization kinetics. (A) Two tensegrity triangle motifs hybridize through two base pairs of sticky ends, and then replicate periodically from three directions (M1, M2, and M3) to form DNA crystals. (B) Schematics of G-C base pairs without and with 5mC modification, using stick model.

We observed that 5mC modification has a minimal impact on the cell parameter of DNA crystals. We used single-crystal X-ray diffraction to analyze the structure of typical 2-turn tensegrity triangle (Δ2T) DNA crystals assembled without (Δ2T) and with (Δ2T-m) the presence of 5mC modification (SI Appendix, Figs. S1B and S2 and Table S2) and found that they belong to the same space group (H3) and have almost the same unit cell parameters (a = b = 104.0 Å, c = 102.0 Å for Δ2T; a = b = 104.3 Å, c = 100.0 Å for Δ2T-m; and both with α = β = 90°, γ = 120°).

Interestingly, we found that 5mC modification changed the dominant growth direction of DNA crystals. As shown in Fig. 2A, by coordinating the 5mC modification on three directions, we designed four combinations of Δ4T^1m (red), Δ4T^12m (purple), Δ4T^13m (yellow), and Δ4T^123m (white) and marked the DNA crystals obtained from these sequences with corresponding colors (Fig. 2 B–G and SI Appendix, Figs. S3 and S4). Without any 5mC modification (Δ4T), the DNA triangle motifs hybridized to form cuboid-like crystals (Fig. 2C). The shape parameters of the DNA crystals (Fig. 2H and SI Appendix, Fig. S4) revealed a three-dimensional vector ratio of 2.2:1.0:2.3. Next, we modified 5mC on the sticky-ends of direction 1 (Δ4T^1m, Fig. 2D) and obtained parallelogram sheet crystals after hybridization of DNA crystal motifs (Fig. 2 D, Bottom panel), with a three-dimensional vector of 1.8:~0:1.0, indicating that direction 1, with 5mC modification, had a faster growth rate compared to directions 2 and 3. Based on these results, we further increased the 5mC modification on the sticky ends of direction 2 (Δ4T^12m) and direction 3 (Δ4T^13m), respectively, and then obtained crystals with a three-dimensional vector ratio of: 5.7:1.0:2.4 (Δ4T^12m) and 1.0:~0:1.2 (Δ4T^13m). These results indicated that the growth rate of DNA crystals increased in directions 1 and 2, and directions 1 and 3, respectively, resulting in cuboid (Fig. 2E) and parallelogram sheet (Fig. 2G) DNA crystals. Finally, we modified 5mC on all three sticky-ends (Δ4T^123m), obtaining cuboid-like DNA crystals with a larger size and a three-dimensional vector ratio of 2.2:1.0:2.8 (Fig. 2F), which was similar to that of unmodified DNA crystals (2.2:1.0:2.3, Fig. 2C). Based on the above experiments, we speculate that 5mC modification enhances the binding force of base pairs on sticky ends, accelerating the crystal motif binding rate in the directions with 5mC modification.

Fig. 2.

The anisotropic growth of DNA crystals based on 5mC regulation. (A) The sticky ends of DNA tensegrity triangle motif in three directions are marked with red (M1, direction 1), blue (M2, direction 2), and green (M3, direction 3) colors after 5mC modification. (B) Different colors are superimposed by the synergistic effect of 5mC modification in different directions: purple (directions 1 and 2, Δ4T^12m), yellow (directions 1 and 3, Δ4T^13m), white (directions 1, 2, and 3, Δ4T^123m), and the corresponding colors are used to label the DNA crystals below. (C) The morphology of DNA crystals and their schematics without Δ4T modification. (D) The schematics and morphology of DNA crystals with single 5mC modification at the sticky end in direction 1 (Δ4T^1m), (E) in directions 1 and 2 simultaneously (Δ4T^12m), (F) in directions 1, 2, and 3 simultaneously (Δ4T^123m), (G) in directions 1 and 3 simultaneously (Δ4T^13m). (H) The statistics of crystal parameters obtained by different direction synergistic regulation. (Scale bars: 100 μm.)

5mC Modification Enhances Thermal Stability of DNA Duplex.

