O2 chemoreceptors elicit cardiorespiratory reflexes in all vertebrates, but consensus on O2-sensing signal transduction mechanism(s) is lacking. We recently proposed that hydrogen sulfide (H2S) metabolism is involved in O2 sensing in vascular smooth muscle. Here, we examined the possibility that H2S is an O2 sensor in trout chemoreceptors where the first pair of gills is a primary site of aquatic O2 sensing and the homolog of the mammalian carotid body. Intrabuccal injection of H2S in unanesthetized trout produced a dose-dependent bradycardia and increased ventilatory frequency and amplitude similar to the hypoxic response. Removal of the first, but not second, pair of gills significantly inhibited H2S-mediated bradycardia, consistent with the loss of aquatic chemoreceptors. mRNA for H2S-synthesizing enzymes, cystathionine β-synthase and cystathionine γ-lyase, was present in branchial tissue. Homogenized gills produced H2S enzymatically, and H2S production was inhibited by O2, whereas mitochondrial H2S consumption was O2 dependent. Ambient hypoxia did not affect plasma H2S in unanesthetized trout, but produced a Po2-dependent increase in a sulfide moiety suggestive of increased H2S production. In isolated zebrafish neuroepithelial cells, the putative chemoreceptive cells of fish, both hypoxia and H2S, produced a similar ∼10-mV depolarization. These studies are consistent with H2S involvement in O2 sensing/signal transduction pathway(s) in chemoreceptive cells, as previously demonstrated in vascular smooth muscle. This novel mechanism, whereby H2S concentration ([H2S]) is governed by the balance between constitutive production and oxidation, tightly couples tissue [H2S] to Po2 and may provide an exquisitely sensitive, yet simple, O2 sensor in a variety of tissues.
- hydrogen sulfide concentration
tissue hypoxia in vertebrates can result from depletion of environmental O2, impairment of O2 uptake and delivery, or an inability of uptake and delivery to keep pace with tissue demand. While in most cells hypoxia has detrimental consequences, a few tissues are designed to sense acute changes in O2 and initiate corrective homeostatic responses when O2 tension (Po2) falls. Notable among these are chemoreceptors that initiate central cardiorespiratory responses, blood vessels that exert local control over tissue perfusion, and chromaffin cells in the adrenal medulla or central veins that provide a humoral source of catecholamines (20, 26, 37, 54).
The carotid body in many mammals appears to be the only tissue to monitor blood oxygenation and initiate cardiorespiratory responses during hypoxia (26). O2 sensing by type 1 glomus cells in the carotid body has been attributed to the plasma membrane, mitochondria, and/or heme-containing proteins (20, 36). Specific elements of the O2-sensing signal transduction cascade include a variety of voltage-dependent K+ (Kv) channels (22, 35), large-conductance Ca2+-dependent K+ channels (BKCa channels; see Ref. 38), TASK-like K+ channels (3, 48), NADPH oxidase (5, 11), increased reactive oxygen species (5), AMP-activated protein kinase (58, 59) and heme oxygenase-2 production of carbon monoxide (18, 38).
Vascular smooth muscle cells also have an intrinsic ability to “sense” O2 and transduce this into a physiologically relevant signal, i.e., hypoxic relaxation of systemic vessels and contraction of pulmonary vessels (24). A number of hypotheses have been offered to explain vascular O2 sensing, and many of these are similar to those proposed to occur in the carotid body. Hypoxic relaxation has been attributed to ATP-sensitive K+ channels (53), loss of Kv1.5 and Kv2.1 channel activity (45), intracellular acidosis (27), run down of ATP (46), redox control of cytosolic NADPH (57), modulation of internal Ca2+ stores (46), Rho-kinase (46), and hemoglobin-mediated catalysis of nitrate to nitric oxide (10). Hypoxic contraction of pulmonary arteries has been attributed to O2-sensitive K+ channels Kv1.5, Kv2.1, and Kv9.3 (2, 53, 33), redox sensors (1), reactive O2 species (21, 52), NADPH and NADPH oxidase (15, 57), cADP ribose (7), Rho kinase (8) and BKCa channels (34).
Although it is evident that hypoxia evokes a number of varied cellular responses, it is not clear how or where Po2 is coupled to cell activation, i.e., the O2 sensor. In fact, none of the hypotheses of the O2 sensor has received unequivocal support (1, 19, 47, 49, 50, 57). Thus it is possible that the O2 sensor remains to be identified.
We recently proposed a novel mechanism of O2 sensing in vascular and nonvascular smooth muscle that involves O2-dependent metabolism of hydrogen sulfide (H2S; see Refs. 6 and 30). In this model, constitutive cellular production of H2S is balanced by mitochondrial H2S oxidation; therefore, the concentration of vasoactive H2S is inversely and tightly coupled to O2 availability. We have also observed an inverse relationship between O2 and H2S production in trout hearts and have proposed that this is the mechanism whereby H2S contributes to ischemic preconditioning in the myocardium (56). Because attributes of this model have been demonstrated in blood vessels from the most ancient vertebrates (cyclostomes) and mammals, it is possible that it is a fundamental property of O2-sensing tissues in general. Thus an examination of this model in other O2-sensing tissues is warranted.
If, as we hypothesize, the metabolism of H2S serves as an O2 sensor in the chemoreceptive neuroepithelial cells (NECs) and it is involved in initiating cardiorespiratory responses to hypoxia, we would expect to see that 1) tissue production of H2S is inversely coupled to Po2 and 2) exogenous application of H2S mimics the cardiorespiratory responses to hypoxia at the level of the whole animal and cellular responses to hypoxia at the level of the NECs. Furthermore, removal of chemoreceptive tissue should have similar inhibitory effects on both hypoxia- and H2S-mediated responses. Fish were chosen for this work because 1) cardiorespiratory control in fish exhibits closer coupling to ambient hypoxia than that of mammals (25), 2) the first pair of fish gills are the primary O2 sensors in trout and are accessible to experimental manipulation (9), and 3) the chemoreceptor cells on the first pair of gills are homologous to the chemoreceptor cells on the carotid arteries, i.e., the carotid bodies, of mammals (26). We monitored the cardiorespiratory responses to intrabuccal H2S injection in intact trout and in trout with either the first or second pair of gills removed. To determine if exogenous H2S mimicked the hypoxic response, we looked for the presence of mRNA for H2S biosynthetic enzymes in a variety of tissues to determine their metabolic capability and then measured H2S production by homogenized gill tissue; and to determine if H2S was enzymatically generated and if H2S production was decreased in the presence of O2, we measured the effects of ambient hypoxia on plasma [H2S], and we evaluated the effects of inhibitors of H2S synthesis on cardiorespiratory responses to hypoxia. We also compared the effects of hypoxia and H2S on transmembrane potential (Em) in cultured zebrafish NECs (a trout model has not yet been developed and validated), the putative branchial O2-sensing cells (16). Our results are consistent with H2S involvement in O2 sensing in fish chemoreceptors.
