Rational design and bioimaging applications of highly selective fluorescence probes for hydrogen polysulfides.

Reactive sulfur species have received considerable attention due to their various biological functions. Among these molecules, hydrogen polysulfides (H2S(n), n > 1) are recently suggested to be the actual signaling molecules derived from hydrogen sulfide (H2S). Hydrogen polysulfides may also have their own biosynthetic pathways. The research on H2S(n) is rapidly growing. However, the detection of H2S(n) is still challenging. In this work we report a H2S(n)-mediated benzodithiolone formation under mild conditions. Based on this reaction, specific fluorescent probes for H2S(n) are prepared and evaluated. The probe DSP-3 shows good selectivity and sensitivity for H2S(n).

Reactive sulfur species (RSS) are a family of sulfur-containing molecules found in biological systems. These molecules include thiols, hydrogen sulfide, persulfides, polysulfides, and S-modified cysteine adducts such as S-nitrosothiols and sulfenic acids. So far many RSS have been demonstrated to exert interesting biological functions.[1−3] Among those, hydrogen sulfide (H2S) is probably most attractive as this gaseous molecule has been recently known as a critical cell signaling molecule, much like nitric oxide. Literature published in the past several years increasingly suggests that H2S is a mediator of many physiological and/or pathological processes, especially in cardiovascular systems.[4−7] In contrast, hydrogen polysulfides (H2S, n > 1) have received much less attention. These species can be considered as oxidized forms of H2S and belong to sulfane sulfur in RSS family. From a chemistry perspective, H2S and H2S are redox partners and therefore very likely coexist in biological systems. On the other hand, H2S may have their own biosynthetic pathways or can be generated from H2S. H2S could also be the precursors of H2S through their degradation.

Because of these properties, some biological mechanisms that were originally attributed to H2S may actually be mediated by H2S. For instance, one of the most interesting reactions of H2S is S-sulfhydration, i.e., converting protein cysteines (-SH) to persulfides (-S-SH). This reaction is significant because it provides a possible mechanism by which H2S alters the functions of a wide range of cellular proteins and enzymes.[8−11] However, how this reaction proceeds is still unclear. Theoretically H2S itself can hardly react with protein cysteine residues or disulfides to form S-sulfhydration. It is possible that H2S reacts with modified cysteines such as S-nitrosothiols (SNO) or S-sulfenic acids (SOH) to form S-sulfhydration.[12,13] Recently the possibility that the reaction is caused by H2S has been revealed.[14−17] From a reactivity point-of-view, H2S should be much more effective in S-sulfhydration than H2S. Kimura found that H2S were indeed hundreds times more potent than H2S in inducing Ca2+ influx in astrocytes via S-sulfhydration on TRPA1 channels.[18] He also found that H2S were very effective in S-sulfhydration on Keap1, the key protein regulating Nrf2 signaling.[19] In another report by Dick and Nagy et al., H2S were found to efficiently sulfhydrate proteins such as roGFP2 and PTEN, while H2S could not cause sulfhydration in the presence of potassium cyanide, an H2S scavenging reagent.[20]

In order to better understand the roles of H2S and differentiate H2S from H2S, it is important to study the fundamental chemistry/reactivity of H2S and develop new methods for their detection. The traditional method for detecting H2S is to measure UV absorption peaks at 290–300 and 370 nm, which is not sensitive and applicable for biological detections.[20] In this respect, fluorescence assays may be useful because of their high sensitivity and spatiotemporal resolution capability. Unfortunately, there is no report on such fluorescent probes for H2S so far. To this end, we have initiated a program to study new reactions of H2S, aiming at developing new fluorescent probes based on these reactions. Herein we report this attempt.

H2S is a combination of polysulfide species. The dissolution of any polysulfide salts should result in similar distribution of these species (this will depend on the relative ratios of sulfide vs the oxidizing equivalents and the applied pH).[14] Hydrogen disulfide (H2S2) may be an active species of H2S, and there should be a dynamic equilibrium between H2S2 and other H2S.[18] Therefore, our focus has been put on the chemistry of H2S2. Taking the advantage of two -SH groups in H2S2, we envisioned that compounds containing bis-electrophilc groups should be able to selectively capture H2S2 (Scheme 1). If one of the electrophilic groups is a latent fluorophore and can be released under nucleophilic reactions (such as A in Scheme 1), the strategy may be suitable in the development of fluorescent probes for H2S2. It is possible that biothiols, i.e., cysteine (Cys) and glutathione (GSH), may compete with H2S2 in reacting with probe A. However, product B should not turn-on the fluorescence. Moreover, upon manipulating electronic properties of the probe, H2S/H2S2 may further react with B (via the SN2Ar reaction) to switch the thioether and turn on the fluorescence (vide infra).

