Development of a Novel Amplifiable System to Quantify Hydrogen Peroxide in Living Cells

Abstract

Although many redox signaling molecules are present at low concentrations, typically ranging from micromolar to sub-micromolar levels, they often play essential roles in a wide range of biological pathways and disease mechanisms. However, accurately measuring low abundant analytes has been a significant challenge due to the lack of sensitivity and quantitative capability of existing measurement methods. In this study, we introduced a novel chemically induced amplifiable system for quantifying low-abundance redox signaling molecules in living cells. We utilized H₂O₂ as a proof-of-concept analyte and developed a probe that quantifies cellular peroxide levels by combining the NanoBiT system with androgen receptor (AR) dimerization as a reporting mechanism. Our system demonstrated a highly sensitive response to cellular peroxide changes induced both endogenously and exogenously. Furthermore, the system can be adapted for the quantification of other signaling molecules if provided with suitable probing chemistry.

Graphical Abstract

Introduction

Fluorescent probes enable the studies of signaling molecules in live cells and have become indispensable tools in cell biology research. The development of the calcium probes by the Tsien group completely revolutionized the field of calcium signaling and took our knowledge about Ca²⁺ biological functions to a new level.¹ These calcium probes reversibly bind Ca²⁺ to shift their absorption maxima, thus producing ratiometric readouts that can quantify calcium concentrations in cells. Following the same concept, our group developed the first reversible reaction-based ratiometric glutathione (GSH) probe (ThiolQuant Green) that can quantify GSH concentrations in live cells,² along with many improved GSH probes with faster reaction kinetics and organelle specificities reported by our and other groups.^3-9 However, the success of GSH probes relies heavily on the high concentrations of GSH (1-10 mM) in cells, enabling fast reactions between GSH and the probes.

Redox signaling molecules, such as hydrogen peroxide (H₂O₂), hydrogen sulfide (H₂S) and nitric oxide (NO), have estimated concentrations in the micromolar or sub-micromolar range.^10-15 The pioneering work by the groups of Chang, Xian, Pluth, Nagano, Ai, Lippert, and many others has led to a myriad of reaction-based fluorescent probes for these redox signaling molecules.^16-22 Based on careful examination of the literature and our in-house experiments, we found that micromolar concentrations of probes are usually necessary to treat cells in the presence of exogenously added excessive redox signaling molecules to obtain meaningful qualitative fluorescence signals using conventional confocal microscopes within a reasonable time frame. The high concentrations of probes can significantly consume the analytes and disturb the biological system. Additionally, these probes are usually qualitative and cannot respond to the endogenous levels of the redox signaling molecules.

We set three criteria to design probes for low concentrations of analytes. (i) The probe concentration used should be less than 1% of the analyte concentration to minimize perturbation of the biological system. (ii) The probe should be able to quantitatively measure the analyte concentrations in live cells. (iii) The assay time should be within a convenient experimental time frame, ideally 1-2 hours.

With these criteria in mind, we initially attempted to follow our work on GSH probes² to identify reversible reactions with redox signaling molecules and realized the technical challenges for this route. For most of the reaction-based probes, one equivalent of probes produces an equal amount of either fluorescent species, or enzyme substrates, or photons from chemiluminescence upon reacting with the analytes. ^22-27 Considering the analytes are in the micromolar range, we would need to keep the probe concentration in the range of low nanomolar. For equal molar conversion from the nanomolar probe to signals, it would be very challenging to detect with either conventional confocal fluorescence microscope or luminescence readouts, which could explain why micromolar concentrations of probes were applied in previous studies.

