Enzyme-Mediated Endogenous and Bioorthogonal Control of a DNAzyme Fluorescent Sensor for Imaging Metal Ions in Living Cells

Abstract

Bioorthogonal control of metal ion sensors for imaging metal ions in living cells is important for better understanding of the distribution and fluctuation of metal ions. While bioorthogonal control using external stimuli such as light has been reported, these stimuli are not suitable for investigating many biological systems where light penetration is limited or quantum yield of activation is low. To overcome this limitation, we herein report the endogenous and bioorthogonal activation of a DNAzyme fluorescent sensor containing an 18-base pair (bp) recognition site of a homing endonuclease (I-SceI), since the recognition site of this length is found by chance only once in 7×10¹⁰ bp of genomic sequences, it can thus form a near bioorthogonal pair with I-SceI for DNAzyme activation with minimal effect on living cells. The formation of the recognition site through double stranded DNA hybridization prevents the formation of an active DNAzyme initially. Once I-SceI is expressed inside cells, it cleaves at the recognition site, which allows the DNAzyme to adopt its active conformation. The activated DNAzyme sensor is then able to specifically catalyze cleavage of a substrate strand in the presence of Mg²⁺ to release the fluorophore-labeled DNA fragment and produce a fluorescent turn-on signal for Mg²⁺. Using this method, we demonstrate the use of I-SceI to activate the 10–23 DNAzyme for imaging of Mg²⁺ in HeLa cells with endogenous and bioorthogonal control.

Graphical Abstract

Introduction

Metal ions are indispensable for sustaining all forms of life, especially their cellular homeostasis associated with different types of human diseases.^[1] To further understand the relationship between biological processes and cellular metal ions, it is essential to gain insights into their distribution and concentration inside cells.^[2] Because these metal ions are distributed in different locations of biological systems and their concentrations can change over time, it is important to design metal ion sensors with spatiotemporal control to activate them only in the desired location at the desired time.^[3] Toward this goal, we and other groups have developed a class of metal ion sensors using DNAzymes, which is a family of single-stranded DNA molecules with catalytic activity.^[4] The metal-specific DNAzymes, such as those for Pb²⁺, Zn²⁺, and Cu²⁺, have been obtained through thein vitro selection technique from a large DNA library that contains as many as 10¹⁵ different sequences.^[5] By conjugating fluorophores and quenchers to these DNAzymes, their metal-dependent catalytic activity can be transformed into a turn-on fluorescence output, using the catalytic beacon approach.^[6]

To achieve precise control of DNAzyme sensors for intracellular imaging of metal ions, we have conjugated photolabile caging groups, such as the 2’-nitrobenzyl group, to the DNAzymes so that the 2’-OH of the scissile ribonucleotide in the middle of the DNAzyme substrate strand is protected from metal-dependent cleavage.^[7] The DNAzyme can be activated with 365 nm light irradiation, which removes the photolabile group from the substrate, allowing metal-ion dependent cleavage of the substrate. This caging strategy has been used for exogenous activation of DNAzymes in vitro, in living cells, and in zebrafish.^[7a,8] Furthermore, another exogenous activation method has been developed based on DNAzyme conjugation to gold nanoshells, which produce a localized temperature increase upon near-infrared (NIR) light irradiation. This temperature increase was used to dehybridize a blocking DNA strand to activate the DNAzyme sensor, allowing for NIR light-based exogenous activation.^[9] While these studies have achieved spatiotemporal control using an external light source in living cells and zebrafish, it may not be suitable for many other systems where light penetration depth can be a limitation. In addition, these photo-caging and photo-thermal caging techniques require complicated synthesis of the caged groups and nanomaterials, as well as the use of high dosages of laser irradiation as a trigger, which could induce damage to living systems.

To overcome the above limitations of exogenous light activation, it is necessary to develop endogenous control mechanisms for DNAzyme activation. However, unlike exogenous controls, endogenous controls must be biologically orthogonal so that the control will not trigger or interfere with other biological events. Toward this goal, ROS-activitable DNAzymes were recently reported by modification of phenylboronate or phosphorothioate to the DNAzymes, in which H₂O₂ or HClO, respectively, are used as triggers to activate the DNAzymes for intracellular metal-ion sensing.^[10] Even though this strategy can achieve activation in the live cell environment, the introduction of exogenous ROS species inevitably induces damage in living cells. Therefore, control of the DNAzyme sensors using an endogenous and bioorthogonal control with less cytotoxicity is desirable, because it can be applied at any location in any system without the above limitations.

