Oxidation of hydrogen sulfide by human liver mitochondria

Nitric Oxide xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Nitric Oxide journal homepage: www.elsevier.com/locate/yniox

Oxidation of hydrogen sulfide by human liver mitochondria Nada Helmy a,b,c,d,e,1, Carina Prip-Buus a,b,c,1, Corinne Vons a,b,c,d,e, Véronique Lenoir a,b,c, Abbas Abou-Hamdan a,b,c, Hala Guedouari-Bounihi a,b,c, Anne Lombès a,b,c, Frédéric Bouillaud a,b,c,⇑ a

Inserm U1016, Institut Cochin, 75014 Paris, France CNRS UMR8104, Institut Cochin, 75014 Paris, France c Université Paris Descartes UMR-S1016, Institut Cochin, 75014 Paris, France d Service de chirurgie digestive et métabolique, APHP, Hôpitaux universitaires de Seine Saint Denis Hôpital Jean Verdier, 93143 Bondy Cedex, France e Université Paris 13, Bobigny, France b

a r t i c l e

i n f o

Article history: Available online xxxx Keywords: Sulfide Mitochondria Liver Obesity Oxygen consumption Respiration Bioenergetics Blood pressure

a b s t r a c t Hydrogen sulfide (H2S) is the third gasotransmitter discovered. Sulfide shares with the two others (NO and CO) the same inhibiting properties towards mitochondrial respiration. However, in contrast with NO or CO, sulfide at concentrations lower than the toxic (lM) level is an hydrogen donor and a substrate for mitochondrial respiration. This is due to the activity of a sulfide quinone reductase found in a large majority of mitochondria. An ongoing study of the metabolic state of liver in obese patients allowed us to evaluate the sulfide oxidation capacity with twelve preparations of human liver mitochondria. The results indicate relatively high rates of sulfide oxidation with a large variability between individuals. These observations made with isolated mitochondria appear in agreement with the main characteristics of sulfide oxidation as established before with the help of cellular models. Ó 2014 Published by Elsevier Inc.

1. Introduction 1.1. Hydrogen sulfide a gasotransmitter, a mitochondrial poison, a mitochondrial substrate Hydrogen sulfide was recognized as the third gasotransmitter in mammals [1]. The two others are NO and CO. The three gases are equally toxic to mitochondrial respiration with an inhibiting power on the mitochondrial complex IV (Scheme 1) similar to that of cyanide [2]. Sulfide brought a further level of complexity because of two opposite/complementary metabolic strategies with regard to sulfide linked to the presence/absence of oxygen. For anaerobic microbes, sulfide is a metabolic waste resulting from the use of an oxidized form of sulfur (such as sulfate) as final electron acceptor. In presence of oxygen, sulfide is used as an electron donor by sulfo-oxidant bacteria. Moreover, a majority of mitochondria in a mammalian organism are able to oxidize sulfide by means of a sulfide quinone reductase (SQR) [3]. Present knowledge shows that this duality exists in a single eukaryotic cell [4] and is sometimes ⇑ Corresponding author at: Inserm U1016 Institut Cochin, 24 rue du Faubourg Saint Jacques, 75014 Paris, France. E-mail address: [email protected] (F. Bouillaud). 1 The contribution of N.H. and C.P.B. is equivalent.

presented as the grounding principle of oxygen sensing [5]. Therefore, and in contrast to the other gasotransmitters, sulfide has two opposite effects on mitochondria: at low concentrations it is a mitochondrial substrate and at high concentrations it is a poison. SQR initiates sulfide oxidation with nanomolar concentrations of sulfide while the transition to toxicity occurs in the micromolar range reviewed in [6]. Another complication with sulfide is that two different sources exist in the organism. The first is the metabolism of eukaryotic cells in which three different enzymes are susceptible to release hydrogen sulfide [7] and are thought to be key elements of the sulfide signaling pathway. The second source is the anaerobic microbiota hosted in the digestive tract where sulfide concentrations could reach values well within the toxic range representing a threat for the surrounding mammalian tissues [8]. Fortunately, the colonic wall has a large capacity to oxidize sulfide based on the SQR activity, restraining diffusion of colonic sulfide in the whole organism [9]. Whether the colonic sulfide is to be considered for signaling is still unclear but at least it contributes to blood levels in sulfide and related molecules [10]. The presence of the SQR in many different cell types/organs of the mammalian organism suggests that the need to oxidize sulfide is not restricted to the gastrointestinal tract and that the spontaneous sulfide release by tissues has to be checked either to avoid toxic accumulation and/or to control sulfide signaling.

http://dx.doi.org/10.1016/j.niox.2014.05.011 1089-8603/Ó 2014 Published by Elsevier Inc.

