Highly Sensitive Fluorescence Assay for miRNA Detection: Investigation of the DNA Spacer Effect on the DSN Enzyme Activity toward Magnetic-Bead-Tethered Probes.

Researchers have recently designed various biosensors combining magnetic beads (MBs) and duplex-specific nuclease (DSN) enzyme to detect miRNAs. Yet, the interfacial mechanisms for surface-based hybridization and DSN-assisted target recycling are relatively not well understood. Thus, herein, we developed a highly sensitive and selective fluorescent biosensor to study the phenomenon that occurs on the local microenvironment surrounding the MB-tethered DNA probe via detecting microRNA-21 as a model. Using the above strategy, we investigated the influence of different DNA spacers, base-pair orientations, and surface densities on DSN-assisted target recycling. As a result, we were able to detect as low as 170 aM of miR-21 under the optimized conditions. Moreover, this approach exhibits a high selectivity in a fully matched target compared to a single-base mismatch, allowing the detection of miRNAs in serum with improved recovery. These results are attributed to the synergetic effect between the DSN enzyme activity and the neutral DNA spacer (triethylene glycol: TEG) to improve the miRNA detection's sensitivity. Finally, our strategy could create new paths for detecting microRNAs since it obliterates the enzyme-mediated cascade reaction used in previous studies, which is more expensive, more time-consuming, less sensitive, and requires double catalytic reactions.

Introduction

Various nanomaterial-based detection platforms have been developed for microRNA detection with improved sensitivity and selectivity.[1,2] Among these nanomaterials, magnetic beads, which have attracted conspicuous interest as a purification and concentration tool, have been effectively harnessed for a variety of sample types and analytes.[3] The exploitation of magnetic beads (MBs) for studying DNA hybridization dates back to the 1980s; since then, a plethora of biosensors combining MBs and DNA have been reported.[4,5] Yet, microRNA detection has remained challenging due to its tiny size and low abundance.[6] Therefore, there is an ultimate need for a signal amplification strategy to boost the detection of target microRNAs in real samples. Of all of the amplification techniques, enzyme-assisted detection strategies have been widely used,[7] which have gone through all of the demanding clinical practice requirements, including reliability, accuracy, efficiency, and ease of use. Among the most promising enzyme-assisted amplifications used in such platforms is the duplex-specific nuclease (DSN).[8] Combining the beneficial properties of magnetic beads and the amalgamation of DSN enzymes may boost the sensitivity of microRNA detection via reducing the background noise and enriching the trace target molecules before analysis.[9] Recently, DSN has been creatively introduced to develop bioassays and biosensors such as fluorescence-based assays,[10] colorimetric assays,[11] surface-enhanced Raman spectroscopy-based sensors,[12] electrochemical sensors,[13] magnetic relaxation switch sensors,[14] and optomagnetic assays.[15] Despite these efforts, the detection limit (LOD) of these proposed assays may still not be good enough to satisfy the microRNA detection challenges. More importantly, to date, there is no conventional and systematic technique for microRNA detection settled by researchers, which combines both magnetic beads and DSN enzymes to further improve the limit of detection since the DSN cleavage-assisted surface-based hybridization displays numerous challenges. Among them, the accessibility of the complementary miRNA target and DSN enzyme may affect the performance of the proposed assays. Several factors that have been implicated in DNA-based surface hybridization have been widely addressed, including accessibility, steric encumbrance, and the surface charge generated by the immobilized DNA probe, along with the solution’s ionic strength.[16,17] Besides, we believe that the local physicochemical microenvironment of the DSN enzyme surrounding the DNA–MB complex can impact the enzyme activity at surface-bound DNA probes. Therefore, to ensure an optimized detection procedure, it is necessary to know the underlying mechanism of DSN cleavage at the MB-tethered DNA substrate, which is unclear to date. Hence, additional studies are needed to understand better the origin of the enhanced/reduced activity of the DSN enzyme.