To explore the reason for 5mC modification that regulating DNA crystal growth, thermodynamic experiments were first applied (Fig. 3 and SI Appendix, Fig. S5). We found that single 5mC modification can enhance the thermal stability of DNA duplex slightly. We designed a DNA duplex with a 5mC modification on a pair of G-C base pairs at the end of duplex C-aa’ (mC-aa’, Fig. 3A) and then used thermal melting experiments to test the duplex melting temperature without and with 5mC modification (Fig. 3B). The results showed that a single 5mC modification increased the duplex melting temperature (T_m) significantly (P-value = 1.38 × 10⁻²²), from 73.6 ± 0.2 °C to 74.3 ± 0.5 °C (ΔT_m = 0.7 °C). Additionally, this thermal stability enhancement effect shown length-independent (Fig. 3 C and D), number-dependent, and position-independent effect (Fig. 3 E and F). Our results could be explained by early theoretical studies that the presence of 5mC in DNA increases the negative charge of DNA, relocates multivalent cations from the DNA major groove to the double helix region, thus enhances the hydrogen bond interaction on the G-C base pair, and finally increases the thermal stability of the duplex (30, 33, 37–39).

Fig. 3.

5mC modification enhanced the thermal stability of DNA duplex. (A) Molecular formula of cytosine and 5mC modification on 20-bp DNA duplexes. (B) Thermal melting experiment for 20-bp DNA duplexes. (C) Molecular formula of cytosine and 5mC modification on 25-bp DNA duplexes. (D) Thermal melting experiment for the 25-bp DNA duplexes. These comparisons show that single 5mC modification can enhance the thermal stability of DNA duplex with length-independent effect. (E) Schematics showing the number and location of 5mC modifications on the DNA duplex. 5mC is indicated by a red triangle. (F) The melting temperatures indicate that the 5mC enhancement has number-dependent and position-independent effect.

DNA-PAINT Study on 5mC Modification.

Next, we found that 5mC modification affected the DNA duplex affinity and hybridization reaction kinetics, using DNA-PAINT. DNA-PAINT is a fluorescence-labeled DNA oligonucleotide chain dynamic binding and dissociation single-molecule detection method developed by Jungmann et al. (40, 42). In general, single 5mC modification increased the affinity constant, k_a, (k_a = k_on/k_off) of the DNA duplex from 9.9 × 10⁶ M⁻¹ to 14.4 × 10⁶ M⁻¹, corresponding to a 45.5% enhancement in binding strength (Fig. 4 and SI Appendix, Figs. S6–S9). First, we designed double or triple boundary marker on two types of square DNA origami to distinguish the un-modified group (“C” group) and the modified group (“5mC” group), respectively (Fig. 4A). Then, we designed an extended single-stranded DNA with or without 5mC modification (“5mC” and “C”) to capture the imager strands (Fig. 4 A, Right panel). We obtained shape-different fluorescent dot arrays by single experiment, dual-channel shooting (Fig. 4B). At the same time, we can track the single-molecule fluorescence blinking kinetics of the hybridization events between the imager and capture strands on DNA origami and characterize the effect of 5mC modification on the frequency, binding, and dissociation time of DNA duplex hybridization reaction (Fig. 4 C and D and SI Appendix, Fig. S7).

Fig. 4.

Effect of 5mC modification on the kinetics of DNA duplex hybridization revealed by DNA-PAINT single-molecule experiments. (A) Schematics of the DNA origami structures with shape-specificity for DNA-PAINT and the single-stranded DNA binding events. (B) Super-resolution images of two DNA origami structures obtained by DNA-PAINT for un-modified group (“C” group) and modified group (“5mC” group). (C) Single-molecule fluorescence blinking kinetics are used to measure the frequency, binding time, and dissociation time of DNA duplex hybridization after 5mC modification. (D) Schematics showing how 5mC modification affects the kinetics of DNA duplex hybridization. (E) 5mC modification increases the constant of binding rate k_on of DNA duplex hybridization by 3.7 × 10⁵ M⁻¹ s⁻¹. (F) 5mC modification strengthens hydrogen bonding interactions, which decreases the dissociation rate constant k_off of DNA duplex hybridization by 0.24 s⁻¹. (G) 5mC modification increases the affinity constant k_a of DNA duplex hybridization by 45.5% or 4.5 × 10⁶ M⁻¹.

