2.1. Reagents

Thiol-modified DNA was purchased from the Japan Gene Research Laboratories (Sendai, Japan). LNA-modified DNA and cyanine 5 (Cy5)-modified DNA were purchased from Integrated DNA Technologies, Inc. (Coralville, IA, USA). miRNAs were obtained from FASMAC Co., Ltd. (Kanagawa, Japan). Unmodified DNA and DNA containing phosphorothioate linkages were purchased from Eurofins Genomics (Tokyo, Japan). The sequences of all oligonucleotides used in this study are listed in Table 1 (in Appendix A).

AuNP solution (20 nm in diameter), borate buffer, and sodium chloride (NaCl) solution were purchased from Nanopartz Inc. (Loveland, CO, USA), Thermo Fisher Scientific (Waltham, MA, USA), and Promega Corp. (Madison, WI, USA), respectively. Phosphate buffer powder (pH 7.2) (FUJIFILM Wako Pure Chemical Corporation, Tokyo, Japan) and polyethylene glycol 20000 powder (Tokyo Chemical Industry Co., Ltd., Tokyo, Japan) were dissolved in ultrapure water (Thermo Fisher Scientific Inc., USA).

The human breast cancer cell line MCF-7 was obtained from the RIKEN BRC Cell Bank (Tsukuba, Japan). Dulbecco’s modified Eagle’s medium (DMEM) and fetal bovine serum (FBS) were purchased from FUJIFILM Wako Pure Chemical Corporation (Osaka, Japan). Opti-MEM, trypsin/EDTA (0.25%), and Lipofectamine-2000 were purchased from Gibco (Gaithersburg, USA), Nacalai Tesque, Inc. (Kyoto, Japan), and Invitrogen (Carlsbad, CA, USA), respectively.

2.2. Preparation of DNA-AuNP Nanomachines

DNA-AuNPs were prepared according to previously reported methods ^16,28. First, the deprotected thiol-modified oligonucleotides were added to the AuNP solution at a molar ratio of \(1000:500:1\) (\(\textrm{A}:\textrm{L}:\textrm{AuNP}\)). Subsequently, borate buffer was added and the mixture was incubated for 15 min with gentle shaking. The NaCl concentration was then gradually increased from 0.05 M to 0.5 M using 5 M NaCl. After each NaCl addition, the nanoparticle solution was sonicated for approximately 10 s and stirred for 15 min. Finally, the solution was incubated overnight with shaking.

Excess oligonucleotides were removed by repeated centrifugation at \(5000 \times {g}\) for 90 min four times. After removing the supernatant, the pellet was resuspended in phosphate buffer. The final concentration of DNA-AuNPs was determined by UV-visible spectroscopy (UV-2600i; Shimadzu Corporation, Tokyo, Japan).

PBS (1\(\times\)) containing B (final concentration, 1 μM) and C (final concentration, 0.6 μM) (final MgCl\(_2\) concentration, 2.5 mM; final NaCl concentration, 50 mM) was added to the DNA-AuNP solution (final concentration, 0.8 nM). The mixture was heated at 65°C for 15 min and then incubated overnight at room temperature in the dark. After incubation, the resulting solution was washed three times by centrifugation at \(5000 \times {g}\) for 90 min at 4°C using 1\(\times\) PBS.

2.3. In Vitro Evaluation of DNA-AuNP Nanomachine Reactions

Reaction solutions were prepared such that the final concentrations of DNA-AuNPs (A/B/C-Cy5), W-DNA, and DNA strand F were 100 pM, 5 nM, and 320 nM, respectively. While incubating at 37°C, the fluorescence intensity for blank samples and target miRNA was monitored for approximately 20 h using a microplate reader (Spark, Tecan Group Ltd., Zurich, Switzerland).

2.4. Intracellular Imaging

The transfection reactions were performed as previously described ¹⁶. An aliquot of Lipofectamine-2000 was diluted in Opti-MEM (solution I). Cy5-labeled DNA-AuNPs assembled with A/B1/C1 (solution II) and a mixture of nucleic acid strand F and W-DNA probes (solution III) were diluted using Opti-MEM. After incubation at room temperature for 10 min, half of solution I was mixed with solution II, and the remaining half of solution I was mixed with solution III. After incubation at 37°C for 20 min, the two mixtures were combined, diluted in DMEM containing 10% FBS, and added to the cells. In the final mixture, the concentrations of Cy5-labeled DNA-AuNPs, DNA strand F, and W-DNA probes were 0.1 nM, 20 nM, and 10 nM, respectively. The cells were incubated at 37°C in a humidified incubator. After washing them twice with 1\(\times\) PBS, fluorescent images of the cells were acquired using a fluorescence microscope (Eclipse Ti, Nikon Corporation, Tokyo, Japan) or a confocal microscope (TCS-SP8, Leica Microsystems GmbH, Wetzlar, Germany). Fluorescence images were analyzed using ImageJ software (NIH). Fluorescence-positive cells were identified using ImageJ software (NIH). The fluorescence channel was thresholded using a fixed intensity range (60–115 nm) to exclude background signals. Regions of interest corresponding to fluorescent signals were extracted using the “Analyze Particles” function with a minimum size threshold of 2000 pixels\(^2\). Cells exhibiting intracellular fluorescence signals above this threshold were classified as fluorescence-positive. The total number of cells was manually counted from the corresponding bright-field images, and the percentage of fluorescence-positive cells was calculated.

