Introduction

DNA nanotechnology harnesses the exquisite specificity of Watson-Crick base pairing to program nucleic acids into modular building blocks, enabling the precise bottom-up assembly of multifunctional nanostructures for various applications1,2,3,4. In 2006, Paul Rothemund brought up the DNA origami technique, utilizing hundreds of short oligo staple strands to direct the folding of a 7-kilobase single-stranded circular DNA molecule from the bacteriophage M13mp18 into predesigned shapes via one-pot annealing5. The DNA origami technique has enabled the programmable fabrication of diverse nanostructures, ranging from geometric forms (triangles, hexagons) to artistic patterns (smiley faces, vases) and complex macroscopic representations like the map of China6,7.

DNA origami structures have been constrained by the inherent limitations of scaffold length. To address the growing demand for higher-order architectures capable of executing sophisticated functions, various attempts have been made to scale up DNA origami structures. Qian et al. demonstrated dynamically programmable 2D/3D self-assembly using symmetry-optimized triangular DNA origami tiles, and also developed a programmable DNA tiling framework that used local rules to control global pattern formation, constructing random loops and mazes on DNA origami arrays8,9. Chen et al. constructed size-controllable DNA origami rings by connecting cross-shaped DNA origami via an oblique linking strategy, and further employed an algorithm-guided hierarchical assembly strategy to achieve precise construction of multi-scale DNA arrays with minimum pairs of DNA connection strands10,11. Moreover, supramolecular DNA assemblies were also fabricated using wireframe origami designed via METIS with parallel crossovers between adjacent DNA origami units, or using chemically conjugated branched staples12,13. Recently, Wang et al. developed DNA origami building block pieces (DOPBs) with eight mutually independent programmable edges and implemented an instruction-driven strategy to achieve efficient assembly of controllable 2D arrays14.

In biomedical applications, DNA nanotechnology has been utilized for drug delivery and identification of cancer biomarkers15. The detection of cancer-associated miRNAs is key for early diagnosis and treatment monitoring in cancer research16. Though RT-PCR remains the most widely used method for miRNA detection due to its high sensitivity and specificity, it involves complex and time-consuming procedures. DNA origami technology can be engineered to create multi-channel or multiplexed detection systems. For example, by extending the staple strands on DNA origami structures, multiple miRNA molecules can be captured and anchored, enabling multiplexed detection capabilities. Fan et al. designed DNA origami-based logic gates (YES and AND gate) for analyzing heart disease-related miRNA-21 and miRNA-19517. Moreover, DNA wireframe paper based molecular sensors were developed for the detection of disease-related miRNA-107 and miRNA-155 by specializing the sequence of orthogonal crease handles18.

In this study, we developed a dynamic nucleic acid detection platform that integrates triangular DNA origami modules with molecular logic gates. Clinically significant biomarkers (cDNA corresponding to miRNA-155, miRNA-182, and miRNA-197) for early lung cancer diagnosis were chosen as the model targets19. Our design leverages edge-specific hybridization sites on DNA origami to encode Boolean logic operations (YES, AND, OR), enabling target-driven hierarchical self-assembly. This architecture not only supports multiplexed detection but also achieves nanoscale resolution of assembly states via atomic force microscopy (AFM). Furthermore, toehold-mediated strand displacement introduces resettable and adaptive functionalities, allowing the system to dynamically respond to changing molecular inputs. Our work highlights the potential of DNA origami for creating programmable molecular circuits and demonstrates its application in molecular computing and biosensing.

Materials and methods

Assembly of the DNA origami triangles

The scaffold strand used in this study is single-stranded M13mp18 DNA which was purchased from New England Biolabs (Catalog # N4040S) at 250 µg/mL in 1× TE buffer (10 mM Tris-HCl, 1 mM EDTA, pH 8.0). Staple stands were purchased from Sangon Biotech and purified by the high affinity purification (HAP) method or the high-performance liquid chromatography (HPLC) method. The DNA dry powder were then dissolved in certain amount of 1× TAE/Mg2+ (40 mM Tris-acetate, 1 mM EDTA, 12.5 mM magnesium acetate) to 100 µM before use. Individual DNA origami triangles were prepared with 5 nM scaffold strand and 25 nM staple strands in 1 × TAE/Mg2+ buffer. The PCT annealing process for the scaffold and staple mixtures was performed in a thermocycler with the following process: heating to 95 °C, holding at 95 °C for 2 min, then annealing from 95 to 4 °C at 6 s/0.1 °C. The annealing time for the individual triangle origami was 90 min.

