Evaluation of borinic acids as new, fast hydrogen peroxide–responsive triggers

Significance

Hydrogen peroxide plays an important role in the fine balance between physiological and pathological processes. To detect this diffusible small molecule, we expanded the scope of organic triggers in developing borinic acids as an alternative and more sensitive trigger than the most conventional boronate-based sensors. We discovered that borinic acid is 10,000-fold more reactive than its boronic counterpart toward H₂O₂-mediated oxidation. An accurate determination of oxidation kinetic constants and computational experiments corroborate this higher reactivity. This improvement also proved effective for in-cell detection of exogenously as well as endogenously produced H₂O₂. We believe borinic acids represent a new and efficient tool allowing for the development of new devices for a better understanding of H₂O₂-mediated signaling processes.

Abstract

Hydrogen peroxide (H₂O₂) is responsible for numerous damages when overproduced, and its detection is crucial for a better understanding of H₂O₂-mediated signaling in physiological and pathological processes. For this purpose, various “off–on” small fluorescent probes relying on a boronate trigger have been prepared, and this design has also been involved in the development of H₂O₂-activated prodrugs or theranostic tools. However, this design suffers from slow kinetics, preventing activation by H₂O₂ with a short response time. Therefore, faster H₂O₂-reactive groups are awaited. To address this issue, we have successfully developed and characterized a prototypic borinic-based fluorescent probe containing a coumarin scaffold. We determined its in vitro kinetic constants toward H₂O₂-promoted oxidation. We measured 1.9 × 10⁴ m⁻¹⋅s⁻¹ as a second-order rate constant, which is 10,000-fold faster than its well-established boronic counterpart (1.8 m⁻¹⋅s⁻¹). This improved reactivity was also effective in a cellular context, rendering borinic acids an advantageous trigger for H₂O₂-mediated release of effectors such as fluorescent moieties.

Reactive oxygen species (ROS) are involved in various physiological processes. In particular, hydrogen peroxide (H₂O₂) plays a critical role in the regulation of numerous biological activities as a signaling molecule (1, 2). However, aberrant production or accumulation of H₂O₂ leads to oxidative stress conditions, which can cause lesions associated with aging, cancer (3), and several neurodegenerative diseases such as Alzheimer’s or Parkinson’s (4, 5). Differentiation of physiological or abnormal conditions is closely connected with slight changes in H₂O₂ levels. However, the generation and degradation of H₂O₂ are variable within different cellular compartments, and this small molecule is highly diffusive, rendering the capture of small H₂O₂ fluctuations and the study of its spatial and temporal dynamics difficult. Therefore, the development of selective and sensitive H₂O₂-reactive tools for applications in a biological context represents a challenge for a better understanding of H₂O₂-mediated signaling in physiological and pathological processes or the use of H₂O₂ activation for the release of biological effectors (6).

Numerous strategies have been developed to implement H₂O₂-reactive molecular triggers, as exemplified by “off–on” small fluorescent probes. Such probes have attracted particular attention due to their easy implementation, high expected signal-to-noise ratio, and compatibility with standard equipment present in cellular biology research environments (7–9). Activation in such a context is triggered or modified by H₂O₂-mediated transformation of a suitable chemical species. Several approaches have been explored including probes based on arylsulfonyl ester perhydrolysis (10), oxidation of arylboronates (11), Baeyer–Villiger oxidation of diketones (12), Tamao oxidation of silanes (13), a tandem Payne–Dakin reaction (14) or a seleno-Mislow–Evans rearrangement (15). Among them, designs based on the boronate esters oxidation pioneered by Chang are the most explored, due to their remarkable stability, low toxicity profile, ease of preparation, and specificity toward H₂O₂, as illustrated in recent reviews (16–18). Upon reaction with H₂O₂, these compounds undergo an oxidative conversion into aryl borate esters that further hydrolyze into the corresponding phenols along with borate esters or boric acid (Scheme 1A). This conversion turns on probe fluorescence or activates drug release either directly or via the degradation of a self-immolative spacer. This chemospecific and biologically compatible reaction allowed, for instance, developing highly selective fluorescent probes for H₂O₂ imaging in cells (19–23). However, H₂O₂-triggered conversion of boronic acids to phenols is still not completely satisfactory in a biological context (24) since most of these probes have second-order reaction rate constants of 0.1 to 1.0 m⁻¹⋅s⁻¹ (14). In cells, H₂O₂ is present in the 1 to 100 nm concentration range in physiological conditions and could reach up to 100 μm under oxidative stress conditions (25). Therefore, most of the boronate-based systems need an incubation time longer than 30 min for activation at an H₂O₂ concentration of 100 μm. At such a time scale, H₂O₂ typically diffuses over a distance of 2 mm (evaluated as (D_H2O2τ)^0.5 with D_H2O2 = 1.7 × 10⁻⁹ m²⋅s⁻¹ from ref. 26 and τ = 30 min). Hence, to improve spatial resolution for H₂O₂ imaging, alternative H₂O₂ triggers with rapid reaction rates allowing real-time activation by H₂O₂ are still required.