Please cite this article in press as: N. Helmy et al., Oxidation of hydrogen sulfide by human liver mitochondria, Nitric Oxide (2014), http://dx.doi.org/ 10.1016/j.niox.2014.05.011

2

N. Helmy et al. / Nitric Oxide xxx (2014) xxx–xxx

2H2S

O2 H2S2O3

Diox

H+

SQR

Glutamate Malate

I

Oligomycin

1/ O 2 2

Rotenone

Q

III

V

IV ADP +Pi

II

H 2O ATP

Succinate Scheme 1. This scheme presents the different partners controlling mitochondrial sulfide oxidation in our experiments. The different complexes of the mitochondrial respiratory chain are shown as squares with roman numerals with the redox enzymes as black and white boxes. The sulfide oxidation pathway is simplified to two enzymes: the sulfide quinone reductase (SQR) and the dioxygenase (Diox). The quinone (Q) is the redox shuttle between the different hydrogen donors shown (complexes I, II and SQR) and one acceptor (complex III). The reactants (sulfide and dioxygen) as well as final products (thiosulfate and water) are shown. The blue circuit divides these complexes into two functional domains: the first (complexes I, III and IV) associates redox reactions to proton pumping, and the second (complex V) uses the proton gradient to phosphorylate ADP into ATP. Therefore, via the proton circuit the activity of complex V controls the redox activity of the other complexes (coupling). Rotenone inhibits complex I redox activity and thus blocks oxidation of glutamate/malate but not that of sulfide or succinate. Oligomycin blocks the phosphorylating activity of complex V. When phosphorylation indeed controls electron transfer (coupling), oligomycin represses oxygen consumption to a low value explained by a proton leakage across the inner membrane, which is low in coupled mitochondria.

Liver mitochondria are the more commonly used to study mitochondrial metabolism. The liver is in a peculiar situation with regard to sulfide. On one side, liver mitochondria are considered as representative of mitochondria on a general basis. On the other side, the liver might be seen as a second barrier against colonic sulfide diffusion as in case sulfide would escape to the oxidation by the colonic wall it would be driven first to the liver through the portal vein [6]. Within the frame of an ongoing study on the mitochondrial state in the liver of obese people an evaluation of the intensity of the sulfide oxidation process was made in a few samples.

1.2. Quantifying mitochondrial tolerance to sulfide: the relative sulfide exposure The toxicity of sulfide makes impossible to probe for sulfide oxidation by adding an amount of this substrate sufficient to sustain oxidation for minutes. Therefore, to establish a stable oxidation rate sulfide has to be provided continuously as it is consumed. The consequences of this infusion protocol are (i) the actual concentration of sulfide is unknown. (ii) If a steady state is maintained without net sulfide accumulation then the infusion rate equals oxidation rate. At the opposite, if effect of known concentrations is to be observed conclusions should be deduced from the subsequent transient oxidation/inhibition states. The sulfide oxidation pathway could be divided into two different blocks: one is SQR itself and associated enzymes (yellow in the Scheme 1). The second block (blue in the Scheme 1) is constituted by the mitochondrial respiratory chain as sulfide oxidation is dependent upon the electron transfer rate in complexes III and IV, which in coupled mitochondria is essentially controlled by the downstream complex V activity. Therefore, unless sulfide would uncouple mitochondria, the electron transfer rate in the mitochondrial respiratory chain is not supposed to change when sulfide is infused and the maximal sulfide oxidation rate results from the interplay between these two blocks. The practical consequence is that the measurement of the respiratory rate (oxygen consumption) in absence of sulfide provides a reference rate

directly proportional to the electron transfer rate in complexes III and IV. The ratio between the sulfide infusion rate and this reference rate (JH2S/JO2) defines a relative sulfide exposure (RSE) imposed to the respiring preparation [3]. Two molecules of sulfide are engaged in the reduction of quinone while only half a molecule of dioxygen is needed to re-oxidize the quinone (scheme 1), this explains why the RSE value could rise above one [3]. By increasing gradually the sulfide infusion rate, the aim of this study was to determine the maximal value of RSE at which the human liver mitochondria could still oxidize all the incoming sulfide. Theoretically, the respiring mitochondria would be able to support indefinitely any sulfide delivery rate lower or equal to this maximal RSE value as no sulfide accumulation occurs. Practically, it has been demonstrated for example that CHO cells could sustain intense sulfide oxidation for a period of 20 min [6].

2. Materials and methods 2.1. Human patients This study was included in an ongoing survey of the mitochondrial state in the liver of obese patients undergoing bariatric procedures and enrolled in a prospective cohort at a specialized university hospital. Morbidly obese patients fulfilled the indications of eligibility for bariatric surgery in accordance with French High Authority of Health (HAS) recommendations [11]. Patients showing evidence of autoimmune, inflammatory, or infectious hepatic diseases including viral hepatitis, cancer, hepatotoxic treatment, or known alcohol consumption (>20 g/day) were excluded from this study. The research protocol obtained the agreement of the regional Person Protection Committee (CPP Ile de France X) in February 2010. All patients, after oral information, signed an informed consent. Preoperative routine workup was performed, and all patients underwent at the morning of the operation: (1)vital signs check-up, of which the blood pressure (systolic/diastolic), was used as reference for this study; (2) a blood test including full liver functions, fasting glycemia and insulinemia. Liver biopsies were realized during laparoscopy, at the beginning

