Garlic oil polysulfides: H₂S- and O₂-independent prooxidants in buffer and antioxidants in cells

Abstract

The health benefits of garlic and other organosulfur-containing foods are well recognized and have been attributed to both prooxidant and antioxidant activities. The effects of garlic are surprisingly similar to those of hydrogen sulfide (H₂S), which is also known to be released from garlic under certain conditions. However, recent evidence suggests that polysulfides, not H₂S, may be the actual mediator of physiological signaling. In this study, we monitored formation of H₂S and polysulfides from garlic oil in buffer and in human embryonic kidney (HEK) 293 cells with fluorescent dyes, 7-azido-4-methylcoumarin and SSP4, respectively and redox activity with two redox indicators redox-sensitive green fluorescent protein (roGFP) and DCF. Our results show that H₂S release from garlic oil in buffer requires other low-molecular-weight thiols, such as cysteine (Cys) or glutathione (GSH), whereas polysulfides are readily detected in garlic oil alone. Administration of garlic oil to cells rapidly increases intracellular polysulfide but has minimal effects on H₂S unless Cys or GSH are also present in the extracellular medium. We also observed that garlic oil and diallyltrisulfide (DATS) potently oxidized roGFP in buffer but did not affect DCF. This appears to be a direct polysulfide-mediated oxidation that does not require a reactive oxygen species intermediate. Conversely, when applied to cells, garlic oil became a significant intracellular reductant independent of extracellular Cys or GSH. This suggests that intracellular metabolism and further processing of the sulfur moieties are necessary to confer antioxidant properties to garlic oil in vivo.

garlic and other organosulfur-bearing plants have been used by humans for medicinal purposes for at least 6,000 years (32). The wisdom of the Ancients has largely been borne out by current research, and in research animals, if not always in humans, there is accumulating evidence that garlic and organosulfur compounds have a myriad of beneficial effects. These have been summarized in a number of recent reviews (1, 3, 4, 7, 18, 38, 48, 52, 61, 64, 66).

Considerable interest in garlic metabolism has focused on the pathway whereby the endogenous precursor alliin (S-allyl-l-cysteine sulfoxide) is metabolized to allicin by the carbon-sulfur lyase enzyme, alliinase, the latter being released when garlic cells are damaged. Allicin rapidly undergoes nonenzymatic decomposition into diallyl monosulfide (DAS) and oil-soluble polysulfides, most notably diallyl disulfide (DADS) and diallyl trisulfide (DATS; 54). Much of the biological activity of these polysulfides has been attributed to DATS; the purported biological activity of DADS being questioned by purported DATS contamination common in commercial preparations (30).

The mechanism through which DATS exerts its biological activity has not been completely resolved, but recent attention has focused on release of hydrogen sulfide (H₂S). This has been demonstrated in buffer (2) and in the presence of red blood cells (2) or homogenized heart tissue (47). H₂S release from DATS in buffer is not spontaneous but can be accomplished nonenzymatically in the presence of another organosulfur such as glutathione (GSH), cysteine (Cys), or homocysteine (Hcys) with the relative potency, GSH > Cys > Hcys (2). H₂S is not released from DAS, and it is not clear whether its release from DADS (2) is also due to DATS contamination.

Interest in H₂S as the active component in garlic not only stems from its release from DATS and other polysulfides but also lies in the almost uncanny similarity in their biological actions. Both H₂S and garlic oil have been shown to be cardioprotective (6, 13, 18, 25, 26, 31, 35, 67) and contribute to ischemic conditioning (45, 65). Both are anti-inflammatory (7, 56), anti-atherosclerotic (29, 33, 62), induce Ca²⁺ influx in astrocytes (22), and have been reported to act as potent antioxidants (5, 9, 15, 21, 49, 53, 59, 60) and decrease oxidative stress (28, 31, 68).

It is clear from the above that many of the actions of the organosulfur compounds have been linked to their antioxidant properties. Paradoxically, however, many of the effects of these compounds (e.g., opening vascular K_ATP channels, nuclear translocation of Nrf2 in activation of antioxidant defenses, reduction in TNF-α in Parkinsonism, to name a few) are best explained by a prooxidant effect on reactive protein cysteines (see below), or by overall increases in oxidative stress (27, 46, 51, 57).

Both prooxidant and antioxidant activities have been ascribed to H₂S and related polysulfides (H₂S_n; n = 2–8), although there is increasing evidence that their prooxidant effects are considerably more prevalent than originally thought (8, 40). Arguably, the greatest interest in H₂S and polysulfide signaling is in their ability to bind to a protein cysteine through a process called sulfhydration, or more appropriately persulfidation (11), and, thereby, change the functional properties of that protein (11, 19, 20, 37, 44). At least 15 structural proteins or enzymes have been shown to be modified through this process (44), and many of these appear to be proteins that are also activated or inactivated by organosulfur compounds. Polysulfides can directly persulfidate protein cysteines, whereas H₂S cannot as both the cysteine and H₂S sulfur atoms are in their most reduced state. H₂S sulfuration requires a two-electron oxidation of either sulfur or a one-electron oxidation of both; the mechanism through which this occurs is only beginning to be revealed, although the outcome is well documented (23). Although H₂S may reduce a very select number of disulfide bridges in proteins (37), there is relatively little evidence to support the common assumption that it is a potent reductant, as H₂S does not readily react with peroxide (39).

