| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,3
* Ecole Pratique des Hautes Etudes and Nurélice UR909, INRA, F-78352 Jouy en Josas, France; and
UPR9078, CNRS Université René Descartes, F-75730 Paris, France
3Correspondence: CNRS-UPR9078, Université René Descartes, Site Necker 156 rue de Vaugirard, 75730 Paris Cedex 15, France. E-mail: bouillaud{at}necker.fr
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: mitochondria • oxidation • cytochrome oxidase • gut epithelium • colon
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2
3
4)
. Sulfide is an inhibitor of the mammalian cytochrome oxidase (5
, 6)
and is therefore considered as a mitochondrial poison for mammalian cells. Poisoning of cytochrome oxidase in cellular homogenates occurs at concentrations within the micromolar range (7)
. High millimolar H2S concentrations have been reported in the lumen of the human large intestine (8)
, as well as in the fecal material of humans receiving a high meat diet (3)
. However, these relatively high concentrations have been recently reexamined given the fact that sulfide can react to or be complexed easily by numerous chemicals (9)
.
Sulfide has been implicated in the etiology of ulcerative colitis (10
11
12)
. Moreover, in experimental colitis animal models, it is possible to induce a pathological state similar to the one observed in ulcerative colitis using two forms of undigestible sulfates (13
14
15)
. It is not known if the inhibition of cytochrome c oxidase activity by sulfide is implicated in this pathology. Regardless, it appears very likely that the colonic mucosa must face concentrations of sulfide that can potentially severely impair mitochondrial respiration. Moreover, since the intestine is permeable to H2S, poisoning of surrounding tissues could also occur. This suggests the existence of a detoxifying mechanism in the colonic mucosa itself, which has been only partially characterized (16
, 17)
.
Inorganic sulfur-reduced compounds can serve as electron donors in many phototrophic and chemotrophic bacteria (18)
. Moreover, sulfide has been shown to donate electrons to the mitochondrial respiratory chain in clams, worms, or fishes adapted to environments showing concentrations of sulfide comparable with those measured inside the mammalian gut [reviewed in (19)
]. Mitochondrial sulfide oxidation has been shown also in chicken liver, suggesting that it is a phenomenon shared by many/all mitochondria and independent of any adaptation to a sulfide-rich environment (20)
. This characteristic is likely to be inherited from the procaryotic ancestor of mitochondria. These existing evidences show that electrons from sulfide could be given to the mitochondrial respiratory chain at the level of coenzyme Q (21)
or cytochrome c (22)
. The inhibition of cytochrome oxidase is overcome in worms by the existence of a cyanide-insensitive alternative oxidase biochemically related to that known in plants (23)
. When sulfide is oxidized by mitochondria the coupling of electron transport to proton pumping and/or ATP production has been shown (21
, 22)
.
In contrast with other organisms (invertebrates, birds), no evidence exists as yet for a mitochondrial oxidation of sulfide coupled with proton transport in mammalian mitochondria. Our previous studies with the human colonic cell line HT29 exposed to 1 mM sulfide (7)
have led to the following conclusions: i) Sulfide readily poisons mitochondrial oxidation of various substrates; ii) long-term (24 h) exposure results in a partial compensation by anaerobic glycolysis and a decline in mitochondrial efficiency as judged from the decrease in cytochrome oxidase expression accompanied with an apparent partial uncoupling; and iii) during short-term exposure (90 min) HT29 cells have a limited but real ability to detoxify a portion of the 1 mM NaHS added to the medium. The present study is aimed to determine the mitochondrial response to short-term exposure to NaHS.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Cellular models
Two human colon adenocarcinoma cell lines were used: The parental CaCo2 cells established in 1974 by Jorgen Fogh (Sloan Kettering Cancer Center, New York, NY, USA) and the HT-29 Glc–/+ cells established in permanent culture in 1975 and selected from parental cells by growing them in glucose-free medium for 36 passages (24)
. Both lines were grown under 10% CO2 in Dulbecco’s modified Eagle medium (DMEM) supplemented with 10% (HT-29) or 20% (CaCo2) (v/v) FBS. The culture medium was changed daily. The cells were seeded at a density of 20 x 103 cells/cm2 on day 0. In all experiments, the cells were used 6 d after cell seeding. Cells in suspension were obtained by the action of trypsin 0.25 g/l in phosphate buffered saline (PBS) containing 1 g/l EDTA. The cells were counted on a hematocytometer. Colonocytes and enterocytes were isolated from rat intestine essentially as described in (25)
.