A possible response to the query of how to well understand the enzymatic activity at MB–substrate conjugates is to perform similar studies for MBs as was done in the case of gold nanoparticles (AuNPs). Driven by the extraordinary properties of these nanoparticles, researchers have studied the interactions between enzymes and surface-bound substrates. The first paper describing the enzymatic extension of the DNA primer conjugated to AuNPs was reported 20 years back by Sheila et al. This pioneering work elucidated the impact of steric effects on hybridization and enzymatic activity. According to the authors, the extension of AuNP-tethered DNA primers by DNA polymerase and the hybridization efficiencies were mainly determined as a function of the spacer length and the primer coverage. Moreover, the extension of primers linked by the longest spacer was as effective as the solution-phase reaction.[18] More recently, Seferos et al. used a deoxyribonuclease I (DNase I) to examine the stability of spherical nucleic acid (SNA) gold nanoparticles (SNA-AuNPs). As a result, SNA exhibited a half-life that is 4.3× longer than that for free DNA with an identical sequence. The authors attributed this stability enhancement to the high local concentration of monovalent cations surrounding SNA-AuNPs, consequently inhibiting the salt-dependent DNase I enzyme.[19] The same group later demonstrated that ribonuclease H, an endonuclease enzyme that cleaves RNA in the RNA/DNA substrate, shows a 2.5 times faster cleavage on the SNA-AuNPs than the same free-substrate in solution, while serum nucleases were strongly inhibited by the AuNp-tethered substrate. Also, they tuned the DNA coverage by introducing different spacer compositions between the gold nanoparticles and the DNA probe. Surprisingly, they found that the activity of RNase H enhanced with the increase of DNA coverage. This phenomenon was attributed to the salt tolerance of the RNase H enzyme, which is relatively unperturbed by the high-salt environment associated with SNA-AuNPs.[20] On the other hand, Degliangeli et al. compared the enzymatic activity of DSN on the AuNP-tethered DNA probe and solution-dispersed DNA at the equal probe and target concentration, and they demonstrated that the reaction rate of DSN for the immobilized probe was only 1.7-fold lower than for the DNA probe in solution.[21] On the whole, these findings suggested that the enzyme’s binding and activity at the nanoparticle surface/solution interface can be significantly affected by several interfacial factors, including steric effects, adsorption, local enzyme microenvironment, and acceleration of substrate turnover.[22] However, to date, no published study has addressed a detailed understanding of DSN enzyme activity for the magnetic-bead-tethered DNA probe. Therefore, the development of coupling methods that can maintain functional oligonucleotides and tune the interfacial environment surrounding magnetic-bead-tethered DNA probes is crucial.

Inspired by previous studies, we herein first perform a systematic investigation of the interaction between the DSN enzyme and the MB-tethered substrate to elucidate the mechanism responsible for DSN enzyme enhancement and inhibition. We examine (DSN)-assisted target recycling as a function of the spacer (TEG, dT10) by which DNA probes were attached to streptavidin-coated MBs, the surface density of the immobilized probe, and the probe pair orientation (3′- or 5′-end immobilized probe). Using a fluorescence detection scheme, we show that the DSN-assisted target recycling process is determined by the local microenvironment of the DSN enzyme surrounding the MB-tethered DNA probe. In addition, the hydrolysis of the DNA probe attached by the TEG spacer led to the highest fluorescence intensity. Overall, our rationally designed amplification strategy using a neutral TEG spacer enables highly enhanced miRNA-21 (as a model) detection with a detection limit of 170 aM and exhibits a high selectivity toward one single-base mismatch.

Materials and Methods

Materials and Instruments

Streptavidin magnetic beads (1 μm, 4 mg/mL) were purchased from New England Biolabs (Beijing, China). From Sangon Biotechnology Co., Ltd. (Shanghai, China), we ordered the oligonucleotides listed in Table S1 (Supporting Information) and the DEPC-treated water (RNase-free). The DSN kit was obtained from Evrogen distributor in Guangdong, China. The RNase inhibitor (2000 U) was bought from Thermo Fisher Scientific Co., Ltd. (Shanghai, China). Human serum and Tween20 were provided by Sigma Life Science Co., Ltd. (Jiangsu, China); sodium hydroxide (NaOH), sodium dihydrogen phosphate dehydrate (NaH2PO4·2H2O), disodium dihydrogen phosphate dedocarbohydrate (Na2H2PO4·12H2O), and sodium chloride (NaCl) were purchased from KESHI, China. Tris-HCl was obtained from Biosharp China.

