Modified cyclodextrins solubilize elemental sulfur in water and enable biological sulfane sulfur delivery

Abstract

An important form of biological sulfur is sulfane sulfur, or S⁰, which is found in polysulfide and persulfide compounds as well as in elemental sulfur. Sulfane sulfur, often in the form of S₈, functions as a key energy source in the metabolic processes of thermophilic Archaean organisms found in sulfur-rich environments and can be metabolized both aerobically and anaerobically by different archaeons. Despite this importance, S₈ has a low solubility in water (∼19 nM), raising questions of how it can be made chemically accessible in complex environments. Motivated by prior crystallographic data showing S₈ binding to hydrophobic motifs in filamentous glycoproteins from the sulfur reducing Staphylothermus marinus anaerobe, we demonstrate that simple macrocyclic hydrophobic motifs, such as 2-hydroxypropyl β-cyclodextrin (2HPβ), are sufficient to solubilize S₈ at concentrations up to 2.0 ± 0.2 mM in aqueous solution. We demonstrate that the solubilized S₈ can be reduced with the common reductant tris(2-carboxyethyl)phosphine (TCEP) and reacts with thiols to generate H₂S. The thiol-mediated conversion of 2HPβ/S₈ to H₂S ranges from 80% to quantitative efficiency for Cys and glutathione (GSH). Moreover, we demonstrate that 2HPβ can catalyze the Cys-mediated reduction of S₈ to H₂S in water. Adding to the biological relevance of the developed systems, we demonstrate that treatment of Raw 264.7 macrophage cells with the 2HPβ/S₈ complex prior to LPS stimulation decreases NO₂⁻ levels, which is consistent with known activities of bioavailable H₂S and sulfane sulfur. Taken together, these investigations provide a new strategy for delivering H₂S and sulfane sulfur in complex systems and more importantly provide new insights into the chemical accessibility and storage of S⁰ and S₈ in biological environments.

---

Introduction

Sulfur is a long-standing critical component to life on Earth. Prior to the Great Oxidation Event approximately 2.4 billion years ago, during which time the Earth's atmosphere became rich in O₂, the atmosphere on Earth was weakly reducing.¹ Volcanic activity was an abundant source of sulfur-containing species, and gases including sulfur dioxide (SO₂) and hydrogen sulfide (H₂S) were released and dissolved in pools or lakes.² Spark discharge experiments using a simulated atmosphere containing H₂S and a mixture of reducing gases thought to be present in the early Earth atmosphere have demonstrated abiotic synthesis of diverse organic compounds including amino acids.³ In conditions reflecting those near deep-sea vents rich in iron-sulfur compounds, both H₂ and organosulfur compounds were generated from a mixture of H₂S and FeS under a N₂/CO₂ atmosphere in acidic conditions.⁴ Sulfane sulfur (S⁰), which is most commonly found as elemental octasulfur (S₈), is also found in these deep-sea environments and is another important source of biologically available sulfur.

Recently, interest in biological and synthetic S⁰ sources has increased significantly due to the connection between such species and the small biological signaling molecule H₂S. H₂S is produced endogenously from cysteine metabolism and serves signaling roles in diverse pathways. Along with carbon monoxide (CO) and nitric oxide (NO), H₂S is now recognized as member of the family of small molecules often referred to as gasotransmitters, which are produced enzymatically and act upon specific molecular targets within cellular environments.^5–7 One unique feature that distinguishes H₂S from CO and NO is that sulfur has biologically-accessible oxidation states ranging from −2 to +6 and participates in a complex redox cellular landscape.⁸ In many eukaryotic organisms, H₂S serves as a source of biologically available sulfur and is intrinsically tied to both organic and inorganic S⁰-containing species, including persulfides and related polysulfides/polysulfanes, in the S⁰ pool. This redox labile pool can generate H₂S upon reduction or participate in transpersulfidation reactions to transfer S⁰ moieties to cysteine residues.⁹

Prior investigations into the reaction between elemental sulfur and H₂S have demonstrated the formation of inorganic polysulfide ions (S²⁻_n>1), which are also important intermediates in sulfur cycling in sediments.¹⁰ Interestingly, biologically produced sulfur particles from Thiobacillus sp. W5 react with HS⁻ to generate mixed polysulfides more efficiently than when compared S₈ (α-S₈) directly.¹¹ All of these examples highlight the important role and interconnected roles of the reactivity of S⁰ and intermediate polysulfide formation upon reaction with sulfhydryl nucleophiles. Connecting with the biological activity of H₂S, the generation, action, and translocation of S⁰-containing species is critical to understanding the intertwined chemistry of reactive sulfur species in biology.