Edge design

To give the individual DNA origami triangle specific binding interactions, some staple strands along their edges were incorporated with single-stranded DNA overhangs. The design details of the YES gate, two-input AND gate, three-input AND gate, three-input OR gate are presented in Figure S1-S4. The overhang part consists of a 3-nt-long poly(T) spacer and a binding site, 11 to 12-nt-long ssDNA which is designed to be complementary to one or the other half of a certain type of miRNA stands. The addition of this miRNA strand or its cDNA strand could activate the interaction between one triangle of which the edge consisted one half of the complementary sequence and another triangle which consisted the other half of the complementary sequence of the miRNA. Thus, bringing the two triangles together and initiated the assembly. On the other hand, glue strands with the sequence of the miRNA strands and 5-nt-long toehold in their terminal were also designed to not only activate the assembly of triangles, but also allow unravelling of the assembled origamis through toehold-mediated DNA strand displacement when the releaser strands which are complementary to the glue strands were added.

Ultrafiltration

The obtained DNA origami triangles were purified from the excess staple strands by ultrafiltration. First, the 50 kDa molecular weight cutoff filters (Millipore) were soaked in 500 µL 1× TAE/Mg2+ buffer and centrifuged at 5000 g for 10 min at 4 °C. After discarding the filtrate, the DNA origamis diluted in 1× TAE/Mg2+ buffer were poured into the filter and centrifuged at 5000 g for 10 min at 4 °C. Buffer was replenished and the centrifugation was repeated twice. Finally, purified DNA origamis were collected by putting the filter upside down into a new tube and spun at 1000 g for 3 min.

AFM imaging

AFM imaging was performed in water tapping mode under 1× TAE/Mg2+ buffer on a Cypher ES Environmental Atomic Force Microscope (Oxford), using a cantilever with the force constant of 0.09 N/m. After self-assembly of the DNA origami for different periods of time (namely, 1 h to 6 h, see Figure S5), samples were prepared for AFM imaging by diluting in 1× TAE/Mg2+ buffer to the concentration of 1 nM. 10 µL of the diluted sample was then deposited onto the freshly cleaved mica surface. After 5 min of absorption, an additional amount of 1× TAE/Mg2+ buffer was added to both the mica surface and the cantilever before they were positioned in the AFM head. As for DNA structures that were unraveled, the preparation steps for AFM imaging were the same as for DNA assembly, except that the mica surface was washed several times with 1× TAE/Mg2+ buffer after the absorption of DNA samples. Compared to the individual DNA origami, there was a much larger excess of short strands (including a high ratio of both the glue strands and the releaser strands) in the samples after unraveling, which needed to be removed with the washing step to provide a cleaner background during imaging.

Results and discussion

Detection of one type of nucleic acid molecules using a YES gate

The programmable detection of nucleic acids using edge-modified triangular DNA origami was demonstrated through structural assembly and logic-gated signal generation. As outlined in Fig. 1A, the system employs two distinct triangular origami units: one functionalized with edge staples complementary to the first half of the target sequence (e.g., miRNA-182 cDNA) and the other with edges matching the second half. These units act as a YES logic gate (Fig. 1B), where the presence of the full-length target bridges the complementary edges, driving the self-assembly of a diamond-shaped nanostructure as the output. AFM imaging confirmed the formation of well-defined dimeric diamond assemblies upon introduction of miRNA-182-specific cDNA (Fig. 1C), with the enlarged view highlighting precise edge-to-edge alignment.

Quantitative analysis of AFM data (Fig. 1D) revealed a high yield of diamond assemblies (80%) for the target miRNA-182 cDNA. This work advances DNA origami-based biosensing by integrating molecular logic operations with programmable nanostructural outputs. Unlike conventional probes that rely on fluorescence or electrochemical signals, this system translates target recognition into geometrically defined, AFM-visible assemblies, providing both spatial and stoichiometric resolution. The diamond structure serves as a direct visual readout, eliminating the need for secondary labeling or amplification steps. However, the reliance on AFM for analysis limits scalability; future implementations could benefit from coupling assembly events with optical barcoding or resistive pulse sensing for high-throughput applications.