Please cite this article in press as: N. Helmy et al., Oxidation of hydrogen sulfide by human liver mitochondria, Nitric Oxide (2014), http://dx.doi.org/ 10.1016/j.niox.2014.05.011

N. Helmy et al. / Nitric Oxide xxx (2014) xxx–xxx

of the surgical procedure, on the edge of the left hepatic lobe, deep in the parenchyma without any tissue trauma or electro coagulation. The biopsy was immediately divided into three portions: (1) for metabolic flux studies and mitochondrial respiration, both realized on fresh liver sample the same day of surgery; (2) was formalin-fixed and paraffin-embedded for histological analysis; (3) was frozen and stored at 80 °C. 2.2. Mitochondrial preparation Human and rat liver mitochondria were isolated using differential centrifugation. Briefly, fresh liver samples were homogenized in an isolation buffer (300 mM Sucrose, 5 mM Tris–HCl, 1 mM EGTA, pH 7.4) at a ratio of 1 g of liver/ 6 mL using glass Teflon homogeniser. After centrifugation at 600 g for 10 min at 4 °C, supernatants containing mitochondria were further centrifuged at 8500 g for 10 min at 4 °C. Pelleted mitochondria were resuspended in the isolation buffer, and protein concentration was determined by the Bradford method with BSA as a standard. 2.3. Measurement of mitochondrial respiration and sulfide oxidation The Oroboros O2k apparatus was used to monitor mitochondrial respiration. Sulfide infusions or injections were made with the Tip2k minipump (Oroboros instrument). The respiration medium contained 100 mM KCl, 20 mM Sucrose, 2 mM EGTA, 10 mM Pi, and 0.4% (w/v) free fatty acid BSA pH 7.4. The mitochondrial suspension was diluted to 0.1–0.4 mg protein/mL in the respiration medium to run experiments. 2.4. Sulfide solution, preparation and use A stock 1 M sulfide solution was prepared from Na2S (Sigma) for each experiment. Ten microliter of this solution were diluted in 2 mL of milliQ water and this 5 mM solution was immediately loaded in the glass syringes of the minipump so that it was not exposed to air more than few tenths of seconds. The pH of sulfide solutions was not equilibrated to a neutral value as it would enhance the volatility of sulfide by increasing H2S content. The pH effect of infusing this 5 mM Na2S is expected to be negligible as the sum of the sulfide infused means few tenth of micromolar in concentration while the respiration medium is buffered by 10 mM Phosphate. A confined sulfide solution becomes anoxic with time (not shown). The likely explanation would be a slow auto-oxidation of sulfide that would be limited by the oxygen content of the water #250 lM dioxygen at room temperature. Maximal oxidation with formation of sulfate (SO4) would mobilize 125 lM sulfide to exhaust 250 lM O2. Alternatively a moderate oxidation with formation of thiosulfate (S2O3) would engage 333 lM. To minimize the relative loss of sulfide due to auto-oxidation it is therefore recommended to use relatively concentrated solutions and to infuse the smallest volumes allowing sufficient precision. The 5 mM solution was a compromise with lowest infusion rates of few nL/s (Figs. 2 and 3). The addition of the anoxic sulfide solution into the 2 mL volume of the chamber causes a decrease in the oxygen concentration proportionate to the volume added. This is interpreted by the software as an oxygen consumption rate, and because of that constitutes a second reason to reduce the volume added. The expected effect could be calculated: fast injection of 2 ll of de-oxygenated solution in 2 ml would decrease oxygen concentration by 1‰. This decrease would be theoretically fully recorded between two successive data point (2s). Consequently, the apparent oxygen consumption rate would be 0.5‰ of the initial oxygen content per second. Firstly, this explanation highlights the complication that the absolute value of this rate is expected to be proportional to the initial oxygen concentration. Secondly, if air saturated water at 25 °C is considered (#250 lM of dioxygen) it means

3

a significant value of 125 pmol/(s.ml). To confirm experimentally the fact that injection of sulfide tallies with the expected consequences of this dilution effect we performed experiments comparing infusion of the «aged in syringe» 5 mM Na2S solution with that of de-oxygenated (Argon saturated) water. With regard to the fast kinetic (few tenth of seconds after injection) we found essentially no difference between the two (Fig. 1). However, in the longer term a low persistent rate could be detected after sulfide injections higher than 30 lM, which is likely to represent the auto-oxidation process. Consequently, while the injections generate significant artefacts in the short term, the low autoxidation rate, which appears dependent on concentrations of tenth of micromolar sulfide, would hardly interfere with respiratory rate measurements. Importantly, the measuring software uses several (5–40) data points to calculate the oxygen consumption rate and therefore on one side it lowers the amplitude of the artefact but on the other side it extends it over time. Oxygen concentration data points are separated by two seconds and in this study 10 data points were used to calculate oxygen consumption rate (Figs. 1–3). Therefore, a delay of 20 s after the injection step appears necessary to eliminate from the rate calculation this dilution artefact. This delay is in fact longer because of the mixing and of the time constant of the electrodes. This is illustrated in the experiment shown in Fig. 1 where the duration of the injection step was set to one second. However, as seen with the use of de-oxygenated water, the complete evaluation of the transition from one stable concentration to another occurred after the recording of several successive data points, which delayed further the return of the rate back to zero (Fig. 1). This dilution artefact is detected with injections above 1 ll (Fig. 1) and clearly visible with injections of more than 2 ll (5 lM) on the right side of Figs. 2 and 3. In contrast, sulfide infusions in the nL/s range (left part of Figs. 2 and 3) induce low artefactual rates that remain usually of marginal importance with regard to measurable mitochondrial oxygen consumption rates.