It has been shown that enzymes are excellent stimuli to control native cellular processes.^[11] Furthermore, a variety of enzyme-responsive sensors have been widely employed for triggered drug release by integration with nanomaterials.^[12] Building on these techniques, we herein report a novel approach of using the I-SceI enzyme to activate the 10–23 DNAzyme for fluorescent imaging of Mg²⁺ in living cells, which doesn’t require complex synthesis of organic molecules or nanomaterials and affords simple and convenient control of the DNAzyme sensor.

We chose I-SceI as the activating enzyme, which is a homing endonuclease derived from Saccharomyces cerevisiae whose normal function is to initiate gene conversion by generating double-stranded breaks at specific genomic invasion sites.^[13] Compared with restriction endonucleases, the target recognition sites of homing endonucleases are much longer, such as 18 base pairs for I-SceI, and thus are expected to be found at much lower rates in any given genome. For instance, the 18 base pair recognition sequence of I-SceI only occurs statistically once in 7×10¹⁰ bp of nucleotides, which is equivalent to one cutting site in a random sequence pool with about twenty times the size of the human genome.^[14] As a result, I-SceI would cause minimal damage to most host genomes, especially for human cells. Therefore, a number of I-SceI expression vectors have been constructed to induce double-stranded breaks for genomic applications (e.g., repair, recombination, deletion) and further therapeutic applications without reported off-target effects common with many other restriction enzymes.^{[14b, 15]} By introducing the I-SceI recognition site into the DNAzyme binding arms and expressing I-SceI in cells, the homing endonuclease and its recognition site can serve as an ideal bioorthogonal pair for control of the sensor. Given this design does not require any complicated chemical synthesis or DNA modification and many DNAzymes with similar secondary structures have already been obtained for many other targets, the method demonstrated in this work will find wide applications for bioorthogonal endogenous control of DNAzyme sensors for imaging many different targets in biological systems with less cytotoxicity than current methods.

Results and Discussion

Mechanism of enzyme-activated DNAzyme sensing of metal ions

The 10–23 DNAzyme is chosen based on its high catalytic activity for Mg²⁺. As shown in Scheme 1, the designed three strand DNA contains a substrate strand (S₁) containing an adenosine ribonucleotide (rA) at the scissile cleavage position. The S₁ sequence is designed so that it can partially hybridize to the conserved sequence of the 10–23 enzyme strand (S₂) responsible for the enzymatic activity of the DNAzyme (in green) and its substrate-binding arm at the 5-prime end, with the resulting S₁₂ complex displaying a calculated melting temperature of 53 °C. To enable activation of DNAzyme activity by I-SceI, the S₂ sequence is extended at the 3-prime end to contain the I-SceI recognition sequence (in blue), which is then hybridized to a linker strand (S₃), with the resulting S₂₃ displaying a calculated melting temperature of 64 °C, to complete the double stranded recognition sequence of the I-SceI endonuclease. Since the melting temperatures of S₁₂ and S₂₃ are above 37 °C, the formation of the S₁₂₃ construct effectively prevents formation of the active conformation of the DNAzyme. As shown later in this work, the key to this successful design is that the calculated melting temperature of S₂₃ is similar to, but higher than that of S₁₂, which further enables efficient I-SceI cleavage of S₂₃ to lower the melting temperature of S₂₃, removing the hybridization locks, and resulting in formation of the active DNAzyme conformation.

Scheme 1.

Schematic of I-SceI activation of the 10–23 DNAzyme for intracellular imaging of Mg²⁺.