We (8) recently observed that H₂S and related polysulfides readily oxidize a number of indicators commonly used to detect redox state and reactive oxygen species, including redox-sensitive green fluorescent protein (roGFP), 2′,7′-dichlorofluorescein (DCF), MitoSox Red, and Amplex Red. In the present studies, we used two of these indicators, roGFP and DCF, along with specific fluorescent indicators of H₂S (7-azido-4-methylcoumarin, AzMC) and polysulfides (SSP4) to examine H₂S and polysulfides release from garlic oil and DATS and to determine their potential to act as an oxidant or reductant. We also examined H₂S and polysulfide production in cells exposed to garlic oil in normoxia and hypoxia. To our knowledge, this study is the first to directly measure the effects of garlic oil on intracellular H₂S and polysulfides and to evaluate the redox properties of garlic oil in buffer and cells.

MATERIALS AND METHODS

Chemicals.

SSP4 was generously provided by Dr. Ming Xian, Washington State University. DCF and LB Broth, Miller were purchased from ThermoFisher Scientific (Grand Island, NY) and roGFP from the University of Oregon, Department of Health Sciences (Eugene OR). All other chemicals were purchased from Sigma-Aldrich (St. Louis, MO).

General protocol.

Buffer and the reporting fluorescent indicator were aliquoted into black 96-well plates in a darkened room. A black cover with a parafilm liner was placed over the plates to further minimize photobleaching and reduce H₂S volatization. The edge of the plate was also wrapped with parafilm. Baseline fluorescence was measured on a SpectraMax M5e plate reader (Molecular Devices, Sunnyvale, CA) every 10 min over 30 min (t = −30, −20, −10 min). The plates were then uncovered, the experimental compounds were added, the plates were resealed, and fluorescence was recorded an additional 100 min (t = 0–90 min) with sampling at 10-min intervals. To reduce well-to-well variation, the samples were normalized to the baseline fluorescence at t = −10 min (10 min prior to addition of drugs). All experiments were done at room temperature in a darkened room. A minimum of three replicates were obtained. Hypoxia experiments were performed with the plate reader in a model 856-HYPO hypoxia chamber (Plas Labs, Lansing, MI) under 100% N₂, which lowered the ambient O₂ to less than 0.35%. Under normal barometric conditions (∼747 ± 2 mmHg), this produced an O₂ concentration less than 3.8 μM in the buffer. Note that because of occasional malfunctions in our plate reader, we used a colleague's plate reader for several of the normoxia experiments. Because this instrument was set up differently, the absolute fluorescence was not identical to ours; however, this did not affect the relative changes in fluorescence, and suitable controls were run with each experiment.

H₂S and polysulfide production from garlic oil and DATS.

Varying concentrations of garlic oil blend consisting of 30–50% by weight of DADS, 10–13% DATS, and 5–13% allyl sulfide were used. Pure DATS was also examined for comparison. AzMC (10 μM) and SSP4 (50 μM) were used to detect H₂S and polysulfides, respectively. 2 mM Cys and 2 mM GSH were added to release H₂S. In pilot experiments, we confirmed that AzMC was insensitive to polysulfides and that SSP4 was insensitive to H₂S.

roGFP.

Plasmid containing a roGFP2 sequence, an ampicillin resistance sequence, and a lac operon promoter sequence were transfected into BL21D3 Escherichia coli cells. Transfection success was verified via purification and spectroscopy.

To purify the roGFP protein, three colonies were chosen at random and grown in LB Broth, Miller in 100 μM ampicillin to an optimum optical density of 0.800 absorbance units at 600 nm. The cells were then pelleted using an Avanti J-30I centrifuge at 6,000 g for 6 min. The pellet was weighed, resuspended in a 5× volume of lysis buffer for 20 min, and then sonicated to complete cell lysis. The lysate was then centrifuged at 50,000 g for 30 min, and the supernatant was successively filtered through 0.8-μm and 0.45-μm syringe filters. Ni-NTA agarose solution was added to the filtered supernatant, and after 2 h, the supernatant was passed through a ThermoScientific 5-ml polypropylene column, washed, and eluted.

From initial studies, the protein appeared to be nearly completely oxidized (fraction oxidized = 1.0) once purified. To obtain a partially reduced protein, the roGFP was incubated with 1 mM DTT for 20 min and then dialyzed with two rinses overnight in 100% N₂ sparged Sorenson's (phosphate) buffer (pH of 7) using Spectra/Por membrane tubing with 3.5-kDa molecular weight cutoff. This yielded protein that was 25–50% oxidized (fraction oxidized 0.2–0.5). The purified protein was diluted 1:50, and a volume of 200 μl of diluted protein was then added to each well of a 96-well plate; final protein concentration was ∼20 μM. Samples were read on a plate reader (General protocol) at excitation wavelengths of 405 and 488 nm and emission wavelength of 510 nm. The fraction oxidation was determined by ratiometric analysis and comparisons to roGFP completely oxidized with 10 mM H₂O₂ and completely reduced with 30 mM DTT, as described previously (34).

DCF.

H₂DCF-DA (2′,7′-dichlorodihydrofluorescein diacetate) is a nonfluorescent ester that when oxidized fluoresces upon exposure to blue light (17). DCF reportedly responds indirectly to H₂O₂ or O₂^·− after their oxidation to HO·(17). DCF (1 or 10 μM) was used with 500/525 nm excitation/emission (e/m) wavelength. Samples were processed as described in General protocol.

Human embryonic kidney 293 cells.

