Significance of ROS in oxygen sensing in cell systems with sensitivy to ohysiological hypoxia

Respiratory Physiology & Neurobiology 132 (2002) 17–41 Significance of ROS in oxygen sensing in cell systems with sensitivity to physiological hypoxia Constancio Gonzalez *, Gloria Sanz-Alfayate, M. Teresa Agapito, Angela Gomez-Nin˜o, Asuncio´n Rocher, Ana Obeso Departamento de Bioquı´mica y Biologia Molecular y Fisiologı´a,Facultad de Medicina, Instituto de Biologı´a y Gene´tica Molecular ( IBGM ) ,Uni6ersidad de Valladolid y CSIC, 47005 Valladolid,Spain Accepted 28 August 2001 Abstract Reactive oxygen species (ROS) are oxygen-containing molecular entities which are more potent and effective oxidizing agents than is molecular oxygen itself. With the exception of phagocytic cells, where ROS play an important physiological role in defense reactions, ROS have classically been considered undesirable byproducts of cell metabolism, existing several cellular mechanisms aimed to dispose them. Recently, however, ROS have been considered important intracellular signaling molecules, which may act as mediators or second messengers in many cell functions. This is the proposed role for ROS in oxygen sensing in systems, such as carotid body chemoreceptor cells, pulmonary artery smooth muscle cells, and erythropoietin-producing cells. These unique cells comprise essential parts of homeostatic loops directed to maintain oxygen levels in multicellular organisms in situations of hypoxia. The present article examines the possible significance of ROS in these three cell systems, and proposes a set of criteria that ROS should satisfy for their consideration as mediators in hypoxic transduction cascades. In none of the three cell types do ROS satisfy these criteria, and thus it appears that alternative mechanisms are responsible for the transduction cascades linking hypoxia to the release of neurotransmitters in chemoreceptor cells, contraction in pulmonary artery smooth muscle cells and erythropoietin secretion in erythropoietin producing cells. © 2002 Elsevier Science B.V. All rights reserved. Keywords : Carotid body, signal transduction; Control of breathing, O 2 sensing; Mediators, erythropoietin; Mediators, ROS; Oxygen, cellular sensing; Signal transduction, carotid body, erythropoietin cell, pulmonary vasculature www.elsevier.com/locate/resphysiol 1. Cell systems with sensitivity to physiological hypoxia The appearance of O 2 on a previously anaerobic earth, forced living organisms to develop antioxidant defense systems. Organisms that evolved tolerance to O 2 also developed mechanisms to use this new element for metabolic transformations * Corresponding author. Tel.: +34-983-423-089; fax: +34983-423-588 E-mail address : [email protected] (C. Gonzalez). 1569-9048/02/$ - see front matter © 2002 Elsevier Science B.V. All rights reserved. PII: S1569-9048(02)00047-2 C.Gonzalez et al. / Respiratory Physiology & Neurobiology 132 (2002) 17–41 18 and efficient energy production by using electron transport chains with O 2 as the final electron acceptor. The development of efficient and secure energy production systems allowed the appearance of multicellular organisms. The acquisition of O 2 and nutrients by diffusion from the surrounding aqueous medium was succeeded in multicellular organisms by the appearance of a circulatory system, a system of tubes to carry and to distribute O 2 and nutrients among all cells of the living organism. Evolution later provided an oxygenating or respiratory system capable of supplying O 2 to the fluid moving in the circulatory system, and the moving fluid itself evolved to blood to gain in capacity to transport O 2 . All these acquisitions provided living organisms with the independence necessary to abandon water and to move freely on land. Nevertheless, full acquisition of terrestrial freedom was only possible by the simultaneous appearance of homeostatic mechanisms. Some of those homeostatic mechanisms match O 2 supply to whole-body energetic needs, by monitoring O 2 levels in the body, and altering the overall function of the respiratory and circulatory systems, as well as the capacity of blood to transport O 2 . There are two main systems regarding O 2 homeostasis. They conform two functional loops with their origin in the O 2 -monitoring cells, the chemoreceptor cells of the carotid body (CB) and the erythropoietin (EPO)-producing cells (Richalet, 1997; Gonzalez, 1998). These two cell types are receptor–effector units, which detect arterial blood O 2 levels and respond with a secretory response. The functional loop originated in the CB works as follows: Chemoreceptors cells detect arterial blood P O 2 and become activated when arterial P O 2 decreases; they release neurotransmitters which activate the sensory nerve endings of the carotid sinus nerve, whose central projections end in the nucleus of the tractus solitarius in the medulla. Integration of the carotid sinus nerve input in the respiratory control centers of the brain stem results in an increased activity of the respiratory muscles, with increased ventilation and increased alveolar and arterial blood P O 2 (Gonzalez et al., 1992, 1994). EPO-producing cells, which in adult animals are mainly located in the kidney, also detect arterial blood O 2 levels and are activated by hypoxia to release EPO; this glycoproteic hormone reaches the bone marrow and activates erythropoiesis. An increase in the red blood cell mass and ultimately in the capacity of the blood to transport O 2 follow (Jelkmann, 1992). These two systems probably are aided in maintaining the homeostasis of O 2 by the pulmonary artery smooth muscle cells (PASMC), which, as the other cell two types, monitor oxygen levels and react to hypoxia with a contractile response leading to vasoconstriction and lung blood flow redistribution which would ameliorate the oxygenation of blood (Marshall et al., 1994). We have previously defined physiological hypoxia and these three cell types as endowed with sensitivity to physiological hypoxia (Gonzalez, 1998). These three cell systems, contributing to the homeostasis of O 2 , share other properties. They have a very low threshold to hypoxia, becoming activated at an arterial P O 2 of :70 mmHg and increasing their effector activity with the intensity of hypoxia. In contrast, the response to hypoxia in most cells is a decrease in their activity (Hochachka et al., 1997) and as the hypoxic threshold for this behavior is higher, a more intense hypoxia is needed to elicit it. The different behavior of the cell types involved in homeostatic loops versus general cells of the organism allows to predict that the detection of O 2 levels (i.e. (O 2 -sensing), if linked to O 2 -consumption rate, should be associated to an increase of O 2 -consumption in the case of the cells involved in homeostatic loops and to a decrease in the rest of the cells. The increase in O 2 -consumption during hypoxia in the former cells is an absolute requirement to support the increased activity they exhibit in these circumstances; the decrease in O 2 -consumption during hypoxia in the rest of the cells is inferred from the near shutdown of all energy-requiring processes. Alternatively, O 2 -sensing could be independent of O 2 -consumption rate in both categories of cells. Whatever the case, there are experimental observations supporting the contention that the cell types involved in homeostatic loops, as exemplified by the CB chemoreceptor cells, exhibit an increased metabolic rate and O 2 - consumption during hypoxia. Thus, ATP levels are not significantly different in normoxia and moderate hypoxia (Verna et al., 1990), and at the same time, hypoxia of comparable intensity in- C.Gonzalez et al. / Respiratory Physiology & Neurobiology 132 (2002) 17–41 19 Fig. 1. Redox models for oxygen sensing in the three cells systems located at the origin of oxygen homeostatic loops. The top lateral arrows with question marks represent alternative oxygen sensing models (e.g. hemoprotein sensing) not discussed in this article. creases glucose oxidation to CO 2 in an ouabaindependent manner (Obeso et al., 1989b, 1993). An alternative way of stating the notions expressed in previous paragraphs is to say that in higher organisms there are a few cell systems devoted to maintaining O 2 levels in the entire organism which appear to have a unique mechanism for O 2 -sensing. The uniqueness of this O 2 - sensing mechanism refers to its threshold and to its ability to become coupled to and to activate cell processes demanding an energy supply. This article will consider the redox models of O 2 -sensing. We shall discuss the available information in favor and against the participation of oxygen reactive species (ROS) in O 2 -sensing and the possible significance of ROS as coupling factors between hypoxia and the effector machinery of the cells involved in the homeostasis of oxygen (Fig. 1). 2. Free radicals, oxygen reactive species (ROS) and allied molecules 2 . 