Insights Into the Detection Selectivity of Redox and Non-redox Based Probes for the Superoxide Anion Using Coumarin and Chromone as the Fluorophores

In this study, we evaluated the applicability of various superoxide anion sensors which were designed based on either redox or non-redox mechanisms. Firstly, both redox- and non–redox-based superoxide anion probes were designed and synthesized using either coumarin or chromone as the fluorophores, and the photophysical properties of these probes were measured. Subsequently, the sensing preference of both types of probes toward various reactive oxygen species (ROS) was evaluated. We found that non–redox-based O2 •− probes exhibited broad sensing ability toward various ROS. By contrast, redox based O2 •− probes showed a clear reactivity hierarchy which was well correlated to the oxidizing strength of the ROS. Lastly, the detection selectivity of redox-based O2 •− recognizing probes was also observed when balancing various factors, such as reactant ROS concentrations, temperature, and changing reaction transformation rates. Herein, we concluded the selectivity advantage of redox-based O2 •− probes.


INTRODUCTION
Reactive oxygen species (ROS) are a group of important oxidizing agents within biological systems, which play a key role in the regulation of homeostasis (Juan et al., 2021;Yang et al., 2019;D'Autréaux and Toledano, 2007). The concentrations of ROS generally remain balanced, and any interruption of this balance results in a cascade of unwanted biological events (Yang et al., 2019;Juan et al., 2021). Therefore, in the clinical setting, it is of utmost importance to accurately detect the concentrations of these ROS, as well as probe the underlying biological mechanism of this dysregulation.
The oxygen of ROS is in a highly oxidizing state, which results in all ROS being highly reactive toward a range of biological substances (Jiao et al., 2018). Thus, ROS are usually found in low concentrations in tissues under regular conditions, and therefore traditionally, it has been difficult to accurately quantify the concentration of ROS. The mitochondria and NADPH oxidase produce the major ROS, superoxide anion, and H 2 O 2 in the cells (Dröge, 2002;Woolley et al., 2013). Usually, under normal conditions, the concentration of superoxide anion and H 2 O 2 is estimated to be about 10 −10 and 5 × 10 −9 M, respectively (Dröge, 2002;Turrens, 2003;Woolley et al., 2013). However, the concentration of these at the cellular level can change in a wide range under stimulated conditions. Moreover, the oxidizing state of ROS ranges from 2 to 0, with degradation from high-oxidizing ROS forming additional low-oxidizing ROS (Jiao et al., 2018). Therefore, several ROS might coexist within a single system, and it will be important to distinguish each ROS during detection (Jiao et al., 2018). Among the ROS, the oxygen of the superoxide anion is in the highest oxidation state, and the superoxide anion is the precursor to several other ROS (Cadenas and Davies, 2000;Jiao et al., 2018). Thus, in regard to superoxide anion detection, selectivity would be a key parameter to be considered.
To this end, various methods have been developed for ROS detection, including the fluorescent dye method, nanoprobe technology, electrochemical biosensors, electron spin resonance method, genetic encoded ROS reporter, and others (Woolley et al., 2013;Mamone et al., 2016) Among them, fluorescent techniques have been widely used in sensing and detecting these biologically important species under certain biological conditions (Fuloria et al., 2021;Duanghathaipornsuk et al., 2021;Wu et al., 2019;Cheng et al., 2019). To date, a range of fluorescent sensors have been developed for various ROS (Chen et al., 2016;Yang et al., 2020). For O 2 •− sensors, based on design principles, these can be classified into two categories: redox and non-redox mechanisms-based O 2 •− fluorescent probes ( Figure 1) (Jiao et al., 2018;Xiao et al., 2020). The redoxbased fluorescent probes have been designed based on the oxidizing ability of O 2 •− (Tang et al., 2004;Zhang et al., 2013;Wang et al., 2020), and the non-redox-based fluorescent probes were designed on the nucleophilicity or other inherent reactivity of O 2 •− (Maeda et al., 2005;Xu et al., 2007;Zielonka et al., 2010).
Although these descriptors exist, there has been no systematic study to understand their key differences in relation to their applicability, particularly, their sensing selectivity.