Human embryonic kidney (HEK) 293 cells were plated in 96-well plates and examined when near confluence. AzMC, SSP4, and DCF (all 10 μM) were added 2 h prior to experimentation. The medium was replaced 2× prior to the start of the experiment.

Transfection of HEK 293 cells was done using a roGFP primer targeted to the cytoplasm (also purchased from the University of Oregon). The cells were transfected following the procedure from Lipofectamine (Fisher Scientific), washed, and plated in 96-well plates. They were used when near confluence.

We were unable to achieve satisfactory reduction of the HEK 293 cells with 10 mM DDT and, therefore, could not accurately calibrate the fraction oxidation of roGFP in the cytoplasm. However, the ratiometric analysis still permitted an evaluation of the relative change in redox status of the cells, and these changes were evaluated.

Data analysis.

Data were analyzed and graphed using QuatroPro (Corel, Ottawa Ontario, Canada) and SigmaPlot 13.0 (Systat Software, San Jose, CA). Statistical significance was determined using one-way ANOVA and the Holm-Sidak test SigmaPlot 13.0. Results are given as means ± SE; significance was assumed when P ≤ 0.05.

RESULTS

Selectivity and sensitivity of AzMC and SSP4.

We first verified that AzMC and SSP4 were selective for H₂S and polysulfides and relatively insensitive to potential interferences from Cys, cystine (CSSC), and reduced and oxidized glutathione (GSH and GSSG, respectively). As shown in Fig. 1, AzMC is only slightly sensitive to 1 and 3 mM H₂S_n, the response to H₂S_n is likely due to H₂S, which exists as an impurity in K₂S_n salts (12). AzMC does not respond to up to 3 mM CSSC or GSSG or to Cys or GSH at the 2 mM concentration used to release H₂S from garlic oil. As shown in Fig. 2, SSP4 is insensitive to H₂S, but responds well to H₂S_n; high H₂S_n concentrations may decrease fluorescence. SSP4 does not respond to Cys or GSH at the 2-mM concentrations used to release H₂S from garlic oil, nor does it respond to GSSG between 1 μM and 3 mM. However, 1 and 3 mM CSSC slightly increase SSP4 fluorescence.

Fig. 1.

Sensitivity and selectivity of AzMC to H₂S, polysulfides (H₂S_n), cysteine (Cys), glutathione (GSH), cystine (CSSC), and oxidized GSH (GSSG). AzMC is selective for H₂S over H₂S_n, but it is insensitive to CSSC, GSSG, and to Cys and GSH at the concentrations used in this study (2 mM). Note that scales are expanded to those of Fig. 3 for comparison. Values are expressed as means ± SE; n = 3 experiments.

Fig. 2.

Sensitivity and selectivity of SSP4 to H₂S, H₂S_n, Cys, GSH, CSSC, and GSSG. SSP4 is sensitive to H₂S_n, somewhat sensitive to CSSC, but insensitive to H₂S or to 2 mM Cys or GSH (concentrations used in this study). Note that the ordinate is expanded for CSSC, GSSG, and 2 mM Cys or GSH to correspond to comparable experiments in Fig. 4. Values are expressed as means ± SE, n = 3 experiments.

H₂S formation from garlic oil and DATS.

H₂S production from garlic oil in normoxia and hypoxia (<5 μM oxygen) is shown in Fig. 3. Garlic oil alone did not produce appreciable amounts of H₂S in either normoxia or hypoxia, whereas a concentration-dependent steady rate of H₂S formation was observed between 1 and 300 μM garlic oil, when in the presence of either 2 mM cysteine (Cys) or 2 mM reduced glutathione (GSH). H₂S production from garlic oil in the presence of either Cys or GSH decreased when the garlic oil concentration was 1 and 3 mM. Hypoxia did not appreciably affect H₂S production from garlic oil in the presence of either Cys or GSH. H₂S was not generated from garlic oil in the presence of either CSSC or GSSG in normoxia or hypoxia.

Fig. 3.

H₂S production from garlic oil and diallyltrisulfide (DATS). In normoxia (top), garlic oil alone did not produce appreciable amounts of H₂S, whereas a concentration-dependent steady rate of H₂S formation was observed between 1 and 300 μM garlic oil in the presence of either 2 mM cysteine (Cys) or 2 mM reduced glutathione (GSH). H₂S production from garlic oil in the presence of either Cys or GSH decreased when the oil concentration exceeded 300 μM. Essentially similar results were observed in hypoxia (<5 μM O₂, middle). H₂S was not generated from garlic oil by either CSSC or GSSG in normoxia or hypoxia (bottom right). H₂S is also produced in normoxia from DATS in the presence of Cys or GSH (bottom left). Dashed lines indicate 100 μM H₂S from simultaneous Na₂S calibration curves. Garlic oil or DATS was added at −10 min; values are expressed as means ± SE; n = 3 experiments.

Because garlic oil contains a mixture of DAS, DADS, and DATS, we used a 10-fold lower concentration of DATS than garlic oil to measure H₂S production in normoxia (bottom left panel in Fig. 3). As with garlic oil, H₂S production from DATS was highly dependent on the presence of Cys and GSH, and these two were essentially equally efficacious. Over the 90-min experimental period, H₂S production from DATS in the presence of Cys or GSH was linear.

Polysulfides in garlic oil and DATS.