To signal I-SceI-induced cleavage of the S₁₂₃ complex, we conjugated FAM at the 3-prime end of S₂ and a quencher (IABkFQ) at the 5-prime end of S₃, and to signal 10–23 DNAzyme sensing of Mg²⁺, we conjugated Cy5 at the 3-prime end of S₁, and another quencher (IAbRQ) at the 5-prime end of S₂. In the absence of I-SceI, formation of the S₁₂₃ complex effectively places the quenchers (IAbRQ and IABkFQ) next to their respective fluorophores (Cy5 and FAM), resulting in minimal fluorescence signal, even in the presence of Mg²⁺, because the 10–23 DNAzyme is unable to fold into its active conformation, rendering it inactive. In the presence of I-SceI, the recognition sequence in S₂₃ is targeted by I-SceI and cleaved into two fragments. After this cleavage, the calculated melting temperature of the left fragment is 41 °C, which is not stably hybridized under normal cellular temperature and thus can partially dehybridize, resulting in the IABkFQ quencher dissociating away from the FAM, leading to an increase in FAM signal. At the same time, the dehybridization of the short fragment of S₃ with a 4 bp hangover at the 3-prime end of S₂ opens the possibility for the left arm of S₁ to hybridize with the left arm of the enzyme strand (S₂) to form the an activate conformation of the 10–23 DNAzyme. Subsequently, in the presence of Mg²⁺, S₁ will be cleaved. Like the catalytic beacon design that our group has demonstrated previously, because the calculated melting temperature of the right arm of the DNAzyme is 26 °C, which is lower than cellular conditions, the DNAzyme cleavage will result in release of the fragment containing Cy5 from the IAbRQ quencher, and thus an increase of the Cy5 signal.

Activation of the DNAzyme

The constructs designed in Scheme 1 were prepared by heating and annealing of the individual ssDNAs in Tris buffer (50 mM Tris, 100 mM NaCl, 50 mM KCl, 1 mM MgCl₂, pH = 7.4) to mimic the typical ionic strength of a cell. When S₁ and S₂ were first annealed in the absence of Mg²⁺, Lane 1 in Figure 1 shows the formation of S₁₂ with a single band. Upon addition of 10 mM Mg²⁺, the S₁₂ band in Lane 2 disappeared while a lower molecular weight (MW) band appeared, suggesting the substrate band was cleaved. This result is not surprising, because the S₁₂ can form the active conformation of 10–23 DNAzyme in the absence of S₃. On the other hand, when S₂ and S₃ were annealed in Lane 3, a single band was observed, showing the formation of S₂₃. Addition of I-SceI in Lane 4 resulted in two lower MW products, revealing the successful cleavage of S₂₃ by I-SceI.

Figure 1.

Native PAGE analysis of S₁₂₃ formation and the cleavage in the presence of I-SceI and Mg²⁺. The gel was stained with ethidium bromide for visualization. Lane 1: S₁₂; Lane 2: S₁₂ + 10 mM Mg²⁺; Lane 3: S₂₃; Lane 4: S₂₃ + I-SceI; Lane 5: S₁₂₃; Lane 6: S₁₂₃+ I-SceI: Lane 7; S₁₂₃ + I-SceI + 10 mM Mg²⁺; Lane 8: dS₁₂₃ (no ribonucleotide); Lane 9: dS₁₂₃ + I-SceI.

Having demonstrated that S₁₂ can form the active DNAzyme and S₂₃ can be cleaved by I-SceI, we annealed all three strands together. As shown in Lane 5, a single band with higher MW than S₁₂ and S₂₃ (Lanes 1 and 3) was observed, suggesting the formation of S₁₂₃. When the S₁₂₃ was incubated with I-SceI, as shown in Lane 6, the band assigned to S₁₂₃ disappeared and two lower MW bands were observed. To confirm that the top band in Lane 6 is indeed the active DNAzyme, we added 10 mM Mg²⁺ and found the top band in Lane 6 disappeared and a new band with a relatively lower MW was observed in Lane 7, which indicates the substrate strand of the activated DNAzyme was cleaved. These results strongly suggest that the DNAzyme in S₁₂₃ can be activated in the presence of I-SceI.