MATERIALS AND METHODS
Chemical Synthesis General 1 H-NMR and 13 C-NMR spectra were recorded with a Varian Mercury 400 or 500 spectrometer using tetramethylsilane as the internal standard in methanol-d 4 , DMSO-d 6 , or chloroform-d. High-resolution mass spectrometry (HRMS) data were measured on a Thermo Exactive Orbitrap Plus spectrometer. Liquid chromatography-mass spectrometry (LC-MS) was conducted on an Agilent 1100 series HPLC and an Agilent LC/MSD FIGURE 1 | The redox-and non-redox-based superoxide anion probes. The reaction mechanisms of the selected probes were also proposed to help to understand this study.
TOF. All of the solvents and chemicals were purchased from commercial sources: Sigma-Aldrich Chemical Co., Beijing Ou-he Reagents Co., Beijing Shiji-Aoke Biotechnology Co., and Shanghai Jingke Chemistry Technology Co. with a purity of more than 95% (LC-MS). All chemicals and solvents used were of reagent grade without further purification or drying before use. All the reactions were monitored by thin-layer chromatography (TLC) under a UV lamp at 254 nm. Column chromatography separations were performed using silica gel (200-300 mesh).

Ra-Rc
The compound 4H-chromen-4-one derivative 4 or 5 or 6 (0.41 mmol) was added to dry THF (15 ml), and then the mixture was stirred and cooled to −20°C. A solution of LiAlH 4 (0.45 ml, 1.0 M solution in THF) diluted with 5 ml dry THF was added dropwise to the above with stirring at −20°C. The mixture was stirred for 2 h at −20°C. The reaction was analyzed by TLC for completion. Then the reaction was quenched with 2 M NH 4 Cl aqueous solution (20 ml), and the solvent was removed in vacuo. The mixture was extracted with ethyl acetate (3 × 20 ml), and the combined organic layers were washed with saturated NaCl aqueous solution (2 × 20 ml). The organic layer was then dried (Na 2 SO 4 ), filtered, and the solvent was removed in vacuo. The products Ra-Rc were purified by silica gel column chromatography.

Measurement of Photophysical Properties of the Probes
The photophysical properties of all compounds were measured. The measurement of the photophysical properties of various compounds was carried out as we described before (Miao et al., 2015;Wen et al., 2018;Yan et al., 2018). All compounds were dissolved in 0.1 M Tris-HCl buffer, pH 8.0 at the concentration of 10 μM. SHIMADU UV-2700, UV-visible spectrophotometer was used to measure UV-visible spectra. HITACHI F-7000 fluorescence spectrophotometer was used to measure excitation and emission spectra. For fluorescence quantum yield calculation, compounds were dissolved in 0.1 M Tris-HCl buffer (pH 8.0) at the concentration of 0.5 μg/ml or less using quinine sulfate (0.5 μg/L in 0.1 M H 2 SO 4 , V 0.54) as a reference (Williams et al., 1983).

Detection of the pH Stability of the Probes
We also investigated the stability of these probes under various pHs. 0.2 M phosphate buffers with desired pHs (pH 3, pH 7, and pH13) were prepared. A 10 μM solution of each probe at different pHs (pH 3, pH 7, and pH13) was prepared, and their fluorescence was scanned (E x , E m ) to see the change.

Determination of the Reactivity Between Fluorescent Probes and Various ROS
Various ROS were also prepared as literature in 0.1 M phosphate buffer, 0.15 M NaCl, pH 7.4, or anhydrous DMSO Chen et al., 2021). Each probe was dissolved in these ROS solutions at the final concentration of 10 μM. After being incubated at 37°C for 5 min, the mixture was scanned for the preferred E x of the desired fluorophore to check if the desired fluorophore was formed. In order to detect the fluorescence change, the E x was set as 340 nm, and the E m was measured between 380 and 600 nm. In addition, the final products of the reactions were also analyzed by LC-MS for confirmation (ESI S2).

