Inorganic hydrogen polysulfides: chemistry, chemical biology and detection

Abstract

Recent studies suggest that inorganic hydrogen polysulfides (H₂S_n, n ≥ 2) play important regulatory roles in redox biology. Modulation of their cellular levels could have potential therapeutic value. This review article focuses on our current understanding of the biosynthesis, biofunctions, fundamental physical/chemical properties, detection methods and delivery techniques of H₂S_n.

Linked Articles

This article is part of a themed section on Chemical Biology of Reactive Sulfur Species. To view the other articles in this section visit http://onlinelibrary.wiley.com/doi/10.1111/bph.v176.4/issuetoc

Abbreviations

3MP

3‐mercaptopyruvate

3MST

3‐mercaptopyruvate sulfurtransferase

BODIPY

boron difluoride dipyrromethene

CARS

cysteinyl‐tRNA synthetase

d‐PET

donor‐excited photoinduced electron transfer

GSSH

glutathione persulfide

H₂S

hydrogen sulfide

H₂S_n

hydrogen polysulfides

HNO

nitroxyl

HSNO

nitrososulfide

ICT

intramolecular charge transfer

Keap1

Kelch ECH associating protein 1

mBB

mono‐bromobimane

MPO

myeloperoxidase

NIR

near‐IR

Nrf2

nuclear factor erythroid 2‐related factor 2

RSS

reactive sulfur species

TP

two‐photon

Introduction

Reactive sulfur species (RSS) are a family of sulfur‐containing molecules found in biological systems and are involved in a variety of biological processes (Giles et al., 2001; Giles and Jacob, 2002; Gruhlke and Slusarenko, 2012; Paulsen and Carroll, 2013; Yang et al., 2017). Representative RSS include thiols, hydrogen sulfide, persulfides, polysulfides and S‐modified cysteine derivatives such as S‐nitrosothiols (SNO) and sulfenic acids (SOH). Among these molecules, hydrogen sulfide (H₂S) has been most well‐studied as this gas molecule has recently been classified as a critical cell signalling molecule, much like nitric oxide (NO) (Li et al., 2011; Wang, 2012; Módis et al., 2014; Polhemus and Lefer, 2014; Szabo et al., 2014). For example, H₂S functions as an endothelial cell‐derived relaxing factor via direct activation of ATP‐sensitive potassium (K_ATP) channels. Deprivation of endogenous production of H₂S contributes to the development of hypertension. Moreover, H₂S has shown beneficial effects on oxidative stress, inflammation and fibrosis. While research on H₂S is still actively ongoing, a hot new topic in this field has emerged that focuses on a series of H₂S‐related reactive sulfane sulfur species (Ono et al., 2014; Toohey and Cooper, 2014; Kimura, 2015; Park et al., 2015; Akaike et al., 2017). Sulfane sulfur refers to sulfur atoms with six valence electrons but no charge (represented as S⁰). H₂S‐related sulfane sulfur compounds include persulfides (R‐S‐SH), polysulfides (R‐S_n‐SH or R‐S‐S_n‐S‐R), inorganic hydrogen polysulfides (H₂S_n, n ≥ 2) and protein‐bound elemental sulfur (S₈). Among these, the hydrogen polysulfides H₂S_n are especially attractive. Recent studies have suggested that H₂S_n exist endogenously and are linked to a number of physiological and pathological processes. H₂S_n can serve as the source of H₂S, as well as the sink of H₂S. Much of what is known as H₂S signalling may be actually due to H₂S_n. In this review, we summarize current knowledge about H₂S_n, focusing on their formation, functions, fundamental physical/chemical properties, detection methods, as well as releasing strategies.

Biosynthesis of H₂S_n

Currently, how H₂S_n are produced endogenously is still a topic under active investigation. The studies by Kimura et al. suggest that H₂S_n are mainly produced from 3‐mercaptopyruvate (3MP) by 3‐mercaptopyruvate sulfurtransferase (3MST) (Kimura et al., 2015). In their studies, the produced H₂S_n were trapped by mono‐bromobimane (mBB) and further identified by LC‐FL and LC‐MS/MS analysis. Based on the identity and ratio of mBB‐trapped products, it was found that the major H₂S_n species was H₂S₃, while H₂S₂ and H₂S₅ were minor products. Kimura et al. also found that H₂S_n were localized in the cytosol of cells, and H₂S_n could be produced from H₂S by 3MST and rhodanese. The basal endogenous H₂S₃ concentration was 3.4 nmol·g⁻¹ protein in the brain, which is comparable to the concentration of H₂S (4.8 nmol·g⁻¹ protein). Kimura et al. later found that 3MST also produces cysteine‐ and glutathione‐persulfides (Cys‐SSH and GSSH) (Kimura et al., 2017). Based on these results, two possible mechanisms describing 3MST‐mediated H₂S_n and persulfide biosynthesis are shown in Scheme 1: (A) 3MST abstracts the sulfur from its substrate 3MP to form a persulfidated enzyme intermediate (3MST‐S_nSH), which then degrades to form H₂S_n. Next, H₂S_n transfer the sulfane sulfur atom to other thiols (Cys, GSH or protein‐SH) to form persulfide species. (B) After the persulfidated enzyme intermediate (3MST‐S_nSH) is formed, it directly transfers the sulfane sulfur atom to other thiols or H₂S to form the corresponding persulfide species, including H₂S_n. It should be noted that in these studies, H₂S_n species were identified by mBB capturing experiments. An assumption in these conclusions is that all H₂S_n species as well as other thiol species like CysSSH are captured by mBB with high effectiveness and the capturing rates are faster than the dynamic interchanges within these sulfur species. This assumption should be further evaluated before the conclusions can be well‐accepted.