More SSP4-reactive polysulfides were observed in garlic oil without either Cys or GSH, and the lowest concentrations were observed in the presence of GSH (Fig. 4). The amount of polysulfide in garlic oil concentration-dependently increased from 1 to 10 μM, was maximal at 10 and 30 μM, and then decreased thereafter. With 2 mM Cys, polysulfides were not apparent until garlic oil reached 30 μM, they were maximal at 100 μM, and they decreased thereafter. Garlic oil polysulfide concentrations were lowest in 2 mM GSH. Estimated maximum polysulfide concentrations in the presence of Cys and GSH were ≥100 μM. SSP4 fluorescence from 1 μM garlic oil alone was similar to that from 30 μM garlic oil with either Cys or GSH. Oxidized low-molecular-weight thiols, CSSC, and GSSG increased SSP4 fluorescence at the lowest garlic oil concentrations. (Note that variations in the y-axis were due to different instruments that altered the absolute fluorescence but did not affect the relative changes.)

Fig. 4.

Polysulfides in garlic oil and DATS. In garlic alone polysulfide, concentration increased between 1 and 10 μM, was maximal at 10 and 30 μM and decreased thereafter. With 2 mM Cys, polysulfides were not apparent until garlic oil reached 30 μM, were maximal at 100 μM and decreased thereafter. Garlic oil polysulfide concentrations were lowest in 2 mM GSH. Oxidized thiols, CSSC (2 mM), and to a lesser extent, GSSG (2 mM), increased SSP4 fluorescence by low (1–30 μM) garlic oil concentrations. Polysulfides in DATS were greatest with 100 μM DATS in 2 mM Cys folowed by 100 μM DATS in 2 mM GSH and 10 μM DATS alone. Garlic oil and DATS were added at −10 min; values are expressed as means ± SE; n = 3 experiments. Use of different instruments produced the variations in the y-axis scale.

Polysulfides in DATS were greatest with 100 μM DATS in 2 mM Cys followed by 100 μM DATS in 2 mM GSH and 10 μM DATS alone (Fig. 4). However, unlike garlic oil, there was considerably more polysulfide with DATS in the presence of Cys and slightly more in the presence of GSH than with DATS alone. Furthermore, 2 mM Cys produced a rapid increase in polysulfide concentration with both 10 and 100 μM DATS, whereas polysulfide concentration increased slowly, nearly linearly, with GSH or DATS alone.

Effects of garlic oil on H₂S and polysulfide in HEK 293 cells.

The effects of garlic oil alone and in combination with Cys or GSH on intracellular H₂S (AzMC fluorescence) and polysulfides (SSP4 fluorescence) in HEK 293 cells exposed to normoxia (21% O₂) or hypoxia (<0.4% O₂) are shown in Figs. 5 and 6. Intracellular H₂S gradually increased in all cells over time. 100 μM garlic oil modestly increased H₂S in cells in normoxia and essentially doubled it in hypoxia; 1 mM garlic oil was without effect in normoxia but also increased H₂S in hypoxia. Conversely, 1 mM garlic oil was more efficacious than 100 μM garlic oil in increasing intracellular polysulfides, and there was no significant effect of hypoxia. The addition of either Cys (2 mM) or GSH (2 mM) to garlic oil greatly increased intracellular H₂S under all conditions, although hypoxia blunted the effect of 1 mM garlic oil in the presence of GSH. Intracellular polysulfides were increased ∼5- and 25-fold by 100 μM and 1 mM garlic oil, respectively, and these responses were unaffected by either Cys or GSH. GSH alone did not affect either H₂S or polysulfide in these cells.

Fig. 5.

Effects of garlic oil (GO; 100 μM and 1 mM) alone and in combinations with cysteine (Cys; 2 mM) or glutathione (GSH; 2 mM) on intracellular H₂S (AzMC fluorescence; A and B) and polysulfides (SSP4 fluorescence; C and D) in human embryonic kidney (HEK) 293 cells exposed to normoxia (21% O₂; A and C) and hypoxia (<0.4% O₂, B and D). A and B: both Cys and GSH increased intracellular H₂S, although hypoxia blunted the effect of 1 mM CO and GSH. C and D: GO alone concentration-dependently increased intracellular polysulfides, and this was not affected by Cys but reduced when GSH was added to 100 μM GO. These responses were unaffected by hypoxia. bkg, autofluorescence of untreated cells; AzMC/SSP4, cells treated with dye only. Values are expressed as means ± SE; n = 3 experiments.

Fig. 6.

Expanded scale of data in Fig. 5 showing the effects of 100 μM and 1 mM garlic oil on H₂S and polysulfides in HEK 293 cells in normoxia and hypoxia. H₂S appeared to steadily increase over time in untreated cells although only t = 0 and t = 24 h were significantly different (P = 0.046 and 0.035 for normoxia and hypoxia, respectively). In all other cells, t = 2–24 h was significantly (P < 0.001) greater than t = 0 h. The addition of 100 μM garlic oil increased intracellular H₂S more than 1 mM garlic oil, and hypoxia augmented the effect of 100 μM garlic oil. Conversely, 1 mM garlic oil was ∼3-fold more efficacious than 100 μM garlic oil in increasing intracellular polysulfides, and this was unaffected by hypoxia. Autofluorescent cells and t = −1 h have been omitted for clarity.

Interaction of reduced/oxidized cysteine and glutathione and garlic oil with the reactive oxygen species indicator, roGFP.

