Key bioactive reaction products of the NO/H₂S interaction are S/N-hybrid species, polysulfides, and nitroxyl

Significance

Reactions of sulfur-centered nucleophiles with nitrogenous species have been studied independently for more than a century for synthetic/industrial purposes; to understand geochemical, atmospheric, and biological processes; and to explain the origins of life. Various products and reaction mechanisms were proposed. We here identify a singular process comprising a network of cascading chemical reactions that form three main bioactive products at physiological pH: nitrosopersulfide, polysulfides, and dinitrososulfite. These anionic products scavenge, transport, and release NO/HNO or sulfide/sulfane sulfur, each displaying distinct chemistries and bioactivities. Our observations provide a chemical foundation for the cross-talk between the NO and H₂S signaling pathways in biology and suggest that the biological actions of these entities can be neither considered nor studied in isolation.

Abstract

Experimental evidence suggests that nitric oxide (NO) and hydrogen sulfide (H₂S) signaling pathways are intimately intertwined, with mutual attenuation or potentiation of biological responses in the cardiovascular system and elsewhere. The chemical basis of this interaction is elusive. Moreover, polysulfides recently emerged as potential mediators of H₂S/sulfide signaling, but their biosynthesis and relationship to NO remain enigmatic. We sought to characterize the nature, chemical biology, and bioactivity of key reaction products formed in the NO/sulfide system. At physiological pH, we find that NO and sulfide form a network of cascading chemical reactions that generate radical intermediates as well as anionic and uncharged solutes, with accumulation of three major products: nitrosopersulfide (SSNO⁻), polysulfides, and dinitrososulfite [N-nitrosohydroxylamine-N-sulfonate (SULFI/NO)], each with a distinct chemical biology and in vitro and in vivo bioactivity. SSNO⁻ is resistant to thiols and cyanolysis, efficiently donates both sulfane sulfur and NO, and potently lowers blood pressure. Polysulfides are both intermediates and products of SSNO⁻ synthesis/decomposition, and they also decrease blood pressure and enhance arterial compliance. SULFI/NO is a weak combined NO/nitroxyl donor that releases mainly N₂O on decomposition; although it affects blood pressure only mildly, it markedly increases cardiac contractility, and formation of its precursor sulfite likely contributes to NO scavenging. Our results unveil an unexpectedly rich network of coupled chemical reactions between NO and H₂S/sulfide, suggesting that the bioactivity of either transmitter is governed by concomitant formation of polysulfides and anionic S/N-hybrid species. This conceptual framework would seem to offer ample opportunities for the modulation of fundamental biological processes governed by redox switching and sulfur trafficking.

Nitrogen and sulfur are essential for all known forms of life on Earth. Our planet’s earliest atmosphere is likely to have contained only traces of O₂ but rather large amounts of hydrogen sulfide (H₂S) (1). Indeed, sulfide may have supported life long before the emergence of O₂ and NO (2, 3).* This notion is consistent with a number of observations: H₂S is essential for efficient abiotic amino acid generation as evidenced by the recent reanalysis of samples of Stanley Miller’s original spark discharge experiments (4), sulfide is an efficient reductant in protometabolic reactions forming RNA, protein, and lipid precursors (5), and sulfide is both a bacterial and mitochondrial substrate (6), enabling even multicellular lifeforms to exist and reproduce under conditions of permanent anoxia (7). Thus, although eukaryotic cells may have originated from the symbiosis of sulfur-reducing and -oxidizing lifeforms within a self-contained sulfur redox metabolome (8), sulfide may have been essential even earlier by providing the basic building blocks of life.

The chemical reactions of sulfur-centered nucleophiles with a range of nitrogen-containing species have been studied for different reasons and as independent processes for more than a century, and early reports indicated complex reaction mechanisms (9–13). The recent surge of interest in this chemistry in the biological community (13–15) was triggered by a growing appreciation that NO and sulfide exert similar and often interdependent biological actions within the cardiovascular system and elsewhere (NO/H₂S “cross-talk”) (16, 17), resulting in mutual attenuation or potentiation of their responses. This cross-talk is possibly mediated by chemical interactions (18–20), but much of the older chemical work seems to have been forgotten. Recently, low concentrations of sulfide were shown to quench NO-mediated vascular responses through formation of an uncharacterized “nitrosothiol” (RSNO) (18–20), assumed to be thionitrous acid (HSNO) (13–15).

A recent report of the detection by MS of the highly unstable HSNO at physiological pH (21) has attracted considerable attention from the biological community, because it could be an intermediate in the reaction of sulfide with RSNOs (22) and a precursor for NO, nitrosonium (NO⁺) equivalents, and nitroxyl (HNO). However, a key aspect of HSNO’s properties that seems to have been overlooked in these discussions is its mobile hydrogen, allowing facile 1,3 hydrogen shift and formation of four isomers with the same chemical equation (13)—a feature described in the seminal studies by Goehring in the 1950s (23) and by Müller and Nonella later on (24, 25) that distinguishes HSNO from all other RSNOs (26). The same feature also contributes to the short half-life of the molecule at ambient temperatures, making it more probable that other yet unknown entities are involved as biological mediators of the NO/H₂S cross-talk. Chemical studies by Seel and Wagner (9, 10) showed that NO readily reacts with HS⁻ in basic aqueous solution or organic solvents under anoxic conditions to form the yellow nitrosopersulfide (SSNO⁻). Accumulation of this product was also observed after reaction of RSNOs with sulfide at pH 7.4 (26, 27); moreover, SSNO⁻-containing mixtures were found to release NO, activate soluble guanylyl cyclase (sGC) (26), and relax vascular tissue (28), although a contribution of other reaction products to these effects cannot be excluded. Meanwhile, other sulfane sulfur molecules, including persulfides (RSSH) and polysulfides (RSS_n⁻ and HS_n⁻), have come to the fore as potential mediators of sulfide’s biological effects (29–31), but little is known about their pathways of formation, prevalence in biological systems, and relationship with NO.

In view of this confusion, we sought to carry out an integrative chemical/pharmacological investigation to study the chemical biology of the reaction of NO with sulfide more thoroughly and systematically identify potentially bioactive reaction products. We here report that the NO/H₂S interaction leads to formation of at least three product classes with distinct in vivo bioactivity profiles: nitrosopersulfide (SSNO⁻), polysulfides (HS_n⁻), and dinitrososulfite [ONN(OH)SO₃⁻ or N-nitrosohydroxylamine-N-sulfonate (SULFI/NO)]; all anions at physiological pH. Their formation is accompanied by both scavenging and release of NO and H₂S and formation of nitrous oxide (N₂O), nitroxyl (HNO), nitrite (NO₂⁻), nitrate (NO₃⁻), and various sulfoxy species. These results not only offer an intriguing explanation for the quenching and potentiating effects of sulfide on NO bioavailability but also, provide a novel framework for modulation of fundamental biological processes governed by redox switching and sulfur trafficking. This chemistry is likely to prevail wherever NO and sulfide are cogenerated.

Results

Sulfide Modulates NO Bioavailability in a Concentration-Dependent Manner.