1 .Primary free radicals Free radicals are chemically defined as atoms or molecules that have one or more unpaired electron; they usually are very reactive (Punchard and Kelly, 1996; Halliwell and Gutteridge, 1999). The primary processes that give rise to free radicals in biological systems are divided into two main categories. Due to the high water content of living organisms, the cleavage of the chemical bonds of water by high-energy radiation (water radiolysis) constitutes an important source of free radicals according to the reaction: H 2 OH(hydrogen radical) +OH(hydroxyl radical) C.Gonzalez et al. / Respiratory Physiology & Neurobiology 132 (2002) 17–41 20 Ultraviolet light and even visible light in the presence of some sensitizer molecules can similarly produce homolytic scission of hydrogen peroxide in the skin to generate hydroxyl radicals (H 2 O 2 2OH). More important, however, (at least quantitatively) is the genesis of free radicals by reactions involving the transference of electrons, and among them the most important are those reactions in which a single electron is transferred to a molecule of O 2 to form the superoxide radical, O 2 − (Davies and Dean, 1997). Phagocytes possess an enzymatic complex containing FAD and a specific b-type cytochrome capable of taking an electron from NADPH and transferring it to O 2 to form O 2 − . This is probably the only reaction in which O 2 − is a beneficial product; this radical and other ROS formed from it are used by phagocytic cells to kill bacteria. There are, however, many unwanted processes producing O 2 − as a byproduct, some of which are briefly commented here. There are auto-oxidation reactions in which several biologically important molecules such as tyrosine, catecholamines and tryptophan slowly reduce O 2 to form O 2 − .O 2 − in turn is capable of further oxidizing those molecules and a cascade of reactions might be started with potentially deleterious effects. Microsomal enzymes involved in reactions of oxidation and hydroxylation of many types of molecules (alcohols, barbiturates, antibiotics, steroid hormone synthesis and metabolism, etc.), which use cytochrome P450 as an electron transport system, can also generate O 2 − . Superoxide radical is also formed in the transport of O 2 by hemoglobin. Although essentially all O 2 binding to hemoglobin (and myoglobin) does not imply the oxidation of Fe 2+ in the heme moiety, there is, however, a certain degree of delocalization of an electron when O 2 is bound and thereby the appearance of some kind of intermediate structure that can be written as an apparent resonant ‘equilibrium’: hemo-Fe 2+ –O 2 lhemo-Fe 3+ –O 2 − . Usually, the Fe 2+ of heme group keeps the delocalized electron and hemoglobin releases O 2 , but occasionally (3%/day) the electron is kept by O 2 and hemoglobin then releases O 2 − , the Fe 2+ of the heme group is transformed into Fe 3+ and methemoglobin is formed. In situations of tissue injury, especially in the processes of ischemiareperfusion, the enzyme xanthine dehydrogenase which oxidizes hypoxanthine and xanthine to uric acid using NAD + as acceptor of electrons, is modified by oxidative or proteolytic processes and converted into xanthine oxidase, which is capable of transferring the electrons of the purines to O 2 to form O 2 − . However, there is no doubt that quantitatively speaking the most important source of O 2 − is the mitochondrial electron transport chain, where univalent reduction of O 2 may take place in some steps of the overall transport system (Fig. 2). In the two places where O 2 − can be produced (complex I and ubiquinone pool) there are electron shuttle molecules (flavin adenin mononucleotide and ubiquinone), which can accept one or two electrons. When they accept a single electron they become free radical molecules themselves capable of slipping the electron to molecular O 2 . It is estimated that a small percentage (B5%) of the total O 2 consumed by mitochondria at normal P O 2 is converted to O 2 − . To complete this overview on radicals it remains to say some words on reactive nitrogen species. Nitric oxide or nitrogen monoxide (NO) is a free radical as it has one unpaired electron. It is a gas molecule capable of freely diffusing through cell membrane systems. NOis physiologically synthesized by a family of different nitric oxide synthases from the amino acid L -arginine. It binds to Fe 2+ of heme groups, and in fact many of its physiological actions are mediated by the soluble guanylate cyclase, which is a hemoprotein that synthesizes cGMP, although cyclase independent physiological effects have been described (Archer et al., 1994; Bolotina et al., 1994; Ohkuma and Katsura, 2001). Due to its ability to react with heme groups, NO can inhibit cytochrome oxidase competitively with O 2 (Boveris et al., 2000). In body fluids NOreacts with O 2 to give the more reactive radical nitrogen dioxide (2NO+O 2 2NO 2 ), but largely it is oxidized to nitrite ion (4NO+O 2 +2H 2 O4NO 2 − ) or reacts with the heme groups of oxyhemoglobin to form metahemoglobin and NO 3 − (nitrate). In fact, the evaluation of NOsynthesis in vivo is quite frequently assessed by the measurement of nitrite and nitrate levels. C.Gonzalez et al. / Respiratory Physiology & Neurobiology 132 (2002) 17–41 21 Fig. 2. General schema of the mitochondrial respiratory chain illustrating the loci where O 2 − can be generated. The schema also shows the most commonly used inhibitors of the respiratory chain, which when acting downstream of the loci of production of O 2 −  would increase, while those acting upstream will inhibit O 2 − production. 2 . 2 .Secondary free radicals and other reacti6e species Once generated, these three main free radicals (OH,O 2 − and NO) can undergo an enormous variety of reactions with other molecules giving rise to new radicals or to species, which are oxidized or reduced. When new radicals are formed they are capable of undergoing further reactions originating chain reactions until one of the radicals reacts with a molecule (scavenger) to generate an unreactive compound. This widespread reactivity is the basis for the toxic effects of free radicals. For example, in their reactions with membrane lipids, OHabstract an atom of hydrogen from polyunsaturated fatty acids forming H 2 O plus a modified fatty acid which is a carbon radical species readily reacting with molecular O 2 . The incorporation of O 2 by the fatty acid carbon radical species transforms it into a peroxyl radical (RCOOCOOH) which in turn is capable of abstracting another hydrogen atom from a nearby polyunsaturated fatty acid, thereby spreading out the alteration of lipids; alternatively, it might attack another double bond in the same molecule to form a cyclic endoperoxide compound which finally would break to generate malonaldehyde and other byproducts whose levels are usually determined to establish the magnitude of lipid peroxidation in circumstances of interest. The final point is that the modifications and breakdown of the lipids of the membranes change their fluidity and many of their basic properties. Aditionally, peroxyl radicals, malonaldehyde and other aldehyde byproducts of lipid peroxidation, as well as OHitself and carbon centered radicals, can react with thiol and with amino groups of proteins or to attack many sites in essentially every amino acid to generate all kinds of distorted molecules (breakdown, intermolecular or intramolecular cross-linking of proteins with the formation of aggregates of membrane proteins, etc.) which ultimately imply the loss of the membrane function and its eventual destruction. The same kind of cross-linking, breakdowns and additional modifications can be made by OHin DNA molecules and between DNA and nuclear proteins leading to alterations in the repair of DNA or in C.Gonzalez et al. / Respiratory Physiology & Neurobiology 132 (2002) 17–41 22 its replication or transcription. In sum, OHis so reactive that it will react in the very same place where it is produced and disappear immediately as such entity, but at the same time new reactive molecules would appear which would spread the effects. The question in the context of the present article is if there exists any mechanism in the cells capable of circumscribing the action of OHfor it to act as the signaling molecule during hypoxia as Kietzmann et al. (2000) suggest. Whatever the mechanism(s) it should be based in two premises: first, the production of OHshould be located in a very restricted area of the cell; this restricted area must be coincident with the location of the cell machinery that initiates the hypoxic transduction cascade (for example, the O 2 -sensitive K + channels or hypoxia inducible factor 1a, HIF-1a), and second, the ‘OH-receptor molecule’itself (e.g. the O 2 -sensitive K + channels or HIF-1a) should act as a very effective OHscavenger to avoid the widespread action of secondary reactive molecules with the inevitable loss of capacity to focus the hypoxic signal. The O 2 − is considerably much less reactive than OH. It can be protonized to form HO 2 (hydroperoxyl radical) which has a greater reducing activity and which in addition is freely permeable to membranes owing to is lack of charge; both characteristics make HO 2 an important free radical in spite of the fact that its concentration in biological systems is 1/100–400 that of O 2 − .In addition, hydroperoxyl radical is involved in the spontaneous dismutation of O 2 − to H 2 O 2 and molecular O 2 . This reaction, which involves the reduction of one O 2 − and the oxidation of another to molecular O 2 , occurs through the formation of hydroperoxyl radical according to the equations: H − +O 2 − HO 2 ; HO 2 +O 2 − +H + H 2 O 2 +O 2 Superoxide radical can donate one electron to transition metals transforming Fe 3+ ,Cu 2+ and Mn 3+ into Fe 2+ ,Cu + and Mn 2+ , respectively, and then it acts as a reducing agent, but it can also act as an oxidizing agent reversing the reaction. These reducing/oxidizing reactions are very important because the enzymes responsible for the cellular elimination of O 2 − , the superoxide dismutases (cytosolic or CuZn superoxide dismutase and mitochondrial or manganese superoxide dismutase), are metalloproteins containing those metals and through cycles of reduction and oxidation of their metals accelerate enormously the disappearance of O 2 − . Other important reactions of O 2 − include oxidation of ascorbate and the catechol ring to form ascorbate and diphenol radicals and H 2 O 2 . Another reaction of interest is that of O 2 − with hypoclorous acid (formed in turn by the action of myeloperoxidase on H 2 O 2 + Cl − HOCl+OH − ) to yield O 2 ,Cl − and OH. Summarizing, we can state that up to this point most of O 2 − ends up as H 2 O 2 due to spontaneous and enzyme catalized dismutation. To some extent, we may say that H 2 O 2 is the final common path of most free radicals in the organism. However, H 2 O 2 is harmful to the cells. It belongs to a category of molecules collectively called reactive oxygen species (ROS), which include free radicals. H 2 O 2 and other ROS such as hypoclorous acid or peroxynitrite (NO+O 2 − ONOO − ; see below) have in common with free radicals being more reactive than molecular O 2 , but unlike free radicals they do not have unpaired electrons. The toxicity of H 2 O 