To examine the redox properties of garlic compounds on roGFP, we first determined the effects of reduced and oxidized cysteine (Cys and CSSC, respectively) and reduced and oxidized glutathione (GSH and GSSG, respectively) on roGFP (Fig. 7, top). Cys reduced roGFP at 1 and 3 mM, the same concentrations of GSH also initially reduced roGFP, although to a lesser extent. roGFP was concentration-dependently oxidized by >3 μM CSSC with 1 and 3 mM producing nearly complete oxidation at 90 min. All GSSG concentrations oxidized roGFP, albeit at a slower rate and lower extent of oxidation at 90 min compared with CSSC.

Fig. 7.

Reduction/oxidation of redox sensitive green fluorescent protein (roGFP) by Cys, GSH, CSSC, and GSSG (top) and by garlic oil alone or in combination with 2 mM of the low-molecular weight thiols (middle and bottom). Cys and CSSC were more efficacious reductants or oxidants, respectively, of roGFP than GSH and GSSG. Garlic oil alone concentration-dependently oxidized roGFP. Simultaneous addition of garlic oil and reduced glutathione (GSH) or cysteine (Cys) decreased the rate and extent of roGFP oxidation. Low concentrations (1–10 μM) of garlic oil were initially unable to overcome the reductive effects of Cys, although roGFP was progressively oxidized thereafter. The effects of GSH were similar to those of Cys, although less pronounced. roGFP was nearly completely oxidized at all garlic concentrations in the presence of CSSC. The effects of GSSG were similar to those of CSSC, although less pronounced. Values are expressed as means ± SE; n = 3 experiments.

roGFP was concentration-dependently oxidized by both garlic oil alone (Fig. 7) and DATS (not shown). Significant roGFP oxidation was produced by 1 μM, the lowest concentration of garlic oil and DATS employed. The maximum responses were attained at 30–100 μM garlic oil and 300 μM DATS. Low concentrations of garlic oil or DATS were initially unable to overcome Cys reduction of roGFP, although the protein became progressively more oxidized thereafter. roGFP was nearly completely oxidized in the presence of 2 mM CSSC masking nearly all of the effects of garlic oil or DATS. GSH and GSSG had similar effects on garlic oil and DATS as Cys and CSSC, respectively, although the effects were less pronounced.

If the roGFP was not reduced prior to experimentation and remained ∼100% oxidized (fraction oxidized ∼1.0; Fig. 8) it was slightly, but concentration-dependently reduced by high concentrations (≥300 μM) of garlic oil. Cys (2 mM) and GSH slightly enhanced garlic oil's efficacy, although the latter to a lesser extent. Neither CSSC nor GSSG affected the garlic oil response.

Fig. 8.

Completely oxidized roGFP was slightly but concentration dependently reduced by garlic oil alone. Cys (2 mM) increased garlic oil efficacy as did 2 mM GSH, although to a lesser extent. Garlic oil oxidization was not affected by either CSSC (2 mM) or GSSG (2 mM). Note that expanded scale relative to Fig. 7. Values are expressed as means ± SE; n = 3 experiments.

Effects of garlic oil on roGFP oxidation in HEK 293 cells.

In spite of our inability to accurately calibrate intracellular redox conditions, we were able to monitor relative changes in cytoplasmic redox after application of garlic oil (Fig. 9). In normoxia, 10 μM garlic oil slightly reduced the cytoplasm after 2 h, whereas 100 μM garlic oil immediately and substantially reduced the cytoplasm, and it remained so for the entire 24-h experimental period. One mM of garlic oil produced the greatest cytoplasmic reduction immediately after application, but the cytoplasm became progressively more oxidized thereafter. These responses were unaffected by the simultaneous application of 1 mM Cys or GSH. Cytoplasmic reduction by 100 μM garlic oil was even more pronounced when cells were exposed to hypoxia, whereas 1 mM garlic oil was a considerably less efficacious reductant. In hypoxia, both Cys and GSH augmented the reductive effects of 100 μM and 1 mM garlic oil.

Fig. 9.

Effects of garlic oil (GO) and GO in combination with 1 mM Cys or 1 mM GSH on oxidation/reduction of roGFP in the cytosol of HEK 293 cells in normoxia or hypoxia. In normoxia, 10 μM GO produced a slight time-dependent reduction of roGFP, 100 μM produced a sustained reduction, whereas after 1 mM, roGFP was initially reduced and then became progressively oxidized. These responses were unaffected by either Cys or GSH. Compared with normoxia, in hypoxia, 10 μM and 1 mM garlic oil were less effective, whereas100 μM was more potent. Both Cys and GSH enhanced the reductive effect of 100 μM and 1 mM garlic oil. Green bars (t = −1 h) are normalized fluorescence 1 h prior to the addition of drugs; bkg is background fluorescence of cells without roGFP. Values are expressed as mean ± SE; n = 3 experiments.

Effects of garlic on the reactive oxygen species indicator, DCF.

Garlic oil alone (1 μM-3 mM) did not produce DCF fluorescence, whereas fluorescence was concentration-dependently increased by Cys, GSH, CSSC, and GSSG (Figs. 10 and 11); the relative potency was CSSC > Cys > GSH >> GSSG. When garlic oil was added in conjunction with Cys or GSH, it concentration-dependently augmented DCF fluorescence between 1 and 30 μM and progressively decreased it thereafter. Low concentrations of garlic oil did not affect maximum CSSC or GSSG fluorescence, while higher garlic oil concentrations progressively inhibited fluorescence. Hypoxia did not affect the inability of garlic oil alone to produce DCF fluorescence nor did it inhibit DCF fluorescence produced by Cys or GSH (not shown).

Fig. 10.