Complicating the design shown in Scheme 1, Mg²⁺ is a cofactor for both the 10–23 DNAzyme and I-SceI enzyme. As a result, the 1 mM Mg²⁺ present in the I-SceI enzyme buffer might provide enough Mg²⁺ to induce cleavage by the 10–23 DNAzyme. To verify whether there is any interference from the 1 mM Mg²⁺, a non-cleavable substrate, dS₁, in which the scissile adenosine ribonucleotide (rA) was replaced with adenosine deoxyribonucleotide (dA), was used to form dS₁₂₃ as a negative control. As shown in Lane 8, dS₁₂₃ displays a single band with a MW that is identical to that of S₁₂₃ (Lane 5). Upon incubation with I-SceI, the dS₁₂₃ band disappeared (see Lane 9) and a lower MW band appeared consistent with the major band from Lane 6 (S₁₂₃ reacted with I-SceI appeared but without addition Mg²⁺). This control experiment suggests that the major bands from Lanes 6 and 9 are consistent with the non-cleaved DNAzyme, rather than a cleaved product, and thus that 1 mM Mg²⁺ is not enough to induce DNAzyme cleavage. Meanwhile, to ensure the design of S₁₂₃ can block the formation of the active DNAzyme, the catalytic activity of S₁₂₃ in the presence of Mg²⁺ was also investigated. As shown in Figure S1 (Supplementary Information), there is no obvious band formation after S₁₂₃ is treated with 10 mM Mg²⁺ but without I-SceI (Lane 2, Figure S1), in comparison with the S₁₂₃ only band shown in Lane 1. These results indicate that the activity of the DNAzyme was blocked initially even in the presence of 10 mM Mg²⁺ for 4 h.

Fluorescence measurements

To maximize the performance of the S₁₂₃ construct, the ratios between S₁, S₂, and S₃, the amount of I-SceI, and the incubation time were all optimized, and the results are shown in Figures S2 and S3. After this optimization, the feasibility of this S₁₂₃ construct as a fluorescent sensor was further evaluated through in vitrofluorescence measurements. As shown in Figure 2, when 2.5 units of I-SceI was added to S₁₂₃ in a buffer containing 1 mM MgCl₂ (50 mM Tris, 100 mM NaCl, 50 mM KCl, 1 mM MgCl₂, pH=7.4) and incubated for 3 h, fluorescence intensity at 520 nm (Figure 2a), due to emission of FAM, increased by 2.5-fold while the fluorescent intensity at 665 nm (Figure 2b), due to emission of Cy5, did not increase as much. This result is consistent with the design in Scheme 1, in which the I-SceI cleaves S₂₃, resulting in dehybridization of S₂₃ and thus the release of FAM from its quencher. The small increase in Cy5 emission at 665 nm is likely due to low activity of the DNAzyme in the presence of 1 mM Mg²⁺, which is in the buffer for I-SceI cleavage activity. Subsequently, when 50 mM Mg²⁺ was added into the above solution for 80 min, the Cy5 emission at 665 nm increased significantly by 5-fold, suggesting formation of the active DNAzyme, cleavage of the S₁ substrate strand, and dissociation of Cy5 from its quencher. While the increase of Cy5 emission was clearly observed in the presence of 50 mM Mg²⁺, we also observed a decrease of the FAM emission at 520 nm from the I-SceI cleavage reaction. We attribute this decrease to enhanced hybridization of the S₂₃ fragment in the presence of the higher (50 mM) concentration of Mg²⁺ in the DNAzyme cleavage reaction than that (1 mM) in the I-SceI cleavage reaction, which results in increased hybridization, and thus less release of FAM from its quencher. To further verify whether S₁₂₃ can effectively block the formation of the active DNAzyme, we measured the Cy5 emission at 665 nm in the presence of Mg²⁺ incubated without I-SceI. As shown in Figure S4, the fluorescence intensity for S₁₂₃ is minimal in the presence of 50 mM Mg²⁺ for 2 hours, suggesting that the DNAzyme was inactive without the I-SceI activation. In contrast, the addition of I-SceI in the above reaction mixture results in 5-fold higher fluorescence intensity than that in the absence of I-SceI, demonstrating that the designed S₁₂₃ can effectively block the formation of the active DNAzyme and its catalytic activity can be activated by I-SceI.

Figure 2.

