¹H-NMR and ¹³C-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 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).

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 H₂SO₄, Ф = 0.54) as a reference (Williams et al., 1983).

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.

Various ROS were also prepared as literature in 0.1 M phosphate buffer, 0.15 M NaCl, pH 7.4, or anhydrous DMSO (Wang et al., 2020; 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).

The enzymatic assay was performed in 0.1 M HEPES buffer, pH 7.4. Please refer to our previous publications (Wang et al., 2020). 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 O₂ ^•− in the system and observed a relatively low fluorescence increase for probes R2 and R3, but no fluorescence change was observed for probe R1.

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 O₂ ^•− probes (series 1, 2) have stable fluorescence intensity in various buffers at differing pH. For the non-redox based O₂ ^•− 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 O₂ ^•− probes are rather stable at differing pHs. However, in basic conditions, some non-redox based O₂ ^•− probes might turn the fluorescence on by OH- group via nucleophilic substitution.

Now that we have a primary understanding of the probes, we would like to explore the reactivity between the probes and various ROS [e.g., tert-butyl hydroperoxide(TBHP), H₂O₂, •OH, ¹O₂, ClO⁻, O₂ ^•−] (Table 2, ESI). Probes were incubated with various oxidizing agents (TBHP, H₂O₂, •OH, ¹O₂, ClO⁻) in 0.1 M phosphate buffer with 0.15 M NaCl (pH 7.4) at 37°C for 5 min (Doura et al., 2012; Xing et al., 2016; Zhan et al., 2017). Because O₂ ^•− cannot exist in an aqueous buffer, the reaction between the probes and O₂ ^•− was carried out in anhydrous DMSO at 37°C for 5 min (Wang et al., 2020). 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 (Table 2) showed that under our conditions: (1) R2 was the most reactive probe in series 1, which can react with ¹O₂, ClO⁻, and O₂ ^•−. The other two probes, R1 and R3, can only react with ClO⁻ and O₂ ^•−. (2) Interestingly, chromone-derived probes (series 2) showed much lower reactivity when compared to series 1, and we unfortunately only saw the reactivity between Rb and O₂ ^•−. (3) However, we observed that the probes from series 3 and 4 were much more reactive toward various ROS than series 1 and 2. The boronate probes (N1 and N2) were successful in sensing TBHP, H₂O₂, ¹O₂, •OH, ClO⁻, and O₂ ^•−. The sulfonyl ester (N3 and N4) successfully reacted with TBHP, H₂O₂, and O₂ ^•−, but not with ¹O₂, •OH, and ClO⁻.

In this section, in order to rank the reactivity hierarchy toward various ROS among probes, we used excessive ROS to react with each probe. Generally speaking, redox-based O₂ ^•− probes exhibited a strong reactivity hierarchy which was well correlated to the oxidizing state of the ROS. The reactivity order of redox-based O₂ ^•− probes is R2 > R3 > R1 > Rb > Ra, Rc. 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.

Next, we explored the selectivity profiles of redox-based O₂ ^•− probes (series 1 and 2). The selectivity of the probe was not only determined by the reactivity but also affected by the transformation rate. The transformation rate can be manipulated by adjusting ROS concentration, reaction temperature, and others. For example, in this study, we found that chromone-derived probe Rb can only react with O₂ ^•− but not with other ROS although we further increased the ROS concentration, reaction time, and temperature. Thus, probe Rb was highly specific to O₂ ^•−. 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 ClO⁻/O₂ ^•−. However, if we set the incubation time for less than 5 min at 37°C, probe R3 can only turn the fluorescence on by O₂ ^•− but not by ClO⁻ (ESI S1.5). Thus, probe R3 can selectively detect O₂ ^•− under certain conditions.

Lastly, we explored the applicability of our redox-based O₂ ^•− probes in a biological system. The O₂ ^•− 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 O₂ ^•− probes (series 1 and 2). We further increased the concentration of the XO enzymes (0.6 U/ml) to produce more concentrated O₂ ^•− 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 studies have shown that O₂ ^•− reacts violently with H₂O (Wang et al., 2020; Tampieri et al., 2019). Thus, in the XO/HPX system, the majority of the produced O₂ ^•− will react with excessive H₂O before they can reach the probes. The reaction between O₂ ^•− and H₂O produces H₂O₂ and OH- (Tampieri et al., 2019), both of which remain relatively stable in the aqueous solution. This will eventually cause the solution contain high concentrations of H₂O₂ and OH-, but rather low concentrations of freshly produced O₂ ^•−. Interestingly, we proved that non–redox-based O₂ ^•− probes can react with almost all oxidizing levels of ROS (Table 2). Thus, both newly produced O₂ ^•− and the degraded low oxidizing ROS can react with non-redox O₂ ^•− probes to turn on the fluorescence. Notably, we also showed that probes of series 4 (the sulfonyl ester) but not series 3 (the boronate) can react with OH- to turn on the fluorescence (Figure 3) (Tampieri et al., 2019), this was accordingly reflected as probes of series 4 were more reactive than those of series 3 in the XO/HPX system (Figure 4A). By contrast, in the XO/HPX system, only O₂ ^•− but not other low oxidizing ROS can turn on the fluorescence of redox-based O₂ ^•− probes (series 1 and 2). Thus, redox-based O₂ ^•− probes reacted more slowly toward O₂ ^•− in the XO/HPX system (Figure 4). When we further increased the concentration of the XO enzyme, a small portion of R2 and R3 slowly turned on the fluorescence. Taken together, we found that in the XO/HPX system, non–redox-based O₂ ^•− probes were more active than redox-based ones. Unfortunately, most of the non–redox-based O₂ ^•− probes unselectively turned on fluorescence by the low oxidizing level of ROS. By contrast, redox-based O₂ ^•− probes R2 and R3 can only be slowly oxidized by O₂ ^•− to turn on fluorescence, and the fluorescence change was directly caused by O₂ ^•− but not other low oxidizing ROS. In summary, we showed the advantage of redox-based O₂ ^•− probes in the detection of O₂ ^•− in a biological system.