Hydrogen sulfide and oxygen sensing: implications in cardiorespiratory control

Hypoxia of external or internal origin is a potential and recurring threat to all cells. Vertebrates have developed complex and specialized oxygen-sensing tissues to monitor environmental hypoxia, blood hypoxia(hypoxemia) and even local tissue hypoxia. Numerous theories have been advanced to explain how these cells monitor oxygen (O₂) levels and how low oxygen signals are transduced into physiological responses. However,no single theory has emerged to date to unequivocally explain the O₂-sensing signal transduction process in any of these tissues. Hydrogen sulfide (H₂S) has recently emerged as a signaling`gasotransmitter' of considerable biological potential and we have proposed that O<sub>2</sub>-dependent metabolism of H<sub>2</sub>S may be the`O₂ sensor' in both vascular and non-vascular smooth muscle as well as putative chemoreceptor cells, the neuroepithelial cells. This commentary identifies current theories and concepts of O₂ sensing during acute hypoxia in fish and mammals, vertebrates at opposite poles of oxygen availability and demand, and provides evidence for H₂S metabolism in the O₂-sensing mechanism. Some of the remaining questions and problems associated with this hypothesis, and of H₂S as a signaling molecule in general, are also considered.

Tissue hypoxia can have an external or internal origin. External (ambient)hypoxia occurs in aquatic environments seasonally, diurnally or episodically,and is a common feature of both freshwater and saltwater environments(Bickler and Buck, 2007). Compared to air, aquatic hypoxia is especially problematic because water has a lower capacity for O₂ (1/30), diffusion through water is slower(Krogh's diffusion coefficients 1/200 000) and O₂ solubility decreases by around 20% for every 10°C increase in temperature and another 20% in seawater compared to freshwater. In addition, the high viscosity of water (60 times that of air) impedes breathing, and even mixing of O₂. Life on land has problems of its own. Terrestrial animals may be exposed to hypoxia when they ascend to higher altitudes due to the fall in barometric pressure (at 3000 m P_O2 is 30% lower than at sea level) or enter burrows where there may be little convective delivery of fresh air (Ramirez et al.,2007). Even the mammalian fetus experiences a form of external hypoxia. Global internal hypoxia can originate from a decrease in the ventilation/perfusion ratio in the respiratory organs resulting in a decrease in O₂ uptake, or it can be more localized following an increase in the metabolism/perfusion ratio in systemic tissues. Low O₂ content in blood is called hypoxemia and if it produces tissue hypoxia this is hypoxemic hypoxia. Ischemia (or ischemic hypoxia) refers to tissue hypoxia resulting from insufficient blood flow.

In order to adequately and rapidly respond to potential shortfalls in oxygen, animals have developed O₂-sensing tissues capable of detecting hypoxia in the environment, in blood, or within tissues. Notable among these are chemoreceptors that initiate central cardio-respiratory responses, chromaffin cells in the adrenal medulla or central veins that provide a humoral source of catecholamines and blood vessels that exert local control over tissue perfusion (Lahiri et al., 2006; Milsom and Burleson, 2007; Reid and Perry, 2003; Weir et al.,2005). While all cells are affected by hypoxia, these`O₂-sensing' cells appear exquisitely sensitive to subtle degrees of hypoxia and thereby serve important homeostatic functions during the initial (acute) hypoxic challenge. Long-term (chronic) effectors of hypoxic responses that regulate genetic responses, such as the hypoxia inducible factor (HIF) transcription factor family, sustain and augment the initial responses but are not considered here.