Fluorescence spectra of S₁₂₃ upon addition of I-SceI and Mg²⁺. 200 nM S₁₂₃, 2.5 unit of I-SceI, 50 mM Mg²⁺ (incubated for 2 h). (a): λ_ex = 495 nm; (b): λ_ex = 644 nm.

To further evaluate the performance of the activated-DNAzyme, the DNAzyme reaction kinetics and Mg²⁺-dependent calibration curves were measured. Figure S5a shows the Mg²⁺-dependent fluorescence changes after S₁₂₃ is incubated with I-SceI. By nonlinear curve fitting, based on reaction kinetics described previously of the observed reaction rate (k_obs) vs. Mg²⁺ concentration (Figure S5b),^[16] we obtained the limit of detection as 2.1 mM (3σ/slope), suggesting the I-SceI activated-DNAzyme maintains its ability to sense Mg²⁺ at physiologically relevant concentrations.

Expression of I-SceI in living cells

Having demonstrated the enzyme-activated DNAzyme for sensing Mg²⁺ in a test tube, we applied the construct for imaging Mg²⁺ in HeLa cells. A commercial plasmid, pCBASceI, was chosen to express I-SceI in mammalian cells. Prior to imaging Mg²⁺ intracellularly, we first investigated the applicability of this plasmid in our I-SceI activated-DNAzyme system, by examining the expression of I-SceI and its subcellular localization after pCBASceI transfection. Since lipofectamine reagent has been widely used to transfect DNA into both the cytoplasm and the nucleus,^[17] we tested its transfection ability for pCBASceI with an immunofluorescence staining assay using an I-SceI specific antibody as the primary antibody and a secondary antibody conjugated with FITC. By varying lipofectamine transfection time for pCBASceI, the presence of FITC fluorescence can be used to measure the expression I-SceI and to indicate its subcellular localization. As shown in Figure 3, using confocal laser scanning fluorescence microscopy, a gradual increasing trend for the green fluorescence emission from FITC response in the range of 8 to 24 h was observed, followed by distribution of green fluorescence throughout the cells at longer transfection times. Through image quantification using ZEN software, the mean intensity of the FITC channel for each transfection group was obtained and shown in Figure S6. These results suggest that I-SceI was expressed within 8 h, and longer transfection times led to higher expression of I-SceI, reaching a plateau in expression by 24 h. Therefore, the I-SceI expression level can be tuned by the length of transfection time of the pCBASceI plasmid, which may further influence the activation process of the sensor inside the cells. Since Hoechst 33258 is known to stain the nucleus of cells, we used it to identify the location of the pCBASceI expression. As shown in Figure S7, the fluorescence from FITC not only showed good colocalization with Hoechst, but also distributed well around the cytoplasm. Together, these results suggest that I-SceI has been successfully expressed in both the cytoplasm and the nucleus.

Figure 3.

Immunofluorescence staining assay in HeLa cells. Under different pCBASceI transfection times, cells were treated with I-SceI antibody and secondary goat anti-rabbit antibody (with FITC). (a) Control, without pCBASceI; Incubation with pCBASceI for (b) 8 h, (c) 16 h, (d) 24 h, (e) 36 h, (f) 48 h transfection times, respectively. Nuclei stained with Hoechst 33258. Scale bar: 50 μm.