All of the obtained signals were acquired by the Tecan infinite M1000 PRO, monochromator-based microplate reader. A tiny volume of 18 μL of each collected supernatant was diluted 10 times with DEPC water. Then, the generated solution of 180 μL, containing the released fluorescent dye, was poured into the well of a 96-well microplate. The spectrum was collected from 560 to 650 nm under an excitation wavelength of 525 nm. All of the measurements were performed in triplicates. Regarding the ζ-potential, the test was executed using the Malvern zetasizer Nano ZS. The sample was prepared with 20 μg of MBs conjugated to a 1 μM DNA probe in a total volume of 2 mL of 1× BW buffer (dilution 100 times).

Detection Probe Immobilization onto Magnetic Beads

The dually tagged DNA detection probe conjugation to the streptavidin-coated MBs was performed according to the recommended procedure provided in the user manual of the streptavidin MBs. Briefly, the desired volume of streptavidin MBs was washed using 1× BW buffer (20 mM Tris-HCl pH 8.0, 0.5 M NaCl, 1 mM EDTA) and was discarded with the help of a permanent magnet. The supernatant was aspirated and removed, and the washed magnetic microspheres were dispersed into the 2× BW buffer. Before coupling, a required amount of preheated detection DNA probes at 95 °C for 5 min was mixed with the suspension of magnetic beads, bringing the final concentration of the capture DNA probe in the mixture to 1 μM. The obtained concoction was then incubated at 37 °C for 20 min with gentle mixing to provide better interaction between biotin and streptavidin. After that, the probe/microsphere conjugates were washed with 0.15 M NaOH to remove any nonspecifically bound probes. Next, the conjugates were rinsed two times with TT buffer (250 mM Tris-HCl pH 8.0, 0.1% Tween20), resuspended in the same buffer, and then incubated at 80 °C for 10 min, followed by decantation to remove any unstable biotin/streptavidin couplings. Finally, the complex MBs/P21 were washed three times with DEPC water and were prepared for DSN reaction.

miRNA Detection and DSN Cleavage Reaction

To achieve the highest performance of the proposed method, all of the variables that affect the reaction, including the time, temperature, binding capacity, probe design, and DSN concentration, were optimized. We used 2 pmol of DNA probes for these experiments and 1 nM miRNA in a reaction mixture of 20 μL. To detect different concentrations of miRNA, 20 μL of a DSN reaction mixture containing 20 μg of magnetic beads was conjugated to 2 pmol of DNA probes 1× DSN master buffer (50 mM Tris-HCl (pH 8.0), 5 mM MgCl2, and 1 mM DTT), 0.2 U DSN, 20 U RNase inhibitor, and different concentrations of the target at 45 °C for 120 min. After incubation in a thermocycler, once the permanent magnet separated the MBs, together with unreacted DNA detection probes, the DSN cleavage process was stopped. The supernatant then underwent fluorescence measurement.

Analysis of miRNA in the Real Sample

To determine whether the proposed method is effective or not, we checked three different concentrations of miRNA in human serum (obtained from Sigma Life Sciences). For that, we prepared four DSN reaction mixtures for different serum concentrations: 1:10, 1:100, 1:500, and 1:1000. Each experiment was done in triplicate.

Results and Discussion

Assay Principle

MicroRNA-21 (miR-21), known as a cancer biomarker, represents one of the first spotted miRNAs in human body fluids and cells. During the past decade, it has been extensively studied, and it was found that its expression is upregulated in a large range of cancers such as brain, lung, breast, and colon.[23] Moreover, it has been designated as a prevailing oncomiRNA since it targets several tumor suppressor genes.[24] Subsequently, miR-21 has been selected as a model to settle on a conventional method for microRNA detection using magnetic beads and a duplex-specific nuclease (DSN) enzyme. Recently, the duplex-specific nuclease has become one of the most explored enzymes as a nucleic acid-based amplification strategy due to its potent activity. It displays a high affinity toward DNA in homo-duplexes DNA/DNA and hetero-duplexes DNA/RNA. The DSN enzyme also has a great selective activity in discriminating fully matched from defectively matched ones.[25] Based on these advantages, we have developed our fluorescent biosensor, as shown in Scheme . Once the dually tagged capture probe conjugated to magnetic beads was hybridized with the target miRNA, the DSN enzyme cleaves the DNA probe. Then, it releases the microRNA as well as the fluorescent dyes in the solution. The discharged target remains intact, and hybridizes with another non-cleaved probe to initiate a new cycle. This cycle was repeated N times within a specified incubation time to amplify the signal because of the accumulation of dyes in the solution. The unreacted DNA probes coupled to magnetic beads were collected through magnetic separation, and the supernatant was separated out and then underwent fluorescence measurement. However, in the presence of nonmatched target miRNA, the DNA probes remain intact due to the depleted activity of DSN toward single strands of DNA, and a negligible signal was obtained. The fluorescence spectra shown in Scheme represent an efficient proof of concept. For instance, we were able to show that the fluorescence intensity and the concentration of the target are proportional, and the DSN enzyme activity is highly selective because of a slight rise of the fluorescence signal in the presence of a high concentration (1 nM) of an off-target (green) as compared to the negative control (violet).