Fish and other aquatic breathing vertebrates such as most larval and a few adult amphibians (e.g. tadpoles, frogs and salamanders and caecilians) are for reasons described above the most susceptible to environmental hypoxia. Individual neuroepithelial cells (NECs) are scattered over the gill surface,but are especially prevalent on the leading (inhalant) and distal edge of the gill filament. In hypoxia-sensitive fish such as trout, they initiate a reflex bradycardia and increase ventilatory rate and amplitude when water P_O2 falls to 75–100 mmHg(Reid and Perry, 2003) (for a review, see Gilmour and Perry,2007). In some fish (e.g. trout) these NECs are predominantly on the first gill arch, the homolog of the mammalian carotid body; in others(bowfin, catfish) they are more evenly distributed over the gills and even the orobranchial cavity (Gilmour and Perry,2007; Milsom and Burleson,2007). These externally oriented receptors are absent in the obligate air-breathing South American lungfish, Lepidosiren paradoxa,and it has been suggested that this mode of O₂ sensing was lost early in vertebrate evolution (Sanchez et al., 2001).

Cells similar to NECs are grouped in clusters called neuroepithelial bodies(NEBs) that line the airways of mammalian lungs especially near bifurcations(Peers and Kemp, 2001; Kemp et al., 2002). NEBs are most prominent in neonatal lungs and are believed to monitor airway P_O2 and help initiate breathing and thereafter optimize ventilation/perfusion matching through central reflexes and by paracrine stimulation of adjacent pulmonary vasculature(Kemp et al., 2002). Although they are more responsive to airway hypoxia than hypoxemia, their function as general monitors of environmental P_O2 in adult animals is unclear.

Monitors of blood O₂ status such as vascularly oriented NECs in fish gills and carotid bodies in tetrapods provide general information on O₂ delivery to systemic tissues and, not unexpectedly, are predominantly associated with the arterial circulation. In addition,O₂-sensitive chromaffin cells are variously found in heart and vasculature of the more ancient vertebrates and condensed in the adrenal medulla of mammals. Their location suggests that they provide information on the balance between O₂ delivery and tissue utilization.

Vascularly oriented NECs are found on all gill arches and in some fish are especially prevalent surrounding the efferent branchial artery, the main vessel draining oxygenated blood from the gill(Milsom and Burleson, 2007). Here, they are ideally situated to monitor blood oxygenation and, unlike mammals, they may better distinguish blood O₂ content than blood P_O2. Like other gill NECs and cells of the NEB they secrete serotonin in response to hypoxemia.

The carotid bodies of mammals are small organelles located at the bifurcation of the internal and external carotid arteries. They are supplied by the carotid artery and have the highest blood flow and O₂consumption per tissue weight of any organ. As such, they are reporters of blood P_O2 rather than O₂ content. Carotid bodies are ideally positioned to monitor O₂ delivery to the brain and are the main, if not only, sensor of acute and chronic arterial hypoxemia in the adult (Milsom and Burleson, 2007). Type I glomus cells are the primary chemosensory cell. Although chemoreceptor sensitivity varies from species to species, there is also considerable difference depending on the site of measurement;chemoreceptor discharge is half-maximal when arterial P_O2 is around 55 mmHg which is equivalent to 18 mmHg in the vicinity of the type I cells(Kumar, 2007). Intracellular P_O2 is probably considerably lower. Acetylcholine and ATP are the primary secretory products, although the carotid body has been reported to have as many neurotransmitters as the brain(Prabhakar, 2006). Arterial hypoxia initiates an increase in ventilation and peripheral vasoconstriction,whereas if breathing is not possible, a fall in P_O2 produces a reflex bradycardia and peripheral vasoconstriction (Marshall,1994).

Chromaffin cells are associated with systemic veins in ancient fishes and the posterior cardinal veins of teleosts. Catecholamines are secreted in response to hypoxemia and stimulate cardiorespiratory responses as well as increase O₂ transport by red blood cells(Perry and Thomas, 1991), but this may not reflect direct O₂ sensing by chromaffin cells(Perry et al., 2000). Conversely, hypoxemia stimulates mammalian chromaffin cells in the adrenal medulla to release catecholamines into the bloodstream where they have general cardiorespiratory effects (Nurse et al.,2006).