Cys, GSH, and GO-induced fluorescence from the ROS indicator, 2′,7′-dichlorofluorescein (DCF). Both low-molecular thiols dose-dependently increased DCF fluorescence, whereas garlic oil alone did not affect DCF fluorescence. The addition of low concentrations of garlic oil augmented 2 mM Cys and 2 mM GSH-induced fluorescence, whereas higher concentrations of garlic oil were inhibitory. Reducing the concentration of either Cys or GSH decreased total fluorescence (right; note change in y-axis scale) without affecting the general influence of garlic oil. Values are expressed as means ± SE; n = 3 experiments.

Fig. 11.

CSSC and GSSG-induced fluorescence from the reactive oxygen species (ROS) indicator, 2′,7′-dichlorofluorescein (DCF) in the presence or absence of garlic oil. Both low-molecular weight thiols dose-dependently increased DCF fluorescence, although GSSG was far less effective (note scale of y-axis). The addition of low concentrations of garlic oil to 2 mM CSSC or 2 mM GSSG did not affect fluorescence compared with 3 mM CSSC or GSSG alone, whereas higher concentrations of garlic oil concentration-dependently increased fluorescence. Reducing the concentration of either CSSC or GSSG decreased total fluorescence (right; note change in y-axis scale) without affecting the general influence of garlic oil. Values are expressed as means ± SE; n = 3 experiments.

DISCUSSION

We examined the production of H₂S and polysulfides from garlic oil and DATS and determined their redox properties using two commonly employed redox indicators, roGFP and DCF. These experiments were performed in buffer to evaluate direct chemical interactions and in cells to ascertain their ultimate fate or effects in a living and dynamic system. A number of conclusions can be drawn from the results. First, H₂S is released from garlic oil by reduced, but not oxidized, low-molecular-weight thiols, as shown previously, but garlic oil alone has modest effects on intracellular H₂S concentration, whereas it greatly increases the concentration of intracellular polysulfides. Second, the large increases in intracellular H₂S following garlic oil application are largely due to extracellular H₂S formation concomitant with application of other low-molecular-weight thiols. Third, the physiological effects of garlic oil that are mediated through reactive protein cysteines may be accomplished without H₂S formation. Fourth, garlic oils are effective oxidants but poor reductants in buffer solutions, and this appears to occur independent of reactive oxygen intermediates. Fifth, when garlic is taken up by cells, the cytoplasm becomes more reduced, and this is likely due to secondary metabolism of the garlic oil polysulfides.

H₂S and polysulfide formation from garlic oil in buffer and cells.

Generation of H₂S from garlic oil in buffer or homogenized tissue has been shown to require the presence of low-molecular-weight thiols, and subsequent H₂S signaling is believed to play a major role in the physiological effects of garlic (2, 47; see also introduction). Our results confirm that H₂S is generated from garlic oils in buffer (Fig. 3), but they also indicate that there is a threshold concentration above which the garlic oil inhibits Cys- and GSH-mediated H₂S release. There was less evidence, however, that increases in H₂S in buffer are directly translated into substantial increases in intracellular H₂S unless low-molecular-weight reduced thiols (Cys or GSH) were also present in the extracellular medium (Fig. 5). We believe that the large (3–7-fold) increase in intracellular H₂S in the presence of extracellular Cys or GSH was due to rapid H₂S production in the extracellular milieu and subsequent diffusion of H₂S into the cells. One would have expected that intracellular GSH, which is also present in millimolar concentrations, would foster similar H₂S production with extracellular garlic oil alone, but that did not occur. Perhaps this is because garlic oil diffusion into the cells is slower than that of H₂S and with the slower delivery of substrate, H₂S metabolism would better be able to keep pace with H₂S formation. This hypothesis is supported by the observation that intracellular H₂S concentration after garlic oil application was greater when the cells were hypoxic (Fig. 6) and by previous observations that cellular H₂S metabolism is to a large extent O₂-dependent (41). Alternatively, the garlic oil sulfides could be metabolized to other persulfides or thiols without H₂S formation (see below).

Our results show that SSP4 clearly measures polysulfides from garlic oils in buffer (Fig. 4) and in cells (Fig. 5) and that this does not require the presence of either Cys or GSH (Fig. 4). High concentrations of the garlic oils inhibit SSP4 fluorescence, but this is not specific for garlic oil, as similar effects were observed using a mixed polysulfide (H₂S_n; n = 1–8). More importantly, our results show that garlic oil polysulfides enter HEK 293 cells and that this is independent of other low-molecular weight thiols (Figs. 5 and 6). We believe these polysulfides are the most relevant signaling moieties of the garlic oils.

The physiological effects of garlic oil mediated through reactive protein cysteines are accomplished without H₂S formation.

While the addition of 2 mM Cys or 2 mM GSH to garlic oil was necessary to substantially increase intracellular H₂S in HEK 293 cells (Fig. 5), neither Cys nor GSH were necessary for the effect of garlic oil on intracellular redox status of roGFP (Fig. 9). This indicates that the effects of garlic oil on cellular redox status is due to garlic oil polysulfides and independent of H₂S.