MATERIALS AND METHODS
Indiana University School of Medicine-South Bend, South Bend.
Rainbow trout (Oncorhynchus mykiss, mixed Kamloops strain; 0.4–0.7 kg) of both sexes were purchased from a hatchery in northern Michigan and kept in circulating 2,000-liter tanks at 12–14°C and under 12:12-h light-dark cycles. Fish were acclimated to these conditions a minimum of 2 wk and were fed up to 48 h before experimentation with commercial trout pellets.
Mitochondria were purified from ventricles of larger steelhead trout (O. mykiss, skamania strain, 3–7 kg, both sexes). These fish were captured by the Indiana Department of Natural Resources (DNR) during the fall migration and kept at the Richard Clay Bodine State Fish Hatchery until January-March. The fish were anesthetized in ethyl m-aminobenzoate methanesulfonate, and after the spawn was collected by the DNR, the ventricles were removed from eight fish, immediately rinsed with 4°C Cortland buffer, and transported on ice back to the laboratory.
University of Ottawa, Ottawa, Canada.
Rainbow trout (0.2–0.4 kg) were obtained from Linwood Acres Trout Farm (Campellcroft, Ontario, Canada). Fish were transported to the University of Ottawa Aquatic Care Facility and were maintained in fiberglass holding tanks (1,275 liters) containing well-aerated, dechloraminated City of Ottawa tap water at 13°C. Fish were subjected to constant 12:12-h light-dark cycles and fed five times a week with commercial trout pellets (Martin Mills 5 PT). All experiments were approved by the respective university's Animal Care Committees in accordance with guidelines provided by the Canadian Council on Animal Care.
Effects of H2S on Cardiorespiratory Reflexes
Methods for cannulation of the dorsal aorta have been described in detail (28). Trout were anesthetized in benzocaine (ethyl-p-aminobenzoate; 1:12,000 wt/vol) before surgery. The dorsal aorta was cannulated percutaneously through the roof of the buccal cavity with heat-tapered polyethylene tubing (PE-60). The cannula was filled with heparinized saline (50 mg/100 ml sodium heparin salt in 0.9 g/100 ml NaCl) and connected to a pressure transducer. A 14-gauge needle trocar was inserted in the buccal cavity via the nares, and a 20-cm-long heat-flared polyethylene cannula (PE-180) was attached to the trocar and pulled back out of the cavity, leaving the flared end in the buccal cavity. A snug rubber “O” ring was slipped over the cannula down to the nares securing the cannula. The buccal cannula was filled with aquarium water and attached to a second pressure transducer. The gills were irrigated briefly (1–2 min) with cold (4°C) aquarium water containing 1:24,000 wt/vol benzocaine between each procedure and continuously irrigated with 4°C aerated water containing 1:24,000 wt/vol benzocaine during placement of the flow probe and, in a few fish, cannulation of the ventral aorta, as described below.
The pericardial cavity was exposed with a midline ventral incision, and the ventral aorta was nonocclusively cannulated with slight modifications from previously described methods (29). Briefly, one end of a 5-cm-long piece of 0.51-mm ID silicone tubing (Dow Corning veterinary grade; Konigsberg Instruments, Pasadena, CA) was slipped over the end of a 2-mm-long piece of beveled polyethylene (PE-60) tubing with the bevel end out. The other end of the silicone tubing was attached to 60 cm of PE-60 connected to a pressure transducer. The cannula was filled with heparinized saline. The bulboventricular junction was occluded with a rubber band snare, and the bulbus was pierced between the fork of the ventral coronary arteries with a sharpened 0.5-mm-diameter wire. The beveled end of the cannula was inserted in the bulbus and backed out until the shoulder of the silicone-beveled PE-60 junction rested against the lumen of the bulbus. The silicone cannula was glued to the outer wall of the bulbus with ∼2 μl of cyanoacrylate glue. This procedure prevented the ventral aorta cannula from interfering with the acoustic window of the flow probe. A 3S Transonic flow probe (Transonic Systems, Ithaca, NY) was placed around the ventral aorta, distal to the site of the cannula insertion, and connected to a Transonic T206 flow meter. The ventral incision was closed with interrupted silk sutures. The fish (N = 15 intact and N = 14 with the first or second pair of gills removed, see below) were revived and placed in black plastic tubes immersed in individual 7-liter chambers with aerated, through-flowing well water at 14°C. Experiments were conducted 24 h after cannulation.
In experiments where the first (N = 8) or second (N = 6) pair of gills was removed, the proximal and distal ends of the gill arches were first ligated with heavy cord to prevent bleeding, and the gills were removed with scissors. This procedure does not affect dorsal aortic Po2 (17, 37). The remaining gills were ventilated intermittently during this procedure.
Blood and buccal pressures were measured with disposable pressure transducers (cdexpress; Maxxim Medical, Athens, TX), and the data were collected electronically using Biopac model MP35 (Biopac Systems, Goleta, CA) and archived at 2 Hz on notebook computers. H2S was administered as dissolved Na2S·9H2O through the buccal cannula in a 1-ml bolus of 1, 5, 10, and 25 mmol/l (1, 5, 10, and 25 μmol H2S injected) followed by a 1-ml rinse of aquarium water. The Na2S was dissolved in aquarium water and titrated to ∼7.0 before injection. Heart rate was obtained post hoc from the recordings 5–10 s before H2S administration, immediately after, and again at ∼30 s after H2S administration. Because H2S was administered via the buccal cannula, there was ∼10 s delay in recording respiratory rate and respiratory amplitude after H2S injection. Blood pressures followed changes in heart rate, and, because they did not appear to provide additional information on the H2S response, were not further analyzed in this study.