Scheme 1

Proposed Strategy for Capturing and Visualizing H2S

With this idea in mind, three 2-fluorobenzoiate derivatives (1–3) were prepared and studied in the reactions of H2S2 (Scheme 2). In this study H2S2 was always used as the primary model compound of H2S. We expected that the activated fluorobenzoiates should undergo nucleophilic aromatic substitution with H2S2 to form the corresponding persulfide intermediates I, which in turn undergoes a cyclization to form benzodithiolone products and release phenol. In these experiments freshly prepared solutions of Na2S2 were used as the equivalent of H2S2. The reactions were carried out in a mixed solution of CH3CN/PBS (pH 7.4, 1:1 v/v). The products were analyzed after 1 h at room temperature. As expected, when the parent compound 1 was treated with H2S2, the desired cyclization product 4 was obtained in low yield (7%). The substrates with electron-withdrawing groups (-CN and -NO2) showed much improved reactivity, and the corresponding cyclization products (5 and 6) were obtained in modest to good yields. As the nitro-substitution (compound 2) was found to be most effective, this compound was selected for further studies.

Scheme 2

Model Reactions of the Probes with H2S

We then tested the reaction between 2 and another hydrogen polysulfide model compound (Na2S4). The reaction worked well, and the desired cyclization product was obtained in good yield (85%). This result confirms that H2S2 may be the major component of H2S or that there is a fast equilibrium between H2S and H2S2.[18] It therefore suggests that compounds like 2 are suitable for capturing H2S. Another concern is that biothiols may also react with the probes that are designed for trapping H2S, leading to the consumption of the probes. To address this concern, we tested the reaction of 2 with a biothiol model 7. The substitution product 8 was obtained in 40% yield under the same conditions. This suggests that biothiols are less reactive (than H2S) toward the substrate. Moreover, compound 8 was able to further react with H2S to give the cyclization product 5. These results indicate that biothiols would not interfere with the detection of H2S.

The reaction shown in Scheme 2 provides a possible application in developing fluorescent probes for H2S. It is known that hydroxyl (-OH) protection (e.g., acylation or alkylation) of many fluorophores can result in fluorescence quenching, and deprotection can restore the fluorescence.[21−28] If -OH sensitive fluorophores are introduced to the benzoiate of 2, the resultant compounds would be specific probes for H2S as they may react with H2S to release the fluorophores. Based on this strategy, three probes (DSP-1, DSP-2, and DSP-3) are synthesized (Scheme 3). Detailed synthetic protocols and structure characterizations are provided in the Supporting Information.

Scheme 3

Structures of New H2S Fluorescent Probes

Next we tested the probes’ fluorescence properties and responses to H2S. We first studied the detection conditions and found that PBS buffer (50 mM, pH 7.4) containing 25 μM cetrimonium bromide (CTAB) was the optimum system (Figure S1). DSP-1 and DSP-3 showed almost no fluorescence emission at 515 nm due to the protection of the two hydroxyl groups of fluorescein, but DSP-2 showed some background fluorescence due to the protection of only one hydroxyl group of fluorescein. Upon treatment with Na2S2, both DSP-1 and DSP-3 gave significant fluorescence enhancements (Figure 1), whereas DSP-2 did not, which may be attributed to its strong background fluorescence. As DSP-3 exhibited a much stronger fluorescence response than DSP-1 (137 vs 57 fold), this probe was selected for further evaluation.

Figure 1

Fluorescence enhancements (F/F0) of probe (10 μM) (1) DSP-1; (2) DSP-2; and (3) DSP-3 with Na2S2 (50 μM) in PBS buffer (50 mM, pH 7.4) containing 25 μM CTAB. Reactions were carried out for 20 min at room temperature. Data were acquired at 515 nm with excitation at 490 nm.

Figure 2 shows the time-dependent fluorescence changes of DSP-3 (10 μM) in the presence of Na2S2 (50 μM). The maximum emission intensity at 515 nm was reached within 5 min, indicating a fast reaction. For the purpose of reproducibility, a reaction time of 20 min was employed in all other experiments. The effects of pH in this reaction were also investigated, and DSP-3 was found to work effectively in neutral to weak basic pH range of 7–8 (Figure S2).

Figure 2

Time-dependent fluorescence intensity changes of DSP-3 (10 μM) in the presence of Na2S2 (50 μM). Reactions were monitored for 40 min at room temperature.

To test the selectivity of the probe for H2S, DSP-3 was treated with a series of RSS including GSH, Cys, Hcy, GSSG, H2S, SO32–, S2O32–, CH3SSSCH3, and S8. As shown in Figure 3A, no significant fluorescence increase was observed for any of these compounds (columns 2–11). Only Na2S2 and Na2S4 gave strong fluorescence increase (columns 12 and 13). We also tested the responses of DSP-3 to other representative amino acids and found no responses (Figure S3). Moreover, when Na2S2 (50 μM) and other RSS coexisted, we still observed obvious fluorescence enhancements (Figure 3B). Compared to the results of Na2S2 only, almost the same levels of fluorescence turn-on responses (without any loss) were observed for most of these compounds. GSH, Cys, and Hcy did cause some fluorescence decrease, presumably due to the reaction between H2S2 and thiols, leading to the decreased concentrations of H2S2 in solutions.[18−20,29] These results demonstrate good selectivity of DSP-3 for H2S2 and hydrogen polysulfides, suggesting that DSP-3 may be useful for monitoring of H2S in biological systems.

