Selenium-Based Catalytic Scavengers for Concurrent Scavenging of H₂S and Reactive Oxygen Species

Abstract

Hydrogen sulfide (H₂S) is an endogenous gasotransmitter that plays important roles in redox signaling. H₂S overproduction has been linked to a variety of disease states and therefore, H₂S-depleting agents, such as scavengers, are needed to understand the significance of H₂S-based therapy. It is known that elevated H₂S can induce oxidative stress with elevated reactive oxygen species (ROS) formation, such as in H₂S acute intoxication. We explored the possibility of developing catalytic scavengers to simultaneously remove H₂S and ROS. Herein, we studied a series of selenium-based molecules as catalytic H₂S/H₂O₂ scavengers. Inspired by the high reactivity of selenoxide compounds towards H₂S, 14 diselenide/monoselenide compounds were tested. Several promising candidates such as S6 were identified. Their activities in buffers, as well as in plasma- and cell lysate-containing solutions were evaluated. We also studied the reaction mechanism of this scavenging process. Finally, the combination of the diselenide catalyst and photosensitizers was used to achieve light-induced H₂S removal. These Se-based scavengers can be useful tools for understanding H₂S/ROS regulations.

Introduction

Hydrogen sulfide (H₂S) is an important signaling molecule that plays crucial roles in various physiological and pathological processes.^[1] Endogenous H₂S production is attributed to enzymes including cystathionine β-synthase (CBS), cystathionine γ-lyase (CSE), 3-mercaptopyruvate sulfur-transferase (3MST), and cysteinyl-tRNA synthetase (CARS). Many studies have revealed that the dysregulation of H₂S is associated with diseases such as hypertension, diabetes, Alzheimer’s, and cancer.^[2] Like other gasotransmitters, H₂S’s biological functions are closely linked to its location and concentration. While a low concentration of H₂S can be beneficial (e.g. antioxidant or anti-inflammatory), a high concentration of H₂S is toxic. Therefore, significant efforts have been made in recent years to develop chemical tools that can regulate intracellular H₂S levels. These tools not only benefit fundamental research, but also show promising therapeutic applications. In this research field, most of the previous work has focused on H₂S-releasing compounds (e.g. donors),^[3] while the development of H₂S suppressors (e.g. chemicals that downregulate H₂S levels in biological systems) has received much less attention.^[4] For example, H₂S overproduction is believed to be critically linked to the development of liver and colon cancer.^[5] However, methods that downregulate H₂S in desired tissues or cells are still lacking. Several inhibitors for H₂S-producing enzymes have been reported, but their potency and specificity are still far from ideal.^[4]

To address the need for H₂S suppressors, our lab has initiated a program to search for H₂S scavengers. We envisioned that H₂S scavengers shared the same criteria as H₂S fluorescent sensors, such as high reactivity and selectivity for H₂S. Inspired by the reported H₂S sensors, we discovered sulfonyl azides were potent H₂S scavengers (Scheme 1).^[6] Sulfonyl azides also showed promising activity as antidotes for H₂S-induced acute poisoning.^[7] In another work by Yi et al., NBD amines were demonstrated to be effective H₂S scavengers, and these compounds were similarly based on known NBD-derived H₂S sensors.^[8] In addition, Szabo et al. reported that the metabolite of disulfiram (CuDDC) acted as both a CBS inhibitor and H₂S scavenger, which was able to suppress intracellular H₂S levels in various cell types.^[9] Several metal oxide and metal-organic framework (MOF)-based nanomaterials were also explored to directly eliminate endogenous H₂S.^[10] While these small molecule H₂S scavengers show interesting activities, most of them rely on stoichiometric reactions with H₂S to remove H₂S. Ideally, we would like to develop catalytic H₂S scavengers as such compounds would be more efficient and can significantly reduce potential cytotoxicity. We expect that this is a valid idea as H₂S and oxidants co-exist under certain circumstances. For instance, in H₂S-induced acute intoxication, it is known that significant amounts of reactive oxygen species (ROS), such as H₂O₂, are produced and represent a major mechanism of H₂S poisoning.^[11] It is also known that the direct reaction between H₂S and H₂O₂ is slow, especially under physiological concentrations.^[12] Therefore, it is possible to find small molecules that can catalyze the reaction between H₂S and H₂O₂, and remove both species from the systems. Such compounds can be considered catalytic scavengers for H₂S. Herein, we report the discovery of selenium-based compounds as such scavengers.