It is well known that in mammals hypoxia contracts pulmonary and relaxes systemic vessels. Although these responses may be modulated by a variety of factors such as paracrine signals from the endothelium or surrounding tissue,the basic responses are intrinsic to the vascular smooth muscle cells (VSMCs)(Madden et al., 1992). Hypoxic pulmonary vasoconstriction (HPV) decreases perfusion of poorly ventilated regions of the lung and thus decreases the potential for partial unsaturation of pulmonary venous (systemic arterial) blood. Hypoxic vasodilation increases tissue perfusion thereby increasing O₂ delivery. Perfused gills and lungs of non-mammalian vertebrates also exhibit hypoxic vasoconstriction, but surprisingly, hypoxia also constricts many systemic conductance (large)vessels in a range of vertebrates from hagfish and lamprey to birds (see Russell et al., 2007). The physiological significance of this is unclear and the effects of hypoxia on resistance-size (0.01–0.2 mm diameter) vessels in many of these animals remains to be determined.

Most effort in identifying the oxygen sensor(s) has focused on mammalian chemoreceptors (carotid body and neuroepithelial bodies) and on pulmonary and systemic arteries. It was thought that since cell activation ultimately appeared to depend on an increase in intracellular calcium, the latter could be achieved through cell depolarization and an increase in calcium influx through voltage-sensitive (L-type) calcium channels. As this can be accomplished by closing potassium channels, it was logical to look for O₂-sensitive potassium channels, and in fact a number of these have been identified in most if not all O₂-sensing tissues, including the carotid and neuroepithelial bodies(Pérez-García et al.,2004; López-López and Pérez-García, 2007), adrenal chromaffin cells(Nurse et al., 2006),pulmonary artery (Weir and Archer,1995; Murry et al., 2006) and even fish gill neuroepithelial cells(Jonz et al., 2004; Burleson et al., 2006). Conversely, hypoxic vasodilation (relaxation) has been proposed to result from the metabolism-sensitive potassium channels (K_ATP channels) that open during hypoxia (Weir and Archer,1995), or a loss of O₂-sensitive k_V1.5 and k_V2.1 channels (Thorne et al.,2002) that hyperpolarize the cell and decrease calcium influx).

It has become evident, however, that while O₂-sensitive K⁺ channels may contribute to, or modulate the hypoxic response,other factors `upstream' from these or other K⁺ channels most likely couple hypoxia to K⁺ channel inactivation(Peers and Kemp, 2001; Peers and Kemp, 2004). Although there are exceptions (see below).

In carotid body type I cells hypoxia has been reported to activate NADPH oxidase and increase reactive oxygen species (ROS), activate AMP-activated protein kinase, inhibit mitochondrial metabolism, or stimulate hemeoxygenase-2(HO-2) production of carbon monoxide, and any of these could serve as the O₂-sensing mechanism (Peers and Kemp, 2001; Dinger et al.,2007; Gonzalez et al.,2007; Lahiri et al.,2006; Prabhakar,2006; Riesco-Fagundo et al.,2001; Kemp, 2005; Wyatt and Evans, 2007; Wyatt et al., 2007). In airway NEB, hypoxia has been proposed to inactivate NADPH oxidase thereby decreasing ROS (Kemp et al., 2002; Nurse et al., 2006), and in adrenal chromaffin cells a mitochondrial mechanism decreases ROS(Nurse et al., 2006). Pulmonary arteries have perhaps the greatest cadre of potential O₂-sensing mechanisms including redox sensors(Archer and Michelakis, 2002),either an increase or decrease in ROS(Ward et al., 2006), NADPH and NADPH oxidase (Jones et al.,2000; Wolin et al.,2005), cyclic ADP ribose(Evans and Dipp, 2002) and Rho kinase (Fagan et al., 2004). Other upstream activators of hypoxic vasodilation, which may or may not be associated with a decrease in intracellular calcium(Thorne et al., 2004), include intracellular acidosis (Nagesetty and Paul, 1994), redox control of cytosolic NADPH(Wolin et al., 2005), direct modulation of internal Ca²⁺ stores and Rho kinase(Thorne et al., 2004).