Measurements in permeabilized cells
After isolation, the cells were incubated at 37°C in a medium (pH 7.4) containing 200 mM mannitol, 10 mM KH2PO4, 2.5 mM MgCl2, 20 mM HEPES, 0.5 mM EGTA enriched with 1 mg/mL BSA. The incubation medium was equilibrated against air before use. Mitochondrial respiration (5x106 cells in 1.5 ml) was monitored by polarography (Oxygraph Hansatech Inst., Norfolk, UK) at 37°C in the presence of the F1/FO ATPase inhibitor oligomycin (0.5 µg/ml), and the membrane potential was measured with a home-made tetraphenyl phosphonium (TPP+) electrode following published procedures (26)
. With no accurate method to calculate the number and vol of mitochondria in situ, an estimated value of 0.03 µl per million of cells was used and gave realistic values of membrane potential. HT-29 cells were permeabilized with digitonin. In the presence of 1 µM rotenone to inhibit oxidation of endogenous substrates, digitonin concentration was adjusted as being the minimal dose (50 µM in our conditions) ensuring a maximal oxidation rate with sulfide or succinate or 
-DL-glycerophosphate in the presence of FCCP as the uncoupler, meaning an optimal permeabilization for succinate. Sulfide was gradually infused in the oxygraph chamber using an adjustable syringe pump (PHD 2000 Havard, Holliston, MA, USA) equipped with a 1 ml glass airtight microsyringe filled with freshly prepared 2 mM NaHS. The infusion range was between 0.4 and 20 nmol/min. The chemical sulfide was calibrated according to Svenson (27)
.
Others
The content in succinic and malic acids of the HT-29 Glc –/+ cells was measured by HPLC associated with mass spectrometry detection using a method previously described. The results were expressed as the mean value (±SEM) of individual experiments performed with cells isolated during different passages or from different animals. Unless otherwise specified, the statistical significance of the differences between mean values was assessed by the Student’s t-test.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
shows a representative experiment where both oxygen consumption and mitochondrial membrane potential were measured. Infusion of sulfide caused oxygen consumption by the permeabilized cells. This oxygen consumption was accompanied by an immediate increase in membrane potential as judged from the TPP+ uptake. Both could be blocked by antimycin and myxothiazol, which are known inhibitors of the complex III (coenzyme Q – cytochrome c reductase) of the mitochondrial respiratory chain (not shown). Cyanide, an inhibitor of the complex IV (cytochrome oxidase), also inhibited oxidation of NaHS (not shown). Moreover, it can be deduced that rotenone did not inhibit this oxidation, since this inhibitor of mitochondrial complex I was present during all the experiments with permeabilized cells (see Materials and Methods). The FoF1 mitochondrial ATP synthase inhibitor oligomycin was also always present in the experiments shown (Figs. 1
2
3
4
). These two inhibitors, rotenone and oligomycin, prevented the use of endogenous substrates and ADP by mitochondria within permeabilized cells. Therefore, no information about the production of ATP in these conditions could be obtained. Actually, when experiments were performed in the absence of oligomycin, and with addition of an excess of ADP the phosphorylating state 3, rate of respiration was not stable but diminished progressively with a concomitant increase in membrane potential. This was not observed with succinate or L-alpha-glycerophosphate (data not shown). An explanation would be that the oxidation of sulfide led to the accumulation of an inhibitor of the FoF1 ATP synthase in the measurement chamber and, therefore, mitochondria progressively returned to state 4.
