An electrochemical biosensor equipped with a logic circuit as a smart automaton for two-miRNA pattern detection†

Benting Xie, Shimao Du, Hejun Gao, Juan Zhang*, Hongquan Fu* and Yunwen Liao
College of Chemistry and Chemical Engineering, Chemical Synthesis and Pollution Control Key Laboratory of Sichuan Province, China West Normal University, Nanchong, Sichuan 637002, China. E-mail: zhangbestone@163.com; fubestone@163.com

First published on 26th August 2024

---

---

Introduction

Refining the logic relationships of disease markers involves determining the specific stage of diseases and identifying numerous physical abnormalities.¹ However, a single biomarker may also participate in various biological events within a complex biological environment. For instance, human epidermal growth factor receptor-2 (HER-2) is overexpressed in lung epidermal cells, and it also serves as a crucial prognostic factor for breast cancer.^2,3 Therefore, a collection of biomarkers characterized by coherent logical relationships is instrumental in facilitating precise judgments. MicroRNAs (miRNAs) are a class of short noncoding RNAs, with 19–24 bases, and many research studies have shown that miRNAs play significant regulatory roles in the post-transcriptional regulation of gene expression and the occurrence and development of malignant tumors.^4–9 Recently, an increasing number of studies have verified that these miRNAs could be used as clinical biomarkers to monitor the development process of malignant diseases.¹⁰ Unfortunately, owing to the high sequence homology and low abundance of miRNAs, it is difficult to profile different types of miRNAs in complex samples.¹¹ Therefore, strategies using intelligent computing modules to accurately identify and sort target miRNAs are urgently needed for disease diagnostics.

As computational blocks, logic gates can implement two or more inputs to perform a specific operation process and generate the corresponding signals, which can be used to detect multiple miRNAs simultaneously.^12,13 DNA molecules are considered to be one of the most exquisite engineering materials in molecular logic computing because of their precise Watson–Crick base pairing principle, remarkable biocompatibility and good operability.^14–18 DNA logic gate-based methods can perform Boolean logic operations initially based on the DNA hybridization reaction and then develop into the basic components of integrated circuits used for information processing and storage, which plays an enlightening and crucial role in DNA computing.^19–22 The surprising programmable power of DNA logic gates has been explored for potential applications in different biochemical fields,^23–25 such as miRNA imaging, cancer cell identification,²⁶ cancer immunotherapy and real-time monitoring of apoptosis.^27–32 In most cases, one malignant disease is often associated with a variety of aberrantly expressed miRNAs, and the latent logical relationship of miRNAs is also connected to various disease stages. So, a highly sensitive and specific DNA-based logic gate could be built via designing a suitable logic operation system with different endogenous miRNAs as inputs, which minimizes false positives as much as possible and further enhances the precision and efficiency of early diagnosis.

Electrochemical biosensors benefit from the advantages of sensitivity, speed, accuracy and low background. Consequently, the electrochemical strategy has garnered significant attention from researchers.^33–40 Inspired by the aforementioned scenario, we developed an electrochemical biosensor endowed with computing capabilities, akin to smart measuring automata. This biosensor seamlessly integrated multiple intelligent functions, including rapid sensing, accurate computing, and direct result output. As a demonstration, we selected logic combinations of miR-21 and miR-122 as the target miRNA patterns. Subsequently, we successfully constructed the corresponding AND and OR measuring automata. Within the AND measuring automaton, the output signal was activated only when both miR-21 and miR-122 were present, whereas in the OR configuration, either miRNA was capable of triggering the measuring signal. In comparison with other miRNA logic sensing probes, these measuring automata largely addressed issues pertaining to design simplicity, reaction speed, and biocompatibility. Moreover, unlike traditional detection methods, where the exact value of each biomarker was exported separately and manually combined for assessment, our system automatically reported the direct result (“0” or “1”) of the logic computing based on the related inputs.