Intracellular imaging

After demonstrating the successful expression of I-SceI inside HeLa cells, we tested the feasibility of I-SceI-activated DNAzymes for intracellular imaging of Mg²⁺. The HeLa cells were first incubated with pCBASceI for 24 h to allow I-SceI expression. The cell viability test via an MTT assay shows no obvious toxicity in the presence of up to 5 μg/mL pCBASceI with 0.5 μL Lipofectamine and 0.25 μL P3000 reagent (Figure S8). Subsequently, the cells were transfected with 200 nM S₁₂₃ (0.5 μL Lipofectamine and 0.25 μL P3000 reagent), for another 4 h and treated with 10 mM Mg²⁺ for 30 min before imaging. As shown in Figure 4b, the fluorescence of the HeLa cells treated with only S₁₂₃ and Mg²⁺ but without pCBASceI is very similar to the sample treated with only S₁₂₃ (Figure 4a), indicating that Mg²⁺-dependent DNAzyme activity of S₁₂₃ within the cells was inhibited by the three-strand construct. These results are also consistent with results from the gel electrophoresis experiments (Figure 1) and fluorescent assays (Figure 2). After incubating the cells with pCBASceI, a significant increase of FAM green fluorescence can be observed (Figure 4c), which is consistent with the I-SceI-induced cleavage and dissociation of the fragment labeled with FAM and quencher as shown in Scheme 1. From colocalization with Hoechst staining, we can observe partial fluorescence of FAM that is present not only in cytoplasm but also in the nucleus, and this result is consistent with the distribution of I-SceI shown in the immunofluorescence staining assay (Figure S7). Together, these results suggest that the S₁₂₃ probes were successfully activated by I-SceI expression. Upon addition of Mg²⁺, a dramatic increase of the red fluorescence signal in the Cy5 channel was observed (Figure 4d), while the FAM green fluorescence channel maintained a similar intensity in comparison with images without addition of Mg²⁺ (Figure 4c). From these confocal images, the distribution of Cy5 intensity was quantified, as shown in Figure S9a, and the obtained significance value (P < 0.0001) reveals the successful activation of the DNAzyme after I-SceI cleavage. The quantification of the Cy5 channel indicates a remarkable 4-fold increase in the Cy5 fluorescence (Figure S9b), which can be attributed to Mg²⁺-dependent DNAzyme cleavage and release of Cy5 from its quencher. To expand upon this experiment, different concentrations of added Mg²⁺, with a range from 5–20 mM, were also tested, as shown in Figure S10 and S11. It can be observed that the Cy5 fluorescence response increases with increasing concentrations of Mg²⁺, indicating the applicability of detection of the change of Mg²⁺ within 5–20 mM using this I-SceI-activated DNAzyme fluorescent sensor.

Figure 4.

Confocal microscopy images of HeLa cells (a) transfected with (a) S₁₂₃ along with (b) 10 mM Mg²⁺, (c) pCBASceI, and (d) pCBASceI + 10 mM Mg²⁺. The green channel is from FAM fluorescence. The red channel is from Cy5 fluorescence, and the blue channel is Hoechst 33258 for nucleus staining. Scale bar: 20 μm.

To further demonstrate that the fluorescence of Cy5 obtained was due to the activity of the activated-DNAzyme, a non-cleavable substrate (dS₁) in which the scissile adenosine ribonucleotide (rA) was replaced with adenosine deoxyribonucleotide (dA), was used as a negative control. As shown in Figure S12a and b, when dS₁ was used, the FAM green fluorescence increased when pCBASceI was expressed and the increase is similar to when the active substrate S₁ was used (Figure 4a and c), because both S₁₂₃ and dS₁₂₃ constructs contain the same I-SceI recognition sequence and thus are both subject to I-SceI-based cleavage and FAM fluorescence increase. Upon addition of Mg²⁺, however, in contrast to the dramatic increase of the Cy5 fluorescence when the S₁ substrate was used (Figure 4d), no obvious Cy5 fluorescence increase was observed when dS₁ substrate was used (Figure S12c), confirming that the Cy5 fluorescence signal enhancement originated from Mg²⁺–dependent catalytic activity of the activated-DNAzyme. Further tests of the stability of the S₁₂₃ complex suggest that it is stable in HeLa cells for at least 16 h (Figure S13), which is much longer than the amount of time S₁₂₃ is incubated with cells before imaging in this study (4 h).

Flow cytometry measurements

Finally, the ability of the S₁₂₃ construct to be used in sensing Mg²⁺ inside of HeLa cells was also confirmed using flow cytometry. As shown in Figure S14, the cells incubated with pCBASceI and without Mg²⁺ displayed a slightly higher level of fluorescence intensity in the FAM channel than those without pCBASceI or Mg²⁺, consistent with I-SceI-induced cleavage and release of FAM from its quencher. Upon addition of Mg²⁺, a dramatic increase of fluorescence intensity in the Cy5 channel was observed. Therefore, the results from gel electrophoresis (Figure 1), fluorescence assays (Figure 2), confocal images (Figure 4), and flow cytometry (Figure S14), together with several negative controls, including the use of dS₁ substrate to replace S₁ substrate, all confirm that the I-SceI can endogenously activate DNAzymes for imaging metal ions inside living cells.