|
|
|
|
Comparison with other mitochondrial substrates
Both oxygen consumption and energization appeared to be extremely sensitive to the infusion rate, i.e., the amount of sulfide delivered instantaneously to mitochondria. At the lowest rate shown here (6 nmol of sulfide per minute) both oxygen consumption and energization ceased immediately when infusion stopped. Increasing the infusion rate to 8 nmol per minute did not accelerate further oxygen consumption; it only allowed it to continue for approximately 1 min after the infusion ceased. Doubling the rate of infusion (12 nmol/min) had deleterious effect both on the respiratory rate (oxygen consumption) and on the energization process (TPP+ uptake). It is noteworthy that in this experiment (12 nmol/min) both energization and oxidation improved before ceasing, i.e., when sulfide concentration was approaching zero. The practical conclusion was that the determination of the maximal oxidation rate of sulfide must be determined after optimization of the infusion rate.
Experiments with inhibitors (see above) indicated that electrons from sulfide follow a route similar to that of the known mitochondrial substrates succinate and L-alpha-glycerophosphate. Therefore, the maximal respiratory rates observable with these two substrates, which could be added in saturating amounts, were compared with that of sulfide (Table 1
). This comparison showed that human HT29 cells, CaCo2 cells, or freshly prepared rat colonocytes showed maximal oxidation rates of sulfide in presence of uncoupler, which compare well with those of succinate or L-alpha-Glycerophosphate, this is not true with rat enterocytes (Table 1)
or with other cell lines such as Chinese hamster ovary cells "CHOs" (not shown).
|
Since both the amount of sulfide provided to mitochondria and the oxygen consumed were known, the stoichiometry of the oxygen consumed to sulfide provided could be determined. With infusion rates that remain below that necessary for an optimal (maximal) oxygen consumption, a value close to 0.8 was observed (Fig. 2)
, which was higher than the expected value of 0.5 (see Discussion).
Top-down analysis of proton leak and respiratory chain in HT-29 cells
Figure 3
shows how the oxygen consumption and the membrane potential are related when different constraints are applied to mitochondria in permeabilized cells. According to the principle of Mitchell’s chemiosmotic theory, the respiratory chain generates a proton motive force measured here by the mitochondrial membrane potential (
). This proton motive force creates a current of protons that are driven back by two qualitatively different ways: The main physiological return pathway is the mitochondrial FoF1 ATP synthase, which converts the energy stored in the gradient into a phosphorylation potential (ATP/ADP ratio). The other pathway(s) are grouped under the term leaks as they authorize proton return without any energetic counterpart. In our experiments the mitochondrial FoF1 ATP synthase was inhibited by oligomycin. Therefore, the respiratory chain and the proton leak determined the membrane potential and proton return. While membrane potential () could be directly evaluated from the TPP+ signal, the proton circulation was deduced from the oxygen consumption. It is, therefore, necessary to consider the stoichiometry between oxygen consumption and proton pumping. In this respect the value of 0.8 obtained above creates a problem (see Discussion). Figure 3A
shows how the respiratory chain of HT29 cells responded to the increase in leaks caused by the addition of increasing amount of the protonophoric uncoupler carbonyl cyanide m-chlorophenylhydrazone (CCCP). Figure 3B
shows how the leaks changed when the respiratory chain activity was changed by varying the availability of substrate.