Scheme 1.

The idea of selenium-based catalytic scavengers for H₂S/H₂O₂.

Results and Discussion

Selenium-containing compounds attract growing interest because of the well-known glutathione peroxidase (GPx) activity and their wide utility as antioxidants.^[13] Notably mimicking the GPx activity center are diaryl selenides (Ar-Se-Ar) that have been used to design fluorescent probes for monitoring the intracellular oxidative stress/H₂S redox cycle.^[14] This design is based on the fact that diaryl selenides can be oxidized by ROS to form the corresponding selenoxide, and H₂S can then reduce selenoxide back to selenide. Inspired by this redox chemistry, we envisioned that selenoxide (SeO)-containing compounds might serve as the proposed catalytic scavengers to remove H₂S in the presence of ROS like H₂O₂. With this idea in mind, three SeO-containing compounds (Figure 1) were selected to test their H₂S-scavenging ability. Diphenyl selenoxide (S1) was prepared via a one-step oxidation from diphenyl selenide. Sodium selenite (S2) is commonly used as a selenium supplement in animal diets and has shown radioprotective potential in clinical trials.^[15] Benzeneseleninic acid (S3) has exhibited potent Phase II enzyme-inducing activity and was regarded as a good candidate for chemoprotection against cancer.^[16] S1 was first used to optimize the protocol for H₂S-scavenging measurements. Briefly, a solution of H₂S (200 μM, using Na₂S as the equivalent) in PBS buffer was freshly prepared and sealed. A Unisense H₂S microsensor was used to monitor the H₂S concentration in real-time. In the absence of the scavenger, the concentration of H₂S was found to be quite stable, and no obvious change was noticed for at least 30 min (Figure S1, the supporting information). However, when one equivalent of S1 was added, the H₂S concentration dropped sharply (Figure 1a), and >95% H₂S was removed within 2 min, indicating the high efficiency of S1 as a scavenger. As a control, diphenyl sulfoxide was also tested and it did not reduce the level of H₂S, which indicated the unique reactivity of Se. We also carried out a model reaction between S1 and Na₂S in a 1:1 mixture of CH₃CN/PBS to verify the mechanism (Scheme S1, the supporting information). As expected, S1 was converted to its reduced form (diphenyl selenide), and the formation of elemental sulfur (S₈) was confirmed. We next tested S2 and S3 under the same conditions. Methanesulfonyl azide (SS20), an established H₂S scavenger,^[6] was also tested for comparison. Both S2 and S3 showed comparable response to S1, and H₂S was completely removed within minutes (Figure 1b/c). The activity of S1-S3 appeared to be comparable to that of SS20 (Figure 1d).

Figure 1.

H₂S-scavenging profiles of compounds S1-S3 and SS20 (200 μM) in PBS buffers (50 mM, pH 7.4). (a) black line: S1, red line: diphenyl sulfoxide, (b) S2, (c) S3, (d) SS20.

To further understand the efficiency of these SeO-based scavengers, their initial rates and T_1/2 (the time required for 50% decrease of H₂S) were determined and listed in Table 1. All the scavengers exhibited high (> 700 μM min⁻¹) reaction rates and 50% H₂S was removed in less than 20 s. The data from S1-S3 appeared to be more potent that the data of SS20, indicating that selenoxides are more promising for H₂S scavenging than sulfonyl azides. Among these three Se-based compounds, S3 was the most potent, and its initial rate was ~3 times faster than SS20.