It is possible that H2S consumed the O2 in the injection solution, and this indirectly produced cardiorespiratory responses. To examine this possibility, aquarium water was deoxygenated by vigorous bubbling with 100% nitrogen, and 1 ml of the deoxygenated water was injected in the buccal chamber (N = 4) following the protocol for H2S administration.
mRNA for two enzymes responsible for H2S synthesis, cystathionine β-synthase (CBS) and cystathionine γ-lyase (CSE), was measured in a variety of trout tissues by RT-PCR. Total RNA from different rainbow trout tissues was isolated using TRIzol Reagent (Molecular Research Center, Cincinnati, OH). Contaminating DNA was removed using the DNA-free kit (Ambion, Austin, TX), and total RNA (2 μl) was reverse-transcribed into cDNA with AMV reverse transcriptase using random hexamer primers according to the manufacturer's protocol (Roche Applied Science). Controls containing no reverse transcriptase were used to check for genomic DNA contamination in each sample. PCRs were performed with a GeneAmp PCR system 2400 (Perkin-Elmer) to detect CBS and CSE gene from the cDNA sample using advantage cDNA polymerase mix (Clontech). The primers of CBS (GenBank accession no. AF266185) were 5′-GCT GTG GAC CTG CTG AAC-3′ (sense) and 5′-CAA ACT GTT GAC TTT AAA TGC TTT C-3′ (antisense). These primers produced a product of 140 bp. The cycling conditions were 94°C for 2 min, followed by 38 cycles of 94°C for 20 s, 63.5°C for 20 s, and 68°C for 30 s. The primers of CSE (GenBank accession no. BC066737) were 5′-TGG CAC CCT TCA CGG AAC AG-3′ (sense) and 5′-ATG GAC AGG CAC AGC GAC TC-3′ (antisense). These primers produced a product of 240 bp. The cycling conditions were 94°C for 2 min followed by 38 cycles of 94°C for 20 s, 62°C for 20 s, and 68°C for 30 s. PCR products were verified by electrophoresis on 2.0% Tris-acetic acid-EDTA agarose gels and visualized under ultraviolet light with ethidium bromide.
H2S Production by Trout Gills, Methylene Blue Method
Trout were killed with a blow to the head, and the branchial basket was immediately perfused via the ventral aorta with cold (4°C) heparinized saline (50 mg/100 ml sodium heparin salt in 0.9 g/100 g NaCl) to remove as much blood as possible. Four replicates were examined; each replica consisted of the first pair of gills and one gill from the second pair. The gills were removed and placed in ice-cold Cortland buffer (pH 7.8) for a maximum of 3 h until homogenization. Immediately before homogenization, the gills were rinsed with cold 100 mM potassium phosphate buffer (pH 7.0), and the gill filaments were cut free of the arch. Gill filaments were then placed in 5 ml of potassium phosphate buffer, minced with scissors, and homogenized on ice until smooth. Homogenates were briefly centrifuged in a microcentrifuge to remove large debris. Sulfide production of the resulting supernatant was measured using the diffusion chamber method from Stipanuk and Beck (43), slightly modified as follows. Aliquots (1.5 ml) of the supernatant were added to three glass diffusion chambers on ice, each chamber containing a 3 × 3 cm piece of Whatman No. 1 filter paper and 0.4 ml 1% zinc acetate suspended over the supernatant by a polyethylene support. Pyridoxal 5′-phosphate (PLP) and l-cysteine (final concentrations 1 and 2 mM, respectively) were added to the first chamber, immediately followed by 750 μl of 50% (vol/vol) TCA, and the chamber was capped. In the second chamber, PLP and l-cysteine were added as above, without TCA, and this chamber was flushed with N2 and capped. In the third chamber, the CBS inhibitor aminooxyacetate (AOA) and the CSE inhibitor propargyl glycine (PPG; see Refs. 14 and 62) were added (final concentration of both 3.3 mM), followed by PLP and l-cysteine. The chamber was then flushed with N2 and sealed. The diffusion chambers were removed from ice and incubated for 2 h at 22°C on a shaker plate. After incubation, 750 μl of 50% TCA were added to the second and third chambers, and all chambers were then placed on a shaker plate at 37°C for an additional hour to trap the H2S on the filter paper. The filter paper and polyethylene support were removed from each chamber and placed in glass cuvettes containing 3.5 ml water. To successive cuvettes, 500 μl of 20 mM N,N-dimethyl-p-phenylenediamine dihydrochloride in 7.2 M HCl were added immediately followed by 400 μl of 30 mM FeCl3 in 1.2 M HCl. Each cuvette was immediately capped and gently shaken. After 10 min, samples were read in a plate reader at 669 nm, and absorbances were compared with a standard curve of Na2S in phosphate buffer run in parallel with the experimental samples. Background H2S from the first experimental chamber was subtracted from the second and third chambers to obtain the rate of H2S production. Protein concentration in the homogenate was measured using the method of Lowry et al. (23).
H2S Production by Trout Gills Measured With the Polarographic H2S Sensor
Four trout were killed by a blow to the head, and the branchial basket was perfused as described above. Gill filaments were then removed from the gill arch, blotted, weighed, and homogenized in 1:3 wt/vol HEPES buffer. Approximately 1 ml of the homogenate was placed in a metabolism chamber (see below), and the chamber was sealed to allow the dissolved O2 to be consumed by gill metabolism. The evolution of H2S gas was then measured with the polarographic sensor under these hypoxic conditions. Cysteine (1 mM) and PLP (1 mM) were added to promote H2S synthesis, and 1 μl of air (∼10 nmol O2) was injected at various intervals to evaluate the effects of O2 on H2S production.
The water-jacketed metabolism chamber consisted of a flat-bottom glass tube ∼1.5 × 3 cm with an acrylic stopper machined to fit tightly and the latter with an injection port and ports for the polarographic H2S sensor and either an O2 sensor or pH electrode as described previously (56).
Effects of Hypoxia on Plasma [H2S], Measurements With the Sulfide Ion Selective Electrode
Trout were anesthetized by immersion in an oxygenated solution of benzocaine (ethyl-p-aminobenzoate; 1: 10,000, wt/vol) and placed on a surgical table that allowed irrigation of the gills with the same anesthetic solution. To permit serial blood removal, a polyethylene cannula (PE-50; Clay Adams) was implanted in the dorsal aorta via percutaneous puncture of the roof of the buccal cavity (42). Fish were allowed to recover for 24 h before experimentation.