2 would explain the existence of specific enzymatic systems to destroy it, including catalases (2H 2 O 2 2H 2 O+O 2 ) and peroxidases, the most important of which is glutathione peroxidase (H 2 O 2 +2GSHGSSG+2H 2 O; GSSG is back reduced to GSH by the action of the NADPH based glutathione reductase). The toxicity of H 2 O 2 is not exerted directly unless it is applied at high concentrations (\10 mm for many cells), but by the OHit can generate upon reaction with transition metals by the Fenton reaction (Fe 2+ +H 2 O 2 Fe 3+ + OH+OH − ) or upon homolytic fission by the action of ultraviolet light. We already have examined the widespread damaging effects of OH (Fig. 3). Nitric oxide (NO) is also readily reactive with many other radicals. For example, it can react rapidly with superoxide to yield peroxynitrite (NO+O 2 − ONOO − ) and with hydroxyl radical to form nitrous acid (NO+OHHNO 2 ). C.Gonzalez et al. / Respiratory Physiology & Neurobiology 132 (2002) 17–41 23 Fig. 3. General schema of ROS generation in the organisms. H 2 O 2 , at the center of the scheme, represents a common final product whose levels are kept low by specific degrading enzymatic systems. Peroxynitrites and nitrous acid are nitrogen reactive species capable of reacting with SH groups of proteins and with iron/sulfur electron transporting proteins of mitochondria. As we have already mentioned, NOcan react in a reversible manner with cytochrome oxidase, thereby competing with O 2 ; this might be a mechanism to control the oxidation rate of the energy-yielding compounds (nutrients). However, mitochondria must harmonize this potentially reversible regulating mechanism of NOwith its ability to react with O 2 − to form peroxynitrite and through it to produce irreversible damaging effects on electron transporting proteins. Peroxynitrite can interact with tyrosine residues in a reaction catalyzed by superoxide dismutase to form the very stable 3-nitrotyroxine, changing the overall charge of the residues of tyrosine present in proteins and making it difficult for 3-nitrotyrosine containing proteins to assemble into multimeric units or to interact with other proteins (Beckman and Koppenol, 1996) and to participate in signaling cascades, by altering the efficiency of tyrosine phosphorylation. The great stability of 3-nitrotyrosine makes the measurement of its level a good index of nitric oxide/ peroxynitrite in biological systems. Peroxynitrite and other nitrogen reactive species can react with thiols, including that in glutathione, to form nitrosothiols. Special interest has recently been given to the nitrosylation of one cysteine residue in the ryanodine receptor/channel, which seems to occur in a P O 2 -dependent manner (Eu et al., 2000). When sarcoplasmic vesicles are exposed to ambient P O 2 (:150 mmHg) six to eight thiol residues per ryanodine receptor become oxidized, the oxidation reverting upon exposure to P O 2 of 10 mmHg, which is in the range of that physiologically encountered in the muscle; on switching back to ambient P O 2 , reoxidation of those thiol residues takes place. These cysteine residues will constitute, according to (Eu et al., 2000), an O 2 - sensor. It is conceptually important to realize that this potential O 2 -sensor needs to be coupled to a redox system, which must be present in the sarcoplasmic reticulum vesicles, and which must be responsible for switching back and forth from the oxidized to the reduced state. Only the reduced form of the channel is susceptible for physiological modulation by NO, which does so by nitrosylating one of the reduced cysteine residues, or in other words, the redox state of the channel set by P O 2 determines the capacity to respond to (or the ability to be regulated by) NO. C.Gonzalez et al. / Respiratory Physiology & Neurobiology 132 (2002) 17–41 24 3. ROS as mediators of the hypoxic transduction cascade 3 . 1 .Criteria to consider ROS mediators of a physiological cellular response Following classical studies on messengers such as neurotransmitters (McLennan, 1963) and the most recent article by Lander (1997) on ROS in signal transduction, it is possible to put forward a set of criteria as requirements to consider ROS as true second messengers with the capacity to control cellular functions under physiological conditions. A first criterion should be that the putative ROS molecule acting as messenger or mediator of the hypoxic response must increase or decrease during hypoxia; a given relationship (modulation can modify that putative relationship) should exist between the ROS level change, the intensity of hypoxia and the strength of the hypoxic response. The second criterion would be that the cellular mechanism responsible for the change in the level of ROS molecule, putative messenger, needs to be identified. The third criterion would be the definition of the primary target of the ROS as well as the definition of the change produced by the ROS molecule on its target. The possibility exists that the primary target of the putative messenger ROS molecule is itself the effector(s) of the cell response or on the contrary, it is possible that the primary target of ROS is a first link of a chain or cascade of events coupling the change in the ROS levels to the effector(s) of the cell response; in this latter case the entire cascade should be defined. The fourth criterion would be to define the cellular responses elicited by the effector(s). The fifth criterion would be the definition of mechanisms responsible for the inactivation of ROS signal. Obviously if the ROS species increase during hypoxia we should be looking for ROS degrading mechanisms, but if they decrease we should be looking for the enzyme or cells process leading to the restoration of normal ROS levels. In any case, if a given stimulus acts via a change in ROS levels, the mechanisms for the elimination of those ROS should be induced by prolonged presentation of the stimulus. The sixth criterion (pharmacological) would be that the manipulation of the ROS levels, whether by reagents that modulate glutathione levels or that directly scavenge ROS, should alter the cell response in a predictable and reproducible manner; a sixth bor seventh criterion would be that overexpression or silencing of the expression of the ROS scavenging enzymes should alter the cell response in a predictable and reproducible manner. 3 . 1 . 1 .First criterion : the signal ROS le6els during hypoxia The very first question that should be addressed is if hypoxia increases or decreases ROS in general or if there is a unique ROS species whose production follows a particular behavior during hypoxia. The classical view is that the rate of production of ROS changes in direct relation to the tissue P O 2 (Chance et al., 1979; Archer et al., 1986, 1989). In fact, this view is generally accepted at present (Archer et al., 1993; Gorlach et al., 1994; Fandrey et al., 1994, 1997; Halliwell and Gutteridge, 1999; Kietzmann et al., 2000; Gnaiger et al., 2000). However, quite recently Schumacker and coworkers (Chandel et al., 1998; Duranteau et al., 1998; Chandel and Schumacker, 2000; Chandel et al., 2000) have found the opposite: hypoxia produces an increase in the cellular rate of ROS production which takes place at the level of mitochondria and which is proportional to the level of hypoxia. Chandel et al. (1998) claim that their measurements using dichlorofluorescein fluorescence, which apparently measures preferentially H 2 O 2 , are the correct ones. However there are many measurements using different methods (see Chance et al., 1979) in which it has been shown that hypoxia decreases the rate of production of ROS, the decrease being attributed to a decrease of production in mitochondria as well as in microsomal and peroxisomal enzymes. Archer et al. (1993) and Paky et al. (1993) using lucigenin chemiluminescence, which apparently measures preferentially O 2 − (Halliwell and Gutteridge, 1999), also found that hypoxia decreases ROS production rate. Similarly contradictory are the observations regarding to CoCl 2 (an inducer of EPO production), which decreases the production of ROS if they are measured with dihydrorhodamine 123 (Fandrey et al., 1997), or increases C.Gonzalez et al. / Respiratory Physiology & Neurobiology 132 (2002) 17–41 25 their production if ROS are measured with dichlorofluorescein (Chandel et al., 1998). Interestingly, both dyes yielded an inhibition of ROS generation under rotenone and diphenileneiodonium (DPI) (Cross et al., 1990; Ehleben et al., 1997; Chandel et al., 1998; Duranteau et al., 1998); rotenone and DPI also decreased lucigenin chemiluminescence (Archer et al., 1993, 1999). However, while antimycin A increased dichlorofluorescein fluorescence (Chandel et al., 1998) it decreased lucigenin chemiluminescence (Archer et al., 1993). With this panorama it is difficult to make a meaningful formulation of the putative ROS molecule(s), which are signals for the hypoxic transduction cascade. We do not know if the signal should be a particular ROS species that decreases during hypoxia, a general decrease in the ROS levels, or if on the contrary, we should be searching for something that increases during hypoxia. We would like to quote (Semenza, 1999, 2000) when he writes that ‘the direct measurements of ROS are so demanding that they generate data supporting opposing views’. However, we want to make explicit some thoughts on the problem, being aware that experimental data might contradict the opinions expressed here. When we consider the general mechanisms for ROS production presented in previous paragraphs, the intuitive notion would be that the rate of production of ROS in cells should be directly proportional to the P O 2 of the cells. Reactions of auto-oxidation as well as oxidations driven by enzymes and the acceptance by molecular O 2 of electrons leaking at the level of the quinone pool in mitochondria (see Fig. 2) would decrease during hypoxia, when P O 2 is diminished. At the same time, however, it would appear that the generation of O 2 − at the quinone pool level would depend on the rate of flow of electrons through the respiratory chain, i.e. on the