Amplification-free detection of Escherichia coli using an acidic deoxyribozyme-based paper device

Abstract

We reported a colorimetric paper-based device by integrating the modified acid RNA-cleaving DNAzymes (MaRCD-EC1) for highly sensitive (detection limit = 10² CFU mL⁻¹), and rapid (within 30 min) detection of E. coli without amplification. This device exhibited a clinical sensitivity of 100% and a specificity of 100% in identifying E. coli-associated urinary tract infections (UTIs) using the clinical urine samples.

---

Escherichia coli (E. coli) is responsible for 75% of urinary tract infections including cystitis and acute pyelonephritis,^1–3 which is a common bacterial infectious disease that poses a threat to public health.⁴ Early accurate detection of E. coli is crucial for controlling the disease process and improving patient treatment. This requires a detection method with high sensitivity and specificity, due to the low abundance of targets at early stages of the disease. Although the polymerase chain reaction (PCR) as a gold-standard method for pathogen detection satisfies these requirements, its disadvantages, such as professional operation and expensive instruments, limit its application in the preliminary diagnosis of diseases in remote and resource-limited regions.^5–7

Point of care testing (POCT) has emerged as a powerful tool for diagnostic testing outside of a laboratory, due to its advantages of portability, disposability, convenient set-up, and easy operation.^8–10 Paper-based devices are ideal candidates for POCT, since they possess the properties of lightweight, disposability, compatibility, and low cost.^11–14 To date, various paper-based POCT devices have been widely used to detect pathogenic bacteria, such as Helicobacter pylori,¹⁵Staphylococcus aureus,¹⁶Salmonella typhi,¹⁷Lactobacillus acidophilus¹⁸ and so on. Among them, the most popular paper-based devices are colorimetric assays,^19,20 which can realize rapid and visual POCT. However, these devices often suffer from poor sensitivity. Incorporating a signal amplification strategy is a general way to improve the sensitivity; nevertheless, it brings more costs and longer analysis time.²¹ Thus, it is essential to develop a novel amplification-free paper-based POCT device for the rapid and highly sensitive detection of E. coli.

An RNA-cleaving DNAzyme (RCD) is a classic DNA catalyst that is capable of performing target-responsive RNA substrate cleavage.²² Most of the bacteria-specific RCDs obtained by in vitro selection need the aid of divalent metal ions to cleave the RNA/DNA chimeric substrate in neutral pH conditions. This may result in nonspecific degradation of nucleic acid by exogenous nucleases, thus generating false-positive results. To address this issue, our group has previously developed an E. coli-activated acidic RCD (aRCD-EC1), which only requires monovalent metal ions as cofactors and exhibits improved catalytic activity at pH = 5.3.²³ aRCD-EC1 was able to detect as low as 10⁴ CFU mL⁻¹ of E. coli using a 10-minute reaction. Interestingly, horseradish peroxidase (HRP) is commonly used in the colorimetric reaction and exhibits the best catalytic performance under acidic conditions.²⁴ Based on these results, we hypothesize that developing an HRP-linked aRCD-EC1 assay may realize the highly sensitive detection of E. coli without amplification. The working principle is illustrated in Scheme 1. The aRCD-EC1 was modified with DNA extensions at the 5′ end for hybridizing with an HRP labelled DNA probe, and with biotin at the 3′ end for immobilizing on agarose beads through streptavidin–biotin interaction. In the presence of E. coli, MaRCD-EC1 can be activated to catalyse the cleavage of the formed HRP@MaRCD-EC1/beads. Subsequently, the released DNA probes containing HRP are separated for performing the H₂O₂-mediated oxidation of 3,3′,5,5′-tetramethylbenzidine (TMB) to generate blue coloured products under acidic conditions. The detection of E. coli is therefore easily converted to detect HRP release.

Scheme 1 Schematic illustration of the HRP@MaRCD-EC1/beads for visual detection of E. coli.