Scheme 1.

3MST‐mediated H₂S_n/persulfide biosynthetic pathways.

Very recently, Akaike et al. reported another interesting enzymic pathway that can contribute to the biosynthesis of H₂S_n (Akaike et al., 2017). They found that prokaryotic and mammalian cysteinyl‐tRNA synthetases (CARS) can very effectively catalyse the production of cysteine persulfide (CysSSH) and polysulfides (CysS_nSH) using cysteine as the substrate (Scheme 1C). CARS was also found to integrate cysteine polysulfides into proteins during translational process. As one can imagine, CysSSH and CysS_nSH are valuable precursors of H₂S_n. Therefore, H₂S_n biosynthesis is likely to be controlled by this CARS‐mediated pathway. Overall, the mechanisms summarized in Scheme 1 indicate that Cys‐SSH/GSSH and H₂S_n are normally formed together.

The crosstalk between H₂S and NO is interesting, and a synergistic effect of these two signalling molecules can be expected. Several groups have studied the direct reactions between H₂S and NO and found the chemistry is rather complicated (Eberhardt et al., 2014; Cortese‐Krott et al., 2015; Miyamoto et al., 2017). In these studies, a mixture of H₂S (using Na₂S as the equivalent) and NO (using a NO donor such as diethylamine NONOate or S‐nitroso‐N‐acetyl‐D,L‐penicillamine) was generated, and the products were trapped and analysed by spectroscopy methods. Indeed, H₂S_n are found to be the products, as well as nitroxyl (HNO), nitrososulfide (HSNO) and even SSNO⁻. The formation of these products can be possibly attributed to the following equations:

$${\mathrm{HS}}^{-}\to {\mathrm{HS}}^{•}+{\mathrm{e}}^{-}$$

$${\mathrm{HS}}^{•}+{\mathrm{NO}}^{•}\to \text{HSNO}$$

$$\text{HSNO}+{\mathrm{H}}_2\mathrm{S}\to \text{HSSH}+\mathrm{HNO}$$

$$\text{HSSH}+\text{HSNO}\to \text{HSSNO}+{\mathrm{H}}_2\mathrm{S}$$

The fast reaction between H₂S and NO to form H₂S_n may have important biological implications. For example, H₂S producing enzymes (3MST and cystathionine β‐synthase) and NO synthase (NOS) are localized to neurons and astrocytes in the CNS. Therefore, endogenously generated H₂S and NO can react with each other to produce H₂S_n, which then activate TRPA1 channels to modify synaptic activity. In the cardiovascular system, eNOS and cystathionine γ‐lyase are localized to vascular endothelium and smooth muscle respectively. NO and H₂S generated from these enzymes may interact and produce H₂S_n to activate protein kinase G (PKG)1α to induce vascular relaxation.

H₂S_n can also be generated from H₂S catalysed by haem proteins. For example, Olson et al. showed that cytosolic copper/zinc SOD catalysed the oxidation of H₂S to produce H₂S₂, as well as H₂S₃ and H₂S₅ (Olson et al., 2018). O₂ or H₂O₂ was used as the electron acceptor in the reaction. Interestingly, SOD‐catalysed H₂S oxidation is found to be inhibited by high H₂S concentration (>1 mM), and the reaction appears to be specific for dissolved H₂S (not the hydrosulfide anion HS⁻). Nagy et al. recently studied the detailed mechanisms of myeloperoxidase (MPO)‐catalysed H₂S oxidation (Garai et al., 2017) and found the Compound III state is formed in the reactions of H₂S with MPO in the presence of oxygen. This enzymic reaction provides a slow flux of sulfane sulfur species generation (which likely involves H₂S_n) that may be important in endogenous signalling.

In addition to their biosynthesis, the degradation pathways of H₂S_n are also important, but these are much less studied. Recently, Nagy and co‐workers investigated the fates of polysulfides under reducing environments (Dóka et al., 2016). They studied how H₂S exposure affects the activity of thioredoxin reductase‐1 (TrxR1). Instead of enzyme inhibition, TrxR1 was found to reduce H₂S_n in a polysulfide concentration‐dependent manner in the presence of NADPH. Polysulfides appear to be good substrates and not inhibitors of TrxR1. Moreover, the GSH system (with NADPH, GSH and glutathione reductase) was also found to catalyse the reduction of polysulfides in a concentration‐dependent manner. As can be expected, the Trx and GSH systems are critical in maintaining sulfane sulfur homeostasis and sulfide signalling.