Scheme 1 illustrates the design of AND and OR measuring automata. In the AND measuring automaton, in order to improve the practicability and flexibility of the measuring automata, we adopt the “plug and play” design in the AND measuring automaton strategy. The Au nanoparticle (dep-Au) covered electrode (dep-Au/GCE) was functionalized with a thiolated strand S, which was a tandem of complementary sequences of miR-122 and miR-21. The Fc labeled flare strand of the AND measuring automaton (F_A) partially paired with S, leaving two segments unpaired: one (1m) was exposed and the other (2m) was sequestered. In the presence of both miRNAs, miR-21 first was capable of displacing F_A partially by a toehold-mediated strand displacement (TMSD) reaction. The resulting newly exposed segment 2m could act as the toehold region for miR-122. Subsequently, another TMSD reaction occurred, completely freeing F_A. To monitor the entire process, a redox probe Fc was labeled at the 5′ end of F_A (F_A-Fc), where the phosphorite of ferrocene reacted with the hydroxyl group on the single DNA F_A to form a phosphoric acid bond (relevant details are presented in the ESI, Fig. S1 and S2†). When F_A-Fc was liberated from the electrode, the anodic peak current of Fc decreased, eventually providing an “on” measuring signal. For the OR measuring automaton, it shared the same S strand with the AND measuring automaton, while changing the Fc labeled flare strand F (F_O-Fc, relevant details are presented in Fig. S3†). The F_O-Fc partially pairs with S, leaving two segments 1a* and 2b* exposed (1a* served as the toehold region for miR-21 and 2b* for miR-122). Either miR-21 or miR-122 could displace F_O via a TMSD reaction. Remarkably, F_O was designed in a hairpin conformation rather than a linear structure. Once a single miRNA molecule bound to the toehold region of the S strand, F_O could fold into the hairpin conformation through an enthalpy-driven force. It is disengaged upstream from the electrode and its signal decreases, ultimately providing a measuring signal of “on”.

Experimental section

Construction of the biosensor

A glassy carbon electrode (GCE) with Φ = 4 mm was polished with alumina powder, sonicated, and electrodeposited with 1% HAuCl₄ for 30 s at −0.2 V to form Au nanoparticles (dep-Au). The GCE/dep-Au/S-F duplex was created by incubating S-F at 4 °C for 14 hours. Next, hexamercaptan (HT) was added at room temperature for 40 min to remove nonspecific DNA adsorption. Subsequently, a mixture of miR-21 and miR-122 containing 10 mM Mg²⁺ was added at 37 °C for 2 hours. Lastly, the electrodes were cleaned with PBS buffer to scavenge the reaction solution and the SWV signal was recorded. The biosensors should be stored at 4 °C when not in use.

Measurement procedure

An electrochemical workstation with a three-electrode setup was employed to analyse the modified electrode. Cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS) were conducted in a 5.0 mM [Fe(CN)₆]^3−/4− solution with 0.1 M KCl. The SWV signal was tested using a CHI 660E electrochemical workstation in phosphate-buffered saline (PBS, pH 7.4) with a sweep potential range of 0.2–0.7 V.

Results and discussion

Feasibility analysis of the logic gate measuring automata

To verify the feasibility of the DNA reactions, we conducted agarose gel electrophoresis experiments. As depicted in Fig. 1A and B, lane 4 represented the F-related band, lane 5 showcased the double-strand formed by combining miR-122 with S, lane 6 exhibited the duplex formed by miR-21 and S, and lane 7 illustrated the triplex complex of miR-21, miR-122, and S. Notably, only a new band of S-F emerged in lane 8, where the inputs of miR-21 and miR-122 were absent (input = 0, 0 and output = 0). In the AND logic gate (Fig. 1A), lanes 9 and 10 demonstrated that incubating miR-21 (input = 1 and 0) or miR-122 (input = 0 and 1) alone did not yield any new bands (output = 0). Only when miR-21 and miR-122 were present simultaneously (input = 1 and 1) did the S-miR-21-miR-122 complex form along with free F_A, indicating successful liberation of F_A from S by the two miRNAs. Next, we evaluated the performance of the OR logic gate, as shown in Fig. 1B. Upon introducing miR-122 (input = 0 and 1), the band in lane 9 migrated slightly faster, accompanied by the appearance of a new band similar to that in lane 4 (output = 1), indicating successful replacement of F_O by miR-122. Similarly, in the presence of miR-21 (input = 1 and 0), lane 10 exhibited faster mobility and a new F_O band (output = 1) appeared. Furthermore, lane 11 demonstrated that the addition of both miRNAs (input = 1 and 1) also generated an F_O band (output = 1), signifying successful separation of the F_O strand from S. Another band in lane 11 moved slower than the S-F_O duplex, indicating the formation of a triplex complex. These results confirmed that the reactions in the AND and OR gates proceeded as anticipated.