Citric cycle intermediates
We also compared succinic and malic acid content in HT-29 cells incubated during 12 min in the respiratory medium at 37°C in the presence of sulfide infusion as substrate. Respiratory rate was normalized to 150% of state 4 with uncoupler. Cells received either 12 nmol HS–/minute/5 x 106 cells (no inhibition of respiration) or 28 nmol HS–/minute/5 x 106 cells (complete inhibition of respiration). The succinate and malate content of these control and treated cells were determined (Fig. 4)
. In the first case (12 nmol), oxidative metabolism of sulfide could take place in mitochondria; in the second (28 nmol), poisoning of the respiratory chain prevented the normal oxidative pathway. This inhibition of mitochondrial respiration by sulfide was accompanied by a significant increase in succinate associated with a slight decrease in malate. Consequently the succinate to malate ratio changed significantly.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
.
|
Sulfide donates electrons to the mitochondrial respiratory chain between the complexes I and III. Therefore, it appears very likely that coenzyme Q, which is the intermediate between complex I and complex III, is the acceptor for the electrons coming from sulfide oxidation. Accordingly, mammalian mitochondria would use a sulfide-quinone oxido reductase linked to the mitochondrial respiratory chain. Actually, a candidate gene exists in animals, including mammals (28
29
30)
. Reduction of a quinone (coenzyme Q) by a molecule of sulfide would require the two hydrogen atoms of sulfide (two electrons and two protons). Electron transfer in the mitochondrial respiratory chain involves a pair of electrons that ultimately reduces a single oxygen atom at the level of cytochrome oxidase (complex IV). Therefore, if a single molecule of sulfide yields two electrons, the stoichiometry within the mitochondrial respiratory chain would be 0.5 (ratio of oxygen molecule used to sulfide oxidized). Disulfide (HSSH) is a known product of sulfide oxidation by the isolated bacterial sulfide-quinone reductase (18)
. In this latter case, two molecules of hydrogen sulfide (H2S) would yield two electrons and two protons and the oxygen-to-sulfide ratio would be 0.25. If disulfide is still usable by the enzymatic machinery, polysulfide would be produced (18)
and intermediate values of the ratio would be found. In any case, 0.5 appears as the higher limit for this ratio as long as the electron transfer from quinone to cytochrome oxidase in the mitochondrial respiratory chain is considered. The higher value of 0.79 we have obtained experimentally (Fig. 2)
suggests that, in addition to the mitochondrial respiratory chain, another mechanism of oxygen consumption is taking place. While reduction of quinone would produce polysulfide or elemental sulfur (see above), the oxidation of sulfide by animal tissues results in accumulation of oxygenated products: mainly thiosulfate (S2O32–), sulfites (SO32–), and sulfates (SO42–) (16
, 19
, 23
, 31)
. Their formation implies the use of some oxygen. We conclude that when sulfide is oxidized by mitochondria in colonocytes there are two sites for oxygen consumption (Fig. 5A
): i) the formation of water at the level of mitochondrial complex IV (cytochrome oxidase); and ii) the formation of sulfur containing oxygenated products.
In procaryotes sulfide quinone reductase activity is associated with other enzymatic activities to oxidize sulfide into sulfite (32)
. Present knowledge indicates that sulfide quinone reductase is inserted in the mitochondrial membrane (28)
. In procaryotes oxidation of sulfide by the sulfide quinone reductase yields as a primary oxidation product elemental sulfur released in the periplasmic space as sulfur globules (33
, 34)
. However, the subsequent oxidation steps (to sulfite) take place in the bacterial cytoplasm and need yet unknown mechanisms of transport (32)
. This leaves open various possibilities in mitochondria: firstly, a conservative mechanism in which the sulfide quinone reductase releases its oxidation product(s) in the intermembrane space, followed by a matricial oxidation to yield oxygenated products, which raises the problem of the permeation of the sulfide quinone reductase oxidation product(s). Secondly semiconservative hypotheses: i) the sulfide quinone reductase is topologically inverted and releases its oxidation products inside mitochondria where they are oxidized further; and ii) the sulfide quinone reductase retains its bacterial topology but, in contrast with bacteria, the formation of oxygenated products occurs in the intermembrane space or the cytosol.