Scheme 1

Illustration of the miRNA Detection Combining MBs and the DSN Enzyme

To the best of our knowledge, no systematic and comprehensive mechanism focusing on the DSN-assisted target recycling process has been yet provided. Additionally, because of the massive publications in this active field, a thorough explanation is needed. Herein, we investigated the effect of tuning surface chemistry on DSN-assisted target recycling. Thus, we first studied the DSN activity according to (1) the spacer composition, (2) the probe pairs’ orientation (3′- or 5′-end immobilized probe), and (3) the surface density of the immobilized probe. Then, we checked the sensitivity and selectivity of the proposed method through the detection of miR-21.

Effect of Probe Design and Concentration on DSN-Mediated Target Recycling

The attainment of high sensitivity and selectivity of the miRNA assay based on functional magnetic beads and DSN-assisted signal amplification requires reducing the steric hindrance, maximizing the accessibility, and mitigating the surface charge of MB-tethered DNA probes. These prerequisites, once conquered, will guarantee higher hybridization with the complementary miRNA, greater DSN enzyme associated with the substrate, and subsequently an enhanced DNA probe cleavage by the DSN enzyme. However, achieving these goals is quite a challenging task. The latter can be overcome by introducing spacer chains conjugated between the probe and the surface. These spacers have several effects on the hybridization efficiency of surface-tethered DNA probes, including (i) lifting the immobilized DNA probe from the surface, thus approaching the hybridization condition of the DNA probe in solution; (ii) impacting the local physicochemical microenvironment of the DNA probe; (iii) enhancing the accessibility of the DNA probe to hybridization; and (iv) reducing the nonspecific adsorption of the complementary target to the surface. While spacers’ beneficial role is generally recognized for DNA hybridization, their length and composition effect on DSN-mediated target recycling is yet to emerge.

In this context, a series of fluorescence measurements was performed to examine the influence of the spacer composition on the DSN activity at MB-tethered DNA probes. Three assays were conducted using three different DNA probes of 22 bases, with biotin labeled at the 3′ end and cy3 labeled at the 5′ end, termed 3′-biotin-TEG-probe (P1), 3′-biotin-T10-probe (P2), and 3′-biotin-probe (P3). The first probe has a 15-atom mixed-polarity triethylene glycol (TEG) spacer intercalated between biotin and DNA sequences, while the second probe has a polyanion (T10), and the last one was synthesized without a spacer. After the immobilization of 1 μM of these DNA probes onto the surface of streptavidin-coated MBs, the obtained conjugates were exposed to 1 nM miRNA-21 in the DSN buffer and 0.4 U DSN enzyme for 2 h at 45 °C. Figure A shows that the normalized fluorescence intensity ΔF% (ΔF% = (F – F0)/F0 × 100, where F0 refers to the blank fluorescence and F stands for the sample fluorescence) was 51% lower for the DNA probe with the T10 spacer than for the DNA probe without a spacer. In contrast, the probe with a TEG spacer exhibited a 42% higher signal than that without a spacer. Given these interesting observations, we quantified the surface densities of P1, P2, and P3 probes on MBs using a fluorescence-based method (Supporting Information). As shown in Table , the DNA surface densities were estimated to be 23.1011 ± 0.6, 27.1011 ± 0.1, and 30.1011 ± 0.2 DNA/μg MBs for probes modified with the T10 spacer, the TEG spacer, and the probe without a spacer, respectively. Interestingly, the beads with a moderate surface density have the highest fluorescent signal, indicating that steric hindrance is not the major factor in reducing DSN-mediated target recycling at the MB-tethered DNA probe. Thus, this unusual behavior observed here for the MB-tethered DNA probe may be attributed to other physicochemical factors rather than a steric hindrance.