Subsequently, the feasibility of the logic gate measuring automata was tested through electrochemical experiments. Initially, Fc-labeled S-F duplexes of the AND and OR gates were separately assembled on the surface of GCE/dep-Au. The presence of miR-21 or miR-122 was considered as input “1”, while their absence was designated as “0”. The current changes are depicted in Fig. 2 using a heatmap. In the AND measuring automaton (Fig. 2A), neither miR-21 alone (input = 1 and 0) nor miR-122 alone (input = 0 and 1) could reduce the electrochemical signal of Fc. However, when both miRNAs were added simultaneously (input = 1 and 1), F_A-Fc could be displaced, leading to a decrease in the electrochemical signal of Fc and an increase in the current difference. Conversely, as shown in Fig. 2C, either miR-21 (input = 1 and 0) or miR-122 (input = 0 and 1) alone facilitated the separation of F_O-Fc from the electrode, resulting in a decrease in current in the OR measuring automaton.

Furthermore, we conducted real-time electrochemical testing for further insight. We introduced both miRNAs (both at 50 nM) into the detection system simultaneously, or added one miRNA first followed by the other after 40 min, and assessed the real-time electrochemical response of the system in three scenarios. Slight increases in the current signal changes were observed upon addition of a single miRNA, such as miR-21 or miR-122. Notably, only when the other miRNA was added did a noticeable current signal change occur within approximately 40 minutes in the AND logic gate system (Fig. 2B), indicating that both miRNAs were necessary to activate the AND measuring automaton. Conversely, as depicted in Fig. 2D, the OR automaton exhibited a sharp current signal change when one or both of the target miRNAs were incubated in the OR logic gate measuring system. Furthermore, we noted that the output signals of the AND and OR logic gates were unaffected by the order of the two input miRNAs. These results underscored the successful operation of the AND and OR logic gate measuring automata, enabling the simultaneous detection of the two miRNAs. Collectively, the AND and OR gate measuring automata with selective responses to miR-21 and miR-122 offer new avenues for employing logic gates in the detection of multiple miRNAs.

Electrochemical characterization of the constructed biosensor

The step-by-step construction of the biosensing platform was effectively monitored using cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS). Initially, CV measurements for the AND (Fig. 3A) and OR (Fig. 3C) measuring automata were conducted. Au nanoparticles (dep-Au) were immobilized on the electrode surface via electrodeposition, leading to a significant increase in the electrochemical signal compared to the bare GCE (curve a and curve f) owing to the excellent conductivity of dep-Au (curve b and curve g). Following the successful immobilization of the S-F double-stranded complex on the electrode, the electrochemical signal markedly decreased (curve c and curve h). To mitigate non-specific binding, HT was introduced, resulting in a further reduction in the electrochemical signal (curve d and curve i). Subsequently, upon the addition of a mixture of miR-21 and miR-122 (in a 1:1 ratio) onto the GCE surface, the electrochemical signal increased (curve e and curve j) due to the reduction of the Fc-labeled F strand. This sequence of changes in the electrochemical signals provides insights into the successful assembly of the biosensing platform and the subsequent interactions between the biomolecules and the sensing elements.