Table 1.

Reaction parameters of SeO-containing compounds with H₂S.

Scavenger	v0 (μM min−1)[a]	Vrel	T1/2 (min)
SS20	512 ± 6	1	0.29
S1	950 ± 60	1.86	0.15
S2	770 ± 21	1.50	0.21
S3	1750 ± 45	3.42	0.12

^[a]The initial rates were reported as averages ± standard deviations from three independent runs.

Since SS20 has shown promising antidotal effects in H₂S poisoning models, we wondered if S1-S3 could have similar activity. We first tested their cytotoxicity in primary cortical neurons. As shown in Figure S13 (the supporting information), these compounds appeared nontoxic in concentrations up to 400 μM. We next compared the effects of S1-S3 with SS20 in an H₂S ongoing exposure model (Figure 2).^[7] Briefly, mice were exposed to a lethal concentration of H₂S (790 ppm in 21% oxygen in a breathing chamber). The concentrations of H₂S and O₂ were continuously monitored by a gas monitor. Fresh gas flow was set at 4 l/min throughout the experiment. After 10 min of H₂S inhalation, mice were randomly assigned to receive an administration of scavengers or vehicle (DMSO) in a volume of 2 ml/kg intraperitoneally. The doses are indicated in Figure 2. Mice were then returned to the chamber and re-exposed to 790 ppm of H₂S gas for another 30 min. After a total of 40-min exposure to H₂S, fresh gas was switched to 79% N₂ and 21% O₂, and mice were allowed to breathe 21% O₂ for another 20 min. Subsequently, the mice were returned to cages in room air. Mice with a respiratory interval 15 sec or longer were deemed to be dead. Mice were monitored for up to 24 h to record the length of survival time during and after H₂S poisoning. The primary outcome of the ongoing exposure model was the survival rate at 24 h after H₂S exposure. As shown in Figure 2, all these selenium-based scavengers showed antidotal effects. S3 was found to be the most effective, with a 100% survival rate (similar to that of SS20).

Figure 2.

Scavengers rescued mice from ongoing H₂S breathing (ongoing exposure model). Scavengers or vehicle was administered to awake mice intraperitoneally at 10 min after breathing 790 ppm of H₂S followed by another 30 min of continuous H₂S inhalation. Percent survival during the first 24 h after H₂S inhalation are presented. *, ** or **** P<0.05, 0.01 or 0.001 vs. vehicle by the log-rank test.

The reactions between these SeO-based scavengers and H₂S should produce the corresponding reduced Se-based compounds (for instance, diphenyl selenide from S1). These reduced forms may rapidly react with ROS to reform the scavengers, so it might create a catalytic cycle to remove H₂S in the presence of ROS. With this consideration in mind, we then examined the catalytic potential of S1-S3. H₂O₂ was selected as the ROS model because it is the most prevalent ROS in biology. It is also known that the direct reaction between H₂S and H₂O₂ is quite slow, especially under physiologically relevant concentrations.^[12] Our study was as follows (Figure 3): 20 μM of the SeO-scavenger was added to 200 μM H₂S solution in PBS buffer and the reaction was monitored for 10 min. Then, H₂O₂ under the same concentration as the remaining H₂S was added. As shown in Figure 3, S1 and S2 (at 20 μM) could remove 1 and 3 equivalents of H₂S, respectively. When H₂O₂ was added, no obvious increased H₂S consumption was noted in these two systems. These results suggested that the H₂S-reduction products of S1 and S2 were inactive toward H₂O₂. This was confirmed by a separate experiment with a 1:1 ratio of diphenyl selenide and H₂O₂, and no reaction was observed (Figure S2, in Supporting Information). With regard to S3, 2 equivalents of H₂S (about 40 μM) was consumed initially. This was followed by a slight and continuous H₂S decrease over 10 min. When H₂O₂ was added, the H₂S scavenging process was effectively promoted and more than half of the remaining H₂S was eliminated in 15 min.