3. Results and discussion 3.1. Maximal rate of sulfide oxidation The primary aim of the experiments was to determine a maximal relative sulfide exposure (RSE) that could be supported by the studied mitochondrial preparation. The followed strategy is visible in the early (left) part of the experiment shown in Fig. 1A. To reproduce a situation relevant to physiology, liver mitochondria were provided with substrate (glutamate/malate, 5 mM each) and with ADP (1.5 mM) to establish a phosphorylating (state 3) respiratory rate. This rate was used as the reference for calculation of the sulfide solution infusion rates necessary to reach the desired RSE values. The sulfide infusions caused an immediate increase in the oxygen consumption rate that within seconds reached a new higher value that remained stable as infusion took place, and dropped back to the reference value as soon as infusion ceased. This sequence of events indicates a steady state where sulfide infusion was compensated by oxidation with no sulfide accumulation. Then a new (higher) infusion rate was realized for the following infusion sequence. The first indication that sulfide infusion exceeds the maximal sulfide oxidation rate is a sulfide accumulation that allowed the sulfide oxidation to continue after the infusion ceased. This is taking place at the end of the fourth infusion in Fig. 1A. Two others indications could be used: a non proportional increase in oxygen consumption with regard to the previous infusion and the occurrence of inhibition of respiration during the infusion caused by sulfide accumulation. The maximal RSE tolerated by the mitochondria could be estimated from the highest infusion rate that produced a stable and synchronous elevation of oxygen consumption (third infusion in Fig. 1 A/B).

Please cite this article in press as: N. Helmy et al., Oxidation of hydrogen sulfide by human liver mitochondria, Nitric Oxide (2014), http://dx.doi.org/ 10.1016/j.niox.2014.05.011

4

N. Helmy et al. / Nitric Oxide xxx (2014) xxx–xxx

Oxygen concentration (µM)

+ identical volumes of: De-oxygenated water 5mM Na2S solution

Oxygen consumption rate (pmol/(s.ml))

Time h:min Volume (µl) [Na2S] (+µM)

0.4

0.8

1.6

3.2

4

4.8

6.4

8

8.8

10

12.8

20

25

30

35

40

50

1

2

4

8

10

12

16

20

22

25

32

50

62

74

86

98

122

Fig. 1. Injection of fixed volumes of deoxygenated water (blue traces) or of 5 mM Na2S solution (red traces). The X axis is the time (h:mm). Top: oxygen concentration, bottom: oxygen consumption rate calculated over a period of 20 s (ten data points). Vertical lines indicate the time of injection of the volumes indicated at the bottom (in dark blue) of de-oxygenated water or of a 5 mM Na2S solution. The increase in Na2S concentration caused by each injection is also indicated at the bottom (red).

This maximal sulfide oxidation rate was determined for twelve human patients. The values are expressed as RSE in the Fig. 4. The values with rat liver mitochondria deduced from the maximal rates of sulfide oxidation determined in [6] meant a RSE value around 0.1. In contrast, mitochondrial preparations from human liver showed values above 0.2 and a large variability with samples able to withstand RSE values above 1. Consequently, few preparations offered unique opportunities to study sulfide oxidation in isolated mitochondria (see Section 3.2). Gaseous signaling is highly relevant to cardiovascular domain, and concerning sulfide it was reported that cystathionine gamma lyase (CSE) knockout mice showed hypertension [12]. Consequently, we attempted to correlate the sulfide oxidation capacity in human liver mitochondria to arterial blood pressure (Fig. 4). The set of samples studied showed a positive relationship between the maximal RSE tolerated by mitochondria and blood pressure. With patients 1 to 11 this correlation reached statistical significance with the Spearman rank order test (correlation coefficient = 0.673, P = 0.021) however the inclusion of the last patient who showed very low values of blood pressure (99–54) deteriorated deeply the correlation and significance disappeared (Fig. 4). 3.2. Few characterisitics of mitochondrial sulfide oxidation The experiments shown in Figs. 2 and 3 were made with the same preparation of human liver mitochondria, which because of its high tolerance to sulfide offered the opportunity to compare