To address these technical issues, strategies to amplify the output signals are needed. The groups of Renslo and Wells reported a seminal study to develop amplifiable ferrous iron (Fe²⁺) probes.²⁸ In this work, a Fe²⁺ responsive group caged puromycin reacts with intracellular Fe²⁺ to regenerate puromycin, which can be incorporated into the C-terminus of elongating nascent peptide chains to terminate translation and detected using horseradish peroxidase (HRP)-based immunofluorescence (IHC) after fixing the cells. Essentially, one equivalent of released puromycin can be converted to an equal amount of HRP, an enzyme that produces amplified signals. The Chang group applied this strategy to develop peroxymycin-1, a H₂O₂ responsive group caged puromycin.²⁹

The caveats of the puromycin system are two-folds. First, micromolar concentrations of the caged puromycin probes are still needed, which could consume significant amount of analytes in a similar concentration range. Considering the Michaelis-Menten constant (K_m) or dissociation equilibrium constant (K_d) for puromycin as a substrate for ribosomes is in the μM range,^30-33 it is not surprising that micromolar of the caged puromycin probes are necessary for this strategy. Ideally, we need to identify small molecules that have K_m or K_d in the low nM or sub-nM range. Second, this strategy can only be applied in fixed cells. However, the fixation process may change the level of these redox analytes.

Hormones are Nature’s chemical biology tools to regulate gene expression through hormone receptors.³⁴ For example, estrogens bind estrogen receptors (ER) with nM binding affinities to activate gene expression.³⁵ This progress can be hijacked as an ER activity reporter assay by expressing a plasmid that has the estrogen responsive element (ERE) with downstream reporter genes, such as fluorescent proteins or luciferases.³⁶ Our initial attempt to develop an amplifiable system for low concentrations of analytes is to develop a caged estrogen (E2) that can be released in the presence of H₂O₂ to subsequently drive firefly luciferase expression. We chose H₂O₂ as the proof-of-concept study because the H₂O₂ responsive chemistry has been well established and the reaction rates between H₂O₂ and boronic esters are relatively fast.^27,37 Unfortunately, our initial design has two major drawbacks. In the absence of estrogens, ER has leaky background activities, which can be as high as 10% of the maximum activity in the presence of saturating concentration of estrogens, thus limiting the dynamic range of the assay (Figure S1A). To avoid the high background noise issue, the AR/ARE (androgen receptor/androgen response element) system was evaluated as well (Figure S1B).³⁸ However, one notable drawback of such systems is that the transcription-based assays take 12-24 hours to accumulate enough reporter proteins for detection. Oxidative stress may change the rate of transcription and translation of the reporter proteins, which renders it difficult to dissect the levels of reporter proteins change is due to the uncaging of pro-estrogens/pro-androgens or secondary effects.^39,40

With the setbacks in mind and inspired by our work in proteolysis targeting chimeras (PROTACs),^41,42 we turned to the chemically induced dimerization (CID) systems,⁴³ in which proteins are expressed before perturbation to the biological system is introduced. Among the CID systems, we chose androgen receptor (AR) as our platform because dihydrotestosterone (DHT) triggers rapid homo-dimerization of AR with a tight intermolecular interaction.^44-46 For reporters, we chose luciferase-based bioluminescence instead of fluorescent proteins due to their high sensitivities. Promega developed nano-luciferase (nLuc), a bioluminescent enzyme with an extremely high catalytic efficiency (k_cat/K_M = 183 μM⁻¹s⁻¹), allowing detection of the enzyme in the picomolar range.^47,48 Additionally, Promega also developed a split version of nLuc, namely NanoBiT.⁴⁹ The complementary fragments of nLuc are called LgBit and SmBit and can be fused to proteins to study protein-protein interactions. AR-LgBit and AR-SmBit fusions have been developed as a reporter assay for androgens, such as DHT (Figure 1A). ^50,51 With these tools in hands, we developed a Chemically Induced Dimerization-based Amplifiable Probe (CIDAP) to quantify H₂O₂ levels in living cells.

Figure 1.

Design of Chemically Induced Dimerization-based Amplifiable Probes (CIDAP). (A) LgBiT or SmBiT are fused to androgen receptor (AR) and transiently co-expressed in mammalian cells. Addition of androgens (such as DHT) triggers AR dimerization and the reconstitution of LgBiT and SmBiT to form nLuc. DHT can be converted to a prodrug (i.e., the probe), which reacts with redox signaling molecules (such as H₂O₂) to regenerate DHT. (B) The structures of DHT, CIDAP, CIDAP-NC1, and CIDAP-NC2. (C) The reaction mechanism of CIDAP and its cascade to release DHT. CIDAP-NC1 is the negative control that cannot regenerate DHT. CIDAP-NC2 is used as the control to investigate the impact of enzymatic hydrolysis on carbonate.