The variety of potential O₂-sensing systems suggests that either this process evolved numerous times and now constitutes a highly redundant system, or that perhaps many of these mechanisms may be either modulators of,or even peripheral to, an even more fundamental process. Not surprisingly,many of these proposed O₂-sensing mechanisms have their detractors and in many instances there are as many arguments against a particular sensing mechanism as there are arguments for it. This has fostered reviews such as`Detecting acute changes in oxygen: will the real sensor please stand up?'(Kemp, 2006), `Hypoxic pulmonary vasoconstriction: mechanisms and controversies'Aaronson et al., 2006),`Hypoxic pulmonary vasoconstriction is/is not mediated by increased production of reactive oxygen species' (Ward et al.,2006), `Oxidant and redox signaling in vascular oxygen-sensing mechanisms: basic concepts, current controversies, and potential importance of cytosolic NADPH' (Wolin et al.,2005), `The mechanisms of hypoxic pulmonary vasoconstriction:potassium channels, redox oxygen sensors, and controversies'(Archer and Michelakis, 2002),`Hypoxia, anoxia, and O₂ sensing: the search continues'(Schumacker, 2002),`Mechanisms of hypoxic pulmonary vasoconstriction: can anyone be right?(Ward and Aaronson, 1999). Are there multiple and redundant O₂ sensors, or does the lack of unequivocal support for a unifying mechanism suggest that it remains to be identified?

Peers and Kemp (Peers and Kemp,2001) state, regarding airway (neuroepithelial) and carotid chemoreceptors, that `...putative O₂ sensors would be drawn from a pool of proteins that naturally underwent oxido-reductive transitions.'However, it is not exactly clear to me why the O₂ sensor must be a macromolecule. In fact we have recently proposed a novel O₂-sensing mechanism that is based on the metabolism of a very simple (but noticeable)gas, hydrogen sulfide (H₂S). This mechanism has both experimental and anecdotal support, as described in the following paragraphs.

There is a rapidly accumulating body of evidence linking H₂S with nitric oxide (NO) and carbon monoxide (CO) as the third `gasotransmitter'(Wang, 2002). Reports of the physiopharmacological effects of H₂S suggest an almost ubiquitous presence in tissues. H₂S has been implicated in neurological,cardiovascular, gastrointestinal, genitourinary and endocrine systems where it may have modulatory and/or cytoprotective effects(Qu et al., 2007; Li and Moore, 2008; Leffler et al., 2006; Li and Moore, 2008; Łowicka and Bełtowski,2007; Szabó,2007). H₂S has also been proposed to be both pro- and anti-inflammatory (Li et al.,2006; Wallace,2007) and it has even been reported to induce suspended animation in mice (Blackstone et al.,2005). In the cardiovascular system, H₂S has long been known to dilate systemic vessels (Hosoki et al., 1997). H₂S is synthesized in many, if not all,tissues by two cytosolic pyridoxyl 5′phosphate-dependent enzymes,cystathionine β-synthase (CBS) and cystathionine λ-lyase (CSE). CBS is thought to be the primary source for H₂S production in the brain, and CSE in the vasculature(Szabó, 2007; Li and Moore, 2008).

During the course of independent studies on the phylogeny of vascular responses to hypoxia (Smith et al.,2001; Olson et al.,2001; Russell et al.,2001; Russell et al.,2007) and H₂S(Dombkowski et al., 2004; Dombkowski et al., 2005; Olson, 2005), it became evident that these two stimuli evoked the same response in isolated blood vessels, irrespective of whether the response was a contraction, relaxation or was multi-phasic (Fig. 1). This prompted further investigation into whether this was coincidental or if the two stimuli were somehow related. We(Olson et al., 2006)subsequently showed that, (1) H₂S was enzymatically generated by many blood vessels, (2) inhibition of H₂S production inhibited vascular responses to hypoxia (Fig. 2), (3) addition of cysteine, the metabolic precursor for H₂S, greatly augmented the hypoxic response, (4) hypoxia and H₂S appeared to compete for a common downstream activation process,and (5) hypoxia and H₂S produce the same degree of cellular depolarization in vascular smooth muscle. We also found evidence for H₂S mediation of hypoxic relaxation of the trout urinary bladder(Dombkowski et al., 2006),suggesting that this mechanism may be intrinsic to smooth muscle in general.

