DNA Triangular Pyramid Frustum Supported miR-21 Assay with Duplex-Specific Nuclease-Assisted Target Recycling

Abstract

MicroRNAs (miRNAs) have the potential to be applied as effective biomarkers for early diagnosis of cancers. Electrochemical techniques exhibit advantages such as high sensitivity and ease of miniaturization. However, electrode interface perturbations may hinder electrochemical responses. To address the limitation, three-dimensional DNA triangular pyramid frustum nanostructures are designed on the surface of the electrode to support miRNA (miR-21 as an example) recognition and following target recycling. Due to the combination of DNA nanostructures and enzyme-mediated signal amplification, the sensitivity and selectivity of this electrochemical biosensor are enhanced. It also performs satisfactorily in human samples, which meets clinical detection requirements. Overall, the strategy provides new possibilities for accurate and reproducible miRNA assays and shows great potential applications in early disease diagnosis.

MicroRNAs (miRNAs) are a class of endogenous noncoding RNAs with essential functions in the modulation of gene expression. , They exhibit great potential to be used as biological indicators of various diseases. , Currently, there are numerous evidence confirming that mutations or abnormal expressions of miRNAs are closely associated with the states of various diseases including cancers. , For instance, miR-21 levels are determined to be repetitively differentiated between healthy participants and lung cancer patients. Due to the limited expressions in clinical samples, researchers have developed different amplification based techniques for the detection of miRNA, such as traditional quantitative reverse transcription PCR (qRT-PCR) and loop-mediated isothermal amplification (LAMP). , Fluorescence-based nucleic acid probes are usually applied with the merits of high sensitivity, convenient operation, and good biocompatibility. However, the primer designs are relatively complicated. Other limitations include the requirements of stringent laboratory conditions, time-consuming reverse translation procedure, expensive equipment, and reagents.

Electrochemical techniques hold great promise to serve as solutions for point-of-care (POC) testing of miRNA. The inherent advantages include ease of miniaturization and facile signal amplification. , Abnormal changes in physical or chemical properties always exist in the region where the electrode contacts the electrolyte. , Such perturbation may be attributed to the adsorption of impurities, interfacial corrosion or external physical stimuli, which hinder the finally obtained electrochemical responses. , For applications in biological systems, adsorption of biomolecules, such as proteins, and cell adhesion should be strictly controlled. Currently, there are several ways to address the problem. For example, the employment of certain molecular layer (e.g., PEG) on the electrode leads to steric hindrance or hydrophilic interactions, which may be used to build the antiadsorption barrier. However, these additional layers may affect the fast charge transfer. DNA materials exhibit the properties of high stability and good programmability. − The double-helical structure of DNA provides potential pathways for charge transport. Therefore, the modification of DNA nanostructures on the surface of electrode can balance antiadsorption efficiency, biocompatibility, and electrochemical response.

In this study, we have designed DNA triangular pyramid frustum (DNA TPF) nanostructures on electrode interface, which support excellent spatial orientation for following reactions with antiperturbation capability. DNA TPF can capture signal probes by hybridization between complementary sequences. It also allows target miRNA induced digestion reactions in the presence of duplex-specific nuclease (DSN). After the collapse of the three-dimensional DNA nanostructures, signal probes cannot be immobilized, generating a decline in electrochemical responses. The innovative sensing strategy provides new possibilities of reproducible and reliable POC testing of miRNA, which exhibits great potential in the applications in early cancer diagnosis.

Detailed assembly principle of DNA TPF is shown in Scheme . Taking miR-21 as a target example, the sequences of the fuel strands are carefully designed (Table S1). Generally, three single-stranded DNA probes (TPF1, TPF2, and TPF3) partially hybridize with each other, forming the basic three-dimensional scaffold (Figure S1). The further-introduced TPF4 strand fixes the bottom segments, making DNA TPF nanostructure very stable. In addition, the three thiol groups modified at the 5′ terminals of TPF1, TPF2, and TPF3 are localized at the vertexes of the bottom triangle. Thus, the synthetic DNA TPF with multiple anchoring sites can be firmly modified on the electrode through gold–thiol covalent bonding. Next, another three independent DNA strands are introduced, which are labeled with Methylene Blue (MB) as the electrochemical species. DNA TPF nanostructures can capture these signal probes with enhanced molecular recognition efficiency and antiperturbation capability. Intense electrochemical responses are thus reported. However, in the presence of target miRNA, DNA/RNA duplex forms between miRNA and the single-stranded region on top of DNA TPF. The duplex can be recognized by DSN and the DNA strand is cleaved by the enzyme. miRNA is thus released for additional reaction cycles, leading to the collapse of multiple DNA TPF nanostructures. Since they lose the top triangle sequences to immobilize signal probes, the significantly declined electrochemical responses are varied with the initial concentration of target miRNA.