We found that 5mC modification enhanced the constant of binding rate (k_on) of DNA duplex, from 8.3 × 10⁶ to 8.6 × 10⁶ M⁻¹ s⁻¹ (Fig. 4E), after statistical analysis of single-molecule events of duplex hybridization (SI Appendix, Figs. S8 and S9). The enhancement effect was modest, possibly because the 5-methyl group introduced steric hindrance that partly interfered with DNA duplex binding and offset the enhancement effect of 5mC modification (46). Remarkably, when the imager strand hybridized with the single capture strand, the 5mC modification boosted the hydrogen bond interaction of the G-C base pair and improved the duplex stability, enabling the imager strand to linger longer on the origami. This was reflected by a marked drop in the constant of dissociation rate (k_off), from 0.84 to 0.60 s⁻¹ (Fig. 4F). Thus, a single 5mC modification raised the DNA double-strand affinity constant k_a from 9.9 × 10⁶ to 14.4 × 10⁶ M⁻¹ (Fig. 4G), which ultimately accelerated the DNA hybridization reaction rate. Overall, the mechanism underlying the enhancement of DNA hybridization reaction rate by 5mC modification stemmed from the larger effect of k_off reduction on k_a.

FRET Study on 5mC Modifications.

We designed a DNA strand displacement reaction FRET model system using a 3-nt sticky-end (dCCC, Fig. 5 and SI Appendix, Fig. S10), in which the Reporter (R1-R2) is composed of reporter strand 1 (R1) labeled with a TET fluorophore (red circle) at the 3′ end and reporter strand 2 (R2) is labeled with a BHQ2 quencher (gray circle) at the 5′ end (Fig. 5A). Then we added the input strand (I, I0-I5), which was fully complementary to R1 and contained 5mC modifications at different positions (Fig. 5B). Due to the higher stability of the fully complementary duplex R1-I than that of R1-R2, and driven by the sticky-end, input strand I gradually hybridized with strand R1 and restore the originally quenched fluorescence.

Fig. 5.

The reaction rate of DNA strand displacement reaction was dependent on the number and site of 5mC modification. (A) Schematic illustration of the DNA strand displacement assay. A red circle denotes the dye TET, and a gray circle denotes the quencher Black Hole Quencher 2 (BHQ2). (B) Schematics showing the number and location of 5mC modifications on the DNA strand. 5mC is indicated by a red triangle. (C) Kinetics traces for different inputs: I0 to I5. (D) Ratio of reaction rates of each 5mC modification input (I1 to I5) to the un-modified input (I0). The reaction without 5mC modification (I0) is slowest. Reaction rates increase as the number of 5mC modification and reaction with 5mC modification in the non-sticky-end (I4) is almost as fast as the reaction with the un-modified input (I0).

We found that 5mC modification on the sticky end enhanced the DNA hybridization rate in a number- and site-dependent fashion. On the one hand, the kinetic plots showed that 5mC modification at the sticky end could accelerate the DNA hybridization reaction rate. The larger the modification number, the faster the hybridization reaction rate: I4 > I2 ≈ I3 > I1 > I0 (^mC^mC^mC > C^mC^mC ≈ ^mCC^mC > CC^mC > CCC, Fig. 5C). As shown in Fig. 5D, the increment of hybridization rate for single 5mC modification (I1) was 1.7 ± 0.1-folds, two 5mCs modification (I2 and I3) were 3.1 ± 0.1-folds, and three 5mCs modification (I4) was 5.4 ± 0.3-folds. On the other hand, 5mC modification at the non-sticky-end region had limited effect on the reaction rate (k(I5)/k(I0) = 1.1 ± 0.1). The possible reason for the site-dependent manner is that the toehold region determines the reaction rate of DNA strand displacement, while the non-toehold region is irrelevant (47). Therefore, by modifying the toehold region with 5mC, the process of DNA strand displacement reaction can be effectively controlled. These results demonstrate the flexibility of nucleobase chemical modification strategy, which is well reflected in our above experimental results for anisotropic growth of DNA crystals.