These studies have led to the formulation of a model for O₂sensing through O₂-dependent regulation of H₂S metabolism (Olson et al.,2006). In this model (Fig. 3) H₂S is constitutively synthesized in the cell cytoplasm and oxidized in the mitochondria. The actual concentration of biologically active H₂S is determined by the simple balance between H₂S production and the amount of O₂ available for H₂S oxidation, i.e. tissue P_O2. In addition CBS and CSE also appear to be sensitive to P_O2 or intracellular redox state(Maclean and Kraus, 2004; Stipanuk, 2004; Banerjee and Zou, 2005) and may either directly contribute to intracellular [H₂S] or may provide a long-term mechanism to bias the rate of constitutive H₂S production. We recently found evidence coupling H₂S production and metabolism to O₂. As shown in Fig. 4A,B, H₂S production by minced trout heart (at P_O2∼0)is increased by cysteine (Cys) and transiently inhibited by injection of micromolar amounts of O₂. Because this only occurs at very low P_O2, it suggests that under normal circumstances intracellular H₂S is very low. Conversely, there is a net H₂S consumption by purified trout mitochondria that is increased in the presence of O₂ and decreased when the mitochondria are gassed with nitrogen (N₂) or heat denatured(Fig. 4C,D). Hildebrandt and Grieshaber (Hildebrandt and Grieshaber,2008) have recently summarized the pathways involved in mitochondrial H₂S oxidation.

Most of our attention to date has focused on H₂S metabolism as the mediator of hypoxic vasoconstriction. However, in many vessels, i.e. trout efferent branchial arteries (Dombkowski et al., 2004) toad and duck pulmonary arteries(Dombkowski et al., 2005) and bovine pulonary arteries (Fig. 1D), it appears that H₂S produces both vasoconstriction and vasodilation and that these occur at different H₂S concentrations. This implies that H₂S has multiple effects on the mechanical properties of vascular smooth muscle, a rare but not unheard of attribute. This too may have physiological significance. For example, it is predicted that low concentrations of H₂S, which would be expected in normoxia, may in fact aid in pulmonary vasodilation; this would then convert to pulmonary vasoconstriction when O₂ levels fall and[H₂S] increases.

Arguments that H₂S-mediated contraction of vascular smooth is only an indirect response have been presented by Koenitzer et al.(Koenitzer et al., 2007). They observed that a H₂S-mediated dilation of rat aortas became more sensitive to H₂S at low P_O2, whereas a H₂S-mediated contraction was only observed at high[H₂S] with concomitantly high P_O2. They proposed that the H₂S-mediated contraction was due to an oxidation product of H₂S rather than a direct vasoconstrictory action of H₂S itself, implying that H₂S is only a vasodilator. However, rat aortas also dilate to hypoxia and may not be a good model with which to examine H₂S vasoconstriction. We(Olson et al., 2008a) repeated these studies in both hagfish and lamprey dorsal aortas, where hypoxia and H₂S only produce vasoconstriction and found that these vessels also became more sensitive to H₂S at low P_O2. Thus it is evident that the H₂S contraction is not the result of an oxidation product. This also provides further support for our hypothesis of H₂S in O₂ sensing,i.e. the vessels become more sensitive to exogenous H₂S at low P_O2 because of the additive affect of endogenous H₂S production.