Effects of sulfide on NO bioavailability and hemodynamics were investigated in anesthetized rats. Pilot studies confirmed that sodium hydrosulfide (NaHS; 1.8–18 µmol/kg) lowers blood pressure and heart rate in a dose-dependent manner; effects were short-lived and accompanied by alterations in NO metabolite status in RBCs and plasma (SI Appendix, Fig. S1 and Table S1). As with inhaled NO (32), higher sulfide doses increased total nitroso (RXNO) levels in RBCs (SI Appendix, Fig. S1). Inhibition of NO synthase by S-ethylisothiourea prolonged the action of sulfide and markedly increased its toxicity (SI Appendix, Fig. S1 A and B), showing that endogenous NO production modulates sulfide bioactivity and attesting to the reciprocal nature of interaction of these signaling molecules. In subsequent experiments, NaHS was administered by continuous infusion (2.8 µmol/kg per minute in PBS, pH 7.4) to counter the rapid rate of sulfide elimination (33), and blood was collected repeatedly for measurement of circulating NO biomarkers (Fig. 1 and SI Appendix, Fig. S2 and Table S2). Consistent with the notion that vascular sulfide levels rise only after inactivation (binding/elimination) pathways become saturated, no significant hemodynamic changes were observed in the first 30 min of infusion (Fig. 1 A and B and SI Appendix, Table S2), albeit NO-heme levels dropped significantly (Fig. 1D). After 1 h of sulfide infusion, blood pressure was significantly lower, whereas heart rate remained constant (Fig. 1 A and B and SI Appendix, Table S2). The lack of a compensatory rise in heart rate and the decrease in respiratory rate that accompanied the fall in mean arterial pressure are consistent with the recognized metabolic effects of sulfide, capable of inducing a state of suspended animation (34). Concomitant with these changes in blood pressure, erythrocyte RXNO and NO-heme levels gradually increased (Fig. 1 C and D).

Fig. 1.

Sulfide affects NO bioavailability in vivo and in vitro. (A and B) Continuous i.v. infusion of sodium hydrosulfide (2.8 µmol/kg per min NaHS in PBS, pH 7.4) progressively decreases blood pressure (BP) in rats. (A) Original recording depicting progressive decrease in BP during ongoing sulfide infusion; incisions (arrows) are caused by interruption of pressure recording during blood collection. (B) Changes in mean arterial blood pressure (MAP; n = 5; ANOVA P = 0.0256) and (Inset) heart rate (HR). *Dunnett's P < 0.05 vs. baseline. (C) Gradual increases in circulating nitroso species (RXNO) levels in RBCs (n = 3; ANOVA P = 0.0032). *Dunnett's P < 0.05 vs. baseline. (D) Concomitant transient decrease followed by an increase in NO-heme levels during continuous sulfide infusion (2.8 µmol/kg per min NaHS in PBS, pH 7.4; n = 3; ANOVA P = 0.0126). *Tuckey P < 0.05 vs. baseline. (E) Sulfide (10 µM Na₂S) decreases Sper/NO (100 µM)-mediated sGC activation in RFL-6 cells pretreated with the phosophdiesterase inhibitor 3-isobutyl-1-metylxanthine (IBMX). The scheme represents the experimental setup (n = 6; ANOVA P < 0.001). CTRL, control. *Tuckey P < 0.01 vs. untreated. # t test P < 0.05 (F) Equimolar concentrations of sulfide (33.4 µM) scavenge NO released form NO donors (33.4 µM DETA/NO) as assessed by time-resolved chemiluminescence detection under both aerated and (Inset) deaerated conditions, whereas excess sulfide (334 µM) transiently elevates NO release (representative of n = 3 independent experiments); DETA/NO, diethylenetriamine NONOate.

The effect of sulfide on NO-induced sGC stimulation was tested in an NO reporter cell line in the presence of a phosphodiesterase (PDE) inhibitor. Low sulfide concentrations (10 µM) inhibited sGC stimulation by the NO donor, spermine NONOate (Sper/NO; 100 µM), whereas cGMP levels at equimolar concentrations of sulfide and Sper/NO did not differ from those of Sper/NO alone (Fig. 1E). Although constitutive cGMP–PDE activity in these cells is very low (26) and cells were pretreated with a PDE inhibitor, 100 µM sulfide increased cGMP on its own (Fig. 1E), prohibiting the use of higher sulfide concentrations to investigate NO responses in these cells. Additional chemical experiments with NO donors/sulfide in cell-free buffer systems confirmed that sulfide, dependent on concentration, either quenches or transiently enhances NO as detected by chemiluminescence (Fig. 1F).

Collectively, these data show a dual effect of sulfide on NO bioavailability, with lower doses inhibiting and higher doses restoring (or in some cases, potentiating) NO bioactivity in cell-free systems, cells in vitro, and rats in vivo.

Sulfide Reacts with NO to Form Polysulfides and Two NO-Containing S/N-Hybrid Species.

Our next efforts were directed toward elucidating whether there is a chemical foundation for this NO/H₂S cross-talk by identifying specific reaction products. UV-visible spectroscopy offered a first glimpse into the chemistry of the NO/sulfide interaction. A dominant product of the reaction of sulfide with NO (Fig. 2A), the NO donor DEA/NO (Fig. 2 B–D and SI Appendix, Fig. S4), or RSNOs (SNAP in Fig. 2E) (other RSNOs are in ref. 26) in aqueous buffer at pH 7.4 in both the absence and the presence of O₂ is a yellow compound (λ_max = 412 nm), in particular when sulfide is in excess (Fig. 2 C and D). We (13, 26) and others (9, 10, 27, 35) attributed this species to SSNO⁻. We here extend those earlier observations with RSNOs to NO itself (as shown by the reaction of sulfide with aqueous NO solution and NO donors). Absorbance increases in the regions of 250–300 nm and below 250 nm are also apparent (Fig. 2, arrows).

Fig. 2.

The reaction of NO with sulfide leads to formation of three major products, which we assign to be SSNO⁻ (λ_max = 412 nm), HS_n⁻ (λ_max = 290–300 nm), and SULFI/NO (λ_max = 259 nm). (A) Reaction of aqueous solutions of NO (200 µM) with sulfide (2 mM) under deaerated conditions in buffer at pH 7.4 leads to formation of a peak with λ_max = 412 nm (SSNO⁻) and increases in absorbance at λ_max < 300 nm (HS_n⁻). Products with λ_max < 250 nm are not discernable from sulfide because of the high concentration of HS⁻ (λ_max = 230 nm) in these experiments; spectra were taken at the reaction start (blue) and every 5 s after the addition of NO. (Inset) kinetics of SSNO⁻ formation. (B) Reaction of the NO donor DEA/NO (1 mM) with sulfide (10 mM) under aerated conditions forms SSNO⁻ and HS_n⁻. The blue line indicates the spectrum before the addition of sulfide. (Inset) Kinetics of formation of SSNO⁻. (C and D) The yield of SSNO⁻ formation depends on both (C) sulfide concentration and (D) the rate of NO release; all spectra were taken 10 min after the start of the reaction. (E) SSNO⁻ (λ_max = 412 nm), HS_n⁻ (λ_max = 290–300 nm), and SULFI/NO (λ_max = 259 nm) are formed in the reaction of sulfide with the S-nitrosothiol SNAP (1 mM SNAP + 10 mM Na₂S; 10 min); SULFI/NO is detectable after removal of sulfide by gassing with N₂ for 10 min. (F) Addition of HS_n⁻ (12.5–200 µM) increases the rate of formation of SSNO⁻ from the reaction of SNAP (200 µM) and sulfide (2 mM); the induction period observed at no/low added HS_n⁻ points to an autocatalytic effect of HS_n⁻ (n = 3). All spectra in A–E are representative of 3–10 independent experiments. Au, arbitrary unit; DEA/NO, dietylamine/NONOate; SNAP, S-nitroso-N-acetyl-penicillamine.