Despite this broad importance in both contemporary and evolutionary chemistry and biology, investigations into S⁰ activity in aqueous systems are challenging due to the complex reactivity of available S⁰ sources. Organic polysulfides, such as diallyl trisulfide (DATS) found in alliums including garlic, or other synthetic organic and inorganic polysulfides can act as sources of biologically available S⁰. These systems present divergent reactivity based on the polysulfide chain length and pendant alkyl group.¹² Inorganic polysulfides in particular are unstable in aqueous conditions and quickly equilibrate to different polysulfide mixtures.¹³ Despite these fundamental challenges, available S⁰ sources demonstrate significant promise in different systems, ranging from anticancer properties in several human cell lines^14–17 to enhanced antioxidative activity.¹⁸ In all of these cases, however, the production of byproducts obfuscates the role of S⁰.

An attractive approach to investigate the chemical biology of S⁰ is to use the most common and simplest form of sulfane sulfur: S₈. However, use of S₈ directly is hindered by its low solubility of 6.4 μg L⁻¹ (19(6) nM) at 25 °C in water¹⁸ and ∼4 μg L⁻¹ in seawater.¹⁹ Despite this low solubility, S⁰ found in volcanic deep-sea environments can be readily metabolized by thermophilic Archaean organisms found in these sulfur-rich habitats.⁵ As one example, species of the order Sulfolobales can derive energy from the metabolism of S⁰via both aerobic and anaerobic pathways.^20,21 Similarly, Acidianus ambivalens can utilize S⁰ as both an electron donor and acceptor. Taken together, such organisms may provide clues into possible mechanisms of stabilizing and activating bioavailable S⁰ in aqueous environments.

One potential strategy for S₈ solubilization and activation can be gleaned from the archaeon Staphylothermus marinus, a strict sulfur reducing anaerobe that requires S⁰ as its terminal electron acceptor. Found near hot deep-sea vents, S. marinus is coated in thermostable filamentous glycoprotein structure called tetrabrachion that protrude from its surface. The tetrabrachion of S. marinus is composed of a four-stranded parallel coiled-coil structure with a hydrophobic core that is particularly stable.²² The 24 kDa right-handed coiled-coil structure of the tetrabrachion contains hydrophobic cavities that have been found to encapsulate two S₈ molecules (PDB: 5JR5). Closer inspection of these S₈-binding cavities revealed that the sulfur motifs were held in place by van der Waals forces with aliphatic amino acid side chains leucine and isoleucine (Fig. 1).²³ Additional computational investigations have investigated the transit pathways for exchange of water and S₈ within these structures.²⁴ Further supporting this observation that hydrophobic motifs can increase S₈ solubility, Steudel and Holdt demonstrated that the solubility of S₈ in water can be increased using surfactants, with S₈ concentrations reaching up to 0.1 mM in saturated hexadecyl(trimethyl)ammonium bromide (CTAB) solutions, although reactivity studies were not reported.²⁵ In addition, the stability of sulfur nanoparticles and sulfur sols, which can be produced both industrially and from sulfur oxidizing bacteria,²⁶ can also be modulated by different types of surfactants.^27,28 Taken together, these prior observations support that this approach to solubilize S₈ in aqueous environments may be more general and could also lead to new approaches to enable chemical accessibility of S⁰-containing species in biological environments.

Fig. 1 Structure of the tetrameric right-handed coiled-coil component of the tetrabrachion of the archaeon S. marinus (PDB: 5JR5) demonstrating two hydrophobic pockets in the core capable of encapsulating S₈. Highlighted residues are colored according to their hydrophobicity, where red indicates stronger hydrophobicity.

Understanding the intrinsic strategies for stabilizing simple S⁰-containing sources in solution remains a key unmet need that could have significant impacts in broad fields ranging from contemporary chemical biology of reactive sulfur species to greener synthetic methods for sulfur-containing compounds. Here we report that hydrophobic interactions within cavity-containing molecules, such as cyclodextrins (CDs), can be used to significantly solubilize S₈ in aqueous solutions, and that this solubilized S⁰ is both chemically and biologically accessible. Specifically, we use 2-hydroxypropyl β-cyclodextrin (2HPβ) to generate 2HPβ/S₈ solutions that are stable and quantifiable, react with thiols to generate H₂S, exert antioxidant activities in cell models of oxidative stress, and increase intracellular S⁰ levels (Scheme 1).

Scheme 1 Solubilization and chemical accessibility of S₈ in 2-hydroxypropyl β-cyclodextrin (2HPβ).

Results and discussion

The simplest form of S⁰, S₈, is readily available in high purity as a sublimed chalky yellow solid. Unfortunately, its use in biological applications is severely hampered by its hydrophobicity and low water solubility. The solubility of S₈ has been calculated to be only 1.9(6) × 10⁻⁸ mol kg⁻¹ (or 1.5(2) × 10⁻⁷ M S⁰), which is multiple orders of magnitude below biologically relevant S⁰ concentrations.¹⁸ These results suggest that even small modifications to increase S₈ solubility in water can result in increased reactivity toward polysulfide formation. Moreover, efficient reduction of S₈ to H₂S will also require stabilization and solubilization of intermediate polysulfides or persulfide intermediates. Despite this low solubility, solid S₈ has been demonstrated to be biologically accessible by erythrocytes to produce H₂S,²⁹ which highlights the potential for increasing the bioavailability of S₈ in different systems. Moreover, these prior results suggest that biological pathways for sulfane sulfur activation from elemental sulfur may be accessible if S₈ could be solubilized in aqueous environments.