observations with isolated mitochondria to the characteristics determined before with cellular models [3,6]. The right (late) part of the experiment in the Fig. 1 used sulfide injections to reach known concentrations. In presence of glutamate/malate and ADP (Fig. 2A) injections in the low micromolar range (2–3 lM) produced transient increase in oxygen consumption rate whose maxima approached the maximal rate observed with infusions. With higher concentrations (5–7 lM) a short period of a stable high rate could be observed. The last addition (12 lM) resulted in a complex kinetics starting with the artefact of addition that contrasted with the subsequent inhibition phase followed by a progressive acceleration of oxygen consumption that culminated few seconds before the return to the «no sulfide rate» indicating the exhaustion of added sulfide. Fig. 3 shows an experiment aiming to provide evidence that sulfide oxidation is coupled to ATP synthesis. In presence of glutamate/malate, the ADP phosphorylation was let to proceed (Fig. 3A) or was inhibited by the addition of oligomycin (Fig. 3B). The increase in oxygen consumption rate was lower when ADP phosphorylation was inhibited and as a consequence the consumption of the same amount of sulfide was significantly longer (compare for example the 2 lM additions). Estimation of the ratio between the oxygen consumption rates in absence/presence of oligomycin after sulfide additions (1.5–5 lM) produced values up to 3.9 (Supplementary Figure). In other words, sulfide oxidation was fast or slow according to the possibility for the mitochondria to phosphorylate ADP into ATP or not. This is the definition of

Please cite this article in press as: N. Helmy et al., Oxidation of hydrogen sulfide by human liver mitochondria, Nitric Oxide (2014), http://dx.doi.org/ 10.1016/j.niox.2014.05.011

5

N. Helmy et al. / Nitric Oxide xxx (2014) xxx–xxx

A 250

125

ADP

a Infusion Na2S

9nl/s

18nl/s

24nl/s

b cde f g

h

i

6µM

7.2µM

12µM

32nl/s

B 250

125

a

ADP Rotenone 9nl/s

18nl/s

24nl/s

b cde f g

h

i

6µM

7.2µM

12µM

32nl/s

Fig. 2. The same human liver mitochondrial preparation was present in the two chambers of the O2k oxygraph: (A) chamber A, (B): chamber B. The oxygen concentration is indicated by the blue trace. The oxygen consumption rate (calculated as in Fig. 1) is shown in red. Events (additions, initiation/stop of infusion) are indicated by vertical lines. The phosphorylating state 3 of respiration was obtained by addition of glutamate/malate 5 mM each (not shown here) followed by ADP (1.5 mM). The resulting respiratory rate was: 43.4 pmol O2/(s.mL) and was the reference to calculate the relative sulfide exposure. Then rotenone (2 lM) was added in the chamber B to inhibit complex I. The same simultaneous sequence of infusions/injections of sulfide (5 mM solution) was made in the two chambers, starting with infusions at 9, 18, 24 and 32 nL/s. The relative sulfide exposure (RSE) is calculated with the formula (((nL/s)  5 mM)/(2 mL))/43.4 pmol/(s.mL) giving for these infusion rates values of 0.5, 1, 1.38 and 1.84, respectively. Since the last infusion (RSE 1.84) allowed accumulation of sulfide, the value of 1.38 was retained as estimation of the maximal RSE for this mitochondrial preparation. Then, succinate (7.5 mM final) was added in chamber B and twelve successive injections of sulfide were performed. For the sake of clarity, the first nine are indicated by letters (values in lM): 0.1(a), 0.3 (b), 0.2 (c), 0.6 (d), 1.2 (e), 1.8 (f), 2.4 (g), 3.6 (h), 4.8 (i) and then 6, 7.2 and 12 lM as indicated.

coupling. No direct demonstration of ATP synthesis is produced here however it should be mentioned that it has been demonstrated in the past with chicken liver mitochondria [13]. In presence of glutamate/malate, the addition of oligomycin reduced the mitochondrial respiration by a factor 4.1, revealing a similar coupled state between sulfide oxidation and that of a classical mitochondrial substrate. Consequently, this experiment provided no evidence for uncoupling of the sulfide driven respiration, which is fully in accordance with our first report showing efficient energization of mitochondrial membrane when sulfide is oxidized [14]. In presence of glutamate/malate, injection of electrons in the respiratory chain occurs at the level of the mitochondrial complex I (scheme 1). Addition of rotenone blocks complex I and thus shutsoff mitochondrial respiration (Fig. 2B left). Infusion of sulfide restores mitochondrial oxygen consumption via the SQR and caused a large increase in oxygen consumption, which is explained by two sites of oxygen consumption: the dioxygenase and the complex IV (cytochrome oxidase). This increase is larger than that

observed when sulfide oxidation takes place at the same time as complex I driven respiration (Fig. 2A left). The respiratory coupling explains the difference. Indeed, in coupled mitochondria the electron transfer in the respiratory chain is limited by the proton return rate through the inner membrane and no increase in the complex IV reaction rate was expected when sulfide was infused over mitochondria oxidizing glutamate/malate. Consequently, the observed increase in oxygen consumption (Fig. 2A) is to be explained exclusively by the dioxygenase reaction. Importantly, the maximal sulfide oxidation rate does not appear different in conditions where sulfide oxidation takes place alone (Fig. 2B left) or when glutamate/malate oxidation in presence of ADP occurs (Fig. 2A left). This observation indicates that, in these conditions, the activity of mitochondrial complex I does not impair in any way SQR to operate at the same fast rate than when alone, and thus that the SQR takes precedence over complex I to reduce quinone. The right of the Fig. 2B allows comparison of the effect of the same known concentrations of sulfide as in Fig. 2A, but in