Scheme 1.

(A) Current boronic acid (R = H) or boronate (R,R = tetramethylethylene) as H₂O₂-responsive group releasing a hydroxyaryl as effector and a boric acid or a borate ester respectively. (B) This study: a borinic acid as H₂O₂-responsive group releasing a hydroxyaryl as effector and a boronic acid.

To address this issue, we envisioned the use of borinic acids, structures in which one of the boron–oxygen bonds of the boronic acid is replaced by a boron–carbon bond. Due to these electronic modifications, borinic acids exhibit more electrophilic properties (27–30) compared to their boronic acid counterparts and could be more prone to rapid oxidation. These structures have been mainly exploited as catalysts in various reactions such as epoxide ring opening (31), hydrosilylation (32), transamidation (33), aldol reaction (34, 35), C–H activation (36, 37), selective monoalkylation, acylation and sulfonation of diols (38, 39), or regioselective glycosylation reactions (40–42). Surprisingly, the reactivity of these borinic species remains underexplored (43–45), probably due to their limited synthetic access (46–49). They were usually obtained through the addition of strong organometallic reagents (RLi/RMgBr) onto boron-based electrophiles such as trialkylborates, boron halides, diborane, or boronate esters. To date, a detailed study of the reactivity of borinic acids toward oxidation including reaction with H₂O₂ has not been reported and their use as triggers for the direct release of a probe or an effector has not been considered.

Herein, we report the design, synthesis, and evaluation of a borinic-triggered prototypic probe prone to direct and rapid activation by the H₂O₂ molecule (Scheme 1B). We establish a detailed kinetic analysis of the H₂O₂-promoted oxidation of this borinic acid as well as a comparative study with its corresponding boronic analog. Furthermore, we demonstrate the shorter response time of the borinic trigger compared to the boronic trigger against H₂O₂-mediated oxidation in a cellular environment.

Results and Discussion

Borinic Acids React Faster with H₂O₂ than Boronic Acids.

First, in order to validate the proof of concept, we compared the reactivity of simple model compounds toward H₂O₂, i.e., diphenylborinic acid, synthesized from addition of phenyl lithium onto pinacol phenylboronate (47, 50), and the commercially available phenylboronic acid. These two compounds were submitted to oxidation with one equivalent of H₂O₂ in deuterated phosphate-buffered saline (PBS, pH 7.4). The reaction progress was monitored using ¹H NMR spectroscopy. These early experiments were extremely encouraging since we observed a complete oxidation of the diphenylborinic acid into phenol and phenylboronic acid within 2 min, whereas 2 h were required for full conversion of the phenylboronic acid into phenol under the same conditions (SI Appendix, Fig. S1). The experimental results revealing the superior reactivity of borinic acid were clearly corroborated by density functional theory (DFT) calculations performed at the M062X/6-311G++(d,p) level (see SI Appendix for details), showing a difference of activation energy for the initial nucleophilic addition of the perhydroxyl anion onto the boron center of both diphenylborinic and phenylboronic acids of 6.5 kcal⋅mol⁻¹ in favor of the former.