We further confirmed the controllability of the 5mC modification strategy through unidirectional multiple modification experiments on DNA crystal. By introducing two 5mC modifications in direction 3 (SI Appendix, Fig. S11), we observed that, in comparison to Δ4T^13m, the dimensions of directions 1 and 3, originally close in ratio (1.0:1.2), exhibited variation (1.0:1.6) in Δ4T^13m2, with significant difference (P-value = 2.92 × 10⁻⁵). This outcome indicates that 5mC modifications can not only coherently regulate from different directions but can also undergo multiple modifications in the same direction.

Programming Crystallization Kinetics with 5-Methylcytosine Modification.

To provide direct evidence for the acceleration of DNA crystallization kinetics through 5mC modification, TEM snapshots of DNA crystals were collected before and after 5mC modification at various growth stages (Fig. 6 A–E and SI Appendix, Figs. S12 and S13) (48). The blank sample exhibited predominately randomly cross-linked cluster structures within the initial 3 h of growth (Fig. 6 A and B). At 3 h, DNA crystals with regular shapes started to emerge in the solution. However, at this point, the crystal lattice was not observable under high-resolution TEM (HR-TEM), possibly indicating insufficient crystallinity at this stage. At 4.5 h, the emergence of a lattice in the DNA crystal was observed, indicating an improvement in the degree of order at this stage. In contrast, in 5mC modified sample (Fig. 6 C and D), blocky structures of a certain thickness were observed at 0.5 h. By 1 h, the boundaries of the crystals became clearer. At 1.5 h, DNA crystals with regular shapes were observed, and the DNA lattice could be discerned under HR-TEM. This time-dependent TEM kinetic monitoring demonstrates that 5mC modification accelerates the binding of DNA motifs in the solution, facilitating faster DNA crystallization and expediting DNA crystallization kinetics.

Fig. 6.

Mechanism of accelerated DNA crystallization kinetics by 5mC modification. (A–D) Schematics (A and C) and time-dependent TEM observations (B and D) of the crystallization of Δ2T without and with 5mC modification. Scale bars: 1 µm, Scale bars for the Insets are 50 nm. (E) The schematics for crystal growth process summarized from TEM results. (F) 5mC modification enhances the hybridization rate between DNA crystal motif and accelerates the nucleation and growth rate of DNA crystals. (G) Schematics of 5mC accelerated hybridization from the single-molecule level (Top), sticky-ends hybridization level (Middle), and DNA crystal motif level (Bottom).

We also discovered that Δ4T^123m could form DNA crystals in lower DNA and reservoir buffer concentration than that for Δ4T (Fig. 6F and SI Appendix, Figs. S14 and S15). For example, under the same conditions in a 3 × TAE buffer, crystallization started at a DNA motif concentration of only 2 μM for Δ4T^123m, in comparison, unmodified Δ4T started at 6 μM. Moreover, 2-turn (Δ2T) and 3-turn (Δ3T) DNA tensegrity triangle crystals also exhibited similar acceleration behavior of nucleation and growth after 5mC modification (SI Appendix, Figs. S16 and S17). These results provide additional support for the notion that 5mC modification enhances the hybridization rate of DNA motifs. This improvement facilitates the assembly of nuclei from DNA triangle motifs to reach the critical size more readily, consequently expediting the nucleation rate of DNA crystals.

At last, we concluded that the 5mC modification enhances the affinity of DNA double-stranded hybridization (Fig. 6 G, Top column), thereby expediting the reaction rate of DNA sticky-end hybridization (Fig. 6 G, Middle column), facilitating the binding between DNA crystal motifs, and ultimately promoting DNA nucleation and the crystallization rate (Fig. 6 G, Bottom column). Although the effect induced by the atomic-level base nucleobase chemical modification is very weak, it can be amplified through periodically replicated motifs in the crystal.

