A rapid and visual turn-off sensor for detecting copper (II) ion based on DNAzyme coupled with HCR-based HRP concatemers

To solve the requirement of on-site, rapid, and visual detection of copper (II) (Cu2+) in aqueous solution, a turn-off sensor for detecting copper (II) ion was developed based on Cu2+-dependent DNAzyme as the recognition element and hybridization chain reaction (HCR)-based horseradish peroxidase (HRP) concatemers as the signal amplifier and the signal report element. The detection unit, which was composed of the immobilized Cu2+-dependent DNAzyme coupled with HCR-based HRP concatemers via Waston-Crick base pairing, could catalyze hydrogen peroxide (H2O2) via TMB, generating obvious green color and turning yellow after sulfuric acid termination with optical absorption at 450 nm. Upon Cu2+ addition, the substrate strand of the Cu2+-dependent DNAzyme concatenated with the HCR-based HRP complex was irreversibly cleaved, efficiently causing dramatic reduction of the detection signal. Under optimal conditions, the detection signal decreased with the concentration of Cu2+ in 5 min, exhibiting a linear calibration from 0.05 to 3 μM with a detection limit of 8 nM. The sensor also displayed a high selectivity for Cu2+ given the specificity and anti-interference of the detection unit, and this system was applicable for monitoring Cu2+ in real water samples. Generally speaking, the proposed sensor exhibits good potential in environment surveys.

to DNA-cleaving DNAzyme with the characteristics of easy synthesis, lower cost and better stability, leads to site-specific cleavage at the guanine position of the substrate in catalytic/substrate multiple strands, which makes it an ideal element for Cu 2+ ion detection due to inherent advantages of outstanding stability under harsh conditions and high selectivity comparable to antibodies 14,39,40 . Over the past decade, different DNAzyme-based biosensors for Cu 2+ using various signal transduction methods have been developed, including fluorescence 41,42 , colorimetry 39,[42][43][44] and lateral flow nucleic acid techniques 45 . Such DNAzyme-based sensing strategies require simpler apparatus and less professional personnel than traditional techniques.
Interestingly, hybridization chain reactions (HCR), which generate linear DNA nanostructures arisen by self-assembly of short DNA fragments through a cross-opening process, have recently become one of the most active research areas in molecular biology and DNA diagnostics relying on several unique advantages 46,47 . Firstly, signal amplification via Watson-Crick base pairing is easier to achieve, which can lead to ultrasensitive detections. Secondly, HCR exhibit the potential of signal reporting via special labels, which pave the way toward multi-target sensing platforms. However, HCR rely on reaction time and are temperature-dependent, so a relatively long time or appropriate temperature likely triggers false positives that interfere with the results. In our research, the HCR was prepared in advance via ingenious hairpin design, effectively eliminating undesirable initiation. Herein, we achieved the proof-of-concept to integrate Cu 2+ -dependent DNAzyme with HCR-based HRP concatemers for rapid detection of Cu 2+ semi-quantitatively observed by the naked eye and quantitatively monitored by a portable spectrophotometer in 5 min, exhibiting immense potential for on-site detection.

Experiment Section
Reagents and Apparatus. All oligonucleotides were HPLC-purified and synthesized at Invitrogen (Life Technologies, Beijing, China). The substrate strand and catalytic strand for copper (II)-dependent DNAzyme were designed according to the literature 42 . The sequences of the oligonucleotides are as follows: Substrate strand for Cu 2+ -dependent DNAzyme (S 1 ):

T T T T T T A GC T T C T T T CT A A TA CG GC T TACCT T TTTTTTTT-Biotin-3′
Catalytic strand for Cu 2+ -dependent DNAzyme (S 2 ): The ingenious design of hairpins should meet the following requirements: Hairpins (H 1 and H 2 ) should possess the same number of bases and the same number of sticky ends. Noticeably, the sticky ends of H 1 locates at the 5′ -end while that of H 2 locates at the 3′ -end. To better promote HCR, nearly half fragment of H 2 at the 5′ -end was truncated as the initiator strand. High binding polystyrene 96-welll microplate was purchased from Costar (NY, USA). In addition, 4-(2-hydroxyethyl) piperazine-1-ethanesulfonic acid (HEPES), ascorbate, streptavidin (SA), bovine serum albumin (BSA), and streptavidin-horseradish peroxidase (SA-HRP) were obtained from Sigma-Aldrich Chemical Co. (St. Louis, MO, USA). Tetramethylbenzidine (TMB) coloring solution was purchased from Biyuntian Co. Ltd (Beijing, China). Ultrapure water was used throughout the experiment and was obtained from a Milli-Q water purifying system (resistivity ≥ 18.2 MΩ cm −1 , Milli-Q, Millipore). Other chemicals were of analytical grade and were used without further purification.