Figure 3.

SeO-based scavengers (20 μM) catalyzed scavenging of H₂S (200 μM) in PBS buffers (50 mM, pH 7.4).

Our results thus far suggested that benzeneseleninic acid S3 might be a catalytic scavenger. Its reaction with H₂S was expected to produce benzeneselenol (PhSeH), which should eventually form diphenyl diselenide (PhSeSePh) under aerobic conditions. It has been reported that the Se-Se bond of diselenides could be cleaved and oxidized to seleninic acid by H₂O₂.^[17] Therefore, we suspected that diselenide compounds might be catalytic scavengers. To test this hypothesis, we first evaluated diphenyl diselenide (S4) for H₂S scavenging using the following protocol: a solution of H₂S (200 μM) in PBS buffer was mixed with H₂O₂ (200 μM) for 2 min, after which S4 (200 μM) was added. The concentration of H₂S was continuously monitored. Control experiments (H₂S only, H₂S+H₂O₂, H₂S+S4) were also carried out for comparison. As shown in Figure 4, H₂S concentration decreased slowly when it was treated with only S4 or H₂O₂, confirming the slow reactions of H₂S toward these two species. However, in the presence of both H₂O₂ and S4, the rate of H₂S scavenging was dramatically enhanced, indicating that S4 was an effective scavenger in this situation.

Figure 4.

H₂S-scavenging profiles of S4 in the presence and absence of H₂O₂ H₂S (200 μM), H₂O₂ (200 μM), S4 (200 μM), in PBS buffers (50 mM, pH 7.4).

Our results indicated that seleninic acids and diselenides were promising as catalytic scavengers. However, seleninic acids only showed high reactivity towards H₂S while their reactivity to H₂O₂ appeared low. Diselenides are the major series of GPx mimetics and are highly sensitive to both oxidative and reductive stimuli. Therefore, we envisioned that diselenides may have greater potential to achieve the catalytic scavenging of H₂S and H₂O₂. We tested 10 diselenide compounds (S4-S13, Figure 5) in H₂S scavenging (in the presence of H₂O₂) to understand the structural effects on activity. We also tested 4 monoselenide compounds (S14-S17) as these four compounds are representative Gpx-like antioxidants. Among them, S14 and S15 are important naturally occurring selenium-containing amino acids which exert chemopreventive and anti-cancer activity in human clinical trials.^[18] The selenium-based sugar mimic S16 has been reported as a potent water-soluble ROS scavenger that could afford effective protection against oxidative stress.^[19] S17 (ebselen), one of the best studied Gpx mimics, was also included. These compounds (at 200 μM) were added into the solution of H₂S and H₂O₂ (both at 200 μM). Their initial rates and T_1/2 were determined and listed in Table 2.

Figure 5.

Structures of diselenides and monoselenides tested.

Table 2.

Diselenides/monoselenides promoted H₂S/H₂O₂ scavenging profiles.

Scavenger	v0 (μM min−1)[a]	Vrel	T1/2 (min)
S4	46 ± 5	1	5.77
S5	501 ± 7	10.89	0.38
S6	270 ± 30	5.87	0.91
S7	8 ± 2	0.17	>20
S8	122 ± 15	2.65	2.28
S9	6.0 ± 0.4	0.13	>20
S10	22 ±2	0.48	7.18
S11	11 ±3	0.24	13.47
S12	422 ± 12	9.17	0.58
S13	7 ± 2	0.15	16.80
S14	38 ±1	0.83	3.76
S15	5 ± 1	0.11	>20
S16	N/A[b]	N/A[b]	>20
S17	480 ± 70	10.43	N/A[c]

^[a]The initial rates were corrected for the uncatalyzed background reaction and reported as averages ± standard deviations from three independent runs.