Results

Design of CIDAP and Evaluation of Its Activity in Cells.

Examining the co-crystal structure of DHT and AR, we found that 17-β-OH in DHT forms a key interaction with AR (PDB: 5JJM).⁵² We developed a CIDAP platform by caging the 17-β-OH in DHT with a H₂O₂ responsive boronic ester group to abolish its binding to AR (Figure 1B and 1C).¹⁶ We initially developed a probe without the carbonate group (CIDAP-NC1) but failed to release DHT in the presence of H₂O₂, given that alcohol is a poor leaving group. CIDAP-NC1 serves as a negative control in the following experiments.

To evaluate whether the carbonate group can be directly hydrolyzed by intracellular enzymes to release DHT in the absence of H₂O₂ (Figure 1C), we developed a CIDAP analog without the boronic ester group, CIDAP-NC2. Density functional theory (DFT) calculations showed that the carbonate carbons in CIDAP and CIDAP-NC2 have similar partial charge densities, suggesting similar hydrolysis rates (Table S1). Moreover, CIDAP was directly incubated with recombinant nLuc protein to confirm that it does not suppress or enhance the luminescence of nano-Luciferase (Figure S4).

We used HeLa cells as a model system to test if CIDAP can respond to endogenous levels of H₂O₂. AR-LgBit and AR-SmBit were transiently transfected and co-expressed in HeLa cells, followed by incubation with CIDAP, a positive control DHT, and two negative controls CIDAP-NC1 and CIDAP-NC2 at 1.1 nM (Figure 2A). As expected, nM levels of DHT and CIDAP trigger bioluminescence with a high signal-to-background ratio (S/N = ~40). The lower signals from the CIDAP group could be attributed to its incomplete reaction with H₂O₂ under the experimental conditions. CIDAP-NC1 and CIDAP-NC2 produce minimal bioluminescence compared to CIDAP, indicating that bioluminescence produced by CIDAP is not contributed from CIDAP direct binding to AR without uncaging or uncaging from direct hydrolysis of the carbonate. Additionally, the reactivity of CIDAP towards other reactive oxygen species (ROS) is insignificant, indicating a high degree of selectivity towards H₂O₂ (Figure S3A), consistent with previous studies.²⁷ We further examined CIDAP's reactivity towards lower concentrations of peroxynitrite and hypochlorite. As anticipated, hypochlorite did not react with CIDAP. In contrast, consistent with previous findings, boronic acid rapidly reacted with peroxynitrite. Interestingly, we observed that, following the oxidation of boronic acid, peroxynitrite proceeded to react with the produced DHT and oxidize the 17-β-OH group (Figure S3B). Despite these results suggesting that peroxynitrite can react with CIDAP and deplete DHT, considering the final product would not induce the reconstitution of nLuc, this side reaction is unlikely to impact the final quantification of peroxide.

Figure 2.

Refinement and assessment of CIDAP system as a quantifying method for cellular peroxide level. (A) The bioluminescence response of DHT, CIDAP, CIDAP-NC1 and CIDAP-NC2. AR-LgBit and AR-SmBit are co-expressed in HeLa cells for 24 h, followed by the incubation with control (ethanol) and each compound at 1.1 nM, separately. The readouts were obtained after 2 h of incubation. Data are mean ± SEM of experimental replicates (n = 4). (B) Simulated Data for DHT and Probe responses. Black trace: DHT response; red trace: Probe response; pale green trace: the ratio of DHT and Probe responses at the same concentrations. (C) The dose-response curves of DHT, CIDAP, and CIDAP-NC1 in HeLa and HEK293T cells. Data are mean ± SEM of experimental replicates (n = 4). Both cell lines were incubated with different concentrations of DHT, CIDAP and CIDAP-NC1 (from 11 nM to 5 pM with 3-fold dilutions) for 2 h. Bioluminescence was measured in the presence of furimazine. (D) H₂O₂ levels in different cell lines. The SEM on the H₂O₂ level calculated using the global fitting program provided by GraphPad Prism (see Supporting Information for details).