3.1. Design of LNA-Modified DNA-AuNP Nanomachines

Figure 1 illustrates the design and operating principles of the entropy-driven DNA-AuNP nanomachines used in this study. The nanomachine consisted of AuNPs functionalized with DNA probes (L-DNA probes) and fluorescence-modified DNA strands (C-Cy5) together with W-DNA probes (walking legs) and a fuel strand (F) in solution. The DNA substrate on the AuNP surface is composed of a three-stranded complex (A/B/C-Cy5), where strand A is immobilized on the AuNP surface and strand B acts as an intermediate strand bridging A and C-Cy5. In this system, the fluorescence of C-Cy5 was quenched because of its proximity to the AuNP surface, and the fluorescence is restored by increasing the distance between the fluorophore and the nanoparticle. When the target miRNA is present, the nanomachine is activated through a strand displacement process mediated by short single-stranded regions (toeholds) within the DNA structure, resulting in the continuous release of C-Cy5 and recovery of the fluorescence signal (Fig. 1(A)).

Figures 1(B) and 7 (in Appendix A) show how the miRNA, W-DNA probe, and L-DNA probe interact near the nanoparticle surface. The L-DNA probe contains a sequence that is partially complementary to the target miRNA, anchoring the miRNA close to the nanoparticle. The W-DNA probe also binds to the same miRNA, forming a ternary complex that brings the W-DNA into proximity to the AuNP surface. This proximity enables W-DNA to initiate strand displacement on the nanoparticle surface, where W-DNA first binds to an exposed toehold region (toehold 1), displacing strand B and exposing a second toehold (toehold 2). Subsequently, the fuel strand binds and hybridizes to the exposed region, triggering branch migration and complete disassembly of the DNA structure. During this process, the W-DNA strand regenerates and participates in the subsequent reactions to form a catalytic cycle that enables repeated signal amplification. Consequently, the fluorescence previously quenched by the AuNP surface is restored. This catalytic cycle repeats as W-DNA “walks” along the AuNP surface, resulting in continuous signal amplification. This intramolecular reaction on the AuNP surface proceeds more efficiently than conventional intermolecular reactions because of the high local concentration of DNA components.

In this study, LNA modifications were introduced into DNA probes that formed complementary duplexes with miRNAs to enhance the double-strand stability under cytosol-mimicking conditions. Specifically, four types of W-DNA probes containing LNA modifications in the miRNA-complementary region were designed and synthesized: probes modified with two LNA bases at the 5\('\) end, two LNA bases in the middle region, two LNA bases at the 3\('\) end, and a probe modified with two LNA bases at both termini (Table 1 in Appendix A). To examine the effect of the number of LNA modifications, two types of W-DNA probes containing three LNA modifications in the miRNA complementary region and two types of L-DNA probes containing three or four LNA modifications were designed and synthesized (Table 1).

3.2. Evaluation of the Effect of LNA Modification on DNA-AuNP Nanomachine Performance under Intracellular-Like Conditions

First, the fluorescence response of the non-modified DNA-AuNP nanomachine was evaluated in the presence and absence of target miRNA-21 (miR-21) under cytosol-mimicking environments (100 mM K\(^{+}\), 10 mM Na\(^{+}\), 0.5 mM Mg\(^{2+}\), and 10% PEG20000) (Fig. 2(A)). Fluorescence recovery was negligible even in the presence of the target miRNA, suggesting a significant loss of function under physiologically relevant conditions. To overcome this issue, experiments were conducted under cytoplasm-mimicking conditions using four types of W-DNA probes with LNA modifications introduced into the miRNA-complementary region: three probes containing two LNA nucleotides at different positions and one probe containing four LNA nucleotides. As illustrated in Fig. 2(B), the LNA-modified DNA-AuNP nanomachines demonstrated a marked improvement in signal generation, reaching a plateau at approximately 400 min, and showing a clear fluorescence signal difference compared to the blank. Furthermore, the reaction kinetics were slightly affected by the position of LNA incorporation (Fig. 8 in Appendix A). However, an increase in the number of LNA modifications did not significantly accelerate this response.