Please cite this article in press as: N. Helmy et al., Oxidation of hydrogen sulfide by human liver mitochondria, Nitric Oxide (2014), http://dx.doi.org/ 10.1016/j.niox.2014.05.011

6

N. Helmy et al. / Nitric Oxide xxx (2014) xxx–xxx

A 250

125

ADP

a b 4nl/s

c

d

e 1µM 2µM

2.5µM

5µM

7.5µM

10µM

15µM

4nl/s

B 250

125

ADP

a b 4nl/s

4nl/s

c d

e 1µM 2µM

2.5µM

5µM

7.5µM

10µM

15µM

Oligomycin

Fig. 3. In presence of glutamate/malate and ADP (first addition shown), two infusion sequences at 4 nL/s were made. Then, oligomycin (0.5 lg/ml #0.6 lM) was added in chamber B and injections of sulfide were performed 0.125 (a), 0.25 (b), 0.5(c), 0.75 (d), 1 (e), 1.5 (f), 2, 2.5, 5, 7.5, 10 and 15 lM. Rest of the legend as in Fig. 1.

presence of another mitochondrial substrate: succinate (in presence of rotenone). Under these conditions, the increase in oxygen consumption after sulfide injection is not as high as in presence of glutamate/malate and evidence for partial inhibition of respiration appeared with a lower concentration of sulfide. On one side a simple explanation could be that the succinate dehydrogenase (complex II) did not yield the SQR reaction to take place at the same rate and thus, in contrast with complex I, complex II opposed resistance to SQR in the competition for quinone reduction. On the other side since complex interactions between sulfide and mitochondrial activity have been evidenced [15,16] it should be borne in mind that the complete explanation might be not that straightforward. The ratio between the increase in oxygen consumption rate and the sulfide infusion rate provides an estimation of the oxygen to sulfide stoichiometry. In the experiment shown, this value was 0.74 in presence of rotenone and 0.53 in its absence. If a larger number of experiments is considered the mean value of the later is 0.57 (n = 8, min 0.44 max 0.75) while that of the former is 0.72 (n = 6, min 0.49 max 1.05). The calculation of these values was made with the assumption that the sulfide weighted was 100% Na2S and therefore would be an underestimation. In contrast, within one experiment and thus with the same sulfide solution, the ratio between the stoichiometry (oxygen consumption) in

presence/absence of rotenone is made independent of the actual sulfide content of the infused solution. This ratio is 1.41 in the experiment shown and 1.31 (n = 6, min 1.04, max 1.49) when more experiments are considered. Consequently, the present study suggests that the contribution of the cytochrome oxidase to oxygen consumption is significantly lower than that of the dioxygenase. This observation fits with the mechanism proposed [17] and with the results obtained with cellular models [3]. 4. Conclusion and perspectives The main conclusion of the present study is that human liver mitochondria show the capacity to oxidize significant amount of hydrogen sulfide. The reference rate used to calculate RSE was the maximal rate of mitochondrial oxygen consumption associated to ATP production (saturating ADP concentration). This suggests that in the human liver the mitochondria would be able to neutralize a flux of incoming sulfide representing 20% or more of the oxygen consumption of the organ. This appears disproportionate to the expected exposure of the liver to sulfide and possible explanations for that are proposed below. It should be recalled here that all the patients were morbidly obese and it is not possible to conclude that the same is true in the general healthy population.

Please cite this article in press as: N. Helmy et al., Oxidation of hydrogen sulfide by human liver mitochondria, Nitric Oxide (2014), http://dx.doi.org/ 10.1016/j.niox.2014.05.011

N. Helmy et al. / Nitric Oxide xxx (2014) xxx–xxx

7

160

The experiments shown (Figs. 2 and 3) indicate that the sulfide oxidation by human liver mitochondria follows the essential rules defined before with cellular models [3]: (i) sulfide oxidation takes precedence over complex I driven respiration, (ii) sulfide oxidation is coupled to complex V activity and presumably ATP synthesis as noticed before [3,14], and (iii) the experimental stoichiometry in mitochondria appears in agreement with the reaction mechanism proposed [17]. However, it should be stressed that there are still uncertainties about the exact reaction scheme [18]. To improve our knowledge in this matter it will be necessary to improve the accuracy of the experimental determination of stoichiometry. This would be dependent on the availability of a representative and reliable mitochondrial preparation with a sufficiently high SQR activity. This report deals with immediate effect of exogenous sulfide on mitochondrial oxygen consumption and cannot pretend to provide an in depth description of sulfide effect on mitochondria and liver cell taking into account other complex effects of sulfide or of sulfide releasing reactions on mitochondrial activity [15,16]. These aspects were merely out of our reach. However, this basic analysis is enough to impose time and concentration constraints.