The hypothesis of two sites of oxygen consumption has important consequences with regard to the possibility of an uncoupled state of mitochondria when sulfide is oxidized. As shown in Fig. 3B
, the experimental points obtained in the presence of sulfide are above the points obtained with succinate or L-alpha-glycerophosphate. In these experiments it is assumed that the oxygen consumption gives an unbiased measurement of the proton circulation at the level of respiratory chain and the leakage pathway(s). If this is the case then the displacement of the curve to top indicates an increase in proton flux that is explained by an increased proton leak (uncoupling). Therefore, in the first instance respiration appears to be uncoupled when sulfide is oxidized. However, the discussion made above indicates that oxygen consumption in the presence of sulfide is not uniquely due to the circulation of electrons from coenzyme Q to cytochrome oxidase. A situation that obviously biased the evaluation of proton circulation on the basis of oxygen consumption when sulfide is oxidized. Accordingly, multiplying the values of oxygen consumption in presence of sulfide by the ratio 0.5/0.79 would provide an evaluation of the maximal amount of oxygen that is used for proton pumping by respiratory complexes III and IV in mitochondria. With this correction factor, the curves obtained in presence of sulfide (dotted lines in Fig. 3
) looked very much like those obtained with L-alpha-glycerophosphate and there is no more evidence for a significantly increased proton leak in presence of sulfide when compared with the other substrates (Fig. 3B
). This conclusion is dependent on a mechanistic hypothesis that allows determination of the 0.5/0.79 correction factor for sulfide. Analysis using uncorrected values within the frame of the metabolic control analysis, which is not dependent on a mechanistic hypothesis, led to the similar conclusion of the absence of uncoupling in presence of sulfide. Details about this analysis could be found in the supplementary material available online. At this point it must be recalled that this study deals with short-term exposure to sulfide of cells grown in its absence. Our previous study in which an increased proton leak was evidenced dealt with the adaptative response after hours of exposure to sulfide, which allowed the induction of gene expression (7)
.
Energization of mitochondria allows phosphorylation of ADP into ATP at the level of FoF1 ATP synthase. This constitutes the major source of energy for cells and the sole source of energy when substrates other than glucose are utilized. While a normal energization of mitochondria was observed it was not possible to record a stable state 3 rate of respiration in presence of ADP. We suspect that accumulation of sulfide oxidation products in the measurement chamber led to a gradual inhibition of the FoF1 ATP synthase. Such an inhibition has been observed with yeast mitochondrial FoF1 ATP synthase and sulfate ion (35)
. This inhibition observed in vitro does not invalidate the hypothesis of sulfide being a usable substrate since in vivo cells are not maintained in a confined environment and the bloodstream would allow elimination of these inhibitors. Sulfide oxidation coupled to ATP production has been observed by others in invertebrates (21
, 22
, 31)
and with chicken mitochondria (20)
. However, in this latter study both the method (luciferase) and the mitochondria (chicken liver) used could not evaluate if a significantly high and stable state 3 rate could be maintained.
It is possible with these experiments to estimate the affinity of mammalian mitochondria to sulfide. For example maximal respiratory rate and complete energization in the Fig. 1
were attained within 30 s with an infusion rate of 0.1 nmol/sec (6 nmol/minute) in a chamber vol of 1.5 ml. This means that in absence of sulfide consumption the concentration would have been 2 µM (0.1 x 30/1.5). Actual concentration was obviously lower and energization started immediately, with which the equipment used meant a few seconds (a maximal concentration of 0.2 µM would be attained in 3 s). Therefore, concentrations as low as the low micromolar range allow a maximal rate of oxidation and concentrations of 
0.1 µM are sufficient for mitochondrial oxidation to take place. A similar evaluation could be deduced when the inhibition by sulfide of the cytochrome oxidase in HT29 cells homogenates is taken into account. This inhibition occurred with an IC50 of 0.32 µM (7)
. Although a significant inhibition of cytochrome oxidase might occur before the maximal respiratory rate would be affected because this enzyme is present in excess (7)
, this would also suggest a maximal value in the low micromolar range around mitochondria for a maximal oxidation rate of sulfide.