Biological functions of H₂S_n

The oxidation state of the sulfane sulfur in H₂S_n is zero. H₂S_n can readily react with protein cysteine residues to form protein persulfides (P‐S‐SH). This reaction, named as S‐persulfidation (also known as S‐perthiolation or S‐sulfhydration), is believed to be the major contributor of the biological functions of H₂S_n. For example, H₂S_n showed strong protective effects against oxidative damage caused by excessive ROS. This is due to H₂S_n promoting the release of nuclear factor erythroid 2‐related factor 2 (Nrf2) and induces the translocation of Nrf2 into the nucleus by persulfidating its binding partner Kelch‐like ECH‐associated protein 1 (Keap1), subsequently causing an increase in intracellular GSH levels and the expression of HO‐1 (a Nrf2‐regulatory gene) (Koike et al., 2013). H₂S_n were also found to regulate the activity of the tumour suppressor protein, lipid phosphatase and tensin homolog (PTEN), by introducing the sulfane sulfur into the active site cysteine of PTEN (Greiner et al., 2013). The study by Mutus and co‐workers showed that H₂S_n could suppress the activity of glyceraldehyde 3‐phosphate dehydrogenase (GAPDH) via persulfidation on Cys 152 (Jarosz et al., 2015). Additionally, H₂S_n were found to induce Ca²⁺ influx in astrocytes and stimulate mouse sensory neurons by activating TRPA1 channels. The molecular mechanism is due to persulfidation of two cysteine residues at the amino terminus of the TRPA1 channels (Yukari et al., 2015). Eaton et al. found that in the presence of oxidants, such as H₂O₂ or O₂, H₂S can be converted into H₂S_n very quickly and afterwards activate PKG1α by thiol‐disulfide exchange reactions. The PKG disulfide formation could then effectively relax vascular smooth muscle (Stubbert et al., 2014). Finally, the recent discovery by Akaike et al. about CARS‐mediated persulfide formation is very interesting (Akaike et al., 2017). Polysulfides generated by CARS (including CysSSH and H₂S_n) contribute significantly to polysulfidation on proteins (via both post‐translational and co‐translational processes). This new pathway plays an important role in the mitochondrial electron transport chain, suggesting a new role of polysulfides/persulfides in maintaining mitochondrial bioenergetics. This unique sulfur respiration is expected to be linked to oxidative stress related diseases.

Fundamental physical/chemical properties of H₂S_n

There are several chemical methods to make mixtures of H₂S_n or relatively pure H₂S₂, H₂S₃ and H₂S₄ (Steudel, 2003). For example, liquid sulfur and H₂S can react to give a mixture of long‐chain H₂S_n (Equation (1)). As H₂S and elemental sulfur occur together in hot underground deposits of natural gas, the formation of H₂S_n under these high‐pressure conditions is very likely.

$${\mathrm{H}}_2\mathrm{S}+{\mathrm{nS}}_{\mathrm{liq}}\to {\mathrm{H}}_2{\mathrm{S}}_{\mathrm{n}+1}$$ (1)

On the other hand, H₂S_n are unstable species and tend to decompose to elemental sulfur (S⁰ or S₈) and H₂S (Equation (2)). In the natural gas field, when the gas is produced and the pressure and temperature are lowered, the decomposition reaction takes place. The precipitated elemental sulfur solid may cause serious problems when it clogs pipelines and valves. H₂S_n decomposition can be catalysed by substances such as alkali, nucleophiles and metals.

$${\mathrm{H}}_2{\mathrm{S}}_{\mathrm{n}}\to {\mathrm{H}}_2\mathrm{S}+{}^{\left(\mathrm{n}-1\right)}/{}_8{\mathrm{S}}_8$$ (2)

Thermolysis and photolysis of H₂S_n are believed to be radical chain reactions (Muller and Hyne, 1969; Gosavi et al., 1973). Take H₂S₂, for example, the first step is assumed to be the homolytic cleavage of the S–S bond to form HS^• (Equation (3)), which then abstracts hydrogen from H₂S₂ to form H₂S and HS₂ ^• (Equation (4)). Recombination of HS^• and HS₂ ^• should form H₂S and elemental sulfur S₂ (Equation (5)). The reaction progress and intermediates can be detected by UV, IR or luminescence spectroscopy. Thiyl radicals generated in this process may also react with other molecules (Dénès et al., 2014) such as NO to form SNO adducts (Madej et al., 2008). This further complicates the signalling mechanisms.

$${\mathrm{H}}_2{\mathrm{S}}_2\to {\text{2HS}}^{•}$$ (3)

$${\mathrm{HS}}^{•}+{\mathrm{H}}_2{\mathrm{S}}_2\to {\mathrm{H}}_2\mathrm{S}+{{\mathrm{HS}}_2}^{•}$$ (4)

$${\mathrm{HS}}^{•}+{{\mathrm{HS}}_2}^{•}\to {\mathrm{H}}_2\mathrm{S}+{\mathrm{S}}_2$$ (5)

The pure forms of H₂S_n (n = 2–8) are found to be yellow liquids at 20°C. The intensity of their colour increases with the sulfur content. The freezing points of H₂S₂, H₂S₃ and H₂S₄ are −90°C, −53°C and −85°C, respectively; while their boiling points (at 1.013 bar) are estimated to be 70°C, 170°C and 240°C respectively.