^[b]The activity of S16 was too low to detect.

^[c]Ebselen could not remove more than half of H₂S due to its poor solubility.

The results shown in Table 2 suggested that the structures of these selenium compounds could significantly impact their scavenging ability. In particular, amines (-NH₂) played an important role. S4 only showed moderate activity. However, when -NH₂ was introduced to the ortho-position, the resulting compound S5 was found to be ~10-folds more potent. Similarly, 3,3′-diselanediyldipropanoic acid (S11) showed weak activity while its amine-substituted analog S12 showed a 30-fold increase in its initial rate. When the -NH₂ group was protected (S7, S9) or replaced by -OH (S10) or -CO₂H (S11), the compounds’ scavenging ability also dropped. Additionally, the location of the -NH₂ relative to the selenium center seems to be critical. When -NH₂ was two carbons away from the selenium atom, the activity was usually higher (S6 vs S8, S14 vs S15). The effects of the -NH₂ group can be explained by the possible intramolecular H-bond between the -NH₂ group and the Se atom, which would facilitate the formation of selenolate during the catalytic cycle. Similar effects were noted in previous studies of diselenide oxidation with H₂O₂.^[20] Interestingly, the cyclic diselenide (S13) and the sugar-based monoselenide (S16) showed relatively low activity. As for ebselen (S17), it showed high reactivity toward H₂S and could rapidly remove H₂S with and without H₂O₂ (Figure S3, supporting information). However, due to its poor solubility, it formed precipitates in the buffers and could not completely remove H₂S (the concentration of H₂S was kept at ~100 μM). This ruled out ebselen as a candidate for further studies. We also found that S5 and S12 had similar solubility problems, so we eventually selected S6 for the following investigations.

The activity of S6 as a scavenger was further investigated. As shown in Figure 6, when H₂S (200 μM) was treated with S6 (200 μM) alone, only ~10% H₂S could be removed and its decrease appeared to be slow. This indicated that the direct reaction between H₂S and the diselenide compound was not very productive. However, in the presence of both S6 and H₂O₂, H₂S was depleted within a few minutes. To demonstrate the specificity of S6 in scavenging H₂S, we tested its ability in removing H₂S in the presence of a series of biologically relevant species including Cys, GSH, homo-Cys, SO₃²⁻, amino acids, etc. As shown in Figure S4 and Table S1 (the supporting information), these species did not cause significant changes to the T_1/2 of H₂S scavenging. In 10% fetal bovine serum (FBS), the efficiency of S6 in removing H₂S was even higher than in pure buffer. These results indicate the high specificity of S6. We next studied the catalytic feature of S6 by varying its dose for the H₂S/H₂O₂ scavenging process. As shown in Figure 7, a dose-dependent H₂S scavenging profile was observed (0.01–1.0 equiv.). S6 showed effective scavenging ability when its dose dropped to 5–10%. However, when the dose was decreased to 1%, its activity was significantly attenuated.

Figure 6.

Representative S6-promoted H₂S scavenging curves in PBS buffer (50 mM, pH 7.4). H₂S (200 μM), H₂O₂ (200 μM), S6 (200 μM).

Figure 7.

Dose-dependent catalytic H₂S scavenging by S6 in PBS buffers (50 mM, pH 7.4). H₂S (200 μM), H₂O₂ (200 μM), S6 (2 μM, 10 μM, 20 μM, 40 μM, 100 μM, 160 μM, and 200 μM).

To further validate the activity of S6, we tested its H₂S-scavenging ability in more complex systems. Bovine plasma and HeLa cell lysate were used in this study. As shown in Figure 8, S6’s H₂S scavenging ability was not significantly affected by the presence of plasma or lysate. In both cases, 0.1 equiv. of S6 was able to efficiently remove H₂S in the presence of H₂O₂. Its activity was even higher in 10% plasma than in PBS buffer. These results suggested that S6 might be a potent scavenger for H₂S in complex biological systems.