Figure 1

Assessing the DNA probe design and concentration effects on the DSN reaction: (A) optimization of the probe design upon the incorporation and absence of a DNA spacer; (B) optimization of the probe end pairs’ orientation. These experiments were carried with 1 nM miR-21, 0.5 U DSN enzyme, and 1 μM probes under incubation at 45 °C for 120 min. (C) Optimization of the probe concentration for the detection of 1 nM miR-21 using a 0.5 U DSN enzyme and different probe concentrations ranging from 250 nM to 2.5 μM with incubation at 45 °C for 120 min.

Table 1

Comparison of the DNA Loading and the ζ-Potential among the Different DNA Probes

sample	probe P1	probe P2	probe P3
ζ-potential	–4.9 ± 0.3 mV	–7.3 ± 0.6 mV	–6.3 ± 0.8 mV
binding capacity	27.1011 ± 0.1 DNA/μg MBs	23.1011 ± 0.6 DNA/μg MBs	30.1011 ± 0.2 DNA/μg MBs

To probe the effect of the local physicochemical microenvironment on DSN-mediated target recycling, we measured the ζ-potential of the MB-tethered DNA probe with the T10 spacer, the TEG spacer, and without a spacer. As mentioned in Table , ζ-potential measurements have shown that the TEG-modified probe, the unmodified probe, and the probe containing the T10 spacer possess negative charges of nearly −4.9 ± 0.3, −6.3 ± 0.8, and −7.3 ± 0.6 mV, respectively. As expected, our finding revealed that the obtained fluorescence signals are reasonably well correlated with the ζ-potential measurements. Modified MBs with a more negative ζ-potential have the lowest fluorescence signal, while the highest fluorescence signal was obtained for modified MBs with a less negative ζ-potential. This behavior could be explained by considering these two facts. First, they reported that a molecular dynamics simulation approach was recently published to understand the correlation between the local ionic and the hydrolysis protein spherical nucleic acid (Pr-DNA) by DNase I.[26] It is worth mentioning that the ionic profiles surrounding the Pr-SNA have a heat map distribution of a monovalent and a divalent cation. Therefore, in our case, highly charged DNA–MB conjugates will attract more counterions toward themselves, generating a sufficiently large number of divalent cations Mg2+ surrounding the conjugates (the DSN buffer does not contain the monovalent cation Na+), which then can enhance or inhibit the enzyme activity. Second, near the DNA chain, the ion densities are strongly modified due to the diffuse counterionic cloud associated with DNA chains. As a consequence, the concentration of H+ will be affected by the DNA charge, resulting in a local pH adjustment.[27] In this context, Zhang et al. have modulated the pH profile near the surface DNA, and it was demonstrated that the local pH shifted toward a more acidic pH value by 0.3–2.8 units.[28] A similar pH profile is expected for MB–DNA conjugates. Therefore, the modulation of the local pH by the highly negative charge of DNA–MBs conjugates might have a critical effect on the enzymatic activity because the enzymes are built up with their optimal pH preferences. According to the manufacturer’s instructions, the duplex-specific nuclease is an Mg2+- and pH-dependent enzyme and requires optimal Mg2+ ion concentration and pH preferences to ensure the desired cleavage activity.[29] Therefore, highly charged conjugates with an A10 spacer and without a spacer have a high local concentration of divalent cations (Mg2+), and an increased shift in local pH toward the acidic pH value leads to a lower fluorescence signal, which may be due to the extra 10 bases in the spacer and the negatively charged surface of streptavidin-coated magnetic beads. However, the conjugate with the TEG spacer provides a better fluorescence signal due to the lower negative charge associated with DNA–MB conjugates, indicating that this spacer plays a significant role by shielding the negative charge of streptavidin-coated magnetic beads (isoelectric point of streptavidin is 6.8–7.5). This finding demonstrated that there is a strong correlation between charge and DSN-mediated target recycling that could be explained by a change in the local salt and local pH surrounding the DNA–MB conjugates, providing further support that the DSN-target recycling is affected by the local microenvironment rather than the steric hindrance, as reported previously by Shen’s group.[30]