In parallel, EIS was utilized to validate the successful construction of the prepared biosensor, with electron transfer resistance (R_et) serving as the main analysis parameter. As illustrated in Fig. 3B (for the AND measuring automaton) and Fig. 3D (for the OR measuring automaton), the naked GCE exhibited a small semicircle field and a long tail (curve a and curve f). Upon the introduction of dep-Au (curve b and curve g) onto the electrode, R_et significantly decreased compared to the bare GCE, owing to the excellent conductivity of dep-Au. Subsequently, with the addition of the S-F double complex and HT dropwise onto the electrode surface, the corresponding R_et (curves c, d and curves h, i) was obtained. Upon incubation with the target miR-21 and miR-122, the modified electrode exhibited a distinctly reduced R_et value (curve e and curve j) due to the escape of Fc-labeled F from the electrode. These EIS results were consistent with the CV findings, providing further evidence that the proposed biosensor was successfully constructed.

Optimization of experimental conditions

In order to achieve the best experiment results, the incubation time of miR-21 and miR-122 for the AND measuring automaton and OR measuring automaton was investigated. As depicted in Fig. S4A,† the electrochemical signal responses for the AND measuring automaton gradually decreased as the reaction time increased, reaching a stationary phase after 120 min. Similarly, in Fig. S4B,† the current signal of the OR measuring automaton decreased with increasing reaction time, also stabilizing after 120 min. Consequently, 120 min was selected as the incubation time for the targets. This duration ensured sufficient interaction between the targets and the biosensing platform, maximizing the sensitivity and accuracy of the detection process.

Analytical performance of the proposed biosensor

Different concentrations of miR-21 and miR-122 with a 1:1 ratio were measured to assess the analytical performance of the dual measuring automata under optimal experimental conditions. The SWV responses for both AND (Fig. 4A) and OR (Fig. 4C) measuring automata consistently decreased with the gradual addition of increasing concentrations of dual miRNAs within the range of 0 fM to 50 nM. The current changes (ΔI) were determined by subtracting the current (I) at different concentrations of miRNAs from the current (I₀) without miRNAs. For the AND measuring automaton (Fig. 4B), a linear relationship between the current changes and the logarithm of the concentrations of two miRNAs was observed, expressed as ΔI = 0.5891lgc + 2.5438 (R² = 0.9969). The detection limit for the detection of both miRNAs was determined to be 3.83 fM (LOD = 3S_b/m), where S_b represents the standard deviation of the blank signals and m is the analytical sensitivity estimated as the slope of the calibration curve in lower concentration ranges (Fig. S5A†). Similarly, for the OR measuring automaton (Fig. 4D), the regression equation was ΔI = 0.4982lgc + 2.3678, with a correlation coefficient of 0.9961 and the detection limit for the detection of the two miRNAs was 4.81 fM (LOD = 3S_b/m), as shown in Fig. S5B.† These results confirmed that the as-prepared dual measuring automata exhibited efficient and sensitive responses to the input of the two miRNAs. Additionally, in a comparison of miR-21 and miR-122 detection with some previously reported bioassays (Table S2†), it was found that our biosensors possessed a lower detection limit.

Selectivity, stability, and reproducibility

To assess the selectivity of the electrochemical biosensor, both AND and OR measuring automata were tested with miRNA mimics, miR-155 and miR-16, using four input modes: (miR-155 and miR-16), (miR-21 and miR-16), (miR-122 and miR-155), and (miR-21 and miR-122). The AND measuring automaton (Fig. 5A) exhibited a significant current change only in the presence of miR-21 and miR-122, while the OR measuring automaton (Fig. 5B) showed a significant increase in current for each interfering group (miR-21 and miR-16, miR-122 and miR-15). The results demonstrated excellent selectivity for the AND/OR logic gate measuring system. Additionally, repeatability was confirmed by recording current peaks from five biosensors in the same batch (intra-assay) or different batches (inter-assay). The peak current intensity remained stable with a relative standard deviation (RSD) of 4.82% (intra-assay) and 4.58% (inter-assay) for the AND measuring automaton in Fig. 5C, and 4.67% (intra-assay) and 5.42% (inter-assay) for the OR measuring automaton in Fig. 5D. Furthermore, stability testing over 7 d, 14 d, 21 d, and 28 d illustrated consistent electrochemical signals for both AND (Fig. S6A†) and OR (Fig. S6B†) measuring automata. In addition, the stability of this biosensor was assessed upon scanning for 16 cycles consecutively. As presented in Fig. S6,† the AND (Fig. S6C†) and OR (Fig. S6D†) measuring automata displayed stable current signals. The above results indicated that these biosensors possessed desirable stability.