The reaction products absorbing in the region around 300 nm (λ = 290–350 nm) seem to be HS_n⁻, because this feature disappeared on addition of the classical sulfane sulfur-reducing reagents DTT or cyanide (SI Appendix, Fig. S5 A and B) or millimolar concentrations of cysteine and glutathione (26). Contrary to HS_n⁻, the SSNO⁻ peak was resistant toward thiols and cyanide (SI Appendix, Fig. S5 A and B), and its decomposition rate was hardly affected by the presence of these chemicals (SI Appendix, Fig. S5C). Like HS_n⁻ but contrary to classical RSNOs, SSNO⁻ is relatively stable at neutral and basic pH levels (SI Appendix, Fig. S5C) but rapidly decomposes with formation of colloidal sulfur and H₂S on acidification as shown previously (26). Interestingly, HS_n⁻s are not only products of SSNO⁻ decomposition (SI Appendix, Fig. S5D) but also, likely intermediates with clear catalytic effects on SSNO⁻ formation (Fig. 2F). Under conditions of no/low added HS_n⁻, an induction period is observed that disappears at higher HS_n⁻ concentrations (Fig. 2F), pointing to possible autocatalytic effects of HS_n⁻ formed during the reaction between sulfide and NO/RSNO. The product yield in the reaction of SNAP with excess sulfide approached 30% for SSNO⁻ as estimated by measuring the concentrations of either H₂S or sulfane sulfur atoms liberated by reduction, cyanolysis, or chloroform extraction during formation and/or decomposition of SSNO⁻ (SI Appendix, Fig. S5) and was calculated to be ∼34% in nonaqueous media (ε = 2,800 M⁻¹ cm⁻¹, λ_max = 448 nm) (35). Furthermore, SSNO⁻ decomposition in the presence of DTT released two times as much sulfide as sulfane sulfur (SI Appendix, Fig. S5 D and E). Therefore, SSNO⁻ contains two sulfur atoms, one of which is a sulfane sulfur.

In the reaction of NO with sulfide, in basic aqueous and nonaqueous conditions under exclusion of air, Seel and Wagner (9) also observed the formation of SULFI/NO ([ONN(OH)SO₃]⁻), a complex formed by the reaction of sulfite with NO. This product is also known as dinitrososulfite (36) (λ_max = 259 nm, ε = 8,198 M⁻¹ cm⁻¹), which we find to be another product of the reaction of NO and sulfide at physiological pH. Indeed, a peak with λ_max = 259 nm appears in RSNOs/sulfide incubates after removal of excess unreacted sulfide by bubbling with N₂ (Fig. 2E). The formation of SULFI/NO from DEA/NO/sulfide mixtures cannot be followed spectrophotometrically, because DEA/NO (like SULFI/NO) is a diazeniumdiolate (36) that absorbs in the same wavelength range. Other putative reaction products are sulfoxy species, including sulfite, sulfate, thiosulfate (S₂O₃²⁻), and polythionates ([O₃S-S_x-SO₃]²⁻), all absorbing at wavelengths <250 nm and spectrophotometrically difficult to distinguish from each other.

Taken together, these results show that the reaction of sulfide with NO or RSNO under physiologically relevant conditions leads to formation of three major reaction products: SSNO⁻, HS_n⁻, and SULFI/NO. These products have been described in different contexts before, but we find that all three are formed in sequential reactions of the same chemical system under physiologically relevant conditions. With high NO fluxes and excess sulfide, SSNO⁻ is a major reaction product.

Mass Spectrometric Identification of the Products of the Sulfide/NO Reaction.

To definitively identify the reaction products of NO (1 mM DEA/NO) or RSNOs (1 mM SNAP) with sulfide (2 mM Na₂S), incubation runs in phosphate or Tris buffer at pH 7.4 were subjected to electrospray ionization (ESI)–high-resolution MS (HRMS) analysis (Fig. 3 and SI Appendix, Table S7). Because many of the key reaction products were suspected to be negatively charged species at pH 7.4, negative ionization mode was used throughout. The S/N-hybrid species SSNO⁻ (compound 1; m/z theoretical = 93.94268, m/z found = 93.9427, error = 0.39 milli mass units, mmu) and SULFI/NO (compound 2; HO₅N₂S⁻; m/z theoretical = 140.96061, m/z found = 140.9612, error = 0.35 mmu) and multiple polysulfide species, including HS₃⁻, HS₄⁻, and HS₅⁻, were identified as reaction products of sulfide with either NO donor (Fig. 3). For SSNO⁻ and SULFI/NO, assigned structures were confirmed by analysis of their fragmentation pattern (Fig. 3 A, Center and B, Center); the former was found to eliminate NO by hemolytic cleavage, forming the persulfide radical (S₂·⁻) (Fig. 3A), whereas N₂O elimination and sulfate formation (detected as HSO₄⁻) characterized the latter (Fig. 3B). The structural assignments were unequivocally confirmed using ¹⁵N labeling, resulting in an m/z shift of 1 for SSNO⁻ and 2 for SULFI/NO (Fig. 3 A, Right and B, Right). Changes in relative abundance of reaction products over time were monitored during direct infusion of the reaction mixture into the ionization chamber. These studies revealed that both SSNO⁻ and SULFI/NO are formed surprisingly quickly (≤2 s, which was evidenced by additional experiments using NO/sulfide coinfusion by a T piece close to the ionization source) from both NO and RSNOs followed by gradual accumulation of medium-chain/long-chain HS_n⁻.

Fig. 3.

Identification by ESI-HRMS of SULFI/NO and SSNO⁻ as S/N-hybrid species formed by the reaction of sulfide with DEA/NO or SNAP. (A) Formation of SSNO⁻ from DEA/NO (1 mM)/sulfide (2 mM) and SNAP (1 mM)/sulfide (2 mM) incubates. (Left) Mass and (Center) fragmentation spectra of SSNO⁻ (compound 1) from DEA-NO-sulfide incubates; (Right) shift in m/z of SSNO⁻ using an equimolar mixture of ¹⁴N/¹⁵N-labeled SNAP with sulfide. (B) Formation of SULFI/NO (compound 2) from DEA/NO (1 mM)/sulfide (2 mM) and SNAP (1 mM)/sulfide (2 mM) incubates. (Left) Mass and (Center) fragmentation spectra of SULFI/NO; (Right) m/z shifts of one and two by reacting an equimolar mixture of ¹⁵N/¹⁴N-SNAP (1 mM) with sulfide (2 mM). (C) Extracted ion chromatograms showing SNAP consumption accompanied by formation of SULFI/NO and SSNO⁻ together with polysulfides (compound 3; n = 2–7), including monoprotonated tri-, tetra-, and pentasulfide (HS₃⁻, HS₄⁻, and HS₅⁻, respectively), sulfite (HSO₃⁻), sulfate (HSO₄⁻), and thiosulfate (HS₂O₃⁻). SI Appendix, Table S7 has details on predicted molecular masses. A.U., arbitrary unit; DEA/NO, dietylamine NONOate; SNAP, S-nitroso-N-acetyl-penicillamine; m/z, mass-to-charge ratio.

More in-depth analysis of reaction mixtures by ESI-HRMS and HPLC revealed the presence of additional anionic products, including nitrite, hyponitrite, nitrate, sulfite, sulfate, thiosulfate, and polythionates (Fig. 3C and SI Appendix, Fig. S6), with evidence for traces of a persulfide NONOate ([ONN(OH)S₂]⁻). The previously reported intermediate in the formation of SSNO⁻, thionitrite (SNO⁻)/thionitrous acid (HSNO) (9, 10, 21, 22, 26), proved impossible to be detected from RSNO/sulfide mixtures using this technique, even with cryospray ionization at −20 °C (SI Appendix, Figs. S8 and S9); although stable at very low temperature (12 K) in a frozen argon matrix (24), detection of HSNO at room temperature as presented in an earlier publication (21) is difficult to understand.

Collectively, these data suggest that the chemical foundation of the NO/sulfide cross-talk is not limited to formation of a single molecular entity but is underpinned by a mixture of compounds, including HS_n⁻ and two S/N-hybrid molecules (SSNO⁻ and SULFI/NO) along with other nitrogen oxides and sulfoxy species. With sulfide in abundance, HS_n⁻ and SSNO⁻ are the major products accumulating.

NO- and HNO-Mediated Bioactivity in Vitro.