Several research groups have studied the DSN activity from both ends. Some of these studies suggest that the DSN enzyme cleaves the DNA probes from its “5” end in the miRNA/DNA heteroduplex into a few DNA fragments.[31,32] However, other works reported that the DSN activity is performed from the “3” end.[33,34] Thus, it is useful to have a comprehensive understanding of the DSN cleavage on a solid support. In this context, we designed two DNA probes with a similar sequence; however, one was modified with the fluorescent dye Cy3 at its 5′ end and biotin/TEG at its 3′ end (P1), while the other was modified with the same dye at the 3’ end and biotin/TEG at its 5′ end (P4). After the hybridization of MB-tethered DNA probes with 1 nM of an miRNA-21 target in DSN buffer at 45 °C for 10 min, the MB conjugates were then exposed to 0.4 U DSN enzyme for 2 h. As shown in Figure B, the normalized fluorescence signals were about 500 and 350% for the “5” end and the 3′ end, respectively, which is exactly what we expected from these two reactions. This could be rationalized by the fact that in the “3” end configuration, the substrate is located within the interior of the MBs, where there is a shift in the local pH to a more acidic pH value and a higher local Mg2+, inducing a reduction in the DSN activity, as we mentioned above. It was concluded that the dye labeling position was a dominant factor of DSN-mediated target recycling.

To study the surface coverage effect, we prepared a series of MBs with different surface densities by varying the probe (P1: biotin-TEG-Probe 21) concentrations from 0.25 to 2.5 μM. After wash, we incubated these beads with 1 nM miRNA-21 target for 10 min in a DSN buffer at 45 °C, and we challenged the mixture with DSN enzyme for 2 h at 45 °C. As shown in Figure C, the normalized fluorescence signal enhanced along with the increased probe concentration and reached a maximum at 1 μM and then decreased upon further increase, indicating that the DSN-mediated target recycling correlation can be affected by the DNA surface coverage. This trend is reasonable and consistent with another study,[34] and it could be mainly attributed to two aspects. First, for the hybridization step, it has been previously shown that the probe surface density strongly affects the rate and the efficiency of nucleic acid hybridization at interfaces.[35] The DSN reaction includes two competing pathways (substrate density and local microenvironment) that occur simultaneously for DSN-mediated target recycling. In our case, low probe surface density on the magnetic bead surface leads to faster hybridization kinetics and smaller variation in the local ionic microenvironment, resulting in a low substrate density (RNA/DNA duplex) limited by the reduced DNA surface coverage, and the DSN enzyme maintains its activity. As the probe surface density gradually increased, the DSN-mediated target recycling was enhanced and reached its maximum at 1 μM; we speculate that the enhancement in fluorescence is most likely a consequence of the substrate density increase and the relatively greater variation in the local microenvironment. When the DNA probe concentration was 1 μM, the normalized fluorescence reached its maximum value, and this is probably due to the highest substrate density that can be reached. The high surface probe density conducts slow hybridization kinetics and a higher variation in the local microenvironment, inducing a low substrate density and gradually inhibiting the DSN enzyme activity. Thus, in this work condition, we chose 1 μM as the optimal DNA probe concentration for the rest of the experiments.

Optimization of the DSN Reaction

Our sensor’s characteristic for miRNA-21 detection was DSN-assisted signal amplification to enhance the detection sensitivity and improve the selectivity. Incubation time, DSN amount, and temperature are the three main parameters expected to affect the hybridization efficiency and DSN activity strongly. First, the influence of incubation temperature ranging from 40 to 55 °C was studied using 1 nM miRNA-21 with a fixed amount of DSN enzyme (0.5 U), and the reaction was carried out for 2 h. As shown in Figure A, the maximum ΔF% value was obtained at 45 °C, while a higher temperature decreased the fluorescence intensity. These findings are due to the melting temperature. In general, the melting temperature of the immobilized probe, outlined as the temperature at which half of the DNA strands are in the single-stranded (ssDNA) or the random coil state, should be slightly lower by 10–15° compared to the melting temperature in the solution. Therefore, in the case of temperatures above 45 °C, the probe/miRNA hybridization is unstable since the melting temperature of our heteroduplex is equal to 59 °C. However, for lower temperatures (below 45 °C), the normalized signal was influenced by the DSN enzyme because its optimum temperature to guarantee 100% activity is 60 °C. Next, we gauged the optimized incubation time. The reaction was performed with 1 nM miR-21 at 45 °C and in the presence of 0.2 U DSN enzyme. As shown in Figure B, the ΔF% value steadily improved with incubation time, which increased from 0 to 120 min. Nevertheless, after 120 min, a slight leveling-off of the normalized signal from 120 to 150 min was observed (less than 10%). Therefore, in the following studies, miRNA detection was carried out at 45 °C for 120 min. In addition to the incubation time and temperature, it was reported that the DSN amount could also affect the DSN reaction performance. In this vein, we have studied the effect of DSN amount on the response of 1 nM miRNA-21. The obtained result in Figure C unveils that the normalized signal increased from 0.1 to 0.4 U; then, a slight increase was observed for 0.5 U. Due to the DSN enzyme’s high cost, 0.4 U was considered the optimal DSN amount for further experiments.