Design and Synthesis of a Model Fluorogenic Borinic Probe.

Capitalizing on these preliminary results, we then designed a borinic acid-based fluorogenic probe in which the boron atom is substituted by a phenyl moiety and a 4-methylcoumarin scaffold as a fluorescent chromophore. The synthetic route involved a four-step sequence in which the key reaction relies on the addition of phenyl lithium onto a coumarin-based pinacol boronic ester (Scheme 2A). Thus, treatment of 4-methylumbelliferone (4-MU) with trifluoromethanesulfonic anhydride afforded the corresponding triflate 1 in 97% yield, which was subjected to a Pd-catalyzed cross-coupling reaction with bis(pinacolato)diboron under microwave irradiation according to Chang’s conditions (20, 51). The resulting pinacol boronic ester 2 was then reacted with a solution of phenyl lithium to afford the borinic acid, which was immediately converted into the corresponding N,N-dimethylaminoethyl ester 3. This compound was isolated by precipitation in 24% yield over two steps. Acidic hydrolysis then provided the expected borinic acid 4 in 67% yield. The corresponding 4-methylcoumarin boronic acid 5 was also synthesized for comparison. It was easily obtained from the pinacol ester using recently reported pinacol exchange reaction with methylboronic acid (Scheme 2B) (52).

Scheme 2.

(A) Synthesis of 4-methylcoumarin–based borinic acid 4. Conditions: a) Tf₂O (1.1 equiv), Pyr (2.2 equiv), CH₂Cl₂, RT, 2 h, 97%; b) bis(pinacolato)diboron (1.2 equiv), KOAc (1.5 equiv), Pd(dppf)Cl₂·CH₂Cl₂ (10 mol %), toluene, microwave irradiation, 120 °C, 3 h, 65%; c) PhLi in n-Bu₂O (1.9 m, 1.1 equiv), −78 °C, 4 h then HCl in Et₂O (2 m, 7 equiv), −78 °C to room temperature (RT); d) N,N-dimethylethanolamine (2 equiv), Et₂O, RT, 12 h, 24%; e) HCl aq 1 m, EtOAc, RT, 30 min, 67%. (B) Synthesis of 4-methylcoumarin boronic acid 5: f) MeB(OH)₂, TFA, CH₂Cl₂, RT, 4 h, 87%.

This Probe Confirms Higher Reactivity of Borinic Acids.

H₂O₂-mediated oxidation of 4 and 5 was first monitored by ¹H NMR spectroscopy in deuterated PBS buffer (pH 7.4). As previously observed with the phenyl-based model compounds, the borinic derivative was found to be more reactive than its boronic counterpart. Thus, the addition of H₂O₂ triggers full oxidation of the borinic acid 4 in less than 2 min. In contrast, the boronic acid 5 is gradually converted into 4-MU to reach complete conversion within 2.5 h (SI Appendix, Fig. S2). Due to the unsymmetrical nature of the borinic acid probe, the oxidative rearrangement of 4 leads to the formation of two different alcohol/boronic acid couples, i.e., either 4-MU/phenylboronic acid as the expected cleavage products or phenol/4-methylcoumarin boronic acid 5 as the undesired cleavage products. The ratio of products was determined by ¹H NMR integration and was evaluated to be 20/80, against the direct release of 4-MU (Fig. 1).

Fig. 1.

Regioselectivity of the oxidation of probe 4 determined by ¹H NMR spectroscopy (600 MHz) in deuterated PBS buffer (pH 7.4, 5% DMSO). ¹H NMR spectra of the borinic acid 4 (A) and 4 with H₂O₂-urea complex (1 equiv) after 2 min (B). (C) The two different couples of alcohol/boronic acid resulting from the oxidation reaction. 4-MU = 4-methylumbelliferone.

Study of the Oxidation Kinetics of the Model Borinic Probe with H₂O₂.