respiratory rate. As stated in the first section of this article, cell systems involved in the general homeostasis of O 2 are activated by hypoxia, this activation requiring an increase in cell respiration to support cell and homeostatic loop function (Obeso et al., 1993, 1997). Thus, it could be expected that in these specialized cells hypoxia produces an increase in electron flow during an ample range of hypoxia (from :70 to 20 mmHg of arterial P O 2 ), and thereby, an increase in the electron leakage and mitochondrial formation of O 2 − is conceivable in those situations (see Fig. 3 in Hoshi and Heinemann, 2001) when hypoxia is intense enough a compromise will occur and in spite of the high affinity of cytochrome oxidase for O 2 , it will be unable to transfer enough electrons to O 2 and the respiratory rate and the genesis of ATP will drop. In the case of the CB in vivo this metabolic compromise occurs at arterial P O 2 :20 mmHg; below this P O 2 the CB is unable to adequately signal the level of hypoxia and chemoreceptor action potential frequency decreases, instead of increasing (see Fidone and Gonzalez, 1996). The majority of the cells of the organism will behave during hypoxia as the CB below the compromise P O 2 : all O 2 − -generating processes including that of mitochondria would decrease as the intensity of hypoxia increases. To state this in a different manner, it would appear that a differential response could be expected: in most cells of the organism hypoxia would tend to decrease ROS production while in CB chemoreceptor cells, in EPO-producing cells and in smooth muscle cells of the pulmonary arteries hypoxia would tend to increase ROS production at least in those ranges of hypoxia compatible with life. In contrast, Schumacker and co-workers measured an increase in ROS production in cardiomyocytes and hepatoma cells (EPO-producing) which is comparable in magnitude and in P O 2 dependence in both cell types (Chandel et al., 1998; Duranteau et al., 1998) and Fandrey et al. (1994) measured a decrease in the rate of production and release of H 2 O 2 to the incubating solution by EPO-producing hepatoma cells precisely in the range of P O 2 where they secrete EPO. In sum, the first criterion is not settled, we do not know whether hypoxia increases or decreases ROS levels. 3 . 1 . 2 .Second criterion : cell mechanisms responsible for the change in ROS production during hypoxia should be defined Those proposing that hypoxia would decrease ROS levels (Cross et al., 1990; Youngson et al., 1993; Acker, 1994a,b; Acker and Xue, 1995; Fan- C.Gonzalez et al. / Respiratory Physiology & Neurobiology 132 (2002) 17–41 32 identification of HIF-1 (Wang and Semenza, 1993a,b) as a factor induced by hypoxia, capable of binding to the 3’enhancer region of EPO and other hypoxia-regulated genes and capable of causing an increase in their rate of transcription. Thereafter, it was demonstrated that in fact HIF1 was a heterodimer composed of HIF-1a, whose levels are regulated by low P O 2 , and by the constitutively expressed HIF-1b[also known as ARNT (aryl hydrocarbon receptor nuclear translocator), as AhR (aryl hydrocarbon receptor) and as dioxin receptor], the dimerization was shown to be necessary, not for the translocation of both components to the nucleus, but for the stable association of HIF-1 within the nuclear compartment (see Gassmann et al., 2000). Additional studies demonstrated that HIF-1aas well as HIF-1b mRNA levels, transcription and translation rates were not modified by hypoxia. Since HIF-1a protein increased during hypoxia it implied that in normoxia the protein is degraded and that hypoxia prevents the degradation (Huang et al., 1996). Therefore, the question we should make at this point is how the change in ROS levels produces such stabilization of HIF-1ato prevent its degradation. Huang et al. (1996) showed that the levels of HIF-1a, which are increased by hypoxia, decreased soon if intact cells were re-exposed to normoxia (21% O 2 ), but exposure of cell extracts to the same normoxic levels of O 2 did not alter the binding of HIF-1ato DNA. In other words, for high P O 2 to destabilize HIF-1athe integrity of the cells is necessary. In additional experiments it was shown that H 2 O 2 (0.1–1.0 mM) resulted in a dose-dependent decrease in the EPO mRNA expression induced by hypoxia which runs parallel to a decrease in HIF-1abinding to DNA and to a decrease in HIF-1aprotein, without modification of HIF-1amRNA levels or translation rate. It was also observed that extracts from hypoxic cells treated with H 2 O 2 showed a normally high binding capacity to DNA. Those findings appear to indicate that the binding of HIF-1ato DNA is controlled primarily at the level of HIF-1a protein. In addition, Huang et al. (1996) also studied some of the modifications of HIF-1a which might alter its binding to DNA. They found that oxidizing sulfhydryl reagents added to cell extracts inhibited the binding in a dose-dependent manner, this inhibition was reversible with reducing thiol reagents except when sulfhydryl alkylating agents were used. In addition, in a cotransfection experiment using a construct expressing thioredoxin under the control of the constitutive cytomegaloviral promoter and enhancer, along with another construct expressing luciferase under the control of the SV40 promoter and EPO-enhancer, they observed that overexpression of thioredoxin resulted in a potentiation of the expression of luciferase both in normoxic and hypoxic conditions, such potentiation appeared to be under the control of the HIF-1 binding because transfection with a mutated EPO enhancer, which was unable to bind HIF-1a, caused the loss of the response. As a whole the results would indicate that H 2 O 2 , like normoxia, produces modifications in the HIF-1amolecule which force it to be directed to degradation, but which does not impede its binding to DNA. Alternatively, hypoxia promotes changes in HIF-1a, which prevent its degradation, leading to accumulation of the transcription factor; the changes induced by hypoxia appear to be redox dependent because they can be mimicked by thioredoxin and counteracted by normoxia and H 2 O 2 ; a similar conclusion was attained by Srinivas et al. (1999). An alternative interpretation to the findings of Huang et al., is that hypoxia (or the decrease of ROS) or the overexpression of thioredoxin produced changes in the degrading enzymes of the proteosoma which make them incapable of degrading a normal (or reduced) HIF 1a. However, in a series of experiments Chandel et al. (2000) found essentially the opposite: hypoxia and CoCl 2 , but not desferriexamine, increased ROS levels, and the three situations augmented the expression of a reporter gene under the control of the EPO-promoter as well as the amount of HIF-1aprotein. Inhibitors of the respiratory chain at the complex I, the thiol reductant pirrolidine dithiocarbamate, and overexpression of catalase, decreased hypoxic ROS levels and prevented the augmented expression of HIF-1a.The thiol reductant and the overexpression of catalase decreased the ROS levels increased by CoCl 2 and C.Gonzalez et al. / Respiratory Physiology & Neurobiology 132 (2002) 17–41 33 also the expression of HIF-1a. Neither maneuver decreased the expression of the reporter gene or decreased the levels of HIF-1apromoted by desferrioxamine. Addition of H 2 O 2 and tert-butylhydroperoxide to normoxic cells promoted an increase of HIF-la protein as well as an increased expression of the reporter gene; these maneuvers were ineffective in cells overexpressing catalase. A phosphatase inhibitor and an inhibitor of phosphatidyl inositol 3-kinase abolished the stabilization of HIF-1ainduced by hypoxia, CoCl 2 and desferrioxamine and a proteasome inhibitor was able to reverse those effects; because treatment of the cells with H 2 O 2 was unable to prevent the effects of the phosphatase and kinase inhibitors, the authors concluded that ROS (including H 2 O 2 ) act to stabilize HIF-1aat a step prior to the action of the phosphatases and kinases. Desferrioxamine would produce the stabilization of HIF-1aby a mechanism independent of ROS and would also act at a step prior to the phosphatases and kinases. Curiously, antimycin A, which increased ROS, did not induce the expression of the reporter gene or accumulation of HIF-1a. Chandel et al. (2000) state that it is because this mitochondrial inhibitor causes a smaller increase in ROS levels than hypoxia. However, it is difficult to understand how hypoxia can produce a more reduced state at the quinone pool than antimycin (see Fig. 2). In addition, antimycin A was applied in normoxia implying that the concentrations of O 2 to accept the electrons from semiquinone are much higher than in hypoxia (20% O 2 vs. 1.5% in their experiments) and, therefore, the formation of O 2 − should be facilitated (Chance et al., 1979). The controversy is obvious: to Huang et al. (1996) it is the decrease of ROS levels which causes the stabilization of HIF-1aand to Chandel et al. (2000) it is the increase of ROS levels that does the same. Therefore, and in keeping with the statement of the third criterion, we do not know if we should be looking for a reduction or an oxidation of HIF-1aas responsible for its stabilization. An additional aspect that we feel forced to restate at this point (which is seldom taken into account; see above) is that in the proposed signalling by ROS the subcellular localization of the ROS signal at the target place should be taken into consideration, otherwise a ‘true’signaling would not exist, Finally, we agree with Halliwell and Gutteridge (2000) when they state regarding signaling by ROS: ‘‘one must be wary: cells in culture are usually hyperoxic and they lack many of the antioxidants they use in vivo…Cells adapt to live in culture. It is not impossible that, with tumor cell lines in particular, adaptations to favor growth may occur by using radicals to trigger pathways that are triggered by other means in vivo’’. 3 . 1 . 