The bond dissociation enthalpies for the S–S and S–H bonds of H₂S_n have been determined by high‐level ab initio MO calculations. For H₂S₂ at 0 k, D(S–H) = 313 kJ·mol⁻¹ and D(S–S) = 271 kJ·mol⁻¹ were obtained by the G2(MP2) method (Antonello et al., 2002). The S–S bond dissociation energies of H₂S₂, H₂S₃ and of the central bond of H₂S₄ were calculated to be 259.5, 211.8 and 168.5 kJ·mol⁻¹, respectively, by the CCSD(T)/6–311++G(2df,p)//MP2/6–311++G** level of theory (Steudel et al., 2001).

The UV‐Vis spectra (200–400 nm) of H₂S_n in cyclohexane have been reported (Fehér and Münzner, 1963). These molecules show strong absorption in the region 200–230 nm, and the extinction coefficients decrease with longer wavelengths. Also, longer sulfur chains resulted in a higher molar absorbance at a given wavelength and more red‐shifted spectra. There is an absorption maximum or a broad plateau at 260–330 nm that becomes more and more pronounced when sulfur atoms increase.

The estimated Gibbs energies of formation of H₂S₂ Δ_fG^o(H₂S₂) are −2 kJ·mol⁻¹ in the gas phase and +5 kJ·mol⁻¹ in water. From the Δ_fG^o values for HSSH, HSS⁻ and SS²⁻, a pK _a of 2.6 is estimated for HSSH, and a pK _a of 13 is estimated for HSS⁻ (Koppenol and Bounds, 2017).

Redox reactions of H₂S/H₂S_n: H₂S can undergo both one‐ and two‐electron oxidations. The HS radical (HS^•) produced by one‐electron oxidation of sulfide (HS⁻) is a strong oxidant. When it is generated, it can participate in radical chain reactions. The reduction potential of the HS^•/HS⁻ redox couple is +920 mv versus NHE at pH 7 (Das et al., 1999). HS^• can react with HS⁻ to form hydrodisulfide radical anion, which can further react with oxygen (O₂) to form HSS⁻. The bimolecular rate constants for H₂S (in the form of HS⁻ in physiological buffers) with biologically important two‐electron oxidants are 0.73 M⁻¹·s⁻¹ for H₂O₂; 4.8 × 10³ M⁻¹·s⁻¹ for peroxynitrite (ONOOH); 8 × 10⁷ M⁻¹·s⁻¹ (Carballal et al., 2011) or 2 × 10⁹ M⁻¹·s⁻¹ (Nagy and Winterbourn, 2010) for hypochlorite (HOCl) (at pH 7.4 and 37°C). The immediate products are HSOH or HSCl, which can further react with HS⁻ to form H₂S₂ (rate constant: 1 × 10⁵ M⁻¹·s⁻¹) (Carballal et al., 2011). The reactivities of H₂S towards these oxidants are comparable to the rate constants of low molecular weight thiols such as cysteine and GSH. Given the very low endogenous H₂S concentration (as compared to thiols) in tissues, it is unlikely these oxidations are physiologically significant. Therefore, it is unlikely H₂S oxidation plays an important role for H₂S_n formation under normal physiological conditions.

If H₂S_n are generated in physiological conditions, it is expected that their aqueous solutions are formed. However, the exact species present in the solutions are still unclear. This is due to the instability and high reactivity of H₂S_n (or their anions) in water. Nevertheless, inorganic salts of H₂S_n are often used as the equivalents of H₂S_n for biological studies. These salts (such as Na₂S_n and K₂S_n) are yellow (or orange‐yellow) hygroscopic crystalline substances, which show a pronounced thermochromic effect. It is worth noting that these salts have different stability. For example, Na₂S₂ and Na₂S₄ are relatively stable, but Na₂S₃ is unstable in the solid state. Na₂S₃ solid readily decomposes to form a eutectic mixture of Na₂S₂ and Na₂S₄ (El Jaroudi et al., 1999). When polysulfide salts are dissolved in water, a complex equilibrium mixture will be established with HS⁻ and elemental sulfur being the products (Equation (6)).

$${{\mathrm{S}}_{\mathrm{n}}}^{2-}+{\mathrm{H}}_2\mathrm{O}\leftrightarrow {\mathrm{S}}_{\mathrm{n}-1}+{\mathrm{HS}}^{-}+{\mathrm{OH}}^{-}$$ (6)