Fluorescence Response of Various Probes Toward XO/HPX System
The enzymatic assay was performed in 0.1 M HEPES buffer, pH 7.4. Please refer to our previous publications . Initially, we began this study at a relatively low concentration of 0.25 U/ml XO enzyme and observed a real-time fluorescence change for non-redox-based probes N1, N2, N3, and N4. However, under these conditions, we did not observe a fluorescence change for redox-based probes R1, R2, and R3. We further increased the concentration of the XO enzymes (0.6 U/ml) to produce more O 2 •− in the system and observed a relatively low fluorescence increase for probes R2 and R3, but no fluorescence change was observed for probe R1.

Series 2
The synthetic routes of compounds Ra, Rb, and Rc are outlined in Scheme 2. Three chromones 4, 5, and 6 were prepared as previously described (Chen et al., 2021). For the reduction, we used LiAlH 4 to reduce the chromones to yield the desired chromanones under a low temperature.

Series 3
The synthetic strategy towards N1 and N2 is shown in Scheme 3.
Compounds 10 and 11 are commercially available. The compounds N1 and N2 were obtained following a previously reported Miyaura borylation protocol (Jana et al., 2014).

Series 4
The synthesis of N3 and N4 is shown in Scheme 4 and was achieved via a simple one-step procedure from the corresponding phenol and sulfonyl chloride. Compound 1 (100 mg, 0.568 mmol) or compound 9 (100 mg, 0.427 mmol) reacted with 2,4-dinitrobenzene-1-sulfonyl chloride (151 mg, 0.568 mmol) in anhydrous DCM (10 ml) and DIPEA (1.704 mmol, 3 eq). The mixture was stirred at room temperature. The reaction was analyzed by TLC for completion. The yields of N3 and N4 were 45.6% or 55.2%.
We, therefore, set out to conduct this study. First, we worked to synthesize ten O 2 •sensors of both redox-and non-redoxbased probes (Figure 2 and Table 1), and 7-donor coumarin and 7-donor chromone were chosen as the fluorophoric portion of the molecule. Coumarin is a known fluorophore and has been frequently used in various studies to a high level of success (Cao et al., 2019). Additionally, chromone derivatives have also been found to exhibit a range of interesting fluorescent properties (Miao et al., 2015;Chen et al., 2021). Series 1 (R1, R2, R3) and series 2 (Ra, Rb, Rc) were designed as redox Frontiers in Chemistry | www.frontiersin.org November 2021 | Volume 9 | Article 753621 mechanism-based probes. For these two series, the broken aromatization of either coumarin or chromone quenched the fluorescence. For the detection, the oxidation of them by certain ROS was expected to recover the aromatization of the coumarin or chromone, and this would in accordance turn on the fluorescence (Doura et al., 2012;Wang et al., 2020). Series 3 (N1, N2) and series 4 (N3, N4) are the probes that were designed with a non-redox mechanism, and the 7-donor groups were modified with the boronate group (series 3) or sulfonyl ester group (series 4). The modification of the 7-donor group of the coumarin and chromone broke the electron transfer between donor and p-conjugatedacceptor, and this dramatically quenched the fluorescence of probes. For the detection, certain ROS will react with the probes to either replace or remove the modified group, and finally turn on the fluorescence (Maeda et al., 2005;Castro-Godoy et al., 2019).