Figure 3

(A) Fluorescence enhancements (F/F0) of DSP-3 (10 μM) in the presence of various RSS. (1) probe alone; (2) 8 mM GSH; (3) 500 μM Cys; (4) 100 μM Hcy; (5) 100 μM GSSG; (6) 100 μM Na2S; (7) 100 μM Na2S2O3; (8) 100 μM Na2SO3; (9) 100 μM Na2SO4; (10) 100 μM CH3SSSCH3; (11) 100 μM S8; (12) 50 μM Na2S2; (13) 50 μM Na2S4. (B) Fluorescence enhancements (F/F0) of DSP-3 (10 μM) to the mixture of various RSS with 50 μM Na2S2. (1) 1 mM GSH; (2) 500 μM Cys; (3) 100 μM Hcy; (4) 100 μM GSSG; (5) 100 μM Na2S; (6) 100 μM Na2S2O3; (7) 100 μM Na2SO3; (8) 100 μM Na2SO4; (9) 100 μM CH3SSSCH3; (10) 100 μM S8; (11) 50 μM Na2S2.

To demonstrate the efficiency of this probe in the measurement of H2S, varying concentrations of Na2S2 (0.5–50 μM) were added to the solutions of DSP-3 (10 μM). The fluorescence intensities were linearly related to the concentrations of Na2S2 in the range of 0.5–15 μM (Figure 4). The detection limit[30,31] was calculated to be around 71 nM, indicating a high sensitivity.

Figure 4

Fluorescence emission spectra of DSP-3 (10 μM) with varied concentrations of Na2S2 (0, 0.5, 1, 3, 6, 10, 15, 20, 30, 40, 50 μM for curves 1–11, respectively). Reactions were carried out for 20 min at room temperature.

It should be noted that the biosynthetic pathways of H2S are still unclear. Recent studies suggested that they may come from H2S in the presence of reactive oxygen species (ROS).[12,14,15,17,18,20,32] We then applied DSP-3 in detecting in situ generated H2S from H2S and ROS. As shown in Figure 5, the probe did not give any response to commonly existing ROS including hydrogen peroxide (H2O2), hypochlorite (ClO–), superoxide (O2–), hydroxyl radical (•OH), and singlet oxygen (1O2) (columns 1–6). However, when H2S was premixed with ROS (columns 7–12), significant fluorescence signals were obtained, suggesting the formation of H2S in these systems. Apparently H2S together with ClO– gave the strongest signals (column 9), indicating that ClO– is the most effective ROS converting H2S to H2S in our in vitro testing systems. This result confirms the discovery by Nagy et al. that hypochlorous acid can rapidly react with H2S to form hydrogen polysulfides.[32]

Figure 5

Fluorescence enhancements (F/F0) of DSP-3 (10 μM) in the presence of various reactive oxygen species (with or without H2S). Reactions were carried out for 20 min at room temperature. (1) 50 μM H2O2; (2) 200 μM H2O2; (3) 50 μM ClO–; (4) 50 μM O2–; (5) 50 μM ·OH; (6) 50 μM 1O2; (7) 50 μM H2O2 + 50 μM Na2S; (8) 200 μM H2O2 + 50 μM Na2S; (9) 50 μM ClO– + 50 μM Na2S; (10) 50 μM O2– + 50 μM Na2S; (11) 50 μM •OH + 50 μM Na2S; (12) 50 μM 1O2 + 50 μM Na2S.

Finally the application of DSP-3 in monitoring H2S in cultured cells was tested. As shown in Figure 6, HeLa cells were first incubated with DSP-3 (10 μM) for 20 min, and no fluorescence was observed. Strong fluorescence in the cells was induced after treating with Na2S2 (100 μM). In comparison, cells treated with H2S (using 100 μM Na2S) did not show obvious fluorescence. In addition, the cell viability assay demonstrated that DSP-3 has almost no cytotoxicity (Figure S4). These results suggest that DSP-3 is cell permeable and can be used in detecting H2S (not H2S) in cells.

Figure 6

Confocal fluorescence images of H2S in HeLa cells. Cells on glass coverslips were incubated with DSP-3 (10 μM) for 20 min, then washed, and subjected to different treatments. (a) control (no Na2S2); cells treated with (b) 100 μM Na2S2 and (c) 100 μM Na2S. Second row shows the corresponding differential interference contrast images for the first row.

In summary, we report in this study a H2S/H2S2-mediated benzodithiolone formation under mild conditions. This reaction proves to be specific for H2S/H2S2 over other RSS such as biothiols and H2S. Based on this reaction, a fluorescent probe, DSP-3, was developed for sensitive and selective detection of H2S/H2S2 in aqueous buffers as well as in cells. With probe DSP-3, we also confirm the possibility of H2S formation from the reaction of H2S with ROS such as ClO–. We are now utilizing these probes to study the contributions of hydrogen polysulfides to physiological and pathological processes. It should be noted that more sensitive fluorescent probes for endogenous hydrogen polysulfides may be needed, and our present design approach should lead to the development of such probes by using more strongly luminescent fluorochrome.