Having demonstrated the higher reactivity of the borinic trigger, we then undertook the accurate determination of the kinetic constants of the oxidation reactions (Fig. 2A). To calculate these parameters, we took advantage of the profluorescent properties of 4 and used real-time measurement of the fluorescence emission performed under pseudo-first-order conditions (excess of oxidant). Initial H₂O₂-promoted oxidation experiments were conducted with boronic acid 5 at different temperatures and H₂O₂ concentrations. As expected, reaction with H₂O₂ leads to an increase in fluorescence emission at 450 nm, indicating the release of the 4-MU chromophore (Fig. 2B).

Fig. 2.

Kinetic study of oxidation of borinic acid 4 and boronic acid 5. (A) Simplified kinetic scheme of oxidation of the borinic acid 4. (B and C) Time-dependent fluorescence intensity evolution of 5 μM 5 (B) or 4 (C) upon addition of 100 μm H₂O₂-urea solution in PBS buffer 1× (pH 7.4, 0.04% DMSO) at 310 K (pseudo-first-order conditions). The fluorescence emission was normalized and was recorded at λ_em = 450 nm (λ_ex = 405 nm). a.u., arbitrary units.

For kinetic analysis of the experimental data (see SI Appendix for the determination), a model including the different steps involved in the 4-MU release was taken into account allowing determination of the second-order rate constants k₃ associated with the boronic acid oxidation as 1.8 m⁻¹⋅s⁻¹ and 6.3 m⁻¹⋅s⁻¹ at 293 and 310 K, respectively (SI Appendix, Fig. S7). These results are consistent with those previously reported in the literature (53–56). In comparison, oxidation of borinic derivative 4 shows a different behavior. Upon addition of H₂O₂, fluorescence rapidly increases, which suggests an oxidation of the borinic acid probe 4 on the minute time scale affording directly 4-MU through the desired, although minor, oxidation pathway. After a few minutes, fluorescence increases more slowly on a time scale, which is in line with the formation of 4-MU from the boronic acid 5 obtained as a major product from oxidizing probe 4 (Fig. 2C). Upon considering that several reactions are rapid on the time scale of the overall oxidation process, we could account for the oxidation of the borinic acid 4 with the reduced mechanism shown in Fig. 2A. After integration of the kinetic results obtained above for the boronic acid 5, we extracted 1.9 × 10⁴ and 4.2 × 10⁴ m⁻¹⋅s⁻¹ for the rate constants k₁ at 293 and 310 K, respectively (SI Appendix, Fig. S10). These results point to an oxidation of the borinic trigger, which is about 10,000 times faster than for its boronic counterpart. This represents a major improvement compared to boronate scaffolds and recently reported allylic selenide trigger oxidation (k₁ 9.2 m⁻¹⋅s⁻¹) (15).

Stability and Selectivity of Borinic Trigger.

In order to explore the potential applicability of borinic acids as new H₂O₂ sensors under physiological conditions, the stability of diphenylborinic acid as a model was evaluated in various biorelevant media. Hence, this model proved to be stable upon shelf storage, as well as in PBS buffer (pH 7.4) and in Dulbecco’s modified Eagle’s medium (DMEM). Addition of glucose, glutathione, or bovine serum albumin (BSA) to the solution in PBS buffer had almost no impact on the stability of borinic acid on an hour time scale (less than 10% of degradation after 8 h; SI Appendix, Fig. S11). In addition, the presence of a reducing agent like glutathione had no influence on the outcome of the H₂O₂-promoted oxidation reaction since only the borinic acid was oxidized in the presence of 1 equivalent of H₂O₂ (SI Appendix, Fig. S14). This result highlights the superior reactivity of the borinic trigger toward H₂O₂ compared to thiol derivatives. Then, the selectivity of the borinic probe 4 toward several ROS was examined in order to verify that this higher reactivity does not result in a decrease in selectivity toward H₂O₂. For numerous relevant ROS, including superoxide, tert-butyl hydroperoxide, as well as hydroxyl and tert-butyl radical, negligible fluorescent changes were detected (Fig. 3A), indicating that borinic acid 4 exhibits a high selectivity toward H₂O₂, which is similar to the corresponding boronic acid 5 (SI Appendix, Fig. S15) as previously reported (57).