The modularity of the edge design also opens avenues for multiplexed detection. By encoding orthogonal staple sequences on additional origami units, the system could simultaneously process multiple targets, each generating distinct structural outputs (e.g., trimers, tetramers). Furthermore, the YES gate’s binary response—diamond formation versus unassembled monomers—provides a straightforward diagnostic metric for single-target detection, ideal for scenarios requiring unambiguous “presence/absence” determinations.

The success of the YES gate paradigm lays the groundwork for developing more complex molecular circuits to address intricate diagnostic challenges requiring multi-parameter analysis.

Fig. 1
figure 1

(A) Schematic representation of the edge design of triangular DNA origami for nuclei acid molecule detection and the corresponding output signal detection. (B) The logic gate symbol corresponding to a YES gate and its truth table. (C) AFM imaging of the output signals after detecting miRNA-182-specific cDNA, with an enlarged dimer detail shown in the upper-right corner. (D) The yield of AFM scanning results corresponding to the detection of different types of nucleic acid molecules.

Full size image

Simultaneous detection of two types of nuclei acid molecules via a two-input AND gate

The triangular DNA origami platform was further engineered to enable simultaneous detection of two nucleic acid targets via an AND logic-gated assembly mechanism. As depicted in Fig. 2A, three distinct sets of triangular origami were designed with edge modifications to recognize complementary regions of two target sequences (e.g., miRNA-155 and miRNA-182 cDNAs). One set of origami featured staples extended from two edges, each complementary to half of one target sequence, while the other two sets were modified on a single edge to hybridize with the remaining halves of the respective targets. Only when both target molecules were present did the three origami units undergo cooperative hybridization, resulting in the formation of a trapezoidal assembly composed of three interconnected triangles. This output signal served as a direct readout of the AND gate operation, where the system returned a “true” state (output “1”) exclusively under dual-target conditions (Fig. 2B).

AFM characterization validated the specificity of the trapezoid assembly. When both miRNA-155 and miRNA-182 cDNAs were introduced, AFM imaging revealed well-defined trimeric trapezoidal structures, with an enlarged view highlighting the geometric alignment of the three origami units (Fig. 2C). Quantitative analysis of the AFM scans demonstrated a high yield of trapezoid formation (66%) in the presence of both targets. The total assembled structures (including diamonds) accounted for ~ 86% of the observed origami, indicating efficient hybridization kinetics under dual-target conditions.

The successful implementation of a two-input AND gate underscores the modularity and scalability of DNA origami for multiplexed molecular logic operations. By partitioning the recognition sequences across three origami units, the system enforced a strict dependency on both targets to stabilize the trapezoidal output—a design that mimics biological cooperativity. This contrasts with earlier single-target YES gate systems, highlighting how geometric constraints in origami architectures can encode programmable Boolean logic. The yield of trapezoid assemblies is lower than the yield of diamonds produced by the YES gate. Moreover, the ~ 20% discrepancy between total assembled structures and correctly formed trapezoids suggests partial aggregation or incomplete hybridization, possibly due to kinetic traps or steric hindrance. Future optimizations, such as tuning staple lengths or introducing toehold-mediated strand displacement, could improve fidelity.

Fig. 2
figure 2

(A) Schematic representation of the edge design of triangular DNA origami for simultaneous detection of two nucleic acid molecules and the corresponding output signal detection. (B) The logic gate symbol corresponding to a two-input AND gate and its truth table. (C) AFM results of output signals after detecting cDNA corresponding to miRNA-155 and miRNA-182, with an enlarged detail of the trimer shown in the upper-right corner. The bar chart illustrates the yields of the trimers and the total assembled structures.