Hypoxia (Po2 130–38 mmHg) was achieved by replacing the air supplying a water/gas equilibration column with N2-air mixtures. Different groups of fish (N = 5–6 fish/group) were used for each level of hypoxia, i.e., each fish experienced normoxia followed by only one level of hypoxia. The desired water Po2 exiting the column was preset and established by adjusting the rate of water and/or N2-air flow through the column. Generally, the desired water Po2 in the experimental box was reached within 10 min and thereafter never varied more than ±2 mmHg. The N2-air mixes were provided by a gas-mixing flowmeter (GF-3/MP; Cameron Instruments, Port Aransas, TX). Water Po2 was monitored continuously by pumping (using a peristaltic pump) water through a temperature-controlled (13°C) chamber housing a Cameron O2 electrode that was connected to blood gas analyzer (Cameron Instruments). All electrodes were calibrated before each individual experiment. The O2 electrode was calibrated by pumping (using a peristaltic pump) a zero solution (2% sodium sulfite) or air-saturated water continuously through the electrode sample compartment until stable readings were recorded. Blood samples (0.5 ml) were withdrawn before (normoxic control) and after 30 min of acute hypoxia. After centrifugation, the plasma was added to antioxidant buffer (0.8 M sodium salicylate, 1.1 M NaOH, 0.2 M ascorbate), and H2S was measured as total S2− with the ion selective electrode (ISE) following the manufacturer's directions (Lazar Research Laboratories, Los Angeles, CA).
After these experiments were completed, it became apparent to us that the ISE may not accurately measure free sulfide in the plasma and, in fact when measured with a polarographic H2S sensor (see below), sulfide appears to be rapidly consumed by blood (56). To determine if blood spiked with exogenous sulfide could be detected with the ISE, whole blood was drawn from trout and incubated at 14°C for 30 min with NaHS sufficient to produce a theoretical concentration of 50, 100, and 250 μmol/l. Samples were then centrifuged, and the plasma was analyzed as above.
Effects of Hypoxia on Plasma [H2S], Measurements With the Polarographic H2S Sensor
Six trout were anesthetized in benzocaine (1:24,000, wt/vol). The dorsal aorta was cannulated percutaneously through the roof of the mouth with polyethylene tubing (PE-60), and the caudal vein was cannulated with PE-50 via a 1-cm incision on the lateral body wall near the caudal peduncle (31). The fish were revived and placed in a black tube inserted in a 10-liter chamber with aerated, through-flowing water at 14°C. The two cannulas were connected to a peristaltic pump that pumped blood at 1.5 ml/min from the dorsal aorta across a thermojacketed polarographic H2S sensor (56) and returned it to the fish via the caudal vein. This extracorporeal loop enabled continuous sampling of plasma sulfide in unanesthetized fish. The fish was made hypoxic by temporarily shutting off the inflowing water and bubbling the chamber with 100% N2 until the Po2 fell to 40 mmHg (ambient Po2 was measured with a polarographic O2 sensor; model 55; YSI, Yellow Springs, OH). The N2 was then turned off, and the Po2 remained at this level for ∼15 min after which the water was turned back on and the chamber was aerated with room air. The H2S sensor was calibrated in position by substituting a vial containing an volume of HEPES buffer (14°C) equivalent to the blood volume of the fish in the extracorporeal loop and serial additions of sulfide as Na2S. Because the polarographic H2S sensor measures only H2S gas, total sulfide was determined from the Henderson-Hasselbalch equation using a pKa of 6.9 and pH of 7.8 (56). The sensor was connected to an Apollo 4000 Free Radical Analyzer (WPI, Sarasota, FL) with 100 mV polarizing voltage, and data were archived on a personal computer (PC). The detection limit of the sensor was 14 nM H2S gas (∼100 nM total sulfide).
Effects of Inhibitors of H2S Synthesis on Cardiorespiratory Responses of Trout to Hypoxia
The dorsal aorta and buccal chamber were cannulated as described above with the exception that, per an anonymous reviewer's suggestion, the buccal cannula was inserted through the bony snout rather than through the nares to minimize irritation to the fish. Hypoxia (Po2 ∼40 mmHg) was adjusted and monitored as described above. Two protocols (hypoxia-hypoxia and hypoxia-inhibitor-hypoxia) were employed. In the first, cardiorespiratory parameters were monitored while trout (N = 6) were subjected to normoxia and 15 min of hypoxia, and then, following a 3-h recovery period, the hypoxia was repeated. These experiments allowed us to verify that there was no difference between the first and second hypoxic exposures. In the second group of experiments, at the end of the 3-h recovery period, AOA, an inhibitor of CBS, and PPG, an inhibitor of CSE (aka CGL) were added to the water (5 mg/l AOA and 10 mg/l PPG) and injected in the dorsal aorta (5 mg/kg body wt AOA and 10 mg/kg body wt PPG). This produced an initial increase in dorsal aortic pressure, a decrease, and then an increase in cardiac output. Because these variables appeared to subside within 20–30 min, in one group of experiments (N = 6), we waited 30 min after the injection of inhibitors to expose the trout to the second bout of hypoxia. In the second group of experiments (N = 6), we waited 90 min between inhibitor injection and the second hypoxia. Cardiorespiratory parameters were recorded and archived as above.
Effects of Hypoxia and H2S on Zebrafish NEC Em
Adult zebrafish (Danio rerio) were obtained from a commercial supplier (MIRDO, Montreal, Quebec, Canada) and transported to the University of Ottawa Aquatic Care Facility where they were maintained in acrylic tanks (4 l) supplied with aerated, dechloraminated City of Ottawa tap water at 28°C. Fish were maintained on a constant 10:14-h light-dark photoperiod. All procedures for animal use were approved and carried out according to institutional guidelines and in accordance with those of the Canadian Council on Animal Care.
Zebrafish (N = 10 for each cell isolation procedure; 3 isolations were performed) were decapitated after a sharp blow to the head. All procedures for isolation of NECs were carried out under sterile conditions in a laminar flow hood (16). Gill baskets were removed and rinsed in a wash solution (2% penicillin/streptomycin in PBS) for 10 min. All four gill arches were then separated, and distal filaments rich in NECs (16) were selectively removed. Tissue was placed in 0.01% hyaluronidase for 10 min and then 0.25% trypsin/EDTA (Invitrogen) for 1 h at room temperature, minced with fine forceps, and triturated in a 15-ml centrifuge tube with a Pasteur pipette. The trypsin reaction was stopped with the addition of 10% FCS. Lower concentrations of trypsin or mechanical separation alone did not work well for dissociation. The cell suspension was centrifuged (140 g) for 5 min, and the pellet was triturated in PBS. Gill cells were centrifuged once more with PBS and suspended in Leibovitz's (L-15, with l-glutamine but without phenol red) culture medium supplemented with 1% penicillin/streptomycin and 5% FCS. Later on, cells were plated in sterile cellbind. NECs that adhered to the culture substrate were identified using 2 mg/ml Neutral Red (Sigma), a vital marker used to identify 5-hydroxytryptamine-containing cells (60).
