Sulfur Dioxide Prodrugs: Triggered Release of SO₂ via a Click Reaction

Abstract

Sulfur dioxide (SO₂) is being recognized as a possible endogenous gasotransmitter with importance on par with that of NO, CO, and H₂S. Herein we describe a series of SO₂ prodrugs that are activated for SO₂ release via a bioorthogonal click reaction. The release rate can be tuned by adjusting the substituents on the prodrug.

SO₂ gas is well known as an air pollutant and may cause respiratory and cardiovascular diseases upon chronic exposure.^1–3 SO₂ has also long been recognized as an antimicrobial agent, and widely used in food industry, especially in the brewery industry. Recent years have seen increasing evidence that SO₂ can be produced endogenously by oxidation of sulfur-containing amino acids (cysteine, homocysteine).^4–6 SO₂ has also been shown to have possible physiological functions. There has been an increasing level of understanding of SO₂’s mechanism of actions at the pathway level,^7–9 and fluorescent probes have been reported that can detect endogenous production of SO₂.^{10, 11} SO₂ has been reported to exacerbate ischemia-reperfusion injury, and suppress hypertension, pulmonary hypertension, vascular remodeling, and inflammation.¹² Thus, interest in exploring SO₂’s functions as a gasotransmitter has increased in recent years. Up until now, knowledge of SO₂’s physiological and pathological roles is still very limited, especially in comparison with the other three gasotransmistters NO, CO, and H₂S.

There are three main methods currently used to provide SO₂ for research purposes: gaseous SO₂, HSO₃⁻/SO₃²⁻ pair, and SO₂ donors. Despite the broad application in research activities, gaseous SO₂ has obvious disadvantages including difficulties in preparing solutions of precise concentrations. Besides, the required equipment can be cumbersome with potential hazard for lab personnel. The HSO₃⁻/SO₃²⁻ ion pair method relies on the equilibrium established between hydrated SO₂ (SO₂·H₂O) and HSO₃⁻/SO₃²⁻ ion pair in aqueous solution to release SO₂.^{13, 14} However, using the HSO₃⁻/SO₃²⁻ ion pair for SO₂ release does not allow easy control of its quantity and relies on conversion kinetics, which can be very complex. Another drawback of this method is that it is hard to de-convolute the biological effects of HSO₃⁻/SO₃²⁻ and molecular SO₂, which have been demonstrated to be different.¹⁵ Thus, there is a need to develop prodrugs with controlled release of SO₂ to imitate the process of endogenous production of SO₂. Several groups have developed SO₂ donor molecules with various triggering mechanisms. Chakrapani and colleagues have reported clever designs that utilize biological thiol (cysteine) and UV light to trigger the release of SO₂ gas from donors.^{16, 17} They also developed elegant work on using benzosultines to generate SO₂ under physiological conditions.¹⁸ Recently, Xian’s group reported benzothiazole sulfinate (BTS) as a pH dependent water soluble SO₂ donor.¹⁹ We are interested in developing SO₂ prodrug systems with controllable release rates under physiological conditions. Such prodrugs would complement what is already available in providing research tools to help advance this area of research.

Our group has a long-standing interest in studying gasotransmitters.^20–26 Previously, we have demonstrated the feasibility of using an extrusion reaction as a way to deliver carbon monoxide (CO).²¹ Herein we describe a strategy to cage SO₂ in thiophene dioxide and use a strained alkyne to trigger the release of SO₂ in a controllable fashion (Scheme 1). By varying the substituents on the thiophene dioxide scaffold, the release rates can also be tuned.

Scheme 1.