Quantitative Method to Convert CIDAP Signals to Peroxide Concentrations.

The next question is how we can quantitatively convert CIDAP signals to H₂O₂ concentrations. We adopt the concept of dose-response shift in pharmacology as the solution. In cells co-expressing AR-LgBiT and AR-SmBiT, addition of DHT triggers nLuc formation and bioluminescence with a classical sigmoidal dose-response curve (black trace in Figure 2B). Addition of the probe with the same concentration series as DHT to cells would also generate a dose response curve (red trace in Figure 2B). When the concentrations of the probe used are significantly lower than those of H₂O₂, the peroxide concentration remains constant throughout the assay. Consequently, the reaction between the probe and peroxide follows the pseudo first-order reaction kinetics. In a first-order reaction, the percentage of conversion remains constant, regardless of the initial concentration of the reactant (grey arrows in Figure 2B). During the measurement, different probe concentrations can be introduced to the cells to initiate partial reactions with peroxide. The constant percentage of CIDAP-to-DHT conversion results in a right shift of the dose-response curve. According to the pseudo first-order reaction law, the relationship between the peroxide concentration and the percentage of probe conversion can be expressed by $\text{ln}(1-\alpha )=-k[{\mathrm{H}}_2{\mathrm{O}}_2]t$, where $k$ is the secondary reaction rate constant between H₂O₂ and CIDAP, $\alpha$ is the percentage of probe conversion, and $t$ is the reaction time. The secondary reaction rate constant $k$ is determined as 196.6 M⁻¹min⁻¹ (Figures S2A and S2B). The percentage of CIDPA conversion α can be calculated using the ratio of the EC₅₀ values of DHT and CIDAP (EC_{50, DHT}/EC_{50, CIDAP}) from the dose-response curves.

Our dose-response shift approach offers a theoretical framework to quantify low concentration analytes using amplifiable systems. In Spangler et al.’s puromycin-based system, a series of concentrations of puromycin and the probe indeed generated dose-response shift curves (Figure 2 in reference 28), consistent with our analysis. However, in the following applications, they only employed a single concentration of puromycin and the probe (Figures 3 and 4 in reference 28). It is worth noting that the response ratios (pale green triangles in Figure 2B) are dependent on the initial concentration used for probe and benchmark (e.g., DHT or puromycin). A single concentration of probe and benchmark can only provide a qualitative measurement.

Quantification of Endogenous Peroxide Levels.

HeLa and HEK293T cells were employed as model systems to refine the experimental protocols. The optimal incubation time was determined to be 2 h to balance signal outputs and prevent complete CIDAP conversion (Figure S5). Following the optimized protocol, cells expressing AR-LgBit and AR-SmBit were incubated with DHT, CIDAP, and CIDAP-NC1 (concentrations ranging from 11 nM to 5 pM with 3-fold dilutions) to generate dose-response curves (Figure 2C). Based on the equation, we determined the concentrations of peroxide in HeLa and HEK293T cells to be 10.09 ± 1.89 and 5.4 ± 1.73 μM, respectively. Moreover, we applied this approach to three additional cell lines, namely AsPC1, HepG2, and MDA-MB-231, and obtained H₂O₂ concentrations ranging from 1.8 to 10 μM (Figure 2D, Figure S6). It should be noted that the fact that the CIDAP concentration applied is <1% of measured H₂O₂ levels supports our pseudo-first-order reaction assumption.

Quantification of Peroxide Level Changes upon Exogeneous H₂O₂ Treatment.