150

4.1. Time

140

Unless sulfide concentration reaches toxic levels able to inhibit durably respiration (incompatible with life unless restricted to specific sites) micromolar concentrations of sulfide are highly unstable and SQR activity is expected to restore in a short time (seconds) the concentration to a steady state in which sulfide elimination compensate for sulfide delivery.

A

B

Systolic Blood Pressure (mmHg)

130 120 110

4.2. Concentration

100 90 0.4

0.8

1.2

1.6

RSEm

1.2

1.6

RSEm

C Diastolic Blood Pressure (mmHg) 110

100

90

80

70 60

50 0.4

0.8

Fig. 4. (A): Box plot for the values of maximal RSE of the 12 preparations of human liver mitochondria compared with the RSE values for rat mitochondria (n = 9) were deduced from the Vmax for SQR activity as determined in [6]. (B): X axis: maximal RSE of the preparations of human liver mitochondria, Y axis: systolic Blood pressure. The correlation between blood pressure parameters and RSE was tested with the Spearman rank order test, the linear regression line (solid) and 95% confidence intervals (dotted traces) are shown: correlation coefficient = 0.309, P value = 0.317 (non significant). (C): X axis: maximal RSE of the preparations of human liver mitochondria, Y axis diastolic Blood pressure. The correlation is shown as above: correlation coefficient = 0.490, P value = 0.100 (non significant).

With the assumption that mitochondrial oxidation is the sole sulfide elimination pathway, in the steady state mentioned above the oxidation rate «R» consumes an amount of sulfide equivalent to the total delivery to the liver cell. The determination of the apparent affinity of rat liver mitochondria for sulfide led to a Michaelian kinetics with a km value close to 2 lM [6]. Then for a given concentration of sulfide «S» (in lM) R expressed relatively to its maximal possible value expressed as RSE (RSEm) is given by the formula R = RSEm.S/(S + 2). In this study RSEm ranged from 27% to 155% of what could be considered as a maximal physiological mitochondrial oxygen consumption rate. Estimation of the liver oxidative metabolism in mammals provided values from 2 (horse) to 18 (mouse) nmol O2 min 1 mg 1 [19]. In contrast, estimation of endogenous (enzymatic) sulfide release rate produced values expressed in pmol H2S min 1 mg 1 with a value of 20 in mouse liver [20], thus three orders of magnitude lower. This large difference is likely to be overestimated: (i) the oxygen consumption in [19] was expressed with reference to dry weight while that of sulfide release [20] used fresh weight, (ii) because in [20] the endogenous sulfide oxidation was unchecked the value of 20 pmol min 1 mg 1 is an underestimation, (iii) the contribution of the sulfide from the gastrointestinal tract brought by the portal vein is to be added to the endogenous liver release rate. It remains however likely that the sulfide influx to be neutralized by mitochondrial oxidation in liver is less than few percent of the oxygen consumption rate. Therefore, at steady state R<
Please cite this article in press as: N. Helmy et al., Oxidation of hydrogen sulfide by human liver mitochondria, Nitric Oxide (2014), http://dx.doi.org/ 10.1016/j.niox.2014.05.011

8

N. Helmy et al. / Nitric Oxide xxx (2014) xxx–xxx

patient showing the highest value of RSEm or 20.01/0.27 = 74 nM when the patient with the lowest RSEm is considered. This calculation indicates that the apparently disproportionately high SQR activity found in this study may reflect the need to maintain intracellular sulfide concentration within acceptable limits that are well below the km value of this enzyme. It also suggests that even if SQR expression remains orders of magnitude higher than apparently needed on the ground of enzymatic activity it still controls intracellular sulfide concentration because of its kinetic properties. Other possible explanations for the high SQR activity are (i) since the molecular activity of SQR is very high [18], variations around a low «background» expression would have large consequences in terms of the possible sulfide oxidation rate, (ii) while the loss of SQR would allow a fatal accumulation of sulfide, there is a negligible gain in repressing its expression to the actual need, (iii) the number of SQR enzymes is determined by other factors than the global enzymatic activity, for example space constraints, (iv) we underestimate the intensity of the sulfide exposure and sulfide redox dynamics in the liver. Finally, the correlation between sulfide oxidation and blood pressure parameters deserves further evaluation in humans. Acknowledgments The authors wish to thank the staff of the hospital Jean Verdier for their help during this study. The research on human liver was funded by a grant from the Association Française pour l’Etude du Foie (AFEF-Gilead Sciences 2012). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.niox.2014.05.011. References [1] R. Wang, Two’s company, three’s a crowd: can H2S be the third endogenous gaseous transmitter?, FASEB J 16 (2002) 1792–1798. [2] C.E. Cooper, G.C. Brown, The inhibition of mitochondrial cytochrome oxidase by the gases carbon monoxide, nitric oxide, hydrogen cyanide and hydrogen sulfide: chemical mechanism and physiological significance, J. Bioenerg. Biomembr. 40 (2008) 533–539.