Motivated by the binding of S₈ to the hydrophobic pockets of the archaeon S. marinus and increased solubility in hydrophobic environments, we envisioned that water-soluble compounds with hydrophobic interiors could enable similar S₈ binding.³⁰ Building from this hypothesis, we utilized CDs, which are cyclic oligosaccharides that contain a hydrophobic interior and that are widely utilized to bind and solubilize hydrophobic compounds.³¹ This solubilization is due, in part, to the hydrophobic interior of CDs, which promotes encapsulation and binding of a non-polar guest, whereas the hydrophilic exterior enables water solubility. CDs are available in different sizes/volumes, and naturally produced CDs include α-CD, β-CD, and γ-CD, which contain 6, 7, and 8 glucose units, respectively. The choice of CD depends upon the size of the nonpolar compound to be solvated and the properties of the system being studied. Although natural CDs have limited water solubility, modification of the ring periphery with hydroxypropyl groups results in significant increases in solubility. In particular, 2-hydroxypropyl β-CD (2HPβ) and 2-hydroxypropyl γ-CD (2HPγ) have been used extensively, including in drug formulations to enable delivery of otherwise hydrophobic and insoluble compounds.³²

To test our general hypothesis, we treated aqueous solutions of 2HPβ with a 10-fold excess of solid S₈. We chose to start our investigations with the β-CD structure because the cavity volume (262 ų) is an ideal match for S₈ (∼149 ų; 57% cavity occupancy), based general preference for encapsulated guests to occupy ∼55% of host volume.³³ After stirring as solution of 2HPβ with S₈ in water for several days and subsequent filtration to remove residual insoluble S₈, we observed a strong absorbance at 263 nm in the UV-vis spectrum, which is a characteristic absorbance of S₈.³⁴ By contrast, stirring S₈ in water under identical conditions but in the absence of 2HPβ failed to produce a significant S₈ absorbance (Fig. 2). To test the stability of the solubilized 2HPβ/S₈, we next assessed whether the solution could be precipitated, filtered, and re-dissolved without loss of S₈. We precipitated the 2HPβ/S₈ complex with acetone, isolated the solid, and re-dissolved the resultant solid in buffer (Fig. S1†). In these experiments, we observed that the same absorbance from the original and re-dissolved solutions, confirming the stability of the solubilized system in both the liquid and solid state.

Fig. 2 Comparison of the UV-vis spectra of S₈ in water with S₈ in aqueous solutions containing 2HPβ. Conditions: 4 mg of S₈ in either 5 mL of PBS 7.4 PBS or 5 mL of 25% w/w 2HPβ in pH 7.4 PBS. Solutions were stirred for one day and then filtered prior to absorbance measurement.

To determine which components of the 2HPβ complex were responsible for S₈ solubilization, we next evaluated S₈ solvation in the presence of glucose and hydroxypropyl cellulose (HPC) as models for the sugar units of the 2HPβ macrocycle and the hydroxypropyl motif, respectively. After stirring an excess of S₈ to solutions of each saccharide in pH 7.4 phosphate buffered saline (PBS) buffer, the solutions were filtered, and UV-vis spectra were recorded (Fig. 3). As shown in Fig. 2, the characteristic absorbance at 263 nm corresponding to S₈ was significantly larger for 2HPβ (orange, 687 μM in this solution) than for glucose (aqua) or HPC (yellow). These data suggest that the cyclic structure and cavity of the CD are key components required for S₈ solvation. We next investigated the importance of the 2HPβ hydroxypropyl groups by testing S₈ solubilization with β-CD (lacking the 2-HP groups). We treated β-CD in pH 7.4 PBS buffer with excess S₈ and stirred for one month (Fig. S2†). After filtering the solution, we failed to observe any solubilized S₈ by UV-vis spectroscopy, which suggests that the hydroxypropyl groups are required for S₈ solubilization.

Fig. 3 UV-vis spectra of 640 mg S₈ in 10 mL pH 7.4 aqueous solutions containing 365 mg of 2HPβ, HPC, or glucose. 2HPβ solvates significantly more S₈ than either HPC or glucose.

To further investigate the requirement of the cyclic structure and presence of a cavity for S₈ solubilization, we also investigated whether 2-hydroxypropyl γ-CD (2HPγ; cavity volume: 427 ų) could solubilize S₈. We compared the S₈ solubilization in 25% w/w solutions of 2HPβ and 2HPγ in both pH. 7.4 PBS buffer and in water and observed significantly less S₈ solubilization from 2HPγ than from 2HPβ (Fig. S3†). Taken together, these experiments strongly support that both the cyclic/cavity-containing structure and the presence of the hydrophobic core and hydrophilic exterior are critical for the observed S₈ solvation by 2HPβ.