This instant reaction has several implications: (i) when polysulfides are formed in aqueous environments, HS⁻ and elemental sulfur will also be present; (ii) if HS⁻ (or H₂S) coexists with elemental sulfur, polysulfides are likely to form; (iii) precipitation of elemental sulfur may be the indicator of polysulfide presence; (iv) polysulfides in aqueous solutions are likely to be the combination of S_n ²⁻, HS⁻, S_n and HS_n ⁻. The actual polysulfide anions (e.g. whether n = 2, 3 or 4 in S_n ²⁻) are difficult to determine. Because there is no direct method to determine single species either analytically or spectroscopically, the results of identification of specific polysulfides are somewhat speculative and rest on certain assumptions. For example, when mBB is used to trap polysulfides, the obtained bimane‐polysulfide adducts (B‐S_n‐B) may not actually reflect the true identity of the polysulfide species presented, as the formation of B‐S_n‐B adducts may be stepwise and thiol exchange reactions could occur. Currently, it is impossible to completely exclude HS⁻ and elemental sulfur from the solutions of polysulfides. This should be taken into consideration when studying the chemistry and biological functions of polysulfides. HS⁻ and elemental sulfur are stable species, and their corresponding clean solutions (without the presence of polysulfides) could be obtained. Therefore, control experiments using HS⁻ and elemental sulfur should always be carried out to confirm the results of polysulflides, especially for in vitro studies.

Detection methods for H₂S_n

UV‐Vis absorption spectroscopy of aqueous polysulfide solutions have been studied (Giggenbach, 1972). Absorption bands in the range of 240–420 nm are assigned to polysulfide anions (S_n ²⁻). In the current polysulfide studies, UV absorption peaks at ~300 and ~370 nm are often used to indicate the presence of polysulfides (Nagy and Winterbourn, 2010; Greiner et al., 2013). However, this method is not particularly sensitive. In recent studies of H₂S_n in biological samples, especially in tissue samples, H₂S_n are normally derivatized with mBB to form the corresponding alkylated products, such as B‐S₂‐B (from H₂S₂) or B‐S₃‐B (from H₂S₃) (Scheme 2). These adducts can then be analysed and quantified by HPLC with a scanning fluorescence detector and tandem MS (MS/MS). The appropriate derivatizing conditions are to incubate H₂S_n with mBB for 30 min in 0.1 M phosphate buffer (pH 7.0) (Koike et al., 2017). Using this method, endogenous H₂S₂ in mouse brain tissues was measured to be 0.026 μmol·g⁻¹ protein. As discussed previously, the use of the mBB assay is based on the assumption that all H₂S_n species and other RSS such as CysSSH are alkylated by mBB with high efficiency, and the alkylation rates are faster than the dynamic interchanges within these sulfur species. This assumption should be carefully validated before it can be considered as the standard method for the identification of H₂S_n.

Scheme 2.

Mono‐bromobimane‐based detection methods for H₂S_n.

Fluorescent sensors for H₂S_n

In order to better understand the functions of H₂S_n, methods that allow real‐time and non‐invasive detection of H₂S_n in biological systems are needed. In this regard, fluorescent sensors are ideal because of their high sensitivity and spatiotemporal resolution ability, as well as their ease of use (Fernández‐Suárez and Ting, 2008; Lin et al., 2015). While a large number of fluorescent sensors for other RSS, such as cysteine, homocysteine, GSH and H₂S, have been developed, only a small selection of H₂S_n sensors have been reported in the last several years (Gupta et al., 2017; Takano et al., 2017). These reported sensors are summarized in this section. Up to date, all known H₂S_n sensors are reaction‐based fluorescent sensors. These sensors react with H₂S_n through certain specific chemical reactions to change their fluorescence properties, thus achieving the selective detection of H₂S_n. So far, three main strategies have been used in the design of H₂S_n sensors, which take advantage of the strong nucleophilicity and reduction ability of H₂S_n. These strategies are (i) H₂S_n‐mediated aromatic substitution‐cyclization reactions; (ii) H₂S_n‐mediated ring‐opening reaction of aziridine; (iii) H₂S_n‐mediated reduction of nitro groups.

Sensors based on H₂S_n‐mediated aromatic substitution‐cyclization (summarized in Scheme 3)

Scheme 3.

Fluorescent sensors based on H₂S_n‐mediated aromatic substitution‐cyclization. The detection limits (D.L.) for each sensor is shown below its structure.

2‐Fluoro‐5‐nitrobenzoic ester and phenyl‐2‐(benzoylthio)benzoate have been used as the specific recognition units for H₂S_n based on aromatic substitution–cyclization reaction. In 2014, our laboratory reported the first selective fluorescent sensors for H₂S_n, for example, DSP1‐3, which employed the 2‐fluoro‐5‐nitrobenzoic ester moiety as the H₂S_n recognition unit (Liu et al., 2014). The sensing mechanism is based on two‐step reactions: (i) the 2‐fluoro‐5‐nitrobenzoic ester is attacked by H₂S₂ via aromatic nucleophilic substitution to form a –SSH‐containing intermediate and (ii) the –SSH moiety undergoes an intramolecular cyclization, leading to the release of the fluorophore. Among the DSP probes, DSP‐3 exhibited high sensitivity and selectivity for H₂S_n. It was almost non‐fluorescent in buffers but when treated with Na₂S₂, a 137‐fold fluorescence increase was observed. The detection limit was 71 nM, and DSP‐3 was successfully applied to image H₂S_n in HeLa cells.