[3] E. Lagoutte, S. Mimoun, M. Andriamihaja, C. Chaumontet, F. Blachier, F. Bouillaud, Oxidation of hydrogen sulfide remains a priority in mammalian cells and causes reverse electron transfer in colonocytes, Biochim. Biophys. Acta 1797 (2010) 1500–1511. [4] D.G. Searcy, Metabolic integration during the evolutionary origin of mitochondria, Cell Res. 13 (2003) 229–238. [5] K.R. Olson, N.L. Whitfield, Hydrogen sulfide and oxygen sensing in the cardiovascular system, Antioxid. Redox Signal. 12 (2010) 1219–1234. [6] C. Szabo, C. Ransy, K. Modis, M. Andriamihaja, B. Murghes, C. Coletta, G. Olah, K. Yanagi, and F. Bouillaud, Regulation of mitochondrial bioenergetic function by hydrogen sulfide. Part I. Biochemical and physiological mechanisms. Br. J. Pharmacol. (in press). [7] O. Kabil, R. Banerjee, Redox biochemistry of hydrogen sulfide, J. Biol. Chem. 285 (2010) 21903–21907. [8] F. Blachier, A.M. Davila, S. Mimoun, P.H. Benetti, C. Atanasiu, M. Andriamihaja, R. Benamouzig, F. Bouillaud, D. Tome, Luminal sulfide and large intestine mucosa: friend or foe?, Amino Acids 39 (2010) 335–347 [9] S. Mimoun, M. Andriamihaja, C. Chaumontet, C. Atanasiu, R. Benamouzig, J.M. Blouin, D. Tome, F. Bouillaud, F. Blachier, Detoxification of H(2)S by differentiated colonic epithelial cells: implication of the sulfide oxidizing unit and of the cell respiratory capacity, Antioxid. Redox Signal. 17 (2012) 1– 10. [10] X. Shen, M. Carlstrom, S. Borniquel, C. Jadert, C.G. Kevil, J.O. Lundberg, Microbial regulation of host hydrogen sulfide bioavailability and metabolism, Free Radical Biol. Med. 60 (2013) 195–200. [11] Recommendation de bonne pratique no. 54: ‘‘Obésité’’ Prise en charge chirurgicale chez l’adulte, in: F.H.A.o.H. HAS (Ed.), Jan 2009. [12] G. Yang, L. Wu, B. Jiang, W. Yang, J. Qi, K. Cao, Q. Meng, A.K. Mustafa, W. Mu, S. Zhang, S.H. Snyder, R. Wang, H2S as a physiologic vasorelaxant: hypertension in mice with deletion of cystathionine gamma-lyase, Science 322 (2008) 587– 590. [13] R. Yong, D.G. Searcy, Sulfide oxidation coupled to ATP synthesis in chicken liver mitochondria, Comp. Biochem. Physiol. B: Biochem. Mol. Biol. 129 (2001) 129–137. [14] M. Goubern, M. Andriamihaja, T. Nubel, F. Blachier, F. Bouillaud, Sulfide, the first inorganic substrate for human cells, FASEB J. 21 (2007) 1699–1706. [15] K. Modis, C. Coletta, K. Erdelyi, A. Papapetropoulos, C. Szabo, Intramitochondrial hydrogen sulfide production by 3-mercaptopyruvate sulfurtransferase maintains mitochondrial electron flow and supports cellular bioenergetics, FASEB J. 27 (2012) 601–611. [16] M. Fu, W. Zhang, L. Wu, G. Yang, H. Li, R. Wang, Hydrogen sulfide (H2S) metabolism in mitochondria and its regulatory role in energy production, Proc. Natl. Acad. Sci. U.S.A. 109 (2012) 2943–2948. [17] T.M. Hildebrandt, M.K. Grieshaber, Three enzymatic activities catalyze the oxidation of sulfide to thiosulfate in mammalian and invertebrate mitochondria, FEBS J. 275 (2008) 3352–3361. [18] M.R. Jackson, S.L. Melideo, M.S. Jorns, Human sulfide:quinone oxidoreductase catalyzes the first step in hydrogen sulfide metabolism and produces a sulfane sulfur metabolite, Biochemistry 51 (2012) 6804–6815. [19] R.K. Porter, Allometry of mammalian cellular oxygen consumption, Cell. Mol. Life Sci. 58 (2001) 815–822. [20] D.R. Linden, L. Sha, A. Mazzone, G.J. Stoltz, C.E. Bernard, J.K. Furne, M.D. Levitt, G. Farrugia, J.H. Szurszewski, Production of the gaseous signal molecule hydrogen sulfide in mouse tissues, J. Neurochem. 106 (2008) 1577–1585.

Please cite this article in press as: N. Helmy et al., Oxidation of hydrogen sulfide by human liver mitochondria, Nitric Oxide (2014), http://dx.doi.org/ 10.1016/j.niox.2014.05.011