Conclusion

In closing, we demonstrated that atomic-level 5mC modification can regulate the crystallization kinetics and the growth habit of DNA crystals. We further verified that the regulatory effect of 5mC modification originates from its ability to enhance the reaction rate of DNA hybridization: single 5mC modification raised the affinity constant by 45.5% and the hybridization rate by 1.7-folds. This subtle modulation effect can be amplified and manifested by changing DNA crystal growth habit and then yielding DNA crystals with different morphology and size. From the perspective of nucleobase structure, this strategy achieves fine-tuning of duplex hybridization kinetics with limited length at the atomic-level, enabling us to more precisely regulate all kinds of systems based on DNA hybridization reactions, especially for DNA nanotechnology, DNA molecular computing, and DNA storage (1, 49). Because of the four nucleobases, deoxyribose and phosphate backbone of DNA can be modified in various ways, which makes our strategy have great potential and important value for the development of DNA and DNA–protein composite crystallography (50–54).

Materials and Methods

Crystallization of Self-Assembled DNA Crystals.

First, 2.0 µM corresponding DNA strands were mixed in 0.2 × TAE/Mg²⁺ buffer and slowly cooled from 95 °C to 22 °C with −0.5 °C/min. Then, 5 µL of the solution was incubated against 600 μL of reservoir buffer in a hanging drop setup at 22 °C for 5 d. Optical images were taken on NIKON Eclipse LV100ND microscope.

Fluorescence Data Acquisition and Analysis for DNA-PAINT.

Super-resolution measurements based on DNA-PAINT technique were carried out on a commercial total internal reflection fluorescence (TIRF) microscope (Leica DMi8). The channel for boundary marker was excited by 15 mW laser at a wavelength of 561 nm. The channel for Imager binding sites was excited at 638 nm by 9 mW laser. The laser beam was focused on the backfocal plane of an oil-immersed objective (100×, NA = 1.4, UPlanSApo, Olympus). The signal was finally acquired by EMCCD camera (Andor, iXon Ultra 897). The calibrated pixel size was 160 nm/pixel. Data acquisition was controlled with the Leica LAS X software. For Imager binding sites measurement, 15,000 frames were acquired at cool temperature (~16 °C). Then, 5,000 frames were acquired for the boundary marker. The exposure time was set to 100 ms each frame for Imager binding sites measurement, and 200 ms each frame for boundary marker measurement. The EMGain value was set to 300 for appropriate signal-to-noise readout. Acquisition was done by setting the camera to a readout sensitivity of 16-bit.

The resulting lif file was converted to raw file format using a custom-written python script. Raw fluorescence data were subjected to super-resolution reconstruction using the Picasso software package. After localization, drift correction and filter, the data were displayed in Picasso Render and more than 400 structures were selected for further binding kinetics analysis. To investigate the kinetics of test sites with single-molecule resolution, we extracted the event time courses of individual spots on origami (sequences for origami are shown in SI Appendix, Table S3), examples of which are shown in SI Appendix, Fig. S7, and the localizations were linked allowing a gap size of 3 frames between localizations within 0.5 pixel to obtain a list of single binding events.

Fluorescence Measurements for the DNA Strand Displacement Reaction.

Synergy H1 Hybrid Multi-Mode Reader (BioTek) and Corning 96-well black assay plate were employed to perform fluorescence measurements. All strand displacement experiments for different Input strands (I0-I5), with different number and location of 5mC modification, were run in parallel on the instrument. Fluorescence kinetics data were collected every 1 min. Excitation wavelength was 510 nm for dye TET, and emission wavelengths was 540 nm. Generally, all the components except the Input strand(s) were mixed in 1 × TE buffer with 12.5 mM MgCl₂. Then, the experiments were performed in the 96-well black assay plate with 98-μL reaction mixture per well for all experiments. The initial fluorescence value before addition of Input strand(s) was recorded as baseline. The experiment was then paused for the addition of 2-μL Input strand(s) and subsequent mixing by shaking. The plate was then put back into the Hybrid Reader and then the experiment was resumed. Experiments were performed with the distinct fluorophores–quenchers pair (TET-BHQ2) at a standard concentration of 50 nM. A standard concentration of 50 nM for Input strand(s) was used. For each experiment, the kinetic data were fitted with a bimolecular rate law and then the reaction rate was obtained. The temperature was kept at 18 °C throughout the measurement.

Supplementary Material

Appendix 01 (PDF)

Data, Materials, and Software Availability

All study data are included in the article and/or SI Appendix.