Several other research groups have adopted the fluoro‐5‐nitrobenzoic ester recognition unit and developed sensors with interesting fluorescence properties. For example, Chen et al. reported near‐IR (NIR) fluorescent probes Mitro‐ss and BD‐ss (Gao et al., 2015a,b). Both probes possessed boron difluoride dipyrromethene (BODIPY) as the fluorophore and 2‐fluoro‐5‐nitrobenzoic ester as a response site. Mitro‐ss bears a triphenylphoniummoiety as the mitochondria targeting group. The probes showed almost no fluorescence due to the donor‐excited photoinduced electron transfer (d‐PET) process from the excited BODIPY to the 2‐fluoro‐5‐nitrobenzoic ester. The addition of Na₂S₂ triggered a significant fluorescence enhancement (λ_em 730 nm), suggesting the d‐PET process was suppressed by deprotection. With the aid of Mitro‐ss, the two possible generation mechanisms of H₂S_n were further demonstrated. With one mechanism, H₂S_n could be generated from the reaction between H₂S and ROS, whereas with the other, H₂S_n could be produced from cystine by cystathionine γ‐lyase and cystathionine β‐synthase. Another NIR probe Cy‐S_n using semiheptamethine as the fluorophore was reported by Peng et al. (Ma et al., 2017). With this probe, the presence of H₂S_n resulted in a remarkable fluorescence enhancement while other RSS or ROS did not induce obvious fluorescence enhancement. The probe was successfully used to image endogenous H₂S_n in RAW264.7 macrophage cells and in mice. Han et al. also reported a NIR probe KB1, in which dicyanomethylene‐benzopyran dye was employed as the fluorophore (Li et al., 2018). KB1 showed more than 30‐fold fluorescence increase (λ_em 682 nm) with H₂S_n and the detection limit of 8.2 nM. The probe was used for imaging H₂S_n in MCF‐7 cells.

Two‐photon (TP) fluorescent probes, with low‐energy NIR wavelength excitation, are very useful for bio‐imaging. These probes have advantages such as deep penetration in tissues, excellent 3D imaging capability and minimum photo‐damage to biological specimens. To date, there are several TP fluorescent probes developed for sensing H₂S_n by utilizing the 2‐fluoro‐5‐nitrobenzoic ester moiety as the recognition site. In 2015, Liu et al. reported a TP probe QS_n using 2‐benzothiazol‐2‐yl‐quinoline‐6‐ol as the fluorophore (Zeng et al., 2015). The addition of H₂S_n to the QS_n solution resulted in a 24‐fold enhancement in fluorescence intensity at 534 nm under both one‐photon (λ_ex 368 nm) and TP (λ_ex 730 nm). QS_n was not only used to visualize exogenous and endogenous H₂S_n in HeLa cells but also applied to investigate the distribution of H₂S_n in living zebrafish embryos. Another TP probe GCTPOC‐H₂S₂ was reported by Lin et al. (Shang et al., 2016). The reaction between GCTPOC‐H₂S₂ and H₂S_n produced a product that exhibited larger TP action cross‐section up to 500 GM (λ_ex 780 nm). With this probe, imaging of H₂S_n in MCF‐7 cells and in situ detecting H₂S_n levels in live mice were achieved. Liu and co‐workers developed a naphthalimide‐based TP ratiometric probe NRT‐HP based on the ICT mechanism (Han et al., 2016). Upon addition of H₂S_n to the NRT‐HP solution, the fluorescence emission at 460 nm decreased, and a new red‐shifted emission at 542 nm increased simultaneously. Probe NRT‐HP was successfully applied to image H₂S_n in different cell lines (MCF‐7, A549) by TP microscopy. Furthermore, upon excitation at 800 nm, probe NRT‐HP could detect H₂S_n in tissue up to 300 μM of penetration depth. In order to get larger emission shift to improve the measuring accuracy, Yin et al. reported a FRET‐based TP probe TPR‐S based on a naphthalene‐rhodol dye (Zhang et al., 2016). Before adding H₂S_n, the rhodol moiety of TPR‐S was in spiro ring‐closing form and the FRET process was off, so the probe emitted only naphthalene emission at 448 nm. With the addition of H₂S_n, the fluorescence emission at 448 nm decreased, and a new emission at 541 nm increased. The reason for this phenomenon was that rhodol moiety was deprotected to form spiro ring‐opening and the FRET was on. The efficiency of FRET from naphthalene to rhodol was estimated to be 91.3%. The probe TPR‐S was able to image H₂S_n in live HeLa cells, rat liver slices and LPS‐induced acute organ injury upon excitation at 740 nm with femtosecond pulses.