We first employed dPAGE analysis to examine the cleavage activity of MaRCD-EC1 at pH 5.3 (Fig. 1a). In the presence of the crude intracellular mixture derived from E. coli (CIM-EC), MaRCD-EC1 exhibited an apparent rate constant (k_obs) of 0.11 min⁻¹ that is lower than that of aRCD-EC1 (k_obs = 1.18 min⁻¹)²¹ and a cleavage yield of 89.2% in 30 min at 25 °C (Fig. 1b and Fig. S1, ESI†). We then tested the response of MaRCD-EC1 towards the CIMs of unspecific pathogens including Klebsiella pneumoniae (K. pneumoniae), Staphylococcus aureus (S. aureus), Pseudomonas aeruginosa (P. aeruginosa), Burkholderia gladioli (B. gladioli), Bacillus cereus (B. cereus), and Bacillus subtilis (B. subtilis), respectively. No cleavage product was observed, suggesting the high specificity of MaRCD-EC1 (Fig. 1c). Its sensitivity was investigated by adding varying concentrations of CIM-EC. MaRCD-EC1 can detect E. coli at a concentration as low as 10² CFU mL⁻¹ (Fig. 1d) within 30 min, demonstrating its high sensitivity. For comparison, MaRCD-EC1 was used to detect E. coli using a 10 min reaction time and exhibited a detection limit of 10⁴ CFU mL⁻¹ (Fig. S2, ESI†), which is consistent with that of aRCD-EC1.²¹ Obviously, the modification of biotin at the 3′ end of RCD-EC1 only extended the analysis time, yet has a negligible effect on the selectivity and sensitivity of MaRCD-EC1. Furthermore, MaRCD-EC1 was hybridized with an HRP labelled DNA probe and fixed on the surface of agarose beads, which was confirmed by the results of Fig. S3a (ESI†). The feasibility of the successfully generated HRP@MaRCD-EC1/beads to release HRP in the presence of CIM-EC was also investigated. Only after adding CIM-EC, the HRP@MaRCD-EC1/beads were cleaved (Fig. S3b, ESI†), and the colour of the supernatant was changed to blue in the presence of TMB and H₂O₂ (Fig. S3c, ESI†), demonstrating HRP release. Taken together, these results provide a basis for developing a highly sensitive colorimetric probe for E. coli detection.

Fig. 1 (a) The working principle of MaRCDs involves activation upon binding to the target in CIM-EC, independent of M²⁺ ions. F, fluorescein-dT and R, adenosine ribonucleotide. (b) Kinetic responses of MaRCD-EC1S to CIM-EC at 25 °C. The k_obs is given in the graph. (c) Responses of MaRCD-EC1 to the CIMs from various bacteria including KP, SA, PA, BG, BC, BS, and EC. M = marker, Unclv = uncleaved, Clv = cleaved, NC = negative control, %Clv = cleavage percentage, which is calculated based on the equation: %Clv = (clv × 100)/(clv + unclv), where clv is the volume of cleaved band and unclv represents the volume of the un-cleaved band. (d) 10% dPAGE analysis of the response of MaRCD-EC1 to CIM-EC prepared from 10–10⁷ CFU mL⁻¹E. coli cells after 30 min reaction at 25 °C.

We next examined the catalytic activity of HRP at different pH values. As shown in Fig. 2a and Fig. S4 (ESI†), HRP exhibits the highest activity in a weakly acidic environment (pH 5–6), which is compatible with the MaRCD-EC1 cleavage buffer. Using simple centrifugation to separate the released HRP from agarose beads, the TMB and H₂O₂ mixture was added to the supernatant for colorimetric determination. The proportion of released HRP after different reaction times can be calculated by a linear relationship between the k_obs of the colorimetric reaction and the concentration of HRP (Fig. S5 and S6, ESI†). In the presence of CIM-EC, the released HRP amount was increased with the MaRCD-EC1-based cleavage reaction time, and reached a plateau within 30 min (Fig. 2b). In contrast, no released HRP was observed in the absence of CIM-EC.