Gel electrophoresis was conducted on a Molecular Imager Gel Doc XR (Bio-Rad, Canada). Visual measurements were carried out with a portable spectrophotometer (NS810, Shenzhen 3nh Technology Co., Ltd). The transmission electron micrograph (TEM) was performed on a JEOL JEM-1400 Transmission electron micrograph (Germany).
Preparation of HCR-based HRP concatemers. Biotinylated hairpins (H 1 and H 2 ) were heated at 95 °C for 5 min and then cooled to 4 °C slowly for 1.5 h before use. The hybridization chain reaction (HCR) proceeded at 37 °C for 1 h, composed of 200 nM S 3 , 1 μ M H 1 and 1 μ M H 2 in Tris-HCl Buffer (10 mM Tris, 500 mM NaCl, 1 mM MgCl 2 , pH 7.4). Afterwards, SA-HRP (500 ng/mL) was added to the HCR system and incubated at 37 °C for 1 h to form the HCR-HRP complex. HCR products were detected by 2% EB-stained agarose gel electrophoresis. HCR-based HRP concatemers was inspected by the transmission electron micrograph 48 . Assembly of the detection unit and copper (II) assay. Immobilized Cu 2+ -dependent DNAzyme was concatenated with HCR-based HRP complex (100 μ L) in each well at 37 °C for 2 h. After washing three times, the detection unit of Cu 2+ was achieved. Then a fixed concentration of Cu 2+ (CuSO 4 , 100 μ L) dispersed in the reaction buffer (1.5 M NaCl, 50 mM HEPES and 50 μ M ascorbate) was incubated with the Cu 2+ -detection unit in each well at 37 °C for 5 min. The same volume of reaction buffer was used as the negative control. After washing three times, 100 μ L TMB coloring solution was added to each well. Then, the samples were incubated at room temperature away from light for 5 min, and 50 μ L H 2 SO 4 (2 M) was added to terminate the chromogenic reaction.
The absorbance values at 450 nm (OD 450 ) were monitored by a portable spectrophotometer. Detection of real water samples. Non-contaminated commercial bottled purified water was purchased from the local supermarket. Lake water was collected from Beijing Olympic Green Park, Beijing, China. Domestic sewage was obtained from the Beijing north river sewage treatment factory, Chaoyang District, Beijing, China. Firstly, the water samples were gently mixed with Cu 2+ standard solution, generating two final Cu 2+ concentrations (0.5 μ M and 2.5 μ M). Then, a 90-μ L water sample was mixed with 10 μ L 10× reaction buffer (15 M NaCl, 500 mM HEPES and 500 μ M ascorbate), and the obtained 100-μ L solution was treated using the same protocol as noted above for the Cu 2+ detection. The ICP-MS analysis for Cu 2+ in real water samples was based on Szpunar's study 6 .

Result and Discussion
Design principle of turn-off sensor. The rapid and visual DNA concatenation sensor for Cu 2+ detection is schematically illustrated in Fig. 1. Beforehand, SA was immobilized onto the high binding polystyrene well via hydrophobic effect. Then, S 1 was captured via biotin-streptavidin combination, which was flanked with spacer sequences of poly-thymine (poly-T) oligonucleotides at the 3′ end and with linker sequence of poly-guanine (poly-G) oligonucleotides at the 5′ end. Via Watson-Crick base pairing at the 3′ terminus and the formation of the DNA triplex at the 5′ terminus, S 2 coupling with S 1 formed the Cu 2+ -dependent DNAzyme. So Cu 2+ -dependent DNAzyme was coated onto the well via biotin-streptavidin linkage to identify the target.