The effects of hypoxia and H2S on NEC Em were measured by whole cell recordings (12) in vitro using zebrafish gill cell culture preparations essentially as described previously (16). Specialized 35-mm culture dishes with sterile cellbind (Corning) were plated with dissociated NECs. After 24 and 48 h, dishes were mounted on a fixed stage of a Zeiss inverted microscope, and a perfusion insert (Warner Instruments) with inlet and outlet was placed on it. Patch electrodes were fabricated from borosilicate glass with filament (1.5 mm OD; 0.86 mm ID; 10 cm length; Sutter Instruments, Novato, CAA). Glass electrodes were pulled on a horizontal pipette puller (P-2000; Sutter Instrument) and filled with pipette (intracellular recording) solution containing (in mM): 135 KCl, 5 NaCl, 0.1 CaCl2, 11 EGTA, 10 HEPES, and 2 MgATP; pH was adjusted to 7.2 with KOH; resulting tip resistances were typically 5–7 MΩ. Bath (extracellular) solution contained (in mM): 135 NaCl, 5 KCl, 2 CaCl2, 2 MgCl2, 10 glucose, and 10 HEPES; pH was adjusted to 7.4 with NaOH. Seal resistance was typically >5 GΩ, and holding current was <5 pA at −60 mV.
Current-clamp protocols were performed using a MultiClamp 700B amplifier (Axon Instruments) interfaced with an IBM PC, data acquisition system (DigiData 1322A; Molecular Devices) and Clampex 9.2 software (Molecular Devices). All recordings were filtered at 5 kHz and digitized at 10 kHz. Data were collected from 17 cells exhibiting stable membrane potentials (∼75% of patched cells). All data were analyzed using Clampfit 9.0 software (Molecular Devices).
The recording chamber was continuously perfused (4 ml/min) using a constant-pressure head with bath solution at room temperature (22–24°C) through a four-channel remote valve control system (ALA Scientific Instruments). Hypoxia was produced by bubbling N2 through standard extracellular recording solution in the perfusion reservoir (60-ml syringe) for at least 10 min until the desired Po2 (20 mmHg) was reached. Measurements of O2 were made using a fiber optic O2 electrode (model Foxy AL300; Ocean Optics, Dunedin, FL) and associated hardware and software (Ocean Optics SD 2000). Different levels of H2S in the form of NaHS (50 and 100 μmol/l final concentration) were added directly to the extracellular recording solution in the perfusion reservoir (60-ml syringe). Gas-impermeable tubing (Tygon; Saint-Gobain Performance Plastics, Akron, OH) was used to transfer the perfusate to the recording chamber. Control experiments were performed in which the perfusate was bubbled with compressed air.
H2S Consumption by Trout Heart Mitochondria Measured With the Polarographic H2S Sensor
Mitochondria were isolated from eight steelhead ventricles as described by West and Driedzic (55). Ventricles (∼3 g) were pooled from two fish and provided sufficient mitochondria for protein determination and a normoxic, hypoxic, and heat-killed experiment. Four replicates were examined. Briefly, the ventricles were minced and homogenized in 9 vol of ice-cold isolation medium using a cold ground-glass homogenizer. The crude homogenate was centrifuged at 600 g (10 min), and the resulting supernatant was centrifuged at 9,000 g (10 min). The mitochondrial pellet was resuspended in 2 ml of isolation medium and centrifuged at 9,000 g (10 min); this step was repeated a second time. The final pellet was resuspended in 0.3–0.6 ml of isolation medium. All steps were at 4°C. Isolation medium containing BSA (fraction V; Sigma) was added to the remaining suspension (final concentration 1 mg/l). Mitochondria were kept on ice until used and tested for viability (respiratory control ratio, RCR, >5; see Ref. 55). Mitochondria protein was determined by Lowry protein assay.
H2S consumption was measured in the metabolism chamber with the polarographic H2S sensor in RCR buffer. After the sensor baseline stabilized, the buffer was spiked to 5 μM with Na2S, and mitochondria were added after the sensor reading had peaked and began to fall. Buffer was made hypoxic by incubation under nitrogen in a separate flask for at least 1 h before experimentation (the presence of BSA in the RCR buffer precluded bubbling). Mitochondria were heat-killed by placing them in hot (50°C) water for 10 min. The rate of mitochondrial H2S consumption was corrected for the rate of H2S disappearance before addition of mitochondria.
The composition of Cortland buffer was (in mM): 124 NaCl, 3 KCl, 2 CaCl2·H2O, 1.1 MgSO4·H2O, 0.09 NaH2PO4, 1.8 Na2HPO4, NaHCO3, and 5.5 glucose, pH 7.8. The mitochondria isolation medium was (in mM): 5 KH2PO4, 5 K2HPO4, 2 EGTA, and 250 sucrose, pH 7.4, and the mitochondrial RCR was (in mM): 12.5 KH2PO4, 12.5 K2HPO4, 10 Tris, 100 KCl, and 2.7 BSA. H2S was prepared by dissolving Na2S·9H2O in aquarium water (intrabuccal injection) or appropriate buffer (in vitro experiments). The aquarium water with Na2S was titrated to pH ∼7.0 with HCl shortly before experimentation. Na2S·9H2O was purchased from Fisher (Chicago, IL); other chemicals were purchased from Sigma-Aldrich (St. Louis, MO).
Comparisons were made with Student's t-test, paired t-test, or one-way ANOVA repeated measures followed by the Holm-Sidak test for multiple comparisons, or linear regression. Wilcoxon Signed Rank test was used when normality failed. Significance was assumed at P ≤ 0.05. Values were expressed as means ± SE or mean + SE in many figures. Data acquired under conditions of current or voltage clamp are presented as means ± 1 SE. Differences between treatments were analyzed using one-way ANOVA repeated measures followed by Bonferroni's t-test for multiple comparisons. Statistical analyses data were performed using commercial software (SigmaStat version 3.1; SPSS).
Effects of H2S on Cardiorespiratory Reflexes
H2S was injected in the buccal cavity of intact, unanesthetized trout to determine if the cardiorespiratory responses were similar to those produced by hypoxia. Intrabuccal injection of H2S produced bradycardia and increased respiratory rate and amplitude in trout (Figs. 1–4). One milliliter of 1 mmol/l H2S produced a slight, but significant, bradycardia that became more pronounced as [H2S] increased. At and above 10 mmol/l H2S, there was often complete cardiac arrest for five or more seconds, yet heart rate recovered 30 s later. The bradycardia was accompanied by transient decreases in ventral and dorsal aortic blood pressures. Blood pressures typically returned to normal before heart rate recovered and well before recovery of respiratory variables. This suggests that venous return was maintained. Respiration rate and amplitude were significantly increased by ≥5 mmol/l H2S immediately after injection, although the changes in rate were modest compared with the apparent dose-dependent increases in amplitude. The effects of H2S on respiratory variables often remained elevated at 30 s following H2S exposure, whereas heart rate had returned to normal by this time (data not shown). Injection of 1 ml of deoxygenated aquarium water in the buccal chamber did not significantly affect any of the cardiorespiratory variables (data not shown).