Although 2‐fluoro‐5‐nitrobenzoic ester has been widely used in the design of H₂S_n probes, this recognition site could also react with biothiols to form thioether adducts. This problem does not affect the selectivity of the probes but could lead to the consumption of the probes in biological systems. High probe loading is usually required in actual applications. To solve this problem, our laboratory designed another strategy for H₂S_n sensing, which utilized phenyl 2‐(benzoylthio)benzoate moiety as the recognition unit. This design takes the advantage of the dual reactivity of H₂S_n. It is known that H₂S_n are stronger nucleophiles than biothiols and H₂S under physiological pH. Moreover, H₂S_n belong to the sulfane sulfur family. A characteristic reaction of sulfane sulfurs is that they can act as electrophiles and react with certain nucleophiles (Chen et al., 2013). Overall, H₂S_n have a unique dual‐reactivity and can be both a nucleophile and an electrophile. Therefore, we predicted phenyl 2‐(benzoylthio)benzoate‐based probes (PSP) should be specific for H₂S_n (Scheme 3B) (Chen et al., 2015a, b). We found the substitution group (R) of the thioester moiety was critical for the probes' selectivity, and PSP3 was the most promising probe. PSP3 showed a fast response to H₂S_n, excellent selectivity and high sensitivity with a detection limit of 3 nM. It was used in the detection of exogenous and endogenous H₂S_n in different cells. Based on this design, Ma et al. further developed an NIR probe HXPI‐1, which employed a hemicyanine derivative as the fluorophore (λ_em 708 nm) and was used to visualize H₂S_n in mice (Fang et al., 2017). The detection limits of each sensor are provided below its structure in Scheme 3.

Sensors based on H₂S_n‐mediated ring‐opening reaction of aziridine (Scheme 4)

Scheme 4.

Aziridine‐based sensor AP and nitro‐reduction based sensors. The detection limits (D.L.) for each sensor is shown below its structure.

H₂S_n are known to be strong nucleophiles and expected to be more reactive in certain nucleophilic reactions than H₂S or thiols. Our group has explored some reactions of H₂S_n and found that H₂S_n could effectively react with aziridines. Based on this reaction, the first off–on twisted intramolecular charge transfer (TICT)‐based fluorescent probe AP was developed (Chen et al., 2015a,b). The probe was constructed by incorporating an aziridine group on dansyl dye to quench the fluorescence. When reacted with H₂S_n, the fluorescence intensity at 530 nm increased due to the suppress of the TICT effect. The reaction product of AP with H₂S_n showed not only good TP photophysical properties but also high luminescence efficiency in solid state. However, attempts to use AP for imaging H₂S_n in live cells were not successful.

Sensors based on H₂S_n‐mediated reduction of nitro groups (Scheme 4)

Based on the strong reducibility of H₂S_n, Chen et al. designed two NIR fluorescent probes Hcy‐Mito and Hcy‐Biot that could be used for monitoring O₂ ^•− and H₂S_n in cells and in vivo (Huang et al., 2016). Two different targeting groups were used on these probes. Hcy‐Mito bearing a benzyl chloride moiety could target the mitochondria, and Hcy‐Biot bearing a biotin moiety could target carcinoma tissues. These probes displayed almost no fluorescence due to the destroyed polymethine conjugated system and the d‐PET process from heptamethine cyanine dye to m‐nitrophenol group. Addition of O₂ ^•− would restore the polymethine conjugated systems of Hcy‐Mito and Hcy‐Biot, resulting in a weak fluorescence at 780 nm, and the detection limit was 0.1 μM. Subsequently, the addition of H₂S_n would reduce the nitro group to amine and trigger strong fluorescence enhancement due to the block of the d‐PET process. The detection limit of H₂S_n was estimated to be 80 nM. Hcy‐Mito was used for in situ detection of O₂ ^•− and H₂S_n levels in mitochondria and endogenous O₂ ⁻ and H₂S_n crosstalk in living cells. Hcy‐Biot showed high tumour‐targeting ability and excellent tissue penetration in live animal models. Yang et al. reported another nitro‐based probe F1 based on the flavylium fluorophore (Gong et al., 2016). The fluorescence of F1 was quenched because attaching a nitro group to flavylium blocked the ICT process. Upon treatment with H₂S_n, the nitro group would be reduced to amine that unblocked the ICT process, leading to restore the strong fluorescence of flavylium. It should be noted that H₂S could also induce some fluorescence enhancement for these nitro‐based probes. Therefore, the selectivity of these probes may not be ideal.