Full size image

Simultaneous detection of three types of nuclei acid molecules via a three-input AND gate

The triangular DNA origami platform was further expanded to enable three-input AND-gated detection of nucleic acid targets through hierarchical self-assembly. As shown in Fig. 3A, four distinct triangular origami units were engineered with edge-specific modifications: one central triangle featured staples extended from all three edges, each complementary to half of one target sequence (e.g., miRNA-155, miRNA-182, or miRNA-197 cDNA), while three peripheral triangles were modified on a single edge to recognize the remaining halves of the respective targets. This design enforced strict dependency on the simultaneous presence of all three targets to drive the assembly of a large triangular structure composed of four interconnected origami units. AFM imaging confirmed the formation of the programmed tetrahedral architecture exclusively in samples containing all three cDNAs (Fig. 3C), with a yield of ~ 54% for the complete four-triangle assembly. The total assembled structures (including partial intermediates) accounted for ~ 69% of observed origami, suggesting cooperative yet kinetically challenged hybridization under triple-target conditions (Fig. 3C bar chart). The system operated as a three-input AND gate (Fig. 3B), where the output signal—a geometrically defined macro-triangle—was generated only when all three inputs (miRNA-155, miRNA-182, and miRNA-197 cDNAs) were present.

The successful implementation of a three-input AND gate exemplifies the scalability of DNA origami for encoding complex molecular logic. By distributing recognition sequences across four origami units, the system required simultaneous hybridization at six independent edges (three on the central triangle and one on each peripheral unit), effectively amplifying specificity through combinatorial binding. This multi-valent approach mirrors biological systems like antibody-antigen interactions, where avidity compensates for individual binding weaknesses. However, the reduced yield of complete assemblies (~ 54%) compared to the two-input AND gate (~ 66% in Fig. 2) highlights kinetic and entropic challenges in coordinating larger numbers of components.

The AFM results (Fig. 3C) reveal the geometric precision of the macro-triangle—evident in the high-resolution AFM image (inset)—demonstrates the programmability of DNA origami for synthesizing nanometer-scale architectures with predefined symmetry. A key advantage of this three-input system lies in its potential for multiplexed diagnostics. For example, simultaneous detection of miRNA-155, miRNA-182, and miRNA-197—biomarkers linked to cancers like lymphoma and ovarian carcinoma—could improve diagnostic accuracy by reducing false positives from single-marker assays. The ~ 15% gap between total assembled structures and correctly formed macro-triangles underscores the need for kinetic optimization. Compared to earlier YES and two-input AND gates, this three-input system illustrates a trade-off between specificity and complexity. While adding inputs enhances diagnostic rigor, it also introduces combinatorial bottlenecks.

Fig. 3
figure 3

(A) Schematic representation of the edge design of triangular DNA origami for simultaneous detection of three nucleic acid molecules and the corresponding output signal detection. (B) The logic gate symbol corresponding to a three-input AND gate and its truth table. (C) AFM output signals after detecting cDNA corresponding to miRNA-155, miRNA-182, and miRNA-197, with an enlarged detail of the trimer shown in the upper-right corner. The bar chart illustrates the yields of the quadruplex and the total assembled structures.

Full size image

Detection of three types of nuclei acid molecules via a three-input OR gate

The triangular DNA origami platform was further engineered to operate as a three-input OR gate, enabling detection of multiple configurations of nucleic acid targets (miRNA-155, miRNA-182, and miRNA-197 cDNA) through geometrically programmable self-assembly. As shown in Fig. 4A, two distinct sets of triangular origami units were designed: one set featured staples extended from all three edges, each complementary to the first half of one target sequence, while the other set contained staples complementary to the second half of the same three targets. This modular design allowed the system to generate hierarchical assemblies—ranging from dimers (rhombus) to trimeric trapezoids and extended macro-triangles—depending on the number of targets present. The OR gate logic (Fig. 4B) was encoded such that the presence of at least one target triggered assembly, with structural complexity scaling with input multiplicity: rhombus (1 input), trapezoid or its derived shapes (2 inputs), or large triangle or its derived shapes (3 inputs).

AFM imaging confirmed the formation of all predicted assemblies (Fig. 4C). For example, samples predominantly formed rhombus structures (~ 29% yield), while dual-target conditions generated trapezoidal assemblies (~ 11%). The full three-input condition produced large triangular architectures, albeit with lower efficiency due to increased entropic penalties (Fig. 4D). Notably, the total assembly yield (all structures combined) reached ~ 55% in three-input conditions, demonstrating robust hybridization activity despite competing kinetic pathways. The inverse correlation between input number and yield highlights kinetic limitations in coordinating multi-edge hybridization.