A number of studies have shown that sulfur signaling via regulatory protein cysteines is effectively accomplished by polysulfides through the process of persulfidation (11, 20, 37, 42–44). In this process, one (or more?) sulfur atoms becomes covalently bound to a protein cysteine, thereby altering protein function. Over 15 proteins have been shown to be regulated via sulfuration, and this number is rapidly increasing (44). However, H₂S cannot persulfidate protein cysteines because sulfur is in its most reduced form (−2) as are the reactive protein cysteines, and their reaction is not favored (11, 55). In order for H₂S to persulfidate, either it, or the protein cystine sulfur, must first undergo a two-electron oxidation to sulfane sulfur (0 formal oxidation state), or both must undergo a one-electron oxidation, which is less likely. The sulfane (thiosulfoxide) sulfur in DADS (RSSR) and DATS (RSSSR), where R denotes the allyl groups, appears to be especially important in this regard, as it can tautomerize to RS(S)R and RS(S)SR, and this sulfur is highly reactive (55).

The sulfane sulfur in garlic oil has the potential to be both nonenzymatically and enzymatically transferred to other thiols or polysulfides. Nonenzymatic transfer is suggested in Fig. 4, whereby SSP4 fluorescence is greatly increased by even the lowest garlic oil concentration when CSSC or GSSG are present. The inability of oxidized thiols, CSSC, and GSSG to release H₂S in either normoxia or hypoxia suggests that this process is a typical thiol-disulfide exchange reaction of the type described by Liang et al. (30) which is O₂-independent and generates a variety of polysulfide intermediates but does not generate H₂S.

Enzymatic processes can also catalyze sulfur transfer and the formation of a variety of polysulfide moieties. Dihydropersulfides (H₂S₂), dihydropolysulfides (H₂S_n), hydropersulfides (RS₂H), and hydropolysulfides (RS_nH), where n = 3–5, have been shown to occur in cells, and they can be endogenously generated. Kimura et al. (23) identified H₂S₂, H₂S₃, and H₂S₅ in mouse brain and showed that the enzymes 3-mercaptopyruvate sulfur transferase (3MST) and rhodanase catalyzed the formation of H₂S and H₂S₃ from 3-mercaptopyruvate, an endogenous substrate. Ida et al. (14) showed that the enzymes originally linked to H₂S biosynthesis, cystathionine γ-lyase (CSE) and cystathionine β-synthase (CBS), catalyze the formation of cysteine and glutathione hydropersulfides and/or hydropolysulfides (RSSH, RSSSH, and RSSSSH, where R = Cys or GSH) from cystine. They also suggest that the intracellular concentration of these GSH hydropersulfides and hydropolysulfides exceeds 100 μM. Recently, Yadav et al. (63) showed that under more physiological conditions CSE and CBS are more likely to generate H₂S than either hydropersulfides or hydropolysulfides from cystine, which questions the intracellular concentrations of the latter two reported by Ida et al. (14) and suggests they are far lower. Yadav et al. (63) also suggested that SSP4 is unlikely to detect low-molecular-weight persulfides in solution but instead detects thiosulfoxides [RS(=S)SR] and that SSP4 itself could bind to cysteine persulfides. The issue is further complicated by the possibility of tautomerization between thiosulfoxides and polysulfides (55). We have shown previously (8) and again in Fig. 2 that SSP4 readily responds to dihydropersulfides and dihydropolysulfides; however, the decrease in fluorescence at high H₂S_n or garlic oil concentrations (see below) may reflect this persulfide-SSP4 interaction. Clearly, additional studies are necessary to fully sort out the reactivity of SSP4.

Although never measured, it is quite possible that similar mono and dihydropersulfides and polysulfides are generated from DADS and DATS by 3MST, CSE, and CBS, and these products would be expected to be quite reactive. The rapid cellular uptake of the garlic polysulfides and their comparatively higher intracellular titers (Figs. 5, 6) and reactivities suggest that they are far more relevant than H₂S to the biological actions of garlic. In addition, these enzymatically generated persulfides and polysulfides may act as either electrophiles or nucleophiles (11, 37, 42, 43).

Garlic oils are effective oxidants but poor reductants.

Our experiments with roGFP, arguably the “gold standard” indicator for ROS (50) show that garlic oil alone acted as a potent oxidant and was effective at concentrations as low as 1 μM (Fig. 7). This is similar to our previous report (8) in which 1 μM of the mixed polysulfide, H₂S_n also effectively oxidized roGFP. Furthermore, while cysteine and glutathione reduced roGFP, these effects could be overcome by garlic oil oxidation at substantially lower concentrations of garlic oil. Conversely, high concentrations (∼1–3 mM) of garlic oil were able to reduce the oxidized roGFP (Fig. 8), which was also similar to the effects of H₂S_n at these same two high concentrations (8). Our previous experiments (8) and those in the present study also suggest that this “oxidative” process is independent of oxygen or its reduced species, superoxide anion, hydrogen peroxide, and hydroxyl radical.

Garlic oil oxidation of roGFP does not require ROS.