Sensors for dual detection of both H₂S and H₂S_n

H₂S_n and H₂S are redox partners. They should work collectively to maintain a sulfur‐related redox balance. In order to better understand their functions and crosstalks, it is highly desirable to develop sensors that can simultaneously detect H₂S_n and H₂S. We recently developed such a probe, DDP‐1 (Scheme 5). It employs two recognition sites in one molecule: phenyl 2‐(benzoylthio)benzoate for H₂S_n and azide for H₂S. Rhodol and coumarin were used as the fluorophores and tethered via a rigid piperazine linker (Chen et al., 2016). H₂S_n selectively reacted with the phenyl 2‐(benzoylthio)benzoate moiety and induced a remarkable green fluorescence increase of rhodol. However, H₂S promoted the reduction reaction of azide −N₃ to −NH₂, which was accompanied by the oxidation of H₂S to H₂S_n. This would lead to blue fluorescence of coumarin. Meanwhile, the resultant H₂S_n would also react with 2‐(benzoylthio)benzoate to release rhodol and trigger the FRET from coumarin to rhodol. As a result, H₂S could be identified by both fluorescence signals in blue and green channels. The detection limit of H₂S_n and H₂S was calculated to be 100 and 24 nM respectively. DDP‐1 is the first fluorescent probe capable of selective discrimination of H₂S_n and H₂S.

Scheme 5.

Dual detection probe DDP‐1.

So far, a handful of H₂S_n fluorescent probes have been reported, and we expect to see more being developed in the coming years. In the study of H₂S_n probes, researchers usually would validate their specificity for H₂S_n (vs. other thiols and H₂S). However, their selectivity for some other important sulfur species, especially recently recognized cysteine hydropolysulfide Cys‐S_nSH (Ida et al., 2014; Akaike et al., 2017), is not well‐defined. As one can imagine, Cys‐S_nSH and H₂S_n are likely to coexist, so one should not expect the probes to differentiate these species. Nevertheless, this should not cause a problem as Cys‐S_nSH and H₂S_n should have similar biological functions.

Releasing agents of H₂S_n

In the current studies of H₂S_n, researchers always use their inorganic salts (Na₂S₂, Na₂S₃, etc.) as the standard H₂S_n equivalents. These salts (Na₂S₂, Na₂S₃, Na₂S₄) are commercially available by vendors like Dojindo Molecular Technologies, Inc. Due to their instability, these salts often contain other sulfur‐based impurities like sulfide (S²⁻) and elemental sulfur (S₈). This problem makes the use of these salts problematic as it is difficult to attribute the observed bioactivity solely to H₂S_n. Moreover, due to the high reactivity of H₂S_n, their biological production is expected to be a slow and steady process, which cannot be mimicked using Na₂S_n (these salts produce H₂S_n immediately upon dissolving in buffers and are considered fast and uncontrollable H₂S_n donors). Due to these concerns, slow and controllable H₂S_n donors should be useful research tools in this field. This is similar to the field of H₂S donors. However, donors of H₂S_n are very underdeveloped. Recently, Wang and co‐workers developed several such donors (Scheme 6) (Yu et al., 2018). In this work, H₂S₂ is caged as two thiol acid groups linked by a disulfide bond, for example, acyl disulfides. A masked phenol hydroxyl acts as a latent nucleophile for initiation of H₂S₂ release through lactonization. A ‘trimethyl lock’ is also employed to facilitate lactonization. Esterase‐triggered donors like BW‐HP‐302 and phosphatase‐triggered donors like BW‐HP‐303 have been prepared. Their enzyme‐catalysed H₂S₂ productions have been demonstrated. Initial studies showed that these donors could induce protein S‐persulfidation on GAPDH. It should be noted that acyl disulfides are used as the core of these donors. It is known that acyl disulfides are highly reactive species and can easily react with nucleophiles like amines (Mali and Gopi, 2014). In biological systems, these donors may react with naturally existing nucleophiles (like amino acids) to cause off‐target effects and undesired H₂S₂ release. This possibility needs to be explored in future studies.

Scheme 6.

Enzyme‐triggered H₂S_n donors.

Concluding remarks

Increasing numbers of publications have suggested H₂S_n are potent physiological mediators and play important roles in sulfur‐related redox biology. Nevertheless, the actual chemical species involved in those biological processes are still unclear as current conclusions are mostly based on derivatization methods, which are viewed as indirect evidence. In terms of their chemistry, H₂S_n are highly reactive and unstable molecules. Tracking their formation and their fates can be very difficult. Moreover, H₂S_n and H₂S often coexist, as H₂S_n are found to be the common impurities in H₂S solutions (Greiner et al., 2013). Therefore, differentiation of the functions of H₂S and H₂S_n can also be challenging. Along with further understanding their biological functions, it is also important to explore their chemical properties and reactions in biological systems. To this end, the development of chemical tools or methods for the delivery and detection of H₂S_n should be valuable. It is gratifying to see the development of H₂S_n detection sensors has made some progress in recent years and the development of H₂S_n donors has started emerging. We expect these will continue to be active research topics, and more interesting work will appear in the future.

Nomenclature of targets and ligands

Key protein targets and ligands in this article are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY (Harding et al., 2018), and are permanently archived in the Concise Guide to PHARMACOLOGY 2017/18 (Alexander et al., 2017a, 2017b).

Conflict of interest

The authors declare no conflicts of interest.

Liu, H. , Radford, M. N. , Yang, C. , Chen, W. , and Xian, M. (2019) Inorganic hydrogen polysulfides: chemistry, chemical biology and detection. British Journal of Pharmacology, 176: 616–627. 10.1111/bph.14330.