Figure 8.

S6 catalyzed H₂S scavenging profiles in: a) w/ 10% bovine plasma, b) w/ 10% HeLa cell lysate. H₂S (200 μM), H₂O₂ (200 μM), S6 (20 μM).

As one can imagine, the scavenging process of S6 should also deplete H₂O₂ from the solutions. We next designed experiments to measure H₂O₂ concentration changes in pure buffer, as well as in plasma and cell lysate-containing mixtures. Briefly, to the mixtures of H₂O₂ (200 μM) and H₂S (200 μM) in the corresponding solutions (pure PBS buffer, w/ 10% plasma, w/ 10% cell lysate), was added S6 (20 μM). After incubating for 15 min, H₂O₂ concentrations were determined by the ferrous oxidation− xylenol orange method (FOX 1 assay). Control experiments (H₂O₂ only, H₂O₂+H₂S, H₂O₂+S6) were also carried out for comparison. As shown in Figure 9, H₂O₂ was found to be quite stable in PBS buffer, and its concentration dropped by <20% when treated with H₂S or S6 alone. In the presence of both S6 and H₂S, >90% H₂O₂ was eliminated. However, >60% H₂O₂ was consumed in plasma and lysate mixtures even in the absence of H₂S and S6. This could be attributed to the presence of catalase or other H₂O₂-reacting biomolecules in both systems. The further decrease of H₂O₂ was not observed when it was treated with H₂S or S6 alone. However, in the presence of both S6 and H₂S, H₂O₂ was significantly depleted, indicating S6 as a H₂O₂ scavenger under these conditions.

Figure 9.

Catalytic H₂O₂-scavenging profiles of S6. H₂O₂ (200 μM), H₂S (200 μM), S6 (20 μM). (a) H₂O₂ only, (b) H₂O₂ + H₂S, (c) H₂O₂ + 10% S6, (d) H₂O₂ + H₂S + 10% S6. The experiments were performed in triplicate, and results are expressed as mean ± SD (n = 3).

Since H₂O₂ is not the only biologically relevant oxidant, we wondered if S6 could also remove H₂S in the presence of other oxidants. tert-Butyl hydroperoxide (t-BuOOH), singlet oxygen (¹O₂), superoxide (O₂^·−), and hydroxyl radical (^•OH) were used in this study. As shown in Figure S5, the direct reactions of H₂S with t-BuOOH, ¹O₂, or O₂^·− were found to be slow and non-productive while the reaction with ^·OH was faster than that of H₂O₂. The initial rate of ^•OH was 10.2 μM min⁻¹, compared to that of H₂O₂ (2.4 μM min⁻¹). We next studied S6-promoted H₂S scavenging with these ROS. The initial rates and T_1/2 were determined and listed in Table 3. It was clear that S6 could dramatically enhance the reactions between H₂S and all four ROS. ^•OH was the most reactive ROS for H₂S, followed by H₂O₂.

Table 3.

S6 promoted H₂S scavenging in the presence of various ROS.

ROS	v0 (μM min−1)[a,b]	T1/2 (min)
H2O2	270 ± 30	0.91
t-BuOOH	208 ± 15	8.62
1O2	130 ± 10	15.2
O2·−	221 ± 17	3.58
·OH	352 ± 23	0.37

^[a]The initial rates were corrected for the uncatalyzed background reaction.

^[b]The reaction curves are shown in Figure S5 in the SI.