Both SSNO⁻ and SULFI/NO may have the potential to generate NO and/or its redox congener, HNO (26, 36). We, therefore, compared the NO and HNO releasing properties of SSNO⁻-enriched mixtures (“SSNO⁻ mix”) with the properties of solutions of authentic SULFI/NO. SSNO⁻ is a potent NO donor as assessed by chemiluminescence (Fig. 4A) and a weak HNO donor as assessed by triarylphosphine trapping (Fig. 4D) (37); however, no signal was obtained using the specific nitroxyl probe P-Rhod (38) (Fig. 4B). Other methods, such as methemoglobin trapping and ferricyanide oxidation of HNO into NO, suffer from artifacts caused by reaction with sulfide (SI Appendix, Fig. S10). In agreement with earlier findings (39), we found SULFI/NO to be a weak combined NO/HNO donor compared with DEA/NO and Angeli’s salt (Fig. 4 B and C and SI Appendix, Figs. S10 and S11). Interestingly, N₂O (the main decomposition product of SULFI/NO) is generated at high yield under aerobic and anaerobic conditions on reaction of sulfide with DEA/NO (Fig. 4C) and RSNOs (SI Appendix, Fig. S11). Thus, as with acidified nitrite and sulfide (12), NO, HNO, and N₂O are common end products of the reaction of sulfide with NO or RSNOs.

Fig. 4.

NO and HNO bioactivity of SSNO⁻ and SULFI/NO in vitro. The scheme shows that the release of NO and HNO from SSNO⁻ and SULFI/NO leads to activation of sGC in cells. (A) Kinetics of NO release from SSNO⁻ after incubation of SNAP (0.1 or 1 mM) and Na₂S (1 or 10 mM) for 1 min as determined by chemiluminescence (final dilution of 1:100). (Inset) NO release from authentic SULFI/NO (100 µM) in the absence of sulfide. (B) Release of nitroxyl (HNO) as assessed by P-Rhod fluorescence from increasing concentrations of SULFI/NO (blue) and SSNO⁻ (orange; 1 mM SNAP, 10 mM Na₂S; gassed). ΔFI, fluorescence intensity-background. (C) The reaction of DEA/NO (10–200 µM) with sulfide (100 µM) generates N₂O over prolonged periods of time. (D) HNO scavenging by triphenylphosphine reveals that part of the N₂O formed during the DEA/NO/sulfide reaction originates from HNO dimerization/dehydration. ***Dunnett’s P < 0.001. (E) SSNO⁻ (20 µM) activates sGC in RFL-6 cells in both the presence and the absence of SOD, whereas equivalent concentrations of SULFI/NO (10 µM) activate sGC only in the presence of SOD after conversion of HNO into NO (n = 6–12; one-way ANOVA, P < 0.001). **Dunnett’s P < 0.01; ***Dunnett’s P < 0.001; ^§t test vs. untreated P < 0.001 (paired t test P = 0.0056); ^#P < 0.01 vs. 10 µM SULFI/NO without SOD. n.s., nonsignificant. (F) Higher concentrations of SULFI/NO (100 µM) activate sGC even in the absence of added SOD, an effect that is abolished by the NO scavenger cPTIO (500 µM) and the HNO scavenger Cys (1 mM; n = 5–10; one-way ANOVA, P < 0.001; F = 27.14). *Sidak’s P < 0.05 vs. CTRL; ***Sidak’s P < 0.001 vs. CTRL; ^#P < 0.01 vs. 100 µM SULFI/NO; ^##P < 0.001 vs. 100 µM SULFI/NO. cPTIO, 2-(4-Carboxyphenyl)-4,4,5,5-tetramethylimidazoline-1-oxyl-3-oxide; CTRL, control; Cys, cysteine; DEA/NO, diethylamine NONOate; IBMX, 3-isobutyl-1-metylxanthine; SOD, superoxide dismutase; TXPTS, triarylphosphine.

As shown recently, the SSNO⁻ mix concentration-dependently increases cGMP in RFL-6 cells (26) and relaxes aortic rings (28) in an NO- and sGC-dependent manner. In Fig. 4E, we show that low concentrations of SULFI/NO do not increase intracellular cGMP levels, except in the presence of very high concentrations of superoxide dismutase (SOD) facilitating noncatalytic, copper-mediated conversion of HNO to NO (40). The effects of SULFI/NO were quenched by trapping either HNO by cysteine or NO by cPTIO (Fig. 4F). HS_n⁻ and colloidal sulfur did not significantly increase cGMP. Considering that SSNO⁻ and SULFI/NO are the major S/N-hybrid molecules formed from sulfide and NO and that SULFI/NO is rather inefficient at releasing NO/HNO, these observations suggest that SSNO⁻ is the main carrier of bioactivity in the NO/sulfide interaction. The formation of sulfite from sulfide may be important in redox switching by virtue of sulfite’s ability to trap NO (SI Appendix, Fig. S3) and form SULFI/NO, allowing conversion of NO into HNO.

In Vivo Bioactivity: Vascular and Cardiac Effects.

The in vivo bioactivity of the SSNO⁻ mix was compared with authentic SULFI/NO by assessing effects on hemodynamics and cardiac function in rats (Fig. 5 and SI Appendix, Tables S3–S6). Acute administration of SSNO⁻ dose-dependently decreases blood pressure, whereas only the highest dose of SULFI/NO tested lowered blood pressure significantly (Fig. 5A). However, in contrast to SSNO⁻, SULFI/NO increases cardiac output, stroke volume, and aortic peak flow velocity without significant changes in heart rate, indicative of increased myocardial contractility (Fig. 5B and SI Appendix, Table S5). Continuous i.v. infusion of SSNO⁻ induces a transient but significant drop in blood pressure followed by an increase in cardiac output (Fig. 5 C and D), whereas heart rate remains constant (SI Appendix, Table S6). By comparison, continuous infusion of SULFI/NO had less of an effect on blood pressure than SSNO⁻ but produced a dramatic increase in cardiac contractility (Fig. 5 C and D); these results suggest that the positive inotropic effects of the SSNO⁻ mix may, in fact, be mediated by the presence of SULFI/NO in the reaction mixture. The i.v. bolus administration of a mixture of HS_n⁻ was found to be of similar potency at lowering blood pressure as sulfide, but HS_n⁻ seems to be endowed with a longer duration of action (SI Appendix, Figs. S12 and 13A); preliminary results suggest that they may also increase vascular compliance (SI Appendix, Fig. S13). Collectively, our results indicate that the three major reaction products of the NO/sulfide interaction all display potent bioactivity, with a distinct cardiovascular profile for each class of species.

Fig. 5.