Proposed SO₂ donor system

As early as 1976, Stille already examined the reaction of thiophene dioxide and a terminal alkyne under reflux condition in toluene. Though conditions were “harsh”, the cheletropic reaction yielded a cyclized product with the release of SO₂.²⁷ We planned to take advantage of this reaction to construct caged SO₂ for release under near physiological conditions (room temperature to 37 °C, pH 7.4, and aqueous solution). A key issue in this project was how to lower the reaction temperature to ambient temperature for applications under near physiological conditions. 2,3,4,5-Tetrachlorothiophene dioxide has been shown to possess high reactivity in cycloaddition reactions with a variety of olefines. Specifically, it was shown to react steadily with ethylene and to give a cyclohexyldiene product and SO₂ at 28 °C.²⁸ We reasoned that a strained alkyne with increased HOMO energy would promote this cycloaddition reaction. The formation of a stable phenyl ring should also help to drive the subsequent cheletropic reaction with the concomitant release of SO_2. Therefore, we synthesized 2,3,4,5-tetrachlorothiophene dioxide 3 from commercially available perchlorothiophene 2. Then we first examined the reaction of 3 (20 mM in MeOH) with endo-BCN (1, 4 equiv.) at room temperature as a proof-of-concept test. Indeed, after 5 min, we saw complete conversion of 3 into a new product, which was spectroscopically characterized as 4 (Scheme 2).

Scheme 2.

Proof-of-concept test of SO₂ donor system

We then examined the reaction kinetics. The UV absorption of the product is significantly lower than that of the reactants at 328 nm. Therefore, we used UV absorbance decrease to monitor the reaction progress. The second order rate constant was determined by first using a large excess of 1 to examine the pseudo-first order reaction followed by plotting the pseudo-first order reaction rate constants against different 1 concentrations. The second order rate constant was determined to be 1.50 M⁻¹ s⁻¹ in MeOH at room temperature, allowing the reaction to be finished in about 3 hours at mid-μM levels (Figure S1B).

To further confirm the formation of SO₂, we performed the DTNB test, which has been applied in SO₂ measurement in atmosphere and food samples.^29–31 DTNB would react with SO₃²⁻ ion to give an reduced product, which can be quantitatively measured by UV absorbance at 412 nm (Figure 1A).²⁹ To allow hydration of SO₂, we performed this test in 5% DMSO/PBS. A reaction mixture was incubated at room temperature or 37 °C for 45 min (about 1 t_1/2) and then the DTNB probe was added. Afterwards, the test solution was incubated at room temperature for another 15 min and subjected to UV absorbance reading at 412 nm. The experimental group showed higher absorbance at both room temperature and 37 °C than the negative control groups, confirming the formation of SO₃²⁻. Both experimental groups and positive control groups showed lower SO₃²⁻ formation at 37 °C than at room temperature, presumably due to increased SO₂ escape at higher temperature (Figure 1B).

Figure 1.

(A) DTNB test mechanism. (B) UV absorption at 412 nm of DTNB test after 45 min incubation at room temperature or 37 °C of (a) blank (5% DMSO/PBS), (b) 250 μM of 3 only, (c) 2.5 mM of 1 only, (d) 250 μM of 4 only, (e) 250 μM of 3 +2.5 mM of 1, (f) 50 μM of Na₂SO₃, (g) 50 μM of NaHSO₃.