The structural diversity observed in AFM images (Fig. 4C) also underscores the programmability of DNA origami as a nanoscale “molecular breadboard.” By reconfiguring edge sequences, the system could be adapted to detect alternative targets while retaining the OR gate’s logical framework. However, current reliance on AFM for readout limits real-time applicability. Compared to earlier YES/AND gates, this three-input OR gate exemplifies a trade-off between functional complexity and operational robustness. While it enables richer diagnostic information, the reduced yield of high-order assemblies necessitates careful balancing between specificity and sensitivity. Future iterations could hybridize AND and OR logic (e.g., “at least two out of three targets”) to prioritize clinically relevant biomarker combinations.

Fig. 4
figure 4

(A) Schematic representation of the edge design of the triangular DNA origami used for detecting multiple configurations of three types of nucleic acid molecules and the detection of output signals. (B) The schematic diagram of a three-input OR gate and its truth table. (C) AFM images of output signals after detecting cDNA corresponding to miRNA-182, miRNA-155, and miRNA-197, with enlarged views of various assemblies on the right side. (D) The yield distribution of different assemblies.

Full size image

Reversibility of the triangular DNA origami assembly

The dynamic reconfigurability of the DNA origami assembly system was demonstrated through controlled disassembly of a preformed dimer structure, as illustrated in Fig. 5A. The dimer, composed of two triangular origami units linked via toehold-modified staples complementary to TmiRNA-182, was designed to undergo strand displacement upon introduction of a full-length cDNA trigger. AFM imaging confirmed the stability of the dimer in the absence of the trigger. Upon adding the cDNA trigger, which competitively hybridized to the toehold region and displaced the inter-origami staples, the dimer disassembled into monomeric triangles within 6 h. This programmable disassembly mechanism underscores the system’s potential for applications requiring spatiotemporal control, such as cargo release in drug delivery or resetting molecular circuits. The toehold-mediated strategy mimics biological processes like enzyme-substrate dissociation, where localized destabilization enables precise structural deconstruction.

The ability to reverse origami assemblies on demand represents a critical advancement over static nanostructures. Unlike previous AND/OR gate systems, which generate irreversible outputs, this design integrates both forward and reverse operations, enabling cyclic computation. For example, in biosensing, disassembly could reset the system for repeated target detection, while in nanofabrication, it could enable error correction during hierarchical assembly. A notable limitation lies in the energy landscape of multi-domain origami interactions. While toehold length and placement govern disassembly efficiency, steric constraints imposed by the origami’s rigid structure may impede full strand separation.

In summary, this work establishes DNA origami as a dynamically tunable platform, bridging the gap between structural programmability and functional reconfigurability. By coupling toehold-mediated logic with macroscopic structural changes, the system opens avenues for adaptive nanomaterials and responsive diagnostic devices.

Fig. 5
figure 5

(A) Schematic representation of the disassembly process of a dimer composed of two DNA origami triangles. (B) AFM images of the DNA dimer connected via TmiRNA-182 with toehold and disassembled via the corresponding complementary sequence.

Full size image

Conclusion

This study establishes a programmable framework for nucleic acid detection by integrating DNA origami nanostructures with molecular logic gates, thereby advancing the frontier of dynamic molecular computation. Triangular DNA origami tiles, engineered as reconfigurable computation modules through edge-specific hybridization sites, were systematically demonstrated to emulate Boolean logic operations—including YES, AND, and OR gates—via target-driven hierarchical assembly. As a proof of concept, this work successfully detected cDNA corresponding to three lung cancer-associated miRNAs, demonstrating the clinical relevance of the platform in biomarker analysis. By coupling the programmability of DNA nanostructures with molecular recognition logic, this work bridges structural nanotechnology with biocomputing, enabling the design of autonomous systems that interpret biological inputs through predefined algorithms. The modular architecture of these origami platforms permits seamless scalability to multi-layered logic circuits, while AFM provides nanoscale-resolution visualization of computational outcomes. Notably, the incorporation of toehold-mediated strand displacement achieved dynamic disassembly of assemblies, extending functionality toward reusable, resettable systems with adaptive feedback capabilities. By harmonizing nanoscale precision with biological signal processing, this platform lays the groundwork for transformative applications in precision diagnostics, synthetic biology, and adaptive nanomedicine.