Figure 2

Histograms of the DSN reaction condition optimization using 1 μM capture probes, 20 μg/mL magnetic beads, and 1 nM miRNA-21: (A) optimization of the reaction temperature (0.5 U DSN enzyme, 120 min reaction time); (B) dependence of the fluorescence signal on the DSN incubation time (0.2 U DSN enzyme, 45 °C reaction temperature); and (C) optimization of the DSN concentration (under incubation at 45 °C for 120 min).

Sensitivity of the Proposed Method toward miR-21 Detection

Via exploiting the optimum conditions, we subsequently analyzed the sensitivity of the proposed method. The obtained fluorescence spectra of different miR-21 concentrations (for 500 aM, 100 fM, 100 pM, and 500 pM) are shown in Scheme . However, to avoid wasting time in the successive experiments, the fluorescence was collected at 570 ± 2 nm. The experiment shown in Figure unveils the significant increase of the fluorescence intensity as the miRNA concentration increased. The plot of ΔF% as a function of the miRNA concentration on the logarithmic scale displays a linear correlation between the normalized signal and the concentration of miRNA-21 in the range from 500 aM to 100 pM. The equation waswith a regression coefficient equal to 0.998. The detection limit was calculated to be 170 aM (3∂: three times the standard deviation of the control/slope).[36]

Figure 3

Calibration curve of miRNA-21 detection: (A) fluorescence spectra issued from the negative control and different miR-21 concentrations {500 aM, 100 fM, 100 pM, and 500 pM}; (B) plot of the fluorescence peak spectra upon addition of different concentrations of miRNA-21 (from 0.5 fM to 1 nM); and (C) linear correlation between the log 10 of ([miR-21]) and the normalized fluorescence intensity in the range from 0.5 fM to 100 pM (1 μM probes, 0.4 U DSN, and a period of 2 h incubation at 45 °C).

Our proposed method reveals a higher sensitivity and a lower detection limit than the previous method in the literature used to detect miRNAs. For instance, Shen et al. developed a fluorescence-based sensor to detect microRNA, wherein they used magnetic beads and the DSN enzyme. The obtained limit of detection was equal to 60 fM.[37] In another report, Zhao et al. reported an assay for miRNA-21 detection combining the DSN enzyme and AuNPs, taking advantage of a (dT)6 spacer. The obtained calibration curve of their proposed biosensors was linear over the range from 0.5 pM to 5 nM with an LOD of 50 fM.[32] However, their method has substantial drawbacks, including the use of AuNPs that may affect the resulting fluorescence signal since it has a quenching activity. For more comparison, another detection strategy for miRNA-21 has been described by Seo’s group. They employed a DSN–RNAse–TdT–T7 Exo probing system as an amplification approach and fluorescence resonance energy transfer (FRET) as a signal detection mechanism. The co-workers revealed that there is a linear correlation between the fluorescence intensity and the logarithm of the miRNA-21 concentration in the range from 10 fM to 100 nM, with the LOD equal to 2.57 fM.[38] Nevertheless, the limitation of this technique is related to the implication of different catalytic cycles. More recently, the combination of DSN and dTd was also investigated to detect miR-21. However, despite the cooperation between both enzymes, the engendered limit of detection remains high in the range of 10 pM.[39] Thus, our findings are the result of the low background signal to be intended for substantial improvement in the signal/noise ratio along with the high and strong hydrolyzing activity of the DSN enzyme without neglecting the involvement of the TEG spacer that helps improve the formation of a strong DNA/miRNA heteroduplex due to the reduced steric hindrance.