Measurement of Photophysical Properties and pH Stability of the Probes
Next, we measured the photophysical properties of all probes ( Table 1). As expected, the majority of the fluorophores (1-9) exhibited moderate to high quantum yields, while the designed probes (R1-N4) had relatively low quantum yields (0.01-0.11). Moreover, given that the fluorescence intensity of a compound corresponds to the quantum yield and the molar extinction coefficient, we calculated the turn-on ratio for each matched pair of probe and fluorophore. As expected, the majority of the synthesized probes have a useful fluorescence turn-on ratio ranging from 10 to several hundred, and are therefore perfectly suited to being used as fluorescence turn-on probes.
We further investigated the stability of these probes under various pHs (Figure 3). A 10 μM solution of each probe at different pHs (pH 3, pH 7 and pH13) was prepared and their fluorescence was measured. We can see that redox based O 2 •− probes (series 1, 2) have stable fluorescence intensity in various buffers at differing pH. For the non-redox based O 2 •− probes, we observed a stable but low fluorescence intensity of series 3 in various buffers (pH 3, pH 7, and pH13). However, we found that probes of series 4 (N3 and N4) exhibited strong fluorescence intensity in a basic buffer (pH13), which was 30-40 times stronger than that of other buffers (pH3 and pH7). This suggests that probes N3 and N4 degraded under basic conditions to turn on the fluorescence. Since N3 and N4 were designed to react with ROS to remove the sulfonyl ester group via nucleophilic substitution, the OH-group in a basic buffer can also react with them to eliminate the modification on 7-hydroxyl (Tampieri et al., 2019). In summary, redox based O 2 •− probes are rather stable at differing pHs. However, in basic conditions, some non-redox based O 2 •− probes might turn the fluorescence on by OH-group via nucleophilic substitution.

Determination of the Reactivity Between Fluorescent Probes and Various ROS
Now that we have a primary understanding of the probes, we would like to explore the reactivity between the probes and FIGURE 2 | The redox-and non-redox-based superoxide anion probes designed and synthesized in this study. Coumarin and chromone were selected as the core fluorophores.  (Doura et al., 2012;Xing et al., 2016;Zhan et al., 2017). Because O 2 •− cannot exist in an aqueous buffer, the reaction between the probes and O 2 •− was carried out in anhydrous DMSO at 37°C for 5 min . The concentrations of the probes were set as 10 μM, but the amounts of various ROS were excessive to promote the reaction (ESI). It was agreed that if the desired fluorophore was detected, regardless of the reaction transformation rate, the reactivity between the probe and ROS would be deemed successful in the study. The results ( Interestingly, although similar reaction mechanisms (the aromatization) were used for series 1 and series 2, series 1 (coumarin derivatives) was more active than series 2 (chromone derivatives) toward various ROS. This indicates that both the reactive group and the structure of the chosen fluorophore affected the reactivity of the probes. This provides the opportunity to further optimize the reactivity of these probes via structural modification. Unfortunately, the non-redox-based probes reacted with almost all ROS without any clear correlation to the oxidizing ability of the ROS. This broad ROS reactivity obviously limits the application of these types of probes. . On the other hand, when we balanced the conditions of ROS concentration, and reaction temperature, the transformation rate between the probe and certain ROS was subsequently changed. If we control the transformation rate to allow the number of reaction products to be below or above the detection line, we can achieve detection selectivity. In this study, we proved the reactivity between R3 and

Exploration of Detection Selectivity and
. However, if we set the incubation time for less than 5 min at 37°C, probe R3 can only turn the fluorescence on by O 2 •− but not by ClO − (ESI S1.5). Thus, probe R3 can selectively detect O 2 •− under certain conditions.
Lastly, we explored the applicability of our redox-based O 2 •− probes in a biological system. The O 2 •− was produced via the more biologically relevant xanthine oxidase (XO)/hypoxanthine (HPX) system ( Figure 4). We began this study at a relatively low concentration of XO enzyme (0.25 U/ml) and 1 mM HPX. We observed a real-time fluorescence change for non-redox-based probes N1, N2, N3, and N4. The sulfonyl ester series (N3, N4) was particularly active under the XO/HPX system, and the fluorescence quickly reached a peak level after several minutes. While the boronate series (N1, N2) was successful, it was much slower than the sulfonyl ester series ( Figure 4A). However, under this enzyme condition, we did not observe a fluorescence change for all redox-based O 2 •− probes (series 1 and 2). We further increased the concentration of the XO enzymes (0.6 U/ml) to produce more concentrated O 2 •− in the system. Then, we observed a slow fluorescence increase for probes R2 and R3 ( Figure 4B), but no fluorescence change was observed for probes R1, Ra, Rb, and Rc.
Previous . Interestingly, we proved that non-redoxbased O 2 •− probes can react with almost all oxidizing levels of ROS (Table 2). Thus, both newly produced O 2 •− and the degraded low