Application of biosensors in real samples

To further investigate the feasibility and accuracy of the constructed AND and OR measuring biosensing platform, we further examined the expression of miR-21 and miR-122 in actual cell lysates including those of HepG2, HeLa and HEK 293T cells. The RNA lysates were obtained using an RNA extraction kit after cell counting (∼10000). Firstly, the expression levels of miR-21 and miR-122 in different cell lysates were detected using a commercial kit based on quantitative reverse transcription-PCR (qRT-PCR) (Fig. S7†). The results of qRT-PCR experiments showed the absolute contents of miR-21 and miR-122 of HepG2, HeLa and L02 cells, respectively (Tables S3 and S4†). The cell lysate from HepG2 had higher expression levels of miR-21 and miR-122, while the cell lysate from HeLa had a higher expression level of miR-21 and a lower expression level of miR-122. Then, the same samples were analyzed using the AND and OR measuring automata. The results in Fig. S8† illustrated that miR-21 and miR-122 had different expression levels with different cells, calculated according to the linear calibration curve equations in Fig. 4B and D. As shown in Fig. S8,† the results obtained using the AND and OR measuring automata were consistent with those of the qRT-PCR method. The above results suggested that both AND and OR measuring automata showed acceptable reliability and accuracy in actual sample analysis.

To evaluate the application of the prepared AND and OR gate biosensors for the detection of miRNAs in real samples, recovery experiments were also conducted with a serum sample (without miR-21 and miR-122). A series of serum samples that were spiked with miR-21/miR-122 = 1:1, with different concentrations of 1, 10, 100, and 1000 pM, were tested using the constructed sensor, and the corresponding results of the recovery experiments with recovery rates ranging from 94% to 107.6% (AND) and 92.9% to 108.6% (OR) are listed in Tables S5 and S6.†

Conclusions

In conclusion, we successfully fabricated an electrochemical biosensor with integrated logic computing functions, serving as an intelligent measuring automaton. By assembling double-stranded complexes modified with Fc and SH on electrodes coated with dep-Au and utilizing miR-21 and miR-122 as inputs, we achieved AND and OR measuring automata. The Fc exhibited robust redox activity, generating SWV signals. Both detection systems displayed exceptional sensitivity and specificity in both buffer solutions and serum samples. The sensor's simple structure allows for easy improvement and scalability based on functional requirements. This work introduces a novel approach to designing an artificial measuring platform for potential in vivo applications, holding promise for clinical analysis and early-stage cancer diagnostics.

Author contributions

Benting Xie: investigation, data curation, formal analysis, and writing – original draft. Shimao Du: methodology, data curation, and formal analysis. Hejun Gao: review & editing. Juan Zhang: conceptualization, resources, writing – review & editing, supervision, project administration, and funding acquisition. Hongquan Fu: review & editing and funding acquisition. Yunwen Liao: supervision and resources.

Data availability

The authors confirm that the data supporting the findings of this study are available within the article and its ESI.†

Conflicts of interest

The authors declare no conflict of interest.

Acknowledgements

This work was financially supported by the Natural Science Foundation of Sichuan, China (2022NSFSC1268, 2023NSFSC0966, and 2022NSFSC0350), the Doctoral Launch Research Project of China West Normal University (20E037 and 21E045), and the Opening Project of the Chemical Synthesis and Pollution Control Key Laboratory of Sichuan Province (CSPC202309).