In order to expand the scope of application for this system and evaluate its responsiveness to changes in the cellular environment, an exogenous H₂O₂ concentration of 50 μM was introduced to elevate the changes of H₂O₂ levels in cells. This resulted in an increase in cellular peroxide levels from 5.4 ± 1.73 μM to 21.9 ± 9.3 μM (Figure 3A). Additionally, a range of H₂O₂ concentrations between 0 and 50 μM were introduced, with the calculated peroxide levels demonstrating an increasing trend that was consistent with the increasing amounts of exogenous H₂O₂ added (Figure S7). These findings confirmed the sensitivity of the method for monitoring changes in H₂O₂ levels within cells. However, it is important to note that the detected changes in overall peroxide concentration in the cellular environment may not necessarily reflect accurate changes in intracellular peroxide levels, as both peroxide and the probe can permeate membranes.^53,54

Figure 3.

Quantification of H₂O₂ levels in cells under various conditions. The response of CIDAP is accurately reflecting the fluctuation of cellular peroxide level. (A) H₂O₂ levels in HEK293T cells upon exogenous 50 μM H₂O₂ treatment. Data on the left are mean ± SEM of experimental replicates (n = 4). (B) H₂O₂ levels in HeLa cells pretreated with 10 μM menadione. Data on the left are mean ± SEM of experimental replicates (n = 4). (C) H₂O₂ levels in HeLa wild type and Grx-3 knockdown cell. Data on the left are mean ± SEM of experimental replicates (n = 3). (D) H₂O₂ levels in 293 cells pretreated with 1 mM NAc. The H₂O₂ level was measured after 1 hour of incubation with DHT/CIDAP. Data on the left are mean ± SEM of experimental replicates (n = 3). The SEM on the H₂O₂ level in (A) – (D) were calculated using the global fitting program provided by GraphPad Prism.

Quantification of Peroxide Level Changes under Biological Perturbations.

First, we assessed the responsiveness of the CIDAP system to fluctuations in endogenous peroxide levels induced by small molecule treatment. Menadione has been reported to induce H₂O₂ through intracellular redox cycling.^55,56 DHT and CIDAP were added to HeLa cells pretreated with menadione. Our findings revealed a significant increase in H₂O₂ level from 10 μM to 21 μM upon menadione treatment (Figure 3B).

Furthermore, we aim to explore the utility of CIDAP to evaluate peroxide level changes due to genetic perturbations. Previous research has demonstrated that mammalian glutaredoxin 3 (Grx3) plays a critical role in maintaining cellular redox homeostasis, and its downregulation results in high oxidative stress.⁵⁷ To assess the potential applicability of our system in such biological contexts, we applied it to both wild-type and Grx3 knockdown HeLa cells. Our results showed a significant increase in peroxide levels from 10 μM in the parental cells to approximately 39 μM in the Grx3 knockdown cells (Figure 3C), thereby demonstrating the feasibility of this system in quantifying peroxide levels upon genetic perturbations.

Moreover, the downregulation of peroxide level was also assessed with CIDAP system. N-acetylcysteine (NAc), a well-established antioxidant molecule, was utilized to scavenger ROS within the cells.^58,59 Pre-treatment of cells within 1 mM NAc prior to the addition of DHT and CIDAP resulted in a proximate one-fold reduction in peroxide levels as determined by our CIDAP system. This demonstrates the versatility of the CIDAP system, which is capapble of quantifying both increases and decreases in peroxide concentration effectively. (Figure 3D)

A Simulation Model for CIDAP to Validate Experimental Measurements.

A previous study using protein-based peroxide sensor, such as HyPer, estimated the endogenous basal level of H₂O₂ is in the low nM range for K562 cells,⁶⁰ which is in disagreement with our measurements of micromolar peroxide levels in typical cancer cell lines. One possible explanation for the discrepancy can be that HyPer is mainly localized in the cytoplasm with no access to peroxide inside organelles like mitochondria or peroxisomes, which generate the majority of H₂O₂ inside cells. As a small molecule, CIDAP can diffuse freely through membranes. The results from CIDAP are considered an overall resting status of peroxide in the whole cellular environment instead of one particular compartment. To corroborate our experimental measurements, we developed Matlab Simbiology models to simulate the entire process, incorporating all possible steps and the corresponding kinetic parameters from the treatment of the compounds to the final reconstitution of nLuc led by AR dimerization (Figures S8 and S9, Tables S2 and S3). The kinetic parameters used in this simulation are based on our measurements and values reported in the literature. If the peroxide level were set to 10 nM (i.e., the pseudo-first-order reaction law is no longer applicable), it would take more than 8 hours to achieve signal saturation at 10 nM of CIDAP concentration, (Figure S10), which contradicted to the experimental results, in which 2 h is sufficient to achieve signal saturation with high concentrations of probes. In another word, if the intracellular H₂O₂ concentrations were in the low nM range under our experimental conditions, we would not observe the results in Figures 2 and 3. The simulation results also revealed that higher peroxide levels in the μM range resulted in a leftward shift of the simulated CIDAP curves at 2 h, which agrees with our experimental data.