Fig. 2 (a) Catalytic activity of HRP at different pH values. (b) Kinetic release of HRP from the beads in the absence and presence of CIM-EC. (c) Relative colour intensity analysis in the presence of different bacteria. Relative colour intensity is represented as I/I₀, where I is colour intensity at t = 6 min and I₀ is the colour intensity at the time when CIM-EC was present. (d) Kinetics of the release of HRP in the presence and absence of different concentrations of CIM-EC. S/B = signal to background ratio. (e) Images of the supernatant from the HRP@MaRCD-EC1-bead solutions in the presence of different bacteria CIMs. (f) Images of the supernatant from the HRP@MaRCD-EC1-bead solutions at different concentrations of CIM-EC.

We also investigated the sensing performance of the HRP@MaRCD-EC1/bead-based biosensor. As shown in Fig. 2c, e and Fig. S7 (ESI†), no obvious colorimetric signal was observed when tested with the CIM obtained from unintended bacteria: K. pneumoniae, S. aureus, P. aeruginosa, B. gladioli, B. cereus, and B. subtilis. The sensitivity of the colorimetric assay was tested using CIMs prepared from different concentrations of E. coli (10–10⁷ CFU mL⁻¹). The results showed that the assay could detect concentrations as low as 10² CFU mL⁻¹ of E. coli (Fig. 2d and f). The colorimetric signal intensity is related to the concentration of E. coli, enabling its semi-quantitative colorimetric detection. These results indicate that the immobilization of HRP@MaRCD-EC1 on the agarose beads has negligible effect on the sensing performance.

After immobilizing agarose beads on the paper surface by pullulan, we created an HRP@MaRCD-EC1-based paper sensor for colorimetric detection of E. coli. The paper-based sensor features three zones (Fig. 3a): a left sensor zone (Z1) that contains HRP@MaRCD-EC1 for target recognition, a middle zone (Z2) covered by pullulan film (5%, wt/v) for controlling the flow rate of liquid, and the right detection zone (Z3) contained TMB for signal output. These three zones were produced on plastic-backed Whatman filter paper (Grade 1) using wax barriers. The whole operating step is described as follows: in a typical test, 50 μL of CIM-EC is added onto Z1 for performing target recognition and cleaving the RNA substrate to release HRP. After 30 min reaction, the pullulan film at Z2 will be dissolved (Fig. S8, ESI†), allowing the flow of the solution to Z3. Since the average diameter of agarose beads (100 μm) is significantly bigger than the average pore size of Whatman filter paper (11 μm), they are not able to be driven to flow by the capillary force. Only the released HRP can flow to Z3 to oxidize TMB in the presence of H₂O₂, thus producing a colorimetric signal.

Fig. 3 (a) Design of the paper-based device and its working principle. Z1 is the sensing zone, Z2 is the valve zone, and Z3 is the detection zone. (b) Average optical density values for the paper device in response to different bacteria. Inset: Images of the Z3 zones upon the addition of bacteria indicated in the figure. (c) Average optical density values for the paper device carried out under different concentrations of CIM-EC. Inset: Images of the Z3 zones at varying CIM-EC concentrations.

We first confirmed the feasibility by performing the visual detection of E. coli using this paper-based device. The colour signal was generated in the presence of HRP, MaRCD-EC1, and CIM-EC (Fig. S9, ESI†). It is demonstrated that HRP@MaRCD-EC1/beads on the device can specifically recognize E. coli and release HRP for colorimetric reaction. Then, the specificity of the device was evaluated using different CIMs prepared from unintended bacteria. No colour signal was observed (Fig. 3b), demonstrating high specificity. Furthermore, it was found that the colour intensity was proportional to the concentration of E. coli from 10² to 10⁷ CFU mL⁻¹ (Fig. 3c). The limit of detection (LOD) of 10² CFU mL⁻¹ was achieved within 30 min. Note that this device provides better sensitivity (LOD of 10² CFU mL⁻¹) and shorter sample-to-answer time (30 min) compared with the reported paper-based device without an amplification process.^15,16 Additionally, we also tested the stability of the paper device (Fig. S10, ESI†). A group of devices was stored at room temperature in the dark. The device remained fully activated within one month, indicating the high stability of bioreagents in pullulan.²⁵