4 .Fourth criterion.Definition of the cellular responses elicited by the effector acti6ated by the changes in ROS le6els At this point in the article it is obvious that we can define the cellular responses elicited by the stimulus, i.e. hypoxia. However, we have serious difficulties in defining the cellular effects produced by the ROS level changes occurring during hypoxia because we do not know what those changes are. And even if we knew the changes in ROS levels produced during hypoxia we do not know if they are cause or consequence of activation of the effector and cellular responses. Hypoxia in CB chemoreceptor cells and PASMC reduces the opening probability of the O 2 regulated K + channels (see criterion 3) leading to cell depolarization, activation of voltage operated Ca 2+ channels and release of neurotransmitters in chemoreceptor cells or contraction in smooth muscle cells (see Fig. 1). In criterion 3 we have already indicated that sulfhydryl reagents cause changes in the opening probability or amplitude of the currents that can be compatible with a decrease or an increase in ROS levels. Referring specifically to ROS, there are many studies in which H 2 O 2 has been applied directly and its effects on several types of K + channels established. For example, Vega-Saenz de Miera and Rudy (1992) showed that H 2 O 2 augmented the current carried by transient type K + channels (like those O 2 -sensitive channels in rabbit chemoreceptor cells) without affecting other K + channels. Similarly the Kv3.3 present in airway chemoreceptors are activated by H 2 O 2 (Wang et al., 1996; Fu et al., 2000). And there are many reports showing that H 2 0 2 or tert-butylhydroper- C.Gonzalez et al. / Respiratory Physiology & Neurobiology 132 (2002) 17–41 34 oxide activate maxi-K + (Hayabuchi et al. 1998; Shin et al., 2000; Guo et al., 2000; Barlow et al., 2000). All these data would favor the notion that hypoxia decreases ROS, and thereby hypoxia would decrease the activity of K + channels and depolarization would occur. However, there are studies in which it has been found that H 2 O 2 inhibits K + channels. That is the case for maxiK + (DiChiara and Reinhart, 1997; Tang et al., 2001). These studies would be in agreement with the notion that hypoxia increases ROS levels and thereby causes inhibition of K + channels, cell depolarization and increased cell activity. When we consider the effects of H 2 O 2 on the overall response of the system, the study of Osanai et al. (1997) in the intact CB is very illustrative as they concluded that maneuvers aimed at altering endogenous H 2 O 2 , including exogenous H 2 O 2 application, did not provide evidence for its supposedly critical role in O 2 chemoreception. On the contrary, in the pulmonary circulation tertbutylhydroperoxide, H 2 O 2 and ROS generating systems (i.e. perfusion of the lungs with xanthine plus xanthine oxidase) clearly inhibited hypoxic pulmonary vasoconstriction (Archer et al., 1986; Mohazzab and Wolin, 2000). In the case of EPOproducing cells we have commented in previous paragraphs on the apparently dual actions of H 2 O 2 since on the one hand it inhibited EPO-production induced by hypoxia (Fandrey et al., 1994) and on the other hand it augmented the stabilization of HIF-1aand promoted the expression of a reporter gene under the control of the EPO enhancer (Chandel et al., 2000). It should also be mentioned that Ueno et al. (1988) working with renal carcinoma EPO-producing cells and Ndele et al. (1996)) working with HEP 3B cells found that H 2 O 2 and ROS generating systems increased EPO production while ROS scavengers have the opposite effects. 3 . 1 . 5 .Fifth criterion.The mechanism ( s ) responsible for the inacti6ation of the ROS signal should be defined Obviously we should consider here both possibilities: that the signal is a decrease and that the signal is an increase in ROS levels. In the first case, if a decrease in ROS levels is considered to be the result of a decrease in the activity of NADPH oxidase, the signal would disappear, i.e. normal ROS levels will be recovered, when P O 2 returns to normoxic levels due to an increase in the activity of the enzyme by increased levels of substrate (it might also require a reassembly of the NADPH subunits promoted by unknown mechanisms; see second criterion). It is conceivable that chronic hypoxia would produce a downregulation of NADPH oxidase. A comparable mechanism to inactivate the ROS signal would be valid whether the decrease of ROS levels is considered to occur at the mitochondrial level, cytochrome P450, etc. If on the contrary, the signal is an increase in ROS levels, a given level of ROS should correspond to a given level of stimulation in order to obtain a response proportional to the intensity of the stimulus. This would imply that simultaneously to the activation of the mechanisms generating ROS there is a parallel activation of the mechanisms degrading them. If the stimulation is long lasting (e.g. chronic hypoxia) an upregulation of the mechanisms degrading ROS should be expected (Halliwell and Gutteridge, 1999). Obviously we are referring here to GSH, glutathione peroxidase, catalase and superoxide dismutase as the main systems to dispose ROS. There are not, to our knowledge, direct measurements of those ROS scavenger systems in the cell types we are discussing, however, there are some fragmentary data obtained in intact animals and humans exposed to high altitude. For example, Costa et al. (1993) obtained evidence indicating that in the liver of rats exposed to a simulated altitude of 4400 m there was a decrease in the rate of ROS production with a concomitant decrease in superoxide dismutase activity. However, Nakanishi et al. (1995) at a simulated altitude of 5500 m found a complex pattern of changes in those enzymes which was organ specific, and which was accompanied by an increase in lipid peroxidation and by a marked decrease in body weight of the animals attributed to an increased metabolic rate due to the hypobaric stress. Although it is difficult from these and similar data to reach sound conclusions, some experts in the area would suggest that there are risks of oxidative stress, independent of UV radiation, at high altitude (Simon-Schnass, 2000). C.Gonzalez et al. / Respiratory Physiology & Neurobiology 132 (2002) 17–41 35 3 . 1 . 6 .Sixth criterion.Manipulation of ROS le6els by alteration of glutathione le6els or by the direct action of ROS sca6engers should produce the predicted change in the cell response O6erexpression and elimination of the enzymes sca6enging ROS should produce the predicted change in the cell responses. The interpretation of the data resulting from the manipulation of ROS and/or glutathione levels is not always straightforward due to the multiple targets of ROS. For example, the activation or mimicking of glutathione peroxidase (e.g. by administration of ebselen) would decrease the GSH/ GSSG ratio, just as a situation in which there is an increase in the rate of ROS production, but contrary to that situation ebselen will produce a decrease in the levels of H 2 O 2 . Since the decrease in the GSH/GSSG ratio would tend to increase the opening probability of K + channels, and thereby to cause membrane hyperpolarization (and vasodilation, in the case of pulmonary circulation), and the decrease of H 2 O 2 would tend to reduce the activity of guanylate cyclase, and thereby to inhibit Ca + dependent K + channels (and to produce vasoconstriction, in the case of pulmonary circulation), the overall effect could be very complex and difficult to define (see Wolin et al., 1999; Mohazzab and Wolin, 2000). In addition, and at least in the case of the lung, ROS can modify the metabolism of arachidonic acidderived metabolism in such a manner that an increase in ROS levels would augment the production of vasoconstrictor metabolites (Seeger et al., 1986; Sanderud et al., 1993; see Wolin et al., 1999). Experiments in the CB directly focused on this last criterion have been performed in our laboratory very recently. We have measured GSH and GSSG levels in calf CB in normoxia, hypoxia and after treatment of the organs with N-acetylcysteine, a precursor of glutathione and direct scavenger of ROS and measured the release of catecholamines from chemoreceptor cells in those experimental situations (Obeso et al., 2000b). Nacetylcysteine would mimic a situation in which there is a decreased production of ROS and would oppose a situation of increased rate of ROS production. In other words, N-acetylcysteine would mimic hypoxia, if we accept that hypoxia decreases ROS levels, and would oppose hypoxia if we accept that it augments ROS levels. We found that hypoxia (P O 2 :46 mmHg) did not alter the absolute levels or the GSH/GSSG quotients. N-acetylcysteine increased absolute glutathione levels by primarily increasing GSH; therefore, it increased the GSH/GSSG ratio. Yet N-acetylcysteine did not modify the basal normoxic or hypoxia-induced release of catecholamines suggesting that ROS are not critical in setting the hypoxic response. In the case of the pulmonary circulation, Archer et al. (1986) showed that diamide (an agent that reacts with GSH and oxidizes it to GSSG) reduced the pressor response to hypoxia. Diamide would mimic a situation in which there is an increased production of ROS: it will cause a decrease in GSH/GSSG ratio and therefore the cellular power for scavenging of ROS leading to an increase in ROS levels. Both the decrease in GSH/GSSG ratio and the increase in ROS levels would tend to oppose the vasoconstriction provoked by hypoxia (see above). In agreement with the expected actions, pretreatment of the animals with N-acetylcysteine prevented the effects of diamide (Archer et al., 1986). In isolated perfused lungs, suplementation of the superfusates with catalase and superoxide dismutase neutralized the inhibition on hypoxic pulmonary vasoconstriction produced by ROS generating systems (xanthine/ xanthine oxidase) (Archer et al., 1986). But in another article it was reported that neither catalase nor superoxide dismutase modified the vasoconstriction elicited by hypoxia (Kjaeve et al., 1990), and more recently Weissmann et al. (1998) also found that superoxide dismutase was similarly ineffective. In this latter study