To understand the mechanism of diselenide (S6)-mediated scavenging of H₂S and H₂O₂, we analyzed the reaction products and potential intermediates. Elemental sulfur (S₈) was found to be the final product from H₂S, which was similar to the results shown in Table 1. We suspected that hydrogen persulfide (H₂S₂) was the key intermediate in this process and then monitored its formation using a specific fluorescent probe PSP-3.^[21] As shown in Figure 10, treating PSP-3 with S3 or S6 alone did not give any notable fluorescence. Weak fluorescence was observed when treating PSP-3 with a mixture H₂S and S3. The background reactions of H₂S with H₂O₂ or H₂S with S6 also gave weak fluorescence due to the slow formation of H₂S₂.^[12] However, much stronger fluorescence was observed when incubating PSP-3 with a mixture of S6, H₂S and H₂O₂, suggesting the formation of H₂S₂ in this process.

Figure 10.

Fluorescent responses of PSP-3 toward different treatments: (1) PSP-3 only, (2) PSP-3+S3, (3) PSP-3+S6, (4) PSP-3+H₂S+H₂O₂, (5) PSP-3+S3+H₂S, (6) PSP-3+S6+H₂S, (7) PSP-3+S6+H₂S+H₂O₂, (8) PSP-3+Na₂S₂. Concentrations: PSP-3 (10 μM), H₂S (200 μM), H₂O₂ (200 μM), S3 (200 μM), S6 (200 μM), Na₂S₂ (100 μM). Incubation time: 20 min. The experiments were performed in triplicate, and results are expressed as mean ± SD (n = 3).

Based on our results thus far, we proposed a mechanism for the diselenide catalyzed H₂S/H₂O₂ scavenging. As shown in Scheme 2, diselenide could react with H₂S to produce selenylsulfide (RSeSH) and selenol (RSeH), similar to a reaction we reported in 2015.^[22a] RSeSH could further react with H₂S to form H₂S₂ and RSeH. The selenol (RSeH) could be oxidized back to diselenide by atmospheric O₂.^[22b] RSeSH could undergo a disproportionation reaction to regenerate diselenide with the release of H₂S₂. In the presence of ROS such as H₂O₂, diselenide could be rapidly oxidized to the corresponding selenenic acid (RSeOH) and seleninic acid (RSeO₂H), ^[22c] both of which should react with H₂S to produce RSeSH and eventually regenerate diselenide to complete the catalytic cycle.

Scheme 2.

The proposed mechanism of H₂S/H₂O₂ scavenging by S6.

Our results demonstrate that diselenides can simultaneously scavenge ROS and H₂S under physiological conditions. In other words, the combination of ROS-diselenide can serve as a superior H₂S scavenging system. However, ROS may not necessarily co-exist with H₂S. In such scenarios it is unrealistic to simply add ROS for the purpose of removing H₂S, as the bolus addition of ROS is often problematic. Precise control of the timing and location of ROS generation would be critical for the ‘in situ’ creation of the catalytic H₂S scavenging system. We envisioned that photosensitizers (PS) could be used to achieve this goal. PS can promote ROS formation upon exposure to visible lights. Previous studies loaded PS into diselenide-containing polymers and found light-induced ROS could effectively oxidize diselenides to seleninic acids.^[23] Thus, we expected that the PS-diselenide combination could serve as a unique light-triggered scavenger to remove H₂S in a controllable manner. With this idea in mind, four PS (Figure 11) were selected to test the light-promoted H₂S scavenging. Nile blue A (NBA) was used to optimize the protocol. Briefly, PS (20 μM) was added to a solution of H₂S (200 μM) in PBS buffer and the mixture was kept in dark for 2 min. Next, S6 (20 μM) was added. After another 1 min, the solution was exposed to red light (660 nm, 120 mW/cm²) for 5 min. The concentration of H₂S was continuously monitored and recorded (Figure S7–S12). Control experiments (H₂S only, H₂S+PS, H₂S+S6, with or without light) were also carried out for comparison. The results were summarized in Figure 11. In general, the concentration of H₂S was found to be stable in the absence of PS and S6, even after 5 min light irradiation. 20 μM S6 could remove ~5% of H₂S, and this process was not affected by light. H₂S was inert to all four PSs in the dark. When exposed to NIR light, NBA and EtNBS-COOH could only remove 5–7% of H₂S, indicating their weak reactions with H₂S. In the presence of both NBA and S6, ~30% H₂S was eliminated after light exposure, which was a 17% increase over the control group. This synergistic effect was further enhanced to 29% by EtNBS-COOH, probably due to its type I PS feature.^[24] Two xanthene-based PS, 2’,7’-dichloro-fluorescein (2’,7’-FlCl₂) and Eosin Y, under white light (400–700 nm, 15 mW/cm²) irradiation were also tested. 2’,7’-FlCl₂ itself could remove ~14% of H₂S upon light exposure and a 20% enhancement was observed with the addition of S6. Eosin Y itself could remove ~50% of H₂S under light. Interestingly, when both Eosin Y and S6 were presented, a significant increase of H₂S scavenging was observed and >95% of H₂S was removed after light exposure. The better performance of Eosin Y may be due to its high ¹O₂ generation capacity (Φ_Δ = 0.52), while other PS’ quantum yields are much lower (Φ_Δ 0.005–0.07).^[25] In addition, a recent work found that Eosin Y-generated ¹O₂ could be effectively converted to H₂O₂ with amine-containing chemicals.^[26] This ‘ROS conversion’ may further contribute to Eosin Y’s efficacy as S6 contains the amine group. These results indicated that light-induced H₂S scavenging could be achieved by combining specific photosensitizers and diselenides, and the synergistic effect was determined by the feature of the photosensitizers.