SSNO⁻ and SULFI/NO show distinct bioactivity in rats in vivo. (A) i.v. Injection of the SSNO⁻ mix (dose range of 0.03–3 µmol/kg) leads to significant dose-dependent decreases of mean arterial pressure (MAP), whereas only the highest dose of SULFI/NO (3 µmol/kg) decreases MAP (n = 3 per group; RM two-way ANOVA, P < 0.01 for treatment and doses). *Sidak’s P < 0.01; **Sidak’s P < 0.01. At comparable doses, NaHS (1.8 and 3.5 µmol/kg) does not significantly affect blood pressure. (B) The SSNO⁻ mix (dose range of 0.03–3 µmol/kg) does not affect cardiac function, whereas higher doses of SULFI/NO (3 µmol/kg) increase cardiac contractility, which was assessed by changes in velocity time index (VTI), cardiac output (CO), and peak flow velocity (PFV; n = 3 per group; RM two-way ANOVA, P < 0.01) (SI Appendix, Tables S3–S5). (Inset) NaHS (dose range of 1.8–18 µmol/kg) does not affect VTI. ^#t Test P < 0.05. (C) Continuous infusion of the SSNO⁻ mix (0.16 µmol/kg per min) significantly decreases MAP compared with baseline, and its effect is more rapid in onset compared with that of NaHS (2.9 µmol/kg per min). Infusion of SULFI/NO equimolar to SSNO⁻ (0.16 µmol/kg per min) has only a mild, nonsignificant effect on blood pressure compared with vehicle control (SI Appendix, Table S6) (n = 3 per group; RM two-way ANOVA treatments, P = 0.00471). *Sidak’s P < 0.05 vs. baseline; **Sidak’s P < 0.01 vs. baseline; ^§P < 0.01 vs. NaHS. (D) Continuous infusion of SSNO⁻ mix (0.16 µmol/kg per minute) increases cardiac contractility (VTI, CO, and PFV) after 60 and 120 min of infusion (n = 3), whereas NaHS (2.9 µmol/kg per minute; shown in Inset) does not have any effect. SULFI/NO markedly increases cardiac contractility already 10 min after the start of infusion (n = 3; RM two-way ANOVA, P < 0.0001). *Sidak’s P < 0.05 vs. vehicle; **Sidak’s P < 0.01 vs. vehicle; ***Sidak’s P < 0.001 vs. vehicle; ^#two-way ANOVA SULFI/NO vs. SSNO, P = 0.0046. RM, repeated measurements.

Discussion

The results of this study show that sulfide reacts with NO under physiologically relevant conditions to form three main bioactive products: SSNO⁻, HS_n⁻ (where n = 2–7), and SULFI/NO. Specifically, we show that (i) in vivo sulfide administration modulates endogenous NO bioavailability; (ii) SSNO⁻ is formed by intermediate polysulfide formation, accumulates at higher sulfide concentrations, is resistant to attack by other thiols, acts as an efficient NO and sulfane sulfur donor, and lowers blood pressure; (iii) HS_n⁻ are also products of the NO/sulfide interaction, are intermediates as well as decomposition products of SSNO⁻, and lower blood pressure; (iv) SULFI/NO is a weak combined NO/HNO donor that has only mild blood pressure lowering but very pronounced positive inotropic effects; and (v) various other reaction products, including nitrogen oxides and sulfoxy species, are formed and likely also contribute to the bioactivity of both NO and sulfide. Importantly, our results suggest that the fates of NO and sulfide are intimately intertwined wherever they are cogenerated in biology; as a corollary, the biological actions of NO and H₂S can be neither considered nor studied in isolation, because they form a network of coupled chemical reactions that gives rise to formation of multiple new chemical entities with distinct bioactivity profiles. This unexpectedly rich chemistry would seem to provide nature with ample opportunities for modulation of various fundamental biological and pathophysiological processes related to, for example, electron transfer, sulfur trafficking, and redox regulation.

Chemistry of the Reaction Between Sulfide and NO.

A reaction between NO and sulfide was first described almost a century ago (refs. 11, 41, and 42 and references therein). The main products of this reaction were found to be nitrogen oxides (NO, N₂, and N₂O; measured in the gas phase) and sulfur-bearing anionic solutes, including SSNO⁻ and HS_n⁻ (9, 10). These are essentially the same products that were later described in the reaction of RSNOs with sulfide (26, 27) without an understanding of detailed reaction mechanisms. SULFI/NO formation was described by others (9, 42–44), although only in one case (9) in the context of sulfide. We here describe a singular process that occurs under biologically relevant conditions and characterizes the interaction between NO or RSNOs and sulfide. This process comprises cascading chemical reactions, giving rise to formation of three main anionic products (SSNO⁻, HS_n⁻, and SULFI/NO) accompanied by the release of H₂S, NO, HNO and N₂O, NO₂⁻, and NO₃⁻ as well as sulfoxy species, such as SO₄²⁻, S₂O₃²⁻, SO₄²⁻, and polythionates. Identification of the S/N-hybrid species was achieved using a variety of techniques, including ESI-HRMS combined with stable isotope labeling.

On the basis of the products identified here and elsewhere (9–11, 41–44), it is possible to propose a unified reaction scheme (summarized in Fig. 6 and explained in more detail in SI Appendix). Although it is not difficult to rationalize how HS⁻ (or the even stronger nucleophile HSS⁻) reacts with RSNOs to form HSSNO/SSNO⁻ (13, 26), formation of the latter from NO and sulfide is less straightforward. The simplest assumption is that, under aerobic conditions, sulfide is nitrosated by NO autoxidation products (NO₂ and N₂O₃), initially leading to transient formation of HSNO with subsequent attack by another HSS⁻ to form SSNO⁻ (reactions 1–3 in Fig. 6). However, our results with DEA/NO and aqueous NO solution clearly show that the reaction also occurs (to about the same extent, although with slower kinetics) under strict exclusion of oxygen, showing that there must be another route to formation of SSNO⁻. A second possibility involves formation of thiyl and perthiyl radicals (S^•− and SS^•−) (reactions 4 and 6 in Fig. 6), which can trap NO directly to form SNO⁻ and SSNO⁻ (reactions 5 and 7 in Fig. 6). Formation of S^•− can take place by one-electron oxidation of sulfide after (i) reaction with O₂ (rather inefficient) or (ii) transition metal catalysis (reaction 4 in Fig. 6); alternatively, S^•− and SS^•− might be also formed by HSNO or SSNO⁻ hemolysis after these species are formed through other routes (reactions 16 and 17 in Fig. 6) or by homolytic cleavage of long-chain HS_n⁻ (e.g., reaction 20 in Fig. 6), although the driving force for HS_n⁻ decomposition to anion radicals is unclear. Nevertheless, as shown here and according to reactions 3 and 7 in Fig. 6, HS_n⁻ (or its radical intermediates) are substrates of SSNO⁻ synthesis. Previous work has shown that relatively pure solutions of SSNO⁻ in acetone can be obtained by passing NO through solutions of inorganic/organic polysulfide as evidenced by ¹⁵N NMR (10) and X-ray crystallographic studies (35).

Fig. 6.

Chemical reaction cascade depicting pathways of formation and decomposition of SSNO⁻, HS_n⁻, and SULFI/NO. Reactions are not mass balanced; numbers in red refer to the reactions described in the text. A more detailed discussion of the reaction mechanisms can be found in SI Appendix. The pK_a values of HS• and HSS• are unknown but likely exceed 7; therefore, at pH 7.4, these species are radical anions.

In addition to HS_n⁻ and SSNO⁻, we identified SULFI/NO (36, 42, 45) as another important product of the reaction of sulfide with either NO or RSNOs. We also found traces of [ONN(OH)S₂]⁻ as predicted by Seel and Wagner (9). A likely precursor of SULFI/NO formation from NO and sulfide is sulfite (SO₃²⁻) (reaction 15 in Fig. 6), which was suggested by others not only for the reaction of sulfide with NO under anoxia (9, 11) but also, as a result of the reaction of NO with thiosulfate (S₂O₃²⁻) and S₂O₄²⁻ (11). In aqueous solution, sulfite may be formed by reaction of sulfide with O₂ by formation of SO₂^•− and S₂O₄²⁻ (46) (reactions 9–12 in Fig. 6), formation of S₂O₃²⁻ originating from radical reactions (by SO₂^•−) (reactions 9–13 in Fig. 6), or hydrolysis of HONS (a product of HSNO isomerization) as proposed by Goehring and Messner (23). Because no SO₃²⁻ was detected by ESI-HRMS in freshly prepared sulfide stock solutions and sulfide autoxidation is a rather slow process, production of S₂O₃²⁻ through radical reactions and/or HONS/HSNO seems to be the most plausible route.