Next, we systematically analyzed the effects of introducing LNA modifications into both W-DNA and L-DNA probes. Fig. 2(C) summarizes the experimental results of the effects of the position and number of LNA modifications on the performance of the DNA-AuNP nanomachines targeting miR-21. \(\Delta\)F represents the fluorescence change, defined as the fluorescence intensity measured in the presence of miRNA minus that measured in its absence, evaluated at the final measurement point. The use of LNA-modified DNA probes resulted in a substantial improvement in performance under cytosol-mimicking conditions. Notably, the extent of this improvement strongly depended on the specific combination of modifications of the two DNA probes. Interestingly, when LNA modifications were introduced at the 3\('\) terminus or the central region of L-DNA, their combination with an LNA modification at the 5\('\) terminus of W-DNA led to a reduced \(\Delta\)F compared with combinations involving modifications at other positions. The decrease in \(\Delta\)F was attributed to an increase in nonspecific fluorescence intensity. Collectively, these results demonstrate that LNA modification is an effective strategy for restoring the function of DNA-AuNP nanomachines under cytosol-mimicking conditions; however, its effect is highly position-dependent. Excessive nonspecific stabilization caused by dual modifications may compromise the specificity.

Furthermore, we examined the effect of the number of LNA modifications introduced into DNA probes (Fig. 3). As shown in Fig. 3(A), pronounced differences in the final \(\Delta\)F were observed depending on the modification position. In the case of W-DNA-modified with LNA at the 5\('\) terminus, a significant increase in fluorescence was observed even in the absence of the target (Fig. 3(B)). As a result, \(\Delta\)F was found to approach zero after approximately 20 h. Therefore, to achieve a robust performance in miRNA detection, optimization of both the number and positional arrangement of LNA modifications is essential. Notably, the requirements for LNA modification are not universal and should be evaluated based on the target miRNA. We also evaluated the effects of introducing LNA on DNA probes targeting miRNA-155 (miR-155). Consequently, even when unmodified DNA probes were used, a target miRNA dependent increase in fluorescence intensity was observed under cytosol-mimicking conditions and the introduction of LNA did not lead to a significant increase in \(\Delta\)F (Fig. 9 in Appendix A). Depending on the target molecule, this variation may be attributed to differences in miRNA sequences. The L-DNA probe for miR-155 has an approximately 1.5-fold higher GC content than that for miR-21, indicating greater duplex stability. Therefore, DNA-AuNP nanomachine targeting miR-155 can likely function under cytosol-mimicking conditions, even without additional stabilization of the duplex provided by LNA modifications.

3.3. Evaluation of Sensitivity

The sensitivity of the proposed LNA-modified DNA-AuNP nanomachine was also evaluated. The fluorescence intensity was monitored over time for miR-21 concentrations ranging from 0 to 5 nM, revealing a clear concentration-dependent increase, which demonstrates the ability of the nanomachine to respond proportionally to target levels (Fig. 4(A)). The fluorescence intensity at the final measurement point was plotted against miR-21 concentration, yielding a strong linear correlation (\(R^2 = 0.987\)) (Fig. 4(B)). Based on this standard curve, the limit of detection was calculated to be approximately 10 pM (3\(\sigma\)b/slope, where \(\sigma\)b is the standard deviation of blank samples) using the standard curve, which corresponds to approximately six copies of miRNA per 10 μm cell.

3.4. Evaluation of Selectivity

The selectivity of the proposed DNA-AuNP nanomachine toward different miRNA sequences was evaluated using miR-21 (target), miR-10b, miR-195, miR-M1 (single-base mutation), miR-M2 (two-base mutation), and a control sample without miRNA (Fig. 5). The nanomachine clearly discriminated miR-21 from miR-10b and miR-195, demonstrating high sequence specificity for these targets. In contrast, miR-M1 and miR-M2, which contain one or two base mutations relative to miR-21, exceeded the threshold and exhibited considerable variability, indicating limited discrimination for closely related sequences and highlighting the need for further optimization of probe design.

3.5. Intracellular Performance of DNA-AuNP Nanomachines for miRNA Imaging

Finally, to evaluate the effectiveness of the DNA-AuNP nanomachines for miRNA imaging in living cells, DNA-AuNP nanomachines complexed with lipofectamine, a cationic liposome-based transfection reagent, were co-incubated with human breast cancer cells (MCF-7), which are known to highly express miR-21. Based on confocal z-stack imaging, the fluorescence was not restricted to the cell surface but was distributed within MCF-7 cells, exhibiting behavior consistent with cellular internalization (Fig. 10 in Appendix A). Furthermore, compared with unmodified DNA-AuNP nanomachines (Fig. 6(A)), the use of LNA-modified DNA-AuNP nanomachines resulted in an increased number of cells exhibiting detectable fluorescence signals, suggesting that LNA modification is effective in enhancing intracellular performance (Fig. 6(B)). The total percentages of fluorescence-positive cells were (A) 4/298 cells \(= 1.3\%\) and (B) 12/317 cells \(= 3.8\%\) (Fig. 11 in Appendix A).