References

1. J. Shen, S. A. Stass and F. Jiang, Cancer Lett., 2013, 329, 125–136 CrossRef CAS .
2. K. T. Roybal, L. J. Rupp, L. Morsut, W. J. Walker, K. A. McNally, J. S. Park and W. A. Lim, Cell, 2016, 164, 770–779 CrossRef CAS PubMed .
3. P. Yang, Y. Wang, Z. Yao, X. Gao, C. Liu, X. Wang, H. Wu, X. Ding, J. Hu, B. Lin, Q. Li, M. Li, X. Li, X. Chen, W. Qi, W. Li, J. Xue and H. Xu, J. Am. Chem. Soc., 2020, 142, 18874–18885 CrossRef CAS PubMed .
4. R. W. Carthew, Curr. Opin. Genet. Dev., 2006, 16, 203–208 CrossRef CAS PubMed .
5. B. P. Lewis, C. B. Burge and D. P. Bartel, Cell, 2005, 120, 15–20 CrossRef CAS .
6. H. Dong, J. Lei, L. Ding, Y. Wen, H. Ju and X. Zhang, Chem. Rev., 2013, 113, 6207–6233 CrossRef CAS .
7. Z. Yang, S. Zhang, H. Zhao, H. Niu, Z. S. Wu and H. T. Chang, Anal. Chem., 2018, 90, 13891–13899 CrossRef CAS PubMed .
8. K. Ruan, X. Fang and G. Ouyang, Cancer Lett., 2009, 285, 116–126 CrossRef CAS PubMed .
9. B. M. Ryan, A. I. Robles and C. C. Harris, Nat. Rev. Cancer, 2010, 10, 389–402 CrossRef CAS PubMed .
10. R. Rupaimoole and F. J. Slack, Nat. Rev. Drug Discovery, 2017, 16, 203–222 CrossRef CAS PubMed .
11. P. He, W. Han, C. Bi, W. Song, S. Niu, H. Zhou and X. Zhang, ACS Nano, 2021, 15, 6961–6976 CrossRef CAS PubMed .
12. P. Miao and Y. Tang, ACS Cent. Sci., 2021, 7, 1036–1044 CrossRef CAS PubMed .
13. H. Yang, D. Chen, Y. Zhang, P. Yuan, N. Xie and Z. Dai, Anal. Chem., 2024, 96, 4154–4162 CrossRef CAS PubMed .
14. F. Zhang, J. Nangreave, Y. Liu and H. Yan, J. Am. Chem. Soc., 2014, 136, 11198–11211 CrossRef CAS PubMed .
15. A. Tandon, Y. Song, S. B. Mitta, S. Yoo, S. Park, S. Lee, M. T. Raza, T. H. Ha and S. H. Park, ACS Nano, 2020, 14, 5260–5267 CrossRef CAS PubMed .
16. S. Bi, J. Y. Ye, Y. Dong, H. T. Li and W. Cao, Chem. Commun., 2016, 52, 402–405 RSC .
17. H. Yang, H. R. Shen, A. Qileng, G. S. Cui, Z. Q. Liang, Y. J. Liu and W. P. Liu, Anal. Chem., 2023, 95, 5764–5772 CrossRef CAS .
18. R. Fu, J. Hou, Z. Wang and Y. Xianyu, ACS Nano, 2024, 18, 14754–14763 CrossRef CAS PubMed .
19. D. Q. Fan, J. Wang, E. K. Wang and S. J. Dong, Adv. Sci., 2020, 7, 2001766 CrossRef CAS PubMed .
20. N. L. Xie, M. Q. Li, Y. Wang, H. Lv, J. Y. Shi, J. Li, Q. Li, F. Wang and C. H. Fan, J. Am. Chem. Soc., 2022, 144, 9479–9488 CrossRef CAS PubMed .
21. Y. C. Guo, J. T. Ren and E. K. Wang, Nano Today, 2022, 44, 101476 CrossRef CAS .
22. N. Xie, M. Li, Y. Wang, H. Lv, J. Shi, J. Li, Q. Li, F. Wang and C. Fan, J. Am. Chem. Soc., 2022, 144, 9479–9488 CrossRef CAS PubMed .