Building from these investigations, our next goal was to determine how much S₈ was solubilized in 2HPβ solutions. However, due to the low solubility of S₈ in water we used the extinction coefficient for S₈ in methanol, 6730 M⁻¹ cm⁻¹ at 263 nm, which has been reported previously.³⁴ To confirm that the extinction coefficient for S₈ in methanol could be used to measure S₈ concentrations in the 2HPβ/S₈ complex in water, we first prepared a solution of S₈ in MeOH at a known concentration (278 μM). We then used the known S₈/methanol extinction coefficient to measure the concentration of S₈ in an existing 2HPβ/S₈ solution. We then diluted this solution to match the calculated value of [S₈] in the methanol stock solution and compared the resulting absorbance traces. If the extinction coefficient remains constant between 2HPβ/S₈ and S₈ in methanol, then the calculated [S₈] for the 2HPβ/S₈ stock solution should be correct, and dilution from this value to the S₈ concentration in the methanol stock solution should yield an identical concentration and thus an identical absorbances.

The resultant curves (Fig. S4†) are similar to the 2HPβ/S₈ solution containing an S₈ concentration of 250 μM, which is ∼12% different between the methanol/S₈ and 2HPβ/S₈ solutions. This observation confirms that the extinction coefficient does not change appreciably between these two systems. On the basis of these comparisons, we have used the molar extinction coefficient of S₈ in MeOH to quantify S₈ in the 2HPβ. These quantified concentrations are further supported by the quantitative S₈ conversion to H₂S by thiols (vide infra), which provides additional support for the value of the reported S₈ concentrations in the 2HPβ system. Applying this extinction coefficient to the 50% w/w 2HPβ solution, provides an S₈ concentration of 2.0 ± 0.2 mM (16 mM S⁰) in water. When compared to the background solubility of S₈ in water, this constitutes ∼10⁵-fold enhancement in S₈ solubility.

Having established that S₈ is readily solubilized in aqueous solution of 2HPβ we next sought to investigate the stoichiometry and magnitude of the interaction between S₈ and 2HPβ. To probe these interactions, we monitored the observed [S₈] in solution as a function of increasing [2HPβ] (from 0–45% w/w) that were prepared with a 10-fold molar excess of solid S₈ in solution. Under these conditions, the activity of S₈ in solution remains constant, and increasing the [2HPβ] should result in a concomitant increase in [S₈] in solution. After stirring each solution for several days to ensure equilibrium, the solutions were filtered, and the S₈ concentration were measured by UV-vis spectrophotometry (Fig. 4a). As expected, the measured [S₈] increased linearly with increasing [2HPβ], which further supports a direct interaction between S₈ and 2HPβ. The above constant activity data can be used to determine the binding stoichiometry and affinity between 2HPβ and S₈ through eqn (1).³⁵ Under these conditions, the total S₈ content is defined as S_t in eqn (1), and the concentration of unbound S₈, held constant throughout experiments to ensure constant activity, is defined as s₀ (1.9(6) × 10⁻⁸ mol kg⁻¹).¹⁸ The constant n represents the binding stoichiometry.

Fig. 4 (a) UV-vis spectra of increasing concentrations of 2HPβ with a 10-fold excess of S₈. (b) Log–log plot of quantified [S₈] as a function of increasing [2HPβ]. Trials were performed in triplicate.

Generating a log–log plot with the parameters of eqn (1) demonstrates that the relationship between 2HPβ and S₈ is consistent with 1 : 1 binding (Fig. 4b). Upon performing linear regression analysis, we obtained a K_a value of 3.4 ± 0.05 × 10⁵ M⁻¹ for the 2HPβ/S₈ complex, which is higher than the typical 10²–10³ M⁻¹ binding affinities observed in βCD systems.³⁶ Although the above analyses are supportive of a 1 : 1 binding stoichiometry, we cannot definitively exclude equimolar higher order interactions from our data.

We repeated these binding stoichiometry experiments with 2HPγ and 2-hydroxypropyl α-CD (2HPα) (Fig. S5†). Our expectation was that these differently-sized CD hosts would not solubilize or bind S₈ as efficiently as 2HPβ For example, if S₈ were bound within 2HPα or 2HPγ, cavity occupancies of 86% and 32% would be observed, respectively, which is outside of the range of most host–guest interactions.³⁰ Consistent with these expectations, 2HPγ solvated less significantly less S₈ and exhibited less linear binding behavior when compared to 2HPβ (Fig. S5a†). Similarly, 2HPα solubilized very little S₈ with significant deviations from a well-defined binding relationship (Fig. S5c and d†). These experiments further support that the size complementarity between S₈ and 2HPβ is an important factor in solubilization.