If H₂S metabolism is involved in O₂ sensing in vascular smooth muscle, we wondered if this could be a general O₂-sensing mechanism in other oxygen-sensitive tissues as well. To examine this, we turned to the external chemoreceptors on the first gill arch of trout. We (Olson et al.,2008b) found that injection of H₂S into the buccal cavity mimicked the classical hypoxic induced bradycardia(Fig. 5) and this response was inhibited when the first, but not second, gill arches, were removed. In addition, both hypoxia and H₂S depolarized isolated zebrafish neuroepithelial cells. This suggests that H₂S metabolism may be an O₂-sensing mechanism in a variety of O₂-sensitive tissues. Interestingly, we also found in unanesthetized trout that whereas hypoxia increased plasma sulfide when measured with a sulfide electrode, it did not affect plasma H₂S measured with a polarographic H₂S sensor. This suggests that some sulfide moiety is carried in the plasma, but it is not H₂S. This is considered further under Curbing the exuberance (below).

There is some anecdotal support for the involvement of H₂S in vascular signaling and in hypoxia sensing in general. First, H₂S has been a common environmental gas since the origin of life and cells have had considerable opportunity to use this gas as a metabolic substrate and deal with potential toxicity (Goubern et al.,2007). Second, there is evidence to support intracellular compartmentalization of H₂S production and oxidation. Searcy(Searcy, 2003) proposed that mitochondria originated from sulfide (H₂S) oxidizing thermophilic bacteria and the nucleocytoplasm from sulfide-reducing Archaea. Thus the mechanism for constitutive production of H₂S and its O₂-dependent inactivation by mitochondria is a common feature of all eukaryotic cells. Third, systemic arterioles consume O₂ at a rate of up to 500 times that of resting skeletal muscle(Shibata et al., 2005). This is far greater than that required to maintain wall tension and the reason for this high level of O₂ consumption is not readily apparent. Could it be due to the continual need for H₂S oxidation in these cells?

Like its predecessor nitric oxide (NO), H₂S has garnered considerable interest as a biologically important regulatory molecule in a seemingly endless number and variety of physiological processes and for its potential therapeutic value (Szabó,2007). Before we get swept away in all the excitement we need to examine several aspects of H₂S signaling that are a bit troublesome.

The first concerns reported levels of H₂S in plasma. In the approximately 25 studies conducted since 2000 all but one have reported plasma or blood H₂S in micromolar concentrations (typically 30–60μmol l^–1 but as high as 300 μmol l^–1)and all but one have used acidic or alkaline modification of the sample prior to measurements [summarized in Whitfield(Whitfield et al., 2008) and detailed in the supplementary material]. These concentrations are typically used in physiological studies, however, they are tens to hundreds of times greater than that which inhibits purified cytochrome c oxidase (1μmol l^–1). Furthermore, at these concentrations, the smell of H₂S should be readily apparent in both plasma and exhaled breath, which, fortunately, is not the case(Whitfield et al., 2008). These 25 studies have also largely ignored reports going back as far as the 1920s (c.f. Haggard, 1921)that showed very little, or no, H₂S in blood and that demonstrated a large capacity for blood to absorb exogenously administered H₂S[summarized by Whitfield (Whitfield et al., 2008)]. We (Whitfield et al., 2008) used a newly developed polarographic sensor to measure H₂S gas in plasma or whole blood at sub-micromolar concentrations,and without any chemical modification, and failed to find measurable amounts of H₂S in blood or plasma from lamprey, trout, mouse, rat, pig or cow, or in an unanesthetized trout fitted with an extracorporeal loop. Furthermore, exogenously administered H₂S rapidly disappeared from whole blood in vitro (Fig. 5A,B) and from blood of the unanesthetized trout(Fig. 5C). We also found that the antioxidant buffer commonly employed with S^2–-selective electrodes generated H₂S from bovine serum albumin, presumably from hydroxyl replacement of cysteine sulfur. Our studies with the polarographic H₂S sensor raise the question of the physiological relevance of studies that employ what appears to be pharmacological concentrations of H₂S. What is too high, and what is physiological in the region of the intracellular effectors? Although it is clear that H₂S does not circulate at 30 μmol l^–1, it remains to be determined how fast exogenous H₂S moves across the cell membrane and how it is distributed within the cell. Does it equilibrate, or does metabolism maintain a lower intracellular concentration, and is H₂S partitioned within the cell? Are seemingly excessive exogenous H₂S concentrations pharmacological or are they necessary to produce a physiological increase in intracellular H₂S? Toxicological studies lend some credence to the latter. Purified cytochrome c oxidase is inhibited by nanomolar H₂S concentrations whereas mitochondria and whole animals are poisoned at micromolar concentrations and sulfide oxidizing bacteria survive 1–2 mmol l^–1 H₂S (Bagarino, 1992). (In our experience some vessels even tolerate >10 mmol l^–1H₂S.) Because there does not appear to be any species-specific differences in cytochrome c oxidase (Bagarino, 1992), it seems likely that differential rates of H₂S metabolism could account for the wide range in H₂S sensitivity, and perhaps, the threshold concentration necessary for physiological responses. Clearly, one of the next breakthrough in H₂S physiology will be measurement of H₂S concentration at the sub-cellular level. Our studies showing rapid removal of exogenous sulfide from the blood, in vivo and in vitro (Whitfield et al.,2008) also raises the question of how numerous studies could show that a variety of experimental manipulations have had lasting effects on plasma [H₂S].