1. Illustrations of the Assembly of DNA TPF and the miRNA Sensing Strategy.

Atomic force microscopy (AFM) characterization is first conducted. The bright dots in the image are ascribed to the formed DNA nanostructures (Figure S2). Polyacrylamide gel electrophoresis (PAGE) is then carried out to reveal the DNA assembly (Figure S3a). Lanes 1–4 represent the bands of mixtures of one, two, three, and four DNA fuel strands of DNA TPF. The migration rate becomes slower with more types of fuel strands, indicating successful hybridization reactions. After further addition of miRNA or DSN, the band barely changed, demonstrating the nanostructures are not changed (lanes 5 and 7). However, after incubating DNA TPF with both miRNA and DSN, a new band appears with significantly decreased migration rate (lane 6). The reason is that the cleavage reaction leads to the generation of multiple branches, which impedes the migration of the DNA nanostructures in the gel. The captures of signal probes can also be evidenced by PAGE (Figure S3b). The band positions of TPF1/S1, TPF2/S2, and TPF3/S3 are slightly higher than those of TPF1, TPF2, and TPF3, respectively. These observations demonstrate the hybridization reactions well, as expected.

The modifications of the electrodes are then characterized by the techniques of electrochemical impedance spectroscopy (EIS) and cyclic voltammetry (CV), respectively. Nyquist diagrams of electrodes after stepwise modifications of DNA TPF, and probes S1/S2/S3 in the absence and presence of miRNA/DSN, are depicted in Figure a. The semicircle domain in a typical Nyquist diagram usually represents the degree of the charge transfer resistance of the electrode interface. After DNA TPF modification, due to the repellant between the DNA monolayer and the electrochemical species of [Fe­(CN)₆]^3–/4–, an obvious semicircle domain appears, compared with that of bare gold electrode. Since the top single-stranded DNA triangle contains partially complementary sequences of probes S1, S2, and S3, the semicircle domain of the Nyquist diagram becomes even larger after the introduction of the three probes. However, miRNA/DSN treatment is able to digest the capture sequences, leading to a decrease in DNA density on the electrode. Meanwhile, probes S1, S2, and S3 cannot be immobilized. As a result, the semicircle domain of the corresponding diagram becomes smaller than that of the DNA TPF modified electrode. Peak intensities of cyclic voltammograms can also reflect the blocking degree of the modified electrode. As shown in Figure b, a pair of well-defined CV peaks can be observed for the bare electrode. With stepwise modifications of DNA TPF and probes S1, S2, and S3, the intensities of the peaks decrease. In the presence of miRNA and DSN, DNA TPF nanostructures collapse, and the electrode interface becomes less negative to repel [Fe­(CN)₆]^3–/4–, leading to the increase of the peak currents. The variations of EIS and CV curves are consistent with each other. Since the signal probes are labeled with MB, square wave votammograms (SWV) are recorded and compared. As shown in Figure c, there are no significant current peaks for the bare electrode and the DNA TPF-modified electrode. After the immobilization of probes S1, S2, and S3, an obvious current peak appears at around −0.3 V, which is ascribed to the localized MB. In addition, the peak current can be significantly reduced with the treatment of miRNA/DSN, which confirms the feasibility of the constructed method for miRNA quantification. Since the top triangle of DNA TPF contains three capture sequences, we have tested the use of one, two, and three signal probes (Figure S4). Apparently, the electrochemical response for the S1/S2/S3 group is the largest, which is thus utilized for the following experiments (Figure d).

1.

(a) Nyquist diagrams, (b) cyclic voltammograms, and (c) square wave voltammograms of the bare electrode before and after modification with DNA TPF, S1/S2/S3 in the absence and presence of miRNA/DSN reactions. (d) SWV peak currents of DNA TPF modified electrodes after further treatments with S1, S1/S2, and S1/S2/S3, respectively. Statistical analysis is performed using a Student’s t-test ((**) p < 0.01, (****) p < 0.0001).

By comparing the SWV peak currents, the experimental conditions are optimized including 0.8 μM DNA TPF, 1.2 μΜ probe S, and 90 min DSN digestion time (Figures S5, S6, and S7). To investigate the analytical performance of the proposed strategy, a series of concentrations of standard miR-21 solutions are prepared, which are used to trigger the digestion reaction and DNA TPF collapse. The corresponding SWV curves are overlaid in Figure a. In the absence of target miR-21, the peak current is the largest. With the increase in miR-21, the peak intensity gradually decreases, demonstrating less-immobilized signal probes. The recorded peak current shows a good linear correlation with logarithm of miR-21 concentration (Figure b). The fitting equation is as follows:

2.