Encouraged by the initial success, we went on to synthesize additional analogues to see whether we can tune the reaction rates for various applications. Thus, we synthesized compounds 5 and 6 with electron withdrawing groups at different positions (Figure 2A; for synthetic scheme see ESI). We reasoned that by varying the electron density on the thiophene ring, different LUMO energy will result in varied reaction rate with BCN. We also tested reaction with a trans-cyclooctene compound (8, equatorial isomer), which is known to have high strain energy and high HOMO, to see if the reaction rate can be further enhanced. In all these reactions at room temperature, we successfully isolated the cheletropic reaction products (see ESI). Interestingly, we had initial difficulties in analyzing the room temperature NMR spectra of the products containing the bicyclo[6.1.0]non-4-ene structure. This was due to peak overlap caused by line broadening in ¹H NMR. Structure elucidation by ¹³C NMR is equally difficult because of missing peak(s) at high field. Given HRMS confirmation, we reasoned that this phenomenon is caused by the slow flipping of eight-member ring between the “chair” and “boat” conformations. We successfully observed two conformations by tracking the −CH₂ group next to the −OH group at low temperature (270 K). As temperature increases, the exchange rate of the two populations increases, and the two peaks observed for the −CH₂ group coalesced and then averaged to give a sharp peak. Unfortunately, within the temperature range tested, the exchange rate is not fast enough to average the −CH₂ and −CH groups from the bicyclo[6.1.0]non-4-ene structure, which became broad above 278 K and hard to track (Figure 2C). However, this phenomenon has been observed exclusively with endo-product. As for product obtained from reaction between 6 and exo-BCN, sharp peaks at high field were observed at room temperature (see ESI). By contrast, with chlorine substitutions, we are able to track the carbons at room temperature, presumably because of its fast flipping rate. Edited HSQC experiment gave clear evidence of geminal protons splitting from the eight-membered ring, confirming the structures of cheletropic products (see ESI). We studied the reaction kinetics for these reactions and obtained k₂ values ranging from 0.02 to 0.33 M⁻¹s⁻¹ (Table 1; Figure S1). However, compound 5 was found to be unstable in MeOH at room temperature, which was presumably ascribed to its strong electrophlicity. Therefore, compound 5 is not a suitable SO₂ prodrug for biological applications due to this stability issue.

Figure 2.

(A) SO₂ donor pairs. (B) Fluorescent SO₂ donor and product after SO₂ release (100 μM in MeOH). (C) Coalescence of cheletropic product (6 + 1) observed in ¹³C NMR (DEPT-135) with increasing temperature.

Table 1.

Kinetics study of SO₂ donor pairs

SO2 donor pairs	3 + 1	3 + 8	5 + 1	6 + 1	7 + 1	6 + exo-BCN
k2 (M−1s−1)	1.50a	0.02a	0.33b	0.05a	0.01c	0.04a

^aMeasured in MeOH under room temperature.

^bMeasured in ACN under room temperature.

^cMeasured in DMSO/PBS 4:1 at 37 °C.

Currently, several elegant fluorescent probes for SO₂ have been developed based on the nucleophilicity of SO₃²⁻/HSO₃⁻. However, few of them are suitable for real-time monitoring of SO₂ generation due to delayed response.³² One alternative strategy for real time monitoring of SO₂ release is to devise such SO₂ prodrugs that would become fluorescent after SO₂ release. Therefore, we were interested in designing SO₂ prodrugs that would lead to the formation of a fluorescent reporter, which allows for real-time monitoring of SO₂ release. We reasoned that attaching an aromatic substituent to the thiophene ring would allow formation of an expanded conjugation system after the cheletropic reaction and thus result in pronounced changes in the spectroscopic behavior of the aromatic system after the reaction. Therefore, we synthesized thiophene dioxide with naphthalene conjugated at the 3,4 positions (7, Figure 2A). The compound itself has weak yellow fluorescence with an excitation wavelength of 299 nm, emission wavelength of 556 nm, and a quantum yield of 0.006. After reaction with BCN and SO₂ release, the resulting product shows strong cyan fluorescence with an emission wavelength of 470 nm, excitation wavelength of 296 nm, and quantum yield of 0.138 (Figure 2B). We also established the kinetics profile of this reaction by monitoring the fluorescence increase at 470 nm. A k₂ of 0.01 M⁻¹s⁻¹ was obtained (Table 1; Figure 3; Figure S2).

Figure 3.

Fluorescence response increases as cycloaddition reaction between 7 and 1 proceeds (λ_ex=350 nm). Condition: 250 μM 7 + 2500 μM 1 in 4:1 DMSO/PBS at 37 °C.

In summary, we have developed a SO₂ donor system, which releases SO₂ through a Diels-Alder reaction with a strained alkyne/alkene as a trigger. We explored the initial tunability of the releasing rate by varying the electron density of thiophene dioxide ring. With a small set of compounds, we were able to tune the reaction rate in the range of 0.01–1.50 M⁻¹s⁻¹, giving a 150-fold range in reactivity.

Additional work is underway in further extending the tunability of reaction rate and in testing the donor system in triggering various biological responses.