Fig. 3.

(A) Selectivity of probe 4 toward H₂O₂ vs. other ROS. Bars represent fluorescence responses of 5 μM probe 4 to 100 μM various ROS at 5 (white), 15 (light gray), 30 (dark gray), and 60 min (black) in PBS buffer 1× (pH 7.4) at 25 °C. Emission intensity was recorded at λ_em = 450 nm (λ_ex = 360 nm). (B) Fluorescence lifetime images. Fluorescence images of COS7^gp91-p22 cells with 10 μm probe 4 (Top) or 5 (Bottom) prior to (control) and upon exposure to 100 μm exogenous H₂O₂ for 30 s, 45 s, and 165 s. Images were acquired using a DAPI filter (λ_em = 447 nm, λ_ex = 387 nm). (Scale bar, 30 μm.) (C) Time-dependent fluorescence intensity of 10 μm borinic acid 4 (red curve) or 10 μm boronic acid 5 (blue curve) in transfected COS7^gp91/p22 cells. Negative controls were carried out in nontransfected COS7^gp91/p22 cells with the borinic acid 4 (light red curve) and the boronic acid 5 (light blue curve). Data were recorded in PBS buffer 1× (pH 7.4, 0.02% DMSO) at 310 K, at λ_em = 450 nm (λ_ex = 360 nm). a.u., arbitrary units.

Probe Activation in Cells Using Exogenously Added H₂O₂.

We next investigated the ability of the borinic trigger to detect H₂O₂ in a cellular environment. First, we compared the real-time responses of probes 4 and 5 upon exposure to exogenous H₂O₂ in COS7^gp91-p22 model cells. Images clearly displayed higher fluorescence intensity with the borinic acid 4 (Fig. 3 B, Top) compared to the boronic acid 5 (Fig. 3 B, Bottom) within 30 s after addition of H₂O₂. These data suggest that the borinic acid trigger is capable of detecting intracellular H₂O₂ with a very fast response time.

Detection of Endogenously Produced H₂O₂.

We then evaluated the capacity of this borinic trigger to release the fluorophore through activation by endogenously produced H₂O₂. For this purpose, we used COS7^gp91-p22 cells transfected with a chimeric protein, rendering these cells capable of continuously producing the superoxide radical anion through an activated NADPH oxidase (see SI Appendix) (58). Since superoxide spontaneously dismutates into H₂O₂ in aqueous buffer (59), these model cells provide an interesting system to compare the reactivity of both fluorogenic probes. We were pleased to observe that the borinic acid 4 rapidly generated a significant increase of the fluorescence signal within the first 15 min of the experiment in contrast to the boronic acid 5 (Fig. 3C). The ratio of the two slopes indicates that 4-MU is released 10-fold faster with the borinic trigger than with the boronic one, which is in a reasonable agreement with the theoretical value determined by the kinetic analysis (see SI Appendix, section 3.2.3). Moreover, the fluorescence increase was arrested in the presence of either diphenyleneiodonium chloride (DPI), a NADPH oxidase inhibitor, or catalase, an enzyme which catalyzes H₂O₂ dismutation (SI Appendix, Fig. S18). These results clearly support the borinic acid trigger as a valuable tool for a real-time and specific monitoring of endogenously produced H₂O₂ in a physiological environment.

Conclusion

We have described an example of a borinic-based molecular device enabling direct and fast fluorophore release upon H₂O₂-mediated oxidation. Through detailed kinetic studies, we demonstrated that this coumarin-containing borinic sensor responded to H₂O₂ in an unprecedented short time. This improved reactivity is also effective when H₂O₂ is endogenously produced by cells. Optimizations are now underway to control the regioselectivity of this oxidation and get improved sensors in terms of spectral properties and brightness for real-time imaging in a biological context. This work has revealed the superior reactivity of the borinic function toward H₂O₂-promoted oxidation compared to the corresponding boronic counterpart, rendering this trigger a promising tool for H₂O₂-mediated activation. We believe that this trigger could find applications not only for H₂O₂ sensors but also in various fields such as drug release and theranostic applications.