Additionally, we tested the effect of compound permeability rate crossing the cell membrane by varying the permeability half-life of the compound from 1 minute to 20 minutes and found that it had a minimal impact on the simulation results (Figure S11). This was also experimentally validated by treating cells with 0.05% digitonin to permeabilize cells (Figure S12). Notably, no significant difference in luminescence signals were observed for DHT induced AR dimerization with or without digitonin treatment, suggesting rapid transmembrane permeability for DHT.

Overall, our simulation model corroborates our experimental results. We tentatively attribute the different H₂O₂ levels measured in our and previous studies to the difference in experimental conditions and cell lines used. Additionally, two key cysteine residues in HyPer form a disulfide bond upon reacting with H₂O₂. And the oxidized form of HyPer can be reduced by thioredoxins in an NADPH dependent manner.⁶¹ So Hyper could potentially serve as a catalyst for H₂O₂ decomposition at the expense of NADPH. Therefore, we are uncertain whether the low nM H₂O₂ levels in cells are indeed perturbed by HyPer. In contrast, CIDAP reacts with H₂O₂ irreversibly and does not catalyze H₂O₂ decomposition. CIDAP is not expected to significantly change the H₂O₂ levels due to the low concentrations of CIDAP used.

Discussion

The Lippert group reported an elegant kinetics-based quantification method for peroxynitrite in solution.⁶² Overall, the fluorescent probe field for reactive species still remains at a qualitative stage. Our work provides a proof-of-concept strategy to quantify these low abundance reactive species.

It should be noted that in all our plots, we assume the DHT and CIDAP concentrations are the same inside and outside of cells. However, these hydrophobic molecules may preferentially accumulate inside cells. If assuming the enrichment ratios for DHT and CIDAP are the same and the probe concentration in cells is less than 10% of H₂O₂ concentration to satisfy the pseudo-first-order assumption, our calculated H₂O₂ levels will still hold. If the enrichment ratios for DHT and CIDAP are not the same, this will introduce a systematic error for the computed H₂O₂ concentrations.

Although we have successfully demonstrated the capability of quantitatively measuring peroxide using the CIDAP system, there is still room for improvement. Primarily, the entire quantification process relies on the rate constant measured in PBS (pH 7.4) at 37°C. However, this rate may vary within a cellular environment due to factors such as salt concentrations and localized pH variations, introducing potential errors. However, measuring the actual rate constant of CIDAP within cells is challenging due to our inability to precisely control intracellular H₂O₂ levels.

The system can only measure the average concentration of peroxide and cannot capture rapid and transient changes in live cells. The main factor underlying this limitation is the slow reaction kinetics between the boronic ester in CIDAP and H₂O₂, with a bimolecular reaction rate constant of only ~200 M⁻¹min⁻¹ (3.33 M⁻¹s⁻¹). It requires 2 hours of incubation to achieve sufficient signal-to-noise ratios. Therefore, the measured H₂O₂ levels are indeed the averaged levels during the experimental time frame. Thus, when applying this system to determine the variation in peroxide levels among different cellular conditions, the fold variations between conditions are more accurate and relevant, as compared to the absolute numerical output produced by the CIDAP system. Ideally, a probe featuring faster kinetics and enabling completion of the assay within 5 minutes, such as Hyper, would provide a more precise measurement of peroxide level changes in real time.