Simultaneously, SA-HRP and the nicked DNA double-helix formed a complex. In the presence of linker-labeled initiator, biotinylated hairpin 1 (H 1 ) first underwent an unbiased strand-displacement reaction to open the hairpin from the 5′ sticky end. The newly exposed sticky nucleates of H 1 opened biotinylated hairpin 2 (H 2 ) and underwent a strand-displacement reaction, exposing another sticky end. Similarly, a chain reaction of hybridization events was propagated to form the biotinylated nicked DNA double-helix, acting as the signal amplifier. In the presence of SA-HRP, which often served as the signal output element, the linker-labeled HCR-based HRP complex formed via biotin-streptavidin linkage. Given that HCR-based HRP complex not only amplified detection signal but also catalyzed hydrogen peroxide (H 2 O 2 ) via TMB, generating obvious yellow color visible to the naked eye and optical intensity at 450 nm (OD 450 ) after sulfuric acid termination, the HCR-based HRP complex acted as a signal amplification element as well as a signal report element.
Finally, Cu 2+ -dependent DNAzyme concatenated with the HCR-based HRP complex was achieved via Watson-Crick base pairing. Upon target Cu 2+ introduction, S 1 of the Cu 2+ -dependent DNAzyme concatenated with the HCR-based HRP complex was irreversibly cleaved. We hypothesized that the cleaved pieces were released and totally washed, leading to the dramatic reduction in OD 450 values and absence of yellow color. Optimization of the immobilized Cu 2+ -dependent DNAzyme. Given that the Cu 2+ -dependent DNAzyme was immobilized onto the surface via biotin-streptavidin linkage, the concentrations of SA, biotinylated substrate strand (S 1 ) and Cu 2+ catalytic strand (S 2 ) were optimized. As illustrated in Fig. 2, the following equation was employed: absorbance ratio (%) = A 0 /A × 100 at 450 nm 7 , where A 0 and A represent the optical intensity after and before Cu 2+ sample addition, respectively. Smaller the ratio, more remarkable the signal variation was, indicating the more appropriate detection conditions. In Fig. 2A, with the SA concentration ranging from 0.5 to 8 μ g/mL, the absorbance ratio (%) exhibited a wave trough and a valley point at 2 μ g/mL of SA, presumably due to stereo-hindrance effects. Small amounts of SA could not capture sufficient amounts of the biotinylated substrate strand, whereas excess SA would lead to steric hindrance, reducing binding with the strand 50 . In Fig. 2B, the ratio gradually declined until reaching a plateau at 120 nM from low dose to high dose, mainly because the combination of streptavidin with biotin conformed to the molar ratio of one to four (SA: 2 μ g/ mL ≈ 30 nM) 51,52 . In Fig. 2C, the ratio was reduced from 100 to 150 nM and subsequently maintained a relatively low value, indicating that slightly excessive S 2 compared with S 1 was ideal for saving experimental materials. In general, 2 μ g/mL of SA, 120 nM of S 1 and 150 nM of S 2 were optimal.
Characterization of the Chromatographic Behavior of the Cu 2+ -dependent DNAzyme on a DNA-NPR Column. Anion exchange chromatography is a widely used technique for separating negatively-charged DNA molecules. The TSK gel DNA-NPR column is a polymer-based anion exchange column for improved DNA resolution over a wide molecular weight range. The capability of the TSK gel DNA-NPR column to separate both high and low molecular weight DNA fragments has been demonstrated by Wilde et al. 49 . Figure 3A,B and C present typical elution profiles of the negative control of the reaction buffer, before and after the Cu 2+ cleavage of the Cu 2+ -dependent DNAzyme, respectively. In Fig. 3A, the negative control showed a high response value at the retention time of 1.187 min with the corresponding peak area of 917. Compared with the negative control, Fig. 3B revealed a relatively high response value at the retention time of 19.835 min with the corresponding peak area of 3073, indicating that the chromatographic peak resulted from the non-dissected Cu 2+ -dependent DNAzyme. In Fig. 3C, there was no response at 19.835 min, whereas a new high chromatographic peak appeared at the retention time of 9.239 min compared with Fig. 3B, suggesting the split Cu 2+ -dependent DNAzyme. Based on these scientific data, we concluded that the Cu 2+ -dependent DNAzyme was cut into small pieces after Cu 2+ sample addition.