several potential ROS scavengers were used and the results were inconclusive as from all the agents tested [superoxide dismutase, aminotriazole (a catalase inhibitor), tiron (a presumed superoxide scavenger) and nitro blue tetrazolium (another superoxide scavenger)] only nitro blue tetrazolium inhibited hypoxic pulmonary vasonstriction. In sum, the experiments aimed to alter ROS levels by manipulation of ROS scavenging systems in the lung have yielded conflicting results, and are therefore not conclusive with regard to the possible C.Gonzalez et al. / Respiratory Physiology & Neurobiology 132 (2002) 17–41 36 significance of ROS levels on hypoxic pulmonary vasoconstriction. The data related to this criterion in EPO-producing cells are likewise inconclusive. In discussing the stabilization of HIF-1a(criterion 3) we have focused on some discrepancies that might relate to this criterion. For example, we have mentioned that according to Huang et al. (1996) overexpression of thioredoxin potentiated the expression of a reporter gene under the control of EPO enhancer implying stabilization of HIF-1a. On the other hand, Chandel et al. (2000) reported that overexpression of catalase prevented stabilization of HIF-1a. A similar controversy exists regarding H 2 O 2 . Thus Fandrey et al. (1994) reported that H 2 O 2 inhibited the production of EPO induced by hypoxia and that catalase reversed the inhibition produced by H 2 O 2 . On the contrary, Ueno et al. (1988), Ndele et al. (1996) (see also Chandel et al., 2000) reported that catalase inhibited the EPO production stimulated by H 2 O 2 and by ROS-generating systems glucose/glucose oxidase and xanthine/xanthine oxidase. 4. Conclusion With the available information we do not know if hypoxia increases or decreases ROS levels, and therefore we do not know if the putative role of ROS, i.e. to inhibit K + channels in the CB chemoreceptor cells and in PASMC and to stabilize HIF-1ain EPO-producing cells, is the result of the reduction or the oxidation of the K + channel proteins or HIF-1a. The uncertainties remain when we consider the effects of agents that increase or decrease ROS levels on the responses of CB chemoreceptor cells, PASMC and EPOproducing cells. There are several mitochondrial poisons, which have opposite effects on the rate of ROS production and levels, and yet all activate chemoreceptor cells and PASMC in normoxia and potentiate their hypoxic response. Similarly, there are mitochondrial agents that can either decrease or increase ROS levels, yet they fail to alter normoxic EPO production, implying that neither the decrease nor the increase in ROS levels mimic hypoxia. The panorama is not clarified by the use of agents or experimental maneuvers that alter the rate of ROS scavenging. There are reports indicating that ROS scavenging enzymes or drugs tend to mimic hypoxia and reports indicating the opposite, both, in PASMC and in EPOproducing cells. In the CB chemoreceptor cells the scavenging of ROS did not alter the normoxic output of the cells nor their hypoxic response. In sum, the available information on O 2 -sensing/hypoxic transduction does not support an unequivocal role for ROS in those cellular processes. The possibility exists that many of the contradictory findings, which have been obtained in tumor cell lines, represent peculiarities of particular clones used in different laboratories. In any case, it should always be kept in mind that the metabolic properties of tumor cells might differ substantially from the parental normal cells. In designing experiments, it should also be remembered that maneuvers aimed at manipulating ROS can do more than simply alter ROS levels: special care should be taken in selecting the potentially pleiotropic ROS scavengers and ROS mimetics. Acknowledgements Supported by Spanish DGICYT Grant PB97/ 0400. Thanks to Professor Benito Herreros for critical reading. References Acker, H., 1994a. Cellular oxygen sensors. Ann. New York Acad. Sci. 718, 3–10. Acker, H., 1994b. Mechanisms and meaning of cellular oxygen sensing in the organism. Respir. Physiol. 95, 1–10. Acker, H., Xue, D., 1995. Mechanisms of O 2 sensing in the carotid body in comparison with other O 2 -sensing cells. News Physiol. Sci. 10, 211–215. Almaraz, L., Gonzalez, C., Obeso, A., 1986. Effects of high potassium on the release of H-dopamine from the cat carotid body in vitro. J. Physiol. London 379, 293–307. Anichkov, S.V., Belen’kii, M.L., 1963. Pharamacology of the Carotid Body Chemoreceptors. MacMillan, New York, USA. Archer, S.L., Will, J.A., Weir, E.K., 1986. Redox status in the control of pulmonary vascular tone. Herz 11, 127–141. C.Gonzalez et al. / Respiratory Physiology & Neurobiology 132 (2002) 17–41 37 Archer, S.L., Nelson, D.P., Weir, E.K., 1989. Simultaneous measurement of O 2 radicals and pulmonary vascular reactivity in rat lung. J. Appl. Physiol. 67, 1903–1911. Archer, S.L., Huang, J., Henry, T., Peterson, D., Weir, E.K., 1993. A redox-based O 2 sensor in rat pulmonary vasculature. Circ. Res. 73, 1100–1112. Archer, S.L., Huang, J.M., Hampl, V., Nelson, D.P., Shultz, P.J., Weir, E.K., 1994. Nitric oxide and cGMP cause vasorelaxation by activation of a charybdotoxin-sensitive K channel by cGMP -dependent protein kinase. Proc. Natl. Acad. Sci. USA 91, 7583–7597. Archer, S.L., Souil, E., Dinh-Xuan, A.T., Schremmer, B., Mercier, J.C., El Yaagoubi, A., Nguyen-Huu, L., Reeve, H.L., Hampl, V., 1998. Molecular identification of the role of voltage-gated K + channels, Kvl.5 and Kv2.1, in hypoxic pulmonary vasoconstriction and control of resting membrane potential in rat pulmonary artery myocytes. J. Clin. Invest. 101, 2319–2330. Archer, S.L., Reeve, H.L., Michelakis, E., Puttagunta, L., Waite, R., Nelson, D.P., Dinauer, M.C., Weir, E.K., 1999. O 2 sensing is preserved in mice lacking the gp91 phox subunit of NADPH oxidase. Proc. Natl. Acad. Sci. USA 96, 7944–7949. Archer, S.L., Weir, E.K., Reeve, H.L., Michelakis, E., 2000. Molecular identification of O 2 sensors and O 2 -sensitive potassium channels in the pulmonary circulation. Adv. Exp. Med. Biol. 475, 219–240. Barlow, R.S., El-Mowafy, A.M., White, R.E., 2000. H 2 O 2 opens BK(Ca) channels via the PLA 2 -arachidonic acid signaling cascade in coronary artery smooth muscle. Am. J. Physiol. 279, H475–H483. Beckman, J.S., Koppenol, W.H., 1996. Nitric oxide, superoxide, and peroxynitrite: the good, the bad, and ugly. Am. J. Physiol. 271, C1424–C1437. Bolotina, V.M., Najibi, S., Palacino, J.J., Pagano, P.J., Cohen, R.A., 1994. Nitric oxide directly activates calcium-dependent potassium channels in vascular smooth muscle. Nature 368, 850–853. Boveris, A., Costa, L.E., Poderoso, J.J., Carreras, M.C., Cadenas, E., 2000. Regulation of mitochondrial respiration by oxygen and nitric oxide. Ann. New York Acad. Sci. USA 899, 121–135. Buckler, K.J., 1997. A novel oxygen-sensitive potassium current in rat carotid body type I cells. J. Physiol. London 498, 649–662. Buckler, K.J., 1999. Background leak K + -currents and oxygen sensing in carotid body type I cells. Respir. Physiol. 115, 179–187. Buckler, K.J., Williams, B.A., Honore, E., 2000. An oxygen-, acidand anaesthetic-sensitive TASK-like background potassium channel in rat arterial chemoreceptor cells. J. Physiol. London 525, 135–142. Chance, B., Sies, H., Boveris, A., 1979. Hydroperoxide metabolism in mammalian organs. Physiol. Rev. 59, 527– 605. Chandel, N.S., Schumacker, P.T., 1999. Cells depleted of mitochondrial DNA (rho0) yield insight into physiological mechanisms. FEBS Lett. 454, 173–176. Chandel, N.S., Schumacker, P.T., 2000. Cellular oxygen sensing by mitochondria: old questions, new insight. J. Appl. Physiol. 88, 1880–1889. Chandel, N.S., Maltepe, E., Goldwasser, E., Mathieu, C.E., Simon, M.C., Schumacker, P.T., 1998. Mitochondrial reactive oxygen species trigger hypoxia-induced transcription. Proc. Natl. Acad. Sci. USA 95, 11715–11720. Chandel, N.S., McClintock, D.S., Feliciano, C.E., Wood, T.M., Melendez, J.A., Rodriguez, A.M., Schumacker, P.T., 2000. Reactive oxygen species generated at mitochondrial complex III stabilize hypoxia-inducible factor-1aduring hypoxia: a mechanism of O 2 sensing. J. Biol. Chem. 275, 25130–25138. Chou, C.L., Shirahata, M., 1996. Two types of voltage-gated K + channels in carotid body cells of adult cats. Brain Res. 742, 34–42. Costa, L.E., Llesuy, S., Boveris, A., 1993. Active oxygen species in the liver of rats submitted to chronic hypobaric hypoxia. Am. J. Physiol. 264, C1395–C1400. Cross, A.R., Henderson, L., Jones, O.T.G., Delpiano, M.A., Hentschel, J., Acker, H., 1990. Involvement on NAD(P)H oxidase as a P O 2 sensor protein in the rat carotid body. Biochem. J. 272, 743–747. Dang, P.M., Cross, A.R., Babior, B.M., 2001. Assembly of the neutrophil respiratory burst oxidase: a direct interaction between p67PHOX and cytochrome b558. Proc. Natl. Acad. Sci. USA 98, 3001–3005. Davies, M.J., Dean, R.T., 1997. Radical-Mediated Protein Oxidation. From Chemistry to Medicine. Oxford Science Publications, Oxford, UK, pp. 25–120. DiChiara, T.J., Reinhart, P.H., 1997. Redox modulation of hslo Ca 2− activated K + channels. J. Neurosci. 