(a) Square wave voltammograms for the analysis of miRNA with a series of concentrations. (b) The linear relationship between the peak current and the logarithm of miRNA concentration. (c) Decreased peak currents in the cases of target miRNA and mismatched sequences. Statistical analysis is performed using a Student’s t-test ((****) p < 0.0001). (d) The electrochemical responses of the method for the detection of various levels of miR-21 in PBS and serum samples.

$$y=-9.392-1.191x(n=4,R^2=0.992)$$ 1

in which y is the peak current (μA) and x represents the logarithm of miR-21 concentration (M). The detection limit is calculated to be 0.03 fM. The linear range is quite wide, and the sensitivity is comparable or higher than most previously reported methods (Table S2). The selectivity of this work is then examined by analyzing the results of six mismatched sequences. Single-base or double-base mismatches are designed and synthesized, which are introduced in the electrochemical system under the same experimental conditions. As shown in Figure c, the decreased current responses at −0.3 V for mismatched cases are negligible, compared to that of target miR-21. After mixing miR-21 with these strands, the responses are comparable with the single miR-21 case (Figure S8). This method is thus demonstrated to be highly selective, which successfully distinguishes target from single-base mismatched sequences. Serum usually contains abundant proteins (e.g., albumin and globulins), metabolites and circulating nucleic acids, which poses significant challenge for the antifouling property of modified electrode. Standard miR-21 solutions are spiked into three independent serum samples, and SWV responses are compared with those in PBS buffers. As shown in Figure d, good consistent values are obtained, demonstrating that complicated biological environments do not disturb the electrochemical sensing interface with the DNA TPF layer.

To verify the practicality of this method in clinical samples, a total of 32 tissue samples were collected from nonsmall cell lung cancer (NSCLC) patients (16) and healthy controls (16) in Suzhou Hospital, Affiliated Hospital of Medical School, Nanjing University (Suzhou, China). Total RNAs are extracted and diluted (10-fold) with 10 mM Tris-HCl before loading into the electrochemical biosensor. miR-21 concentrations in these samples are then calculated according to the obtained SWV peaks and the established equation. The values corresponding to patient and control groups are depicted as high and low hits in the heat maps (Figures a and b). The two datasets are normally distributed (Figure S9). It is obvious that larger concentrations of miR-21 are observed in patients’ samples with a statistically significant difference (p < 0.0001). Thus, the electrochemical response alerts the possibility of cancer disease (Figure c). The obtained miRNA concentrations by the electrochemical approach are also compared with qRT-PCR results. As validated in Figure d, the miRNA concentrations by the two methods are in good agreement. The median levels for the patient group are 65.18 pM (electrochemical method) and 65.58 pM (qRT-PCR). This evidence suggests promising prospects of this method in accurate diagnosis of miRNA-related diseases.

3.

Heat maps of miR-21 analysis by the proposed electrochemical method challenging samples from (a) patients and (b) healthy individuals. (c) Box plot of the calculated miR-21 levels in the samples from patients and healthy individuals. Statistical analysis is performed using a Student’s t-test ((****) p < 0.0001). (d) Comparison of miR-21 quantifications by standard qRT-PCR and the proposed electrochemical method.

In summary, an electrochemical sensing strategy for miRNA is successfully constructed with a rigid and stable DNA TPF as the supporting platform. The three-dimensional nanostructure promises improved DNA hybridization efficiency and declined adsorption of interfering molecules on the electrode interface. Meanwhile, DSN-assisted target recycling occurs for the collapse of the DNA nanostructures, leading to variation of the electrochemical responses. This method achieves highly sensitive and selective detection of miR-21 in not only standard buffer conditions but also real human samples. It holds great potential for miRNA-related early cancer diagnosis and can also be adapted to determine additional targets by facile modification of the recognition sequences.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacsau.5c00597.

• Experimental details including materials, methods and protocols; free energies of secondary structures of DNA probes; AFM image of DNA nanostructures; PAGE analysis of DNA hybridization reactions; SWV curves of DNA modified electrodes after treatments with different signal probes; optimizations of DNA TPF concentration, S1/S2/S3 concentration and DSN reaction time; decreased peak currents in the cases of target miRNA in the absence and presence of mismatched sequences; quantile–quantile plot; RNA and DNA sequences; comparison of the analytical performances of recent miRNA assays (PDF)

#.

Authors H. Chai and L. Yang contributed equally. CRediT: Hua Chai data curation, formal analysis, investigation, writing - original draft; Li Yang data curation, investigation, writing - original draft; Xiaolin Qu investigation, validation; Peng Miao conceptualization, funding acquisition, supervision, writing - review & editing.

The authors declare no competing financial interest.

Published as part of JACS Au special issue “DNA Nanotechnology for Optoelectronics and Biomedicine”.