Materials and Methods

Synthesis.

The syntheses of borinic acids are described in SI Appendix. All chemical reagents were purchased from chemical suppliers and used without further purifications. The 2-aminoethanol was distilled before use. Dry dichloromethane was obtained using a Glass Technology GT S100 device. Tetrahydrofuran (THF) and diethyl ether (Et₂O) were distilled from Na/benzophenone prior to use. Reactions were monitored by NMR spectroscopy or analytical thin-layer chromatography (TLC). TLC was performed over Merck 60 F254 with detection by ultraviolet light and/or by charring with sulfuric acid, KMnO₄, or phosphomolybdic acid solutions. Silica gel 60 (40 to 63 μm) was used for flash column chromatography. NMR spectra were recorded on Bruker Advance I 300, 360 and Advance II 600-MHz spectrometers, using the residual nondeuterated solvent as an internal standard. For ¹¹B experiments, chemical shifts are given in parts per million (ppm) relative to BF₃•Et₂O (0 ppm) as an external standard. High-resolution mass spectra were recorded on a MicroTOFq mass spectrometer equipped with an electrospray interface (electrospray ionization positive or negative mode).

H₂O₂-Mediated Oxidation of Borinic Acids by ¹H NMR.

H₂O₂-triggered oxidations of boronic and borinic derivatives were followed by ¹H NMR spectroscopy performed on a Bruker Advance II 600-MHz spectrometer at 20 °C. Boronic or borinic acids reference samples were prepared in a deuterated PBS buffer 1× (pH 7.4) containing 5% deuterated dimethyl sulfoxide (DMSO-d₆) (vol/vol) at 2 mm concentration (V= 500 µL). The H₂O₂-mediated oxidation was performed using 1 equivalent of oxidant. For this purpose, each sample was prepared by mixing 4 mm solutions of boronic or borinic acids in deuterated PBS buffer 1× (pH 7.4) containing 10% DMSO-d₆ (vol/vol) (250 µL) and a 4 mm deuterated aqueous H₂O₂.urea solution (250 µL). The first spectrum was recorded on Bruker Advance II 600-MHz spectrometer after 2 min upon addition of H₂O₂ and the progress of the reaction was then monitored as a function of time.

DFT Calculations.

An energy diagram for the two first steps of the oxidation addition of H₂O₂ to phenylboronic acid or diphenylborinic acid was established using DFT calculations, which were performed with Gaussian 09. The geometries of all the structures were optimized at the M062X/6-311G++(d,p) level, and the vibrational-frequency calculations confirmed that these conformations were local minima or maxima, as expected. Intrinsic reaction coordinate calculations were carried out from each transition state to verify whether the reactant and the product were connected to the same transition state. The solvation effect was taken into account by including a few water molecules that make hydrogen bonds with the heteroatoms present in each compound. All energies are listed in SI Appendix. Our calculations corroborated the higher reactivity of borinic acid compared to boronic acid toward oxidation reaction since the nucleophilic addition is much more favorable.

Kinetic Studies.

Oxidation reactions were monitored by fluorescence using a Photon Technology International QuantaMaster QM-1 spectrofluorimeter equipped with a Peltier cell holder (TLC50; Quantum Northwest). All measurements of protonation constants were performed in Britton–Robinson buffer 1× and kinetic analyses were performed in PBS buffer 1× (pH 7.4). All solutions were prepared using water purified through a Direct-Q 5 (Millipore). pH measurements were performed on a Standard pH meter PHM210 Radiometer Analytical (calibrated with aqueous buffers at pH 4 and 7 or 10) with a Crison 5208 electrode, which is accurate over the 0 to 14 pH range. Ultraviolet-visible absorption spectra were recorded in 1-cm × 1-cm quartz cuvettes (Hellma) on a diode array ultraviolet-visible spectrophotometer (Evolution array; Thermo Scientific) at 293 K. The molar absorption coefficients were extracted while checking the validity of the Beer–Lambert law. The absorbance evolutions as a function of pH were analyzed with the SPECFIT/32 Global Analysis System (version 3.0 for 32-bit Windows systems) to extract the pKa of the investigated compounds.