17, 4942– 4955. Duchen, M.R., Biscoe, T.J., 1992. Relative mitochondrial membrane potential and [Ca 2+ ] i in type I cells isolated from the rabbit carotid body. J. Physiol. London 450, 33–61. Duranteau, J., Chandel, N.S., Kulisz, A., Shao, Z., Schumacker, P.T., 1998. Intracellular signaling by reactive oxygen species during hypoxia in cardiomyocytes. J. Biol. Chem. 273, 11619–11624. Dvorakova, M., Hohler, B., Vollerthun, R., Fischbach, T., Kummer, W., 2000. Macrophages: a major source of cytochrome b558 in the rat carotid body. Brain Res. 852, 349–354. Ehleben, W., Porwol, T., Fandrey, J., Kummer, W., Acker, H., 1997. Cobalt and desferrioxamine reveal crucial members of the oxygen sensing pathway in HepG2 cells. Kidney Int. 51, 483–491. Eu, J.P., Sun, J., Xu, L., Stamler, J.S., Meissner, G., 2000. The skeletal muscle calcium release channel: coupled O 2 sensor and NO signaling functions. Cell 102, 499–509. Fandrey, J., Frede, S., Jelkmann, W., 1994. Role of hydrogen peroxide in hypoxia-induced erythropoietin production. Biochem. J. 303, 507–510. Fandrey, J., Frede, S., Ehleben, W., Porwol, T., Acker, H., Jelkmann, W., 1997. Cobalt chloride and desferrioxamine C.Gonzalez et al. / Respiratory Physiology & Neurobiology 132 (2002) 17–41 38 antagonize the inhibition of erythropoietin production by reactive oxygen species. Kidney Int. 51, 492–496. Fidone, S., Gonzalez, C, 1996. Initiation and control of chemoreceptor activity in the carotid body. In: Fishman, A.P. (Ed.), Handbook of Physiology. The Respiratory System. American Physiology Society, Bethesda, MD, pp. 247–312. Firth, J.D., Ebert, B.L., Pugh, C.W., Ratcliffe, P.J., 1994. Oxygen-regulated control elements in the phosphogly cerate kinase 1 and lactate dehydrogenase A genes: similarities with the erythropoietin 3%enhancer. Proc. Natl. Acad. Sci. USA 91, 6496–6500. Fu, X.W., Nurse, C.A., Wang, Y.T., Cutz, E., 1999. Selective modulation of membrane currents by hypoxia in intact airway chemoreceptors from neonatal rabbit. J. Physiol. London 514, 139–150. Fu, X.W., Wang, D., Nurse, C.A., Dinauer, M.C., Cutz, E., 2000. NADPH oxidase is an O2 sensor in airway chemoreceptors: evidence from K + current modulation in wild-type and oxidase-deficient mice. Proc. Natl. Acad. Sci. USA 97, 4374–4379. Gabig, T.G., Schervish, E.W., Santinga, J.T., 1982. Functional relationship of the cytochrome b to the superoxide-generating oxidase of human neutrophils. J. Biol. Chem. 257, 4114–4119. Ganfornina, M.D., Lo´pez-Barneo, J., 1991. Single K + channels in membrane patches of arterial chemoreceptor cells are modulated by O 2 tension. Proc. Natl. Acad. Sci. USA 88, 2927–2930. Gassmann, M., Chilov, D., Wenger, R.W., 2000. Regulation of the hypoxia-inducible factor-1a. ARNT is not necessary for hypoxic induction of HIF-1ain the nucleus. Adv. Exp. Med. Biol. 475, 87–99. Geiszt, M., Kopp, J.B., Varnai, P., Leto, T.L., 2000. Identification of renox, an NAD(P)H oxidase in kidney. Proc. Natl. Acad. Sci. USA 97, 8010–8014. Gleadle, J.M., Ebert, B.L., Ratcliffe, P.J., 1995. Diphenyleneiodonium inhibits the induction of erythropoietin and other mammalian genes by hypoxia. Eur. J. Biochem. 234, 92–99. Gnaiger, E., Mendez, G., Hand, S.C., 2000. High phosphorylation efficiency and depression of uncoupled respiration in mitochondria under hypoxia. Proc. Natl. Acad. Sci. USA 97, 11080–11085. Goldberg, M.A., Glass, G.A., Cunningham, J.M., Bunn, H.F., 1987. The regulated expression of erythropoietin by two human hepatoma cell lines. Proc. Natl. Acad. Sci. USA 84, 7972–7976. Goldberg, M.A., Dunning, S.P., Bunn, H.F., 1988. Regulation of the erythropoietin gene: evidence that the oxygen sensor is a heme protein. Science 242, 1412–1415. Gonzalez, C., 1998. Sensitivity to physiologic hypoxia. In: Lopez-Barneo, J., Weir, E.K. (Eds.), Oxygen regulation of Ion Channels and Gene Expression. Futura Publishing Co Inc, Armonk, NY, pp. 321–336. Gonzalez, C., Almaraz, L., Obeso, A., Rigual, R., 1992. Oxygen and acid chemoreception in the carotid body chemoreceptors. Trends Neurosci. 15, 146–153. Gonzalez, C., Almaraz, L., Obeso, A., Rigual, R., 1994. Carotid body chemoreceptors: from natural stimuli to sensory discharges. Physiol. Rev. 74, 829–898. Gorlach, A., Holtermann, G., Jelkmann, W., Hancock, J.T., Jones, S.A., Jones, O.T., Acker, H., 1993. Photometric characteristics of haem proteins in erythropoietin-producing hepatoma cells (HepG2). Biochem. J. 290, 771–776. Gorlach, A., Fandrey, J., Holtermann, G., Acker, H., 1994. Effects of cobalt on haem proteins of erythropoietin-producing HepG2 cells in multicellular spheroid culture. FEBS Lett. 348, 216–218. Grimminger, F., Weissmann, N., Spriestersbach, R., Becker, E., Rosseau, S., Seeger, W., 1995. Effects of NADPH oxidase inhibitors on hypoxic vasoconstriction in bufferperfused rabbit lungs. Am. J. Physiol. 268, L747–L752. Guo, J., Giles, W.R., Ward, C.A., 2000. Effect of hydrogen peroxide on the membrane currents of sinoatrial node cells from rabbit heart. Am. J. Physiol. 279, H992–H999. Halliwell, B., Gutteridge, J.M.C., 1999. Free Radicals in Biology and Medicine, third ed. Oxford University Press, Oxford UK. Halliwell, B., Gutteridge, J.M.C., 2000. Free radicals and antioxidant in the year, a historical look to the future. Ann. NewYork Acad. Sci. 899, 136–147. Hayabuchi, Y., Nakaya, Y., Matsuoka, S., Kuroda, Y., 1998. Hydrogen peroxide-induced vascular relaxation in porcine coronary arteries is mediated by Ca 2+ -activated K + channels. Heart Vessels 13, 9–17. Hochachka, P.W., Land, S.C., Buck, L.T., 1997. Oxygen sensing and signal transduction in metabolic defense against hypoxia: lessons from vertebrate facultative anaerobes. Comp. Biochem. Physiol. A Comp. Physiol. 118, 23–29. Hoshi, T., Heinemann, S.H., 2001. Regulation of cell function by methionine oxidation and reduction. J. Physiol. London 531, 1–11. Huang, L.E., Arany, Z., Livingston, D.M., Bunn, H.F., 1996. Activation of hypoxia-inducible transcription factor depends primarily upon redox-sensitive stabilization of its alpha subunit. J. Biol. Chem. 271, 32253–32259. Iizuka, T., Kanegasaki, S., Makino, R., Tanaka, T., Ishimura, Y., 1985. Studies on neutrophil b-type cytochrome in situ by low temperature absorption spectroscopy. J. Biol. Chem. 260, 12049–12053. Jelkmann, W., 1992. Erythropoietin: structure, control of production, and function. Physiol. Rev. 72, 449–489. Joels, N., Neil, E., 1968. The idea of a sensory transmitter. In: Torrance, R.W. (Ed.), Arterial Chemoreceptors. Blackwell, Oxford, pp. 153–178. Kietzmann, T., Fandrey, J., Acker, H., 2000. Oxygen radicals as messengers in oxygen-dependent gene expression. News Physiol. Sci. 15, 202–208. Kjaeve, J., Veel, T., Bjertnaes, L., 1990. Allopurinol inhibits hypoxic pulmonary vasoconstriction. Role of toxic oxygen metabolites. Acta Anaesthesiol. Scand. 34, 384–388. Kummer, W., Acker, H., 1995. Immunohistochemical demonstration of four subunits of neutrophil NAD(P)H oxidase C.Gonzalez et al. / Respiratory Physiology & Neurobiology 132 (2002) 17–41 39 in type I cells of carotid body. J. Appl. Physiol. 78, 1904–1909. Lahiri, S., Iturriaga, R., Mokashi, A., Ray, D.K., Chugh, D., 1993. CO reveals dual mechanisms of O 2 chemoreception in the cat carotid body. Respir. Physiol. 94, 227–240. Lander, H.M., 1997. An essential role for free radicals and derived species in signal transduction. FASEB J. 11, 118– 124. Liu, H., Moczydlowski, E., Haddad, G.G., 1999. O 2 deprivation inhibits Ca 2+ -activated K + channels via cytosolic factors in mice neocortical neurons. J. Clin. Invest. 104, 577–588. Lo´pez-Barneo, J., Lo´pez-Lo´pez, J.R., Uren˜a, J., Gonzalez, C., 1988. Chemotransduction in the carotid body: K + current modulated by P O 2 in type I chemoreceptor cells. Science 241, 580–582. Lo´pez-Barneo, J., Pardal, R., Montoro, R.J., Smani, T., Garcia-Hirschfeld, J., Urena, J., 1999. K + and Ca 2+ channel activity and cytosolic [Ca 2+ ] in oxygen-sen sing tissues. Respir. Physiol. 115, 215–227. Lo´pez-Lo´pez, J.R., Gonzalez, C., 1992. Time course of K + current inhibition by low oxygen in chemoreceptors cells of adult rabbit carotid body. Effects of carbon monoxide. FEBS Lett. 299, 251–254. Lo´pez-Lo´pez, J., Gonzalez, C., Uren˜a, J., Lo´pez-Barneo, J., 1989. Low P O 2 selectively inhibits K + channel activity in chemoreceptor cells of the mammalian carotid body. J. Gen. Physiol. 93, 1001–1015. Lo´pez-Lo´pez, J.R., Gonzalez, C., Perez-Garcia, M.T., 1997. Properties of ionic currents from isolated adult rat carotid body chemoreceptor cells. Effects of hypoxia. J. Physiol. London. 499, 429–441. Marshall, B.E., Marshall, C., Frasch, F., Hanson, C.W., 1994. Role of hypoxic pulmonary vasoconstriction in pulmonary gas exchange and blood flow distribution. 1. Physiologic concepts. Intensive Care Med. 20, 291–297. Marshall, C., Mamary, A.J., Verhoeven, A.J., Marshall, B.E., 1996. Pulmonary artery NADPH-oxidase is activated in hypoxic pulmonary vasoconstriction. Am. J. Respir. Cell Mol. Biol. 15, 633–644. Matsumoto, S., Nagamine, T., Yakata, H., Kohno, M., 1987. Alteration of chemoreflex phrenic responses by oligomycin in the rabbit. Arch. Int. Pharmacodyn. Ther. 286, 136– 144. McLennan, H., 1963, Synaptic transmission. Saunders. Philadelphia, USA. Mohazzab, K.M., Wolin, M.S., 2000. Oxidant signalling and vascular oxygen sensing. Role of H 2 O 2 in responses of the bovine pulmonary artery to changes in P O 2 . Adv. Exp. Med. Biol. 475, 249–258. Mulligan, E., Lahiri, S., 1982. Separation of carotid body chemoreceptor responses to O 2 and CO 2 by oligomycin and by antimycin A. Am. J. Physiol. 242, C200–C206. Nakanishi, K., Tajima, F., Nakamura, A., Yagura, S., Ookawara, T., Yamashita, H., Suzuki, K., Taniguchi, N., Ohno, H., 1995. Effects of hypobaric hypoxia on antioxidant enzymes in rats. J. Physiol. London 489, 869–876. Ndele, J.K., Yoshioka, K., Fisher, J.W., 1996. Hydrogen peroxide in the regulation of erythropoietin (Epo) gene expression in hepatocellular carcinoma cells. East Afr. Med. J. 73, 143–146. Obeso, A., 1984. Hipotesis metabolica de quimiotansduccion: hechos experimentales. (Ph. D Thesis). Universidad de Valladolid, Valladolid, Spain. Obeso, A., Almaraz, L., Gonzalez, C., 1989a. Effects of cyanide and uncouplers on chemoreceptor activity and ATP content of the cat carotid body. Brain Res. 481, 250–257. Obeso, A., Gonzalez, C., Dinger, B., Fidone, S., 1989b. Metabolic activation of carotid body glomus cells by hypoxia. J. Appl. Physiol. 67, 484–487. Obeso, A., Gonzalez, C., Rigual, R., Dinger, B., Fidone, S., 1993. Effect of low O 2 on glucose uptake in rabbit carotid body. J. Appl. Physiol. 74, 2387–2393. Obeso, A., Rocher, A., Herreros, B., Gonzalez, C., 1997. Oxygen consumption and energy metabolism of the carotid body. In: Gonzalez, C. (Ed.), The Carotid Body Chemoreceptors. Springer, NY, pp. 31–45. Obeso, A., Gomez-Nin˜o, A., Gonzalez, C., 1999. Inhibition of NADPH oxidase does and not interfere with low P O 2 transduction in chemoreceptor cells. Am. J. Physiol. 