The second troublesome aspect of H₂S signaling concerns the lack of specific inhibitors of CBS and CSE. The only readily available enzyme inhibitors, aminooxyacetate (AOA) for CBS, propargylglycine (PPG) andβ-cyanoanalyine (BCA) for CSE and hydroxylamine (HA) for both CBS and CSE lack specificity and are often poorly absorbed by intact tissues(Szabó, 2007). Thus ascribing definitive physiological effects to inhibition of H₂S production is somewhat tenuous.

The third aspect is that H₂S is a good reductant and can potentially react with other signaling molecules (i.e. NO) potentially removing them or modifying their biological attributes. Thus it is not always evident if the effect of H₂S is direct or indirect, although this may not be critical.

Hypoxia, be it of external or internal origin is a potential threat to all animals. Oxygen-sensing abilities of a variety of specialized cells in vertebrate tissues have been known for nearly one hundred years, but the identity of the intracellular Oxygen-sensing mechanism has not yet been resolved. The lack of a unifying hypotheses of O₂-sensing signal transduction has left the door open for yet another potential O₂sensor; this one based on the balance between constitutive production of biologically active H₂S and its metabolism by available O₂. The role of H₂S as a `gasotransmitter' in a seeming endless list of biological mechanisms including O₂ sensing needs additional validation. Important questions regarding the biology of H₂S remain. (1) Does H₂S circulate in the blood bound to some carrier or is it completely metabolized? It is difficult to reconcile studies with the polarographic sensor with those that use other methods and report a change in plasma [H₂S] following experimental manipulation. (2) What is the mechanism of H₂S action in excitable cells? To date, the only known effect is via opening of K_ATP channels, and that relaxes vascular smooth muscle. (3) What is the biologically active form of H₂S? As a weak acid with a pKa close to 7 (Olson,2005), H₂S in solution exists as both H₂S gas and HS^–. As an anion, HS^– is very similar in size to Cl^– and could potentially become involved in Cl^–-mediated processes. (4) Does H₂S participate in O₂ sensing in other oxygen-sensitive tissues, i.e. is there a universal O₂ sensor, or did O₂-sensing mechanisms evolve independently with different tissues? Can H₂S exemplify Occam's razor?

The author wishes to acknowledge the numerous colleagues, Drs R. Dombkowski, D. Duff, D. Emenim, M. Russell, J. Madden, S. Perry, N. Skovgaard and students, M. Healy, E. Kreimier, T. Mudigonda M. Perks, S. Head, N. Whitfield, F. Verdial, who participated in this research. Supported in part by National Science Foundation grant nos IBN 0235223 and IOS 0641436.