Removal of the first pair of gills, the site of much of the external O2-sensitive chemoreceptor activity in trout, significantly attenuated the H2S bradycardia at 5 and 10 mmol/l, but at 25 mmol/l the response was not different from intact trout (Fig. 2). Removal of the second pair of gills had no significant effect on the H2S-mediated bradycardia, and these fish were indistinguishable from intact fish (Fig. 2). Respiration rate and amplitude were not affected by removal of either the first or second pair of gills (Figs. 3 and 4, respectively).
H2S Enzymes in Trout Tissue and H2S Production by Trout Gills
RT-PCR was used to examine a variety of tissues for two enzymes responsible for H2S synthesis, CBS and CSE. Evidence for mRNA of both enzymes was found in all tissues examined, including gill, liver, brain, heart, skeletal muscle, two arteries, and a vein (Fig. 5).
The rate of H2S production by the first pair of trout gills was measured in homogenized filaments in the presence of the H2S precursor l-cysteine and the cofactor PLP using the methylene blue gas diffusion method. Background (preexisting) [H2S] was 173 ± 23 pmol/mg protein. After correction for background H2S, gill filaments produced ∼120 pmol/mg protein in 120 min (Fig. 6A). Addition of the CBS inhibitor AOA and the CSE inhibitor PPG completely inhibited active H2S production (Fig. 6A).
Real-time sulfide production by homogenized gill filaments measured with the polarographic H2S sensor is shown in Fig. 6B. During hypoxia, and in the presence of 1 mM cysteine and 1 mM PLP, homogenized gill tissue continuously produced H2S (expressed as total sulfide). Addition of 1 μl of room air (∼10 nmol O2 or 10 μmol/l if completely mixed in the metabolism chamber) immediately but transiently decreased H2S production. A second injection of air ∼75 min after the first produced similar results. Increasing cysteine to 2 mmol/l greatly increased the rate of sulfide production, and this was also partially inhibited by injection of air, albeit to a lesser extent.
Effects of Hypoxia on Plasma Sulfide Measured With the Ion Selective Sulfide Electrode
To determine the effects of hypoxia on plasma sulfide, blood was taken from normoxic trout and again after exposure to a single level of hypoxia. The average sulfide concentration of all control samples was 391 ± 70 μmol/l (N = 38). Mild hypoxia (Po2 = 130 mmHg) did not affect plasma sulfide concentration, whereas a progressive decrease in Po2 (112, 93, 68, 53, 38 mmHg) produced a significant and apparent dose-dependent (r2 = 0.925, P < 0.001) increase in plasma sulfide (Fig. 7A).
Sulfide concentration in plasma taken from whole trout blood spiked with 50, 100, and 250 μmol/l NaHS was ∼0, 5, and 88 μmol/l. Sulfide concentration in a buffer sample run in parallel with the blood and spiked with 250 μmol/l NaHS was 128 μmol/l.
Effects of Hypoxia on Blood Sulfide, Measured In Vivo With the Polarographic H2S Sensor
H2S gas was not detected in blood from normoxic or hypoxic (Po2 ∼40 mmHg) trout fitted with the extracorporeal pump (Fig. 7B). When a reservoir of HEPES buffer with a volume equivalent to the blood volume of the trout was substituted for the fish, a standard [H2S] response was evident, and the H2S readings converted to total sulfide (based on pH and temperature; see Ref. 56) remained stable (Fig. 7C).
Effects of Inhibitors of H2S Synthesis on Cardiorespiratory Responses of Trout to Hypoxia
The effects of ambient hypoxia (Po2 ∼40 mmHg) on cardiorespiratory parameters in control and inhibitor-treated trout are shown in Fig. 8. Hypoxia did not affect dorsal aortic pressure, whereas it significantly decreased heart rate and increased respiration rate and respiration amplitude. There were no significant differences between the first and second hypoxic exposures in control trout or between the first and second hypoxic exposures in trout treated 30 min (Fig. 8) or 90 min (data not shown) before the second hypoxic exposure by addition of inhibitors of CBS (AOA) and CSE (PPG) to the water and simultaneous injection of these inhibitors in the dorsal aorta.
Effects of Hypoxia and H2S on Em of Zebrafish NECs
The effects of hypoxia and H2S on Em of cultured zebrafish NECs, the putative chemoreceptive cells in fish, were measured by patch-clamp electrophysiology under conditions of current clamp. Hypoxia and 50 μmol/l H2S significantly depolarized zebrafish NECs by ∼10 mV (Fig. 9). H2S (100 μmol/l) produced ∼20 mV depolarization, but this was not significantly different from that produced by hypoxia or 50 μmol/l H2S.
H2S Consumption by Trout Heart Mitochondria Measured With the Polarographic H2S Sensor
O2-dependent consumption of H2S was measured in purified trout heart mitochondria using the polarographic H2S sensor (Fig. 10). Addition of mitochondria to buffer containing 5 μM Na2S initiated a steady decrease in [H2S] indicative of net H2S consumption. The rate of H2S consumption was halved when mitochondria were made hypoxic, or heat-killed.
These experiments were designed to examine the hypothesis that the metabolism of H2S is involved in the O2-sensing signal transduction cascade in trout chemoreceptors. We found that 1) intrabuccal administration of H2S to unanesthetized trout produced bradycardia and increased respiratory rate and amplitude similar to that produced by hypoxia, 2) removal of the first pair of gills substantially attenuated the H2S effect on heart rate, consistent with the predominant location of the externally oriented chemoreceptive cells on the first pair of gills in these fish, 3) gills as well as a variety of other organs and blood vessels appear to possess CBS and CSE and are thus capable of H2S production, 4) H2S was enzymatically generated from cysteine by the first pair of gills, and H2S production was inhibited by O2; 5) although hypoxia did not increase total free sulfide (H2S and HS−) in the plasma, it appeared that sulfide, perhaps from a bound sulfur moiety, did increase, 6) zebrafish NECs, the putative chemoreceptive cells in fish, were depolarized by both hypoxia and H2S. Furthermore, H2S consumption by purified heart mitochondria was O2-dependent. Collectively, these results are consistent with hypoxia and H2S acting through a similar, if not common, pathway in fish O2 chemoreceptors, and they suggest that the metabolism of H2S serves as an O2 sensor.