Selectivity and Reproducibility of the Proposed Method

To gain awareness of the selectivity of the proposed assay, the fluorescence intensity was recorded upon the addition of 1 nM of several miRNA sequences, including one-base mismatch, two-base mismatch, and three-base mismatch of miRNA-21 as well as noncomplementary miRNA and miRNA-10b. As shown in Figure , we are able to discriminate between the perfect and the nonperfectly matched miRNA. Besides, slight rises in fluorescence intensity were recorded on the addition of miRNA-10b and the noncomplementary target (NC). Moreover, we observed that there was a remarkable increase in the response signal in the presence of the complementary target miRNA-21, and it was about five times higher than the single-base mismatched miRNA-21. The high selectivity of the proposed approach should be attributed to the excellent activity of DSN for differentiating, even a single-base mismatch. Therefore, these results demonstrate the capacity to distinguish between miRNA.

Figure 4

Responses of the designed assay to 1 nM miRNA-21, miRNA-10b, NC miRNA, and different mismatches (M1, M2, and M3). Experimental conditions are 1 μM probes, 0.4 U DSN, and a period of 2 h of incubation at 45 °C.

The repeatability and reproducibility of the method were assessed by analyzing 1 nM target miRNA-21 in the presence of 0.4 U of the DSN enzyme, and the relative standard deviation (RSD) values of within-batch (intra-assay) and between-batch (interassay) were generated. The intra-assay test was performed by measuring miRNA-21 at 1 nM with three parallel tests, giving an intra-assay RSD value of 5%. However, the interassay value of 5% was obtained with three separate assays arranged under the same conditions. The obtained results suggest a good reproducibility of the developed miRNA assay.

Detection of miRNA-21 in Serum Samples

To evaluate the practicability and feasibility of our proposed method in real biological samples, a recovery test was performed in 100-fold diluted human serum samples by spiking a series of miRNA-21 at concentrations of 10 fM, 1 pM, and 100 pM and were measured to calculate the recoveries under optimal experimental conditions. As resumed in Table , the recoveries of the miRNA-21 in diluted human serum at 10 fM, 1 pM, and 100 pM were 100.1, 97.6, and 102.01%, respectively. According to the RSD value (n = 3), this assay exhibited an excellent reproducibility for miRNA detection in the real sample. Therefore, the proposed method possesses great potential for detecting miRNAs in the real sample and could be clinically useful for cancer detection.

Table 2

Comparative Table of ΔF (%) in Buffer and Serum in the Presence of 0.01, 1, and 100 pM Concentrations and the Detection Recoveries of miRNA-21 from Human Serum

sample	[miR] (pM)	ΔF% (added)	ΔF% (found)	% recovery	RSD%
1	0.01	123.71 086	123.9316	100. 1	0.15 610
2	1	182.03 563	177.7303	97.6	3.04 433
3	100	251.91 515	256.9801	102.01	3.58 143

Conclusions

Benefitting from a fluorescence-based assay for miR-21 detection, we addressed the effect of surface density and spacer composition on the DSN activity for surface-based nucleic acid-sensing. According to our findings, surface coverage is not a major factor that influences the DSN enzyme activity. However, the linker charge affects the DSN-mediated target recycling attributed to a change in the local salt and local pH surrounding the DNA/MB conjugates, providing further support that the DSN-target recycling is affected by the local microenvironment rather than the steric hindrance. Compared to previous reports in the literature, our results are decent in terms of the limit of detection and discrimination of one single-base mismatch. It should be mentioned that these results are owing to the DSN activity and the benefits of magnetic beads as a promising tool to reduce the false-positive fluorescent signal by removing the unbounded and unreacted probes without neglecting the advantages of the insertion of the TEG as a DNA spacer that plays an undeniable role in increasing the sensitivity. More importantly, our proposed method may open a new opportunity to diagnose different disease biomarkers, including miRNAs. It can also represent a universal platform for a highly sensitive, selective, precise, accurate, rapid, and low-cost fluorescence assay to detect different miRNA biomarkers in real samples. Yet, further in-depth study will be carried to better understand the mechanism on the MB surface and its influence on DSN activity.