The main pathways of decomposition of the key reaction products are as follows (Fig. 6): SSNO⁻ can undergo homolytic cleavage to SS^•− and NO^• (reaction 17 in Fig. 6) or secondary reaction with sulfide (reaction 19 in Fig. 6), which according to Seel and Wagner (10), is not efficient, because they observed an equilibrium between the trisulfide radical and NO in sealed systems, making SSNO⁻ rather stable in excess sulfide (10). HS_n⁻ may undergo polymerization reaction to cyclo-octasulfur (S₈) and/or homolytic cleavage to form sulfur radicals (see above) (reaction 20 in Fig. 6); in the presence of strong nucleophiles (such as DTT), polysulfides undergo decomposition to sulfide, a property that we here used to distinguish them from SSNO⁻. Decomposition of SULFI/NO leads to formation of SO₃²⁻, SO₄²⁻, and N₂O (reactions 21–24 in Fig. 6). Although it has been proposed that this process occurs directly (reaction 21 in Fig. 6) (36), the formation of both HNO and NO from hyponitrite (ON:NOH⁻) has been proposed by others (reactions 22 and 23 in Fig. 6) (42–44). Indeed, evidence for trace levels of ON:NOH⁻ in incubation mixtures of both SULFI/NO and DEA/NO (but not SNAP) with sulfide was observed by us using ESI-HRMS (SI Appendix, Fig. S6). Authentic sodium hyponitrite was found to release NO at pH 7.4, which was markedly enhanced in the presence of the one-electron oxidant ferricyanide, consistent with the generation of both NO and HNO (SI Appendix, Fig. S7) (47). ON = NOH⁻ is known to decompose to N₂O and water after protonation to hyponitrous acid (HNO = NOH).

Parallel reactions lead to formation of HNO, superoxide (O₂^•−), and peroxynitrite (ONOO⁻). These reactions include reduction of NO to HNO by HS_n⁻, very potent radical reducing/trapping agents (48), and oxidation of O₂ or NO to O₂^•− or ONOO⁻, respectively, by reaction with sulfur-centered radicals. Moreover, ONOO⁻ may also be formed after reaction of HNO with O₂; these possibilities are consistent with the rapid consumption of O₂ in the reaction (26). Finally, HNO can be reduced by sulfide to hydroxylamine (NH₂OH), and nitrosation of NH₂OH results in formation of N₂O (12).

Chemical Biology of SSNO⁻, Polysulfides, and SULFI/NO.

The fact that the above reactions occur in aqueous buffers at pH 7.4 does not necessarily mean that they are relevant to biology. What follows is a discussion about whether the chemical properties of SSNO⁻, HS_n⁻, and SULFI/NO are compatible with the biological situation, such that these molecules can conceivably play a role in cell signaling and/or interorgan transport of NO and sulfur equivalents.

SSNO⁻ is a bioactive product with unusual properties: it is stable in the presence of millimolar concentrations of other thiols [sulfide, (homo)cysteine, and glutathione], carries and releases NO, and generates HS_n⁻ on decomposition. These properties suggest it would be sufficiently stable if formed in the cellular milieu to participate in and contribute to NO/sulfur trafficking (13, 26). These properties clearly distinguish SSNO⁻ [and its protonated form perthionitrous acid (HSSNO)] from SNO⁻/HSNO (13, 26). SSNO⁻ is considerably more stable than HSNO, because the latter undergoes rapid isomerization, hemolysis, and polymerization (13). Moreover, HSNO can undergo rapid nucleophilic attack by HS⁻ to give HSSH and NO⁻ (reacting with O₂ to form ONOO⁻), whereas the electron density of the proximate sulfur in HSSNO is increased, making it less susceptible to nucleophilic attack by HS⁻ and reaction with metals (e.g., Cu⁺). Of note, SSNO⁻ exerts faster and more prominent vasorelaxation compared with its precursor and the prototypical nitrosothiol, GSNO (28), indicative of the ease with which it can cross cell membranes and release NO.

HS_n⁻ (and their organic counterparts) are not resistant to reducing conditions and high concentrations of thiols, and contain a highly reactive sulfane sulfur (49, 50). Polysulfides were proposed to be the bioactive molecules responsible for “H₂S signaling” and much of the physiological action of sulfide in cells (31, 48, 49). In addition, they are believed to act as storage or buffer molecules of H₂S, with surprisingly high concentrations measured in murine organs/tissues (31). Although their chemistry is reasonably well-understood (48, 50), biosynthetic pathways, speciation, and modes of action in the biological environment are still poorly defined. The results of this study suggest that HS_n⁻ are also intermediates in SSNO⁻ formation, possibly through homolytic cleavage to radical species and subsequent reaction with NO. Thus, in addition to their enzymatic formation (33) and as a consequence of sulfide oxidation (50), reaction of sulfide with NO or nitrosothiols may provide another pathway to their production in cells and tissues.

SULFI/NO arises as a result of the trapping of two molecules of NO by sulfite (36, 39, 45), a reaction originally described by Davy in 1802. The substance is also known as dinitrososulfite (51), “Pelouze’s salt,” or “Stickoxid-sulfite” in the German literature (ref. 42 and references therein). Its structural characteristics were revealed about 100 years later using IR spectroscopy (51) and confirmed by alternative chemical synthesis (42). It belongs to a class of compounds now better known as diazeniumdiolates or NONOates (36). Consistent with earlier reports, we found that SULFI/NO has similar stability characteristics as other diazeniumdiolates (36, 39, 45, 52), cogenerates small amounts of NO and HNO at pH 7.4, and releases N₂O in high yields. An unexpected yet inevitable consequence of its formation in the NO/sulfide system is that SULFI/NO is formed (in addition to SSNO⁻ and HS_n⁻) whenever NO and sulfide are cogenerated; sulfite generation, thus, results in the scavenging of NO, with consecutive redox conversion of part of this NO to HNO. This unexpected chemistry propels this molecule, a rather ineffective NO donor and vasodilator (39), from the chemists’ curiosity cabinet to the forefront of biological signaling. Proof for its formation in living organisms will require development of sufficiently sensitive analytical techniques to monitor its formation and fate, but the likelihood of its existence in real life is intriguing.

Taken together, the chemical biology of the three major products of the NO/sulfide reaction system ranges from sulfane sulfur signaling (SSNO⁻ and polysulfides) through NO release (SSNO⁻ and SULFI/NO) and NO scavenging (sulfite) to nitroxyl signaling (SULFI/NO), with generation of several other sulfoxy and nitrogen oxide metabolites with known and distinct bioactivity profiles.

Bioactivity of SSNO⁻, Polysulfides, and SULFI/NO.

The different chemical properties attributed to SSNO⁻, HS_n⁻, and SULFI/NO translate into distinct bioactivities in vitro and in vivo. As shown here and elsewhere (26, 28), the SSNO⁻ mixture (produced from either DEA/NO or SNAP with excess sulfide) contains high concentrations of SSNO⁻ and HS_n⁻, releases NO, is a potent sGC stimulator in the NO reporter cell line RFL-6 (26), induces NO-mediated vasorelaxation in isolated aortic tissue (28), and significantly decreases blood pressure in a dose-dependent fashion with minor effects on cardiac function. No changes in any of these parameters were observed by equivalent concentrations/doses of sulfide alone. Control experiments with HS_n⁻ suggest that NO-independent effects on blood pressure may be, in part, dependent on the presence of HS_n⁻ in the mixture but independent of SULFI/NO. Only rather high (>100 µM) concentrations of SULFI/NO (but not HS_n⁻ or S₈) activated sGC in RFL-6 cells; these effects were fully inhibited by addition of cysteine (HNO scavenger) or cPTIO (NO scavenger) and potentiated by high but not low concentrations of SOD converting HNO into NO (40, 47). These results are consistent with its predicted chemical biology as a relatively weak combined NO/HNO donor. However, SULFI/NO dramatically increased cardiac output, stroke volume (measured as the velocity–time integral), and peak blood flow velocity while affecting blood pressure and heart rate only minimally (except at high doses), indicative of its propensity to enhance cardiac contractility, an effect likely mediated by the generation of HNO.