O2-sensitive chemoreceptors in fish are primarily located on the gill filaments (4, 9). In most fish, there are two populations of chemoreceptors. Externally oriented receptors monitor water Po2 and, when stimulated by ambient hypoxia or cyanide, initiate a bradycardia and an increase in ventilation rate and amplitude (4, 37). Internal chemoreceptors are closely associated with branchial vessels, and they appear more suited to monitor blood Po2. When stimulated, the internal chemoreceptors elicit respiratory reflexes, but they generally have little effect on heart rate. In trout, but not all fish, the externally oriented receptors are primarily, but not exclusively, located on the first pair of gills, whereas the internally oriented receptors tend to be distributed throughout all four pairs of gills. Chemoreceptors on filaments of the first pair of gills are homologous to chemoreceptors of the mammalian carotid body (26) and exhibit a similar sensitivity to Po2 (4). Extirpation of the first pair of gills greatly reduces the hypoxia- and cyanide-mediated bradycardia but does prevent the increase in ventilatory rate or amplitude (37).
Our study confirmed previously reported cardiorespiratory responses to hypoxia (4, 9, 26, 37) in that we also observed that a reduction of water Po2 to 40 mmHg produced a bradycardia and an increase in respiratory frequency and amplitude (Fig. 8). Injecting H2S in the buccal cavity produced similar responses (Figs. 1–4). However, the effect of H2S on the bradycardia appeared to be more pronounced than its effect on respiratory variables, since a significant bradycardia was evident at 1 μmol/l H2S (Fig. 2), whereas respiratory rate and amplitude did not significantly change until 5 μmol/l H2S was injected (Figs. 3 and 4). Even then, the effects of H2S on respiration were less pronounced. For instance, hypoxia decreased heart rate by ∼45% and increased respiratory rate by 15% and respiratory amplitude by 140%, whereas 5 μmol/l H2S decreased heart rate by 65% and increased respiratory rate and amplitude by 11 and 33%, respectively. At 10 μmol/l H2S, this discrepancy is even more evident; heart rate decreased by 80%, whereas respiratory frequency and amplitude only increased by 17 and 38%, respectively. It is possible that the difference between hypoxia and H2S is due to the fact that both the external and internal NECs would be exposed to hypoxia, whereas the single bolus of H2S would be rapidly swept away in the respiratory current, and thus the concentration of H2S reaching the surface NECs would be higher than the [H2S] reaching the internal, vascularly oriented, NECs. This would be exacerbated by O2-dependent metabolism of H2S, which is likely to occur as the H2S diffuses through the tissues toward the internal NECs. Although not analyzed in great detail, the hypotension accompanying H2S-mediated bradycardia was relatively transient, and it did not appear to cause the observed changes in respiratory rate or amplitude.
Additional evidence for the prominence of externally oriented NECs on the first pair of gills was observed following removal of these arches, since extirpation of these gill arches profoundly affected the bradycardia response to H2S but had no effect on either respiratory variable (Figs. 2–4). Furthermore, removal of the second pair of gills did not significantly affect any of the cardiorespiratory responses to H2S. It was also evident that the higher concentrations of H2S produced bradycardia in trout in which the first pair of gills were removed. This may be due to activation of less sensitive receptors on other gill arches or higher concentrations of H2S necessary to activate the considerably smaller population of chemoreceptors remaining. Although we could not accurately time the interval between H2S injection and the onset of respiratory responses, the cardiac effects of H2S appeared to recover faster than the respiratory responses, usually within 30 s. This delay would be consistent with two populations of NECs. The cardiorespiratory responses to H2S are also likely to be due to the direct effect of H2S and not due to a lack of O2 in the 1 ml H2S bolus because injection of 1 ml of water deoxygenated with nitrogen did not produce any cardiorespiratory response.
We were not able to directly measure H2S production by NECs. However, most trout tissue appears to have the potential for H2S production via both CBS and CSE pathways (Fig. 5), and H2S was readily generated from cysteine by homogenized gill tissue. Furthermore, H2S production by homogenized gill tissue could be blocked by inhibitors of CBS and CSE (Fig. 6A), supporting the hypothesis that H2S is enzymatically generated. We observed that 120 pmol/mg protein of H2S was produced in 2 h. If the rate of H2S production can be assumed to be linear, then gills produced 1 pmol·mg protein−1·min−1, a value similar to that reported for a variety of rat tissues by Zhao et al. (62, 63), when corrected for tissue protein, which is assumed to be 10–20% of wet weight. Thus H2S production may be a general feature of many cells.
When H2S production by homogenized gill tissue was monitored in real time using the polarographic H2S sensor, it was evident that homogenized gills produced H2S continuously and at a relatively constant rate (Fig. 6B). Furthermore, H2S production was transiently reversed to an apparent H2S consumption by addition of O2. Increasing the level of cysteine increased H2S production, and the inhibitory effect of O2, although still present, was reduced. This suggests that the concentration of H2S in tissues is inversely related to the presence of O2, and, if H2S is constitutively generated in tissues as we proposed (30), it provides a mechanism for inversely coupling tissue Po2 to [H2S]. We have recently observed a similar O2-dependent inhibition of H2S production in heart myocardium (56) that appears to explain the activation process in H2S-mediated ischemic preconditioning in the heart (32, 41), and we have also observed O2-dependent inhibition of H2S production in homogenized bovine lung (N. L. Whitfield and K. R. Olson, unpublished observation) that can explain the excitation step in hypoxic pulmonary vasoconstriction (30). Thus the inverse coupling of H2S production to Po2 appears to be a general feature of O2-“sensing” tissues.
We (30) also proposed that H2S is constitutively generated in the cytoplasm and that the mitochondria are the site of H2S oxidation, which is consistent with the quantitative O2 consumption by rat mitochondria during sulfide metabolism recently demonstrated by Hildebrandt and Grieshaber (13) and the proposed origin of the mitochondria from sulfide-oxidizing thermophilic bacteria (40). As shown in Fig. 10, trout heart mitochondria readily consumed H2S when in normoxic buffer. When the buffer was deoxygenated, the rate of H2S consumption was halved. This provides the O2-dependent sink for H2S generated in the cytoplasm and may also explain the overall coupling of Po2 to intracellular [H2S]. We purified mitochondria from heart tissue to maximize our yield, but it seems reasonable to assume that gill mitochondria would behave similarly.