It is noteworthy that colonocytes that are naturally exposed to sulfide showed maximal rates of sulfide oxidation that matched with that of other mitochondrial substrates. Therefore, in colonocytes it is not possible to conclude whether this limitation reflects the maximal activity of the sulfide-quinone oxido reductase or that of the other complexes of the respiratory chain. In the other cell lines examined so far, the sulfide oxidation is taking place but its maximal rate is of limited importance when compared to other (normal) mitochondrial respiratory substrates. Our study indicates that the oxidative metabolism of colonocytes could eliminate significant amounts of sulfide and lower its concentration to harmless levels when oxygen is available and when sulfide is furnished at a continuous rate remaining below a threshold corresponding to the oxidative ability of cells. Above this threshold, sulfide accumulates and leads to inhibition of the cytochrome oxidase. Although as mentioned before, the bloodstream would be an efficient way to remove the sulfide in excess, it appears likely that the colonocytes are provided with an additional mechanism that has been already observed in invertebrates (19)
: Oxidation of sulfide could continue, but electrons from sulfide would be directed toward the reduction of malate to succinate by reversing the succinate dehydrogenase complex (complex II in Fig. 5B
). Since this oxidation of sulfide takes place in the absence of oxygen consumption, proton pumping, and mitochondrial energization, our study provides no estimation of the possible efficiency of this anaerobic mechanism that depends on malate availability, succinate elimination, and the possibility for cells to obtain energy (ATP) by another anaerobic mechanism (glycolysis). Therefore, it must be recognized that our study provides only an indication that this could occur; the scheme shown in Fig. 5B
should be considered as a speculation. It can be mentioned also that our study failed to provide evidence for an alternative oxidase in mammalian cells. In contrast with another animal model of adaptation to sulfide, the lugworm Arenicola marina (23)
.
In conclusion, we demonstrate that sulfide cannot any longer be considered solely as a toxic compound in the mammalian gut, since a moderate production of this compound would in fact feed the epithelial cells by providing them with a respiratory substrate. Mitochondrial efficiency and high respiratory activity are the characteristics of differentiated cells, whereas anaerobic glycolysis is characteristic of undifferentiated and rapidly dividing cells (24)
. There could be an optimal rate of production of sulfide (36
, 37)
that would feed the differentiated epithelial cells present in the upper part of the colonic crypt, whereas a higher rate would harm them and allow diffusion of sulfide toward the lower part of the colonic crypt where pluripotent and dividing cells are present. A recent study has shown that hydrogen sulfide can act as a genotoxic agent toward dividing colonic epithelial cells (38)
.
H2S is now recognized as an intracellular gaseous mediator like NO or CO (39
, 40)
. It is remarkable that these molecules also share the property of inhibiting mitochondrial cytochrome oxidase. In the case of H2S, oxidation by mitochondria would constitute a mean to extinguish this signal. Accordingly, in cells other than colonocytes, the limited capacity of mitochondria to oxidize sulfide might not be physiologically negligible. Finally, this first demonstration of the oxidative use of sulfide by mammalian mitochondria underlines the strong conservation of the metabolism of sulfur in the living world (30)
.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
2 Present address: UMR 914 Institut National de la Recherche Agronomique- Institut National Agronomique, 16 rue Claude Bernard, 75231 Paris Cedex 05, France.
Received for publication October 23, 2006. Accepted for publication January 4, 2007.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  D. L. Miller and M. B. Roth Hydrogen sulfide increases thermotolerance and lifespan in Caenorhabditis elegans PNAS, December 18, 2007; 104(51): 20618 - 20622. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. L. Wallace, M. Dicay, W. McKnight, and G. R. Martin Hydrogen sulfide enhances ulcer healing in rats FASEB J, December 1, 2007; 21(14): 4070 - 4076. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