Building from our data supporting S₈ solubilization with 2HPβ we next sought to determine whether the solubilized S₈ is chemically accessible. To investigate this question, we first determined whether the reductant tris(2-carboxyethyl)phosphine (TCEP), which is a commonly-used and biologically compatible reductant used to reduce disulfides and other sulfane sulfur species, could access the solubilized S₈ and generate the characteristic P=S and P=O products upon phosphine-mediated reduction and subsequent hydrolysis. Conveniently, this conversion can readily be monitored qualitatively by ³¹P NMR spectroscopy with TCEP (δ = 15.2–15.8 ppm), the resultant TCEP sulfide (δ = 51.5 ppm), and the associated TCEP oxide (δ = 53.0 ppm), all having characteristic NMR resonances. Prior to TCEP addition, the ³¹P NMR spectrum of S₈ in a 25% w/w 2HPβ solution in pH 7.4 buffer only shows a peak at δ = 0 ppm from the PBS (Fig. 5a). After TCEP addition and incubation overnight at room temperature, however, the ³¹P NMR spectrum showed peaks corresponding to unreacted TCEP, as well as the phosphine sulfide and oxide peaks at δ = 53.0 and 51.5 ppm, respectively. To confirm that the peaks were from oxidized TCEP products, we repeated the TCEP incubation with TCEP and S₈ in MeOD (Fig. S6a†) and K₂S₅ in D₂O (Fig. S6b†). In both cases, we observed the same peaks in the ³¹P NMR spectrum at δ = 53.0 and 51.5 ppm, confirming product formation. In the absence of the 2HPβ/S₈, only the TCEP peak is observed, which confirms that TCEP oxide is not formed from adventitious oxidation. Similarly, addition of TCEP to a solution of 2HPβ did not generate TCEP oxide (Fig. 5b). Taken together, these results demonstrate that the solubilized S₈ in the 2HPβ/S₈ solution is chemically accessible and can react directly with reductants.

Fig. 5 ³¹P (¹H} NMR spectra in 2HPβ/S₈ or 2HPβ solutions incubated with TCEP. (a) 25% w/w solution of 2HPβ with 500 mg S₈ before (top) and after (bottom) addition of 10 mg TCEP with overnight incubation. (b) 25% w/w 2HPβ without sulfur in PBS before (top) and after (bottom) TCEP addition and incubation; no oxidized product peaks are observed.

To further the potential biological relevance of the solubilized S₈ in the 2HPβ/S₈ complex, we next determined whether the solubilized S₈ could be reduced by thiols to release H₂S. One important feature of S⁰ that contributes to its role in biology is its ability to release H₂S upon reduction by thiols. Polysulfides, such as DATS, are well established to release H₂S after reaction with thiols and are used broadly as exogenous sources of S⁰. Although prior studies have investigated how different functional groups on polysulfide motifs impacts H₂S release, one limitation of these systems is that all of these compounds also generate organic byproducts upon H₂S release.^12,37 As expected, a higher S⁰ content in organic polysulfides leads to greater H₂S release, although tetrasulfides appear to be the largest synthetically-accessible and consistently-stable polysulfides. Using a similar logic of trying to maximize the S⁰ content per donor motif, we envisioned that the solubilized 2HPβ/S₈ complex could also function as an entirely new approach to deliver S⁰ and/or H₂S. Importantly, since S₈ is comprised entirely of S⁰ it should be an effective donor, with the only byproduct being 2HPβ.

To determine whether the S₈ solubilized in the 2HPβ/S₈ system is accessible to thiols, and to quantify resulting H₂S release, we treated a 2HPβ/S₈ solution (25 μM S₈, 200 μM S⁰ in 50% w/w 2HPβ) with 1 mM (5 equiv. with respect to S⁰ atoms) of cysteine or reduced glutathione (GSH) under air-free conditions in pH 7.4 PBS. We measured H₂S release at different time points during the reaction by using the colorimetric methylene blue assay, which measures H₂S production by the formation of the methylene blue dye (Fig. 6). Calculated efficiency values assume all S⁰ atoms can react to form H₂S. After 45 minutes, we observed 160 ± 5 μM H₂S release (80% efficiency) from the 2HPβ/S₈ in the presence of cysteine (green). Under identical conditions, treatment of the complex with GSH (red), the most abundant biological thiol, yielded 220 ± 7 μM H₂S release after 45 minutes, corresponding to stoichiometric reduction of each S⁰ atom. In the absence of S₈ with only 2HPβ and 500 μM cysteine, no H₂S was observed from the methylene blue assay (yellow), which confirms that 2HPβ and thiols alone do not spontaneously generate H₂S or result in methylene blue formation. Similarly, in the absence of 2HPβ we did not observe any H₂S release from S₈ (50 μM if fully soluble) and cysteine (500 μM, blue), confirming the importance of 2HPβ to the accessibility of S₈. As a whole, these results show that S₈ is made chemically accessible in water by solubilization with 2HPβ, and that this sulfur can be reduced to H₂S with biologically relevant thiols. Because this conversion requires the intermediate generation of persulfides and/or polysulfides en-route to H₂S release, these results may also suggest that the 2HPβ facilitates the solubilization of these reactive intermediates. Alternatively, this solubilizing environment may also help stabilize other forms of elemental sulfur, such as biosulfur or sulfur sols, which could facilitate further reactivity with sulfhydryl-containing nucleophiles.