In aggregate, these results show that SSNO⁻ is a potent NO donor mainly affecting peripheral vascular resistance, whereas SULFI/NO is a rather weak NO/HNO donor that affects the tone of resistant vessels to a lesser degree but potently increases cardiac contractility. Although SSNO⁻ is a major reaction product in NO/sulfide-containing mixtures, one important limitation worth highlighting is that, at present, the biological properties of SSNO⁻ cannot be studied separately from SULFI/NO, sulfide/HS_n⁻, and other reaction products formed. Moreover, the rich possibilities for interaction between defined constituents of the reaction mixture may profoundly affect the chemistry and bioactivity of individual species in the biological setting. For example, by incubating SULFI/NO with sulfide, we observed formation of HS_n⁻. Nevertheless, our results clearly show that NO and sulfide react with each other; wherever cogenerated, their reaction leads to formation of molecules with bioactivities distinct from the parent compounds, resulting in quenching, redox switching, or release/transport of NO bioactivity.

What Is the Significance of These Reactions for the Sulfide/NO Cross-Talk in Vivo?

Our in vivo data suggest that the well-known scavenging and potentiating effects of sulfide on NO bioavailability in vitro (19, 26) and vascular function in vivo (18) correlate with changes in NO bioavailability. In the first 30 min of sulfide infusion, we observed a drop in NO-heme levels within erythrocytes, whereas no significant changes in hemodynamic parameters were apparent; after 1 h of infusion, allowing sulfide accumulation (53), increases in systemic NO bioavailability (i.e., levels of NO-heme and nitroso species) were accompanied by corresponding decreases in blood pressure. These results are consistent with NO scavenging effects after the addition of low concentrations of sulfide as detected by chemiluminescence as well as in RFL-6 cells and other cell types as shown here and elsewhere (19, 26). A decrease in circulating nitrite and RXNO levels was observed in mice lacking the sulfide-producing enzyme cystathionine-γ-lyase (54); however, this result was ascribed to effects of sulfide on endothelial NO synthase activity. However, effects of sulfide on NOS activity in this context are unlikely, considering that the effects of NOS inhibition on circulating and tissue NO metabolites are less rapid than the functional effects of sulfide observed.

The molecular basis that accounts for scavenging and potentiation of NO bioactivity by sulfide is presently unclear. These effects have been ascribed to formation of a nitrosothiol (19), mutual regulation of enzymatic pathways (e.g., endothelial nitric oxide synthase) (54), modulation of PDE (55), and/or targeting of different vascular beds (56). An interesting alternative possibility that would seem to warrant additional investigation is that sulfide might modulate redox-dependent control mechanisms, regulating vascular tone independently of sGC, such as, for example, through modulation of the redox state of PKGIα by oxidative modification of Cys42 (57). This notion would be consistent with the observation that the vasoactive effects of sulfide are partially attenuated in aortic rings from PKG1^−/− mice (58). It is also possible that the effects of sulfide are caused by a chemical interaction by (i) formation of S-nitroso species (such as SSNO⁻) acting as a temporary “NO sink” that is able to release NO farther downstream; (ii) formation of reactive sulfoxy intermediates, such as SO₂^• or SO₃²⁻, forming NO complexes with a low NO-releasing potential, such as SULFI/NO; or (iii) formation of thiyl radicals (such as HS^• and HSS^•) by reaction with O₂. The latter is consistent with our earlier observation that O₂ is consumed in the reaction (26) and would lead to formation of O₂^•− and NO scavenging as a result of peroxynitrite (ONOO⁻) generation. Both formation of sulfoxy species and O₂ consumption are interesting observations not only from a chemical perspective but also, because mitochondrial sulfide oxidation has been linked to physiological oxygen sensing (59) and SO₃²⁻/S₂O₃²⁻ are reduced to HS⁻ in mitochondria (59). Final proof of the relevance of these chemical pathways in biological systems will require the detection of these S/N-hybrid molecules, as shown for RSSH (31), in cells and tissues. Similarly important in this context will be a careful kinetic assessment of the reaction yields depending on the rates of formation of NO and sulfide in subcellular compartments. Because such evaluation does not seem to be possible using current analytical techniques, the development of a novel quantitative/nondestructive methodology for in situ detection and monitoring of those molecules in biological matrices merits additional investigation.

Summary and Conclusions

Understanding the basic chemical principles that govern the interactions between sulfide and NO is essential to interpreting and untangling the conflicting observations about mutual potentiating and inhibitory effects presented in the literature. Here, we show that the reaction between sulfide and NO leads to formation of different bioactive intermediates (including SSNO⁻, HS_n⁻, and SULFI/NO) capable of scavenging, transporting, and releasing NO and generating its redox congeners HNO, N₂O, and sulfane sulfur. Each of these products is characterized by a specific biological chemistry and the potential to release other bioactive mediators. With the exception of SULFI/NO, the bioactivity of HS_n⁻ and SSNO⁻ cannot be studied in isolation and/or the absence of sulfide at present. Nevertheless, by careful comparison of the effects of different mixtures of these components, we conclude that SSNO⁻ is a potent NO donor, resistant to the reducing milieu of the cell, and able to release both NO and HS_n⁻. SULFI/NO is a weak combined NO/HNO donor and generator of N₂O with potent effects on the heart. Formation of its precursor sulfite and generation of sulfur and/or oxygen-centered free radicals may be responsible for the scavenging effects of sulfide on NO bioavailability. Polysulfides may be formed secondary to the reaction of sulfide with NO, either through HSNO or after decomposition of SSNO⁻, and may also contribute to NO scavenging and sulfane sulfur signaling. Although admittedly even more speculative, some of this chemistry may also help explain the “Janus-like” face of NO (i.e., the often opposing biological effects of low and high concentrations observed in physiology and pathophysiology). In any case, our findings open the door to a new field of research with SSNO⁻, HS_n⁻, and SULFI/NO taking center stage as biologically important mediators of both the NO and H₂S transduction pathways. Although the cardiovascular system has been the target of our current efforts, this chemical interaction is likely to be relevant to cell/organelle signaling in many other systems, including neuronal and immune cells, plants, and prokaryotes as exemplified by the recent work on antibiotic resistance by bacterial NO/H₂S production (17). Beyond its likely significance for biology and redox signaling, our results may also be of significance for environmental chemistry pertinent to marine and atmospheric processes.

Materials and Methods

A complete and more detailed description of the materials and methods used is provided in SI Appendix.

Preparation of Stock Solutions of NO Donors, SULFI/NO, Sulfide, and SSNO⁻.

The potassium salt of SULFI/NO was synthesized as described by Drago and coworkers (60). Stock solutions of DEA/NO, DETA/NO, Sper/NO, SULFI/NO, and Angeli’s salt were freshly prepared in 0.01 M NaOH, diluted in PBS, and used immediately. Aqueous stock solutions of ¹⁵N-SNAP or SNAP were freshly prepared either from crystalline material or through reaction of the reduced thiols with acidified nitrite (61) and used immediately. Saturated aqueous solutions of NO were prepared and kept sealed under argon as described (62). Sulfide stock solutions for in vitro use (e.g., in cell culture experiments) were prepared fresh before each experiment by dissolving anhydrous Na₂S in a strong buffer (1 M Tris or phosphate buffer, pH 7.4) and diluting further in 100 mM Tris or 50–100 mM phosphate buffer (pH 7.4) immediately before use. For in vivo experiments, NaHS stock solutions were prepared in 300 mM phosphate buffer (pH 7.4), diluted in PBS (pH 7.4), and used immediately (10 mg/mL = 176.7 mM NaHS; anhydrous; Alfa Aesar). Although all precautions were undertaken to avoid the presence of products of sulfide oxidation in the stock solutions, including polysulfide, their presence in stock solutions cannot be excluded. Stock solutions of SSNO⁻ were prepared by reacting 1 mM SNAP with 10 mM Na₂S in 1 mL PBS or 100 mM Tris (pH 7.4). After 10 min of incubation at RT in the dark, excess sulfide was removed by 10 min of bubbling with N₂. The (theoretical) maximal yield of SSNO⁻ under these conditions is 1 mM, corresponding to the concentration of added nitrosothiol (SI Appendix, Fig. S5 shows the experimental determination of reaction yield). For in vivo experiments, the SSNO⁻ solution was prepared in phosphate buffer (pH 7.4) and further diluted in PBS to obtain the doses indicated for bolus i.v. injection or used directly for continuous i.v. infusion.