Figure 11.

(A) Structures of photosensitizers. (B) Light-induced H₂S scavenging with photosensitizers and S6. H₂S (200 μM), photosensitizers (20 μM), S6 (20 μM). The experiments were performed in triplicate, and results are expressed as mean ± SD (n = 3).

Conclusion

The overproduction of H₂S is believed to be linked to various pathological processes. However, downregulation of endogenous H₂S in certain tissue or cellular locations is still challenging. Recently, some scavengers that allow specific and rapid H₂S clearance were reported. While such scavengers have shown interesting activities, a common problem is that they are based on stoichiometric reactions with H₂S. High doses of substrates may be needed, which could cause unexpected side effects. Since elevated H₂S conditions, such as acute H₂S poisoning, are known to induce oxidative stress with high ROS production, we explored the development of catalytic scavengers to concurrently remove H₂S and ROS in biological settings. In this work, we tested a series of Se-based compounds as possible candidates. Selenoxide (SeO)-containing compounds were first found to be potent H₂S scavengers and exhibit promising efficacy as antidotes for lethal H₂S intoxication in an ongoing exposure mouse model. Since selenoxides could be reduced by H₂S to form diselenides, we then evaluated 14 diselenide/monoselenide compounds and demonstrated that diselenides could simultaneously remove H₂S and ROS. Structural-activity relationship studies led to the identification of a promising compound S6, whose scavenging activity in pure buffers and plasma- or cell lysate-systems were demonstrated. We also analyzed the reaction products and potential intermediates and proposed a mechanism for this catalytic process. This S6-catalyzed reaction of H₂S and ROS may be viewed as a mimic of certain H₂S metabolic reactions, such as sulfide-quinone oxidoreductase (SQR) mediated persulfide/H₂S_n formation. They all convert H₂S to unique sulfane sulfur species including S₈. While S₈ normally shows low reactivity in biological systems, recent studies suggest S₈ could exbibit biological functions like persulfidation under certain conditions.^[27] S₈ may serve as the endogenous reserve pool for H₂S and persulfides. Precisely controlled exchange reactions between H₂S and S₈ may be essential for redox signaling. Moreover, the combination of specific photosensitizers (for the controlled generation of ROS) and the diselenide scavenger was used to achieve light-induced H₂S scavenging. These results suggest that Se-based scavengers are useful tools for investigating H₂S/ROS signaling pathways.