Finally, we validated the applicability of this fully-integrated paper-based device for the detection of E. coli from urine samples. We analyzed 40 urine samples collected from hospital patients that were bacteria-negative (culture-), EC-negative but bacteria-positive (culture+/EC−), and EC-positive with yellow colonies (culture+/EC+). The overall operation includes: (1) cell collection (10 min); (2) cell lysate (10 min); and (3) paper-based detection (35 min) (Fig. 4a). As shown in Fig. 4b, 19 culture-samples and 6 culture+/EC− showed no obvious colour change, while 15 culture+/EC+ samples produced discernible blue colour. We could achieve the semi-quantitative determination of the positive clinical samples using the standard curve derived from the relationship between the average optical density and the number of E. coli cells (Fig. 4c and Fig. S11, ESI†). Compared to the urine culturing gold standard (Fig. S12, ESI†), we achieved a clinical sensitivity of 100% and a specificity of 100% (Fig. 4d). Notably, the terms clinical sensitivity and specificity are often used for evaluating a clinical test.²⁶

Fig. 4 (a) Operation steps of the paper-based device for the detection of E. coli in clinical urine samples. (b) Heat map of the E. coli quantitation in clinical urine samples. (c) Colour response of paper-based devices to CIMs prepared from urine samples after 30 min reaction. (d) Comparison of selectivity and sensitivity between the gold standard culture method and paper-based assay.

In summary, we report a novel amplification-free colourimetric assay using an acidic RNA-cleaving DNAzyme for E. coli detection. In the presence of E. coli, the immobilized MaRCD-EC1 can be activated to cleave the RNA substrate, thus releasing HRP-labelled DNA probes. The released HRP catalyzed the TMB oxidation with H₂O₂ at pH 5.3, allowing naked-eye readout without sophisticated equipment. To the best of our knowledge, this HRP-linked DNAzyme amplification-free strategy has never been reported under acidic conditions. Both MaRCD-EC1 and HRP exhibiting the highest catalytical activities under acidic condition enables the rapid (30 min) and highly sensitive (10² CFU mL⁻¹) detection of E. coli. Based on this, we developed a paper-based device for POCT application, in which HRP@MaRCD-EC1 as the recognition element and TMB as the chromogenic substrate are integrated. Clinical evaluation of this paper device using 40 patient urine samples demonstrated a clinical sensitivity of 100% and specificity of 100%. This device remains stable at room temperature for at least 30 days. Taken together, this device is promising for diagnosis in remote and resource-limited areas with the merits of high sensitivity, portability, simple operation, and rapid detection time.

M. L. applied for the funding and conceived the idea. M. L. and Y. Y. C. supervised the project. G. X. Z. performed the experiments and data collection. G. X. Z., Y. P. W., W. X., D. W. and Y. Y. C. analyzed the data. G. X. Z., Y. P. W., and Y. Y. C. wrote the manuscript.

This work was supported by the National Key R&D Program of China (2023YFC3205804) and the Dalian Science and Technology Innovation Fund (2023YGZD04).

Conflicts of interest

There is no conflict to declare.