Fig. 6 Release of H₂S from S₈ solvated in 2HPβ in the presence of cysteine (green) or GSH (red). 2HPβ alone (yellow) does not release H₂S in the presence of cysteine, and the amount of S₈ in solution at 50 μM without 2HPβ (blue) is not high enough for appreciable release. All data points were collected in triplicate, and the error determined via standard deviation.

In addition to the above experiments, we also determined whether S₈ needed to be pre-solubilized with 2HPβ prior to reaction with thiols, or whether 2HPβ could act as a catalyst for S₈ conversion to H₂S by thiols in water. To answer this question, we added 224 mg S₈ solid and 500 μM cysteine to 180 mg 2HPβ in 40 mL pH 7.4 PBS buffer and monitored H₂S generation using the methylene blue method. Under these conditions, we observed a faster peaking time of 15 minutes, but also a slightly lower overall efficiency of 147 μM H₂S release (74% efficient) (Fig. S6†). These data indicate that S₈ does not need to be pre-solvated to the 2HPβ complex prior to thiol addition in order to facilitate reaction with thiols and subsequent H₂S release. These data may also support the role of 2HPβ as a phase transfer catalyst in these environments and that the rate of thiol-mediated reduction is faster than the rate of S₈ encapsulation. Expanding from the present system, these results suggest that hydrophobic motifs in more complex systems may enable chemical accessibility of transiently formed S₈ from different redox processes.

The accessibility of the solvated S₈ to biological thiols prompted us to investigate whether the solvated S₈ could be taken up into cells. Increasing intracellular levels of S⁰ induces a cytoprotective effect by reducing oxidative stress, making direct S⁰ donation desirable.³⁸ Furthermore, the thiol-mediated reduction of S₈ to H₂S should proceed through persulfide and polysulfide formation, both of which should increase the levels of reactive sulfur species in cells. To evaluate cellular S₈ uptake, we treated HeLa cells with either 10 μM S₈ as 2HPβ/S₈ or an equivalent amount of 2HPβ alone for 24 hours. We then treated cells with the sulfane-sulfur selective fluorescent probe SSP4.³⁹ We observed that cells treated with 2HPβ/S₈ showed a significant increase in fluorescence when compared with those treated with 2HPβ alone, which is consistent with the bioavailability of the solubilized S₈ (Fig. 7).⁴⁰ As a positive control, we repeated these experiments with HeLa cells that were treated with the inorganic polysulfide K₂S₅ as a source of S⁰, which also showed a significant SSP4 fluorescence response. These results support that the 2HPβ/S₈ system can cause significant increases in intracellular S⁰ levels, though the efficiency of this uptake is not yet known.

Fig. 7 Fluorescent images of HeLa cells treated with 2HPβ alone (top), 10 μM S₈ from a 2HPβ/S₈ solution (middle), or the inorganic polysulfide K₂S₅ (bottom) and imaged with the SSP4 probe for S⁰. Scale bar = 50 μm.

Finally, we sought to determine whether the bioavailable solubilized S₈ could be used to access protective effects associated with S⁰/H₂S. Both H₂S and S⁰ species play important antioxidant and anti-inflammatory roles throughout biology and are effective reducing agents able to neutralize damaging oxidants and free radical species.^41,42 Polysulfides and persulfides containing S⁰ have demonstrated antioxidative properties greater than those attributed to H₂S or thiols alone.⁴³ The well-studied antioxidant N-acetyl cysteine has also been shown to enhance the production of S⁰.³⁸ As a whole, a common theme is that polysulfides, and their role as both H₂S and persulfide donor motifs, facilitates their protection against oxidative stress. The increased intracellular S₈ should further produce a diverse range of polysulfides in the cells. Building from this observation, we reasoned that the high S⁰ content of the 2HPβ/S₈ system should therefore make it an effective antioxidant in a cellular environment. With this in mind, we sought to determine whether the 2HPβ/S₈ system provided antioxidant potential in cellular environments.