276, C593–C601. Obeso, A., Sanz-Alfayate, G., Agapito, M.T., Gonzalez, C., 2000a. Significance of ROS in oxygen chemoreception in the carotid body chemoreception. Apparent lack of a role for NADPH oxidase. Adv. Exp. Med. Biol. 475, 425–434. Obeso, A., Sanz-Alfayate, G., Agapito, M.T., Gonzalez, C., 2000b. Glutathione levels and GSH-GSSG ratios in the carotid body and oxygen sensing in chemoreceptor cells. High Altitude Med. Biol. 1, 255. Ohkuma, S., Katsura, M., 2001. Nitric oxide and peroxynitrite as factors to stimulate neurotransmitter release in the CNS. Prog. Neurobiol. 64, 97–108. Osanai, S., Mokashi, A., Rozanov, C., Buerk, D.G., Lahiri, S., 1997. Potential role of H 2 O 2 in chemoreception in the cat carotid body. J. Auton. Nerv. Syst. 63, 39–45. Osipenko, O.N., Tate, R.J., Gurney, A.M., 2000. Potential role for kv3.1b channels as oxygen sensors. Circ. Res. 86, 534–540. Paky, A., Michael, J.R., Burke-Wolin, T.M., Wolin, M.S., Gurtner, G.H., 1993. Endogenous production of superoxide by rabbit lungs: effects of hypoxia or metabolic inhibitors. J. Appl. Physiol. 74, 2868–2874. Patel, A.J., Lazdunski, M., Honore, E., 1997. Kv2.1/Kv9.3, a novel ATP-dependent delayed-rectifier K + channel in oxygen-sensitive pulmonary artery myocytes. EMBO J. 16, 6615–6625. Peers, C., 1990. Hypoxic suppression of K + currents in type I carotid body cells: selective effect on the Ca 2+ -activated K + current. Neurosci. Lett. 119, 253–256. Pe´rez-Garcı´a, M.T., Lo´pez-Lo´pez, J.R., 2000. Are Kv channels the essence of O 2 sensing. Circ. Res. 86, 490–491. Perez-Garcı´a, M.T., Lo´pez-Lo´pez, J.R., Gonzalez, C., 1999. Kvb1.2 subunit i modulates the sensitivity to hypoxia of cloned voltage-gated K + channels in a subfamily-specific manner. J. Gen. Physiol. 113, 897–907. C.Gonzalez et al. / Respiratory Physiology & Neurobiology 132 (2002) 17–41 40 Pe´rez-Garcı´a, M.T., Lo´pez-Lo´pez, J.R., Riesco, A.M., Hoppe, U., Gonzalez, C., Marban, E., Johns, D.C., 2000. Supression of transient outward K + currents in chemoreceptor cells of the rabbit carotid body by viral gene transfer of inducible dominant negative Kv4.3 constructs. J. Neurosci. 20, 5689–5695. Post, J.M., Hume, J.R., Archer, S.L., Weir, E.K., 1992. Direct role for potassium channel inhibition in hypoxic pulmonary vasoconstriction. Am. J. Physiol. 262, C882–C890. Pugh, C.W., Tan, C.C., Jones, R.W., Ratcliffe, P.J., 1991. Functional analysis of an oxygen-regulated transcriptional enhancer lying 3’to the mouse erythropoietin gene. Proc. Natl. Acad. Sci. USA 88, 10553–10557. Punchard, N.A., Kelly, F.J., 1996. Free Radicals. A Practical Approach. IRL Press-Oxford University Press, Oxford, UK, pp. 1–8. Ratcliffe, P.J., Ebert, B.L., Ferguson, D.J., Firth, J.D., Gleadle, J.M., Maxwell, P.H., Pugh, C.W., 1995. Regulation of the erythropoietin gene. Nephrol. Dial. Transplant. 10 (Suppl 29), 18–27. Richalet, J.P., 1997. Oxygen sensors in the organism: examples of regulation under altitude hypoxia in mammals. Comp. Biochem. Physiol. A Physiol. 118, 9–14. Riesco, A., Pere´z-Garcia, M.T., Gonzalez, C., Lo´pez-Lo´pez, J.R., 2001. Ca 2+ -dependent K + channels are modulated by hypoxia and CO in isolated membrane patches of rat chemoreceptor cells. Biophys. J. 80, 206a. Rocher, A., Obeso, A., Herreros, B., Gonzalez, C., 1988. Activation of the release of dopamine in the carotid body by veratridine. Evidence for the presence of voltage-dependent Na + channels in type I cells. Neurosci. Lett. 94, 274–278. Roy, A., Rozanov, C., Mokashi, A., Daudu, P., Al-mehdi, A.B., Shams, H., Lahiri, S., 2000. Mice lacking in gp91 phox subunit of NAD(P)H oxidase showed glomus cell [Ca 2+ ] i and respiratory responses to hypoxia. Brain Res. 872, 188–193. Ruppersberg, J.P., Stocker, M., Pongs, O., Heinemann, S.H., Frank, R., Koenen, M., 1991. Regulation of fast inactivation of cloned mammalian IK(A) channels by cysteine oxidation. Nature 352, 711–714. Sanderud, J., Bjoro, K., Saugstad, O.D., 1993. Oxygen radicals stimulate thromboxane and prostacyclin synthesis and induce vasoconstriction in pig lungs. Scand. J. Clin. Lab. Invest. 53, 447–455. Seeger, W., Suttorp, N., Schmidt, F., Neuhof, H., 1986. The glutathione redox cycle as a defense system against hydrogen-peroxide-induced prostanoid formation and vasoconstriction in rabbit lungs. Am. Rev. Respir. Dis. 133, 1029–1036. Semenza, G.L., 1994. Regulation of erythropoietin production. New insights into molecular mechanisms of oxygen homeostasis. Hematol. Oncol. Clin. North Am. 8, 863– 884. Semenza, G.L., 1999. Perspectives on oxygen sensing. Cell 98, 281–284. Semenza, G.L., 2000. Chairman’s summary: mechanisms of oxygen homeostasis, circa 1999. Adv. Exp. Med. Biol. 475, 303–310. Shin, J.H., Lee, J.E., Park, J.M., Suh, C.K., 2000. T-butyl hydrogen peroxide increases the activities of the Maxi-K channels of rat brain. Life Sci. 67, 2485–2491. Simon-Schnass, I.M., 2000. Risk of oxidative stress at high altitude and possible benefit of antioxidant supplementation. In: Sen, C.K., Packer, L., Hannimen, O. (Eds.), Handbook of Oxidants and Antioxidants in Exercise. Elsevier, Amsterdam, The Netherlands, pp. 555–577. Soh, R., Jung, W., Uhm, D.Y., Chung, S., 2001. Modulation of large conductance calcium-activated potassium channels from rat hippocampal neurons by glutathione. Neurosci. Lett. 298, 115–118. Srinivas, V., Zhang, L.P., Zhu, X.H., Caro, J., 1999. Characterization of an oxygen/redox-dependent degradation domain of hypoxia-inducible factor alpha (HIF-a) proteins. Biochem. Biophys. Res. Commun. 260, 557–561. Tamayo, L., Lo´pez-Lo´pez, J.R., Castan˜eda, J., Gonza´lez, C., 1997. Carbon monoxide inhibits hypoxic pulmonary vasoconstriction by a cGMP-independent mechanism. Pflu¨gers. Arch. Eur. J. Physiol. 434, 698–704. Tan, C.C., Ratcliffe, P.J., 1991. Effect of inhibitors of oxidative phosphorylation on erythropoietin mRNA in isolated perfused rat kidneys. Am. J. Physiol. 261, F982–F987. Tang, X.D., Daggert, H., Hanner, M., Garcia, M.L., McManus, O.B., Brot, N., Weissbach, H., Heinemann, S.H., Hoshi, T., 2001. Oxidative regulation of large conductance calcium-activated potassium channels. J. Gen. Physiol. 117, 253–274. Thomas, H.M., Carson, R.C., Fried, E.D., Novitch, R.S., 1991. Inhibition of hypoxic pulmonary vasoconstriction by diphenyleneiodonium. Biochem. Pharmacol. 42, R9–R12. Thompson, J.S., Jones, R.D., Rogers, T.K., Hancock, J., Morice, A.H., 1998. Inhibition of hypoxic pulmonary vasoconstriction in isolated rat pulmonary arteries by diphenyleneiodonium (DPI). Pulm. Pharmacol. Ther. 11, 71–75. Thuringer, D., Findlay, I., 1997. Contrasting effects of intracellular redox couples on the regulation of maxi-K + channels in isolated myocytes from rabbit pulmonary artery. J. Physiol. London 500, 583–592. Ueno, M., Brookins, J., Beckman, B.S., Fisher, J.W., 1988. Effects of reactive oxygen metabolites on erythropoietin production in renal carcinoma cells. Biochem. Biophys. Res. Commun. 154, 773–780. Vega-Saenz de Miera, E., Rudy, B., 1992. Modulation of K + channels by hydrogen peroxide. Biochem. Biophys. Res. Commun. 186, 1681–1687. Verna, A., Talib, N., Roumy, M., Pradet, A., 1990. Effects of metabolic inhibitors and hypoxia on the ATP, ADP and AMP content of the rabbit carotid body in vitro: the metabolic hypothesis in question. Neurosci. Lett. 116, 156–161. Wang, G.L., Semenza, G.L., 1993a. Characterization of hypoxia-inducible factor 1 and regulation of DNA binding activity by hypoxia. J. Biol. Chem. 268, 21513–21558. C.Gonzalez et al. / Respiratory Physiology & Neurobiology 132 (2002) 17–41 41 Wang, G.L., Semenza, G.L., 1993b. General involvement of hypoxia-inducible factor 1 in transcriptional response to hypoxia. Proc. Natl. Acad. Sci. USA 90, 4304–4308. Wang, D., Youngson, C., Wong, V., Yeger, H., Dinauer, M.C., Vega-Saenz Miera, E., Rudy, B., Cutz, E., 1996. NADPH-oxidase and a hydrogen peroxide-sensitive K + channel may function as an oxygen sensor complex in airway chemoreceptors and small cell lung carcinoma cell lines. Proc. Natl. Acad. Sci. USA 93, 13182–13187. Weir, E.K., Archer, S.L., 1995. The mechanism of acute hypoxic pulmonary vasoconstriction: the tale of two channels. FASEB J. 9, 183–189. Weir, E.K., Wyatt, C.N., Reeve, H.L., Archer, S.L., Peers, C., 1994. Diphenyleneiodonium inhibits both K + and Ca 2+ currents in isolated pulmonary artery smooth muscle cells. J. Appl. Physiol. 76, 2611–2615. Weissmann, N., Grimminger, F., Voswinckel, R., Conzen, J., Seeger, W., 1998. Nitro blue tetrazolium inhibits but does not mimic hypoxic vasoconstriction in isolated rabbit lungs. Am. J. Physiol. 274, L721–L727. Wenger, R.H., Marti, H.H., Schuerer-Maly, C.C., Kvietikova, I., Bauer, C., Gassmann, M., Maly, F.E., 1996. Hypoxic induction of gene expression in chronic granulomatous disease-derived B-cell lines: oxygen sensing is independent of the cytochrome b558-containing nicotinamide adenine dinucleotide phosphate oxidase. Blood 87, 756–761. Wolin, M.S., Burke-Wolin, T.M., Mohazzab-H, K.M., 1999. Roles for NAD(P)H oxidases and reactive oxygen species in vascular oxygen sensing mechanisms. Respir. Physiol. 115, 229–238. Wyatt, C.N., Peers, C., 1995. Ca 2+ -activated K + channels in isolated type I cells of the neonatal rat carotid body. J. Physiol. London 483, 559–565. Wyatt, C.N., Weir, E.K., Peers, C., 1994. Diphenyleneiodonium blocks K + and Ca 2+ currents in type I cells isolated from neonatal rat carotid body. Neurosci. Lett. 172, 63– 66. Youngson, C., Nurse, C., Yeger, H., Cutz, E., 1993. Oxygen sensing in airway chemoreceptors. Nature 365, 153–155. Yuan, X.J., Goldman, W.F., Tod, M.L., Rubin, L.J., Blaustein, M.P., 1993. Hypoxia reduces potassium currents in cultured rat pulmonary but not mesenteric arterial myocytes. Am. J. Physiol. 264, L116–L123. Yuan, X.J., Wang, J., Juhaszova, M., Golovina, V.A., Rubin, L.J., 1998. Molecular basis and function of voltage-gated K + channels in pulmonary arterial smooth muscle cells. Am. J. Physiol. 274, L621–L635. Zhang, F., Carson, R.C., Zhang, H., Gibson, G., Thomas, H.M., 1997. Pulmonary artery smooth muscle cell [Ca 2+ ] i , and contraction; responses to diphenyleneiodonium and hypoxia. Am. J. Physiol. 273, L603–L611. Zhu, H., Qiu, H., Yoon, H.W., Huang, S., Bunn, H.F., 1999. Identification of a cytochrome b-type NAD(P)H oxidoreductase ubiquitously expressed in human cells. Proc. Natl. Acad. Sci. USA 96, 14742–14747.