Selectivity of Probes toward Various ROS.

Fluorescence emission intensity of 5 μM probe in the presence of 100 μm of various ROS in PBS buffer 1× (pH 7.4) was recorded at λ_em = 450 nm (λ_ex = 360 nm) at 25 °C. Experiments were performed in triplicate and the results were averaged. H₂O₂ and superoxide (O₂^•−) were delivered from solid H₂O₂⋅urea and solid KO_2, respectively. tert-butyl hydroperoxide (TBHP) was delivered from 70% aqueous solution. Hydroxyl radical (^•OH) and tert-butoxy radical (^•OtBu) were generated by the Fenton reaction of 2.5 mm Fe²⁺ with 100 µm H₂O₂ or 100 µm TBHP, respectively.

Cell Culture.

COS7^gp91-p22 cells were kindly provided by M. Dinauer, Washington University in St. Louis, St. Louis, MO. These cells stably express two membrane subunits of the NADPH oxidase (gp91 and p22), affording an inactive enzyme. Upon transfection with a chimeric protein named trimera and composed of the domains necessary for the activation of the oxidase, the cells continuously produce superoxide radical anion (see below for details). COS7^gp91-p22 cells were grown in cell DMEM culture medium containing 10% of fetal bovine serum (Gibco) and supplemented with the selecting antibiotics G418 (1.8 mg/mL; Sigma-Aldrich) and Puromycin (1 µg/mL; Sigma-Aldrich) at 37 °C. For cell splitting, cells are detached from the culture surface. We used the commercial Cell Dissociation Buffer Enzyme Free (Gibco) to avoid any damage on the extracellular parts of the membranous gp91 and p22. In order to induce a continuous NADPH oxidase activity, COS7^gp91-p22 cells were transfected with the complementary DNA (cDNA) coding for a chimeric protein, also called trimera, composed of all the domains necessary for the activation of the membranous catalytic core of the NADPH oxidase. The trimera contains the N terminus of p47^phox (amino acids 1 to 286), the N terminus of p67^phox (amino acids 1 to 212), and a full-length Rac1 (amino acids 1 to 192) with the Q61L mutation that mimic the guanosine triphosphate (GTP) bound form. The day before transfection, cells were seeded in 48-well plates and transfected at 85 to 90% of confluency using X-tremeGENE HP DNA transfection reagent following strictly the manufacturer’s protocol (Roche) with the plasmid coding for a fluorescent protein labeled trimera, Citrine-trimera. Twenty-four hours after transfection, the medium was removed and cells were washed twice with Dulbecco’s PBS before the acquisition in 250 µL of PBS with glucose (PBSG).

Cell Imaging: Nontransfected COS7^gp91-p22 Cell with Exogenously Added H₂O₂.

Nontransfected COS7^gp91-p22 cells were plated on an eight-well chamber slide (Ibidi) and loaded with 10 µm of probes for 15 min (295 μL PBSG buffer, pH 7.4, and 0.02% DMSO) prior to imaging the control. Cells were then treated with 5 μL of a 6 mm H₂O₂ solution (final concentration 100 µm). Images were immediately recorded on a Leica inverted SP8 microscope equipped with a ×63 oil immersion objective (numerical aperture 1.4) using a DAPI filter (λ_ex = 387 nm, λ_em = 447 nm). Fiji software was used to analyze all images and to quantify fluorescence.

Detection of Endogenously Produced H₂O₂.

Probes were added at a final concentration of 10 µM to transfected COS7^gp91-p22 cells. Fluorescence was recorded at 450 nm upon excitation at 360 nm with a Synergy H1 microplate reader (BioTek) at 37 °C. Catalase was added at a final concentration of 2.5 μg/mL and DPI at 50 µm.

Data Availability

All study data are included in the article and/or SI Appendix.