23. J. Chen, S. Fu, C. Zhang, H. Liu and X. Su, Small, 2022, 18, 2108008 CrossRef CAS PubMed .
24. L. Ge, W. Wang, X. Sun, T. Hou and F. Li, Anal. Chem., 2016, 88, 9691–9698 CrossRef CAS PubMed .
25. J. Zhu, L. Zhang, Z. Zhou, S. Dong and E. Wang, Anal. Chem., 2014, 86, 312–316 CrossRef CAS .
26. K. Wei, M. Y. He, J. Zhang, C. Zhao, C. Nie, T. Zhang, Y. Liu, T. Chen, J. H. Jiang and X. Chu, Angew. Chem., Int. Ed., 2023, 62, e202307025 CrossRef CAS .
27. J. Zheng, Q. W. Wang, L. Shi, P. Peng, L. L. Shi and T. Li, Angew. Chem., Int. Ed., 2021, 60, 20858–20864 CrossRef CAS PubMed .
28. Y. Yang, J. Xu, Y. Sun, L. T. Mo, B. Liu, X. S. Pan, Z. Liu and W. H. Tan, J. Am. Chem. Soc., 2021, 143, 8391–8401 CrossRef CAS .
29. H. H. Wang, P. Peng, Q. W. Wang, Y. Du, Z. J. Tian and T. Li, Angew. Chem., Int. Ed., 2020, 59, 6099–6107 CrossRef CAS .
30. C. Feng, T. S. Chen, D. S. Mao, F. Zhang, B. Tian and X. L. Zhu, ACS Sens., 2020, 5, 3116–3123 CrossRef CAS .
31. X. Chang, C. Zhang, C. Lv, Y. Sun, M. Z. Zhang, Y. M. Zhao, L. L. Yang, D. Han and W. H. Tan, J. Am. Chem. Soc., 2019, 141, 12738–12743 CrossRef CAS PubMed .
32. V. V. Smirnov, V. S. Drozd, C. K. Patra, Z. Hussein, D. S. Rybalko, A. V. Kozlova, M. A. Y. Nour, T. P. Zemerova, O. S. Kolosova, A. Y. Kalnin and A. A. El-Deeb, Analyst, 2024, 149, 1947–1957 RSC .
33. X. Jiang and M. G. Spencer, Biosens. Bioelectron., 2010, 25, 1622–1628 CrossRef CAS PubMed .
34. S. M. Du, B. T. Xie, H. J. Gao, J. Zhang, H. Q. Fu, F. Liao and Y. W. Liao, Anal. Chem., 2023, 95, 7006–7013 CrossRef CAS PubMed .
35. J. Zhang, Y. Q. Chai, R. Yuan, Y. L. Yuan, L. J. Bai and S. B. Xie, Analyst, 2013, 138, 6938–6945 RSC .
36. J. Zhang, Y. L. Yuan, S. B. Xie, Y. Q. Chai and R. Yuan, Biosens. Bioelectron., 2014, 60, 224–230 CrossRef CAS PubMed .
37. J. Zhang, Y. Q. Chai, R. Yuan, Y. L. Yuan, L. J. Bai, S. B. Xie and L. P. Jiang, Analyst, 2013, 138, 4558–4564 RSC .
38. D. Li, X. L. Zhang, Y. Q. Chai and R. Yuan, Anal. Chem., 2023, 95, 1490–1497 CAS .
39. T. Yao, L. Kong, Y. Liu, H. Li, R. Yuan and Y. Q. Chai, Anal. Chem., 2022, 94, 12256–12262 CrossRef CAS PubMed .
40. M. Qing, S. L. Chen, Z. Sun, Y. Fan, H. Q. Luo and N. B. Li, Anal. Chem., 2021, 93, 7499–7507 CrossRef CAS PubMed .

---

Footnote

† Electronic supplementary information (ESI) available: Materials and reagents, apparatus, Table S1, calculation of the LOD, Table S2, quantitative reverse transcription-PCR (qRT-PCR), analysis of miR-21 and miR-122, Tables S3–S6, and additional figures. See DOI: https://doi.org/10.1039/d4an00719k

This journal is © The Royal Society of Chemistry 2024