We used the colorimetric Griess reagent to track relative levels of NO_x metabolites in RAW 264.7 macrophage cells pretreated with either 2HPβ/S₈ or 2HPβ, and then lipopolysaccharide (LPS). In the presence of proinflammatory cytokines such as LPS, RAW 264.7 cells produce NO from inducible nitric oxide synthase (iNOS).^44,45 When formed, NO is rapidly oxidized to downstream NO_x species, and NO₂⁻ can be quantified directly using the colorimetric Griess assay. Importantly, H₂S-releasing compounds have been previously demonstrated to significantly decrease NO₂⁻ formation in RAW 264.7 cells.^46,47 To investigate the activity of 2HPβ/S₈ in this system, we plated RAW 264.7 macrophage cells on 24 well plates and the following day treated with the delivery vehicle for 2HPβ/S₈ (PBS), different concentrations of 2HPβ/S₈, or equivalent concentrations of 2HPβ alone for 24 hours. The cells were then washed and treated with 1 μg mL⁻¹ LPS for another 24 hours. After incubation, the media was collected from each well and the amount of NO₂⁻ was quantified using the Griess assay. The absorbance values of each treatment condition were normalized to the vehicle. As shown in Fig. 8, addition of the 2HPβ/S₈ complex results in a significant and decrease in NO₂⁻ formation across a range of concentrations, as evidenced by the decrease of absorbance of the Griess product. These data are consistent with H₂S and S⁰ release. The 2HPβ alone also results in a reduction in NO₂⁻ levels, but to a much lower extent than the 2HPβ/S₈ system. It is possible that H₂S, polysulfides, and sulfane sulfur species generated from the solubilized S₈ all play roles in the reduction of NO₂⁻ levels in this assay, which when taken together supports prior work in the field demonstrating that both H₂S and sulfane sulfur provide protective effects toward models of oxidative stress.

Fig. 8 Relative levels of NO₂⁻ in RAW 264.7 macrophage cells treated with different concentrations of 2HPβ/S₈ (orange) or equivalent concentrations of 2HPβ alone (blue). Data is normalized to the vehicle (PBS) treatment. Trials were performed in quadruplicate, and error determined by standard deviation.

Conclusions

We have demonstrated that hydrophobic cyclodextrins can facilitate solubilization and chemical activity of S₈ in water. In addition to providing a new and significant approach to delivering S⁰/sulfane sulfur to aqueous and biological environments, these results provide fundamentally new insights that impact S₈ bioavailability. Building from the solubilization of S₈ by the 2HPβ system, these results may suggest that pools of oxidized sulfur can stably exist in biological hydrophobic structures such as proteins. Moreover, the demonstration that the 2HPβ system can effectively catalyze S₈ reduction to H₂S by thiols in water also highlights this approach as a method to limit S₈ accumulation. We anticipate that this and related systems currently under investigation in our lab will not only find utility as H₂S and sulfane sulfur delivery systems, but also in expanding investigations into how S⁰ is managed in more complex biological systems.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the NSF (CHE-1454747), the Dreyfus Foundation, and the NIH (T32 GM007759, to SGB). Microscopy instrumentation was supported by the NSF (CHE-1531189).