A potential issue for the CIDAP system pertains to the impact on intrinsic peroxide levels due to androgen and the overexpression of AR, especially in hormone-associated systems like prostate cancer and breast cancer.^63-66 In such scenarios, the peroxide levels obtained from the CIDAP system may not accurately represent the natural peroxide levels in these systems. To overcome this limitation, other non-hormone based chemically induced dimerization systems can be considered.

When discussing the measurement and quantification of peroxide, it is impossible to overlook the significance of HyPer and roGFP2-Tsa2ΔC_R fluorescent proteins.^67-71 These genetically encoded, oxidant-sensitive fluorescent probes have revolutionized the field due to their unique features. Hyper reacts rapidly and allows real-time monitoring of H₂O₂ level in vivo, while also enabling compartmental resolution through genetic engineering. The latest version, HyPer7, has surpassed its predecessors by addressing the issue of pH sensitivity and is highly sensitive to small changes in peroxide levels.⁷⁰ With meticulous calibration, HyPer can quantify peroxide levels at the nanomolar level using flow cytometry in K562 cells.⁶⁰ This result is significantly lower than the concentration detected using the CIDAP system. One possible explanation is that HyPer primarily localizes in the cytoplasm, thus potentially reflecting the overall concentration of cytosolic peroxide. In contrast, CIDAP, as a small molecule, can freely distribute between the cytoplasm and subcellular organelles. Therefore, it represents the overall peroxide level in the entire cellular environment, including organelles such as mitochondria and peroxisomes, which have high peroxide levels. To demonstrate CIDAP's capability of penetrating subcellular organelles, CIDAP was incubated with freshly prepared mitochondria from mouse livers for 2 hours. Results showed that CIDAP is permeable to mitochondrial membranes and detected inside the mitochondria (Figure S13). Furthermore, DHT was also detected within the mitochondria, indicating that portion of CIDAP had already reacted with mitochondrial peroxide (Figure S13).

Additionally, HyPer7 can be easily saturated for moderate to high concentrations of H₂O₂, limiting its application in pathological conditions. In a comparative study wherein cells expressing either Hyper7 or AR-NanoBit were exposed to varying peroxide concentrations ranging from 0-50 μM, the results from the Hyper7 group aligned consistently with prior reported findings. The observed signal was easily detectable and dosage-dependent within the 0-10 μM peroxide range (Figure S7). However, at peroxide levels surpassing 20 μM, the signal reached saturation and overlapped with the signal observed at 10 μM. In contrast, the CIDAP system demonstrated unique performance characteristics. At lower peroxide levels, signal distinctions were less discernible, yet they became increasingly marked as the peroxide concentration exceeded 10 μM.

This is where the CIDAP system presents itself as an invaluable tool that can adapt seamlessly across a wide range of peroxide levels – particularly within systems with elevated oxidative stress. It should be noted though that whilst the CIDAP system serves as an exceptional alternative for conditions involving heightened peroxide levels surpassing the detection limits of HyPer, it does not replace them entirely. Furthermore, the main purpose of the CIDAP system is to serve as a platform to demonstrate the potential applications of this system on low-abundance analytes in general, rather than being specifically fine-tuned for H₂O₂ measurement.

Conclusions

In summary, we presented a novel amplifiable system for the quantitative measurement of peroxide. Through rigorous testing and evaluation, our innovative approach effectively met all the criteria we established for developing probes designated to analytes in low abundance. The system demonstrated sufficient flexibility and sensitivity across multiple cell lines, which further emphasizes its adaptability to diverse biological systems. Besides, the low concentration of probe required for the assay introduces minimal disturbance and toxicity in the cellular environment, which suggests potential applications to a broad range of biological systems to investigate H₂O₂ in redox signaling processes and disease development. Our findings provide proof-of-concept evidence for the feasibility of this approach and shed light on the further improvement of probes for low-abundance analytes. With our strategy being extendable to measure other redox such as H₂S and NO, this work serves as a testament to the significance of creating sensitive yet flexible tools capable of investigating essential but elusive analytes involved in intricate and dynamic processes like redox signaling.