Optimization of HCR-based HRP concatemers. As the signal amplifier and signal report element, HCR-based HRP concatemers were optimized, including HCR and the dosage of SA-HRP. To the best of our knowledge, the signal amplification effect of HCR depends on the length of the hybridization chain. Longer the HCR product, the better the effect of signal amplification is. Inspiringly, the longitude of the hybridization chain was hinged on the ratio between the dose of the initiator and two closed hairpins 53 . As proof-of-principle (Fig. 4A), the HCR between the initiator (S 3 ) and H 1 and H 2 was examined by agarose gel electrophoresis. In the absence of S 3 , a bright band of H 1 (lane 2) and H 2 (lane 3) could be observed, and the HCR between H 1 and H 2 was inhibited (lane 4). However, emission bands of high-molecular-weight structures could be observed in lanes 5-10, indicating that higher the ratio of the initiator to hairpins, lower the molecular weight of the HCR product was. A 1:10 ratio of initiator to hairpins triggered the longest hybridization chain, reaching a length of 2 kb. Ratios of 1.5:10 and 2:10 are similar, triggering a longer length of products, which is greater than 1.5 kb and shorter than 2 kb. The 1:10 ratio, corresponding to 100 nM S 3 , 1 μ M H 1 and 1 μ M H 2 was optimal. However, the absorbance ratio at 450 nm was reduced from 100 to 200 nM of S 3 and slightly went up at higher concentrations (Fig. 4B), indicating that 200 nM of S 3 (corresponding to 1 μ M H 1 and H 2 ) was optimal. Comprehensively, 200 nM of S 3 was selected, considering the ideal signal amplification effect, the obvious signal report and no false positive results. Regarding the concentration of SA-HRP (Fig. 4C), the absorbance ratio at 450 nm was reduced until a stable region was achieved when the concentration of SA-HRP was denser than 500 ng/mL, demonstrating that the optimal dosage was 500 ng/mL in the absence of dissipation. Characterization of HCR-based HRP concatemers. HCR products, serving as a typical non-enzymatic signal amplification element, were inspected by agarose gel electrophoresis as shown in Fig. 4A. The emission band demonstrate that HCR products were successfully formed. And the length of HCR products showed a S 3 dose-response effect. When adding SA-HRP into biotinylated HCR nanostructures, an expected color change was appeared. It further confirmed that HCR-based HRP concatemers were successfully fabricated. The color change was attributed to HCR-based HRP complex could catalyze chromogen to produce chromogenic reaction. The detail color image were showed in Fig. S1. Then, transmission electron microscope (TEM) was further employed to observe the morphology of the assembled HCR-based HRP concatemers. As shown in Fig. 5A, morphology with helical concatemers was observed. The result suggested the successful formation of the HCR-based HRP concatemers. In addition, a comparison study with and without HCR-HRP have been conducted to give further validation. As shown in Fig. 5B, OD 450 of the HCR-based system was obviously higher than that of the non-HCR system, and the sensitivity was noticeably improved, clearly indicating that the signal amplification effect of HCR process was significant, which was at least 4-fold greater than that in the non-HCR method.  42 . Therefore, the influence of different concentrations of ascorbate ranging from 1 to 200 μ M on visual readout was studied. As shown in Fig. 6A, the absorbance ratio at 450 nm was reduced and then increased at the concentration of 50 μ M ascorbate. These results indicate that concentrations of ascorbate less than 50 μ M enhanced the cleavage reaction, whereas concentrations greater than 50 μ M suppressed the reaction. We concluded that 50 μ M ascorbate in the slicing buffer was the ideal dosage, and this finding was consistent with Liu's results 42 .
To promote our Cu 2+ turn-off censor for on-site rapid detection, the time of Cu 2+ -mediated cleavage was researched. As shown in Fig. 6B, the following equation was used to characterize the cleavage efficiency: Δ OD 450 (%) = [OD 450 (negative control) −OD 450 (sample)]/OD 450 (negative control) × 100% 7 . The curve of cleavage efficiency rapidly ascended, reaching a plateau after more than 5 min. Three minutes after the addition of Cu 2+ samples, the cleavage efficiency was greater than 80%. After 5 min, the efficiency reached 92%. Given that the immobilized Cu 2+ -dependent DNAzyme coupled with HCR-based HRP concatemers was prepared on the surface in advance, the actual reaction time after Cu 2+ sample introduction was in 5 min, indicating great promise for on-site rapid testing.