The exact site of gill H2S production is unknown because it was not possible to directly measure H2S production by NECs. However, because H2S readily diffuses across cell membranes, neighboring, non-NECs, may also contribute to the O2-sensing signal transduction mechanism in a paracrine fashion. Nevertheless, as discussed below, it is now doubtful that H2S serves as a physiologically relevant signaling molecule in the circulation.
We (56) recently employed a polarographic sensor to measure blood H2S in real time without any sample modification and we found that, contrary to all studies published within the last eight years using other methods, H2S gas or calculated total free sulfide (H2S, HS−, and S2−) was below levels of detection (14 nM H2S gas; 100 nM total sulfide). Furthermore, we measured H2S in unanesthetized trout fitted with an extracorporeal loop that pumped blood from the ventral aorta across the polarographic sensor and into the caudal vein and found that hypoxia did not increase blood H2S to detectable levels and that sulfide injected in the circulation was rapidly removed (56). In the present study, we moved the sampling site of the extracorporeal loop from the ventral aorta to the dorsal aorta in an attempt to determine if hypoxia could increase branchial H2S production enough to increase plasma H2S to detectable levels. Consistent with our previous findings (56), we failed to observe any measurable increase in H2S during hypoxia (Fig. 7B). Thus it appears that the Po2-dependent increase in plasma sulfide measured with the sulfide electrode (Fig. 7A) is not due to an increase in total free sulfide.
Using the ISE, however, we observed an increase in plasma “sulfide” in unanesthetized trout when exposed to progressively lower ambient Po2 (Fig. 7A). If H2S or free sulfide does not circulate in the plasma, as our studies with the polarographic sensor show (Fig. 7B), and it is rapidly removed in the presence of red blood cells (56), then the ISE must have measured sulfide in some form other than free sulfide. When we spiked whole blood with H2S and then measured sulfide in plasma with the ISE, our apparent recovery rates were ∼0 for a 50 μmol/l spike and only 35% for a 88 μmol/l spike, indicative of sulfide binding or metabolism. Warenycia et al. (51) reported that dithiothreitol (DTT) liberated sulfide from a bound form of non-acid-labile sulfide in brain tissue of H2S-poisoned animals. We did not examine this possibility in the present experiments; however, we were unable to identify any sulfide liberation from trout plasma with DTT in a previous study (56). We (56) did, however, observe that the antioxidant buffer routinely used with the ISE appeared to liberate sulfide from plasma and even bovine albumin, albeit at a relatively slower rate. It remains to be determined if the hypoxia-induced increase in plasma sulfide measured with the ion selective sulfide electrode is due to sulfide bound to plasma proteins or carried in the plasma in some other form.
We did not observe any effect of CBS and CSE inhibitors on cardiorespiratory responses of trout to hypoxia either 30 (Fig. 8) or 90 (data not shown) min after injection of the inhibitors. It was evident that the inhibitors were affecting some tissues because of the immediate increase in dorsal aorta pressure and decrease then increase in cardiac output following their administration, but it is unclear where or how this was achieved. Both inhibitors are notoriously difficult to use with intact tissue, since they often fail to penetrate cells (44), which is why most studies on their inhibitory effect (including the present study) use homogenized tissue.
The problems associated with identifying the action of inhibitors of H2S synthesis in vivo are confounded by the technical difficulty in accurately measuring plasma H2S. For example, Zhang et al. (61) used the indirect microdiffusion method and found that plasma [H2S] was 294 μmol/l in normoxic rats, falling to 196 μmol/l when the rats were chronically exposed to hypoxia (6 h of 10% O2 daily for 3 wk). When 30 mg/kg PPG was administered daily before the hypoxic challenge, plasma [H2S] fell even lower to 141 μmol/l. These results suggest that hypoxia lowers plasma H2S (opposite to our findings) and that PPG is an effective inhibitor of H2S biosynthesis in vivo. However, as we recently reported (56), these indirect methods artificially inflate plasma [H2S], and, in fact, it is unlikely that H2S exists in the circulation as a physiologically relevant signaling molecule. Thus it is unclear what actually was measured in the plasma in the study by Zhang et al. (61) or if PPG did in fact inhibit endogenous H2S biosynthesis. Clearly, a resolution of this issue awaits the next generation of enzyme inhibitors and detailed studies on intracellular H2S production and plasma measurements using the polarographic H2S sensor.
NEC are the putative O2-sensing cells in teleosts (reviewed in Ref. 9). NEC in zebrafish are essentially identical to those of trout (39) and have been a good model with which to show an O2-sensitive depolarization (16). Here we confirm the hypoxic depolarization of zebrafish NEC and show that a similar degree of depolarization is achieved by micromolar H2S (Fig. 9). We used concentrations of H2S commonly reported in the literature in other physiological studies. Clearly, we have shown that this concentration is not present in the blood. Whether or not it reflects a physiological level of intracellular H2S remains to be determined. We also recognize that a variety of stimuli, unrelated to hypoxia, may also depolarize these cells. Nevertheless, H2S depolarization of NECs is consistent with our hypothesis that hypoxia and H2S act through a common activation pathway.
Perspectives: A Unifying Model of O2 Sensing
The present findings are consistent with many of our observations on H2S and O2 sensing in vascular (30) and nonvascular (6) smooth muscle and in ischemic preconditioning of the myocardium (56). Collectively, they support a novel, unifying model of O2 sensing based on the metabolism of H2S, as follows. During normoxia, the constitutive production of H2S by cells is offset by mitochondrial oxidation; thus, cellular or tissue H2S is maintained at low levels. However, as tissue O2 falls, H2S oxidation also declines, and as cellular H2S increases it exerts its physiological effects. Thus, in this simple model, a physiological signal is titrated by O2 availability. This hypothesis is supported by recent observations of stoichiometric O2 consumption by rat mitochondria during sulfide consumption (13), and it also has some anecdotal support in that one of the proposed origins of the mitochondria is in sulfur cycling between the sulfur-reducing (H2S-generating) cytoplasm of primitive Archea and sulfur (H2S)-oxidizing bacteria that became the mitochondria (40).
This work was supported in part by National Science Foundation Grant IOS 0641436.
We thank Y. Gao for technical assistance and R. Bell, D. Meunick, and the staff at the Richard Clay Bodine State Fish Hatchery, Indiana Department of Natural Resources for help in obtaining steelhead tissues.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2008 the American Physiological Society