Notes and references

1. K. Zahnle, L. Schaefer and B. Fegley, Cold Spring Harbor Perspect. Biol., 2010, 2, a004895.
2. S. Ranjan, Z. R. Todd, J. D. Sutherland and D. D. Sasselov, Astrobiology, 2018, 18, 1023–1040.
3. E. T. Parker, H. J. Cleaves, J. P. Dworkin, D. P. Glavin, M. Callahan, A. Aubrey, A. Lazcano and J. L. Bada, Proc. Natl. Acad. Sci. U. S. A., 2011, 108, 5526–5531.
4. W. Heinen and A. M. Lauwers, Origins Life Evol. Biospheres, 1996, 26, 131–150.
5. M. Vandiver and S. H. Snyder, J. Mol. Med., 2012, 90, 255–263.
6. R. Wang, Antioxid. Redox Signaling, 2010, 12, 1061–1064.
7. A. K. Mustafa, M. M. Gadalla and S. H. Snyder, Sci. Signaling, 2009, 2, re2.
8. N. Lau and M. D. Pluth, Curr. Opin. Chem. Biol., 2019, 49, 1–8.
9. H. Kimura, Antioxid. Redox Signaling, 2015, 22, 347–349.
10. K. Avetisyan, T. Buchshtav and A. Kamyshny, Geochim. Cosmochim. Acta, 2019, 247, 96–105.
11. W. E. Kleinjan, A. de Keizer and A. J. H. Janssen, Colloids Surf., B, 2005, 43, 228–237.
12. S. G. Bolton, M. M. Cerda, A. K. Gilbert and M. D. Pluth, Free Radicals Biol. Med., 2019, 131, 393–398.
13. H. Liu, M. N. Radford, C. T. Yang, W. Chen and M. Xian, Br. J. Pharmacol., 2019, 176, 616–627.
14. T. Hosono, T. Fukao, J. Ogihara, Y. Ito, H. Shiba, T. Seki and T. Ariga, J. Biol. Chem., 2005, 280, 41487–41493.
15. W. Bat-Chen, T. Golan, I. Peri, Z. Ludmer and B. Schwartz, Nutr. Cancer, 2010, 62, 947–957.
16. D. Xiao and S. V. Singh, Carcinogenesis, 2006, 27, 533–540.
17. M. Murai, T. Inoue, M. Suzuki-Karasaki, T. Ochiai, C. Ra, S. Nishida and Y. Suzuki-Karasaki, Int. J. Oncol., 2012, 41, 2029–2037.
18. J. Boulegue, Phosphorus Sulfur, 2006, 5, 127–128.
19. A. Kamyshny, Geochim. Cosmochim. Acta, 2009, 73, 6022–6028.
20. Y. Liu, L. L. Beer and W. B. Whitman, Environ. Microbiol., 2012, 14, 2632–2644.
21. A. Kletzin, T. Urich, F. Muller, T. M. Bandeiras and C. M. Gomes, J. Bioenerg. Biomembr., 2004, 36, 77–91.
22. J. Peters, W. Baumeister and A. Lupas, J. Mol. Biol., 1996, 257, 1031–1041.
23. M. McDougall, O. Francisco, C. Harder-Viddal, R. Roshko, M. Meier and J. Stetefeld, Proteins, 2017, 85, 2209–2216.
24. C. Harder-Viddal, M. McDougall, R. M. Roshko and J. Stetefeld, Comput. Struct. Biotechnol. J., 2019, 17, 675–683.
25. R. Steudel and G. Holdt, Angew. Chem., Int. Ed., 1988, 27, 1358–1359.
26. R. Steudel, Elemental Sulfur and Sulfur-Rich Compounds I, 2003, 230, 153–166.
27. R. G. Chaudhuri and S. Paria, J. Colloid Interface Sci., 2011, 354, 563–569.
28. A. A. Garcia and G. K. Druschel, Geochem. Trans., 2014, 15, 11.
29. D. G. Searcy and S. H. Lee, J. Exp. Zool., 1998, 282, 310–322.
30. We also note that the 2008 Russian patent RU2321598C1 hypothesizes S₈ binding to cyclodextrin.
31. J. Szejtli, Chem. Rev., 1998, 98, 1743–1754.
32. T. Loftsson, P. Jarho, M. Masson and T. Jarvinen, Expert Opin. Drug Delivery, 2005, 2, 335–351.
33. S. Mecozzi and J. Rebek, Chem. – Eur. J., 1998, 4, 1016–1022.
34. R. Steudel, D. Jensen, P. Gobel and P. Hugo, Ber. Bunsen-Ges. Phys. Chem., 1988, 92, 118–122.
35. K. A. Conners, Binding Constants, The Measurement of Molecular Complex Stability, Wiley, New York, 1987.
36. K. A. Connors, J. Pharm. Sci., 1995, 84, 843–848.
37. M. M. Cerda, M. D. Hammers, M. S. Earp, L. N. Zakharov and M. D. Pluth, Org. Lett., 2017, 19, 2314–2317.
38. D. Ezerina, Y. Takano, K. Hanaoka, Y. Urano and T. P. Dick, Cell Chem. Biol., 2018, 25, 447–459.
39. W. Chen, C. R. Liu, B. Peng, Y. Zhao, A. Pacheco and M. Xian, Chem. Sci., 2013, 4, 2892–2896.
40. The S⁰ appears to localizes particularly in the nucleus and nucleolus of treated cells. This observation is consistent with previous research indicating the production and accumulation of S⁰ in the nucleolus of cells. See: K. R. Olson, Y. Gao, A. K. Steiger, M. D. Pluth, C. R. Tessier, T. A. Markel, D. Boone, R. V. Stahelin, I. Batinic-Haberle and K. D. Straubg, Molecules, 2020, 25, 980.
41. M. Magierowski, K. Magierowska, M. Hubalewska-Mazgaj, M. Surmiak, Z. Sliwowski, M. Wierdak, S. Kwiecien, A. Chmura and T. Brzozowski, Biochem. Pharmacol., 2018, 149, 131–142.
42. L. A. Nicolau, R. O. Silva, S. R. Damasceno, N. S. Carvalho, N. R. Costa, K. S. Aragao, A. L. Barbosa, P. M. Soares, M. H. Souza and J. V. Medeiros, Braz. J. Med. Biol. Res., 2013, 46, 708–714.
43. T. Ida, T. Sawa, H. Ihara, Y. Tsuchiya, Y. Watanabe, Y. Kumagai, M. Suematsu, H. Motohashi, S. Fujii, T. Matsunaga, M. Yamamoto, K. Ono, N. O. Devarie-Baez, M. Xian, J. M. Fukuto and T. Akaike, Proc. Natl. Acad. Sci. U. S. A., 2014, 111, 7606–7611.
44. C. Coletta, A. Papapetropoulos, K. Erdelyi, G. Olah, K. Modis, P. Panopoulos, A. Asimakopoulou, D. Gero, I. Sharina, E. Martin and C. Szabo, Proc. Natl. Acad. Sci. U. S. A., 2012, 109, 9161–9166.
45. F. Aktan, Life Sci., 2004, 75, 639–653.
46. M. Whiteman, L. Li, P. Rose, C. H. Tan, D. B. Parkinson and P. K. Moore, Antioxid. Redox Signaling, 2010, 12, 1147–1154.
47. Y. Zhao, M. M. Cerda and M. D. Pluth, Chem. Sci., 2019, 10, 1873–1878.

---

Footnote

† Electronic supplementary information (ESI) available: Experimental details, UV-vis spectra, NMR spectra, H₂S measurements. See DOI: 10.1039/d0sc04137h

---