Detection performance of the Cu 2+ biosensor. As shown in Fig. 7A, given that the Cu 2+ -dependent DNAzyme was sensitive to the target ion when the concentration of Cu 2+ present in samples increased from 0 nM to 100 μ M, the final visual readout of the Cu 2+ turn-off sensor decreased accordingly until reaching a plateau. The     Scientific RepoRts | 7:43362 | DOI: 10.1038/srep43362 detection limit is much lower than the regulated level. Moreover, the analytical performances of different optical DNAzyme-based assays developed for Cu 2+ analysis are summarized, and the sensitivity of this proposed method is competitive compared with these techniques (Table S1).
The specificity of the Cu 2+ turn-off sensor was evaluated by challenging the system against other metal ions, including Co 2+ , Ti 2+ , Fe 2+ , Fe 3+ , Mn 2+ , Ca 2+ , Ni 2+ , Mg 2+ , Zn 2+ , Cd 2+ , Pb 2+ , Hg 2+ , Ba 2+ , Sr 2+ , UO 2 2+ , Eu 3+ , Tb 2+ , Ag + , Al 3+ and mixed ions. To evaluate the selectivity more accurately and more comprehensively, each metal ion mentioned above was analyzed at three different concentrations of 10 μ M, 100 μ M and 1 mM. As indicated in Fig. 7B and C, in contrast to the significant response of Cu 2+ ion, other metal ions exhibited minimal interference with Cu 2+ detection. Hence, the results exhibited high selectivity toward Cu 2+ compared with other test metal ions given the specific function of the Cu 2+ -dependent DNAzyme.
Determination of Cu 2+ in water samples. To further elucidate the analytical reliability and applicable potential of the Cu 2+ turn-off sensor for testing real samples, three different water samples were monitored by using our developed Cu 2+ sensor and ICP-MS. As shown in Fig. 8, a color chart was designed to assist the qualitative analysis and semi-quantitative LOD of 0.5 μ M was clearly visualized by the naked eye. 0.5 μ M, 1.5 μ M and 2.5 μ M of Cu 2+ could be inspected by the naked eye semi-quantitatively. Therefore, more than 0.5 μ M of Cu 2+ could be inspected by the naked eye. In addition, the detection performance by our strategy for the water samples spiked with Cu 2+ was compared with reference values obtained from ICP-MS, presented in Table S2. Bottled purified water samples, which were considered to contain fewer impurities, exhibited the best recovery results both in the case of low and high Cu 2+ concentrations. Moreover, due to the presence of various mineral components, the detection results obtained from the lake water sample were slightly increased compared with bottled purified water samples. In contrast, the detection results of the water sample from domestic sewage were considerably increased compared with the spiked amount of the Cu 2+ standard mainly due to the inherent Cu 2+ residual in domestic sewage. In general, the detection results were similar between our strategy and ICP-MS in samples at different Cu 2+ concentrations and the developed Cu 2+ turn-off sensor exhibited satisfactory accuracy for Cu 2+ detection in real water samples.

Conclusion
In conclusion, we constructed a rapid and visual turn-off sensor for Cu 2+ detection, displaying several excellent features: (1) Turn-off type. The detection unit was composed of HCR-based HRP concatemers cross-linked with the substrate strand of Cu 2+ -dependent DNAzyme via Watson-Crick base pairing. Upon Cu 2+ addition, Cu 2+ -dependent DNAzyme was irreversibly cleaved, bringing about dramatic decline of the signal output. Thus, the turn-off sensor could effectively get over false positives that interfere with the results. (2) Visualization. As the signal amplifier and signal report element, HCR-based HRP concatemers could catalyze hydrogen peroxide (H 2 O 2 ) via TMB, generating obvious green color and turning yellow after sulfuric acid termination with optical absorption at 450 nm. The signal output could be semi-quantitatively observed by the naked eye with the minimum target of 500 nM and quantitatively monitored by a portable spectrophotometer with the LOD of 8 nM, demonstrating immense potential for on-site detection. (3) Sensitivity, Selectivity and anti-interference. The turn-off sensor possesses outstanding sensitivity, selectivity and anti-interference, exhibiting a linear calibration from 0.05 to 3 μ M with a detection limit of 8 nM and displaying remarkable applicability for monitoring Cu 2+ in real water samples, by virtue of the specificity of DNAzyme and the signal amplification of HCR-based HRP concatemers. (4) Rapidity. Given that the immobilized Cu 2+ -dependent DNAzyme coupled with HCR-based HRP concatemers was prepared on the surface in advance, the actual reaction time after the Cu 2+ sample introduction was 5 min, indicating great promise for on-site rapid testing.