Effects of Sulfide, SSNO⁻, SULFI/NO, and Polysulfides on Cardiovascular Hemodynamics and Circulating NO Metabolites.

All animal experiments were approved either by the Institutional Animal Care and Use Committee (IACUC) at Boston University School of Medicine (Boston, MA), the University College London (London, UK), or the State Veterinary and Food Administration of the Slovak Republic. All procedures were conducted in male Wistar rats (250–300 g; Charles River) anesthetized with 2% (vol/vol) isofluorane. Briefly, the effects of bolus i.v. injection of increasing doses of NaHS (1.8–18 µmol/kg), SSNO⁻ mix (0.03–3 µmol/kg), or SULFI/NO (0.03–3 µmol/kg) in PBS or vehicle alone on cardiovascular hemodynamics and circulating NO stores were assessed at 10-min intervals. Continuous i.v. infusions of NaHS (2.8 µmol/kg per min in PBS), SSNO⁻ (0.16 µmol/kg per min in PBS), SULFI/NO (0.16 µmol/kg per min in 25 mM NaOH, 0.9% NaCl), or the respective vehicle (PBS or 25 mM NaOH, 0.9% NaCl) were performed at a rate of 10 mL/kg per hour through the right internal jugular venous line. Continuous monitoring of mean arterial pressure and blood withdrawal was from an indwelling arterial (left common carotid) line. Blood was taken at defined time points (0, 5, 10, 30, 60, and 120 min) and processed as previously described for determination of NO metabolite concentrations by ion chromatography and gas-phase chemiluminescence (63). Cardiac function and heart rate were assessed by transthoracic echocardiography using a Vivid 7 (GE Healthcare) as described (64). A rectal probe (TES Electrical Electronic Corp.) inserted 3 cm in depth was used to measure core temperature.

Effects of Sulfide on NO Bioactivity and Detection of NO Release and Bioactivity of SSNO⁻ and SULFI/NO in RFL-6 Cells.

The effects of sulfide on Sper/NO-mediated activation of sGC were determined by measuring changes in intracellular cGMP levels in RFL-6 cells pretreated with a phosphodiesterase inhibitor (500 µM 3-isobutyl-1-metylxanthine), and then treated with 100 µM Sper/NO for 20 min in the absence or presence of increasing concentrations of sulfide (1, 10, and 100 µM Na₂S in 100 mM Tris, pH 7.4) as described (26). The NO/HNO bioactivities of SSNO⁻ and SULFI/NO were compared by treating 3-isobutyl-1-metylxanthine–pretreated cells (as above) with either 20 µM SSNO⁻ mix or 1, 10, and 100 µM SULFI/NO for 20 min. To test for HNO bioactivity, ∼7,000 U/mL SOD was added directly to the treatment medium 5 min before addition of SSNO⁻ or SULFI/NO to enable extracellular conversion of HNO into NO. In select experiments, an NO scavenger (cPTIO; 500 µM) or a nitroxyl scavenger (1 mM Cys) was added 5 min before the other treatments. Intracellular cGMP levels and protein content were assessed in cell lysates using a DetectXHigh Sensitivity Direct cGMP Kit (Arbor Assay; Biotrend) and Roti⁻Nanoquant (Carl Roth GmbH + Co. KG), respectively. Data were normalized for protein content and expressed as folds of untreated control to further account for the variability in sGC expression levels of RFL-6 cells of different batches and passages.

Detection of SSNO⁻, SULFI/NO, and Polysulfides by UV-Visible Spectroscopy and HRMS.

The spectroscopic and kinetic behaviors of the reaction between sulfide (1–10 mM) and NO (0.2 mM), DEA/NO (0.5–3 mM), or SNAP (0.2–1 mM) were followed by rapid scanning UV-visible spectroscopy as described (26). The identification of the reaction products was achieved by HRMS using an LTQ Orbitrap XL Hybrid Linear Ion Trap–Orbitrap Mass Spectrometer equipped with a nanospray ionization source controlled with XCalibur 2.1 (Thermo-Fisher). Samples for HRMS were mixed in 50 mM (NH4)₃PO4 buffer (pH 7.4), mixed 1:5 with acetonitrile through a T piece, and infused directly into the ion source. Spectra were acquired in negative ion mode with a spray voltage of 5 kV and nitrogen as sheath gas; capillary temperature was set at 300 °C, and capillary voltage was set at 20 V. Instrument parameters, especially those of the ion optics, were optimized for each individual compound of interest. Elemental analysis based on accurate mass and a priori information of likely elemental composition and isotope distribution simulation were performed using XCalibur 2.1.

Time-Resolved Measurement of NO Trapping and NO Release by Chemiluminescence.

Trapping by Na₂S (33.4 and 334 µM) of NO released from DETA/NO (33.4 µM) and NO released from SSNO-containing mixtures (1–100 µM) after 1 min of incubation of SNAP and sulfide, SULFI/NO (100 µM), and SNAP (10 µM) alone was monitored by gas-phase chemiluminescence (CLD 77:00 AM sp; Ecophysics) using a custom-designed, water-jacketed glass reaction chamber (15 mL total volume) continuously bubbled with nitrogen or air as described elsewhere (26).

Detection of HNO Release by P-Rhod Fluorescence.

HNO release form SSNO⁻ (0.001–1 mM) and SULFI/NO (0.001–1 mM) was determined by using the HNO-specific probe P-Rhod (stock 50 mM in DMSO) (38). Briefly, N₂ gassed SSNO⁻ mix (1 mM SNAP, 10 mM Na₂S), SULFI/NO (1 mM), or Angeli’s salt (1 mM; all in 100 mM Tris⋅HCl, pH 7.4) were serially diluted in Tris⋅HCl in a dark 96-well plate. Buffer alone was used as blank. P-Rhod (5 µM) was added to all wells using an automatic injector, and fluorescence changes were recorded at excitation of 480 nm and emission of 520 nm using a multimode plate reader (FLUOstar Omega; BMG Labtech). Data are reported as percentage increases compared with background signal (blank).

Nitrous Oxide Quantification by GC.

Stock solutions of DEA/NO (10–200 µM) were mixed with Na₂S (100 µM) in phosphate buffer, injected into a 10-mL round-bottom flask sealed with a rubber septum, flushed with either N₂ or air, and incubated at 37 °C, and at the indicated time points, headspace aliquots (100 μL) were injected through a gas-tight syringe onto a 7890 A Agilent Gas Chromatograph equipped with a microelectron capture detector and a 30 × 0.32-m (25 μm) HP-MOLSIV Capillary Column. The retention time of nitrous oxide was 3.4 min, and yields were calculated based on a standard curve for nitrous oxide (Matheson Tri-Gas). Angeli’s salt was used as the reference compound for HNO. In some experiments, 200 µM triarylphosphine was used to trap HNO.

Statistical Analysis.

Data are reported as means ± SEMs. ANOVA followed by an appropriate posthoc multiple comparison test was used to test for statistical significance.