Conclusion

In conclusion, we have demonstrated a simple, endogenous, and bioorthogonal method for activation of the catalytic activity of DNAzymes via the conjugation of the I-SceI enzyme recognition site to the DNAzyme binding arms. Through specific expression of I-SceI in living cells, the enzyme-activated strategy provides a new technique to activate the formation of DNAzymes with endogenous control, and thus prevent the DNAzyme from reacting with extracellular metal ions during the delivery process. In addition, a pCBASceI plasmids that can directly express I-SceI enzymes in mammalian cells is commercially available, making it possible for many labs to apply this method for other targets. Since I-SceI can be modified to be expressed in different subcellular locations, this system can be designed to probe metal ions in specific subcellular locations by using subcellularly expressed enzymes, which will only activate the DNAzyme sensors in these locations. This system also has the potential to use activatable plasmid for more diverse activations of DNAzymes. For example, a plasmid with a promoter that can be activated by small molecule-induced expression of I-Scel will make DNAzyme activation by inducers such as IPTG and arabinose possible.^[18] Therefore, this enzyme-activated DNAzyme strategy not only expands the application of DNAzyme-based sensors for intracellular imaging, but also provides a platform to better understand the localization and distribution of metal ions in the living cell.

Experimental Section

Preparation of S₁₂₃ and activation of DNAzyme

The S₁₂₃ were prepared by heating and annealing the individual ssDNA in Tris buffer (50 mM Tris, 100 mM NaCl, 50 mM KCl, 1 mM MgCl₂, pH = 7.4). The solution was heated at 90 °C for 5 min and slowly cooled down to room temperature. Then, 200 nM S₁₂₃ was incubated with 2.5 unit of I-SceI at 37 °C for 3 h.

Native polyacrylamide gel electrophoresis analysis

12% native polyacrylamide gel (10 × TBE, 40% Acry-Bis, ultrapure water) eletrophoresis was performed to verify the formation of the S₁₂₃ and activated-DNAzyme with the addition of I-SceI and Mg²⁺.10 μL of 200 nM of products were mixed with 2 μL 6 × NEB gel loading dye, and then were loaded on the 0.75 mm thin gel. Electrophoresis was run at 250 V for 30 min at room temperature in 1×TBE buffer. Gel was stained by 0.5 μg/mL ethidium bromide and imaged with a Bio-rad fluorescence gel imaging system.

Fluorescent measurements

After S₁₂₃ was incubated with I-SceI, a 190 μL solution was taken and a 10 μL aliquot of Mg²⁺ was added to measure fluorescence using kinetic analysis mode with excitation at 488 and 644 nm, emission at 520 and 665 nm, respectively. All in vitro fluorescence experiments were carried out at 37 °C.

pCBASceI and S₁₂₃ transfection

A solution of 300 μL Opti-MEM (reduced serum media) containing 0.5 μL Lipofectamine 3000 and 0.25 μL P3000/0.5 μg pCBASceI plasmid complexes was added to HeLa cells and incubated at 37 °C. After 24 h, the cells were washed twice with PBS, after which a solution of 0.5 μL Lipofectamine 3000, 0.25 μL P3000, and S₁₂₃ (200 nM) complexes in 100 μL Opti-MEM was added to the HeLa cells for DNAzyme transfection for 4 h. Subsequently, 100 μL Opti-MEM containing 10 mM Mg²⁺ was added into the dishes and incubated for 30 min.

Confocal images

After the HeLa cell treated with pCBASceI, S₁₂₃, and Mg²⁺, the Hoechst 33258 was used to stain the nucleus. The images were collected using Zeiss LSM 880 confocal microscopy with different objectives. The Hoechst 33258, FAM, and Cy5 were measured with excitation at 405, 488 and 633 nm, respectively. The settings of laser intensity, pinhole, and gain were kept constant throughout the whole imaging process.