References

1. E. Denamur, O. Clermont, S. Bonacorsi and D. Gordon, Nat. Rev. Microbiol., 2021, 19, 37–54.
2. Z. Peng, X. Wang, J. Huang and B. Li, Pathogenic Escherichia coli, in Molecular Medical Microbiology, ed. Y.-W. Tang, M. Y. Hindiyeh, D. Liu, A. Sails, P. Spearman and J.-R. Zhang, Academic Press, 3rd edn, 2024, ch. 53, pp. 1065–1096.
3. R. D. Klein and S. J. Hultgren, Rev. Microbiol., 2020, 18, 211–226.
4. L. Váradi., J. L. Luo, D. E. Hibbs, J. D. Perry, R. J. Anderson, S. Orenga and P. W. Groundwater, Chem. Soc. Rev., 2017, 46, 4818–4832.
5. S. Hameed, L. Xie and Y. Ying, Trends Food Sci. Technol., 2018, 81, 61–73.
6. S. M. Yoo and S. Y. Lee, Trends Biotechnol., 2016, 34, 7–25.
7. A. L. Furst and M. B. Francis, Chem. Rev., 2018, 119, 700–726.
8. X. Chen, X. Wang, Y. Fang, L. Zhang, M. Zhao and Y. Liu, Anal. Chem., 2022, 94, 8382–8391.
9. H. Wang, Y. Wang, W. Wang, Y. Zhang, Q. Yuan and J. Tan, Acc. Mater. Res., 2023, 4, 1083–1094.
10. X. Zhao, Y. Wang, Y. Yao, L. Chen, B. Lin, W. Zheng, Y. Zeng, L. Li, Y. She and L. Guo, Anal. Chem., 2023, 95, 6836–6845.
11. K. J. Land, D. I. Boeras, X. S. Chen, A. R. Ramsay and R. W. Peeling, Nat. Microbiol., 2019, 4, 46–54.
12. J. Liu, J. Shi, Q. Feng, W. Fan and C. Liu, Chem. Commun., 2023, 59, 11851–11854.
13. P. Zhou, F. Lu, J. Wang, K. Wang, B. Liu, N. Li and B. Tang, Chem. Commun., 2020, 56, 8968–8971.
14. W. Xue, Z. Dan, Q. Zhang, Y. Chang and M. Liu, Chem. Commun., 2021, 57, 11465–11468.
15. M. M. Ali, M. Wolfe, K. Tram, J. Gu, C. D. M. Filipe, Y. Li and J. D. Brennan, Angew. Chem., Int. Ed., 2019, 58, 9907–9911.
16. M. M. Ali, R. Silva, D. White, S. Mohammadi, Y. Li, A. Capretta and J. D. Brennan, Angew. Chem., Int. Ed., 2022, 61, 3.
17. J. Chen, S. M. Andler, J. M. Goddard, S. R. Nugen and V. M. Rotello, Chem. Soc. Rev., 2017, 46, 1272–1283.
18. A. L. Furst and M. B. Francis, Chem. Rev., 2019, 119, 700–726.
19. R. Sivakumar, V. P. Dinh and N. Y. Lee, ACS Sustainable Chem. Eng., 2023, 11, 2079–2088.
20. G. Zhang, H. Hu, S. Deng, X. Xiao, Y. Xiong, J. Peng and W. Lai, Biosens. Bioelectrons., 2023, 225, 115090.
21. Y. Sun, Y. Chang, Q. Zhang and M. Liu, Micromachines, 2019, 10, 531.
22. E. M. McConnell, I. Cozma, Q. Mou, J. D. Brennan, Y. Lu and Y. Li, Chem. Soc. Rev., 2021, 50, 8954–8994.
23. Q. Zhou, G. Zhang, Y. Wu, Q. Zhang, Y. Liu, Y. Chang and M. Liu, J. Am. Chem. Soc., 2023, 145, 21370–21377.
24. M. Z. Yang, Y. P. Guan, Y. Yang, T. T. Xia, W. B. Xiong, N. Wang and C. J. Guo, Colloid Interface Sci., 2013, 405, 291–295.
25. S. Jahanshahi-Anbuhi, K. Pennings, V. Leung, M. Liu, C. Carrasquilla, B. Kannan, Y. Li, R. Pelton, J. D. Brennan and C. D. Filipe, Angew. Chem., Int. Ed., 2014, 53, 6155–6158.
26. A. G. Lalkhen and A. McCluskey, Contin. Educ. Anaesth. Crit. Care Pain, 2008, 8, 221–223.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4cc01150c

‡ Contributed equally.

---