Most scenarios of sulfur compounds, especially thiols and polysulfides acting as oxidants indicate that this is ultimately achieved by generation of reactive oxygen species, not by the thiols themselves (10, 16, 36, 58). Munday (36) describes a catalytic cycle, whereby an oxidized transition metal reacts with a thiol to form a thiyl radical (Eq. 1), which reacts with another thiol to form the disulfide radical anion (Eq. 2) that then autooxidizes forming a disulfide and superoxide anion (Eq. 3). Superoxide can further react with another thiol to produce peroxide and regenerate the thiyl radical (Eq. 4). The reduced metal can also reduce oxygen to superoxide or peroxide to the hydroxyl radical.

$${\text{RS}}^{-}+{\text{M}}^{\text{n}+}\to {\text{RS}}^{\mathbf{·}}+{\text{M}}^{(n-1)+}$$ (1)

$${\text{RS}}^{\mathbf{·}}+{\text{RS}}^{-}\to {\text{RSSR}}^{\mathbf{·}-}$$ (2)

$${\text{RSSR}}^{\mathbf{·}-}+{\text{O}}_2\to \text{RSSR}+{{\text{O}}_2}^{\mathbf{·}-}$$ (3)

$${\text{RS}}^{-}+{{\text{O}}_2}^{\mathbf{·}-}\to {\text{RS}}^{\mathbf{·}}+{\text{H}}_2{\text{O}}_2$$ (4)

Oxyhemoglobin can also react with thiols to form peroxide (Eq. 5; Ref 36), and H₂S binding to cytochrome c also leads to superoxide production (58).

$${\text{RS}}^{-}+{\text{Fe}}^{\text{II}}{\text{HbO}}_2+2{\text{H}}^{+}\to {\text{RS}}^{\mathbf{·}}+{\text{Fe}}^{\text{II}}\text{Hb}+{\text{H}}_2{\text{O}}_2$$ (5)

We recently showed that H₂S and H₂S_n mimic ROS activation of a variety of sensors designed to measure ROS, including roGFP, DCF, MitoSox Red, and Amplex Red (8). Furthermore, we showed that sulfide “oxidation” of roGFP was actually increased in hypoxia (O₂ <4 μM). Our present experiments show that garlic oil are also efficacious roGFP oxidants (Fig. 7), and although we did not examine this in hypoxia, parallel experiments with DCF support our conclusion of an oxygen-independent process.

DCF has been shown to be relatively nonspecific for ROS, but because it appears insensitive to direct oxidation by either O₂^·− or H₂O₂, it has been suggested that DCF oxidation requires hydroxyl radicals generated from H₂O₂ and transition metal ions via Fenton-type reactions (17). However, in the present experiments, we found no evidence for garlic oil oxidation of DCF in normoxia and presumably in the presence of trace metals (Fig. 10). This suggests that 1) garlic oil does not produce ROS, and, therefore, 2) garlic oil oxidation of roGFP is accomplished through some ROS-independent mechanism, most likely sulfur-sulfur interactions. Furthermore, the apparent paradoxical ability of garlic oil to greatly enhance DCF fluorescence, which only occurs in the presence of Cys or GSH (Fig. 10), can be readily explained by Cys and GSH liberation of H₂S from garlic oil (Fig. 3) and subsequent H₂S-mediated increase in DCF fluorescence, as we have previously shown (8). The ability of Cys and GSH to reduce roGFP (Fig. 7), yet oxidize DCF (Fig. 10) is further evidence that these reactions do not have ROS generation as the common denominator.

Garlic oil as an intracellular reductant.

To our knowledge, this study is the first to measure the effects of garlic oil on intracellular H₂S, polysulfides and redox conditions. As we found little evidence that garlic oil or DATS is a reductant in buffer (except at high concentrations; Fig. 8), the conundrum is how can it become one in cells (Fig. 9)? Furthermore, why is this reversed at higher garlic oil concentrations and why does hypoxia amplify both the reductive effect of 100 μM garlic oil and the reversal of this effect by 1 mM garlic oil. Although there are a number of possibilities, these must include short- and long-term effects, as we observed both over the 24-h experimental period.

The short-term responses, which were seen immediately upon application of the garlic oil, must be explained by reaction of garlic oil with an intracellular molecule or molecules that subsequently reduce roGFP. The most likely are direct sulfide or disulfide exchange processes and/or enzymatic reactions, as described above. The long-term effects could be explained by persulfidation of Keap1. This dissociates the Keap1-Nrf2 complex, allowing the latter to translocate to the nucleus and activate the antioxidant response elements (24).

Perspectives and Significance

Humans have known for eons that some foods bring more to the table than just calories. Clearly, this is the case with garlic. Sulfur springs also have a long history in human culture, and at the crossroads of these seemingly disparate and malodorous medicinals is sulfur. Since the original finding that hydrogen sulfide is a physiologically relevant signaling molecule, considerable emphasis has been placed on its metabolism and mechanisms of action. Not surprisingly, it has also been demonstrated that hydrogen sulfide is released from garlic under certain conditions. Very recently, however, the ability of hydrogen sulfide to directly interact with regulatory proteins has been questioned, and focus has now turned toward polysulfides as the biologically relevant mediators. In our studies we show that polysulfides derived from garlic readily enter cells and may, in fact, be directly responsible for garlic's attributes. But this is not the final answer. Considerable questions remain regarding intracellular trafficking of the sulfur compounds as well as their metabolism and signaling attributes. With the considerable emphasis now being placed on the therapeutic benefits of synthetic hydrogen sulfide-“donating” compounds, perhaps we should take an even more rigorous look at these natural sulfides, as they can be cultivated under a wide range of conditions with minimal expenditure of energy or labor.

GRANTS

This research was supported by National Science Foundation Grant IOS 1446310.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

E.R.D., Y.G., and K.R.O. conception and design of research; E.R.D., Y.G., and E.H. performed experiments; E.R.D., Y.G., E.H., and K.R.O. analyzed data; E.R.D., Y.G., E.H., and K.R.O. interpreted results of experiments; E.R.D. and K.R.O. prepared figures; E.R.D., Y.G., E.H., and K.R.O. approved final version of manuscript; K.R.O. drafted manuscript; K.R.O. edited and revised manuscript.